Draft - TSpace Repository: Home · 2016-02-11 · Draft 3 RÉSUMÉ Le traitement du drainage minier...
Transcript of Draft - TSpace Repository: Home · 2016-02-11 · Draft 3 RÉSUMÉ Le traitement du drainage minier...
Draft
Preliminary geotechnical assessment of the potential use of
mixtures of soil and acid mine drainage neutralisation sludge as material for the moisture retention layer of covers
with capillary barrier effects
Journal: Canadian Geotechnical Journal
Manuscript ID cgj-2014-0319.R2
Manuscript Type: Article
Date Submitted by the Author: 23-Jul-2015
Complete List of Authors: Mbonimpa, Mamert; Université du Québec en Abitbi-Témiscamingue, Bouda, Médard; UQAT, RIME Demers, Isabelle; UQAT, RIME Benzaazoua, Mostafa; UQAT, RIME Bois, Denis; UQAT, RIME Gagnon, Mario; IMAGOLD,
Keyword: Acid mine drainage, Soil-sludge mixtures, Hydrogeotechnical properties, Mine site reclamation, Neutralisation sludge
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
1
Preliminary geotechnical assessment of the potential use of mixtures of soil and acid mine
drainage neutralisation sludge as material for the moisture retention layer of covers with
capillary barrier effects
Mamert Mbonimpa1,a, Médard Bouda1,b, Isabelle Demers1,c, Mostafa Benzaazoua1,d, Denis Bois1,e
and Mario Gagnon2,f
1Research Institute on Mines and Environment (RIME) – Université du Québec en Abitibi-
Témiscamingue, 445 Boulevard de l’Université, Rouyn-Noranda, Québec, J9X 5E8, Canada
2IAMGOLD, Division Doyon - Mouska –Westwood; Chemin Arthur-Doyon, Preissac, Québec,
J0Y 2E0, Canada
Corresponding author: Mamert Mbonimpa ([email protected])
Submitted in August 2014 for publication in the Canadian Geotechnical journal
Page 1 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
2
ABSTRACT
Lime treatment of acid mine drainage (AMD) generates sludge that is commonly stored in ponds
for dewatering. The use of soil-aged sludge-based mixtures for mine site rehabilitation can allow
emptying existing basins, extending their storage capacity, reducing the volume of borrow soil pit
required for mine site rehabilitation, and consequently reducing the mine footprint. The authors
investigated the geotechnical properties of silty soil–sludge mixtures (SSM) as possible component
of covers with capillary barrier effects (CCBE) to prevent AMD generation from mine waste. SSM
with 10, 15, 20, and 25% sludge (β= wet sludge mass/wet soil mass) were studied. Two water
contents were considered for each of the mixture components: 175 and 200% for the sludge and 7.5
and 12.5% for the soil. Results indicated that saturated hydraulic conductivity values ksat were in the
range of 10-5 cm/s for the soil and SSM mixtures at void ratios ranging from 0.28 to 0.53, with
values slightly decreasing when β was increased from 0 % to 25%. The air entry value (AEV)
increased from 20 kPa for the soil alone to 35 kPa for the SSM with β= 25%. These values of ksat
and AEV are comparable to those of materials used in the moisture retention layers of existing
efficient CCBE. However, the volumetric shrinkage increased from about 2% for the soil alone to
values ranging between 24 and 32% for the SSM with β = 25%, depending on the initial water
contents of the components. Tools are provided to estimate to which extent the use of sludge in
SSM can reduce the volume of borrow natural soil required for a moisture retention layer of a
CCBE.
Keywords: Acid mine drainage, Neutralisation, Soil-sludge mixtures, Hydrogeotechnical
properties, Moisture retention layer, CCBE, Mine site reclamation
Page 2 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
3
RÉSUMÉ
Le traitement du drainage minier acide (DMA) à l'aide de la chaux génère une boue généralement
stockée dans des bassins pour la décantation. L'utilisation de mélanges sol-boue (MSB) pour la
réhabilitation des sites miniers peut permettre la vidange des bassins existants, l’extension de leur
capacité de stockage et la réduction de l'empreinte de la mine. Pour cela, les auteurs ont étudié si
des mélanges d’un sol silteux et de la boue (MSB) peuvent présenter des propriétés géotechniques
appropriées pour être utilisés dans des couvertures avec effets de barrière capillaire (CEBC) visant
empêcher la génération du DMA. Des MSB avec des teneurs en boue β (= masse de boue humide /
de masse de sol humide) de 10, 15, 20 et 25% ont été étudiés. Deux teneurs en eau ont été
considérées pour chacun des constituants des mélanges: 175 et 200% pour les boues et 7,5 et 12,5%
pour le sol. Les résultats indiquent des valeurs de conductivité hydraulique saturés ksat dans la
gamme de 10-5 cm/s pour le sol et les MSB, avec des indices de vides compris entre 0,28 et 0,53.
Les valeurs ksat baissent légèrement lorsque β augmente de 0% à 25%. La pression d'entrée d'air
(AEV) passe de 20 kPa pour le sol seul à 35 kPa pour le MSB avec β = 25%. Ces valeurs de ksat et
AEV sont comparables à celles des matériaux utilisés dans les couches de rétention d'eau de CEBC
existantes et efficaces. Néanmoins, le retrait volumique est passé d'environ 2% pour le seul sol à des
valeurs comprises entre 24 et 32% pour le MSB avec β = 25% selon les teneurs en eau initiales des
composantes. Des outils sont fournis pour estimer dans quelle mesure l'utilisation de la boue dans
des MSB peut réduire le volume de matériaux naturels d'emprunt requis pour une CEBC
Mots clés : Drainage minier acide, neutralisation, mélanges sol-boues, propriétés
hydrogéotechniques, couche de rétention d’eau, CEBC, réhabilitation de sites miniers.
Page 3 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
4
1 INTRODUCTION
Sulphide minerals such as pyrite, pyrrhotite, chalcopyrite, and others contained in tailings and waste
rocks can oxidize in the presence of oxygen and water and generate acid mine drainage (AMD).
Once triggered, AMD is one of the major environmental challenges for the mining industry. A
discharge treatment plant must usually be implemented to meet regulatory effluent loading limits.
The most common method is active AMD treatment, which involves neutralization using different
alkaline chemicals, such as NaOH, Na2CO3, CaO, or CaCO3 (Brown et al. 2002; Younger et al.
2002; Zinck and Griffith 2012a, 2013), and produces large quantities of sludge. As the majority of
metals dissolved in the AMD water precipitate at alkaline pH values, sludge typically consists of
metal hydroxides, sulfates (ettringite, gypsum,..), unreacted lime, and other less abundant phases
(Zinck et al. 1997). Sludge from AMD neutralization requires responsible and effective
management. Sludge is usually stored in ponds for dewatering and permanent disposal. Sometimes
it is stored in the tailings ponds. Alkaline to neutral pH can be sustained within the sludge ponds for
decades, even centuries, due to excess alkalinity induced by lime in the sludge (Zinck et al. 1997).
This alkalinity helps to achieve chemical stability of the sludge (no sludge dissolution and no metal
remobilization) over time. Exposure of the sludge to acidic conditions can allow the re-dissolution
of metals, which constitutes the main environmental disadvantage that should be evaluated for each
case of sludge reuse. Thus, the long term conditions in which the sludge will be placed should be
evaluated to reduce the risk of metal remobilization.
Efficient sludge reuse can provide an environmental advantage because it can allow the extension
of the storage capacity of emptied ponds, which can then be used for ongoing temporary sludge
disposal. The use of soil covers incorporating sludge-based mixtures for mine site rehabilitation can
provide additional environmental and economic advantages: i) reducing the soil borrow pit volume
Page 4 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
5
and the transport distances of the borrow materials (sludge ponds are usually near waste storage
facilities), ii) potential use as substrate for vegetation. The assumption underlying this study is that
silty soil–sludge mixtures (SSM) have the potential for use as earthen cover material for mine site
rehabilitation. To verify this assumption, a laboratory testing program was conducted in this study
to assess the relevant hydrogeotechnical properties of SSM when used as a CCBE component in a
reclamation scenario (see also Bouda et al. 2012). In this paper, material (sludge and soil) sampling,
SSM preparation, and characterization of the soil, sludge and SSM are described. Results are
presented and discussed with respect to the possibility of using SSM as material for the moisture
retention layer in CCBE to prevent AMD generation from mine wastes.
2. BACKGROUND
AMD neutralisation sludge must be disposed of in an environmentally acceptable manner. Besides
the classical method of sludge deposition in ponds, Zinck (2006), Zinck et al. (2010), and Zinck and
Griffith (2012a,b; 2013) reviewed several sludge disposal alternatives and their effectiveness. These
alternatives include sludge mixed with tailings and waste rock, sludge covers over tailings, sludge
disposal in tailings ponds, with waste rock, in mine workings, in pits, under water covers, and as
hazardous waste, and sludge integration in cemented mine backfills and landfilling. Various
stabilization and solidification techniques have also been reported, including lime-based, cement-
based, and thermoplastic polymeric encapsulation (Zinck 2005; 2006; Zinck and Griffith 2012a,b).
However, sludge disposal along with tailings and waste rock is effective only when the mine wastes
are non-acid generating and/or oxidation is prevented. If the wastes oxidize and generate acid,
Page 5 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
6
sludge dissolution and metal remobilization become a serious concern. Zinck et al. (2010) showed
that a sludge cover placed over tailings does not control tailings oxidation.
Benzaazoua et al. (2006) investigated the option of adding a small quantity of sludge (0.15 and 0.3
% by total solid dry weight) to cemented tailings paste backfill and showed that no contaminants
were subsequently released. The contaminants are immobilized in the backfill material matrix
without reducing its strength. Tsang et al. (2013) observed that AMD sludge can be used
successfully for in situ soil stabilisation to minimize the environmental risk of As and Cu release at
a timber treatment site. Some AMD neutralisation sludge required adding compost or lignite to
simultaneously reduce the Cu and As leachability.
Demers et al. (2015a) performed column leaching tests in the laboratory to evaluate the efficiency
of the tailings-sludge mixture and waste rock-sludge mixtures to control AMD produced by tailings
and waste rocks. Results indicated that waste rock–sludge mixtures placed over waste rock (which
represented the worst case scenario because this contact must be avoided) reduces the metal loads in
the column effluent, which however remained acidic, while tailings-sludge mixtures deposited over
tailings reduced sulphide oxidation from tailings. Indeed sludge placed in acidic conditions (such as
in contact with waste rock) were more likely to dissolve and eventually would release metals. These
results were confirmed by results gathered by Demers et al. (2015b) from instrumented test cells
constructed on site to evaluate the performance of these waste-sludge mixtures to control AMD
under natural field conditions.
Water and earthen covers are effective technologies for mine site rehabilitation to control AMD
production. Multilayer earthen covers are widely used as conventional infiltration-limiting cover
Page 6 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
7
systems that typically consist of a low saturated hydraulic conductivity (ksat) layer (clay or
geosynthetic membrane, or a combination of the two) combined with other layers to provide lateral
drainage, resist erosion, prevent moisture loss and desiccation cracking. For this low ksat layer to
effectively control infiltration, the ksat value must be less than 10-7 cm/s to satisfy the regulatory
requirements.
Earthen covers that reduce the oxygen flux towards the tailings appear to be one of the most
efficient ways to prevent AMD generation in humid climates like in Canada. Single- and multi-
layered covers can be used for this purpose. Single-layer covers can consist of non-acid generating
tailings (such as desulfurized tailings) that are slightly reactive with oxygen (Bussière et al. 2004;
Demers et al. 2008, 2009, 2010; Sjoberg Dobchuck 2013). Such materials can contribute to reduce
the oxygen flux through the covers. Natural soils can also be used in single-layer covers (e.g. Pabst
et al. 2011, 2014). The long-term efficiency of single-layer covers as an oxygen barrier is improved
when the covers are used in combination with an elevated water table (MEND 1996; Ouangrawa et
al. 2010; Pabst et al. 2011, 2014). Water table located in the tailings and higher air-entry values of
the cover materials can help to maintain water saturation and to reduce oxygen fluxes. An
additional top layer made of coarse-grained materials can prevent moisture loss by evaporation and
evapotranspiration.
Multi-layer covers such as covers with capillary barrier effects, CCBE, are more efficient than
single-layer soil covers when the water table cannot be raised. Capillary barrier effects are created
under unsaturated conditions when a fine-textured material layer (the moisture retention layer
MRL) is placed over a coarse-textured material layer (the capillary break layer) (e.g., Nicholson et
al. 1989; Aubertin et al. 1994; Woyshner and Yanful 1995; O’Kane et al. 1998; Bussière et al.
Page 7 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
8
2006, 2007). When the underlying coarse material is drained, its unsaturated hydraulic conductivity
becomes much lower than that of the overlying fine-textured material. Downward movement of soil
moisture from the upper fine-grained layer is then prevented. The material used in the moisture
retention layer generally has high water retention potential, with an air entry value (AEV) ≥ 20 to
50 kPa (e.g., Aubertin and Chapuis 1991). An additional coarse-grained top layer provides drainage
and erosion control, and limits revegetation. Keeping the fine-grained layer near saturation reduces
the influx of atmospheric oxygen to reactive wastes and subsequent production of acidic drainage
(e.g., Aubertin et al. 1994; Yanful et al. 1999; Mbonimpa et al. 2003; Bussière et al. 2004; Dagenais
et al. 2012).
The hypothesis underlying this study is that a mixture of soil and sludge (both non-acid generating)
can be a good MRL material in CCBE for the reclamation of mine sites that generate AMD. This
mixture should have appropriate hydrogeotechnical properties and long-term physical and chemical
stability. This paper focuses on the laboratory evaluation of basic hydrogeotechnical properties of
SSM, including compaction behavior, saturated hydraulic conductivity and its evolution under
freeze-thaw cycles, free shrinkage behavior, and water retention curve. The potential use of soil
sludge mixtures (SSM) as MRL material is then evaluated in a preliminary manner by a comparison
between the saturated hydraulic conductivity and air entry values of the mixtures and of soils used
in efficient CCBE.
Page 8 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
9
3. MATERIALS AND METHODS
3.1. Sampling and SSM preparation
The sludge used in this study was sampled from the Doyon mine site operated by IAMGOLD,
located in northwestern Quebec, Canada. At this site, lime is used for AMD neutralization and a
flocculent is added prior to thickening. Two sludge ponds (pond B and North pond) containing a
volume of nearly 1 million cubic meters of sludge were used for sludge disposal over a period of
approximately 15 years (personal communication), and have been inactive for many years. The
sludge layer at the surface was relatively stable, but the sludge appeared wet and plastic at a very
shallow depth. Diverse and dense natural vegetation has grown on the sludge ponds. A survey to
evaluate the risk of metal bioaccumulation by the vegetation was a part of neutralizing lime sludge
valorization assessment (Smirnova et al. 2013).
In August 2010, eight sampling cores were drilled in pond B and seven cores in the North pond at
different accessible locations. At the drilling locations, the sludge thickness varied from 2.5 to 6.0
m in pond B and from 2.5 to 7.5 m in North pond. For each boring, undisturbed sludge was sampled
at intervals of 1 m from the surface to the natural soil substratum. These samples were stored in
tight plastic bags under 4° C until they were tested as described below. Furthermore, disturbed
sludge was sampled using a shovel close to each boring location at a depth of about 1.5 m and
stored in a barrel hermetically sealed after filling. For tests that required dry sludge samples, oven
drying was used in this study. Vertical water content profiles were determined for each boring
location. Sludge water content typically ranged from 100 to 250% (w/w) for the cores in pond B
and from 150 to 300% for the cores in North pond.
Page 9 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
10
The spatial distribution of the chemical composition of the sludge was determined by analyzing a
composite of the sludge cores collected at each sampling point in the two ponds. The elemental
analysis of all the sludge samples (results not presented here) revealed significant homogeneity of
the chemical composition for each boring location. Therefore, the disturbed sludge samples
collected from the different locations in the North pond were mixed to obtain a representative
composite sludge sample (called “sludge” hereafter) for further laboratory investigation.
Homogeneous soil-sludge mixtures (SSM) were prepared in the laboratory using a stand mixer with
10%, 15%, 20%, and 25% wet sludge content (β) with respect to the soil wet mass (i.e., β = wet
sludge mass/wet soil mass). According to the estimated sludge quantity available for potential use
over the tailings ponds at the Doyon site, 25% sludge mixture was identified as the maximum
sludge content. Samples of soil alone (β = 0) and sludge alone (β = ∞) were also studied as
reference materials. The corresponding proportions (%) of wet soil [(=100/(β +100)] and of wet
sludge [( = β /(β +100)] in the wet mixtures are given in Table 1. For the mixture preparation, initial
water content was set at a wi-sl of 175 and 200% for sludge and wi-so of 7.5 and 12.5% for soil.
Indices “sl” and “so” will be used in the following for sludge and soil, respectively. The sludge
contents and wet soil and sludge proportions in the mixtures given in Table 1 can also be expressed
in terms of dry soil, sludge and mixture weight (using β, wi-sl and wi-so). For example, the sludge
content β* based on dry mass mixture ratios (β* = dry sludge mass/dry soil mass) can be expressed
as follows: β*=.(1+ wi-so)β/(1+ wi-sl). With wi-so = 12.5%, wi-sl = 200% and β =25%, β* = 9.37%. In
this case, the corresponding proportions (%) of dry soil [(=100/(β* +100)] and of dry sludge [( =
100×β* /(β* +100)] in the dry mixtures are 91.4% and 8.6 %.
Page 10 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
11
3.2. Soil, sludge and SSM characterisation
The X-ray diffraction (XRD) data for the mineralogical characterization of the soil and sludge were
determined using the quantitative Rietveld method (relative precision of 0.5%) using TOPAS
software (Rietveld 1993; Young 1995). It should be mentioned that only crystalline materials were
identified by XRD while sludge generally contains amorphous or poorly-crystalline phases (e.g.,
MEND 2013). Chemical characterization were performed on the mixture components separately
(i.e., soil and sludge) using Inductively Coupled Plasma and Atomic Emission Spectrometry (ICP-
AES).
Mostly geotechnical properties, including specific gravity of solid grains, grain-size distribution,
dry density/water content relationship using the Proctor test, saturated hydraulic conductivity under
freeze–thaw cycles, free shrinkage, and water retention curves were determined for the soil alone
and for mixtures. The specific gravity of solid grains DR was determined using a Micromeritics
Accupyc 1330 helium pycnometer according to ASTM D5550-06. Grain-size distribution was
obtained by dry or wet sieving combined with a Malvern Mastersizer laser grain-size analyzer,
which provides a volume size distribution for diameters from 0.05 to 900 µm. Modified Proctor
tests were performed on the soil and SSM according to ASTM D1557.
Saturated hydraulic conductivity ksat was determined with a rigid wall permeameter using the
variable head method according to ASTM D5856. Samples were compacted at water contents close
to the Proctor optimum water content in the permeameter (except for the mixture at β =25% for
which a water content close to the natural value was targeted). Samples were then saturated by
subjecting them to an upward flow of distilled and de-aired water for at least 24 h using a vacuum
pump. Degrees of saturation of at least 98% were reached. The effect freeze–thaw cycles on ksat was
Page 11 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
12
also investigated. The rigid permeameter containing the sample was insulated laterally with glass
wool so that the sample underwent one-dimensional freezing from top to bottom. Freezing occurred
in a temperature controlled refrigerator. After 72 h of freezing at a temperature of approximately -
15° C, sample height hj (for each cycle j) was measured at different positions using a caliper and an
average value was calculated. The sample was then allowed to thaw for 48 hours at laboratory
temperature (about 20 °C) before determination of the saturated hydraulic conductivity according to
the ASTM D5856 procedure. For each of the 15 cycles applied, the saturated hydraulic conductivity
(ksat) and swelling ∆hj were determined, with ∆hj = hj - h0, where h0 is the height of the sample at
cycle 0 (before the first freezing step).
The sludge was expected to shrink due to its high water content and fine-grained particle size
distribution. The effect of adding sludge to the soil (assuming negligible shrinkage in the soil alone)
on SSM shrinkage was investigated. Free shrinkage behavior of the soil, sludge, and SSM was
examined in terms of the relationship e(w) between the void ratio e and the water content w. An
initially saturated sample was placed into a mold with smooth walls coated with a thin layer of
grease to prevent cracking. The volume of the mold, the water content and the specific gravity of
solid grains of the material were used to estimate the mass of the wet material to be placed in the
mold under saturated conditions at given initial water content. Air bubbles were removed using
small vibrations. The sample was allowed to dry naturally under laboratory conditions, and changes
in the sample water content and volume with time were monitored. The volume of soft material
that could not be removed from the mold in the early shrinkage stages was determined by
measuring the sample height with a caliper and determining the sample surface area from a photo of
the sample using ImageJ software (Girish and Vijayalakshmi 2004; Saleh-Mbemba et al. 2010). The
Page 12 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
13
ImageJ method consists in digitizing the picture, and converting the number of pixels into an area.
For samples sufficiently stiff for removal from the mold, the volume was determined by underwater
weighing. For this purpose, the sample was first coated with liquid wax of known density (0.84
g/cm3) to guarantee the sample integrity during immersion. Because the wax-coated sample was not
reusable, many molds filled with the same material were used to obtain the shrinkage curve e(w). At
the end of the natural shrinkage process, the last sample was oven dried (at a temperature of 105°C)
to obtain the solid mass and the sample volume was determined to obtain the final void ratio (at w =
0). Most shrinkage tests were performed with molds having a diameter and height of 4.84 cm and
1.52 cm, respectively (called hereafter molds M-I). The mold scale effect on e(w) was also
examined using molds with a diameter of 11.61 cm and heights of 1.52 cm (molds M-II) and 3.02
cm (molds M-III).
Water retention curves (WRC) were determined under drying conditions for the soil and SSM with
the highest sludge content β = 25%. For the soil alone, which was expected to undergo negligible
shrinkage, the WRC was obtained using a Tempe cell (ASTM D6836-02e2) with suctions up to 620
kPa. Given the expected shrinkage under increasing suction for the SSM sample with β = 25%, the
drying WRC test was performed using a 100 bar pressure membrane extractor (Model 1020, Soil
Moisture Equipment Corp.). Twelve saturated samples were prepared into circular flexible retaining
rings with a height of 1 cm and an internal diameter of 5.5 cm. The samples were placed directly in
contact with a cellulose membrane disc previously saturated in the pressure extractor vessel filled
with distilled and de-aired water. This membrane has an air entry value (AEV) > 104 kPa or 100
bars. With this AEV, the pressure in the cell may reach 100 bars, maintaining the cellulose
membrane saturated. Pressure was applied (using a nitrogen cylinder) and maintained constant over
time, and the water expelled from the samples was periodically monitored in a graduated burette.
Page 13 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
14
When no liquid drained out of the cell for 24 h, one sample was removed from the cell and a higher
pressure was subsequently applied to continue drying of the remaining samples. The removed
sample was used to determine the actual water content and volume of the sample. The volume,
required for estimating the actual porosity, n, and void ratio, e, was obtained using the water
immersion method. The change in porosity or void ratio with suction ψ, (i.e., n(ψ) or e(ψ)) is
required for a material that is deformable under suction. In this case, expressing the WRC in terms
of degree of saturation versus suction becomes more appropriate than in terms of gravimetric and
volumetric water contents versus suction (Mbonimpa et al. 2006). The function e(ψ) also represents
a shrinkage curve (Mbonimpa et al. 2006).
Most hydrogeotechnical tests were conducted with SSM prepared with soil and sludge at initial
water contents of 7.5% and 175%, respectively. For the shrinkage tests, initial water content was set
at 7.5% and 12.5% for soil and 175% and 200% for sludge. The initial water content of the mixture
wm was measured for a few SSM samples. Theoretically, wm can be estimated from the initial water
contents of the components as follows:
(1) )w1(w1
)w1(w)w1(ww
soisli
soislislisoim
−−
−−−−
+++
+++=
β
β
In equation (1), wi-sl and wi-so are the initial water contents of sludge and soil, respectively. Because
the wm values calculated with Eq. (1) were in good agreement with values measured on a few
mixtures (results not shown here), Eq. (1) was used for calculations for further mixtures. Table 2
presents the calculated wm values corresponding to the different SSM studied.
Page 14 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
15
4. RESULTS
4.1.Chemical and mineralogical properties
Table 3 presents the main elements of environmental interest according to the Directive 019
(Gouvernement du Québec, 2012) for soil and sludge alone. For all materials, both duplicate
samples had similar chemistry. The soil contains mainly aluminum (6%), iron (4%), and calcium
(2%). The sludge contains mainly calcium (18%), iron (9%), and sulphur (7%).
The mineralogical compositions determined by XRD for the soil and the crystalline materials in the
sludge are summarized in Figure 1. The soil contained mainly albite Na(AlSi3O8) (28%), quartz
SiO2 (22%), hornblende Ca2(Fe2+4Al)(Si7Al)O22(OH)2 (11%), microcline K(AlSi3O8) (11%),
anorthite Ca(Al2Si2O8) (10%), muscovite KAl2(Si3Al)O10(OH)2 (7%), phlogopite
K(Mg,Ti,Fe)3[(Si,Al)4O10](O,F)2 (4%), gypsum Ca(SO4)·2H2O (4%), and talc Mg3Si4O10(OH)2
(4%). The crystalline phases in the sludge contained 80 % of calcium sulphates, mainly ettringite
Ca6Al2(SO4)3(OH)12·26H2O (50%) and gypsum (30%). The remaining 20% consisted of silicates:
quartz (4%), muscovite (7%) and pyrophyllite Al2Si4O10(OH)2 (9%). Although the chemical
analysis indicated that the sludge contained 9.2% Fe, it should be mentioned that amorphous
minerals such as iron-oxyhydroxides and clay minerals may be difficult to analyze by XRD due to
their weak crystallinity. In particular, the clay minerals are difficult to quantify accurately by XRD
due to their compositional variation, variable degrees of structural order/disorder, and their
tendency towards preferred orientation in the whole sample.
According to the chemical and mineralogical properties, neither soil nor sludge contained sulphide
minerals that can oxidize and generate acid. All the sulfur was present in sludge as sulfates and the
Page 15 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
16
sludge possessed a strong neutralization potential (above the limit of the method of Sobek modified
by Lawrence and Scheske; 1997) (Demers et al. 2015a). As already mentioned, excess alkalinity
induced by lime in the sludge can sustain alkaline to neutral pH within the sludge ponds for
decades, even centuries (Zinck et al. 1997).
4.2. Specific gravity of solid grains
The average specific gravity of solid grains DR was 2.10 and 2.74 for the sludge and soil,
respectively. The low specific gravity of the sludge can be explained by the presence of hydrated
phrases such as ettringite (with DR≈ 1.8) and gypsum (with DR ≈ 2.3).
Based on the relative densities of the sludge DR-sl and soil DR-so, the specific gravity DR-m of a given
SSM could be estimated using the following equation:
(2) [ ]
soRsoislRsli
soislislRsoRmR
D)w1(D)w1(
)w1()w1(DDD
−−−−
−−−−−
+++
+++=
β
β
Table 4 compares typical measured and calculated specific gravity of solid grains for different
SSM. The initial water contents of the soil (wi-so) and sludge (wi-sl) used in these mixtures were
7.5% and 175%, respectively. Because the measured and calculated values were in good agreement,
DR-m for the other SSM was estimated using Eq. (2). The specific gravity of the different SSM
decreased with increasing β due to the low value DR-sl compared to DR-so.
Page 16 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
17
4.3. Grain-size distribution
Figure 2 shows the grain-size distributions for the sludge, soil, and SSM with β = 25%, considering
initial water contents of 7.5% for the soil and 175% for the sludge. The grain-size distribution
curves for the mixtures with β = 10, 15, and 20% lie between the curves for the soil and the SSM
with β = 25%, indicating that the grain-size distribution of the mixtures, even up to β = 25%,
remained very close to that of the soil alone. The effective diameter D10, corresponding to 10%
passing on the cumulative grain-size distribution curve, is 0.0026 mm for the sludge alone, 0.009
mm for soil alone and 0.007 mm for the SSM with β =25%. The coefficient of uniformity CU is 17.7
for the sludge alone, 66.7 for soil alone and 68.6 for the SSM with β =25%.
According to the Unified Soil Classification System (USCS) (ASTM D2487-06e1), the soil and
SSM are coarse-grained and can be classified as SM. The sludge is fine-grained and can be
classified as MH considering that the liquid and plastic limits of the sludge were approximately
88% and 62%, respectively. Since the water content of the sludge in the North Pond ranged in
general between 150 and 300%, the water contents are higher than the liquid limit, so the sludge
was extremely soft and in a viscous-liquid state.
4.4. Modified Proctor tests
Modified Proctor tests were performed on the soil and the SSM containing 10, 15, 20, and 25%
sludge. The initial water contents of the soil and sludge in the mixture were 7.5% and 175%,
respectively. Figure 3 shows typical modified Proctor curves (relationship of dry density ρd to water
content w) for the soil and for the different SSM. For the four tested mixtures, when β was
increased from 10% to 25%, maximum dry density ρdmax increased from 2025 kg/m3 to 2052 kg/m3.
Page 17 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
18
The optimum water content wopt ranged from 8.5% to 10.5%. Compared to the results obtained on
the soil alone (ρdmax ≈ 1988 kg/m3 and wopt ≈ 10%), addition of up to 25% sludge to the soil appears
to improve the compactibility because the fine sludge fills the voids between the soil particles. The
porosity (n), void ratio (e), and degree of saturation (Sr) reached at the optimum Proctor for the soil
were 0.27, 0.38, and 0.72, respectively. For the SSM mixtures, 0.23 ≤ n ≤ 0.25, 0.30 ≤ e ≤ 0.33, and
0.79 ≤ Sr ≤ 90 were obtained at the optimum Proctor. In general, n decreased and Sr increased with
increasing sludge content (for β ≤ 25%) at the optimum Proctor.
The change in dry density ρd and void ratio e with respect to water content w for each mixture
sample at a given degree of saturation can be calculated theoretically using the following equations:
(3)
r
mR
wmRd
S
Dw1
D
−
−
+
=ρ
ρ
(4) r
mR
S
Dwe −=
where ρw is water density. The curves for the maximum dry density and minimum void ratio at
saturation can be obtained by setting Sr = 1 in Eqs. (3) and (4). Because the values of DR-m are very
close for the SSM tested (see Table 4), a mean value of DR-m of 2.69 was used in eqs (3) and (4).
The curves for maximum dry density and minimum void ratio are presented in Figure 3. Figure 3
also shows the natural water contents of the different mixtures at sludge contents 10, 15, 20, and
25%, i.e.13.8, 16.8, 19.6 and 22.4 %, respectively (see Table 2). These water contents are higher
than wopt. If these SSMs could be compacted up to saturation at their natural water contents, the
expected maximum dry density and minimum void ratio would be about 1972 kg/m3 and 0.37 for
Page 18 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
19
the mixture with β = 10%, 1857 kg/m3 and 0.45 for β = 15%, 1761 kg/m3 and 0.53 for β = 20%, and
1667 kg/m3 and 0.59 for β = 25%, respectively.
4.5. Saturated hydraulic conductivity ksat
Saturated hydraulic conductivity (ksat), determined on SSM prepared with soil and sludge at water
content wi-so = 7.5% and wi-sl = 175%, is shown in Table 5. It was difficult to compare the values
due to the different sample void ratios, even for duplicates. The soil and SSM mixtures showed a
ksat in the range of 10-5 cm/s at void ratios e ranging from 0.28 to 0.53. The impact of adding sludge
by up to 25% wet weight on ksat seems limited. The studied SSM cannot be used alone as
conventional hydraulic barriers, because the saturated hydraulic conductivity values are higher than
the value required to meet regulatory requirements, i.e., 10-7 cm/s.
The effect of freeze–thaw cycles was examined in terms of swelling ∆hj for each cycle j (with ∆hj =
hj - h0) and change in saturated hydraulic conductivity. To compare swelling between the mixtures,
the swelling ratio ∆hj/h0 was used. Average swelling ratios are presented in Figure 4 for the SSM.
The initial sample height h0 varied from 10.5 to 11.8 cm. As a results, swelling increased generally
rapidly during the three first cycles, followed by a slight stabilization from cycles 3 to 5. Peak
swelling ratios increased with increasing sludge content (from 1.5% for β = 10% to 3.7% for β =
25%). Sample heights subsequently decreased starting from cycle 5, followed by stabilization from
cycle 8 to 9. This decrease could be due to a change in the morphologic structure of the samples, as
described by Othman and Benson (1993) and Wang et al. (2007) in their study on the freezing–
thawing effect in compacted clay membranes. In fact, during the freeze–thaw process, the soil
passes from an unstable state into a stable state. Thus repeated freeze–thaw cycles will change the
Page 19 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
20
soil structure towards a new dynamic stable balance state. This morphologic structure change can
be explained by the formation of ice lenses in the sample. During the first cycles, increasing of the
ice lens size leads to the increase in hydraulic conductivity. However, it appears that from the 7th
freeze-thaw cycle the hydraulic conductivity and swelling stabilized. The swelling ratio after
stabilization increased from approximately 1.1% for β = 10% to 2.5% for β = 25%. Although
freeze–thaw behavior was not studied for the soil alone, the addition of sludge to the mixtures
induced limited swelling under freeze–thaw effects (for the conditions tested).
The change in void ratio ∆ej/e0 associated with the swelling ratio ∆hj/h0 can be estimated with the
following equation:
(5) 0
0
0
j
0
j
e
)e1(
h
h
e
e +=∆∆
The maximum swelling ratio, ∆hj/h0 = 3.7%, was reached for β = 25% (with e0 = 0.53), ∆ej/e0 =
10.7%, or ej= 0.59.
The effect of freeze–thaw cycles was also examined in terms of changes in the saturated hydraulic
conductivity. The ratio of hydraulic conductivity at the end of cycle (i) ksat-j to hydraulic
conductivity at cycle (0) ksat-0 (i.e ksat-j/ksat-0, where ksat-0 is given in Table 5) is used for a comparison
purpose. Figure 5 shows typical results (average values) for the SSM with β = 25%. In this case, the
ratio ksat-j/ksat-0 is around 1.52. This freeze–thaw effect can be considered negligible (less than one
order of magnitude) for the studied SSM.
Page 20 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
21
4.6. Free shrinkage
Free shrinkage tests were performed on the soil alone at initial water contents of 7.5%, 12.5%, 20%
and 25%, and on the sludge alone at initial water contents of 100%, 150%, 175 % and 200%. Free
shrinkage tests were also performed on the SSM with sludge content at β = 25%. The mixtures were
prepared using soil samples with initial water contents wi-so of 7.5% and 12.5% and sludge with wi-sl
of 175% and 200%.
Figure 6 shows photographs of typical shapes of the soil, sludge, and SSM (with β = 25%) samples
at the end of shrinkage. Qualitatively, the sludge samples shrunk considerably compared to the
SSM mixtures, which in turn shrunk slightly more than the soil. Adding up to 25% sludge to the
soil affects shrinkage behavior.
Because the shrinkage curves (i.e., plots of the void ratio e with respect to water content w)
obtained for the triplicate samples were almost similar, average values are presented. As mentioned
previously, the mold scale effect on e(w) was investigated at the beginning of the testing program
using molds M-I (with a diameter of 4.82 cm and a height of 1.52 cm), molds M-II (with a diameter
of 11.61 cm and a height of 1.52 cm) and molds M-III (with a diameter of 11.61 cm and a height of
3.02 cm). Figure 7 shows typical results comparing the shrinkage curves obtained using the three
mold types. These results indicate that the shrinkage curves were almost independent from the mold
dimensions. Accordingly, the small molds (M-I) were mostly used. With these molds, shrinkage
was nearly completed after 4 days for the laboratory testing conditions: during the testing period,
the relative humidity varied between 40 % and 70 % and the temperature between 18°C and 25°C.
Page 21 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
22
Figure 8 shows the influence of the initial water contents on the shrinkage curves for sludge alone.
Also shown is the curve Sr = 1 representing the void ratio at a given water content w when the
material is saturated (e = w×DR for Sr = 1; see Eq. (4)). The sludge shrinkage curves show a wide
saturated shrinkage (or normal) phase where the soil remains saturated (shrinkage curves follow the
saturation line) because the volume decrease was equal to the volume of water lost. This phase
extended from the initial value to about w ≈ 50%. The residual phase with combined desaturation
and shrinkage occurred at water contents from 50% to 15%. The quasi-constant volume, or no-
shrinkage, phase occurred from w ≈ 15% to 0%. These shrinkage phases are described in the
literature (Marshall et al. 1996; Braudeau et al. 1999; Mbonimpa et al. 2006). For the sludge tested
at initial water contents wi-sl of 175 and 200% with ei = 3.7 and 4.2, the final void ratios ef are 0.54
and 0.65, respectively (see Figure 8). The shrinkage limit (ws) defined on the saturation line, as
typically used in geotechnical engineering, is approximately 30% for the sludge.
Figure 9 shows the influence of the initial water contents on the shrinkage curves for the soil alone.
The saturated (normal) and residual shrinkage phases are inexistent for the soil with wi-so ≈ 7.5%
(initial void ratio ei-so = 0.21 and final void ratio ef = 0.21) indicating that the soil dried without
volume change. The shrinkage curve for the soil with wi-so ≈ 12.5% was characterised by very short
saturated shrinkage (or normal) and residual phases (from 12.5 to 10%). However, the quasi-
constant volume or no-shrinkage phase is large: from w ≈ 10% to 0. With increasing initial water
contents, all shrinkage phases became larger. The final void ratio (ef) at full dryness (w = 0)
increased with the initial water content of the soil samples: ef = 0.21 for wi-so = 7.5% (ei = 0.21), ef =
0.32 for wi-so ≈ 12.5% (ei = 0.34), ef = 0.34 for wi-so ≈ 20% (ei =0.55) and ef = 0.42 for wi-so ≈ 25% (ei
= 0.69).
Page 22 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
23
Figure 10 shows the SSM shrinkage curves with β = 25% at different initial sludge water contents
(wi-sl ≈ 175 and 200%) with the soil at wi-so ≈ 7.5% and wi-so ≈ 12.5%. The specific gravity of the
solid grains DR-m in these mixtures varied from 2.665 to 2.673. The resulting line e(w) at Sr = 1
almost merged so that only one saturation line is presented in the figure for a mean value of DR-m.
Some data in the normal and residual shrinkage phases are missing due to factors beyond our
control. The final void ratios increased from around 0.20 to 0.36 when the initial water content of
the soil used in the mixtures was increased from 7.5% to 12.5%. Changing the water content of the
sludge used in the mixtures from 175% to 200% had the lowest impact on the shrinkage path and
final void ratio ef at full dryness (w = 0).
For comparison purposes, the volumetric shrinkage (VS) at full dryness (w = 0) was introduced and
calculated using the initial (Vi) and final (Vf) volumes or the initial (ei) and final (ef) void ratios of
each sample, as follows:
(6) i
fi
i
fi
e1
)ee(100
V
)VV(100(%)VS
+
−=
−=
The calculated VS are given in Table 6. The VS were 0, 2.3, 11.8 and 15.4% for the soil tested at
initial water contents wi-so of 7.5, 12.5, 20 and 25%, respectively. For the sludge material, the VS
were 67.2% and 68.3% at initial water contents wi-b of 175% and 200%, respectively. The VS
obtained for SSM with soil at wi-so ≈ 7.5% including 25% sludge at wi-sl of 175% and 200% were
24.3% and 30.45%, respectively. The VS obtained for SSM with soil at wi-so ≈ 12.5% including 25%
sludge at wi-sl of 175% and 200% were 25.7% and 32.5%, respectively. The volumetric shrinkage
Page 23 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
24
depends on the initial sludge and soil water content. Note that VS increased with the initial sludge
water content for a given initial soil water content. Consequently, adding up to 25% sludge to the
soil at 12.5% water content increased the VS from 2.3% to 25.7% or 32.5%, depending on the
sludge water content (175% or 200%, respectively). All these conclusions are valid for initially
saturated materials.
4.7. Water retention curves
Water retention curves (WRC) were determined for the soil and for the SSM mixtures with β = 25%
and initial water contents of 12.5 % and 175% for the soil and sludge, respectively. Two soil
samples were tested at initial porosities of 0.40 and 0.42 using Tempe Cells. Porosity was assumed
to remain constant during the drainage process, which is acceptable given the very low VS of the
soil. For the SSM, the 100 bar pressure membrane extractor was used due to expected volume
change with increasing suction. Mixing water was added up to a water content of 40.4% to make
sure that the SSM was homogeneous. The initial porosity (n) of the SSM was 0.52, or a void ratio e
= 1.08.
Figure 11 shows the shrinkage curve e(ψ) observed during WRC testing in the 100 bar membrane
extractor. The void ratio of the SSM decreased with increasing suction. Figure 12 presents the
WRC, expressed here in terms of degree of saturation Sr vs suction ψ, for the soil and SSM tested.
The air entry value (AEV) determined using the tangent method as described by Fredlund and Xing
(1994) is 35 kPa for the SSM (with initial water content of 40.4%) versus 20 kPa for soil alone.
Mixing sludge with soil increased the AEV (even for a mixture with a very high initial water
Page 24 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
25
content), which is beneficial in the context of a material to be used in the MRL of a CCBE to limit
oxygen migration towards underlying mine waste.
5. DISCUSSION
In this paper, various hydrogeotechnical properties of soil–sludge mixtures (SSM) were determined
in the laboratory to determine their potential for use as material for the moisture retention layer
(MRL) of CCBE. The mixture with β = 25% presented ksat values in the order of 10-5 cm/s and an
AEV value of 30 kPa, which would be appropriate as a component of a CCBE acting as oxygen
barrier, based on a comparison with materials used in efficient CCBE. For example, non-acid
generating tailings were used in the MRL of the CCBE at the LTA site, located in Abitibi-
Témiscamingue (Québec, Canada). The void ratio, ksat and AEV of this layer was 0.78 (or porosity
n = 0.44), 5×10-5 cm/s and 28 kPa, respectively (Bussière et al. 2003). In the case of the CCBE used
for the reclamation of the Lorraine site, also located in Abitibi-Témiscamingue (Québec, Canada),
laboratory tests on the silt used for the MRL revealed ksat between 2.15×10-6 cm/s and 1.06×10-5
cm/s for e ranging from 0.619 to 0.923, and AEV ranging from 29 kPa to 50 kPa for e from to 0.85
to 0.617 (Dagenais et al. 2012). The void ratio of the moisture retention layer as built was
approximately 0.61 (or a porosity n = 0.38).
Volumetric shrinkage (VS) at full dryness (w = 0) was calculated from free shrinkage tests to assess
the shrinkage susceptibility of the mixtures. In all cases, adding sludge to soil increased VS.
However, the calculated values of VS at complete dryness represent the worst case scenario. In
Page 25 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
26
practice, protective soil layers are required to minimize the impact of desiccation (shrinkage) due to
evaporation, frost action, erosion, animal burrowing, and/or plant rooting. If the initial water
content of the sludge could be lowered before mixing with the soil, the SSM could be easily
compacted to reach relatively low initial water content and void ratio, which may reduce the
susceptibility of the mixture to shrinkage.
One more interesting aspect is to estimate the volume of sludge that can be recovered from an
existing sludge pond for use in soil mixture to obtain the desired properties for mine site
rehabilitation. The total mass of wet sludge Msl that can be used in a MRL with a thickness H and
an area A compacted up to a void ratio em can be estimated with the following equation, where all
parameters have been previously defined (see Eqs. (1), (2) and (3)).
(7) ��� = �����
����� �������
���
If the bulk density the sludge in the sedimentation pond is known, the sludge volume reused Vsl can
then be calculated. Combining equations (1), (2) and (7) can allow choosing optimal water contents
of the sludge and soils to maximize the mass of sludge to be reused (assuming that these initial
water contents can be controlled).
The sludge mass Msl obtained from Eq. (7) can be used to calculate the borrow soil mass Mundist that
does not need to be disturbed or excavated and transported (possibly over longer distance than for
the sludge). Assuming that the void ratio of the MRL made exclusively with the soil is eso, one
obtains:
Page 26 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
27
(8) ������� = ��������� ���������− ����� ��
���������� ���
In the case of the soil and sludge studied, for a 1 ha (104 m2) MRL with a thickness H of 1 m, and
SSM placed at a void ratio em = 0.6, Msl, = 4243 tonnes for the SSM with β = 25% considering
initial water contents wi-so of 12.5 % for the soil and wi-sl of 175% for the sludge (i.e. wm = 27.6 %,
DR-m = 2.66). Assuming an in situ sludge density of 1.5 tonnes/m3 (which is a conservative value),
the sludge mass Msl, given above corresponds to a sludge volume Vsl = 2828 m3. Assuming another
void ratio em = 0.8, Msl = 3771 tonnes and Vsl = 2514 m3 per ha of MRL. In these two cases, the
volume Vsl is significant for a surface of 1 ha. Compared to a MRL made of soil only with eso = 0.8,
wi-so = 12.5 % and DR-so = 2.74, the use of a SSM with β = 25% considering initial water content wi-sl
= 175% for the sludge (i.e. wm = 27.6 %, DR-m = 2.66) will lead to a soil mass Mundist of 2039 tonnes
per hectare that does not need to be disturbed or excavated.
The investigation performed in this study was based on one silty soil and neutralisation sludge from
one pond at the Doyon mine site. It should be kept in mind that the testing procedures presented
here should be repeated for other types of soils and sludge involved in a mixture. Additional work is
required to evaluate the use of different granular soil types and sludge in SSM. Furthermore, an
assessment of the long-term effect of potentially damaging environmental forces (cracking due to
desiccation, swelling and structure changes due to freezing, precipitates remobilization, sludge
dissolution through acidic rain, etc.) should be performed before any use of SSM at a large scale.
Kinetics tests conducted at laboratory scale (with leaching columns tests) and at intermediate scale
(with in situ pilot cells) can be used to evaluate the chemical stability of soil-sludge mixture under
realistic conditions. Based on results presented by Zinck et al. (1997), aged rather than fresh sludge,
Page 27 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
28
sludge with a high neutralizing potential, low initial metal concentrations and stable amorphous
mass has lower risk for metal leaching. Such sludge with a relative chemical stability is preferred.
6. CONCLUSION
Large volumes of sludge are generated by the active treatment of acid mine drainage at different
mine sites in Canada and worldwide. The objective of the paper was to assess the
hydrogeotechnical potential use of silty soil–sludge mixtures as earthen cover material for mine site
rehabilitation using AMD neutralisation sludge from the Doyon mine site –IAMGOLD (Quebec,
Canada). Both the soil and sludge were non-acid generating. For that purpose, the impact of the
sludge content on the hydrogeotechnical properties of soil–sludge mixtures (SSM) was studied, i.e.
grain-size distribution, specific gravity, Proctor curve, freeze-thaw effects on the saturated
hydraulic conductivity, desiccation shrinkage and water retention curves. The results indicated that
adding up to 25% sludge to a silty soil can provide a mixture with appropriate saturated hydraulic
conductivity (about 10-5 cm/s) and water retention properties (AEV about 30 kPa) for the mixture to
be used in the moisture retention layer (MRL) of a cover with capillary barrier effects (CCBE),
based on comparison with existing efficient CCBE. The impact of sludge addition to the silty soil
on freeze–thaw behavior was relatively limited. Volumetric shrinkage at complete drying of the
mixtures (worst case scenario) increased with the sludge content, but shrinkage can be reduced by
covering the mixture with a layer of coarse material (the drainage and protection layer of the
CCBE) to control evaporation. It has been shown that the sludge mass that can be reused in the
MRL of CCBE could be significant, consequently reducing the mass (or volume) of borrow soil
required as well as the footprint of mine sites rehabilitation.
Page 28 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
29
7. ACKNOWLEDGEMENTS
The authors would like to acknowledge the contribution of their partners, the Ministère du
Développement Économique, Innovation et Exportation (MDEIE) du Québec, IAMGOLD,
Division Doyon – Mouska – Westwood, the MISA group, URSTM-UQAT, Golder & Associates,
and ACCORD Abitibi-Témiscamingue.
Page 29 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
30
8. REFERENCES
ASTM standard, D1557-07. 2007. Standard Test Methods for Laboratory Compaction
Characteristics of Soil Using Modified Effort. ASTM International, West Conshohocken, 2007,
PA. DOI: 10.1520/D1557-07.
ASTM Standard D2434 – 68. 2006. Standard Test Method for Permeability of Granular Soils.
ASTM International, West Conshohocken, 2006, PA DOI: 10.1520/D2434-68R06.
ASTM Standard D2487-06e1. 2006. Practice for Classification of Soils for Engineering Purposes
(Unified Soil Classification System). ASTM International, West Conshohocken, 2006, PA.
DOI: 10.1520/D2487-06E01
ASTM Standard D5550 – 06. 2007. Standard Test Method for Specific Gravity of Soil Solids by
Gas Pycnometer. ASTM International, West Conshohocken, 2007, PA. DOI: 10.1520/D5550-
06.
ASTM D5856 – 95. 2007. Standard Test Method for Measurement of Hydraulic Conductivity of
Porous Material Using a Rigid-Wall, Compaction-Mold Permeameter. ASTM International,
West Conshohocken, 2007, PA. DOI: 10.1520/D5856-95R07.
ASTM Standard D6836 – 02e2. 2008. Standard Test Methods for Determination of the Soil Water
Characteristic Curve for Desorption Using a Hanging Column, Pressure Extractor, Chilled
Mirror Hygrometer, and/or Centrifuge. ASTM International, West Conshohocken, 2008, PA.
DOI: 10.1520/D6836-02R08E02
Aubertin, M., and Chapuis, R.P. 1991. Considérations hydro-géotechniques pour l’entreposage de
résidus miniers dans le nord-ouest du Québec. In Proceedings of the Second International
Conference on the Abatement of Acidic Drainage, Sept. 16-18, 1991, Montreal, Canada
Montreal, MEND/CANMET, Vol.3: 1-22.
Aubertin, M., Chapuis, R.P., Aachib, M., Ricard, J.F., Tremblay, L., and Bussière, B. 1994. Cover
technology for acidic tailings: Hydrogeological properties of milling wastes used as capillary
Page 30 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
31
barrier. In proceedings of the 1st International Congress on Environmental Geotechnics, 11-15
July, 1994, Edmonton (AB), 427-432.
Benzaazoua, M., Fiset J-F., Bussière, B, Villeneuve, M., and Plante, B. 2006. Sludge recycling
within cemented paste backfill: study of the mechanical and leachability properties. Minerals
Engineering, 19: 420–432.
Bouda, M., Mbonimpa, M., Demers, I., Benzaazoua, M., and Gagnon, M. 2012. Hydrogeotechnical
characterization of AMD treatment sludge and sludge-based mixtures. In Proceedings of the
Geo-Manitoba. September 30 – October 3, 2012, Winnipeg. Manitoba.
Braudeau, E., Costantini. J. M., Bellier, G., and Colleuille, H. 1999. New device and method for
soil shrinkage curve measurement and characterization. Soil Science Society America Journal,
63: 525–535.
Brown, M., Barley, B., and Wood, H. 2002. Minewater treatment: technology, application and
policy. International Water Association Publishing.
Bussière, B., Aubertin, M., Chapuis, R.P. 2003. The behavior of inclined covers used as oxygen
barriers. Canadian Geotechnical Journal 40(3): 512-535.
Bussière, B., Benzaazoua, M., Aubertin, M., and Mbonimpa, M. 2004. A laboratory study of covers
made of low-sulphide tailings to prevent acid mine drainage. Environmental Geology, 45(5),
609-622.
Bussière, B., Aubertin, M., Mbonimpa, M., Molson, J., and Chapuis, R.P. 2007. Field experimental
cells to evaluate the hydrogeological behaviour of oxygen barriers made of silty materials.
Canadian Geotechnical Journal, 44(3), 1-22
Bussière, B., Maqsoud, A., Aubertin, M., Martschuk, J., McMullen, J., and Julien, M. 2006.
Performance of the oxygen limiting cover at the LTA site, Malartic, Quebec. CIM Bulletin,
1(6), 1-11.
Page 31 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
32
Dagenais, A.M., Mbonimpa, M., Bussière, B. and Aubertin, M. 2012. The modified oxygen
consumption test to evaluate gas flux through oxygen barrier covers. ASCE Geotechnical
Testing Journal, 35(1): 150-158.
Demers, I, Benzaazoua, M, Mbonimpa, M, Bouda, M, Bois, D, Gagnon, M. 2015a. Valorisation of
acid mine drainage treatment sludge as remediation component to control acid generation from
mine wastes, part 1: material characterization and laboratory kinetic testing. Minerals
Engineering, 76:109-116..
Demers, I., Bouda, M., Mbonimpa, M, Benzaazoua, M, Bois, D., Gagnon, M. 2015b). Valorisation
of acid mine drainage treatment sludge as remediation component to control acid generation
from mine wastes, part 2: field experimentation. Minerals Engineering,76: 117-125.
Demers, I., Bussière, B., Mbonimpa, M., and Benzaazoua, M. 2009. Oxygen diffusion and
consumption in low sulphide tailings covers. Canadian Geotechnical Journal, 46: 454–469.
Demers, I., Bussière, B., Benzaazoua, M., Mbonimpa, M., and Blier, A. 2008. Column test
investigation on the performance of monolayer covers made of desulphurized tailings to prevent
acid mine drainage. Mineral Engineering, 21: 317-329.
Demers, I., Bussière, B., Benzaazoua, M., Mbonimpa, M., and Blier, A. 2010. Preliminary
optimization of a single-layer cover made of desulfurized tailings: application to the Doyon
Mine tailings impoundment. Society for Mining, Metallurgy, and Exploration Annual
Transactions, 326: 21-33.
Fredlund, D.G., and Xing, A. 1994. Equations for the soil–water characteristic curve. Canadian
Geotechnical Journal, 31(4): 521–532.
Girish. V., and Vijayalakshmi. A. 2004. Affordable image analysis using NIH Image/ImageJ.
Indian Journal of Cancer, 41(1): 1–47.
Gouvernement du Québec 2012. Directive 019 sur l’industrie minière. Ministère du développement
durable, environement et parcs, Gouvernement du Québec, mars 2012, pp.105.
Page 32 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
33
Lawrence, R.W., and Scheske, M., 1997. A method to calculate the neutralization potential of
mining wastes. Environmental Geology, 32(2), pp. 100-106.
Marshall, T. J., Holmes, J. W., and Rose, C. W. 1996. Soil physics. 3rd Ed.. Cambridge University
Press. Cambridge, U.K.
Mbonimpa, M., Aubertin, M., Aachib, M, and Bussière, B. 2003. Diffusion and consumption of
oxygen in unsaturated cover materials. Canadian Geotechnical Journal, 40(5): 916-932.
Mbonimpa, M., Aubertin, M., Bussière, B., and Maqsoud, A. 2006. A predictive function for the
water retention curve of compressible soils. ASCE Journal of Geotechnical and
Geoenvironmental Engineering, 132(9): 1121–1132.
MEND 1996. Review of use of an elevated water table as a method to control and reduce acidic
drainage from tailings. Report 2.17.1, March 1996.
MEND 2013. Characterization and Prediction of Trace Metal Bearings Phases in ARD
Neutralization Sludge. Report 3.44.1, May 2013.
Nicholson, R.V., Gillham, R.W., Cherry, J.A., and Reardon, E.J. 1989. Reduction of acid
generation in mine tailings through the use of moisture-retaining layers as oxygen barriers.
Canadian Geotechnical Journal, 26: 1-8.
O’Kane, M., Wilson, G.W., and Barbour, S.L. 1998. Instrumentation and monitoring of an
engineered soil cover system for mine waste rock. Canadian Geotechnical Journal, 35: 828-846.
Othman, M.A., and Benson, C.H. 1993. Effect of freeze-thaw on the hydraulic conductivity and
morphology of compacted clay. Canadian Geotechnical Journal, 30: 236-246.
Ouangrawa, M., Aubertin, M., Molson, J., Bussière, B., and Zagury, G.J. 2010. Preventing acid
mine drainage with an elevated water table: Long-term column experiments and parameter
analysis. Water, Air and Soil Pollution, 213: 437–458.
Page 33 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
34
Pabst, T., Aubertin, M., Bussière, B., and Molson, J. 2014. Column tests to characterise the
hydrogeochemical response of pre-oxidised acid-generating tailings with a monolayer cover.
Water Air Soil Pollution: 225:1841.
Pabst, T., Aubertin, M., Bussière, B., and Molson, J. 2011. Column tests to assess water flow and
oxygen transport in monolayer covers for the reclamation of acid-generating tailings. In
Proceedings of the 2011 Mine Closure Conference, Lake Louise, AB, September 18-21, 2011.
Rietveld, H.M. 1993. The Rietveld Method. Oxford University Press.
Saleh-Mbemba, F., Aubertin, M., and Mbonimpa, M. 2010. An investigation on the shrinkage of
low plasticity tailings. In Proceedings of the 63rd Canadian Geotechnical Conference & 1st Joint
CGS/CNC-IPA Permafrost Specialty Conference. Sept 12–16. 2010, Calgary, Alberta, Canada,
pp.1411–1418.
Sjoberg Dobchuck, B., Craig, N., Wilson, G.W., and Aubertin, M. 2013. Evaluation of a single-
layer desulphurized tailings cover. Canadian Geotechnical Journal, 50: 777–792.
Smirnova, E., Demers, I., Bouda, M., Benzaazoua, M., Mbonimpa, M., and Bois, G. 2013. Results
of vegetation survey as a part of neutralizing lime sludge valorization assessment. In
Proceedings of the 2013 Northern Latitudes Mining Reclamation Workshop .and 38th Annual
Meeting of the Canadian Land Reclamation Association. Whitehorse, Yukon September 9 – 12,
2013, pp. 194-204.
Tsang, D.C.W., Olds, W.E., Weber, P.A., and Yip, A.C.K. 2013. Soil stabilisation using AMD
sludge, compost and lignite: TCLP leachability and continuous acid leaching. Chemosphere,
93(11): 2839-47.
Wang, D., Ma, W., Niu, Y., Chang, X., and Wen, Z. 2007. Effects of cyclic freezing and thawing
on mechanical properties of Qinghai–Tibet clay. Cold Regions Science and Technology, 48: 34–
43.
Page 34 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
35
Woyshner, M.R., and Yanful, E.K. 1995. Modelling and field measurements of water percolation
through an experimental soil cover on mine tailings. Canadian Geotechnical Journal, 32(4): 601-
609.
Yanful, E.K., Simms, P.H., and Payant, S.C. 1999. Soil covers for controlling acid generation in
mine tailing, a laboratory evaluation of the physics and geochemistry. Water, Air, & Soil
Pollution, 114: 347-375.
Young, RA 1995. The Rietveld method. Oxford University Press, 312 p.
Younger, P.L., Banwart, S.A., and Hedin, R.S. 2002. Mine water: hydrology, pollution,
remediation. Kluwer Academic Publishers, Dordrecht.
Zinck, J. 2005. Review of Disposal, Reprocessing and Reuse Options for Acidic. Drainage
Treatment Sludge. MEND Report 3.42.3, pp.68.
Zinck, J. 2006. Disposal. reprocessing and reuse options for acidic drainage treatment sludge. 7th
International Conference on Acid Rock Drainage (ICARD). R.I. Barnhisel (ed.). March 26–30.
2006. St. Louis, MO.
Zinck, J., and Griffith, W. 2012a. Review of acidic drainage treatment and sludge management
operations. CANMET-MMSL Report. 10-59 prepared for MEND Project 603–154.
Zinck, J., and Griffith, W. 2012b. Evaluation of sludge management options. In Proceedings of 9th
International Conference on Acid Rock Drainage. May 20–26, 2012. Ottawa, Ontario, Canada.
Proceedings on CD-ROM.
Zinck, J., and Griffith, W. 2013. Review of mine drainage treatment and sludge management
operations. MEND Report 3.43.1, pp. 111.
Zinck, J., Fiset, J.F., and Griffith, W. 2010. Stability of treatment sludge in different disposal
environment: A multi-year leaching Study. In Volkersdorfer. C. and Freund A (ed.). Mine waste
& innovative Thinking operations. IMWA 2010. Sydney, Nova Scotia. p. 527–531.
Page 35 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
36
Zinck, J.M., Wilson, L.J., Chen, T.T., Griffith, W., Mikhail, S., and Turcotte, S. 1997.
Characterization and stability of acid mine drainage sludge. MEND Report 3.42.2a - May, 1997.
Page 36 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
37
Table captions
Table 1. Proportions of wet components in the wet SSM
Table 2. Calculated water content wm (%) of SSM.
Table 3. Content (%) of the main elements of environmental interest in sludge and soil samples.
Table 4. Typical measured and calculated specific gravity of solids for the soil, sludge, and SSM
prepared with soil and sludge at 7.5% and 175% water content, respectively.
Table 5. Saturated hydraulic conductivity ksat (cm/s) for the soil and SSM, with void ratio e in
brackets (mixtures prepared with wi-so = 7.5% and wi-sl = 175%)
Table 6. Calculated volumetric shrinkage values
Page 37 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
38
Figure captions
Figure 1. Mineralogical semi-quantitative composition (wt.%) of a) the soil and b) the crystalline
materials in sludge (amorphous or poorly-crystalline phases were not quantified).
Figure 2. Grain-size distribution curves for the sludge, soil, and SSM with β =25% (mixture
prepared with wi-so = 7.5% and wi-sl = 175%)
Figure 3. Modified Proctor curves, minimum void ratio and maximum dry density curves at
saturation for SSM (mixtures prepared with wi-so = 7.5% and wi-sl = 175%)
Figure 4. Swelling ratio curves for SSM mixtures versus freeze–thaw cycles (mixtures prepared
with wi-so = 7.5% and wi-sl = 175%).
Figure 5. Variation of the ksat /ksat-0 ratio versus the freeze–thaw cycle (SSM with β =25% prepared
with wi-so = 7.5% and wi-sl = 175%).
Figure 6. Typical photographs showing views of the samples (in triplicate) at the end of free
shrinkage in molds M-I : a) soil samples (wi-so at 12.5%), b) sludge samples (wi-sl at 175%) and
c) SSM samples with β=25% ( wi-sl = 200% and wi-so = 12.5%; wi-m = 28.6; see Table 2)
Figure 7. Typical results for the effect of mold scale on the shrinkage curves of SSM (initial water
contents wi-sl = 200% and wi-so = 12.5%; with wi-m = 28.6%; see Table 2).
Figure 8. Shrinkage curves for sludge at initial water contents of 175% and 200%.
Figure 9. Soil shrinkage curves (for initial water contents of 12.5 %, 20%, and 25%).
Figure 10. SSM shrinkage curves with β = 25% at different initial sludge water contents
Figure 11. Shrinkage curves e(ψ) observed during the drying process in the 100 bar membrane
extractor for SSM with β = 25% and wi-so = 12.5% and wi-sl = 175%.
Figure 12. Water retention curves of the soil and SSM with β = 25%, initial water contents of 12.5
% for the soil and 175% for the sludge
Page 38 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
1
Figure 1. Mineralogical semi-quantitative composition (wt.%) of a) the soil and b) the crystalline
materials in sludge (amorphous or poorly-crystalline phases were not quantified).
Page 39 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
2
Figure 2. Grain-size distribution curves for the sludge, soil, and SSM with β =25% (mixture
prepared with wi-so = 7.5% and wi-sl = 175%)
Page 40 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
3
Figure 3. Modified Proctor curves, minimum void ratio and maximum dry density curves at
saturation for SSM (mixtures prepared with wi-so = 7.5% and wi-sl = 175%)
Page 41 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
4
Figure 4. Swelling ratio curves for SSM mixtures versus freeze–thaw cycles (mixtures prepared
with wi-so = 7.5% and wi-sl = 175%).
Figure 5. Variation of the ksat-j /ksat-0 ratio versus the freeze–thaw cycles (SSM with β =25%
prepared with wi-so = 7.5% and wi-sl = 175%).
Page 42 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
5
a)
b)
c)
Figure 6. Typical photographs showing views of the samples (in triplicate) at the end of free
shrinkage in molds M-I : a) soil samples (wi-so at 12.5%), b) sludge samples (wi-sl at 175%) and c)
SSM samples with β=25% ( wi-sl = 200% and wi-so = 12.5%; wi-m = 28.6; see Table 2)
Page 43 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
6
Figure 7. Typical results for the effect of mold scale on the shrinkage curves of SSM (initial water
contents wi-sl = 200% and wi-so = 12.5%; with wi-m = 28.6%; see Table 2).
Figure 8. Shrinkage curves for sludge at initial water contents of 175% and 200%.
Page 44 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
7
Figure 9. Soil shrinkage curves (for initial water contents of 12.5 %, 20%, and 25%).
Figure 10. SSM shrinkage curves with β = 25% at different initial sludge water contents
Page 45 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
8
Figure 11. Shrinkage curves e(ψ) observed during the drying process in the 100 bar membrane
extractor for SSM with β = 25% and wi-so = 12.5% and wi-sl = 175%.
Figure 12. Water retention curves of the soil and SSM with β = 25%, initial water contents of 12.5
% for the soil and 175% for the sludge
Page 46 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
1
Table 1. Proportions of wet components in the wet SSM
β
(%)
Wet sludge
(%)
Wet soil
(%)
0 (soil) 0 100
10 9.1 90.9
15 13.0 87.0
20 16.7 83.3
25 20.0 80.0
∞ (sludge) 100 0
Table 2. Calculated water content wm (%) of SSM.
wi-so (%) 7.5 12.5
wi-sl (%) 175 200 175 200
β = 10% 13.8 14.2 18.9 19.3
β = 15% 16.8 17.3 21.9 22.5
β = 20% 19.6 20.4 24.8 25.6
β = 25% 22.4 23.3 27.6 28.6
Table 3. Content (%) of the main elements of environmental interest in sludge and soil samples.
Element Sludge Soil
Al 2.1 5.940
Ca 18.6 2.31
Mg 1.590 1.155
Fe 9.210 4.135
S 7.370 0.341
Ti 0.014 0.300
As <0.003 <0.003
Cu 0.03 0.030
Ni 0.006 0.018
Pb 0.001 0.000
Zn 0.010 0.004
Page 47 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
2
Table 4. Typical measured and calculated specific gravity DR of solids for the soil, sludge, and SSM
prepared with soil and sludge at 7.5% and 175% water content, respectively.
Material Measured
(-)
Calculated
(-)
β = 0% (Soil) 2.74 -
β = 10 % 2.71 2.71
β = 15 % 2.70 2.69
β = 20 % 2.69 2.68
β = 25 % 2.66 2.67
β = ∞ ( Sludge) 2.10 -
Table 5. Saturated hydraulic conductivity ksat (cm/s) for the soil and SSM, with void ratio e in
brackets (mixtures prepared with wi-so = 7.5% and wi-sl = 175%)
Sample Average w before
saturation (%)
ksat (cm/s)
Sample 1 Sample 2 Average
Soil - β = 0% 9.0 8.4×10-5
(e = 0.46) 3.1×10-5
(e = 0.42) 5.7×10-5
(e = 0.44)
SSM - β = 10% 10.5 4.9×10-5
(e = 0.37) 6.3×10-5
(e = 0.41) 5.6×10-5
(e = 0.39)
SSM - β = 15% 11.9 2.5×10-5
(e = 0.34) 6.2×10-5
(e = 0.38) 4.4×10-5
(e = 0.36)
SSM - β = 20% 10.8 2.3×10-5
(e = 0.32) 1.4×10-5
(e = 0.25) 1.8×10-5
(e = 0.29)
SSM - β = 25% 18.6 3.4×10-5
(e = 0.52) 3.2×10-5
(e = 0.53) 3.3×10-5
(e = 0.52)
Page 48 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
3
Table 6. Calculated volumetric shrinkage values
Material VS (%)
Soil (wi-so ≈ 7.5%) 0
Soil (wi-so ≈ 12.5%) 2.3
Soil (wi-so ≈ 20%) 11.8
Soil (wi-so ≈ 25%) 15.4
Sludge (wi-sl = 175%) 67.2
Sludge (wi-sl =200%) 68.3
SSM (β = 25%)
wi-sl ≈ 175% and wi-so = 7.5% 24.3
wi-sl ≈ 175% and wi-so= 12.5% 25.7
wi-sl ≈ 200% and wi-so= 7.5% 30.4
wi-sl ≈ 200% and wi-so= 12.5% 32.5
Page 49 of 49
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal