Landfill Disposal of Alum Water Treatment Residues:

Some Pertinent Geoengineering Properties

BRENDAN C. O’KELLY*
1
Department of Civil, Structural and Environmental Engineering, Museum Building, Trinity College Dublin, Dublin 2, Ireland

ABSTRACT: This paper presents the geoengineering properties of the alum water treat-
ment residues derived from the production of potable water at three different municipal
works, including the effects of catchment geology; chemical additives; and thixotropic
hardening phenomena, with reducing water content (increasing solids content) from the
viscous slurry state to the semisolid state. The geoengineering behavior was akin to that
of high-plasticity organic clays, with low values of bulk and dry density and high com-
pressibility, although the consolidation rate was low (hydraulic conductivity of the order
of 10-4–10-6 m/day). The data presented in this paper can be used to determine the level
of dewatering necessary for more efficient landfill disposal, including the anticipated
amounts and rates of settlement and the factor of safety against geotechnical instability
of the residue slopes for the short- and long-term conditions. It is recommended that
sludge and residues should be dewatered to achieve minimum shear strengths of 20
and 50 kPa for geotechnical stability at municipal landfills and dedicated monofills, re-
spectively.

INTRODUCTION Increasing quantities of WTRs are being produced

worldwide annually due to the increasing demand for
ATER treatment residues (WTRs) are the gelati- potable water and more stringent regulations: for exam-
W nous slurry by-products from the treatment pro-
cesses used in the production of potable water at munic-
ple, the European Union Drinking Water Directive [3].
Currently, the principal disposal options are: (1) stor-
ipal works. The WTRs comprise sand, silt and clay age, often over an indefinite period, in sludge lagoons;
particles, colloidal organic matter, and chemicals (co- (2) dewatering by mechanical and/or thermal means,
agulants, polyelectrolytes and conditioners) that have followed by landfilling of the residue cake, either at
been added to the source water during the treatment dedicated monofills or co-disposal at municipal land-
processes. Chemical coagulation using ferric chloride, fills; (3) incineration. However, more stringent regula-
or more generally using aluminum sulfate (alum), tion (for example by the US EPA [4] and the European
causes the colloidal particles (i.e. less than 1 mm in size) Union [5,6]) has placed greater restrictions on these
that are suspended in the water entering the treatment disposal options, principally to minimize environmen-
plant to aggregate into flocs that settle out more readily tal impacts. Landfill disposal of large volumes of soft
under gravity. The residue is characterized as alum or residues may also lead to geotechnical problems, in-
iron WTR depending on the coagulant type used. cluding slope instability and excessive differential set-
Polyelectrolytes are synthetic long-chained organic tlement that may damage the landfill capping layer.
molecules that act as binding agents, thereby increasing Hence, the slurry residue must be adequately dewatered
the inherent shear strength of the newly-formed flocs, at the municipal works by mechanical and/or thermal
and hence the viscosity of the slurry residue [1,2]. Con- means, or by allowing the slurry to dry naturally in dry-
ditioners including sulfuric acid, bentonite, calcium or ing beds.
sodium hydroxide, and sodium silicate may also be In this paper, the amount of water within the pore
added, depending on the nature of the source water, in space of the residue materials was quantified in terms of
order to improve the polyelectrolyte performance. the water content (w), defined in the geotechnical litera-
ture as the mass of the pore water to the mass of the dry
*Author to whom correspondence should be addressed.
solids, expressed as a percentage. The water content is
E-mail: bokelly@tced.ie one of the most commonly determined parameters in

Journal of Residuals Science & Technology, Vol. 7, No. 2—April 2010 95

1544-8053/10/02 095-19
© 2010 DEStech Publications, Inc.
96 B. O’KELLY

characterizing geoengineering behavior and is gener-

ally determined using the oven drying method, with the
dry solids mass corresponding to the residual mass after
oven drying the test specimen at 105 ± 5°C for a period
of 24 h. Oven drying removes the free water in the pore
space between the flocs; the internal water in the mi-
cro-channels within the flocs; and the pores within the
constituent solid particles themselves [7]. Bound water
in the form of water of hydration and adsorbed water
held on the surface of the solid particles by electrical at-
traction are considered to be part of the solids and can
not be removed by oven drying. The solids contents
(SC), defined in the water-treatment literature as the Figure 1. Shear strength against water content data for WTR and
sewage sludge materials.
mass of the dry solids fraction expressed as a percent-
age of the bulk wet mass, can be related to the water
content by and levels of chemical treatment, and also biological
100 treatment in the case of the sewage sludge, had been ap-
SC = (1) plied in order to separate the residue by-products at the
Ê w ˆ municipal works. In general, alum WTR tends to have
1+ Á ˜
Ë 100 ¯ slightly higher shear strength than iron WTR at a given
water content, and both alum and iron WTRs tend to
Geoengineering Properties of Alum WTRs from have higher shear strengths than sewage sludge.
the Literature Wet alum WTR is thixotropic, with Wang et al. [11]
reporting strength gains (ratio of the shear strength
Table 1 lists some typical geoengineering data re- measured after a specified curing period to the
ported for alum WTRs in the literature [8–16]. Alum remolded shear strength, under constant external condi-
WTRs have very high Atterberg liquid and plastic lim- tions and specimen composition) of 5.7–8.0 achieved
its; a relatively low specific gravity of solids; high total after a ten-week curing period and for SC = 10–20%.
volatile solids (TVS); and are classified as high-plastic- The effective angle of shearing resistance (f¢) values of
ity organic clays. The wet residue, direct from the treat- 28–44° reported for alum WTRs are towards the upper
ment works, has a low bulk density of 1.0–1.2 end of the range normally associated with high-plastic-
tonne/m3. ity organic clays, although consistent with f¢ = 32–37°
Figure 1 shows data for shear strength against water reported for municipal sewage sludge (TVS = 70–50%)
content for some alum and iron WTRs, and also munici- [17,18].
pal sewage sludge. There is a significant variation in the Alum WTR has a very low hydraulic conductivity,
shear strength value of these materials at a given water which reduces in value from typically 10-4 to 10-6
content, which is expected, due to the natural variability m/day, with increasing effective stress from 2 to 540
of the source waters and hence the mineralogy and or- kPa [12]. Note that the effective stress (s¢), which acts
ganic content of the suspended solids. Different types across the contacts between the solid particles, is de-
fined as the difference between the total/applied stress
Table 1. Typical Geoengineering Properties of (s) and the excess pore water pressure in the saturated
Alum WTRs. pore voids.
Parameter Value
Liquid limit, % 100–550
Volume Change Theory
Plastic limit, % 80–250
Specific gravity of solids 1.8–2.2 Residues and sludge are soil-like materials and, as
Total volatile soilds, % 10–60
Bulk density, tonne/m3 1.0–1.2
such, their settlement behavior in lagoons, monofills or
Dry density, tonne/m3 0.12–0.36 dedicated deposition areas at municipal landfills can be
Effective cohesion, kPa 0 assessed using soil mechanics theory [18], and a more
Effective angle of shearing resistance, degree 28–44
comprehensive explanation of the following can be ob-
 Landfill Disposal of Alum Water Treatment Residues: Some Pertinent Geoengineering Properties 97

tained in undergraduate soil mechanics textbooks, in- Secondary compression (creep) settlement, which is
cluding Craig [19]. due to the gradual rearrangement and compression of
The settlement due to a change in the state of effec- the solids (residue flocs) into a more stable configura-
tive stress comprises two components: primary consoli- tion under the increased effective stress, is a time de-
dation (DHc) and secondary compression (DHs). Pri- pendent process that can be expressed by
mary consolidation can be described by classical 1–D
consolidation theory, whereby the change in the residue t1
DH s = H oC sec log (4)
volume corresponds to the change in the volume of the t2
pore water, assuming fully saturated conditions. The
void ratio (e) is defined as the volume of the pore voids where
to the volume of the solids and, referring to Figure 2, the Csec = coefficient of secondary compression
change in the void ratio (e) can be predicted by t1 = time period to achieve substantial completion
s vo
¢ + Ds ¢v of the primary consolidation phase
De = C c log (2) t2 = time period that extends into the secondary
s vo
¢
compression phase (t2 > t1).
where
Cc = compression index (gradient of the void ratio Objectives of this Study
against logarithm of effective stress curve)
s ¢vo = initial vertical effective stress The objectives of this paper are to study the
s ¢v = increase in vertical effective stress (i.e. applied geoengineering and hydraulic properties of alum
stress) WTRs derived from different catchments, and in partic-
ular to study the effects of several influencing factors:
Wet alum WTR is highly compressible. For example, catchment geology; chemical additives; and
very high compression index values of Cc = 5.3–6.7 thixotropic hardening phenomena. Geotechnical rec-
were reported for alum WTR that had been dewatered ommendations are made regarding the safe and effi-
using a centrifuge and sand-drying method [11]. The cient disposal of these residues in engineered landfills.
primary consolidation settlement, which is due to the In addition, pertinent data are presented that can be used
dissipation of the excess pore water pressure, can be es- to calculate the factor of safety against instability of the
timated as landfill slopes and the time-dependent settlement re-
De Cc s ¢ + s ¢v sponse.
DH c = H o = Ho log vo (3)
1+ eo 1+ eo s ¢vo
TEST MATERIALS
where
Alum WTRs were sourced from three of the larger
eo = initial void ratio
municipal water treatment plants in Ireland: the
Ho = initial thickness of the saturated residue layer
Ballymore Eustace and Leixlip works in County
Kildare, and the Clareville works in County Limerick,
which produce about 91, 73 and 62 Mm3 of potable wa-
ter per annum respectively, and taken together account
for almost 30% of the potable water produced in Ire-
land. The raw water entering these treatment plants is
sourced from three different catchments, thereby pro-
viding a good overall representation of the alum WTRs
produced in the country. The residues are coagulated at
these municipal works using Chemifloc 4140® alum,
supplied in its hydrated form [i.e. Al2(SO4)3◊14H2O],
and Magnafloc LT25® polyelectrolyte, which are man-
ufactured by Allied Colloids and Ciba Zietag. The com-
Figure 2. Theoretical consolidation plot. Note: eo, initial void ratio;
¢ , initial vertical effective stress; De, change in void ratio due to a
s vo bined dosages (given as mg/l source water) that had
change in the vertical effective stress, Ds v¢ ; Cc, compression index. been added during the different treatment processes are
98 B. O’KELLY

Table 2. Chemicals Added During the lowland area of about 15700 km2. The Shannon is a low
Treatment Processes. gradient river, about 260 km in length with a mean an-
Hydrated Dry nual discharge of 186 m3/s, which runs for much of its
Alum Polyelectrolyte H2SO4 course through karstified limestone overlain by raised
Mean % Dry and riverine bogs, many of which has been harvested
Residue mg/l mg/l Residue Mass mg/l for peat over the years. Much of the turbidity and color
WTR 1 40–65 0.6–1.5 3.5 0 in the water is likely to be attributed to the organic mate-
WTR 2a 60–100 0.8–2.0 4.8 0.5 rial associated with the peat.
WTR 2b 0 0 0 0 The raw water is delivered upstream of the
Control 40–60 0.2–1.0 1.6 0
Ardnacrusha dam through an open channel to the Clare-
ville treatment works. After the initial screening pro-
listed in Table 2, with the range of values reflecting the cess, the turbid water passes through 8.0-m deep pri-
necessary adjustments to the dosage appropriate to the mary sedimentation (holding) tanks, before the
level of turbidity of the incoming source water. These coagulation process, where a high alum dosage of
dosages were broadly similar and are at the higher end 60–100 mg/l raw water was added owing to its high tur-
of the ranges normally used in practice, since the source bidity. Sulfuric acid was also added to adjust the pH in
waters were all medium high in turbidity. Also listed in order to improve the polyelectrolyte performance.
Table 2 are the polyelectrolyte concentrations, ex- At the end of the treatment process, the slurry residue
pressed as a percentage dry solids mass of the residue was consolidated to about 700% water content under an
by-product at the different works. These mean concen- applied stress of 800–1000 kPa using a belt-press de-
tration values were calculated on the basis of the annual vice, and the pressed material was then allowed to dry
production of potable water, the mass of dry solids resi- naturally in drying beds, with about 900 tonnes of the
due produced per annum, and the mean polyelectrolyte wet alum residue (SC = 15%) produced at the treatment
dosages as mg/l source water at the different works, as- works in 2006. Residue samples were obtained from the
suming that all of the polyelectrolyte additives were re- drying beds for geotechnical laboratory testing in Octo-
tained in the residues. ber 2006.

Alum WTR 1 Sample Residue WTR 2b Sample

The alum residue from the Ballymore Eustace works Samples of the turbid water were withdrawn from the
(WTR 1) was derived from the treatment of medium mid-height of the primary sedimentation tanks at the
color, medium-turbidity water sourced from the Dublin Clareville works using 22.5 liter drums in February
and Wicklow mountains (upland catchment of peat 2008. These drum samples were allowed to evaporate
over granite bedrock) and which had been stored in and settle out over a period of about four weeks at ambi-
Poulaphouca reservoir, County Wicklow, prior to its ent laboratory temperature of 20°C. The thickened
treatment. The slurry residue was dewatered using a re- slurry residue formed (WTR 2b) was poured from the
cessed-plate filter press device under an applied stress drums into shallow trays and the water content was al-
of 1500 kPa, with about 1200 tonnes of the wet alum lowed to reduce further by fan-assisted evaporation,
residue (SC = 23%) produced at the treatment works in again at ambient temperature, before the saturated
2007. Samples of the pressed residue cake were ob- slurry residue was mechanically dewatered using a
tained from the skip containers at the end of the consolidometer press. Although the solids in the drum
dewatering process at the treatment works in November samples were more than 40% volatile, the fact that the
2005. catchment source mainly comprised natural saturated
peat meant that the bulk of these solids were already in a
Alum WTR 2a Sample near stable condition. Hence, any biological activity
that may have occurred during the course of settling and
The alum residue from the Clareville works (WTR dewatering the drum samples in the laboratory would
2a) was derived from the treatment of high color, turbid not have significantly altered the characteristics of the
water (typically 20–25 NTU) sourced from the lower solids in residue WTR 2b.
Shannon, Ireland’s longest river, draining the central One of the objectives of this study was to determine
 Landfill Disposal of Alum Water Treatment Residues: Some Pertinent Geoengineering Properties 99

the effect of the chemical additives on the the organic material was already in a near stable condi-
geoengineering and hydraulic properties, which was tion and the polyelectrolyte additives only become un-
achieved by comparing the behaviors of alum WTR 2a stable above about 150°C. The aluminum hydroxide
and this non-chemically treated residue WTR 2b. These precipitate degrades at higher temperatures of
materials were identical in composition apart from the 180–600°C, with further calcination occurring above
chemical additives in WTR 2a. 1000°C.
The Atterberg liquid and plastic limits, defined as the
Alum WTR 3 Sample water content values at the transitions from the liquid
state to the plastic state and from the plastic state to the
The alum residue from the Leixlip works (WTR 3) semisolid state respectively, were determined using the
was derived from the treatment of medium color, me- fall-cone penetrometer and Casagrande thread-rolling
dium-turbidity water sourced from the river Liffey (up- methods [20]. The plasticity index was calculated as the
land catchment of limestone bedrock). The residue was numeric difference between the liquid limit and plastic
dewatered using a recessed-plate filter press device un- limit values. The residues had very high liquid limit
der an applied stress of 1500 kPa, with about 1100 (430–550%) and plasticity index (210–290) values.
tonnes of the wet alum residue (SC = 25%) produced at The adhesion limit, determined as the lowest water con-
the treatment works in 2007. Samples of the pressed tent at which the solids adhered to a clean dry spatula,
residue cake were obtained from the skip containers at was also very high; in excess of 240% water content.
the end of the dewatering process at the treatment The bulk density (r) and the dry density (rd) were cal-
works in November 2006. culated as the wet mass and the dry solids mass per unit
volume respectively [20], with the dry density given by
PHYSIOCHEMICAL PROPERTIES
100r
rd = (5)
Table 3 lists the physiochemical properties which 100 + w
were determined using standard geotechnical-labora- where
tory tests [20,21] carried out on fresh residue specimens
obtained directly from the treatment works. w = water content (as %)
The water content values were determined using the The dewatered residue samples from the treatment
oven drying method and are an accurate measure of the works had low bulk and dry density values (1.06–1.10
combined amounts of free and internal water within the and 0.18–0.26 tonne/m3, respectively) and a wide range
pore space of the residue materials [7]. The degree of of water contents, and hence void ratio values, due to
oxidation of the volatile solids was not significant at the inherent differences in their consolidation properties
oven drying temperature of 105 ± 5°C since the bulk of and also differences in the drainage conditions and con-
fining pressures that had been applied by the re-
Table 3. Properties of the Residue Materials Direct cessed-plate filter press and belt press devices. WTR 2a
from the Treatment Works. was of slurry consistency (water content greater than
Parameter WTR 1 WTR 2a WTR 2b WTR 3 the liquid limit value, where the liquid limit condition
Water content, % 340 570 – 300
corresponds to undrained shear strengths of about 1.7
Solids content, % 23 15 – 25 kPa). WTR 1 and WTR 3 were soft to firm in consis-
Bulk density, tonne/m3 1.08 1.06 – 1.10 tency due to greater level of dewatering that had been
Dry density, tonne/m3 0.25 0.18 – 0.26
Void ratio 6.3 11.3 – 5.7
achieved under the higher confining pressure of 1500
Liquid limit, % 490 550 550 430 kPa applied by the recessed-plate filter press device.
Plastic limit, % 240 260 280 220 Nevertheless, their void ratio values of 5.7 and 6.3 were
Plasticity index 250 290 270 210
still very high.
Total volatile soilds, % 57 45 41 46
Specific gravity of solids 1.86 1.99 1.83 1.90 X-ray diffraction analysis indicated that the crystal-
Adhesion limit, % 240 365 355 345 line fraction of the residues comprised quartz and
Linear shrinkage, % 47 45 38 48 manganoan calcite; both common bedrock minerals
Free swell, % 35 10 10 40
pH 8.6 7.1 7.9 7.2 that are present as colloidal particles in the source wa-
WTR 1, WTR 2a and WTR 3 are alum water treatment residues; WTR 2b is a ters. The chemical additives in the alum WTRs did not
non-chemically treated residue. feature in the analysis since the alum was present in its
100 B. O’KELLY

aluminum hydroxide form (disordered and without any of 38–48% were very high, indicating substantial re-
definable crystalline structure), and the ductions in volume would occur on drying the wet resi-
polyelectrolytes are organic molecules. Induc- due materials. The free swell values, defined as the
tively-coupled plasma analysis was also carried out on maximum volumetric expansion that would occur on
six specimens of alum WTR 2a over the water content full re-saturation of the powered oven-dried residue
range of 120% to 1100%, which indicated that this resi- materials, of 10–40% were also high, although consis-
due comprised 24–28% aluminum by dry mass, in line tent with the high plasticity index values. The pH of the
with the range of 29.7 ± 13.3% reported for alum WTRs residues was determined using an electrometric method
[22]. [21] and, as expected, the pH was slightly alkaline due
The total volatile solids (TVS) values of 41–57% to the chemical conditioning of the water during the
were determined by heating dry powered residue speci- treatment processes.
mens in a muffle furnace at a temperature of 440°C The water content of the wet residue samples was re-
[21], which oxidizes both the polyelectrolytes and the duced over time in the geotechnical laboratory by al-
organic solids. The polyelectrolytes, which are unsta- lowing thin layers of the wet material to dry naturally in
ble above 150°C, typically comprised 3.3% dry solids trays at ambient temperature of 20°C.
mass (range 1.6–4.8% dry solids mass) for alum WTRs
1, 2a and 3 (Table 2). Nevertheless, the TVS value is a COMPACTION PROPERTIES
good reflection of the gravimetric organic content since
the crystalline fraction remains stable at the ignition Ordinary Proctor compaction tests were conducted in
temperature of 440°C. For example, calcination to form the one-liter compaction mold [25] on the wet alum
lime occurs at a much higher temperature of about WTR samples that had been allowed to dry naturally
850°C. The 4% higher TVS value measured for alum over different periods, with regular remolding, thereby
WTR 2a, compared to the non-chemically treated resi- obtaining a range of uniform materials with different
due WTR 2b, is most likely due to the 4.8% water content values. The samples were regularly
polyelectrolyte by dry solids mass basis that had been mixed during the drying process, and any clumps were
added during the treatment processes at the municipal disaggregated to pass the 20.0 mm sieve size prior to the
works, and which was ultimately retained in alum WTR compaction tests. Sets of three cylindrical sub-speci-
2a. Some natural variations could also be expected in mens (38 mm in diameter and 76 mm high) were also
the composition of the suspended solids in the source prepared from the wet ordinary Proctor compacted
water since WTR 2a and WTR 2b had been sampled in specimens. The wet mass, bulk volume and water con-
October and February, respectively. Organic matter in tent values of these sub-specimens were measured after
soil is generally responsible for high plasticity; low spe- they had been allowed to slowly air-dry further, and
cific gravity of solids; high shrinkage; high compress- shrink without cracking, at ambient laboratory temper-
ibility; and low hydraulic conductivity [23]. Note that ature. The bulk density was calculated as the wet mass
the organic fractions of WTR and sewage sludge mate- per unit volume, from which the dry density was calcu-
rials mainly comprise colloidal-size particles (i.e. non lated using Equation (5).
fibrous), although the organic fraction of the WTRs is Figure 3(a) shows the density values achieved by or-
usually in a near stable condition whereas sewage dinary Proctor compaction alone. Figure 3(b) shows the
sludge is bioactive, with the organics usually at moder- density values achieved by allowing the sub-specimens
ate to strong levels of biodegradation [17,18,24]. The that had been prepared from the wet ordinary Proctor
specific gravity of solids values of 1.83–1.99, which compacted specimens to reduce in water content by nat-
were measured using the small picnometer method ural air drying. Note that the zero air voids curve in-
[20], were relatively low, and consistent with the high cluded in Figure 3 corresponds to the fully saturated
organic contents. Note that the specific gravity of solids condition and therefore represents an upper bound for
for mineral soils is typically in the range 2.5–2.7. the dry density values. The zero air voids curve was de-
Linear shrinkage was calculated as the percentage re- termined from the measured specific gravity of solids
duction in the length of a bar of the wet residue material, values [19].
which had been prepared at the liquid limit condition The bulk and dry density values of 0.96–1.13 and
using a mold 130 mm in length, and oven dried at a tem- 0.21–0.36 tonne/m3 respectively, achieved by ordinary
perature of 105 ± 5°C [20]. The linear shrinkage values Proctor compaction [Figure 3(a)] were very low com-
 Landfill Disposal of Alum Water Treatment Residues: Some Pertinent Geoengineering Properties 101

nary Proctor compaction alone at the reduced water

content value. Controlled drying and shrinkage without
cracking of the wet sub-specimens had produced lower
air voids contents (A = 2.5–3%), compared with those
achieved by ordinary Proctor compaction alone of the
dried alum WTRs (A = 3.5–5%). Note that the air voids
content is defined as the ratio of the volume of the pore
air voids to the bulk volume, given as a percentage. The
residue materials were brittle below about 160% water
content and crushed to a dust under the impact of the
compaction rammer.
Hence, field compaction of these alum WTRs would
be most efficiently carried out over the water content
range of 200–240% owing to the tendency for these res-
idues to stick to machine plant above the adhesion limit
value of 240%. Insitu air drying and shrinkage of the
compacted residues subsequently produce higher dry
density values than achieved by compaction alone of
material placed below 200% water content, thereby in-
creasing the storage capacity of the monofill. Ordinary
Proctor compactive effort was also found to be exces-
sive for these alum WTRs, evident by some swelling of
the compacted specimens at a constant water content,
and it is suggested that a lighter field-compactive en-
ergy of about one-third ordinary Proctor compactive ef-
fort should be adequate in the case of residue monofills,
following from the recommendations by Loll [26] and
O’Kelly [27]. Hence, residues for landfilling should not
be dried beforehand below 160% water content since
they would easily crush to a dust under the action of the
Figure 3. Ordinary Proctor compaction; (a) Compaction alone, (b) compaction roller, raising additional environmental
Compaction followed by air drying.
concerns.

pared with mineral soils, although in line with the high SHEAR STRENGTH PROPERTIES
water content of 200–400% and the low specific gravity
of solids values. Furthermore, the dry density against The undrained shear strength of the ordinary Proc-
water content curve in Figure 3(a) was relatively flat tor-compacted residue specimens was measured as a
over the test range, which is a distinctive characteristic function of their water content using quick-undrained
of these residue materials, with the dry density decreas- triaxial compression tests [28]. These data are pertinent
ing marginally in value with increasing water content, to the placement and trafficability of the residues in
since the pore water constitutes an increasing propor- monofills, and achieving an adequate factor of safety
tion of the residue volume. The optimum water content against short-term geotechnical instability of the
to achieve the maximum dry density for ordinary Proc- monofill slopes. The thixotropic hardening behavior
tor compaction did not occur within the water content was studied by carrying out laboratory vane and triaxial
range tested, and this trend in the dry density data is compression tests [28] on specimens of alum WTR 2a
consistent with Wang et al. [11]. that had been allowed to cure, undisturbed at a constant
Figure 3(b) shows that below 200% water content, composition, over different periods.
the bulk and dry density values achieved by allowing The effective stress shear strength properties, which
the compacted alum WTR material to air-dry naturally are used in determining the factor of safety against in-
were consistently greater than those achieved by ordi- stability of the monofill slopes for the intermediate and
102 B. O’KELLY

long-term conditions, were measured using isotropic

consolidated-undrained triaxial compression tests, and
with continuous measurement of the pore water pres-
sure response [29].

Undrained Shear Strength

Quick-undrained triaxial compression tests were car-

ried out on the partially saturated specimens (38 mm in
diameter and 76 mm high) that had been prepared from
the one-liter ordinary Proctor compacted specimens
which had air voids contents of 3.5–5%. Some of these
triaxial specimens were allowed to air-dry slowly over
different periods prior to shearing in order to simulate
the strength gain that would occur due to air drying of
the residue used, for example, as daily cover on a mu-
nicipal landfill. A cell confining pressure of 100 kPa
was applied to the triaxial specimens, which were then
sheared quickly at 2% axial strain/min in an undrained
condition. Specimen failure was deemed to have oc-
curred at a limiting 20% axial strain, unless the measured
shear stress value had reached a maximum at a lower
strain value. Figure 4 shows data for shear stress against
axial strain for pairs of WTR 2a and WTR 2b specimens
that had been sheared at similar water contents.
Figure 5(a) shows the data for triaxial undrained
shear strength against water content on a loga-
rithm–logarithm plot. Figure 5(b) shows the water con-
tent values scaled in terms of the liquidity index [IL, Figure 5. Undrained shear strength data. Hollow and shaded sym-
bols denote quick-undrained and consolidated-undrained triaxial
Equation (6)]. Overall, the shear strength values for the compression tests, respectively; (a) Against water content, (b)
three alum WTRs were in good agreement, and consis- Against liquidity index.
tently greater than that measured for the non-chemi-
cally treated residue WTR 2b, with the shear strength logarithm–linear plots, respectively. Note that IL values
approximately inversely related to the water content of unity and zero correspond to the Atterberg liquid
and the liquidity index on logarithm–logarithm and limit and plastic limit conditions respectively, and IL < 0
indicates that the material is in a semisolid state.
w - wP
IL = (6)
wL - wP
where
w = water content
wL = liquid limit
wP = plastic limit

Thixotropy Effects

Laboratory Vane
Figure 4. Shear stress against strain for WTR 2a and WTR 2b (chemi- Saturated alum WTR 2a material prepared at 365%
cally and non-chemically treated residues, respectively) under
triaxial compression. and 520% water content (IL = 0.36 and 0.90, respec-
 Landfill Disposal of Alum Water Treatment Residues: Some Pertinent Geoengineering Properties 103

tively) was pressed into molds, 70 mm square in plan

and 70 mm high, taking care to avoid trapping air voids,
and hermetically sealed by wrapping in cling film. The
vane undrained shear strength of these specimens was
measured [28] after standing periods of up to four
weeks at ambient laboratory temperature of 20°C. The
specimens were sheared quickly using a miniature cru-
ciform vane, 25 mm in both width and length, which
was rotated at 0.1 rev/min. Figures 6(a) and 6(b) show
data for the increase in vane shear strength and the
strength gain ratio against the standing period. For ex-
ample, the vane shear strength measured for IL = 0.90
(i.e. near the liquid limit condition) was found to in-
crease from 3 kPa in the remolded state to about 7 kPa
after a nine-day curing period, with a strength gain ratio
of 3.3 achieved by the end of the four-week test period.
The higher strength gain ratios of 5.7 to 8.0 that have
been reported for alum WTRs by Wang et al. [11] corre-
spond to a ten-week standing period, and for higher li-
quidity index values of 1.6–2.0, which is in line with the
expected trend of increasing thixotropy effects with in-
creasing standing times and liquidity index values.

Triaxial Compression
Three physically-identical triaxial specimens A–C,
each 38 mm in diameter and 76 mm high, were prepared
from a saturated cake of alum WTR 2a that had been
consolidated one-dimensionally from the slurry state
under an applied vertical stress of 60 kPa to a water con-
tent near its plastic limit value (IL = 0.14) using a large
consolidometer press developed by O’Kelly [30,31].
The triaxial specimens A, B and C were allowed to cure,
undisturbed and at a constant composition, for periods
of 0, 5 and 12 days, respectively.
An isotropic cell confining pressure of 260 kPa was
then applied to these triaxial specimens under un-
drained conditions, thereby mobilizing an effective iso-
tropic confining pressure (s c¢ ) of about 60 kPa, which
had been approximately achieved under the applied Figure 6. Thixotropic behavior of alum WTR 2a; (a) Laboratory vane,
(b) Strength gain ratio, (c) Triaxial compression ( IL= 0.14).
vertical stress of 60 kPa and 1–D consolidation condi-
tions in the consolidometer press. Next, the triaxial
specimens were sheared in an undrained condition at content was near the plastic limit condition. Seed and
3.3 ¥ 10-5% axial strain/min, which was sufficiently Chan [32] reported that the effect of thixotropic harden-
slow to allow full equalization of the pore water pres- ing decreases with decreasing liquidity index, and that
sures to occur throughout the specimens at failure. the plastic limit represents a lower bound water content
Similar stress–strain behavior and shear strength val- value for thixotropic behavior.
ues were measured for the triaxial specimens A–C [Fig-
ure 6(c)], indicating that negligible thixotropic strength Effective Stress Shear Strength Properties
gain had occurred over the curing period of up to 12
days, which was expected, since the specimen water A series of isotropic consolidated-undrained triaxial
104 B. O’KELLY

compression tests [29], which included continuous in an undrained condition at rates of the order of 10-5%
measurement of the pore water pressure response, were axial strain/min, determined from curve-fitting analysis
carried out in order to study the effective stress shear of the data from the consolidation stage [29], and which
strength properties. Four physically-identical triaxial were sufficiently slow to allow full equalization of the
specimens (38 mm in diameter and 76 mm high) were pore water pressures to occur throughout the specimens
prepared for each residue material from saturated cakes at failure. The standard corrections for the restraining ef-
that had been pressed from the slurry state in the fects of the filter-paper side drain and the enclosing rub-
consolidometer apparatus and allowed to equilibrate ber membrane on the barreling-type specimen
under an applied vertical stress of sv = 30 kPa. The sets deformation response were applied to the measured
of residue specimens were isotropically consolidated data [29].
under effective cell-confining pressures of 30, 60, 120 Figure 7 shows the s¢–t¢ effective stress path plots;
and 150 kPa in the triaxial apparatus, with the speci- where s¢ = ( s1¢ + s ¢3 ) / 2 and t´ = ( s1¢ + s 3¢ ) / 2 are the
mens allowed to drain radially to filter-paper side Massachusetts Institute of Technology stress path pa-
drains, and to porous discs in contact with both speci- rameters; and s1 and s3 are the major and minor effec-
men ends, against an applied specimen back pressure of tive principal stresses, respectively. The failure lines of
200 kPa over a 24-hour period. An effective confining best fit in the s¢–t¢ plots were drawn passing through the
pressure of at least 30 kPa was applied to these speci- origin (i.e. effective cohesion of zero) and aligned with
mens in order that they would be in a normally consoli- the stress points corresponding to specimen failure,
dated condition (sc ≥ sv) during the triaxial shearing which typically occurred between 2% and 10% axial
stage. The consolidated specimens were sheared slowly strain. Note that the effective stress paths at failure for
the triaxial specimens A–C of alum WTR 2a, which had
been sheared at similar strain rates but after different
curing periods of 0, 5 and 12 days (thixotropy section),
were also coincident with the alum WTR 2a failure line
in Figure 7(b); further evidence that the effects of
thixotropic hardening are not significant at low liquid-
ity index values. Had thixotropy effects been signifi-
cant, the effective stress paths at failure for specimens
A–C would have located above the best-fit failure line
for the uncured specimens. The ratio of the mobilized
shear strength to the effective cell-confining pressure
that had been applied during the triaxial consolidation
stage had a mean value of 1.8. The effective angle of
shearing resistance (f¢) values of 39° to 44°, which
were calculated from the gradient of the s´–t´ failure
lines in Figure 7 using Equation (7), are in line with the
f¢ values of 42° and 44° reported for two alum WTRs by
Wang et al. [11].

sin f ¢ = tan a ¢ (7)

where
a¢ = gradient of the s´–t´ failure line to horizontal s¢
axis

COMPRESSION PROPERTIES

Figure 7. Effective stress path plots; (a) WTR 1 and WTR 3, (b) WTR The compression and consolidation properties were
2a and WTR 2b. determined as a function of the state of effective stress
 Landfill Disposal of Alum Water Treatment Residues: Some Pertinent Geoengineering Properties 105

by testing saturated residue specimens of different size, ping air voids. The specimen was compressed
aspect ratio and consistency under different loading one-dimensionally, with two-way vertical drainage to
and drainage conditions using the oedometer, atmosphere, under an applied vertical stress that was
consolidometer and triaxial apparatus [33]. Table 4 doubled in moving from one load stage to the next,
summarizes the initial dimensions and consistency of over the stress range 3–800 kPa. A displacement trans-
the specimens, drainage conditions and the applied ducer in contact with the oedometer loading cap con-
stresses of between 3 and 800 kPa which covered the di- tinuously measured the specimen deformation re-
verse levels of effective stress achieved in mechani- sponse. Each oedometer load stage was two days in
cally dewatering, storing and landfilling the residue duration in order to record sufficient data covering
materials: that is, low effective stress levels in lagoons, both the specimen consolidation and longer-term
low to medium effective stress levels in landfills and creep settlement responses under successive incre-
the medium to very high stresses applied by mechanical ments of applied stress.
dewatering devices at municipal works. The Figure 8 shows the oedometer data, plotted in the
consolidometer and isotropic triaxial consolidation conventional form of cumulative volumetric strain
tests have already been outlined in the context of de- against the square root of elapsed time, for alum WTRs
scribing the measurement of the effective stress shear 1, 2a and 3. The slurry specimens were highly com-
strength properties of the residue materials. pressible, deforming at slow but steady rates, and by the
end of the final load stage (800 kPa), had undergone
Oedometer Tests very large volumetric strains of up to 62%. The primary
consolidation component due to the dissipation of the
Multiple-increment oedometer tests [33] were car- excess pore water pressure was dominant during the
ried out on saturated materials prepared by thoroughly early load stages, with primary compression ratios (de-
mixing the dewatered alum residues from the munici- fined as the proportion of the total strain due to primary
pal works with distilled water to form uniform slurry consolidation for a particular load stage) of 0.75 to
pastes, which were allowed to equilibrate in sealed 0.90. However, secondary compression became in-
containers over a two-day period. The slurry paste was creasingly significant at higher stress levels, with the
then pressed into the 76-mm diameter confining ring primary compression ratio values for the alum residues
of the oedometer apparatus, taking care to avoid trap- reducing to 0.22–0.40 by the final load stage.

Table 4. Summary of the Consolidation Tests.

Water Content Duration
Dimensionsa Initial Final Applied Stress Void Ratio Load Stage
Test Drainage Compression
Method mm Consistency % % kN/m2 Initial Final day Conditions Index, Cc

Oedometer
WTR 1 76 ¥ 29 Slurry 780 280 sv = 3–800 (MI) 14.5 6.2 2 Two-way vertical 3.2
WTR 2a 76 ¥ 29 Slurry 700 190 sv = 3–800 (MI) 14.1 3.3 2 Two-way vertical 2.5
WTR 3 76 ¥ 29 Slurry 690 310 sv = 3–800 (MI) 11.9 5.4 2 Two-way vertical 2.5
Consolidometer
WTR 1 152 ¥ 139 Slurry 625 475 sv = 10–30 (MI) 14.1 8.8 7 Two-way vertical 3.9
WTR 2a 152 ¥ 139 Slurry 580 420 sv = 10–30 (MI) 10.9 8.4 7 Two-way vertical 3.1
WTR 2a 167 ¥ 139 Very soft 500 330 sv = 7.5–60 (MI) 8.8 6.2 7 Two-way vertical 2.6
WTR 2b 152 ¥ 167 Slurry 760 435 sv = 10–30 (MI) 15.4 8.1 7 Two-way vertical 5.5
WTR 3 152 ¥ 93 Slurry 600 410 sv = 10–30 (MI) 11.1 7.8 7 Two-way vertical 3.1
Isotropic Triaxial
WTR 1 38 ¥ 76 Very soft 475 340 s c¢ = 30–150 (SI) 8.8 7.3 1 All around 2.3
WTR 2a 38 ¥ 76 Very soft 420 350 s ¢c = 30–150 (SI) 8.4 7.9 1 All around 0.8
WTR 2b 38 ¥ 76 Very soft 435 340 s ¢c = 30–150 (SI) 8.1 7.2 1 All around 1.1
WTR 3 38 ¥ 76 Very soft 410 300 s ¢c = 30–150 (SI) 7.8 6.7 1 All around 1.4
MI, multiple increment; SI, single increment; s v , applied vertical stress; s c¢ , effective confining pressure.
aSpecimen dimensions are given as diameter by height.
106 B. O’KELLY

Figure 8. Oedometer data.

Consolidometer Tests

Specimens of very soft and slurry consistencies were

compressed one-dimensionally under applied stresses
Figure 9. Consolidometer data; (a) WTR 1 and WTR 3, (b) WTR 2a
of 10–60 kPa, with two-way vertical drainage to atmo- and WTR 2b. Note: hollow symbols, slurry specimens; solid symbols,
sphere, using a 152-mm diameter consolidometer press very soft specimen; sv , applied vertical stress.
[30,31]. Each load stage was extended to seven days in
duration, compared to the two days that was used as mm high) were prepared for each residue material from
standard for each oedometer load stage, because of the the cakes produced at the end of the consolidometer
greater specimen drainage length in the consolidometer tests. The sets of residue specimens were allowed to
tests. Periodic measurements of the deformation re- drain radially and from both ends under isotropic effec-
sponse were taken using a long-stroke dial gauge in tive confining pressures of 30, 60, 120 and 150 kPa,
contact with the specimen loading platen (Figure 9). with continuous measurement of the volume of pore
Again, the slurry specimens were highly compressible, water that drained from the specimens over a 24 h pe-
and by the end of the final load stage (30 kPa), had un- riod, against an applied back pressure of 200 kPa, in the
dergone large volumetric strains of up to 36%. Water triaxial apparatus [33]. The volumetric strain data are
content tests carried out on the pressed residue cakes shown against the square root of elapsed time in Figure
confirmed that the water content distribution was uni- 10 and against the state of effective stress in Figure
form across the specimen thickness, indicating that full 11(a). The drained bulk modulus (B, as MN/m2) re-
consolidation had been achieved throughout the speci- sponse is shown in Figure 11(b), where B is defined as
mens by the end of the final load stage. the ratio of the effective stress to the volumetric strain
(decimal value). The very soft residue specimens were
Triaxial Consolidation Tests also highly compressible, and by the end of the consoli-
dation stage at 150 kPa, had undergone large volumet-
Four triaxial specimens (38 mm in diameter and 76 ric strains of up to 33%.
 Landfill Disposal of Alum Water Treatment Residues: Some Pertinent Geoengineering Properties 107

Compressibility Data tion load stages (seven days compared to the two days
used as standard for each oedometer stage). Primary
The stress–strain–time data from the oedometer, compression ratio [C *c , Equation (8)] values of 0.19,
consolidometer and triaxial consolidation tests are 0.17 and 0.20, which take into account the initial con-
summarized in the conventional form of void ratio sistency (void ratio) of the specimens, were determined
against logarithm of effective stress in Figure 12. from the oedometer data for alum WTRs 1, 2a and 3, re-
The compression response was quantified in terms of spectively.
the compression index (Cc) values, which are included
in Table 4. Note that the level of compression depends Cc
C c* = (8)
on a range of factors, including: initial water content 1+ eo
(void ratio); hydraulic conductivity; specimen thick- where
ness and drainage length; applied stress; and load-stage
duration. The slurry residues were highly compress- eo = initial void ratio.
ible, with values of Cc = 3.2, 2.5 and 2.5 given by the
gradients of the oedometer void ratio against logarithm
of effective stress curves [Figure 12(a)] for alum WTRs
1, 2a and 3, respectively. Note that for most inorganic
clays, Cc < 1.0, and generally less than 0.5. The
consolidometer data gave higher values of Cc = 2.6–5.5
(Table 4) due to greater amounts of secondary compres-
sion settlement achieved by the end of the longer-dura-

Figure 10. Triaxial consolidation data; (a) WTR 1 and WTR 3, (b) Figure 11. Dewaterability under isotropic confining pressure; (a)
WTR 2a and WTR 2b. Note: s ¢c , effective confining pressure. Volumetric strain, (b) Drained bulk modulus.
108 B. O’KELLY

Figure 13. Coefficient of primary consolidation against effective

stress. Note: hollow symbols, oedometer data; solid symbols, triaxial
data.

k = mv cv g w (9)
where
gw = density of the pore water (assumed 0.98
tonnes/m3)
mv and cv = coefficients of volume change and pri-
mary consolidation, respectively; with
mv defined as the volumetric strain per
unit increase in effective stress
The hydraulic conductivity values were very low and
inversely related to the state of effective stress on a log-
Figure 12. Void ratio against effective stress (primary consolidation
settlement); (a) 1–D compression in oedometer and consolidometer arithm–logarithm plot [Figure 14(a)], with the hydrau-
apparatus, (b) Isotropic triaxial consolidation. Note: CC, compres- lic conductivity decreasing significantly from about 2 ¥
sion index.
10-4 to 1 ¥ 10-6 m/day with increasing effective stress
(reductions in water content and void ratio) from 3 to
Consolidation Rate 800 kPa. This is consistent with the experience of me-
chanically dewatering the wet alum WTRs at the treat-
The rate of consolidation of the residue materials was ment works. Wang and Tseng [12] also reported that the
very slow, and typical of high-plasticity clays. Standard hydraulic conductivity of alum WTR reduced from 10-4
curve-fitting techniques [33] were applied to the to 10-6 m/day over the effective stress range of 2 to 540
oedometer and triaxial consolidation data presented in kPa. Direct comparisons at the same levels of effective
Figures 8 and 10 in order to determine the values of the stress indicate that the hydraulic conductivity values of
coefficient of primary consolidation (cv), which were in WTR 2a and WTR 3 are broadly similar, and about one
good agreement for the residue materials, increasing order of magnitude greater than that measured for WTR
from 0.1 to 0.8 m2/year with increasing effective stress 1 [Figure 14(a)]. It is postulated that this difference is
from 3 to 800 kPa (Figure 13). partly due to the higher organic content of WTR 1 (TVS
= 57%) compared to either WTR 2a or WTR 3 (TVS =
Hydraulic Conductivity 45–46%). The hydraulic conductivity data are also
shown against the water content w in Figure 14(b) and
The hydraulic conductivity [k, Equation (9)] was de- against the void ratio e in Figure 14(c), since these pa-
termined indirectly from the oedometer data presented rameters that can be readily determined in practice us-
in Figure 8 as ing the standard oven-drying method [20] and phase re-
 Landfill Disposal of Alum Water Treatment Residues: Some Pertinent Geoengineering Properties 109

lationships. For example, assuming fully saturated

conditions, the void ratio is given by
e = wG s (10)
where
Gs = specific gravity of solids

Secondary Compression Rate

The rate of secondary compression (creep) for the

alum WTRs was quantified in terms of the coefficient
of secondary compression Csec, defined as the 1–D vol-
umetric strain per ten-fold increase in elapsed time after
the substantial completion of the primary consolidation
phase (i.e. under constant effective stress). High values
of Csec = 0.005–0.010 were determined from the
oedometer data, with a mean value of Csec = 0.006 over
the effective stress range 3–800 kPa (Figure 15). The
Csec values, and hence anticipated long-term settle-
ments, were consistently greater for WTR 1 than for ei-
ther WTR 2a or WTR 3, owing to its higher organic
content [23]. Figure 15 also shows the data values of
0.15–0.18 determined for the secondary compression
index [Cae, Equation (11)] and 0.03–0.05 for the ratio of
Cae/Cc [34], which are typical ranges for organic soils.
C ae = C sec (1 + e o ) (11)

EFFECTS OF CATCHMENT GEOLOGY

AND CHEMICAL ADDITIVES ON THE
ENGINEERING BEHAVIOR

Despite significant differences in catchment geol-

ogy, and allowing for minor differences in the chemical

Figure 15. Secondary compression data. Note: Cc, compression in-

Figure 14. Hydraulic conductivity data; (a) Against effective stress, dex; Csec, coefficient of secondary compression; Cae, secondary
(b) Against water content, w, (c) Against void ratio, e. compression index.
110 B. O’KELLY

WTR 2b. Stiffer responses and values of peak shear

stress of up to 20% greater were also measured for alum
WTR 2a,(compared with residue WTR 2b) from con-
solidated-undrained triaxial compression tests on
specimen pairs at similar water contents [Figure 16(b)].
Hence, the effective angle of shearing resistance of alum
WTR 2a was also slightly greater than that measured for
residue WTR 2b (f¢ = 42° and 40°, respectively).
Although the chemical additives are necessary for the
efficient operation of the treatment processes, they also
have a downside in that the thickened alum WTRs are
potentially more difficult to consolidate, and hence the
larger bulk residue volume will settle to a greater ex-
tent, and typically over a longer period, when placed in
a lagoon or monofill. For example, Figure 12(b) shows
that alum WTR 2a was less compressible than the
non-chemically treated residue WTR 2b, with com-
pression index values of Cc = 3.1 and 5.5 determined
from the slurry consolidometer tests (Table 4). Alum
WTR 2a was also found to consolidate at a slightly
slower rate, reflected by its lower values shown in Fig-
ure 13. Hence, from Equation (9), the hydraulic con-
ductivity of alum WTR 2a is lower than that of the
non-chemically treated residue WTR 2b, owing to the
Figure 16. Effect of polyelectrolyte on the triaxial undrained shear effect of the chemical additives, since the mineralogy;
strength; (a) Ratio of the undrained shear strength of alum WTR 2a to
that of the non-chemically treated residue WTR 2b, (b) Shear stress organic content; and test conditions are similar. This is
against strain data for specimen pairs at similar water contents (w). consistent with Wang and Tseng [12] and O’Kelly [15]
who reported that relatively large amounts of additional
dosages and seasonal variations in the source waters pore water are trapped and absorbed by the aluminum
(the test materials had been sampled at different times hydroxide precipitates and polyelectrolyte molecules
during the year), the three alum WTRs were found to within the constituent flocs, which causes a reduction in
have broadly similar geoengineering properties, most the hydraulic conductivity.
likely due to the fact that the behavior was dominated
by the high organic content (TVS = 41–57%). APPLICATION OF THE TEST DATA
At a given water content, the triaxial undrained shear
strength of the alum WTRs was consistently greater Dewaterability of Slurry Residue at Treatment
than that measured for the non-chemically treated resi- Works
due WTR 2b [Figure 5(b)], which can be explained by
the action of the polyelectrolyte additive in aggregating The alum WTR slurry must be adequately dewatered
and binding the constituent flocs to form floc clusters, at the treatment works in order to reduce its bulk volume,
thereby producing greater inter-particle contact and re- thereby lowering transportation and landfill-disposal
ducing the size of the pore voids and capillary channels costs, and to achieve an adequate shear strength for effi-
(in combination with the deactivation of some pore wa- cient handling, trafficability, and geotechnical stability
ter by the alum coagulant [12]). For example, Figure of the landfill slopes. Dewatering is achieved by me-
16a shows the ratio of the triaxial undrained shear chanical and/or thermal means, or by allowing the slurry
strength values measured for alum WTR 2a and the to dry naturally in drying beds, depending on the size of
non-chemically treated residue WTR 2b. The shear the treatment works. The reduction in the residue vol-
strength of alum WTR 2a, which comprised 4.8% ume, DV, achieved by mechanical dewatering systems
polyelectrolyte by dry solids mass, was between 1.1 can be estimated using Equation (12) and the void ra-
and 1.2 times greater than that measured for residue tio–logarithm of effective stress data shown in Figure 12.
 Landfill Disposal of Alum Water Treatment Residues: Some Pertinent Geoengineering Properties 111

Ê De ˆ shear strength of at least 20 kPa was achieved at 340%

Ê s ¢ˆ
DV = Vo Á ˜ = Vo Á ˜ (12) water content (SC = 22%) for the three alum WTRs
Ë1+ eo ¯ ËB¯
tested in this study [Figure 5(a)], although Figure 1
where shows that for other municipal sludge and residue mate-
rials, the undrained shear strength may be significantly
Vo = initial volume less than 20 kPa [8,11,16] at the maximum 300% water
eo = initial void ratio content value for municipal landfilling.
De = reduction in void ratio under the effective stress No universal relationship exists for soils between the
s¢ = applied by the mechanical dewatering system water content and the undrained shear strength, which
B= drained bulk modulus [Figure 11(b)] is also dependent on a range of other factors, including:
Note that it is technically feasible to recover the alum mineralogy; organic content; chemical dosages; and
from the slurry residue by chemical means, and propri- the type of treatment used to separate the residue
etary full-scale treatment systems are currently under by-product. Hence, it would be more prudent for land-
development in practice. The aluminum hydroxide pre- fill operators to specify a minimum value of shear
cipitate that forms during the coagulation process strength based on sound geotechnical considerations,
readily dissolves in highly acidic solutions [35] and the rather than the current requirement for a maximum wa-
liquid alum can then be decanted and crystallized by ter content of 300% alone, in determining the accept-
evaporation. Alum recovery is most successfully car- ability of sludge and residue materials for landfilling.
ried out using sulfuric acid over the range of pH 2–3, Note that in the case of residue monofills, higher shear
and for retention periods of 10–20 min [36]. However, strengths of at least 50 kPa are recommended for
the effectiveness of these techniques has been varied to geotechnical stability.
date, with the purity of the recovered alum and the over- Data for undrained shear strength against water con-
all economy of the recovery process remaining contro- tent, such as that shown in Figure 5(a) for the alum
versial issues. The data for alum WTR 2a and the WTRs tested in this study, can be used to select an ap-
non-chemically treated residue WTR 2b tested in this propriate mechanical dewatering system in order to
study indicated that the chemical additives had the ef- achieve a specified minimum value of shear strength
fect of reducing the hydraulic conductivity of the slurry before the residue cake is transported for disposal
residue. Further studies are necessary in order to deter- offsite. For example, the recessed-plate filter press de-
mine the extend to which these effects are reversible, vices used at the Ballymore Eustace and Leixlip treat-
since alum recovery may also lead to greater levels of ment works reduced the water content value of the alum
mechanical dewatering of the residue being achieved, slurry residues WTR 1 and WTR 3 to 340% and 300%
and hence reduced landfill-disposal costs. respectively (Table 3), thereby achieving adequate un-
drained shear strengths of about 25 and 31 kPa respec-
Acceptability of Residues for Landfilling Based on tively, and satisfying geotechnical stability criteria for
Shear Strength Criteria municipal landfilling. At the Clareville works, the wa-
ter content value of alum WTR2a was reduced to about
In accordance with current guidelines [4–6], munici- 700% using a belt-press device, and to about 570% by
pal landfill operators will usually not accept sludge or allowing the pressed residue to dry naturally in drying
residues with a water content value greater than 300% beds, although the remolded material was still of slurry
(SC = 25%), which has been set as an indirect measure consistency, and therefore unsuitable for municipal
of the shear strength necessary for efficient handling, landfilling. Belt dryer devices that fully or partially dry
trafficability, and geotechnical stability. In municipal the pressed residue cake at low temperatures; thermal
landfills, the wet residue material is usually placed in treatments or soil-conditioning techniques can be in-
thin layers and mixed and scarified insitu with the solid vestigated as alternative methods to dewater these resi-
waste, which has the effect of reducing the water con- dues sufficiently and expeditiously. The miniature vane
tent and increasing the shear strength of the residue apparatus [28] has been shown to provide a quick and
fraction. Minimum shear strengths of 20 and 25 kPa accurate method of measuring the undrained shear
have been recommended by Loll [26] and strength (O’Kelly [27]; Loll [26]) in order to assess
Siedlungsabfall [37] for the co-disposal of sludge and whether the residue material has been adequately
residue materials at municipal landfills. An undrained dewatered before leaving the treatment plant, and again
112 B. O’KELLY

by the landfill operator before accepting the residue tion of the background to Equations (13)–(15) and their
cake for disposal. application in determining the magnitude and rate of
settlement can be obtained in undergraduate soil me-
Settlement of Sludge Lagoons and Residue chanics textbooks, including Craig [19].
Monofills
SUMMARY AND CONCLUSIONS
The high values of the compression index and the co-
efficient of secondary compression measured for the Despite significant differences in catchment geol-
alum WTRs indicate that residue lagoons and monofills ogy, and allowing for minor differences in the chemical
will consolidate significantly and continue to settle by dosages and seasonal variations in the source waters,
secondary compression (creep) over a long period. This the three alum water treatment residues tested in this
time-dependent settlement response comprises the sum study were found to have similar geoengineering and
of the primary consolidation and secondary compres- hydraulic properties, akin to high-plasticity non-fi-
sion components, DHc and DHsec respectively, which brous organic clays. These viscous slurry residues were
can be quantified for 1–D settlement conditions using highly compressible, although the consolidation rate
Equations (13)–(14). The time period to achieve sub- was low (hydraulic conductivity of the order of
stantial completion of the primary consolidation com- 10-4–10-6 m/day), and underwent thixotropic harden-
ponent, t1, can be quantified using Equation (15). ing in an undisturbed condition. For mechani-
cally-dewatered residue, the ratio of the triaxial shear
De
DH c = H o (13a) strength to the effective confining pressure was about
1+ eo 1.8.
Ordinary Proctor compaction produced low values of
Cc s¢ s¢ s¢ s¢
DH c = H o log vo v = H oC *c log vo v bulk density and dry density of 0.96–1.13 and
1+ eo s ¢vo s vo
¢ 0.21–0.36 tonne/m3 respectively, over the water con-
(13b) tent range 200–400%, although field compaction
would be most efficiently carried out over the water
t1 content range 200–240% using a compactive energy of
DH sec = H oC sec log (14)
t2 about one-third that of ordinary Proctor compactive ef-
fort. It is also recommended that landfill operators
Tv d 2 should specify minimum shear strengths of 20 and 50
t1 = (15)
cv kPa for residue disposal at municipal landfills and dedi-
where cated monofills respectively, in order to satisfy
geotechnical stability criteria. Significant settlements
d = effective drainage length of the residue deposit
of residue lagoons and monofills can be expected to oc-
(initially Ho in thickness)
cur over a long period, owing to the low consolidation
eo = initial void ratio
and high creep rates.
Tv = dimensionless time factor related to the aver-
The triaxial undrained shear strength of the alum
age degree of consolidation
WTRs was found to be 10–20% greater than that mea-
t2 = time period that extends into the secondary
sured for the non-chemically treated residue of similar
compression phase (t2 > t1)
mineralogy and organic content, which can be largely
s vo
¢ = initial vertical effective stress
explained by the action of the polyelectrolyte additive
De = reduction in the void ratio due to the increase
in aggregating and binding the constituent flocs to
in vertical effective stress, s v¢ (i.e. applied
form floc clusters. Although the chemical additives
stress)
are necessary for the efficient operation of the munici-
The values of the compression index Cc [from Figure pal treatment processes, they were also found to have a
12(a)]; the coefficient of primary consolidation cv downside in that the thickened alum residue was po-
(from Figure 13); and the coefficient of secondary com- tentially more difficult to consolidate; hence the larger
pression Csec (from Figure 15) that are used in these cal- bulk volume would settle to a greater extent, and typi-
culations must be consistent with the insitu effective cally over a longer period, when placed in a lagoon or
stress; loading and drainage conditions. A full explana- monofill.
 Landfill Disposal of Alum Water Treatment Residues: Some Pertinent Geoengineering Properties 113

sludges–shear strength”, J. Water Science and Technology, Vol. 36,

ACKNOWLEDGMENTS No. 11, 1997, pp. 43–50.
17. O’Kelly, B. C., “Mechanical properties of dewatered sewage sludge”,
The author would like to thank Martin Carney, Michael J. Waste Management, Vol. 25, No. 1, 2005, pp. 47–52.
Quille, Stephen Conlon, Brian Devaney, Patrick Fogarty, 18. O’Kelly, B. C., “Geotechnical properties of municipal sewage
sludge”, J. Geotechnical and Geological Engineering, Vol. 24, No. 4,
Mbakure Johnson, and Emma Quinn for carrying out the 2006, pp. 833–850.
tests on the residue materials at the geotechnical laborato- 19. Craig, R. F. 2004. Craig’s Soil Mechanics, 7th edn., Taylor & Francis.
ries, Trinity College Dublin; Paul Johnston for his helpful 20. British Standards Institution. 1990. Methods of Test for Soils for Civil
Engineering Purposes. Part 2: Classification Tests, BS1377–2, Lon-
comments during the preparation of this paper; and Mi- don, British Standards Institution.
chael Quille for his assistance in producing the tables and 21. British Standards Institution. 1990. Methods of Test for Soils for Civil
Engineering Purposes. Part 3: Chemical and Electro-Chemical Tests,
figures contained in this paper. BS1377–3, London, British Standards Institution.
22. Babatunde, A. O. and Y. Q. Zhao, “Constructive approaches toward
REFERENCES water treatment works sludge management: An international review of
beneficial reuses”, J. Environmental Science and Technology, Vol. 37,
No. 2, 2007, pp. 129–164.
1. Chih Chao, W.; H. Chihpin and D. J. Lee, “Effects of polymer dosage
on alum sludge dewatering characteristics and physical properties”, J. 23. Mitchell, J. K. and K. Soga. 2005. Fundamentals of Soil Behavior, 3rd
Colloids and Surfaces A: Physicochemical and Engineering Aspects, edn., New Jersey: Wiley and Sons.
Vol. 122, No. 1–3, 1997, pp. 89–96. 24. O’Kelly, B. C., “Effect of biodegradation on the consolidation proper-
2. Turchiuli, C. and C. Fargues, “Influence of structural properties of ties of a dewatered municipal sewage sludge”, J. Waste Management,
alum and ferric flocs on sludge dewaterability”, J. Chemical Engineer- Vol. 28, No. 8, 2008, pp. 1395–1405.
ing, Vol. 103, No. 1–3, 2004, pp. 123–131. 25. British Standards Institution. 1990. Methods of Test for Soils for Civil
3. Council of the European Union. 1998. Council Directive 1998/83/EC Engineering Purposes. Part 4: Compaction-Related Tests, BS1377–4,
of 3rd November 1998 on the Quality of Water Intended for Human London, British Standards Institution.
Consumption, In Official Journal of the European Communities, 26. Loll, U. 1991. “Measurement of laboratory vane shear strength as com-
L330, pp. 32–54. munity standard or reference method for testing the suitability of
4. US Environmental Protection Agency. 1996. Technology Transfer dewatered sludges for use as landfill”, In Treatment and Use of Sewage
Handbook: Management of Water Treatment Plant Residuals, EPA Sludge and Liquid Agricultural Wastes. Ed. L’Hermite P., London:
625-R-95-008, Cincinnati, OH: Environmental Protection Agency. Elsevier, pp. 310–315.
5. Council of the European Union. 1999. Council Directive 1999/31/EC 27. O’Kelly, B. C., “Geotechnical aspects of sewage sludge monofills”, J.
of 26th April 1999 on the Landfill of Waste, In Official Journal of the Municipal Engineer, Proceedings of the Institution of Civil Engineers,
European Communities, L182, pp. 1–19. Vol. 157, No. 3, 2004, pp. 193–197.
6. Council of the European Union. 2006. Council Directive 2006/12/EC 28. British Standards Institution. 1990. Methods of Test for Soils for Civil
of the Council of 5th April 2006 on Waste, In Official Journal of the Engineering Purposes. Part 7: Shear Strength Tests (Total Stress),
European Communities, L114, pp. 1–13. BS1377–7, London, British Standards Institution.
7. O’Kelly, B. C., “Accurate determination of moisture content of or- 29. British Standards Institution. 1990. Methods of Test for Soils for Civil
ganic soils using the oven drying method”, J. Drying Technology, Vol. Engineering Purposes. Part 8: Shear Strength Tests (Effective Stress),
22, No. 7, 2004, pp. 1767–1776. BS1377–8, London, British Standards Institution.
8. Novak, J. T. and D. C. Calkins, “Sludge dewatering and its physical 30. O’Kelly, B.C. 2008. “Development of a large consolidometer-
properties”, J. American Water Works Association, Vol. 67, No. 1, permeameter apparatus for testing soft soils”, In Proceedings of the
1975, pp. 42–45. ASCE GeoCongress on the Challenge of Sustainability in the
Geoenvironment, New Orleans, Louisiana, Geotechnical Special Pub-
9. Raghu, D. and H-N.Hsieh, “Material properties of water treatment lication 179, pp. 60–67.
plant sludges”, J. Civil Engineering for Practicing and Design Engi-
neers, Vol. 5, No. 11–12, 1986, pp. 927–941. 31. O’Kelly, B. C., “Development of a large consolidometer apparatus for
testing peat and other highly organic soils”, J. SUO–Mires and Peat,
10. Geuzens, P. and W. Dieltjens. 1991. “Mechanical strength determination Vol. 60, No. 1–2, 2009, pp. 23–36.
of cohesive sludges–a Belgian research project on sludge consistency”, In
Recent Developments in Sewage Sludge Processing. Eds. Colin, F.; P. J. 32. Seed H. B. and C. K. Chan, “Structure and strength characteristics of
Newman and Y. J. Puolanne, London: Elsevier, pp. 14–23. compacted clay”, ASCE J. Soil Mechanics and Foundations Division,
Vol. 85, No. SM5, 1959, pp. 87–109.
11. Wang, M. C.; J. Q. Hull; M. Jao; B. A. Dempsey and D. A. Cornwell,
“Engineering behavior of water treatment sludge”, ASCE J. Environ- 33. British Standards Institution. 1990. Methods of Test for Soils for
mental Engineering, Vol. 118, No. 6, 1992, pp. 848–864. Civil Engineering Purposes. Part 5: Compressibility, Permeability
12. Wang, M. C. and W. Tseng, “Permeability behavior of a water treat- and Durability Tests, BS1377–5, London, British Standards Institu-
ment sludge”, ASCE J. Geotechnical Engineering, Vol. 119, No. 10, tion.
1993, pp. 1672–1677. 34. Mesri, G.; M. Shahien and T. W. Feng. 1995. “Compressibility param-
13. Lim, S.; W. Jeon; J. Lee; K. Lee and N. Kim, “Engineering properties eters during primary consolidation”, In Proceedings of the Interna-
of water/wastewater-treatment sludge modified by hydrated lime, fly tional Symposium on Compression and Consolidation of Clayey Soils,
ash and loess”, J. Water Research, Vol. 36, No. 17, 2002, pp. Hiroshima, Vol. 2, pp. 1021–1037.
4177–4184. 35. Xu, G. R.; Z. C. Yan; Y. C. Wang and N. Wang, “Recycle of alum re-
14. Roque, A. J. and M. Carvalho. 2006. “Possibility of using the drinking covered from water treatment sludge in chemically enhanced primary
water sludge as geotechnical material”, In Proceedings of the 5th In- treatment”, J. Hazardous Materials, Vol. 161, No. 2–3, 2009, pp.
ternational Congress on Environmental Geotechnics, Cardiff, UK, 663–669.
Vol. 2, pp. 1535–1542. 36. Abodo, M. S. E.; K. T. Ewida and Y. M. Youssef, “Recovery of alum
15. O’Kelly, B. C., “Geotechnical properties of a municipal water treat- from waste sludge produced from water treatment plants”, J. Environ-
ment sludge incorporating a coagulant”, Canadian Geotechnical J., mental Science Health, Vol. 28, No. 6, 1993, pp. 1205–1216.
Vol. 45, No. 5, 2008, pp. 715–725. 37. Siedlungsabfall, T. A. 1993. Dritte Allgemeine Verwaltungsvorschrift
16. Wichmann, K. and A. Riehl, “Mechanical properties of waterwork zum Abfallgesetz, Munich: Bundesanzeiger.