5-3

Clay-Based Pellets for Use in Tunnel

Backfill and as Gap Fill in a Deep
Geological Repository:
Characterisation of Thermal-Mechanical
Properties

NWMO TR-2012-05 December 2012

C-S. Kim1, A. Man2, D. Dixon1, E. Holt3, A. Fritzell4

1
Atomic Energy of Canada Limited
2
Golder Associates Ltd., Winnipeg, Canada
3
VTT Technical Research Centre of Finland, Finland
4
ES Konsult, Sweden
 - ii -

Nuclear Waste Management Organisation

22 St. Clair Avenue East, 6th Floor
Toronto, Ontario
M4T 2S3
Canada

Tel: 416-934-9814
Web: www.nwmo.ca
 - iii -

Clay-Based Pellets for Use in Tunnel Backfill and as Gap Fill in a Deep Geological
Repository: Characterisation of Thermal-Mechanical Properties

NWMO TR-2012-05

December 2012

C-S. Kim1, A. Man2, D. Dixon1, E. Holt3, A. Fritzell4

1
Atomic Energy of Canada Limited
2
Golder Associates Ltd., Winnipeg, Canada
3
VTT Technical Research Centre of Finland, Finland
4
ES Konsult, Sweden
 - iv -

Disclaimer:

This report does not necessarily reflect the views or position of the Nuclear Waste Management
Organisation, its directors, officers, employees and agents (the “NWMO”) and unless otherwise
specifically stated, is made available to the public by the NWMO for information only. The contents of
this report reflect the views of the author(s) who are solely responsible for the text and its conclusions
as well as the accuracy of any data used in its creation. The NWMO does not make any warranty,
express or implied, or assume any legal liability or responsibility for the accuracy, completeness, or
usefulness of any information disclosed, or represent that the use of any information would not infringe
privately owned rights. Any reference to a specific commercial product, process or service by trade
name, trademark, manufacturer, or otherwise, does not constitute or imply its endorsement,
recommendation, or preference by NWMO.
 -v-

ABSTRACT

Title: Clay-Based Pellets for Use in Tunnel Backfill and as Gap Fill in a Deep
Geological Repository: Characterisation of Thermal-Mechanical Properties
Report No.: NWMO TR-2012-05
Authors: C-S. Kim1, A. Man2, D. Dixon1, E. Holt3, A. Fritzell4
1
Company: Atomic Energy of Canada Limited, 2 Golder Associates Ltd. 3 VTT Technical
Research Centre of Finland, Finland, 4 ES Konsult, Sweden
Date: December 2012

Abstract

Pellets composed of bentonite-based materials are being considered by NWMO, Posiva and
SKB for use in filling the rock-buffer annular gap in the In-Floor Borehole placement geometry,
as a component in tunnel backfilling, and as a chamber filling material for the Horizontal Tunnel
Placement (HTP) geometry. The range of locations where pellets could be used in a repository
means that different types of pellets may be used since the primary functional requirements of
the pellets will vary. Preliminary work has revealed that it is difficult to place pellets in
repository simulation trials and achieve as-placed dry density of 1.4 Mg/m3 and target thermal
conductivity for this type of sealing material. Thermal conductivity of these materials tend to be
low (generally <0.5 W/(m·K)), which limits the rate of heat transfer from the used fuel to the
surrounding rock. This has implications on the temperature developed in a repository and the
spacing of the used fuel to prevent excessive temperatures. Improving the thermal conductivity
of clay-based pellet fill material is therefore highly desirable.

The objective of this work was to improve the quality of bentonite based pellets to better meet
the heat transfer requirements of the placement concepts being considered by NWMO, Posiva
and SKB. A range of pellet sizes, shapes and the effects of various volumetrically inert fillers
on the properties of bentonite-based materials were examined. The pellets produced all had
individual pellet densities in the order of 2.0 Mg/m3. When loosely poured, the dry density of the
fill typically was in the order of 1.1 to 1.2 Mg/m3. This could be improved to ~1.4 Mg/m3 using
vibratory compaction. Pellets made with silica sand and illite additives had dry densities slightly
higher than this. However, pellets containing additives have lower EMDDs than 100%
bentonite which adversely affects their hydraulic conductivities and swelling pressures. The
addition of Wyoming bentonite fines (i.e., 80 mesh granules), vibrated into the pore space
between the pellets increased dry density to >1.5 Mg/m3.

The presence of silica sand and illite in the pellets resulted in only a small increase in the
measured thermal conductivity of a mass of dry pellets relative to pellets made with 100%
bentonite. Increasing the overall density through vibratory compaction had a greater effect on
increasing thermal conductivity than the additives. However, maximum values measured on
pellets densified by vibratory compaction were close to 0.6 W/(m·K) and appears to be the
maximum achievable for as-placed pellets without the addition of fines to the void space
between the pellets or artificial wetting. This value can therefore be used as a lower bounding
limit for thermal assessments of the in-floor borehole’s gap fill in the KBS-3V geometry as well
as tunnel backfill where there is no immediate water influx (dry borehole). The vibration of
Wyoming bentonite fines (80 mesh granules) into a pellet-filled volume typically provided a
slight improvement to the thermal conductivity to slightly less than 0.7 W/(m·K).
- vi -
 - vii -

DEFINITION OF TERMS

The discussion contained in this document includes use of terms that are not uniformly applied
in the programs of NWMO, SKB, and Posiva. As a result, a brief description of some of the key
terms and the manner they are used is provided below. It should be noted that unless
otherwise specified in the text, NWMO terminology is used in this document.

AECL – Atomic Energy of Canada Limited.

Backfill – Term is used in this document to define any clay-based material installed in the
placement tunnels, access tunnels, shafts or other locations in a repository. Its composition
may vary depending on its location and functional requirements.

Bentonite – The trade-name used to describe commercially marketed, smectite-rich clay,

typically dominated by the clay mineral Montmorillonite.

BSB – Bentonite-Sand Buffer.

Buffer – Highly compacted bentonite material installed immediately adjacent to the UFC in the
IFB placement geometry. It is typically defined as being a densely pre-compacted bentonite
clay installed within the placement borehole.

Canister – Term used by Posiva/SKB to describe the Used Fuel Container (UFC), also referred
to as Container in NWMO terminology.

EBS – Engineered Barrier System.

EMDD – Effective Montmorillonite Dry Density. A normalising parameter used to express the
density of the swelling clay component after factoring out mass and volume of non-swelling clay
solids. The equation for this parameter is provided on page 18 of the document.

FSI - Free Swell Index. A parameter used to describe the volume that a material will occupy
when allowed unlimited access to free water and unconfined. Parameter is usually expressed
as mL/2gm of dry soil.

Highly Compacted Bentonite (HCB) – Material produced by densely compacting bentonite

clay (a dried and crushed, smectite-rich shale) into desired shapes for use immediately adjacent
to the UFC (buffer rings and cylinders, clay pellets). There is no specific density associated with
this material, rather it is a generic descriptor for an artificially densified bentonite mass.

Horizontal Tunnel Placement (HTP) – The UFC placement geometry proposed by NWMO for
use in a sedimentary environment. It consists of a UFC installed on its side in a deposition
tunnel excavated to hold a series of these UFCs.

In Floor Borehole (IFB) – The UFC placement geometry proposed by NWMO for use in a
crystalline rock environment where the UFC is installed in a placement borehole drilled in the
floor of the placement room. This is the same geometry as the deposition borehole defined for
the KBS-3V concept by SKB and Posiva.

KBS-3V – The in-floor, vertical canister placement geometry proposed for use by SKB/Posiva.
 - viii -

NAGRA – Translates to “National Cooperative for the Disposal of Radioactive Waste”

(Switzerland).

NWMO – Nuclear Waste Management Organisation (Canada).

Pellets – Pre-manufactured, clay-based material, typically composed of bentonite-rich clay that

is processed into uniformly-sized, high-density aggregations that are used to fill spaces and
gaps associated with the buffer and backfill.

Ps – Swelling Pressure.

TDS – Total Dissolved Solids. The quantity of soluble materials in a solution. Typically
expressed as gm/L of solution or % of solution mass.

UFC (Used Fuel Container) – The corrosion-resistant component of the engineered barriers
system used to hold the used fuel assemblies in NWMO’s repository terminology. It is the same
component as that referred to as the Canister in the SKB/Posiva concepts.

URL – Underground Research Laboratory.

 - ix -

TABLE OF CONTENTS
Page
ABSTRACT .................................................................................................................................. v
DEFINITION OF TERMS ............................................................................................................ vii 
1.  BACKGROUND .................................................................................................... 1 
2.  OBJECTIVES OF STUDY .................................................................................... 3 
3.  PREVIOUSLY COMPLETED PELLET DEVELOPMENT STUDIES .................... 5 
3.1  BACKGROUND .................................................................................................... 5 
3.2  IMPROVING AS-PLACED DENSITY OF PELLET FILL: NWMO’S 2010
PELLET STUDY ................................................................................................... 7 
3.2.1  Pellet Materials Examined ..................................................................................... 7 
3.2.2  Pellet Manufacturing Trials .................................................................................... 9 
3.2.3  Densification of Pellet Fill .................................................................................... 11 
3.2.4  Development of Pellet-Granulate Blends to Improve As-placed Density and
Thermal Conductivity ........................................................................................... 12 
3.3  CHARACTERISATION OF THERMAL PROPERTIES OF PELLET
MATERIALS: 2010 NWMO PELLET OPTIMISATION STUDY......................... 13 
3.3.1  Thermal Conductivity Evaluation ......................................................................... 13 
3.3.2  Improving the Thermal Conductivity of Individual Pellets .................................... 13 
3.4  SUMMARY OF RESULTS FROM 2010 STUDIES ............................................. 18 
4.  CURRENT (2011) STUDY .................................................................................. 19 
4.1  BACKGROUND .................................................................................................. 19 
4.2  TEST MATRIX .................................................................................................... 20 
4.3  MATERIALS AND METHODS ............................................................................ 24 
4.3.1  Materials Examined in Study (Bentonite, Illite, Silica Sand) ................................ 24 
4.3.2  Pellet Manufacturing ............................................................................................ 25 
4.3.3  Pellet Testing ....................................................................................................... 33 
5.  RESULTS OF PELLET TESTING ...................................................................... 41 
5.1  VISUAL CHARACTERISATION ......................................................................... 41 
5.2  PELLET DENSITY .............................................................................................. 43 
5.3  FREE SWELL TESTS ......................................................................................... 49 
5.4  CRUSH STRENGTH OF INDIVIDUAL PELLETS .............................................. 52 
5.5  AS-PLACED DENSITY OF POURED PELLETS ............................................... 58 
5.6  ABRASION TESTS ............................................................................................. 65 
5.7  THERMAL PROPERTIES TESTING .................................................................. 67 
5.7.1  Thermal Properties of Poured and Vibratory-Densified Pellet Fills ..................... 67 
5.7.2  Effect of Bentonite Type on Thermal Properties.................................................. 68 
5.7.3  Effect of Additives to Clay ................................................................................... 68 
5.7.4  Effect of Pellet Size on Thermal Conductivity...................................................... 68 
5.7.5  Effect of Fines Addition ....................................................................................... 68 
5.7.6  Summary ............................................................................................................. 70 
6.  DISCUSSION AND CONCLUSIONS .................................................................. 71 
ACKNOWLEDGEMENTS .......................................................................................................... 74 
 -x-

REFERENCES ........................................................................................................................... 75 

APPENDIX A: MATERIAL DESCRIPTIONS AND MANUFACTURERS' DATA SHEETS ...... 79 
APPENDIX B: RESULTS OF PELLET TESTING: DATA TABLES .......................................... 89 
 - xi -

LIST OF TABLES
Page

Table 1: Gap Fill in IFB and HTP Geometries: Parameters and Properties of Primary Interest .. 3 
Table 2: Pellets Produced using Different Compression Pressures ............................................. 9 
Table 3: Comparison of Densification Achieved Using Commercially Available and Specially-
Manufactured Pellets (after Man et al. 2011) ...................................................................... 10 
Table 4: Test and Production Matrix for Pellets for Potential use in Gap-filling (IFB and HTP
geometries) ......................................................................................................................... 22 
Table 5: Test and Production Matrix for Pellets for Possible use in Tunnel Backfill ................... 23 
Table 6: Summary of Products used in Pellet Optimisation Project ............................................ 26 
Table 7: Properties of Materials Investigated in Pellet Optimization Project ............................... 27 
Table 8: Pellet Production Matrix ................................................................................................ 31 
Table 9: Pellet Densities for Wyoming Bentonite and Bentonite-filler Materials Evaluated for
Potential Suitability for use in Gap Filling ............................................................................ 45 
Table 10: Densities Achieved for Individual Clay-only Pellets Evaluated as a Potential
Component in Tunnel Backfill ............................................................................................. 46 
Table 11: Poured and Vibrated Densities of Clay-only Pellets ................................................... 62 
Table 12: Poured and Vibrated Densities of Pellets Containing Non-bentonite Component ...... 63 
Table 13: Density and Moisture Content Results from Tests by VTT, Finland ........................... 64 
Table 14: Results of MicroDeval Test ......................................................................................... 65 
Table 15: Results of Wear Resistance Tests .............................................................................. 66 
Table 16: Thermal Conductivity of Pellet Fills ............................................................................. 69 

LIST OF FIGURES
Page

Figure 1: Vertical Container Placement Geometry (after TKS 2009) being considered by
NWMO (In-floor borehole (IFB)), SKB (KBS-3V), and Posiva. The container, a bentonite
buffer surrounding the canister and the materials used for backfilling the deposition
tunnels are shown. ................................................................................................................ 1 
Figure 2: Horizontal Tunnel Placement (HTP) Geometry considered by NWMO ......................... 2 
Figure 3: Results of an Illustrative Model showing the Effect of Thermal Conductivity of the
Engineered Clay Barrier in the HTP Method on the Required Tunnel Spacing to Prevent
Excessive Heat Build-Up....................................................................................................... 6 
Figure 4: Pellets Examined by Dixon et al. (2005) and Man et al. (2011) Pellet
Characterisation Studies (all photos are to approximately the same scale) ......................... 8 
Figure 5: Laboratory- Scale Roller Compactor used in 2010 (Man et al. 2011) and 2011 Pellet
Manufacturing Trials.............................................................................................................. 8 
Figure 6: Thermal Conductivity as a Function of Poured Dry Density Pellets ............................ 10 
Figure 7: Rebar Shaker used to Aid in Pellet Fill Densification. The vibrating component is
shown attached to a rebar in this photograph. .................................................................... 11 
Figure 8: Fine-grained (30 mesh) Wyoming Bentonite Granules used as Space-Filler in a
Gap Fill Installation Trial using HCB Pellets Produced from MX-80-2001 Bentonite
(Martino and Dixon 2007).................................................................................................... 12 
Figure 9: Thermal Conductivity Testing on Pucks of Compacted Bentonite (a) Pre-Compacted
Specimen Pucks, (b) Sample Frame and Thermal Properties Analyzer, (c) Sensor
Installed in Specimen Frame, (d) Specimen Pucks Installed on either side of Sensor
in a Two-Sided Test Configuration ...................................................................................... 15
 
 - xii -

Figure 10: Measuring Thermal Conductivity of Pellet Fill using a One-sided Test
Configuration ....................................................................................................................... 15 
Figure 11: Thermal Conductivity as a Function of Degree of Saturation for Solid Masses of
100% Bentonite; 50:50 and 60:40 Bentonite:Silica Sand Mixtures (Note Sr = Degree of
Saturation)........................................................................................................................... 16 
Figure 12: Thermal Conductivity of Solid Pucks of Material as a Function of Degree of
Saturation for a Variety of Bentonite-additive Mixtures, Compared to 100% Bentonite
at a Constant Dry Density of 1.8 Mg/m3 .............................................................................. 17 
Figure 13: Pellet Making Machine Showing Rollers .................................................................... 29 
Figure 14: Sizes of Moulds used in Making Pellets .................................................................... 29 
Figure 15: Colour Difference Between a Stock Mixture and Finished Pellets ............................. 30 
Figure 16: Examples of Pellets Produced for Pellet Optimisation Study .................................... 32 
Figure 17: Visual Appearance of Large Pellets Exhibiting Surface Defects ............................... 33 
Figure 18: Device used to Determine the Crush Strength of Pellets .......................................... 34 
Figure 19: Bulk Density Tests using Different Poured Geometries ............................................. 35 
Figure 20: Large Pellet Pouring Test Frame at VTT ................................................................... 36 
Figure 21: Abrasion Testing Equipment and Example of Results of Abrasion Testing of Clay
Pellets ................................................................................................................................. 37 
Figure 22: Test Chamber (upper) and Artificial Wetting of Pellet Fill in the
IFB Geometry (lower) .......................................................................................................... 40 
Figure 23: Types of Defects Observed in Pellets (most resulting from moisture content of
feed material) ...................................................................................................................... 41 
Figure 24: Microscopic Images of Large Pellets showing Nature of Surface Defects................. 42 
Figure 25: Effects of Clay Type and Pellet Size on Dry Density and EMDD .............................. 47 
Figure 26: Effects of Pellet Composition Dry Density and EMDD of Wyoming Bentonite-Based
Pellets ................................................................................................................................. 48 
Figure 27: Free Swell Capacity of Bentonite Clay Pellets and Effect of Solution Salinity ........... 49 
Figure 28: Free Swell Index of Wyoming Bentonite and Blends with Crushed Illitic Shale
and Fine Silica Sand ........................................................................................................... 51 
Figure 29: Changes in Swelling Pressures in Bentonite Clays as Result of Water Salinity
(EMDD of as-placed clay-only pellet fill ~0.6-0.95 Mg/m3) .................................................. 52 
Figure 30: Comparison of Average Crush Strength for Large, Clay-Only Pellets ....................... 53 
Figure 31: Comparison of Average Crush Strength for Small, Clay-Only Pellets ....................... 53 
Figure 32: Crush Strength in Newtons (N) of (a) Large, (b) Medium and (c) Small Pellets,
showing Standard Deviation of Measurements................................................................... 56 
Figure 33: Comparison of Crush Strengths of (a) Large, (b) Medium, and (c) Small, Clay and
Clay – Filler Pellets (Note: the average dry density (ρd) of the pellets are provided in the
legend of each graph) ......................................................................................................... 57 
Figure 34: Poured and Vibrated Density of Various Clay-Only Pellets. (Note: two different
individual dry densities (1.8 and 1.9 Mg/m3) for Milos AC200 small pellets) ....................... 59 
Figure 35: Poured and Vibrated Dry Bulk Densities of Pellets. (Note: two different individual
dry densities (1.8 and 1.9 Mg/m3) for Milos AC200 small pellets) ....................................... 61 
Figure 36: Results of VTT Tests to Determine Loose, As-poured Density of Pellets Installed
in a Narrow Gap .................................................................................................................. 61 
Figure 37: Comparison of Poured Dry Density Achieved in Small-Scale (AECL) and
Large-Scale (VTT) Trials ..................................................................................................... 64 
Figure 38: Relationship between Abrasion Resistance (Mass Loss) and Crush Strength .......... 66 
Figure 39: Comparison of Thermal Conductivity Values for Different Sizes and Compositions
of Pellets ............................................................................................................................. 71 
1. BACKGROUND

Pellets manufactured from bentonite clay have several applications in the placement geometries
considered by NWMO, Posiva and SKB. They can be placed immediately adjacent to the Used
Fuel Container (UFC), filling the perimeter gap between the rock and the buffer; so that they fill
large voids as part of the backfilling process and; as the main component in tunnel filling. Each
of these applications is briefly described below.

Gap Fill in Placement Boreholes

Clay pellets are being considered for two primary applications in association with the buffer
component of the repository sealing system. These applications involve being used in the
KBS-3V being proposed by Posiva-SKB, and NWMO’s In-Floor Borehole (IFB). The exact
dimensions of these gaps vary depending on the specific placement geometry but in the NWMO
concept these pellets may be used to:

1. Fill the inner gap between used fuel containers (NWMO terminology) / canisters (SKB,
Posiva terminology) and highly compacted bentonite (HCB) rings. This dimension may
be between 10 mm and 50 mm depending on the final geometry selected.
2. Fill the outer gap between the HCB buffer rings and the rock wall of the borehole. This
gap fill material must be placed uniformly in a gap approximately 50-60 mm wide. Gap
fill material will be expected to improve transfer of heat from the canister to the
surrounding materials (relative to an unfilled annular gap) and prevent the adjacent HCB
from experiencing excessive decrease in density as it swells into this region.

Backfill blocks

Bentonite
pellets

UFC

Buffer and
pellet-filled
annular gap(s)

Figure 1: Vertical Container Placement Geometry (after TKS 2009) being considered by
NWMO (In-floor borehole (IFB)), SKB (KBS-3V), and Posiva. The container, a bentonite
buffer surrounding the canister and the materials used for backfilling the deposition
tunnels are shown

Tunnel Backfill
A second region where clay pellets are proposed for use is as part of the tunnel backfill
(Figure 1). In the reference concepts for SKB and Posiva backfilling, many regions of a
repository will be backfilled using precompacted blocks of clay or clay-aggregate (filling the
 -2-

majority of the tunnel volume), with clay pellets used to fill the remaining volume. The pellet-
filled void is nominally 150 mm to 200 mm wide but actual gap dimensions will vary depending
on the tunnel dimensions achieved during its excavation.

The backfilling pellets need to be placed efficiently (relatively large volumes in short time),
uniformly and to a density that will ensure that the total sealing system maintains the specified
hydraulic and mechanical properties once water saturation and swelling-induced density
equilibration of the backfill is achieved. Pellets used in backfilling will also need to be able to
protect the other components of the sealing system (e.g., buffer, backfill blocks or in situ
compacted materials) from the erosive effects of inflowing groundwater and if possible it should
also provide a medium that will allow for uniform wetting of the system.

Materials installed in these two regions of the KBS-3V geometry do not have the same size and
thermal conductivity constraints due to the differences in the basic functions. In the buffer-filled
borehole, pellets are intended to aid in heat conduction from the container to the surrounding
rock mass and ensure the isolation function of the adjacent buffer is maintained. Pellets
installed in the tunnels associated with the repository are primarily intended to provide:
(1) physical protection to the backfill blocks; (2) aid in producing a backfill that will resist
extrusion of the buffer from the adjacent boreholes, (3) a means of providing some short-term
water retention (delaying water movement along a newly backfilled tunnel) and (4) ensuring that
the longer-term sealing properties of the overall tunnel backfill meet the requirements set for it.

Horizontal Tunnel Placement Geometry

NWMO is considering placement geometries other than the IFB, particularly for a repository that
is located in a sedimentary environment. For a sedimentary environment, the reference
placement geometry for the used fuel containers (UFCs) is known as the Horizontal Tunnel
Placement (HTP) option. In this geometry the UFCs are installed on their side in a circular
tunnel (Figure 2). In contrast to the IFB, pellets will be the primary filling material adjacent to the
UFC and so needs to have substantial thermal conduction capability as well as hydraulic and
mechanical properties more in line with a buffer material.

Figure 2: Horizontal Tunnel Placement (HTP) Geometry considered by NWMO

 -3-

2. OBJECTIVES OF STUDY

NWMO, SKB and Posiva are considering variations of the KBS-3V geometry. Bentonite-based
pellet materials are being considered as a component in the above-described repository sealing
applications. A few differences exist between the requirements identified for pellet materials in
these repository concepts (Table 1).

Most of the work done on the behaviour and characteristics of pellet materials prior to this study
focused on commercially available roller compacted or extruded bentonite pellets (Dixon and
Keto 2008, Marjavaara and Kivikoski 2011, Dixon et al. 2011b). However, depending on the
specific application and geometry considered, the use of a bentonite-only material in
manufacturing the pellets may not be necessary or desirable. Materials ranging from HCB
pellets to various compositions of HCB pellets, crushed bentonite, extruded bentonite and
admixtures with non-clay materials can be manufactured and used in a repository. There has
not however been a focussed effort to evaluate potential optimization that would improve pellet
performance and compatibility with adjacent materials as well as providing improved short- or
long-term resistance to water movement/clay erosion.

Table 1: Gap Fill in IFB and HTP Geometries: Parameters and

Properties of Primary Interest
Parameter NWMO (IFB) NWMO (HTP) SKB (IFB) Posiva (IFB)
Pellet Density  Optimise to suit  Optimise for Ps  Optimise to suit  Optimise to suit
buffer Ps and TC and TC buffer Ps and TC buffer Ps and TC
requirements requirements requirements
Pellet Size  Optimise for inner  Optimise for  Optimise for  Optimise for
and outer annular tunnel filling buffer-rock gap buffer-rock gaps
gaps in borehole  Optimise for  Optimise for
 Optimise for backfill backfill
backfill placement placement placement
Installed  Optimise to  Maximise  Optimise to  Optimise to
Density ensure buffer ensure buffer ensure buffer
performance performance performance
requirements met requirements requirements
met met
Thermal  Maximise while  Maximise while  Maximise while  Maximise while
Conductivity maintaining other maintaining maintaining maintaining
properties other properties other properties other properties
Pellet  Erosion resistant  Erosion  Erosion resistant  Erosion resistant
Durability  Mechanical resistant  Mechanical  Mechanical
stability  Mechanical stability stability
stability
Pellet  HCB  HCB  HCB-buffer  HCB-buffer
Composition  HCB + filler?  HCB + filler?  HCB-blend  HCB-blend
 Low quality  Low quality
bentonite bentonite
 Backfill clay  Backfill clay
Note: Ps = swelling pressure
TC = thermal conductivity
 -4-

The work reported on in this document has focussed on pellet materials generated using
mechanical roller compaction technologies. This technology produces round, oblong or lozenge
shapes depending on the design of the dies used in the compaction machinery. Preliminary
work done to evaluate options for improving pellets through blending of more conductive
materials with the bentonite prior to pellet manufacture and the potential to improve the
individual pellet densities was completed for NWMO in 2010. Some of these results were
presented by Man et al. (2011) and are briefly summarised in Section 3. These data were also
used in developing the work plan for the current project.

Pellets can also be produced as rod-shaped products of fixed cross-section but varying length
using extrusion technologies. The production, performance and characterisation of these
extruded materials are the subject of a complimentary study being undertaken by Posiva. This
work is in progress and should be published as a Posiva Working Report in 2013.

Trial pellets are also being produced and tested by SKB. Pellets being considered include
extruded pellets and pressed (6 mm diameter) pellets made from bentonite sourced from
IBECO and Asha. Their testing program includes erosion testing, water storage capacity, and
installation trials to evaluate durability. The results of this work will be reported in 2012.

The manufacturing and testing program documented below has produced and tested a variety
of clay-based pellet materials with the goal of identifying if a process of optimizing size, shape,
water content, density and composition of pellets for both gap fill and backfill applications will
yield substantial improvement in system behaviour. The project builds on work conducted by
AECL for NWMO in 2005 (Dixon et al. 2005) and 2010 (Man et al. 2011), which is summarised
in Section 3. The methods used in pellet production and evaluations are described in Section 4
while the results are presented in Section 5.

Specific questions that have been addressed in the current study are:
 What additives (in what proportions) can be included to increase the thermal conductivity
of as-placed pellet materials without sacrificing adequate swelling and hydraulic
behaviour?
 What does varying pellet composition and water content have on their strength?
 What are the properties of pellets made from bentonites of differing sources?
 What are the benefits of rolled pellets versus extruded pellets (if any)?
 -5-

3. PREVIOUSLY COMPLETED PELLET DEVELOPMENT STUDIES

3.1 BACKGROUND

Work to evaluate the basic characteristics of pellet fill materials have been documented in
several reports (Dixon et al. 2005, Dixon and Keto 2008, Dixon et al. 2011b, Hansen et al. 2009,
Man et al. 2011, Marjavaara and Kivikoski 2011). The information contained in these
documents was used in developing the scope of work covered in the current study.

In the IFB geometry, the presence of pellet fill acts to improve the thermal conduction of the
buffer-filled borehole by providing a physical bridge through which heat can flow, as opposed to
having only an insulating air gap. The importance of this fill material can be seen if the
placement geometries, buffer density and conditions developed in the course of system
equilibration for the various IFB geometries are considered. For example:

 NWMO’s IFB geometry calls for a uniform as-placed dry density of 1.61 Mg/m3 (at 65%
initial degree of saturation) for the highly compacted bentonite (HCB) in the volume
surrounding the used fuel container (excluding annular gaps). This HCB will have an
initial thermal conductivity of ~1 W/(m·K) as discussed in Section 3.3. Following
container installation, the regions closest to the heat-generating UFC can be expected to
desiccate to some degree and in a dry borehole condition, the thermal conductivity of the
intact buffer can therefore be expected to decrease to ~0.5 W/(m·K) or less (Man et al.
2011). If left unfilled, the adjacent air-filled gap will only have a thermal conductivity in
the order of 0.024 W/(m·K). Any loss of dry density (e.g., partial swelling, cracking,
erosion) will further reduce the overall thermal conductivity of the region surrounding the
UFC. On completion of HCB hydration and swelling to fill the entire IFB volume not
occupied by the UFC (assuming no pellets were installed in the inner or outer gaps), the
system will reach an equilibrated dry density of approximately 1.6 Mg/m3. This
corresponds to a material having an average thermal conductivity of ~0.7 W/(m·K) after
swelling. Pellets installed in the gaps between the HCB and the adjacent rock/UFC will
provide a thermal bridge allowing better heat transfer away from the UFC. These pellets
will also result in a higher equilibrated dry density of the IFB clay fill and a higher thermal
conductivity of the buffer over the long-term. Hence the consideration for filling the gap
between the UFC and HCB rings in NWMO’s, SKB’s and Posiva’s concepts.

 In Posiva’s Detailed Design of the IFB concept, the HCB is targeted to have an average
buffer dry density 1.60 Mg/m3 at 100% saturation (26.5% gravimetric moisture content at
saturation, void ratio =0.72) on achieving density equilibration (Juvankoski 2011). It
should be noted that Posiva’s design calls for differences to exist in the original densities
of the HCB, depending on the location within the borehole. The disks located above and
below the container are to be 1.99 Mg/m3 dry density (75% initial degree of water
saturation), while the rings surrounding the container are to be 2.05 Mg/m3 at 82% initial
degree of saturation. The pellets placed in the outer gap are defined to be 1.075 Mg/m3
dry density at 24% saturation.

The net result of the use of pellets to fill the gap(s) present in the IFB is an improved
equilibrated density for the materials surrounding the UFC. The degree to which improving or
modifying the density (or thermal characteristics) of the pellets is possible and how this will
affect the thermal conductivity of the buffer surrounding the UFC has not yet been assessed and
is the topic of the study described in Section 4 and 5 of this document.
 -6-

Prior to initiation of the work described in this report, a study whose primary focus was to
evaluate compositional options for pellets to be used in both gap filling in the IFB and the HTP
geometries being evaluated by NWMO was completed (Man et al. 2011). Maximizing the
density to which pellet fill can be placed as part of the backfilling process is also a goal that is
covered by this pellet optimisation study. For the purposes of background documentation and
to put into context the work done in 2011 as NWMO’s contribution to a joint research project
with Posiva and SKB, some of the key results of previous pellet studies are presented below.

The 2010 study included preliminary evaluations of the potential means of manufacturing pellet
materials with superior thermal conduction characteristics. This was done by examining the
effects of blending bentonite clay with other, more thermally conductive materials and then
evaluating their thermal conduction characteristics. The results of this study were presented by
Man et al. (2011) and were used in developing the pellet optimisation work initiated by NWMO,
SKB and Posiva in 2011 (Sections 4 and 5).

To begin the process of determining what the key parameters affecting UFC spacing in the HTP
geometry, a numerical model of the HTP method was constructed for the purpose of estimating
the influence of thermal conductivity of the engineered clay barrier. It should be noted that this
is only a hypothetical model to illustrate the influence of parameters, and some of the values are
not necessarily in accordance with NWMO, SKB and Posiva programs. The results of the
modelling exercise are provided in Figure 3. For a maximum surface temperature at a UFC of
125ºC, a thermal conductivity of 0.5 W/(m·K), will require a spacing of approximately 17 m. If
the thermal conductivity of the engineered clay barrier can be increased into the range of
0.7 W/(m·K) to 0.9 W/(m·K), the required spacing for the HTP concept could be reduced to
approximately 16 m and 14 m, respectively.

300
Maximum Temperature at the Container 

250 k = 0.5
k = 0.7
200 k = 0.9
Surface (oC)

150

100

50
0 10 20 30 40

Spacing (m)

Figure 3: Results of an Illustrative Model showing the Effect of Thermal Conductivity of

the Engineered Clay Barrier in the HTP Method on the Required Tunnel Spacing to
Prevent Excessive Heat Build-Up
 -7-

None of the pellet or pellet granule materials evaluated in the 2010 study achieved a thermal
conductivity greater than 0.7 W/(m·K) (Section 3.2.4). Although a reduction in container spacing
from 17 to 14 m may appear small, when factored over the total number of UFCs in the NWMO
repository, a significant reduction in total tunnelling length and repository area may be realized.
This needs to be compared to the effort and cost of placing tunnel fill at greater densities in a
radiological environment.

3.2 IMPROVING AS-PLACED DENSITY OF PELLET FILL: NWMO’S 2010

PELLET STUDY
3.2.1 Pellet Materials Examined

The NWMO-sponsored evaluation of commercially-available bentonite pellets began with

determining the effects of pellet size, shape and density on the placement density of pellet fill in
the HTP geometry, specifically, their mechanical and thermal characteristics (Man et al. 2011).
The commercial bentonite pellets tested included ~9 mm- and ~13-mm-diameter cylindrical,
oblong shaped pellets of the type examined by Dixon et al. 2005 and subsequently by Man et al.
(2011); square-shaped pellets and extruded (Cebogel) pellets as shown in Figure 4. While
focused on HTP application, much of the information generated has application to materials and
methods for placement of pellet fill in the IFB annular spaces and the gap between backfill
blocks and the surrounding geosphere.

The commercially-sourced materials were compared to pellets produced at AECL’s

Geotechnical Laboratory using a laboratory-scale roller compactor (Figure 5). This device
allowed for production of materials using a range of compressive forces, water contents and
physical compositions. This machine could make two sizes of pellets by using interchangeable
rollers. The smaller rollers produced pellets of 10 mm width and 23 mm in length, and larger
rollers pellets of 20 mm of width and 40 mm in length as shown in Figure 4. Both of these pellet
types were ~10 mm thick.
 -8-

Figure 4: Pellets Examined by Dixon et al. (2005) and Man et al. (2011) Pellet
Characterisation Studies (all photos are to approximately the same scale)

Figure 5: Laboratory- Scale Roller Compactor used in 2010 (Man et al. 2011) and 2011
Pellet Manufacturing Trials
 -9-

3.2.2 Pellet Manufacturing Trials

The production of bentonite pellets using roller compaction allows for variation of several
parameters during the manufacturing process. Trials for manufacturing pellets from bentonite
clay were done to establish the factors controlling the density achieved (e.g., changing feeding
rates, roller speeds and compression pressures (force with which rollers are kept in contact),
water content of the feed material), type of feed material. Additionally, the use of recycled
materials (pellets that are crushed and fed back into the compactor) was evaluated to determine
if an initially denser feed material resulted in improved pellet densities.

Table 2 shows a comparison of the lab-manufactured bentonite pellets produced using a

200-mesh Wyoming bentonite (BC-NSB-200-2008), under three compression pressures. The
bentonite pellets listed as “Recycled” were produced by roller-compacting crushed bentonite
pellets produced in the first compaction cycle. The crushing of the pellets to fine-grained feed
for a second compaction cycle was done to evaluate what effect using a higher-density source
material would have on the pellets.

Table 2: Pellets Produced using Different Compression Pressures

~10 (1450 psi) ~12 (1800 psi) 10.3 (1500 psi)

Compression Pressure
(MPa) Single Recycled Single Recycled Single
cycle pellets cycle pellets cycle
Bulk Density (Mg/m3) 2.12 2.15 2.13 2.17 2.14
3
Dry Density (Mg/m ) 1.95 1.98 1.96 2.00 1.97
Grav. Water Content (%) 9.11 8.71 8.58 8.63 8.76

It was found that there was little difference in achievable density for compression pressure of
~10 MPa (1450 psi) and ~12 MPa (1800 psi) as shown in Table 2. A compression pressure of
~10.3 MPa (1500 psi) was therefore chosen for use in producing pellets in this study. The
production rate of the machine under this pressure was about ~2 kg pellets per minute. The
roller compression process generates considerable heat, resulting in water content decrease of
pellets, a factor that needed to be taken into account with respect to ensuring that the pellets
retained their physical integrity once they had cooled.

Table 3 presents a summary of the as-placed properties of the various pellet materials
examined in the 2010 study (Man et al. 2011). These data show that the laboratory-produced
pellets have higher dry density than commercially available bentonite pellets (1.90 Mg/m3 versus
1.7-1.8 Mg/m3) and higher as-placed bulk thermal conductivities at different densities (0.48 and
0.56 W/(m·K) at loose and shaken conditions, respectively, Figure 6). Even with the improved
density and thermal properties achieved using denser pellets (and addition of fines to fill inter-
pellet gaps), the thermal conductivity of these materials were below 0.7 W/(m·K) (Section 3.2.4).
 - 10 -

0.6
3/8" Volclay Pellet-Fill
1/2" Volclay Pellet-Fill
0.55

Thermal Conductivity
SKB Square MX-80 Pellet-Fill
BACLO Pellet-Fill
0.5

(W/(m·K))
AECL Oblong Narrow Pellet-Fill
AECL Oblong Wide Pellet-Fill
0.45

0.4

0.35

0.3
0.50 0.75 1.00 1.25
Dry Density (Mg/m3)
Figure 6: Thermal Conductivity as a Function of Poured Dry Density Pellets

Table 3: Comparison of Densification Achieved Using Commercially Available and

Specially-Manufactured Pellets (after Man et al. 2011)
1/3-Scale
++
Oblong Mockup
3/8” ½” BACLO BC-NSB-200- BC-NSB-200-
Square
Volclay Volclay Cebogel 2005/8 2008
(26x10x5 mm) (22x14x10
mm)
Bulk Density of each pellet
1.86 1.82 1.92 2.10 2.14 2.11 (1.90+)
(Mg/m3)
Dry Density of each pellet (Mg/m3) 1.72 1.67 1.81 1.83 1.96 1.90 (1.71+)
As-Placed Bulk Density (Mg/m3)* 1.05 1.07 1.08 1.13 1.12 1.12
3
As-Placed Dry Density (Mg/m )* 0.97 0.98 1.01 1.04 1.03 1.03
3
Bulk Density by Shaker (Mg/m )* 1.18 1.20 1.21 1.30 1.32 1.30
Dry Density by Shaker (Mg/m3)* 1.09 1.10 1.12 1.19 1.21 1.20
3
Bulk Density by Vibrator (Mg/m )* - - - - - 1.22 (1.43**)
3
Dry Density by Vibrator (Mg/m )* - - - - - 1.12 (1.31**)
Thermal Conductivity (W/(m·K))
0.41 0.40 0.44 0.46 0.45 0.48
(as-placed condition)
Thermal Conductivity (W/(m·K))
0.36 0.49 0.49 0.53 0.51 0.56
(shaken condition)
* - maximum practically-achievable density using pellet-only fill, ** - achievable dry density using a
modified probe in a mock-up trial and included addition of 80-mesh bentonite fines to pellets to fill large
inter-pellet voids, + - Large size pellets; ++ Extruded bentonite rod-shaped pellets used as gap fill in tunnel
backfilling trials (Dixon et al. 2008a,b; 2011).
 - 11 -

3.2.3 Densification of Pellet Fill

Two compaction techniques (electric shaker and needle-type concrete vibrator), were examined
in an effort to develop a practical means of densifying a large volume of pellet fill. Pellets were
subjected to a shaking agitation and this produced a higher as-placed dry bulk density
(1.20 Mg/m3) than was achieved using a modified small concrete vibrator (1.12 Mg/m3)
(Table 3). Despite achieving a higher as-placed density using a shaker, the shaker-type
densification does not readily transfer to a full-scale tunnel placement or installing pellets into
the annular spaces of the IFB. The vibrating technique was therefore deemed to be more
representative of what can be achieved in a field application. The vibratory technique was
subsequently further developed by use of a concrete rebar vibrator (Figure 7) rather than a
vibrating concrete probe and this modified technique was used in the installation of a 1/3-scale
HTP mock-up. In the 1/3-scale HTP mock-up, rebar lengths were installed horizontally in the
volume of pellets to be backfilled and the pellets were vibrated in layers as the mock-up was
installed. This technique is described further in Section 3.2.4 and a visual example of the
results of this process is presented in Figure 8. This modified technique improved the
densification of the pellet and granule mass to 1.31 Mg/m3, as-placed dry density. This was
slightly lower than the target dry density of 1.41 Mg/m3 (needed to provide the desired thermal
conductivity for the HTP fill).

Figure 7: Rebar Shaker used to Aid in Pellet Fill Densification. The vibrating component
is shown attached to a rebar in this photograph
 - 12 -

a. As-placed Pellets b. Fines and New Layer c. Following Vibratory

of Pellets added Compaction

Figure 8: Fine-grained (30 mesh) Wyoming Bentonite Granules used as Space-Filler in a

Gap Fill Installation Trial using HCB Pellets Produced from MX-80-2001 Bentonite
(Martino and Dixon 2007)

3.2.4 Development of Pellet-Granulate Blends to Improve As-placed Density and

Thermal Conductivity

As noted above, in order to maintain the thermal conduction needed to limit temperatures at the
container surface to acceptable levels and reduce the container spacing needed in the HTP
geometry, the as-placed dry density of bentonite-only fill material needs to exceed ~1.41 Mg/m3.
To accomplish this some means of filling the spaces between the pellets to improve thermal
conductivity is required. The thermal conductivity of the poured pellet-fill material was less than
0.5 W/(m·K) for all of the tested pellets. Light compaction increased the thermal conductivity of
the pellet-fill material to a maximum of 0.56 W/(m·K) for the wide pill-shaped pellets made by
AECL (Section 3.3).

To evaluate the possibility of further increasing the as-placed density and thermal conductivity, a
needle-type concrete vibrator was used in a limited set of trials for the AECL wide oblong
pellets. The vibrating technique is representative of what might be achievable at full-scale and a
dry bulk density of 1.3 Mg/m3 was achieved using this device. While an improvement in as-
placed density, the results indicate that a single-size pellet material will not be able to be
densified sufficiently to reliably achieve the desired thermal conductivity for use in the HTP
geometry. This leaves three additional possible approaches to modifying the pellet fill so that it
achieves the desired thermal characteristics:

(1) Using a graded pellet – fines material that provides better packing density,
(2) Improving the thermal characteristics of the pellets through compositional adjustment of
the pellets themselves (Sections 3.3, 4 and 5), and
(3) Applying a combination of approaches 1 and 2.

In the studies undertaken in 2005 (Dixon et al. 2005) and 2010, Man et al. (2011), approach (2)
above was evaluated by using granules of bentonite (MX-80 bentonite (nominally an 80-mesh
size)) in conjunction with single-sized pellets and vibratory densification. Pellets were installed
in layers, followed by pouring sufficient MX-80 fines on the top of the pellet layers. In a dry
environment these fines will flow through the pellets and fill the open pore spaces when
vibrated. This placement technique was used by Dixon (2005). In that study, the pellet-fine
grained bentonite ratio was 70:30 (pillow-shaped pellets) and 80:20 (elongated pellets) by mass
 - 13 -

and as-placed dry densities in the order of 1.42 to 1.46 Mg/m3 were achieved when a modified
concrete pencil vibrator was used in the installation process to encourage fines movement and
optimised pellet orientation.

3.3 CHARACTERISATION OF THERMAL PROPERTIES OF PELLET MATERIALS:

2010 NWMO PELLET OPTIMISATION STUDY
3.3.1 Thermal Conductivity Evaluation

Section 3.2 identified some of the potential limitations associated with tunnel fill in the HTP
geometry, but also demonstrated that it was possible to place gap fill materials into the relatively
confined spaces associated with the IFB geometry and tunnel backfilling. While thermal
characteristics are less important in the backfill components of the repository sealing system,
they do have an influence in regions close to the UFC. Means to improve the thermal
conduction characteristics of pellet fill while not adversely affecting the other performance
characteristics needed is an important goal (e.g., low hydraulic conductivity, high swelling
pressure, resistance to erosion by flowing water).

The material used to manufacture gap fill and HCB-based buffer materials in the 2010 and 2011
studies is Wyoming bentonite clay (BC-NSB-200-2008: Bentonite Corporation-National
Standard Bentonite, 200 mesh Wyoming-type bentonite). Testing to determine the effects of dry
density and gravimetric water content have on the ability of bentonite-based materials to
conduct heat was undertaken. In these tests solid “pucks” of bentonite-based material were
compacted (dry densities of 1.0 to 1.8 Mg/m3) and tested using the thermal properties
measurement device shown in Figure 10. As Highly Compacted Bentonite (HCB) will be
located next to the container, an understanding of its thermal behaviour is critical, particularly
since a period of thermally-induced desiccation is likely. These data also provide information on
the thermal characteristics of individual pellets that would be used to fill gaps between the
various elements of the sealing system.

3.3.2 Improving the Thermal Conductivity of Individual Pellets

Measurement of Thermal Conductivity

The first step in evaluating potential means of improving the thermal characteristics of pellet
materials is to determine what effect additives will have on the properties of a single pellet.
From this it is possible to evaluate what the effect might be on a mass of pellets. Thermal
properties testing was undertaken using a Hot Disk Thermal Constants Analyzer1 (Figure 9 and
Figure 10). The system operates by supplying a pulse of constant heat to a sample sensor,
which acts as both a heat source for increasing the temperature of the sample and a resistance
thermometer to monitor the change in temperature after the heat pulse. Details of this
technique are provided in Section 4.

Effects of Blending Materials

The scoping studies conducted in 2010 included evaluating a range of materials considered for
use in the repository sealing system. This was in part to provide improved thermal data for the
various components for use in future modelling exercises. As a result considerable information

1
Hot Disk Constants Analyzer, TPS2500, manufactured by Hot Disk AB, Chalmers Science Park, Chalmers University of
Technology, Sven Hultins gata 9 A, SE-412 88 Gothenberg, Sweden
 - 14 -

not necessarily relevant to the development of gap fill or backfill pellets was developed. This
information is not discussed in detail in this report. However these tests did provide some initial
quantification related to the effects of composition and blending on the thermal characteristics of
clay-based materials. From this it was possible to complete an initial assessment of the
potential for improving the thermal characteristics of pellet fill.

Figure 11 shows the thermal conductivity data for a solid mass of 100% Wyoming bentonite as
a function of degree of saturation. These data show that thermal conductivity of bentonite-only
material increases with density and degree of saturation but only by a small amount. The range
of thermal conductivity is only ~0.25 to ~1.0 W/(m·K). When sand is added to the bentonite
prior to compaction of the mass, a dramatic increase in the system’s thermal conductivity is
evident.

The thermal conductivity data for 50:50 and 40:60 mixtures of bentonite and silica sand as a
function of degree of saturation and dry density are also provided in Figure 11. These are
materials that have been examined for use as a Bentonite Sand Buffer (BSB) rather than an
HCB buffer in AECL’s IFB concept in the 1990’s as well as a component in a tunnel plug
(Chandler et al. 2005, Martino et al. 2008) and shaft seals (Martino et al. 2011, Dixon et al.
2011c). Although the grain-size of the sand component used in these materials is larger than
could be used in pellet manufacturing, these data provide valuable insight into what most affects
thermal conductivity in bentonite-based materials and are likely representative of the behaviour
of larger masses of material.

In Figure 11, the thermal conductivity measurement of a 40:60 mixture of bentonite and granitic
sand shows the effect of different types of filler sand on thermal conductivity of bentonite sand
mixtures (BSB contained silica sand while the ESP backfill was produced from sand originating
from granitic rock). Although the aggregate component was higher (60 versus 50%), at similar
dry densities the bentonite-granitic sand mixture has a lower thermal conductivity. This is
attributed to the feldspar fraction in the granitic sand, which has a lower specific gravity and
thermal conductivity (~2 W/(m·K)) than quartz (~3 W/(m·K)).

The low thermal conductivities measured for HCB (at low degrees of saturation) and blends of
bentonite and sand or granite aggregate shown in Figure 11, particularly under low degree of
water saturation led to an investigation of the effects of other potentially compatible additives to
bentonite. Specifically, a natural titanium dioxide (rutile) and copper powder were evaluated for
their effect on the thermal conductivity of a bentonite-based mass. The results of those tests
are shown in Figure 12. Improved thermal conductivity at low degrees of saturation (e.g.,
< 40%), were observed for materials containing silt-sized silica sand or copper powder. It is
important to note that it is at a low degree of saturation where improved thermal conductivity is
needed. At a degree of saturation above ~40%, the effect of increased moisture surpasses the
effect of the copper powder, and copper-amended materials show thermal conductivity values
approximately the same as those of 100% bentonite. Mixtures containing silica sand showed
higher thermal conductivities than bentonite-only materials where the degree of saturation
exceeded ~60% (Figure 12). The relative improvement of silica sand over copper powder for a
given mass ratio is attributed to the smaller volume of copper relative to silica sand for a given
mass ratio (specific gravity of copper vs. silica sand). The bentonite-silica sand blend would
have a higher degree of sand-sand particle contact and hence better thermal conductivity
 - 15 -

(a) (b)

(c) (d)

Figure 9: Thermal Conductivity Testing on Pucks of Compacted Bentonite (a) Pre-

Compacted Specimen Pucks, (b) Sample Frame and Thermal Properties Analyzer,
(c) Sensor Installed in Specimen Frame, (d) Specimen Pucks Installed on either side of
Sensor in a Two-Sided Test Configuration

Figure 10: Measuring Thermal Conductivity of Pellet Fill using a One-sided Test
Configuration
.
 - 16 -

100 % Bentonite 50:50 Bentonite : Sand

40:60 Bentonite : Sand

Figure 11: Thermal Conductivity as a Function of Degree of Saturation for Solid Masses of 100% Bentonite; 50:50 and 60:40
Bentonite:Silica Sand Mixtures (Note Sr = Degree of Saturation)
 - 17 -

Figure 12: Thermal Conductivity of Solid Pucks of Material as a Function of Degree of

Saturation for a Variety of Bentonite-additive Mixtures, Compared to 100% Bentonite at a
Constant Dry Density of 1.8 Mg/m3

Beyond the thermally-related characteristics of bentonite-additive materials, the additives

reduce the mass of bentonite per unit volume at a specific dry density, which in turn reduces the
effective montmorillonite dry density (EMDD). EMDD is defined as:

EMDD = Mm / (Vm + Vv) or EMDD= (MT * M)/ (VT-((Mnm/Gnm))

Where Mm = dry mass of montmorillonite clay; Vm = volume of the montmorillonite minerals; and
Vv = volume of voids, MT is the total mass of the specimen, M is the montmorillonite content, Mnm
is the mass of the non-montmorillonite component(s) and Gnm is the specific density of the non-
montmorillonite component(s). Calculation of this parameter requires that the montmorillonite
content be known. The non-montmorillonite component(s) must also be defined so that their
volume can be removed from the calculation.

EMDD values, provided for each material in the plots, provide a means of estimating the
swelling pressures and hydraulic conductivity of water-saturated materials. Swelling pressure
increases and hydraulic conductivity decreases with increasing EMDD. These changes also
need to be taken into account when evaluating the potential advantages of using a blended fill
material.

Thermal Conductivity of Bentonite Pellets

Based on the result of the screening study to evaluate how thermal conductivity could be
affected by use of additives such as copper and titanium dioxide, it was found that there was
little improvement in the thermal characteristics of bentonite-based materials beyond what could
be provided by more available and less expensive materials such as silica or granitic sand.
Hence, no further evaluations were undertaken for copper or titanium dioxide with respect to
 - 18 -

manufacturing of pellets. The 2010 pellet placement study therefore focussed on bentonite-only
pellet materials and evaluating their thermal characteristics.

As presented in Sections 3.2 and 3.3 (Figure 6), six types of clay pellet-fill materials were
screened during 2010 in order to evaluate their potential for use as gap-fill materials in locations
where thermal conductivity is an important parameter. These included four commercially
available pellets and two types of pellets developed by AECL. The commercially available
bentonite pellets included 3/8” (~9-mm) and 1/2” (~13-mm) cylindrical pellets made of Volclay
(Wyoming) bentonite, square-shaped pellets made of MX80-2008 Wyoming bentonite from SKB
(16-mm long x 16-mm wide x 8-mm thick), and extruded pellets made of activated Milos
bentonite (Cebogel-2010) from the BACLO (“BAckfilling and CLOsure of a deep repository”
project conducted by SKB and Posiva).

3.4 SUMMARY OF RESULTS FROM 2010 STUDIES

Bentonite-based materials were prepared at a range of densities and saturations with distilled
water and their thermal properties measured. The results indicate lower than desired thermal
conductivities for engineered clay barriers that will be in close proximity to heat generating
UFCs for the HTP geometry considered by NWMO. Specifically, the thermal conductivity of
HCB at low saturations and pellet-fill material is close to 0.5 W/(m·K). This will result in the
requirement for larger UFC spacing to prevent excessive temperatures at the surface of a given
UFC. Increasing the thermal conductivity of these materials into the range of 0.7 W/(m·K) to
0.9 W/(m·K) will result in a decrease in UFC spacing in the order of up to 2.5 m. These
relatively low thermal conductivities are not as critical where pellets are to be used to fill annular
gap(s) present in the IFB geometry since the alternative is to leave an open (air) gap which is
much less desirable thermally.

Preliminary work was conducted to examine the effect of adding various admixtures to HCB with
the goal of increasing thermal conductivity. The addition of small amounts of silica sand and
copper powder resulted in an increase in thermal conductivity to between 0.7 W/(m·K) to
0.9 W/(m·K). This occurred at low degrees of saturation, where the increase is needed most
(i.e., during intense heating by UFCs and associated drying). However at higher degrees of
saturation (>60%) copper-amended materials did not show thermal characteristics much
improved from those of bentonite-only materials. Sand-amended bentonite showed a much
improved thermal conductivity over bentonite-only materials for degree of water saturation
>60%. It should be stressed that this increase in thermal conductivity was achieved using puck-
shaped specimens of HCB-additive mixtures. Further testing of pellets with the added variable
of inter-pellet porosity required further investigation.

The used bimodal mixtures of pellets and fine grained bentonite aggregates to fill the porosity
between pellets and therefore increase density and thermal conductivity showed some promise.
Previous work by Dixon et al. (2005) indicated that use of a 2-size material could improve the
density achieved (1.42-1.46 Mg/m3 dry density achieved in laboratory tests) but this is only
marginally acceptable with respect to thermal characteristics needed for the HTP, and not
particularly feasible for the more constrained IFB geometry. This technique has also been
examined by NAGRA in the form of a mixture of pellets, granules and fines as part of their HTP
research (Volckaert 2000, De Bock et al. 2008). This technique achieved some improvement of
the as-placed density (1.38 – 1.51 Mg/m3), but as with the 2010 study described above, this will
provide a fill whose thermal conductivity is in the lower range of acceptability.
 - 19 -

4. CURRENT (2011) STUDY

4.1 BACKGROUND

The objectives of the joint pellet optimization study supported by NWMO, Posiva and SKB
during 2011 have been described in Section 2. The components of the 2011 study are largely
extensions and refinements of the work completed in 2010 described in Section 3 and
published, in-part by Man et al. (2011), as well as additional materials and scope associated
with specific needs of SKB and Posiva.

The work completed in 2010 for NWMO determined that:

1. There was little improvement of the thermal characteristics of pellet materials through
use of exotic materials such as copper powder or minerals such as rutile.
2. The use of a crushed or natural fine-grained silica powder showed substantial
improvements could be achieved in the thermal conductivity of individual pellets and so
further assessment of that option was warranted.
3. Particle packing for pellet materials was very dependent on the size and shape of the
pellets and so a larger range of sizes and shapes should be evaluated.

Questions related to the materials that could physically resist the erosive action of flowing
groundwater in the annular spaces between the HCB and rock in the IFB in the period prior to
buffer swelling and closure of this gap were raised and are initially addressed in the new work
described below. This issue is also relevant to the pellet materials to be placed in the gap
between the backfill blocks and tunnel walls of a repository. The source of the bentonite (and
perhaps also the quality required) to be used in manufacturing of pellets for different
locations/applications is not the same throughout a repository. The various organisations
represented in this joint project are also considering different clay sources/suppliers for their
bentonite (and swelling clay) components. As a result, several bentonite materials were
included in this study in order to determine how the source/quality of bentonite might affect
pellet manufacture and behaviour.

Additionally, there exists the potential to use extruded rather than roller-compacted bentonite
pellets in various locations in a repository. These extruded clay pellets are being evaluated as
part of the Pellet Optimization Project, with most of the work related to that material being
undertaken by Posiva. Many of the same issues and questions related to material composition
options, durability and mechanical performance for the roller-compacted materials are present
for the extruded materials. Some of the initial work conducted AECL’s Geotechnical Laboratory
has been included in this report, but most will be provided in a report to be produced by Posiva
by early 2013.

Beyond the issues described above, consideration of materials that could provide resistance to
longer-term removal of bentonite from the IFB by providing a filter medium along the rock-HCB
interface was also identified as a topic of interest.

Two materials were identified as being potentially suitable as additives to bentonite during the
pellet manufacturing process. Fine-grained silica and illite clay were both evaluated with
respect to their thermal conductivity, as-placed density and ability to resist mechanical erosion.
 - 20 -

4.2 TEST MATRIX

As part of the manufacturing and optimization study, a number of laboratory tests were
conducted in order to identify key parameters and properties of pellets produced at AECL’s
Laboratory. These key parameters and properties of these pellets evaluated by AECL are listed
as follows:
 Determination of density of individual pellets;
 Pouring trials to determine bulk density of pellets and evaluate ability of materials to be
placed effectively;
 Effect of minor vibration on bulk density of poured pellets;
 Durability of pellets (e.g., visual evaluation of pellets, crush strength);
 Thermal conductivity of as-placed, poured pellet mass (using NWMO equipment located
at AECL’s Geotechnical laboratory); and
 Free swell tests involving placing materials in artificial groundwater solutions (0, 35,
70 and 275 g/L TDS, the latter being equivalent to NWMO’s reference groundwater for a
sedimentary environment) to evaluate effect of material type and admixtures on rate of
pellet breakdown.

Several different bentonite materials of interest to NWMO, SKB and Posiva were examined as
well as illite clay and fine silica sand fillers. Filler contents of 10, 25 and 50% by weight were
tested in order to assess the potential utility of such materials in filling gaps between the UFC
and the rock in the In-Floor Borehole geometry as well as in tunnel backfilling applications.
Stock materials selected for pellet production included:
 Wyoming bentonite;
 Asha bentonite;
 Milos AC200 bentonite;
 Milos B bentonite;
 Wyoming bentonite blended with 10, 25 or 50% silica sand;
 Wyoming bentonite blended with 10, 25 or 50% illitic clay.

The above stock materials were used to produce three different sizes (and shapes) of pellets
using a laboratory scale briquetting machine. These included:
 Large (Oblong: 22 mm long x 14 mm wide x 10 mm thick);
 Medium (Oblong: 22 mm long x 11 mm wide x 7 mm thick); and
 Small (Round: 9 mm diameter x 6 mm thick).

A total of 56 batches of pellets were made for the study. Based on preliminary characterisation
results six pellet types were selected for additional advanced testing by VTT in Finland, which is
on-going. The selected pellets included:
 Large pellets made with Wyoming bentonite;
 Small pellets made with Wyoming bentonite;
 Small pellets made with 75% Wyoming bentonite and 25% silica sand;
 Small pellets made with 75% Wyoming bentonite and 25% illite;
 Small pellets made with Milos AC200 bentonite; and
 Small pellets made with Milos B bentonite.

The pellets were assessed at AECL primarily to evaluate potential to improving their density,
strength, durability and heat transfer capability.
 - 21 -

Table 4 presents a test matrix that focuses on materials that could be of use as gap fill pellets
for the IFB or HTP geometries. A total of 38 variations of composition, water content and
compaction load were examined in the course of testing materials for use in “buffer-type”
applications. Values of water content, individual pellet bulk density, crush strength, and free
swell were determined for all pellets produced. On the selected pellets, thermal conductivity
tests were carried out, which also provided poured and vibrated densities for these materials.
Table 5 presents a test matrix for pellets of particular interest in backfilling of tunnels associated
with the block and pellet backfilling concept. This matrix of 18 test variables included three
types of clay and two different pellet sizes. In addition, Cebogel extruded pellets, Pellets
produced by MX-80 in 2008 and MX-80 pellets produced in 2011 (used in Posiva’s Buffer Test
project) were examined as part of the AECL testing program.

Beyond the work completed at AECL, there were other associated tests that were considered as
part of the study scope. Based on the results of the pellet production and testing conducted at
the AECL laboratory, selected pellets were produced and sent for further detailed testing at the
SKB and Posiva laboratories. This work was proposed to be included, as part of the final
project report to provide additional depth to this project. Six of the test batches produced by
AECL (Batch No. 51, 52, 53, 54, 55 and 56), were selected by Posiva and SKB for further
evaluation in their laboratories. Approximately 75 kg of each of these batches were produced
and shipped to Europe.

SKB and Posiva conducted additional material testing and property analyses on pellet materials
produced in the course of this study, including:
 Large wall gap filling tests to evaluate pour density of customized pellets;
 Artificial wetting of gap using customized pellet sizes and composition (assessing
sealing ability/time, swelling pressure, vertical uplift risks, moisture and density
distribution with time);
 Void ratio and water flow ability with customized pellet size (assessing permeability and
ability to use artificial wetting);
 Buffer system design (achievable densities with customized pellet size and composition);
 Durability tests on pellets (MicroDeval Wear Resistance Test) to evaluate handling and
fracture risks;
 Conduct of extruded pellet test production using “optimised” materials identified in roller-
compaction production and evaluation to provide comparison between products
produced by two techniques;
 Erosion tests; and
 Water storage capacity tests.

On completion of the pellet studies, a limited number of materials have been selected for further
evaluation by SKB and Posiva. In particular some materials will be included in an extruded
pellet manufacturing study being undertaken by VTT for Posiva. These tests are intended to
determine if extrusion can produce materials comparable to roller compacted materials with
respect to density and behaviour. This might also require some placement tests using extruded
materials for comparison purposes.

The results obtained from the pellet manufacturing and optimization study have been used for
input into the on-going work of these groups, especially with regards to identifying some
potentially useful materials for inclusion in the on-going testing and demonstration programs of
Posiva and SKB.
 - 22 -

Table 4: Test and Production Matrix for Pellets for Potential use in Gap-filling (IFB and HTP geometries)

Tests 

Material   Batch   Number of  Pellet   Average  Thermal  Free  

Pellet  Average  Crush  Bulk Density 
Composition  Number  Batches  Size  Conductivity  Swell 
Water  Pellet b   Strength  
Content  Poured  Vibrated  Poured  Vibrated    
1,2,3,56  4  Large  √ √ √ 3,56  56  56  56 
100% Wyoming  √  √ √
43  1  Medium  43   43  43  43  √ 
Bentonite 
18,19,20,21,28,29,51  7  Small  √ √ √ 21,29,51  29,51   29,52   29,51  
4  1  Large  √ √ √  ‐   ‐   ‐  ‐  
10% Silica Sand  47  1  Medium  √  √ √  47  47  47  47  √ 
24,30  2  Small  √ √ √  ‐  ‐   ‐   ‐  
5  1  Large  √ √ √ 5  5  5  5 
25% Silica Sand  45  1  Medium  √  √ √ 45  45  45  45  √ 
25,31,54  3  Small  √ √ √ 25,54  54  54  54 
48  1  Medium  √ √ √ ‐   ‐   ‐   ‐  
50% Silica Sand  √ 
40  1  Small  √  √ √  40  40  40  40 
6,7  2  Large  √ √ √  7  7  7  7 
10% Illite  46  1  Medium  √  √ √ 46  46  46  46  √ 
26,32  2  Small  √ √ √  ‐   ‐  ‐   ‐  
8,9  2  Large  √ √ √ 8,9   8  8  8 
25% Illite  44  1  Medium  √  √ √ 44  44  44  44  √ 
27,33,53  3  Small  √ √ √ 33,53  53  53  53 
49,50  2  Medium  √ √ √  50  50  50  50 
50% Illite  √ 
38,39  2  Small  √  √ √  38  38  38  38 
Note: Numbers entered in table are those assigned to pellet batches
 - 23 -

Table 5: Test and Production Matrix for Pellets for Possible use in Tunnel Backfill

Tests 
Material   Batch   Pellet   Average  Thermal  Free  
Bulk Density 
Pellet  Average  Crush  Conductivity  Swell 
Composition  Number  Size 
Water  Pellet b  Strength  Poure
Content        d  Vibrated Poured Vibrated    
41  Large  √ √ √ √  ‐  ‐  ‐ 
42  Large  √  √  √  √ √ √ √  
Milos AC200  √  √  √  √
36  Small  ‐  ‐  ‐  ‐ 
(High Quality) 
37  Small  √  √  √  √  ‐  ‐  ‐ 
52  √ √ √ √ √ √ √
Small 
12  Large  √ √ √ ‐  ‐  ‐  ‐ 
13  Large  √  √  √  ‐  ‐  ‐  ‐   
14  Large  √  √  √  √ √ √ √  
√  √  √  √
15  Large  ‐  ‐  ‐  ‐ 
Asha  √  √  √ 
16  Large  ‐  ‐  ‐  ‐ 
√  √  √  √ 
17  Large  ‐  ‐  ‐ 
√  √  √  √ √ √ √
34  Small 
√  √ √ √ 
35  Small  ‐  ‐  ‐ 
10  Large  √ √ √ ‐  ‐  ‐  ‐ 
11  Large  √  √  √  √ √ √ √  
Milos B  22  Small  √  √  √  ‐  ‐  ‐  ‐  √ 
23  Small  √  √  √  ‐  ‐  ‐  ‐ 
√ √ √  
55  Small  √ √ √ √
Cebogel Pellets*      extruded  √  √  √  √  √ √ √ √ 
Buffer Test MX‐80*     Small   √  √  √  √  √ √ √ √ 
SKB MX‐80*     Small  √  √  √  √  √ √ √ √ 

Note: Numbers are those assigned to pellet batches, * Pellets supplied from others, size differs from AECL-produced materials
 - 24 -

4.3 MATERIALS AND METHODS

4.3.1 Materials Examined in Study (Bentonite, Illite, Silica Sand)

The materials evaluated as part of the Pellet Optimisation Project are listed below with
more details available from the suppliers’ data sheets and properties summaries
provided in Appendix A.

Wyoming Bentonites
 BC-NSB-200-2008: Bentonite Corporation-National Standard Bentonite,
200 mesh Wyoming-type bentonite. Produced in 2008 by the Bentonite
Corporation USA. Referred to as Wyoming bentonite in this study since it is the
material used in the majority of the testing. If other materials are used they are
specifically referenced in the text.
 ACC-MX-80-2001: An 80-mesh granular Wyoming bentonite, marketed by
American Colloid Company through a variety of distributors (e.g., AMCOL
Specialty Minerals and Askiana (see Appendix A)). This product is referred to as
MX-80 in this study. This material was used to manufacture pellets for generic
testing and Posiva’s Buffer Test conducted in 2011.

Other Bentonites
 Milos AC200: Also referred to as Milos HQ (High Quality), Milos A,. This
bentonite is sourced from Milos Greece and is known to have a high (>75%)
montmorillonite content. This material is referred to as Milos AC200 in this
document.
 Milos B: Also known as IBECO RWC-BF. This bentonite is raw, crushed clay
sourced from Milos Greece, and known to have a relatively low (~60%)
montmorillonite content. It is referred to as Milos B in this document.
 Asha-2010: A raw, coarsely crushed bentonite-rich clay from Kutch India.
Referred to as Asha in this document.
 Cebogel QSE-2010: Trade name for commercially-available pellets, made of an
activated Na-bentonite sourced from Milos Greece, and pelletised by Cebo
Holland BV in 2010. This material is extruded bentonite rods of 5-20mm length
and 6.5mm diameter. It is referred to as Cebogel in this document.

The materials and products used in this study have been used in a variety of previous
studies and applications by the participating organisations and in many cases differing
reference names have been used to describe the same (or very similar) product. In an
effort to standardise the naming protocol and link the work in this report to other studies,
a summary of equivalent names is provided in Table 6. A brief summary of their
physical and mineralogical properties of these materials is then provided in Table 7.

The naming protocol includes the producers’ initials followed by the product specification
(1 to 3 sets of initials) and then the year of material production. For example IBECO-
RWC-BR-2008 indicates IBECO as the manufacturer, RWC-BR as the specific product,
produced in 2008.

For ease of reference in the text, the bentonite products are generally referred to by their
short-form names:
Wyoming Bentonite = BC-NSB-200
 - 25 -

MX-80 = ACC-MX-80
Milos AC200 = IBECO-RWC-2008
Milos B = IBECO-RWC-BF-2008
Asha = Asha NW-BFL-L 2010
Cebogel = Cebogel QSE-2010

Two products not commonly referenced in current engineered barriers application, but
which were used in this study are:

Illite Clay
A crushed illitic shale produced and marketed in Canada as a cement plasticizer. The
material was sold under the trade name Sealbond (CB-Sealbond-1985) during the 1980s
and 1990s. A single batch was purchased by AECL during the 1980’s and has been
used for all research work. Although the source deposits still exist, it is no longer
available commercially. The material is 1-3% fine sand, 65-71% silt, and 28-32% clay.
It is referred to as illite in this document and its basic properties have been reported by
Radhakrishna and Chan (1982; 1985).

Silica Sand
A commercially marketed crushed quartzite, fractionated to meet 70-140 mesh size
range (fine sand >99%, silt <1%, clay 0%).

4.3.2 Pellet Manufacturing

Using the raw materials discussed in Section 4.4.1, a total of 56 batches of pellets were
produced at AECL’s Geotechnical Laboratory (Table 8). This work was conducted over
the May 2011 to June 2011 period. Depending on the quality of pellet produced, the
majority of the 56 batches of pellets (~2-5 kg each) were subjected to basic
characterisation tests (e.g., average pellet water content, pellet density, and crush
strength of individual pellets). Based on the results of the basic characterisation tests,
as well as general physical appearance, selected batches of higher quality pellets were
subjected to further testing (e.g., density of a poured mass of pellets, density of a
vibrated mass of pellets, thermal properties of the poured and vibrated masses of
pellets, and free swell testing). The testing procedures are described in Section 4.4.3.

The compositions of the pellets to be manufactured and tested were established by the
partner organisations at the project development stage. The selection of materials was
different for buffer-rock gap fill applications and general tunnel backfill applications.
Blends produced for buffer-rock gap fill applications include:
 100% Wyoming bentonite;
 Wyoming bentonite containing 10, 25 or 50% silica sand;
 Wyoming bentonite containing 10, 25 or 50% illitic clay.

Materials being considered for use in general backfilling applications include the Asha,
(Asha NW-BFL-L 2010); Milos B (IBECO-RWC-BF-2008) ; Milos AC200 (IBECO RWC
2008), Cebogel (QSE-200) and MX-80 bentonites.
 - 26 -

Table 6: Summary of Products used in Pellet Optimisation Project

Name used in Reference Other Names, Product Dominant

Producer Granularity
this report Name(s) Similar Products Type Mineral (%)
Wyoming BC-NSB-200-2008 National Standard Bentonite Bentonite 200 mesh product Montmorillonite 75-90
Bentonite BC-NSB-200-2005 Bentonite Corporation
MX-80 ACC-MX-80-2001 MX-80 American Bentonite 80 mesh product Montmorillonite 75-85
ACC-MX-80-2010 Colloid Co.
Milos AC200 IBECO-RWC-2008 Milos A S&B Bentonite Granules Montmorillonite 75-80
Milos HQ Industrial
Minerals
Milos B IBECO-RWC-BF- NN S&B Bentonite Granules Montmorillonite 50-60
2008 Industrial
Minerals
Asha Asha NW-BFL-L 2010 Asha 230A Kutch India Bentonite Granules Montmorillonite 60-65

Asha 230B*** Kutch India Bentonite Granules Montmorillonite 78-85

Cebogel Cebogel QSE-2010 Cebogel Pellets Cebo Bentonite Extruded bentonite Montmorillonite ~80
Holland BV pellets using activated
IBECO RWC-2010-type
material
Silica Sand Silica Sand Quartz sand Indusmin* Crushed 70-140 mesh Quartz >>95
Quartzite (0.212 - 0.104 mm)
Illite CB-Sealbond -1985 Crushed Illitic shale Canada Illitic Shale Silt-sized product Illite 35-50
Brick** Qtz 20, Kaolin 15-20

* This supplier is no longer in business; there are multiple suppliers of similar products available
** This supplier is no longer in business and product not commercially available. The formation used to supply this material is still accessible.
*** This is provided for information only and was not included in current test series, Literature references exist for this higher-bentonite content
product.
 - 27 -

Table 7: Properties of Materials Investigated in Pellet Optimization Project

Dominant Main Liquid Particle Cation Free Swell
Short Form Name Other Name(s)
Mineral Accessory Limit Density Exchange Freshwater
(used in this report) Sometimes Used
(%) Minerals % (g/cc) (meq/100g) (cc/2g)
Wyoming Bentonite National Standard Bentonite, Montmorillonite Quartz, Feldspar, 514, 2.70 92 26
BC-NSB-200-2008 75-90 Tr. <2% each 536
BC-NSB-200-2005 (75) Calcite, Gypsum, Illite
MX-80 ACC-MX-80-2010(3,4) Montmorillonite Quartz, Feldspar, 450-550 2.78 75, 34-42
ACC-MX-80-2001 75-85 (75) Kaolinite 84-104
(5)

Tr. <2% ea. Calcite,

Gypsum
Milos AC200 Milos A Montmorillonite Illite, Feldspar, 464(14) 2.65 75-111 27, 20
Milos HQ (6,7) 75-80 (80) Kaolinite, Calcite, 2.84(14) >80 >5
Dolomite, Albite,
IBECO-RWC-2008 Quartz, Gypsum
(13)
IBECO Deponit Ca-N-2008
Milos B(3,4) (14)
Milos B Montmorillonite Calcite, Dolomite, 150-157 2.78, 60-72 10, 13
IBECO-RWC-BF-2008 50-60 (58) Feldspar, Illite 115, 2.79(14) 60±10 >7
(14)
215 2.65
Cebogel Cebogel QSE-2010(4,6) Montmorillonite Quartz, Feldspar, 576 99-103 28 (12)
~80 Calcite, Dolomite 107
Asha Asha 230A WR2009-11 Montmorillonite Quartz, Feldspar, 83-93 27
Asha NW-BFL-L 2010 60-65 (65) Calcite, Gypsum, Illite 180 16.8
(3,4,8)
Asha 230B Montmorillonite Quartz, Feldspar, 424, 474 2.90 83-93 27
78-85 (80) Kaolinite
Silica Sand Crushed Quartzite 70-140 Quartz Tr. <1% each NM 2.65 << 1 NS <2
mesh (0.0.074-0.212mm) >>95 Pyrite, Feldspar, Calcite
(11)
Illite Illitic Shale Illite Kaolin 15-20, Quartz 28-31 2.70-2.79 10-18 NS <2
CB-Sealbond-1985
(1 ,2,11)
35-50 ~20, Feldspar, Chlorite,
Vermiculite
Tr. <2% each
Calcite, Horneblende
NM- This parameter is not measurable for the material noted . NS – Non-Swelling
1. Quigley 1984 2. Oscarson and Dixon 1989 3. Johannesson and Nilsson 2006 4. Sandén et al. 2008 5. Keto et al. 2009
6. Ahonen et al. 2008 7. Carlson 2004 8. Johannesson et al. 2008 9. Olesson and Karnland 2009 10. Sandén et al. 2008
11. Dixon 1995 12. Marjavaara and Kivikoski 2011 13. SKB 2010 14. Kiviranta 2011
 - 28 -

Pellets were made from the above-listed materials using a small (laboratory-scale), roller-type
briquetting machine rented for the duration of the project (Figure 13). The same machine used
for the testing program completed in 2010 (Man et al. 2011), was used in this study. Pellets
were manufactured by feeding the stock materials into the briquetting machine. As mentioned
above, this type of machine produces pellets by forcing loose stock material between two rollers
rotating in opposite directions. Each roller is machined with one half of the pellet shape
(Figure 13). Stock material is placed in a hopper located above a feed screw. The feed screw
forces the loose material between the two rollers. The rollers are forced together as stock
material is directed between them. Finished pellets are gravity fed to a bucket via a chute. The
test-scale machine allowed for the production of approximately 40-50 kg of material per hour,
which was adequate for production for laboratory applications. The pressure achievable using
this machine was approximately 4 times that previously required to manufacture pellets to a dry
density of ~1900 kg/m3 so its compaction capacity was adequate.

Three different sizes of pellets were made for buffer-rock gap fill applications. These were
identified as small (S), medium (M) and large (L) (Figure 14). The small pellets were disc-
shaped, with a diameter of 9 mm and a thickness of 6 mm. The medium pellets were oblong-
shaped, measuring 22 mm long by 11 mm wide by 7 mm thick. The large pellets were also
oblong-shaped, measuring 22 mm long by 14 mm wide by 8 mm thick. These large pellets were
manufactured using the same rolls as the preliminary pellet testing program.

The sizes and compositions of the pellets produced in the course of this study are provided in
Table 5. Most of the batches listed in Table 5 represent small production runs (2-5 kg each),
sufficient for basic evaluations (e.g., moisture, density, strength, free swell and in most cases
poured density and thermal properties), but not necessarily for conduct of erosion tests or large-
scale gap filling trials. For the tests requiring larger quantities of material, the results of
preliminary trials were evaluated and a limited number of larger production runs (50-80 kg each)
were completed. For backfill applications, only the small (S) and large (L) sizes were selected
by the project participants for further evaluation.

The preliminary trials with the roller compactor examined a range of variables intended to
identify optimal machine settings for production of good quality pellets. The variables
considered included: rate of material feed into the rollers, rate of roller rotation, compressive
force maintained between the rollers, gravimetric water content of the feed material and the
size/shape of the pellets manufactured. Each of these factors needs to be optimised for each
different material and roller type. For each stock material and pellet size, various production
trials were conducted by changing water content of the stock materials, roller loads, and feed
rates. Details related to the settings used during the optimisation process for each batch
manufactured were recorded and are provided in Appendix A. Visual appearance (darker
colour than stock material, as illustrated in Figure 15) and strength (resistance to be broken by
hand) were used as initial indicators of pellet quality and the need to adjust the production
variables. This generally corresponded well with the basic characterisation tests.
 - 29 -

Figure 13: Pellet Making Machine Showing Rollers

(a) Large and Medium Moulds

(b) Small Mould

Figure 14: Sizes of Moulds used in Making Pellets
 - 30 -

Figure 15: Colour Difference Between a Stock Mixture and Finished Pellets

Evaluation of the variables associated with the manufacturing of the pellets determined that
machine setting for ~6-8 rpm, 1500 psi roller contact pressure and use of relatively low water
content produced the most consistent quality of pellet. Small variations in any of these variables
did not discernibly affect the density or quality of the material produced (see Section 5.1.1) and
so for the most part, these variables were fixed for the course of the pellet manufacturing
process. Details related to the settings for each batch manufactured were recorded and are
provided in Section 5 and in the data tables provided in Appendix B.

Stock materials were blended in a batch mixer at their as-shipped water contents. Water
contents of the stock materials ranged from 4.0% to approximately 16% by weight. Initial trials
indicated that the addition of substantial moisture resulted in the stock material sticking to the
rollers. This produced low density pellets of inconsistent size, and so no moisture amendment
was done during blending of the feed materials. The final moisture content of the materials was
measured prior to manufacturing of pellets, and on the individual batches of finished pellets.
The water content of the individual pellets decreased with increased compression pressure due
to the increased heat associated with higher compression pressures.

Examples of some of the pellets produced are shown in Figure 16. Many of the pellets
produced had a fine webbing of compacted material between the individual pellets (Figure 16j).
This webbing had the potential to adversely affect the ability of the pellets to flow smoothly
during pouring and thereby reduce the packing efficiency of the pellets. Removal of the web
was accomplished at AECL by gently tumbling the pellets in a drum-line sieve. This tumbling
did not cause any damage to the pellets and simply removed the webbing. Similarly, MX-80
pellets produced in Europe for use in Posiva’s Buffer Test were placed in a 50 kg capacity
concrete pan mixer for 2 minutes then screened with a 4 mm mesh, resulting in average of 3.7%
fines removed.

The machine used in the Dixon et al. (2005) study was of slightly different design than the one
used in the 2010 and 2011 studies and this seems to have affected the water content required
for optimised compaction and the cosmetic nature of some of the larger pellets produced. It
was noted early in the current study that the machine used was unable to consistently produce
pellets at the gravimetric water contents initially targeted during the planning for this testing
 - 31 -

program (>12%). When operated and supplied with feed material at ~12% gravimetric water
content the compactor rapidly heated up and in some cases the machine had to be shut down
to prevent overheating and tripping thermal safety devices. Where pellets could be produced,
there were often issues related to the release of the pellets from the mould as the wheels
rotated. As a result the water content of the feed material was decreased substantially and
pellet production was possible. This lower water content in the feed material was attributed as
being the reason for some physical defects observed in the large pellets produced (Figure 17).
This was typically manifested as ribbing along the contact of the two halves (location of web
between pellets) of some of the medium and larger sized pellets (Figure 17), or other cracks in
pellet surfaces. In most cases these cracks appeared to be mostly cosmetic in nature (see
Section 5.1), with the pellets retaining substantial crush strengths (Section 5.4).

Table 8: Pellet Production Matrix

Medium Large
Small
Oblong Oblong
(9 mm
Composition (22 mm long, (22 mm long,
diameter,
11 mm wide, 14 mm wide,
6 mm thick)
7 mm thick) 8 mm thick)
Buffer-Rock Gap Fill:
100% Wyoming bentonite 7 batches 1 batch 4 batches
90% Wyoming bentonite and 10% silica sand 2 batches 1 batch 1 batch
75% Wyoming bentonite and 25% silica sand 3 batches 1 batch 1 batch
50% Wyoming bentonite and 50% silica sand 1 batch 1 batch 0 batches
90% Wyoming bentonite and 10% illite clay 2 batches 1 batch 2 batches
75% Wyoming bentonite and 25% illite clay 3 batches 1 batch 2 batches
50% Wyoming bentonite and 50% illite clay 2 batches 2 batches 0 batches

General Tunnel Backfill:

100% Asha bentonite 2 batches 0 batches 6 batches
100% AC200 bentonite 3 batches 0 batches 2 batches
100% Milos B bentonite 3 batches 0 batches 2 batches
 - 32 -

(b) 90% Wyoming:10% Sand (c) 75% Wyoming : 25% Sand (d) 50% Wyoming : 50% Sand
(a) 100% Wyoming

(e) 90% Wyoming : 10% Illite ((f) 75% Wyoming : 25% Illite (g) 50% Wyoming : 50% Illite

(h) 100% AC200 (k) 100% MX80

(i) 100% Milos B (j) 100% Asha

Figure 16: Examples of Pellets Produced for Pellet Optimisation Study

 - 33 -

Figure 17: Visual Appearance of Large Pellets Exhibiting Surface Defects

4.3.3 Pellet Testing

Pellet Density
The pellets of various sizes, shapes and compositions produced in the course of this
project were all analysed to determine the bulk and dry densities of the individual pellets.
This was done using a standard technique for assessing the bulk density of cohesive soil
materials (ASTM D-1188). The test involves thinly coating a pre-weighed specimen in
paraffin wax. The coated specimen is then weighed again in air and while suspended in
a volume of water. From these measurements it is possible to calculate the volume of
the specimen, the bulk density and following conduct of a gravimetric water content
analysis on the pellets, the dry density of the individual pellets.

In this study, all of the trials that produced pellets that were durable enough to survive
subsequent handling had their bulk and dry densities determined in the laboratory. The
results of these measurements are provided in Section 5.2.

Free Swell Tests

The ability of the pellets to take on water, swell and ultimately entirely fill the volume into
which they were placed is a key behavioural requirement for fill materials. With the
range of clay materials and clay-additive materials examined in this study, measurement
of the ability of the pellets to swell is required. The pellets manufactured for use in this
study were made using the water naturally present in the raw clay, a material that was
slightly dried, or if necessary the water content was increased through use of deionised
water. No saline solutions were used to replace or adjust the water content of the
material being compacted. The swelling capacity of these pellets was measured in a
range of artificial groundwater compositions, representing the full-range of conditions
that might be encountered in the granitic repositories of SKB and Posiva or the
sedimentary rock environment being considered by NWMO.

Free swell measurement was accomplished by placing approximately 5 g of pre-weighed

(and known gravimetric water content), pellet material in the bottom of a volumetric
cylinder (250 or 500 ml volume) and then gently filling the cylinder with the solution being
assessed. As the pellets disaggregate the volume they occupy is measured. On
completion of swelling the volume occupied by the pellets is recorded. Free swell is
defined as the volume occupied by 1 gm of oven-dried clay, so the volume occupied by
 - 34 -

the clay is divided by the dry mass installed in the cylinder (it is typically expressed in
cm3/g). The results of the free swell testing are provided in Section 5.3.

Crush Strength
Crush strength tests were conducted on the all batches of the pellets to determine the
resistance of the individual pellets to compressive force. As strength measurement is
sensitive to strain rate a modified triaxial load frame was used to allow for a consistent
strain rate during testing (Figure 18a) and all tests were repeated at least three times to
provide average and standard deviation values. The strength measured will also be
dependent on the size and shape of the pellets so comparison of results must be done
carefully and results of similarly sized pellets are assessed in terms of their relative
strength.

Crush strength was measured using a load cell that was sitting on a customized metal
plate (Figure 18b). The triaxial load frame was used to compress the pellets at a speed
rate of 1 mm per minute (Figure 18c). A data logger recorded the compressive force
versus time. Four pellets per batch were examined and the average crush strength
value with standard deviation and strain was calculated.

Results were evaluated based on comparison of pellets of the same size and shape.
When comparing identically sized pellets it should be possible to determine relative
changes between pellets of differing composition and water content. It would be
expected that pellets of differing size but otherwise identical composition and density
would exhibit differences in their crush strength requirement. Details of the test results
are provided in Section 5.4.

(b) Load Cell

(a) Crush Strength Testing Apparatus

(c) Crushing a Pellet

Figure 18: Device used to Determine the Crush Strength of Pellets

 - 35 -

Measurement of Bulk Density (Poured and Vibrated)

For selected pellet batches, bulk density of a poured mass of pellets was measured,
using a free fall pouring method. This is a simple test where pellets are poured into a
container of known volume and the mass is measured in order to calculate a bulk
density. No densification beyond that caused by gravity is applied. The AECL test used
two types of containers, a long cylindrical beaker and a Plexiglas box, to measure bulk
density; as shown in Figure 19. Different volume beakers from 200 ml to 500 ml were
used due to a small amount of pellets available for some batches. The Plexiglas box
was also used for measuring the bulk density of the pellets after application of vibratory
densification. The laboratory procedure to determine the vibration-induced densification
of the poured pellets is a low-energy method that is not calibrated against field
techniques but does provide a uniform means of comparing the ability of loosely poured
materials to be further densified. The use of a square placement box to determine the
loose and vibration-densification of the pellets also allowed measurement of the thermal
conductivity of a mass of pellets at a known poured density and compacted density.

(b) Plexiglas Box

(a) Long Cylindrical Beaker

(c) Electric Shaker

Figure 19: Bulk Density Tests using Different Poured Geometries

The tests conducted by AECL involved small volume containers and as noted in
Section 3, there are differences in what can be achieved in small versus large placement
geometries. In conjunction with the small-scale tests undertaken on a large proportion of
the small batches of materials tested in this study, larger scale tests were completed by
 - 36 -

VTT as part of the assessment work undertaken on the large batches of pellets sent to
them by AECL. The VTT tests involved determining the bulk (dry and wet) density of a
larger mass of pellets using a geometry where a narrow (2000 mm long by 35 mm wide)
gap was present and the pellets needed to fall as much as 1.14 m from the point of
pouring (Figure 20). The setup is described in detail by Marjavaara and Kivikoski
(2011). The average density of the large volume of fill was determined by dividing the
measured mass of pellets used to fill a known elevation in the fixed geometry of the
frame. In addition to pellets available at VTT (Cebogel and Buffer Test pellets), AECL
provided large batches (60-75 kg each) of six of the pellet types produced in Canada to
SKB and Posiva for further analysis. The tests undertaken by VTT required ~70 kg of
pellets for each pouring where no additional densification was applied. Due to limitations
in the quantities of materials available it was only possible to do large volume pour tests
on four of the six pellet types from Canada. For each type of pellet examined, the filling
test was repeated three times. The results of the gap filling bulk and dry density
measurements are presented in Section 5.5.

Figure 20: Large Pellet Pouring Test Frame at VTT

Abrasion Resistance Tests at VTT

The abrasion resistance of pellets is tested to provide an indication of how pellets would
react to handling. The risk of cracking, flaking or disintegration due to wear and
movement could be assessed. In order to assess the potential suitability of various
material composition, size and composition options for pellet fill, determining the ability
of selected pellets to withstand physical erosion is an important parameter. Six of the
pellet types produced in the course of this screening study were selected and batches of
approximately 75 kg of each were produced and shipped to VTT Finland for conduct of
large-scale pour testing (see above) as well as abrasion testing.
 - 37 -

Abrasion testing provides a measure of the relative durability of the materials selected
for further evaluation. There is no formally documented standard procedure for conduct
of abrasion testing of clay-based pellet materials but the abrasion test for aggregate
(stone and rock) can provide a relative measure. The abrasion resistance of the
bentonite pellets was assessed using the Micro-Deval test. The test was done
according to the EN standard method (EN 1097-1), using the alternative method of
Appendix A of this technique’s instructions for conduct where no water is added to the
drum. This test equipment is shown in Figure 21.

Micro-Deval test equipment

Large bentonite pellets tested for wear resistance, before (left) and after testing (right).
Figure 21: Abrasion Testing Equipment and Example of Results of Abrasion
Testing of Clay Pellets

Prior to the AECL pellet testing, some background trials were done on bentonite pellets
to establish if the standard limits would be followed or if a modified version would be
used. The test parameters were set to get partial degradation of the pellets, so that the
wear resistance could be compared between pellet types. The goal was not to get full
disintegration or no disintegration, yet to get something between. It was established that
the test method could be used according to the standard, with dried pellets. Using
pellets with the as-delivered moisture content (5-10%) resulted in poor results with
particle agglomerates forming due to compaction caused by the impact of the steel balls.
 - 38 -

In order to test the pellets without risk of particle agglomeration the technique was
modified slightly and used for all the materials evaluated. In these tests, 500 grams of
pellets were first oven-dried for 24 hours at 110°C. They were then cooled and placed in
the drum along with 5000 grams of steel balls. The drum was then sealed and subjected
to 12 000 revolutions with a speed of 100 revolutions per minute. After the test, the lost
material (fines) was screened on a size 1.6 mm sieve and the mass loss percentage was
reported. Figure 21 shows an example of the before and after condition of the pellets.

Measurement of Thermal Properties

The measurement of thermal properties was conducted in conjunction with the
determination of poured and vibrated densities for a given mass of pellets. The
apparatus used to measure the thermal properties of a mass of pellets is shown in
Figure 10. The device used in the thermal properties testing was the same Hot Disk
Thermal Constants Analyzer used in the 2010 study (Section 3.3.2). The system
operates by supplying a pulse of constant heat to a sample sensor, which acts as both a
heat source for increasing the temperature of the sample and a resistance thermometer
to monitor the change in temperature after the heat pulse. The sensor itself consists of
an electrical conducting pattern in the shape of a double spiral etched out of a thin sheet
of nickel. The conducting pattern is supported on both sides with a thin insulating
material consisting of Kapton.

The test configuration shown in Figure 10 and used in this study is known as a one-
sided test. It is performed by taking measurements with the test material in contact with
only one side of the sensor. The other side of the sensor is in contact with an insulating
material with pre-determined thermal properties. In this test series, rigid foam insulating
material was used as the backing material. The material being tested was first poured
into the container (of known volume) and then a thermal properties measurement was
made. After a cooling period, the material was vibrated as discussed above, and then
the thermal properties test was repeated on the compacted mass of pellets.

The solution of the thermal conductivity equation is based on the assumption that the
sensor is located in an infinite material. This means the total time of the transient
recording is limited by the presence of the outside boundaries and the limited size of the
sample. An estimation of how far this thermal wave has proceeded in the sample during
a recording is defined as the probing depth:

∆ 2√
where ∆p is the probing depth (i.e., the shortest distance from sensor edge to specimen
edge), K is the thermal diffusivity; and t is the measuring time.

The distance from any point of the sensor to any point on the surface of the specimens
must exceed ∆p if the total measuring time is t. To determine both the thermal
conductivity and thermal diffusivity with good accuracy, the thickness of a flat sample
should not be less than the radius of the sensor.

The probing depth only provides an estimate of the required sample size as the thermal
diffusivity of the material is unknown, but can be estimated from known properties of
materials. In practice the determination is by an iterative process.
As the sensor is heated, the resistance increase as a function of time is given by:
 - 39 -

1 ∆ ∆

where R0 is the resistance of the disk prior to heating and time (t) = 0, α is the
temperature coefficient of resistivity (TCR), ΔTi is the constant temperature difference
that develops nearly immediately over the insulation on the sensors, and ΔTavg(τ) is the
average temperature increase of the sample surface in contact with the sensor.

The temperature increase recorded by the sensors can be represented by:

∆ 1/ / 1

with ΔTi becoming a constant after a short time Δt. This can be estimated from:

∆ /

where δ is the thickness of the insulating layer, and κj is the thermal diffusivity of the
layer material.

The time dependent temperature increase is given by:

/
∆ /

where P0 is the power output from the sensor, “a” is the overall radius of the disk, “k” is
the thermal conductivity of the sample, and D(τ) is a dimensionless time dependent
function. The dimensionless time dependent function is:

where t is the time measured from the start of measurement, and Θ is the “characteristic
time”. The characteristic time is defined by:

/
By plotting the recorded temperature increase versus D(τ), a straight line is produced,
the intercept of which is ΔTi and the slope is P0/(π3/2 ak) using testing times longer than
Δti. Because thermal diffusivity is not known before testing, the final straight line is
determined through iteration. The results obtained for the pellet materials examined in
this project are presented in Section 5.7.

Water Uptake by Pellets of Differing Size, Shape and Composition

One of the main objectives of the pellet optimisation work is to determine if changing the
size, shape or composition of the pellets will affect the ability of this gap fill material to
take on water and swell against its confinement. This is of particular importance in the
period immediately following pellet installation and before an IFB or backfilled tunnel can
be completely isolated. The more effective the pellet fill is in taking on rather than
transmitting water, the longer will be the period before water movement can become an
issue in the buffer or backfill. It is also possible that artificial wetting of the gap fill may
be desirable in order to improve heat transfer during the period immediately after
container placement. In such a situation a gap fill that allows for a uniform wetting would
be more desirable.

Work to evaluate water movement through block and pellet backfill has been done by
Dixon et al. (2008a, b; 2011a, b) and Riikonen (2009), but these tests involved extruded
 - 40 -

bentonite pellets. As part of the joint pellet project activity, water uptake, density and
radial and axial pressure development have been evaluated by VTT in their test cell
having dimensions of 470 mm height by 360 mm outer diameter. These tests simulated
the 50 mm outer gap associated with the container and the surrounding rock in the IFB
geometry. The test layout is provided in Figure 22 together with a photo showing the
results of one of these tests.

(Photos from VTT)

Figure 22: Test Chamber (upper) and Artificial Wetting of Pellet Fill in the IFB
Geometry (lower)

Wetting tests are continuing at VTT as part of their work on water uptake and swelling in
an IFB geometry where pellets have been used. This work includes evaluation of all six
of the large batches of pellets produced by AECL for Posiva/SKB as well as customized
extruded bentonite pellets, Cebogel pellets and the Buffer Test roller compacted
bentonite materials. The results of these tests will be included in the pellet optimisation
report to be published by Posiva on completion of their project work in early 2013.
 - 41 -

5. RESULTS OF PELLET TESTING

5.1 VISUAL CHARACTERISATION

As each of the batches of pellets was produced they were visually inspected and any
unusual features were recorded (e.g. incomplete formation of pellets (Figure 23), high
fragility (Figure 23), presence of webbing between pellets (Figure 16), or other surface
defects as shown in Figure 23). A summary of some of these observations are provided
in the Tables of Appendix B.

Wyoming B18 : webbing between pellets Asha B35 : webbing between pellets

Wyoming B1 : broken pellets Asha B16 : Poor quality pellets

Wyoming with edge defects Buffer Test Pellets showing minor chips
and edge defects

Figure 23: Types of Defects Observed in Pellets (most resulting from moisture
content of feed material)
 - 42 -

One of the pellet types (Batch 56 consisting of large 100%Wyoming bentonite pellets)
manufactured showed small defects and longitudinal surface cracks along the seams
where the two halves of the rollers met during pellet manufacture. The strength and
durability of these pellets was high so it was subsequently produced as one of the large
batches that were sent to VTT where it was subjected to a more detailed surface
analysis, pouring trials and abrasion testing. Other batches that were produced in large
quantities for more extensive testing did not show the type of surface defects.

The nature of the longitudinal defects in Batch 56 were evaluated using optical
microscopy. A Wild M8 stereo microscope with a QWin image analyzer was used.
Figure 23 and Figure 24 show examples of the visual and microscopy images obtained.
It was determined that all of the pellets were similar, with a substantial number of the
pellets exhibiting some degree of cracking. The cracking did not appear to hinder
performance with respect to strength or swelling. It is expected that in the future the
cracking could be avoided by further optimising the material feed and roller compacting
process.

a) AECL pellets, height of photo 11mm. b) Finland pellets, height of photo 11mm.

c) AECL pellets, height of photo 4 mm. d) Finland pellets, height of photo 4 mm

Figure 24: Microscopic Images of Large Pellets showing Nature of Surface Defects
 - 43 -

5.2 PELLET DENSITY

Using the technique described in Section 4.4.3, each of the pellet types manufactured in
the course of this study had their dry and bulk densities determined. Table 9 presents
the results of these compaction trials on bentonite-only as well as bentonite-sand and
bentonite-illite blends. Table 10 provides the data related to clay-only pellets. As shown
in the tables, these pellets produced at the AECL’s Geotechnical Laboratory have higher
dry density than commercially available bentonite pellets (1.9 - 2.1 Mg/m3 versus 1.7-1.8
Mg/m3). From these data it is possible to begin the process of assessing what effect
adding a more thermally-conductive component to bentonite clay has on the ability to
produce durable pellets or various sizes.

From the large number of compositional and size options examined in the first stage of
the material screening testing, a smaller number of the most-promising material
compositions and sizes were identified based in their observed physical durability and
preliminary strength evaluations. The most promising materials were then manufactured
in larger quantities and these larger batches were subsequently used for larger-scale
placement density and durability evaluations.

The pellet density results in Table 9 and Table 10 are presented in Figure 25 (clay-only)
and Figure 26 (clay-filler blends) and from these data the effects of each additive were
evaluated. In these plots, the pellets are classified by their size with S=Small (disc 9 mm
diameter x 6 mm thick); M=Medium (oval 22 mm long x 11 mm wide x 7 mm thick);
L=Large (oval 22 mm long x 14 mm wide x 8 mm thick).

In Figure 25a a weak trend existed towards increased pellet density in the smaller pellets
relative to the larger sizes. This increase in dry density is less than 5% in the materials
examined, not a particularly substantial improvement. Similarly there is an indication
that there may be a reduction in the achievable dry density with increasing gravimetric
water content; again this represents only a <10% difference when the gravimetric water
content of the feed material goes from ~6% to ~15%. The trend for reduced pellet
density with gravimetric water content increase was also associated with increasing
difficulty in producing pellets with the roller compactor as water content increased. It
was noted that over the range examined, there was relatively little effect of feed moisture
content on the density of pellet produced, although the ease with which they can be
produced decreased with increasing water content. This, if confirmed, indicates that
there is a considerable range of acceptable moisture contents that can be tolerated
during production, making feed material moisture control less important.

When compaction data are presented as gravimetric water content versus Effective
Montmorillonite Dry Density (EMDD) in Figure 25b and Figure 26b, the data trends
observed for the dry density plot remain and are slightly clearer. What is notable
however is, that despite similar dry density values, the EMDD of the Milos B pellets
relative to the other materials manufactured by AECL. This is associated with the lower
smectite content of the Milos B (~60% versus >75% for the other clays). The lower
EMDD values mean that at the same dry density, these materials would develop a lower
swelling pressure than pellets manufactured using the other materials. The relationship
between hydraulic conductivity (and swelling pressure) and EMDD is presented in detail
by Dixon et al. (2011b). It should be noted that this reduction in swelling pressure is not
necessarily significant with respect to the ability of the pellets to fulfill their function since
 - 44 -

they would still retain a substantial swelling capacity (see Section 5.3). Figure 25 also
shows the density values measured for the pellets supplied by Posiva and SKB
(Cebogel, Buffer Test, SKB MX-80). These pellets appear to be of slightly (~10%) lower
density than those manufactured by AECL. These lower densities are likely related to
differences in manufacture (extuded Cebogel) and manufacturing technique (size,
manufacturing technique and perhaps water content).

In Figure 26 the results of compaction trials using fine-grained silica sand and illite
additions to Wyoming bentonite are plotted. Previous trials evaluating admixtures (see
Section 3) determined that the addition of sand to bentonite resulted in substantial
improvement of the thermal conductivity of precompacted bentonite-based materials but
materials such as copper were not particularly practical. Crushed illitic shale has
subsequently been evaluated since it has a substantially higher thermal conductivity
than bentonite. It is also a material often found as a minor constituent in bentonite so it
should be compatible and it might provide a medium that could provide some resistance
to bentonite movement over both the short- and longer-terms. Figure 26 also indicates
that there is a trend for increased pellet dry density as the silica sand component
increased from 0 to 50% of the dry mass proportion of the pellets. Addition of illite to the
bentonite did not seem to affect the compaction characteristics of the pellets and all
mixes generally showed similar degrees of compaction. When these same data are
viewed with respect to their EMDD some more substantial and systematic differences
become evident.

For bentonite-silica sand blends, 10% silica had no discernible effect on the EMDD. At
25% silica content, a very slightly reduction of the EMDD relative to bentonite-only is
evident. At 50% silica content, there is a very substantial reduction of the EMDD of the
pellets. This indicates that the improved compaction characteristics (higher achieved dry
density) for the 10-25% sand component did not result in a reduced EMDD, these mixing
ratios therefore seem to be potentially useful if the pellets prove to be thermally superior
to the clay-only materials (evaluated in Section 5.7), or other advantages to their use
exist. At 50% sand content, the pellets produced are inferior with respect to their EMDD
and have some issues regarding their mechanical durability (Section 5.6).

The addition of illite to the bentonite clay resulted in a reduction in the EMDD of the
pellets as would be expected in a material where the blending did not accomplish any
improvement to the dry density achieved during compaction (Figure 26). Therefore,
unless the addition of illite results in an improvement of some other key behavioural
parameter (e.g., thermal conductivity, erosion resistance (Sections 5.6, 5.7)), this
material is not particularly helpful. The adjacent materials will however have less volume
available to swell into than would be present in an unfilled situation, resulting in a higher
overall buffer/backfill density than would otherwise be present.
 - 45 -

Table 9: Pellet Densities for Wyoming Bentonite and Bentonite-filler Materials

Evaluated for Potential Suitability for use in Gap Filling
Pellet
Bulk
Roller Gravimetric Dry
Material Batch Pellet Density EMDD
Load Water Density
Composition Number Size Averages, (g/cm3)
(psi) Content, (%) d,
(g/cm3)
(g/cm3)
100% Wyoming 1 L 1000 6.2 2.07 1.95 1.79
100% Wyoming 2 L 1500 6.5 2.13 2.00 1.85
100% Wyoming 2010 L 1450 9.7 2.13 1.96 1.81
100% Wyoming 3 L 1750 7.1 2.16 2.02 1.87
100% Wyoming 2010 L 1800 9.5 2.15 1.95 1.79
100% Wyoming 43 M 1500 9.6 2.09 1.91 1.76
100% Wyoming 18 S 1500 7.4 2.24 2.08 1.94
100% Wyoming 51 S 1500 9.7 2.18 1.99 1.84
100% Wyoming 29 S 1500 14.9 2.25 1.96 1.81
100% Wyoming 20 S 1750 7.3 2.28 2.11 1.98
100% Wyoming 28 S 1750 13.8 2.30 2.02 1.85
10% quartz sand 4 L 1500 5.8 2.16 2.04 1.84
10% quartz sand 47 M 1500 7.5 2.19 2.04 1.84
10% quartz sand 24 S 1500 5.6 2.32 2.20 2.04
10% quartz sand 30 S 1500 13.7 2.36 2.08 1.89
25% quartz sand 5 L 1500 5.6 2.19 2.07 1.77
25% quartz sand 45 M 1500 6.1 2.20 2.07 1.77
25% quartz sand 25 S 1500 4.6 2.27 2.17 1.91
25% quartz sand 54 S 1500 7.1 2.27 2.12 1.84
25% quartz sand 31 S 1500 14.3 2.35 2.06 1.76
50% quartz sand 48 M 1750 3.9 2.29 2.20 1.72
50% quartz sand 40 S 1750 3.6 2.04 1.97 1.38
10% Illite 6 L 1500 5.2 2.14 2.03 1.82
10% Illite 7 L 1750 5.3 2.13 2.02 1.81
10% Illite 46 M 1500 7.8 2.15 1.99 1.78
10% Illite 26 S 1500 4.9 2.41 2.30 2.16
10% Illite 32 S 1500 14.3 2.21 1.93 1.71
25% Illite 8 L 1500 4.1 2.10 2.02 1.70
25% Illite 9 L 1750 4.3 2.18 2.09 1.79
25% Illite 44 M 1500 4.7 2.05 1.96 1.62
25% Illite 27 S 1750 3.9 2.18 2.07 1.76
25% Illite 33 S 1750 13.9 2.20 1.93 1.59
25% Illite 53 S 1000 10.2 2.23 2.02 1.70
50% Illite 49 M 1500 5.6 2.09 1.98 1.38
50% Illite 50 M 1750 5.6 2.21 2.09 1.53
50% Illite 38 S 1500 4.6 2.22 2.12 1.57
50% Illite 39 S 1750 4.6 2.23 2.13 1.59

Notes: S=Small = disc 9 mm diameter x 6 mm thick. M=Medium = oval 22 mm long x 11 mm

wide x 7 mm thick, L=Large = oval 22 mm long x 14 mm wide x 8 mm thick
 - 46 -

Table 10: Densities Achieved for Individual Clay-only Pellets Evaluated as a

Potential Component in Tunnel Backfill

Source  Pellet  Pellet    

Material Gravimetric  Bulk  Dry 
Roller  Water  Water  Density  Density 
Batch  Pellet   Load Content   Content,  averages,   d  EMDD 
3 3 3
Material Number Size (psi) (%) (%) (g/cm ) (g/cm ) (g/cm )
100% Wyoming 1 L 1000 as‐is: 7.0 6.2 2.07 1.95 1.79
100% Wyoming 2 L 1500 as‐is: 7.0 6.5 2.13 2.00 1.85
100% Wyoming 2010 L 1450 as‐is: 9 9.7 2.13 1.96 1.81
100% Wyoming 3 L 1750 as‐is: 7.0 7.1 2.16 2.02 1.87
100% Wyoming 2010 L 1800 as‐is: 9 9.5 2.15 1.95 1.79
100% Wyoming 43 M 1500 as‐is: 8.4 9.6 2.09 1.91 1.76
100% Wyoming 18 S 1500 as‐is: 7.6 7.4 2.24 2.08 1.94
100% Wyoming 51 S 1500 as‐is: 10.7 9.7 2.18 1.99 1.84
100% Wyoming 29 S 1500 15.9 14.9 2.25 1.96 1.81
100% Wyoming 20 S 1750 as‐is: 7.6 7.3 2.28 2.11 1.98
100% Wyoming 28 S 1750 15.9 13.8 2.30 2.02 1.85
Milos AC200 41 L 1000 as‐is: 10.8 11.6 2.05 1.83 1.70
Milos AC200 36 S 1000 as‐is: 11.8 11.7 2.08 1.87 1.74
Asha 12 L 750 as‐is: 14.2 13.1 2.09 1.85 1.58
Asha 13 L 1000 as‐is: 14.2 12.9 2.18 1.92 1.66
Asha 16 L 1250 as‐is: 14.2 13.0 2.22 1.97 1.72
Asha 34 S 1250 dried to 9.4 8.6 2.24 2.07 1.84
Milos B 10 L 1000 as‐is: 18.0 15.1 2.19 1.91 1.64
Milos B 22 S 1000 as‐is: 15.2 13.5 2.24 1.97 1.72
SKB MX80 8.8 2.00 1.86 1.69
Buffer Test ‐ MX80 * 14.9 2.06 1.79 1.62
Cebogel Extruded * 15.6 2.07 1.80 1.60

* Pellets were supplied by Posiva and SKB and differ in size from AECL-produced pellets
S=Small = disc 9 mm diameter x 6 mm thick
M=Medium = oval 22 mm long x 11 mm wide x 7 mm thick
L=Large = oval 22 mm long x 14 mm wide x 8 mm thick
 - 47 -

2.25
Wyomining, Large
2.15
Wyoming, Medium

Pellet Dry Density (Mg/m3)
2.05 Wyoming, Small

1.95 Milos AC200, large
Milos AC200, Small
1.85 Asha, Large

1.75 Asha, Small
Milos B, Large
1.65
Milos B, Small
1.55 Posiva Buffer Test
SKB MX80
1.45
Cebogel extruded
1.35
0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0
Gravimetric Water Content (%)
(a) Dry Density – Gravimetric Water Content
2.25
Wyomining, Large
2.15
Wyoming, Medium
2.05 Wyoming, Small
Pellet EMDD (Mg/m3)

Milos AC200, large
1.95
Milos AC200, Small
1.85 Asha, Large

1.75 Asha, Small
Milos B, Large
1.65 Milos B, Small

1.55 Posiva Buffer Test
SKB MX80
1.45
Cebogel extruded
1.35
0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0
Gravimetric Water Content (%)

(b) EMDD – Gravimetric Water Content

Figure 25: Effects of Clay Type and Pellet Size on Dry Density and EMDD
 - 48 -

2.35 2.35
100 % Wyoming Bentonite

2.25 10 % Quartz 2.15

25% quartz
Dry Density (Mg/m3)

2.15 50 % Quartz 1.95

EMDD (Mg/m3 )
2.05 1.75

1.95 1.55 100 % Wyoming Bentonite

10 % Quartz
1.85 1.35 25% quartz
50 % Quartz
1.75 1.15
0 2 4 6 8 10 12 14 16 0 2 4 6 8 10 12 14 16
Gravimetric Water Content (%) Gravimetric Water Content (%)

Bentonite – Silica Sand Blends : Dry Density and EMDD of Pellets

2.35 2.35
100 % Wyoming Bentonite 100 % Wyoming Bentonite

2.25 10 % Illite 2.15 10 % Illite

25% Illite 25% Illite
Dry Density (Mg/m3)

2.15 50 % Illite 1.95 50 % Illite

EMDD (Mg/m3)
2.05 1.75

1.95 1.55

1.85 1.35

1.75 1.15
0 2 4 6 8 10 12 14 16 0 2 4 6 8 10 12 14 16
Gravimetric Water Content (%) Gravimetric Water Content (%)

Bentonite-Illite Blends: Dry Density and EMDD of Pellets

Figure 26: Effects of Pellet Composition Dry Density and EMDD of Wyoming Bentonite-Based Pellets
 - 49 -

5.3 FREE SWELL TESTS

The method used to determine the free swell capacity of each of the pellets
manufactured in the course of this study was briefly described in Section 4.4.3. For
each of the water salinities evaluated in this study (0, 35, 70, 160 and 270 g/L Total
Dissolved Solids, TDS), a free swell test was conducted. For comparison purposes the
materials supplied by Posiva and SKB were also tested to determine their free swell
capacity under fresh water conditions. The detailed results of these tests are provided in
the tables in Appendix B and these data are plotted in Figure 27 and Figure 28.

20
Cebogel extruded
Milos AC200
SKB MX80
Free Swell Index (cc/2g)

15
Posiva Buffer Test
Wyoming
Asha
10
Milos B

Approximate volume occupied by non‐swelling solids

0
0 50 100 150 200 250 300
Solution TDS (g/L)

Figure 27: Free Swell Capacity of Bentonite Clay Pellets and Effect of Solution
Salinity

Figure 27 shows that as per Table 4, two basic qualities of bentonite were examined in
this study. The high montmorillonite-content Wyoming, Milos AC200 and Cebogel
materials all contain >75% swelling clay and exhibit substantially higher Free Swell
Indices (FSI) than the lower smectite Milos B and Asha clays. This difference becomes
particularly evident in the FSI for these materials under high water salinities. The FSI
measured for the compacted pellets are compared to the uncompacted source clays in
Table 4 and the powdered materials show a much higher FSI than the pellets. This is
attributed to the manner in which swelling occurs. In the powdered materials, the clay
has no constraints to swelling and will form a low-density gel. The pellet materials must
take on water from their surroundings, swell and disaggregate and so have no
opportunity to form a gel-like suspension. The FSI values for the powder and pellet
materials are consistent with regards to the relative order of volume change (e.g.,
Cebogel/Wyoming/Milos AC200 > Asha/Milos B), which provides confidence in the ability
to compare results for various materials. These tests do however highlight the need to
be careful when comparing the results obtained from different forms of the clays (e.g.,
powder and pellet).
 - 50 -

Figure 28 shows the effect of the water chemistry on the ability of the pellets to swell in
an unconfined environment. All of the pellets manufactured from swelling clays
examined have a substantial swelling capacity under fresh water conditions (8-16 times
their initial volume), with materials having the highest smectite (montmorillonite) content
showing the highest FSI. As salinity increased to 3.5% TDS, there is a marked reduction
in the FSI (4-7 times the initial pellet volume), with the value again being determined by
the smectite content. Beyond 3.5% TDS water salinity, there is only a gradual decrease
in the FSI of the clays, ultimately at ~27% TDS, the FSI has dropped to between 3 and
6 cm3/2g. It should be noted that the original density of the individual pellets was
approximately 2000 kg/m3 (= 1cm3/2g) and for illite and silica materials the FSI is <
2 cm3/2g. Hence a fine-grained, non-swelling material could be expected to have a FSI
in the order of 2 cm3/2g. The free swell tests confirm that the pellet materials will
disaggregate under the entire range of groundwater salinities examined (0-27% TDS)
and will occupy a volume 1.5 to 3 times that of an individual pellet. Most of this
additional volume represents the void spaces originally between the pellets (as-placed
dry density of pellet mass ~900-1000 kg/m3, is ~0.36 solids and 0.64 voids or a FSI of
~2). The data therefore shows that low smectite content clays have only a very limited
swelling capacity at salinities at or above 3.5% TDS and all materials are strongly
affected by solution chemistry.

The FSI’s measured are consistent with the magnitude of the swelling pressures
developed (a measure of ability of bentonite-based material to expand in a rigidly
confined environment) for low density bentonite-based materials (Figure 29). For a
given dry density, the swelling clay content will therefore decrease in proportion to the
quantity of additive present. The FSI of these materials would also be affected by the
smectite content of the clay used as well as the EMDD. However in this study clay-
additive pellets were all manufactured using Wyoming bentonite as the swelling clay so
the smectite content of the clay will have remained constant. The pellets, when poured
to fill a gap exhibit as-placed dry densities in the order of 0.9 to 1.1 Mg/m3 for clay-only
systems (an EMDD range of ~0.6 to ~0.95 Mg/m3). These densities would result in
swelling pressures of 10 kPa (270 g/L, low quality bentonite) to ~400 kPa (freshwater,
high quality bentonite) if the clay were confined in a fixed volume.

In the bentonite-additive systems the EMDDs achieved during gap filling were never
higher than for the clay-only pellets and in most cases there was a reduction in EMDD
values. Therefore with reduced FSI for the mixed component systems, the swelling
pressures that could develop will also decrease. The effect of an increased dry density,
achieved through vibratory compaction and/or addition of a fines component will improve
both the EMDD and hence the swelling pressure developed by the pellet fill under each
groundwater condition considered.

Addition of filler materials can therefore be concluded to act as an inert component in

bentonite-based pellet materials, which is consistent with the EMDD concept, which
treats non-swelling solids in the bentonite as inert fillers also. The volume of the pellets
on completion of disaggregation and unconfined swelling is proportional to the smectite
content of the clay, the proportion of filler material added and the salinity of the water.
 - 51 -

20 20

Wyoming Wyoming

Free Swell Index (cc/2g)

Free Swell Index (cc/2g)
15 15
10% Sand 10% Illite

25 % Sand 25 % Illite
10 10
50% Sand 50% Illite

5 5

Approximate FSI for illite and quartz Approximate FSI for illite and quartz

0 0
0 50 100 150 200 250 300 0 50 100 150 200 250 300
Solution TDS (g/L) Solution TDS (g/L)

20 20 20

Wyoming Wyoming Wyoming

Free Swell Index (cc/2g)

Free Swell Index (cc/2g)

Free Swell Index (cc/2g)
15 15 15
10% Sand 25 % Sand 50% Sand

10 10% Illite 10 25 % Illite 10 50% Illite

5 5 5

Approximate FSI for illite and quartz Approximate FSI for illite and quartz Approximate FSI for illite and quartz

0 0 0
0 50 100 150 200 250 300 0 50 100 150 200 250 300 0 50 100 150 200 250 300
Solution TDS (g/L) Solution TDS (g/L) Solution TDS (g/L)

Figure 28: Free Swell Index of Wyoming Bentonite and Blends with Crushed Illitic Shale and Fine Silica Sand
 - 52 -

0 % TDS y = 0.003e5.5095x
MX-80 / Wyoming distilled water
R² = 0.9218
1%TDS

10 3.5% TDS

7% TDS
Swelling Pressure (MPa)

10 % TDS

> 30 % TDS

1
1 % TDS y = 0.0007e6.5172x
R² = 0.9684
3.5 % TDS y = 0.0008e6.4543x
R² = 0.8696
7 % TDS y = 0.0007e6.3639x
R² = 0.8946
0.1
10 % TDS y = 0.0004e6.4426x
R² = 0.8268
35 % TDS y = 0.0002e6.6014x
R² = 0.8678

0.01
0.3 0.8 1.3 1.8
Effective Montmorillonite Dry Density (Mg/m3)

Figure 29: Changes in Swelling Pressures in Bentonite Clays as Result of Water

Salinity (EMDD of as-placed clay-only pellet fill ~0.6-0.95 Mg/m3)

5.4 CRUSH STRENGTH OF INDIVIDUAL PELLETS

Mechanical stability is an important parameter for gap fill or backfill materials. These
materials must be durable and able to protect adjacent HCB buffer in the period between
buffer installation and ultimately borehole closure by completion of backfilling operation
in the overlying tunnel. The pellets used in the tunnel backfill or in the HTP geometry
need to be able to fill the voids they are placed in without losing their physical integrity
and continue to provide their filling and thermal functions in the period prior to water
saturation. The strength of individual pellets was evaluated to provide:
 An indication of the effect of pellet composition and manufacturing water content
on the material produced by roller compaction, and
 Information on the ability of the individual pellets to retain their physical integrity
prior to and during installation.

The results obtained from the crush strength testing were evaluated based on the size
and shape of pellets. Figure 30 and Figure 31 show the results of the crush strength
tests and these data are provided in full in the tables provided in Appendix B.
The first step in evaluation of the data looked at the crush strength values of clay-only
pellets produced based on the size and water content. Figure 30 and Figure 31 show
plots of crush strength of large pellets and small pellets respectively based on their
gravimetric water contents (no medium sized pellets were produced using 100% clay). It
should be noted that the values of water content and crush strength shown are the
average values obtained by testing several pellets of each type.
 - 53 -

350

300

250
Crush Strength, (N)

200

150

100% Wyoming, Avg Dry Density of 2.0 Mg/m3

100
100% Wyoming, Avg Dry Density of 1.9 Mg/m3
Milos AC200 (High Quality), Avg Dry Density of 1.8 Mg/m3
50 Milos B, Avg Dry Density of 1.9 Mg/m3
Asha, Avg Dry Density of 1.9 Mg/m3
0
4 6 8 10 12 14 16 18
Water Content, (%)

Figure 30: Comparison of Average Crush Strength for Large, Clay-Only Pellets

350
100% Wyoming, Avg Dry Density of 2.1 Mg/m3

100% Wyoming, Avg Dry Density of 2.0 Mg/m3

300
100% Wyoming, Avg Dry Density of 2.0 Mg/m3

Milos AC200 (High Quality), Avg Dry Density of 1.9 Mg/m3

250
Crush Strength, (N)

Milos B, Avg Dry Density of 2.0 Mg/m3

Asha, Avg Dry Density of 2.1 Mg/m3

200

150

100

50

0
4 6 8 10 12 14 16 18
Water Content, (%)

Figure 31: Comparison of Average Crush Strength for Small, Clay-Only Pellets

In addition to the clay-only pellets manufactured using Wyoming bentonite, small

batches of large- and small-sized pellets were manufactured using Milos AC200, Asha
and Milos B bentonite. These pellets, along with Cebolgel (extruded pellets) and
bentonite pellets used in Posiva’s Buffer Test project were also tested to determine their
crush strength (Figure 32, data provided in Appendix B).
 - 54 -

The results of crush tests on clay-only materials can be summarised as follows:

 The large pellets of Milos AC200 and Milos B materials showed ~25% lower and
~20% higher crush strengths respectively as compared with the pellets
manufactured using Wyoming bentonite. The Asha pellets were essentially the
same as the Wyoming pellets with respect to crush strength.
 Large pellets produced from Milos B clay (lowest smectite content and high water
content) showed the highest crush strength (280 N). 100% Wyoming bentonite
pellets show strengths of 142 N and 235 N for water contents of 7% to 10%
respectively. Asha pellets show a slightly higher strength (approximately 200 N)
than Milos AC200 pellets (175 N) at similar water content.
 In general, for the large pellets, there is a trend for increasing average strength
with increasing water content for the range examined in this study. Minor
differences in the density of the pellets tested accounts for part of the data
scatter while compositional differences (montmorillonite content) will also account
for some differences in the strength of the various pellets, but these should be
secondary effects.
 In some cases when comparing the results of several tests the presence of
micro-cracks in the large pellets can explain some of the variation in results
observed (Figure 32a).
 For small, clay-only pellets, the trend of increasing average strength with
increasing water content found in the large clay-only pellets was not evident.
The strength of all the pellets tested ranged from about 50 to 100 N with Milos B
pellets showing the highest strength (100 N).
 Comparing the relative strengths of the small pellets showed Asha was the
weakest pellet (~20% lower than Wyoming and Milos AC200). Milos B had the
highest strength (~20% higher than Wyoming and Milos AC200). The Cebogel
and Buffer Test (MX-80) pellets both had substantially higher crush strengths
than the small pellets. This is in part due to differences in size and shape relative
to the other materials tested, as well as their higher water content (Cebogel 22%
and Buffer Test pellets 17%).

As was the case with the large pellets, the parameters of montmorillonite content and
range of water content over which pellets could be successfully manufactured have
resulted in some of the data scatter but should not have masked any strong trends in
behaviour.

The effects of adding a filler material to bentonite clay is an important part of determining
what optimisation in pellet composition is realistically achievable. Pellet optimisation is
of interest in development of material options for use as gap fill for the IFB geometry,
placement tunnel fill in the HTP geometry and, to a lesser extent perhaps for
development of pellet materials for use in deposition and central tunnel backfilling using
the block and pellet concept. These applications will each require different pellet sizes
and have different functional demands on the pellets. In order to try and assess
optimisation options relevant to each application, three sizes of pellets were
manufactured with 10, 25 and 50 % of fine quartz sand or Illite (by dry weight %) added
to a Wyoming bentonite. Exactly the same test procedure as was used to determine the
crush strength of the bentonite clay-only pellets was used to determine the strength of
these blended materials. For ease of comparison strengths are discussed relative to
 - 55 -

pellets of the same size produced from100% Wyoming bentonite. The compressive load
required to fail the individual pellets decreased with decreasing pellet size but for the
same size of pellet it will be possible to identify any trends regarding strength that water
content or pellet composition may induce.

The results of compressive strength testing of pellets produced from Wyoming bentonite-
filler materials as well as other clay products are plotted in Figure 32 so as to show their
relative strengths and the variability in the measurements made. Variability was most
noticeable in the large pellets, which tended to have slightly more obvious physical
variability with respect to micro-cracks. In Figure 33 these data are plotted such that the
effects of water content can also be evaluated. The complete set of measurement
results is provided in Appendix B.

The effects of blending bentonite with filler materials can be summarised as follows:
In pellets containing silica sand:
For large pellets,
 the addition of 10-25% by dry weight silica sand had essentially no effect on the
crush strength.
For medium-size pellets,
 the addition of 10-50% by dry weight of silica sand increased strength ~70%, with
maximum densification occurring with a 25% silica sand component,
For small-size pellets
 average strength increased by ~10-20% with a 10% silica sand content;
 at 25% silica sand component, the average strength approximately doubled;
 at 50% silica sand content, the strength of the pellets was less than or equal to
what was observed for the 100% clay pellets. At this combination of pellet size
and filler component there would appear to be interference of the filler with the
compaction process although pellets could still be manufactured.

In pellets containing Illite:

For large pellets,
 With a 10-25% illite component the average pellet strength decreased
discernibly, particularly at the 25% level.
 No pellets were manufactured at 50% illite content.
For medium-size pellets,
 For 10-50% illite content, the pellets showed an increase in crush strength of
~10-25% from the clay-only compositions.
For small pellets
 10-25% illite caused no substantial change in the average strength, and
 50% illite the strength was only slightly lower than for the clay-only.
 100
150
200
250
300
350

0
50

50
100
150
200
250
300
350

0
100% Wyoming(wc~7%)

100% Wyoming(wc~14%)
(a) Large Pellets

100% Wyoming(wc~10%)

10% Sand(wc~6%)

10% Sand(wc~14%)

25% Sand(wc~5%)

25% Sand(wc~7%)

25% Sand(wc~14%)

50% Sand(wc~4%)

10% Illite(wc~5%)
- 56 -

10% Illite(wc~14%)

25% Illite(wc~4%)

(c) Small Pellets

Measurements
100
150
200
250
300
350

0
50

25% Illite(wc~14%)

25% Illite(wc~10%)

50% Illite(wc~5%)

Milos AC200(wc~12%)

Asha(wc~9%)

Milos B(wc~14%)

Cebogel(wc~15%)

Buffer Test Pellets(wc~15%)

(b) Medium Pellets

Figure 32: Crush Strength in Newtons (N) of (a) Large, (b) Medium and (c) Small Pellets, showing Standard Deviation of
 - 57 -

350 350
100% Wyoming, Avg Dry Density of 1.9 Mg/m3
10% Sand, Avg Dry Density of 2.0 Mg/m3
300 300 25% Sand, Avg Dry Density of 2.1 Mg/m3
50% Sand, Avg Dry Density of 2.2 Mg/m3
10% Illite, Avg Dry Density of 2.0 Mg/m3
250 250 25% Illite, Avg Dry Density of 2.0 Mg/m3

Crush Strength, (N)
Crush Strength, (N)

50% Illite, Avg Dry Density of 2.0 Mg/m3

200 200

150 150
100% Wyoming, Avg Dry Density of 2.0 Mg/m3
100 100% 100% Wyoming, Avg Dry Density of 1.9 Mg/m3 100
10% Sand, Avg Dry Density of 2.0 Mg/m3
25% Ssand, Avg Dry Density of 2.1 Mg/m3
50 50
10% Illite, Avg Dry Density of 2.0 Mg/m3
25% Illite, Avg Dry Density of 2.1 Mg/m3
0 0
2 4 6 8 10 12 2 4 6 8 10 12
Water Content, (%) Water Content, (%)
(a) Large Pellets (b) Medium Sized Pellets
350
100% Wyoming, Avg Dry Density of 2.1 Mg/m3 100% Wyoming, Avg Dry Density of 2.0 Mg/m3
100% Wyoming, Avg Dry Density of 2.0 Mg/m3 10% Sand, Avg Dry Density of 2.2 Mg/m3
10% Sand, Avg Dry Density of 2.1 Mg/m3 25% Sand, Avg Dry Density of 2.2 Mg/m3
300 25% Sand, Avg Dry Density of 2.1 Mg/m3 25% Sand, Avg Dry Density of 2.1 Mg/m3
50% Sand, Avg Dry Density of 2.0 Mg/m3 10% Illite, Avg Dry Density of 2.3 Mg/m3
10% Illite, Avg Dry Density of 1.9 Mg/m3 25% Illite, Avg Dry Density of 2.1 Mg/m3
250 25% Illite, Avg Dry Density of 1.9 Mg/m3 25% Illite, Avg Dry Density of 2.0 Mg/m3
Crush Strength, (N)

50% Illite, Avg Dry Density of 2.1 Mg/m3

200

150

100

50

0
2 4 6 8 10 12 14 16
Water Content, (%)
(c) Small Sized Pellets
Figure 33: Comparison of Crush Strengths of (a) Large, (b) Medium, and (c) Small, Clay and Clay – Filler Pellets (Note: the
average dry density (ρd) of the pellets are provided in the legend of each graph)
 - 58 -

5.5 AS-PLACED DENSITY OF POURED PELLETS

The as-placed density of pellet materials used to fill gaps in the IFB geometry or as fill in
the HTP or other backfill applications will determine its thermal, mechanical and
hydraulic performance in the period immediately following its placement. It will also
affect how the fill performs over the longer term. Pellets manufactured in the course of
this study were examined to determine what improvement in both as-placed and vibrated
densities were achievable and were compared to pellets commercially available.

When comparing the poured and densified (vibrated) bulk and dry densities achieved for
the clay-only pellets there is little difference observable for changes in the clay-type
used. The average as-placed dry density of poured pellets is approximately 0.96 Mg/m3
for all of the tests completed (excepting medium sized pellet tests). The large and small
pellets do not seem to have a discernible difference in their as-poured dry densities
(Table 11 and Figure 34). There are small differences in the bulk (wet) densities
achieved but that is entirely the result of pellet water content differences and not any
intrinsic change in pourablity. It is noted that the medium-sized pellets tended to achieve
a poured density that was ~14% higher (1.09 Mg/m3) than was typically achieved using
the small or large pellets. This infers that there is something in their shape/size that
improves their loose packing properties. Vibration of the loosely poured pellets always
results in an increased density.

Vibration on 100% bentonite pellets resulted in an increased as-placed dry density

(maximum 1.34 Mg/m3) by 24% for medium pellets and by 28% for large pellets. These
values of vibrated dry density were slightly higher than those for the commercially
available pellets (maximum 1.21 Mg/m3). However, small pellets of Milos AC200 and
Milos B show 12% and 18% increases in their as-placed dry densities (maximum 1.14
Mg/m3), but these values of vibrated density are still lower than those of commercial
pellets.

Pellets produced with a non-bentonite component (10-50 % illite or silica sand), were
also evaluated for their poured dry density. These pellets all achieved a slightly (~5%)
higher poured density relative to the clay-only materials, consistent with their slightly
higher individual pellet densities (Table 12 and Table 13). As with the clay-only pellets,
there is a discernible (~14%) improvement in the loose poured density of the medium-
sized pellets relative to the small or large pellets but there is no difference in their poured
densities relative to the proportion of additive present. This is to be expected since the
pellets were all of similar dry density.

The void space was calculated on pellets after completion of as-poured density tests.
Table 11 and Table 12 present a summary of poured and vibrated density including
values of void space of batches. All the pellets specially manufactured have a range of
void space from 60% to 70%. The Milos AC200 pellets had the highest void space of
70% at a water content of ~12% and 25% silica sand had the lowest void space of 60%.

Results of VTT tests (Figure 36) for as-poured density of pellets installed in a long
narrow gap were provided for inclusion in this report. As-poured densities of pellets with
100% bentonite, 25% illite, Milos AC200 are close to 1.0 Mg/m3, which is very similar to
results obtained at AECL. Table 13 presents the results of large-scale pouring trials into
long, narrow openings approximating the gap between the buffer and the rock in the IFB
geometry. The correlation to AECL results is presented in Figure 37, based on dry
 - 59 -

density. This was done because there was a slight loss of moisture between the time
the pellets were measured in Canada and Finland (about 1%), as seen from Table 13.
These data are also plotted in Figure 36, and show the effect of scale and confined
geometry on the loose-poured density achievable using these materials. These are
consistent with the types of differences observed in previous pellet pouring and
densification trials (see Section 3). Although not done for these trials because of the risk
of damaging pellets of limited supply, a vibratory action would increase the as-placed
bulk density of the pellet fill (Martino and Dixon 2007, Man et al. 2011, Marjavaara &
Kivikoski 2011). The test results for the current study therefore provide conservative
bounds for use in analysing the effects of using a pellet fill in the IFB geometry.

Poured Dry Density Vibrated Dry Density

1.8

1.6

1.4
Dry Density, (Mg/m3)

1.2

1.0

0.8

0.6

0.4

0.2

0.0

Figure 34: Poured and Vibrated Density of Various Clay-Only Pellets. (Note: two
different individual dry densities (1.8 and 1.9 Mg/m3) for Milos AC200 small pellets)
 - 60 -

Poured Dry Density Vibrated Dry Density

1.8

1.6

1.4
Dry Density, (Mg/m3)

1.2

1.0

0.8

0.6

0.4

0.2

0.0

(a) Large Pellets

Poured Dry Density Vibrated Dry Density
1.8

1.6

1.4
Dry Density, (Mg/m3)

1.2

1.0

0.8

0.6

0.4

0.2

0.0
100% 10% 25% 10% 25% 50%
Wyoming Sand Sand Illite Illite Illite
(wc~10%) (wc~8%) (wc~6%) (wc~8%) (wc~4%) (wc~6%)

(b) Medium Pellets

 - 61 -

Poured Dry Density Vibrated Dry Density

1.8

1.6

1.4
Dry Density, (Mg/m3)

1.2

1.0

0.8

0.6

0.4

0.2

0.0

(c) Small Pellets

Figure 35: Poured and Vibrated Dry Bulk Densities of Pellets. (Note: two different
individual dry densities (1.8 and 1.9 Mg/m3) for Milos AC200 small pellets)

1.2

1.0

0.8
(Mg/m3)

0.6
Dry Density

0.4

0.2

0.0

Figure 36: Results of VTT Tests to Determine Loose, As-poured Density of Pellets
Installed in a Narrow Gap
 - 62 -

Table 11: Poured and Vibrated Densities of Clay-only Pellets

Single  Single 
As‐Placed  As‐Placed  Vibrated  Vibrated 
Batch  Water  Pellet  Pellet 
Size  Bulk  Dry  Bulk  Dry 
Composition  No.  Content  Bulk  Dry 
Density  Density  Density  Density 
Density  Density 
      %  Mg/m3  Mg/m3  Mg/m3  Mg/m3  Mg/m3  Mg/m3 
Wyoming  56  L  10  2.11  1.92  1.12  1.02  1.43  1.30 
Wyoming  3  L  7.1  2.16  2.02  1.07  1.00 
Milos AC200   41  L  11.4  2.04  1.83  1.14  1.02  1.29  1.16 
Asha   17  L  12.9  2.20  1.95  1.11  0.98  1.29  1.14 
Wyoming  43  M  9.6  2.09  1.91  1.19  1.09  1.47  1.34 
Wyoming  51  S   9.7  2.18   1.99   1.14   1.03   1.35  1.23 
Milos AC200   37  S  11.8  2.04  1.82  0.94  0.84 
Milos AC200   52  S  11.8  2.12  1.90  1.08  0.97  1.21  1.08 
Asha   35  S  8.6  2.23  2.05  1.11  1.02  1.20  1.10 
Milos B   55  S  12.6  2.17  1.93  1.09  0.96  1.28  1.14 
Cebogel *      15.2  1.83  1.59  1.05  0.86  1.29  1.12 
Buffer Test MX‐80 *        15.6  1.79  1.55  1.10  0.94  1.31  1.13 
SKB MX80 *  9.3  1.84  1.68  1.09‐1.13  0.97  1.24  1.14 

* These materials were supplied by Posiva and SKB and were of different size and shape from AECL-manufactured pellets
 - 63 -

Table 12: Poured and Vibrated Densities of Pellets Containing Non-bentonite Component

Pellet  Pellet   Pellet   As‐Placed  As‐Placed  Vibrated  Vibrated 

Batch No.  Size  Water  Bulk  Dry  Bulk  Dry   Bulk  Dry 
Composition 
Content  Density  Density  Density  Density  Density  Density 
      %  Mg/m3  Mg/m3  Mg/m3  Mg/m3  Mg/m3  Mg/m3 
100% Wyoming  56  L  10  2.11  1.92  1.12  1.02  1.43  1.30 
100% Wyoming  3  L  7.1  2.16  2.02  1.07  1.00  ‐  ‐ 
25% Silica Sand   5  L  5.6  2.19  2.07  1.09‐1.16  1.03‐1.10  1.27‐1.33  1.19‐1.26 
25% Illite   9  L  4.3  2.18  2.09  1.02  0.98  1.31  1.25 
100% Wyoming   43  M  9.6  2.09  1.91  1.19  1.09  1.47  1.34 
10% Silica Sand   47  M  7.5  2.19  2.04  1.37‐1.49  1.27‐1.39  1.56  1.45 
25% Silica Sand   45  M  6.1  2.20  2.07  1.29‐1.37  1.22‐1.29  1.59‐1.63  1.50‐1.54 
10% Illite   46  M  7.8  2.15  1.99  1.32  1.22  1.33  1.23 
25% Illite   44  M  4.7  2.05  1.96  1.18  1.12  1.29  1.23 
25% Silica Sand   25  S  4.6  2.27  2.17  1.06  1.01  ‐  ‐ 
25% Silica Sand   54  S  7.1  2.27  2.12  1.21  1.13  1.4  1.31 
50% Silica Sand   40  S  3.6  2.04  1.97  1.26  1.21  1.37  1.32 
25% Illite   53  S  10.2  2.23  2.02  1.16  1.05  1.35  1.23 
 - 64 -

Table 13: Density and Moisture Content Results from Tests by VTT, Finland
Water Water VTT VTT AECL
Content Content Gap fill, Gap fill, poured
Material
(%) (%) bulk density dry density dry density
Production Testing (kg/m3) (kg/m3) (kg/m3)
100% Wyoming,
small (Batch 51) 9.7 8.9 1112 1021 1110
100% Wyoming,
large (Batch 56) 10 8.9 1092 1003 1180
75% Wyoming:25%
silica sand, small
(Batch 54) 7.1 5.2 - - 1210
75% Wyoming:25%
illite, small
(Batch 53) 10.2 8.3 1118 1032 1160
Milos AC200, small
(Batch 52) 11.8 10.5 1047 948 1080
Milos B, small
(Batch 55) 12.6 9.6 - - 1070
Cebogel 21.9 21.9 1107 908 1050
Buffer Test MX-80 16.8 16.8 1067 914 993

Figure 37: Comparison of Poured Dry Density Achieved in Small-Scale (AECL) and
Large-Scale (VTT) Trials
 - 65 -

5.6 ABRASION TESTS

Table 14 presents the test results of the abrasion test. The mass loss is presented as
the average of two tests. For comparison, Buffer Test pellets and Cebogel extruded
pellets were also tested. The table also provides references to standard Finnish
aggregates tested in dry conditions (Vuorinen 1999).

Table 14: Results of MicroDeval Test

Sample Name Mass Loss (%)

Reference: Buffer Test pellets (roller compacted, MX-80) 9.78
Reference: Cebogel extruded pellets 45.95
100 % Wyoming bentonite, small pellets 65.31
100 % Wyoming bentonite, large pellets 16.34
75 % Wyoming bentonite + 25 % Quartz Sand 33.04
75 % Wyoming bentonite 25 % Illite 57.12
100% Milos AC200 38.51
100% Milos B 24.08
Reference Good aggregate A [2] an amphibolite composed 2.3
of amphibole and feldspar
Reference Good aggregate B [2] a granite composed 2.3
mainly of quartz and feldspar
Reference Weak aggregate C [2] a limestone 8.6

It was expected that there may be some correlation between the wear resistance and
strength of the pellets. Material having a higher strength may have better resistance to
deterioration (indicated by a low value of mass loss). Table 15 shows the results of both
the abrasion resistance and crush strength tests, along with the water contents of the
various pellets. Figure 38 shows the relationship between these two parameters for the
tested pellets. For equivalent bentonite type and moisture content (Wyoming bentonite
with water content 5-9%), the trend held true that pellets with higher crush strength
values had lower mass loss (better abrasion resistance). The references Cebogel and
Buffer Test pellets had much higher water contents (17-22%) and thus cannot be directly
compared to the other pellets. All three bentonites have similar crush strengths, though
the mass loss was significantly higher for the 100% Wyoming bentonite compared to
Milos AC200 and Milos B.
 - 66 -

Table 15: Results of Wear Resistance Tests

Mass Crush
Water
Sample Loss Strength
content (%)
(%) (N)
Buffer Test MX-80 9.8 136 16.8
100% Wyoming, large 16.3 235 8.9
Milos B, small 24.1 100 9.6
75% Wyoming + 25% Quartz Sand 33.0 169 5.2
Milos AC200, small 38.5 85 10.5
Cebogel 46.0 170 21.9
75% Wyoming + 25% Illite 57.1 129 8.3
100% Wyoming, small 65.3 89 8.9

70 350
Mass Loss (%)
60 300

50 250

Crush Strength (N)
Mass Loss (%)

40 200

30 150

20 100

10 50

0 0

Figure 38: Relationship between Abrasion Resistance (Mass Loss) and Crush
Strength
 - 67 -

5.7 THERMAL PROPERTIES TESTING

5.7.1 Thermal Properties of Poured and Vibratory-Densified Pellet Fills

The thermal conductivity data for the various pellet-fill materials are provided in Table 16
and plotted in Figure 39. Each point represents the mean of several replicate tests. A
pair of points is presented for each type of pellet. The point plotting with lower dry bulk
density represents the dry density of the poured mass of pellets and the corresponding
thermal conductivity measurement for that condition. The point plotting with higher dry
density corresponds to the vibrated mass of pellets and the corresponding thermal
conductivity measurement for that condition. Small symbols represent small pellets,
medium symbols represent medium pellets and large symbols represent large pellets.
Measurements on materials with the addition of Wyoming bentonite fines (i.e., 80 mesh
granules) are noted as individual points. The results for large pellets made with 100%
Wyoming bentonite are included in all graphs as a baseline comparison.

Figure 39 shows the results of tests conducted on the pellets manufactured during the
2011 program in terms of dry density. Figure 6 showed the results of tests conducted
on available commercial pellets tested during the preliminary program and some of this
data is repeated in Figure 39 for comparison to the new data set. The complete data set
including the measured thermal diffusivity and specific heat values for each type of pellet
are provided in Appendix B (Table B-4).

Figure 39(a-e) assists with the evaluation of the effect of vibratory compaction, different
sources of bentonite, the addition of additives and different pellet sizes on thermal
conductivity of the placed material. Figure 39a shows the results for 100% Wyoming
bentonite, and the results for the large pellets are repeated on subsequent plots to
provide a visual benchmark for comparison.

The thermal conductivity of loosely poured pellet-fill materials was less than 0.5 W/(m·K)
for all of the tested pellets. Densification of the pellet mass using light vibratory
compaction increased the thermal conductivity of the pellet-fill material to about
0.6 W/(m·K) for several type of pellets (i.e., the large pellets made with Milos AC200
bentonite, the large pellets made with Milos B bentonite, and the medium sized pellets
made with a mixture of 90% bentonite and 10% silica sand). However, this value is still
slightly less than the lower end of the desired range of 0.7 to 0.9 W/(m·K). For tunnel
backfill or gap fill for use in a repository using the IFB concept, thermal conductivity
improvement is also desirable since the option is to leave an air-gap which is much less
thermally conductive. As shown in Figure 8, the thermal conductivity of the HCB buffer
is also quite low (~0.5 to ~1.1 W/(m·K), depending on density and degree of saturation).
Pellet fill of the type tested in this study would therefore be thermally comparable to the
HCB buffer. For tunnel backfill applications, thermal properties of the gap fill are of
relatively low importance and factors such as mechanical durability, erosion resistance
and swelling and hydraulic properties are of greater importance.
 - 68 -

5.7.2 Effect of Bentonite Type on Thermal Properties

Different sources of bentonite showed no to slight differences in thermal conductivity

(Figure 39b and c). The Milos AC200 bentonite and Wyoming bentonites displayed
similar thermal conductivity values. Milos B bentonite displayed a higher thermal
conductivity (0.58 W/(m·K) and 0.63 W/(m·K) for small and large pellets after vibratory
compaction, respectively), which may be attributed to the higher proportion of more
highly heat conductive impurities in this material. Even then, their thermal conductivity is
not substantially higher than other pellets. In general, the thermal conductivity of the
pellets manufactured during the 2011 program exhibited slightly higher thermal
conductivities than the commercially available bentonite pellets (Figure 6).

5.7.3 Effect of Additives to Clay

Comparing the pellets made with a blend of Wyoming bentonite and silica sand to those
made with 100% Wyoming bentonite (Figure 39a), little improvement in thermal
conductivity is achieved by adding substantial quantities of sand (Figure 39d). The
highest thermal conductivity achieved (with a value of 0.61 W/(m·K)) using silica as an
additive, was with the medium sized pellets made with a mixture of 90% bentonite and
10% silica. This is only slightly higher than the pellets of the same size made with 100%
Wyoming bentonite (0.59 W/(m·K)). It is worth noting that the highest measured thermal
conductivity corresponded to the highest density achieved for a mass of pellets. The
increase in density due to vibratory compaction results in a greater increase in thermal
conductivity than the addition of silica sand to the pellets achieves. Thermal conductivity
is therefore dominated by the large air-voids present at the time of installation and
modifications to the density of the individual pellets are of secondary importance.

Similarly, the addition of illite did not result in increased thermal conductivity of a mass of
pellets relative to the silica-sand materials (Figure 39e). For example, the medium sized
pellets made with 75% Wyoming bentonite and 25% illite had a thermal conductivity of
0.53 W/(m·K) (vibrated mass of pellets), which is slightly less than that measured for the
same sized pellets made with 100% Wyoming bentonite (0.59 W/(m·K) for vibrated mass
of pellets). Higher dry densities were achieved with the addition of illite, but this is at the
expense of a lower overall EMDD.

5.7.4 Effect of Pellet Size on Thermal Conductivity

Comparing different sizes of pellets with the same composition, it appears that the
medium sized pellets generally produced the highest thermal conductivities. This
corresponds to higher densities achieved with the medium pellets relative to the other
sizes. Dry densities in the order of 1.4 Mg/m3, which is the target density for NWMO’s
gap fill material in the HTP geometry, were achieved with the majority of the trials using
medium sized pellets.

5.7.5 Effect of Fines Addition

The addition of Wyoming bentonite fines was conducted later in the program in an
attempt to further increase the dry density of the placed material and perhaps also
 - 69 -

increase thermal conductivity. The fines were in the form of 80 mesh granules, and
were vibrated into the larger pores between the pellets. Figure 39a show that the
addition of fines to the large pellets made with 100% Wyoming bentonite increased the
dry density from 1.2 Mg/m3 to over 1.4 Mg/m3 when 20% fines, by weight, were added.
A further increase in dry density was achieved when 30% fines were added
(1.55 Mg/m3). These densities correspond to thermal conductivities above
0.65 W/(m·K). The best results were obtained with the small pellets made with 100%
Wyoming bentonite with 30% fines. This combination increased the dry density to over
1.6 Mg/m3 and gave a thermal conductivity of 0.69 W/(m·K). When fines were added to
pellets composed with 10% silica sand, the beneficial effects of the silica with respect to
thermal conductivity was masked, and thermal conductivity was reduced (Figure 39d).
The addition of illite to the pellets tended to lower thermal conductivity (Figure 39e). The
addition of fines to the large pellets made with 75% Wyoming bentonite and 25 % illite
simply increased the thermal conductivity into the range representative of pellets made
with 100% bentonite.
Table 16: Thermal Conductivity of Pellet Fills

Thermal Conductivity (W/(m·K))

(for various pellet sizes)
Pellet Composition Densification Small Medium Large
Poured: 0.34-0.45 0.41 0.48
Wyoming Vibrated: 0.52-0.61 0.59 0.56
Vibrated with 20% fines: - - 0.65
Vibrated with 30% fines: 0.62-0.69 - 0.67
Asha Poured: 0.43 - 0.43
Vibrated: 0.49 - 0.54
Milos AC200 Poured: 0.44 - 0.45
Vibrated: 0.54 - 0.61
Poured: 0.47 - 0.43
Milos B
Vibrated: 0.58 - 0.63
Poured: 0.46 - -
Cebogel pellets
Vibrated: 0.53 - -
Poured: 0.52 - -
Buffer Test Pellets
Vibrated: 0.56 - -
Poured: - 0.45 -
10% Silica Sand Vibrated: - 0.61 -
Vibrated with 30% fines: - 0.52 -
Poured: 0.44 0.45 0.56
25% Silica Sand
Vibrated: 0.52 0.59 0.56
Poured: 0.44 - -
50% Silica Sand
Vibrated: 0.52 - -
Poured: - 0.49 0.39
10% Illite
Vibrated: - 0.56 0.49
Poured: 0.42 0.41 0.38
25% Illite Vibrated: 0.48 0.53 0.49
Vibrated with 30% fines: - - 0.54
Poured: 0.46 0.51 -
50% Illite
Vibrated: 0.51 0.52 -
Note: "fines" consist of 80 mesh Wyoming bentonite granules
(similar to MX-80)
 - 70 -

5.7.6 Summary

The overall message provided by this data set is that up to a point, increased as-placed
density has a greater effect on increasing thermal conductivity than the addition of silica
sand or illite to the bentonite in the individual pellets. For pellets composed of 100%
Wyoming bentonite, the addition of fines to fill the voids between pellets increased the
dry density and increased thermal conductivity to just below the desired range of 0.7 to
0.9 W/(m·K).

0.7 0.7
20% Fines
0.65
Thermal Conductivity (W/(m·K))

0.65

Thermal Conductivity (W/(m·K))

30% Fines
0.6 0.6

0.55 0.55

0.5 0.5

0.45 0.45

0.4 0.4

0.35 0.35
(a) 100% Wyoming Bentonite (b) Asha, Milos AC200, Milos B
0.3 0.3
0.75 1.00 1.25 1.50 1.75 2.00 0.75 1.00 1.25 1.50 1.75 2.00
Dry Density (Mg/m3) Dry Density (Mg/m3)
0.7 0.7

0.65
Thermal Conductivity (W/(m·K))

0.65
Thermal Conductivity (W/(m·K))

0.6 0.6

0.55 0.55

0.5 0.5

0.45 30% Fines

0.45

0.4 0.4

0.35 0.35 (d) 90:10, 75:25, 50:50

(c) Cebogel, Buffer Test Wyoming Bentonite: Silica Sand
0.3 0.3
0.75 1.00 1.25 1.50 1.75 2.00 0.75 1.00 1.25 1.50 1.75 2.00
Dry Density (Mg/m3) Dry Density (Mg/m3)
 - 71 -

0.7

0.65

Thermal Conductivity (W/(m·K))

0.6

0.55

0.5

0.45 30% Fines

0.4

0.35 (e) 90:10, 75:25, 50:50

Wyoming Bentonite: Illite
0.3
0.75 1.00 1.25 1.50 1.75 2.00
Dry Density (Mg/m3)

Figure 39: Comparison of Thermal Conductivity Values for Different Sizes and
Compositions of Pellets

6. DISCUSSION AND CONCLUSIONS

The objective of this work was to evaluate the potential of developing bentonite based
pellets that better meet the requirements of NWMO, SKB and Posiva for application as
gap fill as well as tunnel backfill. A total of 56 batches of pellets were made for the
study, representing a range of pellet sizes, shapes and the effects of various
volumetrically inert fillers on the properties of bentonite-based materials.

The pellets were assessed primarily to evaluate potential to improving their density,
strength, durability and heat transfer capability. Several different swelling clay materials
of interest to NWMO, SKB and Posiva were examined as well as mixtures of bentonite
with illitic clay or fine silica sand. Filler contents of 10, 25, and 50% by weight were
tested. Based on preliminary characterisation results six pellet types were selected for
additional advanced testing by Posiva and VTT in Finland and SKB and Clay
Technology in Sweden. These tests are ongoing and are not reported on in this report.

Source material (clay) granularity was determined to be an important factor, influencing

the roller pressure needed to produce good quality pellets. It was found that coarser
feed material increased roller pressure and risked stalling the machine. As a result,
unprocessed “clay” materials such as Milos B bentonite, required pre-treatment of the
materials to screen out the larger aggregates and impurities (e.g. course sand-sized
particles). This demonstrated the need for a consistent feed material during production
of pellets and highlighted a key factor in determining the quality of the product produced.

The as-received moisture (5-13% gravimetric) of most of the raw stock materials
examined was generally sufficient to produce good quality pellets. When moisture was
added to the feed material it tended to stick to the rollers, producing poorly shaped
pellets. In cases where the as-received moisture content was too high, pre-treatment
(partial drying) of the stock material was required. Asha bentonite is an example of
material that required some drying prior to pellet manufacturing.
 - 72 -

In general the dry density of the pellets produced was higher at lower water content
ranges and the EMDD of the individual pellets was not substantially affected by the
presence of modest quantities of filler (typ. <25%). Hence the presence of filler materials
should not adversely affect the swelling pressure developed or the hydraulic conductivity
of the pellets. This has relevance to both the economics of pellet manufacturing as well
as the performance of this barrier component.

The pellets produced all had individual pellet dry densities in the order of 2.0 Mg/m3.
When loosely poured, the dry bulk density of the mass of pellets was generally in the
order of 1.0 to 1.2 Mg/m3. This could be improved with vibratory compaction but the dry
bulk density of most of the pellets after vibration was still less than 1.4 Mg/m3. The
exceptions to these density limits were some of the pellets made with substantial silica
sand and illite additives but the as-placed density of these materials were not
substantially higher (e.g. <1.6 Mg/m3 dry density).

The addition of Wyoming bentonite fines (i.e. 80 mesh granules), vibrated into the pore
space between the pellets was undertaken to evaluate their effect on overall density and
thermal characteristics of the gap fill. The fines typically resulted in improved as-placed
dry density (>1.5 Mg/m3). This would result in improved hydraulic and swelling
behaviour of the fill material but would also limit the ability of the pellet fill to transfer
inflowing water into its internal volume during the initial wetting stages. This could
potentially result in localised variability of swelling pressure and earlier development of
preferential flow paths.

The addition of additives such as silica sand and illite had only a small effect on the
thermal conductivity of a mass of pellets relative to pellets made with 100% bentonite.
Increasing the overall density through vibratory compaction resulted in an increase of the
thermal conductivity that was greater than the effect of additives. This added variable of
inter-pellet porosity dominates relative to the addition of admixtures that showed promise
with pucks of bentonite-additive mixtures. The highest values of thermal conductivity
following vibratory compaction were achieved with the large pellets made with Milos
AC200 bentonite, the large pellets made with Milos B bentonite, and the medium sized
pellets made with a mixture of 90% bentonite and 10% silica sand. The size of these
pellets makes them unsuitable for gap fill use in the IFB geometry, but are of interest in
the HTP geometry or as tunnel backfill. A thermal conductivity value of about 0.6
W/(m·K) appears to represent the maximum achieved without the addition of fines to fill
the void space between the pellets. The maximum thermal conductivity values measured
on pellets densified by vibratory compaction were close to 0.6 W/(m·K) which is below a
more desirable range of 0.7 to 0.9 W/(m·K).

The vibration of Wyoming bentonite fines (80 mesh granules) generally provided a slight
improvement to the thermal conductivity of all the pellet fill materials. The maximum
thermal conductivity achieved through the addition of fines was 0.69 W/(m·K) using the
small 100% Wyoming bentonite pellets with 30% fines vibrated into the pore space
between the pellets. The greatest improvement offered by the addition of fines was a
significant increase in bulk density of the as-placed material (e.g. over 1.5 Mg/m3).

In summary; the implications of this study on the use of pellet materials in a repository
are as follows:
 - 73 -

 The pellet manufacturing process is sensitive to the water content of the source
materials. In the tests completed in this study it was found that low (<15% and
often <10%) gravimetric water content in the feed material was necessary in
order to produce pellets of adequate durability.
 100% bentonite (from various sources) can be used to produce dense pellets of
reasonable durability for placement in a repository.
 The thermal conductivity of the pellet materials evaluated in this study is ~0.5 to
~0.6 W/(m·K) following vibratory compaction. This may not be sufficient for the
HTP geometry but does represent a substantial improvement in the heat transfer
characteristics relative to the air-gap otherwise present in the IFB geometry.
 The presence of silica sand and illite did not significantly increase the thermal
conductivity of the as-placed material relative to the clay-only pellets. The
thermal properties are dominated by the large, air-filled voids between the
pellets. Over the longer term, once the clay has hydrated and swelled to fill
these gaps there will likely be only a limited effect of additives on heat transfer
since the water component dominates the system.
 The addition of fines (80 mesh Wyoming bentonite granules) to the pore space
between pellets (20% to 30% fines by weight) resulted in an increase in thermal
conductivity, but not greater than 0.7 W/(m·K). This is of interest in the HTP
geometry where the fill is relied on to transfer the heat from the UFC to the
surrounding rock mass in a generally dry environment.
 With the addition of fines (80 mesh Wyoming bentonite granules: 20% to 30%
fines by weight), the bulk density of the as-placed material can be substantially
increased. This may be of relevance to applications where pellet fill is placed in
the HTP geometry. This will improve the initial and longer-term thermal
conduction characteristics of the materials surrounding the UFC. In the IFB
geometry, fines could also result in a higher buffer density being maintained
within the borehole.
 The addition of a fine silica sand material resulted in a higher crush strength for
the pellets. This may be of value where the pellets are considered for use in
locations where they need to maintain a discernible crush strength (e.g. flooring
materials where block and pellet backfilling occurs). This strength, if also
associated with a reasonable erosion resistance (mechanical and hydraulic)
could also be of value in pellet materials used in tunnel backfill applications.
There was some correlation between crush strength and abrasion resistance.
 Illitic clay addition did not result in any substantive improvement in the
mechanical or thermal characteristics of the pellets. The presence of this
component may influence the ability of water to erode the pellet fill and adjacent
bentonite materials but this was not evaluated as part of this project.

Ongoing work in Finland and Sweden will provide more insight into the hydraulic and
mechanical performance of pellets as part of the fill materials used in the IFB and HTP
geometries. While this study focussed on use of roller-compacted pellet materials,
ongoing work in Finland and Sweden will include pellets manufactured using extrusion
technology and will provide a valuable basis for comparison.
 - 74 -

ACKNOWLEDGEMENTS
The work described in this document is the result of a co-operative research project
funded by the Nuclear Waste Management Organisation of Canada, the Swedish
nuclear fuel and waste management company (SKB) and Posiva Oy of Finland. The
authors would like to acknowledge the assistance of B. Evenden and F. Johnston of
AECL in the conduct of the laboratory testing done at AECL’s Geotechnical Laboratory.
 - 75 -

REFERENCES

Ahonen, L., P. Korkeakoski, M. Tiljander, H. Kivikoski and R. Laaksonen. 2008. Quality

assurance of the bentonite material, Posiva Oy, Working Report-2008-33,
Olkiluoto.

Carlson, L. 2004. Bentonite mineralogy. Part 1: Methods of Investigation – A Literature

Review. Part 2: Mineralogical Research of Selected Bentonites. Posiva Oy,
Working Report-2004-02, Olkiluoto.

Chandler, N., A. Cournut, D. Dixon, C. Fairhurst, F. Hansen, M. Gray, K. Hara,

Y. Ishijima, E. Kozak, J.Martino, K. Masumoto, G. McCrank, Y. Sugita,
P. Thompson, J. Tillerson and B. Vignal. 2002. The five year report of the Tunnel
Sealing Experiment: an international project of AECL, JNC. Atomic Energy of
Canada Limited Report AECL-12727.

De Bock, C., J-M. Bosgiraud, H. Weber, T. Rothfuchs, J. Verstricht, B. Breen and

M. Johnson. 2008. Achievements of the ESDRED project in buffer construction
technology, (K. Wolf, W.K Seidler and J.M. Bosgiraud, Editors), Euradwaste ’08
conference ESDRED – An Integrated European Project Focused on Technology
Development.

Dixon, D.A. 1995. Towards and understanding of water structure and water movement
through dense clays, Ph.D. Thesis, University of Manitoba, Dept Civil
Engineering, Winnipeg.

Dixon D., S. Stroes-Gascoyne, B. Kjartanson and P. Baumgartner. 2005. Development

of Sealing Materials for Application as Engineered Barriers in a Deep Geologic
Repository for Used Fuel Disposal: Filling the Gaps, Canadian Nuclear Society
Waste Management, Decommissioning and Environmental Restoration for
Canada’s Nuclear Activities: Current and Future Needs Ottawa, Ontario, Canada,
May 8-11, 2005.

Dixon, D.A. and P. Keto. 2008. Backfilling techniques and materials in underground
excavations : Application to Posiva’s spent fuel repository, Posiva Oy, Working
Report-2008 – 56, Olkiluoto.

Dixon, D., S. Anttila, M. Viitanen and P. Keto. 2008a. Tests to determine water uptake
behaviour of tunnel backfill (Baclo Tests at Äspö). Swedish Nuclear Fuel and
Waste Management Company, SKB- R-08-134, Stockholm.

Dixon, D., C. Lundin, E. Örtendahl, M. Hedin and G. Ramqvist. 2008b. Deep repository
– engineered barrier system: Half-scale tests to examine water uptake by
bentonite pellets in a block-pellet backfill system. Swedish Nuclear Fuel and
Waste Management Company, SKB- R-08-132, Stockholm.

Dixon, D.A., J.B. Martino, B. Holowick and D. Priyanto. 2011. Enhanced Sealing
Project: Monitoring the THM response of a full-scale shaft seal. Canadian
Nuclear Society Conference, Waste Management, Decommissioning and
Environmental Restoration for Canada’s Nuclear Activities September 11-14,
2011, Toronto.
 - 76 -

Dixon, D., E. Jonsson, J. Hansen, M. Hedin and G. Ramqvist. 2011a. Effect of localized
water uptake on backfill hydration and water movement within a backfilled tunnel:
half-scale tests at Äspö bentonite laboratory. Swedish Nuclear Fuel and Waste
Management Company, SKB- R-11-27, Stockholm.

Dixon, D., T. Sandén, E. Jonsson and J. Hansen. 2012. Backfilling of deposition

tunnels: Use of bentonite pellets. Swedish Nuclear Fuel and Waste Management
Co, SKB, Stockholm. R-Series report, SKB P-11-44, Stockholm.

EN 1097-1 + A1: Tests for mechanical and physical properties of aggregates. Part 1:
Determination of the resistance to wear (micro-Deval). 2011.

Hansen, J., L. Korkiala-Tanttu, E. Keski-Kuha and P. Keto. 2009. Deposition tunnel

backfill design for a KBS-3V repository. Posiva Oy, Working Report-2009-129,
Olkiluoto.

Johannesson, L-E and U. Nilsson. 2006. Deep repository – Engineered barrier

systems; geotechnical behaviour of candidate backfill materials, laboratory tests
and calculations for determining performance of the backfill. Swedish Nuclear
Fuel and Waste Management Co., SKB, R-06-73, Stockholm.

Johannesson, L.E. 2008. Geotechnical investigations made on unsaturated backfill

materials. Swedish Nuclear Fuel and Waste Management Co., SKB, SKB-R-08-
131, Stockholm.

Johannesson, L-E., T. Sandén, A. Dueck and L. Ohlsson. 2010. Deep repository-

Engineered barrier system: Characterisation of a backfill candidate material,
IBECO-RWC-BF, BACLO PROJECT-PHASE 3 - Laboratory tests. Swedish
Nuclear Fuel and Waste Management Company, SKB-10-44, Stockholm.

Juvankoski, 2010. Description of basic design for buffer. Posiva Oy, Working Report
WR-2009-131, Olkiluoto.

Keto P., D. Dixon, E. Jonsson, D. Gunnarsson, L. Börgesson and J. Hansen. 2009.

Assessment of backfill design for KBS-3V repository. Swedish Nuclear Fuel and
Waste Management Company, SKB-R-09-52, Stockholm.

Kiviranta, L. and S.Kumpulainen. 2011. Quality control and characterisation of

bentonite materials. Posiva Oy, Working Report-2011-84, Olkiluoto.

Kjartanson, B., D. Dixon and C. Kohle. 2005. Placement of bentonite pellets to fill
repository sealing system voids and gaps. Ontario Power Generation, Nuclear
Waste Management Divison, Technical Report, 06819-REP-01200-10136,
Toronto.

Man, A. and J.B. Martino. 2009. Thermal, hydraulic and mechanical properties of
sealing materials. NWMO, Technical Report, NWMO TR-2009-20, Toronto.

Man, A., J. Martino, C-S. Kim and D. Priyanto. 2011. Characterizing and improving the
thermal conductivity of engineered clay barriers for sealing a deep geological
repository. Canadian Nuclear Society Conference, Waste Management,
 - 77 -

Decommissioning and Environmental Restoration for Canada’s Nuclear Activities,

2011 Sept 11-14, Toronto.

Marjavaara, P. and H. Kivikoski. 2011. Filling the gap between buffer and rock in the
deposition hole. Posiva Oy, Working Report-2011-33, Olkiluoto.

Martino, J. and D. Dixon. 2006. Placement and formulation studies on potential light
backfill and gap fill materials for use in repository sealing. Ontario Power
Generation, Nuclear Waste Management Division, Supporting Technical Report,
06819-REP-01300-10011-R00, Toronto.

Martino, J., D. Dixon, S. Stroes-Gascoyne, R. Guo, E. Kozak, M. Gascoyne,

T. Fujita, B. Vignal, Y. Sugita, K. Masumoto, T. Saskura, X. Bourbon, A. Gingras-
Genois and D. Collins. 2008. The Tunnel Sealing Experiment 10 year summary
report. Atomic Energy of Canada Limited, URL-121550-REPT-001, Chalk River.

Martino, J.B., D.A. Dixon, B.E. Holowick and C-S. Kim. 2011. Enhanced Sealing
Project (ESP): Seal Construction and Instrumentation Report. Nuclear Waste
Management Organization, NWMO, APM-REP-01601-0003, Toronto.

Olsson, S. and O. Karnland. 2009. Characteristics of bentonites from Kutch, India and
Milos, Greece – some candidate tunnel backfill materials. Swedish Nuclear Fuel
and Waste Management Company, SKB-R-09-53, Stockholm.

Oscarson, D.W. and D.A. Dixon. 1989. Elemental, mineralogical, and pore-solution
compositions of selected Canadian clays. Atomic Energy of Canada Limited,
Report AECL-9891, Chalk River.

Quigley, R.M. 1984. Quantitative mineralogy and preliminary pore-water chemistry of

candidate buffer and backfill materials for a nuclear fuel waste disposal vault.
Atomic Energy of Canada Limited, Technical Report, AECL 7827, Chalk River.

Radhakrishna, H.S. and H.T. Chan. 1982. Laboratory study of physical properties of
clay buffers for a nuclear fuel waste vault. Proc. 1st Intl. Conference on
Radioactive Waste Management, September 1982, Winnipeg.

Radhakrishna, H.S. and H.T.,Chan. 1985. Strength and hydraulic conductivity of clay-
based buffers for a deep underground nuclear fuel waste disposal vault. Atomic
Energy of Canada Limited, Technical Report, TR-327, Chalk River.

Riikonen, E. 2009. Flow-through and wetting tests of pre-compacted backfill blocks in a

quarter-scale test tunnel. MSc. Thesis, Helsinki University of Technology, Espoo.

Sandén, T., L. Börgesson, A. Dueck, R. Goudarzi and M. Lönnqvist. 2008. Deep

repository-engineered barrier system; erosion and sealing processes in tunnel
backfill materials investigated in laboratory. Swedish Nuclear Fuel and Waste
Management Co., SKB R-08-135, Stockholm.

SKB. 2010. Buffer, backfill and closure process report for the safety assessment SR-
Site, Swedish Nuclear Fuel and Waste Management Co., SKB, TR-10-47,
Stockholm.
 - 78 -

TKS. 2009. Nuclear Waste Management at Olkiluoto and Loviisa Power Plants: Review
of current status and future plans for 2010–2012. Posiva Oy, Olkiluoto.

Volckaert, G. 2000. A large scale in situ demonstration test for repository sealing in an
argillaceous host rock - Reseal Project Phase 1. EUR report 19612.

Vuorinen, J. 1999. Papers of Road Office in Finland 30/1999. The comparability of

Nordic abrasion test and Micro-Deval test. Finnish Road Administration, 2000,
28 pp. in Finnish.
 - 79 -

APPENDIX A: MATERIAL DESCRIPTIONS AND MANUFACTURERS'

DATA SHEETS
- 80 -
 - 81 -

CEBOGEL QSE (pg 2/2)

Typical values
Montmorillonite level X-ray diffraction 80 %
Moisture content DIN 18121 16 %

Chemical and physical properties

Composition High-quality activated sodium bentonite
Colour Grey green
Form Cylindrical rods
Dimensions Diameter 6.5 mm Length 5 – 20 mm
Density 2100 kg/m3
Bulk density 1100 kg/m3

Packaging
• 1000 kg packed in 25 kg polyethylene bags on a pallet with shrink film
• 1000 kg big bags

Revision date : 10-07-2003

Document no : CQ03IP

In so far as we can ascertain the above-stated information is correct. However, we are unable to provide any
guarantees with regard to the results that you will achieve with this. This specification is provided on the
condition that you determine yourself to what degree it is suitable for your purposes.

Page 2 of 2
 - 82 -

Fine Granular
Industrial Sodium Bentonite
MX 80
General Fine granular Sodium Bentonite with an average particle size
Description ranging between 16 and 200 mesh.

Functional Use Multi-purpose product noted for rapid dispersion in water.

Employed in a wide variety of industrial applications.

Purity Hydrous aluminium silicate comprised principally of the clay

mineral Montmorillonite. Montmorillonite content 90% minimum. Contains
small portions of feldspar, biotite, selenite, etc.

Chemical
Composition Typical Analysis (moisture free)
SiO2 63.02% MgO 2.67%
A12O3 21.08% Na2O 2.57%
Fe2O3 3.25% CaO 0.65%
FeO 0.35% H2O 5.64%
Trace 0.72%

Chemical
Formula A tri-layer expanding mineral structure of approximately:
(Al, Fe1.67, Mg 0.33) Si4O10 (OH2) Na+Ca2+0.33

Moisture
Content Maximum 12% as shipped.

Dry Particle Maximum 10% retained on 18 mesh (850 microns )

Size Maximum 15% passing 200 mesh (75 microns )

pH 5% solids dispersion 8.5 to 10.5

Viscosity 1 part bentonite to 15 parts deionised water (6.25% solids)

Dispersed on high-speed mixer. Fann viscometer , 8cps. Minimum

Packaging Multi – wall paper bags, (25 kg), big-bags or bulk

Data from website: masco.net/pdf/catalogs/liquidand powder/datasheets/TD_MX_80.pdf

- 83 -
 - 84 -

Wyoming Bentonite Used in Pellet Manufacture: NSB-BC200-2008

 - 85 -

IBECO RWC-2008: Milos HQ : AC200

 - 86 -

IBECO RWC-BF : Milos B : IBECO DEPONIT-BF

 - 87 -

IBECO DEPONIT CA : IBECO RWC-BF : Milos B

- 88 -
 - 89 -

APPENDIX B: RESULTS OF PELLET TESTING: DATA TABLES

CONTENTS

Page

Table B-1: Results of Free Swell Tests .......................................................................... 91

Table B-2: Density Achieved in Small Volume Pouring Tests for Clay-Only Materials .. 94
Table B-3: Density Achieved in Small Volume Pouring Tests Using Wyoming
Bentonite-Additive Pellets ............................................................................ 95
Table B-4: Thermal Properties of Pellets ....................................................................... 96
Table B-5: Crush Strength of Clay-Only Pellets ........................................................... 107
Table B-6: Crush Strength of Pellets for use as Gap fill in IFB or as Tunnel Fill
in HTP Geometry ....................................................................................... 109
 - 90 -

Table B-1: Results of Free Swell Tests

Free swell  Free swell 
index  index 
Test # Group ID Batch Number Liquid TDS (g/L) (cc/2g) (cc/g)
1 100% Wyom. Bent. Pellet 43 DW 0 11.8 5.9
Wyoming 35 g/L solution 35 7.3 3.7
70 g/L solution 70 7.5 3.8
275 g/L solution 275 5.7 2.8
2 90%‐10% B‐S Pellet 4 Distilled water 0 16.2 8.1
10% Sand 35 g/L solution 35 6.4 3.2
70 g/L solution 70 6.4 3.2
275 g/L solution 275 5.1 2.6
3 90%‐10% B‐S Pellet 47 DW 0 16.5 8.2
10% Sand 35 g/L solution 35 6.8 3.4
70 g/L solution 70 6.9 3.4
275 g/L solution 275 5.4 2.7
4 90%‐10% B‐S Pellet 30 DW 0 13.1 6.6
10% Sand 35 g/L solution 35 6.5 3.3
70 g/L solution 70 7.2 3.6
275 g/L solution 275 4.9 2.4
5 90%‐10% B‐I Pellet 6 DW 0 12.1 6.0
10% Illite 35 g/L solution 35 5.9 2.9
70 g/L solution 70 6.3 3.1
275 g/L solution 275 4.9 2.5
6 90%‐10% B‐I Pellet 46 DW 0 12.2 6.1
10% Illite 35 g/L solution 35 7.0 3.5
70 g/L solution 70 6.7 3.3
275 g/L solution 275 5.6 2.8
7 90%‐10% B‐I Pellet 32 DW 0 13.6 6.8
10% Illite 35 g/L solution 35 6.4 3.2
70 g/L solution 70 7.1 3.6
275 g/L solution 275 5.2 2.6
8 75%‐25% B‐S Pellet 5 DW 0 10.8 5.4
25 % Sand 35 g/L solution 35 7.0 3.5
70 g/L solution 70 7.0 3.5
275 g/L solution 275 5.9 2.9
9 75%‐25% B‐S Pellet 45 DW 0 10.1 5.0
25 % Sand 35 g/L solution 35 5.4 2.7
70 g/L solution 70 6.1 3.1
275 g/L solution 275 4.4 2.2
10 75%‐25% B‐S Pellet 54 DW 0 11.9 5.9
25 % Sand 35 g/L solution 35 6.0 3.0
70 g/L solution 70 5.9 2.9
275 g/L solution 275 4.6 2.3
 - 91 -

Free swell  Free swell 
index  index 
Test # Group ID Batch Number Liquid TDS (g/L) (cc/2g) (cc/g)
11 75%‐25% B‐I Pellet 9 DW 0 11.1 5.5
25 % Illite 35 g/L solution 35 5.7 2.9
70 g/L solution 70 7.0 3.5
275 g/L solution 275 5.0 2.5
12 75%‐25% B‐I Pellet 44 DW 0 10.0 5.0
25 % Illite 35 g/L solution 35 5.5 2.7
70 g/L solution 70 4.6 2.3
275 g/L solution 275 4.4 2.2
13 75%‐25% B‐I Pellet 53 DW 0 11.0 5.5
25 % Illite 35 g/L solution 35 5.3 2.7
70 g/L solution 70 6.0 3.0
275 g/L solution 275 4.3 2.2
14 50%‐50% B‐S Pellet 48 DW 0 11.3 5.6
50% Sand 35 g/L solution 35 4.1 2.1
70 g/L solution 70 4.8 2.4
275 g/L solution 275 3.6 1.8
15 50%‐50% B‐S Pellet 40 DW 0 9.9 5.0
50% Sand 35 g/L solution 35 6.0 3.0
70 g/L solution 70 5.8 2.9
275 g/L solution 275 4.4 2.2
16 50%‐50% B‐I Pellet 50 DW 0 13.6 6.8
50% Illite 35 g/L solution 35 4.3 2.2
70 g/L solution 70 4.9 2.4
275 g/L solution 275 3.8 1.9
17 50%‐50% B‐I Pellet 39 DW 0 12.1 6.1
50% Illite 35 g/L solution 35 4.8 2.4
70 g/L solution 70 5.3 2.6
275 g/L solution 275 4.0 2.0
18 Milos B Pellet 11 DW 0 9.2 4.6
35 g/L solution 35 4.2 2.1
70 g/L solution 70 4.2 2.1
275 g/L solution 275 3.9 2.0
19 Milos B Pellet 55 DW 0 8.6 4.3
35 g/L solution 35 4.6 2.3
70 g/L solution 70 4.8 2.4
275 g/L solution 275 3.0 1.5
20 Milos AC200 Pellet 41 DW 0 11.2 5.6
35 g/L solution 35 6.6 3.3
70 g/L solution 70 7.2 3.6
275 g/L solution 275 5.5 2.7
 - 92 -

Free swell  Free swell 
index  index 
Test # Group ID Batch Number Liquid TDS (g/L) (cc/2g) (cc/g)
21 Milos AC200 Pellet 52 DW 0 16.3 8.1
35 g/L solution 35 7.4 3.7
70 g/L solution 70 7.5 3.8
275 g/L solution 275 5.7 2.9
22 Asha Pellet 17 DW 0 10.4 5.2
35 g/L solution 35 4.6 2.3
70 g/L solution 70 4.8 2.4
275 g/L solution 275 3.9 1.9
23 Asha Pellet 35 DW 0 10.1 5.0
35 g/L solution 35 4.7 2.3
70 g/L solution 70 5.0 2.5
275 g/L solution 275 3.7 1.9
SKB MX‐80 pellets DW 0 15 7.5
Posiva Buffer Test DW 0 13.9 7
Cebogel  DW 0 16.5 8.3

Note: Bentonite used in production of mixed component pellets is Wyoming bentonite =

NSB-BC200-2008
 - 93 -

Table B-2: Density Achieved in Small Volume Pouring Tests for Clay-Only Materials

Single  Poured 
Roller  Roller  Average  Average  Average  Pellet  Crush  Bulk  Poured  Poured  Vibrated  Vibrated  Vibrated 
2  STDEV
Batch  Pellet  Load Speed  Pellet Pellet  b  Pellet  d  EMDD   Strength Density Dry  EMDD   Bulk  Dry  EMDD  
Material Number Size (psi) Setting w (%) (g/cc) (g/cc) (Mg/m3) (N) (Mg/m3) Density (Mg/m3) Density Density (Mg/m3)
Wyoming 1 L 1000 3 6.2 2.07 1.95 1.79 95 ±36.6
Wyoming 2 L 1500 3 6.5 2.13 2.00 1.85 144 ±25.7
Wyoming 56=1/3 HTP L 1500 3 10 2.11 1.92 1.76 235 ±29.8 1.12‐1.17 1.02‐1.06 0.85‐0.88 1.33‐1.43 1.21‐1.30 1.02‐1.11
Wyoming 2010 L 1450 10.0 2.12 1.93 1.77
Wyoming 3 L 1750 3 7.1 2.16 2.02 1.87 187 ±45.4 1.07 1.00 0.83
Wyoming 2010 L 1800 9.4 2.17 1.98 1.83
Wyoming 43 M 1500 3 9.6 2.09 1.91 1.76 119 ±31.1 1.17‐1.30 1.07‐1.16 0.89‐1.01 1.47 1.34 1.15
Wyoming 18 S 1500 4 7.4 2.23 2.08 1.94 61 ±10.0
Wyoming 19 S 1500 3 7.3 2.24 2.09 1.95 54 ±17.7
Wyoming 51 S 1500 4 9.7 2.18 1.99 1.84 89 ±8.2 1.16 1.06 0.88 1.35 1.23 1.04
Wyoming 29 S 1500 3 14.9 2.25 1.96 1.81 78 ±21.6 1.19 1.04 0.87 1.39 1.21 1.02
Wyoming 20 S 1750 4 7.2 2.26 2.11 1.98 79 ±18.0
Wyoming 21 S 1750 3 7.3 2.28 2.12 1.99 52 ±14.4 1.06 0.99 0.82
Wyoming 28 S 1750 3 13.8 2.30 2.02 1.85 79 ±22.3
Milos AC200 41 L 1000 3 11.4 2.04 1.83 1.70 113 ±14.8 1.05‐1.14 0.94‐1.02 0.77‐0.85 1.29 1.16 0.98
Milos AC200 42 L 1000 4.25 11.7 2.05 1.84 1.71 238 ±31.0
Milos AC200 36 S 1000 3 11.6 2.09 1.87 1.75 50 ±20.4
Milos AC200 37 S 1000 4.25 11.8 2.04 1.82 1.69 72 ±9.2 0.94 0.84 0.68
Milos AC200 52 S 1000 4.25 11.8 2.12 1.90 1.78 85 ±11.3 1.08 0.97 0.80 1.21 1.08 0.94
Asha 12 L 750 4.25 13.1 2.09 1.85 1.58 91 ±19.4
Asha 13 L 1000 4.25 13.0 2.17 1.92 1.66 195 ±40.8
Asha 14 L 1000 4 12.8 2.18 1.93 1.67 285 ±63.2 1.11 0.98 0.74 1.2 1.06 0.81
Asha 15 L 1000 3.5 13.0 2.18 1.93 1.67 220 ±28.0
Asha 16 L 1250 3.25 13.0 2.23 1.97 1.72 273 ±30.2
Asha 17 L 1250 3.5 12.9 2.20 1.95 1.70 246 ±30.4 1.01 0.89 0.67
Asha 34 S 1250 3 8.6 2.25 2.07 1.84 48 ±2.5
Asha 35 S 1250 3.25 8.6 2.23 2.05 1.82 69 ±14.1 1.01‐1.11 0.93‐1.02 0.70‐0.77 1.2 1.10 0.85
Milos B 10 L 1000 3 14.0 2.18 1.91 1.60 299 ±36.6
Milos B 11 L 500 4.25 16.2 2.20 1.89 1.58 261 ±11.6 1.07 0.92 0.63 1.29 1.11 0.8
Milos B 22 S 1000 3 13.5 2.24 1.97 1.67 100 ±16.0
Milos B 23 S 500 4.25 14.3 2.30 2.01 1.72 101 ±20.2
Milos B 55 S 1000 4.25 12.6 2.17 1.93 1.62 100 ±27.7 1.07‐1.10 0.95‐0.98 0.65‐0.68 1.28 1.14 0.81
Cebogel Extruded * 15.2 1.83 1.59 1.44 170 ±50.7 1.04/1.05 0.91 0.87 1.29 1.12 0.98
SKB‐MX80 pellets * 9.3 1.84 1.69 1.51 NM NM 1.1 1.00 0.83 1.244 1.14 0.96
Posiva Buffer Test * 15.6 1.79 1.55 1.36 136 ±20.7 1.09‐1.13 0.97 0.80 1.31 1.13 0.95

Note: Bentonite used in AECL pellets is Wyoming bentonite = NSB-BC200-2008 *. Materials supplied by SKB and Posiva, different sizes than AECL pellets
 - 94 -

Table B-3: Density Achieved in Small Volume Pouring Tests Using Wyoming Bentonite-Additive Pellets

Single  Poured  Poured  Poured  Vibrated  Vibrated 

Roller  Roller  Average  Average  Average  Pellet  Crush  Strength   Bulk  Dry Bulk  Pellet  Bulk  Dry  Vibrated 
Batch  Load Speed  Pellet Pellet  b  Pellet  d  EMDD   Strength  STDEV Density Density   EMDD   Density   Density   EMDD  
Material Composition Number Pellet Size (psi) Setting w (%) (g/cc) (g/cc) (Mg/m3) (N) (Mg/m3) (Mg/m3) (Mg/m3) (Mg/m3) (Mg/m3) (Mg/m3)
Wyoming Bentonite 1 L 1000 3 6.2 2.07 1.95 1.79 95 ±36.6
Wyoming Bentonite 2 L 1500 3 6.5 2.13 2.00 1.85 144 ±25.7
Wyoming Bentonite 2010 L 1500 9.3 2.14 1.96 1.80
Wyoming Bentonite 56=1/3 HTP L 1500 3 10 2.11 1.92 1.76 235 ±29.8 1.12‐1.17 1.02‐1.06 0.85‐0.88 1.33‐1.43 1.21‐1.30 1.02‐1.11
Wyoming Bentonite 3 L 1750 3 7.1 2.16 2.02 1.87 187 ±45.4 1.07 1.00 0.83
Wyoming Bentonite 2010 L 1800 9.4 2.17 1.98 1.83
Wyoming Bentonite 43 M 1500 3 9.6 2.09 1.91 1.76 119 ±31.1 1.17‐1.30 1.07‐1.19 0.89‐1.01 1.47 1.34 1.15
Wyoming Bentonite 18 S 1500 4 7.4 2.23 2.08 1.94 61 ±10.0
Wyoming Bentonite 19 S 1500 3 7.3 2.24 2.09 1.95 54 ±17.7
Wyoming Bentonite 51 S 1500 4 9.7 2.18 1.99 1.84 89 ±8.2 1.16‐1.17 1.06 0.88 1.35 1.23 1.04
Wyoming Bentonite 29 S 1500 3 14.9 2.25 1.96 1.81 78 ±21.6 1.19 1.04 0.87 1.39 1.21 1.02
Wyoming Bentonite 20 S 1750 4 7.2 2.26 2.11 1.98 79 ±18.0
Wyoming Bentonite 21 S 1750 3 7.3 2.28 2.12 1.99 52 ±14.4 1.06 0.99 0.82
Wyoming Bentonite 28 S 1750 3 13.8 2.30 2.02 1.85 79 ±22.3
10 % Quartz Sand 4 L 1500 3 5.8 2.16 2.04 1.84 217 ±25.5
10 % Quartz Sand 47 M 1500 3 7.5 2.19 2.04 1.84 189 ±26.6 1.37‐1.49 1.27‐1.39 1.02‐1.14 1.56 1.45 1.19
10 % Quartz Sand 24 S 1500 3 5.6 2.32 2.20 2.04 116 ±7.1
10 % Quartz Sand 30 S 1500 3 13.7 2.36 2.08 1.89 87 ±10.9
25% Quartz sand 5 L 1500 3 5.6 2.19 2.07 1.77 213 ±17.0 1.08‐1.16 1.02‐1.10 0.69‐0.81 1.26‐1.33 1.19‐1.26 0.83‐0.90
25% Quartz sand 45 M 1500 3 6.1 2.20 2.07 1.77 223 ±21.3 1.29‐1.37 1.22‐1.29 0.86‐0.92 1.59‐1.63 1.50‐1.54 1.12‐1.16
25% Quartz sand 25 S 1500 3 4.6 2.27 2.17 1.91 95 ±9.4 1.06 1.01 0.68
25% Quartz sand 54 S 1500 4 7.1 2.27 2.12 1.84 169 ±25.7 1.21 1.13 0.78 1.4 1.31 0.94
25% Quartz sand 31 S 1500 3 14.3 2.35 2.06 1.76 119 ±9.6
50% Quartz sand 48 M 1750 3 3.9 2.29 2.20 1.72 190 ±37.2
50% Quartz sand 40 S 1750 3 3.6 2.04 1.97 1.38 65 ±13.7 1.26 1.22 0.64 1.37 1.32 0.72
10% Illite 6 L 1500 3 5.2 2.14 2.03 1.82 239 ±34.6
10% Illite 7 L 1750 3 5.3 2.13 2.02 1.81 100 ±28.8 1.23 1.17 0.92 1.41 1.34 1.08
10% Illite 46 M 1500 3 7.8 2.15 1.99 1.78 149 ±31.3 1.32 1.22 0.94 1.33 1.23 0.94
10% Illite 26 S 1500 3 4.9 2.41 2.30 2.16 67 ±22.6
10% Illite 32 S 1500 3 14.3 2.21 1.93 1.71 95 ±20.6
25% Illite 8 L 1500 3 4.1 2.10 2.02 1.70 101 ±14.0 1.14 1.10 0.79 1.31 1.26 0.89
25% Illite 9 L 1750 3 4.3 2.18 2.09 1.79 166 ±80.0 1.02 0.98 0.66
25% Illite 44 M 1500 3 4.7 2.05 1.96 1.62 136 ±19.6 1.18 1.13 0.78 1.29 1.23 0.87
25% Illite 27 S 1750 3 3.9 2.15 2.07 1.76 39 ±8.9
25% Illite 33 S 1750 3 13.9 2.20 1.93 1.59 92 ±8.3 1.02 0.90 0.65
25% Illite 53 S 1000 4 10.2 2.23 2.02 1.70 129 ±14.5 1.13‐1.16 0.87 0.57 1.35 1.23 0.87
50% Illite 49 M 1500 3 5.6 2.09 1.98 1.38 85 ±15.0 1.21 1.15 0.59 1.36 1.29 0.69
50% Illite 50 M 1750 3 5.6 2.21 2.09 1.53 179 ±14.3
50% Illite 38 S 1500 3 4.6 2.22 2.12 1.57 78 ±31.0 1.17 1.12 0.57 1.36 1.30 0.7
50% Illite 39 S 1750 3 4.6 2.23 2.13 1.59 72 ±13.9

Note: Bentonite used in production of mixed component pellets is Wyoming bentonite = NSB-BC200-2008
 - 95 -

Table B-4: Thermal Properties of Pellets

(Note: TC = Thermal Conductivity, TD = Thermal Diffusivity, SH = Specific Heat)

Batch  Rep  Poured  Vibrated  Mass of  bulk 

   TC  TD  SH  TC  TD  SH  Pellets    
2 3 2
      (W/(m∙K))  (mm /s)  (Mj/m K)  (W/(m∙K))  (mm /s)  (Mj/m3K)  (kg)  (Mg/m3) 
100% Wyoming Bentonite ‐ Large                   
56  1  0.295  4.764  0.062  1.792  1.179 
   2  0.223  1.854  0.120  1.760  1.158 
   3  0.307  4.963  0.062  1.816  1.195 
   Average:  0.275  3.860  0.081  1.177 
   1  0.318  8.046  0.040  2.035  1.339 
   2  0.365  9.976  0.037  2.019  1.328 
   3  0.347  6.240  0.056  1.988  1.308 
   Average:           0.344  8.087  0.044     1.325 
100% Wyoming Bentonite ‐ Medium                   
43  1  0.360  1.843  0.196  1.341  1.355 
   2  0.449  3.271  0.137  1.219  1.232 
   3  0.431  6.341  0.068  1.228  1.240 
   Average:  0.414  3.818  0.134  1.276 
   1  0.586  2.722  0.215  1.447  1.462 
   2  0.583  3.713  0.157  1.466  1.481 
   3  0.600  1.768  0.339  1.463  1.478 
   Average:           0.590  2.734  0.237     1.474 
100% Wyoming Bentonite ‐ Small                   
29  1  0.363  2.370  0.153  1.800  1.184 
   2  0.319  1.854  0.172  1.817  1.195 
   Average:  0.341  2.112  0.163  1.190 
   1  0.592  2.598  0.228  2.111  1.389 
   2  0.634  2.855  0.222  2.107  1.386 
   Average:  0.613  2.727  0.225  1.388 
 - 96 -

  Poured  Vibrated  Mass of  bulk 

Batch  Rep 
TC  TD  SH  TC  TD  SH  Pellets   
   
(W/(m∙K))  (mm2/s)  (Mj/m3K)  (W/(m∙K))  (mm2/s)  (Mj/m3K)  (kg)  (Mg/m3) 
 
100% Wyoming Bentonite ‐ Small 
51  1  0.424  1.683  0.252  1.768  1.163 
   2  0.513  2.146  0.239  1.723  1.133 
   3  0.425  1.032  0.412  1.778  1.170 
   Average:  0.454  1.620  0.301  1.155 
100% Wyoming Bentonite ‐ Large 
   1  0.425  1.253  0.339  4.675  1.172 
   2  0.455  1.126  0.404  4.675  1.172 
   3  0.462  1.434  0.322  4.675  1.172 
   Average:  0.447  1.271  0.355  1.172 
100% Wyoming Bentonite Using Different Sized Testing Boxes           
   1        0.522  1.246  0.419  5.400  1.353 
   1        0.491  0.614  0.800  2.072  1.363 
   2        0.498  1.327  0.375  2.044  1.345 
   Average:           0.503  1.062  0.531     1.354 
100% Wyoming Bentonite ‐ Large, with 20% Bentonite Fines             
56  1  0.658  2.292  0.288  2.400  1.579 
   2  0.644  1.854  0.364  2.400  1.579 
   Average:           0.651  2.073  0.326     1.579 
100% Wyoming Bentonite ‐ Large, with 30% Bentonite Fines             
56  1  0.697  1.491  0.468  2.600  1.711 
   2  0.652  1.803  0.361  2.600  1.711 
   3  0.657  1.193  0.551  2.600  1.711 
   Average:           0.669  1.496  0.460     1.711 
 - 97 -

  Poured  Vibrated  Mass of  bulk 

Batch  Rep 
TC  TD  SH  TC  TD  SH  Pellets   
   
(W/(m∙K))  (mm2/s)  (Mj/m3K)  (W/(m∙K))  (mm2/s)  (Mj/m3K)  (kg)  (Mg/m3) 
 
100% Wyoming Bentonite ‐ Small, with 30% Bentonite Fines             
29  1  0.628  0.697  0.901  2.730  1.796 
   2  0.599  0.732  0.819  2.730  1.796 
   3  0.620  0.662  0.936  2.730  1.796 
   Average:     0.616 0.697  0.885 1.796
100% Wyoming Bentonite ‐ Small, with 30% Bentonite Fines             
51  1  0.696  1.143  0.609  2.730  1.796 
   2  0.696  1.324  0.525  2.730  1.796 
   3  0.676  0.925  0.731  2.730  1.796 
   Average:     0.689 1.131  0.622 1.796
Asha ‐ Large                         
14 & 15  1  0.406  0.268  0.152  1.697  1.116 
   2  0.446  3.547  0.126  1.691  1.113 
   Average:  0.426  1.908  0.139  1.114 
   1  0.543  2.236  0.243  1.958  1.288 
   2  0.541  4.564  0.119  1.948  1.282 
   Average:     0.542 3.400  0.181 1.285
Asha ‐ Small                         
34 & 35  1  0.405  3.929  0.103  1.693  1.114 
   2  0.446  1.923  0.232  1.684  1.108 
   Average:  0.425  2.926  0.167  1.111 
   1  0.458  1.178  0.389  1.825  1.201 
   2  0.528  1.554  0.340  1.834  1.207 
   Average:           0.493  1.366  0.364     1.204 
 - 98 -

  Poured  Vibrated  Mass of  bulk 

Batch  Rep 
TC  TD  SH  TC  TD  SH  Pellets   
   
(W/(m∙K))  (mm2/s)  (Mj/m3K)  (W/(m∙K))  (mm2/s)  (Mj/m3K)  (kg)  (Mg/m3) 
 
Milos AC200 ‐ Large                         
41 & 42  1  0.454  4.640  0.098  1.732  1.139 
   2  0.438  4.828  0.091  1.732  1.140 
   Average:  0.446  4.734  0.094  1.140 
   1  0.604  1.928  0.313  1.978  1.301 
   2  0.618  2.580  0.240  1.957  1.287 
   Average:           0.611  2.254  0.276     1.294 
Milos AC200 ‐ Small 
52  1  0.441  1.959  0.225  1.638  1.078 
   2  0.435  2.371  0.184  1.628  1.071 
   3  0.434  2.367  0.183  1.635  1.076 
   Average:  0.436  2.232  0.197  1.075 
   1  0.542  2.427  0.223  1.866  1.227 
   2  0.543  2.152  0.252  1.813  1.193 
   3  0.538  2.284  0.235  1.859  1.223 
   Average:           0.541  2.288  0.237     1.214 
Milos B ‐ Large                         
11  1  0.439  5.456  0.081  1.595  1.049 
   2  0.426  4.945  0.086  1.665  1.095 
   Average:  0.433  5.201  0.083  1.072 
   1  0.622  2.478  0.251  1.922  1.264 
   2  0.645  2.673  0.241  1.983  1.305 
   Average:           0.634  2.576  0.246     1.285 
 - 99 -

  Poured  Vibrated  Mass of  bulk 

Batch  Rep 
TC  TD  SH  TC  TD  SH  Pellets   
   
(W/(m∙K))  (mm2/s)  (Mj/m3K)  (W/(m∙K))  (mm2/s)  (Mj/m3K)  (kg)  (Mg/m3) 
 
Milos B ‐ Small                         
55  1  0.472  1.230  0.384  1.634  1.075 
   2  0.472  2.438  0.194  1.698  1.117 
   3  0.468  1.863  0.251  1.671  1.099 
   Average:  0.471  1.844  0.276  1.097 
   1  0.609  1.710  0.356  1.958  1.288 
   2  0.585  2.162  0.271  1.956  1.287 
   3  0.557  1.803  0.309  1.925  1.267 
   Average:           0.584  1.892  0.312     1.280 
90% Wyoming Bentonite:10% Silica Sand ‐ Medium             
47  1  0.441  5.348  0.082  0.861  1.471 
   2  0.432  10.190  0.042  0.865  1.479 
   3  0.490  4.979  0.098  0.886  1.514 
   Average:  0.454  6.839  0.074  1.488 
   1  0.684  2.366  0.289  1.016  1.737 
   2  0.576  2.740  0.210  1.010  1.726 
   3  0.584  1.794  0.326  1.009  1.725 
   Average:           0.614  2.300  0.275     1.729 
90% Wyoming Bentonite:10% Silica Sand ‐ Medium, with 30% Bentonite Fines       
47  1  0.515  2.743  0.188  1.300  2.222 
   2  0.518  3.478  0.149  1.300  2.222 
   3  0.514  2.790  0.184  1.300  2.222 
   Average:           0.516  3.004  0.174     2.222 
 - 100 -

  Poured  Vibrated  Mass of  bulk 

Batch  Rep 
TC  TD  SH  TC  TD  SH  Pellets   
   
(W/(m∙K))  (mm2/s)  (Mj/m3K)  (W/(m∙K))  (mm2/s)  (Mj/m3K)  (kg)  (Mg/m3) 
 
75% Wyoming Bentonite:25% Silica Sand ‐ Large                
5  1  0.550  1.824  0.300  1.535  1.077 
   2  0.588  6.321  0.093  1.532  1.075 
   3  0.548  4.688  0.117  1.530  1.074 
   Average:  0.562  4.278  0.170  1.076 
   1  0.562  1.833  0.307  2.060  1.445 
   2    
   3    
   Average:           0.562  1.833  0.307     1.445 
75% Wyoming Bentonite:25% Silica Sand ‐ Medium             
45  1  0.390  7.751  0.050  0.878  1.501 
   2  0.481  1.326  0.363  0.875  1.496 
   3  0.484  5.460  0.089  0.876  1.497 
   Average:  0.452  4.846  0.167  1.498 
   1  0.583  2.663  0.219  0.963  1.646 
   2  0.561  4.526  0.124  0.949  1.623 
   3  0.622  2.231  0.279  0.944  1.614 
   Average:           0.588  3.140  0.207     1.628 
75% Wyoming Bentonite:25% Silica Sand ‐ Small                
54  1  0.436  1.349  0.323  1.807  1.189 
   2  0.440  1.055  0.417  1.861  1.224 
   3  0.446  1.231  0.362  1.841  1.211 
   Average:  0.441  1.212  0.368  1.208 
   1  0.496  0.992  0.500  2.087  1.373 
   2  0.561  1.638  0.343  2.142  1.409 
   3  0.510  1.051  0.485  2.143  1.410 
   Average:           0.522  1.227  0.443     1.397 
 - 101 -

  Poured  Vibrated  Mass of  bulk 

Batch  Rep 
TC  TD  SH  TC  TD  SH  Pellets   
   
(W/(m∙K))  (mm2/s)  (Mj/m3K)  (W/(m∙K))  (mm2/s)  (Mj/m3K)  (kg)  (Mg/m3) 
 
50% Wyoming Bentonite: 50% Silica Sand ‐ Small                
40  1  0.461  2.361  0.195  0.596  1.419 
   2  0.430  2.669  0.161  0.607  1.446 
   3  0.421  1.299  0.324  0.604  1.438 
   Average:  0.438  2.110  0.227  1.434 
   1  0.549  1.270  0.432  0.722  1.718 
   2  0.494  1.119  0.441  0.694  1.652 
   3  0.509  1.474  0.345  0.681  1.621 
   Average:           0.517  1.288  0.406     1.664 
90% Wyoming Bentonite:10% Illite ‐ Large                
7  1  0.397  5.262  0.075  1.784  1.239 
   2  0.390  7.157  0.055  1.744  1.211 
   Average:  0.394  6.210  0.065  1.225 
   1  0.479 6.246  0.077 2.031 1.410
   2  0.494 2.851  0.173 2.032 1.411
   Average:           0.487  4.549  0.125     1.411 
90% Wyoming Bentonite:10% Illite ‐ Medium                
46  1  0.440  4.869  0.090  0.774  1.323 
   2  0.490  3.034  0.161  0.771  1.318 
   3  0.553  3.286  0.168  0.772  1.320 
   Average:  0.494  3.730  0.140  1.320 
   1  0.543  2.496  0.217  0.789  1.349 
   2  0.635  2.474  0.257  0.790  1.351 
   3  0.504  3.182  0.158  0.762  1.302 
   Average:           0.561  2.717  0.211     1.334 
 - 102 -

  Poured  Vibrated  Mass of  bulk 

Batch  Rep 
TC  TD  SH  TC  TD  SH  Pellets   
   
(W/(m∙K))  (mm2/s)  (Mj/m3K)  (W/(m∙K))  (mm2/s)  (Mj/m3K)  (kg)  (Mg/m3) 
 
75% Wyoming Bentonite:25% Illite ‐ Large                
8  1  0.409  4.493  0.091  1.750  1.151 
   2  0.359  5.226  0.069  1.701  1.119 
   Average:  0.384 4.860 0.080 1.135
   1  0.474  4.823  0.098  1.996  1.313 
   2  0.508  1.952  0.260  1.970  1.296 
   Average:           0.491  3.388  0.179     1.305 
75% Wyoming Bentonite : 25% Illite ‐ Medium Pellets             
44  1  0.413  3.182  0.130  0.791  1.351 
   2  0.399  4.470  0.089  0.793  1.356 
   3  0.404  5.183  0.078  0.819  1.400 
   Average:  0.405 4.278 0.099 1.369
   1  0.506  1.211  0.418  0.899  1.537 
   2  0.522  1.145  0.456  0.893  1.526 
   3  0.547  1.620  0.338  0.877  1.499 
   Average:           0.525  1.325  0.404     1.520 
75% Wyoming Bentonite:25% Illite ‐ Small                
53  1  0.407  1.386  0.294  1.774  1.167 
   2  0.434  2.677  0.162  1.760  1.158 
   3  0.426  1.594  0.267  1.766  1.162 
   Average:  0.422  1.886  0.241  1.162 
   1  0.467  0.913  0.512  2.009  1.322 
   2  0.483  0.933  0.517  2.054  1.351 
   3  0.494  0.956  0.517  2.076  1.365 
   Average:     0.481 0.934  0.515 1.346
 - 103 -

  Poured  Vibrated  Mass of  bulk 

Batch  Rep 
TC  TD  SH  TC  TD  SH  Pellets   
   
(W/(m∙K))  (mm2/s)  (Mj/m3K)  (W/(m∙K))  (mm2/s)  (Mj/m3K)  (kg)  (Mg/m3) 
 
75% Wyoming Bentonite:25% Illite ‐ Large, with 30% Bentonite Fines          
8  1  0.538  0.884  0.609  2.600  1.711 
   2  0.552  2.024  0.273  2.600  1.711 
   3  0.531  1.275  0.416  2.600  1.711 
   Average:           0.540  1.394  0.433     1.711 
50% Wyoming Bentonite:50% Illite ‐ Medium                
49 & 50  1  0.546  5.871  0.093  1.829  1.203 
   2  0.602  3.466  0.174  1.823  1.200 
   3  0.424  8.306  0.051  1.854  1.220 
   4  0.461  2.616  0.176  1.831  1.205 
   Average:  0.508  5.065  0.124  1.207 
   1  0.523  1.629  0.321  2.084  1.371 
   2  0.523  1.067  0.490  2.047  1.347 
   Average:           0.523  1.348  0.406     1.359 
50% Wyoming Bentonite:50% Illite ‐ Small                
38 & 39  1  0.461  3.221  0.143  1.783  1.173 
   2  0.456  3.450  0.132  1.782  1.173 
   Average:  0.459  3.336  0.138  1.173 
   1  0.517  1.817  0.285  2.082  1.370 
   2  0.512  1.793  0.296  2.059  1.355 
   Average:           0.514  1.805  0.290     1.362 
Buffer Test Pellets: Trial #1                         
   1  0.571  2.150  0.265  1.102 
   2  0.496  2.931  0.169  1.101 
   3  0.519  1.951  0.266  1.100 
   4  0.461  4.688  0.098  1.100 
   Average:  0.512  2.930  0.200  1.101 
 - 104 -

  Poured  Vibrated  Mass of  bulk 

Batch  Rep 
TC  TD  SH  TC  TD  SH  Pellets   
   
(W/(m∙K))  (mm2/s)  (Mj/m3K)  (W/(m∙K))  (mm2/s)  (Mj/m3K)  (kg)  (Mg/m3) 
 
Buffer Test Pellets: Trial #2                       
   1  0.520  2.712  0.235  1.663  1.094 
   2  0.513  5.126  0.100  1.672  1.100 
   Average:  0.517  3.919  0.168  1.097 
   1  0.540  1.255  0.430  1.924  1.266 
   2  0.585  1.182  0.495  1.981  1.303 
   Average:           0.562  1.219  0.462     1.284 
Buffer Test Pellets: Trial #3; mc=15.6%                         
   1  0.297  1.648  0.18  1.133 
   2  0.333  2.151  0.155  1.132 
   3  0.264  1.103  0.240  1.124 
   Average  0.287  1.634  0.192  1.130 
   1  0.622  2.69  0.231  1.294 
   2  0.658  2.569  0.256  1.290 
   3  0.664  2.397  0.277  1.286 
   Average  0.648  2.552  0.255  1.290 
Cebogel  mc=15.2% 
1  0.501  2.553  0.196  1.133 
2  0.500  1.794  0.279  1.132 
3  0.521  1.687  0.309  1.124 
Average  0.503  2.011  0.261  1.130 
0.602  1.578  0.382  1.294 
0.604  1.353  0.447  1.290 
0.616  1.694  0.364  1.286 
Average  0.607  1.542  0.398  1.290 
 - 105 -

  Poured  Vibrated  Mass of  bulk 

Batch  Rep 
TC  TD  SH  TC  TD  SH  Pellets   
   
(W/(m∙K))  (mm2/s)  (Mj/m3K)  (W/(m∙K))  (mm2/s)  (Mj/m3K)  (kg)  (Mg/m3) 
 
SKB‐MX80  mc=9.3% 
1  0.360  1.086 
  2  0.390              1.100 
  3  0.329              1.100 
  Average  0.360              1.095 
  1        0.505  3.672  0.138    1.233 
  2        0.425  2.193  0.194    1.256 
  3        0.440  1.359  0.323    1.243 
  Average        0.457  2.408  0.218    1.244 
 - 106 -

Table B-5: Crush Strength of Clay-only Pellets

Water Dry Crush

Batch Content Density Strength
Size STDEV
No. 3
% Mg/m N
100% Wyoming  1  L  6.2  1.9  95  37 
100% Wyoming  2  L  6.5  2.0  144  26 
100% Wyoming  3  L  7.1  2.0  187  45 
   Average  6.6  2.0  142  52 
100% Wyoming  56  L  10  1.9  235  30 
Milos AC200  41  L  11.4  1.8  113  15 
Milos AC200  42  L  11.7  1.8  238  31 
   Average  11.6  1.8  175  70 
Milos B 10  L  14.0  1.9  299  37 
Milos B 11  L  16.2  1.9  261  12 
   Average  15.1  1.9  280  32 
Asha 12  L  13.1  1.8  91  19 
Asha 13  L  13.0  1.9  195  41 
Asha 14  L  12.8  1.9  285  63 
Asha 15  L  13.0  1.9  220  28 
Asha 16  L  13.0  2.0  273  30 
Asha 17  L  12.9  1.9  246  30 
   Average  13.0  1.9  215  73 
Cebogel       15.2    1.59  170 51 
Buffer Test ‐ MX‐80         15.6  1.55   136 21 
Note: Wyoming bentonite = NSB-BC200-2008
 - 107 -

Table B-5: Crush Strength of Clay-only Pellets (continued)

Water Dry Crush

Batch Content Density Strength
Size STDEV
No.
3
% Mg/m N
100% Wyoming  18  S  7.4  2.08  61  10 
100% Wyoming  19  S  7.3  2.09  54  18 
100% Wyoming  20  S  7.2  2.11  78  18 
100% Wyoming  21  S  7.3  2.12  52  14 
   Average  7.3  2.1  61  17 
100% Wyoming  28  S  13.8  2.02  79  22 
100% Wyoming  29  S  14.9  1.96  78  22 
      Average  14.4  2.0  79  20 
100% Wyoming  51  S  9.7  1.99  89  8 
Milos AC200  36  S  11.6  1.87  50  20 
Milos AC200  37  S  11.8  1.82  72  9 
Milos AC200  52  S  11.8  1.90  85  11 
   Average  11.7  1.9  69  20 
Asha 34  S  8.6  2.07  48  3 
Asha 35  S  8.6  2.05  69  14 
      Average  8.6  2.1  59  15 
Milos B 22  S  13.5  1.97  100  16 
Milos B 23  S  14.3  2.01  101  20 
Milos B 55  S  12.6  1.93  100  28 
Average  13.5  2.0  101  20 
Note: Wyoming bentonite = NSB-BC200-2008
 - 108 -

Table B-6: Crush Strength of Pellets for use as Gap Fill in IFB or as Tunnel Fill
in HTP Geometry

Water Dry Crush

Batch Content Density Strength
Size STDEV
No. 3
% Mg/m N
100% Wyoming  1 L 6.2 1.9 95 37 
100% Wyoming  2 L 6.5 2.0 144 26 
100% Wyoming  3 L 7.1 2.0 187 45 
Average 6.6 2.0 142 52
100% Wyoming  56 L 10 1.9 235 30 
10% Sand 4 L 5.8 2.0 217 25 
25% Sand 5 L 5.6 2.1 213 17 
10% Illite 6 L 5.2 2.0 239 35 
10% Illite 7 L 5.3 2.0 100 29 
Average 5.3 2.0 169 80
25% Illite 8 L 4.1 2.0 101 14 
25% Illite 9 L 4.3 2.1 166 80 
Average 4.2 2.1 133 63
100% Wyoming  43 M 9.6 1.9 119 31 
10% Sand 47 M 7.5 2.0 189 27 
25% Sand 45 M 6.1 2.1 223 21 
50% Sand 48 M 3.9 2.2 190 37 
10% Illite 46 M 7.8 2.0 149 31 
25% Illite 44 M 4.7 2.0 136 20 
50% Illite 49 M 5.6 2.0 85 15 
50% Illite 50 M 5.6 2.1 179 14 
Average 5.6 2.0 132 52
Note: Wyoming bentonite = NSB-200-2008
 - 109 -

Table B-6: Crush Strength of Pellets for use as Gap Fill in IFB or as Tunnel Fill in HTP
Geometry (continued)

Batch Water Dry Crush STDEV

Size
No. Content Density Strength
(%) (Mg/m3) (N) (N)
100% Wyoming  18 S 7.4 2.1 61 10 
100% Wyoming  19 S 7.3 2.1 54 18 
100% Wyoming  20 S 7.2 2.1 78 18 
100% Wyoming  21 S 7.3 2.1 52 14 
Average 7.3 2.1 61 17
100% Wyoming  28 S 13.8 2.0 79 22 
100% Wyoming  29 S 14.9 2.0 78 22 
Average 14.4 2.0 79 20
100% Wyoming  51 S 9.7 2.0 89 8 
10% Sand 24 S 5.6 2.2 116 7 
10% Sand 30 S 13.7 2.1 87 11 
25% Sand 25 S 4.6 2.2 95 9 
25% Sand 54 S 7.1 2.1 169 26 
25% Sand 31 S 14.3 2.1 119 10 
50% Sand 40 S 3.6 2.0 65 14 
10% Illite 26 S 4.9 2.3 67 23 
10% Illite 32 S 14.3 1.9 95 21 
25% Illite 27 S 3.9 2.1 39 9 
25% Illite 33 S 13.9 1.9 92 8 
25% Illite 53 S 10.2 2.0 129 15 
50% Illite 38 S 4.6 2.1 78 31 
50% Illite 39 S 4.6 2.1 72 14 
Average 4.6 2.1 75 22
Note: Wyoming bentonite = BC- NSB-200-2008