Science of the Total Environment 598 (2017) 916–924

Contents lists available at ScienceDirect

Science of the Total Environment

journal homepage: www.elsevier.com/locate/scitotenv

Effect of naphtha diluent on greenhouse gases and reduced sulfur

compounds emissions from oil sands tailings
Kathleen F. Gee, Ho Yin Poon, Zaher Hashisho, Ania C. Ulrich ⁎
Department of Civil and Environmental Engineering, University of Alberta, Canada

H I G H L I G H T S G R A P H I C A L A B S T R A C T

• Indigenous microbes degraded residual

diluent in tailings thus producing emis-
sions.
• Reduced sulfur compounds were pro-
duced first at a max. of 0.12 μmol
RSCs/mL MFT.
• H2S and 2-methylthiophene contribut-
ed 81% of the RSCs gas produced.
• CH4 and CO2 production occurred after
week 5.
• A max. of 40.7 μmol CH4/mL MFT and
5.9 μmol CO2/mL MFT was produced.

a r t i c l e i n f o a b s t r a c t

Article history: The long-term storage of oil sands tailings has resulted in the evolution of greenhouse gases (CH4 and CO2) as a
Received 12 January 2017 result of residual organics biodegradation. Recent studies have identified black, sulfidic zones below the tailings-
Received in revised form 14 April 2017 water interface, which may be producing toxic sulfur-containing gases. An anaerobic mesocosm study was con-
Accepted 14 April 2017
ducted over an 11-week period to characterize the evolution of CH4, CO2 and reduced sulfur compounds (RSCs)
Available online xxxx
(including H2S) in tailings as it relates to naphtha-containing diluent concentrations (0.2, 0.8, and 1.5% w/v) and
Editor: D. Barcelo microbial activity. Our results showed that RSCs were produced first at 0.12 μmol°RSCs/mL MFT (1.5% w/v diluent
treatment). RSCs contribution (from highest to lowest) was H2S and 2-methylthiophene N 2.5-
Keywords: dimethylthiophene N 3-methylthiophene N thiofuran N butyl mercaptan N carbonyl sulfide, where H2S and 2-
Greenhouse gases methylthiophene contributed 81% of the gas produced. CH4 and CO2 production occurred after week 5 at
Reduced sulfur compounds 40.7 μmol CH4/mL MFT and 5.9 μmol CO2/mL MFT (1.5% w/v diluent treatment). The amount of H2S and CH4 gen-
Methanogens erated is correlated to the amount of diluent present and to microbial activity as shown by corresponding in-
Sulfate reducing bacteria creases in sulfate-reducers' Dissimilatory sulfite reductase (DsrAB) gene and methanogens' methyl-coenzyme
Oil sands tailings
M reductase (MCR) gene.
Diluent
© 2017 Elsevier B.V. All rights reserved.

1. Introduction

⁎ Corresponding author at: Department of Civil and Environmental Engineering,
University of Alberta, Edmonton, Alberta T6G2W2, Canada.
E-mail address: aulrich@ualberta.ca (A.C. Ulrich).
The oil sands deposits in Alberta are the third largest in the world
after Saudi Arabia and Venezuela, with an estimated oil reserve of 166
billion barrels of oil (Government of Alberta DoE, 2014). Bitumen is
commonly extracted from surface-mined oil sands ore using the Clark

http://dx.doi.org/10.1016/j.scitotenv.2017.04.107
0048-9697/© 2017 Elsevier B.V. All rights reserved.
 K.F. Gee et al. / Science of the Total Environment 598 (2017) 916–924
Hot Water Extraction Process, which, in turn, has generated enormous
quantities of fluid tailings waste (Arkell et al., 2015). These fluid tailings,
an alkaline slurry mixture of process water, sand, silt, clays, organics, in-
organics, and unrecovered bitumen and diluent (Small et al., 2015) are
deposited in large tailings ponds under a zero discharge policy
(Government of Alberta DoE, 2013). Following the initial settling of
coarse solids (N 44 μm) (Chalaturnyk et al., 2002), the suspended fines
in the fluid fine tailings (FFT, ~ 10 wt% solids) gradually settle over
2–4 years to become mature fine tailings (MFT, ~ 30 wt% solids)
(Arkell et al., 2015). Due to their poor consolidation properties, it may
require several decades or more in situ before tailings sufficiently con-
solidate for reclamation (BGC Engineering Inc., 2010). Currently, the
surface area of these ponds is approximately 182 km2 and contains an
estimated 976 million m3 of accumulated tailings (Government of
Alberta DoE, 2013; Government of Alberta DoE, 2014).
One effect of long-term storage of tailings in the ponds is the evolu-
tion of gases. Greenhouse gases (GHG), methane (CH4) and carbon di-
oxide (CO2), are now emitted in most tailings ponds tested to date
(Siddique et al., 2012; Small et al., 2015). Tailings ponds that receive
froth treatment tailings (FTT), tailings that contain hydrocarbons and
diluent, were reported to have higher GHG emissions and were more
prone to intensive bubbling compared to using composite tailings and
thickened tailings (Burkus et al., 2014). Diluents are added to improve
recovery rates of bitumen (Small et al., 2015). The n-alkanes and BTEX
(benzene, toluene, ethylbenzene, and xylenes) compounds found in
naphtha diluent have been observed to stimulate the biological produc-
tion of GHG (Siddique et al., 2006; Siddique et al., 2007). Recent re-
search has also revealed the presence of black, sulfidic zones beneath
the tailings-water interface (Chen et al., 2013; Chi Fru et al., 2013;
Ramos-Padron et al., 2011; Stasik and Wendt-Potthoff, 2014) that are
abundant in sulfate reducing bacteria (SRB) (Stasik and Wendt-
Potthoff, 2014). Potentially, the SRB in these zones are consuming sul-
fate and producing considerable amounts of HS− and toxic, hydrogen
sulfide (H2S) gas (Stasik and Wendt-Potthoff, 2014). The addition of
gypsum to increase MFT densification increases the sulfate concentra-
tion in tailings, which in the presence of organics such as naphtha dilu-
ent, provides SRB with substrates for H2S production. Beside
precipitation, H2S emissions are most likely prevented by the chemical
(Ramos-Padron et al., 2011) and microbial re-oxidation (Boudens
et al., 2016; Stasik et al., 2014). However, there is a lack of peer-
reviewed literature regarding H2S emissions and other reduced sulfur
compounds (RSCs) in tailings. Indeed, H2S is a well-known toxic gas
that impacts human health, and other RSCs (including H2S) compounds
can be transformed into sulfur dioxide and/or sulfuric acid in the atmo-
sphere (Bates et al., 1992). These transformed products can contribute
to acidic precipitation and aerosol formation, which further affect
human and environmental health (Bates et al., 1992; Small et al.,
2015). Additionally, there is little peer-reviewed literature on the tem-
poral relationship between CH4, CO2 and RSCs (including H2S)
productions.
As such, the overall objective of this study is to understand the evo-
lution of CH4, CO2 and RSCs (including H2S) in tailings as it relates to:
i) diluent concentration and ii) microbial activity. Studying the produc-
tion of these compounds in tailings will aid in further understanding the
emissions currently being released from tailings, and yield further in-
sight on the effect of adding sulfate-containing substrates on gas pro-
duction that require consideration in current or future tailings
remediation plans.
2. Materials and methods
2.1. Mature fine tailings, pond water, and naphtha diluent
MFT and pond water (PW) samples were provided by Canada's Oil
Sands Innovation Alliance (COSIA). The samples were taken in June
2012 at a depth of 12 m below the surface and were stored at 4 °C
917
until use. Due to limited availability of pond water, surrogate pond
water (SPW) was also prepared and used in this experiment (see
Table S1 for recipe). Naphtha diluent was also provided by COSIA and
was determined in our laboratory to have a specific gravity of 0.77
(T = 20 °C and P = 1 atm).
2.2. Mesocosm set-up
Mesocosms were constructed using 1 L Pyrex® glass bottles and
modified caps fitted with butyl rubber septa. MFT and pond water sam-
ples were oxygen-purged using.
Argon gas (Praxair) prior to use. MFT were amended with typical dil-
uent concentration in the tailings ponds – 0.8% w/v (b 1% mass) (Penner
and Foght, 2010), whereas a range of diluent concentrations (0.2% w/v,
0.8% w/v, and 1.5% w/v) were studied using the SPW. All experimental
mesocosms contained 400 mL MFT and 400 mL pond water (PW or
SPW) with diluent. Two types of controls were used in both the PW
and SPW groups: no-diluent control (400 mL MFT and 400 mL pond
water) as a 0% w/v baseline, and a no-MFT control (800 mL pond
water and 0.8% w/v diluent) to account for the presence of MFT in this
study. All mesocosms were assembled in an anaerobic chamber (5%
CO2, 5% H2, N2 balance) and incubated in the dark at 24 °C. Mesocosms
were monitored for 14 weeks (for RSCs production) or 11 weeks (for all
other parameters).
2.3. Gas analysis
CH4 and CO2 were measured with a 7890A gas chromatograph with
a thermal conductivity detector (GC-TCD), and RSCs (hydrogen sulfide,
carbonyl sulfide, thiofuran, butyl mercapatan, 2-methylthiophene, 3-
methylthiophene, and 2,5-dimethylthiophene) were analyzed using a
7890A GC and an Agilent Technologies 355 sulfur chemiluminescence
detector (GC-SCD) (See Supplementary information for detail analysis
conditions).
2.4. Chemical analyses
Prior to the experiment the solids, bitumen, and water content of the
MFT were characterized using the Dean Stark procedure (Cao et al.,
2014). The chemical analyses in this study were conducted on a mixture
of released water and pore water but for the purposes of this study will
be referred to as oil sands process water (OSPW). Prior to taking a liquid
sample, the mesocosms were inverted to ensure complete mixing thus
reducing sampling bias. After inversion, the mesocosms were placed
upright and allowed to sit for approximately 1 h to minimize solids
from clogging the sampling syringe. At each analytical time point,
20 mL of liquid sample was removed from the OSPW layer and mea-
sured for redox using an Accumet redox probe (4 M KCl internal solu-
tion) in the anaerobic chamber. The OSPW samples were then
removed from the chamber, centrifuged at 8000 ×g for 5 min at 20 °C,
and the supernatant was removed and filtered with 0.45 μm filters. Con-
ductivity and pH were measured using an ExStik®II pH/Conductivity/
TDS Meter. Alkalinity was measured using a Mettler Toledo DL53 with
0.02 N H2SO4 as a titrant. Dissolved organic carbon samples were ana-
lyzed with a Shimadzu TOC-L CPH using the Non-Purgeable Organic
Carbon (NPOC) method. OSPW samples for sulfate and nitrate analysis
were filtered (0.22 μm) and analyzed using a Dionex ICS-2100 ion chro-
matography, or by the Department of Biological Sciences at the Univer-
sity of Alberta using a Dionex DX600 ion chromatography and EPA
method 300.1 (USEPA 1997). Detailed explanation of the sulfate reduc-
tion rate calculation is included in the Supplementary information.
2.5. Microbial analysis
A 1.5 mL MFT and OSPW mixture was withdrawn from each
mesocosm under anaerobic conditions in week 5 and 11. In addition,
918 K.F. Gee et al. / Science of the Total Environment 598 (2017) 916–924
MFT samples collected prior to commencing the experiment was used
as a week 0 control. The MFT and OSPW mixture (1.5 mL) were then
centrifuged for 30 s at 10000 × g, and supernatant was discarded. A
0.25 g MFT sample was used for the metagenomic DNA extraction
with Power Soil DNA Extraction Kit (MoBio Laboratories, USA) accord-
ing to the manufacturer recommended procedure. The Dissimilatory
sulfite reductase (DsrAB) and methyl-coenzyme M reductase (MCR)
genes were amplified from 0.5 μL of template DNA. PCR was performed
with Phusion High-fidelity DNA polymerase (Thermo scientific, USA)
with the initial denaturation 98 °C for 2 min and final extension 72 °C
for 2 min. The amplification conditions, Dsr (Wagner et al., 1998) and
mcrA (Springer et al., 1995) primers are listed in Table S2. The resulting
PCR products were analyzed on 2% agarose gels. Microbial communities
were characterized by 16S rRNA gene with Illumina MiSeq sequencing
platform (detailed information can be found in the Supplementary
information).
3. Results and discussion
3.1. Sample properties prior to diluent amendment
Physical and chemical properties of the no-diluent controls at t = 0,
which represent the initial sample properties in the MFT-containing
mesocosms prior to diluent amendment, are summarized in Table 1.
The physical and chemical parameters of MFT were characterized
prior to the experiment (Table 1), and the 9.5% discrepancy observed
was due to water evaporation during the Dean Stark extraction process.
Regardless of pond water group, pH of the mesocosms were slightly al-
kaline and conductivity values ranged from 3.29 to 3.36 mS/cm. High,
initial alkalinity concentrations were found in the PW and SPW groups,
which provided buffering capacity in the tailings (Table 1). Positive
redox potential values were observed for PW and SPW groups
(Table 1). These values suggest that a small amount of oxygen was in-
troduced into the tailings sample as the standard redox potential
(pH 7 and T = 25 °C) for an oxygen to water coupling is significantly
more positive at +820 mV (Liebensteiner et al., 2014). The initial DOC
concentrations in the PW no-diluent controls (92 ± 0.3 mg/L) were
nearly three times that in the SPW samples. As the SPW does not con-
tain DOC, the 33 ± 3.8 mg/L found in the SPW samples was contributed
by the MFT. The OSPW in both pond water groups were also analyzed
for initial nitrate and sulfate concentrations. Nitrate concentrations in
both PW and SPW mesocosms at t = 0 were b2 μmol, indicating that ni-
trate reduction by microorganisms was unlikely to be significant at this
point in the study (Penner and Foght, 2010). The amount of sulfate in
the tailings samples was found to be similar to those previously report-
ed (ranged between 907.8 and 1207.6 μmol) (Allen, 2008). The SPW
group had 633 μmol more sulfate than the PW samples at time 0, how-
ever, this discrepancy was likely attributed to the SPW being based on
historical OSPW data received from the oil sands industry. Regardless,
Table 1
Initial physical and chemical properties for the MFT and pond water samples prior to
amendment.
Parameter PW with MFT SPW with MFT
Bitumen (%) – –
Solids (%) – –
Water (%) – –
Redox, EH (mV) 369 ± 12 190 ± 22
pH 7.84 ± 0.01 7.91 ± 0.02
Conductivity (mS/cm) 3.29 ± 0.01 3.36 ± 0.04
Alkalinity (mg CaCO3/L) 779 ± 7 895 ± 6
DOC (mg/L) 92 ± 0.3 33 ± 3.8
NO−3 (μmol) 1.58 ± 0.31 1.23 ± 0.09
SO2−
4 (μmol) 932.3 ± 20.2 1565.8 ± 9.2
Note: Measurements for PW and SPW with MFT are presented as averaged values from
duplicate mesocosms. MFT, mature fine tailings; DOC, dissolved organic carbon; NO−
3 , ni-
trate; SO2−
4 , sulfate.
as a result of the high sulfate levels, sulfate reduction is likely to be a sig-
nificant microbial process in both the PW and SPW mesocosms.
3.2. Changes in chemical parameters
The chemical parameters outlined in Table 1 were monitored over
an 11-week period to determine whether there were any observed
trends during gas production.1 Over the course of 11 weeks, pH and con-
ductivity values did not alter significantly: pH ranged from 7.59 (±na)
to 7.95 (± 0.09), and conductivity ranged from 2.78 (± na) to 3.59
(±0.28) mS/cm (Fig. 1a–d). This indicates that any microbial processes
occurring within the tailings was buffered by the high alkalinity of the
samples, and that no significant ion release was detected during our
study.
A decrease in nitrate concentrations was observed in the PW, no-
MFT controls. Initially containing 94.6 μmol of nitrate, concentrations
gradually declined to 2.7 μmol by week 11 (Fig. 2a & b). As the redox po-
tentials in these mesocosms were largely positive throughout the study
(254 ± 20 mg/L to 395 ± 0.8 mg/L) it is likely that microbial nitrate re-
duction was occurring during this time (Fig. 2e & f). It is possible that
there was similar initial nitrate concentrations within all of the PW
mesocosms prior to t = 0 but the nitrate may have been utilized in
the days between mesocosm assembly and withdrawing the t = 0 liq-
uid samples. Therefore these results indicate that microbial nitrate re-
duction likely occurred in the MFT-containing mesocosms prior to the
t = 0 liquid sampling and was followed by microbial sulfate reduction
until sulfate had become depleted around week 2.
In contrast, there was a significant shift in redox potentials and sul-
fate concentrations within the first few weeks of the study (Fig. 2c–f).
While the redox potentials in the no-MFT controls remained largely
positive throughout the 11 weeks, the redox potentials in all of the
MFT-containing mesocosms markedly decreased and remained nega-
tive for the remainder of the study (Fig. 2e & f). This decrease in redox
potentials coincided with a decrease in sulfate concentrations and in-
crease in alkalinity (Fig. 2c, d, g & h), suggesting that there was an in-
crease in microbial sulfate reduction. A similar drop in redox values
were also observed after one week in unamended fluid fine tailings
(FFT) samples from Syncrude Canada Ltd.'s WIP pond (Chen et al.,
2013). The authors suspected this was due to the production of highly
reducing sulfide compounds by SRB (Chen et al., 2013). The standard re-
duction potential (E0) for SO24 −/HSO− 3 occurs at E° of approximately
− 516 mV (Liebensteiner et al., 2014). Although our redox values for
the MFT mesocosms did not appear to decrease below − 145 (± na)
in the first two weeks during sulfate depletion (Fig. 2e & f), our redox
measurements were conducted on OSPW and not a whole MFT sample.
Salloum et al. (2002) reported that redox values in pore water samples
could be up to 215 mV higher than those in a whole MFT sample. There-
fore the redox values in the MFT of our mesocosms, where microbial ac-
tivity is occurring, is likely lower than what was measured in OSPW in
our study. Given the shifts in redox, sulfate, and alkalinity concentra-
tions, and the high initial sulfate concentrations and low nitrate concen-
trations in the MFT-mesocosms at t = 0, the results of our study indicate
that microbial sulfate reduction occurred during the first two weeks of
MFT this study.
With respect to the DOC content of the samples, the no-MFT
2.5
51 mesocosms had relatively stable DOC concentrations between 42
37 (± 0.9) mg/L and 79 (±7.6) mg/L throughout the 11 weeks (Fig. S1a
– & b). However, there is no clear trend concerning the fluctuating DOC
– concentrations between the different treatment levels (Fig. S1a & b).
–
–
In some cases, there were significantly large differences in DOC content
– within duplicate mesocosms. For example, at week 11 in the SPW
– group, half of the no-diluent controls, 0.8% w/v and 1.5% w/v bottles
–
1
Due to sampling difficulties between week 1 and week 4 of this experiment, these
measurements represent a single mesocosm and are denoted by a standard deviation of
±na [not available].
 K.F. Gee et al. / Science of the Total Environment 598 (2017) 916–924
Fig. 1. pH (a and b) and conductivity (c and d) measurements for the PW and SPW groups over the 11-week study period. The measurements for the MFT containing mesocosms at weeks
1, 2, and 4 represent one measurement. All other values are averages from duplicate measurements and standard deviation bars, where visible, are plus and minus one standard deviation.
had a DOC concentration b 45 mg/L whereas the duplicate mesocosms
contained DOC concentrations N 920 mg/L (Fig. S1a & b). In addition to
this large range of fluctuating DOC, the DOC concentrations decreased
very slightly unless a significant spike in DOC concentrations preceded
it (Fig. S1a & b). One possibility is that the DOC content in the samples
at t = 0 are largely recalcitrant (Penner and Foght, 2010) and the in-
creases in DOC are due to microbial activity within the tailings. Some
microorganisms have the ability to excrete extracellular polymeric sub-
stances (EPS) (Flemming and Wingender, 2001) that can be used by the
microorganisms to emulsify hydrocarbons, thereby increasing hydro-
carbon bioavailability (Vasconcellos et al., 2011). Bordenave et al.
(2010) reported that tailings aggregates formed under various condi-
tions (methanogenesis, nitrate reduction, and sulfate reduction)
contained fine clays, microbial cells, and EPS. It is postulated that the
EPS emulsifies hydrocarbons in the tailings to form readily available
DOC, and this represents the subsequent spike of DOC in the OSPW be-
fore being taken up by microorganisms.
3.3. Sulfate reduction rates
Sulfate reduction rates (SRR) were determined for mesocosms
that contained MFT. In general, SRRs were found to be higher in
mesocosms amended with greater amounts of diluent. In the SPW
group, the maximum SRR observed in the no-diluent controls was
0.08 μmol/mL MFT/day (Fig. 3b). However, when amended with dil-
uent, maximum SRRs increased by a factor of 2.6 to 3.4 times that of
the no-diluent controls, ranging from 0.21 μmol/mL MFT/day (0.2% w/v)
to 0.27 μmol/mL MFT/day (0.8% w/v) (Fig. 3b). Similar maximum SRRs
were observed in the PW no-diluent controls (0.07 μmol/mL MFT/day)
and 0.8% w/v mesocosms (0.29 μmol/mL MFT/day) but sulfate reduction
in the 0.8% w/v mesocosms did not peak until week 2 (Fig. 3a). One
possibility for the one-week lag between the PW and SPW pond water
groups is due to the presence of toxic compounds in the PW water that
temporarily inhibited the SRB. Regardless of this, after week 2, when
919
sulfate had become depleted (b 29 μmol), SRRs in all MFT-diluent
amended mesocosms promptly declined.
3.4. Reduced sulfur compound production
Six of the most abundant biogenic, reduced sulfur compounds
(RSCs) in the environment are reported to be hydrogen sulfide (H2S),
carbonyl sulfide (COS), methane thiol (or methyl mercaptan, MeSH), di-
methyl sulfide (DMS), carbon disulfide (CS2), and dimethyl disulfide
(DMDS) (Pandey and Kim, 2009; Wardencki, 1998). Originally, there
were twelve RSCs to be analyzed in this study, including H2S, COS,
MeSH, DMS, and CS2. RSC concentrations were anticipated to be low
as those in the environment are typically bppb levels (Pandey and
Kim, 2009; Wardencki, 1998). However by week 4, we switched our
focus on these seven RSCs, H2S, COS, thiofuran, butyl mercaptan, 2-
methylthiophene, 3-methylthiophene, and 2,5-dimethylthiophene, as
their concentrations greatly exceeded the 1 ppmv concentrations in
the calibration standard.
The RSC gas productions are plotted in Fig. 4 and Figs. S2 and S3. Of
the seven RSCs analyzed, H2S was distinct as there was no H2S detected
in the no-MFT controls; only mesocosms containing MFT produced H2S.
This indicates that H2S production in tailings is associated with the mi-
crobial activity in the MFT (Fig. 4a–d). Indeed, PCR results of the dissim-
ilatory sulfite reductase (DsrAB) gene showed that the SRB community
was detected in the MFT samples at week 5 (MFT-PW-0.8%, MFT-SPW-
0.8%, and MFT-SPW-1.5%), and higher band intensity was observed in
the samples with higher concentration of diluent (MFT-SPW-1.5%)
(Fig. 4c). DsrAB gene was also detected in the week 11 MFT samples,
but a lower band intensity was observed for the MFT-SPW with 1.5% dil-
uent, which correspond to the decline observed in the H2S gas produc-
tion data (Fig. 4c). A faint DsrAB band was observed on the week 0 MFT
sample, which suggests that endogenous SRB community exists in
the MFT prior the experiment and that the effect observed is likely
due to the amendment stimulation of the endogenous community.
Several SRB families were identified through the 16S rRNA gene
920 K.F. Gee et al. / Science of the Total Environment 598 (2017) 916–924

Fig. 2. Nitrate (a and b), sulfate (c and d), redox (e and f), and alkalinity (g and h) measurements for the PW and SPW groups over the 11 weeks study period. The measurements for the
MFT containing mesocosms at weeks 1, 2, and 4 represent one measurement. All other values are averages from duplicate measurements and standard deviation bars, where visible, are
plus and minus one standard deviation. Filled circles: PW or SPW + 0.8% diluent; open diamonds: +MFT; open triangle: +MFT / +0.2% diluent, open square: +MFT / +0.8% diluent; x:
+MFT / +1.5% diluent.

amplicon sequencing, including Desulfobulbaceae, Desulfobacteraceae,
Desulfuromonadaceae, and Desulfomicrobiaceae (Fig. 6). The relative
abundance of the Desulfobulbaceae family mirrored the pattern of the
DsrAB gene PCR results (Figs. 4c & 6 and Table S3). The majority of
H2S production occurred within the first six weeks, before CH4 gas pro-
duction began to increase (Figs. 4 and 5). Therefore, week 0 to week 6
were considered the RSC gas production timeframe and were the
focus for RSC analysis. Following elevated SRRs in the first several
weeks (Fig. 3), H2S production in the MFT-diluent amended mesocosms
also increased with increasing diluent concentrations. In the SPW
group, mesocosms amended with 0.2%, 0.8%, and 1.5% w/v diluent re-
sulted in 9.1 μmol, 16.3 μmol, and 19.7 μmol H2S, respectively, an in-
crease of 1.3 to 2.9 times the H2S concentrations found in the no-
diluent control (6.8 μmol). H2S generated in the PW-0.8% w/v diluent
mesocosms resulted in 5.5 μmol H2S after six weeks. Although this
amount is approximately one-third of the H2S found in the SPW-0.8%
w/v diluent treatment, the increase in H2S generation between the
PW-0.8% w/v and no-diluent control (2.1 μmol) yielded a similar factor
 K.F. Gee et al. / Science of the Total Environment 598 (2017) 916–924 921

Fig. 3. Sulfate reduction rate for the (a) PW and (b) SPW groups over the 11 weeks study period. The amount of change in sulfate concentration (μmol SO4/mL of MFT/d) between 0 and 1,
1–2, 2–4, 4–6, and 6–11 weeks were plotted. Note that negative values indicating reduction in sulfate concentration. Open diamonds: PW or SPW + MFT; open triangle: +MFT / +0.2%
diluent, open square: +MFT / +0.8% diluent; x: +MFT / +1.5% diluent.

increase of 2.7. These results further support that diluent stimulates sul-
fate reduction by the SRB and this in turn results in increased H2S gen-
eration from tailings. The higher than anticipated H2S concentrations
in the mesocosm headspace (bppb levels) (Wardencki, 1998) may be
due to several reasons. Salloum et al. (2002) reported that the release
of H2S from tailings may become an issue if the tailings pH is low or if
there is an insufficient amount of metals available to form metal sul-
fides. As the pH of our samples was slightly alkaline throughout the
study, the generation of H2S may be due to the latter. Additionally, our
RSC gas generation estimates are those yielded under anaerobic condi-
tions; when exposed to oxygen, as found in the water cap of tailings
ponds, H2S may be removed by conversion into sulfate (Pisz, 2008).
In contrast to H2S, the other five RSCs measured in this study (ex-
cluding COS, b0.1 μmol throughout the study) 2-methylthiophene, 2-
methylthiophene, 2,5-dimethylthiophene, thiofuran, and butyl mercap-
tan (Fig. S2 and S3) were not only detected in the no-MFT controls at
t = 0 but these concentrations were noticeably higher than the t = 0
production found in the MFT mesocosms. It is likely that these RSCs
originated from the diluent, and is further supported by two other ob-
servations in this study. Firstly, at the end of 14-weeks, the RSC concen-
trations in the MFT mesocosms were similar to the t = 0 concentrations
found in the no-MFT controls. Secondly, there was a significant differ-
ence in the gases generated between the MFT-diluent amended
mesocosms and the no-diluent controls. In the case of H2S, concentra-
tions increased by a factor of 2.9 times in a worst-case diluent scenario
of 1.5% w/v compared to that of no-diluent control. For 2-
methylthiophene, 3-methylthiophene, 2,5-dimethylthiophene, and
thiofuran, the total amount of RSC production increased by a minimal
factor of 4 at 0.2% w/v and reached up to 45 times that of the no-
diluent control emissions at 1.5% w/v. Therefore, the marked difference
in emissions produced between the treatments may be due to these
RSCs originating from the diluent and that there is a greater amount of
diluent in some of the treatments. Regarding the process behind the re-
lease of these RSCs from tailings, one possibility is that there is a sorp-
tion and release mechanism occurring between the RSCs and the MFT.
However, as the partitioning coefficients for the RSCs analyzed in this

Fig. 4. Hydrogen sulfide measurement (upper panel) for the (a) PW and (b) SPW groups over the 11-week study period. Filled circles: PW or SPW + 0.8% diluent; open diamonds: +MFT;
open triangle: +MFT / +0.2% diluent, open square: +MFT / +0.8% diluent; x: +MFT / +1.5% diluent. (c) Dissimilatory sulfite reductase (Dsr) (lower panel) gene was amplified from each
of the duplicate MFT samples (indicated in parenthesis). MFT obtained prior to experiment (week 0) was used as control for both week 5 and 11 PCR. (d) The total reduced sulfur gas
production between week 0 and 6.
922 K.F. Gee et al. / Science of the Total Environment 598 (2017) 916–924

Fig. 5. Methane measurement (upper panel) for the (a) PW and (b) SPW groups over the 11-week study period. Filled circles: PW or SPW + 0.8% diluent; open diamonds: +MFT; open
triangle: +MFT / +0.2% diluent, open square: +MFT /+0.8% diluent; x: +MFT / +1.5% diluent. (c) Methyl-coenzyme M reductase (MCR) (lower panel) gene was amplified from each of
the duplicate MFT samples (indicated in parenthesis). MFT obtained prior to experiment (week 0) was used as control.

study are largely unavailable in literature, further research would be re-
quired to verify whether sorption of these compounds is an important
environmental fate process. Several studies have also found that micro-
bial activity coincides with changes in tailings properties and increased
MFT consolidation rates (Arkell et al., 2015; Bordenave et al., 2010). As
RSC gases generation were the greatest in diluent amended mesocosms
where microbial activity is stimulated, perhaps these changes in tailings
properties also influence the release of RSCs over time. Further research
would be required to determine the role microbial processes may play
in these particular RSCs emissions release.
Overall, the total amount of RSCs produced was higher in MFT
mesocosms with greater concentrations of diluent (Fig. 4). The total
combined RSC production rates between weeks 0 and 6 for the no-
diluent controls, regardless of pond water group, ranged from 0.01–
0.02 μmol RSCs/mL MFT (Fig. 4d). Considering the worst-case diluent
scenario of 1.5% w/v diluent with 400 mL of MFT, approximately
49 μmol total RSCs was produced in the first six weeks. The RSCs contri-
bution to this amount from highest to lowest was H2S and 2-
methylthiophene N 2.5-dimethylthiophene N 3-methylthiophene N
thiofuran N butyl mercaptan N COS, where H2S and 2-methylthiophene
combined contributed to 81% of the gas produced. Based on our results,
further reducing the concentration of residual diluent in the tailings
ponds is anticipated to reduce the amount of RSCs produced in the
tailings.
3.5. Methane and carbon dioxide production
Based on the results of Fedorak et al. (2002), it was hypothesized
that methanogenesis would begin when there was approximately 17–
20 mg/L of sulfate remaining within the tailings samples. Although ini-
tial sulfate concentrations in our samples were N 200 mg/L (between
932 and 1566 μmol), there was small amounts of CH4 (b 205 μmol) de-
tected in all of the MFT mesocosms at t = 0. Given that CH4 was unde-
tected in the no-MFT controls throughout the 11 weeks, the initial CH4
concentrations in the MFT mesocosms were likely due to suppressed
microbial methanogenic activity in the MFT. Following t = 0, sulfate
concentrations immediately declined and became depleted in the
MFT-diluent amended mesocosms around week 2. CH4 production rap-
idly increased after week 5, indicating there was a lag period of approx-
imately 3 weeks between sulfate depletion and the start of
methanogenesis (Fig. 5a & b). The PCR results of the methyl-
coenzyme M reductase (MCR) gene showed that the methanogenic
 K.F. Gee et al. / Science of the Total Environment 598 (2017) 916–924
community was not detected in week 5 except for the MFT-SPW-1.5%
samples (Fig. 5c). The MCR gene was detected in week 11 diluent-
containing samples (MFT-PW-0.8%, MFT-SPW-0.8%, and MFT-SPW-
1.5%), increasing band intensity was observed in the SPW samples
with increased diluent concentration (MFT-SPW-0.8% and − 1.5%)
(Fig. 5c). Two methanogenic families were identified from 16S
rRNA sequencing, Methanoregulaceae and Methanotrichaceae
(Fig. 6). The relative abundance observed for Methanoregulaceae
and Methanotrichaceae was similar to the MCR gene PCR results
(Figs. 5 & 6 and Table S3). Further evidence of this system transition
from sulfur gas production to methanogenesis was denoted by the
stabilization or decline of H2 S as CH4 spiked after week 5 (Figs. 4
and 5). Of the CH4 generated over the 11-week study period, N75%
of CH4 in all MFT mesocosms were generated between week 5 and
week 11. Therefore, this period was considered the CH4 production
timeframe and was the focus of CH4 analysis.
The molecular biology data and gas analysis data showed that the
quantity of CH4 produced was found to increase with increasing diluent
concentrations (Fig. 5a & b). The estimated amount of CH4 produced be-
tween weeks 5 and 11 in the PW no-diluent controls and 0.8% w/v dil-
uent amended mesocosms was 751 μmol and 9928 μmol, respectively.
In the SPW no-diluent controls, 0.2% w/v, 0.8% w/v, and 1.5%w/v diluent
amended mesocosms, there was approximately 455 μmol, 766 μmol,
11,280 μmol, and 16,264 μmol CH4, respectively. In comparison to the
no-diluent controls, the presence of diluent appears to be a greater in-
fluence on the generation of CH4 than that of H2S. CH4 production dou-
bled at 0.2% w/v diluent amended and was up to 36 times the no-diluent
controls at the 1.5% w/v worst-case diluent scenario. CH4 production
Fig. 6. Microbial communities were characterized by 16S rRNA gene with Illumina MiSeq sequencing platform. The relative abundance (%) of microbial communities within the
mesocosms were plotted at family levels.
923
rates for the treatments were also estimated for the week 5 to week
11 timeframe. There was little difference between CH4 production
rates in the no-diluent controls and the SPW-0.2% w/v diluent treat-
ment, with rates ranging from 1.1 to 1.9 μmol CH4/mL MFT. Rates
were markedly higher in the 0.8% w/v diluent treatments at 24.8–
28.2 μmol CH4/mL MFT (PW-SPW) and 1.5% w/v diluent treatments at
40.7 μmol CH4/mL MFT.
Regarding CO2 generation, the majority of CO2 (N 55%) in the MFT
mesocosms was also produced after week 5 (Fig. S1c & d). Between
weeks 5 and week 11, the amount of CO2 produced in the PW no-
diluent controls and 0.8% w/v diluent treatments was approximately
682 μmol and 1259 μmol CO2, respectively (Fig. S1c). In the SPW no-
diluent controls, 0.2% w/v, 0.8% w/v, and 1.5% w/v diluent treatments,
approximately 549 μmol, 609 μmol, 1760 μmol, and 2361 μmol CO2 re-
spectively, was produced (Fig. S1d). Small amounts of CO2 were present
in the mesocosms at t = 0 (up to 249 ± 122 μmol) due to the anaerobic
chamber gas mixture where the mesocosms were assembled. However
this amount is smaller than the maximum expected amount of CO2 in
the mesocosm headspace at t = 0 (687 μmol), indicating that the in-
crease in CO2 production in the MFT mesocosms were likely due to mi-
crobial activity.
Similarly with CH4, CO2 production appeared to increase with in-
creasing diluent concentrations but to a lesser extent. CO2 production
increased by a minimum factor of 2 times the no-diluent controls at
0.8% w/v diluent amended and increased by a factor of 4 at the worst
case 1.5% w/v diluent level (Fig. S1c & d). CO2 production rates followed
similarly at 1.4 to 1.7 μmol CO2/mL MFT in the no-diluent controls and
0.2% w/v diluent treatments, 3.1 to 4.4 μmol CO2/mL MFT (PW-SPW)
924 K.F. Gee et al. / Science of the Total Environment 598 (2017) 916–924
at the 0.8% w/v treatments, and a maximum CO2 production rate
of 5.9 μmol CO2/mL MFT in a worse-case scenario diluent level of
1.5% w/v. Interestingly, the quantities of CH4 found in this study rela-
tive to CO2 was opposite to that reported (Small et al., 2015). Small
et al. (2015) compiled CH4 and CO2 emissions data obtained from
flux chamber measurements by oil sands companies between 2010
and 2011, and found that the quantities of CO2 emissions from tailings
ponds typically exceed CH4 emissions. Several potential factors that led
to this difference in results between studies may include difference in
solubility between CO2 and CH4 (Arkell et al., 2015), the types of me-
thanogenic microorganisms active in the tailings samples (Siddique
et al., 2011), and our gas generation estimates were produced
under strictly anaerobic conditions. Additionally, within the water
cap of tailings ponds there is evidence that CH4 may consumed by
methanotrophs, therefore lowering CH4 emissions relative to CO2
before release to the atmosphere (Saidi-Mehrabad et al., 2013).
4. Conclusion
Taken together, our data suggested there were three distinct stages
that occurred in the MFT mesocosms over the 11 weeks: 1) Nitrate re-
duction prior to t = 0; 2) High SRRs, followed by RSCs production
between weeks 0 to week 6; and 3) Methanogenesis from weeks 5
to week 11. In general, gases production collectively increased at
higher diluent concentrations, and the genetic data presented here im-
plicate two of the anaerobic microbial communities' involvement
(methanogen and SRB). Furthermore, the microbial activities, influ-
enced by the presence of diluent, could lead to the potential release of
RSC under anoxic conditions. Finally, our results demonstrate that
there is potential to reduce gas productions from tailings by further re-
ducing the amount of diluent being lost to the tailings ponds. However,
further research is required to determine to what extent these gas gen-
eration estimates, which are produced under anaerobic conditions, are
released to the atmosphere after exposure to aerobic conditions.
Acknowledgements
This investigation was supported by funds from NSERC-CRD
(CRDPJ386934-09), OSTRF for partnership support, and COSIA for pro-
viding materials. We would also like to thank Dena Cologgi, Emily Cao,
Bin Ma, and Christine Hereygers for their technical assistance.
Appendix A. Supplementary data
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.scitotenv.2017.04.107.
References
Allen, E.W., 2008. Process water treatment in Canada's oil sands industry: I. Target pollut-
ants and treatment objectives. J. Environ. Eng. Sci. 7, 123–138.
Arkell, N., Kuznetsov, P., Kuznetsova, A., Foght, J.M., Siddique, T., 2015. Microbial metabo-
lism alters pore water chemistry and increases consolidation of oil sands tailings.
J. Environ. Qual. 44, 145–153.
Bates, T.S., Lamb, B.K., Guenther, A., Dignon, J., Stoiber, R.E., 1992. Sulfur emissions to the
atmosphere from natural sources. J. Atmos. Chem. 14, 315–337.
BGC Engineering Inc., 2010. Review of Reclamation Options for Oil Sands Tailings
Substrates.
Bordenave, S., Kostenko, V., Dutkoski, M., Grigoryan, A., Martinuzzi, R.J., Voordouw, G.,
2010. Relation between the activity of anaerobic microbial populations in oil sands
tailings ponds and the sedimentation of tailings. Chemosphere 81, 663–668.
Boudens, R., Reid, T., VanMensel, D., Prakasan, M.R.S., Ciborowski, J.J., Weisener, C.G., 2016.
Bio-physicochemical effects of gamma irradiation treatment for naphthenic acids in
oil sands fluid fine tailings. Sci. Total Environ. 539, 114–124.
Burkus, Z., Wheler, J., Pletcher, S., 2014. GHG Emissions from Oil Sands Tailings Ponds:
Overview and Modelling Based on Fermentable Substrates.
Cao, E.S.R., Olubodun, A., Burns, T., Ulrich, A., 2014. Bioamendment of oil sands mature
fine tailings accelerated dewatering and improved release water quality. IOSTC
(Lake Louise, AB).
Chalaturnyk, R.J., Scott, J.D., Ozum, B., 2002. Management of oil sands tailings. Pet. Sci.
Technol. 20, 1025–1046.
Chen, M., Walshe, G., Fru, E.C., Ciborowski, J.J.H., Weisener, C.G., 2013. Microcosm assess-
ment of the biogeochemical development of sulfur and oxygen in oil sands fluid fine
tailings. Appl. Geochem. 37, 1–11.
Chi Fru, E., Chen, M., Walshe, G., Penner, T., Weisener, C., 2013. Bioreactor studies predict
whole microbial population dynamics in oil sands tailings ponds. Appl. Microbiol.
Biotechnol. 97, 3215–3224.
Fedorak, P.M., Coy, D.L., Salloum, M.J., Dudas, M.J., 2002. Methanogenic potential of tail-
ings samples from oil sands extraction plants. J Microbiol]–>Can. J. Microbiol. 48,
21–33.
Flemming, H.C., Wingender, J., 2001. Relevance of microbial extracellular polymeric sub-
stances (EPSs)—part I: structural and ecological aspects. Water Sci. Technol. 43, 1–8.
Government of Alberta DoE, 2013. Oil Sands Tailings.
Government of Alberta DoE, 2014. Oil Sands: Facts and Statistics. 2016.
Liebensteiner, M.G., Tsesmetzis, N., Stams, A.J., Lomans, B.P., 2014. Microbial redox pro-
cesses in deep subsurface environments and the potential application of (per) chlo-
rate in oil reservoirs. Front. Microbiol. 5, 428.
Pandey, S.K., Kim, K.H., 2009. A Review of methods for the determination of reduced sul-
fur compounds (RSCs) in air. Environ. Sci. Technol. 43, 3020–3029.
Penner, T.J., Foght, J.M., 2010. Mature fine tailings from oil sands processing harbour di-
verse methanogenic communities. J Microbiol]–>Can. J. Microbiol. 56, 459–470.
Pisz, J., 2008. Characterization of Extremophilic Sulfur Oxidizing Microbial Communities
Inhabiting the Sulfur Blocks Of Alberta's Oil Sands. University of Saskatchewan
Saskatoon.
Ramos-Padron, E., Bordenave, S., Lin, S., Bhaskar, I.M., Dong, X., Sensen, C.W., et al., 2011.
Carbon and sulfur cycling by microbial communities in a gypsum-treated oil sands
tailings pond. Environ. Sci. Technol. 45, 439–446.
Saidi-Mehrabad, A., He, Z., Tamas, I., Sharp, C.E., Brady, A.L., Rochman, F.F., et al., 2013.
Methanotrophic bacteria in oilsands tailings ponds of northern Alberta. ISME J. 7,
908–921.
Salloum, M.J., Dudas, M.J., Fedorak, P.M., 2002. Microbial reduction of amended sulfate in
anaerobic mature fine tailings from oil sand. Waste Manag. Res. 20, 162–171.
Siddique, T., Fedorak, P.M., Foght, J.M., 2006. Biodegradation of short-chain n-alkanes in
oil sands tailings under methanogenic conditions. Environ. Sci. Technol. 40,
5459–5464.
Siddique, T., Fedorak, P.M., MacKinnon, M.D., Foght, J.M., 2007. Metabolism of BTEX and
naphtha compounds to methane in oil sands tailings. Environ. Sci. Technol. 41,
2350–2356.
Siddique, T., Penner, T., Semple, K., Foght, J.M., 2011. Anaerobic biodegradation of longer-
chain n-alkanes coupled to methane production in oil sands tailings. Environ. Sci.
Technol. 45, 5892–5899.
Siddique, T., Penner, T., Klassen, J., Nesbo, C., Foght, J.M., 2012. Microbial communities in-
volved in methane production from hydrocarbons in oil sands tailings. Environ. Sci.
Technol. 46, 9802–9810.
Small, C.C., Cho, S., Hashisho, Z., Ulrich, A.C., 2015. Emissions from oil sands tailings ponds:
Review of tailings pond parameters and emission estimates. J. Pet. Sci. Eng. 127,
490–501.
Springer, E., Sachs, M.S., Woese, C.R., Boone, D.R., 1995. Partial gene sequences for the a
subunit of methyl-coenzyme M reductase (mcrI) as a phylogenetic tool for the family
Methanosarcinaceae. J Syst Bacteriol]–>Int. J. Syst. Bacteriol. 45, 554–559.
Stasik, S., Wendt-Potthoff, K., 2014. Interaction of microbial sulphate reduction and
methanogenesis in oil sands tailings ponds. Chemosphere 103, 59–66.
Stasik, S., Loick, N., Knoller, K., Weisener, C., Wendt-Potthoff, K., 2014. Understanding bio-
geochemical gradients of sulfur, iron and carbon in an oil sands tailings pond. Chem.
Geol. 382, 44–53.
Vasconcellos, S.P., Dellagnezze, B.M., Wieland, A., Klock, J.H., Santos Neto, E.V., Marsaioli,
A.J., et al., 2011. The potential for hydrocarbon biodegradation and production of ex-
tracellular polymeric substances by aerobic bacteria isolated from a Brazilian petro-
leum reservoir. J Microbiol Biotechnol]–>World J. Microbiol. Biotechnol. 27,
1513–1518.
Wagner, M., Roger, A.J., Flax, J.L., Brusseau, G.A., Stahl, D.A., 1998. Phylogeny of dissimila-
tory sulfite reductases supports an early origin of sulfate respiration. J. Bacteriol. 180,
2975–2982.
Wardencki, W., 1998. Problems with the determination of environmental sulphur com-
pounds by gas chromatography. J. Chromatogr. A 793, 1–19.