Sunteți pe pagina 1din 8

Water Research 124 (2017) 513e520

Contents lists available at ScienceDirect

Water Research
journal homepage:

The role of sulfate in aerobic granular sludge process for emerging

sulfate-laden wastewater treatment
Weiqi Xue a, Tianwei Hao a, b, *, Hamish R. Mackey c, Xiling Li a, Richard C. Chan a,
Guanghao Chen a, d, e, f
Department of Civil & Environmental Engineering, The Hong Kong University of Science and Technology, Hong Kong, China
Institute for Advanced Study, Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong
Division of Sustainable Development, College of Science and Engineering, Hamad Bin Khalifa University, Qatar Foundation, Doha, Qatar
Water Technology Center, The Hong Kong University of Science and Technology, Hong Kong, China
Hong Kong Branch of Chinese National Engineering Research Center for Control & Treatment of Heavy Metal Pollution, The Hong Kong University of
Science and Technology, Hong Kong, China
Wastewater Treatment Laboratory, FYT Graduate School, The Hong Kong University of Science and Technology, Nansha, Guangzhou, China

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

Article history: Sulfate-rich wastewaters pose a major threat to mainstream wastewater treatment due to the unpre-
Received 6 April 2017 ventable production of sulfide and associated shift in functional bacteria. Aerobic granular sludge could
Received in revised form mitigate these challenges in view of its high tolerance and resilience against changes in various envi-
1 August 2017
ronmental conditions. This study aims to confirm the feasibility of aerobic granular sludge in the
Accepted 3 August 2017
Available online 4 August 2017
treatment of sulfate containing wastewater, investigate the impact of sulfate on nutrient removal and
granulation, and reveal metabolic relationships in the above processes. Experiments were conducted
using five sequencing batch reactors with different sulfate concentrations operated under alternating
Sulfate metabolism
anoxic/aerobic condition. Results showed that effect of sulfate on chemical oxygen demand (COD)
Nutrients removal removal is negligible, while phosphate removal was enhanced from 12% to 87% with an increase in
Granulation sulfate from 0 to 200 mg/L. However, a long acclimatization of the biomass (more than 70 days) is needed
Quorum sensing at a sulfate concentration of 500 mg/L and a total deterioration of phosphate removal at 1000 mg/L. Batch
tests revealed that sulfide promoted volatile fatty acids (VFAs) uptake, producing more energy for
phosphate uptake when sulfate concentrations were beneath 200 mg/L. However, sulfide detoxification
became energy dominating, leaving insufficient energy for Polyhydroxyalkanoate (PHA) synthesis and
phosphate uptake when sulfate content was further increased. Granulation accelerated with increasing
sulfate levels by enhanced production of N-Acyl homoserine lactones (AHLs), a kind of quorum sensing
(QS) auto-inducer, using S-Adenosyl Methionine (SAM) as primer. The current study demonstrates in-
teractions among sulfate metabolism, nutrients removal and granulation, and confirms the feasibility of
using the aerobic granular sludge process for sulfate-laden wastewaters treatment with low to medium
sulfate content.
© 2017 Elsevier Ltd. All rights reserved.

1. Introduction alternative water source for municipal non-potable water purposes

(e.g., cooling, toilet flushing, etc.) is a preferable approach in view of
Successful water management is vital for the development of energy efficiency and brine discharge (Grant et al., 2012). For
coastal cities as the number of people who live within 200 kms of instance, Hong Kong has implemented seawater toilet flushing for
the coast are estimated to account for 75% of the world's population nearly 60 years, achieving a 22% reduction in freshwater demand
by the end of 2050 (Gladwin, 2008). The use of seawater as an (Lu et al., 2011). Extensive application of seawater as an alternative
water resource in Hong Kong results in 1.7 million cubic meters/day
saline wastewater containing chloride and sulfate concentrations of
* Corresponding author. Department of Civil & Environmental Engineering, The
5000 and 550 mg/L on average, respectively (Ding, 2010). Both
Hong Kong University of Science and Technology, Hong Kong, China. salinity and sulfide generation caused by sulfate reduction pose
E-mail address: (T. Hao). noteworthy operational problems and challenges to conventional
0043-1354/© 2017 Elsevier Ltd. All rights reserved.
514 W. Xue et al. / Water Research 124 (2017) 513e520

activated sludge processes when treating these brackish or saline 10 g/L. Sludge retention time (SRT) was not deliberately controlled
wastewaters. In light of these considerations, aerobic granular through sludge wastage but was naturally stable at around 35 days
sludge (AGS), a rapidly growing wastewater treatment technology, due to natural washout of biomass in the effluent. All reactors were
might provide a suitable solution for saline/brackish wastewater operated for 130 days, during which hydraulic retention time (HRT)
treatment in coastal regions in view of its excellent settling-ability, was 8 h and temperature was controlled at 23 ± 1  C.
high biomass retention, reactor compactness, and high resilience
against variations in pollutants and environmental conditions (de 2.2. Feeding
Bruin et al., 2004; Beun et al., 1999; de Kreuk and Van
Loosdrecht, 2006; Pronk et al., 2015). Successful application of Synthetic wastewater was prepared from a stock solution con-
AGS in coastal areas can alleviate space and energy shortage as well taining organics, mainly comprising of acetate and glucose,
as potentially tackle the threats posed by sulfate-laden wastewater. ammonia and phosphate as well as micronutrients and alkalinity.
However, the influence of sulfate on biological nutrients removal Detailed components of the stock solution are shown in Table S1
(particularly phosphate) and sludge granulation is yet to be un- and S2 of Supplementary Information (SI). The stock was diluted
derstood fully, which is necessary for facilitating optimization of with tap water 100 times. Then, additional sodium acetate, NH4Cl,
AGS for sulfate-laden saline/brackish wastewater treatment. K2HPO4 and KH2PO4 were added to set COD, ammonia-N and
For AGS systems, alternative anaerobic/aerobic operation is phosphate concentrations of the influent at 600, 100 and 40 mg/L
commonly applied in a Sequencing Batch Reactor (SBR) to achieve respectively. The ratio of bicarbonate (NaHCO3) to ammonium-
simultaneous removal of organics and nutrients (Beun et al., 1999). nitrogen (N) in each reactor influent was set at 4 in order to
The anaerobic phase enables acidification, hydrolysis, denitrifica- maintain pH at moderate level. The designated sulfate concentra-
tion and phosphate release, while organics removal, nitrification tion for each reactor was achieved by addition of sodium sulfate to
and phosphate uptake occur during the aerobic phase. However, the necessary level.
anaerobic treatment of high sulfate-containing (>500 mg/L)
wastewater inevitably leads to a shift in the functional microbial 2.3. Supplementary experiments
community and change in the metabolic pathways (Guo et al.,
2016). Conventional biological nutrient removal processes are Three supplementary experiments were conducted after 130
threatened by out- competition of sulfate-reducing bacteria (SRB), days of operation. The first experiment (A) aimed to distinguish the
leading to inhibition of ordinary heterotrophic organisms (OHO), impacts of sulfate concentrations from changes in ionic strength on
including phosphate-accumulating organisms (PAO), by sulfide nutrient removal efficiency. In the second (B) and third (C) tests,
generated from sulfate reduction (Muyzer and Stams, 2008; VFA and phosphate uptake were monitored after using 2e4 dini-
Karhadkar et al., 1987; Jin et al., 2013). Recently, sulfate- trophenol to eliminate proton motive force (PMF) (Morley et al.,
associated phosphate removal has been observed in a full-scale 1992) and adding sodium molybdate for sulfate reduction inhibi-
enhanced biological phosphorus removal (EBPR) process tion (de Jesus et al., 2015). In this way, the chain effects imposed by
(Ginestet et al., 2015) and a lab-scale sulfate-driven AGS process sulfate reduction on proton motive force (PMF), volatile fatty acid
named DS-EBPR, consisting of completely different microorgan- (VFA) intake, PHA storage, energy availability, and phosphate up-
isms from that of conventional AGS (Wu et al., 2013). However, the take were investigated. Detailed descriptions of these supplemen-
exact mechanism for phosphate removal under high sulfate content tary experiments are supplied in the Supplementary information
as well as the necessary process conditions of the DS-EBPR are not (SI).
well understood.
To evaluate the feasibility of applying aerobic granular sludge for 2.4. Analytical methods
sulfate-laden wastewater treatment, the overall role of sulfate in
the AGS process requires a systematic investigation. This work thus Total organic carbon (TOC), ammonia-N and phosphate of the
focuses on studying the interrelations among sulfur metabolism, influent and effluent were measured regularly for reactor perfor-
sludge granulation and nutrients removal, especially phosphate mance monitoring. Batch tests of five reactors were conducted at
removal under alternating anoxic/aerobic conditions. In- day 60, 90 and 120. In these tests, TOC, ammonia-N, phosphate,
vestigations of five lab-scale granular sludge systems with different nitrate, nitrite, sulfate and sulfide concentrations were analyzed
sulfate concentrations was conducted by continuous examination according to the Standard Methods (APHA, 2005). Detailed
of intermediates, internal storage compounds and internal electron analytical procedures are provided in the SI.
shuttles generated during the operation. From the above study, key Polyhydroxyalkanoate (PHA), including poly-b-hydroxybutyrate
metabolic relationships among sulfate metabolism, nutrient (PHB) and poly-b-hydroxyvalerate (PHV), was quantified by gas
removal and granule formation are elucidated. chromatography mass spectrometry (GC-MS) (Agilent 7890A-
5975C) following Wu et al. (2013). Particle size distribution was
2. Materials and methods determined using laser diffraction particle size analyzer (LSI3 320,
Beckman Coulter). MLSS, Mixed Liquid Volatile Suspended Solid
2.1. Reactor setup and operational condition (MLVSS), and Sludge Volume Index at 5 and 30 min (SVI5 & SVI30)
were measured periodically according to the Standard Methods
Five lab-scale air-lift SBRs, each having a working volume of 1.1 L (APHA, 2005). Extracellular Polymeric Substances (EPSs) were
(100 cm in height and 5 cm in diameter), were operated in parallel extracted and analyzed for polysaccharides and protein quantity
with different influent sulfate concentrations of 0, 50, 200, 500 and according to Liu and Fang (2002). Polysaccharides components of
1000 mg/L respectively. The operation cycle of these reactors were EPSs were analyzed using Fourier transformation infrared spec-
identical, consisting of 10 min feeding, 100 min anoxic phase with troscopy (FTIR) (Vertex 70 Hyperion 1000, Bruker). Detailed
dissolved oxygen (DO) concentration of 0.1e0.3 mg/L, 100 min extraction and analysis methods are provided in the SI.
aeration with DO of 3e6 mg/L, 10 min settling, 10 min decant and Adenosine triphosphate (ATP) and zeta-potential of sludge were
10 min idle. All reactors were inoculated with activated sludge measured on day 85 and 110 at the beginning and end of the anoxic
taken from a local saline wastewater treatment plant in Hong Kong phase as well as the end of the entire cycle using an ATP Deter-
at an initial mixed liquid suspended solid (MLSS) concentration of mination Kit (Thermo Fisher) and Zetasizer Nano Z (Malvern, UK),
W. Xue et al. / Water Research 124 (2017) 513e520 515

respectively (Su et al., 2014). N-acyl homoserine lactones (AHLs) For reactor 1 to 3, faster volatile fatty acid (VFA) uptake and
were bio-assayed using C. violaceum CV026 as a bio-detector (Ravn more intensive phosphate release were observed together with
et al., 2001) to investigate how sulfate metabolism linked to sludge increasing sulfide production during anoxic phase. Specific phos-
granulation. Acetyl-CoA and S-Adenosyl Methionine (SAM) quan- phate release to VFA uptake ratio as well as aerobic phosphate
tification in biomass, required for metabolic studies of nutrient uptake increased with increasing sulfate content in these three
removal and granulation, were performed during the batch tests reactors. However, both anaerobic VFA uptake and aerobic phos-
conducted at day 120. Detailed methods for those above mentioned phate uptake rates decreased when sulfate content was larger than
measurements are described in the SI. 200 mg/L, and a sharp decrease was witnessed when sulfate con-
centration equaled to 1000 mg/L. Analysis of PHA content at the
3. Results and discussions end of the anoxic phase and Poly-P content at the end of the aerobic
phase further supported the observed behavior in phosphate
3.1. System performance and batch tests removal capacity as the sulfate content increased from 0 to
1000 mg/L. Saad et al. (2013) reported similar phenomenon con-
Key results of long-term system performance as well as batch cerning sulfide effect on PAOs, i.e. increasing sulfide concentration
tests are shown in Fig. 1. Varying amounts of sulfate concentrations (<100 mg/L) can initially enhance anaerobic phosphate release by
were present in all five reactors, however, no significant difference PAOs for additional energy to drive detoxification but would inhibit
was observed for organics removal, with all reactors achieving phosphate release when total dissolved sulfide exceeded certain
greater than 97% TOC removal over the course of the experiment as value between 100 and 300 mg/L depending on pH. Specific
shown in Fig. S2. The average ammonia-N conversion efficiency phosphate release and uptake were calculated based on measure-
was inversely correlated with sulfate concentration with values of ments of one cycle of all reactors on day 85 as shown in Tables 1 and
53%, 43%, 32%, 32% and 31% in reactor 1 to 5 respectively. However, 2. Phosphate release was recorded at the end of anoxic phase, while
all reactors displayed steady improvement in ammonia-N removal phosphate uptake was recorded at the end of aerobic phase. The
efficiency throughout the study and all achieved higher than 50% PHA quantification was conducted on day 90. Data showed in both
conversion by day 130 as shown in Fig. S3 in the SI. Fluctuations of tables indicate that energy storage for phosphate removal and Poly-
the ammonia-N concentration during day 20e40 resulted from the P formation increased from R1 to R3, and later decreased when
temporary usage of industrial grade ammonium chloride in feed- sulfate content was further increased. pH values of all 5 reactors
ings due to a supply issue. Ammonium chloride with analytical were recorded every 20 min during the batch tests and all followed
grade was resumed afterwards. For all five reactors, no nitrite an initial increase from 6.8 to 7.4e7.6 in the anoxic phase and
accumulation was detected at the end of aerobic phase at any point decline to around neutral during the aerobic phase. This similar
throughout the study. behavior was due to the sufficient buffering by bicarbonate
Long-term average phosphate removal efficiency, calculated (238 mg/L as CaCO3 in the influent).
from day 45 onwards when biomass concentrations were stable Long-term performance and batch tests indicate that beneath a
(~3500 mg VSS/L), were 12%, 28%, 87%, 59% and 1.3% for reactors 1 critical sulfate concentration (200 mg/L in this study), VFA uptake,
to 5, as shown in Fig. 1(a). The influence of sulfate concentration on PHA formation, poly-P storage and phosphate uptake were inten-
phosphate removal was much more significant and complicated. At sified with increasing sulfate concentration. Further increase in
low sulfate concentrations (<200 mg/L), phosphate removal was sulfate concentration beyond the level slowed down and inhibited
enhanced from 12% to 87% with an increase in sulfate concentration these processes. To investigate whether these observations were
from 0 to 200 mg/L. However, inhibition of phosphate removal associated with changes in ionic strength, supplementary experi-
efficiency was observed when sulfate concentrations exceeded ment A was performed at the end of the study by replacing sodium
200 mg/L. More than 70 days acclimatization was required to reach sulfate with sodium chloride while maintaining the same ionic
phosphate removal efficiency of 76% and became stable with sulfate strengths for each reactor as created by sodium sulfate before. In
concentration of 500 mg/L. Reactor 5 under the highest sulfate the presence of varying sodium chloride concentrations, all re-
concentration showed that phosphate concentration in the effluent actors showed similar net phosphate removal efficiency of around
was sometimes higher than that of influent, indicating no net 25% (as shown in Fig. S4), indicating that former differences
aerobic phosphate uptake in some cycles (Fig. 1(b)). Such behavior observed in phosphate removal among the five reactors were
was also confirmed in batch tests afterwards. These results imply associated with sulfate concentration rather than ionic strength.
that at sulfate concentrations typical of seawater toilet flushing
(500 mg/L), treatment performance will be inhibited compared 3.2. Phosphate removal influenced by sulfate metabolism
with low sulfate EBPR, but still allow plants to meet or be close to
typical marine discharge standards (Water Pollution Control In a normal phosphate removal cycle, microorganisms degrade
Ordinance, Cap 358, section 21). their intracellular Poly-P into inorganic phosphorous (P) under
Such differences in phosphate removal were attributed to sul- anaerobic conditions, resulting in phosphate release, which is
fate because: (1) the concentrations of other salts/ions that may coupled with ATP generation. Meanwhile, VFA uptake takes place
affect phosphate removal are exactly the same among all reactors driven by proton motive force (PMF). Subsequently, the generated
(influence of Cl on phosphate removal was proved to be negligible ATP will be used to convert VFA into acetyl-CoA. NADH will be
by supplementary experiment A); and (2) chemical precipitation of produced after partial oxidation of Acetyl-CoA in the TCA cycle
phosphate is only dominant at high pH (>8.5) (Man ~ as et al., 2012), (Minoa et al., 1998) to supports the synthesis of PHA (McMahon
which is much higher than the pH in the current system. To further et al., 2007), PHA then acts as the energy supplier in the aerobic
analyze how sulfate affects phosphate removal, batch tests were phase for both the phosphate inorganic transport (Pit) system and
conducted to trace the changes of sulfide, sulfate, phosphate, and phosphate-specific transport (Pst) system (Willsky and Malamy,
organic carbon (TOC) once a month from day 60. Ammonia-N, ni- 1980) that are responsible for phosphate uptake afterwards.
trite and nitrate were also measured throughout the batch tests Therefore, anaerobic and aerobic metabolisms are closely linked
since accumulation of nitrite is also detrimental to phosphate with energy balance.
removal (Meinhold et al., 1999). Fig. 1(b) shows the results of one It is known that sulfide, especially hydrogen sulfide, is toxic at
typical batch test. certain concentrations to microorganisms. Hence, sulfide needs to
516 W. Xue et al. / Water Research 124 (2017) 513e520

Fig. 1. (a) Long-term phosphate removal performance of the five SBRs, and (b) Batch test profiles of the five reactors based on the average of three tests conducted at days 60, 90 and
120 (error bars are the standard deviation from the three batch tests).
W. Xue et al. / Water Research 124 (2017) 513e520 517

be transported from inside of the cell to outside, which is coupled current study as proven by measured changes in sulfate and sulfide
with proton translocation (Dworkin et al., 2006). Such trans- concentration. The sulfur conversion intensity to maintain phos-
portation, termed as sulfide detoxification, is an energy consuming phate removal was 16.7e166.7 mgS/L in this study, while the most
process (Larry, 1995), which involves ATP utilization and poly-P favorable intensity was 75 mgS/L (R3).
degradation when ATP is insufficient. Saad et al. (2013) suggested Zeta-potential measurement revealed that the zeta-potential of
that this energy consuming detoxification is the reason responsible granules increased from 10.8 to 4.6 mV at the end of anoxic
for the increased release of phosphate under anaerobic phase by stage with increasing sulfate concentration (Fig. S6). Such behavior
PAOs when exposed to high sulfide content. However, no sub- provides support of increased proton transportation occurring in
stantial experimental evidence is available to prove that enhanced order to remove intracellular sulfide. However, cations, like Mg2þ
removal of phosphate is due to the increase in sulfide content when and Kþ, are produced during Poly-P degradation, which also
beneath a critical level. contribute to the elevation of zeta-potential. Therefore, supple-
In this experiment, ATP as well as zeta-potential of mature mentary experiments B and C, in which the PMF and sulfate
granules were measured as shown in the SI (Fig. S5 & S6). In the reduction were inhibited were conducted. After elimination of PMF
anoxic phase, there was a net generation of ATP in reactor 1 and 5, using 2-4-dinitrophenol, a similar VFA uptake was observed in all
while in the other reactors (2, 3, 4) ATP was consumed. It was five reactors (Fig. S7), supporting the hypothesis that the difference
observed that ATP consumption increased as sulfate concentration in PMF caused by sulfide detoxification influenced VFA uptake and
increased from 0 to 200 mg/L in the aerobic phase and the highest subsequently phosphate removal. In supplementary experiment C,
average consumption of ATP occurred in reactor 3 with sulfate where sulfate reduction was inhibited by sodium molybdate
concentration of 200 mg/L. The ATP utilization is partially attrib- (20 mg/L), VFA uptake and phosphate removal were diminished in
uted to phosphate uptake, for instance in reactor 3, an average 87% all reactors (Fig. S8), confirming that sulfate reduction (i.e. presence
phosphate removal was shown accompanied by the sharp decrease of sulfide) was influential in promoting VFA uptake. These results in
in ATP during the aerobic phase. In contrast, less aerobic ATP con- combination with zeta-potential measurements illustrate that
sumption was observed in reactors 4 and 5. sulfate reduction enhances VFA uptake through proton trans-
In the anoxic phase, sulfide is generated within the core of location and increased PMF that is coupled with sulfide detoxifi-
granules due to the anaerobic conditions caused by the layered cation. The slight decrease in VFA uptake and sharp decrease
structure of the granule. This leads to an increased requirement for observed in reactors 4 and 5 under normal (uninhibited) conditions
sulfide detoxification coupled with proton pumping. In reactors 2 is probably related to the inhibition of intracellular carbon utiliza-
and 3 under higher sulfate concentrations, partial ATP was allo- tion (organic carbon cannot be converted into PHA due to the lack
cated to drive transportation of sulfide out of the cell. Therefore, of ATP which supports conversion of carbon into Acetyl-CoA) as
both cross-cell sulfide transportation and phosphate uptake demonstrated by slight and a shape reduction of Acetyl-CoA con-
contribute to the higher ATP consumption during the anoxic stage tent measured in R4 and R5 (SI, Table S3).
in reactors 2 and 3. In addition, the higher amounts of PHA syn-
thesized (as shown later) also consumed more ATP. Meanwhile, 3.3. Sulfate metabolism and granulation
hydrogen ions will be pumped out together with sulfide. These
hydrogen ions will create a proton motive force (PMF) which drives Particle size distribution and SVI of sludge in all reactors were
VFA uptake (Oehmen et al., 2007). Therefore, increasing sulfide periodically detected during the entire study as shown in Fig S9. de
concentration within a certain limit leads to a higher PMF, and Kreuk et al. (2007) suggested that two fundamental requirements
results in more VFA uptake and PHA storage that is later used for should be satisfied in judging the formation of granular sludge: 1)
phosphate uptake. However, if the sulfate concentration is too high mature granules should have a minimum particle size of 200 mm;
(500 mg/L and beyond as is the case in this study), more ATP must and 2) the SVI5/SVI30 ratio of mature granules should be close to 1.
be used to release sulfide out of the cell, leaving less or even Sludge in the five reactors took 66, 45, 30, 20 and 20 days,
insufficient ATP for PHA synthesis, which is fatal for phosphate respectively, to fulfill both requirements and achieve complete
uptake. The above phenomenon is depicted in Fig. 2. Therefore, granulation. Hence, the granulation process was accelerated by the
sulfide, which comes from sulfate reduction, influences the phos- presence of sodium sulfate in this study. Additionally, particle size
phate uptake process by affecting the energy allocation between of mature granules increased from 230 to 1050 mm with increasing
sulfate and phosphate metabolism. Interestingly, an increase in ATP sulfate concentrations by the end of the experiment.
during the aerobic phase was witnessed in reactors 4 and 5, which To investigate the co-relation between sulfate metabolism and
might be due to: 1) the presence of VFA throughout the whole cycle, bio-aggregation, processes through which sulfate is utilized were
which is responsible for continuous ATP generation in aerobic explored. The first step in sulfate reduction is conversion of sulfate
phase; 2) the inhibition of phosphate uptake capacity of PAOs by into adenosine phosphosulfate (APS) with the consumption of in-
high concentration of sulfide; and 3) occurrence of autotrophic ternal ATP. Later, APS will be converted into Cysteine (Cys) and
denitrification using sulfide as electron acceptor, which is sup- Methionine (Met) in assimilatory sulfate reduction which then
ported by enhanced nitrate loss during anoxic phase indicated in participate in the Activated Methyl Cycle to form diverse sulfur
batch tests and reported by Fajardo et al. (2012). containing chemicals, including S-adenosylmethionine (SAM), S-
Wu et al. (2013) reported a novel Sulfur Conversion-associated adenosylhomocysteine (SAH), S-ribosylhomocysteine (SRH) and
Enhanced Biological Phosphorus Removal (DS-EBPR) system un- homocysteine (Hansen, 1994). Subsequently, a protein encoded by
der alternating anaerobic and anoxic conditions. It is hypothesized LuxI gene utilizes SAM as primer in the synthesis of AHLs, one kind
that various internal polymers associated with energy flow and of quorum sensing (QS) auto-inducer which facilitates aggregation
storage (i.e. polysulfide or elemental sulfur) are synthesized during of microorganisms via quorum sensing (Churchill and Chen, 2011).
the anaerobic stage and are combined with PHA to supply energy As an important intermediate of both sulfate metabolism and
for phosphate uptake later. Analogously, phosphate removal only quorum sensing, endogenous SAM was measured with an ELISA kit
occurs when sulfur conversion intensity is maintained at moderate and AHLs were also determined using bioassay with CV026 as in-
levels (15e40 mgS/L). Both lower and higher conversion intensity dicator (Fig. 3(a)).
causes decline in phosphate removal. However, no polysulfide or Purple pigment production occurs when CV026 senses the
elemental sulfur was generated during the anoxic phase in the presence of AHLs. The diffusion area of AHLs, indicated by the area
518 W. Xue et al. / Water Research 124 (2017) 513e520

Table 1
Specific phosphate release, phosphate uptake and PHA/phosphate release ratios of one typical cycle on day 85 (PHA/phosphate release values are calculated based on PHA data
of day 90 and phosphate release data of day 85).

Phosphate release (mg Phosphate/g VSS/cycle) phosphate uptake (mg Phosphate/g VSS/cycle) PHA/phosphate release (mmol-C/mg Phosphate)

R1 3.3 4.7 0.109

R2 2.7 6.0 0.922
R3 7.9 20.4 0.408
R4 5.2 5.7 0.488
R5 5.5 0.7 0.016

Table 2
PHA quantity (mmol-C/g-MLVSS) at the end of the anoxic phase & calculated Poly-P quantity (mg-P/g-MLVSS) at the end of the aerobic phase.

R1 R2 R3 R4 R5

PHA Poly-P PHA Poly-P PHA Poly-P PHA Poly-P PHA Poly-P

Day 60 0.97 26.0 1.83 44.5 3.36 130 2.07 38.5 0.52 15.0
Day 90 0.36 46.5 2.49 55.2 3.22 169.5 2.33 72.0 0.09 27.0

Fig. 2. Metabolic pathway of microbes in anoxic phase.

of purple pigmentation, has a positive relationship with the con- when bulk sulfate content increased (Fig. S10). Specifically, syn-
centrations of AHLs. Fig. 3(a) shows the result of the bioassay test. thesis of one kind of polysaccharides related with sulfate meta-
With increasing sulfate content, the ratio of pigment covered plate bolism, named sulfated polysaccharides, is enhanced with
area increased from 6.25% to 63% as calculated by ImageJ. This increasing bulk sulfate as indicated by the increasing signal in FTIR
shows that AHLs production has a positive relation with bulk sul- result (820e1250 cm1). The improved production of EPS also fa-
fate concentration. The SAM quantity also demonstrated a positive cilitates granule formation and keeps the stable structure of mature
relation with sulfate concentration, as shown in Fig. 3(b). The granule.
synthesis of APS and SAM are both ATP consuming bioprocesses,
which may also result in limited ATP resources for phosphate up-
4. Conclusions
take under high sulfate conditions. Consequently, quorum sensing
(QS) based on AHLs enhance the aggregation of microorganisms to
In the current AGS system consisting of sulfur metabolic mi-
form granules as shown in Fig. 3(c).
croorganisms, the presence of sulfate accelerated the granulation
Detailed EPS analysis, in which quantities of protein and poly-
process by stimulating the formation of AHLs, a group of QS auto-
saccharides as well as their types were analyzed, showed that the
inducers. Sulfate concentration also influenced phosphate
contents of protein and polysaccharides in the sludge EPS increased
removal efficiency via affecting available energy for phosphate
W. Xue et al. / Water Research 124 (2017) 513e520 519

Fig. 3. Sulfate metabolites and granulation mechanism. (a) AHLs bio-assay using CV026 as bio-indicator; (b) SAM quantification; (c) metabolic relations between sulfate metabolism
and granulation.

uptake with poor phosphate removal at low (0 mg/L) and high Appendix A. Supplementary data
(1000 mg/L) concentrations of sulfate and almost 90% removal
achieved in the presence of 200 mg/L sulfate. The increase in Supplementary data related to this article can be found at http://
driving force for VFA uptake, i.e. PMF, caused by proton trans-
portation associated sulfide detoxification and PHA synthesis leads
to the enhancement of phosphate removal when sulfate content is
increased from 0 to 200 mg/L. However, further increase in sulfate
concentration inhibited phosphate removal due to the allocation of APHA, 2005. Standard Methods for the Examination of Water and Wastewater,
the majority of ATP to sulfide detoxification, leaving less energy for twenty-first ed. American Public Health Association/American Water Works
acetate uptake and PHA production. This resulted in lower phos- Association/Water Environment Federation, Washington. ISBN - 10:
phate removal and long acclimatization of the biomass (more than Beun, J.J., Hendriks, A., van Loosdrecht, M.C.M., Morgenroth, E., Wilderer, P.A.,
70 days) at a sulfate concentration of 500 mg/L and a total deteri- Heijnen, J.J., 1999. Aerobic granulation in a sequencing batch reactor. Water Res.
oration of phosphate removal at 1000 mg/L. Based on the above- 33 (10), 2283e2290.
Churchill, M.E.A., Chen, L.L., 2011. Structural basis of acyl-homoserine lactone-
mentioned metabolic relationships, the current study
dependent signaling. Chem. Rev. 111 (1), 68e85.
demonstrates the feasibility of using AGS systems consisting of de Bruin, L.M.M., de Kreuk, M.K., van der Roest, H.F.R., Uijterlinde, C., van
sulfate utilizing microbes for sulfate laden wastewater treatment Loosdrecht, M.C.M., 2004. Aerobic granular sludge technology: and alternative
with low to moderate (50e500 mg/L) sulfate content. to activated sludge. Water Sci. Technol. 49 (11e12), 1e7.
de Jesus, E.B., de Andrade Lima, L.R.P., Bernardez, L.A., Almeida, P.F., 2015. Inhibition
of microbial sulfate reduction by molybdate. Braz. J. Petrol. Gas 9 (No 3).
Ding, G.Y., 2010. Detection, Formation and Algal Toxicity of Polar Brominated
Acknowledgments Disinfection Byproducts in Chlorinated Saline Sewage Effluents. Thesis, Ph.D.
Hong Kong University of Science and Technology.
Dworkin, M., Falkow, S., Rosenberg, E., Schleifer, K.H., Stackebrandt, E., 2006. The
This work was supported by Chinese National Engineering Prokaryotes. In: Bacteria: Firmicutes, Cyanobacteria, vol. 4.
Research Center (ITC-CNERC14EG03), the National Natural Science de Kreuk, M.K., Kishida, N., van Loosdrecht, M.C.M., 2007. Aerobic granular sludg-
Foundation of China (51638005), Guangdong Provincial Science eestate of the art. Water Sci. Technol. 55 (8e9), 75e81.
de Kreuk, M., Van Loosdrecht, M.C.M., 2006. Formation of aerobic granules with
and Technology Planning Project (grant no. 2016A050503041) and domestic sewage. J. Environ. Eng. 132 (6), 694e697.
the Research Grants Council (RGC) of Hong Kong (Project 6214117). Fajardo, C., Mosquera-Corral, A., Campos, J.L., Me ndez, R., 2012. Autotrophic
520 W. Xue et al. / Water Research 124 (2017) 513e520

denitrification with sulphide in a sequencing batch reactor. J. Environ. Manag. Meinhold, J., Arnold, E., Isaacs, S., 1999. Effect of nitrite on anoxic phosphate uptake
113, 552e556. in biological phosphorus removal activated sludge. Water Res. 33 (8),
Gladwin, T.N., 2008. Doomsday alert: megachallenges confronting urban moder- 1871e1883.
nity. J. Int. Inst. 16, 14e16. Morley, G.E., Anumonwo, J.M., Delmar, M., 1992. Effects of 2,4-dinitrophenol or low
Ginestet, P., Nicol, R., Holst, T., Lebosse , X., 2015. Evidence for sulfide associated ATP on cell excitability and action potential propagation in Guinea pig ven-
autotrophic biological phosphorus removal in a full scale wastewater treatment tricular myocytes. Circ. Res. 71 (4), 821e830.
plant. In: Specialist Conference on Nutrient Removal and Recovery: Moving Minoa, T., van Loosdrecht, M.C.M., Heijnen, J.J., 1998. Microbiology and biochemistry
Innovations into Practice. 18e21 May 2015. of the enhanced biological phosphate removal process. Water Res. 32 (11),
Grant, S.B., Saphores, J.D., Feldman, D.L., Hamilton, A.J., Fletcher, T.D., Cook, P.L.M., 3193e3207.
Stewardson, M., Sanders, B.F., Levin, L.A., Ambrose, R.F., Deletic, A., Brown, R., Muyzer, G., Stams, A.J.M., 2008. The ecology and biotechnology of sulphate-
Jiang, S.C., Rosso, D., Cooper, W.J., Marusic, I., 2012. Taking the “Waste” out of reducing bacteria. Nat. Rev. Microbiol. 6 (6), 441e454.
“wastewater” for human water security and ecosystem sustainability. Science Oehmen, A., Lemos, P.C., Carvalho, G., Yuan, Z.G., Keller, J., Blackall, L.L., Reis, M.A.,
337 (6095), 681e686. 2007. Advances in enhanced biological phosphorus removal: from micro to
Guo, G., Wu, D., Hao, T.W., Mackey, H.R., Wei, L., Wang, H.G., Chen, G.H., 2016. macro scale. Water Res. 41 (11), 2271e2300.
Functional bacteria and process metabolism of the Denitrifying Sulfur Pronk, M., de Kreuk, M.K., de Bruin, B., Kamminga, P., Kleerebezem, R., van
conversion-associated Enhanced Biological Phosphorus Removal (DS-EBPR) Loosdrecht, M.C.M., 2015. Full scale performance of the aerobic granular sludge
system: an investigation by operating the system from deterioration to resto- process for sewage treatment. Water Res. 84, 207e217.
ration. Water Res. 95, 289e299. Ravn, L., Christensen, A.B., Molin, S., Givskov, M., Gram, L., 2001. Methods for
Hansen, T.A., 1994. Metabolism of sulfate-reducing prokaryotes. Antonie Leeu- detecting acylated homoserine lactones produced by Gram-negative bacteria
wenhoek 66 (1), 165e185. and their application in studies of AHL-production kinetics. J. Microbiol.
Jin, R.C., Yang, G.F., Zhang, Q.Q., Ma, C., Yu, J.J., Xing, B.S., 2013. The effect of sulfide Methods 44 (3), 239e251.
inhibition on the ANAMMOX process. Water Res. 47 (3), 1459e1469. Su, B.S., Qu, Z., Song, Y.H., Jia, L.M., Zhu, J.R., 2014. Investigation of measurement
Karhadkar, P.P., Audic, J.M., Faup, G.M., Khanna, P., 1987. Sulfide and sulfate inhibi- methods and characterization of zeta potential for aerobic granular sludge.
tion of methanogenesis. Water Res. 21 (9), 1061e1066. J. Environ. Chem. Eng. 2 (2), 1142e1147.
Larry, B., 1995. Sulfate-reducing Bacteria. Springer Science & Business Media. Saad, S.A., Welles, L., Lopez-Vazquez, C.M., van Loosdrecht, M.C.M., Brdjanovic, D.,
Liu, H., Fang, H.H., 2002. Extraction of extracellular polymeric substances (EPS) of 2013. Sulfide effects on the anaerobic kinetics of phosphorus- accumulating
sludges. J. Biotechnol. 95 (3), 249e256. organisms. In: Conference: 13th World Congress on Anaerobic Digestion, At
Lu, H., Wu, D., Tang, W.T., Chen, G.H., van Loosdrecht, M.C.M., Ekama, G., 2011. Pilot Santiago de Compostella, Spain.
scale evaluation of SANI® process for sludge minimization and greenhouse gas Water Pollution Control Ordinance. CAP 358AK. Technical memorandum standards
reduction in saline sewage treatment. Water Sci. Technol. 63 (10), 2149e2154. for effluents discharged into drainage and sewerage system, inland and coastal
Man ~ as, A., Pocquet, M., Biscans, B., Sperandio, M., 2012. Parameters influencing wasters.
calcium phosphate precipitation in granular sludge sequencing batch reactor. Willsky, G.R., Malamy, M.H., 1980. Characterization of two genetically separable
Chem. Eng. Sci. 77 (1), 165e175. inorganic phosphate transport systems in Escherichia coli. J. Bacteriol. 144 (1),
McMahon, K.D., He, S.M., Flowers, J.J., Kunin, V., Martin, H.G., Hugenholtz, P., 2007. 356e365.
Phosphorus accumulating organisms reveal their secrets a genome level un- Wu, D., Ekama, G.A., Lu, H., Chui, H.K., Liu, W.T., Brdjanovic, D., van
derstanding of enhanced biological phosphorus removal. Proceedings of the Loosdrecht, M.C.M., Chen, G.H., 2013. A new biological phosphorus removal
Water Environment Federation. WEFTEC 2007 Sess. 61 through Sess. 70, process in association with sulfur cycle. Water Res. 47 (9), 3057e3069.
4920e4932 (13).