THE EFFECT OF SAMPLE PREPARATION AND TESTING

PROCEDURE ON THE GEOTECHNICAL PROPERTIES OF

TROPICALLY WEATHERED RESIDUAL LATERITE SOILS OF
ETHIOPIA
Addiszemen Teklay1, Messele Haile2, Alemayehu Teferra2, and E. J. Murray3
1
School of Civil & Water Resources Engineering, Bahirdar Institute of Technology,
2
School of Civil & Environmental Engineering, Addis Ababa Institute of Technology,
3
Murray Rix Geotechnical, UK
Corresponding Author’s E-mail: [email protected]

ABSTRACT geo-morphological impact on soil formation

together with conventional classification systems
Identification and classification of the various is suggested and can result in a good indication
grades of tropically weathered residual soils for
of the engineering properties of laterites.
engineering purposes is a problem that
geotechnical engineers repeatedly face in Key Words: laterite, air-dried, as-received, oven-
Ethiopia, as well as in other tropical countries. dried, sesquioxide, mineralogy
This paper investigates the applicability of
conventional laboratory testing, classification INTRODUCTION
and evaluation systems, which were principally Urban and rural infrastructure development in
developed for temperate climate sedimentary Ethiopia is expanding at a rapid pace (21). The
soils, to the tropically weathered residual southern and western parts of Ethiopia are
laterite soils of Ethiopia. The construction of a characterized by residually formed laterites (7, 8,
new highway project from Bako to Nejo in 9, 10, 11, 12) that are widely used as
western Ethiopia, with extensive laterite soils, construction materials and have been found
forms a study site that has allowed detailed satisfactory for many earthworks applications.
inspection and sampling of soils for this For example, they form foundation materials for
research project. Testing of these laterite major roads and are used as fill materials for
deposits has revealed that they are sensitive to highway profiles and low-permeability layers in
handling and disturbance. Revisions to sample earth fill dams. The major roads constructed in
preparation and testing procedures are Western Ethiopia, such as the Nedjo-Jarso-Begi
considered necessary and are proposed in road, and the earth fill dam in Gilgel-Gibe for
utilizing these soils for civil engineering hydroelectric power generation are projects that
purposes. utilize these types of soil (11, 12). The laterites
It is concluded that the plasticity chart should be sampled and tested in this research are from a
regarded as a guidance tool only for evaluating road project in western Ethiopia.
the likely properties of laterite soils. It should Tropical weathered residual soils (TWR) are
not be used to classify them based on derived from the in-situ weathering and
conventional methods without due deference to decomposition of rocks in a tropical environment
their genesis. Unlike soils for which the and can have characteristics distinctively
Casagrande Plastcity Chart, as adopted in the different from those of transported sedimentary
United States Classification System (USCS), was soils (1, 2). Only recently have they been the
developed, laterite soils that plot well below the subject of concerted studies in the field of
A-line generally have good engineering geotechnics. Increasingly findings indicate that
properties. A combination of classification based the application of laboratory testing and
on structure, mineralogical composition, and classification methods, developed mainly for

Journal of EEA, Vol. 33, December 2015 45

 Addiszemen Teklay, et al.,

soils from temperate climates, are either partly or REVIEW OF DEFINITIONS AND SOIL
wholly inappropriate in assessing the properties PROPERTIES
and suitability of TWR laterite soils for roads,
earthworks and other purposes. TWR laterite The paper focuses on TWR laterite soils that are
soils are a specific group of TWR soils which found in the Ethiopian tropical environment and
include varying amounts of aggregates or are strongly influenced by the clay minerals
crystals of weathered mineral matter that break present. They include the halloysite subgroup
down and become progressively finer under soils and the sesquioxide influenced subgroup
continuous manipulation (1, 3, 5). This soils in Wesley‟s residual soil classification
influences the results of laboratory tests and their scheme (1, 14). The latter type of residual soils
interpretation. From an engineering viewpoint, are abundant in Ethiopia (7, 8, 11, 15). In humid
laterite soils frequently display significant tropical regions, as decomposition leaching and
property differences from those predicted from dehydration proceed during chemical
conventional laboratory test results. For weathering, a wide variety of laterite soils can be
example, despite relatively high liquid limits and produced from the parent rocks (17). Some
moderate particle size they have good common terminology used for these materials,
engineering properties. They are remarkably though some are misleading or incorrect, are red
stable, exhibit high shear strength, low clay, tropical red clay, lateritic clay and laterite,
compressibility and do not normally give rise to or the soils are referred to as reddish brown
swelling or shrinkage problems. Secondary lateritic, ferruginous tropical, ferrallitic,
properties, which may be considered somewhat ferralsols, acrisols or nitosols (16).
abnormal, are low plasticity index despite high Laterite soils are rich in sesquioxide (Fe2O3 +
liquid limit, and irreversible changes that can Al2O3) and low in bases and primary silicates
take place on drying, affecting particularly the but may contain appreciable amounts of quartz
results of Atterberg limit and particle size tests and kaolinite (18, 19, 20). Due to the presence of
(5). iron oxides, these soils are red in color ranging
The identification of the suitability of TWR soils from light through bright to brown shades.
for use in earthworks operations is imperative. Reddish color, well drained topography and
However, soil sample preparation, laboratory volcanic parent rocks are useful indicators of
testing procedures and classification of TWR in tropical red laterite soils (1). They are heavily
Ethiopia is not consistent. Differences in leached soils having a granular or nodular
interpretation are apparent in the evaluation of appearance and range from silty clay to silty
the geotechnical properties, leading to different gravel with sand texture (3).
conclusions by engineers on the suitability of the Historically, the first systematic study of TWR
same type of soil. The sensitivity of TWR soils soils in Ethiopia was a study of the different
to drying and remolding has been shown to be a properties of tropical residual volcanic soils by
result of destructuring, the influence of Morin and Parry in Lyon Associates Institute
dehydration (desiccation) and the mineralogy of Inc. (1971). Using different soil sample
the soil (6). Hence, a generalized approach to the pretreatments and laboratory procedures, the
identification and evaluation of the different properties were correlated with the mineralogy
grades of TWR materials without regard to the and the origin of the residually formed volcanic
origin, degree of decomposition, laterization or soils for road construction. After several years,
desiccation may be misleading and lead to further studies were initiated into the
erroneous conclusions (13). More precise geotechnical properties of TWR soils in different
definitions, evaluation and standardization of parts of Ethiopia (7-12, 22-24). Studies of TWR
appropriate laboratory testing in TWR soils in different parts of Ethiopia have increased
characterizations is imperative in deciding on the in recent years because of increased construction
acceptability of these materials in earthworks operations, but these have adopted conventional
construction in Ethiopia. testing and classification systems in engineering
property characterization. Several researchers

46 Journal of EEA, Vol. 33, December 2015

 The Effect of Sample Preparation and Testing Procedure on the Geotechnical Properties…

agree that TWR soils require careful the study site and Fig. 3 shows soil sampling at
identification and characterization with special study site C.
field and laboratory procedures to obtain reliable
and consistent test data (25). The study of TWR
laterite soils depends primarily on the selection
of appropriate physical properties as indicators
of engineering behavior and the establishment of
suitable tests for their investigation (25). Despite
the large coverage and use of the residually
formed soils in Ethiopia, there has been no
systematic research, standardization in testing,
classification or description of these materials.
These gaps in knowledge have led to this
research that is aimed at characterization of the
basic geotechnical properties of TWR laterite
soils, by assessing the effects of sample
preparation and testing protocols, and the
significance of standard classification
procedures.

MATERIALS, TESTING METHODS

AND PROCEDURES
Laterite samples for laboratory testing were
obtained from a new highway project along the
Bako-Nekemet-Gimbi-Nejo road in western
Ethiopia as shown in Figs. 1 and 2. Detailed
Fig. 1. The main study site location along
field studies of the geology, climate and
highway alignments from: Bako-
topography of this area were also made. A
Nekemte-Gimbi-Nejo
number of test pits were excavated at different
sections of the project and two distinct residual
soils were exposed. To supplement the data,
more than sixty nine selected laboratory test
results from similar studies on neighboring sites,
such as the Nejo-Mendi road, Assosa, Weliyta-
Sodo, Nedjo-Jarso-Begi road and Nejo-Mendi-
Assossa road areas have also been taken into
account.
Site Description and Soil Sampling
Fig. 2. Partial Elevation versus distance along
The study site includes the section from Bako to the study site: Nekemet–Gimbi-Nedjo-
Nekemte (study site A: where the first four pits Mendi-Assosa
Pit 1 to Pit 4 were excavated), the section from
Nekemte to Gimbi (study site B: where Pit 5 and Laboratory Experimental Methods and
Pit 6 were excavated), and the section from Procedures
Gimbi to Nejo (study site C: where Pits 7 to 10 Suites of tests were carried out in the
were excavated). The relief of the route in the investigation of the properties of the laterite
study area comprises several plateaus with soils. These included mineralogical analysis
altitude varying from 1100 to 2250m cut with a using a Philips X-Ray Diffractometer (PW-
V-shaped valley. Figure 1 shows the location of 1729) and geochemical analysis. Other tests

Journal of EEA, Vol. 33, December 2015 47

 Addiszemen Teklay, et al.,

reported herein are moisture content Dry sample preparation was carried out for all
determinations; liquid limit and plastic limit AD samples in general accordance with ASTM
determinations (on material passing a 425μm D421-85. A similar procedure was also adopted
sieve); specific gravity determinations and for all OD samples. For the index properties, the
particle size distribution analyses. samples were prepared from the portion of the
soil samples passing a 425μm sieve and were
Tests were undertaken on air-dried (AD), oven-
kept wet for a period of 24hrs before carrying
dried (OD) and as-received (AR) soil samples
out Atterberg limit tests) .
using different preparation procedures to
investigate the influence on the results and Plastic and Liquid Limit tests – Both plastic and
classification of the soils based on the testing liquid limit detrminations were carried out by
methods prescribed by the ASTM. The AR reducing the moisture content (from wet to dry)
samples were carefully stored so as to maintain (22). The mixing times for the liquid limit tests
their in-situ moisture content prior to sample were first limited to 5 minutes, and the mixed
preparation and testing. The AD soil samples samples were left to dry overnight before testing.
were prepared in accordance with the After the moisture content for each liquid limit
methodology of Lyon Associates Institute Inc. test point was determined, the remainder was
(1971), Fookes, (1997) and Blight (1997) using mixed for a further 25 minutes before
an oven at a temperature of 50ºC and relative determining the liquid limit again. The
humidity not exceeding 30% (for at least 5 days) Casagrande cup method was used for the liquid
until the soils were sufficiently dry to achieve limit determinations.
constant weight under normal room Specific Gravity determinations - The specific
temperature. The OD samples were prepared by gravity determinations were used in the particle
drying the soils thoroughly at 105ºC. size distribution analyses and to assess any pre-
test drying effects. They were carried out in
accordance with ASTM D854-92.
Particle Size Distribution (PSD) - The potential
mechanical stability of the lateritic soils was
studied using particle size analyses tests.
Separate specific gravity tests were first made
for fine and coarse fractions on each soil sample
and if significantly different, the grading was
calculated by modifying the mass proportions as
well as adjusting the mass proportions for
analysis (27).
Wet preparation was used for the AR samples.
Fig. 3 Typical undisturbed sampling using a thin
For the AR samples, no form of pre-test drying
wall tube sampler around Nejo in western
was used. During the sample preparation, the test
Ethiopia
samples were soaked until the coating material
was fully softened noting that it was important
Sample Preparation - Wet sample preparation
was carried out on all AR soil samples in that the fines adhering to coarser particles were
removed and the fracturing of weak coarse
accordance with ASTM D2217- 85, procedure
B. For index properties that portion of soil particles was prevented (22, 27). The soaked
samples were washed in the manner stated in
passed while it was washed on the 425μm sieve
section 6.1.2 of ASTM D 2217-85 and
was used. The samples were allowed to dry at
procedures recommended by Lyon Associates
room temperature until the mass reached a putty-
(1971), using 2.00 mm and 0.075-mm sieves.
like consistency (15-25 drops of the cup in the
Hydrometer testing was used in the
liquid limit test).
determination of particle-sizes less than 0.075
mm in accordance with the procedures stated in

48 Journal of EEA, Vol. 33, December 2015

 The Effect of Sample Preparation and Testing Procedure on the Geotechnical Properties…

ASTM D 422-63. Sodium Hexa-Metaphosphate altitude of the study area is generally above
was used as a dispersing agent and proved 1600m and occasionally above 2000m. The
successful in dispersing the Ethiopian lateritic yearly mean monthly temperature is above 20ºC
soil particles (7, 22). with long periods of high temperatures. Using
local metereological data, the annual
Dry preparation was used for the AD and OD
precipitation for the Bako-Nekemte, Study Site
samples. The aggregations were broken down
A, is above 1839.37mm/yr. For Nekemte –
thoroughly in a mortar with a rubber-covered Gimbi (study site B) and Gimbi–Nejo (study site
pestle before they were screened through the C) it is 1860mm/yr. According to the Köppen-
nest of sieves. Then the test samples were climate classification, the areas are grouped as
separated into two portions by sieving with a Bsk (i.e. B = dry/arid, evaporation exceeds
2.00mm sieve. The retained particles were precipitation, s = semi-arid climates; steppe and
used for coarse sieve analysis, while the k = mid-latitude dry climate/cold) with humid
particles passing were subject to hydrometer aridity dominating from 25 April to 26 October.
analysis as for the AR samples. About 80% of the mean annual rainfall occurs
during a period of 4 months. This indicates
CLASSIFICATION OF THE SOILS cyclic wet and dry periods. The climate in these
areas is favorable for the formation of TWR
The effects of the different sample preparation laterite soils.
and testing procedures on the soil classification Topography and Drainage - Topographical and
has been investigated. In particular, the drainage characteristics seem to have a strong
applicability of the Unified Soil Classification and consistent influence on the weathering
System (UCS) (ASTM D 2487-93) and the process, and thus on the type of clay minerals
AASHTO Soil Classification System (AASHTO formed, especially in the wet tropics of western
M 145-87) has been explored. Soil grouping Ethiopia. This area is characterized by hilly and
based on mineralogical composition, as proposed mountainous land with plateaus and valleys. The
by Wesley and Irfan (1), and on the genetic basis soil in the higher ground is generally well
and soil-forming factors as proposed by drained and seepage flow has a strong downward
Duchaufour (1982) have also been investigated. component. This leads to the formation of low-
activity clay minerals, especially kaolinite. The
RESULTS AND DISCUSSIONS well-drained upland soils are frequently reddish,
red or brown or brown-red and such red colors
Field Study: Effects of Climate, Topography denote a non-hydrated iron oxide (Fe2O3) in the
and Geology on Laterite Soil Formation soil. In the middle and lower slopes, soil
drainage is poorer than in the upper slopes and
The identification and evaluation of laterites for summit and hydrated iron oxides are produced,
construction appears to depend on simultaneous mainly goethite (Fe2O3.H2O) and lomonite
consideration of all the major factors that affect (Fe2O3.1.5H2O). Their presence is responsible
the behavior of the soils. Although parent for the change in color down-slope from reddish
material, topography and climate are together brown to warm brown or orange brown and then
regarded in the tropics as the factors controlling to yellow or even brown yellow on lower
the soil formation, the climate has the ground.
predominant effect on a continental scale (13).
Granites, basalts and metamorphic rocks
Climate: Temperature and Rainfall - The including gneisses and schists characterize
Ethiopian climate is mainly controlled by the western Ethiopia. Very deep and intense
seasonal migration of the Inter Tropical weathering processes have taken place and rock
Convergence Zone (ITCZ) and the associated exposures are mainly evident only in the areas
atmospheric circulations of air as well as by the underlain by granites and gneisses (7, 12).
complex topography of the country (28). The

Journal of EEA, Vol. 33, December 2015 49

 Addiszemen Teklay, et al.,

Results of Laboratory Testing i 1.347 1.381 2.238 2.247

Geochemical and Mineralogical Soil ki 0.669 1.382 0.986 0.910

Characteristics - In accordance with the USCS
Remark Laterite Laterite Laterite Laterite
system, the laterite soils encountered in the
fieldwork may be classed as Silty Gravel with 
Group A: Test results on Silty/Clayey Gravel with Sand laterite soils
Sand texture (Group A) other than those soils in
Pit 5 and Pit 6 that comply with Elastic 
Group B: Test results on Elastic Silt/Silty Clayey textured laterite soils
Silt/Clayey Silty textured (Group B); Group A
soils having <35% fines of silt and clay and
Group B >35% fines. The two groups of soil are The results of analysis in Table 1 and the
derived from granite gneiss. It is these soils and Lateralisation Indices Fig. 4 indicate that the
their weathering with depth for which testing has soils tested are true laterites. The lowest ratio
been carried out. Average oxide compositions of (i.e. molecular silica: sesquioxide ratio) is
Group A and B soils are shown in Table 1. The recorded for the granite gneiss derived Group A
extent of soil laterisation and weathering laterite soils at a depth less than 1.5m, and
conditions were evaluated using the following indicates more lateritization than the granite
three indices. gneiss derived Group B laterite soils at the same
depth. However, while weathering decreases
SiO2
[1] i  , with depth for both groups, laterisation increases
Al2O3 with depth for Group B laterites. The presence of
SiO2 bases and feldspar in some soils shows that the
[2] i  ,
Fe2O3 soils are capable of undergoing further
SiO2 weathering. Silicon is the most abundant element
[3] ki 
 Al2O3  Fe2O3  found in both textured soils within both depths.
Apart from silicon, only iron, aluminum, and
lithium are expressed significantly. Combined
Table 1. Average geo-chemical oxide water is also present in appreciable amounts. The
composition in percentage for triangular representation of major oxides (SiO2,
Western Ethiopian laterite soils with Al2O2, Fe2O3), as shown in Fig. 5, indicates all
depth laterites are localized toward the Fe2O3 pole
Group A Soils  Group B Soils  confirming their ferruginous nature. It is also
noted that all points stretch on a straight line
0.0-1.5m 1.5-3.m 0.0-1.5m 1.5-3.m
Oxides perpendicular to the SiO2 and Al2O3 axis,
% % % %
SiO2 32.990 33.80 40.37 38.23
confirming the importance of alumino-silicate
minerals identified as kaolinite in mineralogical
Al2O3 24.800 24.77 22.89 25.00
tests.
Fe2O3 24.495 24.46 18.04 17.02
Mineralogical identification reveals that, these
LiO 8.860 7.360 11.17 11.97
soils are composed of kaolinite, quartz, hematite,
H 2O 4.870 6.240 3.530 3.460
vermiculite, illite, borax, dickite, and nacrite.
TiO2 2.111 1.760 2.070 2.310 The clay mineralogy is dominated highly by
K 2O 0.371 0.500 0.700 0.711 weathered 'mature' kaolinite (Al4Si4O10(OH)8)
Na2O 0.369 0.220 0.490 0.333 and quartz (SiO2), generally associated with
MgO 0.273 0.250 0.430 0.360
smaller amounts of hematite and sometimes
gibbsite.
P2O3 0.250 0.280 0.410 0.154
CaO 0.230 0.240 0.311 0.063
MnO 0.200 0.210 0.124 0.131
i 1.330 1.365 1.763 1.529

50 Journal of EEA, Vol. 33, December 2015

 The Effect of Sample Preparation and Testing Procedure on the Geotechnical Properties…

There were no allophane or halloysite clay

minerals present in the soil, and the relatively
high plasticity is due to the high water retention
capacity of the minerals present and the micro
aggregate structure. The study has shown that
there is a great concentration of sesquioxide of
aluminum and iron, no amorphous materials, and
limited feldspar particularly associated with the
coarser soils. The dominance of acidic rocks
such as granite, granitic gneiss and schists from
geological studies is supported by the
mineralogy test results. Moreover, the
weathering has led to congruent dissolution of
kaolinite and quartz with a less pronounced
formation of gibbsite.
Fig. 4. Average Laterisation indices of Group A Moisture Content and the Effect of Testing
and B Ethiopian laterite soils with depth. Temperature - In-situ moisture contents were
determined by oven drying at 50ºC and 105ºC.
Soil rich in poorly ordered kaolinite is typical for As expected, the measured moisture contents
a mature tropical soil formed by intense showed a general increase with increase in
weathering and leaching and is rarer in drying temperature. Both testing temperatures
temperate climates. However, quartz-rich parent led to a normal distributed plot (within 0.05
rocks tend to produce gravelly materials significant level of accuracy). Figure 6 shows the
containing more coarse quartz particles than the results with all results in the scatter plot below
concretionary ones. The laterites have been the line of equality towards drying temperatures
formed in the absence of humus/organic matter. of 105ºC. Moreover Group B laterite soils show
The Laterisation Indices for Group A and B generally higher values for both drying
Ethiopian laterite soils with depth in Fig 5, temperatures. The natural moisture contents for
support the above conclusions. Group A laterite soils range from 24.4 to 28.3%
and for Group B laterites range from 29.1 to
31.6% .
The difference in moisture content obtained
between the two drying temperatures was less
than around 4% for both textured laterite soils.
This indicates that Ethiopian laterite soils do not
contain a considerable amount of „water of
hydration‟ or „structural water‟ which leads to
irreversible changes on drying. It is concluded
that for general purposes the moisture content
tests for the soils can be conducted using the
conventional oven drying (105ºC) method. As a
result, ASTM D:2216–98, standard test method
for laboratory determination of moisture content
is considered acceptable for determining the
natural moisture content of Ethiopian laterites. It
Fig. 5. SiO2 - Al2O3 - Fe2O3 triangular diagram is recommended to use three or more duplicate
indicating the position of the degree of samples of AR soil with weighted average to
laterization of Ethiopian laterites. obtain the most accurate results.

Journal of EEA, Vol. 33, December 2015 51

 Addiszemen Teklay, et al.,

western and southern Ethiopian laterites (7, 9,

10, 12, 23).
Regarding objective (ii), comparison of the
differences of the liquid limits for 5 and 30
minutes remolding time for each of the three
pretesting preparation cases (AR, AD and OD)
indicates increased mixing increases the liquid
limit by around 5%. Excessive manipulation
prior to testing leads to excessive breakdown of
the soil structure and disaggregation resulting in
loss of lateralization that would not be expected
to occur under field conditions. Disaggregation
results in increased fines and higher liquid limit
values. The data suggest that the time of mixing
causes a greater change in the liquid limit data
than the pretest drying proceedures considered.
Fig. 6. The relationship between natural moisture
content using oven drying of 500C and Regarding objective (iii), it is recommended that
1050C for a range of TWR laterite soils of AR soils are used for plasticity index
Ethiopia. determinations and that mixing times are limited
to not more than 5 minutes with fresh soil used
Effect of Pretest Drying, Remolding and Testing for each moisture content point in the Atterberg
Procedures on Atterberg Limits - Atterberg limit limit tests. The soils should be broken-down by
laboratory tests were conducted to study: soaking in water, and not by drying and
grinding, as is suggested for temperate zone
(i) the effects of pretest drying of the soils prior soils. Pretesting treatment should reflect the field
to remolding and testing; conditions at the time of construction as
represented by AR conditions. The test samples
(ii) the effect of the degree of remodeling of the should not be over-dried prior to testing, as the
soil samples prior to testing; and soils would not be expected to experience such
high temperatures in the field (1, 7, 12, 22). In
(iii) to assess the most effective method of support of this, the plots of Figs. 11 and 12,
achieving meaningful and repeatable limit based on compilation of similar studies made in
results applicable to site conditions. Ethiopia, show that greater liquid limits are
obtained for AR samples than for AD and OD
It is generally recognized that limit tests are samples for Group A and B soils tested at 5min
normally reproducible within 5 to 6% (22, 26). mixing time.
While bearing this in mind, the following
substantial body of results (Figs. 7, 8, 9 and 10) The relationship between liquid limits using 5
indicate consistent trends for a mix of Group A and 30 min mixing time on AR, AD and OD for
and Group B soils. a range of TWR laterite soils presented in Fig.
13 supports the conclusion that increased
Regarding objective (i), the AR samples give manipulation results in an increase in liquid
liquid limites somewhat greater than those of the limit. Average PI value of 21.8% and 22.3% for
AD samples, which in turn give values greater the AR soil samples of Group a and Group B
than those of the OD samples. The difference respectively, obtained for 5 minutes mixing time,
between the AR and OD samples is typically have been used for classification purposes (7-12,
around 5%. The results are in general agreement 23 and 24).
with the conclusions made by Blight (1997) and
the results of previous research conducted on

52 Journal of EEA, Vol. 33, December 2015

 The Effect of Sample Preparation and Testing Procedure on the Geotechnical Properties…

The Effect of Pretest Drying on Specific

Gravity

The in-situ specific gravity of the laterite soils

studied ranged from 2.69 to 3.02 with the
maximum value higher than that of most
temperate soils, which are typically in the range
2.65-2.70. The specific gravity of coarse-grained
Group A laterite soil is greater than that of fine-
grained Group B textured soil and there is a
general trend of increase in specific gravity with
depth. This is interpreted as due to a high
concentration of iron oxide. Specific gravity
values for different pretreatment conditions are
indicated in Fig. 14. There is a clear trend of
higher specific gravity for the AR (no pre-
drying) samples. The reason for this possibly lies
in the loss of water of crystallization (or intra-
particle or inter-layer water) from the clay (or Fig. 8. Variations of liquid limit using AD soil
other) minerals on oven drying, and hence a sample tested at 5 and 30min mixing time
reduction in the density of the solid matter. with natural moisture content for TWR
laterite soils tested.
It is concluded that laterite soils should be at
their natural moisture content for specific gravity
analysis; avoiding pretest drying as it tends to
reduce the measured specific gravity compared
with that at the natural moisture content.

Fig. 9. Variations between liquid limits using

OD soil sample tested at 5 and 30min
mixing time with natural moisture content
for laterite soils tested
Fig. 7. Variations of liquid limit using AR soil
sample tested at 5 and 30min mixing time
with natural moisture content for a range
of TWR laterite soils tested

Journal of EEA, Vol. 33, December 2015 53

 Addiszemen Teklay, et al.,

Fig.12. Variations in liquid limits on AR, AD

and OD Group B Ethiopian laterite soils
Fig. 10. The relationship between liquid limits tested at 5min mixing time
using 5 and 30 min mixing time on AR,
AD and OD for laterite soils tested Effect of Soil Pretreatments, Testing
Procedures, Specific Gravity and Depth on
Particle-Size Distribution

The particle-size distribution results were

analysed to assess:

(i) the effects of different pre-test drying (i.e.

AR, AD and OD) while using the different
testing procedures (i.e. dry versus wet sieving);

(ii) to assess the most effective method of

achieving meaningful and repeatable grading
results applicable to site conditions;

(iii) the effect of using separate specific gravities

for the fine and coarse fractions of the test
samples; and

(iv) the effect of depth of sampling.

Fig. 11. Variations in liquid limits on AR,AD

and OD Group A Ethiopian laterite
soils tested at 5min mixing time

54 Journal of EEA, Vol. 33, December 2015

 The Effect of Sample Preparation and Testing Procedure on the Geotechnical Properties…

particles. Generally, the particle-size

distributions of lateritie soils change as a result
of hydration reaction when soils are dried
especially at high temperatures. There is also a
possible influence from adherence of fines to
coarser particles which were not dislodged
during washing and that some laterites
experience fracturing of the coarser fractions
during sample preparation and testing.
Regarding objective (ii), the variation in
percentage of partcle sizes in the different test
proceedures is not significant and comparable to
that which might be expecetd in testing of
temperate soils. Although, pre-drying did not
significantly affect the results, it is
recommended that pre-test preparation and
testing is performed on soils in their natural state
Fig. 13. Liquid limit between 5 and 30min (AR soil). It is also recommended that the wet
mixing time on AR, AD and OD preparation (i.e. soaking the soil until the coating
samples for Group A and Group B material is fully softened) and the wet sieving
TWR laterite soils of Ethiopia procedures should be used in practice.
Dispersion of the soils was effectively achieved
using 'standard' sodium hexametaphosphate
dispersant, provided the dispersant with the soil
suspension was shaken well (7). Testing the soil
without pre-drying and using a wet sieving
procedure is consistent with normal
recommended preparation procedure. In general,
consistent and meaningful particle size
distributions were not difficult to obtain for
Ethiopian laterite soils. The results of Western
Ethiopian laterites tests by several researchers
also support this conclusion (7, 12).
Regarding objective (iii), though it has been
shown that speciifc gravity values vary
somewhat with pre-treatmnet conditions, the
grading curves using mass and modified mass
Fig. 14. Line of equality between specific proportions do not vary significantly, especially
gravity values for ifferent pretest for samples tested under wet seiving (i.e. AR
drying treatments pretreatment condition). The need to modify the
Representative particle size ditribution plots are grading curves using modified mass proportion
shown in Figs. 15 and 16. is not considered important when the wet sieving
method on AR soil is used.
Regarding objective (i), the AD and OD soil pre- Regarding objective (iv), the study has shown a
treatments and testing procedures for Group A general increase in the percentage of coarse
and Group B laterites resulted in an increase of gravel particles with increasing depth and a
the finer particles compared to the AR samples decrease in the clay and silt fractions consistent
with a concomitant decrease in the coarser with a decrease of weathering. testing

Journal of EEA, Vol. 33, December 2015 55

 Addiszemen Teklay, et al.,

procedures considered. The range of grading

curves of the soil samples is as shown in Fig. 17
for various AR sample pretreatment conditions.

Fig. 16. Variation of the % of the different sizes

at different predrying temperature and
testing conditions with depth for Group
B soils.
Fig. 15. Variation of the % of the different sizes
at different predrying temperature and
testing conditions with depth for Group
A soils.

Effects of Sample Preparations and Testing

Procedures on Classification of TWR
Laterite Soils
The laterite soils tested are typical of western
Ethiopia having activity indices (AI) of < 1.25
indicating them as inactive or normal. This is
attributed to the mode of weathering which
involves the coating of the soil particles with Fig.17. Grading envelop for silty gravel with
sesquioxide resulting in suppression of the sand textured laterites with depth for
surface activity of the clay particles. Density various AR pretreatment conditions
determinations resulted in a range of values with The purpose of a soil classification system is to
bulk unit weights ranging from 17 to 19.2 describe the characteristics of a soil that give
and dry unit weight ranging from 13.7 to guidance on how it will behave as a construction
14.9 with a volumetric water content material. Soil classification should not be used as
of the soils generally ranging from 0.4 to an end in itself but as a guidance tool to material
0.5% for Group A and Group B materials within properties and usage. Soil properties such as
the depth investigated. It is an assessment of the strength and compressibility also need to be
appropriate testing techniques and classification taken into account as does the sensitivity of the
of these soils based on the test results that is a soil to testing technique and experience of using
prime purpose of this paper. the materials in practice. It is these latter two
points that are addressed in this paper.

56 Journal of EEA, Vol. 33, December 2015

 The Effect of Sample Preparation and Testing Procedure on the Geotechnical Properties…

This is counteracted to an extent by increase in

Based on USCS, the majority of the granular plasticity if the longer mixing time of 30 minutes
soils are classified as SM, GM, GW-GM and is used as the aggregates tend to break down.
GW- GC (Silty Gravel with Sand or Silty Sand Depending on the pre-treatment, the data shifts
with Gravel textured). Figures 18 and 19 indicate parallel to the A-Line within the same group. It
the relative position of the soils on the plasticity is stressed that the changes in plasticity because
chart. All samples for material <425μm lie of the pretreatment procedure can have an
below the A-line in the region of silts, MH influence on the selection of earthworks
category. In general the soils are Silty Gravel compaction plant and assessment of other soil
with Sand textured [Group A] and few of them properties such as strength, compressibility and
are of Silty Clay textured [Group B] laterites. volume change characteristics when correlation
Both groups lie in the region of kaolinite- techniques are used. It is important that
dominated soils on the plasticity chart in laboratory techniques yield results that reflect
agreement with the mineralogical tests likely site conditions.
undertaken. Based purely on the plasticity chart Also of great importance is the sensitivity of the
of the USCS the laterites would be classed as laterite soils to pretreatment. This sheds doubt on
having „poor‟ engineering properties and the applicability of the USCS for TWR laterite
considered troublesome and potentially soils at least without due consideration to their
unsuitable for construction purposes. However, true nature and in particular to the finer Group B
according to [5], TWR red clays/or clayey slit soils. According to the AASHTO M-145 system,
soils are considered to possess unusual or the majority of the soils are granular
abnormal properties, at least when compared (Silty/Clayey gravel and Sand). These granular
with what might be termed normal sedimentary soils fall under subgroup A-2-7 with group index
clays of more temperate regions. The most less than three and are characterized as a good to
abnormal property from an engineering excellent subgrade materials, and few samples
viewpoint is that despite relatively high liquid are fine grained soils (Silty/Clayey Soils) and
limits and small particle size they have good located under subgroup A-7-5 with group index
engineering properties and are stable usually less than 21, characterized as fair to poor road
having a high shear strength and low construction materials. The AASHTO system,
compressibility. As a result, in practice, even though it appears to appropriately classify
Ethiopian laterite soils generally show good the coarse-grained Group A soils, classifies the
engineering properties especially for tropical Group B soils as potentially unsuitable
construction of subgrades and embankments (5, for construction materials. However, these
29). The classification of these laterite residual materials for the reasons given previously are
soils, using the USCS has shown contradictory good engineering materials in practice. Vargas
results and several studies have questioned the (1982) showed in the plasticity chart, that most
use of USCS for classifying TWR soils (1). of the soil obtained from decomposed granite,
Figures 18 and 19 show the Casagrande plasticity gneiss and schist lay below the A-Line. Hence,
chart for AR, AD and OD samples for 5min and the use of the AASHTO system is not advised
30 minute mixing times in determining the liquid for classifying Group B textured laterite soils for
limit for both Group A and B soils. The effect of road construction purposes, though it may work
the pretreatment conditions on the classification well for Group A laterite soils.
of the soils using the USCS results in no major
change in the group name and symbol. However, It appears that the most practical way to evaluate
laboratory results show that drying at high the likely engineering behavior is to use the
temperatures tends to aggregate soil particles, classification systems that utilize plasticity data,
either through oxidation of free iron or structure, and mineralogy and geo-
dehydration of clay minerals, or both, and thus morphological impacts on soil formation, such
decreases the liquid limit and plasticity index. as the Wesley and Irfan (1997) classification
system, combined with AASHTO system. Such

Journal of EEA, Vol. 33, December 2015 57

 Addiszemen Teklay, et al.,

a classification together with material properties chemical weathering of the soils. In this regard
such as strength and compressibility would arm using the Duchaufour (1982) scheme, the
the engineer with the data needed to decide on development of these soils should be seen as
the suitability of the materials, compaction phases in the same weathering process, forming
requirements and the design parameters in part of a weathering continuum from fersiallitic
earthworks. through to ferrallitic soils. According to the
prevailing soil forming factors and genetic basis,
the silty clayey (Group B) textured laterite soils
(or A-7-5) are grouped under a ferrisol group.
Ferrisols seem to cover several varied soils and
are generally shallow (about 2m), with higher
silt, free iron contents and a horizon of clay or
silica gel accumulation.
However, Group A textured soils (A-2-7), with a
lesser silica to sesquioxide ratio, are grouped as
ferrallitic soils. The ferrallitic process is
associated with intensified deep, pale, sandy
surface horizons that have evolved especially on
quartz-rich granites or sands. Moreover, the soils
are all true laterites based on the texture,
mineralogy and geo-chemical test results which
Fig. 18. The locations of the laterite soils studied
give a silica/sesquioxide ratio below 1.33
on the Casagrande Plasticity Chart.
(Laterization Index in Fig. 4) True laterites
are very old soils from which practically all the
soluble materials have been leached. As a result,
the soils based on Wesley (1988) and Wesley
and Irfan (1997) classification system is grouped
under “Group C: residual soils strongly
influenced by special clay minerals not found in
sedimentary soils”. The influence of sesquioxide
was detected during chemical and mineralogy
tests; as a result, the soils are sub-grouped into
“Soils influenced by the presence of
Sesquioxide”. The principal role of the
sesquioxides appears to be as a cementing agent
that binds the other mineral constituents into
clusters or aggregations.
Fig. 19. Locations of Ethiopian laterite soils on
CONCLUSIONS AND
the Casagrande Plasticity Chart based
on several studies (7, 8, 9, 10, 11, 12, 22,
RECOMMENDATIONS
23, 24) The major conclusions arrived at in this study in
relation to the use of laterite soils in earthworks
The soils formed in the southern and the western are presented below:
parts of Ethiopia are characterized with a mean
a) Mineralogical identification reveals that the
annual temperature ranging from of
soils are composed of kaolinite, quartz, hematite,
and mean annual precipitation greater than
vermiculite, illite, borax, dickite, and nacrite.
1500mm, therefore according to Summerfield
The clay mineralogy is dominated by highly
climate zonation cited in (29), the region is a
weathered 'mature' kaolinite and quartz,
humid tropical zone. The geomorphological
associated with smaller amounts of hematite and
setting strongly supports the formation of in situ

58 Journal of EEA, Vol. 33, December 2015

 The Effect of Sample Preparation and Testing Procedure on the Geotechnical Properties…

sometimes-small proportions of goethite and/or road construction purposes. However, it appears

gibbsite. No humus/organic matter, no allophane to work well for coarse-grained laterites.
or halloysite minerals were present. h) Classification systems that take account of
b) No significant water of crystallization plasticity data, soil structure, mineralogical
(structural water/ chemically combined water, compositions and geo-morphological impacts on
was detected that could lead to irreversible soil formation, together with the ASSHTO
changes on drying) was driven off samples dried system, can result to give a good indication of
at 50 or 105°C. As a result, a conventional the engineering properties of laterites.
drying temperature of 105°c as specified in
ASTM D-2216 for moisture content RECOGNITION
determination was adopted for the TWR laterites The work described in this paper was prepared
tested. as a part of the PhD thesis submitted in AAU,
c) It is considered appropriate that laterite soil AAiT in 2012.We wish to acknowledge the
samples for Atterberg limit testing should be contributions of AAU and BDU lab technicians
broken down by soaking in water and not by and MH Engineering Plc.
drying and grinding as is conventional practice
for temperate zone soils. REFERENCE
d) The laterite soils in Ethiopia are sensitive to [1]. Abebaw, Z. “Basic Engineering Properties
handling and manipulation. To obtain consistent of Lateritic Soils found in Nedjo-Mendi
and repeatable plasticity data it is recommended Road Construction Project”, Addis
that remolding of samples prior to Atterberg Ababa, Ethiopia: M.Sc. thesis, Addis
limit testing is restricted to no more than 5 Ababa University, Department of Civil
minutes with fresh soil used for each moisture Engineering, AAiT, 2005.
content point in the Atterberg limit test on as
received soil using the 'drying-back technique. [2]. Amare G., “Compaction Properties of
Lateritic Soils” Addis Ababa, Ethiopia:
e) It is recommended to use soils in their natural M.Sc. thesis, Addis Ababa University,
state (no pretest drying) for particle size tests, as Department of Civil Engineering, AAiT,
the soils are sensitive to manipulation and 2008.
pulverization. The use of the wet sieving method
is recommended with the soil soaked until the [3]. Blight, G. E. Mechanics of Residual soils,
coating material is fully softened. The “A Guide to the Formation, Classification
modification of the grading curves using and Geotechnical Properties of Residual
modified mass proportion is not considered Soils, with Advice for Geotechnical
important if the wet sieving method is used. Design”, Balkema A.A., Rotterdam:
Netherlands. 1997.
f) Specific gravity tests should be carried out on
samples at their as-received moisture content, [4]. Charman, J. H. Laterite in road pavements:
rather than on pre-dried samples as temperature Transport Research Laboratory,
can significantly influence the results. Department of Transport. London:
Construction Industry Research and
g) The laterite soils plot below the A-line in the
Information Association (CIRIA), 1995.
USCS classification system casting doubt on
their suitability in earthworks. Such soils are [5]. David, A.A. Geology and Engineering
however frequently and satisfactorily adopted Properties of laterites from Ilorin, Nigeria.
for such usage and accordingly the plasticity Amsterdam, Netherlands: Elsevier Science
chart should be used with causation. Moreover, Publishers B.V., 1982. pp. 111-118
it has also been found that the AASHTO system (1982/83).
is unsuitable for classifying the finer tropical
Elastic Silt/Silty Clayey textured laterite soils for

Journal of EEA, Vol. 33, December 2015 59

 Addiszemen Teklay, et al.,

[6]. Dresai, M. D. Geotechnical aspects of department of Civil Engineering, AAiT,

Residual Soils of India, Sampling and 2008.
testing of Residual Tests. 1985.
[16]. Harwant, Singh and Bujang, B.K Huat.
[7]. Duchaufour, P. Pedology, Pedogenesis and Origin, formation and occurrence of
Classification. (English edition Trans.). T tropical residual soils. [Book auth.] See-
R Paton London: George Allen & Unwin, Sew & Ali Huat. Tropical Residual Soils
1982. Engineering. London: Taylor & Francis
Group, 2004.
[8]. Ethiopian Ministry of Water Resources,
National Meteorological Services Agency. [17]. Huat, B. K., Sew, G. S. and Ali, F. H.
Addis Ababa, Ethiopia: National Tropical Residual Soils Engineering. s.l.:
Meteorological Services Agency under the Taylor & Francis Rutledge, 2004.
GEF supported Climate Change Enabling
[18]. Jiregna, D. Detailed investigation of the
Activities Project of Ethiopia, 2001.
properties of lateritic soils, the case of
[9]. Fasil, A. Investigation into some of the Nedjo – Mendi – Assosa. Addis Ababa,
Engineering properties of Red clay soils in Ethiopia: M.Sc. thesis, Addis Ababa
Bahirdar. Addis Ababa, Ethiopia: M.Sc. University, department of Civil
thesis. Addis Ababa University, Engineering, AAiT, 2008.
department of Civil Engineering, AAiT,
[19]. Lyon Associates Institute Inc. Laterite and
2003.
Lateritic soil and other problem soils of
[10]. Fekede, W. Investigating the Index Africa. USA: Lyon Associates Institute
properties of Residual Tropical Soils of Inc. USA & Building and research
western Ethiopia (The case of Asossa). Institute, Kumasi Ghana., 1971.
Addis Ababa, Ethiopia: M.Sc. thesis.
[20]. Morin, W. J. and Parry, W. T.
Addis Ababa University, department of
Geotechnical properties of Ethiopian
Civil Engineering, AAiT, 2007.
Volcanic Soils. USA, Geotechnique 21.
[11]. Fookes, P.G. Tropical residual soils. A No.3. 1971, pp. 223-232.
geological society engineering group
[21]. Nortlunor, K. J., et al., Engineering
working party revised report. s.l.: The
geology of tropical red clay soils:
Geological Society, 1997.
Geotechnical characterization: index
[12]. Gidigasu, M. D. Laterite soil Engineering, properties and testing
Pedo genesis and Engineering Principles.
[22]. procedures. Report WN/93/12. Key worth,
Ghana: s.n., 1976.
Nottingham: British Geological survey.
[13]. Gidigasu, M. D. Degree of weathering in 1992.
the identification of laterite materials for
[23]. Sherman, D. G. Factors Influencing the
engineering purposes -- a review. 1974.
Development of Lateritic and Laterite
pp. 8:213-266.
Soils in the Hawaiian Islands'. Honolulu,
[14]. Hailemariam, A. Investigation into Shear T. H.: University of Hawaii Agricultural
Strength Characteristics of Red Clay Soil Experiment Station, 1949.
of Addis Ababa. Ethiopia: M.Sc. thesis.
[24]. Sintayehu, H. Investigation of Influence of
Addis Ababa University, department of
Compaction on the Stability of Earth-fill
Civil Engineering, AAiT, 1992.
Dams of Tropical Soils. Addis Ababa,
[15]. Hanna, T. Study of Index properties and Ethiopia. : M.Sc. thesis, Addis Ababa
shear Strength parameters of Laterite soils University., 2003.
in southern part of Ethiopia, the case of
[25]. Suhaimi, J., et al., Sampling and testing of
Wolayita - Sodo. Addis Ababa, Ethiopia. :
tropical residual soils. [Book auth.] See-
M.Sc. thesis, Addis Ababa University,
Sew & Ali (eds. Huat. Tropical Residual

60 Journal of EEA, Vol. 33, December 2015

 The Effect of Sample Preparation and Testing Procedure on the Geotechnical Properties…

Soils Engineering. London: Taylor &

Francis Group, 2004, p. 21.
[26]. Townsend, F.C., Phillip, G.M. and
Parcher. Effect of Remoulding on the
Properties of a Lateritic Soil. 1969.
[27]. Varghese, T. Laterite soils and their
management. Vellayani, India:
Department of Soil Science, Kerala
Agricultural University, 1987.
[28]. Wesley, L. D. Cluster hypothesis and the
shear strength of tropical red clay.
Indonesia: Department of Civil
Engineering, Imperial College, formerly
Department of Public Works, 1973.
[29]. Wesley, L. D. Engineering classification
of residual soils. Singapore: s.n., 1988, pp.
77-84.
[30]. Wesley, L. D. Geotechnical engineering in
residual soils. Hoboken, New Jersey: John
Wiley & Sons, Inc., 2010.
[31]. Wossen, M. Investigation into the
appropriate laboratory testing procedures
for the determination of the index
properties of the Lateritic Soils of Western
Ethiopia (Nedjo-Jarso-Begi Road area).
Addis Ababa, Ethiopia: M.Sc. thesis.
Addis Ababa University, department of
Civil Engineering, AAiT, 2009

Journal of EEA, Vol. 33, December 2015 61

 Addiszemen Teklay, et al.,

62 Journal of EEA, Vol. 33, December 2015