Proceedings of the 19th Japan-Korea-France-Canada

Joint Seminar on Geoenvironmental Engineering

Geo-Environmental Engineering 2021

May 20-21, 2021

ESITC Caen - Caen National University, France

Organized by
ESITC Caen, France
Kyoto University, Japan
Fukuoka University, Japan
Hokkaido University, Japan
Seoul National University, South Korea
Concordia University, Montreal, Canada
University of Normandy, Caen National University, France
 Front cover photos with the courtesy of Valgo company

Former refinery of Petit-Couronne (France) beyond rehabilitation; the industrial activity

started in 1929 and stopped in 2013. Since 2014, VALGO is dismantling the units
(superstructures up to 170m and underground piping networks), remediating the soils from
the hydrocarbon pollutions and welcoming new industrial activities. Depending of the
subzones, the LNAPL layer is pumped, the soil is cleaned or removed and replaced with
recycled materials. A clean layer of soils coming from Grand Paris’ huge works is
compacted to finalize the new industrial zone.
 Proceedings of the 19th Japan-Korea-France-Canada

Joint Seminar on Geoenvironmental Engineering

Geo-Environmental Engineering 2021

May 20-21, 2021
ESITC Caen - Caen National University, France

Organized by
ESITC Caen, France
Kyoto University, Japan
Fukuoka University, Japan
Hokkaido University, Japan
Seoul National University, South Korea
Concordia University, Montreal, Canada
University of Normandy, Caen National University, France

With the support of

Centre Français du Littoral Association
Communauté urbaine Caen-la-Mer
Laboratory Morphodynamique Continentale et Côtière - M2C, Caen
SMÉDAR Le Grand Quevilly
VALGO company, Portet-sur-Garonne
Preface

Like anyone, the GEE2021 conference has suffered from the covid19 pandemic.
Originally planned for May 2020, it had to be postponed when the pandemic was at its
worst. Postponed to May 2021 in a less pronounced but still active pandemic situation,
this conference turned into a virtual conference. This is how the GEE2021 conference
came about.

Geo-Environmental Engineering 2021 (GEE2021 or 19th Japan-Korea-France-Canada

Joint Seminar on Geo-Environmental Engineering) was held on May 20-21 at the
ESITC High school for civil and construction engineers of Caen in the campus 2 of
Caen Normandy University. This annual seminar was organized under the auspices of
the French National Center for Scientific Research CNRS, the International
Geosynthetics Society IGS, the Pole Mer which is a French Sea Innovation and
Business Cluster. It was supported by the ESITC Caen for local organization, the Centre
Français du Littoral Association for the editorial work, Caen La Mer Urban Community,
VALGO company (Paris and Petit-Couronne agency near Rouen) and the Laboratory
“Continental and Coastal Morphodynamics - M2C” of Caen.

The first Korea-Japan Joint Seminar took place in Seoul in 2001. Prof. Junboum Park
from Seoul University and Prof. Masashi Kamon from Kyoto University initiated it. For
the fifth seminar in 2005, the Korean Institute of Construction Technology has joined
the seminar group. And the seventh seminar was for the first time held in France, at
University of Grenoble with the cooperation of Prof. Jean-Pierre Gourc. One year after,
in 2008, the joint seminar held in Kyoto University with the Canadian cooperation
through the venue of Prof. Loretta Li from University of British Columbia and then with
the participation of the Prof. Catherine Mulligan of Concordia University. Then, it took
place continuously in these 4 countries Japan-Korea-France and Canada: Nantes (2016),
Seoul (2017), Fukuoka (2018) and Montreal (2019). To ensure the continuity of the
joint seminar, with the both cooperation of Prof. Jean-Pierre Gourc (Grenoble
University) and Prof. Daniel Levacher (Caen University), the 19th joint seminar was the
fourth one held in France, in Caen at ESITC with cooperation of Caen Normandy
University. Previously prepared for May 2020 it was postponed in 2021, May 20-21 as
a virtual conference.

And so far from Japan and Korea, we highly appreciate the papers coming from
overseas and particularly from Brazil and India, these both countries could follow the
GEE in the future. During the 2 days, the symposium has provided interesting
discussions and participants have debated on updated questions relative to the waste and
landfill management, the geotechnical reuse of solid waste, the remediation of
contaminated sites, the environmental risk assessment and about lessons from the field
and case studies. The exchange of ideas on research and experience in these fields was
rich. For these reasons, the Geo-Environmental Engineering 2021 has demonstrated that
the symposium should continue in 2022 and in future.

While the virtual conference was well suited to the pandemic situation, it could not
provide the traditional technical field visits, which participants regretted. But the hope
that the pandemic will end in the near future should reassure all participants for the next
GEE.

The symposium would not have been possible without the full contribution of Em. Prof.
Daniel Levacher (Caen University), member of GEE committee and Dr Mohamed
Boutouil (ESITC Caen) for the local organization. All the GEE committee members,
greatly appreciate their efforts.

Daniel Levacher, Civil Engineering Emeritus Professor, Caen University

 TABLE OF CONTENTS

Preface

KEYNOTE LECTURE

1 Remediation of a refinery’s site by LNAPL pumping adapt to the constraints of the matrix and of the pollutant 1
Laurent THANNBERGER, Marius DRAGA, Théo DECLERCQ

ENVIRONMENTAL RISK ASSESSMENT

REMEDIATION MANAGEMENT

2 Assessment of marine submersion's risk on the coastal morpho-dynamics of Bejaia, Algeria 23

Cherif AOUDJ, Mokhtar GUERFI, Khoudir MEZOUAR
3 3D geophysical characterization of LNAPL volume extension 29
Théo DE CLERCQ, Laurent THANNBERGER, Abderrahim JARDANI
4 Possibility of well clogging by recharging tunnel seepage in the Kanto region, Japan 49
Sho HASEGAWA 1, Toshifumi IGARASHI
5 Change in groundwater flow in a limestone quarry by drainage tunnel excavation 57
Keisuke INOUE, Kouta FUJIMAKI, Chika UMEDA, Toshifumi IGARASHI
6 Assessment of heavy metal contamination through sediment cores from shallow harbour in Quebec, Canada 69
Masoumeh JAVID, Catherine N. MULLIGAN
7 A new prediction method for the rainfall-induced landslides and debris flows 81
Sangseom JEONG, Moonhyun HONG
8 Development of a model to quantify dust emissions from truck traffic on earthmoving sites - Approach and 87
preliminary results
Mickael LE VERN, Ouardia SEDIKI, Andry RAZAKAMANANTSOA, Frédéric MURZYN, Frédérique
LARRARTE, Pascal INSENGA, Philippe GOTTELAND
9 Leaching behaviour of arsenic from excavated rock under anaerobic conditions 95
Yusuke MASAKI, Tomohiro KATO, Lincoln Waweru GATHUKA, Atsushi TAKAI, Takeshi KATSUMI
10 Oral bio-accessibility of Pb and Zn from playground soils in Kabwe, Zambia 103
Walubita MUFALO, Pawit TANGVIROON, Toshifumi IGARASHI, Mayumi ITO, Tsutomu SATO, Meki
CHIRWA, Imasiku NYAMBE, Hokuto NAKATA, Shouta NAKAYAMA, Mayumi ISHIZUKA
11 Desorption parameters to evaluate arsenic leaching behaviour 111
Torataro NASAHARA, Jiajie TANG, Tomohiro KATO, Lincoln Waweru GATHUKA, Atsushi TAKAI, Takeshi
KATSUMI
12 Serial batch tests and up-flow column tests to evaluate the sorption performance of soil amended with a 119
stabilising agent
Takaomi OKADA, Tomohiro KATO, Lincoln Waweru GATHUKA, Atsushi TAKAI, Takeshi KATSUMI
13 Removal of heavy metals using tire derived activated carbon (TAC) vs commercial activated carbon (CAC) 131
Rahim SHAHROKHI, Junboum PARK
14 Arsenic stabilization by oyster shell, zeolite and the mixture of them as binders evaluated in an As- 139
contaminated soil and an As-aqueous solution
Cecilia TORRES QUIROZ, Junboum PARK

REUSE OF SOLID WASTE & WASTE ENGINEERING

15 Characterization and mechanical proprieties of soil-vegetal fibre material 147

Athmane AZIL, Tuan Anh PHUNG, Malo LE GUERN, Nassim SEBAIBI
16 Mechanical and leaching characteristics of carbonated MSW IBA using exhaust gas and CO2 discharged from 153
 waste incineration facilities
Takuro FUJIKAWA, Kenichi SATO, Chikashi KOGA, Hirofumi SAKANAKURA, Hiroshi KUBOTA, Yosuke
NAGAYAMA
17 Effect of fine and water contents of soil on effective separation of mixed wastes generated by huge disasters 163
Kansei HIRAOKA, Junichiro SHIOIRI, Shogo NAKAGAWA, Atsushi TAKAI, Takeshi KATSUMI
18 Analysis of sediment-based fired bricks strengths: A case study of fluvial sediments from Mexico 171
Mazhar HUSSAIN, Daniel LEVACHER, Nathalie LEBLANC, Hafida ZMAMOU, Jean-Baptiste BESNIER,
Irini DJERAN-MAIGRE, Andry RAZAKAMANANTSOA
19 Grain size characterisation - Study of the sediment dynamics of Monastir bay, Tunisia 179
Nouha KHIARI, Abdelfattah ATOUI, Nadia KHALIL, Abdelkrim CHAREF
20 Comparative study of the compressive strength of hollow concrete blocks made in two factories in the city of 185
Yaoundé based on local and imported cement
Bertille Ilalie K. MANEFOUET
21 Mechanical properties of the biomass ash-based binder 201
Désiré NDAHIRWA, Hélène LENORMAND, Hafida ZMAMOU, Nathalie LEBLANC
22 Valorization of sunflower bark in agropellets 213
Anaelle REIX, Hafida ZMAMOU, Hélène LENORMAND, Nathalie LEBLANC
23 A case study of solidification technology applied to soft soils and sediments: Construction of a railway 221
platform in Ningde (China)
Xiao WANG, Daniel LEVACHER, Tianxue ZHONG
24 Cartography of agricultural by-products available in the World for a potential use in building materials 229
Hafida ZMAMOU, Hélène LENORMAND, David ABOUZEID, Nathalie LEBLANC

GROUND AND CONSTRUCTION IMPROVEMENT TECHNOLOGIES

GEOSYNTHETICS & SUSTAINABLE GEOENVIRONMENTAL ENGINEERING

25 Influence of fiber crushing on light earth hygrothermal properties 237

Farjallah ALASSAAD, Karim TOUATI, Daniel LEVACHER, Nassim SEBAIBI
26 Reinforcement design using geosynthetics for foundations on karstic zones: analytical calculation and 247
numerical stress-strain model
Paulo CASTRO, Isabelle SILVA, Paula MARTINS, Filipe XAVIER, Rodrigo FONSECA
27 Management of harbor sediments to reuse as backfill 257
Ali HUSSAN, Daniel LEVACHER, Salim MEZAZIGH
28 Dewatering cone tests: Evaluation of chemical conditioning and geotextile efficiencies for a water treatment 265
plant sludge
Matheus MÜLLER, Gabriel L. A. de OLIVEIRA, Delma VIDAL
29 Evaluating geotechnical properties and hydraulic performance of soil-bentonite cutoff wall using in situ 277
samples
Kazuki NISHIMURA, Yan TIAN, Atsushi TAKAI, Toru INUI, Takeshi KATSUMI
30 Pilot-scale heating of soft clays using solar collectors for the thermal acceleration of consolidation: 289
experimental and numerical studies
Mohammed Tarek Sayed SAKR, Shuhei NISHI, Atsushi OGAWA, Atsushi TAKAI, Takeshi KATSUMI
31 Towards low carbon binders: binder’s evolution and use of wood ash as mineral filler 297
Léo SAOUTI , Daniel LEVACHER, Hafida ZMAMOU, Nathalie LEBLANC, Louis JARDIN
32 Oyster shell powder, zeolite and red mud as a binder stabilizer to remediate heavy metal contaminated soil 305
Cecilia TORRES QUIROZ, Junboum PARK
33 Studies on the use of polymer blended MSW as a structural fill material in embankments 315
B.V.S. VISWANADHAM, Ankit KUMAR
34 Studies on verification of dynamic compaction induced densification of MSW landfills using shear wave 331
velocity profiling
B.V.S. VISWANADHAM, Saptarshi KUNDU
35 Economic viability of using iron ore tailings to make blocks to be reinforced with geosynthetics 341
 Filipe XAVIER, Gabriela PRINZ, Paulo CASTRO, Rodrigo FONSECA, Isabelle SILVA
36 Effect of adding slags on strength and leaching properties of soft soil 351
Aye Cho Cho ZAW, Lincoln Waweru GATHUKA, Yan TIAN, Atsushi TAKAI, Takeshi KATSUMI
37 Attenuation performance of geosynthetic sorption sheet against arsenic under different compressive stresses 359
Yu ZHANG, Yosuke KINOSHITA, Tomohiro KATO, Lincoln W. GATHUKA, Atsushi TAKAI, Takeshi
KATSUMI
38 An innovative process for continuous dehydration and reuse of sediment, sands washdown waters and 369
contaminated soil
Mohammed BOUMAHDI, Sébastien DESCHULTER
39 Acid buffering and arsenic leaching behaviours of excavated acid rock treated by the MgO based 377
immobilization material
Tsutomu TAKATA, Xun DU, Toru INUI, Sho OGATA, Hirotoshi MORI
40 Analysis of soil subsidence from the loading cycle using a concrete mattress with woven bamboo 385
reinforcement
Lusmeilia AFRIANI, Daniel LEVACHER, Ryzal PERDANA

ABSTRACTS

A1 Permeable pavements 397

Philippe DHERVILLY, Jose OLIVEIRA
A2 Contribution to the assessment of landslide factors in the Bushwira area 398
Ilalie MANEFOUET BERTILLE, Christian KALIKONE BUZERA, Claude CUBAKA RUGENDABANGA,
Célestin LUNGERE BARHADOSANYA
KEYNOTE LECTURE
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Remediation of a refinery’s site by LNAPL pumping adapt to

the constraints of the matrix and of the pollutant

Dépollution du site d’une ancienne raffinerie : pomper les

huiles en s’adaptant aux contraintes produit et matrice

Laurent THANNBERGER 1, Marius DRAGA 1, Théo DECLERCQ 1

1. VALGO, PiC, 72, rue Aristide Briand, 76650 PETIT-COURONNE, France.

[email protected]

Abstract:
In 2014, VALGO bought the site of the closed Petit-Couronne refinery (Normandy,
France), former Petroplus, former Shell, with the objective of converting this brownfield
to welcome the 21st century industry. In addition to the superstructures visible for miles,
the refining activities, for more than 80 years, have left deep marks in the subsoil,
particularly on the upper side of the water table. Refining transforms crude oil into a
multitude of finished or semi-finished products. Each of these compounds, when in contact
with the ground, presents a different behaviour, as for the dynamics of penetration,
accumulation, and possible attenuation.
The geology of the site, composed of alluvial layers on a chalky base, gives it a particular
hydrogeological functioning, reinforced by the sensitivity to the tide, transmitted by the
nearby Seine River.
This presentation will review 4 years of studies and tests, necessary for the qualification of
the best techniques, different in 3 main areas, because each one is adapted to the nature of
the pollutant (always mixtures of petroleum hydrocarbons, but with viscosities and
densities ranging from light gasoline to thick tars) and to the surrounding matrix (sandy or
clayey alluvium, flint bed, karstic massif of faulted chalk).

Keywords:
Remediation, Refinery, Oils and fuels, LNAPL, Pumping, Skimming, Polluted soils,
Groundwater, Capillary fringe zone.

Corresponding author mail: [email protected]

Résumé:
En 2014, VALGO a acquis le site de l’ancienne raffinerie de Petit-Couronne (Normandie,
France), anciennement propriété de Shell, puis Petroplus, dans le but de reconvertir cette
friche industrielle pour accueillir l’industrie du 21ème siècle. En plus des superstructures,
très élevées et visibles à des kilomètres à la ronde, les activités de raffinages, pendant plus
de 80 ans, ont laissé des marques profondes dans le sous-sol, notamment sur la face
supérieur (le toit) de la nappe phréatique. Le raffinage transforme du pétrole brut en une
multitude de produits finis ou semi-finis. Chacun de ces composés, lorsqu’il entre en
contact avec le sol, présente un comportement différent, quant à sa dynamique de
pénétration, son accumulation et sa possible dégradation.
La géologie du site, composée d’alluvions déposées sur le socle crayeux, génère un
fonctionnement hydrogéologique particulier, renforcé par la sensibilité du site à la marée,
transmise par la Seine toute proche.
Cette présentation passe en revue 4 ans d’essais et d’études, qui ont été nécessaires pour la
qualification des meilleures techniques à appliquer, avec des variantes dans 3 zones
majeures, puisque chacune prend en compte la nature du polluant (toujours des
hydrocarbures pétroliers, mais avec des viscosités et densités variant des essences légères
au bitumes lourds) et la nature de la matrice encaissante (alluvions sableuses ou argileuses,
lit de silex, ou massif karstique de la craie faillée).

Mots clés :
Dépollution, Raffinerie, Huiles et carburants, Pompage, Ecrémage, Sols pollués, Nappe
phréatique, Zone capillaire.

1. Introduction

1.1 Un long contexte historique

La raffinerie de Petit-Couronne était une raffinerie de pétrole ouverte en 1929 par
« Pétroles Jupiter » avant d’être exploitée par Shell puis cédée à Petroplus pour une
fermeture du site en avril 2013. Elle s’étendait sur une surface de 225 hectares, avec un
maximum de 1 342 employés (1938) et fut capable de produire 161 000 barils de pétroles
par jours à son apogée.
- De 1927 à 1929, les travaux de la première usine (40 ha) démarrent par les
terrassements, des bacs de stockages et le montage des distilleries, puis des
recrutements de centaines d’ouvriers.
- Dès février 1929, commence le raffinage d’environ 200 000 tonnes par an de
production d'asphaltes et d'huiles de graissage à partir des bruts lourds du Venezuela.
- En 1933, une seconde usine est construite, la plus moderne d'Europe, portant la capacité
à plus de 600 000 tonnes de pétrole brut.
- En 1948, création de la Compagnie de Raffinage Shell Berre.
2
- En 1950, la capacité de traitement de la raffinerie monte à 2 000 000 tonnes par an.
- En 1969, avec une capacité de distillation de 10 millions de tonnes de pétrole brut par
an, la raffinerie de Petit-Couronne devient la deuxième raffinerie de France.
- En 1974, la raffinerie double sa capacité de production en ouvrant une nouvelle unité.
- Les années 1980 sont marquées par une stagnation du marché pétrolier et une
modernisation des équipements, entrainant, à travers différents plans de restructuration,
une diminution importante des effectifs.
- Entre 1980 et 1985, les unités en surcapacités sont démontées, dont les unités de
distillation, d’hydrotraitement et d’hydrodésulfuration. Cette étape de démantèlement
a pu entrainer des déversements de fluides polluants dans le sous-sol de la raffinerie.
- En 1994, l’unité de distillation brut/lourd vénézuélien est démontée.
- En 2008, la raffinerie est rachetée par le groupe suisse Petroplus qui déposera le bilan
en 2012 et la société sera placée en redressement judiciaire puis en liquidation judiciaire
(BLOG EKLABLOG, WOKUZY, 2012).
- En 2014, VALGO rachète la friche au liquidateur judiciaire et procède à la séparation
des lots : le stockage est vendu à Bolloré Energy (140 ha), VALGO se chargeant de
démanteler et dépolluer les 90 ha dédiés à la production industrielle.

1.2 Un tournant dans la prise en compte environnementale

Au cours de ces 84 années de fonctionnement, de nombreux déversements ont pu se
produire, de manière intentionnelle (fonctionnement révolu) ou involontaire, entrainant une
contamination du sol et de la nappe sous-jacente. C’est dans ce contexte historique que
VALGO, entreprise de désamiantage et de dépollution des sols, a été choisie, un an après
la fermeture du site, pour prendre en charge la réindustrialisation du nouveau pôle industriel
de Petit-Couronne. Ce projet tient compte, dans son plan de réindustrialisation, de la
pollution présente sur le site. Depuis sa création, VALGO a articulé son développement
dans le but de maîtriser l'ensemble de la chaîne des valeurs inhérentes à la réhabilitation de
sites et sols pollués. VALGO met en œuvre son modèle 3D de Déconstruction, Dépollution
et Développement afin de parvenir aux objectifs affichés.

1.3 Caractéristiques géologiques de la zone

D’après la base de données du site Info Terre, l’ancienne raffinerie se situe en majorité sur
des alluvions fluviatiles provenant de la Seine. Les alluvions actuelles et subactuelles sont
à dominance limono-sableuses. Elles reposent sur les alluvions anciennes, composées
principalement de sable graveleux à silex. Sous ces alluvions, le socle est composé de craie
blanche à silex, formée de biozones de foraminifères datées entre le Santonien et le
Campanien (86 à 72 millions années), Figure 1.
Les coupes géologiques ont, d’une manière générale, permis (selon la position et
profondeur de l’ouvrage) d’identifier les 4 formations connues sur le secteur :

3
- Les remblais : sables, graviers, limons avec parfois de nombreux débris anthropiques
tels que des tuyaux, des fondations et du béton.
- Les alluvions modernes de la Seine : composées le plus souvent d’éléments fins (sables
à limons).
- Les alluvions anciennes : généralement composées d’éléments plus grossiers (sable,
graviers, silex).
- La craie : blanche à grise et souvent assez compacte et fissurée, Figure 2.

Figure 1. Extrait de la carte géologique au 1/25 000,

Note : 1 : Alluvions modernes ; 2 : Craie (Santonien, campanien) ; 3 : Formation à silex ; 4 : Alluvions anciennes.

Figure 2. Représentation des géologies affleurantes sur le relief.

4
La raffinerie repose sur une formation d’alluvions d’une épaisseur moyenne de 10 m. En
dessous, se trouve une couche de craie dépassement localement 100 m d’épaisseur. La craie
affleure au niveau de la zone Stockage Est

1.4 Types de pollutions concernées

Plusieurs zones de la raffinerie étant impactées par du LNAPL (Light Non Aquaous Phase
Liquid) mobile au-dessus du toit de la nappe, elle servira à l’étude d’un tel milieu. Les
enjeux du projet sont d’identifier et de suivre les secteurs impactés par le LNAPL dans un
contexte de battement de nappe (marnage journalier, mensuel et saisonnier). D’après
l’agence Américaine de Protection de l’Environnement (U.S. EPA, 1995), les liquides en
phase non aqueuse (NAPL) sont des hydrocarbures qui existent sous forme de phase non
miscible séparée, lorsqu'ils sont en contact avec l'eau et/ou l'air. Les différences dans les
propriétés physiques et chimiques de l'eau et de la couche NAPL entraînent la formation
d'une interface physique entre les liquides qui empêche les deux fluides de se mélanger.
Les liquides en phase non aqueuse sont généralement classés comme des liquides en phase
non aqueuse légère (LNAPL) qui ont des densités inférieures à celle de l'eau, ou des
liquides en phase non aqueuse denses (DNAPL) qui ont des densités supérieures à celle de
l'eau. Dans le cadre de ces travaux, la définition retenue dans le cadre des travaux présentés
est que les LNAPL sont des contaminants des eaux souterraines non solubles dans l'eau
ayant une densité plus faible que l'eau, Figure 3.
La synthèse réalisée à partir des diagnostics antérieurs (ANTEA, 2014 ; GOLDER, 2015),
complétés des données acquises par VALGO, a permis de recenser les principales
connaissances concernant les caractéristiques de la zone sur les aspects de réservoirs
(géologiques, pétrophysiques...) et de pollutions (sols, eaux, air). Ces informations ont été
compilées dans une base de données afin d'aider à la compréhension de la répartition des
hétérogénéités.

5
Figure 3. Schéma des structures géologiques de la raffinerie et répartition des LNAPL et
DNAPL en phases dissoutes, libres et résiduelles, d’après ABBAS et al., 2018.

La Seine passant à proximité n’est pas impactée par la situation sur la zone de raffinage du
fait de plusieurs facteurs, Figure 4 :
- Une couche limoneuse le long des berges, jusqu’à l’emprise du site dans sa partie sud
(ANTEA, 2013 ; données internes VALGO).
- Des palplanches le long des berges installées lors de la construction des aménagements.
- Echanges de matières entre nappe de la craie et Seine très faibles (<1 kg/h) (ANTEA,
2013).
- Faible mobilité des hydrocarbures (données VALGO).

Figure 4. Représentation schématique des couches limoneuses limitant le transfert de

LNAPL vers la Seine.

6
Les mesures révèlent que la nappe de la craie n’est pas impactée par la pollution présente
au sein de la nappe alluviale, (Tableau 1) :
- Bien qu’elle soit en continuité hydraulique avec la nappe alluviale, les échanges sont
très faibles ;
- Bien qu’elle soit vulnérable aux HAP (densité >1) et benzène (très soluble dans l’eau),
les concentrations mesurées pour ces deux types de composés sont très faibles.

Tableau 1. Impacts en hydrocarbures de la nappe phréatique sous la raffinerie.

Nappe alluviale Nappe de la craie Ref. OMS

Eau
Valeurs Minimales Maximales Minimales Maximales
potable

Hydrocarbures C5-C40
0,742 156,7 0,03 0,987 -
(mg/l)

Concentration 16 HAP
0,06 94 0,06 6,4 -
(µg/l)

Concentration 6 HAP*
<0,0575 <0,0875 <0,0575 <0,0875 0,2
(µg/l)

Benzène
<0,5 174 <0,5 1,31 10
(µg/l)
*somme des concentrations des HAP suivants : fluoranthène, benzo(b)fluoranthène, benzo(k)fluoranthène,
benzo(a)pyrène, benzo(ghi)pérylène, indéno(1,2,3, cd)pyrène, suivant les recommandations de l’OMS.

2. Travaux réalisés et résultats

2.1 Problématique de renforcement des données

Lorsque VALGO a pris en main le site en 2014, les rapports de suivi les plus récents
(rapport ANTEA, 2014, sur des données de 2012) étaient fondés sur 137 ouvrages de type
piézomètres, dont 75 présentaient une épaisseur de flottants (LNAPL), comprise entre 1 et
184 cm. Les ouvrages impactés étaient principalement situés le long d’une ligne médiane
de la raffinerie, parallèle à la Seine, Figure 5.

7
 Figure 5. Carte d’implantation des ouvrages initiaux (haut), Représentation graphique
des épaisseurs relevées dans les ouvrages (bas), (ANTEA, 2012).

Ces données initiales montraient l’étendue des zones à traiter et un volume indicatif
d’hydrocarbures à éliminer.
Mais, après des années durant lesquelles ces ouvrages servaient uniquement à la
surveillance réglementaire des impacts de la raffinerie, VALGO avait besoin d’augmenter
la densité d’information, pour adapter sa stratégie de dépollution, et a commandé, au
Bureau d’Etudes Golder Associates, de réaliser une centaine d’ouvrages supplémentaires,
tout en apportant des précisions sur la nature des hydrocarbures relevés (Figure 6). Cette
campagne d’investigation a rapporté des informations complémentaires sur la nature des
8
polluants présents, leur densité, leur viscosité, données indispensables pour définir les
modes de dépollution les plus pertinents.
Cette étude a également fait ressortir des résultats partiellement discordants avec l’étude
précédente, quant à la localisation des lentilles de LNAPL et à leurs épaisseurs.

Figure 6. Estimation des emprises et de la nature des couches d’hydrocarbures flottants

sur la nappe phréatique sous la raffinerie, (GOLDER, 2015).

Pour réduire ces incertitudes et augmenter l’efficacité de ses traitements, VALGO a

poursuivi les campagnes d’acquisitions de données, selon 3 voies principales :
- 4 ans d’essais récurrents et itératifs, notamment sur une zone modèle de 1 ha, au centre
du site (Figure 7).
- Un projet de recherche partenarial sur l’utilisation des outils géophysiques pour la
détection des hydrocarbures et l’imagerie des milieux fracturés (DE CLERCQ et al.,
2021).
- Des découvertes incidentes lors des précédentes recherches, portant sur l’influence des
cycles de marées sur les niveaux de l’aquifère, en lien avec la Seine.

9
 Figure 7. Localisation de la zone expérimentale, au centre du site.

2.2 Variations des niveaux piézométriques - hydrogéologie

Différentes études ont montré (KEMBLOWSKI & CHIANG 1990 ; ARAL & LIAO, 2002
; ITRC, 2009 ; DESKA & OCIEPA, 2013 ; ATTEIA et al., 2019), que, lorsque qu’un site
pollué est soumis à des battements périodiques de niveau de nappe, les épaisseurs
d’hydrocarbures sont corrélées négativement avec le niveau piézométrique : lors des hautes
eaux, l’épaisseur de LNAPL diminue fortement. Ce comportement est la résultante des
différences de viscosité et de densité entre l’eau et les hydrocarbures, ainsi qu’à des
phénomènes de « mouillabilité » différentielles des grains de sol, dont le comportement
hydrophile/hydrophobe évolue en fonction du fluide environnant.
Or, la Seine à Petit-Couronne est soumise à la marée de la Manche, qui remonte son cours,
jusqu’au barrage de Poses, premier ouvrage de régulation du débit depuis l’embouchure.
Les informations du marégraphe de Rouen montrent nettement des variations pouvant
atteindre 2 à 3 m par jour, avec un décalage d’environ 6 h par rapport à l’embouchure à
Honfleur.
Par la communication entre les masses d’eau souterraine, le niveau piézométrique du site
subi également des variations, très atténuées au fur et à mesure que l’on s’éloigne du lit du
fleuve. En prenant en compte cet état de fait observé, les mesures dans les ouvrages
piézométriques, relatifs tant à la hauteur d’eau qu’à l’épaisseur d’éventuels LNAPL,
devaient être complétées d’un relevé des marées du jour, voire synchronisés avec elles,
lorsque des comparaisons étaient requises. En effet, des relevés piézométriques effectués
en continu, pendant plusieurs heures, montrent bien une variation du niveau, deux fois par
jour.
Cependant, lors de relevés à plusieurs jours ou semaines d’intervalle, les différences de
niveau observé étaient mises sur le compte d’une recharge hydraulique par le plateau
10
crayeux, en lien, mais avec un fort effet de retard, avec les épisodes pluvieux.
Ce point de vue a radicalement changé lorsqu’un relevé a été mené en continu sur une
période de plusieurs semaines, Figure 8.
On observe très aisément 2 variations sinusoïdales enchevêtrées :
- Une de fréquence d’environ 12 h, correspondant à l’alternance de marées hautes et
basses.
- L’autre de fréquence 28 j, correspondant au cycle des lunaisons et du rythme de
l’alternance de grandes marées (de pleine lune et de nouvelle lune).
Une observation plus fine montre également que la hauteur moyenne sous le site (moyenne
mobile sur 12 h) est quasi synchrone avec la variation de coefficient du jour pendant une
partie du cycle et très décalée au moment des marées de plus forts coefficients (NL).

Figure 8. Mesure continue des variations piézométriques dans un ouvrage du site.

Pour faire ressortir ces périodes avec un effet retard plus ou moins prononcé, la Figure 9
présente les mêmes données, mais en exprimant la hauteur piézométrique en fonction du
coefficient de chaque marée et non plus en fonction du temps.

11
 Figure 9. Relation entre les hauteurs d’eau et les coefficients de marée. En bleu de morte-
eau (MH) vers marée de PL, en vert retour de PL vers MH, en orange de MH vers NL et retour en violet.

Cette connaissance accrue du comportement des hauteurs piézométriques a permis de bien

meilleures prédictions sur les niveaux rencontrés, les périodes propices aux mesures, aux
pompages, etc…

2.3 Qualifications des techniques de récupération

Outre les considérations sur le comportement hydrogéologique vu au 2.2, les équipes
œuvrant sur le site ont mené de longs essais (plusieurs semaines à plusieurs mois) pour
adapter les matériels et leurs mises en œuvre aux réalités du terrain.
Les techniques utilisées pour enlever une couche de LNAPL sont fondées sur différents
principes :
- L’écrémage pur, qui consiste à pomper uniquement la couche d’hydrocarbure, très
précisément, pour aspirer le moins d’eau possible. 5 pompes spécialement dédiées à
l’écrémage ont été mises en œuvre (modèles ou modes de fonctionnement différents)
- Le pompage écrémage qui vise à augmenter l’épaisseur d’hydrocarbures à éliminer, en
créant un cône de rabattement, par un pompage d’eau de la nappe sous-jacente. Lorsque
celle-ci n’est pas impactée par la pollution, l’eau peut être rejetée sans traitement. Des
pompes immergées servent au rabattement, tandis que des pompes écrémeuses
éliminent les hydrocarbures.
- Le pompage du mélange eau-hydrocarbure est moins complexe à mettre en œuvre, mais
nécessite d’être suivi d’un séparateur en surface. Différents modèles de pompes,
immergées ou de surface peuvent être utilisées, Figure 10.

12
 Figure 10. Exemples de matériels utilisés pour les essais.
De gauche à droite : 2 pompes écrémeuses, pompe immergée, séparateur coalesceur.

Toutes les techniques utilisées ont montré une diminution très rapide des hydrocarbures
dans les puits, puis un retour du niveau d’huile en quelques jours. En fonction de chaque
puits, cette vitesse de réalimentation était très variable, mais toujours très inférieure à la
vitesse d’écrémage. Il a donc fallu ajuster des durées de cycles alternant pompage et arrêt
pour réalimentation, pour chacun des puits à traiter, Figure 11.
Figure 11

Figure 11. Suivi des épaisseurs d’hydrocarbures lors des essais d’écrémage,
(Pompage séquentiel de 3 à 5 cycles / jour / pompe).

13
2.4 Elargissement à l’ensemble du site
Du fait de la variabilité des comportements de réalimentation de chaque puits, l’extension
des connaissances acquises sur la zone d’essai, n’a pu se faire par généralisation de la
« meilleure » technique, soit celle apportant le meilleur rapport efficience/coût, Figure 12.

Figure 12. Diversité des moyens de pompage des hydrocarbures employés et répartition
sur une carte du site.

L’optimisation des séquences de pompage, avec des ajustements pour chaque puits, en
fonction de la matrice minérale encaissante et de la nature de LNAPL présent, a permis de
maîtriser les vitesses de réalimentation des puits, par retour à l’équilibre des forces en jeu.
Ceci a permis d’améliorer la récupération des masses d’hydrocarbures flottant sur la nappe
phréatique.
Cependant, la diversité des matériels et des techniques de pompage en forage n’a pas suffi
à créer des conditions optimales pour chaque configuration polluant/matrice. Une
technique mal adaptée conduit à des déséquilibres financiers, mais également à des
aberrations environnementales, l’énergie dépensée par quantité de polluant éliminé
croissant rapidement quand peu de LNAPL est pompé.
D’autres techniques de récupération des hydrocarbures ont été explorées, adaptées et mises
en œuvre, en fonction des zones du site.
Par exemple, dans une ancienne zone dédiée à la production d’hydrocarbures de type huiles
épaisses et bitumes, le LNAPL était trop visqueux pour les pompes et la captation a été
réalisée par bande écrémeuse oléophile, avec un cycle de 15 min par heure.

14
De larges puits (>60 cm) ont également été forés, permettant un contrôle visuel et un
pompage sous vide, par camion hydrocureur ou pompe à vide et réservoir fixe, à la
demande.
Dans la zone sud-ouest du site, grâce à la proximité de la nappe (environ 2 m sous le niveau
du sol), une vingtaine de fosses a été creusée, révélant visuellement le comportement des
hydrocarbures. Les dispositifs de pompage utilisés ont été également de l’hydrocurage par
camion, mais également des déshuileurs flottants, actionnés par une pompe à double
membrane.
Dans ces cas, le contrôle visuel permet d’agir dès la réalimentation en hydrocarbures,
même si la périodicité est très aléatoire et d’observer les comportements. Notamment le
fait, récurrent en plusieurs emplacements du site, que la réalimentation s’arrête après 4 ou
5 campagnes de pompage (Tableau 2), alors que les piézomètres présents sur la zone, à
proximité des fosses ouvertes, montrent toujours des épaisseurs conséquentes de LNAPL
(40 à 60 cm), Figure 13.

Tableau 2. Récapitulatif des pompages-écrémages effectués sur 3 zones du site.

Zone Durée V pompé Réalimentation Séquences de pompage
(m3) flottant
F n°1 3 mois 1 3-5 L /j 4 campagnes
Hydrocureur / Pompe à membrane
F n°2 3 mois 2 0,5-5 L/j 5 campagnes
F n°3 Hydrocureur / Pompe à membrane

Figure 13. Diversité des aspects des affleurements de la nappe dans des fosses ; la flèche
représente une distance de 100 m.

15
2.5 Révision des objectifs fonctionnels
Les nombreuses observations de comportement des hydrocarbures, réalisées lors de la
qualification des techniques et lors de leurs applications sur l’ensemble du terrain à
dépolluer, ont conduit à reconsidérer le modèle classique ayant présidé à l’établissement
des objectifs initiaux.
Les objectifs de dépollution sont établis, par un processus technique et administratif, pour
atteindre des objectifs sanitaires. Les enjeux attachés à la gestion de la zone de raffinage,
exposés dans le plan de gestion, concernent :
- La solubilisation des hydrocarbures.
- La migration verticale des hydrocarbures, notamment vers la craie.
- Le risque de déplacement selon un axe nord-sud de la phase LNAPL, en direction des
zones dédiées à de l’habitat.
Le traitement des terres impactées concernerait 270 000 m3 ; mais la réalisation d’un
biotertre de 10 à 20 ha est incompatible avec les objectifs de développement du site et la
solution de traitement par voie biologique n’a pas été retenue comme pertinente pour
l’atteinte des objectifs.
Lorsque les quantités de sols et de polluants sont très importantes, il est parfois utile, et
reconnu, d’appliquer le principe de Pareto, afin d’optimiser l’effort fourni pour un résultat
maximisé. Les prévisions par calculs d’une application possible de ce principe, font
ressortir que traiter les zones les plus concentrées ne permet pas de traiter 80% des
polluants selon le principe de Pareto (voir Tableau 3). En effet, traiter ces 80% demanderait
de traiter une partie de la catégorie la moins impactée des sols du site : ceux qui ont été
classés sous la dénomination « 500 à 2500 mg/kg de MS », ce qui conduirait à traiter plus
de 50 % de la masse de sols impactés.

Tableau 3. Répartition de la masse des polluant en fonction des concentrations.

Plage de concentration Vsol concerné Mhydrocarbure dans ce Part de cette masse dans la Part
(g/kg) (m3) volume (t) masse totale de polluants cumulée
10 – 20 15 584 350,64 4,31 % 4,31 %
7,5 – 10 35 952 404,46 4,98 % 9,29 %
5 – 7,5 175 654 1 646,76 20,26 % 29,54 %
2,5 – 5 604 858 3 402,33 41,85 % 71,39 %
0,5 – 2,5 1 033 606 2 325,61 28,61 % 100,0 %

Nous avons mené plusieurs modélisations innovantes, comme :

- L’études des oscillations à différentes fréquences (marées, voir section 0).
- Les relevés périodiques pendant plusieurs années des valeurs d’hydrocarbures dissous.
- L’étude des variations des épaisseurs de flottant et leurs liens avec la hauteur d’eau et
la SOR.

16
Ces connaissances acquises par modélisations ont renforcé l’accumulation des
observations sur ces comportements, pour écarter les conclusions simplistes du modèle
classique:
- La hauteur de flottant dans un puits n’est pas un indicateur fiable de sa mobilité.
- La mobilité est contrôlée par le type de sol, l’épaisseur de surnageant.
- La mobilité est contrôlée également par la saturation du sol aux HC et leur viscosité.
Nous avons établi que la transmissivité du sol aux HC intègre l’ensemble de ces variables,
en représentant d’une façon plus fiable le comportement observé, déterminé sur site, lors
des essais d’épuisement.
En s’appuyant sur les données tirées du guide de l’ITRC (2009 ; Figure 14), relatif à la
définition des critères d’arrêt dans le cadre de la dépollution de sites contaminées par les
HC flottants, nous avons proposé une gestion focalisée sur un nombre restreints de zones
à traite, en appliquant notamment l’hypothèse qu’une transitivité inférieure à 10-7m2/s
correspond à des HC flottants non mobilisables.

Figure 14. Diagramme théorique de calcul de la part d’hydrocarbures mobilisables

(ITRC, 2009).

2.6 Mise en place d’un nouveau protocole intégratif

Les lithologies typiques du site (sable, sable fin, limono-argilo-sableux), ont été définies
par leurs caractéristiques intrinsèques (permissivité et porosité efficace).
Les catégories d’hydrocarbures, retrouvées dans le sous-sol, ont été caractérisées par leur
viscosité et leur masse volumique.
14 configurations, combinant une lithologie et un type de polluant, représentatives des
différentes conditions réelles retrouvés sur le site de l’ancienne raffinerie, ont été testées
au cas par cas, dans des colonnes de laboratoire et comparées, entre elles et avec les
modèles théoriques, Figure 15.

17
 Figure 15. Essai de détermination de la SOR en colonnes.

Toutes les valeurs obtenues ont été intégrées dans un protocole de gestion des épaisseurs
résiduelle, validé par les autorités compétentes et qui comprend les étapes suivantes :
- Les tests d’épuisement sur site (bail-down tests), par les données de réalimentation,
informent sur la transmissivité du sol aux hydrocarbures ;
- Si cette valeur est inférieure à la limite ITRC de 10-7m2/s, des échantillons d’eau,
d’hydrocarbures et de matrice encaissante sont prélevés et portés au laboratoire ;
- Sinon, les méthodes de pompage classiques sont appliquées ;
- Les essais en colonne permettent de déterminer l’Epaisseur Résiduelle Objectif pour
chaque couple matrice/polluant :
- Cette ERO est appliquée lors de la mise en œuvre de la dépollution sur site.
L’amélioration de la maîtrise des comportements de LNAPL a ainsi permis d’adapter les
objectifs de pompage appliqués à chaque zone. Nous donnons ici pour exemple, les
données réelles tirées des travaux de réhabilitation sur l’ancienne zone de fractionnement
catalytique et distillation sous vide, dite FCC-DSV. L’équipement était constitué de deux
unités de pompage :
- 1 Extraction double phase (250 m3/h).
- 1 Pompage - écrémage muni de 12 pompes pneumatiques.
Les volumes d’hydrocarbures pompés de 2016 à 2019 dans cette zone sont illustrés dans
la Figure 16, tandis que le suivi des épaisseurs de LNAPL durant le traitement est montré
en Figure 17, constituant la conséquence des pompages.

18
 Figure 16. Bilan du pompage de la zone FCC-DSV ; EDP : extraction double phase ;
PdM : Pompe à double membrane ; PE : pompage-écrémage.

Figure 17. Evolution des épaisseurs de LNAPL, dans 3 puits de la raffinerie.

3. Conclusion - discussion
L’application du concept classique de pompage ou pompage-écrémage, au site de Petit-
Couronne, arrivait à ses limites techniques et ne permettait pas d’atteindre les objectifs de
dépollution, propres à assurer une qualité sanitaire pour les usages futurs. Cette difficulté

19
était notamment liée aux particularités géologiques et hydrogéologiques du site, comme la
susceptibilité à la marée, transmise par la Seine.
La multiplicité des gammes de polluants pétrolier apportait également un niveau
supplémentaire de complexité.
Quatre années d’essais sur site, de tests en laboratoire, de relevés de terrain et de
modélisation ont conduit à changer de concept et à abandonner l’idée commode d’une
lentille hydrophobe flottant à la surface de la nappe phréatique. Ce concept, issu des
observations et applicable en eau libre, ne l’est plus lorsque les interactions avec la matrice
encaissante priment sur les propriétés favorisant la séparation des 2 fluides, notamment la
densité, Figure 18.
Dans le nouveau modèle, dit intégratif, la notion de « couches » distinctes de composés
« purs » doit alors être remplacée par des « volumes » contenant des proportions variables
de chaque espèce chimique, ouvrant la réflexion à de nouveaux paradigmes.

Modèle « classique » Modèle « intégratif »

⚫ Equilibre ⚫ Dispersion hétérogène
thermodynamique des phases
vertical ⚫ Non Equilibre local
⚫ LNAPL surnageant ⚫ Distribution en fonction
⚫ Séparation de phases de l’équilibre des forces
homogènes capillaires et de gravité
Figure 18. Illustration du concept d’interface non-plane entre eau et LNAPL.

Il en résulte notamment une amélioration du modèle de percolation biphasique, dont la

version classique serait applicable, si le temps de relaxation était très grand au regard des
cinétiques de migration des fluides. Or, dans le cas étudié ici, l’interface eau/LNAPL est
en perpétuel équilibre dynamique, la couche d’hydrocarbure étant soumise périodiquement
à des forces ascendantes (poussée d’Archimède) et descendantes (gravité), alternativement
à co-courant et à contre-courant des mouvements de l’eau.
La Figure 19 illustre ainsi les différentes notions de « saturations » qui résultent de ces
nouveaux concepts et expliquent la relation complexe, non-linéaire, entre la quantité
d’hydrocarbures présente dans la matrice et la part de cette masse susceptible de s’écouler
vers les points de pompage lors des opérations de dépollution. La mobilité prime alors sur
le simple aspect quantitatif.

20
 Figure 19. Représentation des dispositions du biphase eau/huile dans le sol,
(haut : BRGM, 2016 ; bas : ADEME, 2010 ).

Sur le site de l’ancienne raffinerie de Petit-Couronne, c’est la prise en compte de cette

complexité qui a permis d’atteindre les objectifs sanitaires et de réceptionner les travaux
des parcelles passant aujourd’hui à l’étape d’aménagement, pour accueillir les activités
industrielles et commerciales de demain.

4. Références bibliographiques
ABBAS M., JARDANI A., MACHOUR N., DUPONT J.-P. (2018). Geophysical and
geochemical characterization of a site impacted by hydrocarbon contamination
undergoing biodegradation. Near Surface Geophysics, Vol. 16(2), pp. 176-192,
doi.org/10.3997/1873-0604.2017061
ADEME (2010). Elaboration d’une base de données de saturations résiduelles et de
relations perméabilité relative / saturation et pression capillaires / saturation. Rapport
final n°RRD03120b / CRDZ071063 / A6273, Coordination technique de DUCLOS Y.,
KERGARAVAT O., ROUSSEL H.
ANTEA (2013). Rapport sur la raffinerie de Petit-Couronne n°A71576/A.
ANTEA (2014). Rapport sur la raffinerie de Petit-Couronne.
ARAL M.M., LIAO B. (2002). Effect of groundwater table fluctuations on LNAPL
thickness in monitoring wells. Environmental Geology, Vol. 42 (2), pp. 151-161,
doi.org/10.1007/s00254-001-0485-8
ATTEIA O., PALMIER C., SCHAEFER G. (2019). On the influence of groundwater table
fluctuations on oil thickness in a well related to an LNAPL contaminated aquifer. Journal
of Contaminant Hydrology, Vol. 223, 103476, doi.org/10.1016/j.jconhyd.2019.03.008

21
BRGM (2016). Définir une stratégie de dépollution : approche basée sur la masse de
polluant et la capacité de relargage d’une pollution, Rapport n°64350.
DE CLERCQ T., THANNBERGER L., JARDANI A. (2021). 3D geophysical
characterization of LNAPL volume extension, Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, (in press).
DESKA I., OCIEPA E. (2013). Impact of the water table fluctuations on the apparent
thickness of Light Non-aqueous Phase Liquids. Ecological Chemistry and Engineering,
Vol. 20, pp. 771-778, doi. org/10.2428/ecea.2013.20(07)070.
GOLDER (2015). Rapport sur la raffinerie de Petit-Couronne.
ITRC - Interstate Technology & Regulatory Council. (2009). Evaluating natural source
zone depletion at sites with LNAPL. Washington, D.C.
KEMBLOWSKI, M. W., C.Y. CHIANG. (1990). Hydrocarbon Thickness Fluctuations in
Monitoring Wells. Ground Water, Vol. 28, pp. 244–252, doi.org/10.1111/j.1745-6584.1990.tb02252.x
U.S. EPE – US Environmental Protection Agency. (1995). Ground Water Issue: Light Non-
Aqueous Phase Liquids.
BLOG EKLABLOG, WOKUZY. (2012), 31 Décembre. http://petroplus.eklablog.com/histoire-de-la-
raffinerie-de-petit-couronne-c19234325

22
ENVIRONMENTAL RISK ASSESSMENT
&
REMEDIATION MANAGEMENT
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Assessment of marine submersion's risk on the coastal

morpho-dynamics of Bejaia, Algeria

Cherif AOUDJ 1, 2, Mokhtar GUERFI 2, 3, Khoudir MEZOUAR 1, 2

1. National Higher School of Marine Sciences and Coastal Management, ENSSMAL,

Bois des Cars, B.P. 19, Dely Ibrahim 16320, Algeria.
[email protected]
2. Marine and Coastal Ecosystem Laboratory, EcoSys, Chemin du complexe de Sidi
Fredj, Staoueli 16096, Algeria.
3. University of Sciences and Technologies Houari Boumediene, USTHB, BP 32, El
Alia, Bab Ezzouar 16111, Algeria.

Abstract:
Submergence, the temporary inundation of the coastal zone by sea water, is a major risk
for the coast. This phenomenon can be much more brutal, even if the risk of a tidal
wave can be considered very unlikely.
The phenomena of flooding are all the more to be considered as they concern vast
urbanized areas of the coast, including all or part of urban areas, such as the Bay of
Bejaia for example.
The knowledge of waves and swells at the coast can inform on the coastal sensitivity to
marine erosion.
Our work was to process wave data off the study area (Bejaia coastline) in order to
study, through ArcGis software, the effect of waves for each sector of the Bejaia
coastline. Once known, the waves at the coast will be used to identify the areas most
impacted by the energy of erosive waves.
The eastern part of the coastline is the most sensitive to be threatened by marine
submersion.

Keywords: ArcGis, Assessment, Coast, GIS, Marine submersion, Morpho-dynamics.

Corresponding author mail: [email protected]

1. Introduction
In the last century, surface temperatures have shown a gradual increase based on
National Weather Service data, and tide gauges have recorded about a 30 cm rise along
the U.S. Atlantic coast (National Research Council, 1987). Some of this relative rise in
sea level (relative to the land surface) can be explained by the natural compaction and
subsidence of unconsolidated coastal sediments. However, important regional
variability has also been reported. Knowledge of waves and swells at the coast can,
among other things, provide information on coastal sensitivity to marine erosion
(AOUDJ et al., 2017). Currently, the problems of coastal areas in the Mediterranean
come much more from the effect of climate change, in this regard the study of sea level
rise on the Bejaia Bay zone, is a tool that allows us to have a vision on coastal risks, and
to predict the risk we agreed to make a simulation of sea level rise to facilitate the
development of a management plan for natural disasters.

2. Study area
The Bay Coast of Bejaia is located in the central-eastern part of the Algerian coast, 220
km East of the capital Algiers. It extends on 100 km of coastline between Cape Carbon
to the west and Cape Ziama Mansouria to the East. The study area lies between parallels
36° 37′ 37″ N and 36° 46′ 32″ N and meridians 5° 00′ 55″ E and 5° 25′ 13″ E (Figure 1).

Figure 1. Localization of the study area (Bejaia bay coast).

Overall, the mountains, the valley of Soummam and the coastal area are the main
topographic sets that conceal the wilaya of Bejaia. The first ones dominate very largely

24
the others to the point of becoming without question authentic landmarks and symbols
of Kabylia.

3. Data and methods

The simulation of sea level rise was carried out by means of two software programs,
namely:
- ArcGis, a geographic information software developed by the American company Esri,
this system is composed of different platforms that allow GIS users to collaborate and
share geographic information.
- Global Mapper, which is more than a visualization tool capable of displaying raster
images, elevation data and vector data. It converts, edits, prints, acquires GPS data, and
allows you to use GIS functionalities on your datasets in a cheap and simple way.
Global Mapper also provides direct access to the entire USGS TerraServer database of
satellite imagery and topographic maps, as well as the visualization of elevation data in
true 3D with draping of any raster image or vector data (Figure 2).

Figure 2. Interface of the Global Mapper with the Digital Elevation Model of Bejaia.

25
 Figure 3. Methodology.

Based on both a satellite image of 2019 from the National Institute of Cartography and
Remote Sensing and a digital terrain model downloaded from the vertex site, we were
able to process the data in ArcGis and subsequently assess the consequences of sea level
rise in our study area (Figure 3).

Figure 4. Sea level and other parameters observed from satellite altimetry for the
period 1993 – 2010.

The above Figure 4 represents Sea level observed from satellite altimetry for the period
1993 - 2010 (blue solid curve) Thermal expansion (red curve, average value based on
temperature data from LEVITUS et al., (2012) and ISHII & KIMOTO (2009)).
Contribution from Greenland and Antarctica (cyan curves) and glaciers (green curve).
The black curve represents the total land ice contribution while the blue dashed curve
represents the total climate contribution (sum of thermal expansion and land ice)
(updated from CAZENAVE & LLOVEL (2010)). According to this diagram, we have
opted for a sea level rise scenario of 0.76 meters.
26
4. Results and discussion
For each of the two parts of the coast of Bejaia Bay, the areas of land likely to be lost by
marine submersion to Mediterranean marine waters, for the 0.76m sea level rise
scenario and for each time horizon.
The scenario 0.76 meters, informs us of an estimated surface loss of 976 hectares. It can
be seen on the map (Figure 5), that the area at risk is located in the central and western
part of the bay and this is due to the morphology of the region (beaches), where there
are low slopes, while the eastern region shows a resistance to sea level rise due to its
morphology (cliff area).

Figure 5. Vulnerability of the Bejaia bay to sea level rise.

The estimate of the various damages caused by the rise of the sea level in the bay of
Bejaia are summarized in the table below:

Table 1. Monetizing the negative impacts of sea level rise.

Type of impact Losses

Erosion-related Damage to land and residential or commercial buildings.

Submergence-related Emergency evacuation and infrastructure damage.
Economic Loss of merchandise and tourist revenues.
Environmental Loss of natural habitat.
Social Loss of access to the sea and sea views.
Decrease of the quality of life.
Deterioration of the historical and cultural heritage.

27
The value of the tide, for the study area, is more or less 34cm with a maximum level of
0.5 m NGA (General Level of Algeria) and the lowest sea is at -0.34 m NGA.
Therefore, the coastal area of Bejaia, like all the Algerian coast, is under the influence
of a micro tidal environment that makes it at very high risk to the rise in sea level.

5. Conclusions
In conclusion, the results of the study of sea level rise on the coastline of Bejaia showed
that it is microtidal type. Therefore, the absence of the intertidal zone strongly increases
the coastal sensitivity of the coastline of Bejaia.
The rise in the average level of the Mediterranean Sea as a result of global warming
does not help the current situation, which is worrying in many respects. These
phenomena add an additional lot to the constraints and pressures that are related to the
evolution of this area.
Climate change will have different impacts on human activities and other drivers of
change on the coast (economic development, tourism, agriculture, fisheries...),
knowledge of the impact of climate change and its associated risks are essential in
strategies in the current and future coastal development.

6. References
AOUDJ C., MEZHOUD A., GUERFI M., HEMDANE Y. (2017). Analyse des
variations spatio-temporelles du littoral sableux : Est Béjaoui (Algérie). Conférence
Méditerranéenne CM2, Split Croatie, pp. 7-10. https://dx.doi.org/10.5150/cmcm.2017.002
CAZENAVE A., LLOVEL W. (2010). Contemporary sea level rise. Annual Review of
Marine Science, pp. 145-173, Annual Reviews, Palo Alto, Calif., doi.org/10.1146/annurev-
marine-120308-081105
ISHII M., KIMOTO M. (2009). Reevaluation of historical ocean heat content
variations with time-varying XBT and MBT depth bias corrections. J Oceanogr 65, pp.
287–299, doi.org/10.1007/s10872-009-0027-7
LEVITUS S., ANTONOV J.I., BOYER T.P., BARANOVA O.K., GARCIA H.E.,
LOCARNINI R.A., MISHONOV A.V., REAGAN J.R., SEIDOV D., YAROSH E.S.,
ZWENG M.M. (2012). World ocean heat content and thermosteric sea level change (0-
2000 m), 1955-2010. Geophysical Research Letters 39(10). doi :10.1019/2012GL051106

28
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

3D geophysical characterization of LNAPL volume extension

Caractérisation 3D par géophysique de l’extension d’une

lentille de LNAPL

Théo DE CLERCQ 1, Laurent THANNBERGER 1, Abderrahim JARDANI 2

1. VALGO, PIC, 72, rue Aristide Briand, 76650 Petit-Couronne, France.

[email protected]
2. Université de Rouen Normandie, UFR Sciences et Techniques, UMR 6143 CNRS -
M2C, Place Emile Blondel, 76130 Mont-Saint-Aignan, France.

Abstract:
The measurements required to obtain an overview of the parameters needed for
designing the management plan is essential for cost-effectiveness optimization. This
work is often complex due to the heterogeneity of environmental properties and
pollution. In this context, geophysical and hydro-geophysical methods, intrinsically
spatialized and considered as non-intrusive, offer an alternative to conventional
investigative techniques for characterizing a hydrocarbon pollution type.
In this context, the objective of the GEOCARE project is to assess the potential of these
methods to characterize a free LNAPL volume extension and to follow up on the
pumping/skimming treatment technique.
Several geophysical methods were tested in cubitainers and on a pilot site, previously
characterized, in order to correlate geophysical measurements to field observations.
The geophysical method studied in the proposed presentation is the electrical resistance
tomography, in a version with the peculiarity of having electrodes arranged on the
surface in 2D, but also vertically: either in the wells or directly in contact with the soil.
This characterization method allows to highlight 3D high electrical resistances in
correlation with the quantity of hydrocarbons observed in wells.
The work was also carried out on GPR (Ground Penetrating Radar) investigation to
characterise the extension of a LNAPL contaminated site.

Keywords:
3D characterisation, Geophysics, Electrical resistivity, Georadar, Refinery, LNAPL,
Polluted soils, Groundwater, Capillary fringe, Pumping / skimming.

Corresponding author mail: [email protected]

Résumé :
Les mesures nécessaires à la définition des caractéristiques servant à la conception d’un
plan de gestion sont essentielles dans une optique d'optimisation de la rentabilité. Ce
travail est complexe en raison de l'hétérogénéité des propriétés environnementales et de
la pollution. Dans ce contexte, les méthodes géophysiques et hydro-géophysiques,
intrinsèquement spatialisées et principalement non intrusives, offrent une alternative aux
techniques d'investigation conventionnelles pour la caractérisation de pollutions aux
hydrocarbures.
Dans ce contexte, l'objectif du projet GEOCARE est d'évaluer le potentiel de ces
méthodes, pour caractériser une extension de volume de LNAPL libre et de suivre son
traitement par pompage / écrémage par méthodes géophysiques.
La corrélation des mesures géophysiques aux observations de terrain a été réalisée par
tests de plusieurs méthodes géophysiques dans des cubitainers et sur un site pilote,
préalablement caractérisé.
Ce rendu s’attarde tout particulièrement sur la tomographie par résistivité électrique
dont les électrodes sont disposées en surface en 2D mais aussi verticalement : soit dans
les puits, soit directement en contact avec le sol. Cette méthode de caractérisation
permet de mettre en évidence des résistances électriques plus importantes en corrélation
tri-dimensionnelle avec la quantité d'hydrocarbures observée dans les puits.
Les travaux se sont également intéressés au GPR (Géoradar) dans le but de caractériser
une extension de panache de pollution au LNAPL disséminé.

Mots clés :
Caractérisation 3D, Géophysique, Résistivité électrique, Géoradar, Raffinerie, Sols
pollués, LNAPL, Nappe souterraine, Zone capillaire, Pompage/ écrémage.

1. Introduction
Ce projet de recherche se déroule au niveau de l’ancienne raffinerie de Petit-Couronne.
La zone d’étude, impactée par du LNAPL mobile sur le toit de la nappe, sert de cas
modèle d’un tel milieu. Les enjeux du projet sont d’identifier et de suivre le LNAPL
dans un contexte de battement de nappe (marnage journalier, mensuel et saisonnier), par
méthodes géophysiques.
D’après l’agence Américaine de Protection de l’Environnement (MCCAULOU et al.,
1995), les liquides en phase non aqueuse (LNAPL) sont des hydrocarbures qui existent
sous forme de phase non miscible séparée lorsqu'ils sont en contact avec l'eau et/ou l'air.
Les différences de propriétés physiques et chimiques, entre l'eau et la couche NAPL,
entraînent la formation d'une interface physique entre les liquides.

30
1.1 Une contamination intense et étalée par le marnage
La compilation des données d’échantillons sur une même verticale permet de préciser
les variations d’intensité de pollution avec la profondeur. Dans la zone d’étude du
projet, la zone entre 4 et 6 m sous la surface, correspondant à la zone de battement de la
nappe, est celle qui présente les concentrations d'hydrocarbures les plus importantes.
Ces profils sont caractéristiques des empreintes laissées par le comportement des coupes
pétrolières classiques (essence, gazole …) qui surnagent au-dessus de la nappe et
imprègnent les terrains de la zone de battement de nappe.
Dans cette zone, la moyenne des concentrations en C5-C40 est de 16 g/kg de matière
sèche et la médiane de 6. La zone superficielle présente des concentrations assez
variables. En dessous de -6 m, les concentrations diminuent progressivement et sont
globalement inférieures à 1 g/kg MS à partir de -10 m.

1.2. Le LNAPL dans le sol et dans les puits

L'épaisseur apparente de LNAPL dans un puits est liée aux conditions hydrogéologiques
et aux caractéristiques du LNAPL et du sol. Les fluctuations de la nappe phréatique, en
lien avec la saisonnalité, les coefficients de marée et les fluctuations tidales
bijournalières, influencent la présence de phase libre de LNAPL dans le sol et dans les
puits. Les changements d'épaisseurs observables dans les puits sont le résultat de
changements dans la distribution verticale des saturations en LNAPL et n'indiquent
généralement pas de changement dans l'étendue latérale de la lentille de LNAPL
(MARINELLI & DURNFORD, 1996).

Figure 1. Aperçu des expériences de colonnes de drainage par gravité et distribution

théorique de la pression (modifié de TANAHASHI et al. 2006).

L'épaisseur de LNAPL apparente dans un puits est communément exagérée, comparée à

l'épaisseur de NAPL mobile dans le sol (Figure 1). En aquifère non confiné et soumis à
la marée, comme c'est le cas sur le terrain d'étude, l'épaisseur de LNAPL dans un puits
est corrélée de façon négative à la hauteur de l'aquifère. L'épaisseur de LNAPL dans les
31
puits augmente lorsque la nappe phréatique s'abaisse. A l'inverse, à mesure que la nappe
phréatique monte, le LNAPL est piégé dans le sol et son épaisseur apparente dans le
puits diminue.
La caractérisation d’une lentille de LNAPL libre dans un milieu très hétérogène,
nécessite un maillage serré pour la réalisation d’analyses chimiques des échantillons
provenant de forages (Figure 2). La réalisation de puits et piézomètres apporte une
information limitée dans l'espace, mais il est possible d'utiliser ces ouvrages pour un
suivi temporel. Exploiter l’espace entre les puits via des mesures géophysiques permet
de diminuer le nombre d’ouvrages réalisés et d’avoir accès à une information
tridimensionnelle, sur l'état de pollution du sol.

Figure 2. Profils verticaux de concentration en C5-C40 de 7 ouvrages avec en bleu

clair la zone de battement de nappe observée et le ruban bleu foncé la profondeur de la
nappe au moment de la réalisation du sondage.

2. Calibration de la réponse de résistivité électrique

De nombreuses études se sont penchées sur la détection d'hydrocarbures par des
méthodes géophysiques sensibles aux propriétés des fluides, avec notamment le radar
géologique (BRADFORD, 2003 ; BERTOLLA et al., 2014 ; BABCOCK et al., 2015) la
tomographie de résistivité électrique (DAILY & OWEN, 1991 ; SHIMA, 1992 ;
SAUCK et al., 1998 ; ATEKWANA et al., 2004 ; DENG et al., 2018 ; IRAVANI et al.,
2020) ou encore les méthodes de polarisation induite (IP) (OROZCO et al., 2012).
Cependant, aucune d'entre elles n'a cherché à établir un lien entre une épaisseur
d'hydrocarbures sur la nappe (quantifiable dans les puits) et un signal géophysique.

2.1 Matériel et méthode

Les mesures de tomographie de résistivité électrique sont réalisées par 4 tiges
d’électrodes verticales (nommées TEV par la suite), composées chacune de 15
électrodes utilisées pour la mesure de la résistivité, réparties verticalement et espacées
de 5 cm les unes par rapport aux autres entre la surface et 70 cm de profondeur. Trois
32
des tiges d’électrodes verticales sont disposées suivant une forme triangulaire autour
d’une TEV centrale) (Figure 3). La configuration prévue sur le terrain lors de l’essai
pilote de suivi d’un pompage écrémage de LNAPL a été reproduite dans le cubitainer.

Figure 3. Plan de la surface du cubitainer, positionnement des tubes (piézomètres et

électrodes de résistivité verticales) et du point d’injection d’hydrocarbures. La
délimitation rouge représente la zone de mesures des résistivités électriques. La
photographie de droite représentant les électrodes localisées tous les 5cm sur les TEV.

Les 60 électrodes sont utilisées pour la réalisation de 3495 quadripôles suivant la

configuration dipôle-dipôle AM-BN. La durée moyenne de réalisation des mesures à
l’aide du Syscal Pro Switch est de 3h30 pour ce pilote. Le modèle d’inversion est basé
sur un maillage de 2,5 cm (moitié de l’espacement inter-électrodes). La représentation
des parois du cubitainer a été réalisée dans le modèle, en fixant à 0 les flux de courant
par la condition de limite de Neumann (CHAPKO et al., 1999 ; LI, 2011 ; BROGGINI
et al. 2017).
Du LNAPL en phase pure a été déversé progressivement (1,5 L/h) directement au
niveau de la zone saturée via un puit d’injection. Cette injection s’est déroulée en deux
phases de 8 h chacune. Au total, 24 L de LNAPL en phase pure ont été injectés. Une
semaine après l’injection, l’épaisseur moyenne de LNAPL dans les piézomètres était de
6,5 cm. Par transparence des parois du cubitainer, l’épaisseur de LNAPL mobile dans le
cubitainer a été évaluée à 5 cm.

33
Un total de 6 tomographies de résistivité électrique verticale ont été effectuées au cours
des quatre jours de suivi de l’évolution de la répartition du LNAPL injecté. Ces mesures
ont été comparées à l’état initial, réalisé juste avant le début de l’injection de LNAPL.

2.2 Suivi d’injection de LNAPL en cubitainer pour calibration des mesures de résistivité
électrique
Afin de comprendre comment réagissent les signaux géophysiques à la présence de
pollution de type NAPL dans un aquifère, la réalisation de mesures dans un milieu
contrôlé permet de calibrer la réponse attendue par tomographie de résistivité électrique
verticale. Dans la présente étude, la représentation de la zone d’intérêt a été réalisée à
l’échelle d’un cubitainer, rempli de matériaux limoneux provenant de la zone de
battement de nappe de la zone pilote. Le volume du cubitainer, proche de 1 m3, permet
d’observer les variations spatiales et temporelles, en réponse à une pollution évoluant en
trois dimensions à la surface de la nappe, dont le niveau a été fixé à 45 cm du fond du
cubitainer.

Figure 4. Variations de résistivités (en %) au cours de l’étalement d’une lentille de

LNAPL par injection progressive de LNAPL à hauteur de la partie saturée.

La mesure de piézométrie réalisée 8 heures après l’injection de LNAPL n’a pas permis
de détecter de produit pur au niveau des piézomètres. Cependant, la tomographie
réalisée en parallèle montre un changement de résistivité est notable dans l’angle de
l’injection de LNAPL. Les tomographies suivantes illustrent l’avancée progressive de la
lentille de LNAPL en phase libre à la surface de la zone saturée, confirmée par le suivi

34
des niveaux piézométriques. Le suivi temporel permet d’illustrer, en trois dimensions, la
zone de plus forte variation de résistivités liées à l’avancée de la phase pure. Les plus
fortes variations de résistivité (Figure 4) sont visibles sur 5 mailles verticales
(équivalent à 10 cm). Les épaisseurs obtenues lors de l’inversion sont donc supérieures
aux épaisseurs de LNAPL mesurées dans les piézomètres du cubitainer et dans le sol
(par transparence).
La réalisation d’une carotte dans la partie centrale du cubitainer a permis de comprendre
la répartition d’hydrocarbures dans le sol :
- 5 cm de limons sableux saturés en hydrocarbures et une phase mobile visible
- 15 cm de limons sableux saturés en eau et colorés de façon graduelle (gradient
décroissant de la zone saturé en LNAPL vers le fond du cubitainer, visible sur la
photographie de la Figure 5, indiquant une migration graduelle du LNAPL dans la
zone saturée avec 10 g/kg de C5-C40).

Figure 5. Photographie de la carotte réalisée au centre du cubitainer montrant la

répartition verticale du LNAPL.

Ce carottage confirme les résultats obtenus par tomographie de résistivité électrique

verticale. Cette méthode semble donc adaptée à la caractérisation d’une lentille
d’hydrocarbures en phase libre sur la nappe. Les variations de résistivité, obtenues après
inversion des données collectées, ont permis d’isoler la zone saturée en hydrocarbures,
ainsi que la partie supérieure de la zone saturée en eau, polluée à une très forte teneur (>
10 g/kg d’hydrocarbures).

3. Modélisation de la réponse de résistivité électrique in situ

3.1 Matériel et méthode

Les mesures de tomographie de résistivité électrique sont réalisées par 3 TEV,
composées chacune de 12 électrodes, utilisées pour la mesure de la résistivité réparties
entre -4,25 et -7 m tous les 0,25 m (Figure 6.A). Les puits étudiés (P16, B4, P12 et P19 -
sur la Figure 7) sont mesurés par des électrodes immergées tous les 0,5 m, directement
dans les puits en question (Figure 6.B). Afin de compléter ces électrodes, placées à la
verticale et en profondeur, 36 électrodes de surface, utilisées classiquement lors

35
d’investigations de résistivité électrique, sont placées selon une maille carrée de 5 m
(Figure 6.C).

Figure 6. Photographie des différents types d’électrodes utilisées pour les mesures de
résistivité électrique 3D. A : TEV ; B : électrodes immergées ; C : électrodes classique.

Un total de 108 électrodes sont ainsi réparties sur la zone investiguée (225 m²) selon des
dispositifs dipôle-dipôle équatorial (BING et al., 2000). Les quadripôles seront formulés
de façon à toujours respecter le parallélisme entre les électrodes d’injection et les
électrodes de mesure de potentiel (BING et al., 2000 ; DECEUSTER et al., 2006 ;
OCHS & KLITZSCH, 2020).

3.2 Modélisation numérique du positionnement des TEV

Les TEV sont positionnées sur les axes, entre le puits central P16 et les puits de la zone
d’étude les plus proches du P16 (B4, P12 et P19), afin d’obtenir une plus forte densité
d’information suivant des axes 2D (BORDING et al. 2020). Cette plus forte densité
permettra une meilleure caractérisation 2D, afin de corréler les inversions aux données
de LNAPL observé dans les puits.
Le positionnement des TEV a été déterminé comme étant optimal à 2,5 m du puits
central de l’étude (Figure 7). Une trop petite distance entre les TEV et le puits central ne
permet pas de caractériser l’intégralité de la surface d’étude, tandis qu’une trop grande
distance ne permet pas de caractériser correctement le centre de la zone d’étude et plus
particulièrement la zone autour du puits principal. La détermination de la position
optimale des TEV pour la caractérisation de la répartition des hydrocarbures dans la
zone de battement de nappe est déterminée par comparaison des analyses de sensibilité
en fonction des différentes positions testées.

36
Figure 7. Analyse de sensibilité du dispositif de mesures avec la position réelle des TEV
(chaque point rouge représentant une électrode de résistivité électrique).

Les TEV ont été implantées comme illustrées sur le schéma suivant (Figure 8) avec une
distance entre les TEV (T4, T12 et T19) et le puits de pompage P16 de respectivement
2,9 m ; 2,42 m et 2,81 m.

Figure 8. Cartographie de l’implantation des TEV par rapport aux puits de la zone
étudiée et zoom, en vignette, sur la partie centrale de la zone pilote, affichant les
distances entre les TEV et le puits de pompage (P16).

37
3.3 Modélisation numérique de la réponse de résistivité électrique à différentes tailles de
lentilles d’hydrocarbures
La réalisation de modèles numériques permet également de s’assurer que le dispositif de
mesures est adapté à la caractérisation recherchée. Ainsi, une seconde modélisation
numérique permettra de s’assurer de la bonne répartition des quadripôles, afin de
caractériser la lentille d’hydrocarbures située à la surface de la nappe, mais également la
zone de battement de nappe, où le LNAPL s’est étalé, au gré des marnages.
Par sa forte épaisseur observée dans les puits, le LNAPL libre sera considéré comme
formant un continuum résistant, non suffisamment dégradé pour engendrer une
augmentation de la conductivité électrique de la zone d’étude. La conductivité
électrique des hydrocarbures écrémés a été mesurée à 0 mS/m en laboratoire confirmant
cette hypothèse. La résistivité du compartiment contenant les hydrocarbures sera fixée à
60 Ω.m.
Différentes simulations de tailles de lentilles d’hydrocarbures de forme carrée (seule
possibilité du logiciel ERTLab 64) ont été testées, afin de valider le dispositif de
mesures des quadripôles en dipôle-dipôle équatorial (BING et al. 2000 ; BORDING et
al. 2020). Afin de reproduire l’incertitude liée aux mesures de terrain et de tester le
dispositif dans des conditions réalistes, un bruit (V/I) constant de 5% est superposé à un
bruit (V/I) aléatoire maximum de 10%. L’écartement des électrodes situées en
profondeur (TEV + puits) permet de caractériser avec acuité les lentilles de 25 cm de
hauteur, quelles que soient leurs étendues (Figure 9).
Diverses inversions numériques ont permis de se rendre compte que le dispositif de
mesure est adapté à la caractérisation d’une lentille d’hydrocarbures de forte résistivité
situé au niveau de la nappe. L’espacement entre les électrodes immergées étant de 50
cm, la caractérisation de lentilles d’hydrocarbures d’une épaisseur inférieure à la moitié
de cet espacement ne donne pas résultat satisfaisant.

Figure 9. Modèles inversés des résistivités électriques numériques (Ω.m) de lentilles

d’hydrocarbures de forme carrée, mesurant de 10 m à 0,5 m de côté et d’épaisseur 0,25 m.

38
4. Suivi d’un pompage écrémage par résistivité électrique 3D in situ

4.1 Etat initial des valeurs de résistivité avant le pompage/écrémage

L’inversion 3D des mesures de résistivité électrique réalisées avant le début du
pompage/écrémage permet de corréler, sans écoulements hydrauliques forcé, la
présence de LNAPL dans les puits à une zone de plus forte résistivité (Figure 10.A).
Cette zone de plus forte résistivité est plus épaisse au niveau du puits B4 que dans le
puits P16, corroborant les épaisseurs mesurées dans les puits B4 et P16 au cours de la
mesure des quadripôles de résistivité (Figure 10.B).

Figure 10. Coupe B4-P16 de la résistivité inversée (A) comparée aux profondeurs des
interfaces air/LNAPL (A/O) et LNAPL/H20 (O/W) dans les puits (B).

La représentation des résistivités électriques calculées en fonction de la profondeur,

permet de faire ressortir le lien entre les fortes résistivités et les fortes concentrations en
LNAPL corrélées avec la présence de LNAPL en phase pure dans les puits P16 et B4
(Figure 11.B). Cette comparaison permet de fixer une valeur limite à 73 Ω.m
correspondant à la présence de LNAPL en phase pure aux mêmes profondeurs dans les
puits d’observation (Figure 11.A). La représentation en carte de ces fortes résistivités
électriques (Figure 11.C) permet d’estimer l’extension de la lentille d’hydrocarbures à la
surface de la nappe. Ainsi, une faible épaisseur de LNAPL est visible dans le puits P19
et aucune trace de LNAPL n’a été enregistrée dans le puits P12 lors des 3 dernières
années de suivi des épaisseurs de LNAPL. La visualisation en carte illustre cette
tendance même si le modèle ne s’étend pas jusqu’aux puits P19 et P12, avec une
extension des fortes valeurs de résistivités vers le nord-ouest de la zone modélisée.
L’extraction des données du modèles à proximité des puits (Figure 11.D) confirme le
lien entre la présence de LNAPL et les fortes résistivités identifiées sur la Figure 10
avec des fortes valeurs de résistivité électrique concentrées aux profondeurs de présence
39
de LNAPL dans les puits et ainsi en corrélation avec l’épaisseur de LNAPL dans les
puits.

Figure 11. Comparaison des mesures réalisées à l’état initial (16 juillet). Les valeurs
de résistivité supérieures à 73 Ω.m sont identifiées en jaune (A) et reportées sur une
carte (C). Les résistivités des mailles du modèle inversé, situées autour du puits P16,
sont extraites en gris tandis que celles autour du puits B4 sont extraites en noir (C et D)
et comparées aux profondeurs des interfaces mesurées dans les puits (B).

Figure 12. Vue 3D de la résistivité inversée à l'état initial de la zone pilote avant le
démarrage du pompage avec une extraction des résistivité supérieures à 73 Ω.m.

La comparaison des données de présence de LNAPL dans les puits et de la répartition

des valeurs de résistivité électrique supérieures à 73 Ω.m et la corrélation entre les

40
épaisseurs de LNAPL dans les puits et les fortes valeurs de résistivité à proximité des
puits permet de s’intéresser à la vue en trois dimensions des variations des valeurs de
résistivité. Ainsi, la visualisation tridimensionnelle des résistivités ayant été calculées
comme supérieures à 73 Ω.m délimite un volume semblant corréler la présence et la
quantité de LNAPL dans les puits (Figure 12).

4.2 Suivi des valeurs de résistivité pendant le pompage/écrémage

La première inversion des mesures de résistivité électrique réalisé après le démarrage du
pompage (22 juillet) est validée par l’observation de la diminution de la quantité de
LNAPL observée dans les puits. En effet, l’épaisseur de LNAPL mesuré dans le puits
de pompage/écrémage est nulle et une moyenne de 29 cm de LNAPL a été mesuré dans
le puits B4 au cours de la mesure de résistivité (Figure 13.B). Les valeurs de résistivité
sont plus faibles et la représentation en carte des données observées confirment la
présence de LNAPL autour du puits B4 mais pas dans le puits P16 (Figure 13.C).

Figure 13. Comparaison des mesures réalisées après 5 jours de pompage/écrémage (22
juillet). Les valeurs de résistivité supérieures à 73 Ω.m sont identifiées en jaune (A) et
reportées sur une carte (C). Les résistivités des mailles du modèle inversé, situées
autour du puits P16 (extraites en gris) et celles autour du puits B4 (extraites en noir) (C
et D) sont comparées aux profondeurs des interfaces mesurées dans les puits (B).

Le pompage s’est poursuivi jusqu’au 22 aout avec une variation des épaisseurs de
LNAPL au cours du temps. L’écrémage a permis de récupérer un total de 410 L
d’hydrocarbures. Les mesures de résistivité ont permis d’imager les variations

41
d’épaisseur de LNAPL observées dans les puits et de suivre les déplacements du
LNAPL engendré par le pompage de rabattement et d’expliquer la réalimentation
constante en phase pure du puits écrémé.

4.3 Suivi des valeurs de résistivité après le pompage/écrémage

A l’arrêt du pompage (22 août), une grande épaisseur de LNAPL a été mesurée dans les
puits (Figure 14.B). Les calculs de résistivité électrique indiquent des valeurs de
résistivité élevées dans un rayon de 5 m autour du puits de pompage (Figure 14.C).
Localement, autour des puits B4 et P16, les valeurs de résistivité sont corrélées à la
présence d’hydrocarbures à la surface de la nappe.

Figure 14. Comparaison des mesures réalisées 24h après arrêt du pompage/écrémage.
Les valeurs de résistivité supérieures à 73 Ω.m sont identifiées en jaune (A) et reportées
sur une carte (C). Les résistivités des mailles du modèle inversé, situées autour du puits
P16 (extraites en gris) et celles autour du puits B4 (extraites en noir) (C et D) sont
comparées aux profondeurs des interfaces mesurées dans les puits (B.)

Cette grande quantité de LNAPL dans les puits et dans le sol perdure dans le temps et
est visible plus d’une semaine après l’arrêt du pompage (30 août). Les valeurs de
résistivité électrique mesurées restent très élevées et la visualisation en carte indique que
la répartition de ces fortes valeurs est orientée selon un axe nord-est sud-ouest.

42
5. Détermination de l’extension d’une zone polluée aux hydrocarbures dégradés
La réalisation de mesures au géoradar sur la zone investiguée, par la méthode de
résistivité électrique, n’a pas permis d’obtenir des résultats exploitables, puisque l’onde
électromagnétique émise est très vite dégradée, par la trop forte conductivité électrique
du milieu découlant de l’étalement des hydrocarbures, disséminés et dégradés sur toute
la hauteur de la zone de battement de nappe.
Une zone située plus au sud dans la raffinerie, moins polluée dans sa partie non saturée,
a permis de tester le radar dans de meilleures conditions. Une mesure de la piézométrie
de la zone a révélé une profondeur de nappe autour de 6 m. Les radargrammes réalisés
sur la zone ont permis de mettre en évidence deux comportements différents de la zone
non saturée, en réponse aux ondes électromagnétiques du géoradar. La forte réflexion
censée indiquer le passage à la zone saturée n’est visible que sur la partie est du profil
présenté sur la Figure 15. Sur l’autre portion du profil, le signal est très largement
dégradé à partir de 3,5 m. Cette analyse permet de faire ressortir un fractionnement en
deux zones dont l’une est contaminée (à l’ouest). La réponse aux ondes radar de l’autre
zone est différente. La réalisation de plusieurs profils au niveau de cette nouvelle zone
d’étude, a permis de fixer les limites de la zone polluée (Figure 16) (REDMAN et al.,
1994)

Figure 15. Radargramme avec une distinction de la forte réflexion correspondant à la

nappe (portion est) et à la dégradation du signal radar (portion ouest).

Figure 16. Cartographie des radargrammes en vue 2D et détermination des extensions

de la pollution en hydrocarbures.

43
 La réalisation d’un sondage a permis de prélever des échantillons pour analyse de la
contamination des sols en hydrocarbures tous les mètres entre -2 et -8 m. Ces
échantillons ont révélé des indices de contamination (somme C5-C40) inférieurs à
1500 mg/kg. L’un des deux ouvrages piézométriques situés dans la zone déterminée
comme polluée ne contenait pas de LNAPL en phase libre (P1 métal), tandis que l’autre
ouvrage (PTS-Métal) contenait plus de 20 cm de hauteur de LNAPL au moment de la
réalisation des mesures.

6. Conclusion – discussion
La répétition des mesures dans le temps accorde de la robustesse aux données acquises.
Le suivi temporel de l’évolution d’un phénomène peut ainsi être caractérisé et isolé.
Dans le cas de l’étude présentée, plusieurs variables sont superposées avec le marnage
de la nappe et le pompage-écrémage du produit en phase pure (l’impact de la
pluviométrie étant considérée comme négligeable). Le marnage n’entraine pas de
diminution de la quantité de LNAPL dans le sol, mais entraine un déplacement vertical
de la zone de plus forte densité de LNAPL dans la porosité.
Cette zone, de signature résistive, suit les variations de la nappe et se caractérise par une
variation verticale de la profondeur des plus fortes résistivités après inversion. Comme
observé dans les puits, la différence de vitesse de déplacement de la nappe entre la
marée montante (plus rapide) et la marée descendante (plus lente) entraine une
accumulation plus importante de LNAPL à la surface de la nappe, à marée basse et lors
des faibles coefficients de marnage. En marée montante, les hydrocarbures sont
disséminés dans la porosité, car leur conductivité hydraulique (lié à la viscosité du
fluide), plus faible que celle de l’eau, ne leur permet pas de suivre les mouvements de
nappe.
La quantité d’hydrocarbures retirés en un mois (410 L) n’est pas significative par
rapport à la quantité d’hydrocarbures présents dans le sol. La quantité d’hydrocarbures
dans les puits avant la réalisation du pompage (0,49 m dans P16) est très similaire à
celle mesurée après l’arrêt du pompage (0,42 m dans P16). Un suivi sur plus long terme
de l’avancement de la dépollution permettra d’identifier la diminution des fortes
résistivités électriques mesurées à la surface de la nappe.
Dans le cas d’un compartiment de sol saturé en hydrocarbures en phase libre, la
résistivité électrique mesurée est plus résistive que l’encaissant ne présentant pas
d’hydrocarbures. La profondeur des lentilles d’hydrocarbures, combinée à leurs faibles
épaisseurs, rend difficile la caractérisation de ces zones résistantes depuis la surface.
C’est pourquoi, il est nécessaire de réaliser des mesures depuis la profondeur à
investiguer.
Dans l’ensemble, la réalisation des mesures de résistivité électrique en trois dimensions
a permis de corréler la présence d’hydrocarbures aux fortes valeurs de résistivité
électriques aussi bien verticalement que spatialement. Ainsi les variations d’épaisseurs
44
de LNAPL dans le puits B4 s’est accompagné d’une variation de l’intensité des valeurs
de résistivité électrique élevées. A l’arrêt du pompage/écrémage, une accumulation
progressive de LNAPL a été observée dans les puits et les calculs ont mis en avant une
augmentation des résistivité mesurées.
La réalisation des profils verticaux, en extrayant les résistivités des mailles autour des
puits, permet de confronter directement les données des modèles inversés aux données
de profondeur des interfaces mesurées dans les puits. Cependant, cette extraction des
données, selon leur position dans le modèle, souligne une différence de profondeur
d’environ 1 m entre les valeurs de résistivité élevées et la présence de LNAPL dans le
puits B4.
Les hydrocarbures disséminés ont tendance à se dégrader plus facilement, entrainant
une accumulation de résidus liés à leur biodégradation (RAHMAN et al., 2003). Cette
accumulation entraine une augmentation de la conductivité électrique du milieu qui ne
permet pas au signal du géoradar de pénétrer en profondeur (SAUCK, 2000 ; CASSIDY
et al., 2001). Cependant, cette propriété peut être utilisée pour déterminer des limites de
zones contaminée avec un hydrocarbures âgé (ayant déjà commencé le processus de
biodégradation) (Figure 17). Les contrastes de profondeurs d’investigations permettent
ainsi de délimiter, par une réponse qualitative, des zones de pollution en hydrocarbures
dégradés de zones moins polluées, ou dont la pollution en hydrocarbures est récente et
dont le signal serait plus résistant.

Figure 17. Modèle conceptuel de contamination en hydrocarbures, illustrant le

comportement temporel de la conductivité électrique en fonction de la biodégradation,
de la réduction de la masse contaminante, de l’atténuation naturelle, ou de
l’assainissement réalisé (CHE-ALOTA et al., 2009).

7. A retenir/Key Point
La présence d’hydrocarbures dans le sol et dans l’eau a un impact sur les propriétés du
milieu mesuré par les méthodes géophysiques. Ainsi, la présence de LNAPL en phase

45
pure, à la surface de la nappe, entraine une augmentation de la résistivité électrique
mesurée par tomographie (de résistivité électrique). Le positionnement des électrodes
autour de la profondeur d’investigation permet de caractériser cette augmentation de
résistivité sur une faible épaisseur. Ce dispositif, dont la géométrie est spécifique à
chaque site investigué (LABRECQUE et al. 1996), notamment du fait de la profondeur
visée, de l’hétérogénéité du sous-sol et des dimensions de la zone à caractériser, permet
un gain proportionnel à la profondeur, des sensibilités des mesures à forte profondeur.

8. Références
ATEKWANA E.A., WERKEMA D.D., DURIS J.W., ROSSBACH S., ATEKWANA
E.A., SAUCK W.A., CASSIDY J.P., MEANS J., LEGALL F.D. (2004). In-situ
apparent conductivity measurements and microbial population distribution at a
hydrocarbon-contaminated site, Geophysics, volume 69, pp. 56-63,
doi.org/10.1190/1.1649375
BABCOCK E., BRADFORD J.H. (2015). Reflection waveform inversion of ground-
penetrating radar data for characterizing thin and ultrathin layers of nonaqueous phase
liquid contaminants in stratified media. Geophysics, volume 80(2): H1–H11.
doi.org/10.1190/GEO2014-0037.1
BERTOLLA L. PORSANI J.L., SOLDOVIERI F., CATAPANO I. (2014). GPR-4D
monitoring a controlled LNAPL spill in a masonry tank at USP, Brazil Journal of
applied geophysics, volume 103, pp. 237-244, doi.org/10.1016/j.jappgeo.2014.02.006
BING Z., GREENHALGH S.A. (2000). Cross-hole resistivity tomography using
different electrode configurations. Geophysical Prospecting, volume 48 (5), pp. 887-
912, doi.org/10.1046/j.1365-2478.2000.00220.x
BORDING T., KUHL A.K., FIANDACA G., CHRISTENSEN J. F., CHRISTIANSEN
A.V., AUKEN E. (2020). Cross-borehole geoelectrical time-lapse monitoring of in situ
chemical oxidation and permeability estimation through induced polarization. Near
Surface Geophysics. Volume 19 (1), pp. 43-58, doi.org/10.1002/nsg.12131
BRADFORD J. H. (2003). GPR offset-dependent reflectivity analysis for
characterization of a high-conductivity LNAPL plume. Symposium on the Application
of Geophysics to Environmental and Engineering Problems (SAGEEP), Environmental
and Engineering Geophysical Society, pp. 238-252. doi.org/10.3997/2214-4609-pdb.190.con08
BROGGINI F., VASMEL M., ROBERTSSON J., VAN MANEN D. J. (2016).
Immersive boundary conditions: Theory, implementation, and examples. Geophysics.
Volume 82 (3), doi.org/10.1190/geo2016-0458.1
CASSIDY D. P., WERKEMA D. D., SAUCK W. A., ATEKWANA E. A.,
ROSSBACH S., DURIS J. (2001). The effects of LNAPL biodegradation products on
electrical conductivity measurements. Journal of Environmental and Engineering
Geophysics, volume 6, pp. 47–52, doi.org/ 10.4133/JEEG6.1.47

46
CHAPKO R., KRESS R., YOON J.R. (1999). An inverse boundary value problem for
the heat equation: the Neumann condition. Inverse Problems, volume 15 (4): 1033.
CHE‐ALOTA V., ATEKWANA E. A., ATEKWANA E. A., SAUCK W. A.,
WERKEMA D. D. Jr. (2009). Temporal geophysical signatures from contaminant‐
mass remediation, Geophysics, 74(4), B113–B123, doi.org/10.1190/1.3139769
DAILY W., OWEN E. (1991). Cross-borehole resistivity tomography. Geophysics,
volume 56, pp. 1228–1235. doi.org/10.1190/1.1443142
DECEUSTER J., DELGRANCHE J., KAUFMANN O. (2006). 2D cross-borehole
resistivity tomographies below foundations as a tool to design proper remedial actions
in covered karst. Journal of applied geophysics, volume 60, pp. 68-86.
doi.org/10.1016/j.jappgeo.2005.12.005
DENG Y., SHI X., REVIL A., WU J., GHORBANI A. (2018). Complex conductivity of
oil-contaminated clayey soils. Journal of hydrology., volume 561, pp. 930-942,
doi.org/10.1016/j.jhydrol.2018.04.055
IRAVANI M. A., DEPARIS J., DAVARZANI H., COLOMBANO S., GUERIN R.,
MAINEULT A. (2020). The influence of temperature on the dielectric permittivity and
complex electrical resistivity of porous media saturated with DNAPLs: A laboratory
study, Journal of Applied Geophysics, volume 172: 103921, ISSN 0926-9851,
doi.org/10.1016/j.jappgeo.2019.103921
LABRECQUE D., RAMIREZ A., DAILY W., BINLEY A.M., SCHIMA S. (1996).
ERT monitoring of environmental remediation processes. Measurement Science and
Technology, volume 7, pp. 375-383, doi.org/10.1088/0957-0233/7/3/019
LI K. (2011). Multiple solutions for an asymptotically linear Duffing equation with
Newmann boundary value condition. Nonlinear Analysis: Theory, Methods and
Applications, volume 74 (8), pp. 2819-2830, doi.org/10.1016/j.na.2011.01.003
MARINELLI F., DURNFORD D.S. (1996). LNAPL thickness in monitoring wells
considering hysteresis and entrapment. Ground Water, volume 34, pp. 405–414,
doi.org/10.1111/j.1745-6584.1996.tb02021.x
MCCAULOU D. R., JEWETTE D.G., HULING S. G. (1995). Nonaqueous phase
liquids compatibility with materials used in well construction, sampling and
remediation. Ground Water Issue, EPA/540/S-95/503
OCHS J., KLITZSCH N. (2020). Considerations regarding small-scale surface and
borehole-to-surface electrical resistivity tomography. Journal of Applied Geophysics,
volume 172: 103862, doi.org/10.1016/j.jappgeo.2019.103862
OROZCO A.F., KEMNA A., OBERDÖRSTER C., ZSCHORNACK L., LEVEN C.,
DIETRICH P., WEISS H. (2012). Delineation of subsurface hydrocarbon
contamination at a former hydrogenation plant using spectral induced polarization
imaging. Journal of Contaminant Hydrology, volume 136, pp. 131–144.
doi.org/10.1016/j.jconhyd.2012.06.001

47
RAHMAN K. S. M., RAHMAN T. J., KOURKOUTAS Y., PETSAS I., MARCHANT
R., BANAT I. M. (2003). Enhanced bioremediation of n-alkane in petroleum sludge
using bacterial consortium amended with rhamnolipid and micronutrients, Bioresource
Technology, volume 90 (2), pp. 159–168, doi.org/10.1016/S0960-8524(03)00114-7
REDMAN J.D., DE RYCK S.M., ANNAN A.P. (1994). Detection of LNAPL pools with
GPR. Theoretical modelling and surveys of a controlled spill. Proceedings of the Fifth
International Conference on Ground Penetrating Radar GPR, pp. 1283–1294,
doi.org/10.1109/ICGPR.2012.6254964
SAUCK W.A., ATEKWANA E.A., NASH M.S. (1998). High conductivities associated
with an LNAPL plume imaged by integrated geophysical techniques. Journal of
Environmental and Engineering Geophysics, volume 2, pp. 203-212.
doi.org/10.1080/02533839.2007.9671230
SAUCK W. A. (2000). A model for the resistivity structure of LNAPL plumes and their
environs in sandy sediments, Journal of Applied Geophysics, Volume 44 (2–3), pp. 151-
165, ISSN 0926-9851, doi.org/10.1016/S0926-9851(99)00021-X
SHIMA H. (1992). 2D and 3D resistivity imaging reconstruction using cross-hole data.
Geophysics, volume 55, pp. 682–694, doi.org/10.1190/1.1443195
TANAHASHI H., SATO T., KONISHI J. (2006). Study on water–NAPL–air three-
phase capillary pressure-saturation relationships of sand based on the pressure-plate
and the gravity drainage column experiments. Journal of Japan Society of Civil
Engineers, volume C62(2), pp. 320–334.

48
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Possibility of well clogging by recharging tunnel seepage in

the Kanto region, Japan

Sho HASEGAWA 1, Toshifumi IGARASHI 2

1. Cooperative Program for Resources Engineering, Graduate School of Engineering,

Hokkaido University, Sapporo, Japan.
[email protected]
2. Faculty of Engineering, Hokkaido University, Sapporo, Japan.
[email protected]

Abstract:
Several tunnels have been excavated for expressway construction in the Kanto region,
Japan. Groundwater levels may commonly be lowered by the tunnel excavation. Thus,
recharge of tunnel seepage into groundwater is occasionally required. However, there
may be a possibility of clogging recharge boreholes. The objective of this research was
to evaluate the possibility of clogging of recharge boreholes by injecting tunnel seepage.
There are three major factors of clogging; physical, biological, and chemical ones.
These factors were evaluated by collecting and analyzing the tunnel seepage,
groundwater and river water at the site. The results showed that physical and biological
factors did not significantly affect the clogging because the samples had the low total
organic carbon concentration and turbidity. The results of dissolved chemical
compositions showed that calcite might be formed when the tunnel seepage was
recharged into groundwater. The mixing of the tunnel seepage with groundwater was
effective in reducing the saturation index of calcite although aeration of the tunnel
seepage was not effective. This indicates that the tunnel seepage should be injected into
the recharge borehole without aeration at a specific rate to prevent the formation of
calcite in recharge boreholes.

Keywords:
Groundwater, Tunnel seepage, Clogging, Calcite, Saturation index, Precipitation

Corresponding author mail: [email protected]

1. Introduction
Several tunnels have been excavated for expressways in the Kanto region, Japan.
Groundwater is released out of the tunnels due to excavation, and groundwater levels
around the tunnels may commonly be lowered (TAKAHASHI, 1965; MIKI &
YOSHIZAWA, 1979). However, groundwater is used as tasty drinking water by
residents in the surrounding area. Thus, recharge of tunnel seepage into groundwater is
occasionally required when the tunnel is excavated in the same basin. However, there
are a variety of factors affecting clogging during recharge of tunnel seepage. In this
research, possibility of clogging of boreholes during recharge of tunnel seepage was
evaluated by analyzing chemistry of groundwater, river water and tunnel seepage.

2. Materials and methods

2.1 Study site and water sampling

The study site is located in the Kanto region, Japan. The geological formation of the site
consists of a series of highly permeable sand, gravel, and loamy layers. Several tunnels
for expressway were constructed upstream of the alluvial fan at the site.

Figure 1. Location of sampling points.

Three types of water samples, groundwater (GW1, 2 and 3), tunnel seepage (TS1, 2 and
Treated water), and river water (RW), were collected 14 times in 2019 at the site.
Sampling points are shown in Figure 1. The pH of the tunnel seepage exceeded the
effluent standard (pH 8.6) during and immediately after tunnel excavation. Thus,
liquified carbon dioxide was injected to the tunnel seepage to neutralize pH (Treated
water). However, after October, 2019 the pH was reduced less than the effluent
standard, and then the neutralization was stopped.

50
2.2 Analysis of water samples and geochemical calculation
The pH, electrical conductivity (EC), oxidation-reduction potential (ORP), turbidity and
temperature of collected water samples were measured in situ. Major ion concentrations
were analyzed by ICP-AES and anion chromatograph. In addition, total organic carbon
was measured by the TOC meter. Bicarbonate ion concentration was measured by
titration with sulfuric acid (H2SO4). Collected water samples were filtered by 0.45 µm
before chemical analysis.
Geochemical modelling software called PHREEQC was applied to evaluate the
saturation index (SI) of calcite (PARKHURST & APPELO, 1999; TOKORO, 2016).

3. Results

3.1 Chemistry of water samples

The analyzed results of TOC and turbidity are shown in Figures 2 and 3, respectively.
These figures showed that TOC and turbidity of tunnel seepage and groundwater were
almost the same as those of river water, except for several data of turbidity. This means
that physical and biological factors do not significantly affect the clogging.
16
GW1
14 RW
GW2
12 GW3
TOC (mg/L)

TS1
10 TS2
Treated water
8

0
4/11 5/13 5/28 7/29 8/26 11/29
Date (M/D)
Figure 2. Change in total organic carbon concentration in 2019.
500
GW1
RW
400 GW2
GW3
Turbidity (ntu)

300 TS1
TS2
Treated water
200

100

0
4/11 5/13 5/28 6/11 6/24 7/9 7/29 8/6 8/26 9/10 10/28 11/15 11/29
Date (M/D)

Figure 3. Change in turbidity in 2019.

51
The piper diagram of water samples is shown in Figure 4. From this figure, Ca2+ and
HCO3- were dominant irrespective of the location of sampling.

GW1

1.0
1.0
RW

0.8
0.8

Ca
O 2-
4
GW2

2+
0.6
0.6
+S

+M0.4
GW3

Cl -
0.4

g
2+
TS1

0.2
0.2
TS2

0.0
0.0

Treated water

0.0
0.0

1.0
1.0

0.2
0.2

0.8
0.8

Na 0.6

O-
+

SO 4
Mg 2

+ +K

0.4
0.4

0.6
HC
0.6

2-
0.6
+

0.4
0.4

0.8
0.8

0.2
0.2

1.0
1.0

0.0
0.0

1.0 0.8 0.6 0.4 0.2 0.0 0.0 0.2 0.4 0.6 0.8 1.0

Ca²⁺ Cl -

Figure 4. Piper diagram of water samples.

Changes of Ca2+ and HCO3- concentrations are shown in Figures 5 and 6, respectively.
Both ion concentrations of tunnel seepage and groundwater were higher than those of
river water. This means that when the tunnel seepage is recharged into groundwater,
there is a possibility that calcite may be formed.

70
GW1
60
RW
GW2
50 GW3
Ca2+(mg/L)

TS1
40 TS2
Treated water
30

20

10

0
2/22 4/11 5/13 5/28 6/11 6/24 7/9 7/29 8/6 8/26 9/10 10/28 11/15 11/29
Date (M/D)

Figure 5. Change of Ca2+ concentration in 2019.

52
 GW1
220
RW
200 GW2
GW3
180
HCO₃⁻ (mg/L) TS1
160 TS2
treated water
140

120

100

80

60

40
2/22 4/11 5/13 5/28 6/11 6/24 7/9 7/29 8/6 8/26 9/10 10/28 11/15 11/29
Date (M/D)
Figure 6. Change of HCO3- concentration in 2019.

3.2 PHREEQC calculation

The calculated results of SI of calcite by PHREEQCE is shown in Figure 7. The SI
values of the tunnel seepage became positive, indicating that calcite may be formed
when the quality of the tunnel seepage remains unchanged. Therefore, it is necessary to
take countermeasures against clogging in recharge boreholes when tunnel seepage is
recharged into groundwater.

2.0 GW1
RW
GW2
1.5
GW3
TS1
Saturation Index

1.0 TS2
Treated water
0.5

0.0

-0.5

-1.0

-1.5

-2.0
2/22 4/11 5/13 5/28 6/11 6/24 7/9 7/29 8/6 8/26 9/10 10/28 11/15 11/29
Date (M/D)

Figure 7. SI of calcite for tunnel seepage, groundwater, and river water.

4. Discussion
Two countermeasures against clogging were examined. One was aeration of tunnel
seepage and the other was mixing of the tunnel seepage with groundwater (GW1).

4.1 Aeration
Aeration was applied to decrease the partial pressure of CO2 of tunnel seepage in the
calculation using PHREEQC because partial pressure of CO2 of the tunnel seepage was

53
higher than the atmospheric one (AOKI et al., 2009; AOKI et al., 2010). The
relationship between Ca2+ activity and pH after aeration is shown Figure 8. When the
plot below the straight line, calcite was not saturated. On the other hand, when the plot
position is above the straight line, calcite is supersaturated. In this figure, the SI of
calcite of tunnel seepage became lower or higher by aeration depending on the
chemistry. This means that the SI depends on not only partial pressure of CO2 but also
pH.

-2.5
GW1
RW
GW2
-3.0 GW3
TS1
TS2
log(Ca²⁺)

Treated water
-3.5

-4.0

-4.5
6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0 10.5
pH

Figure 8. The relationship between Ca2+ activity and pH after aeration.

4.2 Mixing with groundwater

Mixing of the tunnel seepage with the surrounding groundwater was examined in the
calculation using PHREEQC. The calculated SI values of calcite for the mixture of
tunnel seepage and groundwater are shown in Figure 9.

1.5
2/22
4/11
1.0 5/13
5/28
6/11
Saturation Index

0.5 6/24
7/9
7/29
0.0 9/10
10/28
11/29
-0.5

-1.0

1:9 2:8 3:7 4:6 5:5 6:4 7:3 8:2 9:1

Mixing ratio

Figure 9. SI of calcite for mixture of tunnel seepage and groundwater.

54
The SI decreased with the mixing ratio of groundwater because neutral pH of
groundwater enhanced the solubility of calcite. This indicates that minimum mixing
ratio of groundwater can be estimated using this figure. This knowledge can be applied
to design the recharge system of tunnel seepage into groundwater. (Mixing ratio means
the ratio of groundwater to tunnel seepage.)
Recharging tunnel seepage into groundwater is of importance when enough water
resources are not obtained. Thus, the mixing method of tunnel seepage and groundwater
is effective in managing sustainable water resources.

5. Conclusion
The followings were obtained in this research.
(1) Clogging induced by physical and biological factors was insignificant because TOC
and turbidity of tunnel seepage were almost same as those of groundwater.
(2) Ca2+ and HCO3- were dominant in the tunnel seepage and groundwater. Therefore,
calcite may be formed in the recharge borehole when the tunnel seepage is
recharged into groundwater.
(3) Aeration reduced the partial pressure of CO2 of the tunnel seepage. However,
aeration was not a good method for preventing the formation of calcite because the
pH of the tunnel seepage was changed and approached 8.0 to 8.5.
(4) The mixing of the tunnel seepage with the groundwater reduced the pH of mixed
water and changed the SI of calcite to negative values.

6. References
TAKAHASHI H. (1965). Characters and some problems on the tunnel water, Journal
of the Japan Society of Engineering Geology, Vol. 6, No. 1, pp. 25-52.
MIKI K., YOSHIZAWA H. (1979). On the fluctuation of water contents due to civil-
geological environment of tunnel seepage, Journal of the Japan Society of Civil
Engineers, No. 282, pp. 31-43.
PARKHURST D. L., APPELO C. A. J. (1999). User’s guide to PHREEQC (version 2)
– A computer program for speciation, batch-reaction, one-dimensional transport, and
inverse geochemical calculations, U.S. Geological Survey Water-Resources
Investigation Report, No. 99-4259, http://pubs.er.usgs.gov/publication/wri994259
TOKORO C. (2016). Introduction to Reaction Analysis by PHREEQC for Beginners, R
& D Support Center.
AOKI T., IGARASHI T., IIO Y. (2009). Properties of alkaline groundwater seepage
from a tunnel and its neutralization process by self-purification mechanisms, Journal of
the Japan Society of Engineering Geology, Vol. 50, No. 5, pp. 273-279.

55
AOKI T., IGARASHI T., IIO Y., NISHIO H. (2010). Reduction in pH of alkaline
groundwater seepage from tunnel by dissolving atmospheric carbon dioxide, Journal of
the Japan Society of Engineering Geology, Vol. 51, No. 5, pp. 220-228.

56
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Change in groundwater flow in a limestone quarry

by drainage tunnel excavation

Keisuke INOUE 1, Kouta FUJIMAKI 2, Chika UMEDA 2, Toshifumi IGARASHI 3

1. Division of Sustainable Resources Engineering, Graduate School of Engineering,

Hokkaido University, Sapporo, Japan.
[email protected]
2. Ryokolime Industry Co., Ltd., Tokyo, Japan.
3. Faculty of Engineering, Hokkaido University, Sapporo, Japan.
[email protected]

Abstract:
Limestone has been excavated by the bench cut method in many quarries. In this quarry,
rock displacement has been observed since September 2007, and it is reported that the
displacement is related to the groundwater level. Therefore, a drainage tunnel was
excavated a couples of years ago for reducing the groundwater level so that the rock
displacement could be stabilized. The purpose of this study was to evaluate the change in
groundwater flow in the quarry by drainage tunnel excavation.
Change in groundwater quality before and after tunnel excavation was examined by
collecting and analyzing groundwater samples. The results showed that the
concentrations of calcium and bicarbonate ions in groundwater near the interbedded layer
were reduced by the excavation. This indicates that the residence time of the groundwater
was reduced due to the excavation. This means that the excavated tunnel is promising for
confirming the stability of the rock slope by draining the groundwater near the
interbedded layer.

Keywords:
Groundwater flow, Water quality, Quarry, Drainage tunnel.

Corresponding author mail: [email protected]

1. Introduction
It is common that limestone is excavated by the bench cut method in Japan. It is important
to preserve the stability of rock slope in mine operation. However, collapses of rock slope
have often occurred in Japan. A large-scale displacement of rock slope has been observed
since 2007 in a quarry. Thus, the displacement of the slope and groundwater chemistry in
and around the slope have been monitored (OZAWA, 2018; KONDO et al., (2020);
EANG et al., (2018a); EANG et al., 2018b). It is found that the rock displacement is
related to the groundwater level. Thus, a drainage tunnel at an elevation of 970 m above
the sea was excavated from 2017 to 2018 to stabilize the rock slope. After that, the rock
slope became almost stable. The purpose of this study was to evaluate the effect of the
tunnel excavation on the groundwater quality in the quarry by comparing it before and
after excavation.

2. Materials and methods

2.1 Outline of quarry and sampling of groundwater

Figure 1 shows a cross section of the quarry. The base rock is green rock while the surface
is limestone. There is an interbedded layer between the rocks. There may be groundwater
flow pathways around the interbedded layer. In this figure, groundwater sampling points
are also shown. Figure 2 (left and centered) presents the representative sampling points
in the drainage tunnel. The tunnel is located at an elevation of 970 m, which drains the
groundwater in the interbedded layer.

Figure 1. Cross section of the quarry and sampling points.

When taking groundwater samples, pH, electrical conductivity (EC), oxidation-reduction

2
58
potential (ORP), and temperature were measured in situ as shown in Figure 2 (right).
Groundwater flow rates were also measured by the volumetric method. The collected
samples were filtered using a 0.45 μm membrane filter in situ, and then the filtrates were
stored in polyethylene bottles for analysis. The field campaign was conducted from May
2016 to December 2019.

Figure 2. Sampling points in the drainage tunnel and in situ measurements.

2.2 Analysis of groundwater samples

Major elements (Ca2+, Mg2+, Na+, K+, Cl-, HCO3-, NO3-, SO42- and Si) were analyzed in
the laboratory. Concentrations of Ca2+, Mg2+, Na+, K+ and Si were measured by ICP-AES
whereas concentrations of Cl-, NO3- and SO42- were measured by ion chromatograph (IC).
Concentration of HCO3- was measured with titration by 0.01 M sulfuric acid (H2SO4).

2.3 Geochemical analysis

Saturation index of calcite (SIcalcite) and the partial pressure of carbon dioxide (PCO2(g))
were calculated by using the geochemical modeling software PHREEQC (PARKHURST
& APPELO, 2013). SIcalcite and PCO2(g) are expressed in equations (1) and (2) respectively,
and the hydrolysis of inorganic carbon dioxide is expressed in equations (3) to (6).
[Ca2+ ][CO2−
3 ]
𝑆𝐼calcite = log 𝐾c
(1)

[H+ ][HCO−
3]
𝑃CO2(g) = 𝐾0 𝐾1
(2)

𝐾c = [Ca2+ ][CO2−
3 ] = 10
−8.48
a (t = 25 °C) (3)
[CO2(aq) ]
𝐾0 = 𝑃CO2
(4)

[H+ ][HCO−
3]
𝐾1 = [CO2 (aq)]
(5)

3
59
 [H+ ][CO2−
3 ]
𝐾2 = [HCO−
(6)
3]

[-] is activity, KC is solubility of calcite, and K0, K1 and K2 are equilibrium constants of
CaCO3-H2O-CO2 system.

3. Results and discussion

All groundwater samples in the tunnels were of Ca-HCO3 type (Figure 3). Some samples
had higher concentration of SO42-. This is due to the effect of blasting for tunnel
excavation.

Figure 3. Piper diagram of groundwater collected in drainage tunnel in December

2019.

The groundwater in the limestone layer and the interbedded layer tended to decrease the
flow rate after excavation (Figures 4 and 5). It is because groundwater in the limestone
and interbedded layers was drained by excavation of the drainage tunnel.

60
4
 During excavation of drainage tunnel
L-N-1
L-N-2
L-N-3
L-N-4
100 L-N-5
L-N-6
L-N-7
Flow rate (L/min)

10 L-N-8
L-N-9
L-N-10
L-N-11
1 L-N-12

0.1

0.01

0.001
16

17

17

19

19
08/

09/

12/

05/

10/
Date (Month/Year)
Figure 4. Change in flow rate of groundwater samples in drainage tunnel.

L-1
L-2
L-3
L-4
100 L-7
L-8
L-9
Flow rate (L/min)

10 L-10
L-11
LP-1
LP-10
1 S-1
S-2

0.1

0.01

0.001
16

17

17

19

19
08/

09/

12/

05/

10/

Date (Month/Year)

Figure 5. Change in flow rate of groundwater samples in the other tunnels.

It can be seen that the alkalinity and Ca2+ concentration of most of groundwater samples
in the limestone and interbedded layers decreased after the excavation (Figures 6 to 9).
After excavation, only in October 2019, the Ca2+ concentration increased due to heavy
rain.

61
5
Figure 6. Change in alkalinity of groundwater samples in drainage tunnel.

Figure 7. Change in alkalinity of groundwater samples in the other tunnels.

62
6
Figure 8. Change in concentration of Ca2+ of groundwater samples in drainage tunnel.

Figure 9. Change in concentration of Ca2+ of groundwater samples in the other tunnels.

Focusing on SIcalcite and logPco2 (Figures 10 to 13), these parameters did not change
dramatically before and after excavation. It is expected that the decrease in Ca2+
concentration caused the change in pH and led to new equilibrium of calcite dissolution.
In October 2019, samples were collected immediately after a typhoon passed, so the
heavy rain induced a large amount of infiltration into the quarry, causing a drop in SIcalcite.

763
Figure 10. Change in SIcalcite of groundwater samples in drainage tunnel.

Figure 11. Change in SIcalcite of groundwater samples in the other tunnels.

64
8
 Figure 12. Change in logPco2 of groundwater samples in drainage tunnel.

Figure 13. Change in logPco2 of groundwater samples in the other tunnels.

Both Ca2+ concentration and flow rate decreased after excavation (Figures 14 and 15).
This is because the average retention time of groundwater decreased. This means that the
excavation of the drainage tunnel was effective in reducing the retention time of
groundwater. The reduction in the residence time of groundwater helps prevent the
displacement of the rock slope in response to rainfall.

65
9
To confirm the stability of rock slope in quarries is required. The excavation of drainage
tunnels was found to be one of the promising countermeasures. However, it is important
to evaluate the groundwater flow in quarries by climate changes in the future.

Figure 14. Relationship between Ca2+ concentration and flow rate of groundwater
samples in the other tunnels (before excavation).

Figure 15. Relationship between Ca2+ concentration and flow rate of groundwater
samples in the other tunnels (after excavation).

66
10
4. Summary and future work
Based on the results of groundwater quality before and after the excavation of a drainage
tunnel, the followings were revealed.
(1) Groundwater in the limestone and interbedded layers was effectively drained by the
excavation of a drainage tunnel by considering the changes in flow rate and groundwater
chemistry.
(2) However, it is necessary to continue to monitor groundwater quality to quantify the
relationship between the deformation of rock slope and groundwater flow chemistry.
In addition, the residence time of groundwater by tritium analysis should be examined to
identify the flow pathway of groundwater.

5. References
OZAWA N. (2018). Analyzing the impacts of rainfall on rock slope behavior in Une mine
and evaluating the effectiveness of countermeasure methods, Journal of Mining and
Materials Processing Institute of Japan (MMIJ), Vol. 134, No.11, pp. 208-221,
KONDO M., OZAWA N., FUJIMAKI K, IGARASHI T. (2020). Water quality survey and
its application to rock slope management at the Une mine, Journal of Mining and
Materials Processing Institute of Japan, Vol. 134, No. 10, pp. 117-126, doi:
10.2473/journalofmmij.136.117
EANG K. E., FUJINAGA R., IGARASHI T., KONDO M., TABELIN C. B. (2018a).
Groundwater monitoring of an open-pit limestone quarry: Groundwater characteristics,
evolution and their connections to rock slopes, Environmental Monitoring and
Assessment, Vol. 190, 193, doi: 10.1007/s10661-018-6561-2
EANG K. E., IGARASHI T., KONDO M., NAKATANI T., TABELIN C. B., FUJINAGA
R. (2018b). Groundwater monitoring of open-pit limestone quarry: Water-rock
interaction and mixing estimation within the rock layers by geochemical and statistical
analysis, International Journal of Mining Science and Technology Vol. 28, pp. 849-857,
doi.org/10.1016/j.ijmst.2018.04.002
PARKHURST D. L., APPELO C. A. J. (2013). Description of input and examples for
PHREEQC Version 3 - A computer program for speciation, batch-reaction, one-
dimensional transport, and inverse geochemical calculations, U.S. Geological Survey
techniques and methods, Book 6, Chap. A43, 497p.

67
11
68
12
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Assessment of heavy metal contamination through sediment

cores from shallow harbour in Quebec, Canada

Masoumeh JAVID 1, Catherine N. MULLIGAN 1

1. Department of Building, Civil and Environmental Engineering, Concordia University,

1455 de Maisonneuve Blvd. W., Montreal, Qc, H3G 1M8, Canada.
[email protected]

Abstract:
Sediment cores from a shallow harbour on the bank of the St. Lawrence River, in
Quebec, Canada, were sectioned and analyzed for total metal content (Cr, Mn, Ni, Cu,
Zn, As, Cd and Pb) and organic matter. To evaluate the metal pollution in this area, the
total concentrations of heavy metals in 6 core sediments were analyzed. Strong and
similar sediment profiles were observed for Cr, Zn, Pb and Cd in the sediment. They
increased with increasing depth from top to the deepest part. In contrast, the Mn profile
for four cores showed lower concentrations in the deepest part, except that two of the
profiles showed a maximum concentration at a slightly shallower depth than the other
four profiles. The profile of Cu showed an increase with increasing the depth in four of
the cores and showed less concentration in the deepest part of the two other cores.
Profiles for Ni differed substantially from the other profiles; in two cores the
concentration was maximum at the deepest part. On the other hand, it remained
approximately constant, with only a slight trend towards lower concentrations with
depth. All of the concentrations were below the Frequent Effect Level (FEL) except for
Pb but some of the concentrations of elements exceeding the OEL (Occasional Effect
Level). Therefore, the sediment samples can be categorized as contaminated sediment,
which need to be managed.

Keywords:
Core sediment, Heavy metals, Yacht club, Contamination.

Corresponding author mail: [email protected]

1. Introduction
Nowadays, heavy metal contamination is a common, important, and growing problem
around the world (MULLIGAN et al., 2009). Large quantities of hazardous
contaminants such as heavy metals originating from anthropogenic activities are
accumulated within the sediments of lakes, rivers and marine areas world-wide
(GERBERSDORF et al., 2005). Heavy metals enter the environment in various forms
during metal mining, smelting, processing, recycling and disposal of wastes (HAN et
al., 2002). Although they are naturally occurring, anthropogenic activities such as
mining and smelting operations, industrial production and domestic and agricultural use
of metals and metal-containing compounds can introduce them in large quantities in
different environmental media (DURUIBE et al., 2007; TCHOUNWOU et al., 2014).
The problem of heavy metal pollution is a serious concern at various scales (local,
regional, and global) (RAI, 2008). In the aquatic environment, heavy metals are usually
distributed as follows: water-soluble species, colloids, suspended forms and
sedimentary phases (PENG et al., 2009). Whether the source of heavy metals is natural
or anthropogenic, they are absorbed by suspended solids and can accumulate in
sediments. Thus, sediment can act as a sink for accumulating contaminants and may act
as a source of contaminants as well, through the transferring of the heavy metals from
the sediment to water (KHAN et al., 2011; ZHANG & SHAO, 2013).
A large amount of heavy metals is discharged into the aquatic environment and
accumulated in sediments, some of them can directly pollute the raw water, resulting in
sublethal effects or death in local fish populations and accumulate in crops through
irrigation; some are released to the water through sediment resuspension, adsorption or
desorption reactions, reduction or oxidation reactions and the degradation of organisms.
Such processes will increase the dissolved concentration of metals in the environment
and threaten the ecosystem and human health (LIAO et al., 2017). The objective of this
work is to examine the level of contamination in the sediments near a yacht club.

2. Materials and methods

2.1. Study area and sampling

The core sediments were compacted and collected in a stainless-steel cylinder with a
diameter of 6.3 cm and a height of 21 cm. Sediment samples were transferred and kept
in the airtight poly- ethylene bottles and placed in an ice-cooled box. In total, 6 core
sediment samples were obtained from selected stations (Figure 1). The sediment cores
were sectioned into 2 cm intervals using a thin plastic cutter. The samples were
immediately stored in plastic container and then transferred to a freezer (-18° C) and
were homogenized before analysis.
All plastic- and glassware used during the experiment process were soaked in 5% (v/v)
nitric acid and 2.5% (v/v) hydrochloric acid (trace metal grade) for at least 8 h followed
70
by two rinses with deionized water (prepared using a Milli-Q 18 mΏ cm). Also, all
sediment samples were analyzed using a blank, control and duplicates.

Figure 1. Study area on the north bank of the St. Lawrence River with 6 selected
stations (Source: Google Earth).

2.2 Chemical analysis

The total metals were extracted from the sediment samples using an acid digestion
method (EPA 3050B). The total metals were extracted from the sediment samples with
repeated additions of nitric acid (HNO3, 70% - trace metal grade) and hydrogen
peroxide (H2O2, 30%) as well as hydrochloric acid (HCl, 35% - trace metal grade) at the
end of the digestion (USEPA, 1996). Samples were kept in the fridge around 4°C and
were analyzed with Inductively Coupled Plasma Mass Spectrometry (ICP-MS, Agilent
7700x). Seven metals and metalloids including Cr, Ni, Cu, Zn, As, Cd and Pb were
selected for analysis.

2.3 Determination of organic matter

2.3.1 Loss on Ignition (LOI)

Sediment samples were first dried at 105°C until a constant weight was obtained
(W105oC). Samples were weighed and then were placed in a furnace at 550°C for 4 h
(W550oC). Percentage mass loss was calculated based on the following equation:

LOI%= [(W105°C-W550°C)/ W105°C] *100 (1)

2.4 Sediment quality guidelines (SQG)

Sediment quality assessment guidelines (SQGs) are very useful in terms of revealing
sediment contamination by comparing the sediment concentration with the
corresponding quality guideline. These guidelines evaluate the degree to which the

71
sediment-associated chemical status might adversely affect aquatic biota and are
designed to assist in the interpretation of sediment quality (MAANAN et al., 2015;
SESHAN et al., 2010). SQGs include occasional effect level (OEL), the probable effect
level (PEL) and the frequent effect level (FEL), from low to high concentrations of the
substance level, respectively (Table 1) (ENVIRONMENT CANADA & MDDEP,
2007). The OEL and FEL were intended to evaluate the level of contamination for the
management of dredged sediment. For substances above the OEL, adverse effects are
predicted in many benthic species. Thus, open-water disposal is prohibited unless the
toxicity test shows there is no threat to aquatic biota. For heavy metal contents equal or
exceeding the FEL, open-water disposal is banned without any further tests. Sediments
containing elements exceeding the FEL are highly contaminated and the site must be
treated before any action as adverse effects are expected for the majority of benthic
species. Probable effect level (PEL) above which adverse biological effects are expected
to occur more often than not. Based on the SQG, the level of contamination below the
PEL does not justify initiation of site remediation. Therefore, for trace metal levels
above the PEL, evaluation of the contamination, risk assessment and determination of
the remediation is required (ENVIRONMENT CANADA,1993). The presence of one
trace metal in sediment samples that exceeds the standard level is sufficient to
categorize sediments as contaminated (POURABADEHEI & MULLIGAN, 2016).

Table 1. Criteria for assessment of sediment quality in St. Lawrence River [mg/kg]
(ENVIRONMENT CANADA & MDDEP, 2007).
Cr Ni Cu Zn As Cd Pb

OEL 57 47 63 170 8 1.7 52

PEL 90 - 200 310 17 3.5 91

FEL 120 - 700 770 23 12 150

3. Results and discussion

The sample properties are summarized in Table 2. S1 had the lowest sand and organic
contents. The texture of this area and S5 were less coarse in comparison to the rest of
the stations. S1 is the area that usually receives the runoffs from the maintenance area
(POURABADEHEI & MULLIGAN, 2016). S4 and S1 were representative of the
coarsest and the fines texture of sediment samples, respectively. S5 and S6 significant
amounts of organic materials as the loss on ignition (LOI) implies. The presence of
organic materials in these stations that are located at the stationary parts of the harbour
was more than other stations.

72
Table 2. Core sediment properties.
Core Grain size (mean±SDa)
Station LOI%
length
(S) Clay (av.) % Silt (av.) % Sand (av.) % (mean±SD)
(cm)
0.5-9.5 5.8-29.3 62.2-93.7
4 33 10.3±17.5
(3.2±2.4) (14.2±6.8) (82.6±9.1)
1.8-10.3 11.4-29.5 62.9-86.9
6 23 16.6±3.9
(5.8±2) (21.5±4.7) (72.8±6.5)
0.8-10.7 10.8-45.8 45.4-88.4
2 28 8.5±3.1
(4.9±2.8) (26.1±7.6) (69±10.1)
0.8-22.4 7.7-32.6 48.8-91.5
3 37 9±3
(7.9±5.3) (21.5±6.4) (70.6±11)
5.5-19.5 22.6-46.2 43.3-70.2
1 33 6.2±2.1
(10.2±4.2) (31.7±5.5) (58.1±6.5)
1.2-20.5 15.2-41.9 43.9-83.6
5 22 15.5±10.5
(8.1±5.2) (32±9.2) (59.9±13.7)
a Standard Deviation.

3.1 Variation in total concentration of heavy metals with sediment depth

Figures 2(a)-(e) show the variations in heavy metal accumulation with different
sediment depth of the six sediment cores. Cr in all stations tended to increase with an
increase in sediment depth. The Mn concentrations increased in stations 1, 3 and 6,
fluctuated in stations 3 and 4 and it did not change significantly in station 2 with an
increase in sediment depth. The total concentration of Ni did not vary with different
sediment depths for the stations 4 and 6. However, the Ni concentration gradually
increases from the surface with increasing the depth. The concentration of copper varied
with depth and has a maximum value of 52 mg/kg at a depth of 30–32 cm from the
surface in S4. The Cu concentration in S6 varied with depth, has a maximum value of
87 mg/kg at a depth of 0–2 cm, whereas the concentration of copper at the bottom of the
core is 71 mg/kg. The total concentration of Cu in stations 3, 1 and 5 tended to decrease
with an increase in sediment depth.
The total concentration of arsenic in stations 4, 3 and 5 tended to increase with an
increase in sediment depth. The concentration of arsenic at the bottom of the core is 6
mg/kg (32–33 cm), 10 mg/kg (36–37 cm) and 16 mg/kg (20–22 cm), respectively.
Arsenic concentration did not change significantly in station 6 with an increase in
sediment depth and has a maximum value of 10 mg/kg in the core fraction at a depth of
16–20 cm from the surface. Arsenic concentrations in station 2 fluctuated around an
average value of 8 mg/kg with no particular trend as a function of depth. It has a
maximum value of 12 mg/kg in the core fraction at 26–28 cm from the surface.

73
The total concentration of cadmium in stations 6, 3 and 5 tended to increase with an
increase in sediment depth. Cadmium concentrations in station 4 did not change
significantly with depth. The concentration of cadmium varied with depth, has
maximum value of 0.9 mg/kg and 1 mg/kg at a depth of 18–20 cm and 26-28 cm from
the surface in stations 1 and 2, respectively.

74
75
Figure 2. Variations in total concentration of heavy metals with different sediment
depths at the various stations.

76
The total concentration of Pb in stations 3, 4 and 5 tended to increase with an increase
in sediment depth. The concentration of lead varied with depth, has a maximum value of
87 mg/kg, 210 mg/kg and 221 mg/kg at a depth of 18–20 cm, 0-2 cm and 14-16 cm
from the surface in station 1, 2 and 5 respectively.

3.2 Spatial distributions of heavy metals in surface sediments from the Yacht club
Spatial distribution of Cr, Cu, Zn, Pb, Cd, Ni, and As in the core sediments from the
yacht club are shown in Table 3 using the sediment quality guideline (SQG). In this
study, heavy metal concentrations in the assessed sediment samples were compared
with OEL, PEL and FEL values (Table 1).
Mean concentrations of six heavy metals (Cr, Ni, Cu, Zn, Pb, Cd, and As) were higher
than their corresponding OEL except for Cd. Cr was higher than OEL in 80, 11.9, 3.3,
21.8 and 100 % of samples in S6, S2, S3 and S5 respectively. It decreased in the order
of S5 > S6 > S1 > S2 > S3 > S4. Concentrations of Ni in 2% of the samples were higher
than OEL in station 2. Cu was higher than OEL, average concentrations of each
sampling site for Cu decreased in the order S1=S2>S5>S6>S3. According to Table 3,
average concentrations of each sampling site for Zn, As, and Pb that exceed the OEL
were in the descending order of S6=1 > S2 > S5 > S3>S4, S6 > S5 > S2 > S1 > S3 and
S2>S6>S1>S5>S3.
As can be seen in Table 3, almost all the metals are above the PEL level, the exceptions
are Cd, Ni, indicating that the concentrations of these metals are likely to result in
frequent occurrence of harmful effects on sediment-dwelling organisms (LIAO et al.,
2017). Chromium and arsenic were higher than the PEL in 9.3% of samples in station 4.
As and Cd are proven carcinogens to human beings and have potential damage to the
ecological communities from the US Environmental Protection Agency (USEPA) (LI et
al., 2013). Copper was higher than the PEL in 84 and 36.25 % of samples in stations 2
and 1 respectively. The zinc concentrations exceeded the PEL limit in 16.7 and 2.38%
of the samples at S6 and S2 sampling points, respectively. Cu and Zn are two
micronutrients for aquatic life in all-natural water and sediments. Although both of them
are minor nutrients at low concentrations, they can become toxic to aquatic life at higher
concentrations than the threshold required levels (LI et al., 2013). Pb values were higher
than the PEL in 3.1, 9.4, 26, 73.8 % of the samples at S5, S1, S6 and S2 sampling
points, respectively. SADIQ et al., (2003) stated that the low concentrations of Pb still
might pose a threat to life in an aquatic environment in comparison with other heavy
metals.
All of the concentrations were below the FEL except for Pb, stations 6 and 2 which
presented Pb levels that are above the FEL in 10 or 21.4 % of the samples. Therefore,
the sediment samples can be categorized as highly contaminated and the site must be
treated before any action. Adverse effects are expected for the majority of benthic
species in the contaminated sediments which need to be managed.
77
Table 3. Comparison between sediment quality guidelines (SQGs) and heavy metal
concentrations (mg/kg) of all samples at the Yacht club.
Station 4 Cr Ni Cu Zn As Cd Pb
Minimum 14.9 0.37 12.3 51.5 2 0.12 10.9

Measured values in study (N= 46) Maximum 51.2 0.57 62.3 216.4 7.5 0.95 48.2

Median 32 0.51 27 109.7 3.9 0.36 21.4

Sample>OEL 0 0 0 7 0 0 0
SQGs (total%) Sample>PEL 0 0 0 0 0 0 0
Sample>FEL 0 0 0 0 0 0 0
Station 6
Minimum 50.5 29.6 51.2 188.8 7.5 0.67 41.1

Measured values in study (N= 30) Maximum 77.5 36.1 119 712.6 13.1 1.2 235.2

Median 61.8 33.1 65.7 285.7 10.3 0.9 81.4

Sample>OEL 80 0 50 100 90 0 80
SQGs (total%) Sample>PEL 0 0 0 16.7 0 0 26
Sample>FEL 0 0 0 0 0 0 10
Station 2
Minimum 26.8 15.4 165.5 101.5 4.2 0.31 46.1

Measured values in study (N= 42) Maximum 66 55.7 628.2 397.6 13.3 1.1 258.3

Median 44.9 22.5 277 201.3 7.7 0.7 121

Sample>OEL 11.9 2.38 100 71 45.2 0 90.4
SQGs (total%) Sample>PEL 0 0 83.3 2.38 0 0 73.8
Sample>FEL 0 0 0 0 0 0 21.4
Station 3
Minimum 18.6 11.9 23.7 66.7 2.4 0.2 13.8

Measured values in study (N= 57) Maximum 61.8 34.6 112.1 255.5 10.4 1.4 54.4

Median 39.4 23.5 55.4 149.6 5.6 0.5 33.2

Sample>OEL 5.3 0 29.8 28 14 0 8.7
SQGs (total%) Sample>PEL 0 0 0 0 0 0 0
Sample>FEL 0 0 0 0 0 0 0
Station 1
Minimum 24.8 15.6 111.7 103.4 3.7 0.22 37.6

Measured values in study (N= 32) Maximum 77.3 32.1 587.5 308.4 12.9 1.1 101.4

Median 43.3 22.8 242 195.9 6.3 0.5 61.4

Sample>OEL 21.8 0 100 56.25 21.8 0 56.25
SQGs (total%) Sample>PEL 0 0 56.25 0 0 0 9.4
Sample>FEL 0 0 0 0 0 0 0

78
 Station 5
Minimum 62.8 27.2 51.9 189 4.9 0.64 37.5

Measured values in study (N= 32) Maximum 115.8 44.8 107.9 305 19.7 1.2 98.6

Median 73.4 34.4 67 235.7 10.5 0.9 52.3

Sample>OEL 100 0 65.6 100 56.25 0 46.8
SQGs (total%) Sample>PEL 9.3 0 0 0 9.37 0 3.1
Sample>FEL 0 0 0 0 0 0 0

Conclusions
In this work, 7 heavy metals in 6 core sediments from the shallow harbour on the bank
of the St. Lawrence River, in Quebec, Canada, were investigated. The metal profiles in
the cores from different areas of this area suggest that contaminant accumulations are
similar throughout the region for Cr, Zn, Pb, Cu and Cd. An increase in metal
concentrations with increasing depth from top to the deepest part can be seen.
Interestingly, the Mn profile for the four cores showed less concentration in the deepest
part. The Ni profiles showed that concentrations were maximum in the deepest part in
the two cores. On the other hand, they remained approximately constant, with only a
slight trend towards lower concentrations with depth. The significantly high
concentrations of As, Pb, Cd, Zn, Cu and Cr in sediments were found in the mixed
region by comparing with SQGs, indicating that the target metals are likely to result in
harmful effects, and the risk is expected to occur frequently for aquatic organisms.

References
DURUIBE J.O., OGWUEGBU M.O.C., EGWURUGWU J.N. (2007). Heavy metal
pollution and human biotoxic effects. Int. J. Phys. Sci. 2, 112,
https://doi.org/10.1016/j.proenv.2011.09.146
ENVIRONMENT CANADA (1993). Sediment contamination in Montreal harbor:
Overview and remediation proposals. Minister of Supply and Service Canada.
ENVIRONMENT CANADA – MDDEP (2007). Criteria for the assessment of sediment
quality in Quebec and application frameworks: Prevention, dredging and remediation,
Ministère du Développement durable, de l'Environnement et des Parcs du Québec, 39 p.
GERBERSDORF S.U., JANCKE T., WESTRICH B., (2005). Physico-chemical and
biological sediment properties determining erosion resistance of contaminated riverine
sediments – Temporal and vertical pattern at the Lauffen reservoir / River Neckar,
Germany. Limnologica 35, pp. 132–144, https://doi.org/10.1016/j.limno.2005.05.001
HAN F.X., BANIN A., SU Y., MONTS D.L., PLODINEC M.J., KINGERY W.L.,
TRIPLETT G.E. (2002). Industrial age anthropogenic inputs of heavy metals into the
pedosphere. Naturwissenschaften 89, pp. 497–504, https://doiorg/10.1007/s00114-002-0373-4
KHAN M.S., ZAIDI A., GOEL R., MUSARRAT J. (2011). Biomanagement of metal-

79
contaminated soils. Springer.
LI F., HUANG J., ZENG G., YUAN X., LI X., LIANG J., WANG X., TANG X., BAI
B. (2013). Spatial risk assessment and sources identification of heavy metals in surface
sediments from the Dongting Lake, Middle China. J. Geochemical Explor. 132, pp. 75–
83, https://doi.org/10.1016/j.gexplo.2013.05.007
LIAO J., CHEN J., RU X., CHEN J., WU H., WEI C. (2017). Heavy metals in river
surface sediments affected with multiple pollution sources, South China: Distribution,
enrichment and source apportionment. J. Geochemical Explor. 176, pp. 9–19,
https://doi.org/10.1016/j.gexplo.2016.08.013
MAANAN M., SADDIK M., MAANAN M., CHAIBI M., ASSOBHEI O.,
ZOURARAH B. (2015). Environmental and ecological risk assessment of heavy metals
in sediments of Nador lagoon, Morocco. Ecol. Indic. 48, pp. 616–626,
https://doi.org/10.1016/j.ecolind.2014.09.034
MULLIGAN C.N., DAVARPANAH N., FUKUE M., INOUE T. (2009). Filtration of
contaminated suspended solids for the treatment of surface water. Chemosphere 74, pp.
779–786, https://doi.org/10.1016/j.chemosphere.2008.10.055
PENG J., SONG Y., YUAN P., CUI X., QIU G. (2009). The remediation of heavy
metals contaminated sediment. J. Hazard. Mater. 161, pp. 633–640,
https://doi.org/10.1016/j.jhazmat.2008.04.061
POURABADEHEI M., MULLIGAN C.N. (2016). Effect of the resuspension technique
on distribution of the heavy metals in sediment and suspended particulate matter.
Chemosphere, 153, pp. 58–67, https://doi.org/10.1016/j.chemosphere.2016.03.026
RAI P.K. (2008). Heavy metal pollution in aquatic ecosystems and its phytoremediation
using wetland plants : An ecosustainable approach. Int. J. Phytoremediation 10, pp. 37–
41, https://doi.org/10.1080/15226510801913918
SADIQ R., HUSAIN T., BOSE N., VEITCH B. (2003). Distribution of heavy metals in
sediment pore water due to offshore discharges : an ecological risk assessment,
Environmental Modelling & Software, vol.18, pp. 451–461, https://doi.org/10.1016/S1364-
8152(03)00010-0
SESHAN B.R.R., NATESAN U., DEEPTHI K. (2010). Geochemical and statistical
approach for evaluation of heavy metal pollution in core sediments in southeast coast of
India. Int. J. Environ. Sci. Tech 7, pp. 291–306, https://doi.org/10.1007/BF03326139
TCHOUNWOU P.B., YEDJOU C.G., PATLOLLA A.K., SUTTON D.J. (2014). Heavy
metals toxicity and the environment. natl. institutes health, pp.1–30.
https://doi.org/10.1007/978-3-7643-8340-4
ZHANG L., SHAO H. (2013). Heavy metal pollution in sediments from aquatic
ecosystems in China. Clean – Soil, Air, Water 41, pp. 878–882.
https://doi.org/10.1002/clen.201200565

80
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

A new prediction method for the rainfall-induced landslides

and debris flows

Sangseom JEONG 1, Moonhyun HONG 1

1. Yonsei Univ., Civil and Environmental Eng., Seoul, 03722, Republic of Korea.
[email protected]; [email protected]

Abstract:
In this study, a unified method was proposed for the analysis of rainfall-induced
landslides and subsequently debris flows in a regional scale. Special attention is given to
the calculation of the propagation of debris flows by considering rainfall infiltration into
soil slopes and soil entrainments by debris flows. A susceptibility assessment of
landslides and debris flows in Umyeonsan (Mt.), South Korea was conducted as a case
study by using a combined method for rainfall-induced landslides and debris flows. It is
shown that the result of the analysis has a good agreement with historical landslides and
debris flows. Based on the case study for Umyeonsan (Mt.) landslides, failure behaviors
of landslides and debris flows are well predicted by the method described in this study.
This procedure can be applied to the prediction of landslides and debris flows of another
study area.

Keywords:
Rainfall-induced landslide, Debris flow, Regional scale, Soil entrainment, Combined
numerical method, Umyeonsan mountain.

1. Introduction
Large-mass and high-velocity debris flows can be fatal to human society and
infrastructures. Solving these dynamic problems with extremely large deformations over
a short event time are slightly different from traditional problems in geotechnical
engineering. It is highly difficult to reproduce debris flows in the field, so lab- or large-
scale experiments or numerical simulations are usually carried out to better understand
the mechanisms and assess the risks (CHEN & LEE, 2000).
This paper presents a simplified depth-averaged debris flow model for tracking density
evolution. The developed model uses Hershel-Buckley rheology in internal and basal
frictions and considers complex terrains and entrainments. In particular, the interaction
between solid-fluid phases in the mixture is ignored. A finite volume formulation of the
proposed model is presented with relevant numerical schemes to obtain stable and
accurate solutions.

Corresponding author mail: [email protected]

2. Prediction method for the rainfall-induced landslides and debris flows
The governing equations derived by depth-averaging the Navier-Stokes equations can
facilitate debris flow simulations at a large scale by reducing the computational cost
compared to full-scale 3D simulations. The continuity equations for the mixture and solid
phases and the momentum equations of the mixture phase for the x- and y-axes are given
as:

𝜕ℎ𝜌 𝜕 𝜕 𝜕𝑧𝑏
+ (ℎ𝜌𝑣𝑥 ) + (ℎ𝜌𝑣𝑦 ) = −𝜌𝑏 (1)
𝜕𝑡 𝜕𝑥 𝜕𝑦 𝜕𝑡
𝜕ℎ𝑐𝑠 𝜕 𝜕 𝜕𝑧𝑏
+ 𝜕𝑥 (ℎ𝑐𝑠 𝑣𝑠,𝑥 ) + 𝜕𝑦 (ℎ𝑐𝑠 𝑣𝑠,𝑦 ) = −𝑐𝑏𝑠 (2)
𝜕𝑡 𝜕𝑡
𝜕ℎ𝜌𝑣𝑥 𝜕ℎ𝜌𝛼𝑚 𝑣𝑥 𝑣𝑥 𝜕ℎ𝜌𝛼𝑚 𝑣𝑦 𝑣𝑥 𝜕ℎ𝜏̄ 𝑥𝑥 𝜕ℎ𝜏̄ 𝑦𝑥 𝜕ℎ𝑝̄ 𝜕𝑧𝑏
+ + = + 𝜕𝑥 + 𝜕𝑦 − 𝜕𝑥 − 𝑝𝑏 + 𝜔𝜏𝑏 𝑡𝑏,𝑥 (3)
𝜕𝑡 𝜕𝑥 𝜕𝑦 𝜕𝑥
𝜕ℎ𝜌𝑣𝑦 𝜕ℎ𝜌𝛼𝑚 𝑣𝑥 𝑣𝑦 𝜕ℎ𝜌𝛼𝑚 𝑣𝑦 𝑣𝑦 𝜕ℎ𝜏̄ 𝑥𝑦 𝜕ℎ𝜏̄ 𝑦𝑦 𝜕ℎ𝑝̄ 𝜕𝑧𝑏
+ + =+ + − − 𝑝𝑏 + 𝜔𝜏𝑏 𝑡𝑏,𝑦 (4)
𝜕𝑡 𝜕𝑥 𝜕𝑦 𝜕𝑥 𝜕𝑦 𝜕𝑦 𝜕𝑦

where ℎ 𝜌 nd 𝑐𝑠 nre the ebris height mixture edsity nd soli volume frnctiod
respectively. 𝑣𝑘 nd 𝑣𝑘,𝑠 nre the k-nxis compodedts of the epth-nvernge velocity for the
mixture nd soli phnses respectively. 𝜌𝑏 nd 𝑐𝑏𝑠 nre the mixture edsity nd soli
volume frnctiod of nd edtrnide bnsnl se imedt. 𝑧𝑏 is the surfnce level of bnsnl se imedt.
𝛼𝑚 is n momedtum correctiod fnctor. 𝜏̄𝑖𝑗 is the epth-nvernge shenr stress. 𝑝̄
𝑝̄ nd 𝑝𝑏 nre epth-nvernge nd bnsnl pressures respectively. 𝜏𝑏 is the bnsnl frictiod
stress. 𝑡𝑏,𝑖 is n irectiodnl vector of bnsnl frictiod.
The z-nxis compodedt of the epth-nvernge velocity nd the irectiodnl vector of bnsnl
frictiod nre expresse ns:

𝜕𝑧𝑏 𝜕𝑧𝑏
𝑣𝑧 = 𝑣𝑥 + 𝑣𝑦 (5)
𝜕𝑥 𝜕𝑦
𝑣𝑖
𝑡𝑏,𝑖 = − (6)
√𝑣𝑥2 +𝑣𝑦2 +𝑣𝑧2

This stu y uses the Herschel-Buckley mo el for the rheology of ebris flow ns

𝜇 = 𝑚𝑖𝑛([𝜏𝑦 + 𝑘0 𝛾 𝑛 ]𝛾 −1 , 𝜇𝑚𝑎𝑥 ) (7)

where 𝜏𝑦 is the yiel stress 𝛾 is the mngditu e of the shenr rnte 𝑘0 is n codsistedcy id ex
𝑛 is n flow id ex nd 𝜇𝑚𝑎𝑥 is the mnximum viscosity thnt prevedts idfiditely high
viscosity whed the shenr rnte nppronches zero.
The epth-nvernge shenr stnid nd stress tedsors nre npproximnte ns:

82
 1
𝛾̄ 𝑖𝑗 ≈ 2 (𝜕𝑖 𝑣𝑗 + 𝜕𝑗 𝑣𝑖 ) for 𝑖, 𝑗 = 𝑥, 𝑦 nd 𝑧 (8)
𝜏̄𝑖𝑗 = 𝜇𝛾̄ 𝑖𝑗 for 𝑖, 𝑗 = 𝑥 nd 𝑦 (9)

The x- nd y- irectiodnl erivntives of the epth-nvernge velocities nre npproximnte by

the cedtrnl ifferedce scheme. The z- irectiodnl erivntive is nlso nssume ns:

𝜕𝑣𝑘 𝑢𝑘 |𝑧𝑏 +ℎ −𝑢𝑘 |𝑧𝑏 𝑣𝑘

≈ = (2 − 𝛼) (10)
𝜕𝑧 ℎ ℎ

This stu y uses the ydnmic equilibrium nppronch for the edtrnidmedt of bnsnl se imedt
propose by FRACCAROLLO & CAPART (2002) nd MEDINA et al. (2008). The
edtrnidmedt rnte is gived ns:

𝜕𝑧𝑏 𝜏𝑏 −𝜎′ 𝑏 𝑡𝑎𝑛𝜙𝑏𝑠 −𝑐𝑏𝑠

=− (11)
𝜕𝑡
𝜌√𝑣𝑥2 +𝑣𝑦2 +𝑣𝑧2

where 𝜙𝑏𝑠 nd 𝑐𝑏𝑠 nre the idterdnl frictiod ndgle nd cohesiod of the bnsnl se imedt
respectively.
By implemedtidg n time idtegrntiod of Eq. (11) we cnd obtnid the erosiod positiod 𝑧𝑒
which becomes the dew bnsnl surfnce level nfter the edtrnidmedt ns:

𝑡+𝛥𝑡 𝜕𝑧𝑏
𝑧𝑏 (𝑡 + 𝛥𝑡) = 𝑧𝑒 = 𝑧𝑏 (𝑡) + ∫𝑡 𝑑𝑡 (12)
𝜕𝑡

3. Case study

3.1 Study area description

The study area of Umyeonsan (Mt.) is located in the south-east of Seoul, Korea with a
total area of 418 ha and an altitude ranging from 50 to 312 m. Between 8:30 to 8:50 a.m.
on July 27, 2011, landslides and debris flows occurred simultaneously in the study area.
A total of 151 landslides and 33 debris flows expanded from one or more landslides were
reported (JEONG et al., 2015). Figure 1 shows the location of the landslides and traces
of the debris flows in the satellite image. In this study, simulation of the Umyeonsan (Mt.)
landslides and debris flows was conducted based on the observed rainfall information at
the time of the landslide occurrence, the site investigation results, and the simulation
results were compared with the measured data.

83
 Seocho AWS
126.990280 E 127.020152 E
N

37.478027 N
Namhyeon AWS

37.460413 N
China Seoul Umyeonsan (Mt.) Landslide
South Debris flow path
Korea
Japan 0 0.5 1.0 km

Figure 1. Umyeonsan (Mt.) landslides and traces.

3.2 Modeling and input parameters

Two automatic weather stations are located near the study area: one is Seocho station and
the other is Namhyun station (Figure 1). The landslides occurred on the morning of July
27, and the hourly rainfall recorded ranged from 70 to 85 mm/h maximum. The
cumulative rainfall was recorded as 424.5 mm at Namhyun station and 364.5 mm at
Seocho station. These rainfall records were applied as rainfall input data for the rainfall-
infiltration analysis and slope stability analysis.
Quercus mongolica was the most distributed species (43 %) and had a diameter at breast
height of 23.1 cm and an average height of 10.7 m. From these investigations, the
representative additional shear strength at the roots of the tree was determined as 1.0 kPa,
and the uniform load by tree was estimated as 0.253 kPa (KIM et al., 2017).
Ground investigation was also conducted to determine the geotechnical and hydrological
properties of the study area. Seven boreholes were drilled and in situ tests were conducted,
including borehole shear tests. A series of laboratory tests including the soil water
characteristic curve (SWCC) tests, direct shear tests and density tests were conducted to
estimate the geotechnical and hydrological properties of the colluvium soil of the study
area (KIM et al., 2017). Table 1 summarizes the input parameters applied to the landslide
and debris flow simulations determined from the field and laboratory test results.

84
Table 1. Input parameters for the numerical simulation.
Parameters Values
Permeability (m/s) 8×10-6
Initial water content 0.3
Deficit water content 0.2
Matric suction head (m) 0.83
Soil cohesion (kPa) 6.9-18.5 (11.7)
Soil internal friction angle (°) 21.7-32.1 (25.3)
3
Total unit weight of soil (kN/m ) 18
Additional strength by roots of tree (kPa) 1.0
Additional load by tree (kPa) 0.253

3.3 Results and discussion

Figure 2 shows the debris flow propagations of the entire study area after the landslide
initiation (t = 0). The debris flow from the upstream to the middle of the mountain reaches
the downstream after approximately 20 seconds. The sporadic landslides merge into the
valley and flow downstream, and the thickness of the debris flow increases due to
entrainment with time.
A regional-scale analysis of rainfall-induced landslide and debris flows was conducted,
and it confirmed that the proposed numerical method can offer meaningful prediction of
rainfall-induced landslide and debris flows.

339 m Time = 0 sec N 339 m Time = 10 sec N 339 m Time = 30 sec N

Thickness (m) Thickness (m) Thickness (m)

0~1 0~1 0~1
1~2 1~2 1~2
2~3 2~3 2~3
3~4 3~4 3~4
4~5 4~5 4~5

0m 0 0.5 1.0 km 0m 0 0.5 1.0 km 0m 0 0.5 1.0 km

(a) Time = 0.5 sec (b) Time = 10 sec (c) Time = 30 sec
Figure 2. Debris flow propagations from the numerical simulation.

4. Conclusion
This work developed a simplified debris flow model with Hershel-Buckley rheology for
tracking density evolution. A finite volume formulation of the debris flow model was also
proposed for accurate and stable numerical simulations. Both the internal and basal
frictions of the debris flow were considered in the model as well as the basal topology
effect. By using the model developed in this study, it is possible to simulate not only
reverse analysis after events but also the expandable debris flow induced by the input
rainfall applied in engineering practice.

85
5. References
CHEN H., LEE C. F. (2000). Numerical simulation of debris flows. Canadian
Geotechnical Journal, Vol. 37(1), pp. 146–160, https://doi.org/10.1139/t99-089
FRACCAROLLO L., CAPART H. (2002). Riemann wave description of erosional dam-
break flows. Journal of Fluid Mechanics, Vol. 461, pp. 183-228,
doi: 10.1017/s0022112002008455
MEDINA V., HÜRLIMANN M., BATEMAN A. (2008). Application of FLATModel, a
2D finite volume code, to debris flows in the northeastern part of the Iberian Peninsula.
Landslides, Vol. 5(1), pp. 127–142, doi: 10.1007/s10346-007-0102-3
JEONG S., KIM Y., LEE J. K., KIM J. (2015). The 27 July 2011 debris flows at
Umyeonsan, Seoul, Korea. Landslides, Vol. 12(4), pp. 799–813, https://doi.org/10.1007/s10346-
015-0595-0
KIM J., KIM Y., JEONG S., HONG, M. (2017). Rainfall-induced landslides by deficit
field matric suction in unsaturated soil slopes. Environmental Earth Sciences, Vol.
76(23), 808, 17p., doi:10.1007/s12665-017-7127-2

86
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Development of a model to quantify dust emissions from

truck traffic on earthmoving sites
Approach and preliminary results
Mickael LE VERN 1, Ouardia SEDIKI 1, Andry RAZAKAMANANTSOA 1,
Frédéric MURZYN 2, Frédérique LARRARTE 3, Pascal INSENGA 4, Philippe
GOTTELAND 5

1. University Gustave Eiffel, GERS Department, Allée des Ponts et Chaussées, 44344,
Bouguenais, France.
[email protected]; [email protected];
[email protected]
2. Department of Mechanical Engineering, ESTACA West Campus, Rue Georges
Charpak, 53000, Laval, France.
[email protected];
3. University Gustave Eiffel, GERS Department, Boulevard Newton, 77447, Champs sur
Marne, France.
[email protected]
4. VINCI Construction Terrassement, 61 Avenue Jules Quentin, 92730, Nanterre, France.
[email protected]
5. Fédération Nationale des Travaux Publics, 3 Rue de Berri, 75008, Paris, France.
[email protected]

Abstract:
Dust emission is a major issue in general but even more for civil engineering
professionals. This is particularly important on earthmoving sites with truck traffic on
compacted soils. Suspended dust causes many troubles, including visibility mitigation,
which leads to safety risks. To prevent the lift of particles, soils are watered abundantly.
When carried out in an arbitrary way, this solution can lead to a significant waste of this
natural resource. Savings could be achieved by anticipating the necessary and sufficient
quantities for dust suppression. For this purpose, a model is being developed to highlight
and quantify the main factors involved in the lifting of soil particles by truck traffic. The
approach adopted for the development of the model together with the preliminary results
are presented. Soils degradation by traffic are studied in laboratory in order to assess the
accumulation of particles on the surface. These particles are lift due to mechanical and
aerodynamic forces, which are studied in test track and wind tunnel, respectively. This
study allows to characterize the mechanism of dust emission on construction sites by

Corresponding author mail: [email protected]

relying on physical parameters that are not considered in the empirical models currently
available.

Keywords:
Dust emission, Truck traffic, Earthworks, Soil mechanics, Wake flows, Water savings.

1. Introduction
In an increasing stringent context of improving air quality, limiting dust emissions
generated by human activities is an overriding objective for environmental and public
health issues. Soil degradation is a major factor of particulate matter suspension in the
atmosphere. It is caused in particular by vehicle traffic on unpaved roads. Due to the
contact between the wheel and the ground, the aggregates of soil particles can be torn off
and then crumbled. This leads to the formation of fine particles which are prone to fly
away when vehicles pass by or because of the wind and related aerodynamic processes
(ETYEMEZIAN et al., 2003; GILLIES et al., 2005; KUHNS et al., 2010).
Airborne dust suspension leads to health issues (POPE & DOCKERY, 2006) as well as
visibility mitigation (MOOSMÜLLER et al., 2005; BADDOCK et al., 2014). This
involves significant safety risks, particularly in areas with vehicle traffic (ASHLEY et al.,
2015).
On earthmoving sites, the truck traffic on runways is a major source of dust (MULESKI
et al., 2005). The most common method to minimize particle lift is to spray the ground
with a huge amount of water. This watering was estimated to be about 12 liters of water
per cubic meter of earthworks in France in 2010 (SPTF, 2011). This can lead to a
significant waste of water.
Resource savings could be made by anticipating the necessary and sufficient quantities to
prevent dust emission. In order to achieve this goal, it is require to study the mechanisms
of soil degradation and particle lift by traffic. To this end, research works have been
undertaken in recent years at the University Gustave Eiffel (SEDIKI et al., 2018; LE
VERN et al., 2020a; LE VERN et al., 2020b). This paper aims at presenting the
corresponding approach as well as some preliminary results.

2. Methodology
The dust emission mechanisms are assessed according to three main lines of study:
- The soil surface degradation under traffic stress is analyzed at the laboratory scale.
- The physical, mechanical and environmental parameters that govern dust emission when
a vehicle is in motion are highlighted with a test track study.
- The transport of particles in the wake of trucks circulating on the earthworks runways
is characterize with scale models in a wind tunnel.
Figure 1 illustrates the steps that lead to the dust plume formation. Above a certain
threshold of particulate concentration that needs to be identified, suspended dust will
88
significantly reduce visibility. This nuisance requires the watering of the runways on the
construction site. The aim of the study is to develop a model based on physical laws and
experimental measurements for the quantification of dust concentration generated by
trucks traffic on runways. A better knowledge of dust emission mechanisms would allow
to limit them and to forecast the quantities of water needed during the construction
project.

Figure 1. Processes related to dust emissions, from soil compaction to its watering.

3. Preliminary results
The dust emission model under development is based on three parameters:
- The number of particles likely to fly away as they are no longer bound to the granular
matrix. They appear due to soil surface degradation by traffic;
- The Particle Size Distribution (PSD) of these particles;
- The mechanical and aerodynamic stresses on particles when vehicles are in motion.
These three points are discussed below.

3.1 Origin of dust: mechanical degradation of the soil surface

Earthworks runways are made by soil moistening and compaction. This process forms a
granular matrix consisting of aggregates. Dust is generated by trucks traffic as these
aggregates are torn off by the wheels. This phenomenon was studied with a traffic
simulator reproducing the motion of a wheel on laboratory-prepared soils. They were
made with two materials: Speswhite kaolin clay (denoted “K”) and Hostun HN38 sand
(denoted “S”). Three samples were degraded in the traffic simulator: S0K100 (100%

89
clay), S50K50 (50% sand and 50% clay) and S75K25 (75% sand and 25% clay). Table 1
summarises their geotechnical properties.

Table 1 : Soil geotechnical properties

Liquid Plastic Methylene blue Specific surface Unified Soil
Soil 2
Limit (%) Index (%) value (g/100g) area (cm /g) Classification System
CH: High plasticity
S0K100 55.00 25.00 1.56 10.65
clay
CL: Low plasticity
S50K50 29.90 12.60 0.74 5.28
clay
S75K25 23.10 7.60 0.43 2.47 SC: Clayed sand

Figure 2 shows the experimental results of their surface degradation as a function of the
number of wheel passes. They represent the cumulative mass of aggregates torn off per
unit of ground surface covered by the tire during its movement. The results show
deviations of several orders of magnitude from one soil to another. For the same number
of wheel passes, high-plasticity clay generates ten times more dust than clayed sand and
hundred times more than low-plasticity clay. This latter has the least propensity to
generate dust under traffic because it is a well graded soil with a high shear strength (LE
VERN et al., 2020a).

Figure 2. Soil surface degradation as a function of the number of wheel passes, with
50% uncertainty (shading).

3.2 Particle size distribution of dust during traffic degradation

By measuring the Particle Size Distribution (PSD) of the aggregates torn off during soil
degradation by traffic, a typical trend can be observed. An example is given in Figure 3
for S75K25 sample.
90
 Figure 3. Particle size distributions of initial soil (S75K25) and dust particles after N
passes of the wheel.

For the three soils, similar evolutions of aggregate PSD as a function of the number of
wheel passes are observed. They are empirically modelled by the difference between the
PSD curve of aggregates torn off after N wheel passes and the PSD curve of the parent
soil, using the Δ function (equation (1):
0.111 × e[0.004(%clay × %sand)]
Δ(d) = × (9.72 × 10-10 N2 -4.92 × 10-5 N
0,49
(1)
d 2π d
+ 1.09) × ln( ) × sin [ × ln ( ) + π]
dmin d dmin
ln ( max )
dmin

where d is the diameter of particle (μm), %clay is the mass percentage of particles smaller
than 2 μm, %sand is the mass percentage of particles between 20 and 2000 μm, dmin and
dmax represent the minimum and maximum soil particle diameter (μm) respectively (LE
VERN et al., 2020b). This empirical function highlights the capacity of a soil to form
aggregates (characterized by the product %clay x %sand) and their gradual crumbling
when the number of wheel passes N increases. This makes it possible to know the PSD
evolution of the dust to be put into suspension during the soil degradation by traffic.
3.3 Particle lift by vehicle motion
When a vehicle is in motion, the lift of particles depends jointly on physical (type and
quantity of particles, vehicle geometry), mechanical (vehicle speed, tyre/surface contact
dynamics) and environmental (wind, relative humidity and air temperature) parameters.

91
An experimental study on a test track was carried out to assess the impact of two different
moving vehicles (Renault Clio III and Ford Ranger) on particle lift. Our attention was
particularly focused on the effect of the above-mentioned parameters on airborne
particulate matter concentration when vehicle traffic occurs over a bed made of four
different types of particles:
- Montmorillonite clay
- Clayed silt from Marche des Dames (France)
- Silt from Val d’Europe (France)
- Loess from Strasbourg (France)
Each test consisted of running one of the two vehicles at a given speed (30, 45 or 60 km.h-
1
) over one of the four types of particles previously placed on the ground at a given surface
concentration (200, 400 or 600 g.m-2). This latter is directly related to the surface
degradation (Figure 2). Figure 4-a shows the experimental setup. The concentration of
resuspended particles was measured by suction of the dust plume behind the rear tire
using a particle counter. Figure 4-b shows an example of measurements for one of the
tests (PMX being the concentration of particulate matter having a diameter less than X
μm). Temperature, relative humidity and solar radiation were monitored with a weather
station. The Renault Clio III was equipped with a dynamometric hub enabling real-time
measurements of the longitudinal, vertical and transverse forces and torques applied to
the right front wheel. Overall, 74 tests were conducted corresponding to a wide range of
experimental conditions. The results are currently being analyzed.

Figure 4. a) Configuration of test track experiments; b) Particulate concentration

measurements during the vehicle traffic over the bed of particles.

4. Future works
Future works will be undertaken to characterize the transport of dust in the wake of trucks
circulating on earthworks runways. The study will take place in a wind tunnel with two
models corresponding to the types of trucks used to carry materials on construction sites:
articulated dump trucks (Figure 5-a) and cab-over trucks (Figure 5-b). Wake flows (mean

92
and turbulent velocity fields) will be characterized by LASER Doppler Velocimetry and
Particle Image Velocimetry. The final goal is to couple the results obtained in the wind
tunnel with those observed on the test track in order to quantify the stresses on suspended
particles.

Figure 5. Wind tunnel scale model of a) Caterpillar articulated dump truck and b)
Mercedes 8x4 cab-over truck.

5. Conclusions
To date, works carried out have made it possible to characterize the appearance of dust
during traffic on compacted soils. Models of surface degradation and PSD evolutions
have been developed. They will need to be validated by testing a wider range of soils.
This work represents a novelty in the field of dust emission studies and is a first step in
the development of a robust model for the quantification of dust emissions. A particular
attention will be drawn to the lift of particles due to mechanical and aerodynamic forces.
Analysis of the test-track experiments results and wind-tunnel studies will be undertaken
to this end.

7. Acknowledgments
The authors would like to express their gratitude to the Fédération Nationale des Travaux
Publics (grant number RP2_E18101) and the Region Pays de la Loire (grant number
GHPAI3) for their financial support. Special thanks to Erwann Rayssac, Sébastien
Buisson and Olivier Burban for their help during the experiments.

93
8. References
ASHLEY W.S., STRADER S., DZIUBLA D.C., HABERLIE A. (2015). Driving blind:
Weather-related vision hazard and fatal motor vehicle crashes. Bulletin of the American
Meteorological Society, Vol. 96, pp. 755-779. doi: 10.1175/BAMS-D-14-00026.1
BADDOCK M.C., STRONG C.L., LEYS J.F., HEIDENREICH S.K., TEWS E.K.,
MCTAINSH G.H. (1992). A visibility and total suspended dust relationship. Atmospheric
Environment, Vol. 89, pp. 329-336. doi: 10.1016/j.atmosenv.2014.02.038
ETYEMEZIAN V., KUHNS H., GILLIES J., GREEN M., PITCHFORD M., WATSON
J. (2003). Vehicle-based road dust emission measurements: I-methods and calibration,
Atmospheric Environment Vol. 37, pp. 4559-4571. doi:10.1016/S1352-2310(03)00528-4
GILLIES J.A., ETYEMEZIAN V., KUHNS H., NIKOLIC D., GILLETTE D.A. (2005).
Effect of vehicle characteristics on unpaved road dust emissions. Atmospheric
Environment Vol. 39, pp. 2341-2347. doi: 10.1016/j.atmosenv.2004.05.064
KUHNS H., GILLIES J., ETYEMEZIAN V., NIKOLICH G., KING J., ZHU D.,
UPPAPALLI S., ENGELBRECHTJ., KOHL S. (2010). Effect of soil type and momentum
on unpaved road particulate matter emissions from wheeled and tracked vehicles,
Aerosol Science and Technology, Vol. 44, pp. 187-196, doi:10.1080/02786820903516844
LE VERN M., SEDIKI O., RAZAKAMANANTSOA A.R., MURZYN F., LARRARTE
F. (2020a). Experimental study of particle lift initiation on roller compacted sand-clay
mixtures. Environmental Geotechnics. doi: 10.1680/jenge.19.00172
LE VERN M., SEDIKI O., RAZAKAMANANTSOA A.R., MURZYN F., LARRARTE
F. (2020b). Experimental assessment of dust emissions on compacted soils degraded by
traffic. Atmosphere Vol. 11, 369, doi: 10.3390/atmos11040369
MOOSMÜLLER H., VARMA R., ARNOTT W.P., KUHNS H.D., ETYEMEZIAN V.,
GILLIES J.A. (2005). Scattering cross-section emission factors for visibility and
radiative transfer applications: Military vehicles traveling on unpaved roads, Journal of
the Air & Waste Management Association, Vol. 55, pp. 1743-1750.
DOI:10.1080/10473289.2005.10464763
MULESKI G.E., COWHERD C., KINSEY J.S. (2005). Particulate emissions from
construction activities. Journal of Air and Waste Management Association, Vol. 56, pp.
709-742. doi:10.1080/10473289.2005.10464669
POPE C.A., DOCKERY D.W. (2006). Health effects of particulate air pollution: Lines
that connect. Journal of Air and Waste Management Association, Vol. 55, pp. 772-783.
doi:10.1080/10473289.2006.10464485
SEDIKI O., RAZAKAMANANTSOA A., HATTAB M., LE BORGNE T., FLEUREAU
J.M., GOTTELAND P. (2018). Degradability of unpaved roads submitted to traffic and
environmental solicitations: laboratory scale. 7th International Conference of
Unsaturated Soils, HKUST, Hong Kong, 3-5 August 2018.
SPTF (2011). Intensifier l’engagement pour la route durable en 2012. COTITA Sud-
Ouest, in French, http://cotita.fr/IMG/pdf/11_VINCI_pour_SPTF_COTITA_SO_111103.pdf
94
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Leaching behaviour of arsenic from excavated rock

under anaerobic conditions

Yusuke MASAKI 1, Tomohiro KATO 1,

Lincoln Waweru GATHUKA 1, Atsushi TAKAI 1, Takeshi KATSUMI 1

1. Graduate School of Global Environmental Studies, Kyoto University, Yoshida-

honmachi, Sakyo-ku, 606-8501, Kyoto, Japan.
[email protected]; [email protected];
[email protected]; [email protected];
[email protected]

Abstract:
For utilisation of excavated rocks as geomaterials, the leaching behaviour of geogenic
contaminants sometimes should be evaluated. This is because toxic metals and metalloids
such as arsenic and lead may be naturally contained in the excavated rocks and may
sometimes be leached in concentrations exceeding the mandated limit. Leaching
characteristics and behaviours of soils or rocks of concern are usually evaluated using
conventional batch leaching tests. Although the soils and rocks of concern would exist
under anaerobic conditions in the embankment, their leaching behaviour under anaerobic
conditions is uncertain. The present work evaluates the arsenic leaching behaviour of an
excavated hornfels and shale rock using modified batch leaching tests. The results show
that dissolved oxygen (DO) concentrations were approximately 3.6 and 6.1 mg/L when
the headspace in the tests was 0 and 500 mL, respectively. Also, the leaching
concentrations of shale rock decrease as the DO decreases.

Keywords:
Batch leaching test, Dissolved oxygen, Oxidation-reduction potential.

Corresponding author mail: [email protected]

1. Introduction
Proper management of soils and rocks with geogenic (natural) contamination has been an
important consideration in geotechnical and geoenvironmental engineering in Japan,
particularly since 2010 (KATSUMI, 2017; TABELIN et al., 2018). In 2010, the Soil
Contamination Countermeasures Law (SCCL), which is the national law related to soil
contamination, was revised to cover natural contamination. Evaluating the leaching
characteristics and behaviour of the soils and rocks of concern is essential for their proper
management. SCCL describes regulatory limits of leaching values for 26 items (9 metals
and metalloids, 12 organic compounds, 5 pesticides and others) and composition values
for 9 items (metals and metalloids) with standardised testing methods. Leaching values
are determined by the batch leaching test prescribed in Japanese Leaching Test Method
No.46 (JLT46).
While JLT46 is applied to assess the environmental impact of the soils and rocks of
concern (as mentioned above), the test may not be scientifically reasonable. Several
discussions have been raised concerning the regulatory test. One crucial issue is that the
soils and rocks, when utilised in an embankment, would exist under anaerobic conditions
(PAWIT et al., 2017; TAMOTO et al., 2008). However, the leaching behaviour under
anaerobic conditions cannot be clarified with the regulatory test. Because leaching
behaviour can change dramatically over time and under different leaching conditions,
tests that more closely represents the actual site are required (TAMOTO &
KURAHASHI, 2017).
The present work seeks to establish a batch leaching test for evaluating the leaching
behaviour under anaerobic conditions. Here, various modified batch leaching tests were
used. First, batch leaching tests using de-aired water. Second, batch leaching tests
with/without shaking and in which headspace is minimal. Headspace is the air space
between the surface of the solution and the cap of the container.

2. Materials and methods

2.1 Materials
Shale rock and hornfels were used in this study and were generated in construction works
such as excavation and tunnelling. The construction works were executed in the
mountainous areas far from industrial and urban areas in Japan. Therefore, considering
the site conditions, the arsenic and other hazardous constituents in the rocks are
considered to be derived through natural phenomena. Hornfels with a particle size of <2
mm (obtained through sieving only) was used for the tests. Shale rock was crushed with
a jaw crusher to particle sizes of <2 mm and used for the tests.
Their particle density was determined as per JIS A 1202:2009, and particle size
distribution as per JIS A 1204:2009. Figure 1 shows the particle size distribution curves
of the rocks after sieving them. For the chemical analysis, the rocks were manually ground
96
to <75 m using an agate mortar and pestle and then analysed by X-ray fluorescence
(XRF) (EDX-720, Shimadzu). Table 1 summarises the physical and chemical properties
of the rocks.

Figure 1. Particle size distribution curves of the sieved rocks.

Table 1. Physical and chemical properties of the rocks.

Hornfels Shale rock
Particle density 2.38 g/cm3 2.73 g/cm3
Natural water content 3.41% 0.06%
Particle size distribution Sand [2.0-0.075 mm] 95.2% 85.5%
Fines [<0.075 mm] 4.8% 14.5%
Chemical composition As: 35.2 mg/kg As: 20.0 mg/kg
SiO2: 38.8% SiO2: 35.3%
Fe2O3: 33.7% CaO: 30.5%
Al2O3: 16.0% Fe2O3: 17.0%
CaO: 1.14% Al2O3: 10.8%
Others: 10.4% Others: 6.4%

2.2 Batch leaching tests

Here, regulatory tests and various modified batch leaching tests were used, as summarised
in Table 2. The modified batch leaching tests include the batch leaching tests using de-
aired water and those with/without shaking and in which headspace is minimal.

2.2.1 Regulatory test

Batch leaching tests corresponding to JLT46 were carried out. Here the distilled water
was poured into a 1000-mL polypropylene bottle. Sieved rock was gently poured into the
water, and mixing done by horizontal shaking at 200 rpm for 6 hours. After the shaking

97
 Table 2. Summary of experimental conditions.
Non-shaking test without
Regulatory test Shaking test with Shaking test Non-shaking test
headspace in a vacuum
(or JLT46) de-aired water without headspace without headspace
deaerator
Liquid-to-solid ratio 10 10 10 10 10

Volume of solution 500 mL 500 mL 570 mL 570 mL 570 mL

Mass of soil 50 g 50 g 57 g 57 g 57 g

Shaking time 200 rpm 200 rpm 200 rpm - -

Leaching duration 6 hours 6 hours 6 hours 6 hours 6 hours

Headspace (volume) 500 mL 500 mL 0 mL 0 mL 0 mL

3,000 rpm 3,000 rpm 3,000 rpm 3,000 rpm 3,000 rpm
Centrifugation
20 minutes 20 minutes 20 minutes 20 minutes 20 minutes
Membrane filter 0.45 m 0.45 m 0.45 m 0.45 m 0.45 m
De-aired De-aired De-aired De-aired
Type of solution Distilled water
distilled water distilled water distilled water distilled water
Hornfels Hornfels
Type of rock Shale rock only Shale rock only Shale rock only
Shale rock Shale rock
The test is The test was carried out in
Remarks - - -
specified in SCCL a vacuum deaerator

98
step, approximately 50 mL supernatant was collected in a centrifuge tube. The
supernatant was centrifuged at 3000 rpm for 20 minutes and then filtered using a 0.45-
m membrane filter. ORP, pH and arsenic concentrations were then measured. Atomic
Absorption Spectrophotometer (Shimadzu, AA-6800) was used for measuring the arsenic
concentrations.

2.2.2 Shaking test with de-aired water

The shaking test with de-aired water is similar to the regulatory test. The only difference
being the distilled water was deaerated before use in the tests. A vacuum desiccator was
used for this purpose. The deaeration step was carried out for ~24 hours. Afterwards, an
appropriate amount of rock specimen was immediately added to the water in the 1000-
mL bottle. The goal was to prevent an increase in O2 concentration in the bottle. After the
completion of the shaking step, the solution’s DO was measured immediately. However,
ORP and arsenic concentration were not measured immediately after the test but rather
after solid-liquid separation because they may not be stable if the turbidity of the solution
is high. Also, if colloidal particles remain in the solution, in that case, arsenic may be re-
sorbed, and its concentration may not be accurately measured.

2.2.3 Batch leaching tests without headspace

Shaking tests without headspace were carried out. De-aired water was used for tests. To
achieve a no-headspace condition, the amount of de-aired water was adjusted so that the
water surface of the solution comes to the mouth surface of the 500-mL bottle when the
rock specimen and water are mixed in the bottle. Because of adjusting the volume of the
solution to unify the liquid-to-solid ratio to 10, the amount of de-aired water was 570 mL.
Therefore, 570-mL de-aired water and 57-g rock specimen were used for the tests.
In addition to the effect of headspace, it was considered that DO might be affected by the
dissolution of O2 inside and outside the container into the solution by shaking. Therefore,
non-shaking tests without headspace were also carried out. Non-shaking tests without
headspace were also conducted in a vacuum desiccator while deaeration was carried out.
This is because there is a possibility that O2 in the air may flow through slight gaps in the
lid during the leaching tests. Chemical analyses were carried out the same way mentioned
in section 2.2.2.

3. Results and discussion

3.1 Relationship between DO and arsenic leaching concentration

Figure 3 shows the relationship between DO and concentrations of arsenic leached from
the rock specimens. The results for regulatory tests and shaking test with de-aired water
show that the DO of solutions after the tests was 5.5–7.0 mg/L. It was challenging to
maintain a low DO just by changing the solvent. It can be seen that the batch test without

99
headspace performed on shale rock can keep DO relatively low even after the test
compared to the result without headspace. For example, comparing the results of the
shaking test with de-aired water and shaking test without headspace showed that
removing headspace could reduce the DO after the test by ~2.5 mg/L. Furthermore,
comparing the shaking test without headspace and the non-shaking test without headspace
showed that shaking increased the DO after the test by ~2.0 mg/L. In addition, comparing
the non-shaking test with the non-shaking test using a vacuum desiccator revealed that
the DO after the test could be reduced by ~0.5 mg/L due to deaeration during the test.
Based on the above results, it is presumed that changes in DO during the test under the
regulatory test include three effects. First, the effects of the gas phase inside the container.
Second, the effects of horizontal shaking. Third, the effects of oxygen outside the
container that flows into the container through a slight gap. Also, regarding the
relationship between DO and arsenic leaching concentration, a clear relationship between
the two was not confirmed in hornfels because the change in DO was slight. However, in
shale rock, it can be seen that the arsenic leaching concentration decreased as DO
decreased. One of the causes is that the decrease in DO suppressed the oxidative
decomposition of pyrite in rocks.

3.2 Relationship between ORP and arsenic leaching concentration

The relationship between pH and arsenic leaching concentration is shown in Figure 3(d).
The ORP of distilled water and deaerated water before the test were respectively 275.4
mV and 227.6 mV. In hornfels, the ORP after the test differed by ~200 mV between when
distilled water was used and when de-aired water was used. On the other hand, in shale
rock, the value was close to 100 mV in all test cases. Except for the test cases of the non-
shaking test with the non-shaking test using vacuum desiccator, which had lower DO than
others, a negative correlation was found between ORP and arsenic leaching
concentration. In addition, the relationship between DO and ORP in this study was
unclear, and the linear relationship between DO and ORP, as seen in the previous study
(MUTO et al., 1986), was not confirmed. The pH was approximately 5.35 to 9.89
regardless of the test method. Hornfels and shale rock were not highly acidic in the
anaerobic leaching test conducted in this study. These results indicate that the pyrite
components of hornfels and shale rock are low. That arsenic may have been leached by
other mechanisms rather than the oxidative decomposition of pyrite.

100
 a) b)

c) d)

Figure 2. Correlations of DO, ORP, pH, and As concentration.

3.3 Maintaining anaerobic conditions

From 3.1, the influence of the gas phase part inside the container, horizontal shaking, and
oxygen inflow from the outside of the container cannot be ignored to maintain the
anaerobic condition. To eliminate the effects of the gas phase part in the container,
performing a non-shaking test with the container lid open in a glove box under deaerating
is conceivable. There is a possibility that the dissolution of oxygen in the headspace and
the inflow of oxygen from the outside of the container may be activated by horizontal
shaking. However, suppose the influence of the headspace is eliminated, the shaking
effect will be relatively small. Also, regarding the effects of the inflow of oxygen from
the outside of the container, a possible method for this might be to apply silicone grease
to the joint between the container lid and the container and perform a batch leaching test.

101
4. Conclusions
In this study, arsenic leaching behaviour was investigated using the regulatory and five
modified batch leaching tests that simulate anaerobic conditions. The results support the
following conclusions:
1) Results of the shaking test using de-aired water show that the DO increases from 0.89
mg/L to 5.5–7.0 mg/L. This finding shows that maintaining anaerobic conditions
during the test is challenging by only changing the solvent.
2) Modified batch leaching tests without headspace were conducted on shale rock and the
DO in three tests was 1.0 to 3.6 mg/L. The results show that tests without headspace
are more effective tests to simulate anaerobic conditions than the shaking tests with
de-aired water.
3) Considering the three types of modified batch leaching tests without headspace, the
increase in DO after the test may be attributed to the effect of headspace, horizontal
shaking and inflow of oxygen from outside the container.
4) While similar value of DO were obtained regardless of the type of rocks, different
value of ORP were obtained depending on the type of rocks.

5. References
KATSUMI T. (2017). Utilization of excavated soils with natural contamination,
Environmental Information Science, Vol. 46, Issue 2, pp. 38–42.
PAWIT T., HAYASHI R., IGARASHI T. (2017). Effects of additional layer(s) on the
mobility of arsenic from hydrothermally altered rock in laboratory column experiments.
Water Air and Soil Pollution, 228, Vol. 5, https://doi.org/10.1007/s11270-017-3295-7
MUTO N., KIM KS (1986). Experimental studies of factors influencing the value of ORP
-Water Temperature, DO, H2S and pH-. Japan Journal of Water Pollution Research, Vol.
9, Issue 10, pp. 661–667, https://doi.org/10.2965/jswe1978.9.661
TABELIN C.B., IGARASHI T., VILLACORTE-TABELIN M. et al. (2018). Arsenic,
selenium, boron, lead, cadmium, copper, and zinc in naturally contaminated rocks: a
review of their sources, modes of enrichment, mechanisms of release, and mitigation
strategies. Science of the Total Environment, Vol. 645, pp. 1522–1553,
https://doi.org/10.1016/j.scitotenv.2018.07.103
TAMOTO S., ITO Y., OKAZAKI K. (2008). The leaching characteristics of arsenic from
rocks (Part 3). Proceedings of the 43rd Symposium on Japanese Geotechnical Society,
Hiroshima, Japan, pp. 559–560 (in Japanese).
TAMOTO S., KURAHASHI T. (2017). Preliminary study on batch leaching test method
in reductive condition. Proceedings of the 23rd Symposium on Soil and Groundwater
Contamination and Remediation, Okinawa, Japan, pp. 2079–2080, (in Japanese).

102
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Oral bio-accessibility of Pb and Zn from playground soils in

Kabwe, Zambia

Walubita MUFALO 1, Pawit TANGVIROON 2, Toshifumi IGARASHI 2,

Mayumi ITO 2, Tsutomu SATO 2, Meki CHIRWA 3, Imasiku NYAMBE 3,
Hokuto NAKATA 4, Shouta NAKAYAMA 4, Mayumi ISHIZUKA 4

1. Graduate School of Engineering, Hokkaido University, Sapporo 060-8628, Japan.

[email protected]
2. Division of Sustainable Resources Engineering, Faculty of Engineering, Hokkaido
University, Sapporo 0608-628, Japan.
[email protected]; [email protected]; [email protected];
[email protected]
3. IWRM Centre/Geology Department, School of Mines, The University of Zambia,
Lusaka 32379, Zambia.
[email protected]; [email protected]
4. Faculty of Veterinary Medicine, Hokkaido University, Kita 18, Nishi 9, Kita-Ku,
Sapporo 060-0818, Japan.
[email protected]; [email protected];
[email protected]

Abstract:
Zambia’s Kabwe Mine Wastes (KMWs) are responsible for heavy metals contamination
of the surrounding soil and dust. Unfortunately, these wastes arise from the historical
mining activities of lead (Pb) and Zinc (Zn), which lacked adequate waste management
strategies. As a result, the dust containing Pb and Zn is easily dispersed to the
surrounding areas by seasonal variation of the wind. To assess human exposure and
potential risks of Pb and Zn, we characterised school playground soils (SPs)(n = 8)
located within a 10 km radius from the old mining area. In vitro bio-accessibility tests
using artificial lysosomal fluid (ALF) and hydrochloric acid (HCl) were combined with
the determination of pseudo-total elemental extraction and particle size distribution. The
total content in the majority of the samples was high with Pb (up to 3,020 mg/kg) and
Zn (up to 1,820 mg/kg). Bio-accessible fractions with ALF recorded up to 54 % for Pb
and 68 % for Zn. Particle size distribution showed that 10 % effective diameter (D10) of
the SPs ranged from 3 - 20 µm, therefore considering accidental ingestion of dust
particles (< 10 µm), substantial amounts of geochemically mobile phases in the soil
could be bio-accessible in the body.
Keywords:
Lead, Zinc, Bio-accessibility, Soil, Dust, Mine wastes.
Corresponding author mail: [email protected]
1. Introduction
Zambia being a low-income country has had severe Pb poisoning in the town of Kabwe.
Kabwe District is the most significant Pb-Zn deposit in Zambia. Mining of Pb-Zn took
place between 1902-1994, thereafter operations of the mine were decommissioned
(SOUTHWOOD et al., 2019). Throughout the mining operations, except in the later
years after 1990, no pollution law regulated the discharge of mine wastes into the open
environment (EPPCA, 1990). Huge stockpiles of mine wastes were disposed of and left
unattended after decommissioning the mines in 1994. Consequently, 74.9 % of the
residents have blood lead levels (BLLs) greater than 5 μg/dL, the standard reference
level for Pb poisoning (YAMADA, et al., 2020)
Because of the long dry season, the dust fallout from the mine wastes easily disperses
and mixes with adjacent soils. Inevitably, soil and water contamination of the heavy
metals in the mine wastes pose a serious environmental health issue. KŘÍBEK et al.,
(2019) reported that substantial amounts of Pb and Zn in the soils surrounding the mine
area. Therefore, it is beyond reasonable doubt that heavy metals as components in acid
mine drainage may leach into the groundwater (TOMIYAMA et al., 2019).
Besides, the Pb and Zn are also bio-accessible through inhalation or accidental ingestion
of the dust from the soil. However, children are the most susceptible to Pb poisoning.
This is because play activities for children involve hand to mouth pathway thus the Pb
in contaminated soil is easily ingested (BELLO et al., 2016). Because of this, one of the
potential exposure sources of Pb intoxication in children is SPs. Due to various play
activities, the contaminated soil/dust can easily be deposited in playgrounds,
subsequently taken up either through the respiratory or digestive system. Long-term
exposure to such an environment may cause memory disintegration, behavioural
changes such as reduced attention span, and increased antisocial behaviour
(AKINWUNMI et al., 2017). Therefore, this study seeks to investigate the contents of
toxic metals, Pb and Zn, in the SPs and their bio-accessible fractions in case of
accidental ingestion or inhalation of the dust in the SPs.

2. Materials and methods

2.1 Sampling location

Kabwe is the provincial headquarters of central Zambia and has a population of at least
202,000 as of the 2010 census. Sampling points were selected approximately within a
radius of 10 km from the mine area. Topsoil samples (about 5 cm deep) with 3-4 mixing
points for each sample were collected from 8 sites; Community ground in Kasanda area
(S1), Makululu Primary School (S2), Malumbo Community ground (S3) Makululu Day
Secondary School (S4), Makululu Day Secondary garden (S5), David Ramushu
Secondary School (S6), David Ramushu Secondary School garden (S7) and Kebar
Christian Academy (S8). These samples were air-dried indoors and then shipped to
104
Hokkaido University, Japan for characterization, sequential extraction, and leaching
experiments.

2.2 Leaching experiments and sequential extraction of Pb and Zn

Leaching tests with artificial lysosomal fluid (ALF) were conducted to extract
bioavailable Pb and Zn from SPs through the alveoli. Three grams of soil were mixed
with 30 mL of the ALF for 6 h. The leachate was separated with 0.45µm membrane
filters. For the total content of Pb and Zn, 1g of soil was digested with 20 ml of 60 %
HNO3 and then brought to boil at 120 °C on a hot sand plate for 1 h. The residue was
then diluted to a 100 ml volumetric flask. To mimic the conditions of extreme case
scenario of accidental ingestion of soil, 3 g of soil samples were extracted with 100 ml
of HCl for 2 h. The concentrations of Pb, Zn and co-existing ions were then analyzed by
inductively coupled plasma atomic emission spectroscopy (ICP-AES) (ICPE-9000,
Shimadzu Corporation, Japan) (margin of error = ± 2~3 %).
The sequential extraction procedure was based on the earlier works of TESSIER et al.,
(1979). The procedure partitioned the target elements into five relevant phases: (1)
exchangeable (2) carbonates (3) Fe-Mn oxides (4) sulfide phases and (5) residue phase.
One gram of sample and pre-determined volumes of the extractants were mixed in 20
mL centrifuge containers using a thermostat water bath shaker. After the pre-determined
extraction time, the suspension was centrifuged at 3,500 rpm for 30 minutes to separate
the leachate from the residue. The leachate and washing solution, which was also
collected by centrifugation decantation, were mixed, diluted to 50 mL, filtered through
0.45 μm Millex® syringe-driven sterile membrane filters (Merck Millipore, USA) and
provided for analysis by ICP-AES.
The chemical and mineralogical properties of the collected soils were measured by X-
ray fluorescence spectroscopy (XRF) and X-ray powder diffraction (XRD),
respectively. The particle size distribution of the soils was measured by the laser
diffraction particle size analyzer.

3. Results and discussion

3.1 Chemical composition, mineralogy and particle size distribution

The soil samples were classified as either sandy loam (S1, S5, S6, and S7), silt loam
(S3, S4, and S8), or loam (S2) (Table 1). The chemical composition of the soil was
mainly comprised of SiO2, (56.4˗83.7 wt. %), Al2O3 (6.4˗29.7 wt. %), and Fe2O3
(7.6˗2.2 wt. %). Figure 1 shows the XRD patterns of the SPs. Quartz (SiO2) also known
as silica is the major mineral in all SPs (Figure 1). Noticeable peaks of kaolinite
(Al2SiO5(OH)) were detected in samples S1, S2, S3, S4 and S5, which might be
attributed to the weathering of the primary silicate minerals from the mine wastes and
this agrees with the clay fractions in the samples (Table 1).
105
 Figure 1. XRD patterns showing the major minerals found in playgrounds.

Table 1. A brief description of the type of soil samples collected: Particle size
distribution into clay, silt and sand according to The United States Department of
Agriculture for the size groups of mineral particles (USDA,1987).

Figure 2 illustrates the particle size distribution of the SPs. The results of the diameter
of the SPs showed that the D10 ranged from 3 to 20 µm. Particles less than 20 µm are
likely to remain in suspension or airborne for a long period (KOK et al., 2012).
Children exposed to such an environment with high Pb and Zn contents in soil are at
high risk of absorbing substantial amounts of Pb in the bloodstream. Moreover, dust
particles that are >10 µm are likely to enter the tracheobronchial region and those < 4
µm are further deposited in the alveolar region of the lung (GUNEY et al., 2017).
106
However, D50 of the SPs was > 20 µm and thus only remain airborne for a short time
(KOK et al., 2012). The play activities that generate substantial dust in the SPs pose a
serious threat to the children because accidental ingestion or inhalation of the dust or
soil is inevitable.

Figure 2. Descriptive simulation of accessible dust in the playgrounds with 10 and 50 %

effective diameter (D10 and D50, respectively).

3.2 Total content of Pb and Zn

The playground soils from S1, S2, S3, S7, and S8 had leaching concentrations of Pb
above the permissible limit for drinking water in Zambia (Pb < 0.01 mg/L). Meanwhile,
Zn was within the tolerable limits for all the SPs (Zn <3 mg/L). The pseudo total
content of the SPs from HNO3 digestion ranged from 102 to 3,020 mg/kg for Pb and
180 to 1,920 mg/kg for Zn. The contents of Pb and Zn were much higher than the
average background content in sub-Saharan Africa; Pb of 37 mg/kg and Zn of 29 mg/kg
(TOWETT et al., 2015). In comparison with the Japanese standard for the quality of the
soil, the SPs in Kabwe are classified as hazardous wastes (Pb > 150 mg/kg). The
profound Pb and Zn in the SPs could be attributed to the deposition of aerosol particles
from historical mining activities. Moreover, high contents of Pb and Zn were recorded
near the mine area.

3.3 In-vitro bio-accessibility tests with ALF and HCl

In-vitro bio-accessibility tests with HCl extraction was compared with the solid phase
partitioning of the SPs. Interestingly, HCl extraction was slightly aggressive in
removing the exchangeable (SE 1), weak acid-soluble (SE 2) and reducible phases (SE
3) of Pb (R2 = 0.98) and Zn (R2 = 0.88) (Figure 3). In case of accidental ingestion into
the digestive system, a substantial amount from the geochemically mobile phases could

107
be absorbed in the stomach. Consequently, substantial Pb could easily enter the
bloodstream. Moreover, children are the most susceptible to Pb poisoning. Because
children have a higher gastrointestinal absorption of Pb than adults (LJUNG et al.,
2006, YABE et al., 2020).

Figure 3. Comparison of single leach extraction with HCl and sequential extraction
phases (SE: 1+2+3).

Meanwhile, bio-accessible fractions with ALF from the geochemically mobile phases
(SE 1+2+3) yielded high bio-accessible fractions;14 to 52 % for Pb and 17 to 77 % for
Zn while the bio-accessible fraction of Pb from the total contents in the SPs was 12 to
43 % for Pb, and Zn was recorded at 12 to 68 % (Table 2). The bio-accessible fractions
were high for Pb and Zn for the geochemically mobile phases. It should be noted that a
higher release with ALF implies that a larger percentage of the inhaled particles could
be adsorbed in the lung environment. This poses a serious threat to children who have
108
access to the SPs. Lead can cause acute and chronic intoxication. Therefore, long term
exposure of the children to the accidental ingestion and inhalation of soil or dust may
cause severe abnormal neurological symptoms, seizures, encephalopathy and finally
death (WORLD HEALTH ORGANISATION, 2019).

Table 2. Bio-accessible fractions (BF) of artificial lysosomal fluid (ALF) for Pb and Zn
compared acid digestion of HNO3 and sequential extraction phases SE (1+2+3).
ALF Total content SE(1+2+3) ALF Total content SE(1+2+3)
BF BF
(Pb) (Pb) (Pb) (Zn) (Zn) BF (Zn) BF
(%) (%)
mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg
S1 871 3020 29 1799 48 646 1920 34 1587 69
S2 122 583 21 316 39 147 853 17 834 60
S3 226 881 26 682 33 59 493 12 561 19
S4 42 102 41 115 36 26 180 14 209 32
S5 102 190 54 238 43 53 279 19 244 56
S6 51 427 12 362 14 99 817 12 1373 17
S7 462 1400 33 968 48 545 1300 42 1184 64
S8 709 1640 43 1368 52 617 904 68 924 77

4. Conclusions
The playgrounds were extremely contaminated with Pb and Zn. Extraction with HCl
was slightly aggressive in removing 3 mobile fractions of sequential extraction;
exchangeable, carbonates and iron/manganese oxides. Hence, substantial amounts of Pb
and Zn could be bio-available in the stomach in case of accidental ingestion of Pb and
Zn. Bio-accessible fractions from ALF digestion for the geochemically mobile phases
recorded up to 52 % for Pb and 77 % for Zn.

5. References
AKINWUNMI F., AKINHANMI T. F., ATOBATELE Z. A., ADEWOLE O.,
ODEKUNLE K., AROGUNDADE L. A., ADEMUYIWA O. (2017). Heavy metal
burdens of public primary school children related to playground soils and classroom
dusts in Ibadan North-West local government area, Nigeria. Environmental Toxicology
and Pharmacology, 49, pp. 21-26. doi: 10.1016/j.etap.2016.11.006
BELLO O., NAIDU R., RAHMAN M. M., LIU Y., DONG Z. (2016). Lead
concentration in the blood of the general population living near a lead-zinc mine site,
Nigeria: Exposure pathways. Science of the Total Environment, 542, pp. 908-914. doi:
10.1016/j.scitotenv.2015.10.143
EPPCA (1990). Environmental Protection and Pollution Control Act, 1990.
Environmental Council of Zambia (ECZ), Act No. 12.

109
GUNEY M., BOURGES C. M. J., CHAPUISR. P., ZAGURY G. J., (2017). Lung
bioaccessibility of As, Cu, Fe, Mn, Ni, Pb, and Zn in fine fraction (< 20 μm) from
contaminated soils and mine tailings. Science of the Total Environment, 579, pp. 378-
386, doi: 10.1016/j.scitotenv.2016.11.086
KOK J. F., PARTELI E. J., MICHAELS T. I., KARAM D. B. (2012). The physics of
wind-blown sand and dust. Reports on Progress in Physics, 75(10), 106901, doi:
10.1088/0034-4885/75/10/106901
KŘÍBEK B., NYAMBE I., MAJER V., KNÉSL I., MIHALJEVIČ M., ETTLER V.,
SRACEK O. (2019). Soil contamination near the Kabwe Pb-Zn smelter in Zambia:
Environmental impacts and remediation measures proposal. Journal of Geochemical
Exploration, 197, pp. 159-173. doi.org/10.1016/j.gexplo.2018.11.018
LJUNG K., SELINUS O., OTABBONG E. (2006). Metals in soils of children's urban
environments in the small northern European city of Uppsala. Science of the Total
Environment, 366(2-3), pp. 749-759, doi: 10.1016/j.scitotenv.2005.09.073
SOUTHWOOD M., CAIRNCROSS B., RUMSEY M. S. (2019). Minerals of the
Kabwe (“Broken Hill”) mine, central province, Zambia. Rocks & Minerals, 94(2),
pp.114-149. doi.org/10.1080/00357529.2019.1530038
TESSIER, A., CAMPBELL, P. G., BISSON, M. J. A. C. (1979). Sequential extraction
procedure for the speciation of particulate trace metals. Analytical Chemistry, 51(7),
pp. 844-851, doi.org/10.1021/ac50043a017
TOMIYAMA S., IGARASHI T., TABELIN C. B., TANGVIROON P., II H., (2019).
Acid mine drainage sources and hydrogeochemistry at the Yatani mine, Yamagata,
Japan: A geochemical and isotopic study. Journal of Contaminant Hydrology, 225,
103502, doi: 10.1016/j.jconhyd.2019.103502
TOWETT E. K., SHEPHERD K. D., TONDOH J. E., WINOWIECKI, L. A.,
LULSEGED T., NYAMBURA M., CADISCH G. (2015). Total elemental composition
of soils in Sub-Saharan Africa and relationship with soil-forming factors. Geoderma
Regional, 5, pp. 157-168, doi.org/10.1016/j.geodrs.2015.06.002
YAMADA D., HIWATARI M., HANGOMA P., NARITA D., MPHUKA C., CHITAH
B., YABE J., NAKAYAMA S.M., NAKATA H., CHOONGO K., ISHIZUKA M.
(2020). Assessing the population-wide exposure to lead pollution in Kabwe, Zambia: An
econometric estimation based on survey data. Scientific Reports, 10(1), pp.1-11.
YABE J., NAKAYAMA S. M., NAKATA H., TOYOMAKI H., YOHANNES Y. B.,
MUZANDU K., KATABA A., ZYAMBO G., HIWATARI M., NARITA D.,
YAMADA D., HANGOMA P., MUNYINDA N. S., MUFUNE T., IKENAKA Y.,
CHOONGO K., ISHIZUKA M. (2020). Current trends of blood lead levels, distribution
patterns and exposure variations among household members in Kabwe,
Zambia. Chemosphere, 243, 125412, doi: 10.1016/j.chemosphere.2019.125412
WORLD HEALTH ORGANIZATION (2019). Lead poisoning, and health.
https://www.who.int/news-room/fact-sheets/detail/lead-poisoning-and-health

110
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Desorption parameters to evaluate arsenic leaching behaviour

Torataro NASAHARA 1, Jiajie TANG 2, Tomohiro KATO 2,

Lincoln Waweru GATHUKA 2, Atsushi TAKAI 2, Takeshi KATSUMI 2

1. Graduate School of Engineering, Kyoto University, Kyotodaigaku-katsura, Nishikyo-

ku 615-8530, Kyoto, Japan.
[email protected]
2. Graduate School of Global Environmental Studies, Kyoto University, Yoshida-
honmachi, Sakyo-ku 606-8501, Kyoto, Japan.
[email protected]; [email protected]
[email protected]; [email protected]
[email protected]

Abstract:
Leaching of toxic metals and metalloids from geogenic contaminated soils may be
attributed to two different mechanisms—intraparticle diffusion and desorption from the
surface of the soil particles. However, these two mechanisms are not clarified when
evaluating the leaching behaviour. In this study, batch leaching tests were conducted
using different liquid-to-solid ratios (L/S) with laboratory-prepared artificially
contaminated soils. Artificially contaminated soils were prepared by soaking ‘clean’ soils
in the arsenic solutions. The arsenic solutions had concentrations of 5, 10, or 50 mg/L.
Comparing the adsorption mass of artificially contaminated soils and the desorption
parameters obtained from batch leaching tests, 85-95% arsenic was considered
immobilised.

Keywords:
Geogenic contaminated soil, Adsorption, Partition coefficient, Arsenic.

Corresponding author mail: [email protected]

1. Introduction
Leaching of toxic metals and metalloids from geogenic contaminated soils is attributed
to two different mechanisms—intraparticle diffusion and desorption from the surface of
the soil particles (SAKANAKURA & ITO, 2018). Intraparticle diffusion is a mechanism
in which contaminants move inside the solid phase and changes as time proceeds.
Desorption is a mechanism on the surface of a solid phase and is defined together with a
paired adsorption. Desorption means that once adsorbed, substance molecules become
uniform again. However, both mechanisms―intraparticle diffusion and desorption―are
not clarified when evaluating leaching characteristics.
Desorption parameters are proposed to model desorption by distinguishing surface
leaching from intraparticle diffusion (ISHIMORI et al., 2020). This model assumes that
desorption occurs instantaneously between the contaminants adsorbed on the particle
surface and the contaminants present in the liquid phase. These parameters are described
by Figure 1 and Equation (1):

Figure 1. Concept of ML, q, and MT.

MT = ML + q (1)
where: MT (mg/kg) is the total mass of contaminants per mass of the solid participating
in desorption equilibrium, ML (mg/kg) is the mass of contaminants present in the liquid
phase per mass of the solid, and q (mg/kg) is the adsorption mass. Values of q can be
calculated with Equation (2) by assuming Henry type ad/desorption isotherm:
q = −Kd C (2)
where: C (mg/L) is the concentration of contaminants in the liquid phase, and Kd (L/kg)
is the partition coefficient, and represents the equilibrium constant. From Equation (1)
and (2), ML is expressed as Equation (3) and Figure 2, assuming Henry type ad/desorption
isotherm.

112
M L = −Kd C + M T (3)

Figure 2. Relationship of desorption parameters assuming Henry type.

Batch leaching tests with different liquid-to-solid ratios (L/S) are proposed to obtain these
parameters (SAKANAKURA & ITO, 2018). Data sets of C and ML can be obtained by
conducting the tests and a regression line determined to obtain MT, as shown in Figure 2
(ITO et al., 2011; SEKITANI et al., 2018; OKAMOTO et al., 2019).
In this study, batch leaching tests with different L/S were conducted using artificially
contaminated soils. From the tests, desorption parameters were obtained and used to
evaluate the arsenic leaching behaviour of the soil. A comparison between the desorption
parameters and adsorption mass was made.

2. Materials and methods

2.1 Preparation of artificially contaminated soil

Decomposed granite soil with a particle size of 0.075–2 mm was used in this study.
Arsenic (As) solutions were prepared using a reagent of sodium arsenite (NaAsO2) and
used for contaminating the soils. Artificially contaminated soils were prepared by mixing
500-g soil with a 500-mL arsenic solution with a concentration of 1, 10, or 50 mg/L in a
1000-mL plastic container. The mixture was horizontally shaken at 200 rpm for 24 hours
using a mechanical shaker (TAITEC, TS-10). After the shaking step, the solution and the
soils were separated by filtering the specimens using a nylon filter net of 0.075-mm
opening (ASONE, 5-3441-10, N-No.200HD). The collected filtrate was centrifuged at
3000 rpm for 10 minutes. The supernatant was filtered using membrane filters with a pore
size of 0.22-mm. Afterwards, the As concentration was measured with an Inductively
Coupled Spectrophotometer―ICP (Agilent Technologies, 700 Series ICP-OES) ―and
Atomic Absorption Spectrophotometer―AAS (SHIMADZU, AA-6800). The remained
solid-water mixtures were placed on a plastic tray with a thin layer (~1 cm) and oven-
113
dried at 110°C for ~18 hours. Completely dried samples were used as the artificially
contaminated soils. Herein the expressions A-1, A-10 and A-50 refer to the artificially
contaminated soils with concentrations of 1, 10, 50 mg/L, respectively.

2.2 Batch leaching test

Batch leaching tests were conducted in triplicate using the artificially contaminated soils
with the conditions summarised in Table 1. The soils and distilled water were put into a
plastic container and shaken horizontally at 200 rpm for 6 hours using a mechanical
shaker. Then the supernatant in the container was collected and centrifuged at 3000 rpm
for 10 minutes. After centrifugation, the supernatant was filtered using a membrane filter
with a pore size of 0.22 mm. Arsenic concentration was then measured using ICP and
AAS.

Table 1. Batch leaching tests with different L/S.

L/S (L/kg) 1 5 10 20 50 100
Soil’s weight (g) 50 10 5 5 5 5
Liquid volume (mL) 50 50 50 100 250 500
Container capacity (mL) 50 50 50 250 500 1000

3. Results and discussion

3.1 Adsorption mass of artificially contaminated soil

Table 2 sumarises the obtained value of adsorption mass, q, and adsorption rate of arsenic
in A-1, A-10, and A-50. Values of q were calculated using Equation (4):

Table 2. Adsorption mass in artificially contaminated soil.

Case A-1 A-10 A-50
Initial concentration of liquid phase: C0 (mg/L) 0.80 9.43 56.4
Equilibrium concentration of liquid phase: C (mg/L) 0.01 0.27 4.40
Adsorption mass: q (mg/kg) 0.79 9.16 52.0
Adsorption rate (%) 98.7 97.1 92.2

q=
( C0 − C )V (4)
M
where: C0 (mg/L) is the initial concentration of the solution, C (mg/L) is the equilibrium
concentration of solution after shaking step, V (L) is the volume of the solution, and M
(kg) is the mass of soil. It was confirmed that the lower the initial concentration of the
solution, the higher the adsorption rate. In each case, the adsorption rate was 90% or more,
which means most of the arsenic in the solution was adsorbed by the soil.
3.2 Results of Batch leaching test
114
Figure 3 shows the relationship between the L/S and average leaching concentrations of
the artificially contaminated soils: A-1, A-10, and A-50. It was confirmed that the
leaching concentration of A-50 at low L/S was remarkably high considering the ratio of
the adsorption mass of A-1 and A-10. Therefore, suggesting that there is more soluble As
in A-50 than in A-1 and A-10. The relationship between the leaching concentration, C,
and the leaching mass calculated from the leaching concentration, ML, is shown in Figure.
4, 5 and 6 for each case. In each case, the leaching concentration and the leaching mass
were found to have a negative correlation. In A-1 and A-10, the leaching concentration
was low at L/S 1, and the leaching mass was small at L/S 100. This suggests that at low
L/S, the leaching of As was suppressed because other dissolved ions saturated the liquid
phase. At high L/S, the leaching amount was calculated low by dilution (ITO et al., 2017).

Figure 3. Average leaching concentration. Figure 4. Leaching amount of A-1.

Figure 5.Leaching amount of A-10. Figure 6. Leaching amount of A-50.

115
By linearly approximating each plot assuming the Henry type isotherm, two desorption
parameters were obtained: the partition coefficient, Kd, from the negative gradient and the
total mass involved in desorption, MT, from the intercept. Table 3 shows the obtained
value of Kd and MT with the adsorption mass. From Tables 3, the change in Kd was small
for the changes in q. Therefore, suggesting no apparent correlation between the adsorption
mass and the partition coefficient in desorption. On the other hand, MT increased as q
increased. Since MT represents the total mass of desorbed contaminants, it is assumed that
the value will increase as the adsorption mass increases. Comparing q and MT in each
case confirmed that the ratio of MT to q was 5-15%. This suggests that most of the arsenic
adsorbed on the soil (85-95%) was immobilised on the surface of soil particles.

Table 3. Desorption parameters.

Case A-1 A-10 A-50
Adsorption mass: q (mg/kg) 0.79 9.16 52.0
Partition coefficient: Kd (L/kg) 2.34 2.92 3.55
Total mass involved in desorption: MT (mg/kg) 0.04 0.44 5.81

4. Conclusions
In this study, the leaching behaviour of arsenic was evaluated by preparing artificially
contaminated soils and conducting batch leaching tests with different L/S. Desorption
parameters proposed by ISHIMORI et al. (2020) to model the surface leaching separately
from intraparticle diffusion are discussed. The findings of this study support the following
conclusions:
1) Desorption parameters―Kd and MT―of artificially contaminated soils can be obtained
using batch leaching tests with different L/S.
2) The total mass desorbed, MT, increased as the adsorption mass, q, increased.
3) Not clear correlation is confirmed between the adsorption mass, q, and the partition
coefficient, Kd, in desorption.
4) Approximately 85–95% of arsenic adsorbed on decomposed granite soil was insoluble.
In addition, the phenomenon that the leaching mass becomes inconsistent at the low and
high L/S confirmed in the previous studies was also confirmed in this study using
artificially contaminated soil. It is suggested that to model the leaching behaviour by
batch leaching test. It is necessary to give priority to this issue, which is considered to
influence the value of the desorption parameters.

5. References
ISHIMORI H., TANG J., NAKAGAWA M., SAKANAKURA H. (2020). Effects of
nonlinear adsorption/desorption mathematical models on chemical concentration
profiles of leachate from tsunami deposits: simplified calculation method for advection-

116
dispersion analysis. Japanese Geotechnical Journal, Vol. 15, Issue 3, pp. 497–508.
https://doi.org/10.3208/jgs.15.497
ITO K., HIRATA T., KITABATAKE Y., MIYAGUCHI S., KAWANO K., YOKOTA
H. (2011). Risk assessment and characteristics of arsenic contaminated soil of Nichinan
group in Miyazaki. Proceedings of the Ninth National Symposium on Ground
Improvement, pp. 181–188.
ITO K., TAKAHASHI H., SAEKI T., OKAMOTO N., NODA N. (2017). Understanding
the concentration and the total amount of heavy metal leaching from soil using soil batch
leaching test. Twenty-third Symposium on Soil and Groundwater Contamination and
Remediation, pp. 289–294.
OKAMOTO N., HOSOBORI K., NODA N., ITO K. (2019). Study on application of
liquid-solid ratio batch test. Proceedings of the Thirteenth National Symposium on
Ground Improvement, pp. 199–206.
SAKANAKURA H., ITO K. (2018). Calculation method of adsorption/desorption
parameters for pollutants-holding materials by liquid-solid ratio batch test. Fifty-third
Japan National Conference on Geotechnical Engineering, pp. 2167–2168.
SEKITANI M., SAKANAKURA H., SUZUKI T., INUI T. (2018). Evaluation of
adsorption and desorption behavior of toxic substances from secondary contaminated
tsunami deposit. Fifty-third Japan National Conference on Geotechnical Engineering, pp.
2177–2178.

117
118
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Serial batch tests and up-flow column tests to evaluate the

sorption performance of soil amended with a stabilising agent

Takaomi OKADA 1, Tomohiro KATO 2, Lincoln Waweru GATHUKA 2,

Atsushi TAKAI 2, Takeshi KATSUMI 2

1. Kyoto University, Graduate School of Engineering, Kyotodaigaku-katsura, Nishikyo-

ku, Kyoto, 615-8530, Japan.
[email protected]
2. Kyoto University, Graduate School of Global Environmental Studies, Yoshida-
honmachi, Sakyo-ku, Kyoto, 606-8501, Japan.
[email protected]; [email protected];
[email protected]; [email protected]

Abstract:
For preventing contamination of adjacent ground where excavated soils and/or rocks
containing geogenic contaminants are utilised, there have been researches focused on
installing an attenuation layer between the contaminated materials and the ground. This
study focuses on using soil amended with a MgO-based stabilising agent as the
attenuation layer material. Up-flow column sorption tests, which can simulate in situ
groundwater flow, were conducted after compaction of the mixtures with 1, 5 and 10%
agent contents, using fluoride solutions. Serial batch sorption tests were also conducted
to evaluate the possibility of predicting the results of the column sorption test. In the
column sorption tests, breakthrough (C/C0 > 0.05) occurred after an approximately liquid-
to-solid ratio of 0.6, 8 and 17 when the agent content was 1, 5, and 10%, respectively.
Increasing the agent content delayed breakthrough. The attenuation layer's design based
on the sorbed mass obtained from the single batch tests could be on the safe side.
However, based on the serial batch tests, it could be on the dangerous side depending on
the shaking time.

Keywords:
Fluoride, Geogenic contamination, Magnesium oxide, Attenuation layer.

Corresponding author mail: [email protected]

1. Introduction
To reduce the disposal of soils, excavation of new materials, carbon footprint, etc.,
reutilisation of soils and rocks excavated at construction sites is highly encouraged (e.g.,
MAGNUSSON et al., 2019). A certain percentage of these excavated materials contain
toxic geogenic chemicals such as arsenic (As), fluorine (F), and lead (Pb) (NAKA et al.,
2016; TABELIN et al., 2018; TAMOTO et al., 2015). However, sometimes toxic
chemicals are leached in concentrations that slightly exceed the mandated limits under
the Soil Contamination Countermeasures Law (SCCL) (ITO & KATSUMI, 2020) which
is a Japanese law enacted by the Ministry of the Environment in 2002 to take
countermeasures against soil contamination to prevent human health damage. Also,
significantly large volumes of excavated soils and rocks are generated at construction
sites (MINISTRY OF LAND, INFRASTRUCTURE, AND TRANSPORT, 2010).
Therefore, implementing cost-effective countermeasures is desirable to utilise these
contaminated materials properly (KATSUMI, 2015).
One possible countermeasure to prevent the contamination of adjacent ground where
naturally contaminated soils and rocks are utilised is the attenuation layer method.
Benefits of this method are its low material costs and reduced management efforts
(TATSUHARA et al., 2012, 2015; NOZAKI et al., 2013a). To prevent (or reduce to
acceptable levels) infiltration of toxic chemicals from these contaminated materials into
adjacent ground, an attenuation layer is installed on the embankment foundation, as
shown in Figure 1 (MO et al., 2020; NOZAKI et al., 2013b; TABELIN et al., 2013). A
typical attenuation layer material is clean sandy soil mixed with a stabilising agent.
The sorption performance should be evaluated for the attenuation layer. Column sorption
test is used to evaluate the sorption performance under much closer in-field conditions
(TATSUHARA et al., 2012). The advantage of this test is that it can simulate in situ
groundwater flow and monitor the temporal changes in the effluent concentration.
However, the column tests take a long testing time and has a relatively complex method
compared to the batch tests. (WATANABE et al., 2009; KIDA et al., 2003).

Figure 1. Schematic of the conventional design of the attenuation layer method.

120
Serial batch tests can evaluate sorption performance more easily than the column tests. In
the serial batch tests, the concentration profiles at certain intervals of liquid to solid ratio
(L/S) can be obtained by exchanging the solution and reputation of the shaking. The
advantage of this test is its simple testing protocol and shorter testing time compared to
the column tests. The serial batch tests were used in the previous studies to evaluate the
long-term leaching behaviour of the waste (PETER et al., 2010; WATANABE et al.,
2009; YAMAOKA et al., 2006; KIDA et al., 2003). However, very few studies have been
applied to evaluate the sorption performance of the stabilising agent.
In this study, three types of sorption tests were conducted to discuss the sorption
performance of soil amended with a stabilising agent. Column sorption tests and single
batch sorption tests were conducted. Also, the applicability of the serial batch sorption
tests was investigated to compare the column tests. Sorption behaviours were obtained
from each test up to L/S 50. Fluoride (F-) was selected as the target contaminant because
it is a geogenic contaminant whose concentration exceeds the acceptable limit regulated
under SCCL (ITO & KATSUMI, 2020; NAKA et al., 2016). Magnesium oxide (MgO)-
based stabilising agent used because it is a promising agent for the attenuation layer
(NOZAKI et al., 2013b).

2. Methodologies

2.1 Materials
Silica sand #7 was used as the clean parent material. Table 1 shows its physical properties.
The stabilising agent was manufactured by Taiheiyo Cement. The main component was
MgO. It had a Blaine specific surface area of 5970 cm2/g, determined as per JIS R 5201
(2015). The residue on a 90-μm sieve was 21.2 wt.%. It had a particle density of 3.21
g/cm3, determined as per JIS A 5201 (2015).
In the sorption tests, specimens had agent contents of 1, 5, or 10%. Agent contents of 1,
5, and 10% corresponded to 10 g-, 50 g-, and 100 g-stabilising agent per kg-dry soil,
respectively. In the experiment, silica sand was poured into a steel bowl. Then the
appropriate amount of stabilising agent was added. Finally, the sand and stabilising agent
were manually mixed with care to prepare homogenous mixtures.

Table 1. Physical properties of the silica sand.

Parameter Standard Value
Particle density JIS A 1202 (2009) 2.62 g/cm3
Particle size distribution JIS A 1204 (2009)
Sand [2–0.075 mm] 98%
Clay [<0.075 mm] 2%

121
2.2 Hydration kinetics
Understanding the hydration kinetics is important to evaluate the sorption performance
of MgO. In these tests, 0.6 g agent was put in contact with 30 mL distilled water using
plastic tubes. The specimens stood for a given time. After hydration tests, solid and liquid
were separated using centrifugation. The pH values of the liquid were measured using a
pH/EC meter (F-54, Horiba). The X-ray diffraction (XRD) profiles with CuKα, 40 kV,
100 mA (RINT-2500, Rigaku in GSJ-Lab, AIST) were measured after 1, 4, and 27 days.

2.3 Column sorption tests

Column tests evaluated the sorption behaviours of the specimens. Column tests were
conducted using acrylic columns ( 5 cm × h 10 cm) at room temperature of ~20 ºC. A
dried soil-agent mixture was placed in the column. Each specimen was compacted in the
acrylic column in five layers with equal heights. Dry density of the specimen was ~1.4
g/cm3. During compaction, a 125-g rammer was dropped freely from a height of 20 cm.
This method corresponds to ISO 21268-3 (2019). The specimen was placed between filter
papers to prevent channel clogging due to the fine soil particles. Then distilled water was
percolated in an up-flow direction using a peristaltic pump at a flow rate of ~36 mL/h
until the specimen reached saturation. Finally, percolation was interrupted for 24 hours
to achieve a saturated condition.
The influent was a fluoride solution prepared using sodium fluoride (NaF) with a
concentration, C0 = 80 mg/L F-. The solution was continuously percolated in an up-flow
direction via a peristaltic pump at 36 mL/h, equivalent to a Darcian velocity of 44 cm/d.
Influents were permeated by L/S 50. Effluents were collected periodically in plastic
bottles and filtered using a 0.45-m membrane filter. The filtrate pH was measured. The
fluoride concentration was measured using a fluoride selective electrode (6561S-10C,
Horiba).

2.4 Single batch sorption tests

Batch tests were conducted by referring to the guideline in SCCL (MINISTRY OF THE
ENVIRONMENT, 2019). The solution of C0 = 80 mg/L F- was prepared. The 240 mL F-
solution and the 4.8 g soil-agent mixture were put in the plastic bottles (L/S 50). The
specimen was stirred at a shaking speed of 200 rpm, a shaking width of 2 to 4 cm, and a
shaking time of 24 hours using a horizontal shaker (NR-30, TAITEC). After shaking, the
bottles were stood for 30 minutes and centrifuged at the rotation speed of 3000 rpm for
20 minutes. The collected solutions were filtered using a 0.45-m membrane filter. The
fluoride concentration of the effluent was measured using ion chromatography device
(ICS-1500, Nippon Dionex).

122
2.5 Serial batch sorption tests
Serial batch tests were conducted to investigate the applicability of predicting the results
of the column tests. The tests started from L/S 10, and the L/S was increased by 10 until
the cumulative L/S reached 50.
The 80 mg/L fluoride solution, silica sand and stabilising agent were prepared. The
polypropylene tube with a capacity of 50 mL was used as a container. The 30 mL solution
and the 3.0 g soil-agent mixture were poured into the container. Using a rotary shaker
(MX-RL-Pro, AS ONE), the specimen was stirred at a rotation speed of 60 rpm, a turning
radius of 8 cm, and a shaking time of 4.8 or 21 hours. After the shaking, the specimen
was stood for 30 minutes and centrifuged at the centrifugal acceleration of 2300 g for 90
minutes. In serial batch tests, higher acceleration was applied because not to lose the mass
of the stabilising agent. After centrifugation, the new 30 mL fluoride solution was added
to the specimen while the soil-agent mixture remained. Same procedures were repeated
five times. Here, five solutions are collected in one specimen, and cumulative L/S reached
50. The collected solutions were filtered using a 0.45-m membrane filter. The fluoride
concentration of the filtrate was measured using ion chromatography.
As shown in Figure 2, it was impossible to separate all solution from the soil-agent
mixture using syringes. In each step, the solution that remained in the specimen was about
2–3 mL. Therefore, in the second to fifth shaking step, batch shaking tests were conducted
with a solution volume of 32–33 mL. In this study, the extra volume (2–3 mL) was
considered to be negligible because the extra volume was very small.

2.6 Calculation of cumulative sorbed mass

For column tests, the cumulative fluoride sorbed mass per unit mass of soil-agent mixture,
S1 was calculated as
𝑛
(𝐶0 − 𝐶𝑠 )𝑉𝑠
𝑆1 = ∑ (1)
𝑚
𝑠=1
where: C0 is the initial solution concentration (80 mg/L F-), m (kg) is the mass of soil-
agent mixture, Cs (mg/L) is the effluent concentration and Vs (L) is the volume of effluent.

Figure 2. Image of exchanging liquid phase in the serial batch sorption test.

123
For single batch tests, the sorbed fluoride mass per unit mass of mixture at L/S 50, S2 was
calculated as
(𝐶0 − 𝐶)𝑉
𝑆2 = (2)
𝑚
where: C is the equilibrium concentration, V is the volume of the solution and m is the
mass of soil-agent mixture.
For serial batch tests, the fluoride concentration was obtained from each L/S. The
cumulative fluoride sorbed mass per unit mass of soil-agent mixture, S3 was calculated as
𝑘
(𝐶0 − 𝐶𝑖 )𝑉
𝑆3 = ∑ (3)
𝑚
𝑖=1
where: Ci is the equilibrium concentration at each L/S. For example, k = 5 shows the
cumulative sorbed mass up to L/S 50 is calculated.

3. Results and discussion

3.1 Hydration kinetics

The hydration kinetics of the stabilising agent was investigated based on Section 2.2.
Figure 3 shows the XRD patterns. Before hydration, MgO peaks were significant.
However, during hydration, MgO peaks become smaller, and magnesium hydroxide
[Mg(OH)2] was predominant. Even after 27 days of contact with water, the peak of MgO
was confirmed. It is considered that the progress of the hydration reaction for MgO to
change to Mg(OH)2 is relatively slow, and the production rate of Mg(OH)2 varies.
Although this test did not consider the chemical reactions occurring in the column tests
and batch tests, it can reveal the general trend of hydration kinetics. After 1, 4, and 27
days, the pH values were 11.1, 11.7, and 9.89, respectively.

Figure 3. XRD patterns of the stabilising agent during hydration test.

124
3.2 Sorption behaviour of column tests
Figure 4 shows breakthrough curves obtained from the column tests. The changes in
fluoride concentration of effluents, C, were evaluated with respect to the L/S. L/S was
calculated by dividing the cumulative volume of the effluent collected during the test by
the mass of the soil-agent mixture in the specimen. Previous research defined the
breakthrough point as the concentration in the effluents where C exceeds 5% of C0, that
is, C/C0 = 0.05 (CHEN et al., 2011; TOR et al., 2009). The breakthrough occurred after
approximately L/S 0.6, 8 and 17 when using agent contents of 1, 5 and 10%, respectively.
These results suggest that increasing the agent content can delay breakthrough.
Comparing the shapes of the breakthrough curves, the larger the agent content, the lower
the fluoride concentration of effluents, which means the larger the sorbed mass.
The pH profiles were obtained from the column tests, and all effluents were alkaline, as
shown in Figure 4. For an agent content of 1%, the pH decreased from 11.4 to 10.4, and
for an agent content of 5 and 10%, the pH decreased from 11.8 to 11.2 and 11.5,
respectively. The difference in pH profiles is due to the CaCO3 content, as shown in
Figure 3.

a) 1 12 b) 1 12

0.8 0.8

Column
0.6 21-h serial batch 0.6
C/C0

C/C0

4.8-h serial batch

pH

pH
11 11
pH
0.4 0.4
Column
0.2 0.2 21-h serial batch
4.8-h serial batch
pH
0 10 0 10
0 10 20 30 40 50 60 0 10 20 30 40 50 60
L/S L/S

c) 1 12

0.8
Column
21-h serial batch
0.6 4.8-h serial batch
C/C0

pH

pH 11
0.4

0.2

0 10
0 10 20 30 40 50 60
L/S

Figure 4. Sorption behaviors obtained from column tests and serial batch tests for a)
1%, b) 5% and c) 10% agent content, and pH profiles obtained from column tests.

125
3.3 Comparison of sorption behaviours
Sorption behaviours between the column tests and serial batch tests were compared to
evaluate the possibility of predicting the sorption behaviour of the column tests. Figure 4
shows the profiles of fluoride concentration obtained from the column tests, the 4.8-h and
21-h serial batch tests.
The sorption behaviours of the 4.8-h serial batch tests and the column tests were
compared. In the case of all agent contents, the sorption behaviour from the 4.8-h serial
batch tests were flat. Thus, particularly in the small L/S, the difference between the
fluoride concentrations of both tests tended to be large.
The sorption behaviours of the 21-h serial batch tests and the column tests were compared.
As shown in Figure 4 a), for an agent content of 1%, C/C0 obtained from the 21-h serial
batch tests was lower than that from the column tests at L/S 10. As the L/S increased, the
C/C0 of both tests become closer, but this is expected because the effluent concentration
almost reaches the initial concentration. As shown in Figure 4 b) and c), for an agent 5%
and 10%, C/C0 obtained from the 21-h serial batch tests were higher than that from the
column tests at L/S 10. However, C/C0 obtained from the 21-h serial batch tests were
lower than that from the column tests at L/S 50. These two sorption behaviours obtained
from both tests are different.
The sorption behaviours of the 4.8-h and 21-h serial batch tests were compared. As shown
in Figure 4, the difference of C/C0 was the largest when the L/S was 10 for all agent
contents. Then, as the L/S increased, C/C0 obtained from both tests tended to become
closer. Sorption performance differs depending on the shaking time.
The possibility of predicting the column tests results from the serial batch tests was
verified. However, the results did not match under the test conditions in this study. The
reason cannot be elucidated in this study. It is considered that the sorption performance
of MgO in the stabilising agent changes over time due to the test method, such as the
presence or absence of pre-saturation step and length of shaking time. Because the
progress of the hydration reaction of MgO is slow, as discussed in section 3.1. To predict
the result of the column tests by conducting the serial batch tests using stabilising agent
containing MgO as the main component, the influence of the test method should be
carefully investigated.

3.4 Applicability of batch sorption tests

Figure 5 shows the cumulative sorbed mass calculated from the results of each test. For
all agent contents, the cumulative sorbed mass in the 4.8-h serial batch tests tended to
increase linearly as the L/S increased. Since it was lower than in other tests, the time
required for sorption was not sufficient with shaking for 4.8 hours at each L/S. The rate
of increase in the cumulative sorbed mass in the 21-h serial batch tests were larger than
the column tests for all agent contents. Notably, the cumulative sorbed mass at L/S 50 in

126
the 21-h serial batch tests was 33% higher for 1% agent content, 6.6% for 5% agent
content, and 9.6% higher for 10% agent content than that in the column tests.

a) 1 b) 3
Cumulative sorbed mass (mg/g)

Cumulative sorbed mass (mg/g)

2.5
0.8

2
0.6
1.5
0.4
1
Column Column
0.2 21-h serial batch 21-h serial batch
0.5
4.8-h serial batch 4.8-h serial batch
0 0
0 10 20 30 40 50 60 0 10 20 30 40 50 60
L/S L/S
c) 4
Cumulative sorbed mass (mg/g)

Column
1
21-h serial batch
4.8-h serial batch
0
0 10 20 30 40 50 60
L/S
Figure 5. Cumulative sorbed mass obtained from column tests and serial batch tests for
a) 1%, b) 5% and c) 10% agent content.

4
Cumulative sorbed mass (mg/g)

Column
1 21-h serial batch
4.8-h serial batch
Single batch
0
0 2 4 6 8 10
Agent content (%)
Figure 6. Cumulative sorbed mass at L/S 50 obtained from column tests, serial batch
tests and single batch tests.

127
For the design of the attenuation layer, the agent content of the stabilising agent is mainly
determined based on the sorbed mass obtained from the single batch tests. However,
different sorbed mass was obtained from the several methods as shown in Figure 5. When
evaluating the long-term sorption performance, if the short testing time such as a single
batch tests is used, the result could differ from the sorption performance exhibited on site.
Thus, it is necessary to compare the sorbed mass obtained from the single batch tests, the
serial batch tests, and the column tests.
The cumulative sorbed mass up to L/S 50 calculated from each test result was compared
as shown in Figure 6. For all agent contents, the cumulative sorbed mass in the order of
21-h serial batch, column, 4.8-h serial batch, and single batch tests. Thus, the design of
the attenuation layer based on the sorbed mass obtained from the single batch tests could
be on the safe side. Because the sorption performance by MgO of the agent was not
sufficiently exhibited in the single batch tests with shaking once for 24 hours. On the
other hand, since the cumulative sorbed mass from the 21-h serial batch tests exceeded
that from the column tests, it could be on the dangerous side. Sorption performance
evaluation considering the hydration of MgO is a future work for the attenuation layer.

4. Conclusions
This study evaluated the sorption performance as the attenuation layer using clean sandy
soil amended with a stabilising agent by column tests, single batch tests and serial batch
tests. Fluoride was considered a target geogenic chemical, and MgO-based stabilising
agent was used. The results support the following conclusions.
1) In the column sorption tests, breakthrough (C/C0 > 0.05) occurred after
approximately L/S 0.6, 8, and 17 when the agent content was 1%, 5%, and 10%,
respectively. Increasing the agent content delayed breakthrough.
2) For an agent content of 1%, 5% and 10%, the cumulative sorbed mass up to L/S 50
increased in the order of 21-h serial batch tests, column tests, 4.8-h serial batch tests,
and single batch tests. it was found that the design of the attenuation layer based on
the sorbed mass obtained from the single batch tests is on the safe side. However,
when based on the serial batch tests, it could be on the dangerous side.
3) The results of the column tests have possibility to be predicted using the serial batch
tests, while the results did not match under the test conditions in this study.
4) According to the XRD patterns, even after 27 days of contact with distilled water,
the peak of MgO was confirmed. Therefore, the progress of the hydration reaction
for MgO to change to Mg(OH)2 is slow, and the production rate of Mg(OH)2 varies.
5) To predict the result of the column tests applying the serial batch tests using MgO-
based stabilising agent, the hydration of MgO should be carefully investigated.

5. Acknowledgement
The authors sincerely appreciate Taiheiyo Cement for providing the stabilising agent used
128
in this study. The authors also acknowledge the support provided by the JSPS KAKENHI
(Grant Nos. 18H04141).

6. References
CHEN N., ZHANG Z., C. FENG M. LI R. CHEN N. SUGIURA (2011). Investigations
on the batch and fixed-bed column performance of fluoride adsorption by Kanuma mud.
Desalination, Vol. 268, pp. 76-82, https://doi.org/10.1016/j.desal.2010.09.053
ITO H., KATSUMI T. (2020). Leaching characteristics of naturally derived toxic
elements from soils in the western Osaka area: considerations from the analytical results
under the Soil Contamination Countermeasures Act. Japanese Geotechnical Journal, Vol.
15, pp. 119-130, https://doi.org/10.3208/jgs.15.119
KATSUMI T. (2015). Soil excavation and reclamation in civil engineering:
environmental aspects. Soil Science and Plant Nutrition Vol. 61, pp. 22-29,
https://doi.org/10.1080/00380768.2015.1020506
KIDA A., OSAKO M., SAKAI S. (2003). Evaluation of soil and groundwater
contamination from molten slag generated from solid waste by column tests.
Environmental science, Vol. 16, pp. 497-516, https://doi.org/10.11353/sesj1988.16.497
MAGUSSON S., JOHANSSON M., FROSTH S., LUNDBERG K. (2019). Coordinating
soil and rock material in urban construction—scenario analysis of material flows and
greenhouse gas emissions. Journal of Cleaner Production, Vol. 241,
https://doi.org/10.1016/j.jclepro.2019.118236
MINISTRY OF LAND, INFRASTRUCTURE AND TRANSPORT, JAPAN (2010).
Technical manual on the countermeasure against soils and rocks containing naturally
occurring heavy metals in construction works, (Draft).
http://www.mlit.go.jp/sogoseisaku/region/recycle/recyclehou/manual/index.htm
MINISTRY OF THE ENVIRONMENT, JAPAN (2019). Soil contamination
countermeasures law guidelines, Vol. 3, pp. 283-291,
https://www.env.go.jp/water/dojo/gl-man/dojogl2019_3.pdf
MO J., FLORES G., INUI T., KATSUMI T. (2020). Hydraulic and sorption
performances of soil amended with calcium-magnesium composite powder against
natural contamination of arsenic. Soils and Foundations, Vol. 60, pp. 1084-1096.
https://doi.org/10.1016/j.sandf.2020.05.007
NAKA A., YASUTAKA T., SAKANAKURA H., KALBE U., WATANABE Y.,
INOBA S., TAEKO M., INUI T., KATSUMI T., FUJIKAWA T., SATO K.,
HIGASHINO K., SOMEYA M. (2016). Column percolation test for contaminated soils:
key factors for standardisation. Journal of Hazardous Materials, Vol. 320, pp. 326-340.
https://doi.org/10.1016/j.jhazmat.2016.08.046
NOZAKI F., SHIMIZU Y., ITO K. (2013a). Discussion on construction method of heavy
metals adsorption layer. In: Proceedings of the 19th Symposium on Groundwater and
Soil Contamination and Countermeasures, Kyoto, Japan.
129
NOZAKI T., MATSUYAMA Y., SUGIYAMA A., MORIYA M., KOMUKAI Y.,
NAGASE T. (2013b). Fundamental study of soil materials for absorbent construction
method using MgO material. In: Proceedings of the 19th Symposium on Groundwater
and Soil Contamination and Countermeasures, Kyoto, Japan.
PETER A., ANN G., ALEXANDER S., STEVEN C. (2010). Serial batch leaching
procedure for characterization of coal fly ash. Journal of the Environmental Monitoring
and Assessment, Vol. 168, pp. 523-545. https://doi.org/10.1007/s10661-009-1132-1
TABELIN C.B., IGARASHI T., YONEDA T., TAMAMURA S. (2013). Utilisation of
natural and artificial adsorbents in the mitigation of arsenic leached from hydrothermally
altered rock. Engineering Geology, Vol. 156, pp. 58-67,
https://doi.org/10.1016/j.enggeo.2013.02.001
TABELIN C.B., IGARASHI T., VILLACORTE-TABELIN M., PARK I., OPISO E.M.,
ITO M., HIROYOSHI N. (2018). Arsenic, selenium, boron, lead, cadmium, copper, and
zinc in naturally contaminated rocks: a review of their sources, modes of enrichment,
mechanisms of release, and mitigation strategies. Science of the Total Environment, Vol.
645, pp. 1522-1553, https://doi.org/10.1016/j.scitotenv.2018.07.103
TAMOTO S., TABELIN C.B., IGARASHI T., ITO M., HIROYOSHI N. (2015). Short
and long term release mechanisms of arsenic, selenium and boron from a tunnel-
excavated sedimentary rock under in situ conditions. Journal of Contaminant Hydrology,
Vol. 175-176, pp. 60-71, https://doi.org/10.1016/j.jconhyd.2015.01.003
TOR A., DANAOGLUA N., ARSLAN G., CENGELOGLU Y. (2009). Removal of
fluoride from water by using granular red mud: Batch and column studies. Journal of
Hazardous Materials, Vol. 164, pp. 271-278, https://doi.org/10.1016/j.jhazmat.2008.08.011
TATSUHARA T., ARIMA T., IGARASHI T., TABELIN C.B. (2012). Combined
neutralisation–adsorption system for the disposal of hydrothermally altered excavated
rock producing acidic leachate with hazardous elements. Engineering Geology, Vol. 139,
pp. 76-84, https://doi.org/10.1016/j.enggeo.2012.04.006
TATSUHARA T., JIKIHARA, S., TATSUMI, T., IGARASHI, T. (2015). Effects of the
layout of adsorption layer on immobilising arsenic leached from excavated rocks.
Japanese Geotechnical Journal, Vol. 10, pp. 635-640, https://doi.org/10.3208/jgs.10.635
WATANABE T., KOMINE H., YASUHARA K., MURAKAMI S., TOYODA K.
(2009). Environmental impact of aluminum leaching on drinking water sludge.
Proceedings of Japan Society of Civil Engineers, Vol. 65, pp. 188-201,
https://doi.org/10.2208/jscejg.65.188
YAMAOKA S., MORISHITA T., OHTSUBO M., HIGASHI T., HAGIO T. (2006).
Leachability of heavy metals and soluble ions in the incineration ash by serial batch tests.
Transactions of the Agricultural Engineering Society, Japan, Vol. 246, pp. 41-47,
https://doi.org/10.11408/jsidre1965.2006.849

130
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Removal of heavy metals using tire derived activated carbon

(TAC) vs commercial activated carbon (CAC)

Rahim SHAHROKHI 1, Junboum PARK 1

1. Department of Civil and Environmental Engineering, Seoul National University, Seoul,

South Korea.
[email protected]

Abstract:
In this study, a supreme activated carbon was derived from pulverized waste tires by
carbonization and chemical activation techniques in sequence. Adsorption-desorption batch
experiments were performed to evaluate the efficiency of the Tire derived Activated Carbon
(TAC) in comparison of commercial activated carbon (CAC) as a reference material for
removing three heavy metals ions (pb2+, Cu2+ and Zn+2) from aqueous solution. The TAC
showed a great potential to adsorb heavy metals with a monolayer capacity as high as
322.5,185.2, and 71.9 mg. g-1 for Pb, Cu and Zn, respectively. These values were much higher
than deduced uptake capacities for CAC with 42.55, 15.01, and 14.0 mg. g-1 for Pb, Cu and
Zn, receptively. Sorption kinetics data, with better fit to the pseudo-second order kinetics
model, revealed that TAC had faster sorption rate in comparison to CAC. Also, the adsorption
– desorption tests revealed the adsorption capacities of TAC and CAC were reduced to half
of their initial capacities after three successive adsorption-desorption cycles.

Keywords: TAC, CAC, Adsorption, Heavy metals removal.

Corresponding author mail: [email protected]

1. Introduction
Exponential growth of industrial, agricultural, and mining activities has significantly
increased the amounts of inorganic and organic contaminants, including heavy metals, being
discharged into surface water and groundwater. For instance, in the Republic of Korea
(hereafter, Korea), 15.3% of industrial areas release notable levels of heavy metals, among
which lead (Pb), copper (Cu), and zinc (Zn) are reported as highly persistent pollutant metals
in the groundwater (MINISTRY OF ENVIRONMENT KOREA, 2017). These heavy metals
are poisonous, nonbiologically degradable, and cause significant threats to the ecosystem and
human health by accumulating in the groundwater (CHOWDHURY et al., 2017). Therefore,
inadvertent discharge of heavy metals is a pressing environmental issue that must be
remediated before releasing any type of contaminated water back into the ecosystem.
Adsorption is considered to be the most effective physicochemical technique for heavy metal
removal due to the simplicity of operation, cost-effectiveness, and regenerative nature of the
adsorbents (BAIMENOV et al., 2020). Activated carbon has distinctive characteristics (high
pore volume and large surface area) that make it a valuable adsorbent for heavy metal
removal (DO, 1998). Although numerous kinds of carbonaceous materials can be used for
making activated carbon, charcoal is commonly used in industrial manufacturing (PRAHAS
et al., 2008) to make commercial activated carbon, which, despite its ideal adsorption
efficiency, requires high investment costs for large scale applications (SHAHROKHI-
SHAHRAKI et al., 2020). Therefore, increased attention has been aimed at finding
affordable and unconventional precursors, such as agricultural and industrial wastes, for the
production of efficient activated carbons (LI et al., 2019).
Worldwide production and distribution of waste tires has been steadily increasing over the
past two decades, and one billion tires were produced worldwide in 2014 (SEKHAR, 2014).
Due to a lack of sufficient landfills and the nondegradable nature of tires, disposal of waste
tires has become a worldwide environmental concern (ABBASPOUR et al., 2019). Pyrolysis
has recently been considered the state of the art recycling method for dealing with waste tires,
wherein waste tires are converted to char, hydrophobic liquids, and volatile fractions
(ACOSTA et al., 2016). Tire pyrolysis char is an extremely carbonaceous material and can
be used as a precursor to make activated carbon for heavy metal adsorption (SALEH &
GUPTA, 2014), which is an economical way to reduce waste tire disposal challenges.
Therefore, in this study, the efficiency of tire-derived activated carbon (TAC) produced by
sequential carbonization and chemical activation techniques was evaluated in comparison
with that of commercial activated carbon (CAC) for removing heavy metals. A series of batch
adsorption–desorption experiments using TAC and CAC samples were carried out for
removing three heavy metals (Pb, Cu, Zn) from aqueous solutions.

132
2. Materials and methods
2.1. Preparation of Tire-derived Activated Carbon (TAC)
TAC was synthesized using a two-step sequence: carbonization and chemical activation.
First, the PWT material was washed using DI water and dried at ambient laboratory
temperature for 24 h. Then, the carbonization step was initiated by heating the dried PWT in
a tube furnace (OTF-1200X) to 570 ºC at a rate of 30 ºC∙min-1. The PWT was held at 570 °C
for 100 min in a N2 environment, with nitrogen gas flow in the furnace at a rate of 200
mL∙min-1 (ACOSTA et al., 2016). The subsequent chemical activation step was performed
by physically mixing the carbonized PWT with KOH at ratio of 1 to 6 (this impregnation
ratio was selected based on the work of ACOSTA et al. (2016)). To enhance the chemical
reaction, the mixture was diluted with DI water, mixed for 1 h and soaked overnight.
Afterwards, the dried mixture was placed into the tube furnace and heated at a rate of 3º
C∙min-1 to 750 ºC and was maintained for 60 min in a N2 environment with a gas flow of 200
mL∙min -1 (ACOSTA et al., 2016). After cooling the activated sample under nitrogen flow,
it was washed with 1 N HCl and then rinsed using DI water to remove any remaining KOH.
Finally, the product was dried at 105 ºC for 24 h and passed through a 75-µm sieve (U.S.
standard sieve NO. 200). The final product was denoted as TAC and tested as an adsorbent
for heavy metals.

2.2. Isotherm and kinetic batch adsorption tests

The adsorption capacity of TAC was determined by performing batch adsorption tests in
accordance with the EPA/530-SW-87-006-F standard (ROY et al., 1986). Batch samples
were prepared by adding 0.2 g of TAC to 50 mL of each of the heavy metal ion solutions at
the desired concentration. Samples were mechanically agitated at 180 rpm at ambient
laboratory temperature (20 ± 2 °C) for 24 h. After agitation, the samples were centrifuged at
3900 rpm for 20 min and then filtered (pore size 0.45 µm). The equilibrium concentrations
of heavy metal ions in the supernatant were determined by atomic absorption spectroscopy
(Analytik Jena AG, DE/novAA 300). The same batch adsorption method was also performed
for CAC, which acted as the reference. Since the used CAC had a larger granular size than
TAC, a higher weight of CAC was used for the batch adsorption tests (2.5 g CAC in 50 mL
solution). To check the reproducibility, each test specimen was prepared in triplicate.
The kinetics of adsorption was investigated for the following contact times of 1, 3, 6, 12, 24,
48, 72 and 96 h at initial concentration of 500 mg∙L-1 for each ion (Pb, Cu, and Zn).
The sorption capacities of the TAC and CAC samples were calculated by subtracting their
initial and equilibrium concentrations using the following formula:

133
 (𝐶𝑖 −𝐶𝑒 )𝑉
𝑞𝑡 = (1)
𝑚

where qt (mg∙g-1) is the amount of each metal ion adsorbed by TAC or CAC, Ci (mg∙L-1) and
Ce (mg∙L-1) are the initial and equilibrium concentration of the adsorbate, respectively, V (L)
is the volume of the metal ion solution, and m (g) is the dry mass of the TAC or CAC.

2.3. Reusability tests

To evaluate the possibility of recovering heavy metals from the loaded TAC and CAC
surfaces as well as their reusability for subsequent use, one-step desorption experiments were
performed at the end of the sorption experiments. After batch adsorption tests, the supernatant
was decanted from each sample, and the metal-loaded samples were dried at 60 °C for 24 h.
A total of 50 mL of 0.01 M HCl (pH ~ 2) was added to the dried samples, which were placed
on a shaker for 24 h at 180 rpm and then centrifuged at 3900 rpm for 20 min to desorbed the
loaded ions on the sorbents. At the end of one cycle of desorption tests, the exhausted sorbents
with heavy metals with initial concentration of 500 mg∙L-1 were used to assess the reusability
potential of the sorbents. For this aim, after finishing the first desorption cycle, the next
adsorption cycle began by drying the used sorbents in an oven at 60 ºC for 24 h. The used
and dried adsorbent material was placed into a fresh solution containing 500 mg∙L-1 of all
three metals. The adsorption-desorption tests were continued for three cycles. Each test was
performed twice and the mean was calculated.

3. Results and discussion

3.1. Isotherm and kinetic adsorption tests
Figure 1 (a-b) shows adsorption curves for TAC and CAC as a function of equilibrium
concentration versus adsorbed amount of heavy metals. For all three examined heavy metals,
the adsorption capacity increased with increasing equilibrium concentration of metals.
However, at higher concentrations, the adsorption rate reached a relatively constant value.
The experimental adsorption data for CAC and TAC samples were fit to the Langmuir and
Freundlich, adsorption isotherm models, and the model parameters are presented in Table 1.
Comparison of the coefficients of determination (R2) values for adsorption isotherm models
of TAC shows that the Langmuir model fit the experimental data better than other three
models. Whereas, for CAC the Freundlich model fit the data the best.

134
 Figure 1. (a-b) Heavy metals adsorption by TAC and CAC samples.

For TAC, the maximum adsorption capacities estimated by the Langmuir isotherm were
322.5, 185.2 and 71.9 mg∙g-1 for lead, copper, and zinc, respectively. When converted to
mass per surface area (qmax/BET surface area), the calculated values are 3.9, 2.2 and 0.87
mg∙m-2 for lead, copper and zinc, respectively. For CAC, the maximum adsorption capacities
were 42.5, 15.0, and 14.0 mg∙g-1 for lead, copper and zinc, respectively. Which correspond
to mass per surface area values of 0.034, 0.012 and 0.011 mg∙m-2 for lead, copper, and zinc,
respectively. It can be observed that for the three examined heavy metals (Pb, Cu, Zn), TAC
displayed a higher adsorption capacity than CAC.

Table 1. Langmuir and Freundlich isotherm parameters for Pb, Cu and Zn sorption by
TAC and CAC samples.
TAC CAC
Isotherm model
Pb Cu Zn Pb Cu Zn
Langmuir model qmax
(mg∙g-1) 322.50 185.20 71.90 42.50 15.0 14.0
KL (L∙mg-1) 0.015 0.005 0.052 0.008 0.003 0.001
R2 0.978 0.975 0.992 0.962 0.971 0.908
Freundlich model
KF (mg∙g-1) 78.50 20.20 21.80 3.87 2.31 1.38
1/n 0.19 0.27 0.15 0.31 0.21 0.34
R2 0.748 0.906 0.840 0.977 0.987 0.982

From the kinetic adsorption results (Figure 2 (a-b)) as the adsorption time increased, the
adsorption capacity was increased until the adsorption reached to the equilibrium phase.
However, the adsorbents sites were almost occupied with the metal ions at the early hours of
contacting time. The TAC sample was saturated faster than CAC, and its equilibrium sorption
was almost achieved within 6 h, while the sorption equilibrium of the CAC was reached after
24 h. Also, the analyzed data with pseudo first-order and pseudo second-order models,
showed that the pseudo second-order model was the best fit for adsorption data.
135
 Figure 2. (a-b) Kinetics of heavy metals adsorption by TAC and CAC samples.

3.2. Reusability results

The reusability results (Figure 3) showed that after three adsorption-desorption cycles the
adsorption capacities of both TAC and CAC sorbents were reduced as compared to their
initial maximum adsorption capacities. This can be due to poor recovery of loaded ions
during desorption process, which decreases the available active adsorption sites on the
sorbents surface. However, after three successive adsorption-desorption cycles, the TAC and
CAC samples retained almost half of their initial adsorption capacities (especially for Cu and
Zn ions), which confirms that both the TAC and CAC have potential for renewal and
subsequent reuse as adsorbent materials.
140
120 First Cycle Second Cycle Third Cycle
100
qt (mg g-1)

80
60
40
20
0
Pb-loaded Pb-loaded Cu-loaded Cu-loaded Zn-loaded Zn-loaded
TAC CAC TAC CAC TAC CAC
Ion-loaded sorbent

Figure 3. TAC and CAC adsorption capacities for heavy metals after three successive
adsorption-desorption cycles.

136
4. Conclusion
In this study, tire- derived activated carbon (TAC) and commercial activated carbon (CAC)
were compared in term of their adsorption/desorption capacities for select heavy metals. Key
findings from this study are as follows:
- Adsorption capacity of TAC was much more than CAC (almost 8 times higher).
- The adsorption sequence of the examined heavy metals was Pb > Cu > Zn.
- The kinetic sorption results showed that TAC had faster adsorption rate compared to CAC.
- For both TAC and CAC sorbents, the maximum adsorption capacities of the three ions
under competitive condition were reduced compared to those under single sorption.
- The loaded ions on TAC and CAC surfaces could not be completely recovered using 0.01 M
HCl, and after three successive adsorption-desorption cycles, the maximum adsorption
capacities of both TAC and CAC were reduced to approximately half of their initial values.

5. Acknowledgment
This research is supporting by the National Research Foundation of Korea NRF
2021R1A2C2011180, and also Bk21 program of Seoul National University.

6. References
ABBASPOUR M., AFLAKI E., NEJAD F.M. (2019). Reuse of waste tire textile fibers as
soil reinforcement. J. Clean Prod. 207, pp. 1059–1071. https://doi.org/10.1016/j.jclepro.2018.09.253.
ACOSTA R., FIERRO V., DE YUSO A.M., NABARLATZ D., CELZARD A. (2016).
Tetracycline adsorption onto activated carbons produced by KOH activation of tyre
pyrolysis char. Chemosphere 149, pp. 168–176. https://doi.org/10.1016/j.chemosphere.2016.01.093.
BAIMENOV A., BERILLO D.A., POULOPOULOS S.G., INGLEZAKIS V.J. (2020). A
review of cryogels synthesis, characterization and applications on the removal of heavy
metals from aqueous solutions. Adv. Colloid Interface 276, 102088.
https://doi.org/10.1016/j.cis.2019.102088
CHOWDHURY P., ATHAPATHTHU S., ELKAMEL A., RAY A.K. (2017). Visible-solar-
light-driven photo reduction and removal of cadmium ion with Eosin Y-sensitized TiO2 in
aqueous solution of triethanolamine. Sep. Purif. Technol. 174, pp. 109–115.
https://doi.org/10.1016/j.seppur.2016.10.011
DO D.D. (1998). Adsorption analysis: equilibria and kinetics. Imperial College press
London, https://doi.org/10.1142/p111
LI L., ZOU D., XIAO Z., ZENG X., ZHANG L., JIANG L., WANG A., GE D., ZHANG G.,
LIU F. (2019). Biochar as a sorbent for emerging contaminants enables improvements in

137
waste management and sustainable resource use. J. Clean. Prod. 210, pp. 1324–1342.
https://doi.org/10.1016/j.jclepro.2018.11.087.
MINISTRY OF ENVIRONMENT KOREA. (2017). Soil environmental conservation act.
Gwacheon: Ministry of Environment. https://elaw.klri.re.kr/eng_service/lawView. do?hseq=46236&lang=ENG
PRAHAS D., KARTIKA Y., INDRASWATI N., ISMADJI S. (2008). Activated carbon from
jackfruit peel waste by H3PO4 chemical activation: Pore structure and surface chemistry
characterization. Chem. Eng. J.140, pp. 32–42, https://doi.org/10.1016/j.cej.2007.08.032.
ROY W.R., KRAPAC I.G., CHOU S.F.J., GRIFFIN R.A. (1986). Batch-type adsorption
procedures for estimating soil attenuation of chemicals. U.S. Environmental Protection
Agency, Washington, DC.
SALEH T.A., GUPTA V.K. (2014). Processing methods, characteristics and adsorption
behavior of tire derived carbons: a review. Advances in Colloid and Interface Science, 211,
pp. 93–101. https://doi.org/10.1016/j.cis.2014.06.006
SEKHAR G. (2014). Proceedings of the Tire Technology Expo. Cologne, Germany.
SHAHROKHI-SHAHRAKI R., KWON P.S., PARK J., O’KELLY B.C., REZANIA S.
(2020). BTEX and heavy metals removal using pulverized waste tires in engineered fill
materials. Chemosphere 242, 125281. https://doi.org/10.1016/j.chemosphere.2019.125281

138
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Arsenic stabilization by oyster shell, zeolite and the mixture of

them as binders evaluated in an As-contaminated soil and an
As-aqueous solution
Cecilia TORRES QUIROZ 1, Junboum PARK 1

1. Seoul National University, Department of Civil and Environmental Engineering, 1

Gwanak-ro, Gwanak-gu, Seoul 08826, Korea
[email protected]; [email protected]

Abstract:
Arsenic (As) is well known for its biological toxicity against any living organisms. It has
also been classified as a carcinogen by the United States Environmental Protection
Agency. Thus, soil contaminated with As is a hazard for humans because it can be
bioaccumulated in plants and enter the body through food ingestion. The stabilization–
solidification technique involves adding an absorbent material called a binder for
contaminants to attach to it. In this research, oyster shell (OS), and synthetic zeolite 4A
(Z), were evaluated to determine its capacity to treat As. First, the binders were applied
in the As-contaminated soil and for 2 months, where changes in the As concentration was
controlled by Mehlich-3 extraction (Pot cultivation test). Then the As-removal capacity
in As-aqueous solution was described by sorption test results at different As
concentrations. As a result, the As immobilization efficiency defined as the difference
between final concentration and initial concentration in percentage reported that OS only
reached 20%, Z was 60% and the combination of them (OS/Z) was 50%. In an Aqueous
solution, where the absorbent was directly in contact with As, using a binder-to-liquid
ratio of 1:5, both only reduced 15% of the initial concentration. In addition, Freundlich
isotherms fitted suggesting that the adsorption behaviour of arsenic was single-layered on
the surface for both binders.

Keywords:
As-contaminated soil, Oyster shell, Zeolite, Immobilization efficiency, Sorption.

1. Introduction
Arsenic, like lead, copper, zinc, and cadmium, is classified as a potentially toxic metal
because it can lead to cancer once it enters the human body. Mining, industries, and
pesticides or herbicides in agricultural soils are cited as anthropogenic sources of arsenic
pollution (KAHAKACHCHI et al., 2004). It is also awarded that it can occur naturally in
the soil in the range of 1-40 mg/kg. Regulations regarding toxic trace elements limit As-

Corresponding author mail: [email protected]

concentration’s presence to 5.0 mg/L or 100mg/kg for Hazardous waste in US. However,
this limitation can vary from one country to another. For instance, agricultural soils in
Canada should be under 12 mg/kg and for Australia, and countries in the EU, is 20 mg/kg
(CHEN et al., 2018; LIU et al., 2018). South Korea established a value limit of 1.5 mg/L
(EHTI, 2017).
The Arsenic presented in the contaminated soils can be removed by using alkaline
solutions through the washing treatment technique (ALAM et al., 2001). However, As
(III) is less efficiently removed than As (V) from aqueous solutions by almost all of the
Arsenic removal technologies. In terms of solidification and stabilisation technique, the
binders seek to immobilise the contaminant (Arsenic), preventing it from spreading by
leaching. Sustainable binders such as oyster shell and zeolite should be evaluated. Oyster
shell (OS), a by-product of shellfish-farming in Korea and containing a high amount of
calcium carbonate CaCO3 (BONNARD et al., 2020; LEE et al., 2008), which has a high
potential to be used as a liming material in agricultural (LEE et al., 2008) or alkaline
amendment (PALANSOORIYA et al., 2020). Meanwhile, Zeolite(Z) corresponds to the
chemical group of hydrated aluminosilicates of Sodium, Calcium, Magnesium,
Potassium, and other alkaline minerals, whose formation is attributable to pyroclastic
deposits. It has a tectosilicate 3D structure, considered as an open structure. Due to this,
Z has voids like tunnels that can loosely hold water molecules, and these tunnels can
adsorb molecules smaller than tunnel diameter (GEORGIANNOU et al., 2017).
In this opportunity, the authors continue evaluating OS (94.5% CaCO3) and Z as
sustainable binders, which previously show good binding performance for toxic metal
such as Pb, Cu, Zn and Cd (TORRES-QUIROZ et al., 2021). This will now be applied to
evaluate their capacity facing Arsenic in As-contaminated soil and As-aqueous solution.
Being a third binder option, the combination of them in equal proportion. This paper
presents the results and discussions from two different tests to evaluate the binder
performance, the sorption test and the pot cultivation test.

2. Materials and methods

2.1 Materials
Soil samples were collected from an industrial zone contaminated with As in Ipjang-
myeon Hongcheon-ri 33-8, Chungcheong Nam-do, Cheonan-si, South Korea; classified
as well-graded sand with Fine Content < 5% (denoted as ‘As-soil’). All the samples were
air-dried and passed through a 2 mm mesh before the preliminary analysis and
experiments.
Binder materials obtained from domestic suppliers; OS (Jisan Industrial Co. Ltd) come
in the shape of small balls for agricultural purpose application, it will be crushed, and the
particle size will be reduced prior to analysis. Furthermore, the synthetic Zeolite 4A (Z),
is an enhanced adsorbent produced chemically (Table 1 and Figure 1).
140
Table 1. Physio-chemical properties of the contaminated soil and binders used in this study.
Specific
Classification Cation Exchange
gravity
Property Chemical Form pH (USCS Capacity Source
(ASTM D
ASTM D 422) (meq/100g)
854)
As-soil from Quartz (67.35%),
Industrial Microcline, Albite, 7.9 SP 2.68 9.58 -
zone Muscovite
Oyster Shell
CaCO3 8.5 SP 2.71 4.81 -
(OS)
Na12[(AlO2)12
Zeolite 4A (Z) 10.87 ML 2.10 311* *Provider
(SiO2)12] 27H2O

Oyster Shell Powder Synthetic Zeolite 4A

OS Z

CaO – 94.17% SiO2 – 71.73%

Al2O3 – 17.55%
Figure 1. Binder characteristics, texture and porosity (by SEM analysis), quantity
analysis by XRF for identification of the main chemical compounds.

2.1.1 XRD analysis

XRD performed to identify the structure of each binder and the new binder from the
mixture of OS/Z was conducted in Chemical and Biological Engineering Research
Facilities – Seoul National University (Smart Lab by Rigaku), with Cu radiation. The
XRD patterns were collected at 2θ values in the range of 5°– 70°, with a 2θ step size of
0.02° and a count time of 4s per step. The qualitative analyses of the XRD patterns were
conducted using the Match! Version 3 (Developed by Crystal Impact launched in
Germany 2021) with reference to the patterns present in the International Centre for
Diffraction Data database (2002).

2.2 Methods

2.2.1 Initial toxic metal concentrations

The concentrations measured in leachate (mg/L) using ICP-OES after conducted TCLP
extraction was converted as a fraction of mass (mg/kg), as it is presented in Table 2.
Additionally, Mehlich-3 extraction method was used to measure the bioavailability of As
and Agua regia extraction to quantify the total As content.
141
Table 2. Initial toxic metal concentrations.
Extraction Initial concentrations
Source
Method Pb Cu Zn Cd As
DI (mg/L) - 0.015 0.019 - 0.101
TCLP (mg/L) - - 0.388 - 0.318
As-soil
TCLP (mg/kg) - 7.760 - 6.360
from an
M-3 (mg/L) 0.445 0.560 0.637
industrial zone
M-3 (mg/kg) 4.45 5.600 6.370
Aqua Regia 40.653 74.454 134.00

2.2.2 Pot cultivation test

Pot cultivation test recreates field conditions such as moisture content (around 30%) and
the lack of sun light into the soil. Control samples and treated soil was preserved under
dark conditions in polypropylene amber wide neck round bottles. The steps are described
as follow:
a) In 14 polypropylene amber wide neck round bottles of 250ml capacity was set up
150g of soil with 30% moisture content. First, 2 bottles for the control samples, and in
the remaining 12 bottles, OS, Z and OS/Z was added in a proportion of 3% and 5% of
the total weight by duplicate. Soil, water and binder were thoroughly mixed.
b) Samples were collected each week starting with time control 0 (t0) immediately after
mixing and then settled in the bottles. The bottles were put down in a dark space and
covered to keep their temperature constant. The collection of samples continues for
seven weeks (49 days).
d) After removing a small sample from each bottle, the wet soil weight was measured,
then samples were carried to the oven for drying them at 60°C for 24hrs. After that,
moisture content was calculated.
e) Finally, bioavailability of As was examined by using Mehlich-3 extraction. In a flask
glass of 100ml capacity, 5g of the sample was collocated in a 50ml (L/S=10) of
extraction solution, later agitated by 5min at 120rpm with control of final pH control
after extraction.

2.2.3 Batch sorption test

Batch sorption experiments were conducted to evaluate the binders in an As-aqueous
solution. It consists of evaluating the adsorption of “As” (solute) from a solution by an
absorbent (binder: OS and Z) at equilibrium. Using As-standard solution of 1000mg/L
(As2O3・NaCl(0.05%)・HCl), authors prepared the solution of As of 250ppm, 100ppm,
50ppm and 5ppm. For determining the adsorption capacity of each binder independently,
50mL of As-solution were set up in 100ml flask glass, and the binder was added at
different binder-to-liquid (S/L) ratio 1:20 (2.5 g), 1:10 (5 g) and 1:5 (10 g). The
suspensions were agitated at 180 rpm for 24 h, centrifuged at 3500 rpm for 20 min, and
142
filtered by a 0.45μm membrane filter. The initial pH value for all suspensions was
measured before and after binder addition. Since the pH level changed during the As
adsorption recorded as final pH using a Thermo Scientific Orion Series – Model Meter
121900 (USA).

3. Results and discussions

3.1 XRD analysis for binders

Conventional XRD has been conducted to analyse the new binder from the physical
mixture of OS/Z. The measurements indicate that the use of OS combined with Zeolite
4A promotes the crystallisation of the zeolite A. OS/Z display a general decrease in peaks
intensity (around half) compared with the XRD pattern of Z. The peaks related to CaCO3
in OS also become reduced after combined it with Z (Figure 2).

Z OS OS/Z Zeolite 4A
CaCO3 (Oyster shell)

0 10 20 30 40 50 60 70
2θ-theta
Figure 2. XRD patterns of commercial zeolite 4A (Zeobuilder 4A), Oster shell (Jisan
Industrial).

3.2 Effects of dosage and time contact in pot cultivation test

We observed that the As-Soil, which is a granular soil, had an initial As-concentration of
0.318 mg/L measured by TCLP test and 0.637 mg/L measured by Mehlich-3 extraction
methods, both results are below the limits from South Korean regulation value for As
concentration in the leachate from waste (<1.5mg/L), or the concentration cited by US
EPA, to be considered as a hazard (<5mg/L) (EHTI, 2017; U.S. EPA, 2011). However,
the Korean regulation also regulates the total arsenic content in soils as a maximum
75mg/kg (EHTI, 2018). The total As content is 134 mg/kg determined by aqua regia
extraction. The initial water content was 30%, and it was controlled for seven weeks
during the pot cultivation test. Then, the moisture content decreases slightly during the
first month (4 weeks) from 30% to 28.2 ± 1.3% in soil samples, to 28.3 ± 1.7% in soil
treated with OS, and to 28.4 ± 1.3% in soil treated with OS/Z, meanwhile, soil treated
with Z reported a slight increase to 31.1± 2.2%. In the remaining time (3 weeks), the

143
moisture content showed a more significant reduction attributed to the evaporation during
the removal of samples. All the removed samples were dried at 70°C for 24h prior
Mehlich-3 extraction test (M-3 solution pH is 2.5). Figure 3 shows the variation of the
pH per time and per case. In general, all the binders increased the pH at least by 1 point
and greater dosage also contributes to increasing the final pH per binder case.
The As-concentration measured in the soil without binder along the period of evaluation
was 0.67 mg/L. After adding OS in dosage of 3% and 5%, it was reduced to 0.55mg/L
and 0.48 mg/L. Then, it is reduced to 0.25 mg/L and 0.24 mg/L when 3 and 5% of Z were
used. The use of the OS/Z does not represent a significant improvement in the binding
process. The As concentration is reduced to 0.32 mg/L and 0.28 mg/L when 3 and 5% of
OS/Z were added. As Figure 4 displays, at the final of the period, OS immobilized only
20% of As; meanwhile, Z immobilized 60%. The combination of OS and Z showed
similar results than Z used alone; it immobilized around 50%.

pH after M- 3 extraction As concentration

4,5 S 0,8
4,0 +OS3 0,6
+OS5
C (mg/L)

3,5
pH

+OS/Z3 0,4
3,0 +OS/Z5
+Z3 0,2
2,5
+Z5 0
2,0
0 7 14 21 28 35 42 49 0 5 10 15 20 25 30 35 40 45 50
Days Days
Figure 3. pH after Mehlich-3 extraction and As concentration after adding binder.

Binder immobilization capacity

100% +OS3
80% +OS5
60% +OS/Z3
40% +OS/Z5
20%
+Z3
0%
+Z5
0 5 10 15 20 25 30 35 40 45 50
Days
Figure 4. As-immobilization capacity of OS, Z and OS/Z.

3.3 Sorption batch experiments

The most important physiochemical aspects in evaluating the adsorption process are the
kinetics and equilibrium of adsorption. Equilibrium studies give the capacity of the
adsorbent for specific contaminants. For example, Freundlich model is characterized by
sorption that continues as the concentration of sorbate increases in the aqueous phase.
Figure 5 (left) shows that the mass of pollutant sorbed (As) is proportional to the aqueous
phase concentration at low sorbate concentrations and decreases as the sorbate
accumulates on the sorbent surface. The fitting parameters are listed in Table 3. We can
see that all the R2 values were between 0.8819 and 0.9841, revealing that arsenic

144
adsorption process was suitable for modelling by Freundlich isotherm. This indicated that
the adsorption behaviour of arsenic was single-layered on the surface for both binders.
The results also can be expressed as relative final concentration versus the binder-to-
liquid ratio (S/L) used in this bath sorption test, as it is observed in Figure 5 (right); using
a S/L of 1/5 both binders reached only 15% of As removal.

300
OS 0,98
250 Z 0,96
Cs (mg/kg)

200 Freundlich isotherm for OS 0,94

Freundlich isotherm for Z 0,92
150 Langmuir isotherm for OS
100 0,86
0,85
50 OS Z
0
0 100 200 300 1/20 1/10 1/5
Ce (mg/L) Binder-to-liquid (S/L) ratio
Figure 5. Absorption of arsenic by Oyster Shell Powder and Zeolite 4A at room
temperature, described by Freundlich equation (Left). Relative final As concentration
for different S/L ratio. The mean pH of the OS-solute and Z-solute were 8.61±0.10 and
10.42 ±0.04, respectively.

Table 3. Isotherm models, adsorption parameters for OS and Z.

Freundlich isotherm Langmuir isotherm
KF
1/n R2 1/a 1/ab R2
(mg/kg)
Oyster shell 5.287 0.5946 0.9841 0.0051 0.6018 0.8819
Zeolite 4A 0.2371 1.1986 0.9730 - - -

4. Conclusions
As a result, the As immobilization efficiency defined as the difference between final
concentration and initial concentration in percentage reported that OS only reached 20%,
Z was 60% and their combination (OS/Z) was 50%. In an aqueous solution, where the
absorbent was directly in contact with As, using a binder-to-liquid ratio of 1:5 both only
reduced 15% of the initial concentration. In addition, Freundlich isotherms fitted
suggesting that the adsorption behaviour of arsenic was single-layered on the surface for
both binders. Overall, OS and Z are not a suitable binder for As contamination. When
they were evaluated in As-aqueous solution, they reached only 15% of As removal
efficiency and according to the isotherm analysis, the As adsorption occurred on the
single-layered on the surface for both binders. However, the better performance of Z in
the pot cultivation test can be attributed to the high cation exchange capacity, instead of
absorption mechanism.

5. Acknowledgment
This research was supported by the Korea Ministry of Environment as part of the

145
‘Subsurface Environmental Management’ SEM project: Grant No. 2019002470002.

6. References
ALAM M.G.M., TOKUNAGA S., MAEKAWA T. (2001). Extraction of arsenic in a
synthetic arsenic-contaminated soil using phosphate. Chemosphere 43 (8), pp. 1035-
1041, doi.org/10.1016/S0045-6535(00)00205-8
BONNARD M., BOURY B., PARROT I. (2020). Key Insights, Tools, and Future
Prospects on Oyster Shell End-of-Life: A Critical Analysis of Sustainable Solutions.
Environmental Science & Technology 54 (1), pp. 26-38, doi.org/10.1021/acs.est.9b03736
CHEN S.-B., WANG, M., LI S.-S., ZHAO Z.-Q., E W.-D. (2018). Overview on current
criteria for heavy metals and its hint for the revision of soil environmental quality
standards in China. J. of Integ. Agric., 17, pp. 765-774, doi.org/10.1016/S2095-3119(17)61892-6
EHTI (2017). No. 2017-54 Waste Process Test Standard, ES 06002.a Standards, labeling
limits and result indication of hazardous substances contained in designated waste -
2014. Environmental Health Technology Institute, South Korea.
EHTI (2018). No. 2018-53 Standards for Soil Pollution Process, ES 07400.2c Metals-
inductively coupled plasma-atomic emission spectrometry. Environmental Health
Technology Institute, South Korea.
GEORGIANNOU V.N., LEFAS D., KONSTADINOU M., PERRAKI M. (2017).
Geotechnical properties of a natural zeolite. Proceedings of the Institution of Civil
Engineers - Geotechnical Engineering 170, pp. 395-406. doi.org/10.1680/jgeen.16.00157
KAHAKACHCHI C., UDEN P.C., TYSON J.F. (2004). Extraction of arsenic species
from spiked soils and standard reference materials. Analyst 129 (8), pp. 714-718, doi:
10.1039/b406050d
LEE C.H., LEE D.K., ALI M.A., KIM P.J. (2008). Effects of oyster shell on soil chemical
and biological properties and cabbage productivity as a liming materials. Waste
Management 28, pp. 2702-2708, doi: 10.1016/j.wasman.2007.12.005
LIU L., LI W., SONG W., GUO M. (2018). Remediation techniques for heavy metal-
contaminated soils: Principles and applicability. Science of The Total Environment 633,
pp. 206-219, doi: 10.1016/j.scitotenv.2018.03.161
PALANSOORIYA K.N., SHAHEEN S.M., CHEN S.S., TSANG D.C.W.,
HASHIMOTO Y., HOU D., BOLAN N.S., RINKLEBE J., OK Y.S. (2020). Soil
amendments for immobilization of potentially toxic elements in contaminated soils: A
critical review. Environment International 134, 105046, doi.org/10.1016/j.envint.2019.105046
TORRES-QUIROZ C., DISSANAYAKE J., PARK J. (2021). Oyster Shell Powder,
Zeolite and Red Mud as Binders for Immobilising Toxic Metals in Fine Granular
Contaminated Soils (from Industrial Zones in South Korea). International Journal of
Environmental Research and Public Health, 18 (5), doi.org/10.3390/ijerph18052530
U.S. EPA (2011). 40 CFR 261.24 - Toxicity Characteristic, in: Regulations, C.o.F. (Ed.).
U.S. Environmental Protection Agency, United States.
146
REUSE OF SOLID WASTE
&
WASTE ENGINEERING
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Characterization and mechanical proprieties of soil-vegetal

fibre material

Athmane AZIL 1, Tuan Anh PHUNG 1, Malo LE GUERN 1, Nassim SEBAIBI 1

1. COMUE Normandie Université - Laboratoire de Recherche ESITC Caen, France

[email protected]

Abstract:
The construction in cob is an ancestral technique that consists to mix soil, vegetable
fibres and water. This construction technique exists in the northwest of France and the
south of the United Kingdom. This technique enables an optimal using of natural
resources with low environmental impact. The aim of this study is to understand the
influence of soil and vegetal fibre characteristics on the mechanical strength of different
cob formulations. Cob specimens of size 150×150×150 mm, were prepared. Two kinds
of soils were used: sandy-silt soil (F2), and a silty soil of medium plasticity (UK6); also,
two kinds of vegetal fibres: flax straw and reed. These fibres have different physical and
water absorption characteristics. Then, the mechanical strength of cob materials under
simple compression was determined.
The results reveal that the F2 soil-flax straw mix (2%) provides a higher mechanical
strength than the other mix (1.95 MPa). The lowest behavior is obtained for the F2 soil-
flax straw mix (1.45 MPa). The water content and the density are the parameters
influencing the mechanical strength of the cob materials.

Key words: Soil, Vegetal fibre, Cob, Water absorption, Mechanical strength.

1. Introduction
European Buildings consume most of the energy, responsible for 40% of final energy
consumption (DIXIT et al., 2010). The use of local materials, as well as the absence of
hydraulic cement, reduces the consumption of grey energy due to the construction. For a
long time, people have built their houses and cities using earth material, it is one of the
most used materials in construction. We have different construction methods using
earth: daub, rammed earth, compressed earth blocks (CEB), and cob. The cob method is
very common in northwestern of France and southern of the United Kingdom, the cob
construction consists to mix soil, vegetal fibre, and water. It allows a use of natural
resources with low environmental impact (MOREL et al., 2001). The cob material has
been less studied than other earthen construction methods (WEISMANN & BRYCE,
2010; LEBAS et al., 2007) and often confused with these methods. Recent studies

Corresponding author mail: [email protected]

(RÖHLEN & ZIEGERT, 2013; MICCOLI et al., 2014) have presented answers to some
problems about the cob technique, by comparison with other methods of earth
construction. The objective of this study is to investigate the influence of soil and
vegetal fibre characteristics on the compressive strength of the cob material composed
of different mixes.

2. Materials and methods

2.1 Methods of material characterization
For the characterization of the materials, two different soils were selected for this study:
F2 from Normandy (France), and UK6 from England, see Figure 1. The properties of
the soils were determined by particle size analysis (XP P94-041, 1995), the Atterberg
limit test (NF EN ISO 17892-12, 2018), the methylene blue test (NF P94-068, 1998),
and the Proctor test (NF P94-093, 2014). The vegetal fibre used in this study is natural
flax straw and reed as shown in Figure.1. Regarding the characterization of vegetal
fibres, the parameters determined are the absolute density and the water absorption
coefficient.

Figure 1. UK6 and F2 soils (left) - vegetal fibres (right).

2.2 Cob formulation and mechanical behavior

Specimens of size 150×150×150 mm were prepared with different formulations, mixing
F2 soils with 2% and 4% flax straw and reed, the formulations are presented in Table 1.
In a tank, the mix of the soils and fibres was done manually, by adding the necessary
water until a homogeneous mix was obtained. The cubic molds were filled layer by
layer, with manual compaction rod.

Table 1. Formulations of cob mixes

Samples Fibre Water
Mixes Fibre type
number (%) content (%)
F2 A2 6 Flax straw 2 20
F2 R2 6 Reed 2 20
F2 A4 6 Flax straw 4 20
F2 R4 6 Reed 4 20

148
After two weeks of drying in ambient air and humidity (T°C= 20°C± 2 and RH= 50%
±5), the specimens were removed from the molds and the uniaxial compression test was
carried out on the F2 A2% and F2 R2% specimen mixes, using an IGM press with a
capacity of 250 kN. The press is force-controlled, with an imposed loading rate of 0.01
kN/s. The stress-strain curve was obtained using a 20 mm displacement sensor.

3. Results
3.1 Material characterization
The soil F2 has a high fine fraction (≈ 81 %), but Soil UK6 present higher clay activity
than soil F2 (Figure 2 and Tables 2 and 3). The soils used were classified according to
the LCPC-USCS classification, soil F2 is sandy-silt soil, and UK6 is silt soil.

Table 2. Granulometric properties of soils.

Soil Dmax (mm) Passing to 2 mm (%) ∅ < 80μm (%)
F2 31.5 89.13 7.84
UK6 20.0 92.78 81.79

According to proctor test, the maximum dry densities of soils F2 and UK6 range: 1700
to 1800 kg/m3, see Figure 3 and Table 3.

Figure 2. Soil particle size distribution, (PHUNG, 2018).

Table 3. Properties of soils.

Nature of soil VBs WL WP IP WOPN Dry density
Soils
(USCS classification) (g/100g) (%) (%) (%) (%) (kg/m3)
F2 Sandy-silt soil 0,55 31 20,5 10,5 14 1827
UK6 Silt soil 2,62 30,2 23 7,2 18 1708

149
 Figure 3. Proctor curves, (PHUNG, 2018).

The natural flax straw and reed fibres were used in the mixes (see Figure 1). The raw
fibres were cut, within 5 to 10 cm for flax straw and 4 to 6 cm for reed. The water
absorption capacity was performed according to RILEM protocol (AMZIANE et al.,
2017). After 24 hours, Flax straw shows water absorption of 431 % and reed 203 %, see
Figure 4. The water absorption capacity of fibres influences the mix properties at fresh
state (absorption of available water), and at long term (hygrometric balance). The
maximum tensile strength is 112 MPa (PHUNG, 2018). The average absolute density of
flaw straw is 1360 kg/m3 and 1300 kg/m3 for reed.

Figure 4. Absorption coefficient curve of fibres.

3.2 Compressive strength

The results of the compressive strength show that the F2-flax straw soil mixture (2%),
with a yield stress of 1.95 MPa (see Figure 5a), provides a higher mechanical strength

150
than the F2-Reed mixture (2%), which presents an elastic and then plastic behavior, see
Figure 5b.

Figure 5. Stress-strain curve of mixes F2 A2 (a) and F2 R2 (b).

4. Conclusions and perspective

In this paper, the main factor that impacts the mechanical strength of the cob mixes is
the initial water content (20 %). The physical characteristics of the vegetal fibres are
water absorption, and density. The soil type has a role in the connection between the
components of the soil-fibre mixes. The soil mixed with water allows better covering of
the fibre area, creating a great soil-fibre connection. Based on the analysis of the results
obtained from the characterization of the different soils, vegetal fibres, and the
compression strength tests of the F2-flax straw (2%) and F2-reed (2%) mixes, this work
contributed to observe that:
- When the initial water content decreases, the mechanical strength increases.
- The cohesion of the soil-vegetal fibre mixes is related to the clay activity of the soil.
- The flax straw in the soil makes good compressive strength and the reed increases the
ductility of the material.

151
Other mixes will be made and tested with compressive strength, with the UK6 soil and
the same 2% and 4% vegetal fibres.

5. References
AMZIANE S., COLLET F., LAWRENCE M., MAGNIONT C., PICANDET V.,
SONEBI M. (2017). Recommendation of the RILEM TC 236-BBM: characterization
testing of hemp shiv to determine the initial water content, water absorption, dry
density, particle size distribution and thermal conductivity, Materials and Structures. 50.
https://doi.org/10.1617/s11527-017-1029-3
DIXIT M.K., FERNANDEZ-SOLIS J., LAVY S., CULP C.H. (2010). Identification of
parameters for embodied energy measurement: a literature review, Energy and
Building, 42, pp. 1238–1247, https://doi.org/10.1016/j.enbuild.2010.02.016
LEBAS P., LACHERAY C., PONTVIANNE C., SAVARY X., SCHMIT, P.,
STREIFF, F. (2007). La terre crue en Basse-Normandie, De la matière à la manière de
bâtir, OREP Edition & Communication, 76p., EAN 13 : 9782951884519
MICCOLI L., MULLER U., FONTANA P. (2014). Mechanical behaviour of earthen
materials: A comparison between earth block masonry, rammed earth and cob,
Construction and Building Materials, Volume 61, 30 June 2014, pp. 327-339,
https://doi.org/10.1016/j.conbuildmat.2014.03.009
MOREL J.C., MESBAH A., OGGERO M., WALKER P. (2001). Building houses with
local materials means to drastically reduce the environmental impact of construction,
Building and Environment 36, pp. 1119–1126. https://doi.org/10.1016/S0360-1323(00)00054-8
NF EN ISO 17892-12 (2018). P94-512-12, Geotechnical investigation and testing -
Laboratory testing of soil - Part 12: determination of liquid and plastic limits.
NF P94-051 AFNOR. (1993). Soil: investigation and testing. Determination of
Atterberg's limits. Liquid limit test using Casagrande apparatus. Plastic limit test on
rolled thread.
NF P94-068 AFNOR. (1998). Soils: investigation and testing. Measuring of the
methylene blue adsorption capacity of a rocky soil. Determination of the methylene blue
of à soil by means of the stain test.
NF P94-093 AFNOR. (2014). Soils: investigation and testing - Determination of the
compaction reference values of a soil type - Standard proctor test - Modified proctor test.
PHUNG T. (2018). Formulation and characterization of an earth-plant fibre composite,
PhD thesis, Normandy University, 178p., tel-01938827, version 1
RÖHLEN U., ZIEGERT C. (2013). Construire en terre crue, Ed. Le Moniteur, 332p.,
ISBN 978-2-281-11567-3
WEISMANN A., BRYCE K. (2010). Construire en terre facilement - La technique du
cob, Ed. Eyrolles, 228p., ISSN 978-2-84221-216-2
XP P94-041 AFNOR. (1995). Soil: Investigation and testing. Granulometric
description. Wet sieving method.
152
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Mechanical and leaching characteristics of

carbonated MSW IBA using exhaust gas and CO2
discharged from waste incineration facilities
Takuro FUJIKAWA 1, Kenichi SATO 1, Chikashi KOGA 1,
Hirofumi SAKANAKURA 2, Hiroshi KUBOTA 3, Yosuke NAGAYAMA 4

1. Fukuoka University, Department of Civil Engineering, 8-19-1 Nanakuma, Jonan,

Fukuoka-city, 8140180, Japan.
[email protected]; [email protected]; [email protected]
2. National Institute for Environmental Studies, 16-2, Onogawa, Tsukuba-city, Ibaraki,
3058506, Japan.
[email protected]
3. Fujita Corporation, 2025-1 Ono, Attugi-city, Kanagawa, 2430125, Japan.
[email protected]
4. Fukuoka University Graduate School, Major in Construction Engineering, 8-19-1
Nanakuma, Jonan, Fukuoka, 8140180, Japan.
[email protected]

Abstract:
In Japan, approximatively 4 million tons of Incineration Bottom Ash (IBA) from
municipal solid waste (MSW) are discharged a year. Since Japan doesn’t have enough
space for landfilling, it is necessary to reuse it. However, since bottom ash includes lot
of toxic heavy metals, immobilization is needed for effective use. Considering to these
back grounds, this study is focused on the carbonating treatment which immobilizes the
bottom ash by exhaust gas and carbon dioxide (CO2) generated from Saga city
incineration plant. In this incineration plant, there is equipment that separates and
collects CO2 from the exhaust gas generated when incineration of garbage, this is
Japan's first Carbon dioxide capture and utilization plant at a waste incineration facility.
This paper describes the mechanical characteristics and leaching properties of the
carbonated IBA as a base course material using the CBR test and leaching test. As a
result, it was revealed that the carbonated IBA can be reused as a sub-base course
material. Also, the leaching concentrations of heavy metals from carbonated IBA were
influenced by carbonating treatment processing. It was revealed that the concentration
of lead meets the environmental quality standard values for soils set by the Japanese
Ministry of the Environment.
Keywords:
Incineration bottom ash, Carbonating treatment, CBR, Leaching test.

Corresponding author mail: [email protected]

1. Introduction
In Japan, approximately 4 million tons of incineration bottom ash (IBA) from municipal
solid waste (MSW) is discharged per year. IBA from MSW has to be landfilled at the
final disposal site to protect the environment. There is a serious capacity shortage at
final disposal sites because of limited space. Therefore, alternative uses of IBA,
especially as a construction material, are being sought.
The material characteristics of IBA generated from MSW vary due to regional
characteristics (e.g., type of waste, volume, place, and season). Even though the
mechanical and geotechnical characteristics of IBA have been explored by many
researchers, there are a limited number of studies considering the changing trends of
IBA components based on its decomposition over time. Moreover, since IBA contains
toxic heavy metals, it is necessary to understand not only geotechnical characteristics
but also leaching properties to use it as a ground geo-material. However, the effect of
aging on not only shear strength of IBA but also leaching properties of IBA has been
limited and further studies are needed on the effects of long-term aging to promote
effective utilization. Regarding the immobilization of metals in IBA by natural
precipitation, JIANG et al., (2009) and QUEK et al., (2016) revealed the effectiveness
of washing IBA before utilization and that it is necessary not only to reduce heavy metal
content, but also to improve mechanical properties. Also, regarding the immobilization
of metals in IBA by carbonation and/or weathering, BRENNE & WEIGAND (2015)
show that the aging of bottom ash can be accelerated by enhanced carbonation.
Moreover, KUBOTA et al., (2020) examined the effect of long-term of leaching of IBA
by means of on-site bottom ash stabilization treatment that combined water washing and
CO2. FUJIKAWA et al., (2020) revealed the possibility of effective utilizing IBA by
means of an aging method in the landfill. In addition, geotechnical reuse of carbonated
material may even lead to revenue when conventional construction materials are
substituted by stabilized bottom ash.
Based on the abovementioned background information, this study is focused on a
carbonating treatment that immobilizes bottom ash using exhaust gas (EG) and
recovered carbon dioxide (CO2) generated from the Saga city incineration plant in Japan.
In this incineration plant, equipment is available to separate and collect CO2 from EG
generated when garbage is incinerated. This is Japan's first carbon dioxide capture and
utilization (CCU) plant at a waste incineration facility. In this study, we examined a
carbonating treatment that immobilizes IBA by means of two types of carbon dioxide
from the CCU plant. This paper describes (1) the bearing capacity of IBA from MSW
obtained by the California bearing ratio (CBR) test to investigate the possibility of
usage of IBA as a base course material, and (2) the leaching properties of the IBA from
MSW, obtained by the single batch leaching test (Notifications No. 46 Japanese
Ministry of the Environment), to evaluate the effect of the carbonating treatments.
154
2. Testing procedure

2.1 Carbonating treatment

According to the Saga city official homepage (SAGA CITY, 2020), in the Saga
cleaning plant (garbage incineration facilities), their facility for separation and recovery
of carbon dioxide (CO2 only) from exhaust gas generated during incineration is Japan’s
first CCU plant in a waste incineration facility. Carbon dioxide (CO2), which has been
said to be the cause of global warming, can actually help foster the promotion of raw
materials of carbonated beverages and dry ice and help vegetables and micro-algae
grow as a photosynthesis resource. In Saga, to take advantage of carbon dioxide uses in
vegetable and algae culture, the plant was allowed to operate the carbon dioxide
separation and recovery facilities from August 2016. In this study, we focused on EG
and CO2 and examined a carbonating treatment for immobilizing IBA. Moreover, the
effects of the difference between these two types of gases on geotechnical
characteristics, such as mechanical properties and leaching, of IBA were investigated.
Figure 1 shows the facility flow and gas sampling locations.

Chimney

Exhaust gas CO2 separation

Incineration CO2
pretreatment and recovery
Facility storage tank
equipment equipment

Exhaust gas Recovered CO2

(EG) (CO2)
Figure 1. Facility flow gas sampling locations.

2.2 Materials
The IBA discharged from the Saga incineration facility was used as the experimental
samples. This type of IBA at this facility is dry ash discharged from a stoker- type
furnace. Therefore, to promote reaction with CO2, IBA that was previously mixed with
water to have a water content of 17.6% was used in the experiment. The reason for
setting the IBA water content to 17.6% here is based on a previous study (KUBOTA et
al., 2020). After the mixed IBA was packed loosely (less than  t = 1.0 g/cm3) in the
container shown in Figure 2, two types of gases were ventilated. The lower part of the
container has a hollow structure and pumped gas passes through the IBA and is
discharged to the upper part. The carbonated IBA are defined as EG-IBA and CO2 -
IBA.

155
The processing conditions for each gas are shown in Table 1. Since EG and CO2 have
different carbon dioxide concentrations, the total CO2 amount is set in each condition by
adjusting the ventilation speed and ventilation time. As a comparative sample, IBA
without carbonating treatment was also used. The physical properties of these IBA are
shown in Table 2 Fine contents (Fc) is slightly increased with the carbonation treatment.

IBA
CO2 gas
cylinder

Thermocouple sensor
Container size: 1,200×750×400 mm
(a) Photograph of the equipment. (b) Schematic of the experimental equipment
Figure 2. Experimental equipment container.

Table 1. Processing conditions for EG and CO2.

CO2 Ventilation Ventilation Amount of Packing Amount of
Type of density
Concentration speed time IBA CO2
gas
(%) (L/min) (h) (kgDW) (g/cm3 ) (g/kgDW)

EG 7 85 19 138 0.7 90
CO2 100 38 6.5 293 0.86 93

Table 2. Physical properties of carbonated IBA.

s Fc
Sample Cu Cc
(g/cm3) (%)
Untreated 2.385 24.7 3.0 11.7
EG-IBA 2.368 49.2 4.4 14.7
CO2-IBA 2.360 28.5 3.4 14.9
Note: ρ s = Particle density, Cu = Coefficient of uniformity, Cc = Coefficient of curvature, Fc = Concent rate of fine grain fraction

2.3 Experimental details of California bearing ratio (CBR) test

CBR tests (JIS A 1211, 2009) were carried out to investigate the possibility of usage of
carbonated IBA as a base course material. The CBR index provides a measurement of
the impact resistance of compacted carbonated IBA by the E-b method (using 4.5 kg
rammer and drying method with non-repeating sample). Table 3 shows the testing
conditions of CBR test. The water content of carbonated IBA was adjusted to the
optimum water content considering compaction test results.

156
Table 3. Testing conditions for the CBR test (JIS A 1211, 2009).
Weight of Number Number Maximum
Water content Compaction
Sample rammer of of grain size
(%) method
(kg) layers blows (mm)
Untreated Optimum 17
EG-IBA water E-b 4.5 3 42 19
CO2-IBA content 92

2.4 Experimental details and testing conditions of the batch leaching test
It is important to understand the leaching properties of heavy metals in the carbonated
IBA. The Japanese Leaching Test No. 46 (Notifications No. 46 by the Japanese
Ministry of the Environment, JLT 46) was used to measure concentrations of heavy
metals in the carbonated IBA. Table 4 shows the testing conditions of JLT 46. The
samples are sieved to less than 2 mm and a solvent is added so that the liquid-solid ratio
(L/S) is 10. Inductively coupled plasma atomic emission spectrometry (ICP-AES) was
used for analysis of lead (Pb), cadmium (Cd), and boron (B). Ion chromatography was
used for analysis of fluorine (F). For the analysis of hexavalent chromium (Cr(VI)), a
spectrophotometer (UVmini-1240) was used by means of diphenyl carbazide
absorptiometry. The detection limits of Pb, Cd, B, F and Cr(VI) are 0.01 mg/L, 0.001
mg/L, 0.01 mg/L, 0.08 mg/L, and is 0.02 mg/L, respectively. The reasons for focusing
on these heavy metals are that they are included in the Japanese environmental standard
values for soils and are in IBA.

Table 4. Testing conditions for the batch leaching test (JLT 46).
Liquid-solid Maximum Shaking
Eluent Measurement of heavy
Sample ratio grain size time
pH metals and metalloid
(L/S) (mm) (hour)
Untreated
EG-IBA 5.8 - 6.3 10 2 6 Pb, Cd, Cr(VI), B, F
CO2-IBA

3. Results and discussion

3.1 Mechanical characteristics of carbonated IBA as base course material

The Figure 3 shows compaction curve relationships between the dry density of each
carbonated IBA and water contents obtained from compaction test using the E-b
method. The IBA used in this study has a clear peak under each condition. These data
indicate that the carbonation treatment slightly reduces the maximum dry density and
increases the optimum water content by approximately 5%. Figure 4 shows the results
of a CBR test performed using the maximum dry density and the optimum water
content obtained by the compaction test results. The CBR index provides a

157
measurement of the impact resistance of the compacted aggregate. It is determined as
the ratio between the impact load causing a given penetration in the samples and a fixed
pattern (FORTEZA et al., 2004). Focusing on the untreated results, the IBA satisfied the
criteria for base course material (more than 80%: JGS, 2010) regardless of the
compaction degree. Regarding the conditions using EG and CO2, the results satisfied
the quality standard of sub base course material (more than 30%: JGS, 2010) regardless
of the degree of compaction. These CBR test results indicate that the carbonation
treatment tends to lower the CBR value of IBA. The decrease in the CBR value is
attributed to the change in unreacted CaO and Ca(OH) 2 to CaCO3 due to carbonation
and the reduction of compaction force. FUJIKAWA et al., (2020) showed that each
CBR value decrease with increasing carbonation is considered to change the grain size
distribution curve of IBA and is also caused by decreased cohesion of IBA. Other
factors that reduced CBR is attributed to the effect of reaction heat generated in the
carbonation treatment. Figure 5 shows the monitoring results of temperature changes in
the IBA during the carbonation process. The temperature of IBA rose to 80-100 °C, and
water vapor was observed in the container. Considering that the fine grain fraction
content increased before and after the carbonation treatment, it is possible that some
deformation such as refining, or something occurred in the IBA particles. As a result, it
is possible that the CBR decreased due to the change in the compaction characteristics
of IBA. From our results it is revealed that the IBA can be effectively used as sub-base
course material based on its mechanical properties, even though the CBR value
decreased due to carbonation.

1.5
dmax : 1.46 g/cm3 dmax : 1.45 g/cm3
w opt: 21.4 %
1.45 wopt: 16.6 %
Dry density  (g/cm )
3

1.4
d

dmax : 1.42 g/cm3

1.35
wopt: 20.7 %

1.3
Untreated
1.25 EG-IBA
Compaction:
CO2-IBA E-b mehod
1.2
0 10 20 30 40
Water content w (%)
Figure 3. Compaction curve relationships between dry density of each carbonated IBA
and water content.

158
 150

Compaction Deg: 90%

120 Compaction Deg: 95%

CBR value (%) 90 Standard for base-course: more than 80%

60

30
Standard for sub base-course: more than 30%

0
1
Untreated EG-IBA CO2-IBA

Figure 4. Results of CBR test performed using the maximum dry density and the
optimum water content.
P1-Upper
P1-Upper P1-Midle
150 P1-Modle 150 P1-Bottom
P1-Bottom P3-Midle
EG P3-Midle
CO2
P3-Bottom
Temperture ( C)
Temperture ( C)

P3-Bottom P4-Middle
100 100
o

P4-Midle P4-Bottom
P4-Bottom Out side
Out side
50 50

0 0
0 5 10 15 20 0 5 10 15 20
Time (hour) Time (hour)
(a) Exhaust gas. (b) CO2
Figure 5. Monitoring results of temperature changes in the IBA
during the carbonation process.

3.2 Leaching properties and environmental safety assessment of carbonated IBA

Table 5 shows the batch leaching test results (JLT 46). Focusing on Pb, even though
untreated IBA has a high concentration of elution, the concentration of the elution of
carbonated IBA by means of CO2 is suppressed to below the soil environmental
standard value set by the Ministry of the Environment in Japan. In addition, it was
revealed that even though carbon dioxide concentration, ventilation speed, and
ventilation time were adjusted so that the total amount of carbon dioxide gas to be
passed was equal, the immobilizing effect was particularly high in CO2 compared with
EG. This is because CO2 has a higher gas concentration and a slower aeration rate than
EG, therefore the reaction was spread slowly throughout. The elution of B in the CO2
condition is attributed to a decrease in pH value.

159
Regarding other heavy metals and metalloids except Cr(VI), all satisfied the soil
environmental standards regardless of the difference in carbonation treatment. However,
the carbonation of Cr(VI) exceeded the soil environmental standard value. The reason
why the concentration of Cr increased was attributed to shift of the oxidation-reduction
potential of Cr(III) in IBA to the acidic side, changing it into Cr(VI). Alternatively, it is
possible that Cr(VI) was present in IBA from the beginning (fixed to a hydrate), etc.)
and that hydrate was destroyed by carbonation to release Cr(VI). Regarding this point, it
is necessary to correctly quantify and confirm the initial content of hexavalent
chromium and the content after carbonation. Therefore, further studies to suppress the
elution of Cr(VI) due to carbonation are necessary.

Table 5. Batch leaching test results (JLT 46).

Pb Cd Cr(VI) B F
Sample pH
(mg/l) (mg/l) (mg/l) (mg/l) (mg/l)
Untreated 12.78 4.7 N.D. 0.04 N.D. 0.35
EG-IBA 11.92 0.02 N.D. 0.19 0.19 0.18
CO2-IBA 10.94 N.D. N.D. 0.15 0.93 0.29

Soil environmental standard

< 0.01 < 0.003 < 0.05 <1 < 0.8
values (Japan)

Lower limit of quantification in

0.01 0.001 0.02 0.01 0.08
this experiment

4. Conclusions
In this study, we examined the applicability of IBA, using two types of carbonation
treatment, as construction materials, such as base course and embankment materials,
and its environmental safety.
Even though the carbonation treatment tends to lower the CBR value of IBA,
carbonated IBA can be effectively used as a sub-base course material based on its
mechanical properties and regardless of the carbonation conditions. Regarding heavy
metal and metalloid leaching behavior, it was revealed that all IBA samples satisfied the
soil environmental standards value set by the Ministry of the Environment in Japan,
except for Cr(VI). The immobilizing effect was particularly high for Pb in CO2
compared with EG. However, the carbonation of Cr(VI) exceeded the soil
environmental standard value. Therefore, it is necessary to take measures to immobilize
Cr(VI) in future studies.

160
5. Acknowledgments
This research was supported by the Environment Research and Technology
Development Fund (3-1804) of the Environmental Restoration and Conservation
Agency of Japan. Also, this research was supported by Saga city by helping to carry out
the experiment.

6. References
QUEK A., XU W., GUO L., WU D. (2016). Heavy Metal removal from Incineration
Bottom Ash through Washing with Rainwater and Seawater, International Journal of
Waste Resources, Vol.6, Issue 1, doi: 10.4172/2252-5211.1000203
JGS - JAPANESE GEOTECHNICAL SOCIETY (2010). Basic and Companion of Soil
test, Doshitsusiken Kihon to Tebiki, 89 p., Japanese Geotechnical Society, (in
Japanese).
JIS A 1211. (2009). Method for the California Bearing Ratio (CBR) of soils in
laboratory. Japan Industrial Standard, Japanese Industrial Standards Committee, Tokyo,
Japan.
KUBOTA H., SHIGEIZUMI K., FUJIKAWA T., KOGA C., SATO K. (2020). An
accelerated, on-site bottom ash aging and washing treatment and its effect for long-
term leaching, Waste Biomass Valorization, Vol.11, pp. 7067-7077, doi.org/10.1007/s12649-
020-01136-9
FORTEZA R., FAR M., SEGUI C., CERDA V. (2004). Characterization of bottom ash
in municipal solid waste incinerators for its use in road base, Waste Management 24
(9), pp. 809-909, doi: 10.1016/j.wasman.2004.07.004
SAGA CITY. (2020). For carbon dioxide separation and recovery equipment, Official
homepage. (Accessed on 6th Mar. 2020). https://www.city.saga.lg.jp/main/44494.html
FUJIKAWA T. SATO K., KOGA C., SAKANAKURA H. (2020). Effect of Aging on
Material Characteristics and Leaching Properties of Incineration Bottom Ash from
Municipal Solid Waste, Waste Biomass Valorization, Vol.11, pp. 7097-7107, doi:
10.1007/s12649-020-01058-6
BRENNE T., WEIGAND H. (2015). Stabilization of heavy metals in MSWI bottom ash
by dynamic carbonation, Proceedings of the 14th International Conference on
Environmental Science and Technology, CEST2005_00227.
JIANG Y., XI B., LI X., ZHANG L., WEI Z. (2009). Effect of water-extraction on
characteristics of melting and solidification of fly ash from municipal solid waste
incinerator, Journal of Hazardous Materials, 161 (2-3), pp. 871-877,
doi.org/10.1016/j.jhazmat.2008.04.033

161
162
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Effect of fine and water contents of soil on effective separation

of mixed wastes generated by huge disasters

Kansei HIRAOKA 1, Junichiro SHIOIRI 2, Shogo NAKAGAWA 3,

Atsushi TAKAI 1, Takeshi KATSUMI 1

1. Graduate School of Global Environmental Studies, Kyoto University, Yoshida-

honmachi, Sakyo-ku, 606-8501, Kyoto, Japan.
[email protected]; [email protected];
[email protected]
2. Okumura Corporation, 2-2-2 Matsuzaki-cho Abeno-ku 545-8555, Osaka, Japan.
[email protected]
3. Graduate School of Engineering, Kyoto University, Kyotodaigaku-katsura, Nishikyo-
ku 615-8530, Kyoto, Japan.
[email protected]

Abstract:
Large volumes of mixed wastes generated from huge natural disasters become a big issue
for the restoration of the disaster-stricken area. Mixed wastes should be separated into
individual materials for utilisation of the materials. However, separating mixed wastes
indiviualy is difficult due to limited time and budget. Consequently, impurities are
contained in the separated materials. This study evaluates the effect of fine and water
contents of soil on the quality and quantity of separated materials. Sieving experiments
were conducted using a vibrating sieve machine. The results show that high fine and water
content result in low separation accuracy if the plastic limit is not exceeded because a
large amount of soil adhered to the woodchips and formed clumps. Soil consistency can
be considered one of the critical factors affecting the quality and quantity of separated
materials.

Keywords:
Great East Japan Earthquake, Disaster wastes, Separation treatment, Coefficient of
separation.

Corresponding author mail: [email protected]

1. Introduction
Recently, large-scale disasters have been increasing worldwide (VINOD & RAMON,
2015). In Japan, the Great East Japan Earthquake of March 11, 2011, generated 30 million
tons of disaster wastes (MOE, 2018). About two months after the disaster, the Ministry
of the Environment (MOE) announced the ‘Master Plan for the Disposal of Disaster
wastess from the Great East Japan Earthquake’ (MOE, 2011). The plan set a target of
completing the disposal of disaster wastes by the end of March 2014. However, due to
the Fukushima Daiichi Nuclear Power Plant accident, disaster wastes management in
Fukushima Prefecture was not completed on schedule. However, this goal was achieved
in Iwate and Miyagi Prefectures (JSCE, 2014). This experience reaffirmed disaster wastes
management as essential for early recovery and reconstruction of the affected areas.
Disaster wastes is transported from the disaster-affected areas to the primary temporary
storage site. At the primary storage site the wastes are roughly separated, and the
separated wastes are transported to secondary temporary storage site. After that, the
materials undergo a more detailed secondary treatment that combines crushing and
sieving, and are classified as either reconstruction materials which will be recycled or
waste which will be incinerated.
Soil-based mixed wastes account for 1/3 of all disaster wastess by weight (TAKAI et al.,
2016), and its effective utilisation is required. Considering the Great East Japan
Earthquake, the treatment is done ad hoc, and the quality of separated materials depends
on the technique used in the waste treatment site (JSCE, 2014). In order to increase the
effective utilisation of separated materials, a quantitative evaluation of the technical
factors affecting the quality and quantity of separated materials is needed.
This study focused on how aggregation of soil-based mixed wastes affects the quality and
quantity of separated materials. Since the fine and water contents of the soil influence the
aggregation of soil-based mixed wastes, they were considerd to be parameters for waste
separation. Sieving experiments for the soil-based mixed wastes were conducted using a
vibrating sieve machine. The sieving experiment was conducted to simulate the sieving
process, which is the main process of disaster wastes treatment systems.

2. Methodologies

2.1 Materials and equipment

The simulated soil-based mixed wastes constituted of soil, woodchips and crushed rock.
The woodchips and crushed rocks were used to simulate mixed wastes containing
combustible materials and those containing non-combustible materials, respectively.
Only two materials were mixed in this study to simplify the testing conditions and data
analyses. Decomposed granite soil and Kasaoka clay were used as soil. Table 1
summarises the soils’ physical properties. The decomposed granite soil was selected since

164
Table 1. Physical properties of decomposed granite soil and kasaoka clay.
Particle size (%) Soil particle Liquid limit Plastic limit
Gravel Sand Clay 3
density (g/cm ) (%) (%)
Decomposed
42.8 51.2 6.0 2.59 ND ND
granite soil
Kasaoka clay 0 0 100 2.68 58.5 24.6
Note: ND = No Data

Table 2. Mixing ratio of the simulated mixed wastes.

Fine Water Woodchips Soil
content, content, or Decomposed Kasaoka
FC (%) w (%) Crushed stones granite soil Clay
7.80 1 9.0 0.38
12.6 1 9.0 0.38
10
17.3 1 9.0 0.38
22.0 1 9.0 0.38
7.80 1 7.2 1.84
34
17.3 1 7.2 1.84
Combustible
7.80 1 6.1 3.25
12.6 1 6.1 3.25
40
17.3 1 6.1 3.25
22.0 1 6.1 3.25
7.80 1 6.0 3.05
56
17.3 1 6.0 3.05
7.80 1 2.0 0.08
12.6 1 2.0 0.08
10
17.3 1 2.0 0.08
22.0 1 2.0 0.08
10.0 1 1.8 0.20
15.0 1 1.8 0.20
34 20.0 1 1.8 0.20
30.0 1 1.8 0.20
Non- 37.5 1 1.8 0.20
combustible 7.80 1 1.4 0.72
12.6 1 1.5 0.79
40
17.3 1 1.4 0.72
22.0 1 1.4 0.72
15.0 1 1.2 0.80
22.5 1 1.2 0.80
56 30.0 1 1.2 0.80
37.5 1 1.2 0.80
45.0 1 1.2 0.80
Note: The values of soil is the ratios of soil to woodchips or crushed stones.

165
it is widely distributed in Japan. Kasaoka clay was selected to control fine content of the
soil-based mixed wastes. Woodchips were used to simulate practical conditions, since
disaster wastess contain a lot of combustible materials (TAKAI et al., 2016). The
woodchips were thin and ~10-cm long. Crushed stones of 30–40 cm in diameter were
used.
Simulated mixed wastes were prepared by mixing soil, woodchips or crushed stones, and
water in a 1-L steel bowl. Approximately 1 kg simulated wastes was prepared for each
case listed in Table 2. Water content shown in this table is the water content of the
recovered soil.

2.2 Methods
The simulated mixed wastes were sieved with a vibrating sieve machine (RETSCH AS
200) using an steel sieve with a mesh opening of 19 mm at an amplitude of ~1.2 mm and
a frequency of 3600 vibrations/min. The materials passing the screen and those that were
retained were oven-dried for 24 hours. Woodchips contained in the oven-dried materials
were removed by washing the materials with water. The woodchips, crushed stones, and
soil were oven-dried again for 24 hours. Then, the mass of each separated materials was
measured.

3. Results and discussion

3.1 Soil recovery rate

Figure 1 shows the relation between soil water content and soil recovery rate. Here the
soil recovery rate is used as an index for evaluating the amount of soil that was recovered
from the sieving process. The soil recovery rate is calculated by dividing the amount of
recovered soil by that of the original soil.

a) Combustible b) Non-combustible
Figure 1. Soil recovery rate for the simulated mixed wastes.

166
For both kinds of simulated mixed wastes, the soil recovery rate was ~90% at lower water
content while the soil recovery rate decreased as the water content increased. However,
after exceeding the plastic limit, the soil recovery rate increased with increasing water
content. For instance, the plastic limit of FC = 34% is 26.6% (Table 3). After water content
exceeds this plastic limit (wp), the soil recovery rate increases and reaches ~60%.

Table 3. Liquid, plastic and shrinkage limits for original soil.

FC(%) Liquid limit(%) Plastic limit(%) Shrinkage limit(%)
10 ND ND 21.6
34 39.4 26.6 18.8
40 42.1 22.8 15.4
56 47.2 33.1 16.6

Liquid index (IL) is an index of the relative hardness and softness of the soil in the each
water content state, expressed as:
wn − wp
IL = (1)
wL − wp
where: wn is the water content of the recovered soil, wp is the plastic limit of the recovered
soil, and wL is the liquid limit of the recovered soil.
IL of the soil for different fine content is given in Table 3. Figure 2 shows the relation
between soil recovery rate and IL. The figure shows that the soil recovery rate is the lowest
at the point relatively close to wp and confirms the results in Fig. 1. The findings suggest
that fine content changes the plastic limit, and as a consequence, affects the lowest soil
recovery rate.

a) Combustible b) Non-combustible
Figure 2. Liquid index and soil recovery rate.

167
3.2 Quality of recovered soil
Ignition loss which is affected by the presence of combustible materials has been a
criterion for the use of separated soil (TAKAI et al., 2016). Therefore, this section
discusses quality of soil recovered from soil-based mixed wastes containing combustible
materials by determining how much percentage of woodchips the soil contains. Figure 3
shows the percentage of woodchips in the recovered soil against the water content. The
results show that the recovered soil contains ~4–10% woodchips for different water
contents. These findings suggest that it is technically difficult to recover pure soil from
soil-based mixed wastes.

Figure 3. Percentage of woodchips in recovered soil.

4. Conclusions
This study focused on the fine and water contents of the soil on quality and quantity of
separated materials. Sieving experiments for the soil-based mixed wastes were conducted
using a vibrating sieve machine. The results support the following conclusions:
1) For both kinds of soil-based mixed wastes, the soil recovery rate was ~90% at lower
water content while the soil recovery rate decreased as the water content increased.
However, after exceeding the plastic limit, the soil recovery rate increased with
increasing water content. For instance, the plastic limit of FC = 34% is 26.6% (Table
3). After water content exceeds this plastic limit (wp), the soil recovery rate increases
and reaches ~60%.
2) The findings suggest that fine content changes the plastic limit, and as a consequence,
it affects the lowest soil recovery rate.
3) The results show that the recovered soil contains ~4–10% woodchips for different
water contents. These findings suggest that it is technically difficult to recover pure
soil from mixed wastes.

168
5. Acknowledgments
The authors gratefully appreciate and acknowledge the generous cooperation extended by
Dr Lincoln W. Gathuka (Kyoto University) during the preparation of this manuscript.

6. References
JSCE - JAPAN SOCIETY OF CIVIL ENGINEERS (2014). Disposal and effective
utilization of disaster waste. Records and lessons from the East Japan Earthquake, No.
142.
MOE - MINISTRY OF THE ENVIRONMENT (2011). Master plan for the disposal of
disaster wastes from the Great East Japan earthquake,
https://www.env.go.jp/jishin/attach/haiki_masterplan.pdf
MOE - MINISTRY OF THE ENVIRONMENT (2020). Damage caused by the East
Japan Earthquake. Disaster Waste Management Information Site,
http://kouikishori.env.go.jp/archive/h23_shinsai/damage_situation/
MOE - MINISTRY OF THE ENVIRONMENT (2020). Status of formulation of disaster
waste treatment plans and study on the promotion of formulation,
https://www.env.go.jp/recycle/waste/disaster/earthquake/committee2/r1-02/R1_2_02_keikakusakutei.pdf
TAKAI A., KAWASHIMA M., KATSUMI T., INUI T., IWASHITA S.-I., OKAWARA
M. (2016). Quality and its variation of soils recovered from disaster debris in Iwate
prefecture after the 2011 East Japan Earthquake. Journal of Japan Society of Civil
Engineers (Ser. C), Vol. 72, No. 3, pp. 252–264, http://doi.org/10.2208/jscejge.72.252
VINOD T., RAMON L. (2015). Global increase in climate-related disasters. Asian
Development Bank economics working paper series No. 466, http://doi.org/10.2139/ssrn.2709331

169
170
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Analysis of sediment-based fired bricks strengths: A case

study of fluvial sediments from Mexico

Mazhar HUSSAIN 1,2, Daniel LEVACHER 1,2, Nathalie LEBLANC 2, Hafida

ZMAMOU 2, Jean-Baptiste BESNIER 2, Irini DJERAN-MAIGRE 3, Andry
RAZAKAMANANTSOA 4

1. Université Caen Normandie, M2C, 14000, Caen, France.

[email protected]; [email protected]
2. UniLaSalle, Transformations & Agroressources EA7519, Rouen, France.
3. Université Lyon, INSA Lyon, GEOMAS, 69621 Villeurbanne, France.
4. Université Gustave Eiffel, Département GERS, 44344 Bouguenais, France.

Abstract:
Ceramics for construction appears to constitute an interesting beneficial use for dredged
sediments. Many case studies were reported from different countries. As any non-
renewable resource is not used, this recovery issue could be considered as sustainable
and eco-friendly in sediments management, even if the firing energy is needed.
A feasibility study of manufacturing bricks was conducted with Mexican fluvial
sediments at laboratory scale. Sediments were sampled from 2 main sites along the
Usumacinta River (Mexico) and stockpiled in barrels. Sediments from each barrel were
characterized to obtain the properties useful for ceramics manufacturing.
A large experimental plan was undertaken on specific specimens (prismatic and cubic)
prepared at molding water content according to Sembenelli diagram and fired at 850°C.
The fired specimens were tested for flexural and compressive strengths. The results are
presented and discussed regarding the sediment properties. Following the results
obtained from separate sediments, sediments from each site were mixed and finally all
the sediments from two sites were mixed, fired and tested. Analysis of the results on
strengths are given and concluding remarks are drawn according to the approach
considered and bricks standards.

Keywords:
Sediment, Characterization, Beneficial uses, Fired bricks, Compressive strength,
Flexural strength, Sustainable management.

Corresponding author mail: [email protected]

1. Introduction
Dredged sediments are used in landfills, beach nourishments, roads and ceramics etc. In
ceramics, dredged sediments can be used in making fired bricks which are key
construction material. Dredged sediments use in ceramics gives an alternate to non-
renewable clay resources. Clay and other minerals additive are sometimes used to make
suitable grain mix and eliminate contaminants from dredged sediments (XU et al., 2014,
AGOSTINI et al., 2007). Physical, chemical, and mineralogical characteristics of
sediments are determined to use them in fired bricks. Some important properties are
grain size, Atterberg limits, chemical composition, organic matter and carbonate content
etc. Characteristics of sediments help to define sediments suitability for fired bricks
through industrial approaches such as Winkler diagram, Sembenelli interpretation, clay
workability chart, Casagrande diagram and Augustinik diagram. Winkler diagram is
used in fired bricks industry for sediments selection on the base of clay, silt and sand
percentage. Sembenelli diagram defines moulding moisture content of fired bricks on
the base of Atterberg limits. Clay workability chart describes sediments suitability for
moulding with liquidity and plasticity limits. Casagrande diagram (based on Gippini)
describes extraction characteristics of sediments with Atterberg limits. Augustinik
diagram recommends zones suitable for fired bricks with oxide contents of sediments
(HUSSAIN et al., 2020). Chemical analysis of sediments helps to determine the
presence of oxides such as silica, alumina, lime, iron oxide, magnesia and pollutants.
Pollutants are neutralized with treatment and addition of minerals. Organic matter of
sediments is found as its higher amount induces pores in fired bricks which effects the
strength and water absorption capacity of bricks. Carbonate content effects the
durability of bricks due to chemical reactions at high temperature. Excessive carbonate
content cause swelling and disintegration of bricks. Procedure of making fired bricks
involve sediments mixing, moulding, drying and firing. Sediments are mixed with
water. Quantity of water usually varies between liquidity limit and plasticity limit.
Sediment mix is moulded into different shapes. Moulding is followed by drying. Oven
drying is done at temperature ranges of 40°C to 105°C. Finally, bricks are fired at
temperature ranging from 700°C to 1100°C. Fired bricks characteristics include
compressive strength, tensile strength, water absorption, density, color and shrinkage in
bricks (AL-FAKIH et al., 2019). Compressive and tensile strength of fired bricks varies
with particles size, oxide contents, moulding moisture content, minerals presence and
firing temperature. Fired bricks compressive strength values found in literature varies
from 2-50 MPa (HUSSAIN et al., 2020). Common range of compressive strength is 5-
15MPa. Indirect tensile strength of bricks is usually found by flexural strength test by
three-point bending test.
Sediments used in this study are dredged from Usumacinta River in Tabasco State of
Mexico. Sediments were used in manufacturing fired bricks after studying their
characteristics and suitability for fired bricks.
172
2. Materials and methods

2.1 Sediments
Usumacinta River sediments were dredged from different sites in Tenosique and Jonuta
towns in Tabasco state of Mexico and are named as T1, T2, T3, T5, T6, J1, J3, J4 and
J5. Characteristics of sediments were determined through different tests and are
summarized in Table 1.

Table 1. Characteristics of Usumacinta River sediments.

LL Clay Silt Sand Particle density Organic matter CaCO3 MBV Wopt
Sediments
(%) (%) (%) (%) (g/cm3) (%) (%) (g/100g) (%)
T1 39.01 6.90 40.2 52.9 2.70 3.77 8.21 2.53 17.3
T2 28.75 8.97 55.9 35.1 2.71 3.5 7.19 2 15.5
T5 36.75 8.75 45.9 45. 2.61 3.46 1.9 5.9 20.8
T6 30.71 6.09 40.7 53.2 2.68 4.9 8.73 1.93 17.9
J3 37.62 4.66 36.3 59.0 2.70 4.98 8.55 4.5 18.3
J4 59.77 13.4 62.5 24.1 2.53 5.72 8.49 >8 -

Different industrial approaches such as Sembenelli interpretation, clay workability

chart, Winkler diagram and Augustinik diagram are presented in Figure 1 to see
Usumacinta River sediments suitability for fired bricks.

a b

c d

Figure 1. Common approaches used for manufacturing fired bricks.

173
Usumacinta River sediments have intermediate plasticity, low clay, carbonate and
organic matter content. MBV value of J4 indicates that these sediments have high
plasticity while for T2, plasticity is low. From Figure 1-a we can see that most of the
sediments have optimum moulding characteristics. In Augustinik diagram (Figure 1-d),
most of the sediments lies with in the zone recommended for bricks. Winkler diagram
(Figure 1-c) show that J4 sediments lies near suitable zone for manufacturing bricks
while other sediments have less clay and are outside recommended zone.

2.2 Manufacturing of bricks

Manufacturing of bricks consists of sediments mixing, moulding, drying and firing.
Sediments were grinded and passed through 2mm sieve (XU et al., 2014). Moulding
moisture content was taken at midpoint between the liquidity and plasticity limit of
sediments in Sembenelli diagram shown in Figure 1-b. Mixing of sediments was
followed by moulding. Due to sediments and oven limitation, laboratory scale wooden
bottomless mould of cubic and prismatic shapes of size 20x20x20mm3 and
15x15x60mm3 were used for compressive and flexural strength tests. Prismatic and
cubic bricks specimens are shown in Figure 2. Moulding of sediments was followed by
oven drying at 60°C for 4-12 hours duration. Drying of bricks removes the moisture,
preserves the shape and prevents crack development during firing. After drying, bricks
were fired at 850°C in oven for six hours. During firing of bricks, moisture evaporation,
oxidation, CO2 emission and vitrification takes place at different temperatures. After
firing bricks were left in oven for 12 hours to cool. Different steps of manufacturing
fired bricks are shown in Figure 2.

Figure 2. Fired bricks manufacturing steps.

174
3. Mechanical testing
Compressive and tensile strength of fired bricks was found with Shimadzu AGS-X
model machine by using 200 N and 50 KN sensors.

3.1 Compressive strength

Compressive strength tests were done on cubic specimens, see Figure 3a. Usumacinta
bricks testing and typical compressive load deflection behaviour are shown in Figure 3.

a b

Figure 3. Compressive strength test and load deflection curve.

Compression curve in Figure 3-b indicates the linear increase in strength before failure.
Failure in compression is not brittle; a post peak load behavior is observed on the curve
in Figure 3-b.

3.2 Tensile strength

Flexural strength was obtained with three-point bending test on prismatic samples, see
Figure 4a. Usumacinta bricks testing and typical flexural load deflection behaviour is
shown in Figure 4. In Figure 4b, a sudden drop in flexural force at failure is observed.

a b

Figure 4. Tensile strength test and load deflection curve.

175
4. Analysis of results
Compressive and flexural strengths of fired bricks are shown in Figure 5. The fired
brick strengths of sediments J4 and T3 are largely over the values obtained for other
sediments. This can be explained by different parameters such as quantity of clay, silt
and sand, moulding characteristics, oxides and Atterberg limits that influences the
strength.

Figure 5. Compressive and flexural strengths for sediment bricks fired at 850°C.

3.2 Relationship between compressive and tensile strength

Compressive strength relation with modulus of elasticity and flexural strength for
Usumacinta River sediments from different sites at temperature range of 700°C to
1100°C is shown in figure 6a and 6b respectively. Modulus of elasticity ranges from
around 10MPa to 1000 MPa depending on firing temperature.

a b

Figure 6. UCS vs modulus of elasticity (a) and flexural tensile strength (b).

176
5. Effect of mixing
Additives like clay are mixed with dredged sediments to make them suitable for fired
bricks (BAKSA et al., 2017). However, sediments suitability can also be gained by
mixing dredged sediments from different sites with different proportions with respect to
their fine particles, oxide content and specific surface area. Winkler diagram in Figure
1c suggests that percentage of clay is smaller in both Tenosique and Jonuta sediments.
To optimize bricks strength and propose practical approach, different sediments mixes
were prepared. Sediments from Jonuta sites were mixed and named as J. Sediments
from Tenosique sites sediments and named as T. In global mix, both Tenosique and
Jonuta sediments were mixed and named as T+J.

Figure 7. Simulations on Winkler and Augustinik diagram (YAMAGUCHI, 2018).

177
Winkler diagram recommends zones suitable for bricks and tiles on the base of sand, silt
and clay content. Suitable sediment mixes within the recommended zone for
Usumacinta River sediments were found. Similarly based on oxide content of
Usumacinta River sediments suitable sediments mix for bricks recommended zones in
Augustinik diagram were found (YAMAGUCHI, 2018). Results for compressive
strength and tensile strength for T, J and T+J sediments are shown in figure 5. In figure
7-b, J1-2N, J5-15N and T3-8N are mixture suggested with Winkler diagram. Mixture
suggested with Augustinik diagram are J5-15N, T3-8N, J1+J5, J3+T5, T1+T2+T6 and
J1-2N (YAMAGUCHI, 2018). We can observe from the figure 5 that mixture suggested
with Winkler approach have relatively high compressive and tensile strength.

6. Conclusions
In this study, Usumacinta River sediments use in fired bricks was assessed through
different industrial approaches. Compressive and tensile strength results of fired bricks
suggest that, bricks with T3, T5, J4, J5 sites sediments have good compressive and
tensile strength. Sediments mix based on Winkler diagram also show encouraging
results. If mixing of sediments could improve mechanical behaviour, the temperature
must be optimized so as to meet the standards required for bricks use in construction.

7. References
AGOSTINI F., SKOCZYLAS F., LAFHAJ. Z. (2007). About a possible valorisation in
cementitious materials of polluted sediments after treatment. Cement & Concrete
Composites 29, pp. 270–278, https://doi.org/10.1016/j.cemconcomp.2006.11.012
AL-FAKIH A.A., MOHAMMED B.S., LIEW M.S., NIKBAKHT E. (2019).
Incorporation of waste materials in the manufacture of masonry bricks: An update
review. Journal of Building Engineering 2, pp. 37–54. https://doi.org/10.1016/j.jobe.2018.09.023
BAKSA P., CEPAK F., LUKMAN R.K., DUCMAN V. (2017). An evaluation of
marine sediments in terms of their usability in the brick industry: Case study Port of
Koper, Journal of Sustainable Development of Energy, Water and Environment
Systems, 6(1), pp. 78-88, https://doi.org/10.13044/j.sdewes.d5.0183
HUSSAIN M., LEVACHER D., LEBLANC N., ZMAMOU H., DJERAN-MAIGRE I.,
RAZAKAMANANTSOA A. (2020). Sediment-based fired brick strength optimization
A discussion on different approaches. XVIème Journées Nationales Génie Côtier –
Génie Civil Le Havre, 2020, DOI:10.5150/jngcgc.2020.072
XU Y., YANN C., XU B., RUAN, X., WEI Z. (2014). The use of urban river sediments
as a primary raw material in the production of highly insulating brick. Ceramic
International, 40, pp. 8833-8840, https://doi.org/10.1016/j.ceramint.2014.01.105
YAMAGUCHI K. (2018). Consideration of the sustainable utilization of the sediments
in Usumacinta River. Master thesis report, Kyoto University-IFSTTAR-Université de
Caen-Normandie, 20p.
178
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Grain size characterisation

Study of the sediment dynamics of Monastir bay, Tunisia

Nouha KHIARI 1, Abdelfattah ATOUI 2, Nadia KHALIL 1,

Abdelkrim CHAREF 1

1. Water Technology and Research Centre, Borj Cedria, Soliman Tourist Route BP273,
8020 Soliman, Tunisia.
[email protected]
2. Ministère de l'Équipement, de l'Habitat et de l'Aménagement du Territoire, Direction
de l'Hydraulique Urbaine (DHU), Rue de Hadi Chaker Sakiet Ezzit, Sfax 3021,
Tunisia.

Abstract:
The Monastir Bay is located in the south of the Gulf of Hammamet, is an environment
with strong anthropogenic activity. the terrigenous inputs are limited by development
works linked mainly to the conservation of water and soil in the various coastal
catchment areas of the region, Khniss, Ksibet El Mediouni, Syada, Lamta, Teboulba and
Ras-Dimas. Several sources of industrial pollution are installed along the coast and/or
on the rivers with exoreic outlets. A contribution to the study of sedimentary dynamics
through a particle size approach is essential for any ecosystem study of the region.
Particle size analysis of the surface sediments shows that Monastir Bay is covered by
two types of facies. They are fine sands and medium sands which are variable between
moderately classified to poorly classified. The average diameter is growth from center
of the bay to wide and all along the coast; from El Enf arrow to Teboulba. Only the area
between Teboulba and Ras Dimas, protected by the sandy arrow of Ras Dimas, is
covered by the same texture (fine sand) during these years. The sands are transported in
a more at least agitated environment with the discharge of particles from the fishing port
of Monastir towards the arrow of Ras Dimas by the currents of coastal drift generated
by the prevailing swell of direction Nord-Northeast and coast-wide by the back currents
near the bottom. Recent sedimentary dynamics in Monastir Bay are closely related to
the effects of north-to-northeast swells and the regression of detritic inputs by exoreic
wadis.

Keywords: Monastir Bay, Sediment dynamics, Surface sediment.

Corresponding author mail: [email protected]

1. Introduction
The study area is part of the coastline of eastern Tunisia. It is the bay of Monastir that
extends from the Marina Monastir to Ras-Dimas. Located between latitude 35°38'00’’
and 35°46'00’’ and longitude 10°49'00’’ and 11°03'00’’ (Figure 1). The bay of Monastir
is virtually closed to the East by the shoals that join the peninsula of Thapsus to the
Kuriat Islands. On the west side, the bay closes on the Monastir peninsula
(KAMMOUN, 1981). The Monastir Bay is characterized by relatively flat landscapes
and the most remarkable relief is the cliff of Monastir and the infralittoral zone
belonging to the carbonate platform developed under temperate climate.

2. Material and method

2.1 Method of sampling

Two sampling campaigns for surface sediments were conducted. A sea campaign that
was conducted to collect surface sediment in April 1999 (Figure 1). A total of 42
samples were collected, with a mesh size of approximately 1 km. A second campaign
was carried out in August 2015 (Figure 2). Small bottom sediments were collected from
15 radials, ranging from 1 to 7 m deep, using a diver. A total of 97 sediment samples are
collected.

Figure 1. Location of surface sediment samples (1999).

180
 Figure 2. Location of surface sediment samples (2015).

2.2 Sediment analysis method

The coarse fraction is dry sieved for 20 minutes on a series of AFNOR type sieves with
mesh sizes of 2000, 630, 500, 250, 125 and 63. For each sample, a particle size curve
was established in a semi-logarithmic diagram in which the ordered represents the
cumulative percentage of refusal and the abscissa the corresponding diameter. Some
size indices are determined.

3. Results and discussion

3.1. Grain distribution of surface sediment in the Monastir bay (1999)

There are two forms of cumulative surface sediment curves (LAFOND 1953; LAFOND
1965; RIVIERE 1953).
- the cumulative curves (P10, P12 and H17) are characterized by a form of S
indicating that they are sands transported in a more or less agitated environment
with discharge of fine particles towards the sea by the return currents.
- the cumulative curves (H11, H15, H18 and H16) are less regular and spread out,
indicating poor classified.

181
Figure 3 confirms that sediments collected from the coastal fringe between Khniss and
Ksibet El Mediouni show a particle size composition dominated by sandy materials that
are fine to very fine in size towards the open sea.

Figure 3. Spatial evolution of average sediment size in sand (1999).

This particle size distribution indicates that sedimentation is sluggish and settles in a
shallow and very quiet environment. For the coastal fringe between Ksibet El Mediouni
and the port of Teboulba, the middle sands dominate the seabed with a decreasing
gradient going from the coast to the offshore. The sandy fraction is more heterogeneous
and poorly classified following a very limited hydrodynamism. Finally, the sediments
collected from the coastal fringe of Teboulba to Ras Dimas, one finds the particle size
fractions of the fringe Khniss-Ksibet El Mediouni where the sands are also fine that
tends to decrease towards the large

3.2. Grain distribution of surface sediment in the Monastir bay (Campaign 2015)
The cumulative curves have the spread S-shape (Figure 4) indicating that the sands are
transported in a more or less agitated environment.
The cumulative curves have a hyperbolic shape (Figure 4). This form presents a
heterogeneity of the sandy stock and a medium is more or less agitated with discharge
of fine particles towards the sea by the currents of return.

182
 Figure 4. Cumulative surface sediment curves.

Figure 5. Spatial evolution of the average size of sand sediments (Campaign 2015).

Sediments collected from the shoreline of Monastir Bay show that 90% of the total
sediment is sandy (Figure 5). Such a particle size distribution means that the sediments
deposited in the littoral fringe of Khniss are composed of fine to medium sand.
Therefore, it is a sedimentation with a sluggish character depositing by settling in a
shallow and very calm environment. For the coastal strip, there are medium sands which

183
are abundant thanks to the relatively high hydrodynamic intensity in the vicinity of the
wadi which allows the large amount of the fine fraction to be swept farther out.

3.3. Comparison
In 2015, the distribution of the sandy fraction shows that there is a particle gradient
from North to South between Monastir and Lamta, and from South to North between
Bekalta and Sayada and a coast gradient to Offshore in the center of the bay between
Ksibet El Mediouni and Teboulba. So, we have a set of drift currents that move the
sediment from north to south and south to north parallel to the coast, which creates a
return current in the central part of the bay in front of Kuirate island. In 1999, the
distribution of grain sizes is mostly fine and very fine. There is a decreasing tri-particle
size from the coast to the wide passing from the average size on the coast to the very
fine depth this is due to a low hydrodynamism and following the terrigenous
contributions. Therefore, the coastal fringe Monastir-Bekalta is the seat of a mainly
sandy sedimentation. This sedimentary hydrodynamics shows an increase in the size of
grains lining the bottom between 1999 and 2015 following the decrease in terrigenous
inputs.

4. Conclusion
Particle size analysis of surface sediments in Monastir Bay shows that the undersea
beach is covered by two types of facies which are fine sands and medium sands and
moderately to poorly classified. The particle size result of the surface sediments of
Monastir Bay shows overall a decreasing grain classification from coast to coast. The
sedimentary hydrodynamics of the study area is closely related to natural factors such as
the energies of the environment, the typology of coasts, the underwater morphology and
the presence of rivers and man-made rivers, as well as the construction of harbours,
have an influence on this distribution of sandy facies.

5. References
KAMMOUN Y. (1981). Etude néotectonique de la région de Monastir-Mahdia (Tunisie
orientale), Thèse de 3éme cycle, Université Paris-Sud, Orsay, 175p.
LAFOND L.R. (1953). Sur l’évolution granulométrique des sédiments dans le bassin de
la Vilaine, Bulletin de la Société Géologique de France, Vol. 6, 75 p.
LAFOND L.R. (1965). Etudes littorales et estuariennes en zone tropicale humide.
Thèse de Doctorat d’Etat, Université Paris-Sud, Orsay.
RIVIERE A. (1953). Méthode d’interprétation des granulométries des sédiments
meubles. Rev. Inst. Fr. du Pétrole et Annales des combustibles liquides, n° spécial,
pp. 102-107.

184
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Comparative study of the compressive strength of hollow

concrete blocks made in two factories in the city of Yaoundé
based on local and imported cement

Bertille Ilalie K. MANEFOUET

1. University of Dschang, Department of Earth science, BP 96, Dschang, Cameroon.

[email protected]

Abstract
This article is the comparative study of mechanical properties of perforated concrete
blocks manufactured in two factories in the city of Yaoundé. It has two goals: (1) the
study of compressive strength of 28 days old hollow concrete blocks and (2) the better
understanding of the effect of this on the sustainability and workability of constructions.
Based on the combination of particle size analysis, identification tests and mechanical
tests performed on samples taken, the following results were obtained: (1) chipboard
concrete manufactured in factories are more favourable to workability than sustainability;
(2) the dosage D7 (Pm = 10.8 %; E/C = 1.5) may be recommended to the leaders of
factories due to its high resistance. Otherwise, it is advisable to obtain a workability and
durability while avoiding cement wastage, using dosage from D4 while increasing the
thickness of the cover.

Keywords:
Hollow concrete block, Yaoundé, Dosage, Compressive strength, Durability,
Workability.

Corresponding author mail: [email protected]

1. Introduction
Yaoundé, Capital of Cameroon, is both the capital of the Center region and the Mfoundi
division. Located on the southern Cameroonian plateau, in the middle of the equatorial
forest, it is about 200 km from the Atlantic coast. It stretches for 23.4 km on its main axis
(EW) and covers an urbanized area of nearly 18,000 ha.
According to Cameroon's national contribution report to the 3rd conference (2015) on
housing and sustainable development in urban areas (habitat III), the growth rate of the
Cameroonian population has been 2.6 % since 2010. 67 % of Cameroon's urban
population live in slums. The latter are experiencing an annual growth of 5.5%.
According to UN-Habitat (2007), the population of Yaoundé is estimated at 1.5 million
inhabitants, with the average annual growth rate in sharp decline, which stabilizes around
5.8 %. According to the urbanistic audit report of the city of Yaoundé (2001), the annual
growth rate was 9% in 1980, 6.7 % until the 1990s and 5.4 % in 2001. In 2003, 50 % of
Cameroon's population lived in urban areas. Among them, about 67 % live in so-called
spontaneous neighbourhoods. In Yaoundé, this proportion is between 70 and 80 %. These
neighbourhoods are characterized by an "over densification" of space.
With more than 500,000 settlements, the habitat survey in urban areas (1989) reveals that
56.9 % of houses are built with perforated concrete blocks against 21.8 % for mud brick
houses. In 1990, a study by the International Research Group Consultancy (IRGC, 1990)
noted the mediocrity of the construction industry in all cities in Cameroon. As part of a
study in concrete factories in Yaoundé, LIKIBY et al., (2009) corroborate this
observation by highlighting the inadequacies in the typology and quality of prefabricated
concrete blocks. In general, the production of concrete blocks is neither regulated nor
controlled; the dosages applied to date do not meet any established standard.
The strength of concrete - an artificial stone obtained by hardening a mixture of binders,
water and aggregates (sand, crushed rock or gravel) - depends on its formulation. This is
why many countries have established standards governing manufacturing based on their
climatic and socio-economic environment. In Cameroon, where no known standard is
applied, it is necessary to establish an endogenous one to govern this sector which is very
determining on the quality of housing. This study aims at contributing towards
establishing this standard by studying the mechanical resistance of perforated concrete
blocks manufactured in two sand pits in the city of Yaoundé.
The quality of concrete blocks depends on both the ingredients used and the working
environment. Thus, a postulate can be made by linking its resistance to three conditions
of realization: (1) setting time and the fineness of grinding of the cement, (2) the evolution
of the granularity and (3) the cement/sand (C/S) ratio in mixtures. In the event that one of
these parameters is not respected, the manufactured concrete easily undergoes shrinkage
and desiccation quickly leading to its destruction.
Of the two sand pits studied, a set of 28 concrete blocks was treated in order to obtain the
evolution of grain size, setting time, fineness of grinding and compressive strength of the
186
concrete blocks. The samples were obtained in 2 steps. The first is to obtain ranges by
considering the use of two types of cement: CPJ 35 and CPA 42.5R. The second step is
the dosage by varying the amounts of cement in the manufacture.
This work will be presented in two parts: methodology and results.

2. Methodology

2.1 Study area

Yaoundé, a metropolitan city in full urbanization, is between latitudes 3 ° and 5 ° North
and longitudes 11 ° and 12 ° East (Figure 1).

Figure 1. Yaoundé geological location map: (a) Ntem complex, (b) Yaoundé group.

It is part of the vast South Cameroonian plateau (BITOM, 1982; MBARGA BINDI, 1992;
ONGUENE MALA, 1993; YONGUE FOUATEU, 1986). The geology of the region
under consideration consists of Precambrian metamorphic rocks (migmatites) para and

187
ortho derivates (CHAMPETIER DE RIBES & AUBAGUE, 1956; NZENTI, 1987;
OWONA, 1998). The quarry sands used in construction come from the crushing of
paramigmatite massifs (Nkomkana, Nkomtou). The Yaoundé region, whose
hydrographic regime is of equatorial type (ONGUENE MALA, 1993) and the dendritic
hydrographic network, is made up of two large watersheds: the Nyong basin in the
southern part and the Sanaga basin in the northern part. The so-called Sanaga sand, sand
used in construction sites originates from the Mbam-Sanaga confluence in the locality of
Ebebda. The lithological substratum of this confluence consists of migmatite with
embrechitic facies (DUBREUIL et al., 1975). The fine sand used in the sites comes from
the localities of Monatélé and Olembé. The soil profile in these villages is made up of
three horizons. So, from bottom to top: gravels, sands and clays.

2.2. Materials and method in construction sites

2.2.1. Used materials

After a preliminary investigation in the city of Yaoundé, two sand pits (concrete factory
site) were selected: one in Mvan and the other in Obili. Subsequently, the concrete blocks
were manufactured according to a range of seven different dosages, including the D3
dosage used by the manufacturers on the construction site (Table 1). The concrete mixes
are obtained with a mixture of sand (Sanaga sand + quarry sand + fine sand), cement
(imported or from Cimencam) and water. The imported cement has the following
characteristics: Ordinary Portland Cement, grade 42.5R: CPA, china standard GB175, net
weight: 50 kg. The cement from the cement factories of Cameroon (CIMENCAM) has
the following characteristics: Portland cement CPJ35, 35-NF, average P15-301, weight:
50 kg.

2.2.2. Manufacture of concrete blocks

On site, in addition to the dosages of the mixes used on site, concrete blocks of 15 (20 x
15 x 40 cm) were manufactured with six different dosages and in two different sets of
cement.
- The technicians at Mvan, to produce 60 blocks of 15, make the following D3 dosage:
2 wheelbarrows of Sanaga sand + 1 wheelbarrow of quarry sand + 1 wheelbarrow of
fine sand + 10 buckets of 10L of water + 1 bag of cement. That means: 600 kg (400
L) of sand + 100 kg (100 L) of water + 50 kg (20L) of cement.
- The technicians in Obili, to produce 75 blocks of 15, make the following D3 dosage:
2 wheelbarrows of Sanaga sand + 2 wheelbarrows of quarry sand + 1 wheelbarrow of
fine sand + 10 buckets of 10L of water + 1 bag of cement. That is: 750 kg (500L) of
sand + 100 kg (100L) of water + 50 kg (20L) of cement.
To manufacture the ranges of concrete blocks with local cement or with imported cement,
the dosages according to the following formulas were used:
188
Mc = Pm.Ms (1)
Vc = Pv.Vs (2)
with Mc = mass of cement, Vc = volume of cement, Pm = mass percentage, Pv =
volumetric percentage, Ms = mass of sand, Vs = volume of sand. The quantities of water
and sand in our samples were set respectively at: Mvan: 2.1 L and 8.4 L; Obili: 1.68 L
and 8.4 L. Subsequently, the amounts of cement varied according to the dosages in Table
1, with E = water; C = cement; W/C = water to cement mass ratio; Di = dosage numerated
i; the other parameters are defined in the text. The fresh breeze blocks are wetted from
time to time to avoid shrinkage after drying. The 28 perforated concrete blocks
manufactured, as well as the cements and sands used were transported to laboratory
LABOGÉNIE (National Laboratory of Civil Engineering, Yaoundé) for testing.

Table 1. Presentation of the different dosages.

Mvan Obili
Dosage

Vc (L) Mc (kg) Pv (%) Pm (%) E/C Vc (L) Mc (kg) Pv (%) Pm (%) E/C
D1 15.6 39 3.9 6.5 2.7 16 40 3.2 5.3 2.5
D2 18.21 45.5 4.6 7.6 2.2 19.2 48 3.8 6.4 2.1
D3 20 50 5.0 8.3 2 20 50 4 6.7 2
D4 20.8 52 5.2 8.7 1.9 22.4 56 4.5 7.5 1.8
D5 22.1 55.3 5.5 9.2 1.8 25.6 64 5.1 8.5 1.6
D6 23.4 58.5 5.9 9.8 1.7 28.8 72 5.8 9.6 1.4
D7 26 65 6.5 10.8 1.5 32 80 6.4 10.7 1.25

2.3. Laboratory work

As the study was based on the strength of perforated concrete blocks, the work took place
in concrete manufacturing sites and at the National Civil Engineering Laboratory
(LABOGÉNIE). In the laboratory, four series of tests were carried out.

2.3.1. Particle size analysis of sands by sieving and sedimentometry

In order to know the ranges of variation of the diameters of the grains of sand and their
weight, particle size analyzes by sieving and by sedimentation were carried out according
to standards NF P 18 - 560 and NF P 94 - 093 respectively.

2.3.2. Sand equivalent (ES)

The sample washed under standard NF P 18 - 598 conditions is left at rest. Subsequently,
the heights H1 and H2 of the clean sand containing fine elements and only clean sand
respectively are measured. By definition:
ES = (H2 / H1) x 100 (3)

189
2.3.3. Tests on cements
The setting test was carried out using the Vicat apparatus and a cement paste of normal
consistency according to standard EN 196 - 3. Identification tests are used to acquire data
on particle size analysis (GA), Blaine specific surface area (SSB, standard EN 196 - 6),
specific weight (PS) and bulk density (DA). The different values obtained are shown in
Table 2.

2.3.4. Simple compressive strength test on concrete blocks and cements

The compressive strength of the concrete blocks or hardened cement samples was carried
out using a press according to standard EN 196 - 1.

3. Presentation and interpretation of the results

3.1. Cements
3.1.1. Setting time
The start of setting corresponds to the moment when an increase in the viscosity or
stiffening of the cement paste is observed. It is 2h 45mn for the CPJ 35 cement and 2h
55mn for the CPA 42.5R cement. The mixture sets when after a while the hydrate crystals
become more and more important. The mixture having changed viscosity stiffens. The
CPJ 35 cement end of setting time is 4h 07mn, and of CPA 42.5R cement it is 4h 33mn.
The setting times of the cements used for making the blocks are therefore 1h 22mn for
the CPJ 35 cement and 1h 38mn for the CPA 42.5R cement.
It is necessary to know the beginning and the end of setting of the cement pastes, in order
to be able to assess the time available for the correct placement of the mortars or
concretes. Depending on their resistance class (VENUAT, 1989), the standards specify a
minimum setting time, which at a temperature of 20 ° C, is 1 hour 30 minutes for class
35 and 45 cements. The cements are more or less close (1h 30 ± 8mn) from the minimum
value. In general, the setting time for cements is up to 2h 30mn and 3h. Knowing on the
one hand that heat accelerates the hardening and that the hardening promotes desiccation,
and on the other hand that cold tends to stop the process of hydration of the cements, the
setting times obtained from these cements are acceptable for the manufacture of sand
concrete, even more that of CPA 42.5R cement.

3.1.2. Fineness of grinding

The cement identification tests gave the results in Table 2.

Table 2. Identification test data.

Type of cement AG (% 80µm) PS (T/m3) SSB (cm2/g)
CPJ 35 3,8 3,05 3038
CPA 42,5R 4,1 2,95 3177

190
The SSB characterizes the fineness of the grind. The domain of definition of this varies
according to the authors. According to NEIL & REVINDRA (1996), practical values are
between 3250 and 3850 cm2/g. VENUAT (1989) proposes values between 2900 and 3500
cm2/g. The values of SSB are included in the range of values according to the
classification of Venuat and close to the lower limit according to Neil and Rev Will. CPA
42.5R cement is a little better than other cement.
3.1.3. Mechanical properties of cements
The summary of the results of the mechanical tests is given in Table 3.
Table 3. Average data of the mechanical strengths of cements at 28 days.
Type of cement Flexural strength (MPa) Compressive strength (MPa)
CPJ 35 6.3 80.2
CPA 42.5R 6.9 78.7

According to British Standard requirements for strength of the principal Portland cements
(NEIL & REVINDRA, 1996), the values of the minimum strengths at 28 days are variable
according to their class. Class 35 has nominal values of 35 MPa and 55 MPa for the lower
and upper limits respectively. The 42.5R class has values of 42.5 MPa and 62.5 MPa
respectively for lower and upper nominal limits.
According to the NF P 18-360 standard, current directives require minimum 28-day
strengths of 4 MPa in flexural tension and 30 MPa in compression. The values of tensile
and compressive strengths of the cements used to make our concretes are much higher
than these minimum values. This gives these cements (CPJ 35 and CPA 42.5R) a very
good ability to consolidate the granular skeleton of concretes. They are therefore likely
to have a good influence on the quality of concrete resistance.

3.2. Sands

3.2.1. Sand equivalent

The degree of cleanliness of sand is characterized by a visual (ESv) or piston (ESp)
numerical value. This degree influences the quantities of water in the dosages, hinders
the adhesion of the binder and thus creates the risk of cracks in hardened concrete. In the
present case, its values are recorded in table 4.

Table 4. Sand equivalent values (ES).

Fine sand Sanaga sand Quarry sand
ES
Mvan Obili Mvan Obili Mvan Obili
ESv 80.7 80.7 95.8 91.6 96.9 95
ESp 71 71 91 88.7 99.5 97.5

DREUX (1985) qualifies concretes according to the degree of cleanliness of the sands
used. In view of its qualifications, these sands are fine, clean, with a low percentage of
191
fine clayey (75 ≤ ESv < 85) and perfectly suitable for quality concretes, meaning common
concretes of relatively high resistance (ESV ≥ 75). The Sanaga and quarry sands are very
clean sands (ESv ≥ 75), suitable for high-strength quality concretes, meaning exceptional
concretes (ESv ≥ 80).

3.2.2. Granulometry
The results of the particle size analysis have been graphically translated by the particle
size curves in Figures 2, 3 and 4 (Note: X and Y-axes legend are in French).
Sieving Sedimentation
Gravel Coarse Sand Fine sand Silt Clay
Cumulative sieve (%)

Aggregate dividing
line

Granular reference
curve

0.1 0.01 0.001

Particle diameter (mm)

Figure 2. Grain size curves of the sands from Mvan quarry.

Cumulative sieve (%)

Aggregate dividing
line

Granular reference
curve

0.1 0.01 0.001

Particle diameter (mm)

Figure 3. Grain size curves of the sands from Obili sand pit.

192
 Cumulative sieve (%)

0.1 0.01 0.001

Particle diameter (mm)

Figure 4. Particle size curves of the different sands.

OAB reference particle size compositions are plotted there. According to Simonnet's rule
(BAUDIN, 1986), the coordinates of these different points are:
- 0 % on the ordinate and 0.08 mm on the abscissa for point O.
- 100 % on the ordinate and Dmax on the abscissa for point B.
- For point A, the coordinates are defined as follows: x = Dmax / 2 if Dmax ≤ 20 mm
and y = 50 - √Dmax + Ks (with Ks = 8 for normal rolled sand; Ks = 6 for crushed
normal sand).
The values of the maximum diameter (Dmax) correspond to the diameter of the sieve at
D closest to 92 % of the particle size of the largest aggregate. The diameter d of the
particle size classes corresponds to the diameter of sieve at d closest to 8 % of the fine
particle size. Thus, from the shape of the grain size curves, the following information can
be drawn:
- The different granular classes (D / d) of sands used for concrete are as follows:
SF-Mv : 1,25/0,05 SF-Ob : 2,5/0,01
SS-Mv : 2,5/0,315 SS-Ob : 5/0,63
SC-Mv : 2,5/0,02 SC-Ob : 5/0,005
This allows us to obtain the values of the coordinates of A: (1.25; 55.9) at Mvan and (2.5;
55.3) at Obili.
- Through the dividing line of the aggregates (95 % - 5 %), we have the greater or lesser
proportion of the elements of the different granular classes: 40 % fine sand and 60 %
coarse sand at Mvan; 34 % fine sand and 56 % coarse sand in Obili.
- The granulometry of the sands is continuous and increasing in the case of sands and
more or less static in the case of silts and clays.
These different results make it possible to retain that the sands used are suitable for the
manufacture of micro-concretes. According to DREUX (1985), they should be used for

193
making small construction parts. As for the fineness modulus (MF), below 2.2, sand has
a majority of fine and very fine elements. Above the value of 2.8 there is a lack of fine
elements. Between 2.2 and 2.8, the MF defines a good concrete sand. The fine sands used
at Mvan (MF = 1.658) and at Obili (1.792) mainly have fine elements. Mvan Sanaga sand
(MF = 2.887), Obili Sanaga sand (MF = 3.521) and Obili quarry sand (2.854) lack fine
elements. Only the Mvan quarry sand (MF = 2.49) seems to suit as a good concrete sand.

3.3. Compressive strength Rc28 of perforated concrete blocks

The compressive strength measured at 28 days of age (Rc28) of the blocks guides the
choice of dosage (see APPENDIX 1) in the best indicated mixtures [sand (S) + cement
(C) + water (E)].

3.3.1. Rc28 depending on C/S

Figures 5 and 6 respectively show the change in Rc28 as a function of volumetric and
mass percentages. We observe an increasing parallelism of the resistance curves. The
curves of CPJ 35 cement are above those of other cement regardless of location. Likewise,
the curves of the Mvan samples are above those of Obili. The crossing point being due to
the fact that the dosages on site (D3) are different in the two sand pits. We can deduce:
- The CPJ 35 cement is better in this respect than the CPA 42.5R cement;
- The on-site dosages D3 at Mvan (Rc of CPJ 35 = 2.93 MPa and Rc of CPA 42.5R =
2.97 MPa) are better than those of Obili (Rc of CPJ 35 = 2.42 MPa and Rc CPA 42.5R
= 2.15 MPa).
- The compressive strength increases with the C/S ratio, i.e. with increasing cement. In
the present case, the maximum value (5.50 MPa) is that of the D7 dosage (Pv = 6.5
%, Pm = 10.8 %, W/C = 1.5) with the CPJ 35 cement at Mvan.
Rc (MPa)

Figure 5. Rc28 according to the volumetric percentage.

194
 Rc (MPa)

Figure 6. Rc28 as a function of mass percentage.

3.3.2. Rc28 according to the mass ratio E/C

Figure 7 shows the compressive strength as a function of the W/C mass ratio. It reveals
the parallelism of the curves as well as the decrease in Rc as a function of the E/C ratio.
Once again, the curves of CPJ 35 cement are slightly above those of CPA 42.5R cement.
We can deduce that:
- CPJ 35 cement is better in this respect than CPA 42.5R cement;
- the Rc decreases with the evolution of the E/C ratio; in other words, the compressive
strength decreases when the amount of water increases compared to the cement. In
the case of this study, the minimum value (1.16 MPa) is that of the D1 dosage carried
out in Obili (W / C = 2.5; Pv = 3.2 %; Pm = 5.3 %) with the CPA cement 42.5R.
Rc (MPa)

1.5 2.5

Figure 7. Rc28 as a function of the E/C mass ratio.

195
3.3.3. Dosage of cement, water and aggregate
The results obtained pose the problem of the most suitable dosage likely to be fixed. With
the above, we realize that the greater resistance corresponds to the greater dosage of
cement, that is, per 100 L of water, 600 kg of sand and 80 kg of cement. This is in line
with the experiences of DREUX (1985) and BERGA et al., (1994). Thus, for a defined
quantity of cement, the Rc will be low or high depending on whether the quantity of
aggregates is respectively large or low, and depending on whether the quantity of water
is high or not respectively.
As a practical matter, it is advisable to use the following mixture whenever possible in
concrete factory sites:
5 wheelbarrows of sand [(2 of Sanaga sand + 2 of quarry sand + 1 of fine sand) + 1 and a
half and ¼ of a 10L bucket of cement + 12*10L buckets and 2 L of water]. Mixing the
sands helps reduce the pores in the concrete.

3.3.4. Resistance, workability and durability

The study of the composition of concrete almost always involves jointly investigating its
two essential qualities of strength and durability. In fact, the workability of concrete is its
quality based on its tightening in good conditions. It depends on the quality of the
aggregates and the cement dosage (C/E). Durability, on the other hand, expresses the
quality of concrete to be durable. According to DREUX (1985), DREUX & FESTA,
(1998) and VENUAT (1989), the latter are closely related, but they vary in the opposite
direction. So according to APPENDIX 2, we can conclude that:
- experienced concretes are more suitable to workability than to durability;
- to improve resistance, it would be necessary either to:
- increase the cement dosage;
- increase the G/S ratio (gravel/sand);
- slightly increase the discontinuity of the granularity;
- increase the maximum diameter (Dmax) of the aggregates.
On the other hand, durability evolves with resistance. Usually the choice of concrete type
is influenced by durability. NEIL & RAVINDRA (1996) propose the minimum durability
qualities of perforated concrete blocks. According to them, the latter must have at least a
minimum compressive strength of 3.5 to 7 MPa for the walls of buildings. Compared to
this value, the Rc of the materials of the sandpits studied (D3) are well below the required
durability threshold. It would be necessary to have at least the Rc of D7 at Mvan (5.50
MPa for the CPJ 35 cement and 5.33 MPa for the CPA 42.5R cement) to be at the lower
limit. Owners of sand pits who complain about the high blow of the cements, can increase
the Rc as shown above. But this method is less favourable to workability.
The problem thus posed can be solved in another way by considering the alterability of
the components. In general, it can be assumed that the aggregates are inert and that it is
the cement that risks being altered in the long or short term depending on climatic
196
conditions (humidity and temperature). To overcome this handicap, masons can improve
durability by increasing the quality and thickness of the facing concrete during plastering.
In accordance with the statements of Professor Campus in the work of DREUX (1985),
it is proposed to improve durability by making a cover of minimum thickness:
- e ≤ 1 cm, for any structure not exposed to bad weather;
- e ≤ 2 cm, for any structure exposed to bad weather.
At this level, complaints of an economic nature would emanate much more from the
consumer.

4. Conclusion
The objective of this study as defined at the beginning was to better identify the dosages
of the components of the concrete manufactured in the city of Yaoundé, with the intention
of improving the durability of the concretes manufactured at a lower cost. Thus, the
results obtained made it possible to retain that:
- CPJ 35 and CPA 42.5R cements according to the setting time values (2h 45mn and 2h
55mn), SSB grind fineness (3038 cm2/g and 3177 cm2/g) and compressive strength (80.2
MPa and 78.7 MPa) are good for concrete. A low shrinkage tendency is noted, which is
why it is always advisable to sprinkle the manufactured concrete blocks.
- Granular classes are of the order of 2.5/0.01; 5/0.063; 5/0.005, respectively for fine sand,
Sanaga sand and quarry sand. Depending on the fineness modulus, no sand alone
constitutes good concrete sand. That’s why they have to be mixed.
Regarding perforated concrete blocks (breeze blocks):
- The compressive strength of the dosages (D3) generally carried out in sand pits (< 3
MPa) hardly testifies to a good durability.
- The best indicated dosage for the concrete of the walls of buildings is that of D7 to
Mvan: Pv = 6.5 %; Mp = 10.8 %; E/C = 1.5. This dosage gives the concrete blocks a
compressive strength of 5.50 MPa and 5.33 MPa respectively for the cement CPJ 35
and CPA 42.5R. It is therefore a question of advising the owners of sand pits to use
mixtures with at least the following quantities: 5 wheelbarrows of sand (2 of Sanaga
sand + 2 of quarry sand + 1 of fine sand) + 1 bag and half and ¼ bucket of 10L of
cement + 12 buckets of 10L and 2L of water.
Owners of sand pits who complain about the high cost of cements will be able to improve
strengths, durability and workability: either by slightly increasing the gravel/sand ratio,
or by increasing the cement dosage to D4 Mvan (Pm = 8.7 %; Pv = 5.2%; W/C = 1.9; RC
28 = 4.18 MPa with CPJ 35 cement), but in this case the level of durability decreases;
masons will be able to overcome this handicap by increasing the thickness of the
plastering (e ≤ 2 cm). This resulted in good durability and good workability. Thus, in
concrete block factories, owners must never go below the following formulation: 5
wheelbarrows of sand (2 of Sanaga sand + 2 of quarry sand + 1 of fine sand) + 1 bag and
2.5L of cement + 10 buckets of 10L and 1.25 L of water.
197
These dosages thus proposed are indicated for hollow blocks of walls elevation.
Formulations from strength D7 are the most suitable. But if the means are more limited,
we could go down to D4 and increase the thickness of the blanket.
However, to obtain these dosages, the amounts of water and sand were fixed. Work on
the variation of water within a known cement/sand dosage has not been done. Doing so
would be necessary to improve strength in the dosages offered and possibly lead to better
durability at lower cost.

5. Bibliography
BAUDIN D. (1986). Cours de béton pour stagiaire. LABOGÉNIE.
BERGA P., MARMOTTAN C., BONNET G., COSTE DOAT C., JEAUGET P.,
LESTAGE G. (1994). Bétons de sable. Projet national recherche/développement
SABLOCRETE, Presse de l’École Nationale des Ponts et Chaussées, LRPC de Bordeaux,
SERMA.
BITOM L.D. (1982). Contribution à l’étude structurale des séquences et facies
d’altération des gneiss migmatitiques de Yaoundé. Mémoire Maîtrise, Université de
Yaoundé I.
CHAMPETIER DE RIBES G., AUBAGUE M. (1956). Carte géologique de
reconnaissance, Yaoundé Est, Ech. 1/500 000, Paris.
DREUX G. (1985). Nouveau guide de béton. 4e éd. Eyrolles.
DREUX G., FESTA J. (1998). Nouveau guide du béton et de ses constituants. Ed.
Eyrolles, 416p, ISBN 978-2-212-10231-4.
DUBREUIL P., GUISCAFRE J., NOUVELAT J.F. et OLIVRY J.C. (1975). Le bassin
de la rivière de Sanaga. Ed. ORSTOM, Col. Monographies hydrologiques, n°3, Paris,
350p.
INFO CIMENTS. Retrieved 9 May 2021 on www.infociment.fr
IRGC - INTERNATIONAL RESEARCH GROUP CONSULTANCY (1990).
Evaluation des besoins réels en logement au Cameroun – Indicateurs de la politique
nationale de logement – stratégies pour une politique nationale de l’habitat. Ministère
de l’Urbanisme et de l’Habitat (MINUH), République du Cameroun.
LIKIBY B., SOFACK J.C., TAMO T.T., KAMGANG K.B.V., MELO C.U. (2009).
Qualité des parpaings préfabriqués en béton dans l’agglomération de Yaoundé,
Cameroun. International Journal of Pure and Applied Science, vol.2, n°1, pp. 19-26.
MBARGA BINDI A. (1992). Processus de reconstruction de la forêt mésophile
guinéenne, cas du secteur forestier de la région de Yaoundé (Cameroun). Thèse
Université Paris Sud, Centre d’Orsay, 167p.
NEIL J., RAVINDRA K. (1996). Civil engineering materials. Fifth edition, D.HIR.
NZENTI J.P. (1987). Pétrogenèse des migmatites de Yaoundé (Cameroun), élément pour
un modèle géodynamique de la chaîne panafricaine Nord-équatoriale. Thèse de doctorat,
Université de Nancy I, 147p.
198
ONGUENE MALA (1993). Différentiation pédologique dans la région de Yaoundé
(Cameroun), transformation d’un sol rouge ferralitique en un sol à horizon jaune en
relation avec l’évolution du modélé. Thèse de doctorat. Université de Paris VI, Spécialité
Pédologie, 254p.
UN-HABITAT (2007). Profil urbain de Yaoundé. Programme des Nations unies pour les
Établissements Humains, HS/960/07F, Printing: UNON, Publishing Services Section,
Nairobi, Kenya. ISO 14001:2004-certified, ISBN 978-92-113-1959-0
OWONA S. (1998). Contribution à l’étude pétrostructurale de la signature
morphologique des métamorphytes du Sud de Yaoundé. Mémoire Maîtrise, Université de
Yaoundé, 80p.
VENUAT M. (1989). La pratique des ciments, mortiers et béton. Edition Le Moniteur.
YONGUE FOUATEU R. (1986). Contribution à l’étude pétrologique de l’altération et
du cuirassement ferrugineux des gneiss migmatitiques de la région de Yaoundé. Thèse de
doctorat de 3e cycle, Université de Yaoundé.

6. Appendices

Note: abbreviations used in the paper: SF = Fine Sand; SS = Sanaga sand; SC = Quarry
sand; Mv = Mvan; Ob = Obili.

APPENDIX 1. Compressive strength (in MPa) depending on the dosage.

Mvan Obili
Dosage
CPJ 35 CPA 42,5R CPJ 35 CPA 42,5R
D1 2 1,63 1,36 1,16
D2 2,33 2,16 1,88 1,62
D3 2,97 2,93 2,42 2,15
D4 4,18 4,16 3,01 2,93
D5 4,63 4,55 3,91 3,78
D6 5,10 4,93 4,43 4,25
D7 5,50 5,33 4,87 4,60

APPENDIX 2. Quality of various parameters depending on workability and strength

researched for concrete, (DREUX, 1985).
Concrete composition For good
For good resistance
factors workability
Finesse of sand Rather fine Rather coarse
G/S ratio To decrease To increase
Water dosage To increase To decrease
Continuous Discontinuous slightly
Granularity
preferable preferable
Maximum size of
Rather small Rather strong
aggregates
199
200
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Mechanical properties of the biomass ash-based binder

Désiré NDAHIRWA 1, Hélène LENORMAND 1, Hafida ZMAMOU 1,

Nathalie LEBLANC 1

1. UniLaSalle, Univ. Artois, ULR7519 - Transformations & Agro-ressources, Normandie

Université, 3 rue du Tronquet, 76130 Mont-Saint-Aignan, France.
[email protected]

Abstract:
The prompt demand in cement manufacturing and applications is one of the main sources
of greenhouses gases emissions. Besides, the transition from fossil to renewable energy
has led to increased production of biomass ash, the disposal of which represents a serious
environmental concern. This study was undertaken as the first phase of a broad research
project considering biomass ash (BA) valorization in lightweight bio-based concrete. The
objective of the first phase was to optimize the binding matrix mainly comprising of
biomass ash. In the experimental program, small amounts of ordinary Portland cement
(OPC) and natural hydraulic lime (NHL) varying from 5%, 10% and 20% (by BA mass)
were added to the mixture as partial BA replacement. After 7,14 and 28 days of curing in
laboratory ambient conditions, the compressive and flexural strengths were tested. The
results showed that the incorporation of OPC and NHL effectively improved the
mechanical properties of the BA1, BA3 and BA4 pastes. In fact, at 28days, the BA
mixture containing OPC exhibited higher compressive and flexural strengths compared
to that of NHL. Excluding the reference mixtures (100%NHL and 100%OPC), the highest
compressive and flexural strengths being equal to 5.6 MPa, and 2.7 MPa, respectively
were achieved by BA3-OPC20. On the contrarily, the lowest compressive and flexural
strengths were observed on BA2-NHL and were less than 0.1MPa.

Keywords:
Biomass ash, Lightweight bio-based concrete, Compressive and flexural strengths.

Corresponding author mail: [email protected]

1. Introduction
Fly ash is a byproduct discharged from the furnace as a result of coal and wood
combustion at thermal power plants (LIN et al., 2020). At present, only 30% of biomass
fly ash (BFA) is recycled and re-used in different applications whereas 70% remains
landfilled (FORT et al., 2021). The recognition of human activities role in degradation of
the environment has led to implementation of tighter regulations. Hence, some coal-fired
power stations are transitioning from coal to biomass to produce green energy. In the
future, coal fly ash (CFA) is most likely to be replaced by biomass fly ash (FORT et al.,
2021).
The application of BFA in mortar/concrete has promptly increased in recent years. BFA
have been used as a supplementary cementitious material and as a replacement of
currently used pozzolans in construction materials production (FORT et al., 2021; PARK,
2020). Besides, due to high potassium (K) content and high alkalinity, BFA are used as
soil fertilization agent OBERNBERGER & SUPANCIC, 2009; LIN et al., 2020;
ANDAVAN & PAGADALA, 2020). BFA can be used for CO2 capture (DINDI et al.,
2019) and also as an adsorbent to remove soil and water contaminants (PARK, 2020). In
the United Kingdom (UK), the recycling of coal fly ash in production of blended cement
contributes to the reduction of about 600 000 tons of CO2 per year (SAMAD & SHAH,
2017). Besides, fly ash is widely available at low cost compared to conventional cement
and concrete raw materials. Eventually, the recycling of fly ash does not only have
environmental benefits but also economic profits.
The chemical composition of BFA depends on biomass species, source and combustion
technology and storage conditions (OMRAN et al., 2018; LIN et al., 2020). The fly ash
particle size distribution and shape influence the hydration of mortar/concrete. For
example, CFA with finer particles have better reactivity and pozzolanic index compared
to coarser CFA (MYADRABOINA et al., 2017). The angular form and fineness of CFA
particles enhance the density and durability of mortar/concrete (ALQAHTANI et al.,
2021). According to CHEN et al., (2016), ultrafine fly ash (grain size <60µm)
significantly improve the compressive strength, bulk density, thermal conductivity and
linear shrinkage of the lightweight insulation material. The compressive strength of CFA
mortar/concrete is generally lower in early ages but increases with time due to CFA slow
pozzolanic reactions (CHO et al., 2019).
The partial replacement of cement with CFA at 5% by weight increased the compressive
strength of unmodified CFA cement mortar from 38.3 MPa to 60.5 MPa and from 36.8
MPa to 63.2 MPa for surface-modified CFA cement mortar at the age of 2days and
28days, respectively (MIRONYUK et al., 2019). However, (KAYALI & SHARFUDDIN
AHMED, 2013) produced a concrete by replacing cement with different fly ash contents
20%, 40%, 60% and 80%. The mechanical performance results showed a decrease in
compressive strength, tensile strength and modulus of elasticity of fly ash concrete with
increase in fly ash replacement level. A similar tendency was reported by
202
(MYADRABOINA et al., 2017) who pointed out that incorporation of CFA at higher
content (> 40%) has adverse impact on the mechanical performance of concrete.
The comparison between compressive strengths (Rc) of cement mortars incorporated with
CFA from 20% to 60% contents as cement replacement showed that the mortars
containing between 20% and 30% CFA had comparable results as the reference mortar,
whereas a significant reduction in Rc value was observed on mortars containing 35% to
60% CFA content (ALQAHTANI et al., 2021). Nevertheless, AREZOUMANDI &
VOLZ (2013), reported that the addition of CFA class C in the range of 50% to 70% as
cement replacement slightly increased the mechanical performance of the high-volume
fly ash concrete (HVFAC). At 28days, the compressive strength (Rc) of the reference
concrete being equal to 29 MPa, for HVFAC-50 was 32.4 MPa and 30.7 MPa for
HVFAC-70. The flexural strengths (Rf) were 3.2 MPa, 3.5MPa and 3.3MPa for the
reference concrete, HVFAC-50 and HVFAC-70, respectively.
Previous findings associated with fly ash applications in construction materials tend to
focus on their feasibility in the production of blended cement and as partial cement
replacement in mortars and concrete. However, the replacement levels of fly ash remain
limited to avoid a probable compromise in mechanical performance of the resulted
composite. Therefore, the present study investigates the mechanical properties of an
alternative binder consisting of high-volume biomass ash, which could be used in non-
structural lightweight mortar/concrete. For comparison purposes, conventional binders
such as ordinary Portland cement (OPC) and natural hydraulic lime (NHL) were added to
the mixture as BA partial replacements.

2. Materials and methods

The collected biomass ashes were originated from wood chips and pallets combustion.
Four different biomass ashes comprising of fly ash BA1, BA3, BA4 and bottom ash BA2
were collected from biomass-fired power plant and local wastes treatment center. BA1
and BA2 were dried for 48h in oven at 105°C until the constant weight. After removal of
the moisture, a hammer was used to grind BA1 and BA2 into smaller particles diameter
≤ 500 µm. Given that BA3 and BA4 were initially dry and had fine particles, they were
directly used in the mixtures without any pretreatment. Natural hydraulic lime (NHL3.5)
and ordinary Portland cement (OPC) CEMII 32.5N used in the study, were purchased
from a local supplier.
The mixture proportions of BA-NHL and BA-OPC pastes are presented in Table 1.

203
Table 1. Mixture proportions of BA-NHL and BA-OPC pastes.
Mixture W/BA Mixture W/BA
BA (%) NHL (%) BA (%) OPC (%)
designation ratio designation ratio
NHL100 100 0 0.5 OPC100 100 0 0.5
BA1-NHL0 100 0 0.7 BA1-OPC0 100 0 0.7
BA2-NHL0 100 0 0.7 BA2-OPC0 100 0 0.7
BA3-NHL0 100 0 0.5 BA3-OPC0 100 0 0.5
BA4-NHL0 100 0 0.5 BA4-OPC0 100 0 0.5
BA1-NHL5 95 5 0.7 BA1-OPC5 95 5 0.7
BA2-NHL5 95 5 0.7 BA2-OPC5 95 5 0.7
BA3-NHL5 95 5 0.5 BA3-OPC5 95 5 0.5
BA4-NHL5 95 5 0.5 BA4-OPC5 95 5 0.5
BA1-NHL10 90 10 0.7 BA1-OPC10 90 10 0.7
BA2-NHL10 90 10 0.7 BA2-OPC10 90 10 0.7
BA3-NHL10 90 10 0.5 BA3-OPC10 90 10 0.5
BA4-NHL10 90 10 0.5 BA4-OPC10 90 10 0.5
BA1-NHL20 80 20 0.7 BA1-OPC20 80 20 0.7
BA2-NHL20 80 20 0.7 BA2-OPC20 80 20 0.7
BA3-NHL20 80 20 0.5 BA3-OPC20 80 20 0.5
BA4-NHL20 80 20 0.5 BA4-OPC20 80 20 0.5

To prepare the mixtures, dry BA1, BA2, BA3 and BA4 were each mixed with NHL3.5
or OPC 32.5N and tap water using automatic cement mixer. The W/BA ratios were fixed
at 0.5 for the BA3 and BA4 pastes and 0.7 for the BA1 and BA2 pastes. For comparison
purposes, BA1, BA2, BA3 and BA4 were each partially replaced by 5%, 10% and 20%
of NHL and/or OPC. A proctor method was used to compact the BA mixtures with a
compacting energy of about 600 kJ/m3. The BA paste specimens were casted in steel
molds. After 24h, the specimens were demolded and cured in the laboratory at ambient
temperature and humidity conditions until the desired testing age.

Moreover, the mechanical performance tests were conducted using the Shimadzu
universal testing machine. The flexural strength (Rf) test was carried out on prismatic
BA1, BA2, BA3 and BA4 pastes specimens, dimensions 40 × 40 × 160 𝑚𝑚3 .The
obtained half prisms were cut into 40 × 40 × 40 𝑚𝑚3 cubes on which the compressive
strengths (Rc) were measured at the curing ages of 3, 7, 14 and 28 days. At least three
cubes were made for every mixture and the Rc test was conducted on them over an area
of 40 × 40 𝑚𝑚2 at the loading rate of 5 mm/min. The average of three test results was
considered as Rf and Rc values of the paste specimen.

204
 100

80
OPC
Percent passing (%)

60 NHL
BA1
40 BA2
BA3
20 BA4

0
0.1 1 10 100 1000
Grain size (µm)

Figure 1. Particle size distribution of the OPC, NHL and BA.

The particle size distribution of the materials was measured using laser diffraction
method. As shown on Figure 1, OPC and NHL particles are finer than BA1, BA2, BA3
and BA4. The particles of BA1 and BA2 are coarser compared to BA3 and BA4 ones.
The graph shows that most particles of OPC, NHL, BA3 and BA4 passes through 100 µm
sieve, whilst only about 50% of BA1 and 40% of BA2 particles are able to get through
similar sized sieve. Generally, almost all particles of the raw materials used in this study
range between 0.1 µm and 500 µm.
The chemical composition the OPC, NHL and BA, was analyzed by inductively coupled
plasma-atomic emission spectrometry (ICP-AES) as listed in Table 2. An important
quantity of CaO was observed in examined BA, which corresponds to the previously
reported results (LOTHENBACH et al., 2011; FUSADE et al., 2019).

Table 2. Chemical composition of OPC, NHL and BA.

Compound OPC NHL BA1 BA2 BA3 BA4
SiO2 22.46 24.57 14.52 33.14 8.97 6.10
Al2O3 4.62 0.48 3.84 7.22 2.26 2.00
Fe2O3 4.27 0.29 1.96 4.41 2.51 3.99
CaO 61.17 72.55 43.67 34.86 32.00 34.16
MgO 1.41 0.82 5.10 4.46 4.85 5.20
TiO2 0.28 0.04 0.28 0.62 0.21 0.17
MnO 0.07 0.01 0.67 0.73 1.71 2.79
P2O5 0.32 0.05 4.42 3.39 4.94 4.84
K2 O 1.16 0.10 11.26 8.26 14.72 16.22
SO3 4.25 1.09 14.28 2.91 27.84 24.53

205
The bulk density results are given in Table 3. It can be seen that OPC and BA2 had high
bulk densities corresponding to 1303 kg/m3 and 1035 kg/m3, respectively while the lowest
values were observed on BA3 and BA4, equaled to 374 kg/m3 and 379 kg/m3,
respectively.

Table 3. Bulk density of OPC, NHL and BA.

Raw materials OPC NHL BA1 BA2 BA3 BA4
3
Bulk density (kg/m ) 1303 715 631 1035 374 379

5
Skeletal density (g/cm³)

0
OPC NHL BA1 BA1* BA2* BA2 BA3 BA4
Raw materials

Note: BA1* and BA2*: uncrushed; BA1 and BA2: crushed

Figure 2. Skeletal density of OPC, NHL and BA.

The skeletal densities of OPC, NHL and BA were computed using a Pycnomatic Evo
pycnometer, which uses Argon. As shown on Figure 2, the skeletal densities of the raw
materials vary from 2 to 4.5 g/cm3. The crushed biomass ashes BA1 and BA2 exhibited
higher skeletal density values compared to uncrushed biomass ashes BA1* and BA2*.
The increase in skeletal density of BA1* and BA2* can be attributed to the grinding
process which led to the reduction of closed pores in biomass ash particles.

3. Results and discussion

3.1 Flexural and compressive strength of BA-NHL pastes

The flexural and compressive strength results of the pastes containing BA-NHL are
shown in Figure 3 and Figure 4. At 7days and 14days, the flexural strength values of all
BA-NHL specimens varied from 0.1 MPa to 1.1 MPa. Moreover, the results showed that
at 28days, the flexural strength of BA3 and BA4 mixtures gradually increased, except for
BA1 and BA2 mixtures that decreased. Indeed, the highest flexural strength was achieved
by BA3-NHL20 (1.7 MPa) and followed by BA4-NHL20 (1.4 MPa) at 28days. As far as
BA3 and BA4 mixtures are concerned, flexural strength increased with increasing NHL
206
content and curing age. In addition, the impact of NHL on BA1 and BA2 mixtures
remained negligible.

2 2
Flexural strength (MPa)

Flexural strength (MPa)

7 days 7 days

1.5 14 days 1.5 14 days

28 days
1 28 days
1

0.5 0.5

0 0

BA1-NHL mixtures BA2-NHL mixtures

2 7 days 2 7 days
Flexural strength (MPa)
Flexural strength (MPa)

14 days 14 days
1.5 1.5
28 days 28 days
1 1

0.5 0.5

0 0

BA3-NHL mixtures BA4-NHL mixtures

Figure 3. Flexural strength (Rf) of BA-NHL pastes at 7days, 14days and 28days.

In Figure 4, the compressive strength results of BA-NHL specimens are presented.

Obviously, adding NHL at the replacement level of 5% negatively affected the
compressive strength of BA2-NHL. However, incorporating NHL up to 10% displayed a
minimal improvement of the mechanical performance of the BA mixtures. Furthermore,
a substantial increase in compressive strength was observed on BA-NHL mixtures that
contained up to 20% NHL content. At 7days, it appears that apart from the reference
mixture (NHL100), the remaining specimens exhibited low compressive strengths.
Nevertheless, the considerable compressive strengths were achieved by BA3-NHL20 and
BA4-NHL20 with 3.3 MPa and 2.8 MPa, respectively at 28days. Therefore, it is
concluded that the partial replacement of BA with NHL at 5%,10% and 20% contents
brings moderate improvements in terms of the mechanical properties.

207
 4 4
7 days
Compressive strength (MPa)

Compressive strength (MPa)

7 days
3 14 days 3 14 days

28 days 28 days
2 2

1 1

0 0

BA1-NHL mixtures BA2-NHL mixtures

4 4
7 days 7 days

Compressive strength (MPa)

Compressive strength (MPa)

14 days 14 days
3 3
28 days 28 days

2 2

1 1

0 0

BA3-NHL mixtures BA4-NHL mixtures

Figure 4. Compressive strength (Rc) of BA-NHL pastes at 7days, 14days and 28days.

3.2 Flexural and compressive strength of BA-OPC pastes

The flexural and compressive strength results of BA-OPC pastes are presented in Figure
5 and 6, respectively. It can be seen that the flexural strength of the reference mixture
(OPC100) was relatively higher compared to BA-OPC specimens at all curing ages. In
fact, at 7days, the flexural strength of OPC100 was about 1.6 MPa, but at 28days it
increased up to 3.1 MPa. Note that, the lowest flexural strength was obtained from BA2-
OPC5 at 14days. As shown in Figure 5, the flexural strength of BA3 and BA4 pastes
steadily increased with increased curing age and OPC content. A pronounced rise in
strength was perceived between 14days and 28days on all BA3-OPC. Concerning BA1
and BA4, only mixtures containing OPC at 10% to 20% replacement levels exhibited a
slight increase in flexural strength at the same ages. Besides, regardless the addition of
OPC to BA2 mixtures, the flexural strength remained almost identical at 28days.

208
 3.5 3.5
7d 7d
Flexural strength (MPa)

Flexural strength (MPa)

3
14d 14d
2.5 2.5
28d 28d
2 2
1.5 1.5
1 1
0.5 0.5
0 0

BA1-OPC mixtures BA2-OPC mixtures

3.5
7d 3.5
3 7d

Flexural strength (MPa)

3
Flexural strength (MPa)

14d 14d
2.5 2.5
2 28d 28d
2
1.5 1.5
1 1
0.5 0.5
0 0

BA3-OPC mixtures BA4-OPC mixtures

Figure 5. Flexural strength (Rf) of BA-OPC paste specimens at 7days, 14days and
28days.

In Figure 6, the BA-OPC compressive strength results are illustrated. According to the
bar charts, apart from BA2-OPC, all BA-OPC pastes exhibited a gradual augmentation in
compressive strength from the early age up to 28days. Indeed, the compressive strength
increased with increase in OPC content and curing age. The incorporation of OPC at 5%
and 10% contents had an unsignificant impact on BA1-OPC as the compressive strength
remained almost similar to that of the reference mixture (BA1-OPC0). Moreover,
unexpected tendency was seen on BA2-OPC specimens where OPC marginally decreased
the compressive strength as the curing age increased. It is noteworthy to mention that
BA3-OPC20 and BA4-OPC20 showed the highest compressive strengths of 5.6 MPa and
4.3 MPa, respectively.

209
 6 6
7d 7d
Compressive strength (MPa)

Compressive strength (MPa)

5 14d 5 14d
28d 28d
4 4

3 3

2 2

1 1

0 0
BA1-OPC0 BA1-OPC5 BA1-OPC10 BA1-OPC20 BA2-OPC0 BA2-OPC5 BA2-OPC10 BA2-OPC20
BA1-OPC mixtures BA2-OPC mixtures

6 6
7d 7d

Compressive strength (MPa)

Compressive strength (MPa)

5 14d 5 14d
28d 28d
4 4

3 3

2 2

1 1

0 0
BA3-OPC0 BA3-OPC5 BA3-OPC10 BA3-OPC20 BA4-OPC0 BA4-OPC5 BA4-OPC10 BA4-OPC20
BA3-OPC mixtures BA4-OPC mixtures

Figure 6. Compressive strength (Rc) of BA-OPC paste specimens at 7days, 14days and
28days.

4. Conclusion
In this phase of the study, a binding matrix comprising of BA was optimized and tested
for the mechanical performance. Note that, chemical activating agents were not
considered in this experiment. To improve the strength of BA pastes, small contents of
OPC and NHL (5%, 10% and 20%) were added as BA partial replacements.
Based on the obtained results, the following conclusions can be made:
- The partial replacement of BA with NHL content ranging from 5% to 20% shows
negligible influence on flexural and compressive strength. In some mixtures, for
example, BA1 and BA2, the flexural strength softly dropped, whereas for BA3 and
BA4, it marginally increased.
- Apart from BA2-NHL, all mixtures exhibited a similar trend in compressive strength.
At 28days, the highest compressive strength being equal to 3.3 MPa, was spotted on
BA3-NHL20 paste. Contrariwise, BA2-NHL5 had the lowest value which was
equivalent to 0.1 MPa at the same age.
- The incorporation of OPC at 5% and 10% content exhibits a constant growth in both
flexural and compressive strengths of BA3 and BA4 pastes however, slight variations
were seen in BA1 and BA2 pastes.
- Adding 20% OPC to BA pastes did not only improve the flexural strength of BA3 and
BA4 pastes but also the compressive strength of all BA pastes.

210
- At 28days, the greatest compressive strength which is equal to 5.6 MPa was obtained
from BA3-OPC20, whereas the lowest was measured on BA2-OPC5 and is about 0.1
MPa.
- The optimal NHL and OPC replacement levels are fixed at 20% for BA-NHL and
10% to 20% for BA-OPC.
- Overall, BA3 and BA4 showed promising flexural and compressive results thus, they
can be used as an alternative binder in the production of lightweight bio-based
mortar/concrete.

5. References
ALQAHTANI F.K., RASHID K., ZAFAR I., KHAN M.I., ABABTAIN A.A. (2021).
Production of sustainable green mortar by ultrahigh utilization of fly ash: Technical,
economic and environmental assessment. Construction and Building Materials 281,
122617, https://doi.org/10.1016/j.conbuildmat.2021.122617
ANDAVAN S., PAGADALA V.K. (2020). A study on soil stabilization by addition of
fly ash and lime. Materials Today: Proceedings 22, pp. 1125–1129,
https://doi.org/10.1016/j.matpr.2019.11.323
AREZOUMANDI M., VOLZ J.S. (2013). Effect of fly ash replacement level on the shear
strength of high-volume fly ash concrete beams. Journal of Cleaner Production 59, pp.
120–130, https://doi.org/10.1016/j.jclepro.2013.06.043
CHEN R., LI Y., XIANG R., LI S. (2016). Effect of particle size of fly ash on the
properties of lightweight insulation materials. Construction and Building Materials 123,
pp. 120–126, https://doi.org/10.1016/j.conbuildmat.2016.06.140
CHO Y.K., JUNG S.H., CHOI Y.C. (2019). Effects of chemical composition of fly ash on
compressive strength of fly ash cement mortar. Construction and Building Materials 204,
pp. 255–264, https://doi.org/10.1016/j.conbuildmat.2019.01.208
DINDI A., QUANG D.V., VEGA L.F., NASHEF E., ABU-ZAHRA M.R.M. (2019).
Applications of fly ash for CO2 capture, utilization, and storage. Journal of CO2
Utilization 29, pp.82–102, https://doi.org/10.1016/j.jcou.2018.11.011
FOŘT J., ŠÁL J., ŠEVČÍK R., DOLEŽELOVÁ M., KEPPERT M., JERMAN M.,
ZÁLESKÁ M., STEHEL V., ČERNÝ R. (2021). Biomass fly ash as an alternative to coal
fly ash in blended cements: Functional aspects. Construction and Building Materials 271,
121544, https://doi.org/10.1016/j.conbuildmat.2020.121544
FUSADE L., VILES H., WOOD C., BURNS C. (2019). The effect of wood ash on the
properties and durability of lime mortar for repointing damp historic buildings.
Construction and Building Materials, Volume 212, pp. 500-513,
https://doi.org/10.1016/j.conbuildmat.2019.03.326
KAYALI O., SHARFUDDIN AHMED M. (2013). Assessment of high-volume
replacement fly ash concrete – Concept of performance index. Construction and Building
Materials 39, pp. 71–76, https://doi.org/10.1016/j.conbuildmat.2012.05.009
211
LIN W.Y., PRABHAKAR A.K., MOHAN B.C., WANG C.-H. (2020). A factorial
experimental analysis of using wood fly ash as an alkaline activator along with coal fly
ash for production of geopolymer-cementitious hybrids. Science of The Total
Environment 718, 135289, https://doi.org/10.1016/j.scitotenv.2019.135289
LOTHENBACH B., SCRIVENER K., HOOTON R.D. (2011). Supplementary
cementitious materials. Cement and Concrete Research 41, pp. 1244–1256,
https://doi.org/10.1016/j.cemconres.2010.12.001
MIRONYUK I., TATARCHUK T., PALIYCHUK N., HEVIUK I., HORPYNKO A.,
YAREMA O., MYKYTYN I. (2019). Effect of surface-modified fly ash on compressive
strength of cement mortar. Materials Today: Proceedings S2214785319335163,
https://doi.org/10.1016/j.matpr.2019.10.016
MYADRABOINA H., SETUNGE S., PATNAIKUNI I. (2017). Pozzolanic Index and
lime requirement of low calcium fly ashes in high volume fly ash mortar. Construction
and Building Materials 131, pp. 690–695, https://doi.org/10.1016/j.conbuildmat.2016.11.038
OBERNBERGER I., SUPANCIC K. (2009). Possibilities of ash utilisation from biomass
combustion plants. Proceedings of the 17th European Biomass Conference & Exhibition.
Hamburg.
OMRAN A., SOLIMAN N., XIE A., DAVIDENKO T., TAGNIT-HAMOU A. (2018).
Field trials with concrete incorporating biomass-fly ash. Construction and Building
Materials 186, pp. 660–669, https://doi.org/10.1016/j.conbuildmat.2018.07.084
PARK J.-H., EOM J.-H., LEE S.-L., HWANG S.-W., KIM S.-H., KANG S.-W., YUN
J.-J., CHO J.-S., LEE Y.-H., SEO D.-C. (2020). Exploration of the potential capacity of
fly ash and bottom ash derived from wood pellet-based thermal power plant for heavy
metal removal. Science of the Total Environment, 740:140205, doi:
10.1016/j.scitotenv.2020.140205
SAMAD S., SHAH A. (2017). Role of binary cement including Supplementary
Cementitious Material (SCM), in production of environmentally sustainable concrete: A
critical review. International Journal of Sustainable Built Environment 6, pp. 663–674,
https://doi.org/10.1016/j.ijsbe.2017.07.003

212
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Valorization of sunflower bark in agropellets

Anaelle REIX 1, Hafida ZMAMOU 1, Hélène LENORMAND 1,

Nathalie LEBLANC 1

1. UniLaSalle, Univ. Artois, ULR7519 - Transformations & Agro-ressources.

Normandie Université, 3 rue du Tronquet, 76130 Mont-Saint-Aignan, France.
[email protected]

Abstract:
Sunflower stems are agricultural by-products considered as wastes. Once the seeds are
collected to make oil, the stem remains in the field. The stem of sunflower is composed
of a pith, i.e. internal part and a bark, i.e. external part. UniLaSalle is one of the partners
of DIVA project, for which the sunflower sector is developing the process to separate
pith and bark to valorise sunflower pith in bio-based mortars. So, finding ways to
valorise sunflower bark (SB) is essential to balance the sector of sunflower stem.
The aim of this paper is to study the use of SB to elaborate agropellets without binder
and additives. An investigation on the existing pellets on the market from agricultural
by-products was performed. Raw materials components influence various
characteristics of the pellets, in particular mechanical properties. Standard AFNOR
(2012) describe pellets properties to be respected certifying good quality pellets. One of
the main factors in the formation of pellets is the lignin concentration of the raw
material (DÖRING, 2012). Lignin plays the greatest role of binder during processing
thank to its ability to make pellets resistant to humidity. In addition, the ash
concentration of the raw material must be enough low to avoid damaging the stove after
combustion and to limit the emission of acid fumes that have a negative impact on the
environment (BODINEAU & POUET, 2006). The study is divided into several steps: 1/
characterisation of the raw material, then 2/ optimization of the press parameters
(temperature, pressure, grain size, humidity.) and 3/ formulation and characterization of
agropellets. Different formulations of pellets have been tested and compared to the solid
biofuels standard NF by FCBA technological institute. According to the standard, two
categories of pellets exist: (i) Agro-industrial quality and (ii) High-Performance quality.
Properties of agropellets such as humidity, length and diameter dimensions, ash content,
fines content, density and PCI (low calorific value) have been investigated. The
production of agropellets from SB without additives has been conclusive, thanks to its
intrinsic qualities.

Keywords: Agropellets, Wood pellets, Sunflower bark, Physical properties.

Corresponding author mail: [email protected]

Table 1. Nomenclature.
Symbol Definition
ADEME Agence de l’environnement et maitrise de l’environnement.
DIVA « Industrial demonstrator to valorize agro-residues in
ecoconstruction and bio renovation » Project.
DU Durabilité mécanique (Mechanical durability)
IDT Initial Deformation Temperature
MC Moisture Content
PCI Pouvoir Calorifique Inférieur (low calorific value)
SB Sunflower Bark

1. Introduction
Certain energy sources will gradually deplete in future. Thus, while the world reserves
of petroleum and gas would be set to run out within next forty to fifty years. At the
same time, energy needs continue to grow, given the global economic and demographic
development. Reducing energy cost and finding new sources of energy and heat
constitute a challenge for researchers all over the world. Hence, agricultural by-products
seem to be realistic alternative fuels, providing benefits such as reducing CO2 emissions
from fossil fuels combustion (BRAM et al, 2009).
During the year 2020, in France, 779.000 hectares of sunflowers plant were cultivated
for a harvest of nearly 1.74 million tons across the entire country (POGNAT, 2020).
However, today, only the sunflower seed is valued: The production of sunflower oil is
the fourth largest in the world, and its co-products are valued as part of animal or even
human food thanks to their high protein content (MARTINEZ-FORCE et al, 2015).
Sunflower stems, considered as waste, remain free to degrade in the fields, serving only
as humic fertilizer for the soil. However, this restitution of humus is low, it is therefore
today more interesting to a new way of valuing these stems, as well the surface
represented by the cultivation of sunflower plant is considerable and tends to increase.
According to POGNAT, the surface used is increasing by 34% between 2019 and 2020
(POGNAT, 2020). Thus. under the DIVA project funded by ADEME French agency
(see Table 1 for all symbols) and coordinated by ParexGroup company with various
partners including UniLaSalle, ways of valorization of sunflower pith were highlighted.
In fact, its high porosity and low thermal conductivity make it a very insulating material
for building. However, if the qualities sought-after in the building sector make it
possible to use the pith as a raw material, i.e. only the internal part of the stem, the
production of sunflower pith induces the production of a new co-product: the external
part, also called sunflower bark (SB). So, finding ways to valorise sunflower bark is
essential to balance the sector of sunflower stem. It is in this context that transformation
of sunflower bark into agropellets i.e. compressed pellets of combustible biomass was

214
initiated. The formulation of agropellets is similar to elaboration of wood-based pellets.
The biomass combustible from agricultural crops is condensed in the form of pellets.
Pellets can be produced using straw from cereals, oilseed cakes or even from dedicated
crops such as miscanthus (GAUTHIER, 2012). These agropellets are economically
more interesting than wood pellets thanks to their origin: A ton of wood pellets cost €
200 per ton in 2011, a ton of agropellets cost on average € 160 for domestic use and €
110 for industrial use (RAGT, 2021). Due to the differences in qualities between wood
and other biomass, agropellets will be subject to different specifications. with their own
standards. Table 2 shows the details of these specifications according to French
standard.

Table 2. Specifications NF certification agropellets (AFNOR. 2012).

Characteristics Agro-industrial quality High-Performance quality
Non-biobased additives (%) ≤ 2%
D = 6 to 16 ± 1 D = 6 to 8 ± 1
Dimensions (mm)
L = [3.15 to 40*] L = [3.15 to 40*]
Moisiture content (%) ≤ 15% ≤ 11%
PCI (MJ.kgˉ¹) ≥ 14.9 ≥ 15.8
PCI (kWh.kgˉ¹) ≥ 4.1 ≥ 4.4
Bulk density (kg.mˉ³) ≥ 650
DU (%) ≥ 92% ≥ 95%
Amount fines particles (% weight) ≤ 3% ≤ 2%
Ash content (%) ≤ 7% ≤ 5%
Chlorine content (%) ≤ 0.3% ≤ 0.2%
Sulfur content (%) ≤ 0.2%
Nitrogen content (%) ≤ 2% ≤ 1.5%
IDT (°C) ≥ 800 ≥ 1000
Arsenic (mg.kgˉ¹) ≤1
Copper (mg.kgˉ¹) ≤ 40
Chromium (mg.kgˉ¹) ≤ 10
Cadmium (mg.kgˉ¹) ≤ 0.5
Mercury (mg.kgˉ¹) ≤ 0.1
Nickel (mg.kgˉ¹) ≤ 15
Lead (mg.kgˉ¹) ≤ 10
Zinc (mg.kgˉ¹) ≤ 60

Note:* It accepted a maximum 1% of weight of pellet range 40mm ≤ L ≤ 45mm.

215
In this study, the relevance of using sunflower bark in elaboration of agropellets was
studied. Crushing of sunflower bark was mandatory to transform it then into pellets by
using a machine for individuals. Once the feasibility has been verified. the agropellets
properties have been studied and compared to French standard. The analysis of heavy
metals in ashes of agropellets was not studied (second part of Table 2). The only
parameters highlighted in Table 2 have been studied.

2. Materials and methods

2.1 Sunflower bark

Two types of sunflower bark are used: SF-1, very thick and regular and SF-2, very thin.
Characteristics of sunflower bark were determined through different tests and are
summarized in Table 3 and Figure 1. The biochemical fraction of raw materials has
been determined by Van soest method and the bulk density has been determined by
using a glass cylinder (diameter 10 and length 20 cm).

Table 3. Characteristics of sunflower bark.

Biochemical fraction (%) Water
Sunflower Bulk density Skeletal density
Soluble absorption
bark (kg/m3) Cellulose hemicellulose lignin (g/cm3)
fraction (%)
SB-1 132.9 57.2 15.8 13.2 13.8 1.19 272
SB-2 107.5 46.2 14.6 19.5 19.7 1.38 311

Figure 1. Particles size of both sunflower barks (SB-1 and SB-2) and their photos.

The particle size distribution by mechanical sieving is based on the particle width. The
particle sizes of SB-1 and SB-2 were mainly included respectively between 2 and 4mm,
and 0.5 and 2mm. By 2D imaging, the particle size analysis of SB-1 and SB-2 particles
present a length between respectively, 0.2 and 1.1 cm and 0.2 and 1.4 cm. The

216
difference between these raw materials is explained by the presence of particles of
sunflower pith in SB-2. The low value of SB-2 bulk density confirms the composition.
According to DÖRING, one of the main factors in the formation of pellets is the lignin
concentration of the raw material (DÖRING, 2012). The higher this content, the higher
the resistance of the pellets. Table 4 shows chemical compositions of various types of
woods and sunflower. Softwoods contain more lignin that hardwoods and are the more
used for agropellets production. Lignin plays the greatest role of binder during
processing. Lignin, under the effect of heat and pressure, will activate and migrate to the
surface of pellets to form a protective layer and increase their resistance to humidity.

Table 4. Chemical composition of hardwood and softwood species. sunflower stem and
sunflower bark (SUN et al, 2013; DORING, 2012).
Type Specie Cellulose Lignin Resin Ash content Hemicellulose
(%) (%) (%) (%) (%)
Softwood Fir 42.3 28.6 2.3 1.2 22.5
Softwood Pine 41.9 29.5 3.2 1.3 21.5
Softwood Spruce 41 30 - - 24.3
Hardwood Ash 40.2 26 2.2 1.3 25
Hardwood Beech 45.4 22.7 0.7 1.6 22.2
Hardwood Birch 40.9 27.3 2.2 1.8 27.1
Hardwood Poplar 48.4 21.6 2.4 1.3 18.2
Hardwood Willow 42.9 24.7 2 1.2 21.9
Sunflower Bark 38.5 25 - - -
Sunflower Stem 48 17.5 - - -

2.2 Manufacturing of agropellets from sunflower bark

First of all, in order to be conform with the French standard, we have crushed the two
types of sunflower bark to obtain a particle size fraction between 0.5 and 1mm, as we
can see in the Figure 2-a. To transform the raw material into pellet, we therefore
introduced it gradually into the press (Figure 2-b). As illustrated by STELTE et al,
(2012) in the Figure 2-3, the press forms pellets by compressing layers of raw materials
between a grid and a roll. Due to friction forces, the temperature of the machine
increases as well as the pressure created by the passage of the roll which press the
material into the perforations of the grid. In experimental program, we have adopted
two approaches : (i) as well as possible optimizing of parameters of formulation such
ash, the temperature, pressure, and humidity. (ii) we tried some formulation by
introducing only the water, thus by addition of sawdust. Some examples of formulation
are shown in Figure 2-d.

217
 (a)

SB-2 SB-1 (d)

(b)

(c)

Figure 2. Manufacturing of agropellets : (a) Crushed SB-1 and 2, (b) pellets press used
(c) schema of pellets press processus (STELTE et al. 2012), (d) example of some
formulations tested.

To obtain the agropellet from SB, we found that the optimal parameters were a
temperature range from 75 to 105 °C, and a Moisiture Content (MC) of raw materials
range from 8 to 15 %. The pression of the press was not measurable. After several
attempts to elaborate agropellet we can conclude that the formulation without any
additves is the best formulation for sunflower bark. Its intrinsec properties like lignine
content and percentage of humidity is enough.

2.3 Brief description of analysis

To analyse the agropellets obtained, muffle furnace has been used to determine the ash
content. The pellet was exposed to a temperature of 550 °C for 6 hours. We have used a
sealed glass cylinder (diameter 10 and length 20 cm) and a balance to estimate the bulk
density of pellet. Moisiture content has been measured by Infrared moisture
determination balance. It determines the moisture content of a sample by detecting
weight loss during heating and drying. The dimensions of pellet were controlled and
measured on a 100g of pellet with 3 repetitions, by using a caliper with a precision
around 10µm. The particles were meseared one by one. To measure pellet PCI, it is
necessary to be able to measure the quantity of heat which this one releases during its
combustion. A bomb calorimeter of type 6200 Parr Calorimeter, was used. The
calibration has been carried out with benzoic acid (known calorific value of 26.5 kJ / g).
Measuring the temperature of the water before the reaction and throughout the

218
combustion will allow precise determination of the amount of heat produced by the
substance studied. The mechanical durability (DU) has been determined by other
laboratory only on the best sunflower bark pellets : SB-1.

3. Results and discussion

We have characterised the two-formulation 100%SB-1 and 100%SB-2 (without any
additives). Thus, we have verified the conformity of these agropellets with industrial
quality and high-performance quality from French standard, see Table 5.

Table 5. Results of analysis of agropellets from sunflower bark.

Characteristics 100% SB-1 100% SB-2
Additives (%) Without any additives
Dimensions : d = 6.004 mm d = 6.050 mm
L = 17.46 mm L = 16.56 mm
Moisture content (%) 5 to 7 6
PCI (MJ/kg) 18.3 16.7
Bulk density (kg.mˉ³) 650 -
Mechanical durability 88.9 -
Amount of fine particles (% weight) : 0.77 1.61
Ash content (%) ≤5 ≤ 7%
(2.50+/-0.12%) (6.4+/-0.4)
Note: *d: diameter; L: length.

If durability is not considered, 100% SB-1 can be upgraded in both industrial and high-
performance qualities of agropellets standard. However, the 100%SB-2 could not be
used in the form of high-performance quality pellets. because of its ash content.
Mechanical durability can be improved by the use of industrial press (the equipment
used lacks power and it was difficult to control the formulation parameters). Several
parameters are missing to conclude. We have listed all these parameters to complete this
study in Table 2.

4. Conclusions
The study highlighted the possibility of using sunflowers bark (SB-1 and SB-2) in the
production of agropellets following the French agropellets standard. Despite the use of
the pellet press, which intended for exclusively forest biomass, we were able to obtain
sunflower-based pellets without any additives or water. Today, several crucial data are
missing to confirm that the production of agropellets is a means of sustainable
development of sunflower bark. In order to complete this study, it is necessary to
determine the IDT, and proportions of sulfur or nitrogen. The recommendation that we

219
can formulate for the future studies is to use a press where you can control with
precision the temperature and pression parameters.

5. References
AFNOR (2012). Référentiel certification NF : Biocombustibles solides. FCBA. 12 mars
2012.
BODINEAU L., POUET J-C. (2006). Etude bibliographique sur la combustion de
produits issus de cultures annuelles (blé, paille, maïs). AXENNE, mars 2006.
BRAM S., DE RUYCK J., LAVRIC D. (2009). Using biomass: a system perturbation
analysis. Applied Energy, 86 (2), pp. 194-201. doi:10.1016/j.apenergy.2007.12.002
DÖRING S. (2012). Biomass types for pellet production. In Power from Pellets:
Technology and Application, pp. 13‑30, Springer, ISBN 978-3-642-19961-5
GAUTHIER G. (2012). Essais de production de pellets composés à 100% de
miscanthus, Valbiom.
MARTINEZ-FORCE E., DUNFORD N.T., SALAS J.J. (2015). Preface. In Sunflower
– Chemistry, Production, Processing, and Utilization. Elsevier Inc., ix‑x. AOCS Press.
2015.
POGNAT J.-M. (2020). Grandes cultures (Version modifiée). Forts reculs des récoltes
2020, à l’exception du maïs et du tournesol. N° 127, (septembre 2020), 5p.,
http://www.epsilon.insee.fr:80/jspui/handle/1/134447
RAGT ENERGIE. (2021). Les combustibles biomasse - RAGT Energie. Retrieved
13/01/2021 on https://www.ragt-energie.fr/fr/biomasse/combustibles-biomasse.php
STELTE W., SANADI A., SHANG L., HOLM J., AHRENFELDT J., HENRIKSEN U.
(2012). Recent developments in biomass pelletization - A review. Bioresources 7 (3), pp.
4451-4490, doi: 10.15376/biores.7.3.4451-4490
SUN S., MATHIAS J-D., TOUSSAINT E., GREDIAC M. (2013). Hygromechanical
characterization of sunflower stems. Industrial Crops and Products 46, pp.50‑59, HAL Id:
hal-00923258

220
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

A case study of solidification technology applied to soft soils

and sediments: Construction of a railway platform in Ningde
(China)
Xiao WANG 1, 2, Daniel LEVACHER 1, Tianxue ZHONG 2

1. Normandie Université, Unicaen, UMR 6143 CNRS - M2C, 24 rue des Tilleuls, 14000
Caen, France.
[email protected]; [email protected]
2. Jiangsu Kun Ze Technologie, Floor 5, Building B1, Ruan Jian Da Dao street, quartier
Du Yu Hua Tai district, Nanjing, Chine.

Abstract:
Most of the time, researchers are mainly studying the effect of a single hardening
additive on an inorganic material for fine sediments solidification, and most of them
only use laboratory tests. In this paper the hardening additive was composed and
applied to the construction of railway platform at Ningde (China) not only through
laboratory research work, but also in an innovative way for the construction of a railway
foundation site. The study of the type and dosage of the additive (hardening agent), the
long-term performance and durability of the solidified sediment was conducted
considering the project design and the construction process. That allowed clarifying the
relative requirements for the use of the solidification technology with sediments that can
serve as foundation soils for railways station. The main conclusions are as follows. This
solidification technology was applied to sediments having a medium to low water
content. By modifying the hardening agent formula, a composed additive ADC2 is
obtained which met the design requirements, which improved the strength of the
solidified sediment and allowed the use of sediments and industrial by-products (partial
components in the hardening agent). The unconfined compressive strength of the
solidified sediment at 28 days was ≥ 200 kPa, the cohesion ≥ 100 kPa, internal friction
angle ≥ 25 °, compressibility modulus ≥ 6.0 MPa. Then, all the mechanical properties of
the solidified sediment increase with time (age). The durability of the solidified
sediment is acceptable. The hardened sediment prepared using the ADC2 composed
hardening agent has a relatively low relative mass loss rate and a high resistance
retention rate after cycles and meets the durability requirements. The solidified sediment
was qualified by inspection and acceptance tests on site.
Keywords:
Beneficial use, Sustainable sediment management, Solidification/stabilization
technique, Binders, Hardening additives, Sediment, Civil engineering.

Corresponding author mail: [email protected]

1. Introduction
China has been facing a shortage of aggregates for more than a decade, or at least a
slower and more expensive supply of aggregates. This has resulted in a slowdown in
construction and development in some parts of China. In addition, China is paying more
and more attention to the strategy of sustainable development by monitoring quarry
management and trying to limit their exploitation in view of the non-renewal of the rock
material resource. However, at present, filling materials used as foundations in
construction and site development in China are increasingly difficult to obtain (WANG,
2019; WANG et al., 2020). There is therefore a strong demand for new foundation
materials i.e. alternative or substitute materials that can replace and supplement
granular materials (sand, gravel, pebbles) from quarries. For the development of near-
shore sites in wetlands or areas with high sediment deposits, the use of soil or sediment
solidification technology to provide alternative foundation materials is receiving
increasing attention. It is all the more recommended as methods of removing these soils
or sediments such as dumping at sea and landfills inevitably lead to serious
environmental pollution (JUNG et al., 2019), occupation of farmland and other
environmental and social problems. This solidification technique allows on the one
hand to valorise soils and sediments with a low to medium water content resulting from
dredging to make them a materials resource for the development of sites, and on the
other hand, it can also solve the problem of the non-renewability of the resource and the
environmental pollution caused by the production of natural aggregates. This has
become a concern for all countries in the world.
This low- to medium- water content soil or sediment solidification technology for the
development and construction of infrastructures can solve both the problem of the non-
renewable resources of aggregates or filling materials and to eliminate environmental
pollution that generates the production of traditional aggregates. These kinds of
materials can improve the degree of compaction of gravelly soils difficult to control by
mixing them and help to decrease or avoid differential settlement of the foundation
layers of infrastructure. This solidification technology has already been applied to silty
soils considered to be waste (sludge). This solidification technology has certain
advantages such as low cost of implementation on site with high efficiency and
generating low pollution. An application of this solidification technology on the site of
Ningde (China) was gradually encouraged and a project was carried out occupying a
leading position in the construction of the Ningde site dedicated to a railway platform.
Through its long experience of building the company JSKZ Technologie proposed an
innovative solidification technique in the construction of the site to accommodate the
Zhangzhou-Ningde Railway Station Project (Fujian Railway Section), hereinafter
referred to as Ningde project. The project needed to provide reliable design methods and
test procedures for the large-scale application solidification technology on soft silts for
embankment layers as foundation. The Ningde project is first presented on the concept

222
of the railway station embankments and the site conditions, then the results of the tests
undertaken on the sediment-binder mixtures are discussed according to the values
required at the project.

2. Presentation of the Ningde project

The Ningde city is in the North-east of the Fujian province, see Figure 1. The project
concerns the construction of the railway station foundations on soft soils.

China

Figure 1. Ningde city in Fujian province in South-East of China

The Ningde project involves the treatment of silty surface sediments over an area of
73839m2. Two areas can be distinguished, namely (i) an area where piles are driven
(zone 1, 50817m2). The solidification treatment of the surface layer is planned for a
thickness of 0.8 m. It supports the pile drivers and must therefore meet the load-bearing
capacity requirements of specific pile drivers and, (ii) a second fill area where the base
is excavated and replaced by a solidified silty sediment but the thickness of the fill
varies from 2.0 to 3.0 m (zone 2, 6500m2). For these two areas, the silty sediments are
first excavated, treated to solidify and then replaced.
The site of the railway embankments is relatively flat with a thin layer of artificial fill of
about 1m. Underneath this fill, the surface layer is a 2m thick, plastic, grey silty clay
overlying a 7m thick layer of dark grey, fluid to plastic silt. The latter overlies a mille-
feuilles of alternating thin metric layers of clay, sand and/or gravel over a highly
weathered granitic substrate at an average depth of 25m. Excavated sediment is clayey-
silty soil with a medium water content and negligible organic content, see Table 1.

Table 1. Initial clayey-silty sediment characteristics in the Ningde project

Clay content Silt content Liquid limit Plastic limit Plastic index Water content Organic matter
<5μm (%) 5-75μm (%) LL (%) PL (%) PI (%) W (%) OM (%)
51.1 48.9 46.6 27.4 19.2 56.2 0.5

223
 Schematic diagram of the lay out of the foundation treatment of
the Ningde station – scale 1:1000

Figure 2. Sites of sediment solidification and type of embankments in Ningde project

3. Site construction requirements in the Ningde project

These include both design requirements and reference specifications. In accordance
with the project, there is excavation of the surface sediments in place and exchange of
these silty-clayey sediments into solidified sediments for a thickness of 0.8m in (Zone
1) or 2-3 m (Zone 2), see Figure 2. For the solidified sediment, required geotechnical
indicators are the following: a density γ ≥ 17.0 kN/m3, a cohesion C ≥ 100 kPa, a
friction angle φ≥ 25°, a deformation modulus Es ≥ 6.0 MPa and a compressive strength
at 28days UCS >200kPa. They are obtained from either laboratory samples or from site
samples. The site engineering indicators are performed on the embankment sites,
namely from cone penetration test (CPT - point resistance ≥ 200 kPa) and plate testing
(a 30cm plate diameter thick of 2.5cm, stiffness K30 ≥ 100 MPa/m). In addition to the
geotechnical requirements (laboratory and in situ tests), there are also requirements for
the solidification process. The mixture of sediments and binders must be homogeneous
whatever the machine used (excavator or rotary cutter). The volume mixed depends on
the machine (5 to 10m3 for an excavator) and the mixing time is limited (15 to 30
minutes). It is laid in layers of approximately 0.5m by light compaction and then more
important depending on the hardening. The time required for the second layer is
between 24 and 72 hours, the control operated concerns the water content of the
solidified sediment. Control tests are carried out, the frequency of sampling is

224
approximately 1 per 1000m2 and take place in the lower and central part of the backfill
layer. The embankment construction must be ended within 28 days.
.
4. Sediment-mixtures investigated and laboratory test results
Excavated sediments were solidified with five types of binders/hardening agents as
follows: OPC cement, lime and 3 hardening agents ADC1, ADC2 and ADC3 produced
by Jiangsu Kunze Technology Co (China). The ADCs agents are activated with one or
two additives such as sodium silicate and sodium sulphate. Laboratory tests are
performed with different hardening agent dosages i.e. 80, 90 and 100 kg/m3. The main
parameters that governs the solidification of sediments in this project are unconfined
compressive strength requirements and cohesive values.

Table 2. Sediment-mixtures and main parameters from laboratory tests

Hardening agent Dosage UCS (kPa) Cohesion (kPa) Hardening agent Dosage UCS (kPa) Cohesion (kPa)

reference (kg/m3) 7days 28days 7days 28days reference (kg/m3) 7days 28days 7days 28days

80 89 150 16 32 80 92 151 17 34

Cement 90 102 159 18 38 Lime 90 100 164 22 43

100 118 172 20 44 100 124 195 24 46

80 122 254 40 82 80 154 302 43 87

ADC1 90 167 353 51 90 ADC2 90 205 396 75 103

100 219 405 72 102 100 253 431 78 105

80 182 361 54 96

ADC3 90 224 406 76 106

100 279 453 80 108

The Table 2 shows the unconfined compressive strength (UCS) with cement and lime
and a dosage of 100 kg/m3 does not meet the design requirements. Similarly, the
cohesion is also far from the required values. Hardening agents ADC1 to 3 overcome
the shortcomings of a single binder such as OPPC cement or lime. For these agents,
different types of modified micro-powder mixtures with high specific surface area and
high activity are introduced, and by adding an appropriate amount of activator, a
significant hardening effect on solidified sediment with medium to low water content is
achieved. According to the data in Table 2, when the ADC1 agent reaches 100 kg/m3, it
can meet the design requirements as the agents ADC2 and ADC3 with a dosage at
90kg/m3. Considering the performance and cost factors, the agent ADC2 is finally
chosen for the sediment solidification in the Ningde project. The determination of the
long-term mechanical properties was carried out in the laboratory as a function of the
time for the ADC2 hardener with a dosage at 90kg/m3. The mechanical properties of the
solidified sediment evolve with time before stabilization as shown in Figure 3. This
increase is mainly due to the hydration reaction of the hardening agent which continues

225
to occur as the solidified sediment ages. Although the degree of reaction is much less as
time goes on, the gelling agent within the solidified sediment continues to increase. The
soil particles are continuously agglomerated and cemented, so that all indicators are
growing.

Figure 3. Evolution of main mechanical parameters vs time in ADC2-sediment mixture

The problem of durability of solidified sediment materials under alternating dry and wet
conditions (climatic conditions) should be considered. It ensures the long-term
application of a solidified sediment. Wetting/drying tests based on ASTM D 4843
(2016) were performed on the ADC2-sediment mixture. The test includes 10 dry and
wet cycles over 20 days. After the specimen had been subjected to 0, 2, 4, 6, 8 and 10
dry and wet cycles, the unconfined compressive strength UCS and mass of the specimen
were measured and the relative mass loss rate with the change in strength of the
specimen throughout the cycle was deduced. Figure 4 shows the relative mass loss rate
and strength retention rate of the solidified sediment subjected to the dry-wet cycle after
28 days of curing. Both relative parameters are tending towards stabilisation. According
to the ASTM D4843 (2016) standard, if the rate of relative mass loss is greater than
15% or if the rate of strength loss is greater than 20% of the specimen, it is considered
that the drying-wetting durability of the specimen cannot meet the standard
requirements. The sediment solidified with the hardening agent ADC2 was subjected to
this test in dry-wet cycles between 20 and 60°C, the rate of mass loss of the solidified

226
sediment is still less than 15% and the strength drop is less than the standard 20% which
meets the durability requirements, see Figure 4.

Figure 4. Influence of wetting-drying tests at 28days on ADC2-sediment mixture

5. On-site sediment solidification process and control test results

The construction process is as follows: excavation, mixing (15min.) with the ADC2
agent at 90kg/m3, compaction of a layer of 0.5m thick with 4 roller passages.

Figure 5. Sediment solidification technology on site

227
The execution time of the construction on site was 28 days and acceptance tests were
carried out at the same time. It should be noted that control tests were also carried out
during construction. 20 series of cohesion (C) and internal friction angle (φ) control
tests have shown a variation from 102.2kPa to 111.3kPa for C and from 25.4° to 28.6°
for φ. Series of acceptance tests have confirmed that the solidification of sediments with
90kg/m3 of ADC2 hardening agent is satisfying as shown in Table 3.

Table 3. Acceptance test results at 28 days for the ADC2 solidified sediments
Series number Compressibility modulus Es (MPa) Cohesion (kPa) Internal friction angle (°)
1 29.19 141.7 25.6
2 31.85 168.9 28.8
3 25.77 146.9 27.6
Required value >6 >100 >25

6. Conclusions
The use of ADC2 hardening agent for solidifying low to medium water content clayey-
silty sediments in the Ningde project was justified with laboratory and in-situ tests. The
study focused on type of agent and its optimal dosage, the long-term performance and
durability of the solidified sediment. But the project design as the construction process
were also considered. The requirements for the use of solidification technology for
sediments as foundation soil for railways have been satisfied and showing the mastery
of key technologies for the preparation of solidified sediments with medium to low
initial water content and the acceptance in terms of durability of the solidified sediment.
A satisfactory design of the construction process for the implementation of the
treatment technology for sediment solidification in the construction of a railway
platform foundation has been also demonstrated

7. References
ASTM D4843 – 88. (2016). Standard test method for wetting and drying test of solid
wastes, http://www.astm.org/cgi-bin/resolver.cgi?D4843-88(2016)
JUNG J.M., CHOI K.Y., CHUNG C.S., KIM C.J., KIM S.H. (2019). Fractionation and
risk assessment of metals in sediments of an ocean dumping site. Marine Pollution
Bulletin, volume 141, pp. 227-235, doi: 10.1016/j.marpolbul.2019.02.041
WANG X. (2019). Méthodologies de valorisation de sédiments mises en œuvre dans
des ouvrages géotechniques en Chine. PhD thesis, Normandie Université, 202p., tel-
02520015, version 2
WANG X., LEVACHER D., ZHONG T (2020). Valorisation de sédiments à teneur en
eau élevée en génie côtier : construction de la digue de Xiapu (Chine), XVIème Journées
Nationales Génie Côtier – Génie Civil Le Havre, 2020, pp. 683-690,
DOI:10.5150/jngcgc.2020.076

228
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Cartography of agricultural by-products available in the

World for a potential use in building materials

Hafida ZMAMOU 1, Hélène LENORMAND 1, David ABOUZEID1,

Nathalie LEBLANC 1

1. UniLaSalle, Univ. Artois, ULR7519 - Transformations & Agro-ressources,

Normandie Université, 3 rue du Tronquet, 76130 Mont-Saint-Aignan, France.
[email protected]

Abstract:
As the world became more and more urbanized, the consequences of this transformation
have led to major impacts on the environment. One of those impacts have been the
materials used in construction. That applies in particular to non-renewable materials
used in construction. On the other hand, agriculture produces large amounts of by-
products every year. and the problem is that there is not a lot of ways to recycle such
abundance in the number of agricultural by-products, most of them either return to the
soil or is burnt, which have a strong impact environmentally. Identify potential crop by-
products to be recycled or to be valorised will help construction industry to find more
eco-friendly solutions.
In this study, a focus has been made on crops presenting by-products with a potential
use in the construction industry. We were interested in a lot of plant aggregates, for
example: coconut shell, oil palm kernel shell, bagasse, tobacco waste, rice, maize,
sunflower, miscanthus, wild giant reed… The distribution of the crops in a regional-to-
world context, using geo-spatial data, have been analysed to pinpoint the most of them.
Data of harvested areas and the potential production quantity of each crop have allowed
to estimate the potential quantities of plant aggregates. This study can also be used to
understand the global patterns of farming systems as well the abundance and the
similarities of crops production that can be found very close regionally.

Keywords:
Agricultural by-products, Thermal performance, Bio based materials, Cartography.

Corresponding author mail: [email protected]

1. Introduction
The high environmental impact of the construction and building sectors has given rise to
an increased research on alternative materials last thirty years. In fact, using lightweight
building materials from ecological resources reduces the environmental impact of
buildings. Using local agricultural resources as lightweight aggregates in combination
with inorganic binder can be a good solution to elaborate the lightweight building
materials that can be a bio-based mortars/concrete (also named vegetal concrete, agro-
concrete, or bio-aggregate-based concretes).
Agricultural areas make up over one third of the worlds’ terrestrial area
(RAMANKUTTY et al., 2018) and this area are continuously increasing. Although the
majority of agricultural crops are produced to feed people and animals, land is
increasingly dedicated to the production of renewable energy sources, such as biogas
(DE WIT et al, 2011; TORRIJOS, 2016). This development has led to increased excess
farming agricultural by-product to valorize. Crop production has more than kept pace
with population growth over the last 50 years. For example, cereal production per capita
increasing from 0.29 to 0.39 tonnes per person between 1961 and 2014,
(RAMANKUTTY et al., 2018). So, Agricultural productions generate many co-
products that are potential sources of renewable raw materials to produce
environmentally friendly materials.
In this paper we focus only on plant aggregates (wood particles coming from stem of
plant, or fruit) and not on plant fiber such as flax or hemp fibers.
Many studies use agricultural by-products as raw material for building. For example,
PANTAWEE et al., (2017), HUSSAIN et al., (2019), GLÉ et al., (2021) used hemp
shiv, NTIMUGURA et al., (2021) miscanthus shives, CHABANNES et al., (2014),
DEPAA et al., (2021) rice husk or rice husk ash, AHMAD et al., (2018) maize,
BARBIERI et al., (2021) wheat husk, NOZAHIC et al., (2012) sunflower stems,
GUNASEKARAN et al., (2011) coconut shell aggregates, to produce bio-based
construction materials. The use of these aggregates can be used for one of these goals:
using a plant aggregates as filler to replace a part of the cementitious materials (non-
renewable and polluting source) or to improve a thermal and acoustic properties for old
or new building. Other studies use the ashes of plant aggregates (biomass) as filler.
In this research, we used the statistical data to estimate agricultural by-product
abundance and the yield. Our main objective was to investigate the plant aggregates
present a good potential for building materials. Specifically, we predicted that
agricultural by-product abundance increases and there are a lot of materials not
valorising. The main aim of DIVA project "Industrial demonstrator to valorize agro-
residues in eco-construction and bio renovation" is to develop new bio-based mortars
from maize stalk, sunflower pith, and sunflower bark. Another aim is to imagine
duplicating these solutions in the other parts of the world.

230
2. Methodology

2.1 How to select the potential plant aggregates for building?

Plant crops are often highly dependent on climate and the nature of the soil, but some
crops can grow everywhere. The applied methodology refers to four continents: Asia,
Europe, America, and Africa. The data for area cultivated were taken from (FAO,
2019), and the data for production quantity were taken from the FAOSTAT (FAO,
2019). The percentage of by-product per production was estimated. The methodology to
have access to the estimate of the deposits is as follows:
• Identify the most present crops in the world RAMANKUTTY et al., (2018).
• Identify the potential by-products for eco-construction of the identified crops.
• Estimate the amount (mass) of plant aggregates thanks to the yields
Only the 10 most important deposits in each area will be keep in this study.

2.2 Overview of plant aggregates selected

The global area is full of plant aggregates identified as potential sources of construction
and building sectors. This plant aggregates can be maize stalk, wheat, rice, rapeseek,
sugarcane, coconut.. Figure 1 shows photographs of types of plant aggregates in the
world.

Figure 1. The most plant aggregates present in the world for a valorisation in bio-based
materials.

3. Mapping and discussion

Due to the selection criteria described in section 2.1 and FAO data, fourteen plant
aggregates were identified as the most agricultural by-product present in the world. The
area occupied by each plant aggregates and the straw yield per hectare were estimated,
see Table 1. Figure 2 shows the result of estimation of plant aggregates available on the
area of study.

231
 Table 1. Crop areas, yield, and distribution of plant aggregates on 4 continents.
Estimation Total of
of by- plant
World product aggregates
Crops Area cultivated (ha) Yield (tons/ha)
region per per
production production
(tons) (tons)
Bagasse (sugarcane) 38139 23,6 (SOLOMON, 2011; YADAV & SOLOMON, 2006) 900080
Maize straw 18353189 2 (LENORMAND &LEBLANC, 2020) 36706378
Wheat straw 62385441 3 (ADEME, 2013) 187156323
Barley straw 24222012 3 (ADEME, retrieved on website, May 2021) 72666036
Europe Rice husk/straw 623432 2 1246864 357104145
Rapeseed straw 8797198 2 (ADEME, 2013) 17594396
Sunflower stalk 19301171 1,5 (LENORMAND & LEBLANC, 2020) 28951757
Miscanthus 20000 (LEWANDOWSKI et al., 2018) 19 (ZUB et al,, 2011) 380000
Reed 361000 ((WICHMANN & KÖBBING, 2015) 30 (LE BARTZ et al., 2009) 10830000
Flax shiv 240111 2,8 (LENORMAND & LEBLANC, 2020) 672311
Bagasse (sugarcane) 10961549 23,6 258692556
Maize straw 66474093 2 132948186
Wheat straw 98638462 3 295915386
Barley straw 12582974 3 37748922
Rice husk/straw 138605217 2 (GIDDE & JIVANI, 2007; ESA et al., 2013) 277210434
Asia 1084903324
Rapeseed straw 13703088 2 27406176
Miscanthus 100000* (XUE et al., 2015) 19 1900000
Palm Kernel Shell 20918728 1 (DIT, 2007) 20918728
Coconut shell 9618092 1 (KOOPMANS & KOPPEJAN, 1998; SUDHIR, 2005) 9618092
Sorghum straw 5636211 4 (GANYO et al,, 2018) 22544844
Bagasse (sugarcane) 1580586 23,6 37301830
Maize straw 40711874 2 81423748
Wheat straw 9765237 3 29295711
Barley straw 4116152 3 12348456
Rice husk/straw 17110769 2 34221538
Africa 318574729
Rapeseed straw 115251 2 230502
Sunflower stalk 2208861 1,5 3313292
Palm Kernel Shell 5577245 1 5577245
Coconut 1155720 1 1155720
Sorghum 28426672 4 113706688
Bagasse (sugarcane) 13759956 23,6 324734962
Maize straw 71584010 2 143168020
Wheat straw 34665394 3 103996182
Barley straw 5736686 3 17210058
Rice husk/straw 5704659 2 11409318
America 645809509
Rapeseed straw 9295110 2 18590220
Sunflower stalk 2657759 1,5 3986639
Palm Kernel Shell 1596159 1 1596159
Coconut 566440 1 566440
Sorghum 5137878 4 20551512
Note: * it is not easy to find the area occupied by miscanthus in Asia. the value 100,000 ha corresponding only to
China.

232
 Figure 2. Estimation of the quantity of plant aggregates and their distribution
in (a) Europe, (b) Asia, (c) Africa and (d) America.

The plant aggregates available, exceed 357 million tons in Europe, 1.08 billion tons in
Asia, 318 million tons in Africa and more than 645 million tons in America every year.
They are mostly constituted by cereal straw (wheat, maize, barley, sorghum, rice),
which represents the most abundant biomass available in the world. This quantification
of plant aggregates is underestimated because the surface area of some crops is not
identified on FAO website. From data in Table 2, it is clear that a good part of this
agricultural by-product is used for animal feed. But all the excess can be a good solution
for the building to reduce its environmental impact by offering a lightweight concrete.
In fact, replacing a part of cement by agricultural by-product reduce the amount of CO2
emission resultant from the manufacturing process of the cement.

4. Focus on plant aggregates available in France

According to LENORMAND & LEBLANC (2020), the French deposit of plant
aggregates with a potential use for eco-construction would be about 15 million tons

233
from cereal crops (50%) and other crops such as hemp, flax, sunflower, maize,
rapeseed, miscanthus, rice and reed (LENORMAND & LEBLANC, 2020). This
estimation considered other valorisation ways such as mulching, bioenergy, straw
construction, animal litter. Figure 3 lists the main sources of plant aggregates in France
as well as information relating to their area, biomass yield per ha, geographical area,
harvest period and the types of plant particles that can be generated.

Figure 3. Summary of information relating to French crops able to produce plant

aggregates for bio-based materials for construction (LENORMAND & LEBLANC,
2020).

5. Conclusions
Today, a lot of agricultural by-products are left and are mostly burnt in open fields. The
valorisation of these by-products could help the conception of eco-friendly building
materials. Another advantage is to add economic value to agricultural by-products, thus
creating an incentive for the farmer to refrain from open burning of this materials.
For the future work, characterisation of these identified plant aggregates can be
investigated in order to identify their thermal or acoustical performances.

7. References
ADEME, Paille de céréale. Fiche n°1 Coproduits riches en ligno-cellulose,16p.
ADEME (2013). Estimation des gisements potentiels de substrats utilisables en
méthanisation. 117p.
AHMAD M.R, CHEN B., ODERJI S.Y., MOHSAN M. (2018). Development of a new
bio-composite for building insulation and structural purpose using corn stalk and
magnesium phosphate cement. Energy and Buildings. Volume 173, 15, pp. 719-733.
https://doi.org/10.1016/j.enbuild.2018.06.007

234
BARBIERI V., GUALTIERI M.L., MANFREDINI T., SILIGARDI C. (2021).
Lightweight concretes based on wheat husk and hemp hurd as bio-aggregates and
modified magnesium oxysulfate binder: Microstructure and technological
performances. Construction and Building Materials. Volume 284, 122751.
https://doi.org/10.1016/j.conbuildmat.2021.122751
CHABANNES M., BENEZET J-C., CLERC L., GARCIA-DIAZ E. (2014). Use of raw
rice husk as natural aggregate in a lightweight insulating concrete: An innovative
application. Construction and Building Materials. Volume 70, pp. 428-438.
https://doi.org/10.1016/j.conbuildmat.2014.07.025
DEPAA RA.B., PRIYADARSHINI V., HEMAMALINIE A., XAVIER J.F.,
SURENDRABABU K. (2021). Assessment of strength properties of concrete made with
rice husk ash. Materials Today: Proceedings, https://doi.org/10.1016/j.matpr.2020.12.605
DE WIT M., LONDO M., FAAIJ A. (2011). Productivity developments in European
agriculture: relations to and opportunities for biomass production. Renew. Sustain.
Energy Rev., 15 (5), pp. 2397-2412, https://doi.org/10.1016/j.rser.2011.02.022
DIT M., (2007). Palm kernel shell (PKS) is more than biomass for alternative fuel after
2005. Proceedings of the PIPOC 2007, International Palm Oil Congress (Chemistry &
Technology), pp. 275-287.
ESA N.M., LING T.B., PENG L.S. (2013) By-products of rice processing: An overview
of health benefits and applications. J Rice Res 1:107, doi: 10.4172/jrr.1000107
FAO (2019). FAOSTAT website. http://www.fao.org/faostat/fr/?#data/QC.
GANYO K.K., MULLER B., GAGLO E.K., GUISSE A., CISSE N., ADAM M.
(2018). Optimisation du NPK et urée basée sur les informations climatiques pour
accroitre la production du sorgho en zones soudano-sahéliennes du Sénégal. Journal of
Applied Biosciences, 131, pp. 13293-13307. doi: 10.4314/jab.v131i1.5
GIDDE M.R., JIVANI A. P. (2007). Waste to wealth potential of rice husk in India, a
literature review. Proc. of the Int. Conf. On Cleaner Technologies and Environmental
Management PEC, Pondicherry, India, pp. 586-590.
GLÉ P., MASSOSSA-TELO G., HELLOUIN DE MENIBUS A., DEGRAVE-
LEMEURS M., GOURDOND E. (2021). Characterization and modelling of the sound
reduction of hemp-clay walls in buildings. Journal of Building Engineering, 40, 102315.
https://doi.org/10.1016/j.jobe.2021.102315
GUNASEKARAN K., KUMAR P.S., LAKSHMIPATHY M. (2011). Mechanical and
bond properties of coconut shell concrete. Construction and Building Materials, 25 (1),
pp. 92-98. https://doi.org/10.1016/j.conbuildmat.2010.06.053
HUSSAIN A., CALABRIA-HOLLEY J., LAWRENCE M., JIANG Y. (2019).
Hygrothermal and mechanical characterisation of novel hemp shiv based thermal
insulation composites. Construction and Building Materials, 212, pp. 561-568.
KOOPMANS A., KOPPEJAN J., (1998). Agricultural and forest residues—generation,
utilization and availability. Paper to Regional Expert Consultation on Modern
235
Applications of Biomass Energy. Bangkok: Regional Wood Energy Development
Programme in Asia. http://rwedp.org/p_residues.html
LE BARTZ C., MICHAS M., FOUQUE C. (2009). Les roselières en France
métropolitaine premier inventaire (1998-2008). Faune Sauvage n°283. 13p.
LENORMAND H., LEBLANC N. (2020). Des particules végétales en abondance et
renouvelables. 6p. https://www.construction21.org.
LEWANDOWSKI I., CLIFTON-BROWN J., KIESEL A., HASTINGS A., IQBAL Y.
(2018). Miscanthus. E. Alexopoulou (Ed.), Perennial Grasses for Bioenergy and
Bioproducts, Academic Press, pp. 35-59. https://doi.org/10.1016/B978-0-12-812900-5.00002-3.
NOZAHIC V., AMZIANE S., TORRENT G., SAÏDI K., DE BAYNAST H. (2012).
Design of green concrete made of plant-derived aggregates and a pumice–lime binder.
Cement and Concrete Composites, 34, 2, pp. 231-241.
https://doi.org/10.1016/j.cemconcomp.2011.09.002 .
NTIMUGURA F., VINAI R., HARPER A.B., WALKER P. (2021). Environmental
performance of miscanthus-lime lightweight concrete using life cycle assessment:
Application in external wall assemblies. Sustainable Materials and Technologies, 28, doi:
10.1016/j.susmat.2021.e00253
PANTAWEE S., SINSIRI T., JATURAPITAKKUL C., CHINDAPRASIRT P. (2017).
Utilization of hemp concrete using hemp shiv as coarse aggregate with aluminium
sulfate [Al2(SO4)3] and hydrated lime [Ca(OH)2] treatment. Construction and Building
Materials, 156, 15, pp. 435-442, https://doi.org/10.1016/j.conbuildmat.2017.08.181
RAMANKUTTY N., MEHRABI Z., WAHA K., JARVIS L., KREMEN C., HERRERO
M., RIESEBERG L.H. (2018). Trends in global agricultural land use: implications for
environmental health and food security. Ann. Rev. Plant Biol., 69, pp. 789-815.
SOLOMON S. (2011). Sugarcane by-products based industries in India. Sugar Tech,
13(4), pp. 408-416.
SUDHIR C.R. (2005). Advancing bioenergy for sustainable development. guideline for
policymakers and investors. Vol. I, II, and III. Stockholm Environment Institute, 236p.
TORRIJOS M. (2016). State of development of biogas production in Europe. Procedia
Environmental Sciences, 35, pp. 881-889. https://doi.org/10.1016/j.proenv.2016.07.043.
WICHMANN S., KÖBBING J.F. (2015). Common reed for thatching - A first review of
the European market. Industrial Crops and Products, 77, 23, pp. 1063-1073.
XUE S., XIAO L., ZILI Y. (2015). Miscanthus production and utilization in Dongting
Lake Region, China. International Conference on Perennial biomass crops for a
resource-constrained world.
YADAV R.L., SOLOMON S. (2006). Potential of developing sugarcane by-product
based industries in India. Sugar Tech 8, pp. 104–111. doi 10.1007/s12355-011-0114-0.
ZUB H.W., ARNOULT S., BRANCOURT-HULMEL M. (2011). Key traits for
biomass production identified in different Miscanthus species at two harvest dates.
Biomass Bioenergy, 35, pp. 637-651, https://doi.org/10.1016/j.biombioe.2010.10.020
236
GROUND AND CONSTRUCTION IMPROVEMENT
TECHNOLOGIES
GEOSYNTHETICS & SUSTAINABLE
GEOENVIRONMENTAL ENGINEERING
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Influence of fiber crushing on light earth hygrothermal

properties

Farjallah ALASSAAD1, Karim TOUATI1, Daniel LEVACHER2,

Nassim SEBAIBI1

1. COMUE NU, Laboratoire de Recherche, ESITC Caen, 1 Rue Pierre et Marie Curie,
14610, Epron, France.
[email protected]
2. COMUE NU, M2C UMR 6143 CNRS, UNICAEN, 24 Rue des Tilleuls, 14000,
Caen, France.

Abstract:
Present study aims to evaluate the effect of fiber crushing on the structural and
hygrothermal properties of light earth. For this purpose, reed fibers were treated
differently and incorporated in the light earth. The fibers were first cut, then partially
crushed. After that, crushed and uncrushed fiber were mixed into a clay soil. The
obtained light earth samples were structurally and hygro-thermally characterized. Fiber
grinding showed a decrease in the light earth hygrothermal properties, including thermal
conductivity, sorption, and water vapor permeability. This was assumed due to the
change in porosity.

Keywords:
Light earth, Fiber crushing, Porosity, Hygrothermal properties, Building insulation.

1. Introduction
In the context of the global energy transition, several approaches are being explored to
reduce the energy consumption and carbon footprint of buildings without compromising
occupant comfort. Some researchers are attempting to improve the thermal performance
of the walls without disturbing the building envelope by using heat storage system
(ALASSAAD et al., 2021, ANGO, 2011) or adding insulation layers (ZACH et al.,
2013). Otherwise, new innovative solutions targeting building wall improvement have
begun to emerge, mainly in concrete. It involves adding components to the cement
matrix which can enhance the thermal behavior, such as foaming to increase the voids
in the concrete (RUIWEN, 2004) or including light aggregates (CHIDIGHIKAOBI,
2019, NGUYEN et al., 2014, YUN et al., 2013). The porosity offered by the foam or
lightweight aggregates decreases the thermal conductivity of the concrete as a function
of their dosage in the mix. In compliance with thermal and environmental regulations,

Corresponding author mail: [email protected]

the well-known earth construction techniques (adobe, cob, light earth, etc.) have been
revised. So many investigations have been undertaken for a better understanding of
earth-based materials thermal, hygroscopic and mechanical behavior, in order to find an
innovative solution limiting energy consumption (COLINART et al., 2020,
GOODHEW et al., 2019, HAMARD, 2017, HOLZHUETER & ITONAGA, 2017,
LABOREL-PRÉNERON et al., 2016, PHUNG, 2018). The use of earth-based materials
in construction is a true alternative. Earth is a naturally occurring resource that is often
widely available. Most earth that contain clay can be used for construction (HOUBEN
& GUILLAUD, 2006). Materials from renewable raw materials (such as plant
materials) are an answer to the problem of depletion of natural resources. Together with
recyclable materials (such as earth), these materials provide an answer to the problem of
waste at the end of the building's life. Earth-based materials are characterized by their
good thermal mass and capacity to regulate the buildings hygrometry with a low
economic and environmental cost. It retains or releases moisture in the air in accordance
with the ambient humidity, contributing to a healthier indoor environment (ANGER et
al., 2011, CAGNON et al., 2014, GIUFFRIDA et al., 2019, MEDJELEKH et al., 2017,
TOURÉ et al., 2017, ZHANG et al., 2020). This natural moisture regulation is a quality
missing in conventional building materials such as concrete. Many studies have been
carried out to improve the thermal properties of earth-based materials. For example, the
density impact on the hygrothermal properties of earth-based materials has attracted
particular interest (CAGNON et al., 2014, COLINART et al., 2020, LABAT et al.,
2016, MEDJELEKH et al., 2017, NIANG et al., 2018, PHUNG, 2018, PHUNG et al.,
2019). Furthermore, these studies have demonstrated the effect of fiber type and shape
on density, porosity and subsequently on hygrothermal properties. Present study aims to
evaluate the effect of fiber shape on porosity and hygrothermal properties. For this
purpose, reed straws were processed differently. One part is simply cut into length of 4
cm keeping the straw aspect, the other part is cut but also crushed. These straws are
used in a mixture of light earth for thermal insulation purposes.

2. Earth-based material preparation

Present study was inspired by the CobBauge research project in which different
formulations were tested. Thus, the most thermally efficient earth-fiber mixture,
reported within the project, was selected (DOCUMENTATION TECHNIQUE, 2018).
Thereby, the material mix is composed of silty-clay earth and 25 % wt. of reed.
As for soil characterization, plasticity, consistency and clay content were determined
based on the Atterberg limits and the methylene blue value (MBV) according to the
standards NF P94-051 (NF P94-051, 1993) and NF P94-068 (NF P94-068, 1998),
respectively. The limits (liquidity limit (LL) and plasticity limit (PL) serve as an
indicator of an earth's plasticity by giving the plasticity index (PI). The methylene blue
test offers an indicator of the clay content and reveals the clay's activity. Results of soil
238
characterization are given in Table 1. Soil classification is performed based on
conventional geotechnical analysis and considering the applicable standards. Besides,
the grain size distribution was investigated following the standard XP P94-041 (XP
P94-041, 1995). The methylene blue value and grain size distribution allow the soil
classification as sandy, silty, or clayey.

Table 1. Soil characterization.

Soil distribution [%] LL LP PI MBV
Parameter Particle size (<2 Particle size (<80
[%] [%] [%] [g/100g]
mm) µm)
Soil 100 95 57.8 42.5 5.64 15.3

As mentioned in the introduction, the fibers are processed differently. Both fibers are
cut at 4 cm, but a part is then crushed. Fibers aspect is shown in Figure 1.

Figure 1. Aspect of fibers before and after crushing.

The formulation i.e. each dosage, is defined according to the mass of soil mass. 25 %
mass of fiber and 100 % mass of water is added. Water is added first. With this amount
of water, the earth gets to its liquid state. Then, the fiber is added and mixed for about 2
min. The mixing procedure is carried out within a concrete mixer (PROVITEQ
Concrete Mixer 65 L). The molds are filled and placed in an oven at 40 °C.
Characterization of the prepared samples begins when the samples are dry. Table 2
summarizes the performed tests, the standards met, the size, and the number of samples
tested.

Table 2. Standard, number, and size of samples for testing.

Test Standard Sample dimensions [cm] Number of samples
Moisture sorption isotherm (NF EN ISO 12571, 2013) 3x3x3 3
Water vapor permeability (NF EN ISO 12572, 2016) Ø15x5 3
Thermal conductivity (ISO 8301, 1991) 22x22x4 3
Porosity (NF ISO 5017, 2013) Ø11x3 3

239
3. Experimental method

3.1 Porosity
Porosity is an essential physical parameter in material characterization. According to the
NF ISO 5017 standard (NF ISO 5017, 2013), this parameter can be measured by
immersing small samples in a liquid, here it is a non-wetting oil, dearomatized oil.
The samples undergo vacuum saturation in a desiccator for at least 24 hours, allowing
the liquid to replace the air in the pores without interacting with the sample. Then they
are weighed with the oil and then weighed in air. Finally, the samples are oven-dried at
105 ± 5 °C. This allows the knowledge of porosity volume initially filled with oil. The
accessible porosity p0 is given by:

(1)

where 𝑀dry is the mass of dry specimen, 𝑀𝑎𝑖𝑟, the mass of saturated specimen in air and
𝑀oil, the mass of saturated specimen in oil. Light earth, like all hygroscopic materials, is
known to have a significant impact on building’s interior air quality and energy
consumption. In the following section, light earth hygrothermal properties are described
along with the methodologies used to study them.

3.2 Hygroscopic behavior

The first step is the study of the interaction between light earth and moisture using the
Dynamic Vapor Sorption technique (ProUmid SPSx-1μ). This equipment offers
accurate monitoring of sample mass and sorption kinetics with a precision balance and
careful temperature and humidity control. In this study, sorption isotherms of raw
materials (earth and fibers) and mixtures are examined according to the standard ISO
12571 (NF EN ISO 12571, 2013).
The pre-dried samples were exposed to an environment with relative humidity varying
from 10 % to 90 % in 5 steps while keeping the operating temperature constant at 23°C.
Afterward, the dry cup method is used to study the capacity of a material to let water
vapor pass through. This property called water vapor permeability was measured
according to the standard ISO 12572 (NF EN ISO 12572, 2016). The mass tracking of
the sample is done under a humidity gradient (0 % RH inside, 50 % RH outside).
Measurements of water vapor permeability by the dry cup method provide insight into
the material's behavior when moisture transfer is dominated by vapor diffusion. The
influence of type of fibers used on the hygroscopic behavior of earth-fiber mixtures can
be assessed by determining moisture sorption isotherms and water vapor permeability.

240
3.3 Thermal conductivity
Thermal conductivity is an important parameter to consider when developing a
building’s insulating materials. This parameter characterizes the ability of a material to
conduct heat from a hot spot to a cold one. In this work, it is measured with a Heat Flow
Meter HFM 436 Lambda using Fourier’s law (GOOCH, 2011) and used for calculation.
The measurements were performed at 14 °C, 24 °C, 34 °C with a temperature difference
between the two sides of the samples fixed at 10 °C.

4. Results and discussion

4.1 Porosity
To observe fiber's porosity, scanning electron microscopy technique has been used.
From SEM images reported in Figure 2, it can be seen that crushed and uncrushed fibers
present relatively the same structure. However, the grinding degrades slightly the fibers
pores in the different layers and destroys cavities, see Figure 2. Crushing induces a
slight light earth porosity variation.

Figure 2. MEB image of crushed (right) and uncrushed fiber (left).

241
As it can be seen in Table 3, light earth with uncrushed fibers is more porous than the
one with crushed fibers. Furthermore, this decrease in porosity will surely impact the
hygrothermal behavior of the light earth.

Table 3. Light earth porosity when crushed and uncrushed fibers are incorporated.
Fiber used Crushed fiber Uncrushed fiber
Porosity [%] 54 58

4.2 Hygroscopic behavior

In present study, two hygroscopic properties have been studied. The first one is related
to the material's ability to adsorb/absorb moisture (sorption). This property is directly
linked to the material's constituents. In present work, light earth constituents are kept
unchanged. Consequently, regarding its sorption behavior, no significant change will be
held. This assumption is confirmed with the results obtained experimentally, see Figure
3. However, water vapor permeability is largely affected by porosity as moisture seeps
through the pores. The light earth's water vapor permeability decreases, and its water
vapor resistance increases when the fibers are crushed, see Table 4. This slight
resistance increase is supposed due to the small reduction in porosity.

Figure 3. Moisture sorption isotherm of light earth when crushed and uncrushed fibers
are considered.

Table 4. Water vapor resistance factor of light earth when crushed and uncrushed
fibers are considered.
Formulation Crushed fiber Uncrushed fiber
Water vapor resistance factor μ 7,237 7,101

242
4.3 Thermal conductivity
Materials thermal conductivity depends, among others, on the constituent nature, their
density as well as porosity. With the decrease of porosity, the mixture loses air-filled
pores and consequently insulating properties are affected. When the fibers are grinded,
the porosity decreases, and the thermal conductivity increases. As a result, light earth
loses approximately 23 % of its thermal conductivity, see Figure 4.

Figure 4. Thermal conductivity of light earth at three different temperatures.

5. Conclusion
In present study, we were interested in the effect of fiber crushing on light earth thermal
and hygroscopic properties. Samples with different cut and/or crushed fibers have been
prepared. Thermal conductivity, sorption behavior, and water vapor permeability have
been experimentally studied. Regarding thermal properties, the study showed that light
earth thermal conductivity increases when the fibers are crushed. This is supposed due
to the decrease in the material’s porosity resulting from the fiber crushing. The fiber
crushing has also affected the light earth water vapor permeability. Nevertheless, the
sorption behavior does not change since the mixture constituents have been kept
unchanged.

6. References
ALASSAAD F., TOUATI K., LEVACHER D., SEBAIBI N. (2021). Impact of phase
change materials on lightened earth hygroscopic, thermal and mechanical properties. J.
Build. Eng. 41, 102417. doi:10.1016/j.jobe.2021.102417

243
ANGER R., FONTAINE L., JOFFROY T., RUIZ E. (2011). Construire en terre, une
autre voie pour loger la planète. Sect. Privé Dév. Rev. Bimest. Proparco 10, 18–21.
ANGO S. E. (2011). Contribution au stockage d’énergie thermique en bâtiment :
développement d’un système actif à matériaux à changement de phase. PhD, Arts et
Métiers ParisTech, https://pastel.archives-ouvertes.fr/pastel-00650275
CAGNON H., AUBERT J. E., COUTAND M., MAGNIONT C. (2014). Hygrothermal
properties of earth bricks. Energy Build. 80, 208–217. doi: 10.1016/j.enbuild.2014.05.024
CHIDIGHIKAOBI P. C. (2019). Thermal effect on the flexural strength of expanded
clay lightweight basalt fiber reinforced concrete. Mater. Today Proc. International
Conference on Modern Trends in Manufacturing Technologies and Equipment 2019 19,
2467–2470. doi: 10.1016/j.matpr.2019.08.110
COLINART T., VINCESLAS T., LENORMAND H., MENIBUS A. H. D., HAMARD
E., LECOMPTE T. (2020). Hygrothermal properties of light-earth building materials.
J. Build. Eng. 29, 101134. doi: 10.1016/j.jobe.2019.101134
DOCUMENTATION TECHNIQUE. (2018). CobBauge. http://www.cobbauge.eu/documentation-
technique/
GIUFFRIDA G., CAPONETTO R., NOCERA F. (2019). Hygrothermal Properties of
Raw Earth Materials: A Literature Review. Sustainability. doi:10.3390/su11195342
GOOCH J. W. (2011). Fourier’s law of heat conduction. In: Encyclopedic Dictionary
of Polymers, J. W. Gooch, ed., 323–323. New York, Springer. doi:10.1007/978-1-4419-6247-
8_5262
GOODHEW S., CARFRAE J., HOOD-CREE K., FOX M., BOUTOUIL M., STREIFF
F. (2019). Building with earth: How we are working to revive an ancient, sustainable
building technique. Constr. Res. Innov. 10(4), 105–108. Taylor & Francis.
doi:10.1080/20450249.2019.1700077
HAMARD E. (2017). Rediscovering of vernacular adaptative construction strategies
for sustainable modern building: application to cob and rammed earth. Université de
Lyon,
https://www.researchgate.net/publication/323387530_Rediscovering_of_vernacular_adaptative_construction_strategi
es_for_sustainable_modern_building_application_to_cob_and_rammed_earth
HOLZHUETER K., ITONAGA K. (2017). The potential for light straw clay
construction in Japan: An examination of the building method and thermal
performance. J. Asian Archit. Build. Eng. 16(1), 209–213. doi:10.3130/jaabe.16.209
HOUBEN H., GUILLAUD H. (2006). Traité de construction en terre, CRATerre,
Parenthèses, Marseille, https://www.eyrolles.com/BTP/Livre/traite-de-construction-en-terre-9782863641613/
ISO 8301. (1991) Thermal insulation - Determination of steady-state thermal resistance
and related properties - Heat flow meter apparatus.
https://www.iso.org/cms/render/live/en/sites/isoorg/contents/data/standard/01/54/15421.html
LABAT M., MAGNIONT C., OUDHOF N., AUBERT J.-E. (2016). From the
experimental characterization of the hygrothermal properties of straw-clay mixtures to
244
the numerical assessment of their buffering potential. Build. Environ. 97, 69–81. doi:
10.1016/j.buildenv.2015.12.004
LABOREL-PRÉNERON, A., AUBERT J. E., MAGNIONT C., TRIBOUT C.,
BERTRON A. (2016). Plant aggregates and fibers in earth construction materials: A
review. Constr. Build. Mater. 111, 719–734. doi: 10.1016/j.conbuildmat.2016.02.119
MEDJELEKH D., ULMET L., DUBOIS, F. (2017). Characterization of hygrothermal
transfers in the unfired earth. Energy Procedia Materials & Energy I (2015) 139, 487–
492. doi: 10.1016/j.egypro.2017.11.242
NF EN ISO 12571. (2013). Hygrothermal performance of building materials and
products - Determination of hygroscopic sorption properties. https://sagaweb.afnor.org
NF EN ISO 12572. (2016). Hygrothermal performance of building materials and
products - Determination of water vapour transmission properties - Cup method.
https://sagaweb.afnor.org
NF ISO 5017. (2013). Produits réfractaires façonnés denses — Détermination de la
masse volumique apparente, de la porosité ouverte et de la porosité totale.
https://sagaweb.afnor.org
NF P94-051. (1993). Soil: investigation and testing. Determination of Atterberg’s
limits. Liquid limit test using Casagrande apparatus. Plastic limit test on rolled thread.
https://cobaz.afnor.org/notice/norme/nf-p94-
051/FA020765?rechercheID=795218&searchIndex=1&activeTab=all#id_lang_1_descripteur
NF P94-068. (1998). Soils: investigation and testing. Measuring of the methylene blue
adsorption capacity of à rocky soil. Determination of the methylene blue of a soil by
means of the stain test. https://sagaweb.afnor.org
NGUYEN L. H., BEAUCOUR A.-L., ORTOLA S., NOUMOWÉ A. (2014). Influence
of the volume fraction and the nature of fine lightweight aggregates on the thermal and
mechanical properties of structural concrete. Constr. Build. Mater. 51, 121–132. doi:
10.1016/j.conbuildmat.2013.11.019
NIANG I., MAALOUF C., MOUSSA T., BLIARD C., SAMIN E., THOMACHOT-
SCHNEIDER C., LACHI M., PRON H., MAI T. H., GAYE S. (2018). Hygrothermal
performance of various Typha–clay composite. J. Build. Phys. 42(3), 316–335.
doi:10.1177/1744259118759677
PHUNG T. A. (2018). Formulation et caractérisation d’un composite terre-fibres
végétales: la bauge, PhD thesis, Normandie Université, https://tel.archives-ouvertes.fr/tel-
01938827
PHUNG T. A., LE GUERN M., BOUTOUIL M., LOUAHLIA H. (2019).
Hygrothermal behaviour of cob material. In: Earthen Dwellings and Structures Springer
Transactions in Civil and Environmental Engineering, B. V. V. Reddy, M. Mani & P.
Walker, eds., 345–356, Singapore, Springer, doi:10.1007/978-981-13-5883-8_30
RUIWEN K. (2004). Properties of high-strength foam concrete, Master thesis, National
University of Singapore. https://scholarbank.nus.edu.sg/handle/10635/27683
245
TOURÉ P. M., SAMBOU V., FAYE M., THIAM A., ADJ M., AZILINON D. (2017).
Mechanical and hygrothermal properties of compressed stabilized earth bricks (CSEB).
J. Build. Eng. 13, 266–271. doi: 10.1016/j.jobe.2017.08.012
XP P94-041. (1995). Soil: investigation and testing. Granulometric description. Wet
sieving method. https://sagaweb.afnor.org
YUN T. S., JEONG Y. J., HAN T.-S., YOUM K.-S. (2013). Evaluation of thermal
conductivity for thermally insulated concretes. Energy Build. 61, 125–132. doi:
10.1016/j.enbuild.2013.01.043
ZACH J., HROUDOVÁ J., BROŽOVSKÝ J., KREJZA Z., GAILIUS A. (2013).
Development of thermal insulating materials on natural base for thermal insulation
systems. Procedia Eng. Modern Building Materials, Structures and Techniques 57,
1288–1294. doi: 10.1016/j.proeng.2013.04.162
ZHANG L., SANG G., HAN W. (2020). Effect of hygrothermal behaviour of earth
brick on indoor environment in a desert climate. Sustain. Cities Soc. 55, 102070. doi:
10.1016/j.scs.2020.102070

246
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Reinforcement design using geosynthetics for foundations on

karstic zones: analytical calculation and numerical stress-
strain model

Paulo CASTRO 1, Isabelle SILVA 1, Paula MARTINS 1, Filipe XAVIER 1,

Rodrigo FONSECA1

1. Golder Associates Brasil, R. Antonio de Alburquerque, 194 – Funcionarios, Belo

Horizonte, MG, 30112-010, Brazil.
[email protected]; [email protected]; [email protected];
[email protected]; [email protected]

Abstract:
Karst zones are defined by their high solubility, porosity, and consequence of subsidence
occurrences. Considering that about 20% of of ice-free land is made up of karst regions,
the present work aimed to reinforce by geosynthetics a hypothetical foundation subject to
windings through the analytical methodology of BRIANÇON & VILLARD (2008) and
by modeling numerical stress-strain in RS2 software (ROCSCIENCE, 2018). Initially, a
hypothetical foundation to be strengthened was defined, in which there is the presence of
circular dolines of a maximum of 3 meters in diameter. For the determination of the
mechanical properties of the reinforcement, the methodology proposed by BRIANÇON
& VILLARD (2008) was applied, having as a premise a surface deformation of, at most,
5%, whose results generated dimensioning abacuses to define the length parameters
anchoring and traction effort of the geogrid. Numerical analyzes of the stress-strain type
were performed to verify the level of strain related to the resistance of the foundation
materials, considering the values obtained by the analytical method. The reinforcement
by a geogrid set, for reinforcement and a woven geotextile, for physical separation, in
elastoplastic regime, obtained resistance reduction factors (SRF) corresponding to the
maximum surface deformations of 5%. The SRF obtained was consistent with good
practices. It was observed that the deformations caused in the geosynthetics did not lead
to their rupture.

Keywords:
Karst Zone, Geosynthetics reinforcement, Stress-strain analysis, Analytical method.

Corresponding author mail: [email protected]

1. Introduction
All voids in the soil are elements of weakness within a rocky mass, and a karst zone is
distinguished by having the largest natural voids, where failure in the ceiling can create a
significant geological hazard. The natural consequence of the progressive failure of the
top is the upward migration of voids, which can reach the surface where it causes
sinkholes due to a collapse caused by a cavity (WALTHAM et al.,2005).
According to FORD & WILLIAMS (2007), karst zones arise from the combination of
high solubility in rock and well-developed secondary porosity. Also, according to FORD
& WILLIAMS (2007), approximately 20% of the Earth's continental surface, except for
polar zones, is occupied by karst zones, with 20 to 25% of the world's population
depending on the aquifers contained in these regions.
Given the probability of formation of dolines bellow structures foundations, there is some
methodology proposed to minimize damage by design a geosynthetic reinforcement. One
of the most used is that proposed by BRIANÇON & VILLARD (2008), which is
summarized in three stages: determination of the load applied to the geosynthetic,
determination of the strain of the geosynthetic, and determination of the settlement of the
ground. In addition to this analytical design, numerical modeling has been a useful tool
to incorporate the differential behavior of elements subject to stresses and deformations.
The present work has the objective of design reinforcement of a hypothetical foundation
in which there is a probability of sinkholes, using geosynthetics, through the analytical
methodology proposed by BRIANÇON & VILLARD (2008), and its verification by
stress-strain analysis numeric model in RS2 software (ROCSCIENCE, 2018).

2. Analytical methodology
Quantitative calculations for the design of a hypothetical foundation, considering the
occurrence of sinkholes, were carried out based on the analytical methodology applied by
BRIANÇON & VILLARD (2008). The general design procedure necessary to determine
the mechanical properties of geosynthetic reinforcement includes (BRIANÇON;
VILLARD, 2008):
- Determination of the load applied to the geosynthetics, depending on the overload
and the characteristics of the covering soil.
- Determination of the strain of the geosynthetic, considering its membrane effect and
its elongation in the anchorage areas.
- Determination of soil settlement, depending on its physical properties.
It is noteworthy that, at the junctions between the geosynthetic meshes, a minimum length
of overlap is required. In this area, when a collapse occurs, the free ends of the
geosynthetics and the weak friction of the interface lead to large deformations of the soil
surface and reduction of the traction forces acting on the geosynthetics concerning those
obtained in the continuous section. In this way, the displacement of the soil surface is

248
calculated in the overlapping sections and the maximum stresses are determined in
continuous sections. The following sub-items describe the methodology proposed by
BRIANÇON & VILLARD (2008).

2.1 Determination of the load applied to the geosynthetic

The load applied to the geosynthetic is calculated considering the cylindrical collapse of
the soil over the cavity using the method proposed by TERZAGHI et al., (1996). Thus,
through the geotechnical parameters for the foundation soils and the geometry of the
subsidence, the applied load is calculated. The soil parameters adopted for the
reinforcement landfill, used to calculate the load on the subsidence and on the anchored
areas are shown in Table 1.

Table 1. Input data for the methodology by BRIANÇON & VILLARD (2008).
Circular sinkhole geometry Soil Parameters
L [m] 3 𝜑 [°] 35
D [m] 3 𝐶𝑒 1,1
𝑝 [kN/m] 20 𝑘𝑎 0,27
𝛾 [kN/m³] 20 Q [kPa] 47,68

2.2 Geosynthetic Behavior due traction

The vertical strain of the geosynthetic over the sinkhole, linked to its horizontal strain in
the anchorage zones, and the membrane effect, result from the load applied. This
generates an increase in the length of the geosynthetic. Thus, in a continuous section
where the highest stress is found, a complex relationship is established between the
following factors: the membrane effect of the geosynthetic, the applied loads, the problem
geometry, the soil, and the geosynthetic behaviors. So, it is assumed that the
reinforcement is unidirectional, and the applied load is uniform. The requested traction
effort of the geosynthetic is then calculated using Equation 1.

𝑞𝐿 2𝛽𝑥 2
𝑇𝑀 = 2𝛽 √1 + ( ) (1)
𝐿

where 𝑇𝑀 is the tractive effort at a point M, in N/m; q is the vertical load above the
sinkhole, in kPa; 𝐿 is the length of the sinkhole, in m; 𝛽 the characteristic parameter of
the change in orientation of the geosynthetic at its initial point and 𝑥 the horizontal initial
position of a point M, in m.

2.3 Influence of changes in geosynthetic orientation

The change in the orientation of the geosynthetic at the end of the subsidence generates a
reduction of the traction effort in the geosynthetic, as illustrated in Figure 1. This

249
mechanism, represented by the characteristic parameter β, is calculated by interaction
based on Equation 2 considering Figure 1.

β = tan 𝜑𝐴 (2)

Figure 1. Influence of changes in geosynthetic orientation,

(BRIANÇON & VILLARD, 2008).

2.4 Geosynthetic behavior due strain

In the anchorage area, three possible situations are verified in a segment (Figure 2)
considering the horizontal strain of the geosynthetic at the border of the sink, at the end
of the geosynthetic and corresponding to the maximum friction stress.
In the overlapping sections, in addition to the friction between the soil and the
geosynthetic, the friction between the geosynthetic meshes is considered. For this case, it
is necessary to consider the equations obtained on the left and on the right for equilibrium.
Considering the limit balance of a section that the reinforcement is unidirectional, and the
load applied is uniform, there is a vertical deformation of the geosynthetic through
Equation 3.

x2 −(L2 ⁄4)
zM = β (3)
L

where 𝑧𝑀 the vertical strain of the geosynthetic at a point M.

2.5 Assessment of surface deformation

The displacement of the surface is less than the vertical displacement of the geosynthetics,
due to the expansion of the soil during the collapse or possible arching of stresses.
Considering that the largest settlement of the soil above the subsidence is observed in the
geosynthetic joint, if the deformation of the soil and the geosynthetic are assumed
parabolic, it is possible to find a relationship between the maximum displacement of the
surface (𝑠), expansion coefficient (𝐶𝑒 ) and the height of the ground cover (𝐻). This
relationship is expressed by Equation 4 to a circular sinkhole.

s = f − 2H(Ce − 1) (4)

250
2.6 Determination of anchoring length
The optimum anchorage length of the geosynthetic (𝐿𝑅 ) is defined as the length between
the middle and the end of the geosynthetic plus the length necessary to sustain the tensile
and strain effort of the geosynthetic.

2.7 Design
The design of the reinforcement by geosynthetic type geogrid, was carried out following
the procedure proposed by BRIANÇON & VILLARD (2008) for 3 m circular sinkhole.
Many values were set for the following parameters: geogrid stiffness, geogrid
deformation, and height of the protective landfill, according to acceptable design values.
Thus, corresponding values of surface deformation, anchorage length, and tensile strength
of the geosynthetic were calculated. It is noteworthy that the same procedure previously
detailed was done for circular sinkholes of 2 meters and 4 meters, as an evaluation of
sensitivity regarding the parameters of the geosynthetic.
The results are illustrated from Figure 2 to Figure 6, which were used as abacuses for the
design, each curve being representative of one valuated height of the protective layer.
As a premise, it was defined that the surface deformations must be minimal so that a
maximum deformation of the surface of 5 cm was established as the initial design point.
Thus, in Figure 2, an economically viable stiffness of the geosynthetic was determined
from the defined surface deformation of 5 cm. A 2.5 m protective landfill was then found
to provide the required geosynthetic stiffness of approximately 3,000 kN / m.
In Figure 3, the anchorage length of 5.5 m was determined from the height of the
embankment and the rigidity of the geosynthetic defined in the previous step. In Figure
4, a 130 kN / m geosynthetic stress load was applied to the geosynthetic, based on the
height of the landfill and the geosynthetic stiffness according to the first stage. Finally, in
Figure 5 it was verified that the expected deformation in the geosynthetic approaches 5%.

251
Figure 2. Determination of geosynthetic Figure 3. Determination of anchorage
stiffness and H from 5 cm surface length based on the geosynthetic stiffness
deformation. and the height of the landfill.

Figure 4. Determination of the Figure 5. Determination of the strain in

geosynthetic traction effort based on the the geosynthetic from the stiffness of the
geosynthetic stiffness and the height of geosynthetic and the height of the
the landfill. landfill.

252
3. Numerical modeling
In order to verify the foundation strain and stress states caused by sinkholes and resisted
by geosynthetics, as dimensioned in item 2.4 by the methodology of BRIANÇON &
VILLARD (2008), two-dimensional stress-strain models were performed, using the
software RS2 version 2019 10.002 by ROCSCIENCE (2018).

3.1 Analysis premises

As an initial modeling premise, there is a probability of natural sinkholes occurrence. The
representative section, illustrated in Figure 6, shows the reinforcement arrangement by
geosynthetics for stabilizing a hypothetical foundation with a circular sinkhole of 3 m.
The sinkhole formation leads to the modeling of two representative instances of the case.
A first instant when the in-situ condition is prior to the dissolution of the rock, in which
the observed stratigraphy develops its confining tensions and vertical loads. In the second
instant (Figure 6), the dissolution of the rock allows the sinkhole insertion in this
environment established by the conditioning loads of the first instant. It is noteworthy
that, for the representative model, the water table was disregarded.

Figure 6. Modeling section considering the sinkhole outcrop.

Due to the safety assessment, the option Shear Strength Reduction was performed in RS2,
which performs analyzes of the stress-strain interaction by finite elements, in order to
calculate a critical strength reduction factor (SRF) for each section. Therefore, the model
does not seek to access the conditions of stability, but to verify the strain level related to
the foundation material’s resistance in contrast to the existence of sinkholes.

3.2 Material parameters

The parameters and constitutive models adopted for the materials were based on those
adopted for the development of the BRIANÇON & VILLARD (2008) methodology.

253
Therefore, the resistance parameters and the elastic-plastic behavior of the resistive
materials related to the sinkhole were defined.
For the representative section, it was considered a foundation soil where the sinkhole is
located. The compacted soil represents the protection layer, and the cement soil serves as
protection for the geomembrane, as seen in Table 2.

Table 2. Soils parameters.

Ei c Ф  ϒ
Soil Model Parameter υ
(MPa) (kPa) (˚) (˚) (kN/m³)
Compacted
Plastic 7 0,30 10,0 35 35 20,00
soil
Effective
Foundation
Elastic stresses 20 0,30 10,5 35 - 20,00
soil
Soil Cement Plastic 50 0,30 20,0 35 35 20,00

Geosynthetics were treated as structural lines in the stress and strain modeling. It is
noteworthy that the contact interface between geosynthetic and soil is considered a
fragility zone. Thus, a 70% reduction in resistance in these regions was considered in
comparison to the resistance of the soil in question. Thus, Table 3 presents the parameters
associated with geosynthetics.

Table 3. Geosynthetics parameters.

Traction Characteristic Residual
Soil Model module resistance resistance
(kN/m) (kN/m) (kN/m)
Geogrid Plastic 3000 130 100
Geomembrane Elastic 100 - -
Geotextile Plastic 400 100 100

3.3 Numerical modeling results

According to WALTHAM et al., (2005), in the case of sinkholes, values corresponding
to safety factors revolve around 5, by analogy, the value found as SRF is consistent with
what the good practices indicate, as observed in Figure 8. For the SRF close to 5, there
are strains of 6 to 7 cm (Figure 7), close to 5 cm of the analytical dimension (item 2.4).
It is also possible to observe the geosynthetics displacements, represented by lower and
upper bars along its extension, caused by traction due to the sinkhole presence. Figure 8
illustrates the sliding of the geomembrane, geotextile and geogrid in relation to the contact
soils. It is verified that the traction solicitation did not lead to the rupture of these
elements, with displacements in the order of 10 cm close to the sinkholes and millimeters
at the extremities.

254
 on er en on er en fail

Figure 7. Convergence for SRF variation to shear for 3m sinkhole.

Note: The bars along the geosynthetics represent their slips related to the contact soils.

Figure 8. Vertical strains expected for 3m sinkhole, SRF = 13.85 with strain of 5.2m.

4. Conclusion
The dimensioning of a hypothetical foundation reinforcement with sinkholes probability
was presented using geosynthetics through the analytical methodology proposed by
BRIANÇON & VILLARD (2008). Verification is performed by stress-strain analysis
using a numerical model in RS2 software (ROCSCIENCE, 2018).
The analytical and numerical methodologies showed levels that comply with the
foundation reinforcement propitious to sinkholes of up to 3 meters in diameter, in relation
to a surface deformation of 5%. However, such sizing does not prevent the outbreak of

255
larger sinkholes. For this reason, risk analysis is recommended in order to map the cavities
and validate the expected diameter of subsidence in non-hypothetical cases.
Therefore, the described methodologies show good adherence related to the behavior of
the proposed system for the scenario of subsidies appearance on the surface with
diameters up to 3 m.

5. References
BRIANÇON L., VILLARD P. (2008). Design of geosynthetic-reinforced platforms
spanning localized sinkholes. Geotextiles and Geomembranes, v.26, n.5, pp. 416–428.
doi: 10.1016/j.geotexmem.2007.12.005
FORD D., WILLIAMS P. (2007). Karst hydrogeology and geomorphology. John Wiley
& Sons Ltd. Chichester.
ROCSCIENCE. (2018). RS2 - 2D geotechnical finite element analysis. Toronto:
Rocscience Inc.
TERZAGHI K., PECK R. B., MESRI G. (1996). Soil mechanics in engineering practice.
3 ed. New York, NY: John Wiley & Sons, Inc.
WALTHAM A.C., BELL F.G., CULSHAW M.G. (2005). Sinkholes and subsidence:
karst and cavernous rocks in engineering and construction. Springer, 2005. 300p.

256
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Management of harbor sediments to reuse as backfill

Ali HUSSAN 1, Daniel LEVACHER 1, Salim MEZAZIGH 1

1. Comue Normandy University, UniCaen, M2C Caen, 24 rue des Tilleuls, 14000 Caen,
France.
[email protected]; [email protected]; [email protected]

Abstract:
Many studies have been carried out on the reuse of sediments as fill soils, and for the
most part they have only been conducted in the laboratory. They are always organized in
the same way and most of them follow well-established recommendations of national
authorities. The basic organizational diagram proposed is the following: sampling,
characterization, stabilization, on-site application. On-site application can be considered
later.
But what is the most efficient organizational scheme to adopt for the reuse of organic
sediments in backfill soils from a practical, economic, and environmental point of view?
If the sampling phase is unavoidable, the characterization tests can be minimized in order
to concentrate on the stabilization study, which is a hindrance to recovery and subsequent
application. A stabilization study with adapted and low-carbon binders on organic
sediments is used as an example in this paper.

Keywords:
Sediment management, Beneficial uses, Sediment stabilization, Methodology.

Corresponding author mail: [email protected]

1. Introduction
Dredging is necessary to maintain navigation in ports, harbors, and other water bodies.
Dredging produces a considerable amount of sediments. France for example generate
about 50 million cubic meter of these sediments per year, (ALZIEU et al., 1999), but also
sediment brings organic matters and other contaminations. Most of the time these
sediments are considered as waste due to their poor engineering properties. Therefore, a
handsome amount of money and time is needed to dispose them without effecting nature
too much. Moreover, with the new European Union directives (OSPAR CONVENTION,
1992), managers are encouraged to find new environmentally friendly solutions for
dredged materials. To reduce the climate impacts by the construction works, it is
necessary to increase the resource efficiency by reusing waste material or excavated
material, (BLENGINI & GARBARINO, 2010). Sediment management is becoming a
critical issue around the world, particularly in areas where the construction of harbor
facilities, the protection of coastal habitats, and the needs of tourism all compete for the
sustainable use of sediment resources. Contaminated sediments pose potential risks to
humans, wildlife, and aquatic organisms. Therefore, the management of these sediments
requires a special attention (APITZ et al., 2007; VELOSO-GOMES & TAVEIRA-
PINTO, 2003; XUE, 1999). Figure 1 gives the simple sediment management process.

Figure 1. Sediment recovery process.

2. Methodology
Dredged harbor sediments are increasingly being used as backfill and other construction
applications to save money and environment compared to the disposal of these sediments.
However, the sustainability of these soft organic material as backfill or other construction
applications is unclear (SCHLUE et al., 2009). Therefore, valorization/stabilization study
258
requires a lot of time and number of testing to understand the sustainability of these
organic sediments as backfill. The characterizations of the dragged soil with regards of
physical properties are significant in decision making either to be reused or disposed
(PEBBLES & THORP, 2001). Sediments can be characterized in all kinds as given in
Table 1, but what are the minimum requirements for them to be used as backfill?

Table 1. Summary of tests and applied standards.

Characteristics Methods and tests Test standards Findings
Water content Method by boiling NF P94-050 W (%)
Granularity Laser diffraction method ISO 13320-1 Cu, Cc
Methylene blue adsorption
Clay contents NF P94-068 MBV (g/100g)
ability
Density and specific
Helium gas pycnometer (g/cm3),
weight of solid DIN 51913
and water pycnometer γs(N/m3)
particles
Liquidity at the cup LL, PL, PI, CI
Atterberg Limits NF P94-051
Plasticity with a roll (%)
NF XP P94-
Organic matter Calcination method OM (%)
047
Analysis by X-ray Sediment
Mineralogy NF 13925-1
Diffraction (DRX). Minerals
Carbonates Calcimeter method NF P94-048 CaCO3 (%)
Material compaction
Compaction test NF P94-093 ρdOPM, wOPM
Miniature Proctor test
Settling down of sediments Concentration
NF XP P94-
Decantation test in the absence of effective Permeabilty K
060-2
stress Settlement
Settling of sediments in the NF XP P94- K, Cv, Cc, Es
Œdometer test
presence of effective stress 090-1 Settlement
Lime fixation test By using pH meter NF-ISO 10390 pH

The characterization of the sediments is the first step in any recovery process. It allows
to determine a set of properties necessary to choose a valuation chain to optimize this
process. Therefore, the physicochemical characteristics of sediment, the contaminants
interaction with the sediment matrix, the amount of sediment, as well as local
environmental factors, must all be considered when designing an appropriate sediment
recovery strategy (KRIBI, 2005). Once the sediments are characterized, they are than
added with different eco-binders. Different types of dredged sediments require different
amount of binder and impurities present in them also affect the amount of binder to be
added. The presence of large amount of organic matter can cause problems regarding the
cement hydration. The presence of acidic compounds can rise the pH of sediments and

259
therefore hurdle the pozzolanic reactions (KUJALA et al., 1996). The need of a large
amount of hydraulic binder can make the use of polluted dredged sediments non-feasible
from an economical point of view. The main motive of this paper is to understand how
characterization step can be minimized to earn more time for valorization process. What
are the useful tests in characterization for stabilization?
- Measurement of the water content of a soil is the most common action in Geotech and
allows expressing many other values by bringing it to an expression of dry mass as per
French standard NF P 94-050 (1995).
- The determination of the mineralogical composition allows us to understand the
interactions between the sediment and the hydraulic binder (BEL HADJ ALI et al.,
2014).
- The particle-size-distribution analysis is used to determine the size of grain and the
granularity (dimensional distribution of grains) of sediments as per French standard NF
P 94-056 (1996).
- Calcination: According to the GTR (LCPC-SETRA, 2000), materials with low organic
matter contents (OM <3%) are considered to be good Backfill materials and can be
treated with hydraulic binders. Otherwise (MO ≥ 3%), the materials must be the subject
of a specific study to evaluate their behavior to be treated with hydraulic binders.
- The determination of the organic content in sediment is an important step for the
identification of characteristics and the track of valorization. In effect, the organic matter
is a component undesirable in a material of construction (DUBOIS et al., 2009).
- Lime fixation test: To determine the percentage of lime needed, lime fixation test is
used. This test consists of measuring the pH of a lime-sediment-water mixture (NF ISO
10390) and determining for which lime dosage the pH of the suspension reaches a value
of more than 12, favorable to the development of cement hydrates. GTS also
recommends the percentage of hydraulic binder with respect to the percentage of lime
(LCPC-SETRA, 2000).
- The determination of the Atterberg limits allows us to determine the plasticity index PI
and the consistency index CI as per the French standard NF P 94-051 (1996). PI gives
the knowledge about the percentage of clay present in the sediments. The value of the
methylene blue (French standard NF P94-068, 1998) is the second parameter to
understand the “amount of clay present in sediments. The goal is to evaluate overall clay
richness of the sediment.
- Once the treatment formulation has been set, the parameters for compacting these
sediments at normal Proctor energy must be defined according to the NF P94-093
standard (1999), i.e., the optimum water content wopt for which the maximum dry
density ρ is obtained ρdopt.
These properties of sediment should be well known before making samples for
mechanical testing. The samples should prepare by compacting them using normal
Proctor energy defined by Proctor test.
260
3. Application
The example of sediments from the port of Cherbourg (a city in France) was utilized to
well understand the essential characterization needed before valorization process.
Sediments were dredged from three different points of Cherbourg. Four buckets were
filled with dredged sediments at point 1 (49.6394901, -1.6193592), 5 buckets were filled
at point 2 (49.6396503, -1.6191782) and 3 buckets were filled at point 3 (49.6409674, -
1.6189857). These three points are mentioned as small red circle on Figure 2. Point
coordinates are linked to the WGS system.

Figure 2. Three collecting points at Cherbourg port.

Table 2. Characteristics of Cherbourg sediments.

Parameters CHER-1 CHER-2 CHER-3
Dmax (mm) 20 20 20
Size < 80μm (%) >70 >70 >70
Size < 2mm (%) >99 >99 >99
Cc 10.29 11.78 12.91
Cu 0.93 0.95 0.97
3
Dry density (g/cm ) 2.27 2.34 2.25
MBV (g/cm3)) 1.42 1 0.94
wn (%) >150 >150 >150
PI (%) 20.84 22.97 23.53
OM (%) 16.75 15.68 19.68
ρopt (g/cm )
3 1.32 1.34 1.34
wopt (%) 40.80 34.50 37.50
Relation wn and wopt wn> 1.3wopt wn> 1.3wopt wn> 1.3wopt

261
After the determination of characteristics mentioned in Table 2 for Cherbourg sediments,
valorization process can be started. Preparation of samples through compaction at normal
Proctor energy (Proctor test) is the next step before mechanical testing.

4. Strategy for obtaining backfill material

Mandatory characteristics of the sediments are now well known. Laboratory scale
experimentation can be commenced once the choice of eco-binder is made. If laboratory
scale studies give good results in both technical and economical point of view, pilot-scale
testing in the next stage. Otherwise repeat the laboratory scale experiments with different
binders with increase dosage or with pre-treatment of sediment prior to the addition of
eco-binders. Figure 3 gives the strategy for reusing the sediments.

Figure 3. Work flow chart and decision-making steps.

Previous studies (WILK, 2004; SAID et al., 2015; KUJALA et al., 1996) have shown that
the addition of hydraulic binder can fulfill the required mechanical characteristics criteria
for the material to be used as backfill or other applications. However, the chemistry of
dredged sediments is very complex and variable. Therefore, type of binder and its amount
required to make them feasible to be used as construction material is inconstant as well.
The management of polluted sediment has become the environmental and economical
concern for today’s world. Therefore, harbor and dam managers are encouraged to find
environmentally friendly solution to these materials. The choice of binders depending
upon their impacts on the environment are the following:
- Low carbon binders e.g., eco-binder NeoliX (use an activator with a mineral charge).
- Then pozzolanic by-product as sideline and gardalithe (slags)
- GGBS binders e.g., Ecocem (an ecofriendly solution)

262
- Specific binders e.g., Neutracem or CSA cement (with high clinkers)
- Conventional binders used as reference e.g., CEM-I (cement mostly consisted of clinker)
with or without the addition of lime, depending upon the percentage of organic matter
present in the sediments as in Figure 4.

Figure 4. Selection of binders: from standard binders to low carbon and eco-binders.

5. Conclusion
After obtaining the sediments through dredging process, find the geotechnical
characteristics of sediments. Using Proctor test (miniature Proctor to save the sediments)
determine the water content at which sediments have the maximum dry density and use
this water content to make samples (cylindrical or cubical) for mechanical testing. Pilot
scale studies are the next step once mechanical testing gives promising results. Choose
binder by avoiding clinker component as they impose high environmental risks. Binders
consisting of industrial by-products (activated or not) should be preferred as they are
environmentally friendly and economically stable. Pre-treatment of sediments with lime
(to tackle organic matter) or sand (size corrector) is sometime unavoidable.

6. References
ALZIEU C. (1999). Dragages et environnement marin: état des connaissances - Dredging
and marine environment: state of the art, Ifremer, https://archimer.ifremer.fr/doc/00000/1040/.
APITZ S., CARLON C., OEN A., WHITE S. (2007). Strategic framework for managing
sediment risk at the basin and site-specific scale. Sustainable management of sediment
resources, Vol. 3, pp. 77-106, https://doi.org/10.1016/S1872-1990(07)80064-X
BEL HADJ ALI I., LAFHAJ Z., BOUASSIDA M., SAID I. (2014). Characterization of
Tunisian marine sediments in Rades and Gabes harbors, International Journal of
Sediment Research, 29(3), pp. 391-401. https://doi.org/10.1016/S1001-6279(14)60053-6

263
BLENGINI G. A., GARBARINO E. (2010). Resources and waste management in Turin
(Italy): the role of recycled aggregates in the sustainable supply mix. Journal of Cleaner
Production, 18(10-11), pp. 1021-1030, https://doi.org/10.1016/j.jclepro.2010.01.027
DUBOIS V., ABRIAK N. E., ZENTAR R., BALLIVY G. (2009). The use of marine
sediments as a pavement base material, Waste Management, 29(2), pp. 774-782.
https://doi.org/10.1016/j.wasman.2008.05.004
KRIBI S. (2005). Décomposition des matières organiques et stabilisation des métaux
lourds dans les sédiments de dragage. PhD thesis, INSA Lyon, 220p., http://theses.insa-
lyon.fr/publication/2005ISAL0064/these.pdf
KUJALA K., MÄKIKYRÖ M., LEHTO O. (1996). Effect of humus on the binding
reaction in stabilized soils. Grouting and Deep Mixing, In Proceeding of IS-TOKYO 96,
The Second International Conference on Ground Improvement Geosystems, Tokyo, pp.
14-17.
LCPC-SETRA. (2000). Realization of backfill and subgrade. Technical guide for road
earthworks (GTR).
OSPAR Convention. (1992). Convention for the Protection of the Marine Environment
of the North-East Atlantic, Paris, 33p.
PEBBLES V., THORP S. (2001). Waste to resource: beneficial use of great lakes
dredged material. Great Lakes Commission, USA. 16p. www.glc.org/dredging
SAID I., MISSAOUI A., LAFHAJ Z. (2015). Reuse of Tunisian marine sediments in
paving blocks: factory scale experiment, Journal of Cleaner Production, 102(2015), 66-
77. doi.org/10.1016/j.jclepro.2015.04.138
SCHLUE B. F., KREITER S., MOERZ T. (2009). Time-dependent deformation of
dredged harbor mud used as backfilling material, Journal of waterway, port, coastal, and
ocean engineering, 135(4), pp. 154-163. https://doi.org/10.1061/(ASCE)0733-950X(2009)135:4(154)
VELOSO-GOMES F., TAVEIRA-PINTO F. (2003). Portuguese coastal zones and the
new coastal management plans, Journal of Coastal Conservation, 9(1), pp. 25-34,
https://doi.org/10.1652/1400-0350(2003)009[0025:PCZATN]2.0.CO;2
WILK C. M. (2004). Solidification/stabilization treatment and examples of use at port
facilities, In: Curtis SA (ed) Port development in the changing world; ASCE conference
proceedings, Houston, Ports 2004, p 1–10. doi:10.1061/40727(2004)92
XUE C. (1999). Coastal sedimentation, erosion, and management on the north coast of
Kosrae, Federated States of Micronesia, Journal of Coastal Research, pp. 927-935,
http://www.jstor.org/stable/4299012

264
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Dewatering cone tests: Evaluation of chemical conditioning

and geotextile efficiencies for a water treatment plant sludge

Matheus MÜLLER 1, Gabriel L. A. de OLIVEIRA 1, Delma VIDAL 1

1. Instituto Tecnológico de Aeronáutica – ITA, Praça Marechal Eduardo Gomes, 50, Vila
das Acácias, 12228-900 São José dos Campos/SP, Brasil.
[email protected]

Abstract:
Treatment plants generate sludge through its processes of cleaning, this sludge having a
high-water content and so needing correct handle and disposal. In that matter dewatering
technics are vital, separating the solid from the liquid phase, decreasing its moisture
content, and facilitating its handle. This paper discourses a not standardized test, the cone
test, as a possible preliminary test that helps to indicate the best chemical conditioning
and geotextile to use in a real case of dewatering with geotextile tubes. A new essay
procedure was proposed, and cone tests were conducted with sludge from a water
treatment plant in Brazil, testing four different geosynthetics, three woven and one
nonwoven geotextile. Results pointed that woven geotextiles have better efficiencies for
chemical conditioned sludge, reaching more than 214% in dewatering and 98% in
filtration efficiency, and nonwoven for sludge without conditioning, reaching more than
65% in dewatering and 95% in filtration efficiency. Also, a much faster drainage was
observed for woven geotextiles, while the nonwoven clogged easily, retaining moisture.
As a general conclusion it is possible to observe that for the tests made if polymerization
of the sludge is not possible, a nonwoven geotextile may be recommended, otherwise a
woven geotextile may be better.

Keywords:
Cone test, Dewatering, Woven geotextile, Nonwoven geotextile, Water treatment plant
sludge.

Corresponding author mail: [email protected]

1. Introduction
The world faces a time of consumerism and, with this, an imminent scarcity of resources,
where, with the population increase, the demand for products and inputs grows. As the
planet is limited by finite resources, a scenario of probable diverse shortcomings is
predicted for the future. As pointed out by GHISI et al., (2006), a resource marked by this
depletion/insufficiency is water.
Thus, the use of techniques that contribute to the preservation of natural resources is of
great value. Processes such as the reuse of resources, improvement of water treatment
stages, less disposal of sludges with high water content, recycling, among others, are
relevant. In that scenario the dewatering technique is interesting.
Dewatering is a technique to increase the solids content of an effluent/sludge. Thus, this
technique corresponds to the operation where the moisture content is reduced so that the
resultant residue can be handled as a semi-solid/solid (MÜLLER, 2019).
With the dewatering, the water drained can be recovered and reused in other activities or
correctly returned to water bodies. With this, the final, drier residue offers less danger to
the environment, with less potential for contamination and, due to water loss, the drier
residue has less volume, resulting in easier transportation and other benefits. Thus, this
volume reduction provides economic advantages to its transportation, correct disposal,
and carbon footprint (GUIMARAES et al., 2014).
Finally, the dewatered residue, depending on its composition, can be seen as a material
that can be reused and can be valued and used for various purposes, such as in agriculture,
as a soil corrector and fertilizer, in infrastructure works, as for landfills and roads
constructions, in the manufacture of ceramic products and others (COLIN, 2003;
DUBOIS, 2006; SEMCHA, 2006; CEDEX, 2007; DUAN, 2008; ANGER, 2014).
A wide variety of techniques for dewatering residue can be found, divided into natural,
mechanical, and mixed dewatering techniques, and further divided into continuous or
discontinuous dewatering processes (BOULLOSA ALLARIZ & LEVACHER, 2016). In
this study the dewatering technique in geotextile tubes is approached, focusing on
preliminary tests for choosing the geotextile and the chemical conditioning for the sludge.
In that way, cone tests and chemical conditioning tests were performed, compared
through qualitative and efficiencies analysis.
The cone test used is a test without standardization, widely used due to its simplicity and
speed of execution. This test consists of evaluating in a practical way the dewatering
performance of sludge in front of a geotextile, being carried out through the dumping of
sludge volume inside a geotextile cone positioned on a percolate collection container
(VERTEMATI, 2015).

2. Literature review
Correlation to aid in the dosage of the polymeric solution is proposed and filtration and
dewatering efficiencies are presented.
266
Where the working dosage for the polymeric solution can be checked with the aid of the
correlation described in Equation 1.

𝑉𝑠𝑜𝑙𝑢𝑡𝑖𝑜𝑛 𝑥 𝐶𝑠𝑜𝑙𝑢𝑡𝑖𝑜𝑛
𝑃𝑃𝑀 = (1)
𝑉𝑠𝑙𝑢𝑑𝑔𝑒

Where, PPM is the dosage of polymeric solution in parts per million (mg/l); Vsolution is the
volume of polymeric solution (l); Csolution is the concentration of the polymeric solution
(mg/l); and Vsludge is the volume of the sludge sample (l). MOO-YOUNG et al., (2002)
suggest the analysis of Filtration Efficiency, and with adjustments by AVANCINI (2017)
and TOMINAGA (2010), it can be obtained by comparing the amount of Total Solids
(TSfinal) of the filtrate, that is, the percolate liquid, with the initial amount of Total Solids
(TSinitial) of the sample of sludge, given by Equation 2.

𝑇𝑆𝑖𝑛𝑖𝑡𝑖𝑎𝑙 −𝑇𝑆𝑓𝑖𝑛𝑎𝑙
𝐹𝐸 = 𝑥 100 (%) (2)
𝑇𝑆𝑖𝑛𝑖𝑡𝑖𝑎𝑙

where: FE is the filtration efficiency (%); TSinital is the initial total solids (mg/l); and TSfinal
is the total solids after filtration (mg/l).
Dewatering efficiency can be obtained by comparing the percentage of final solids with
the percentage of initial solids (MOO-YOUNG et al., 2002):

𝑃𝑆𝑓𝑖𝑛𝑎𝑙 −𝑃𝑆𝑖𝑛𝑖𝑡𝑖𝑎𝑙
𝐷𝐸 = 𝑥 100 (%) (3)
𝑃𝑆𝑖𝑛𝑖𝑡𝑖𝑎𝑙

Where, DE is the dewatering efficiency (%); PSinitial is the percentage of initial solids by
mass (%); and PSfinal is the percentage of final solids by mass of the retained cake (%).

3. Materials and methods

3.1 Sludge
The sludge sample was collected in the Water Treatment Plant (WTP) of DCTA - São
José dos Campos/SP, Brazil, from the bottom of on decanter by pumping.
The sample was stored in a plastic opaque 40 liters gallon, in-room temperature and
protected from sunlight. This sample was later used in lab tests. The physical-chemical
characterization of the sample is shown in Table 1.

267
Table 1. Sludge hysical-chemical characterization.
Parameter Measurement method Sample Unit

pH Direct, Potentiometric 6,2 -------------

PS Standard Methods, Gravimetric 2,64 %

W Standard Methods, Gravimetric 3691,47 %

Al SEM 20,8 + 0,1 %MS

Fe SEM 8,6 + 0,4 %MS

O SEM 52,8 + 0,3 %MS

Si SEM 17,6 + 0,1 %MS

Ni SEM 0,2 + 0,1 %MS

3.2 Geotextile
Four different geotextile types were used in cone tests, three woven geotextiles (GTX
W1, GTX W2, and GTX W3) and one non-woven geotextile (GTX NW1). Table 2 shows
the characterization values provided by the product manufacturer.

Table 2. Characterization of the geotextiles used.

Parameter Standard GTX W 1 GTX W 2 GTX W 3 GTX NW 1

Da (mm) NBR ISO 12.956 0,200 0,400 0,200 0,060

Vnb (m/s) NBR ISO 11.058 20 55 15 39

Tnc (kN/m) NBR ISO 10.319 105 105 55 27

a
characteristic opening size; b index velocity; c tensile strength in both directions.

3.3 Chemical conditioning tests

3.3.1 Scope test in watch glass

As a preliminary qualitative test, to determine in a practical way which polymer is more
efficient for the formation of flakes in the studied sludge, mixtures in watch glasses were
performed. These mixtures containing 10ml of sludge to 1ml of polymeric solution with
a concentration of 1000mg/l. Four tests were performed in this way, one for each type of
polymer available, two cationic and two anionics, each pair representing one with low
and one with medium molecular weight.

268
With the completion of the mixture, the polymerization was evaluated qualitatively,
observing the size of the flakes, the resistance of the flakes formed, and the quality of the
supernatant. The selected polymer was then evaluated in concentration tests.

3.3.2 Polymer concentration test

With the previously selected polymer, concentration tests were carried out, to find the
best concentration of polymeric solution for the studied sludge. In this way, 19 samples
of 100ml of sludge were separated and, for each one, the volume of polymeric solution
in variable concentration was admitted and evaluated.
The concentrations of the polymeric solution in the sludge varied in such a way that in
each sample it was obtained from one to one from 1 to 10 ppm and from ten to ten from
10 to 100 PPM (mg/l). The samples were then mixed using the eight-drop technique
(VERTEMATTI, 2015), that is, with the use of two beakers the volume of sludge with
polymer solution was passed vigorously from container to container a total of eight times,
providing its total mixing and polymerization.
After the mixing process, the polymerization was evaluated qualitatively, observing the
size of the flakes, the resistance of the flakes, and the quality of the supernatant.

3.4 Cone Test

Cone tests were performed for the 4 geotextiles studied, and for each geotextile, tests with
and without polymerization were conducted. In this way, trying to evaluate the pertinence
and necessity of polymerization. After preliminary tests for all geotextiles, the best results
were confirmed through triplicate tests for each selected geotextile.
New methodology for carrying out cone tests is proposed, being:
1. Take a circular geotextile sample and fold it to make it in a cone shape, holding it
together with a metal clip. Place the cone in a ceramic dish and place it on a stove at 65°C
until it reaches mass constancy.
2. Reserve sufficient and properly identified ceramic diches to collect the effluent drained
from the geotextile specimen.
3. Homogenize the sludge sample with the help of a glass rod and separate 400g in a
beaker using an analytic scale to perform the cone test. From the same homogenized
sample, separate enough mass to perform a solid content test.
4. If sludge chemical conditioning is intended, take another empty beaker and place the
polymer solution with the aid of a pipette. Transfer the sludge to the beaker with polymer
solution and perform the eight-drop technique.
5. Arrange the geotextile specimen on an adjustable stand over the ceramic dish. The
ceramic dish must be placed over an analytic scale and the measurement of mass must be
taken each minute.
6. Pour the 400g of sludge from the beaker to the geotextile specimen and measure the
mass drained for 25 min.
269
7. Remove the geotextile cone specimen with the retained material from the support and
place it on another ceramic dish.
8. After the end of the drainage time, measure the mass of the retained material in the
geotextile cone and the mass of the liquid percolate in the ceramic dish.
9. Take both dishes to the stove, set in 65 °C, and measure its masses after they reach
mass constancy.
10. Evaluate efficiencies and solids content for the test.
11. Repeat the previous procedures for duplicate and triplicate.

4. Results and discussion

4.1 Sludge conditioning

The polymer with the best results in flake formation, solid-liquid phase separation, and
liquid phase clarification was both cationic and anionic medium molecular weight, as
shown in Figure 1, in b and d.

Figure 1. Scope test in watch glass results: a- cationic low molecular weight; b-
cationic medium molecular weight; c- anionic low molecular weight; d- anionic
medium molecular weight.

As a tiebreaker criterion, both watch glasses were subjected to a new mixture with the aid
of a glass rod and a new qualitative analysis was performed. In this second mixture, it
was observed that the flakes formed by the medium molecular weight cationic polymer
resisted more, maintaining their shape and suffering little fragmentation about those
formed by the medium molecular weight anionic polymer, so being the chosen one.

270
The results of the polymer concentration test are shown in Figure 2. Two distinct intervals
were analyzed, 1ppm to 10ppm and 10ppm to 100ppm.
It is possible to observe that for tests below 10 ppm the flakes are large or the polymer
concentration is insufficient for an efficient formation of flakes. For tests greater than
20ppm, the large flakes broke in the mixing process of eight drops, resulting in very small
flakes. Therefore, through the tests, it was found that the optimal relationship between
sludge volume and polymeric solution is between 10ppm and 20ppm.

Figure 2. Concentration test.

The dosage of 20ppm was selected for the cone tests, the criterion of choice was due to
the medium size flakes, resistance, and good separation of the solid phase from the liquid.

4.2 Cone tests without polymerization

Figure 3 shows the execution of the four cone tests without polymerization. It clearly
shows that woven geotextiles, having larger pores, allowed the sludge to pass through
with almost no retention of solid particles. It also demonstrates that the non-woven
geotextile was able to retain the solid particulate more efficiently, however, due to the
clogging of the geotextile's filtering surface, it trapped moisture.

271
 Figure 3. Cone test without polymer conditioning.

By performing the tests without polymerization, Table 3 is presented, showing the

filtration and dewatering efficiencies obtained.

Table 3. Cone test efficiencies without polymer conditioning.

Sample FE (%) DE (%)

GTX NW 1-1 97,16 65,57

GTX NW 1-2 95,37 82,43

GTX NW 1-3 97,44 82,59

GTX W 1 0,52 27,97

GTX W 2 0,16 57,29

GTX W 3 9,87 20,53

Also, during the tests, constant weighing of the percolate was carried out. Thus, Figure 4
presents the drainage curves for the tests without polymerization, illustrating that the
GTX W 1 and 2 retain virtually no solid particles, while the GTX W 3 retains some and
the GTX NW 1 retains more than the others geotextiles, even retaining moisture by
clogging.

272
 Figure 4. Mass drainage curves by time for cone tests without polymerization.

4.3 Cone tests with polymerization

Figure 5 shows the execution of the four cone tests with polymerization. It demonstrates
that the use of an appropriate polymeric solution and in its correct dosage results in better
dewatering than the non-use of chemical conditioning.

Figure 5. Cone tests with polymerization.

273
By performing the tests with polymerization, Table 4 is presented, showing the filtration
and dewatering efficiencies obtained.
Table 4. Cone test efficiencies with polymer conditioning.
Ensaio FE (%) DE (%)

GTX NW 1 99,31 113,35

GTX W 1-1 98,79 214,88

GTX W 1-2 99,48 228,13

GTX W 1-3 99,87 238,54

GTX W 2 98,57 240,37

GTX W 3 98,92 214,61

Also, during the tests, constant weighing of the percolate was carried out. Thus, Figure
6 presents the drainage curves for the cone tests with polymerization, illustrating that all
geotextiles obtained better drainage performance and speed through the chemical
conditioning. Still, it shows that the woven geotextiles, by its larger pore diameters,
obtained faster drainages.

Figure 6. Mass drainage curves by time for cone tests with polymerization

274
5. Conclusions
Through the polymerization tests, it was possible to realize that the choice of the polymer
through a simple test on watch glass is possible, avoiding the performance of
concentration tests for all available polymers. Thus, saving time and focusing on the most
detailed and time-consuming tests for the relevant polymer solution.
In the concentration tests it is noticed that lower dosages of polymeric solution often
resulted in larger flakes, while higher dosages result in smaller flakes. This phenomenon
can be explained by the speed and turbulence of this mixing procedure, where a lower
dosage provided the formation of late flakes, and a higher dosage resulted in the formation
of flakes earlier. In this way, the flakes formed by a lower dosage suffered less falls and
turbulence, whereas the flakes formed by a higher dosage suffered more falls and
turbulence, breaking in the process.
It is noticed that the cone and mixture tests by eight-drops are widely disseminated among
professionals in the area, not only due to the fidelity of the mixture process to the Chicane
used in the field but also due to the practicality of its execution, representing an easy and
quick way to be aware of the dewatering that is to be carried out.
Through the cone tests without polymerization, it was possible to observe that the woven
geotextiles due to the pore diameter, allowed the passage of almost all the solid particles
of the sludge. In contrast, the non-woven geotextile has clogged, retaining the solid
particulate and moisture.
In the cone tests with polymerization, it was observed that the retention of the solid
particulate was efficient for all the tested geotextiles. In this sense, drainage speed and
greater retention were observed, where the GTX W1 woven geotextile was more efficient
than the others.
Moreover, it is observed that the use of chemical conditioning in the chosen dosage of
polymeric solution, that promoted greater retention of solid particles and obtained better
results in the filtration and dewatering efficiencies.
Through the cone tests with and without polymerization, it was possible to observe that
the use of chemical conditioning prevented the clogging of the pores of the non-woven
geotextile, providing more efficient drainage and higher levels of final solids for the
retained material. It is also noticed that for all the cone tests carried out, the levels of
solids of the retained material were higher with the use of chemical conditioning.
Finally, it is observed that the chemical conditioning with the use of polymeric solution
for the formation of flakes in the sludge was fundamental for efficient retention of the
solid particulate, drainage speed, and effectiveness of the desired dewatering.

6. References
ANGER B. (2014). Caractérisation de sédiments fins de retenues hydroélectriques en
vue d’une orientation vers des filières de valorisation matière. Doctoral thesis of Caen
Normandy University, Unicaen, France, 316 p.
275
AVANCINI L.L.S. (2017). Análise do desaguamento do lodo de estação de tratamento
de água em sistemas de confinamento em geotêxtil. Dissertação de mestrado em
engenharia de infraestrutura aeronáutica – ITA, São José dos Campos, 135 f.
BOULLOSA ALLARIZ, B. LEVACHER, D. (2016) Mechanical dewatering solutions
for sediments. 15th Geo-Environmental Engineering – GEE, pp. 23-32.
CEDEX - Centro de Estudios y Experimentación de Obras Públicas (2007). Dredged
sediments. Technical sheet of the Ministry of Equipment of Spain, 18 p.
COLIN D. (2003). Valorisation de sédiments fins de dragage en technique routière.
Doctoral thesis of Caen Normandy University, Unicaen, France, 180 p.
DUAN Z. (2008). Caractérisation, stabilisation et solidification de sédiments fins
marins. Doctoral thesis of Caen Normandy University, Unicaen, France 198 p.
DUBOIS V. (2006). Étude du comportement physico-mécanique et caractérisation
environnementale des sédiments marins – Valorisation en technique routière. Doctoral
thesis of Artois University, France, 298 p.
GHISI E., MONTIBELLER A., SCHMIDT R W. (2006). Potential for potable water
savings by using rainwater: An analysis over 62 cities in southern Brazil. Building and
Environment, Volume 41, Issue 2, pp. 204-210, doi.org/10.1016/j.buildenv.2005.01.014
GUIMARAES M.G., URASHIMA D.C., VIDAL D. (2014). Dewatering of sludge from
a water treatment plant in geotextile closed systems. Geosynthetics International, v. 21,
pp. 310-320, doi.org/10.1680/gein.14.00018
MOO-YOUNG H. K., GAFFNEY D. A., MO X., (2002). Testing procedures to assess
the viability of dewatering with geotextiles tubes, Geotextiles and Geomembranes, Vol.
20, pp. 289-303, doi: 10.1016/s0266-1144(02)00028-6
MÜLLER, M. (2019) Efeito de diferentes configurações no deságue de lodos em sistemas
de confinamento de resíduos em geotêxtil, Dissertação de mestrado em engenharia de
infraestrutura aeronáutica – ITA, São José dos Campos, 102 f.
SEMCHA A. (2006) Valorisation des sédiments de dragage: Applications dans le BTP,
cas du barrage de Fergoug. Doctoral thesis of Reims Champagne-Ardenne, France, 175
p.
TOMINAGA, E. (2010) Análise dos procedimentos para avaliação de desempenho de
sistemas fechados com geotêxtil para desaguamento. São José dos Campos, 2010.
Dissertação de mestrado em infraestrutura aeroportuária do ITA.
VERTEMATTI, J.C. (2015) Manual brasileiro de geossintéticos. Blucher, 2. Ed. p. 570.

276
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Evaluating geotechnical properties and hydraulic performance

of soil-bentonite cutoff wall using in situ samples

Kazuki NISHIMURA 1, Yan TIAN 2, Atsushi TAKAI 2,

Toru INUI 3, Takeshi KATSUMI 2

1. Kyoto University, Graduate School of Engineering, Yoshida-honmachi, Sakyo-ku,

Kyoto, Kyoto, 606-8501, Japan.
[email protected]
2. Kyoto University, Graduate School of Global Environment Studies, Yoshida-honmachi,
Sakyo-ku, Kyoto, Kyoto, 606-8501, Japan.
[email protected]; [email protected];
[email protected]
3. Osaka University, Graduate School of Engineering, 1-1, Yamadaoka, Suita, Osaka,
565-0871, Japan.
[email protected]

Abstract:
Soil-bentonite (SB) is a mixture of bentonite and in-situ soil and has extremely low
hydraulic conductivity. The SB cutoff wall is one of the most effective ways for the
containment of ground contamination which is extensive or exists directly under existing
structures. Adequate barrier performance of SB cutoff walls needs to be maintained for a
long period to prevent contaminants from spreading into the surroundings. Since soil and
bentonite are mixed in situ without excavation during the construction of SB cutoff walls,
it is essential to assess the homogeneity and performance using undisturbed soil samples
collected after the construction. This paper reports the hydraulic and sorption properties
of a 40-m deep SB cutoff wall. The authors analyzed hydraulic conductivity, sorption
performance and geotechnical properties such as Atterberg limits at several depths of SB.
Experimental results demonstrate that the properties of the SB cutoff walls meet the
performance-based criterion in terms of barrier performance and are homogeneous at any
depth. In addition, high applicability as a hydraulic barrier wall for the in-situ containment
method was implied.

Keywords:
Soil-bentonite, Cutoff wall, Ground contaminant, Post-construction quality, Hydraulic
conductivity, Sorption performance, Arsenic.

Corresponding author mail: nishimura.kazuki.47a @st.kyoto-u.ac.jp

1. Introduction
When soil is contaminated by toxic substances leaked from factories and other facilities,
there is a risk that human health may be threatened by direct ingestion of the soil (e.g.,
through ingestion or skin contact) or by drinking the groundwater in which the toxic
substances are dissolved. Because excavation and/or active remediation of contaminated
soil is costly and technically difficult when ground contamination exists over a wide area
or directly under existing structures, in situ containment technique using vertical cutoff
walls can be an effective countermeasure to prevent the migration of contamination.
This study targeted a vertical cutoff wall whose material is a soil-bentonite (SB) mixture.
SB is a type of low-permeability material and is widely used in in-situ containment
methods because of the following merits. First, montmorillonite contained in sodium
bentonite swells when it comes in contact with polar fluids such as water. This swollen
bentonite fills the pores between soil particles, and the hydraulic conductivity, k, of SB
becomes extremely low (e.g., TAKAI et al., 2013). Second, SB also has high
attenuation/sorption capacity against contamination including arsenic (MINJA & EBINA,
2002), lead (DANIELS et al. 2004), and cesium (INUI et al. 2016). Third, SB wall itself
has the capability to adsorb heavy metals and other contaminants since SB is made of
clay minerals (IKEDA, 2014).
While SB mixtures have advantages relative to other materials as impervious materials in
vertical cutoff walls, it is unclear that SB wall is being created vertically evenly and has
an adequate hydraulic performance relative to the performance-based criterion since the
in-situ soil and bentonite are directly mixed and stirred under the ground in situ (TAKAI
et al., 2016). Thus, it is essential to confirm post-construction quality. Regarding the post-
construction quality, SB cutoff wall is expected to have the adequate hydraulic and
sorption performance (IKEDA, 2014), and also homogeneity of SB may have a strong
influence on the outflux of vertical cutoff walls (BRITTON et al., 2005). However, it is
challenging to collect undisturbed core samples of the SB cutoff wall because the strength
stiffness of SB is relatively low. So, there are few cases in the world where the post-
construction quality of the SB cutoff wall has been clarified.
In this study, physical property tests, hydraulic conductivity tests and batch sorption tests
were conducted on core samples from a 40 m long SB cutoff wall to examine the vertical
homogeneity, hydraulic and sorption performance of the SB cutoff wall. We analysed the
sorption performance against Arsenic (As) since As contamination is reported as one of
the most widely distributed contamination in 2018 by Ministry of the Environment
Government of Japan (2018).

2. SB cutoff wall and materials

The SB cutoff wall targeted in this study was constructed by the equal thickness trench
cutting machine used in the TRD (Trench cutting Re-mixing Deep wall) method, and the
machine has the merit to erect the homogeneous vertical wall by mixing slurry and
278
powdered additive with the in-situ soil while excavating the ground with a chainsaw-like
cutter. In the construction of the SB cutoff wall, 3.25 kg/m3 of 5% bentonite slurry and
125 kg/m3 of powdered bentonite were added to 1981 kg/m3 of in-situ soil with an average
natural water content of about 24.7%, resulting in a mass ratio of bentonite to the SB
mixture of about 6.5%. Na-bentonite with free swelling index over 16mL/2g was used as
bentonite.
Figure 1 shows the soil layer distributions of the original ground. The groundwater level
was GL -1.86 m. A rotary drilling machine collected the samples with a hydraulic feed
control system. The samples were covered with a damp cloth and stored in nylon bags
during sample collection to prevent drying. The sample collection point is located in the
construction site of the river improvement project in the Kanto region, where the SB
cutoff wall was constructed in March 2019.

3. Methods

3.1 Physical property tests

Laboratory tests of SB were conducted using core samples for various depths of SB.
Physical properties such as density of soil particles, water content, particle size
distribution, Atterberg limits and wet density were obtained as per JIS A 1202 (2009), JIS
A 1203 (2009), JIS A 1204 (2009), JIS A 1205 (2009) and JIS A 1225 (2009), JIS A 1225
(2009), respectively.

3.2 Hydraulic conductivity tests

In this present work, flexible-wall permeability tests with a falling head system are used
to measure the hydraulic conductivity. The flexible-wall permeameter is more appropriate
SB samples used in the hydraulic conductivity test were cut to  6.0 cm × h 7.0 cm using
a cutter and wire-saw not to damage them. After cutting, specimens were submerged in
distilled water and saturated by vacuum deaeration for around 48 hours. Since the
specimens were undisturbed, no pre-consolidation was performed. Then, the saturated
specimen was set on the flexible-wall permeameter and tested using distilled water as the
permeant. The hydraulic gradient, i, was set to approximately 15. Cell pressure of 40 kPa
was applied as confining pressure since it is known that the vertical stress of the SB cutoff
wall is constant with the depth increasing to a relatively large value if the other parameters
do not change because of the arching effect (EVANS & RUFFING, 2019). The processes
and tests were implemented under constant temperature conditions at 20 °C.

279
 Vent Port

Acrylic cylinder
O-Ring seal Water
Membrane Geotextile and
Filter paper
Air pressure
Influent
Regulator
Specimen

Water
Gauge

Effluent Cell
pressure

Figure 2. A flexible-wall permeameter.

Figure 1. Boring log of the original ground.

3.3 Batch sorption tests

Batch sorption tests were conducted for SB for 3 depths to quantitatively evaluate the
sorption performance of the SBM cutoff wall for contaminants.
The SB samples were oven-dried at 40°C of relatively low temperature for 3 days to
prevent thermal denaturation of bentonite. Then, the dried samples were crushed into
smaller particles by hammer and the fraction under 2.0 mm was used for the tests. The
test solvent was prepared by bringing NaAsO2 powder to a 1000 mg/L solution, and then
diluting it with distilled water in 20, 10, 5, 1, 0.5, and 0.1 mg/L to make the NaAsO2
solutions. For the experiment, 180 mL NaAsO2 solution and 18 g of the above 2 mm sieve
passage sample were put into a 250 mL polyethene container, horizontally shaken at 150
rpm for 24 hours with a shaking width of 4–5 cm. The suspension was transferred to a
centrifuge tube and centrifuged (3000 rpm, 60 min). After that, the suspension was again
transferred to a new centrifuge tube, centrifuged (3000 rpm, 30 min), filtered through a
0.45 μm membrane filter, and then used for analysis. Values of pH and EC were measured

280
by pH/conductivity meter, and concentrations of arsenic (As), Natrium (Na) and Calcium
(Ca) were determined by ICP mass spectrometry. Regarding pH, EC and concentrations
of the elements except for As, they were also measured for solutions of each concentration
before shaking. The above test conditions are summarized in Table 1, and all the processes
were performed at about 20℃ and 1 atm.

Table 1. Preparation and conditions for batch sorption tests.

Operations Amount and Conditions
Soil sample 18 g of SB passing through 2 mm sieve
solution 180 mL of NaAsO2 (0.1, 0.5, 1, 5, 10, 20 mg/L)
Shake 24 h, 150 rpm, shaking width 4–5 cm, horizontal direction
1st centrifugation 60 min, 3000 rpm
2nd centrifugation 30 min, 3000 rpm

4. Results and discussion

4.1 Particle size distribution

The particle size distribution curves of the original ground and SB wall and percent of
gravel, sand, and fine contents in different depths of core samples are shown in Figure 3
and Figure 4, respectively. As shown in Figure 3, the distribution curves of the original
ground and SB wall have a clear difference. The curve of the original ground suggests
that fine contents and gravel contents are concentrated in different layers, while the
distribution curve of the SB wall is much more similar.

Figure 3. Particle size distribution curves of the core samples.

Similarly, Figure 4 shows that gravel contents are not nearly contained below G.L.-21.4
m in the original ground while any depth of SB wall contains gravel contents

281
approximately over 20%. That may be because, during the construction of SB wall, the
chain of the machine brought fines and gravel to each layer. It implies that less variability
of particle size distribution is achieved by using an equal thickness cutting machine.
Figure 5 illustrates the changes in coefficient of variations (CV) of fine, sand and gravel
from the core samples to see the movement of each content of soil and its degrees. Figure
5(a) shows the changes in a 11-m deep SB cutoff wall in TIAN et al. (2019) and (b) shows
the changes in this study. According to Figure 5, the decrease in CV of this study
compared to that of TIAN et al. (2019) is generally more significant despite the
approximately same values before construction, indicating that the SB wall of this study
may be mixed more homogeneously.

(a) Original ground (b) SB wall

Figure 4. Percent of gravel, sand and fine contents in the core samples.

(a) TIAN et al. (2019) (b) This present work

Figure 5. Changes in coefficient of variations of fine, sand, gravel
from the core samples in (a)TIAN et al. (2019) and (b)this present work.

282
According to TIAN et al. (2019), the number of soil types on soil classification contained
in the original ground is smaller relative to this study, and this less complexity of the
original ground might lead to the larger values of CV. This fact suggests that the vertical
homogeneity of the SB wall can be significantly improved by the equal thickness trench
cutting machine regardless of the sites, but it may be affected by the original ground.

4.2 Void ratio

Figure 6 shows the depth distribution of the void ratio. The coefficient of variation
decreased from 0.24 in the original ground to 0.15 in the SB wall. This indicates an
improvement in vertical homogeneity, even more when G.L.-36–37 of the outlier value
is considered. The void ratio near the surface did not change, while the void ratio at depths
below G.L.-8 m increased compared to the original ground. This may be attributed to the
arching effect of the frictional resistance from the adjacent soil, which reduced the vertical
consolidation stress in the SB wall, and the incorporation of adsorbed water into the
interlayer due to the hydrated swelling of bentonite.

Figure 6. Depth distribution of void ratio - (CV: coefficient of variation).

Note: In calculating CV of Original ground, G.L.-36–37 m is excluded.

4.3 Hydraulic conductivity tests

Figure 7 shows temporal changes in hydraulic conductivity of SB wall. The hydraulic
conductivity of the SB at all depths was 1.0 × 10-11 m/s at 2 weeks, but it decreased to 1.0
× 10-12 m/s at 6 weeks. This decreasing trend might be due to the closing of pores by the
confining pressure. At the end of 8 weeks, the hydraulic conductivity was <1.0 × 10-9 m/s,
which is the standard of performance for hydraulic performance at all depths. Therefore,
it can be said that the SB cutoff wall can provide high hydraulic performance regardless

283
of the depth. To determine the mass ratio of bentonite, hydraulic conductivity tests were
conducted several times before construction and hydraulic conductivity gained under the
same mass ratio as this study was 1.29 × 10-10 m/s. In Figure 7 the values of hydraulic
conductivity at all depths were lower than that value. This fact indicates that the SB at all
depth has the adequate hydraulic performance above the minimum expected and the
bentonite was evenly mixed vertically without gathering locally. This decrease of
hydraulic conductivity may also be attributed that pores in SB wall were closed by
consolidation which was not applied to the hydraulic conductivity tests before
construction. Therefore, considering the performance-based criterion, the mass ratio of
bentonite of about 6.5% in the SB cutoff wall of this study is sufficient.

Figure 7. Temporal changes in hydraulic conductivity of SB wall.

4.4 Batch sorption tests

Figure 8 shows the relationship between the initial arsenic (As) concentration and the
concentrations of Natrium (Na) and calcium (Ca). Figure 8 demonstrates that a higher
concentration of both ions is leached from G.L.-2–3 m relative to the other layers, despite
the less variability of particle size distribution achieved by using an equal thickness
cutting machine as explained in 4.1. As shown in Figure 1, on the other hand, the layers
around G.L.-2–3 m, G.L.-20–21 m, and G.L.-36–37 m contain mainly clay, sand and silt,
respectively. So, there is the possibility that most of the clay which was contained before
construction remains at the same layers and is contributed to the high concentration of
ions in Figure 8 due to the high ion exchangeability of clays. According to these facts,
clays in the layers near the ground surface are not transported much. However, Figure 4
shows that the fine and sand contents move respectively, and the gravel contents move
from the layer near the surface, and thus it can be inferred that the movement of sand and
gravel contents contributes to the vertical homogeneity of the SB wall.
In Figure 8, the leaching concentration of Na was high at all depth. This might be due to
the fact that the bentonite was Na-bentonite, which contains a large amount of Na+ in the

284
interlayer cations of montmorillonite, and the test solution was NaAsO2 solution.
Therefore, it cannot be determined from this result whether the bentonite was mixed
evenly in the vertical direction.

Figure 8. Relationships between the initial As concentration and Na, Ca concentration.

Figure 9 illustrates the Freundlich-type sorption isotherms approximated by the values

from SB wall at different depths and the Freundlich-type sorption isotherms obtained
from the batch sorption test of As on decomposed granite soil by MO et al., (2015). It can
be seen from Figure 9 that both G.L.-20–21 m and G.L.-36–37 m have higher sorption
performance than G.L.-2–3 m in the range of the As concentrations in this study. There is
no significant difference between SB in G.L.-2–3 m and decomposed granite soil in the
range of the As concentrations in this study, but in the higher concentrations of As, SB in
G.L.-2–3 has the trend to show a higher sorption performance relative to decomposed
granite soil. As shown in Figure 7, G.L.-2–3 m also shows an adequate hydraulic
performance, so it can be judged that the bentonite reached the surface layer.
Generally comparing the sorption isotherms of SB and decomposed granite soil in Figure
9, it can be read that there is no significant difference between the sorption capacity of
them in the low equilibrium concentration range, but at higher equilibrium concentrations,
the sorption capacity of all depths of SB wall is higher than that of decomposed granite
soil. Therefore, under the condition of As concentration around 0.01 mg/L, which is the
environmental standard, no increase in sorption capacity can be expected by adding
bentonite to the in-situ soil, but under the condition of higher As concentration, which
means the level of anthropogenic pollution, an increase in sorption capacity can be
expected by adding bentonite to the in-situ soil. Therefore, it can be said that the SB cutoff
wall has high applicability as a barrier wall to contain contaminated soil due to
anthropogenic pollution.

285
 Figure 9. Arsenic sorption isotherms between SB and decomposed granite soil.

5. Conclusions
In this study, physical property tests, hydraulic conductivity tests, and batch sorption tests
were conducted on the soil-bentonite (SB) taken from core samples of a 40 m long SB
cutoff wall to investigate the vertical homogeneity, hydraulic and sorption performance,
effects of bentonite addition and mixing effect by an equal thickness cutting machine. The
results support the following conclusions:
a. The vertical homogeneity of physical properties can be improved by an installation of
SB wall using an equal thickness cutting machine regardless of the wall length.
However, the original ground may still have some influence on the particle size
distribution of SB since the degree of change in vertical homogeneity of this study
differed significantly from the previous study.
b. The hydraulic conductivity of SB at all depths is less than 1.0 × 10-11 m/s, which is
less than the standard requirement (1.0 × 10-9 m/s). So, the SB wall has an adequate
vertical homogeneity and hydraulic performance.
c. The mass ratio of bentonite to the SB mixture is about 6.5% and the SB wall has the
lower hydraulic conductivities of 1.0 × 10-11 − 1.0 × 10-13 m/s at the end of 8 weeks
at all depth than the performance standard of 1.0 × 10-9 m/s. Thus, hydraulic
performance of a 40 m-depth SB wall can be ensured with a bentonite mass ratio of
6.5%.
d. When As concentration ~0.01 mg/L, the As sorption performance of SB is comparable
to that of decomposed granite soil, but at 5−10 mg/L, the sorption capacity of SB in
G.L.-20–21 m and G.L.-36–37 m against As is higher than that of decomposed granite
soil, which confirms the high applicability of SB cutoff wall as a hydraulic barrier
wall for in-situ containment.

286
6. Acknowledgements
The authors would like to acknowledge the support of Mr. G. Araki and Mr. N. Ukaji
(Raito Kogyo, Co., Ltd.) for providing the samples. We also would like to extend our
thanks to Dr. L.W. Gathuka for his help to review our paper.

7. References
BRITTON J.P., FILZ G.M., LITTLE J.C. (2005). The effect of variability in hydraulic
conductivity on contaminant transport through soil–bentonite cutoff walls. Journal of
geotechnical and geoenvironmental engineering, Vol. 131(8), pp. 951–957,
https://doi.org/10.1061/(ASCE)1090-0241(2005)131:8(951)
DANIELS J.L., INYANG H.I., CHIEN C.C. (2004). Verification of contaminant sorption
by soil-bentonite barrier materials using scanning electron microscopy/energy dispersive
X-ray spectrometry. Journal of Environmental Engineering, Vol. 130(8), pp. 910–917.
https://doi.org/10.1061/(ASCE)0733-9372(2004)130:8(910)
EVANS J.C., RUFFING, D. (2019). Stresses in soil-bentonite slurry trench cutoff wall.
In: Geo-Congress 2019. ASCE, Philadelphia, US, pp. 177–184,
https://doi.org/10.1061/9780784482148.018
IKEDA K. (2014). EcoClay Wall-Clay Cutoff Wal1 Installation Without Producing Waste
Sludge. Geotechnical Engineering Magazine, Japanese Geotechnical Society, Vol. 62(10),
pp. 24–25.
INUI T., KATSUMI T., TAKAI A. (2016). Cesium sorption/desorption characteristics of
sodium bentonite affected by major cations in leachate from MSW incinerator ash.
Japanese Geotechnical Society Special Publication, Vol. 2(153), pp. 1841–1844,
https://doi.org/10.3208/jgssp.JPN-071
MINISTRY OF THE ENVIRONMENT GOVERNMENT OF JAPAN (2018). 平成 30
年度土壌汚染対策法の施行状況及び土壌汚染調査・対策事例等に関する調査結
果, 水・土壌・地盤・海洋環境の保全 報告書, (in Japanese),
http://www.env.go.jp/water/report/r1-01/index.html
MINJA R.J., EBINA T. (2002). Arsenic sorption capabilities of soil-bentonite mixtures
as buffer materials for landfills. Clay Science, Vol. 12(1), pp. 41–47,
https://doi.org/10.11362/jcssjclayscience1960.12.41
MO J., INUI T., KATSUMI T., KUNINISHI K., HAYASHI S. (2015). Effectiveness of
immobilizing agent used as a sorption layer against natural contamination, Japanese
Geotechnical Society Special Publication, Vol. 1(4), pp.19–24,
https://doi.org/10.3208/jgssp.JPN-22
TAKAI A., INUI T., KATSUMI T. (2016). Evaluating the hydraulic barrier performance
of soil-bentonite cutoff walls using the piezocone penetration test. Soils and Foundations,
Vol. 56 (2), pp. 277–290, https://doi.org/10.1016/j.sandf.2016.02.010
TAKAI A., INUI T., KATSUMI T., KAMON M., ARAKI S. (2013). Hydraulic barrier
performance of soil bentonite mixture cutoff wall. In: MANASSERO M.,
287
DOMINIJANNI A., FOTI S., MUSSO G. (Eds.). Coupled Phenomena in Environmental
Geotechnics. Taylor and Francis, Politecnico di Torino, Italy, pp. 707–714,
https://doi.org/10.1201/b15004-96
TIAN Y., TAKAI A., INUI T., KATSUMI T., WEI G. (2019). Geotechnical properties of
a 15-year-old soil-bentonite mixture cutoff wall. Geo-Environmental Engineering 2019.
Concordia University, Montreal, Canada, May 30–31, 2019.

288
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Pilot-scale heating of soft clays using solar collectors

for the thermal acceleration of consolidation:
experimental and numerical studies
Mohammed Tarek Sayed SAKR 1, Shuhei NISHI 2, Atsushi OGAWA 3,
Atsushi TAKAI 4, Takeshi KATSUMI 4

1. Kyoto University, Graduate School of Engineering, Kyotodaigaku-katsura, Nishikyo-

ku 615-8530, Kyoto, Japan.
[email protected]
2. Kyoto University, Faculty of Engineering, Kyotodaigaku-katsura, Nishikyo-ku 615-
8530, Kyoto, Japan.
3. Okumura Corporation, Matsuzaki-cho, Abeno-ku 545-0053, Osaka, Japan.
[email protected]
4. Kyoto University, Graduate School of Global Environmental Studies, Yoshida-
honmachi, Sakyo-ku 606-8501, Kyoto, Japan.
[email protected]; [email protected]

Abstract:
Various techniques have been utilized to accelerate the consolidation of reclaimed land.
Recently, the thermally accelerated consolidation technique has been proposed to
accelerate consolidation. However, the energy-intensive nature of this technique requires
further research into the modelling and implementation of large-scale, economical ground
heating methods. Therefore, this paper investigates the potential of using solar thermal
collectors to heat clayey ground for accelerating consolidation. A pilot-scale test has been
conducted in which water is heated using a solar thermal collector and circulated through
pipes inserted in a tank of kaolin clay. The test has been conducted almost continuously
for about 18 months to clarify how seasonal variations affect the heating process. Finite
element analysis was conducted using the experimental results to identify appropriate
boundary conditions for modelling the heating process. The heating process could be
modelled with some reasonable accuracy using publicly available weather data (total
daily solar irradiance and hourly averages of wind speed and temperature). Overall, the
test results suggest that supplementary heating sources coupled with solar collectors can
be used for soil heating.

Keywords:
Ground improvement, Solar thermal collector, Thermal consolidation, Ground heating.

Corresponding author mail: [email protected]

1. Introduction
Clays with high water content have insufficient bearing capacities for construction and
experience large, long-term settlements that can damage their overlying structures.
Therefore, such soils must be consolidated to increase their bearing capacities and
minimize their long-term settlements. However, the consolidation of clays is a slow
process due to their low hydraulic conductivity. The lost opportunity for construction due
to long consolidation processes is particularly costly for reclaimed coastal land in dense
metropolitan areas with high land values. Therefore, several techniques have been
developed to accelerate consolidation, such as sand drains, prefabricated vertical drains,
and vacuum consolidation (ROWE, 2001). Nonetheless, consolidation remains a lengthy
process even when such traditional consolidation acceleration techniques are employed.
For example, the consolidation of the reclaimed land for Kansai International Airport
employed over 2 million sand drains and lasted for several years (MESRI & FUNK,
2015). The development of faster consolidation techniques could lead to significant cost
savings for construction over clayey soil, especially reclaimed land.
One promising approach for a new consolidation technique based on the thermal
consolidation effect, whereby raising the soil temperature induces additional irreversible
volume change and accelerates consolidation. Accordingly, a consolidation technique
utilizing this effect may be called thermally accelerated consolidation. The thermal
consolidation effect has been attributed to various mechanisms in the literature, such as
the increase in hydraulic conductivity caused by the lowering of pore-water viscosity with
temperature increase (ABUEL-NAGA et al., 2006), the induction of additional pore-
water pressure due to differential thermal expansion of soil and water particles
(GHAAOWD et al., 2017), and thermal softening of the soil skeleton (LALOUI &
CEKEREVAC, 2003). Several thermo-hydro-mechanical models have been introduced
to combine the mechanisms mentioned above (ABED & SOLOWSKI, 2017), and a field-
scale embankment consolidation test has been conducted, where it was observed that
raising the soil temperature notably accelerated the consolidation process and induced
additional settlement (POTHIRAKSANON et al., 2010).
One key requirement for thermal consolidation is appropriate ground heating methods to
raise the ground temperature sufficiently for achieving the desired consolidation
acceleration. Preferably, such heating methods would use renewable or alternative energy
sources such as solar energy to minimize their cost. The analysis and modelling of large-
scale ground heating have been investigated in the context of underground thermal energy
storage (DIERSCH & BAUER, 2015). The use of solar collectors for the heating of
residential areas was studied by NORDELL & HELLSTRÖM (2000), where they stored
heat energy in bedrock and used grid power to cover peak demand. However, little
research has been done on ground heating in the context of thermally accelerated
consolidation, where different ground composition, time constraints and the necessary
high temperatures require additional research. Therefore, it is necessary to model the
290
ground heating performance of solar collectors to design heating systems that can heat
the ground sufficiently within the necessary time frames to reach the desired
consolidation acceleration. In this work, a pilot-scale soil heating experiment is conducted
to explore the potential of using solar thermal collectors for thermally accelerated
consolidation and to evaluate the necessary boundary conditions for modelling the
heating process.

2. Pilot-scale test
Figures 1 and 2 show a photograph and diagrams of the test setup. A topless steel box
with 4.5-mm thickness and a 1.5-m inner side length was installed into a sandy ground
and internally lined with a 6-cm thick thermal insulation layer of polyethylene (PE) foam.
The resulting container was filled with a 100% water content mixture of kaolin clay and
covered with a 6-cm thick PE foam cover. The kaolin clay had a liquid limit of 78.1%,
plastic limit of 30.7%, specific gravity of 2.59, assumed saturation of 100%, porosity of
0.721, bulk density of 1442.90 kg/m3, and thermal conductivity of 1.13 W/m∙K. To heat
the kaolin, a solar collector (Noritz SJQ-420-BL) was used to heat the water inside the
attached 200-L tank, and a pump then circulated the water through 3 PVC U-pipes
inserted into the kaolin clay and arranged in a triangular layout with a side length of 70
cm. All pipes in this experiment had inner and outer diameters of 1.6 cm and 2.6 cm,
respectively. Table 1 shows the properties of the materials used in this test and the model.
To evaluate the heating of the kaolin, temperature sensors were installed at the locations
and depths shown in Figures 2(b) and 2(c). Temperature sensors were also installed in the
external ground at the same depths of the kaolin sensors 5-cm away from the tank corner.
The water temperature was measured at the main inflow pipe and 3 outflow sub-pipes,
and the flow rate was measured at the 3 inflow sub-pipes using ultrasonic flow sensors.
A relief valve was installed to protect the circulation system from pressure build-up.
Three valves were installed in the 3 inflow sub-pipes to adjust their flow rates to 7-L/min.
The experiment was conducted from August 2019 to March 2021. In this paper, a period
of 9 months, i.e. March 17, 2020, to December 17, 2020, was used for the analysis.

Figure 1. Side view of the test setup.

291
 (a) Diagram of the test setup.

(b) Top view (c) Front view

Figure 2. A diagram of the test setup is shown in (a), while (b) and (c) show the sensor
and U-pipe arrangement inside the kaolin tank.

Table 1. Material properties used in the model.

Thermal Specific heat
Density
Material conductivity capacity Domains
[𝒌𝒈/𝒎𝟑 ]
[W/m∙K] [J/kg∙K]
Kaolin 1.13 2511.47 1442.90 Kaolin
Ground 1.81 1634.27 1951.94 External ground
PE 0.043 2403.7 64 Kaolin tank cover
PVC 0.2 1010 1402.83 Pipes
Combined
0.0462 601.18 648.45 Kaolin tank
steel-PE

3. Numerical simulation
The goal of this simulation is to model the heating of the kaolin using commonly available
weather data and the inflow temperature of the water from the solar collector tank as
boundary conditions. The commercial finite element analysis software COMSOL
Multiphysics was used in this simulation.

292
3.1 Governing equations
Thermal conduction through the pipe walls, kaolin, tank, and external ground was
modelled using the conductive heat transfer equation as follows:
𝜕𝑇
𝜌𝑐𝑝 𝜕𝑡 − 𝛻 ∙ (𝑘𝛻𝑇) = 0 (1)
where: 𝜌 , 𝑐𝑝 , 𝑘 , and 𝑇 are the material’s density, specific heat capacity, thermal
conductivity, and temperature; and 𝑡 is time. Convective heat transfer by the ambient
wind was modelled as external forced convection over a horizontal plate as follows:
𝑞conv = ℎ𝐴(𝑇a − 𝑇) (2)
where: 𝑞conv is the convective heat flux, 𝐴 is the plate area, 𝑇 is the plate surface
temperature, 𝑇𝑎 is the air temperature, and ℎ is the convection coefficient. The following
relationship was used to calculate the convection coefficient based on INCROPERA et
al. (2006):
𝑘 0.3387Pr1/3 Re1/2
2𝐿 1/4 if Re𝐿 ≤ 5 ∙ 105
0.0468 2/3
ℎ= (1+( ) )
Pr
𝑘 1/3 4/5 5 (3)
{2 𝐿 Pr (0.037Re𝐿 − 871) if Re𝐿 > 5 ∙ 10
where: Pr = ν/α is Prandtl’s number; Re𝐿 = 𝑢𝐿/𝜈 is Reynold’s number; 𝑢, 𝜈, 𝛼, and 𝑘
are the wind speed, kinematic visocosity, thermal diffusivity, and thermal conductivity;
and 𝐿 is the length of the plate parallel to the flow direction. 𝛼, 𝜈, and 𝑘 were evaluated
at the average of 𝑇𝑎 and 𝑇. Lastly, the heating of the external ground by solar irradiance
was modelled as a radiative heat flux as follows:
𝑞𝑟𝑎𝑑 = 𝛼𝐼 (4)
where: 𝑞𝑟𝑎𝑑 is the radiative heat flux absorbed by the material, 𝐼 is the incident heat flux,
and 𝛼 is the solar absorb factor of the material.

3.2 Initial conditions

The initial temperature distribution of the kaolin clay and external ground were linearly
interpolated and constantly extrapolated from the experimental results.

3.3 Boundary conditions

The lateral and bottom boundaries of the external ground were set to thermal insulation.
To simplify the analysis, the boundary condition along the inner walls of the pipes was
set as the average of the inflow and outflow temperatures from the experimental data (see
Figure 3). To estimate the heating effect of solar irradiance (𝐼) on the external ground,
solar irradiance was assumed to be proportional to the sine of the solar elevation angle
(𝛼). This assumption ignores the effects of the thickness of the atmospheric layer and
weather conditions such as clouds. Accordingly, the temporal distribution of solar
irradiance may be expressed as:
𝐼(𝑡) ∝ sin[𝛼(𝑡)] (5)
293
where: 𝐼(𝑡) and 𝛼(𝑡) are the incident solar irradiance and solar elevation angle as
functions of time. For a given day, Eq. (5) was scaled by 𝐶 = 1/ ∫ sin[𝛼(𝑡)]𝑑𝑡, such that
proportionality constant for that day becomes its total solar irradiance as follows:
𝐼(𝑡) = 𝐼𝑇 𝐶 sin[𝛼(𝑡)] (6)
Total solar irradiance data from Nara weather station were used for this analysis, and the
solar absorb factor was assumed to be 0.5. The convective heat flux by the ambient wind
was modelled with Eqs. (2) and (3), using the ambient temperature measured on-site and
the wind speed data measured at Ueno weather station. Figure 4 shows the
aforementioned solar irradiance and ambient temperature data.

Figure 3. Boundary temperature of heating pipe 1 used in the simulation.

Figure 4. Ambient temperature measured at experiment site and total daily solar
irradiance measured at Nara city.

4. Results and discussion

The water temperature reached a peak range of about 55-78°C in the summer and 20-
45°C in the winter. The average temperature of sensors located within 35-cm from a

294
heating pipe reached a maximum of 43°C, versus 30°C for sensors located at the corners
of the tank. The kaolin temperature declined relatively quickly after August because the
water temperature during the night became colder than the kaolin adjacent to the pipes.
Therefore, water flow should be regulated to stop circulation when the water is colder
than the surrounding soil to prevent heat loss and maintain higher soil temperatures for
longer durations. To compare the experimental and numerical results, the average
temperature of all kaolin sensors and the average temperature at the sensor locations in
the model are plotted versus total daily rainfall at the Ueno weather station in Figure 5.
While rainfall seems to correlate with the test results, its exclusion from the model did
not significantly affect its accuracy. In other words, rainfall infiltration in the external
ground minimally affected the temperature inside the kaolin tank. The initial gap between
the experimental and numerical results is likely because the linear interpolation
underestimated the initial temperature distribution of the kaolin clay. This gap eventually
evens out until the week between June 17 and June 25. During this period, no
experimental data was recorded due to technical difficulties, so the temperature of the
inner pipe surface was linearly interpolated over that period (see Figure 4).

50 200

Total daily rainfall at Ueno

Rainfall
Experiment
40
temperature [°C]

150
Average sensor

Simulation
30

[mm]
100
20
50
10

0 0
mars-17 mai-17 juil-17 sept-16 nov-16

Figure 5. Average temperature of all temperature sensors inside the kaolin tank versus
average temperature at all sensor locations in the simulation. Rainfall data is included
as an environmental factor not accounted for in the model.

5. Conclusions
Heating the ground by circulating hot water through pipes embedded in the ground can
be modelled using the relatively easy-to-obtain inputs of water temperature, total daily
solar irradiance, ambient wind temperature and speed. These inputs may be obtained as
typical meteorological year data and used as the required boundary conditions for
consolidation design. The results suggest that solar collectors show promise for soil
heating, especially during the summer months. The heating performance may be

295
significantly improved if pipes with more thermally conductive materials and solar
collectors with better specifications are used for heating. Winter heating on the other hand
may require the use of supplemental heating methods depending on the required
consolidation schedule. Future work regarding ground heating for thermally accelerated
consolidation should expand the model to include the solar collector unit, as well as
consider other potential heating methods.

5. References
ABED A.A., SOLOWSKI W.T. (2017). A study on how to couple thermo-hydro-
mechanical behaviour of unsaturated soils: Physical equations, numerical
implementation and examples. Computers and Geotechnics. Vol. 92, pp. 132–155,
https://doi.org/10.1016/j.compgeo.2017.07.021
ABUEL-NAGA H.M., BERGADO D.T., RAMANA G.V., GRINO L., RUJIVIPAT P.,
THET Y. (2006). Experimental evaluation of engineering behavior of soft Bangkok clay
under elevated temperature. Journal of Geotechnical and Geoenvironmental Engineering,
Vol. 132, Issue 7, pp. 902–910, https://doi.org/10.1061/(ASCE)1090-0241(2006)132:7(902)
DIERSCH H.J., BAUER D. (2015). Analysis, modeling and simulation of underground
thermal energy storage (UTES) systems. Advances in Thermal Energy Storage Systems:
Methods and Applications, pp. 149–183, https://doi.org/10.1533/9781782420965.1.149
GHAAOWD I., TAKAI A., KATSUMI T., MCCARTNEY J.S. (2017). Pore water
pressure prediction for undrained heating of soils. Environmental Geotechnics, Vol. 4,
Issue 2, pp. 70–78, https://doi.org/10.1680/jenge.15.00041
INCROPERA F.P., DEWITT D.P., BERGMAN T.L., LAVINE A.S. (2006).
Fundamentals of heat and mass transfer. 6th ed., John Wiley & Sons.
LALOUI L., CEKEREVAC C. (2003). Thermo-plasticity of clays: An isotropic
yield mechanism. Computers and Geotechnics, Vol. 30, Issue 8, pp. 649–660,
https://doi.org/10.1016/j.compgeo.2003.09.001
MESRI G., FUNK J.R. (2015). Settlement of the Kansai International Airport islands.
Journal of Geotechnical and Geoenvironmental Engineering, Vol. 141, Issue 2,
https://doi.org/10.1061/(asce)gt.1943-5606.0001224
NORDELL B., HELLSTRÖM G. (2000). High temperature solar heated seasonal
storage system for low temperature heating of buildings. Solar Energy, Vol. 69, pp. 511–
523, https://doi.org/10.1016/s0038-092x(00)00120-1
ROWE R.K. (2001). Geotechnical and Geoenvironmental Engineering Handbook.
Springer US, https://doi.org/10.1007/978-1-4615-1729-0
POTHIRAKSANON C., BERGADO D.T., ABUEL-NAGA H.M. (2010). Full-scale
embankment consolidation test using prefabricated vertical thermal drains. Soils and
Foundations, Vol. 50, Issue 5, pp. 599–608, https://doi.org/10.3208/sandf.50.599

296
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Towards low carbon binders: binder’s evolution and use of

wood ash as mineral filler

Léo SAOUTI 1,2,3*, Daniel LEVACHER 1, Hafida ZMAMOU 2, Nathalie

LEBLANC 2, Louis JARDIN 3

1. M2C UMR 6143 CNRS, Univ. Caen Normandie, ComUE Normandie Université, 24
rue des Tilleuls 14000 Caen, France.
2. UniLaSalle, Univ. Artois, ULR7519 - Transformations & Agro-ressources, Normandie
Université, 3 rue du Tronquet, 76130 Mont-Saint-Aignan, France.
[email protected]
3. Néolithe, 5 rue des Ateliers, 49290 Chalonnes-sur-Loire, France.

Abstract:
The increasing urbanization of cities implies a high demand for construction materials,
especially aggregates and binders. Aggregates are becoming scarce natural resources,
while binders are a source of greenhouse gas emissions. In the context of sustainable
development and in order to reduce the carbon footprint of the most widely used binder
in the world, i.e. cement, many investigations have led to the development of other binders
that consume less energy and emit less CO2.
Cements have been composed while lowering the clinker content, replacements of cement
in mortars and concretes have been widely studied. Geopolymers or the activation of
mineral fillers are in full development and biopolymers are currently being studied. All
this shows that the use of cementitious binders is moving towards low carbon binders. A
review of these different binders and their evolution is first proposed in this paper.
The activation of mineral fillers is an interesting approach in the development of low
carbon binders. Indeed, the selected mineral fillers are industrial byproducts such as slags,
fly ash from coal, from paper sludge or from wood waste. These byproducts have
apparently a low carbon impact and their recycling would be ensured by these geopolymer
developments. Alkaline activation, which is the most used for this type of by-products,
allows a better reactivity. The same mechanism was observed in the activation of meta-
kaolin that occurs in the geo-polymerization process. These new materials, called
geopolymers, can develop a mechanical strength similar or even superior to those
obtained in conventional Portland cement-based materials. Activation of wood fly ash are
suggested in the development of aggregates.

Keywords:
Cement, Binder, Activation, Geopolymer, Wood ash, Low carbon binder.

Corresponding author mail: [email protected]

Table 1. Nomenclature.
Symbol Definition
w/c or w/b Water cement ratio or water binder ratio
AAC Alkali activated cement
CCS Carbon capture and storage
FA Fly ash
GGBS/GGBFS Ground granulated blast furnace slag
MK Meta-kaolin
OPC Ordinary Portland cement
SF Silica fume
WSA Wastepaper sludge ash

1. Introduction
One of the main causes of global warming is the CO2 emission. To reduce these
emissions, the development of new building materials focuses on the valorisation of co-
product and waste from the construction industry. The concrete composition is composed
by aggregates (gravel, sand, or crushed stone), water and cement. The manufacturing of
the cement is one of the main energy consumers in the concrete production. In 2013, the
production of cement led to 36Gt of global CO2 emission. The clinker production process
is responsible of approximately 58% of the total impact of cement production (SALAS et
al., 2016). To improve the cement production's environmental performance, diverse fields
have been investigated: the kiln improvement in energy efficiency and the use of
alternative fuels. It is possible to reduce the sintering temperature from 1450°C to 1400°C
to save energy without any significant decrease on alite content, which is an important
quality parameter for clinker. Moreover, the clinker substitution is used to reduce clinker
consumption and finally, the carbon capture and storage (CCS, see Table 1 for symbols)
has a high improvement potential. The carbonation is a reaction between CO2, calcium
silicate hydrate and calcium which creates calcite (CaCO3). This phenomenon helps to
capture CO2, even at the end of the concrete's structural life. At this point some calcium
oxide do not carbonate and the demolition material can be reused, as aggregates which
can rise the CO2 capture up to 47% of cement production CO2 emission. Furthermore, the
carbonation depth is better with fly ash concrete than limestone powder concrete. Carbon
capture and storage is about the capture and storage of CO2, especially for clinker
substitution it presents good opportunities to achieve major CO2 emission decrease. Lime
is a binder which can have a carbon capture and storage role. First the kiln technology
has reached an optimal thermodynamic efficiency through improved fuel use and heat
conservation (PROVIS & VAN DEVENTER, 2009). Another way to decrease the carbon
footprint is to reduce firing temperatures. The authors also point out that the local
production and use of binders are a great way to lessen transport’s CO2 emission.

298
2. Supplementary cementitious material
Material substitution in cement is a good way to reduce clinker consumption. Different
materials can be used as partial replacement for Portland cement as, fly ash (FA); Ground
granulated blast furnace slag (GGBS), silica fume (SF), meta-kaolin (MK) or rice husk
ash (RH). As stated, SALAS et al., (2016), regarding fly ash, the replacement of a portion
of clinker may improve concrete durability while reducing energy requirements. Also,
this material emits less CO2 than the GGBS which are used in replacement. Usually used
as filler, supplementary cementitious material, and as healing agent; the Silica fume
incorporation into concrete may increase pozzolanic activity which results in an
improvement in compressive and flexural strengths. The chloride reinforcement against
corrosion due to marine environment can be achieved by the use of SF as SCM. The water
to binder ratio controls the quality of the reinforcement; in the case of low w/c ratio this
reinforcement is moderate. The main disadvantage of using silica fume is that addition
helps to reduce the workability. The use of ground granulated blast furnace slag in alkali
activated slags and in production of geopolymer concrete is under study. As reported by
CHIDIAC & PANESAR, (2008), the optimum water/binder ratio is from 0.31 to 0.38.
The OPC/GGBS optimum ratio is 4:1 and the cement sand ratio is 1:15. The use of fly
ash and GGBS is reported to initiate corrosion. In the other hand, it has been demonstrated
that the achievement of durable structures can be done with fly ash and GGBFS.
Furthermore, the combination of these two materials should be appropriate to make
concrete with good compressive strength. Waste glass (WG) is used as a supplementary
cementitious replacement and in fine aggregate replacement. The positive aspects of the
utilization of WG are the resistance to freezing and thawing, chloride penetration, good
resistance to Na2CO3 and H2SO4, and it increases the compressive strength. The optimum
replacement of cement with glass powder 5-10%. When waste glass is used as a fine
aggregate the optimum value varies between 7.5-25%. Like other supplementary
cementitious the use of high quantity may lead to a decrease in compressive strength. The
use of rice husk ash (RHA) as supplementary cementitious material and sand replacement
has been studied by several research teams. Some relationships have been discovered on
this material, the pozzolanic activity and Blaine specific surface area (SSA) are linked by
a positive relationship. On the other hand, there is an inverse relationship between
pozzolanic activity and the median particle size. The positive effect to use FA as SCM is
that it helps to increase the bulk density, increase in compressive strength, and reduce the
porosity. But it will also increase the drying shrinkage slightly. The curing time,
temperature and the type of FA are crucial parameters that have been reported by LIEW
et al. (2017). The fly ash inclusion in cement product can develop delay in setting time,
increase mechanical strength, it can also decrease early-age strength in the case of
incorporating high volume of fly ash. Furthermore, the use of high content of class C
(ASTM-618, 2019) fly ash in silica fume concrete may increase alkali silica reaction.

299
3. Activation of mineral fillers
The use of Alkali activated cements (AAC) in concrete may decrease de CO2 emission
by 55%-75% in comparison to Portland cement (YANG et al., 2013). In the production
of AACs, the FA and GGBS are the most used industrial by-products. AACs showed
good durability and mechanical properties. The production of GGBS based geopolymer
could be an alternative to cement to reduce CO2 emissions. It has been reported that the
production of one tone of GGBS creates 0.07t CO2 emission. In comparison, one tone of
Portland cement leads to the creation of one tone CO2 (CABEZA et al., 2013). But, to
make these geopolymers, activation is preferred. In order to increase SCM finesse and
their specific surface area, SCM can be crushed to get a smaller fraction. Temperature
activation can be from air, water, or the combination of both. It has been reported (LIEW
et al., 2017), that it is possible to encourage pozzolanic behaviour (of glass particles) with
curing temperature at 50°C. The particle size recommended are smaller than 25µm. Also,
a combination of thermal and mechanical activation allows a better activation. The most
effective activation technique is the chemical activation. The benefit of chemical
activation includes early strength development. One ratio has been studied on fly ash
binders: SiO2/Na2O. To reach the greatest compressive strength SiO2/Na2O molar ratio
should be 0.40 (DE VARGAS et al., 2011). In addition, this ratio correlated to the pH are
parameters that control microstructure, morphology and final compressive strength of Fa
based geopolymer system. Furthermore, the Ca/Si ratio should be between 0.85 and 1, to
obtain good compressive strength. In terms of compressive strength, (LIEW et al., 2017)
have made a preference order of alkali activator with NaOH+WG>NaO>Na2CO3. On
other hand FARIS et al., (2017), created another classification which focuses on the
combination SCM and alkali activators. Here, ability to form combinations between
activators and most of binders are classified as: alkali silicates > alkaline hydroxide >
alkali carbonates > alkali sulphates. This is in agreement with the work of
KOMLJENOVIĆ et al., (2010) who classified the alkali activator as: n
SiO2>NaOH>Na2CO3+NaOH>KOH.
The setting time is an important parameter for the concrete application. This time is
divided into two parts. The initial setting time and the final setting time. The value of
these times depends on the SCM type and the amount used. GGBFS-based geopolymer
exhibits initial setting time from 4:50-7:45h and final setting time from 6:30 to 10:15h.
FA- based geopolymer seems to get faster setting time, with 20min to 4h20min and 1h-
5h15min respectively for initial and final setting time (DE VARGAS et al., 2011). The
alkali contents and sulphate contents seem to be the origin of the setting time. Also, the
investigation reported that good deformability and stability can be reached with FA-based
self-consolidating concrete. ASSAEDI et al., (2015) had worked on composites that are
activated by an alkaline activator for geo-polymerization which is an association between
sodium silicate and sodium hydroxide. The composite is composed by an alkaline
solution to fly ash ratio of 0.75. The alkaline solution is a mix between sodium silicate
300
solution and sodium hydroxide solution with a ratio of 2.5. As reported (ABDALQADER
et al., 2015), most activators used are alkali hydroxides with NaOH and Alkali silicate
with Na2SiO3. These two activators are not sustainable due to the high greenhouse gases
emission necessary for their making. Also, these materials are the most expensive
component in the composition of AACs. ABDALQADER et al., (2015) used sodium
carbonate (Na2CO3) as an activator, which is 2 or 3 times cheaper than two previous cited
activators. This activator may give better results at late ages due to CO3 2- ions (WANG
et al., 1994). ABDALQADER et al., (2015) reported that there are 3 different reaction
stages. The first is the dissolution of the slag and formation of zeolite and gaylussite. The
second is the conversion of gaylussite to CaCO3 and hydrotalcite formation. Finally, the
formation of CASH gel. The dissolution of aluminosilicate precursors leads to the
formation of cementitious material in alkali activated cements. The authors conclusions
are that increasing the activator concentration from 5 to 10% is increasing the mechanical
strength. This behaviour involves that alkali treatment have an important role in the
mechanical strength development. Indeed, the increasing pH leads to an increase in
dissolving reactive species (Al2O3, SiO2). Using Na2CO3 as an activator for GGBS and
fly ash with a w/b ratio of 0.31 can give, with a curing at room temperature, good
compressive strength (<60MPa). Also, the use of fly ash can be done until it reaches 25%,
above this limit the samples showed mechanical strength decrease. ABDALQADER et
al., (2015), stated on the curing out of water is better due to less alkali ions dissolved in
water. SEIFI et al., (2019) worked on dry mortar without cement based on wastepaper
sludge ash (WSA) and GGBS. They conclude that the best water ratio is 0.6% for the
Na2CO3 and Na2O3Si activators. This water to binder ratio gives better unconfined
strength than the one with a ratio of 0.5. They also conclude the type of activator, they
have used two widely used alkaline activators: alkaline silicates (Na2SiO3), alkali
Carbonates (Na2CO3) and last activator used was calcium chloride (CaCl2). The reaction
process is unique for each activator used. The alkaline silico-aluminate gel is the main
product formed with the fly ash activation. Alkali carbonates have shown a fast reaction
leading to a hardening before the compaction process. This activator is not appropriate
for dry mortars which required a long compaction process. The authors also concluded
that this activator is not leading to the development of strength-giving phases and it leads
to a decrease in compression resistance. The quantity used for each activator was also
studied: 2, 3, and 6% for the alkali silicate and 2 and 3% for CaCl2 activator. The optimum
dosage for the alkali silicate is 2% with a 0.5 ratio of water/binder (SEIFI et al., 2019). A
conclusion can be drawn, higher the percentage of activator, higher will be the
compressive strength. The use of CaCl2 activator leads to a basic pH, which causes the
GGBS's dissolution and helps the formation of n CaO.SiO2 phases in favor of
compressive strength. That behavior promotes higher resistance in short terms. The
alkaline activation of fly ash may provoke the release of reactive phases such as Al2O3
and SiO2. ALOMAYRI et al., (2014) worked on alkaline solution prepared with sodium
301
silicate solution to sodium hydroxide ratio fixed at 2.5. The hydroxide and sodium silicate
solution were used with a ratio of 0.35 for fly ash to alkaline solution. RAJAMMA et al.,
(2012) have studied biomass fly ash activation with meta-kaolin, and Na2SiO3:NaOH
ratio of 2:1. Best compressive strength performance (38MPa) is achieved when 60% fly
ash is mixed with 40% Metakaolin.

4. Geopolymers
Aluminosilicate polymers are emerging today as an alternative to cements. These
polymers are also known as geopolymers. Theses geopolymers are known for their ability
to be prepared at low temperature which give them smaller carbon footprint than Portland
cement. ALOMAYRI et al., (2014) reported the procedure to prepare layered geopolymer
composite. To characterize the mechanical properties ASSAEDI et al., (2015), used
flexural tests according to the standard ASTM D790 (2017). The pure geopolymers
composites get an improvement from 4.5MPa to 23MPa in comparison to layered
composites with 4.1 wt.% flax fiber added to the mixture. The compressive strength also
presents an improvement with an increase from 19 to 91MPa, for neat geopolymer and
4.1 wt.% flax fibre’s composite. The neat composites show a compressive strength close
to 20 MPa when 8.3 wt.% cotton fibres composite evolve between values 50 to 90 MPa,
depending on the orientation of the cotton fibers.
The ability of cotton fibers to absorb the stress transmitted by the matrix, may explain the
increase in compressive strength. The highest compressive strength found by (JALAL et
al., 2015) was 92.1MPa for cement silica fume and nano silica concrete addition.
According to (HWANG & HUYNH, 2015), with high Si/Al ratio, alkali activated
geopolymer showed low compressive strength, around 35MPa. Moreover, water
absorption and apparent porosity are related with FA/cement ratio and curing ages. It has
been reported that the increasing FA content will lead to an increasing water content
absorption. LIEW et al., (2017), proposed on a by-product classification in terms of their
characteristics. For compressive strength at elevated temperature: FA>GGBS>SF. The
resistance to sulphate attack: WG>SF>GGBS>FA>RHA. And finally, for chloride
penetration: GGBS>RHA>SF>FA>WG.

5. Conclusion
It is important for construction industry to use green concrete to reduce CO2 emission. To
reach this goal, the use of different supplementary cementitious material as, fly ash (FA)
Ground granulated blast furnace slag (GGBS), silica fume (SF), meta-kaolin (MK) or rice
husk ash (RH) can be done. The use of wood ash as mineral filler can act as a binder in
concrete and does not decrease the mechanical strength as long as the proportion is less
than 25% of cement replacement. The use of fly ash is preferred to ground ash. The fly
ash alkali activation is a way to increase pozzolanic activity, which lead to better late age
compressive strength. SiO2/Na2O and Ca/Si ratios are important in alkali activators, and
302
they should be around 0.4 and between 0.5 to 1. The activator concentration from 5 to
10% is preferred to increase the mechanical strength. Furthermore FARIS et al., (2017),
made a classification about ability to form combinations between activators and most of
binders which are classified as alkali silicates > alkaline hydroxide> alkali carbonates>
alkali sulphates. The use of fly ash as cement supplementary material gives good
compressive strength. Furthermore as showed with wastepaper sludge ash (WSA) and
GGBS, it is possible to make dry mortar without cement (SEIFI et al., 2019). More
research work is needed to establish the capacity of wood fly ash to make this kind of
mortar.

6. References
ABDALQADER A.F., JIN F., AL-TABBAA A., (2015). Characterisation of reactive
magnesia and sodium carbonate-activated fly ash/slag paste blends. Construction and
Building Materials, 93, pp. 506–513. https://doi.org/10.1016/j.conbuildmat.2015.06.015
ALOMAYRI T., SHAIKH F.U.A., LOW I.M. (2014). Effect of fabric orientation on
mechanical properties of cotton fabric reinforced geopolymer composites. Materials &
Design 57, pp. 360–365. https://doi.org/10.1016/j.matdes.2014.01.036
ASSAEDI H., ALOMAYRI T., SHAIKH F.U.A., LOW I. M. (2015). Characterisation
of mechanical and thermal properties in flax fabric reinforced geopolymer composites. J
Advanced Ceramics 4, pp. 272–281, https://doi.org/10.1007/s40145-015-0161-1
ASTM C618. (2019). Standard specification for coal fly ash and raw or calcined natural
pozzolan for use in concrete.
ASTM D790. (2017). Standard test methods for flexural properties of unreinforced and
reinforced plastics and electrical insulating materials.
CABEZA L.F., BARRENECHE C., MIRÓ L., MORERA J.M., BARTOLÍ E., INÉS
FERNÁNDEZ A. (2013). Low carbon and low embodied energy materials in buildings:
A review. Renewable and Sustainable Energy Reviews, Volume 23, pp. 536–542.
https://doi.org/10.1016/j.rser.2013.03.017
CHIDIAC S.E., PANESAR D.K. (2008). Evolution of mechanical properties of concrete
containing ground granulated blast furnace slag and effects on the scaling resistance test
at 28days. Cement and Concrete Composites, Volume 30, Issue 2, pp. 63–71.
https://doi.org/10.1016/j.cemconcomp.2007.09.003
DE VARGAS A.S., DAL MOLIN D.C.C., VILELA A.C.F., DA SILVA F.J., PAVÃO
B., VEIT H. (2011). The effects of Na2O/SiO2 molar ratio, curing temperature and age
on compressive strength, morphology and microstructure of alkali-activated fly ash-
based geopolymers. Cement and Concrete Composites, Volume 33, pp. 653–660.
https://doi.org/10.1016/j.cemconcomp.2011.03.006
FARIS M.A., ABDULLAH M.M.A., SANDU A.V., ISMAIL K.N., MOGA L.M.,
NECULAI O., MUNIANDY R. (2017). Assessment of alkali activated geopolymer

303
binders as an alternative of Portland cement. Materiale Plastice, 54, pp. 145–154.
http://www.revmaterialeplastice.ro
HWANG C.-L., HUYNH T.-P. (2015). Effect of alkali-activator and rice husk ash
content on strength development of fly ash and residual rice husk ash-based geopolymers.
Construction and Building Materials, pp. 101, 1–9. https://doi.org/10.1016/j.conbuildmat.2015.10.025
JALAL M., POULADKHAN A., HARANDI O.F., JAFARI D. (2015). Comparative
study on effects of class F fly ash, nano silica and silica fume on properties of high
performance self-compacting concrete. Construction and Building Materials 94, pp. 90–
104, https://doi.org/10.1016/j.conbuildmat.2015.07.001
KOMLJENOVIĆ M., BAŠČAREVIĆ Z., BRADIĆ V. (2010). Mechanical and
microstructural properties of alkali-activated fly ash geopolymers. Journal of Hazardous
Materials, 181, pp. 35–42, https://doi.org/10.1016/j.jhazmat.2010.04.064
LIEW K.M., SOJOBI A.O., ZHANG L.W., (2017). Green concrete: Prospects and
challenges. Construction and Building Materials 156, pp. 1063–1095.
https://doi.org/10.1016/j.conbuildmat.2017.09.008
PROVIS J.L., VAN DEVENTER J.S.J. (2009). Geopolymers: Structures, processing,
properties and industrial applications. Woodhead Publishing, 464p. ISBN: 9781845694494
RAJAMMA R., LABRINCHA J.A., FERREIRA V.M. (2012). Alkali activation of
biomass fly ash–metakaolin blends. Fuel 98, pp. 265–271. https://doi.org/10.1016/j.fuel.2012.04.006
SALAS D. A., RAMIREZ A. D., RODRIGUEZ C. R., PETROCHE D. M., BOERO A.
J., DUQUE-RIVERA J. (2016). Environmental impacts, life cycle assessment and
potential improvement measures for cement production: A literature review. Journal of
Cleaner Production 113, pp. 114‑122, https://doi.org/10.1016/j.jclepro.2015.11.078
SEIFI S., SEBAIBI N., LEVACHER D., BOUTOUIL M. (2019). Mechanical
performance of a dry mortar without cement, based on paper fly ash and blast furnace
slag. Journal of Building Engineering, 22, pp. 113–121. https://doi.org/10.1016/j.jobe.2018.11.004
WANG S.D., SCRIVENER K.L., PRATT P.L. (1994). Factors affecting the strength of
alkali-activated slag. Cement and Concrete Research 24, pp. 1033–1043.
https://doi.org/10.1016/0008-8846(94)90026-4
YANG K.-H., SONG J.-K., SONG K.-I. (2013). Assessment of CO2 reduction of alkali-
activated concrete. Journal of Cleaner Production 39, pp. 265–272.
https://doi.org/10.1016/j.jclepro.2012.08.001

304
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Oyster shell powder, zeolite and red mud as a binder

stabilizer to remediate heavy metal contaminated soil

Cecilia TORRES QUIROZ 1, Junboum PARK 1

1. Seoul National University, Department of Civil and Environmental Engineering, 1

Gwanak-ro, Gwanak-gu, Seoul 08826, Korea
[email protected]; [email protected]

Abstract:
Conventional methods such as capping or soil washing are difficult to execute in Urban
area due to its space limitations. So, the stabilization-solidification technique which
involves the addition of binders for heavy metals becomes attractive. This caption
phenomenon could occur by sorption capacity (physical property) or cation exchange
capacity (chemical property) of the binder material. This study evaluated Oyster Shell
(OS), Zeolite (Z), and Red Mud (RM) as binders for chemically stabilizing heavy metals
into the soil. Each binder was mixed with a soil sample case using DI water, after
completed mixing and dried the samples, Toxicity Characteristic Leaching Procedure
(TCLP) was performed to determine heavy metal concentrations. For this purpose, soil
samples were obtained from two real cases scenarios from a deposit of metallic waste
named “Soil 1” and a military area called “Soil 2”. Additionally, authors evaluated the
binders on the artificially high Pb-contaminated soil. OS bound approximately 82% of Pb
and 78% of Cu in real cases scenario. While Z was highly effective in stabilizing Pb in
highly polluted artificial soil (>50% of Pb) at lower dosages than OS and RM. RM did
not perform consistently stabilizing toxic metals in soils from contaminated sites, but it
demonstrated a remarkable Pb-immobilization under dosages over than 5% in the
artificial soil. Further, authors observed that OS removal efficiency reached up to 94%
after 10 days. The results suggest that OS is the best low-cost adsorbent material to
stabilize soils contaminated with toxic metals considered in the study.

Keywords:
Low-cost adsorbent, Heavy metals, Soil remediation, Leaching test, TCLP.

Corresponding author mail: [email protected]

1. Introduction
Heavy metals are defined as elements with an atomic number greater than that of Fe, and
as metal elements with densities greater than 5.0 g/cm3, they exert adverse effects on soil
quality (BÜNEMANN et al., 2018; MA et al., 2020) and agricultural productivity
(WANG et al., 2020) and ultimately become a hazard to human and animal health
(SEMENKOV & KOROLEVA, 2020; U.S. EPA, 2001; WATTS, 1998). Therefore,
comprehending the mobilization and solubility of heavy metals through soil is important
in preventing spread pollution of them, these contaminants could reach the groundwater
level being dispersed into deep ecosystems (CHUAN et al., 1996; KIM et al., 2001;
WANG et al., 2019). Among the heavy metal releasing from mine tailing, mainly Pb, Cd,
Cu and Zn are the most common heavy metal founded.
To remediate soil contaminated by heavy metals, there are techniques catalogued as in-
situ and ex-situ as well it can be divided according to the remediation mechanism, which
can be physical, chemical, thermal or biologic (LIU et al., 2018). Soil stabilization–
solidification is an easily applicable and low-cost technique that involves adding an
absorbent material that can preserve long-term stability (BABEL & KURNIAWAN,
2003). Researchers found on it the opportunity to explore new materials which can
interact directly with the contaminant immobilizing it. Through this technique, treatment
mainly can occur by three mechanisms, namely; Chemical fixation of contaminants by
interactions between the hydration products of binders and contaminants, physical
adsorption of the pollutants on the surface of binder hydration products, or physical
encapsulation of contaminants (PARIA & YUET, 2006). Through this technique,
treatment mainly can occur by three mechanisms, namely; Chemical fixation of
contaminants by interactions between the hydration products of binders and
contaminants, physical adsorption of the pollutants on the surface of binder hydration
products, or physical encapsulation of contaminants. In the discussion section, the binding
mechanisms occurring in the remediation process and the main factors that influence the
toxic metal immobilization will be pointed out.

2. Materials and methods

Soil samples were collected from two sites contaminated with toxic metals, their initial
concentrations are shown in Table 1, namely, soil from surroundings of an abandoned
metal mine site classified as silty sand (SM) with a fine content of 22% (denoted as “Soil
1”) and soil from a military service area classified as well-graded sand (SW) with a fine
content lower than 5% (denoted as “Soil 2”). In addition, an artificially contaminated soil
was made, dissolving PbCl2 with DI water and mixed it with soil thoroughly using an
electric mixer (TORRES-QUIROZ et al., 2021).

306
Table 1. Initial toxic metal concentrations.
Initial concentrations
Source USCS 1 pH TCLP
Pb Cu Zn Cd Ni
Soil 1: mg/L 0.639 3.954 102.784 0.316 0.432
SM 8.1
Mine area mg Kg 12.780 79.080 2055.680 6.320 8.640
Soil 2: mg/L 0.079 2.235 10.053 0.046 -
SW 6.7
Military area mg Kg 1.580 44.700 201.060 0.920 -
Soil 3: mg/L 159.802 0.444 - - -
SW 4.9
HCS mg Kg 3196.04 8.880 - - -
1
Unified Soil Classification System—USCS

Regarding to the binders, they were obtained from domestic suppliers in South Korea as
part of the research project, where authors are participating. This project aims to establish
the effectiveness of novel materials, recycling materials or well-known absorbents to treat
real scenarios of contaminated soil by heavy metals. Figure 1 shows the binder physical
and chemical characteristics.

Oyster Shell Powder (OS) Zeolite (Z) Red Mud (RM)

Fe2O3 – 39.47%
CaO – 94.59% SiO2 – 70.75%
Al2O3 – 13.93% Al2O3- 17.47
SiO2 15.18
Figure 1. Binder characteristics, texture and porosity (by SEM analysis) and the main
chemical compounds obtained by XRF analysis.

The experimental procedure consisted of taking a triplicate sample of 50 g of air-dried

soil as a control point and measuring the initial concentration by leaching with DI water
and by TCLP. Then, different dosages of the binders were added to the soil (1, 3, and 5
wt.%), the mixture was prepared at least in triplicate per case as a minimum. Finally, the
307
final concentration after binder addition is measured and compared to evaluate the binder
efficiency. Figure 2 shows the overall experimental procedure.

Figure 2. An overview of the experimental procedure setting of control sample and

treated soil (TORRES-QUIROZ et al., 2021).

3. Results

3.1. Initial heavy metal concentrations

We observed that Soil 1, which is a silty sand soil, had an initial Cu concentration (Table
1) that exceeded the South Korean regulation value for leachate and the World Health
Organization (WHO, 1996), Australian and Canadian guideline values for soil
(>90mg/Kg) (CCME, 2020; LIU et al., 2018). Moreover, the silty sand soil contained Cd
and Zn concentrations beyond the recommended guideline values. Meanwhile, Soil 2,
which is a sandy soil, was slightly contaminated with Zn (201.060 mg/kg) based on the
Korean, Canadian, Australian (<200 mg/kg) (EHTI, 2017; LIU et al., 2018) and WHO
(<50 mg/kg) (WHO, 1996) regulations. HCS, which is also sandy soil, had an initial pH
between 4.75 and 5.05 due to the high content of lead (100 times the minimum
concentration for leaching waste according to the Korean regulation).

3.2 Effects of dosage increment by binder

Soil 1 (silty sand soil) had an initial pH of 8.1 due to the presence of CaO (around 10%).
When it mixed with different dosages of OS, Z and RM, the pH of the soil changed to
7.87–8.18, 7.60–7.81 and 7.89–9.12, respectively. The dosages of OS and Z exerted no
significant effect on the pH of this soil, whereas RM dosage had an increasing effect on
pH (up to 9.12). After adding 5 wt.% of OS, the concentration of Cu in the leachate was
reduced from 3.954 mg/L to 0.937 mg/L (Figure 3), i.e., a reduction of 76%, which is
below the Korean regulation (<3 mg/L) and the limits stipulated by WHO (1996). Zn was
reduced from 102.784 mg/L to 68.657 mg/L, which is also under relevant limits.

308
 Moreover, Cd concentration in the leachate decreased from 0.316 mg/L to 0.182 mg/L,
satisfying the condition for toxic metal presence in wastewater (U.S. EPA, 2011).
9,0 1,5 0,6 150 1,0
7,2 Pb (mg/ L) 1,2 120 0,8
Cu (mg/ L)

0,4

Cd (mg/ L)

Zn (mg/ L)

Ni (mg/ L)
5,4 0,9 90 0,6
3,6 0,6 60 0,4
0,2
1,8 0,3 30 0,2
0,0 0,0 0,0 0 0,0
IC 1% 3% 5% IC 1% 3% 5% IC 1% 3% 5% IC 1%3%5% IC 1%3%5%

(a) Abandoned metal mine site (silty sand soil)

3,6 0,16 0,06 20
Cu (mg/ L)

2,7 0,12

Cd (mg/ L)

Zn (mg/ L)
Pb (mg/ L)

0,04
1,8 0,08 10
0,04 0,02
0,9
0,0 0,00 0,00 0
IC 1% 3% 5% IC 1% 3% 5% IC 1% 3% 5% IC 1% 3% 5%

(b) Military service area (sandy soil)

0,6 375
300
Cu (mg/ L)

Pb (mg/ L)

0,4 225
0,2 150
75
0,0 0
IC 1% 3% 5% 7.5% 10% IC 1% 3% 5% 7.5% 10%
(c) Handmade contaminated soils (HCS)

Figure 3. Toxic metal concentrations in mg/L from the TCLP test results of
contaminated soil after binder addition (TORRES-QUIROZ et al, 2021).

The pH of Soil 2 is 6.7. When mixed with OS, Z and RM, its pH changed to 7.59–8.04,
6.25–6.58 and 7.90–9.48, respectively. OS and RM acted as alkalinizes while Z reduced
the pH value slightly. With a 5 wt.% dosage of OS, Z and RM, Zn immobilization rates
of 50%, 37% and 24% were observed, respectively. Also, shown in Figure 3, the
increment of the dosage from 1% to 5% positively impacts on the toxic metal
immobilization, mainly when OS is used. Even though Cu was below the relevant
guideline values, our findings implied that applying OS, Z and RM immobilized 86%,
42% and 37% of Cu, respectively (selectivity binder preference).
Finally, in Soil 3 (HCS, sand soil), the initial acid pH changed from 4.9 to 6.91–7.51,
4.42–5.18 and 6.23–9.63, after soil treatment with OS, Z and RM, respectively. Like
occurred with the soil 2, OS and RM alkalinize the medium. Meanwhile, the addition of
Z at dosages lower than 5% decreased pH slightly whilst dosages over 5% increased pH
marginally. The leachate with DI water presented an initial Pb concentration of 301.65
mg/L. Conversely, the leaching test with TCLP reported a mean Pb concentration of
159.802 mg/L as the initial concentration of HCS. This value is lower than the
concentrations obtained by leaching with DI water. This finding can be attributed to the

309
ageing effect (short time). The results suggested that Pb concentration can be reduced by
62% by adding 5 wt.% of Z. However, the final concentration (113.825 mg/L) was still
higher than Korean and international regulations (<5 mg/L). Using the same proportion
of RM reduced the concentration to 177.637 (41%), while OS only reached 269.942
(11%). However, when the dosage was increased to 10 wt.%, the immobilization rate via
OS addition significantly improved. By contrast, such an improvement in binder
performance was not observed with Z when its dosage was increased (p>0.05). After
adding 10 wt.% of OS, Z and RM, Pb immobilization rates of 53%, 64% and 59%,
respectively, were observed in Figure 3.

3.3 Effects of contact time

Considering the effect of the contact time in Soil 3, we observed that the high Pb
concentration in HCS stabilized with OS and measured on the leachate was reduced to
38.769 mg/L (87% of Pb was immobilized) after 1 day of contact time, and 94% of Pb
was immobilized after 10 days at dosages above 5%. However, these values were still
above the regulation level. When HCS was treated with OS at 10 wt.%, the Pb
concentration in the leachate collected after 1 day was 3.67 mg/L, which is slightly above
the Korean regulation but under the maximum value of 5 mg/L stipulated by U.S. EPA.
At short contact times (e.g., 12 h), OS with dosages up to 5% immobilized less than 10%
of Pb. After 1 day, OS reached maximum Cu and Pb stabilization because the authors did
not observe any significant improvement in immobilization.

100 100
80 80
C/Co

60 60
C/Co

40 40
20 20
0 0

OS (wt%) OS (wt%)
Pb (12 h) Pb (1 day) Pb (10 days) Cu (12 h) Cu (1 day) Cu (10 days)

Figure 4. Pb and Cu concentrations in the HCS treated with OS obtained through the
TCLP test at different contact times (12 h, 1 day and 10 days).

4. Discussions

4.1 Oyster Shell Powder (OS)

OS powder demonstrated the best performance in binding nearly all the considered toxic
metals, particularly Pb and Cu, for the silty sand and sandy soil samples. Furthermore,
our observations suggested that the preference for sorption of OS was in the following
order: Pb2+ > Cu2+ > Zn2+ > Cd2+ > Ni2+. This order is similar to the observations of SHIN

310
et al. (2014) in their kinetic model. In addition, when HCS was treated with OS, all the
samples had a pH above 7 (alkaline). CaCO3 and CaO in OS were dissolved in water to
produce hydroxyl ions (OH−), increasing the pH of the medium (CaCO3 + H2O → Ca2+
+ CO32−; CO32− + H2O → HCO3 + OH−) (OK et al., 2010). This alkaline condition can
promote the precipitation of metals as metal hydroxides [Mn+ + n(OH) − → M (OH)n,
where M denotes metal] (BONNARD et al., 2020; OK et al., 2010) and can be linked to
the reduction of toxic metals in leachate.
The number of sorption sites and reactive hydroxide ions increased with the OS dosage,
significantly reducing toxic metal concentrations (p < 0.05), as illustrated in Figure 3.
Furthermore, the results of Pb and Cu concentrations in the HCS leachate stabilized with
OS over time. Similar to the preceding results were found by XU et al. (2019) and DESTA
(2013), who observed the roles played by the adsorption characteristics and ion exchange
capacity of binders and the precipitation of ions in achieving stabilization over time.

4.2 Zeolite (Z)

In the current study, Z-treated soil had a neutral pH of approximately 6.5, and Z did not
drastically change the pH of the medium. A slight increase in pH was observed when Z
dosage was increased. A slight increase in pH was observed when Z dosage was
increased. Such pH conditions are beneficial for stabilization by Z because the major
binding mechanisms of Z are adsorption and cation exchange (BONNARD et al., 2020;
OK et al., 2010). This phenomenon can be observed in HCS soil (a slightly acidic soil),
wherein Z achieved the best immobilization compared with the other binders at a contact
time of 12 h. Moreover, given that the immobilization rate does not improve with dosage,
we can argue that ion exchange is the primary binding mechanism of Z in our study. The
surface of Z is negatively charged through the isomorphous replacement of Al+ by Al3+.
This negative charge can be balanced by exchangeable cations, such as metal ions: Pb,
Cd, Zn and Cu (ZHANG et al., 2021). The final concentrations of toxic metals in the
leachates of Soil 1: silty sand and Soil 2: sandy soil at 5 wt.% were in the following order:
Pb < Cu < Zn. This finding corroborates zeolite’s selectivity for cation exchange, i.e.,
Pb2+ > Cu2+ > Zn2+ (BABEL & KURNIAWAN, 2003).

4.3 Red Mud (RM)

RM has an alkaline nature because of the presence of NaOH, a strong base used in
producing alumina (CUI et al., 2019); it can increase pH drastically, as we observed in
this study. RM was rarely more effective for highly contaminated soil than for lower
contamination. RM has a larger surface area than OS and Z (SHIN et al., 2014); thus, this
characteristic may improve the adsorption capacity (primary components: silica, alumina
and Fe2O3) and ion exchange capacity of this material, and it might probably be the reason
for the previous observation. Furthermore, RM can be used to precipitate soluble toxic
metals in their hydroxide form due to its alkaline effect.
311
5. Conclusions
OS demonstrated the best performance in binding Pb and Cu in Soil 1 and Soil 2 (the silty
sand soil and sandy soil); its effect was immediately observable after adding 3% of the
binder to soil samples catalogized as soils with low-to-medium contamination levels.
Moreover, OS proved to be a suitable binder even for soil with a high Pb concentration
(i.e., HCS) under certain conditions. However, higher dosages (>5%) and longer contact
times (>1 day) are required to achieve the desired immobilization rates. Therefore, OS
can stabilize soils contaminated with Pb and Cu. Z is also an excellent alternative binder
for highly contaminated soils because its cation exchange capacity with toxic metals and
sorption properties can bind even under acidic conditions. RM can be used alternatively
to Z. However, caution should be taken during its application because of the risk of
releasing other metalloids (associated with the presence of water-soluble Al, which can
be released into the surrounding environment; in its acutely toxic form, i.e., [Al(OH)4]−).
The final concentration of toxic metals measuring in the leachate can be dropped below
the maximum permissible limits; however, we should keep clear that the toxic metal has
not been removed. Thus, depending on the case, additional soil treatments might be
necessary to eliminate any adverse situation that can lead to spreading the contaminant
caused by the alteration of the medium (stable condition).

6. Acknowledgment
This research was supported by the Korea Ministry of Environment as part of the
‘Subsurface Environmental Management’ SEM project: Grant No. 2019002470002.

7. References
BABEL S., KURNIAWAN T.A. (2003). Low-cost adsorbents for heavy metals uptake
from contaminated water: a review. Journal of Hazardous Materials 97 (1-3), pp. 219-
243, doi.org/10.1016/S0304-3894(02)00263-7
BONNARD M., BOURY B., PARROT I. (2020). Key Insights, Tools, and Future
Prospects on Oyster Shell End-of-Life: A Critical Analysis of Sustainable Solutions.
Environmental Science & Technology 54 (1), pp. 26-38, doi.org/10.1021/acs.est.9b03736
BÜNEMANN E.K., BONGIORNO G., BAI Z., CREAMER R.E., DE DEYN G., DE
GOEDE R., FLESKENS L., GEISSEN V., KUYPER T.W., MÄDER P., PULLEMAN
M., SUKKEL W., VAN GROENIGEN J.W., BRUSSAARD L. (2018). Soil quality – A
critical review. Soil Biology and Biochemistry 120, pp. 105-125.
doi.org/10.1016/j.soilbio.2018.01.030
CCME (2020). Canadian Environmental Quality Guideline. Canadian Council of
Ministers of the Environment.
CHUAN M.C., SHU G.Y., LIU J.C. (1996). Solubility of heavy metals in a contaminated
soil: Effects of redox potential and pH. Water, Air, and Soil Pollution 90, pp. 543-556,
doi.org/10.1007/BF00282668

312
CUI Y., CHEN J., ZHANG Y., PENG D., HUANG T., SUN C. (2019). pH-dependent
leaching characteristics of major and toxic elements from red mud. International Journal
of Environmental Research and Public Health 16 (11), doi: 10.3390/ijerph16112046
DESTA M.B. (2013). Batch Sorption Experiments: Langmuir and Freundlich Isotherm
Studies for the Adsorption of Textile Metal Ions onto Teff Straw (Eragrostis tef)
Agricultural Waste. Journal of Thermodynamics 2013, 375830, doi.org/10.1155/2013/375830
EHTI (2017). No. 2017-54 Waste Process Test Standard, ES 06002.a Standards, labeling
limits and result indication of hazardous substances contained in designated waste -
2014. Environmental Health Technology Institute, South Korea.
KIM K.-K., KIM K.-W., KIM J.-Y., KIM I.S., CHEONG Y.-W., MIN J.-S. (2001).
Characteristics of tailings from the closed metal mines as potential contamination source
in South Korea. Environmental Geology 41, pp. 358-364, doi.org/10.1007/s002540100396
LIU L., LI W., SONG W., GUO M. (2018). Remediation techniques for heavy metal-
contaminated soils: Principles and applicability. Science of The Total Environment 633,
pp. 206-219, doi: 10.1016/j.scitotenv.2018.03.161
MA J., CHEN Y., ANTONIADIS V., WANG K., HUANG Y., TIAN H. (2020)
Assessment of heavy metal(loid)s contamination risk and grain nutritional quality in
organic waste-amended soil. Journal of Hazardous Materials 399, 123095. doi:
10.1016/j.jhazmat.2020.123095
OK Y.S., OH S.-E., AHMAD M., HYUN S., KIM K.-R., MOON D.H., LEE S.S., LIM
K.J., JEON W.-T., YANG J.E. (2010). Effects of natural and calcined oyster shells on
Cd and Pb immobilization in contaminated soils. Environmental Earth Sciences 61, pp.
1301-1308, doi.org/10.1007/s12665-010-0674-4
PARIA S., YUET P.K. (2006). Solidification–stabilization of organic and inorganic
contaminants using Portland cement: a literature review. Environmental Reviews 14, pp.
217-255, doi.org/10.1139/a06-004
SEMENKOV I., KOROLEVA T. (2020). Heavy metals content in soils of Western
Siberia in relation to international soil quality standards. Geoderma Regional 21,
doi.org/10.1016/j.geodrs.2020.e00283
SHIN, W.-S., KANG K., KIM Y.-K. (2014). Adsorption characteristics of multi-metal
ions by red mud, zeolite, limestone, and oyster shell. Environmental Engineering
Research 19, pp. 15-22,.doi.org/10.4491/eer.2014.19.1.015
TORRES-QUIROZ C., DISSANAYAKE J., PARK J. (2021). Oyster Shell Powder,
Zeolite and Red Mud as Binders for Immobilising Toxic Metals in Fine Granular
Contaminated Soils (from Industrial Zones in South Korea). International Journal of
Environmental Research and Public Health, 18 (5), doi.org/10.3390/ijerph18052530
U.S. EPA (2001). General principles for performing aggregate exposure and risk
assessments. U.S. Environmental Protection Agency, United States.
U.S. EPA (2011). 40 CFR 261.24 - Toxicity Characteristic, in: Regulations, C.o.F. (Ed.).
U.S. Environmental Protection Agency, United States.
313
WANG F., GUAN Q., TIAN J., LIN J., YANG Y., YANG L., PAN N. (2020).
Contamination characteristics, source apportionment, and health risk assessment of
heavy metals in agricultural soil in the Hexi Corridor. CATENA 191, 104573.
doi.org/10.1016/j.catena.2020.104573
WANG P., SUN Z., HU Y., CHENG H. (2019). Leaching of heavy metals from
abandoned mine tailings brought by precipitation and the associated environmental
impact. Science of The Total Environment 695, doi: 10.1016/j.scitotenv.2019.133893
WATTS R.J. (1998). Hazardous Wastes: Sources, Pathways, Receptors. Wiley.
WHO (1996). Permessible limits for heavy metals in soils and plants. World Health
Organization (WHO),, Geneva, Switzerland.
Xu, X., Liu, X., Oh, M., Park, J. (2019) Oyster Shell as a Low-Cost Adsorbent for
Removing Heavy Metal Ions from Wastewater. Polish Journal of Environmental Studies
28, 2949-2959.
ZHANG L., ZHANG L., DONG X., ZHANG S., ZHAO Y., CEN Q., ZHANG K. (2021).
The effect and mechanism of Si/Al ratio on microstructure of zeolite modified ceramsite
derived from industrial wastes. Microporous and Mesoporous Materials 311,
doi.org/10.1016/j.micromeso.2020.110667

314
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Studies on the use of polymer blended MSW as a structural

fill material in embankments

B.V.S. VISWANADHAM 1, Ankit KUMAR 1

1. Department of Civil Engineering, Indian Institute of Technology Bombay, Powai,

Mumbai-400076, India.
[email protected]

Abstract:
Nowadays, availability of natural materials is in scarce. In view of the above, it is required
to evolve alternate structural fill materials for meeting ever demanding requirement of
construction of embankments. The objective of the present study is to explore the use of
aged MSW as a structural fill material in embankments. In the present study, MSW of 6
years old was collected from Mulund dumping yard of Mumbai suburbs, India. Some
initial tests were done at the site itself for characterizing its unit weight and permeability.
As the MSW collected from the site has large size multi-components, the sun-dried MSW
was shredded to achieve a size less than 4.75 mm for conducting experiments in the
laboratory. Laboratory study of MSW includes basic geotechnical characterisation and
effect of polymers on geotechnical characterisation. Further, in order to enhance shear
strength and compressibility characteristics of MSW, the use of Bondex-J400 polymer
with dosages ranging from 1% to 5% and with curing periods of 1 day, 3 days, 7 days and
14 days was investigated by performing unconfined compressive strength tests and
consolidation tests in the laboratory. With the use of polymer dosage from 1% to 5%, the
maximum dry unit weight decreased from 11.90 kN/m3 to 10.73 kN/m3 and optimum
moisture content increased from 28.1% to 37.2% respectively (Standard Proctor
compaction). For a polymer dosage of 2 % (by dry weight of MSW) and 14 days of curing,
the unconfined compressive strength was found to be 1.8 times more than MSW without
any polymer content. The primary compression index of MSW blended with 2% polymer
and 14 days of curing was found to decrease from 0.189 to 0.090. This indicates that the
MSW can be used as a structural fill material in embankments by blending with an
appropriate dosage and type of polymer for ensuring adequate shear strength and reduced
compressibility. Finally, MSW utilization with all safety regulations results in reducing
landfill space requirements and providing much needed alternative structural fill material
for the construction of embankments.

Keywords:
Embankments, MSW, Polymer, Structural fill, Laboratory tests.

Corresponding author mail: [email protected]

1. Introduction
In the urban areas an increase in the waste generation rate was recorded due to the
improvement in the standard of living and due to the increased urbanisation. This results
in scarcity of land and excessive burden on the landfill (SHARHOLY et al., 2007,
KAUSHAL et al., 2012, CPCB, 2016). To overcome this issue, either MSW needs to be
compacted to it maximum capacity, with a suitable design, so that the threat of failure of
landfill can be avoided, and this requires proper knowledge of geotechnical
characterisation of MSW. As an alternative, aged MSW buried in the landfill needs to be
utilized so that more space can be allowed in the old landfills for the upcoming MSW. To
overcome the issue related to the availability of space in old landfills for future MSW and
also non-availability of natural materials for embankment construction, the present study
aims to probe the potential of utilization of aged MSW as a filling material in geotechnical
projects.

Formation width Formation width

Shoulder Pavement Shoulder Pavement

Wheel load

Wheel load Soil cover of nH

Side slope Intermediate separation 0.5-1.5 m thickness 1V
Soil cover of nH barrier layer
0.5-1.5 m thickness 1V
Anticipated Aged MSW placed Rutting
Anticipated deformation
deformation at MDD and OMC h>4 m 0.4 m
Rutting Side
h<4 m drainage
Side drainage MSW placed at h=2 m
Aged MSW placed at MDD and OMC
MDU and OMC
Ground level Ground level

Figure 1. Embankment (height < 4 m) Figure 2. Embankment (height > 4 m)

constructed using aged waste as a filling constructed using aged waste as a filling
material. material.

This will also minimise the scarcity of filling material. But using aged MSW as a
structural fill may lead to excessive settlement due to primary and secondary
compression. Expected deformation in embankments constructed using aged MSW are
shown in Figure 1 and Figure 2. Total settlement due to primary compression could be
up to 20-30% of initial thickness of MSW layer (LANDVA & CLARK, 1990). Therefore,
before constructing the embankment it is necessary to properly analyse the settlement due
to primary and secondary compression, and to adopt some technique to minimise the
long-term settlement under traffic loading.

2. Methodology
From the geotechnical point of view, MSW can be characterized based on its age i.e.
fresh/young, average age and old age/aged MSW. Aged MSW is the waste which was
dumped some years ago and most of the organic matter has been stabilized by biological
degradation. This MSW has low compressibility and low concentration of hazardous
316
materials, as compared to fresh and average age waste, which make its more suitable to
be used as a structural fill as compared to fresh and average aged waste. Aged MSW from
Mulund dump yard was collected from one of the heaps, which was receiving waste since
2012, for the study so that the waste could find its utilization with or without minimum
modifications. This excavated waste is then filled in polythene bags and transported to
the geotechnical engineering laboratory of IIT Bombay for further processing. The
collected MSW was then sundried for 7 days so that all the moisture evaporates and the
shredding of waste could be done followed by segregation. MSW is segregated by hand
picking. Following are the major components found in the waste during segregation with
proportion of each component in the MSW by its dry weight is presented in Table 1.

Table 1. Components of MSW collected from Mulund dump yard with its fraction.
Sr. No. Components Fraction by weight (%)
1 Soil like and other materials 65.60
2 Ceramics and stones 13.56
3 Cloth 3.57
4 Metal 0.90
5 Glass 0.90
6 Coconut shell and fibres 0.88
7 Thermocoal 0.15
8 Polythene 9.47
9 Paper 0.78
10 Wood 3.91
11 Plastic 0.70

After segregation, waste is then cut down into small pieces with the help of a butcher
knife, jaw crusher and a plastic cutter. This shredded waste is then mixed thoroughly for
90 minutes by hand to get a uniform composition of the waste. MSW obtained after
mixing is shown in Figure 3 and is legended as MSW-A.

Figure 3. Shredded MSW-A, scale in cm).

317
All the tests (Proctor compaction, UCS (unconfined compression strength) test and
Consolidation tests were conducted without polymer and with polymer doses of 1,2,3 and
5% on the dry mass of MSW. For 1% of polymer 100 g of dry MSW, 1 g of polymer was
used. In the present study, Bondex-J400 polymer (commercially available in Mumbai)
was used. Sample was cured for 1,3,7 and 14 days. Sample was first tested for standard
Proctor test and then UCS and consolidation tests were conducted at respective OMC
(Optimum moisture content) and MDU (Maximum dry unit weight) for different
percentages of polymers.

2.1 Methodology for standard Proctor test

Standard Proctor test was used for the determination of compaction characteristics of the
MSW. Around 2.5 kg of MSW was taken initially and 25g of polymer was mixed in the
MSW for making 1% of MSW polymer mixture. First a solution of water and polymer
was made and then this solution was mixed to the waste and transferred to desiccator.
This MSW is then left for 24 hrs, so that the reaction between particles of MSW and
polymer can take place. After 24 hrs of curing, compaction test was conducted. This
mixture of MSW and water is then compacted in the compaction mould with different
percentage of moisture by adding water and graphs are plotted for 1, 2, 3, and 5% polymer
dosages to find its maximum dry unit weight and corresponding moisture content.

2.2 Methodology for UCS tests

UCS sample of diameter 75 mm and height 150 mm was prepared with the help of a split
mould, as shown in Figure 4. The sample was then compacted in 4 layers to achieve
uniform density throughout the height. Each layer is scratched by spatula before placing
the next layer so that the layers could meet proper interlocking. Water and polymer are
mixed properly, and the water-polymer solution is then mixed with MSW and bulk weight
of this MSW was recorded. After that, the sample was taken out from the mould and
placed in the desiccator, packed with a polythene sheet so no loss of moisture and the
reaction between the polymer and the particles of MSW-A, can take place with time.
Compacted UCS samples were taken out after 1, 3, 7, and 14 days of curing for testing
respectively.

2.3 Methodology for Consolidation tests

For consolidation tests, sample was prepared in the same way as prepared in for the UCS
test except that the sample of MSW was not extracted out of the oedometer ring of 75
mm diameter and 25 mm thickness, as shown in Figure 5. And sample was placed in the
desiccator with ring for curing. After 1, 7, and 14 days of curing, samples were taken out
for testing.

318
 Top plate cm

Collar

Split
mould

30 cm
Clamp

Base plate
hammer

(a) Mould after closing (b) Mould after opening

Figure 4. Mould used for making 15 cm x 7.5 cm UCS samples.

cm

Figure 5. View of compacted MSW-A sample for consolidation test.

3. Results and discussion

MSW was tested with and without polymer for its compaction, unconfined compressive
strength and compression characteristics.

319
3.1 Compaction tests
Evaluation of compaction characteristics is of prime importance for geotechnical
properties of MSW. As the land available for the purpose of landfilling is limited and the
waste generation keeps on increasing, it is required to compact MSW to its maximum
possible density so that maximum utilization of land could be achieved. Standard Proctor
test is conducted on aged MSW without polymer and maximum dry unit weight obtained
was 11.90 kN/m3 with optimum moisture content of 28.12%. Establishment of zero air
void line shows that at an OMC value of 28.12%, maximum dry unit that can be achieved
with zero air voids is 12.85 kN/m3, as shown in Figure 6. GABR & VALERO (1995)
reported compaction characteristics of 15-30 years old waste and reported an MDU value
of 9.3 kN/m3 and the OMC value as 31%.

Figure 6. Variation of maximum dry unit weight with moisture content for MSW-A
(Standard Proctor Compaction).

Summary of standard Proctor compaction tests on MSW-A sample with 1, 2, 3 and 5%

dosages of polymer is given in Table 2. Figure 7 shows the comparison of compaction
characteristics of MSW with 1, 2, 3 and 5% of polymer. It could be seen from Figure 7
that for the same amount of compactive effort, with an increase in the percentage of
polymer in MSW, maximum dry unit weight starts to reduce and the optimum moisture
content increases. Lubrication effect of fluids depends on its viscosity. As viscosity of
fluid increases lubrication effect decreases. Water has viscosity of 1 centi-stoke and J-
400 polymer has viscosity of 1932.36 centi-stoke at 20C. As the percentage of polymer
increases viscosity of fluid increases, which leads to decrease in lubrication effect. Water
works as a lubricant between the particles of MSW. During the compaction test, addition

320
of polymer would change the lubrication effect of it and this would then change the value
of maximum dry unit weight and optimum moisture content.

Table 2. Comparison of MDU and OMC for different percentage of polymers.

Sr. No. Polymer (%) MDU (kN/m3) OMC (%)
1 0 11.90 28.12
2 1 11.54 31.29
3 2 11.14 34.05
4 3 10.87 35.89
5 5 10.73 37.21

Therefore, OMC required to pack the MSW particles in the best configuration increases.
This decrease in the MDU with increase of polymer will affect the UCS and compression
of the MSW. As fraction of polymer in the MSW increases, its compactness reduces,
therefore without any curing UCS value will reduce and compressibility will increase.

Figure 7. Variation of maximum dry unit weight with moisture content for MSW-A with
(a) 1% Polymer, (b) 2% Polymer, (c) 3% Polymer and (d) 5% Polymer.

3.2 Unconfined Compressive strength tests

Unconfined compressive strength tests without polymer was used to determine cohesion
of the sample in undrained condition. Test was conducted on the MSW-A sample
compacted at its MDU and OMC without addition of polymer at a constant strain rate of
1.2 mm/min. Failure of sample occurred at a strain value of 18% with bulging. Figures
8a-8c show the status of the UCS MSW-A sample without any polymer.

321
 (a) Before failure (b) At failure (c) After large strain
Figure 8. Deformation of UCS sample at displacement rate 1.2 mm/min (a) before
failure, (b) at failure and (c) after collapse (MSW).

Figures 9a-9d show the stress vs strain of UCS samples with polymer dosages ranging
from 1%, 2%, 3% and 5% and curing periods of 1, 3, 7 and 14 days of curing respectively.
Figure 9a shows the change in UCS value for 1% of polymer with curing time. This
increased shear strength of MSW-A is attributed to the reaction of polymer with different
components of MSW. With an increase in curing period, increase in unconfined
compressive strength of MSW-A can be noted. This increase is found to be significant
for MSW-A sample with polymer dosage of 2% by dry weight. This is attributed to the
binding action induced by the polymer. As can be noted from Figure 9b, for MSW-A
sample blended with 2% polymer content (i.e. for 14 days of curing), a significant
increase in stiffness can also be noted.

322
 Figure 9. Axial stress vs axial strain of MSW with 1, 2, 3, and 5% polymer.

Figure 10 gives the variation of stress vs strain obtained from UCS tests on 14 day cured
MSW-A samples blended with 1%, 2%, 3% and 5% polymer dosages. As can be noted,
with an increase in polymer dosage beyond 3%, there is a decrease in UCS. This could
be attributed to the reduction in MDU and increased OMC (Figure 7).

Figure 10. Axial stress vs axial strain for 0, 1, 2, 3 and 5% polymer after 14 days for
MSW-A.
323
3.3 Compression characteristics
Settlement of MSW takes place due to immediate settlement, primary compression,
secondary compression resulting due to biological degradation. After the end of primary
compression, additional settlement starts to occur due to readjustment of particles of
MSW. This settlement which occurs at a constant stress level with an increase in time is
known as creep or secondary compression. In the case of MSW settlement, creep comes
along with biological decomposition and continues to take place until entire MSW has
stabilized. In the aged MSW biological decomposition is less as most of biodegradable
components might have degraded already.
Compression characteristics were determined in this study by performing 1-D
consolidation tests using 75 mm diameter and 25 mm thickness oedometer cell Sample
placed in the desiccator for curing with oedometer rings were taken out after 1, 7 and 14
days respectively, and results were compared for with and without polymer doses. The
main parameters obtained from this study were primary compression index, primary
compression ratio, coefficient of consolidation, coefficient of volume change, and
coefficient of permeability. In Figure 11, time vs settlement (%) variation was plotted for
MSW-A samples with 1%, 2%, 3% and 5% polymer (after curing of 14 days).

Figure 11. Settlement (%) vs time for MSW-A with 1, 2, 3 and 5% polymer.

324
Coefficient of compression ratio Cce was calculated as the ratio of percentage settlement
to the change in the log of effective stresses at base 10. Cce is defined in equation 1 given
below.
ΔH Δe (1)
Cce = =
H×log(σ1 /σ 2 ) (1+eo )×log(σ1 /σ 2 )

where Δh is change in thickness and H, is initial height, Δe is change in void ratio, eo is

initial void ratio, σ1 is final effective stress and σ2 is initial effective stress.
Value of Cce can also be calculated from the slope of the curve of Figure12. From
Figure12, maximum settlement after the end of primary consolidation at the end of 400
kPa was found to be 22.21% without polymer which reduces by 15.02% for 2% of
polymer after 14 days of curing. Table 3 presents comparison of compression ratio
obtained in the presented study with the published literature.

Figure 12. Compression stress vs percent settlement for MSW-A with and without
polymer dosages.

Figure 13. Coefficient of permeability vs effective stress for1, 2, 3, and 5% polymer.

325
Table 3. Comparison of compression index with others.
Compression index, Age
References Remarks
Cce (years)
Waste was synthesised in the
Thusyanthan et al. (2006) 0.25 -
laboratory
0.17 3 -
Landva et al. (2000)
0.22 6 -

0.18 14.6 m depth of landfill

-
0.18 19.2 m depth of landfill
Gabr and Valero (1995) 15-30
0.22 11.6 m depth of landfill

0.35 Degree of degradation = 0%

0.26 --a Degree of degradation = 50%

Reddy et al. (2011)
0.15 Degree of degradation = 82%
Vilar and Carvalhod (2004) 0.22 15 Degradable organic 14%
Present study MSW without
0.11 6 Waste collected at 3 m depth
polymer
--a Not reported.

The value of compression index of MSW-A sample without any polymer is obtained as
0.189 and it is found to decrease to 0.09 for MSW-A blended with 2% polymer.
Further, results of consolidation tests were interpreted by calculating permeability of
MSW-A blended with various dosages of polymer (tested after 14 days of curing). With
an increase in effective stress, decrease in permeability could be obtained. But distinct
variation was not registered. However, as the MSW-A is relative aged, it is found to have
a permeability of 1 x10-7 m/s at an effective stress of 10 kPa, see Figure 13.

3.4 Effect of curing on shear strength and compressibility

Curing to the process of allowing materials to react under favourable conditions. Curing
is widely used for hydration of concrete and lime. Adhesion due to addition of some
blended material takes time to react. In this case of Styrene Acrylic Co-polymer was used
to react with MSW-A and samples were placed for 1, 3, 7 and 14 days for reaction to take
place between MSW and polymer. As percentage of polymer increases thickness of
polymer layer over the MSW particles also increases. This polymer layer takes time to
react with MSW and complete picture of rate of completion of reaction between polymer
and MSW can be seen through the increase in the UCS value for 1 and 2% of polymer
with time of curing.

326
 In Figure14 variation of UCS value with time of curing is shown. In Figure-14, U1 is the
UCS strength without curing and as the time of reaction between polymer and MSW
increases UCS value increases to U2 with a slow rate and this phase and be called as Initial
or slow curing phase up to time t1. After end of initial curing phase, a phase with
accelerated curing occurs and UCS value increases rapidly in short time and at the end
UCS value reached to value U3. After end of accelerated phase of curing a deaccelerated
phase of curing started at time td and further increase in UCS occurs at very slow rate up
to maximum UCS value U4. The time length of initial curing phase, accelerated curing
phase and deaccelerated curing phase depends on the amount of polymer used. As the
percentage of polymer increase thickness of polymer layer over the MSW particle also
increases and as thickness of polymer layer increases time required to different phases of
curing also increases.

Figure 14. Variation of UCS value with days of curing for MSW-A.

Variation of UCS values and compression index with days of curing and various dosages
of polymer is shown in Figures 14 and 15 respectively. For 1% of polymer, UCS has
increased by 8%, 25% and 28% for 3, 7 and 14 days, when compared with UCS value
obtained on MSW-A sample tested after one day of curing. In Figure 15, a decrease in
trend of compression index with curing period was registered for MSW-A blended with
2% polymer. This could be due to increase in stiffness of the sample and which inhibits
compression under stress.

327
 Figure 15. Variation of Compression index with days of curing for MSW-A.

4. Conclusions
Based on the analysis and interpretation of test results carried-out on MSW sample with
and without polymer dosages, the following conclusions can be drawn:
-The MSW-A sample blended with 2% polymer was registered to have increase in
unconfined compressive strength and reduction in primary compression index. This
implies that the MSW like the one tested in the present study can be used a structural
fill material, provided it is blended uniformly with a polymer dosage of 2%.
- With an increase in curing period, an increase in unconfined strength was registered
and along with a decrease in compression index.
- Blending of polymer with MSW-A sample has not shown much variation on the
permeability.
The results of preliminary study on MSW-A sample blended with polymer has shown
encouraging results and showing a way for evolving a scheme for effective utilization of
aged MSW samples. Further detailed laboratory testing is warranted to quantify further
in this regard.

Acknowledgements
Thanks are due to solid waste department of Municipal Corporation of Greater Mumbai
(MCGM) for providing the MSW used in the present study. Authors would like to thank
M/s Jesons Industries Ltd., Mumbai, India for providing polymer used in the present
study. Also, authors would like to thank staff at geotechnical engineering laboratory and
centrifuge laboratory of IIT Bombay for their untiring support throughout the study.

5. References
CPCB - CENTRAL POLLUTION CONTROL BOARD. (2016). Solid waste generation
in 46 metrocities, Central Pollution Control Board, Ministry of Environmental, Forest and
Climate Change, Govt. of India.
328
GABR M. A., VALERO N. S. (1995). Geotechnical properties of municipal solid waste,
Geotechnical Testing Journal, ASTM, 18(2), pp. 241-251, doi.org/10.1520/GTJ10324J
KAUSHAL R. K., VARGHESE G. K., CHABUKDHARA M. (2012). Municipal solid
waste management in India- current state and future challenges : A review municipal
solid waste management in india-current state and future challenges : A Review,
International Journal of Engineering Science and Technology, 4(4), pp. 1473-1489.
LANDVA A.O., CLARK J.I. (1990). Geotechnics of waste fill. In Geotechnics of waste
fills -theory and practice. Edited by A. Landva and G.D. Knowles. ASTM, Special
Technical Publication STP 1070, pp. 86–103, doi.org/10.1520/STP25301S
LANDVA A. O., VALSANGKAR A. J., PELKEY S. G. (2000). Lateral earth pressure
at rest and compressibility, Canadian Geotechnical Journal., 37(6), pp. 1157-1165,
doi.org/10.1139/t00-057
REDDY K. R., HETTIARACHCHI H., GANGATHULASI J., BOGNER J. E. (2011).
Geotechnical properties of municipal solid waste at different phases of biodegradation,
Waste Management, 31(11), pp. 2275-2286, doi: 10.1016/j.wasman.2011.06.002
SHARHOLY M., AHMAD K., MAHMOOD G., TRIVEDI R. C. (2007). Municipal solid
waste management in Indian cities - A review, Waste Management, 28(2), pp. 459-467.
THUSYANTHAN I., MADABHUSHI S. P. G., SINGH S. (2006). Centrifuge modeling
of solid waste landfill systems-part 1: development of a model municipal solid waste,
Geotechnical Testing Journal, ASTM, 29(3), pp. 217-222, doi.org/10.1520/GTJ14314
VILAR O. M., CARVALHO M. de F. (2004). Mechanical properties of municipal solid
waste, Journal of Testing and Evaluation, ASTM, 32(6), pp. 438-449, doi: 10.1520/jte11945

329
330
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Studies on verification of dynamic compaction induced

densification of MSW landfills using shear wave velocity
profiling
B.V.S. VISWANADHAM 1, Saptarshi KUNDU 2

1. Department of Civil Engineering, Indian Institute of Technology Bombay, Powai,

Mumbai - 400076, India.
[email protected]
2. Department of Civil Engineering, Indian Institute of Engineering Science and
Technology, Shibpur, Howrah 711103, India.
[email protected]

Abstract:
Recently rapid industrialization has resulted in a massive generation of MSW in various
cities across the world. The current global scenario reveals that the existing landfills are
over-flowing with waste, and acquiring new sites within a city or on the outskirts is
becoming increasingly difficult owing to space constraints and high transportation costs
involved in individual cases. One of the viable option of compressing the existing waste
volume is through dynamic compaction (DC) based ground improvement technology.
Hence, the objective of this study is to examine the densification of municipal solid
waste (MSW) landfill sites under the application of dynamic compaction (DC). The
study is especially relevant to the MSW landfills located in congested urban cities
which are overflowing with wastes, and are on the brink of their design life. Finite
element (FE) based analysis is carried out using commercially available software
ABAQUS (Explicit module). The response of MSW is modelled using Drucker-Prager
constitutive law with a hardening cap. The increase in stiffness and densification of
MSW under the impact of DC is studied using shear wave velocity profiling.
Subsequent parametric studies are carried out for two different compression ratios of
MSW and for different tamper energies to observe the impact of high-energy and low-
energy DC processes on young and moderately old landfill sites. The results indicated
that the degree of improvement monitored in terms of Vs increased substantially by
about 3.4-5.5 times, depending on the waste compressibility and energy level. Further,
the depth of improvement achieved varied on an average from 11 m - 14.4 m. The
utility of this study is to aid in decision making regarding the implementation of DC to
MSW landfills and to focus the efforts of full-scale field trials.

Keywords: Dynamic compaction, Municipal solid waste, Landfills, FE modeling.

Corresponding author mail: [email protected]

1. Introduction
Over the past few decades, Dynamic compaction (DC) has gained popularity as an
effective improvement technique for geomaterials in view of its simplicity and cost
effectiveness. In addition to being economically viable, the carbon footprint associated
with DC (both for direct and indirect emissions) is significantly less when compared to
conventional deep foundation techniques. In the literature, field tests, case studies and
laboratory simulations on DC are widely documented (MAYNE et al., 1984; KUMAR
& PURI, 2001; ZOU et al., 2005; BO et al., 2009; FENG et al., 2011; BONAB &
ZARE, 2014; KUNDU & VISWANADHAM, 2016; KUNDU & VISWANADHAM,
2018). However, despite extensive field application, limited research has been done to
date in quantifying and optimizing the design parameters for effective implementation
of DC in densifying Municipal Solid Waste (MSW) landfill sites. Recently, rapid
industrialization has resulted in a massive generation of MSW in various cities across
the world. Since the year 2006, the annual rate of waste generation in developing
country like India has increased by 1.3% (Ministry of Urban Development (CPHEEO
2016)). The global waste management scenario reveals that the existing landfills are
overflowing with wastes, while acquiring new landfill sites has legal complications and
associated space constraints. These problems have triggered interest in extending the
design life of existing MSW landfill systems.
In the literature, limited studies exist on the settlement and strains induced within young
and old MSW landfills as an aftermath of the energy applied during DC (VAN IMPE &
BOUAZZA, 1996; ZEKKOS et al., 2013). However, analysis of waste settlements
based on landfill age yields unpredictable trends with widely scattered data points
(ZEKKOS et al., 2006; ZEKKOS et al., 2013). Hence, the coefficient of the primary
compression index, also known as the compression ratio (Cce), is adopted in this study
to analyse landfill performance, which represents the slope of strain (ɛv) versus an
effective vertical stress (log σ′v) curve. The primary objective is to estimate numerically
DC induced increased stiffness and densification of MSW using shear wave velocity
profiling. This technique is selected in the present study owing to its non-invasive
nature and its ability to overcome the logistic effort of drilling expensive boreholes in
urbanized settings. The constitutive relationship applicable for modeling DC on
granular material is extended to that of MSW. Finite element (FE) analyses (ABAQUS,
(2012) Dassault Systems, ver. 6.14), is conducted to assess the shear wave velocity of
soil before and after tamper blow based on the value of shear modulus, and subsequent
improvement in loose soils post DC is investigated.

2. Modelling considerations
Numerical simulation of DC on MSW necessitates a complex class of formulation
involving massive generation of surface strains. This is accounted for by adopting
arbitrary Lagrangian Eulerian (ALE) remeshing approach within ABAQUS explicit
332
module. The field DC process is numerically simulated by the application of contact
formulation between the moving tamper and MSW.

2.1 Geometry and mesh details

An axisymmetric FE model is adopted, consisting of 2 separate entities, viz. the tamper
of width equivalent to the radius (r) and MSW domain consisting of 18 m depth and 14
m width (measured from the centreline of tamper). The tamper is assumed to be a flat
based mass with a circular cross section. The classical Coulomb friction law is
employed to define contact formulation between the moving tamper and MSW. The
model mesh adopted along with relevant boundary conditions is depicted in Figure 1.

r
Axis of symmetry
z
Rt Waste (MSW)

Steel tamper
60 x 0.2 = 12 m
15 x 0.4 = 6 m

0.2 x 20 = 4 m 0.4 x 10 = 4 m 0.6 x 10 = 6 m

14 m

Figure 1. Model mesh and boundary conditions involved in FE analyses.

333
2.2 Constitutive law
In the present study, MSW is modelled as a two-phase heterogeneous system consisting
of a solid phase (waste particles and fibres) and a liquid phase (leachate). Further, as
conventional landfill sites are usually designed above the water table to prevent leachate
contamination, it is assumed that pore water pressures are negligible. Hence, the study
is applicable primarily to relatively dry landfills but not bioreactor landfills (ZEKKOS
et al., 2006). Further, the possibility of rebound of the landfill surface resulting from
post DC-induced settlements is not considered in the present context of study.
GU & LEE (2002) and GHASSEMI et al., 2010 proposed an elastoplastic constitutive
law based on DIMAGGIO & SANDLER (1971) model for simulating impact loading
on soils. This constitutive law has been chosen in this study, wherein the plastic
response is modelled according to DIMAGGIO & SANDLER (1971) cap model to
account for the permanent deformations of MSW taking place during DC, Figure 2. The
yield surface is composed of primarily two parts, a non-hardening Drucker-Prager shear
yield surface (fs) and an elliptical volumetric hardening cap (fh). Within the yield
surface, the MSW is modelled as a nonlinear elastic material using bulk modulus (K)
(Equation 1) and Poisson’s ratio,
V 
K= (1)

where V = specific volume; ′ = mean stress; and κ = the slope of the unloading line of
e-log p′ curve. Details of the constitutive law and its validation are discussed in
KUNDU & VISWANADHAM, 2020. This elastoplastic constitutive law with a
volumetric hardening cap is not inbuilt in ABAQUS, therefore, a user-defined
subroutine coded in FORTRAN language was fed as input in the analysis.

Transition surface
B
Shear yield surface (fs)

A Hardening cap (x-l)/R

tan - 1 a surface (fh)

D C I1
l
x

Figure 2. Drucker-Prager yield surface with hardening cap.

334
2.3 Material Properties
The constitutive law discussed previously involves parameters related to hardening cap
(D, W and R), elastic behaviour (ν, κ), Drucker-Prager yield surface (α, k), and initial
void ratio (e). These constitutive parameters were determined in this study through
numerical simulation of 1D compression tests on MSW, by following the procedure
outlined by GU & LEE, 2002 and LEE & GU, 2004. The gradient of the line fitted to
the loading part of the numerical 1D compression curves represent corresponding
compression ratios (Cce). The compression ratio (Cce) presented in Equation (2)
(HOLTZ & KOVACS, 1981) is widely adopted in the literature to date for predicting
primary settlement (Δh) of MSW under surcharge.
h  (2)
Cce = = v

   +  v     +  v 
H log  v 0  log  v 0 
  v0    v0 
where σ′v0 = existing vertical effective stress; Δσ′v = additional vertical effective stress
inducing compression; Δh = the resultant settlement for initial depth H; and ɛv = the
vertical strain. As reported by STOLTZ et al., 2010, a single value of Cce can model
MSW for stresses from 20 to 300 kPa. Hence, in this study, Cce is adopted as the
engineering parameter for investigating landfill performance. As explained in KUNDU
& VISWANADHAM, 2020, the slope of stress–strain curves (ɛv versus log σv) obtained
from numerical 1D compression tests on MSW coincide with Cce values of 0.41 and
0.21 corresponding to MSW of high compressibility (MSW_HC) and moderate
compressibility (MSW_MC) respectively. The material parameters involved in the
constitutive model have been defined separately for each MSW type, as given in Table
1. The mild steel tamper is modelled as an elastic material with Young’s modulus of
210 GPa and Poisson’s ratio of 0.303, as observed in the case of standard steel
(KUNDU & VISWANADHAM, 2018).

Table 1. Constitutive model parameters adopted in FE analysis.

Properties
*
D *
W *
R #
α #
k κ υ
Type of waste
(m2/kN) - a
-a -a (kPa) -a -a
MSW_HC (Cce = 0.41) 0.008 0.55 4.33 0.22 40.96 0.0066 0.2
MSW_MC (Cce = 0.21) 0.004 0.24 4.33 0.22 40.96 0.0050 0.2
*
Work-hardening parameters; #Drucker-Prager parameters; κ: Slope of the unloading line in the e-log
p΄ curve; υ: Poisson’s ratio; -a: Dimensionless parameter.

335
2.4 Simulation of DC
The approach of simulating DC on MSW is similar to that suggested by PAN & SELBY
(2002) and GHASSEMI et al., (2010) while modeling the impact of DC on granular
soils. In the first step of the analysis, gravity load is applied to achieve a state of
equilibrium with exiting geostatic stresses at zero deformations. This is followed by the
simulation of the first tamper blow on MSW, wherein, the tamper nodes are assigned an
initial velocity computed from the free-fall equation presented in Equation (3).

V0 = 2 gh (3)
where in, Vo is the tamper velocity at the moment of impact, h represents the drop
height and g is the acceleration due to gravity (i.e., 9.81 m/sec2). In the third step, the
tamper is allowed to create a blow on the MSW surface, and densifies it. In the process,
the tamper decelerates from initial velocity (V0) to a stationary position (zero velocity),
thus completing the first blow. Multiple blows on MSW is next simulated by the
repetition of step two and three of the analyses.

2.5 Modelling of shear wave velocity profiling

In the present study, the soil improvement after DC is quantified using shear wave
velocity measurements (Vs). In the field, this is usually evaluated by conducting SASW
testing. However, a major problem encountered in this connection concerns numerical
prediction of the field dispersion curve (especially in case of layered soils), inversion of
which yields the shear wave velocity profile. Hence, in the present study, shear wave
velocity profiling is calculated using the shear modulus (G) and density of soil (ρ) as
presented in Equation (4).
G
Vs = (4)

Further, the value of shear modulus (G) at any depth is calculated using the standard
relationship involving the bulk modulus (K) (refer Equation (1)) and Poisson’s ratio (υ)
as shown in Equation (5).
3K (1 − 2 )
G= (5)
2(1 +  )
The soil improvement is studied in this case by plotting the shear wave velocity (Vs) at
various depths along the center of tamper. The robustness of the above method of
predicting shear wave velocity profiling of soil deposits is established earlier by
KUNDU & VISWANADHAM, 2018 by comparing the numerically computed (Vs)
values against field measurements of FENG et al., 2011.

336
3. Results and discussion
The impact of DC on MSW is investigated in this section by observing the influence of
waste compressibility and tamper energy on the extent of densification. Another
important design aspect to be considered in this aspect is the determination of the
optimum number of blows to be delivered in the field for different MSW
compressibilities. As the efficiency of the blows reduce gradually with successive
drops, an optimization is necessary. As evaluated by KUNDU & VISWANADHAM,
2020, the optimum number of impacts for MSW_HC (Cce = 0.41) and MSW_MC (Cce =
0.21) are about 12 and 8 respectively, as the resultant crater depth exhibits negligible
change beyond this point. Detailed information regarding the selection of DC
parameters adopted in this study is presented in Table 2.

Table 2. DC parameters adopted in FE analyses.

Parameters
E M R m H
Type of DC
(t-m) (t-m/s) (m) (t) (m)
High energy DC 400 396.2 1.6 20 20
Low energy DC 200 198.1 1.6 10 20
E: Tamper energy; M: Tamper momentum; R: Tamper base radius; m: Mass of tamper; H: Height of fall.

As can be observed from Table 2, two different energy levels are considered in the
study, viz. 200 t-m and 400 t-m, representing low-energy and high-energy DC
processes. The variation of Vs with depth (D) for the said energy levels are plotted in
Figure 3(a)-Figure 3(b) respectively for highly compressible (MSW_HC) and
moderately compressible (MSW_MC) landfill wastes. Figure 3(a) -Figure 3(b) indicate
that a substantial increase in Vs occurs along the depth of MSW after DC-induced
densification. The degree of improvement ranges from a maximum of 3-3.4 times for
highly compressible wastes (MSW_HC) irrespective of the energy level. In the case of
moderately compressible wastes (MSW_MC), the average degree of improvement is
observed to be higher. The maximum extent of improvement is in the tune of 4.4 times
for low-energy DC process and about 5.5 times for high-energy DC. Further, as can be
observed from Figure 3(a)-Figure 3(b), the resultant depth of improvement depicted in
terms of shear wave velocity profiling is considerably high for both the waste types. The
magnitude is about 12.2 m (MSW_HC) and 11 m (MSW_MC) respectively for low-
energy DC simulation, and about 14.4 m (MSW_HC) and 13.5 m (MSW_MC)
respectively for high-energy DC simulation. On the basis of the above results, it can be
inferred that the shear wave velocity profiling can be effectively applied at the site as a
useful tool for the verification of DC-induced densification of MSW landfills. The
numerical study thus contributes in focussing the efforts of expensive field trials.

337
 Shear wave velocity (m/s)
0 50 100 150 200 250 300
0
Crater depth
3
Depth (m)

9
Low energy DC (200 t-m)
12 High energy DC (400 t-m)
Initial shear velocity profile
Initial
15
(a) Highly compressible wastes (MSW_HC).

Shear wave velocity (m/s)

0 50 100 150 200 250 300
0

3 Crater
depth
Depth (m)

9
Low energy DC (200 t-m)
12 High energy DC (400 t-m)
Initial shear velocity profile
Initial
15

(b) Moderately compressible wastes (MSW_MC).

Figure 3. Shear wave velocity profile before and after DC for MSW_HC.

4. Conclusions
The study highlights the applicability of dynamic compaction (DC) based ground
improvement technology in compressing the waste volume accumulated within MSW
landfill systems. Further, the study establishes that shear wave velocity profiling is an
effective site investigation technology for the verification of DC-induced settlement of
MSW landfills. The study is especially relevant in the global scenario considering that
majority of the existing MSW landfills have attained their design capacity. Finite

338
element (FE) based analysis was conducted using ABAQUS (Explicit module using
Drucker-Prager constitutive law with hardening cap. The effect of waste compressibility
and varying tamper energy was investigated numerically. The results indicated
substantial increase in the degree and depth of improvement monitored in terms of shear
wave velocity profiling. However, further studies are warranted through physical model
tests. The study can be further extended to observe the influence of tamper drops at
multiple drop locations and by considering the presence of daily cover systems within
the MSW landfill.

5. References
ABAQUS (2012). Abaqus analysis user’s manual, Dassault Systems, Providence, RI:
Simula Corp.
BO M.W., NA Y.M., ARULRAJAH A., CHANG M.F. (2009). Densification of
granular soil by dynamic compaction. Ground Improvement, Vol. 162(3), pp. 121–132,
https://doi.org/10.1680/grim.2009.162.3.121
BONAB M.H., ZARE F.S. (2014). Investigation on tamping spacing in dynamic
compaction using model tests. Ground Improvement, Vol. 167(3), pp. 219–231.
https://doi.org/10.1680/grim.12.00018
CPHEEO (Central Public Health and Environmental Engineering Organization). (2016).
Municipal solid waste management manual. Ministry of Urban Development,
Government of India, New Delhi, India.
DIMAGGIO F.L., SANDLER I.S. (1971). Material models for granular soils. Journal
of the Engineering Mechanics Division, Vol. 97 (3), pp. 935-950.
FENG S.J., SHUI W.H., Tan K., Gao L.Y., He L.J. (2011). Field evaluation of dynamic
compaction on granular deposits. Journal of Performance of Construction Facilities,
Vol. 25 (3), pp. 241–249. https://doi.org/10.1061/(ASCE)CF.1943-5509.0000160
GHASSEMI A., PAK A., SHAHIR H. (2010). Numerical study of the coupled hydro-
mechanical effects in dynamic compaction of saturated granular soils. Computers and
Geotechnics. Vol. 37 (1–2), pp. 10-24. https://doi.org/10.1016/j.compgeo.2009.06.009.
GU Q., LEE F.H. (2002). Ground responses to dynamic compaction of dry sand.
Géotechnique, Vol. 52 (7), pp. 481-493. https://doi.org/10.1680/geot.2002.52.7.481
HOLTZ R. D., KOVACS W.D. (1981). An introduction to geotechnical engineering.
Englewood Cliffs, NJ: Prentice Hall.
KUMAR S., PURI V.K. (2001). Soil improvement using heavy tamping—a case history.
ISET Journal of Earthquake Technology, Vol. 38(2-4), pp. 123–133.
KUNDU S., VISWANADHAM B.V.S. (2016). Studies to evaluate the impact of tamper
on the depth of improvement in dynamic compaction. Japanese Geotechnical Society
Special Publication, Vol. 2(59), pp. 2033–2037. https://doi.org/10.3208/jgssp.IND-20

339
KUNDU S., VISWANADHAM B.V.S. (2018). Numerical studies on the effectiveness
of dynamic compaction in loose granular deposits using shear wave velocity profiling.
Indian Geotechnical Journal, Vol. 48 (2), pp. 305-315. https://doi.org/10.1007/s40098-018-0298-2
KUNDU S., VISWANADHAM B.V.S. (2020). Numerical modelling of dynamic
compaction induced settlement of MSW landfills. International Journal of
Geomechanics, ASCE, Vol. 20(8), pp. 04020125-1:12. doi: 10.1061/(ASCE)GM.1943-5622.0001754
LEE F. H., GU Q. (2004). Method for estimating dynamic compaction effect on sand.
Journal of Geotechnical and Geoenvironmental Engineering, Vol. 130(2), pp. 139-152.
https:// doi.org/10.1061/(ASCE)1090-0241(2004)130:2(139).
MAYNE P.W., JONES J.S., DUMAS J.C. (1984). Ground response to dynamic
compaction. Journal of Geotechnical Engineering 110 (6), pp. 757–774. https://doi.org
/10.1061/(ASCE)0733-9410(1984)110:6(757)
PAN J.L., SELBY A.R. (2002). Simulation of dynamic compaction of loose granular
soils. Advances in Engineering Software, Volume 33, Issues 7–10, pp. 631-640,
https://doi.org/10.1016/S0965-9978(02)00067-4
STOLTZ G., GOURC J.P., OXARANGO L. (2010). Characterization of the physico-
mechanical parameters of MSW. Waste Management (Oxford), Vol. 30(8–9), pp. 1439-
1449. https://doi.org/10.1016/j.wasman.2010.03.016
VAN IMPE W.F., BOUAZZA A. (1996). Densification of domestic waste fills by
dynamic compaction. Canadian Geotechnical Journal, Vol. 33 (6), pp. 879–887. https://
doi.org/10.1139/t96-118
ZEKKOS D., BRAY J.D., KAVAZANJIAN E, MATASOVIC N., RATHJE E. M.,
RIEMER M.F., STOKOE K.H. (2006). Unit weight of municipal solid waste. Journal of
Geotechnical and Geoenvironmental Engineering, Vol. 132(10), pp. 1250-1261.
https://doi.org/10.1061/(ASCE)1090-0241(2006) 132:10(1250)
ZEKKOS D., KABALAN M., FLANAGAN M. (2013). Lessons learned from case
histories of dynamic compaction at municipal solid waste sites. Journal of Geotechnical
and Geoenvironmental Engineering, Vol. 139 (5), pp. 738-751.
https://doi.org/10.1061/(ASCE)GT.1943-5606.0000804
ZOU W.L., WANG Z., YAO Z.F. (2005). Effect of dynamic compaction on placement
of high-road embankment. Journal of Performance of Construction Facilities, Vol.
19(4), pp 316-323.

340
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Economic viability of using iron ore tailings to make blocks to

be reinforced with geosynthetics

Filipe XAVIER 1, Gabriela PRINZ 1, Paulo CASTRO 1, Rodrigo FONSECA 1,

Isabelle SILVA 1

1. Golder Associates Brasil, R. Antonio de Alburquerque, 194 – Funcionarios, Belo

Horizonte – MG, 30112-010, Brazil.
[email protected]; [email protected]; [email protected];
[email protected]; [email protected]

Abstract:
The objective of this work is to show the economic feasibility of using containments
made of soil reinforced with geosynthetics, using block faces that replace the
conventional fine aggregate with iron ore tailings. The analyzed aspect is to highlight a
city like Belo Horizonte (Brazil), in which it has high landslide rates and large dams.
The reuse of the tailings potentiates and facilitates the decommissioning actions of
mining dams, which is currently being discussed due to the demands of regulatory
bodies. Geosynthetics are materials produced from synthetic or natural polymers and are
increasingly being used in the mining area by several factors, among them, because they
have a shorter execution time (easy to apply) and cause less environmental damage.
Through the study of the technological potential of tailings in civil construction shown
in several articles and a simple market research, trying to prove the economy of this
substitution in the region, he concludes that the volumes found to become economically
viable. The availability of tailings and the proximity to urban centers show that there is
an economical and ecologically correct solution, and that in practice it becomes another
way to mitigate mass movements in addition to a solution for decommissioning dams.

Keywords:
Economy, Sustainability, Block, Tailings, Viability.

Corresponding author mail: [email protected]

1. Introduction
The consumption of raw materials occurs rampant in the current society, harming future
generations (PORTELLA, 2015). The iron ore mining process generates large volumes
of waste, known as tailings. Such tailings at first have no economic value besides being
an environmental obstacle, as they must be stored in large structures such as dams.
The breakdown of structures like these in Brazil has occurred in recent years, causing
several problems such as socioeconomic, socio-territorial, and environmental impacts.
According to the Public Civil Inquiry (2016), mining companies are required to reuse
about 70% of their iron ore tailings before the year 2025, in 2019 the Law was enacted
23.291 / 19 in which it determines that dams that are intended for the accumulation or
for the final or temporary disposal of tailings over 10 meters (counting from the lowest
point to the crest) or with a volume equal to or greater than one million cubic meters
must be emptied; in the case of inactive ones, in the case of active ones within three
years, promote the migration to an alternative technology.
With this, due to the recent accidents and the demands of the regulatory bodies, the
decommissioning of mining dams has been widely ruled and discussed in the mining
industry.
According to Resolution No. 32 of May 11, 2020 (BRASIL, 2020b) decommissioning
means the “closing of operations with the removal of associated infrastructures, such as,
but not limited to, spikes, pipes, except those intended to guarantee safety of the
structure.”
According to Law No. 14.066 of September 30, 2020 (BRASIL, 2020a), an
uncharacterized dam is "one that does not operate as a sediment or tailings containment
structure, and does not have dam characteristics, and is intended for another purpose".
Therefore, the reuse of the tailings can be considered as a positive aspect in the
decommissioning or de-characterization of mining dams.
Currently, there is a paradox within the use of aggregates in civil construction, sand
production meets national demand, however such demand has been growing and
availability decreasing due to factors such as scarcity of these resources close to large
urban centers, environmental problems, legislative restrictions, and soil dispute.
Another recurring obstacle in urban centers is landslides, caused by several factors such
as the geometry of slopes and embankments, geological and geotechnical conditioning,
presence or absence of vegetation, position of the water level in the terrain, local
climatic conditions (such as excess precipitation), urban occupation, among others.
With this the need to create intervention, to avoid the movement of embankments.
The use of geosynthetics came to assist in several aspects’ projects both in geotechnical
works and in projects related to environmental protection, among its various properties
such as watertightness, filtration, drainage, containment, its action as a reinforcement
element when inserted into the soil stands out, improving the resistance and deformation
of natural soil (SANTOS & SILVA, 2019). Thus, the use of geosynthetics makes it
342
possible to build containments that challenge factors such as the elevation variation of a
terrain.

2. Objective
The general objective is to show the economic feasibility of using containments made of
soil reinforced with geosynthetics, using block faces that replace the conventional fine
aggregate with iron ore tailings. Making an ecologically correct structure, reducing the
volume of dams, exploring new deposits, in addition to creating a viable solution for
mass movements, achieving economic and environmental gains.
The specific objective is to assess the viability by exemplifying the city of Belo
Horizonte (Brazil), a capital that has a variable topographic elevation, which needs
intervention measures on slopes and slopes, and which have a relatively large number,
when compared with other capitals, from iron ore dams in its vicinity. To estimate this
possible use, an example of application will be made with the proposed methodology,
real dams will be analyzed, close to a region of Belo Horizonte where there are large
cases of mass movements, and through logistics a minimum volume of substitution was
determined. which makes changing the fine aggregate into blocks economically
attractive.

3. Development and methodology

3.1 Assessment of the tailing

To evaluate the potential use of Iron Ore Dam Tailing (IODT) in slope wall blocks
structured by soil-reinforced with geosynthetics, a review and analysis of the
characteristics of the IODT and the IODT concrete blocks was carried out. Several
researchers have used the tailings in other production process cycles, such as mortar
aggregates (FONTES et al, 2016), ceramics (BEHERA et al., 2019) and road
infrastructure (SANT’ANA FILHO et al., 2016). However, for these solutions to
become viable, it is essential to carry out tests to know the behaviour of the IODT with
the cementitious matrix.
Generally, the tests performed for characterization are (X-Ray Fluorescence (XRF), X-
Ray Diffraction (XRD), Scanning Electron Microscopy (SEM), particle size
distribution, moisture content, specific mass, unit mass and testing of harmful
substances. Such tests aim to determine the possible use of the tailing by substituting the
fine aggregate. In addition, the dosage of this substitution is another important factor,
for which it is evaluated the expandability, water absorption and resistance to
compression (SANT’ANA FILHO, 2013).
Through the XRF, the presence of no harmful elements such as lead, mercury and
arsenic are evaluated. Presence of chlorides that generates pH variation leading to

343
carbonation or expansive substances such as sodium and potassium oxides that
generates the alkali-aggregates reaction (XAVIER, 2019).
Through the XRD it is possible to identify possible minerals such as Quartz (SiO2),
Hematite (Fe2O3), Kaolinite (SO3) and Gypsum (CaO). Quartz is used in various
products in civil construction as a fine aggregate. Gypsum is used mainly in the
manufacture of cement, kaolinite has ideal pozzolanic functions for concrete and
hematite it is a denser element than the natural fine aggregate and has a reddish-brown
color, that is, they are minerals that can be used together with the cement matrix. The
granulometric curve of the tailing must be adjusted next to a trace of aggregates for a
composition close to that of the Usable Zone (XAVIER, 2019).
In the Figure 1 there is an example of the results of the compressive strength of
interlocking blocks for paving. The values presented show the replacement in volume of
conventional aggregate by IODT.

Figure 1. Compressive strength of blocks with replacement of natural fine aggregate by

IODT, broken after 7 and 70 days of molding, reference blocks (REF), 10% 50% and
80% replacement (SANT’ANA, 2013).

It is observed, the variation of the resistance to compression, there is a decrease as the

volume of replacement increases, however the vestment blocks have no structural
function, they work as an adornment, so the replacement of up to 80% of the aggregate
does not become a hindrance.

344
3.2 Case study
As shown in Figure 2 and Table 1, the Center South and East regions contain the
highest number of buildings at risk by region, in addition to having the highest
comparative of geological risk classified as high and very high. Such level of alert (very
high) portrays an evolutionary stage in which there is no need for pluviometric events
for there to be immediate destruction of homes (URBEL - URBANIZING AND
HOUSING COMPANY OF BELO HORIZONTE).

Table 1. Regions with geotechnical risks in the Belo Horizonte region,

(Adapted from diagnostics of the geological risk situation – 2015/16, URBEL (2016)).
High and very high geological risk
1994 2004 2009 2011 2015/16
Barreiro 801 798 652 418 162
South Center 2133 2049 627 638 319
East 1281 3216 798 354 225
North 2531 381 257 256 135
North East 1996 1089 234 365 180
Northwest 1180 904 278 78 74
West 1180 1261 765 320 196
Pampulha 694 223 27 207 76
Venda Nova 2553 733 151 125 138
Total 14349 10659 3789 2761 1505

Figure 2. Risky buildings in 2015/16, adapted from diagnostics of the geological risk
situation – 2015/16, URBEL, 2016).

345
In a casual way, a large part of the dams near Belo Horizonte are in municipalities
adjacent to the regions where there is the greatest number of mass movements. Table 2
shows the volume, class, and distance of these dams to the center of the most affected
regions.

Table 2. Table assembled through the interactive map of dams registered in the system
(NWA – National Water Agency) and classification (NMA – National Mining Agency),
average distances from the regions determined through Google Earth.
Distance from the region
Dams 3
Volume (m ) Class South Center (km) East (km)
5 (MAC) 13,718,217.00 B 31.3 35
5 7,100,000 B 32 36
7B 79,431.00 B 14.2 17.2
8B 92,635.00 B 13.8 19.3
Galego 1,261,209.47 B 24.4 18.6

In this perspective, the work proposes to evaluate the cost of logistics between the
studied dams and the storage in pile. It is noteworthy that the cost of transportation
between the pile and the possible work was not considered, since the regions include an
extensive area in the city. In order to estimate economic viability, the price of the
aggregate commonly used for making blocks was evaluated, but without accounting for
freight, only the price taken from the deposit, the values were determined through
SINAPI tables of April 2021 (CAIXA ECONÔMICA FEDERAL, 2021), the average
price per cubic meter of fine sand is US$ 14,693. Thus, it is determined from which
volume the logistics (removal of the tailings stored in the dam and transport to the plant
where the material will be dried) will present economic gains compared to the cost of
the raw material from deposits.
Thus, to determine the necessary volume, several costs will be scored, such as the
purchase of equipment for the transportation and removal of tailing, fuel costs,
laboratory tests for characterization of the material (to validate its possible replacement)
and execution of a shed where the drying process will be carried out. For the execution
of the methodology, it is necessary that the excavator in the dam reservoir previously
select the IODT, selecting the real pulp of the tailing in which there is the most granular
tailing material. The suggested average transport distance (ATD) is 5.0 km between the
dam and the possible shed, such a distance was inferred to avoid interference between
the operation of the dam and the equipment flow. It is worth mentioning that it would be
interesting to build the shed next to the work to be carried out, thereby reducing future
freight expenses even more, remembering that the construction of structures in regions
downstream from dams should be avoided. When doing the first cycle between dam and

346
shed, the tailings will be dumped and previously spread by a bulldozer in an area with
sufficient dimensions to meet the demand for ventilation. The use of additives to help
stabilize the material may come to be used, such as hydration stabilizers that provide
greater water flow by capillarity (ANTONIAZZI et al., 2020).

4. Results
The equipment for the simulation of tailings transport was the 8x4 trucks and a
compatible excavator for loading. Conservatively, the ATD time was estimated at 5.0
km, 2 trips per hour, with a cycle reaching almost 30 min. Thus, the stretch would be
covered at a safe speed (below 30 km) since this logistics is taking place inside a mine,
in addition to assuming a tortuous route between the dam and the installation of the
drying plant. Diesel oil consumption was estimated based on the manufacturers' table,
the purchase price based on a market survey through the ERIF table (Economic
Research Institute Foundation) and other sources that were blunt with the market.
It is worth mentioning that the capacity of the proposed trucks is 37.0 t, and the iron ore
tailing is composed of hematite and quartz that have densities between 5.25 g / cm3 and
2.65 g / cm3 respectively (PORTES, 2013), so the tailing tends to have density between
both extremes (PEREIRA, 2005). With that, we considered 3.5g / cm3 as a reference
value, so in each trip the truck would take 10.57 m3 of tailing. During an 8-hour shift
with two trucks, it would be possible to transport 338.24 m3 of tailing, according to the
Ministry of Mines and Energy of Brazil, the average price of diesel oil is US$ 0.67.
The Table 3 and Table 4 below show the initial investment, average daily cost with
diesel, estimated reserve for possible maintenance, laboratory costs for tailing assessing
and the operators' salary in approximately 4 months.

Table 3. Production costs of the tailing destined for replacement.

Costs Unity Total
Shovel Libher 954 US$ 180,000 1 US$ 180,000
Truck 4844 US$ 90,000 2 US$ 180,000
Bulldozer D6-T US$ 90,000 1 US$ 90,000
Drying shed US$ 18,000 1 US$ 18,000
Diesel liter of oil US$ 0.67 49,208 L US$ 32,969
Laboratory US$ 5,400 1 US$ 5,400
Daily maintenance US$ 180 108 US$ 19,440
Operators’ salary US$ 2,160 4 US$ 8,640
Expenses US$ 534.449
Note: 0.18 US$. = 1 R$ (Brazilian Real)

347
Table 4. The volume needed to become economically viable.
Daily production 338.24 (m3)
Unit cost US$ 14,693
Service days 108
Profit US$ 536,734.11
Volume tailings required 36,529.92 (m3)

Thinking about the use of all that volume as a substitute for aggregate in wall blocks, it
must be evaluated how many square meters of wall face can be produced with such
volume. Below in Table 5, a possible mass trace is shown to make a concrete block
replacing the aggregate.

Table 5. Viable replacement of the aggregate, meeting the minimum characteristics of

the wall block (XAVIER, 2019).
Daily production Block Ref. Block 80%
Crushed aggregate 8 kg 8kg
CPV-ARI 7.72 kg 7.72 kg
Sand / Tailings 13.9 / 0 kg 2.78 / 11.12 kg
Additive 0.14 L 0.14 L
Water 1.23 L 1.9 L
Yield 1 block 1 block

The block for the commercialized wall has an average of 30 kg, and each one with 0.08
m2 of wall face, that is, for each 1.0 m2 of wall approximately 13 blocks are used,
according to a viable proposal for the substitution of raw material 1, 01m2 would lead to
the reuse of 144 kg of IODT.
Considering the 3.5g / cm3 density of the tailing, 72,817.99 kg of tailing would be
available, which would make 505.68 m2 of wall face. This number is feasible in view of
the need for interventions in mass movements in the region. The advantages of having a
wall with a block face are that it avoids the depreciation of geosynthetics and
vandalism, which could occur in large centers like Belo Horizonte.

5. Conclusion
All the dams have volumes greater than the minimum for the viability of the tailing.
However, to replace the aggregate, a complete characterization of the material is
necessary, a cost that was raised within the estimates of production expenses.
The present paper carried out a general analysis of a possible logistics, an optimum
number of trucks per daily cost was not counted, to further reduce this expense, perhaps
there may be an optimization of values.

348
Possible logistics reports a new reuse of iron ore tailing, but now on commercial scales,
the purpose of the paper is to stimulate new substitution research, thus creating
materials that accept a greater volume of reuse of tailing.
It is also notable that there is a geotechnical solution for landslides in the regions, in
addition to being viable, it is economical and sustainable, partly solving the problem of
the lack of mischaracterization of nearby dams.

6. References
ANTONIAZZI J.P., MOHAMAD G., CASALI J.M., SCHMIDT R.P.B., TRINDADE
B. (2020). Ação dos aditivos estabilizador de hidratação e incorporador de ar em
pastas de cimento Portland, Ambiente construido, Porto Alegre, July/Sept. 2020,
vol.20, n°3, pp. 249-262, http://dx.doi.org/10.1590/s1678-86212020000300427
BEHERA K., BOSE B.P., MONDAL M.K. (2019). Production of construction bricks
using iron ore tailings and clay, in Ghosh SK (ed.), Waste management and resource
efficiency, Proceedings of 6th IconSWM 2016, pp. 583-596, doi.org/10.1007/978-981-10-7290-1_49
BRASIL (2020a). Lei n. 14.066, de 30 de setembro de 2020. Altera a Lei nº 12.334, de
20 de setembro de 2010, que estabelece a Política Nacional de Segurança de Barragens
(PNSB), a Lei nº 7.797, de 10 de julho de 1989, que cria o Fundo Nacional do Meio
Ambiente (FNMA), a Lei nº 9.433, de 8 de janeiro de 1997, que institui a Política
Nacional de Recursos Hídricos, e o Decreto-Lei nº 227, de 28 de fevereiro de 1967, 30
de set. 2020.
BRASIL (2020b). Resolução Nº 32, de 11 de maio de 2020. Altera a Portaria nº
70.389, de 17 de maio de 2017 e dá outras providências. Diário Oficial da União.
Brasilia – DF.
CAIXA ECONOMICA FEDERAL (2021) Tudo sobre o SINAPI Disponível em:
https://www.caixa.gov.br/site/Paginas/downloads.aspx#categoria_648 (acessado em 9 de abril de 2021).
FONTES W.C., CASTRO J., DA SILVA S.N., PEIXOTO R.A.F. (2016). Mortars for
laying and coating produced with iron ore tailings from tailing dams. Construction and
Building Materials, 112(1), pp. 988–995. https://doi.org/10.1016/j.conbuildmat.2016.03.027
PEREIRA E.L. (2005). Estudo do potencial de liquefação de rejeitos de minério de
ferro sob carregamento estático. Tese de D.Sc, UFOP, Ouro Preto, MG, Brasil.
PORTELLA M.O. (2015). Efeitos colaterais da mineração no meio ambiente, Revista
Brasileira de Políticas Públicas, v.5, n.3.
PORTES A.M.C. (2013). Avaliação da disposição de rejeitos de minério de ferro nas
consistências polpa e torta. Dissertação de M.Sc., UFMG, Belo Horizonte, MG, Brasil.
SANT’ANA FILHO J.N.de (2013). Estudos de reaproveitamento dos resíduos das
barragens de minério de ferro para uso nas rodovias e fabricação de blocos
intertravados, Dissertação de Mestrado, CEFET-MG, Belo Horizonte, MG, Brasil,
130p.

349
SANT'ANA FILHO J.N.de, DA SILVA S.N., SILVA G.C., MENDES J.C., PEIXOTO
R.A.F. (2016) Técnico e ambiental viabilidade funcional de pavimentadores de
concreto intertravados com cauda de minério de ferro de barragens de rejeitos. Jornal
de Materiais em Engenharia Civil 29: 04017104.
SANTOS L.R.R., SILVA L.C.A.F.M. (2019). Aplicações de geossintéticos na
geotecnia ambiental : análise de publicações Brasileiras nos ultimos 20 anos (1999 a
2018), UFERSA, Brazil.
URBEL - URBANIZING AND HOUSING COMPANY OF BELO HORIZONTE -
(2016). Diagnóstico da Situação de Risco Geológico da Vilas, Favelas e Conjuntos
Habitacionais de Belo Horizonte – 2015/16, URBEL, Prefeitura de Belo Horizonte.
XAVIER F.G. (2019). Estudo do potencial tecnológico de uso de rejeito de minério de
ferro em blocos para paramento de estruturas reforçadas por geossínteticos em Ouro
Preto - MG. Trabalho de Conclusão de Curso – UFOP – Ouro Preto. Disponível em:
https://www.monografias.ufop.br/handle/35400000/2118

350
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Effect of adding slags on strength and leaching properties of

soft soil

Aye Cho Cho ZAW 1, Lincoln Waweru GATHUKA 1, Yan TIAN 1,

Atsushi TAKAI 1, Takeshi KATSUMI 1

1. Graduate School of Global Environmental Studies, Kyoto University, Yoshida-

honmachi, Sakyo-ku 606-8501, Kyoto, Japan.
[email protected], [email protected],
[email protected], [email protected],
[email protected]

Abstract:
Reusing dredged soil is an essential approach for reducing the amount of geo-wastes.
However, it is crucial to improve its mechanical property and control potential
contaminants before using the dredged soil as engineering materials. The pre-treatment
method of mixing dredged soil with additives has been commonly used. In this study,
blast furnace slag and steelmaking slag were used as additives to understand their
performance in improving both the strength and leaching properties of the soft soil.
Each slag was mixed with Kasaoka clay simulating dredged soft soil with slag contents,
CS, of 10, 30 or 50%. The specimens were cured for 7, 14 or 28 days. Unconfined
compression strength (UCS) tests were conducted to evaluate the strength of the
specimens. Batch leaching tests were conducted to study the leaching behaviour of the
specimens. For batch leaching test, laboratory-prepared artificially contaminated soil
(with an arsenic content of 20 mg/kg) was used to evaluate the leaching property of
slags in the amended soil. UCS results show that the soft soil’s strength was
significantly improved because of mixing the soil with steelmaking slag. For CS = 50%
and 28-days curing resulted in strength of ~250 kPa. However, no noticeable
improvement in strength resulted from blast furnace slag addition. Batch leaching test
results show that the arsenic leaching concentration of the artificially contaminated soil
was ~1.9 mg/L. This concentration was reduced to values of 0.2–1.0 mg/L because of
blast furnace slag addition. Much lower concentrations of <0.02 mg/L could be obtained
due to steelmaking slag addition.

Keywords:
Ground improvement, Blast furnace slag, Steelmaking slag, Dredged soil, Unconfined
compression strength, Immobilisation, Arsenic.

Corresponding author mail: [email protected]

1. Introduction
Utilising the large volumes of dredged soil generated every year is much encouraged to
reduce the amount of geo-wastes. However, the poor mechanical properties and/or
contamination make the dredged soil not suitable to be reused for engineering purposes
without treatment (CHAN & JALIL, 2014; WEERAKOON et al., 2020). Pre-treatment
is required to improve its mechanical properties and environmental safety of the soft
soils. A common treatment method is using additives for improving mechanical
properties as well as immobilising contaminants. Industrial by-products such as slags
have been introduced to be used as additives for soil improvement (REDDY et al.,
2019; ZHANG et al., 2020). This study discusses two iron and steel slags, namely: blast
furnace slag and steelmaking slag. The blast furnace slag is generated as by-products
from the stage of iron production, and the steelmaking slag was from the stage of steel
production. The steelmaking slag has high potential of hydration capacity similar to
cement due to the significant content of Ca(OH)2. Several studies have evaluated the
slags amendment for strength improvement of soft soil (KANG et al., 2019;
WEERAKOON et al., 2020; ZHANG et al., 2020). However, very few studies have
evaluated slag amendment for ensuring the environmental safety of contaminated soil
(REDDY et al., 2019). Moreover, many previous studies conducted the combined use
of slags with other materials such as cement or admixing agents.
This present work intends to evaluate the individual performance of the two slags
mentioned above in amending a soft soil by focusing on the strength and leaching
properties. This approach may successfully reduce the amount of cement consuming in
ground improvement, reducing the emission of CO2. This study may support
understanding the applicability of each slag as an additive for ground improvement and
the benefit of resource utilization from the environmental perspective. Each slag was
mixed with Kasaoka clay using slag contents, CS, of 10, 30 or 50 wt.%. The specimens
were cured for 7, 14 or 28 days. Unconfined compression strength (UCS) tests were
conducted to evaluate the strength of the amended soil. Batch leaching tests were
conducted to study the leaching behaviour of the amended soil.

2. Methodologies

2.1 Materials
A commercially available Kasaoka clay was used as host soil. Table 1 shows the basic
physical properties of the clay. The clay constitutes ~90% fines content (<0.075 mm in
diameter). X-ray fluorescence was conducted for the clay. Results show that the clay
consists of 66.2% SiO2, 19.3% Al2O3, and 5.9% Fe2O3. The remaining 8.6% includes
K2O, CaO, TiO2, MgO, etc.
Artifically contaminated Kasaoka clay was prepared to be used in the batch leaching
tests. Since arsenic is one of the most commonly found geogenic heavy metals, it is
352
considered in this study. The target arsenic (As) concentration of Kasaoka clay is 20
mg/kg referring to composition of As in silt and clay of marine deposit (MITAMURA
& MASUDA, 1998). The clay was soaked with sodium arsenite (NaAsO2) solution for
1 week. The initial concentration of arsenic in the solution and the concentration of
arsenic in the As-clay mixture after keeping for 1 week were investigated by Atomic
Absorption Spectroscopy (AAS) (SHIMADZU, AA-6800). The concentration of As in
prepared Kasaoka clay was calculated by differentiating the initial concentration and the
concentration after 1 week. That is ~20.45 mg/kg which is in the similar range of
intended concentration. This clay containing artificial arsenic is hereafter called as
"contaminated clay" in this manuscript.

Table 1. Physical properties of Kasaoka clay.

Parameter Method Value
Particle density JIS A1202:2009 2.64 g/cm3
Particle size distribution JIS A1204:2009
Gravel fraction (>2.0 mm) 0%
Sand fraction (0.075–2.0 mm) 5.3%
Fine fraction (<0.075 mm) 94.7%
Atterberg limits JIS A1205:2009
Plastic limit 23.7%
Liquid limit 51.9%
Plastic index 28.2%

Blast furnace slag and steelmaking slag from Nippon Steel Corporation were used as
additives. Figure 1 shows the particle size distribution of the two slags and Kasaoka
clay. Only steelmaking slag has certain amounts of gravel contents, while blast furnace
slag mainly consists of sand contents. Table 2 lists the chemical components of two
slags. Both slags contain certain content of calcium oxide (CaO). Blast furnace slag
contains a higher amount of SiO2 than in steelmaking slag, while iron and Ca(OH)2 are
composed in steelmaking slag with a significant amount.

Figure 1. Particle size distribution of materials.

353
Table 2. Chemical components of slags (unit: wt.%).
Component Method Blast furnace slag Steelmaking slag
CaO 41.1 38.6
SiO2 33.1 13.7
T-Fe 0.95 16.8
MgO 4.98 6.27
MnO ICP analysis 0.21 3.27
P 2O 5 0.02 2.86
f-CaO ≤0.30 2.35
Al2O3 12.4 1.86
f-MgO 0.02 0.06

T-S Heating infrared adsorption method 1.11 0.03

Ca(OH)2 Thermogravimetric analysis ≤0.2 6.9

2.2 Preparation of specimens

Test specimens were prepared by mixing Kasaoka clay or contaminated clay with each
slag at three slag contents. The contents of slag, CS, are 10, 30 or 50%, which
correspond to 100, 300, or 500g/kg-soil in dry weight, respectively. The initial water
content was 1.2wL of Kasaoka clay. Three different curing times of 7, 14 or 28 days
were considered. Plastic moulds with 50 mm in diameter and 100 mm in height were
used for specimen preparation. First, clay, slag and distilled water slurry were prepared
using a mixer and filled into the moulds by pouring three layers. Tapping was applied
during the filling once after each layer for removing air bubbles. Then the top of the
moulds were sealed with plastic film, and the moulds were cured within a vacuum bag
containing a cup of distilled water to keep 100% humidity for relevant curing times.
Specimens were prepared in triplicate. The specimens for the batch leaching tests were
prepared by using the contaminated clay.

2.3 Unconfined compression tests

UCS tests were conducted by referring to JIS A 1216:2009. This test is commonly used
for clay specimens and applied to self-standing specimens without the presence of
confining pressure. The purpose is to determine the unconfined compressive strength of
a cohesive soil sample. Two loading rings of different capacity (500 and 2000 N) were
applied depending on the condition of specimens (e.g., CS value and the curing times,
etc.). The compressive rate of the test was 0.1 cm/s, which is equal to 0.1% of specimen
height in this study.

2.4 Batch leaching tests

Conventional batch leaching tests were conducted at a liquid to solid ratio of 10 to
evaluate the leaching behaviour of As in amended soil. The appropriate volume of
distilled water was poured into a 500 mL plastic bottle, and an appropriate mass of
354
samples was added to the plastic bottle. The mixtures were horizontally shaken at 200
rpm for 6 hours. After that, centrifugation was done at 3,000 rpm for 20 minutes, and
the sediments were filtered by using 0.45 μm membrane filters. Then, the pH, electrical
conductivity (EC), and redox potential (Eh) of the solutions were measured. After that,
the concentration of metal ions (Ca, Al, Fe, Mg, Zn, etc.) was measured using
inductively coupled plasma mass spectrometry (ICP) analysis (710 ICP-OES). The
concentration of As was measured using atomic absorption spectroscopy (AAS)
analysis (Shimadzu, AA-6800).

3. Results and discussion

3.1 Strength property

After reaching each targeted curing time, the specimens with the mixture of clay and
blast furnace slag for all mixing ratios showed slight improvment of strength. However,
the strength was not high enough to form as self-standing specimens. Therefore, the
specimens were not able to be conducted the UCS tests. The possible reasons are that
the slag does not contain sufficient content of CaO to develop reactions, such as the
formation of carbonates and cementitious materials and containing less number of fine
particles for strength improvement. The specimens with clay-steelmaking slag mixture
were solidified, and all the specimens could be subjected to the UCS test. The results of
the stress-strain relationship of the specimens with different curing days and slag
contents are shown in Figure 2. As shown in this figure, the strength enhancement and
deformation of specimens with CS = 10% were similar with soil. The characteristics of
specimens with higher slag content shows more similarity with soil amended with
cement (see Figure 3). The improvement of strength in the soil amended with
steelmaking slag can be due to three reasons. First, change in particle size distribution
since steelmaking slag contains over 50% of coarse aggregates (>2 mm in diameter).
Second, the presence of free-CaO and free-MgO in steelmaking slag affects strength
development since those materials may form CaCO3 after long-term curing. CaCO3 can
fill the voids between soil particles which benefits strength performance (KANG et al.,
2019). Third, the cementitious materials like calcium-silicate-hydrate (C-S-H) and
calcium-silica-alumina (C-A-S-H) generated during hydration be admitted as gel. C-S-
H and C-A-S-H bind the soil particles together and change aggregates size of amended
soil (WEERAKOON et al., 2020). The specimens were cured with the same condition
of 100% humidity. However, the water content was significantly reduced in specimens
with higher slag content. The water content in a specimen with CS = 50% for 28 days
was reduced to 71.5% from the initial water content. Since the strength and stiffness of
clay may have a strong relationship with matric suction, reduction of water content may
impact strength, which is not considered in this study. This factor shall be studied in the
following research.
355
 Figure 2. UCS test results for the specimens with different CS and curing days.

(a) CS = 10% (b) CS = 30%

Figure 3. Deformation of specimens for 14-days curing.

3.2 Leaching behaviour

Batch leaching tests were conducted for the contaminated clay (CS = 0%), soil amended
with blast furnace slag and soil amended with steelmaking slag. The average leaching
concentration value of arsenic for contaminated clay is 1.9 mg/L. Figure 4 shows the
leaching concentration of arsenic from the soil amended with blast furnace slag. The
values were decreased compared to the contaminated clay-only cases. Figure 5 descibes
the results of batch tests for soil amended with steelmaking slag. The concentration of
arsenic was significantly reduced, which can meet the environmental quality standard of
sample solution for soil pollution in Japan. That is 0.01 mg/L. One of the mechanisms
of arsenic immobilisation is due to CaO in slags, which may form calcium arsenate
(BOTHE & BROWN, 1999). The pH value had been increased to alkaline condition
(see Figure 6), which favours the formation of carbonates. Compared to blast furnace
slag, the pH value of leachate with steelmaking slag amended soil increased much
significantly. This phenomenon may be explained by the difference of Ca(OH)2 in two

356
slags, as reported in Table 2. Moreover, iron in steelmaking slag affects the sorption of
arsenic (DEXIT & HERING, 2003). However, the reactions among the cations were not
explicitly confirmed in this present work to prove this mechanism.

Figure 4. As leaching concentration from Figure 5. As leaching concentration from

soil amended with blast furnace slag. soil amended with steelmaking slag.

Figure 6. pH values of specimens with 14-days curing.

4. Conclusions
The individual performance of two slags on strength improvement and arsenic leaching
of amended soil were considered in this present work. The findings of the study include:
(i) The resutls of UCS test shows that the strength of soil amended with steelmaking
slag is significantly improved by obtaining maximum strength of up to 250 kPa in
specimen with CS = 50% for 28-days curing. However, there is no noticible strength
improvement in the soil amended with blast furnace slag.
(ii) As leaching concentration from contaminated clay-only specimen is 1.9 mg/L. The
As concentration from soil amended with blast furnace slag reduces approximately
90% of initial concentration. The As concentration from soil amended with
steelmaking slag decreases reaching the value less than 0.01 mg/L, which meets the
acceptable leaching value in Japan for contaminated soil.

357
(iii) The UCS results with CS = 10%, 30%, and 50% are ~15, ~160, and ~250 kPa,
respectively. The results with CS = 50% for 7, 14, and 28-days curing are ~200,
~220, and ~250 kPa, respectively. This is showing that the results with different
slag conent have higher variance than those with different curing days. Hence, the
strength enhancement seems more relying on the slag content rather than curing
day.

5. Acknowledgement
The authors sincerely acknowledge to Mr. Yuzo Akashi and his colleagues from Nippon
Steel Corporation for their contributions to this research.

6. References
BOTHE J.V., BROWN B.W. (1999). Arsenic immobilisation by calcium arsenate
formation. Environmental Science and Technology, Vol. 33, Issue 21, pp. 3806–3811,
doi.org/10.1021/es980998m
CHAN C.M., JALIL A. (2014). Some insights to the reuse of dredged marine soils by
admixing with activated steel slag. Advances in Civil Engineering, Vol. 2014,
doi.org/10.1155/2014/345134
DIXIT S., HERING J.G. (2003). Comparison of arsenic(V) and arsenic (III) sorption
onto iron oxide minerals: implications for Arsenic mobility. Environmental Science and
Technology, Vol. 37, Issue 18, doi.org/10.1021/es030309t
KANG G., CIKMIT A., TSUCHIDA T., HONDA H., KIM S. (2019). Strength
development and microstructural characteristics of soft dredged clay stabilised with
basic oxygen furnace steel slag. Construction and Building Materials, 203, pp. 501–513,
doi: 10.1016/j.conbuildmat.2019.01.106
MITAMURA M., MASUDA H. (1998). Arsenic concentration of boring cores at
Osaka area. Eighth Symposium on Geo-Environments and Geo-Technics, pp. 85–88.
REDDY K.R., GOPAKUMAR A., CHETRI H.K. (2019). Critical review of
applications of iron and steel slags for carbon sequestration and environmental
remediation. Reviews in Environmental Science and Bio/Technology, Vol. 18, pp. 127–
152, doi.org/10.1007/s11157-018-09490-w
WEERAKOON N.R., NISHIMURA S., SATO H., TODA K., SATO T., ARAI Y.
(2020). Stiffness and strength mobilisation in steel-slag-mixed dredged clays in early
curing. Proceedings of the Institution of Civil Engineers-Ground Improvement, Vol.
173, Issue 2, pp. 65–81.
ZHANG W.L., ZHAO L.Y., BRYAN A.M., CHEN Y.H., MORRISON L. (2020).
Dredged marine sediments stabilised/solidified with cement and GGBS: factors
affecting mechanical behaviour and leachability. Science of the Total Environment,
Vol. 733, doi: 10.1016/j.scitotenv.2020.138551

358
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Attenuation performance of geosynthetic sorption sheet

against arsenic under different compressive stresses

Yu ZHANG 1, Yosuke KINOSHITA 2, Tomohiro KATO 1, Lincoln W. GATHUKA 1,

Atsushi TAKAI 1, Takeshi KATSUMI 1

1. Kyoto University, Graduate School of Global Environment Studies, Yoshida-

honmachi, Sakyo-ku, Kyoto, Kyoto, 606-8501, Japan.
[email protected]; [email protected]
[email protected]; [email protected]
[email protected]
2. Kyoto University, Graduate School of Engineering, Yoshida-honmachi, Sakyo-ku,
Kyoto, Kyoto, 606-8501, Japan.
[email protected]

Abstract:
Excavated soils and rocks generated in the construction contain a relatively low content
of arsenic (As), which may pose a potential threat to the environment because the As
leachability is sometimes higher than the environmental standard in Japan. A new
method, the attenuation layer method, was considered as an effective countermeasure to
address the risk of As leaching. This method applies geosynthetic sheets with sorption
capacity beneath the embankment to adsorb pollutants dissolved in leachate. A
geosynthetic sheet with sorption capacity has recently been developed as a promising
material for the attenuation layer method. However, previous researches were conducted
without considerations to the compressive forces which should be given to the sheets in
the field. Therefore, the present study examined the sorption performances of the sheet
against arsenic solution under compressive stresses of 10 and 200 kPa using improved
oedometers. The results illustrated that the cumulative flow volumes corresponding to
the breakthrough point for 10 kPa and 200 kPa are 6.62 L and 6.58 L, respectively.
Therefore, the mechanical loading has a slight effect on the attenuation performance of
the sheet. Further, the sorption mass of As(V) onto the sheet is around 20.6 times that of
As(Ⅲ).

Keywords:
Arsenic, Attenuation performance, Compressive stress, Attenuation layer method,
Geosynthetic sheet

Corresponding author mail: [email protected]

1. Introduction
Large volumes of excavated soils and rocks are generated from the various construction
works infrastructure development. For example, data from the Japanese Ministry of
Land, Infrastructure, Transport, and Tourism (MLIT 2014) reported that in 2012, the
total yield of excavated soil nationwide exceeded 140 million m3 (LI et al., 2017;
KATSUMI, 2015). How to properly manage the surplus soils/rocks after they are
excavated is an important concern in many countries (QIAN et al., 2019). The
management alternatives of excavated soils/rocks vary between construction projects,
which have been summarized as 1) use on-site 2) use in other projects 3) pre-treated
before use in other projects 4) store for later use, 5) use as landfill cover or dispose at
the landfill (MAGNUSSON et al., 2015). Nevertheless, the quantity of these surplus
soils/rocks is considerable, unless proper managed, may result in spatiotemporal land-
use hazards (BECKER et al., 2020). Therefore, to elevate material coordination and
recycling possibilities between construction fields, it is essential to provide temporary
storage areas close to each other for excavated soils and rocks (MAGNUSSON et al.,
2019).
These excavated solid materials from deep underground are rarely subjected to
anthropogenic disturbance, and the natural arsenic (As) content in them is relatively low
(KATSUMI, 2015; LI et al., 2018). However, the leaching concentration of As from
such soils/rocks could often exceed the Japanese environmental standard (10 μg/L),
because excavation-induced changes in environmental conditions may lead to enhanced
As release (TABELIN et al., 2014; LI et al., 2018). This situation effectively indicates
that soil/rock materials from deep excavations may negatively impact the surrounding
ecosystem and humans if dumped on the land surface without proper management.
Unlike artificially contaminated soils, excavated soils belong to lower-lever
contamination, and its leaching concentration just narrowly above the value in
environmental standard (KATSUMI, 2015). For effectively and sustainably managing
the low-level contaminated soils/rocks with large excavation, cost-effective
countermeasures should be seriously considered to better prevent soil contamination.
The attenuation layer method was recently developed. The method is a low
contamination-oriented countermeasure for managing large quantities of excavated
soils/rocks, achieving contaminants attenuation by installing a permeable soil layer with
sorption capacity at the bottom of contaminations (Technical Committee of
Environmental Conservation Technology Association, Hokkaido 2012). The major
attenuation process is that contaminants of concern can be partially removed by the
constituent material in the attenuation layer before reaching groundwater aquifers.
Hence, the design and selection of sorption material is an essential part of the
attenuation layer method. Local soil is regarded as a potential component material after
its mechanical and sorption properties are improved by stabilizing and/or immobilizing
agents (MO et al., 2020; INUI et al., 2013). However, this method needs to overcome
360
the obstacle that using natural soil as filtration media will ultimately lead to blockage
problems (KONG et al., 2014; SU et al., 2020). To replace the soil materials, media
with higher permeability should be experimentally studied, for the primary purpose of
remaining relatively high removal efficiency and alleviating clogging in the system.
Geosynthetics made from nonwoven fabric have proven to maintain its satisfactory
hydraulic performance after 21-years service as a filter (SARAN & VISWANADHAM,
2018). Over the past several decades, the development of geosynthetics has achieved
significant progress. Compared with the previously used natural construction materials,
geosynthetics can economically provide the desired functions required to construct
hazardous waste storage facilities (ORTEGO et al., 1995). A geosynthetic sheet with
sorption capacity has recently been developed as a promising material for the
attenuation layer method. In previous studies, this geosynthetic shows the high sorption
capacity against hazardous items like As (INATOMI et al., 2019; SHIMODA et al.,
2019). However, few studies have focused on the attenuation performance of sorption
materials subjected to compressive stress. Load action is a factor that has to be
considered in practical engineering. For the temporary storage site, during its service,
due to the continuous stacking and reuse of excavated soils/rocks, the thickness of
which shows dynamic changes. This means the earth pressure on the geosynthetic
sheets is inevitably different. Therefore, the influence of mechanical loading on sorption
performance of the sheets needs to be evaluated to guarantee the storage project safe
operation.
Given the deficiencies of current research, this study investigates the effectiveness of
the attenuation layer made from geosynthetic sheet materials under different
compressive stresses. Two stress levels 10 and 200 kPa were considered during column
tests, which were carried out in modified oedometers. Arsenic was used as the object
contaminant for this research because it is one of the most common contaminants in
naturally contaminated soils/rocks (MO et al., 2020).

2. Materials and methods

2.1 Materials
All reagents were of analytical grade. As(III) and As(V) solutions were prepared by
dissolving sodium dihydrogen arsenate (Na2HAsO4) and sodium arsenite (NaAsO2)
(Nacalai Tesque Inc., Kyoto, Japan) in distilled water, respectively. The geosynthetic
sorption sheet material (Figure 1) used was provided by Toyobo Co., Ltd., Kyoto,
Japan. The sorption sheet shows a closely-linked two-layer structure, with a thin and
compact sorption layer containing hydrotalcite and a non-woven fabric layer with long
polyester fiber. The basic physical properties of the geosynthetic sorption sheet
provided by the manufacturer are shown in Table 1. The thickness of the sorption sheet

361
is 2.5 mm. The hydraulic conductivity is 1  10-3 m/s. The mass per unit area is 450
g/m2. The mass per unit area of the adsorbent is 50 g/m2.

Figure 1. Appearance and composition of the geosynthetic sorption sheet.

Table 1. Physical properties of geosynthetic sorption sheet.

Parameters Values
Thickness of the geosynthetic sorption sheet (mm) 2.5
Mass per unit area of the non-woven fabric (g/m2) 450
Mass per unit area of the adsorbent (g/m2) 50
Hydraulic conductivity of the geosynthetic sorption sheet (m/s) 1×10-3

2.2 Experimental procedure

A novel column experimental setup (Figure 2), transformed from the traditional
consolidation instrument (MIS-232-1-35, MARUI & Co., Ltd.), can simultaneously
perform mechanical loading and column percolation tests on geosynthetic materials to
evaluate the As removal effect and estimate the loading-sorption relationship. The
columns, 6 cm in diameter and 2 cm in height, were packed with the tailored sorption
sheet sandwiched with porous stones used to diffuse the flow rate uniformly. After
installation, each specimen was submerged in distilled water and saturated by a vacuum
saturator (ASONE) for 24 h. Then the column was subjected to continuous injection of
As solution through the bottom. A peristaltic pump (Thermo Fisher Scientific) was used
to ensure a stable upward flow rate of 13-16 mL/h. The initial concentration for As(III)
or As(V) solution used in the sorption test was 0.1 mg/L. Compressive stress was
directly loaded on the specimens to simulate the earth pressure of a certain height of
excavated soils/rocks in temporary storage sites or embankments. In this investigation,
the compressive stresses were taken 10 and 200 kPa, assuming two simplified field
conditions employing a soils/rocks layer of 0 and 10 m over the sheet material,
respectively. The effluent samples were collected every several hours to measure the pH
value and the concentration of ions in the effluent. The experimental conditions of

362
column sorption tests are designed as presented in Table 2. All the experiments were
performed under room temperature (20℃).

Table 2. Experimental design of column tests.

Sample Initial concentration Compressive stress Flow rate
Influent
ID (mg/L) (kPa) (mL/h)
Case 1 As(III) solution 0.1 10 13-16
Case 2 As(III) solution 0.1 200 13-16
Case 3 As(V) solution 0.1 200 13-16

Figure 2. Schematic diagram of the up-flow column test apparatus.

2.3 Analytical methods

The surface morphology of the sorption sheet material was investigated using a
scanning electron microscope (SEM) (JSM-6390LVS, JEOL). The arsenic
concentration was analyzed by flame atomic adsorption spectrometry (AAS) (AA-6800,
SHIMADZU). The sorption capacity was calculated using Eq. (1):
(C − C )V
q= 0 (1)
S
where: q (mg/m2) represents the quantity adsorbed of As onto the sorption sheet. C0
(mg/L) and C (mg/L) are the feed and effluent concentrations, respectively. V (L) is the
volume of permeated solution, and S (m2) is the contact area of the sorption sheet.

3. Results and discussions

The removal efficiency of As(Ⅲ) by a geosynthetic sheet under two loading conditions
was compared to studied the loading-sorption relationship. The breakthrough curves
were obtained by recording the effluent to influent concentration ratio against the
permeated influent volume, as shown in Figure 3. The breakthrough point may be
defined as the cumulative flow volume at which the effluent reaches the initial
concentration. Figure 4 shows the comparison of As sorption onto the sheet material at
363
different test conditions. All column sorption tests were carried out at a fixed flow rate,
which allows a similar interaction time between the contaminant and the adsorbent (PEI
et al., 2021). As presented in Figure 3, the breakthrough curves of As(Ⅲ) influent are
steep in the former stage before the influent volume reaches 3 L. Then, the C/C0 value
trends to be flat after approximate 6 L, showing completion of the sorption equilibrium.
For the samples subjected to 10 kPa (case 1) and 200 kPa (case 2), the cumulative flow
volumes corresponding to the breakthrough point are 6.62 L and 6.58 L respectively. In
Figure 4, the sorption amount for case 1 and 2 vary in the same profile, reach a plateau
at the cumulative flow volume of around 6 L, and then become essentially stable, which
is in accord to the result presented in Figure 3. As shown in Figure 4, the maximum
sorption mass for the sheets bearing 10 kPa and 200 kPa are 41.58 and 33.61 mg/m2,
respectively. These obtained results suggested that the effluent concentrations were little
changed by the presence of different compressive stresses, and sorption mass of the
sheets decreased only slightly when compressive stress increases from 10 kPa to 200
kPa. Mechanical loading therefore has a little demonstrable effect on arsenic removal. It
might because that as a dominant role in the sorption process, the physico-chemical
properties of adsorbent in sheet materials remain unchanged under vertical loading.
Accumulative sorption mass (mg/m2)

1.0 700
Case 1 (As(Ⅲ), 10 kPa) Case 1 (As(Ⅲ), 10 kPa)
Case 2 (As(Ⅲ), 200 kPa) Case 2 (As(Ⅲ), 200 kPa)
0.8 Case 3 (As(Ⅴ), 200 kPa) 560 Case 3 (As(Ⅴ), 200 kPa)
1.0

0.8
0.6 420
C/C0

0.6

0.4 45
0.4 280
0.2 Case 1 (As(Ⅲ), 10 kPa) 36
Case 2 (As(Ⅲ), 200 kPa) 27
0.0
0.2 0 2 4 6 8
140 18

9 Case 1 (As(Ⅲ), 10 kPa)

Case 2 (As(Ⅲ), 200 kPa)
0
0.0 0 2 4 6 8
0 4 8 12 16 20 0
Effluent volume (L) 0 4 8 12 16 20
Effluent volume (L)
Figure 3. Breakthrough curves of C/C0 Figure 4. Comparison of As sorption mass
versus permeated solution volume. onto geosynthetic sheets.

364
 9.0
Case 1 (As(Ⅲ), 10 kPa)
Case 2 (As(Ⅲ), 200 kPa)
8.4 Case 3 (As(Ⅴ), 200 kPa)

7.8

pH 7.2

6.6
Influent
6.0
0 4 8 12 16 20
Effluent volume (L)
Figure 5. The variation of effluent pH in column tests.

Figures 3 and 4 also examine the comparison of As(Ⅲ) and As(Ⅴ) sorption onto the
geosynthetic sheet. As illustrated in Figure 3, it can be seen that when As(Ⅲ) (case 2)
reaches the breakthrough point, the C/C0 value of As(Ⅴ) (case 3) is still relatively low,
only 0.05. This could be interpreted that the adsorbent loaded As(V) better than As(Ⅲ).
Figure 4 shows the quantity of As(V) removal is about 20.6 times that of As(Ⅲ)
removal and presents an increasing tendency continuously. Based on its current
increasing trend, it would require a longer period to reach sorption saturation, as more
As(Ⅴ) ions are still being removed gradually, indicating that a steady state had not been
reached. These results suggest that the geosynthetic sheet is more suitable for As(V)
sorption. Figure 5 shows the monitoring of effluent pH in column tests. The initial pH
values of the influent are marked by shadows, ranging from 6.50 to 6.69. As observed,
for all cases, the pH values tend to increase to ~pH 8 initially and then decrease to ~pH
7, which may be attributed to hydrotalcite dissolution, releasing Mg and Al ions to the
solution. Some studies monitored soluble Al and Mg in the residual solution during
sorption experiments (DIAS & FONTES, 2020; MAZIARZ et al., 2019). According to
GOH and LIM (2010), under the neutral condition, the release rate of Mg ranged
between 0.1 and 1.3%, while that of Al was below 0.2%. However, hydrotalcites would
gradually become stable with negligible releases of Mg and Al during the continuous
use (GOH et al., 2009). This explains that the pH values of effluent drop back to the
initial value at the later stage of the tests.

4. Conclusions
In this work, the effect of compressive stress on the attenuation performance of the
geosynthetic sorption sheet was evaluated. In the column sorption tests, traditional
oedometer were modified to conduct column tests under mechanical loading. The
breakthrough volumes are 6.62 and 6.58 L for 10 kPa and 200 kPa. Therefore, loading
has almost no influence on the sorption properties. In addition, the geosynthetic sheet is
found to be suitable for the sorption of As(V). The adsorbent in this sheet is
365
hydrotalcite, which is considered as an excellent sorption material. Regarding the
sorption mass onto the sheet, the amount of As(V) removal is about 20.6 times greater
than that of As(Ⅲ) removal.

5. Acknowledgement
The authors wish to acknowledge to Mr. Tanaka, Mr. Shimoda and Mr. Nakamura
(Toyobo, Co., Ltd.) for providing the geosynthetic sorption materials.

6. References
BECKER N., KIMHI A., ARGAMAN E. (2020). Costs and benefits of waste soils
removal. Land Use Policy, Vol. 99, 104877, doi: 10.1016/j.landusepol.2020.104877
DIAS A.C., FONTES M. (2020). Arsenic (v) removal from water using hydrotalcites as
adsorbents: a critical review. Applied Clay Science, Vol. 191, 105615,
doi.org/10.1016/j.clay.2020.105615
GOH K.H., LIM T.T. (2010). Influences of co-existing species on the sorption of toxic
oxyanions from aqueous solution by nanocrystalline Mg/Al layered double hydroxide.
Journal of hazardous materials, Vol. 180, pp. 401–408, doi: 10.1016/j.jhazmat.2010.04.045
GOH K.H., LIM T.T., DONG Z. (2009). Enhanced arsenic removal by hydrothermally
treated nanocrystalline Mg/Al layered double hydroxide with nitrate intercalation.
Environmental Science & Technology, Vol. 43, pp. 2537–2543, doi.org/10.1021/es802811n
INATOMI S., ISHIKAWA M., SHIMODA K., NISHIOKA K., MIYAWAKI K.
(2019). Evaluation of adsorption characteristics by sheet adsorption materials (3), The
25th Symposium on Soil and Groundwater Contamination and Remediation. (In
Japanese).
INUI T., OYAMA S., MIURA T. (2013). Recent trends in ground improvement
technologies: (1) geotechnical utilization of excavated materials containing natural-
derived heavy metals. Japanese Society of Materials Science, Vol. 62, pp. 224–230.
KATSUMI T. (2015). Soil excavation and reclamation in civil engineering:
environmental aspects. Soil Science Plant Nutrition, Vol. 61, pp. 22–29.
doi.org/10.1080/00380768.2015.1020506
KONG Z., FENG C.P., CHEN, N., TONG, S., ZHANG, B.G., HAO, C.B., CHEN, K.
(2014). A soil infiltration system incorporated with sulphur-utilizing autotrophic
denitrification (SISSAD) for domestic wastewater treatment. Bioresource Technology,
Vol. 159, pp. 272–279, doe: 10.1016/j.biortech.2014.02.079
LI J., KOSUGI T., RIYA S., HASHIMOTO Y., HOU H., TERADA A., HOSOMI M.
(2017). Use of batch leaching tests to quantify arsenic release from excavated urban
soils with relatively low levels of arsenic. Journal of Soils and Sediments, Vol. 17, pp.
2136–2143, doe: 10.1007/s11368-017-1669-5
LI J., KOSUGI T., RIYA S., HASHIMOTO Y., HOU H., TERADA A., HOSOMI M.
(2018). Pollution potential leaching index as a tool to assess water leaching risk of
366
arsenic in excavated urban soils. Ecotoxicology and Environmental Safety, Vol. 147,
pp. 72–79, doe: 10.1016/j.ecoenv.2017.08.002
MAGNUSSON S., JOHANSSON M., FROSTH S., LUNDBERG K. (2019).
Coordinating soil and rock material in urban construction: scenario analysis of
material flows and greenhouse gas emissions. Journal of Cleaner Production, Vol. 241,
118236, doi.org/10.1016/j.jclepro.2019.118236
MAGNUSSON S., LUNDBERG K., SWEDBERG B., KNUTSSON S. (2015).
Sustainable management of excavated soil and rock in urban areas: a literature review.
Journal of Cleaner Production, Vol. 93, pp. 18–25, doi.org/10.1016/j.jclepro.2015.01.010
MAZIARZ P., MATUSIK J., STRACZEK T., KAPUSTA C., WOCH W.M., TOKARZ
W., RADZISZEWSKA A., LEIVISKA T. (2019). Highly effective magnet-responsive
LDH-Fe oxide composite adsorbents for as(v) removal. Chemical Engineering Journal,
Vol. 362, pp. 207–216, doi.org/10.1016/j.cej.2019.01.017
MO J., FLORES G., INUI T., KATSUMI T. (2020). Hydraulic and sorption
performances of soil amended with calcium-magnesium composite powder against
natural arsenic contamination. Soils and Foundations, Vol. 60, No. 5, pp. 1084–1096,
doi.org/10.1016/j.sandf.2020.05.007
ORTEGO J.D., AMINABHAVI T.M., HARLAPUR S.F., BALUNDG, R.H. (1995). A
review of polymeric geosynthetics used in hazardous waste facilities. Journal of
Hazardous Materials, Vol. 42, pp. 115–156, doi.org/10.1016/0304-3894(95)00008-I
PEI X., GAN L., TONG Z., GAO H., CHEN Y. (2021). Robust cellulose-based
composite adsorption membrane for heavy metal removal. Journal of Hazardous
Materials, Vol. 406, 124746, doe: 10.1016/j.jhazmat.2020.124746
QIAN J., HU Y., ZHANG J., XIAO W., LING J. (2019). Evaluation the performance of
controlled low strength material made of excess excavated soil. Journal of Cleaner
Production, Vol. 214, pp. 79–88, doi.org/10.1016/j.jclepro.2018.12.171
SARAN R.K., VISWANADHAM B. (2018). Centrifuge model tests on the use of
geosynthetic layer as an internal drain in levees. Geotextiles and Geomembranes, Vol.
46, pp. 257–276, doe: 10.1016/j.geotexmem.2017.12.004
SHIMODA K., ISHIKAWA M, NISHIOKA K., KENTARO M. (2019).
Countermeasure for heavy metal contaminated soil by natural origin. Geotechnical
Engineering Magazine, Vol. 67, pp. 34–35. (In Japanese).
SU C., ZHU X., SHI X., XIE Y., CHEN M. (2020). Removal efficiency and pathways of
phosphorus from wastewater in a modified constructed rapid infiltration system. Journal
of Cleaner Production, Vol. 267, 122063, doi.org/10.1016/j.jclepro.2020.122063
TABELIN C.B., HASHIMOTO A., IGARASHI T., YONEDA T. (2014). Leaching of
boron, arsenic and selenium from sedimentary rocks: ii. pH dependence, speciation and
mechanisms of release. Science of the Total Environment, Vol. 473–474, pp. 244–253,
doe: 10.1016/j.scitotenv.2013.12.029

367
368
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

An innovative process for continuous dehydration and reuse

of sediment, sands washdown waters and contaminated soil

Mohammed BOUMAHDI 1, Sébastien DESCHUTTER 2

1. CDI Technologies, 23 avenue de la Créativité, 59650 Villeneuve d’Ascq, France.

[email protected]
2. Ixsane, 23 avenue de la Créativité, 59650 Villeneuve d’Ascq, France
[email protected]

Abstract:
The SURICATES Interreg NWE project aims to highlight circular economy projects in
territories where sediments can be recovered as a material to limit flooding and erosion
and initiate the creation of an industrial sector for the recovery of sediments in the
region.
The first obstacle to the recovery of freshly dredged sediments is the presence inside of
a large quantity of water which makes them difficult to transport as they are. Settling
lagoons are used to allow their dehydration. This proven solution has the disadvantage
of mobilizing land by occupying ground surfaces and the drying process could take
several months.
One of the objectives of the SURICATES project (2017-2021) is the design and
construction of an innovative continuous dehydration system. After having carried out a
state of the art on the relevant technologies and processes considering energy sobriety,
the robustness of the system and the level of dryness obtained, a mobile continuous
dehydration unit was designed allowing both the classification of sediments and their
drying.
The machine has been designed in such a way that it can operate as a standalone. It
consists of three containers of 40 feet each which can be either placed in a barge or
placed on the ground.
The flow rate of the demonstrator was limited to 60 m3 / hour in order to be able to
easily carry out the tests while being representative and to understand the problem of
recovery at the scale of the pilot in situ.

Keywords:
SURICATES, Fresh dredged sediment, Reuse, Dryness, Flooding, Dehydration, Coastal
erosion.

Corresponding author mail: [email protected]

1. Introduction
SURICATES aims to increase sediment reuse for erosion and flood protection. It will
provide authorities, ports and waterways managers and erosion experts with new large-
scale options for sediments reuse in NWE ports, waterways and coastlines
(SURICATES, 2017-2021).

1.1 Foreword
Climate change, erosion and flood risk increase require increased protective measures
that consume huge amounts of natural resources.
However, sediments to be dredged are a problem for European ports providing
navigation access with 200 million m3/year (80 million tons dry weight). Increased
social and environmental pressures shows the critical need for new solutions for ports to
maintain and develop activities. Currently, more than 99% of marine sediment dredged
in Europe is dumped at sea and managed as waste, with only 1% (800 000 t/year)
reused.
Sediment is not seen as a resource for flood risk and erosion by land managers, erosion
and civil engineering experts and society. The goal of this WP is to make the increase in
sediment reuse in such applications possible and to increase the willingness for such
solutions in public authorities, sectorial agencies, infrastructure and service providers,
interest groups, SMEs and enterprises and the general public (SURICATES, 2017-
2021).

1.2 Preliminary study

To address sediment natural dehydration time (around 6 months) within the project
period and civil works issues (lack of time and space), we considered developing and
testing compact machine for continuous dehydration coupled with innovative
decontamination techniques, (BOUMAHDI et al., 2020).
A state of the art was carried out on the various existing techniques of dehydration and
depollution of sediments (BOULLOSA ALLARIZ, 2018; MANCIOPPI et al., 2013).
Once the machine was built, we carried out the first tests to evaluate the functionality of
the machine as well as its performance.
The first tests carried out on the sediments have given complete satisfaction. We were
able to obtain from sediments composed of 80% water and 20% liquid solid matter, a
dry sand with a particle size between 2 mm and 48 µm at the exit of the classification
module and a fine and compact sediment with a dryness greater than 50% at the outlet
of the dehydration module
The agility of the machine's design made it possible, by inserting technological bricks,
to broaden the field of possible uses. It was thus possible to use it for the dehydration of

370
and recovery of sand washing water in quarries. We are also considering adapting it for
the treatment of polluted soil in a deposit platform.

2. Machine design
Different design solutions have been studied. The choice of the machine was motivated
by considering the following criteria (BOUMAHDI, 2019; BOUMAHDI, 2021):
- Compactness and robustness of the machine;
- Stability in the event of installation on a barge;
- Particle size classification of sediments;
- The dryness of the fine part must be greater than 50%.
The machine was designed in the spirit of a demonstrator easily adaptable according to
the materials to be treated. It is made up of 3 modules as shown on Figure 1:
- The first module ensures independent operation of the entire machine. It contains a
generator, a water reserve, toilets and a cloakroom.
- The module N°2 allows a granular classification of the material into several categories,
for example:
- greater than 10 mm,
- between 2 and 10 mm,
- between 40 µm and 2 mm
- and less than 40 µm).
Each category is transported outside the mobile unit and is immediately ready for
recovery.

Figure 1. Machine block diagram.

371
- Module N°3 allows dehydration. It treats the fine part of the material below the
threshold predefined in module N°2 (for example 40 µm).
After a flocculation process, the material goes through a drainage step followed by a
pressing operation making it possible to have at the output a material whose dryness is
greater than 50%, Figure 2.

Figure 2. Dehydration module drip table.

3. Sediment validation test

The machine was used to dehydrate fine river sediment with a particle size of 50% less
than 20 µm, Figure 3.

Figure 3. Particle size distribution of the sediment.

The results obtained demonstrated the efficiency of the process of dehydration and
separation of the granular part from the fine part. The easily reusable sand fraction was
372
first extracted by classifying it into two categories: greater than 2.5 mm and between 40
µm and 2.5 mm, Figure 4.

Figure 4. Sand outlet (left) and dehydrated sediment outlet (right).

The fine part of the material less than 40 µm has been efficiently dehydrated after
passing through the dewatering and pressing table. The sediment coming out of the
dehydration module has reached a dryness greater than 60%, Figure 4.

4. Validation test on washing waters from sands in quarry

The sand production quarries have washing units to rid the sands of unwanted fines.
These washing waters generally contain 2 to 10% solid matter. We used the machine in
a circular economy logic to add value to this washing water. The objective was twofold,
to recover the water to reinject it into the process and to recover the residual solid
material in reusable products, see installation on Figure 5.

Figure 5. Installation of the three machine modules in the quarry.

373
The quarry object of the intervention is located in Vendee (France) and produces gravel
and sand. The dry matter contained in the washing water is less than 5%. The machine
has been adapted to the specificities of sand washing units.
The trials took place over a period of 2 weeks. The classification made it possible to
recover sands with a particle size greater than 40 µm. The dehydration module took care
of the rest.
The tests were conclusive and gave very encouraging results, see Figure 6.

Figure 6. Dryness of fines after dehydration.

The fine part of the material below 40 µm was efficiently dehydrated after passing
through the dehydration unit. The exiting product represented an average dryness
evaluated at 50%. The filtrate at the outlet of the dehydration module reaches levels of 5
mg / l of suspended matter, Figure 7.

374
 Figure 7. Quality of the filtrate.

5. Valorisation test for polluted soils

After the first validation tests of the machine, we realized that the field of possible uses
of the machine is very wide. We could tackle the valorisation and recovery of deposits
in stock. Several activities, in the absence of a recovery channel, store their waste while
waiting for better days.
These land deposits mobilize land and give a negative image with regard to respect for
the environment.
The machine could process and recycle these deposits and even partially reduce their
pollution.
Several interventions are in the planning phase concerning the treatment and
valorisation of sediments and soils from several types of deposit:
- sediment land deposits,
- settling lagoons,
- polluted soil storage platforms,
- pits for settled sludge.

6. Conclusion
The demonstrator developed within the framework of the SURICATES Interreg NWE
project has fully fulfilled its function of treating sediments. It allows the sediments to be
classified and dehydrated quickly and continuously, thus facilitating their recovery. It
allows treatment beyond the issue of sediment recovery by broadening the concept to
the recovery of polluted soils, sand-washing water, sludge…

7. References
BOULLOSA ALLARIZ B. (2018). Déshydratation naturelle et mécanisée de
sédiments. Étude des processus mis en jeu et applications. Thèse de doctorat, Université
de Caen Normandie, 407 p.

375
BOUMAHDI M. (2019). Etude et conception d’une machine de déshydratation des
sédiments, Rapport technique Ixsane.
BOUMAHDI M., et al. (2020). Livrable Work Package 2 WP2, Projet SURICATES
Interreg NWE.
BOUMAHDI M. (2021). Projet européen SURICATES: procédé expérimental de
déshydratation accélérée des sédiments, Webinaire 19 janvier 2021, Valorisation des
sediments de la Rance, EPTB Rance Frémur & Baie de Beaussais, diaporama 16p.
MANCIOPPI L., DHERVILLY P., LEVACHER D. (2013). Process Nemeau,
WODCON XX, Congress and Exhibition: The Art of Dredging, Square Meeting Centre
Brussels, Belgium, June 3-7, 2013, version CD, 7 p.
SURICATES - Sediment Uses as Resources In Circular And Territorial EconomieS
(2017-2021). Interreg North West Europe SURICATES project,
https://www.linkedin.com/company/suricates-project/

376
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Acid buffering and arsenic leaching behaviours of excavated

acid rock treated by the MgO based immobilization material

Tsutomu TAKATA 1, Xun DU 1, Toru INUI 2, Sho OGATA 2, Hirotoshi MORI 3

1. Graduate School of Engineering, Osaka University,

2-1, Yamadaoka, Suita, Osaka, 565-0871, Japan.
[email protected], [email protected]
2. Department of Civil Engineering, Osaka University,
2-1, Yamadaoka, Suita, Osaka, 565-0871, Japan.
[email protected], [email protected]
3. Department of Civil and Environmental Engineering, Yamaguchi University
2-16-1, Tokiwadai, Ube, Yamaguchi, 755-8611, Japan.
[email protected]

Abstract:
Soils and rocks often contain non-anthropogenic heavy metals/metalloids such as lead
and arsenic due to their geogenic histories. When soils and rocks contain sulphide
minerals, leaching of these toxic metals is triggered by generation of acidic drainage
resulting from the oxidation of sulphide minerals associated with excavation processes.
Treatment with the immobilizing additives is one of the effective remediation
technologies for mitigating acidic leachate with unacceptable concentrations of heavy
metal and metalloids. In this study, long-term tank leaching tests were conducted for the
acid rock samples containing sulphide minerals with and without the immobilization
treatment by the magnesium oxide (MgO) based material to assess acid buffering
performance and arsenic leaching behaviour. The rock samples with three different
particle sizes (0.85-2 mm, 2-19 mm and 19-37.5 mm) were tested to evaluate the
particle size effects on the pH values and arsenic leaching concentrations. Among the
raw samples with different gradings, dissolution of iron sulphide by the oxidization was
most effectively limited in the 2-19 mm sample, since the degree of oxidization was
influenced by the combined effects of the supecific surface as well as how much fresh
surface was exposured by crushing, which determines the leaching of Ca and other
buffering ion. Initial leaching of arsenic with high concentrations as well as acid rock
drainage was successfully limited by the immobilization treatment. After long-term
immersion, arsenic leaching concentrations were increased by gradual surface oxidation
and detachment of the immobilization agent from the samples.

Keywords: Acid drainage, Stabilizing agent, Immobilization, Arsenic leaching.

Corresponding author mail: [email protected]

1. Introduction
Excavated soils and rocks generated from construction works are sometimes subjected
to leaching of toxic heavy metals (e.g. lead, cadmium) and/or metalloids (e.g. arsenic,
selenium) with unacceptable concentration due to their geogenic histories (e.g.
HATTORI et al., 2003, OKUMURA et al., 2007). Leaching of these metals/metalloids
as well as acid drainage is triggered by the oxidization of sulphide minerals contained in
the soils and rocks. Various remedial options including containment in the embankment,
immobilization treatment and installation of sorption layer have been employed
(MLITT, 2010). Among them, immobilization treatment mixes soils or crashed rocks
with the solidification or stabilization agents in fine graded or liquid forms to limit the
leaching of the metals/metalloids. Its long-term performance under field conditions is
not fully verified and has been a major concern. Particularly, applicability of the
immobilization treatment using the fine-graded immobilization material is questionable
to poor-graded rocks, since loss and/or erosion of the immobilizing material associate
with infiltration and immersion are concerned. Conversely, particle size control into
smaller sizes may promote the leaching of toxic metals/metalloids associated with
promoting oxidation of sulphide minerals (INUI et al., 2010).
This manuscript addresses the results of long-term tank leaching tests, which were
conducted for different grading acid rock samples containing sulphide minerals with
and without the immobilization treatment by the magnesium oxide (MgO) based
immobilization material (e.g. OYAMA, 2020). The main objective is to assess the grain
size effects on the oxidation and leaching behaviours of arsenic and the applicability of
the immobilization treatment.

2. Materials and methods

2.1 Materials
The rock sample used was rhyolitic tuff, collected in a mountainous area, Japan. It can
be considered that arsenic contained is derived from geogenic resources. The sample
was stored with its natural water content in a fully-sealed condition. Table 1 shows
chemical properties of the rock sample, including the chemical composition which were
determined by the X-ray fluorescence analysis and Japanese regulatory leaching tests
for soil quality.
MgO based powder-like material was used as the immobilization agent. Its average
particle diameter was about 15 m. Titration curve against the addition of sulfuric acid
solution is shown in Figure 1, to determine the acid buffering potential. Initially, pH
values were gradually decreased from 10.5 to 9.5, and when the additive amount of acid
reached 40 mmol/g-sample, pH valued dropped rapidly. The MgO-based material has
an acid buffering capacity of 40 mmol/g.

378
Table 1. Chemical properties of the rock sample.
Chemical composition
Fe2O3 49.4
SiO2 22.7
Al2O3 10.5
SO3 3.7
ZnO 3.2
K2 O 2.5
As2O3 2.4
Japanese standard leaching test for soil quality*
pH 4.10 4.10
Electrical conductivity (mS/m) 52.3 58.8
Arsenic concentration (mg/L) 0.413 0.329
*Tested in duplicate

Figure 1. Titration curve of the MgO based immobilization material.

Figure 2. Schematic view of the tank leaching test.

379
 Photo 1. Crashed rock samples, (left: < 2 mm, center: 2-19 mm, right: 19-37.5 mm).

2.2 Sample preparation

The rock sample was hit manually by a non-metallic hammer until the maximum grain
size was less than 37.5 mm. Thereafter, crushed rock was sieved and classified into
three fractions; 0.85 to 2 mm, 2 to 19 mm, and 19-37.5 mm in grain size (Photo 1). To
prepare the rock-immobilizing agent mixture, MgO powder was mixed with the rock
samples in its natural water content by 5% in dry weight basis.

2.3 Tank leaching test

Tank leaching tests were carried out to determine the rate of arsenic release associated
with the oxidization of sulphide minerals (Figure 2). Normal tank leaching and
intermittent tank leaching (i.e. leaching with interspersed periods of drying process)
were carried out on both raw and treated rock samples. The schedule of the intermittent
test consisted of cycles of leaching periods followed by the exposure to the oven-drying
condition at 110 ± 3˚C for 21 hours and cooling to the room temperature for 3 hours to
accelerate the oxidization of sulphide minerals. The intermittent tank leaching was
conducted only for 2 to 19 mm samples.
During the leaching intervals for both tank leaching tests, 400 g of the sample was
immersed in the solvent (ion exchanged water), as shown in Figure 2, with the LS
(liquid to solid) ratio of 5 mL/g (2,000 mL). The solvent was fully replaced periodically
(18 times) to reach the cumulative immersion period of 7200 hours. In other words, the
cumulative volume of water contacting with the sample was 95 mL/g.
The leachates obtained in both tank leaching tests were filtered through a 0.45 m
opening polypropylene membrane. pH and electric conductivity of the leachate were
measured with the probes. As, Fe, Ca, Mg and Al concentrations were analysed using
the microwave plasma atomic emission spectrometer (Agilent 4200 MP-AES).

380
3. Results and discussion

3.1 Raw rock sample

Figure 3(a) shows the pH profiles with the immersion period for raw samples with
different grading and subjected to the accelerated oxidization. pH values of the 0.85-2
mm and 2-19 mm samples were relatilely higher than 19-37.5 mm sample. This is
because the leaching of Ca was promoted from fresh surface of the sample by crushing
resulted in acid buffering of the leachate. Indeed, the higher concentrations of Ca were
observed in the finer samples, particularly immediately after starting the leaching test.
By the accelerated oxidization, pH values of the 2-19 mm sample were lower by
approximately 1 than the normal sample.
Figure 3(b) shows the cumulative mass of As released from the raw samples with the
immersion period. The accelerated oxidation promoted the As leaching associated with
the oxidization of sulphide minerals. For raw samples, 0.85-2 mm sample with its larger
specific surface and 19-37.5 mm sample with less fresh surface were sujected to the
larger leaching amount of As by about 2.5 times after 7200 hours of immersion than 2-
19 mm sample.

(a) (b)
Figures 3. Tank leaching test results for raw samples: (a) pH, (b) Arsenic release.
Figure 4(a) shows the Fe concentration profiles with the immersion period for raw
samples. Fe concentration can be considerd as an index of degree of oxidization of iron
sulphide minerals and generation of acid drainage (see Figure 4(b)). Profiles of Fe
concentrations were significantly influenced by the sample grading. Fe concentrations
gradually increased for 0.85-2 mm sample, except the initial wash-off. For 19-37.5 mm
sample, high Fe concentrations were maintained but slightly decreased after 1000 hours
immersion, and, for 2-19 mm sample, Fe concentrations were gradually decreased and
after 1000 hours, the concentrations were lower than 0.1 mg/L. This profile indicates
that decompositon of iron sulphide was limited on 2-19 mm sample after 1000 hours,
probably because the oxidization was decelerated by the combined effects of its

381
relatively larger supecific surface and generation of fresh surface by crushing, which
promotes the leaching of Ca and other buffering ions

(a) (b)
Figures 4. Tank leaching test results for raw samples:
(a) Fe release, (b) Fe concentrations vs. pH values.

3.2 Treated samples

Figure 5(a) shows the pH profiles with the immersion period for treated samples with
different grading and subjected to the accelerated oxidization. In all the cases, pH
profiles were consistent. Owing to the acid buffering potenital of the MgO based
immobiliation material, pH vaules were stabilizing at 10.2 to 10.5 until 1000 hours, and
thereafter pH vaules were gradually decreased but higher than 9.5 in all the cases. These
facts support that the MaO based immobilization material could act as the buffering
agent for acid rock drainage even in the accelerated oxidization case.
Figure 5(b) shows the cumulative mass of As released from the treated samples with the
immersion period. Until 1000 hours, which was equivalent to the cumulative L/S of 50,
As leaching was sucessfully limited. However, As leaching concentrations rapidly
increased after 1000 hours, particularly for the 0.85-2 mm sample and 2-19 mm sample
with the accelerated oxidization. Meanwhile, the 2-19 mm sample without the
accelerated oxidization exhibited the smallest leaching amount of As, similar to the raw
sample.
For the 2-19 mm samples, the cumulative mass of released As after 7200 hours reached
4.4×10-4 mg/g (raw), 9.9×10-3 mg/g (raw + accelerated oxidisation), 4.2×10-4 mg/g
(treated), and 1.9×10-3 mg/kg (treated+accelerated oxidization). It can be considered
that two processes caused unsuccessful immobilization of As after 1000 hours
immersion. The one is dissolution of hydration products associated with acid buffering,
resulting in the release of As encapsulated in Mg hydration products. The other is loss
of the MgO based immobilization material during a long-term immersion, as shown in
Photo 2. From this observation, applicability of the immobilization treatment using the

382
fine-graded immobilization material is questionable to poor-graded and high-permeable
soils and rocks, since loss and/or erosion of the immobilizing material associate with
infiltration and immersion are concerned.

Figures 5. Tank leaching test results for treated samples: (a) pH, (b) Arsenic release

Photo 2. Appearances of the treated samples during the tank leaching test
(19-37.5 mm sample, Left: 264 hours later, Right: 1536 hours later).

4. Conclusions
This study discusses leaching behaviours of non-anthropogenic arsenic contained in the
acid rock. Long-term tank leaching tests were conducted for both raw and treated rock
samples with the magnesium oxide (MgO) based immobilization material, to assess the
grain size effects on the oxidation and leaching behaviours of arsenic and the
applicability of the immobilization treatment.
Experimental results indicate that the finer crashed samples with a larger specific
surface was not initially acidified due to the leaching of buffering ions such as Ca, and
then oxidized at a higher rate. Conversely, the coarser samples with less fresh surface
exhibited the lowest pH values. For the treated samples, the MgO based material acted
as the acid buffering agent owing to its high acid buffering capacity. However, long-

383
term immobilization performance was concerned, due to dissolution of hydration
products associated with acid buffering and loss of the MgO based immobilization
material from poor-graded samples.

5. Acknowledgement
This research was supported by Mr. Seiichi OBANA and Mr. Fumio OKADA (Ube
Material Industries, Ltd.) and funded by JSPS scientific research B (19H02236).

6. References
HATTORI S., OHTA T., KIYA H. (2003). Engineering geological study on exudation
of acid water from rock mucks - Evaluation methods of rocks at the Hakkouda Tunnel
near mine area, Jour. Japan Soc. Eng. Geol., 43 (6), pp. 359-371.
INUI T., KATSUMI T., KATAYAMA M. KAMON M. (2010). Effects of friability and
grain size on the leaching of heavy metals in excavated rock materials, Environmental
Geotechnics for Sustainable Development, M. Datta et al. (eds.), Tata McGraw Hill,
New Delhi, pp. 730-733.
MLITT - MINISTRY OF LAND, INFRASTRUCTURE, TRANSPORT AND
TOURISM, JAPAN. (2010). Technical Manual on the Countermeasures against Soils
and Rocks Containing Natural-Derived Heavy Metals in Construction Works (Draft),
90 p.
OKUMURA K., SAKURAI K., NAKAMURA N., MORIMOTO Y. (2007).
Environmental impacts of naturally occurring heavy metals and countermeasures,
Journal of Geology, 116 (6), pp. 892-905.
OYAMA S. (2020). Solidification/stabilization treatment with MgO-based material for
mercury-contaminated soil, Japanese Geotechnical Journal, 15(3), pp. 465-470,
https://doi.org/10.3208/jgs.15.465

384
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Analysis of soil subsidence from the loading cycle using a

concrete mattress with woven bamboo reinforcement
Lusmeilia AFRIANI 1, Daniel LEVACHER 2, Ryzal PERDANA 3

1. Universitas Lampung, Faculty of Engineering, Jl. Prof. Dr. Soemantri Brojonegoro

No. 1, Bandar Lampung, Indonesia.
[email protected]
2. Université de Caen Normandie, Faculté des Sciences, UMR 6143 CNRS - M2C,
Esplanade de la Paix, 14032 Caen, France.
[email protected]
3. Universitas Lampung, Faculty of Teacher Training and Education, Jl. Prof. Dr.
Soemantri Brojonegoro No. 1, Bandar Lampung, Indonesia.
[email protected]

Abstract:
The water content of soft clay soils and organic clay soils is very large, their weight is
greater than the soil grains, so that when the soil is given a load, it will experience a
problem of land subsidence which causes the pore water pressure to rise, exit and
experience a decrease in consolidation. Therefore, it is necessary to conduct soil
subsidence research in the laboratory using a test box and a bamboo concrete mat as a
means of lowering the soil. The amount of land subsidence is detected by data
acquisition.
This research was conducted on a box-shaped test instrument measuring 80 x 90 x 100
cm3. Soil is put into the box and soil subsidence test is carried out with a concrete slab
using bamboo mats, after which a pressure of 0.2 kg/cm2 is applied; 0.3 kg/cm2; 0.4
kg/cm2 and 0.5 kg/cm2. This test is carried out until it is no longer reduced. The results
of the clay soil subsidence test using the test box modeling and bamboo concrete mats
show that the soft clay soil is experiencing the reduction rate (Cv) is faster than organic
clay soils. The amount of reduction (Cc) in soft clay soils is smaller than that of organic
clay soils. Soft clay soils also experience volume changes (Mv) which are faster than
organic clay soils. The compressibility (Av) of soft clay soils is also faster than organic
clay soils. The amount of total soil subsidence in a period of 20 years in soft clay soils is
smaller than in organic clay soils.

Keywords:
Soft clay soil, Soil organic clays, Consolidation, Concrete mattress.

Corresponding author mail: [email protected]

1. Introduction
Clay has a weak grain structure. One of the causes is the high-water content and the fine
grain size. Therefore, if the soft soil is used as a construction foundation, the soil must
be compacted. If the soil is not compressed, there will be a settlement. In other cases,
the soil supporting the foundation has been consolidated before loading, but there will
still be the settlement. It is probably because of excessive load factors or unstable soil
conditions.
It is recommended, before carrying out any construction work, compaction of the soil
and calculation of its settlement should be carried out. The subsidence occurs in loam
soils and organic or semi-organic soils, significantly reducing their magnitude compared
to loam soils. Therefore, this study will use soft clay soil as an experimental sample to
obtain the reduction coefficient. Compaction will occur because the weight of the
structural load above the soil will compact and destroy the soil grains in each layer of
soil and push water out of the soil layer due to the shrinkage of soil pores (TERZAGHI
& PECK, 1987). The structural weight here is assumed to be the concrete mattress and
the working floors and wood reinforcement. This wood is used to reduce soil problems.
The research will use land from Belimbing Sari Village, Jabung Regency, East
Lampung, Indonesia.
If the soil carries a load above it, the soil will increase vertical stress, horizontal stress,
shear stress, and pore water pressure. The result of this increase in tension will cause a
decrease in subgrade elevation. This loading causes particle deformation, particle
relocation, and the release of air from the soil pores will reduce the volume of the soil. It
can cause soil displacement if it is on a slope, (DAY, 2006). Soil subsidence, i.e., loss of
soil thickness, is for either clay or organic soil after drainage. It is a widespread
phenomenon in Southeast Asia (WÖSTEN et al., 1997; HOOIJER et al., 2012) and
New Zealand (PRONGER et al., 2014). In the United States, the best-known cases of
subsidence are the Sacramento-San Joaquin Delta in California and the Florida
Everglades (GALLOWAY et al., 1999). Several environmental factors such as type of
peat, decomposition rate, density and thickness of peat, climate, water table depth,
temperature, and land use history affect subsidence (ARMENTANO, 1980; WÖSTEN
et al., 1997). Water table depth is considered the dominant factor controlling soil
subsidence because it regulates peat growth and formation (CLYMO, 1984). In the
EAA, studies with subsidence lines established in 1913 indicate that rapid subsidence
rates (ca. 9 cm yr−1) followed the initial drainage of the Everglades (STEPHENS,
1956). After this rapid rate, subsidence remained constant at 3 cm yr−1, dominated by
the oxidation process (STEPHENS, 1956; STEPHENS & SPEIR, 1970). During the
past few decades, subsidence rates have decreased to 1.45 cm yr−1 (SHIH et al., 1998).
Possible causes for this decrease in subsidence rates include increases in the water table,

386
possibly due to best management practices, in the mineral content of the soils, and in
recalcitrant forms of organic carbon (C) (SHIH et al., 1998; GALLOWAY et al., 1999).
Peat soils will rapidly decline in sustainable soil surface due to densification (shrinkage
and consolidation) and oxidation of the peat substrate. Shrinkage occurs when water
leaks from the soil layer. This settlement rate is consistent with the literature synthesis
of moderate zone subsidence rates reported for the same period since drainage occurred.
This settlement rate is consistent with the literature synthesis of reasonable zone
subsidence rates reported for the same period since drainage occurred (STEPHENS et
al., 1984; PRONGER et al., 2014). Shrinkage, consolidation, and oxidation have been
the dominant processes controlling subsidence in the Everglades Agricultural Area
(EAA). STEPHENS (1956) and STEPHENS et al. (1984) showed that subsidence
accelerated after installation of water pumps that quickly lowered the water table in the
EAA. Soil subsidence in the EAA is currently dominated by oxidation (TATE, 1980).
Before drainage, EAA soils had limited oxygen availability due to their saturated
condition and low oxygen diffusion rates in water (REDDY & DELAUNE, 2008).
Following drainage, oxygen became available to aerobic bacteria that decomposes
organic matter more rapidly, resulting in subsidence due to oxidation (STEPHENS &
JOHNSON, 1951). VOLK (1973) found that carbon dioxide (CO2) efflux from EAA
soils accounted for 58% – 73% of the average soil subsidence estimated by field studies
with subsidence lines, while STEPHENS & SPEIR (1970) found that oxidation losses
accounted for 75% of field subsidence estimates. Higher water tables result in lower
subsidence rates (STEPHENS & SPEIR, 1970; SHIH et al., 1998). Furthermore,
lowering the water table depth by a factor of two in the EAA doubles the subsidence
rate (SNYDER et al., 1978). Increases in bulk density due to compaction and shrinkage
increased subsidence rates, particularly during the initial years after drainage. The
estimated bulk density of EAA peats before drainage (0.1g/cm3), (AICH et al., 2013)
doubled 10 years after drainage (STEPHENS & JOHNSON, 1951). Studies in the EAA
show that oxidation of these soils is affected by soil temperature with a Q10 of 2
(KNIPLING et al., 1970; VOLK, 1973), where Q10 is the factorial increase in
decomposition for every 10°C increase in temperature.
Consolidation will not only occur in the soil layer, but the pile foundation will also
undergo settlement. Some analysis to see the decline in the pole. One of them is the
numerical method. They also provide a relatively fast way of carrying out a parametric
study of the effects of soil characteristics. This article will also examine the numerical
settlement of the pile. They also provide a relatively fast way to perform parametric
analysis of the effects of pile and soil characteristics and prepare various solutions,
which design purposes can use.

387
Damages resulting from settlement can range from total structural failure to minor
damage. Damage due to settlement in buildings causes cracks in the structure and
frequent cracks in partitions such as windows and doors, which do not stand in their
original position. It can be overcome if there are no more settings.
Settling can create problems in loam soils, so a large drop in numbers will occur when
the soil is loaded. The issue of subsidence will continue after the construction period has
been completed, often for several years. It happens because there is an immediate or
elastic settlement of soil in which there is no water flow. It was at that time when clay
soil deforms to accommodate the imposed shear stress. Based on the stress history of
saturated cohesive soils, they are considered either under consolidated, ordinarily
consolidated, or over consolidated. The over consolidation ratio (OCR) is used to
describe the stress history of cohesive soil, for under consolidated soil OCR <1, (DAY,
2006).
The primary use of this research is to design a large-scale consolidation tool. It is a
square box. This innovative tool is made for the benefit of consolidated research with a
field-scale approach. This tool is called the modification consolidation tool. Inside this
tool will be filled with soil samples, above which are given a load. This tool will be
linked by data acquisition during the research process, which will record all the research
processes. The parameters recorded are vertical subsidence due to concentrated load and
even loading. This tool made it easier for planners from the government and the private
sector to determine the magnitude of deterioration in the soil type of clay and organic
clays when given a load. In building planning, a consolidation test is required. The
consolidation tool in the laboratory is small in size. Later, it will correlate the laboratory
scale consolidation value and the approach scale in the field. So that when the research
uses a small tool, the researcher can estimate and correct the actual consolidation value
in the area.

2. Materials and method

This research uses one type of soil, namely soft clay soil. The sampling location is 2 - 3
hours from Bandar Lampung, Lampung Province, namely. Belimbing Sari Village,
Jabung District, East Lampung Regency, Indonesia. The test is carried out in 3 stages.
The first is physical properties testing. Second, the consolidation test using standard
laboratory equipment. The third stage uses the design of a large consolidation test
instrument. This tool is made to resemble the original conditions in the field. The testing
phase was carried out at the Laboratory of Soil Mechanics, Faculty of Engineering,
University of Lampung. Figure 1 shows the location for taking soil samples.
Previously conducting experiments in the third stage, researchers had made a new
consolidation tool. The size is quite large, 90 cm long x 80 cm wide x 100 cm high. This

388
tool is made of 0.5 cm and 0.1 cm thick steel plate, equipped with glass that is 1.2 cm
thick and 100 cm high. This new consolidation device is added with a water tap which
is located at the bottom. It has the function of removing water when the experiment is
carried out. Testing procedures of physical properties in accordance with
standardization by the American Society for Testing Materials (ASTM), (GOGOT,
2011; HEAD, 1994; BOWLES, 1991). As an alternative to bamboo matting, the
researcher can also use other fibers such as raw or processed coconut fiber, such as local
coconut fiber research showing natural fibers such as coconut fiber mixed with mortar
for its construction (BUI et al., 2020).

Figure 1. Location for taking soil samples.

Figure 2 is a plan of the working process, and vertical loads are placed in the middle of
the bamboo concrete slab. At the same time, the measuring instrument for vertical
displacement (strain gauge/sensor) is installed on four sides (ISWAN et al., 2016,
NAWAWI et al., 2017). They have done the same tests but used the gelam wood
reinforcement. Gelam wood is the name of a type of wood that is widely available in
Indonesia. Its nature is that the longer it is immersed in water, the wood is strong and
not rotten.

389
This study used a modified consolidation test instrument with a concrete mat reinforced
by 80 mm thick bamboo wire. The quality of the concrete used was K-225, see Figure 2.
As a validation experiment, a concrete mat with wire mesh reinforcement was also
carried out. The test procedure is as follows :
1- The physical and mechanical properties of the soil were tested before and after the
consolidation test. The soil samples were tested for their property index.
2- Soil is put into the consolidation test instrument, then saturation and soaking for 24
hours, soaking so that the soil is homogeneous.
3- After soaking, the water content is tested. The water content used is the same as the
water content in the field. All tests used the same moisture content.
4- Consolidation testing using standard laboratory consolidation tools is also used as
data validation, (HEAD, 1994).
5- The placement of vertical loads, sensors is as in the design Figure 2.

Figure 2. Design of test equipment size.

From Figures 3 and 4, a concrete mattress is placed and then on the mattress the load is
placed with a gradual load, namely: 0.2 kg/cm2; 0.3 kg/cm2; 0.4 kg/cm2; 0.5 kg/cm2.
Take measurements of the decline in bamboo mats with a proximity sensor measuring
device.
Record settlement yields and makes comparisons of subsidence between soft clay soils
and organic loam soils. The data obtained from the laboratory results in the laboratory
are processed according to the data classification by using valid equations and formulas.
The results of the data processing are described in tables and graphs. Then analyzed and
compared the results with previous research.

390
 Figure 3. Size and shape of a concrete mat with bamboo / wire mesh reinforcement.

Figure 4. a). The bamboo mat is ready to be cast; b) Placement of strain gauge and
pressure / hydraulic jack, (ISWAN et al., 2016; NAWAWI et al 2017).

3. Results

3.1. Results and discussion of physical soil test

Testing the physical properties of soil is a consideration in planning and implementing a
construction. The results of this test are also used in papers from ISWAN et al., (2016)
and NAWAWI et al., (2017). The results obtained from the physical properties test of
the water content in the Belimbing Sari Village, Jabung Timur District, East Lampung,
amounting to 64.48%. The water content contained in the Belimbing Sari soil is
relatively high. The water content is greater than the soil grains. The level of water
content in the soil dramatically affects other soil properties (SOSRODARSONO &
TAKEDA, 2016). Based on the Unified Soil Classification System (USCS) soil
classification system, the percentage value of soil grains that pass the analysis number
200 is 82.77%. The value of soil grain content, which gives a sieve n° 200, is superior to
50%, according to the USCS soil classification table, generally categorized as fine-
grained (loam) soils. In this case, the soil samples were taken from Belimbing Sari
Jabung Subdistrict, East Lampung Regency, are fine-grained soil, and the water content
is greater than the soil grains. Apart from being fine-grained, the properties of the two
soils are different. The soil from Benteng Sari village can be semi-organic soil because

391
it still contains soil grains even though it is very fine. Soil conditions are taken in a
location near the rice fields. Visually, the soil contains a lot of crude fiber derived from
plants. Previous research has also been carried out in the exact location (SETYANTO,
1992).
While the weight of the soil volume of Belimbing Sari is 1.56 gr/cm3, the dry weight of
the soil will decrease with the addition of water. It will reduce the concentration of solid
soil particles per volume. The following things are obtained from the specific gravity
test of Belimbing Sari Village 2.54. Figure 5 shows the type of soil as a sample of the
consolidation study. DAS (1995) states that soil has a GS value ≤ of 2.68 - 2.75 is soft
clay soil.

Figure 5. Soil types at the sampling location in the Belimbing Sari area.

The water content in the two soil samples shows that the soil will expand, and the
effective stress of the soil will decrease along with the increase in pore water tension.
Likewise, when there is a process of shrinkage on the ground. Soil that loses water
suddenly will experience shrinkage in pore volume due to water loss. This will cause
the soil to experience large shrinkage. To improve the properties of expansive soil,
generally, expansive soil is stabilized with materials that are by the properties of clay to
become better and qualify as a construction material. Clay soils are mainly composed of
microscopic particles in flat plates and are particles of mica and fine-grained soil
minerals or colloidal grains with the size of soil particles (DAS, 1995). The original soil
image in the Belimbing Sari area is given in Figure 5.

3.2 Consolidated test results

Soil shrinkage conditions will cause soil subsidence when loaded. It can occur, for
example, on the road; thus, there will be soil development. This research looks for a
392
solution by placing a plate on the ground with a concrete plate but with bamboo
reinforcement and validated with steel fiber reinforcement. After the test is complete, a
graph is made with the time root method used to determine Cv by describing the
consolidation test results on the time root relationship to the decline. The characteristic
of this time root is to assess the degree of consolidation U = 90%, and T90. The graph's
results are used to determine the values of Cv, Cc, Mv and Av, see Table 1. This study
obtained the following products:

Figure 6. The relationship between time and settlement with 6 trials using concrete
mats with bamboo and wire mesh (bendrat) reinforcement with variations in pressure.

The graph in Figure 6 shows that the different descent stages for each are different.
From the chart above, starting at 0.2 kg/cm2; 0.3 kg/cm2; 0.4 kg/cm2; 0.5 kg/cm2 shows
a steeper decline in the initial minutes at all loading stages, this is because in the early
minutes there is an immediate decrease. After all, the soil is not too dense, and there are
still many pore cavities, and the water content is still high so that the time to accept the
initial loading of the soil is significantly reduced. After that, there is a gradual decline
(consolidation completion). After that, the land consolidation occurs gradually
(consolidation completion). This condition applies to all types of soil and concrete
mattress.
At ∆δ1, the initial drop on the graph looks steep because this is an immediate settlement
caused by the initial loading of the specimen. For ∆δ-initial on soft clay soil, there is an
393
average decrease of 46 - 55%. In the middle, there has been a consolidation settlement.
This process depends on the time it occurs in saturated fine-grained soils and has a
small permeability coefficient. This condition causes the subsidence to appear not so
steeply, but there is still a gradual decline because the soil has begun to harden and
compress. For ∆δ2 or ∆δ, the middle subsidence in soft clay soil is 75% - 86%. In the
final ∆δ, you can see the graph starting to slope. This happens because in this part, the
soil has undergone secondary compression, at a loading of 0.2 kg/cm2; 0.3 kg/cm2; 0.4
kg/cm2 and 0.5 kg/cm2. The soil has not decreased anymore, meaning that the soil has
consolidated. For ∆δ-end subsidence in soft clay soil is 2%- 4% from the initial loading.

Table 1. Cv, Cc, Av and Mv values.

Mattress Bamboo Mattress Bendrat
Testing Testing with standard Testing using a large Testing with standard Testing using a large
laboratory equipment* box-shaped tool laboratory equipment box-shaped tool
Cv (cm2/det) 0,00035 0,0019 0.00044 0.00039
Cc 0,7997 0,8018 0.8300 0.8500
Av (x10-4 m2/kN) 0,0861 0,0857 0.0812 0.0824
Mv (x10-4 m2/kN) 0,0746 0,0745 0.2965 0.3009
*: Head, K.H., 1994.

4. Conclusion
Consolidation experiments conducted in the laboratory have used 2 testing instruments
(see Table 1). The consolidation tool in the form of a box measuring 900 x 800 x 1000
mm3 is an innovation of a large consolidation device to resemble field conditions and a
standard consolidation test tool which is often used for practicum and laboratory scale
testing. The samples chosen were soft clay soil samples from Belimbing Sari, East
Lampung. The location was chosen because the area contains a lot of soft clay and
organic soils.
From the study results, the values of Mv, Cv, Cc, and Av in soft clay soil samples using
all consolidation tools get the importance of Mv, Cv, Cc, and Av almost the same from all
consolidation tests. At the beginning of loading, the soil immediately receives a big load
and undergoes a significant consolidation so that the soil becomes denser.

5. References
AICH S., MCVOY C.W., DRESCHEL T.W., SANTAMARIA F. (2013). Estimating
soil subsidence and carbon loss in the Everglades agricultural area, Florida using
geospatial techniques. Agriculture, Ecosystems & Environment, Volume 171 (2013),
pp. 124-133, https://doi.org/10.1016/j.agee.2013.03.017

394
ARMENTANO, T.V. (1980). Drainage of organic soils as a factor in the world carbon
cycle. BioScience 30 (12), pp. 825–830, https://doi.org/10.2307/1308375
BOWLES J. E. (1991). Sifat-sifat fisik dan geoteknis tanah. Penerbit Erlangga, Jakarta.
BUI H., SEBAIBI N., BOUTOUIL M., LEVACHER D. (2020). Determination and
review of physical and mechanical properties of raw and treated coconut fibers for their
recycling in construction materials, Fibers, 8 (6), 37; doi:10.3390/fib8060037
CLYMO, R.S. (1984). The limits to peat bog growth. Philos. Trans. R. Soc. Lond., B,
Biol. Sci. 303 (1117), pp. 605–654, doi.org/10.1098/rstb.1984.0002
DAS M. B. (1995). Mekanika Tanah (Prinsip – Prinsip Rekayasa Geoteknis), Jilid I,
Erlangga, Jakarta, 283p.
DAY R.W. (2006). Foundation engineering handbook – Design and construction with
2006 international building code, McGraw-Hill editors, 930p., ISBN 0-07-144769-5.
GALLOWAY D.L., JONES D.R., INGEBRITSEN S.E. (1999). Land Subsidence in the
United States. US Geological Survey 1182, 177p., https://doi.org/10.3133/cir1182
GOGOT S. B. (2011). Laboratory soil testing, explanations and guidelines, Graha Ilmu
Yogyakarta Press, 154p.
HEAD K.H. (1994). Manual of soil laboratory testing: permeability, shear strength and
compressibility test, Vol. 2, Second Edition, 436 p., ISBN: 0-470-23362-1
HOOIJER A., PAGE S., JAUHIAINEN J., LEE W.A., LU X.X., IDRIS A., ANSHARI
G. (2012). Subsidence and carbon loss in drained tropical peatlands. Biogeosciences 9
(3), pp. 1053–1071, https://doi.org/10.5194/bg-9-1053-2012
ISWAN I., AFRIANI L., ADHA I., IKRATUL H. (2016). Experiments and analysis of
the behavior of soft clay and organic clays in terms of the loading cycle using bamboo
concrete mattress, Proceeding SN-SMIAP, National Seminar on Science, Mathematics,
Informatics and Its Applications, Vol. 4, Book 3, pp. 178-188, ISSN: 2086–2342,
http://repository.lppm.unila.ac.id/27149/.
KNIPLING E.B., SCHRODER V.N., DUNCAN W.G. (1970). Soil and Crop Science
Society of Florida. CO2 evolution from Florida organic soils Proceedings 30, pp. 320–
326.
NAWAWI N., AFRIANI L., ISWAN I. (2017). The study of derivation analysis of soft
clay and organic clay soil using bamboo concrete mat modeling, Journal of Civil
Engineering and Design (JRSDD), Vol. 5, N°3, E-ISSN:2715-0690 (Online).
PRONGER J., SCHIPPER L.A., HILL R.B., CAMPBELL D.I., MCLEOD M. (2014).
Subsidence rates of drained agricultural peatlands in New Zealand and the relationship
with time since drainage. Journal of Environment Quality 43 (4), pp. 1442-1449,
http://doi.org/10.2134/jeq2013.12.0505
REDDY K.R., DELAUNE R.D. (2008). Biogeochemistry of wetlands: Science and
applications, CRC Press, 800p., ISBN 9781566706780

395
SETYANTO (1992). Study of the analysis of the shear strength of peat soil in South
Sumatra with a triaxial experiment, Thesis of Master degree, ITB Bandung, 98p.
SHIH S.F., GLAZ B., BARNES R.E. Jr. (1998). Subsidence of organic soils in the
everglades agricultural Area during the past 19 years, Proceedings, ISSN: 0096-4522.
SNYDER G. H., BURDINE W. H., CROCKET J. R., GASCHO G., HARRISON D. S.,
KIDDER G., MISHOE J. W., MYHRE D. L., PATE F. M., SHIH S. F. (1978). Water
table management for organic soil conservation and crop production in the Florida
everglades. Florida Agricultural Experiment Station Bulletin 801, pp. 1–22, ISSN: 0096-
607X
SOSRODARSONO S., TAKEDA K. (2016). Fill type dam, Balai Pustaka Press, 385 p.
ISBN: 978-602-260-093-0
STEPHENS J.C. (1956). Subsidence of organic soils in the Florida everglades. Soil
Science Soc. Am. J. 20 (1), pp. 77-80, https://doi.org/10.2136/sssaj1956.03615995002000010019x
STEPHENS J.C., ALLEN L.H., CHEN E. (1984). Organic soil subsidence. Rev. Eng.
Geol. Vol. 6, pp.107-122, https://doi.org/10.1130/REG6-p107
STEPHENS J.C., JOHNSON L. (1951). Subsidence of organic soils in the upper
Everglades region of Florida. Soil Science Society of Florida, Proceedings 11, pp. 191–
237.
STEPHENS J.C., SPEIR W.H. (1970). Subsidence of organic soils in the USA. Publ.
int. Ass. sci. Hydrol., Symp. Tokyo, pp. 523–534.
TATE R.L. (1980). Microbial oxidation of organic matter of histosols. Advances in
Microbial Ecology. Springer, pp. 169–201.
TERZAGHI K., RALPH B.P. (1987). Mekanika tanah dalam praktek rekayasa, Jilid 1.
395p.
VOLK B.G. (1973). Everglades histosol subsidence: 1. CO2 evolution as affected by
soil type, temperature, and moisture. Proc Soil Crop Sci Soc Fla, vol. 32, pp. 132-135.
WÖSTEN J.H.M., ISMAIL A.B., VAN WIJK A.L.M. (1997). Peat subsidence and its
practical implications: a case study in Malaysia. Geoderma 78 (1–2), pp. 25–36.
https://doi.org/10.1016/S0016-7061(97)00013-X

396
ABSTRACTS
 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Permeable pavements

Philippe DHERVILLY 1, Jose OLIVEIRA 2

1. ARD Normandie Technologies, Espace Jean Monnet, 14200 Hérouville-Saint-Clair,

France.
[email protected]
2. Jardins e Afins, Av. Manuel Violas 476, San Felix de Marinha, Portugal.

Abstract:
The use of permeable block pavement has been acknowledged as one of the
promising Low Impact Development (LID) strategy to mitigate the harmful effects of
depletion of natural surfaces due to uncontrollable development in infrastructures and
buildings. Numerous studies associated with the drainage properties and long-term
performance of this traditional pavement alternative have been conducted in the
past 30 years. Nevertheless, standardized equipment and methodologies are still
limited, specifically for small-scale laboratory models. This paper suggests an
equipment that is capable of evaluating the hydraulic performance of permeable
pavement materials in a laboratory set-up by monitoring permeability and
simulating physical clogging process. Constant head permeability tests with systematic
application of fine clogging particles were conducted on three identical permeable
pavement systems composed of four stone pavers. Each test system received an
equivalent amount of eight years’ particle loading of silica sand having different size
distributions. The experiment results reveal that all models showed permeability
degradation trend similar to other literatures. An average of 77% permeability
loss related to particle accumulation was evaluated.
The use of permeable pavements has been an important tool for flood reduction and
minimization of the greenhouse effect in urban centers. This type of pavement
allows the passage of water and air through its layers and, when designed and
applied correctly, may contribute to the reduction of surface runoff during the
occurrence of peak flows.

Keywords:
Permeable pavement, Hydraulic conductivity, Constant head test, Permeability
equipment, Low impact development.

Corresponding author mail: [email protected]

 Geo-Environmental Engineering GEE2021
ESITC-UniCaen, Caen, France, May 20-21, 2021

Contribution to the assessment of landslide factors in the

Bushwira area

Ilalie MANEFOUET BERTILLE 1, Christian KALIKONE BUZERA 2, Claude

CUBAKA RUGENDABANGA 2, Célestin LUNGERE BARHADOSANYA 2

1. University of Dschang, Department of Earth science, BP 96, Dschang, Cameroon.

[email protected]
2. Département de Géologie, Université officielle de Bukavu, C. Kadutu karhale,
Campus U.O.B, Bukavu, R.D. Congo.
[email protected]

Abstract:
This work is part of the study of geohazards in the Great Lakes region. It contributes to
the understanding of the study of natural, physical and anthropogenic phenomena
leading to the triggering of landslides in the Bushwira sector in Kabare territory (South
Kivu, DRC). To do this, the mapping and diachronic analysis of land use was carried
out. This investigation began with the inventory and determination of the ongoing
processes and morphometric parameters of landslides. Twelve (12) landslides were
inventoried, including 6 rotational, 4 translational and 2 old. The diachronic analysis
carried out on the basis of observations from Google Earth and LANDSAT satellite
images revealed a significant change in land use in the study area. The amplification of
landslides in the study area is due to the presence of clay soil; steep slopes; slopes
oriented to the North-West, the South and the North; untimely deforestation and the
exploitation of construction materials. Lithology, topography, soil and hydrology are the
natural physical factors that predispose the Bushwira area to landslide hazards.

Keywords:
Landslide, Google Earth, Land use, Predisposing factors, Bushwira.

Corresponding author mail: [email protected]

399
400
 Editeur et Distributeur

D.Levacher & Centre Français du Littoral

Trésorerie/Comptabilité : 30 Avenue Félix Vincent

44700 ORVAULT – Tél 00 33 (0)6 41 05 17 19 – Email : [email protected]

ISBN 978-2-35921-022-4