w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 6 6 6 1 e6 6 6 7

Available online at www.sciencedirect.com

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

Simultaneous heterotrophic and sulfur-oxidizing autotrophic

denitrification process for drinking water treatment: Control
of sulfate production

Erkan Sahinkaya a,*, Nesrin Dursun a, Adem Kilic a, Sevgi Demirel a,

Sinan Uyanik a, Ozer Cinar b
a
Harran University, Environmental Engineering Department, Osmanbey Campus, 63000 Sanliurfa, Turkey
b
Kahramanmaras Sutcu Imam University, Environmental Engineering Department, Kahramanmaras, Turkey

article info abstract

Article history: A long-term performance of a packed-bed bioreactor containing sulfur and limestone was
Received 12 July 2011 evaluated for the denitrification of drinking water. Autotrophic denitrification rate was
Received in revised form limited by the slow dissolution rate of sulfur and limestone. Dissolution of limestone for
21 September 2011 alkalinity supplementation increased hardness due to release of Ca2þ. Sulfate production is
Accepted 28 September 2011 the main disadvantage of the sulfur autotrophic denitrification process. The effluent
Available online 19 October 2011 sulfate concentration was reduced to values below drinking water guidelines by stimu-
lating the simultaneous heterotrophic and autotrophic denitrification with methanol
Keywords: supplementation. Complete removal of 75 mg/L NO3eN with effluent sulfate concentration
Denitrification of around 225 mg/L was achieved when methanol was supplemented at methanol/NO3eN
Sulfur-limestone autotrophic deni- ratio of 1.67 (mg/mg), which was much lower than the theoretical value of 2.47 for
trification heterotrophic denitrification. Batch studies showed that sulfur-based autotrophic NO2eN
Sulfate reduction rate was around three times lower than the reduction rate of NO3eN, which led
Drinking water to NO2eN accumulation at high loadings.
ª 2011 Elsevier Ltd. All rights reserved.

1. Introduction being treated properly. Reverse osmosis, ion exchange,

In many countries, nitrate concentration in ground water has
exceeded the maximum allowable limits for the drinking
water. In the USA, between 10 and 25% of the ground waters
used as a drinking water source has nitrate concentration
above the maximum allowable concentration of 10 mg/L
NO3eN (Liu et al., 2009). Some wells in Harran Plain, Sanliurfa,
Turkey contain nitrate as high as 180 mg/L NO3eN and the
average concentration for whole plain is 35 mg/L NO3eN
(Yesilnacar et al., 2008). The most important sources of nitrate
in ground waters are nitrogen containing fertilizers, and
industrial and domestic wastewaters discharged without
distillation, and electrodialysis are the physico/chemical
methods used for the removal of nitrate from drinking waters.
The main disadvantages of these processes are high opera-
tional cost, low selectivity, and the formation of secondary
brine wastes after treatment. Moreover, these processes are
expensive and not proper for in-situ applications. For these
reasons, biological denitrification should be considered as an
alternative process. Heterotrophic denitrifiers utilize simple
organic compounds as a carbon and energy source. The main
advantages of the process are high denitrification rate and
treatment capacity. However, nitrite accumulates when
organic is stoichiometrically insufficient and organic remains

* Corresponding author. Tel.: þ90 414 344 00 20; fax: þ90 414 344 00 31.
E-mail address: erkansahinkaya@yahoo.com (E. Sahinkaya).
0043-1354/$ e see front matter ª 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.watres.2011.09.056
6662 w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 6 6 6 1 e6 6 6 7
in the effluent when supplemented in excess amount (Liu
et al., 2009 and Sierra-Alvarez et al., 2007). In the practice,
the addition of exactly the right amount of organic is difficult
to achieve (Oh et al., 2001). Organic supplementation is not
required in autotrophic denitrification process, which leads to
low biomass production, decreased risk of bacterial contami-
nation and reduced operational cost. Elemental sulfur is an
attractive source of energy for autotrophic denitrification of
nitrate contaminated ground water (Moon et al., 2008 and
Soares, 2002). Furthermore, elemental-sulfur is non-toxic,
water insoluble, stable under normal conditions, and readily
available (Soares, 2002). In this process, the elemental sulfur
and nitrate act as an electron donor and an acceptor, respec-
tively. Hence, when nitrate is reduced to nitrogen gas, sulfur is
oxidized to sulfate (reaction (1)).
55S0 þ 50NO þ
3 þ 38H2 O þ 20CO2 þ 4NH4 /4C5 H7 O2 N
(1)
þ 55SO2 þ 2.1.
4 þ 25N2 þ 64H
Although sulfur-based autotrophic denitrification has
several advantages, its main disadvantages are sulfate and
acid formation (Park et al., 2002; Liu and Koenig, 2002; Moon
et al., 2008). Lime stone is the most commonly used low-cost
alkalinity source (Liu and Koenig, 2002). The main drawback
of using limestone is the increased hardness of treated water
due to Ca2þ dissolution. In sulfur-based autotrophic denitri-
fication process, 7.54 mg sulfate is formed for per mg NO3eN
removal (reaction (1)). Theoretically, around 33 mg/L NO3eN
(or 150 mg/L NO 3 ) can be denitrified using sulfur-based auto-
trophic denitrification without exceeding sulfate limit value of
250 mg/L, set by US EPA (Oh et al., 2001), if treated water has
not background sulfate. For the waters with higher nitrate
concentrations, heterotrophic and autotrophic denitrification
processes can be combined (mixotrophic process) to control
the sulfate formation (Lee et al., 2001; Liu et al., 2009 and Oh
et al., 2001). During heterotrophic denitrification, 3.57 g
CaCO3 is produced for each gram NO3eN denitrified (reaction
(2)) (Oh et al., 2001), which can be used by sulfur-oxidizing
autotrophic denitrifiers (reaction (1)) in mixotrophic denitrifi-
cation processes. Hence, mixotrophic process requires less
alkalinity compared to autotrophic process. In an effort to
reduce alkalinity requirement of autotrophic denitrification,
mixotrophic denitrification of nitrified industrial effluents has
been studied previously (Kim and Bae, 2000 and Lee et al.,
2001).
NO3 þ 1:08CH3 OH þ 0:24H2 CO3 /0:056C5 H7 NO2 þ 0:47N2
(2)
þ 1:68H2 O þ HCO3
Table 1 e Operational conditions of the column reactors.
Periods 1 2 3 4
Days 0e35 35e52 52e68
NO3eN (mg/L) 50 50 50
HRT (h) 15 8.4 5.6
Loading (g NO3eN/(L.d)) 0.080 0.143 0.214
Methanol (mg/L)a e e e e
a Values in parenthesis shows the methanol concentrations as DOC.
Despite the high denitrification potential of sulfur-based
mixotrophic processes, limited studies have been conducted
for the denitrification of drinking water (Liu et al., 2009;
Soares, 2002). The first study on combined heterotrophic and
sulfur-based autotrophic process for drinking water denitri-
fication was conducted by Liu et al. (2009). In this study, they
used separate reactors for heterotrophic and autotrophic
denitrification processes, which may increase the process
cost. In this context, the present study aims at promoting
simultaneous autotrophic and heterotrophic denitrification
processes (mixotrophic) in one-reactor for drinking water
treatment to decrease sulfate formation and alkalinity
requirement of sulfur-oxidizing denitrification process.
2. Materials and methods
Bioreactor
A laboratory-scale glass column reactor with an empty bed
volume of 350 mL was used. The column reactor was filled
with sulfur (0.5e1 mm), lime-stone (0.5e1 mm) and activated
carbon (1e1.5 mm) particles with equivalent volume of
117 mL. The activated carbon granules were used to improve
the biofilm formation. The use of small sulfur and lime-stone
particles was not to limit the denitrification rate as the
dissolution of sulfur depends on surface area and decreasing
the diameter of sulfur granules would increase the process
performance. The reactor was covered with aluminum foil to
prevent the growth of phototrophic bacteria. A denitrifying
activated sludge obtained from the first anoxic tank of a 5-
stage Bardenpho process located in Harran University
Campus (Sanliurfa, Turkey) was used as inoculum. The
reactor was operated in batch mode for 3 days after inocula-
tion, and then it was operated continuously in up-flow mode
at 28e30  C in a temperature controlled room. In order to
prevent biological activity, the feed container was kept
refrigerated at 4  C. The freshly prepared feed solution was
deoxygenated by passing through the N2 gas for 20 min. Then,
the feed was kept under anaerobic conditions in collapsible
feed containers.
The reactor was fed with tap water supplemented with
50 mg/L K2HPO4 as a source of phosphorus, and different
concentrations of KNO3 to obtain predetermined NO3eN
concentrations. During the study of mixotrophic process,
methanol was supplemented as an external organic carbon at
varying concentrations (15e50 mg/L dissolved organic carbon
(DOC)) (Table 1).
5 6 7 8
68e80 80e122 122e132 132e157 157e190
50 75 75 75 75
11.0 11.0 11.0 11.0 11.0
0.109 0.164 0.164 0.164 0.164
e 37.0 (14) 75 (28) 125 (47)
 w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 6 6 6 1 e6 6 6 7
2.2. Experiments
2.2.1. Continuous tests with column reactor
A laboratory-scale glass column reactor was operated at
different NO3eN, hydraulic retention time (HRT), and auto-
trophic and mixotrophic conditions (Table 1) to evaluate their
effects on denitrification performance, and sulfate produc-
tion. HRT was calculated considering the empty bed volume.
Each operating period shown in Table 1 was changed after
reaching steady-state conditions, which was indicated less
than 10% differences in at least three successive samplings.
The reactor effluent was sampled at least three times
a week for the measurement of NO3eN, NO2eN, sulfate,
sulfide, DOC, pH, and alkalinity. The feed solution was
sampled once a week for the determination of NO3eN, NO2eN,
sulfate, DOC, pH, and alkalinity. Additionally, both the feed
and the effluent of the reactor were sampled once a week for
the measurement of Na, Ca, Mg, and Fe ions, and NH4eN. All
measurements were at least double and the standard devia-
tions were within 5% for NO3eN, NO2eN, sulfate and pH, and
within 10% for DOC, alkalinity, and ion analysis.
In order to promote simultaneous heterotrophic and
autotrophic denitrification, the feed was supplemented with
methanol after period 5. According to reaction (2), 2.47 g
methanol (or 0.93 g methanol as DOC) is required to denitrify
each g of NO3eN. Therefore, the theoretical fraction of NO3eN
denitrified by heterotrophic bacteria under mixotrophic
conditions were 20, 40, and 68%, respectively, for periods 6, 7
and 8 (Table 1).
On day 139, 70 mL (20% of the bed) of the reactor content
was replaced with fresh sulfur particles to improve the
process performance.
The daily gas production was measured using liquid
displacement method and the gas production rate was
compared with the theoretical value using the following
equation (Moon et al., 2008).
Theorethical N2 gas production rate ðml=dÞ
22:4ml Tempð KÞ
¼ removed NO3  N conc:ðmg=LÞ  
28mg 273:15ð KÞ
 Flow rateðL=dÞ
2.2.2. Activity tests
The autotrophic denitrification activity was tested on day 140 in
parallel 150 mL serum bottles at 30  C, filled with 100 mL
medium supplemented with 50 mg/L NO3eN. The serum bottles
were consisted of 3 g sulfur and lime-stone particles. After
addition of sulfur and lime-stone, N2 gas was passed through
the reactors for 5 min to ensure the removal of dissolved
oxygen. Then, the serum bottles were inoculated with the 5 mL
column bed. The initial biomass concentration in the serum
bottles was 127  18 mg volatile suspended solids (VSS)/L.
The serum bottles were sampled at regular intervals for the
measurement of NO3eN, NO2eN, and sulfate. At the end of the
study, biomass was analyzed for organic nitrogen for the
calculation of biomass concentration.
2.2.3. Kinetic tests
The autotrophic denitrification rates were tested in 150 mL
parallel batch serum bottles at 30  C, similar to the activity
6663
tests. Again, 3 g sulfur and 3 g lime-stone particles were added
to the batch bottles. Then, medium containing varying
concentrations of NO3eN (30 mg/L, 80 mg/L, and 100 mg/L)
was added to the batch bottles. Five mL suspension (127  18
mg VSS/L medium) of serum bottles used for activity tests
(Section 2.2.2) were added as inoculum after passing N2 gas for
5 min.
Elemental sulfur is insoluble and apolar mineral, which
may make mass transfer an important rate-limiting factor
(Sierra-Alvarez et al., 2007). In order to avoid the slow disso-
lution of sulfur become rate-limiting in batch assays, sulfur
was added at much higher amount (3 g) than the theoretical
requirement for complete reduction of NO3eN according to
reaction (1). The contents in serum bottles were sampled and
analyzed similar to the activity tests.
2.3. Analytical methods
Samples were filtered using cellulose acetate syringe filters
with pore size of 0.45 mm before the measurements of NO3eN,
NO2eN, sulfate, DOC, and dissolved sulfide in the superna-
tant. NO3eN, NO2eN, and sulfate were measured using ion
chromatography, Schimadzu, Prominence HIC-NS. Sulfide
was analyzed spectrometrically using a Shimadzu UV-1601
Spectrophotometer following the method described by Cord-
Ruwish (1985). Alkalinity was measured according to Stan-
dard Methods (APHA AWWA WEF, 2005) in unfiltered samples
titrated with 0.1 N HCI to a pH 4.5 end point. DOC was
measured using TOC analyzer (Shimadzu, Japan). Organic
nitrogen of biomass at the end of batch tests was measured
according to Standard Methods (APHA AWWA WEF, 2005).
Then, the biomass concentrations in the batch serum bottles
were estimated assuming average cell formula C5H7O2N
(Rittmann and McCarty, 2001). Cations were measured with an
Inductively Coupled Plasma (ICP) combined with atomic
emission. When the standard deviation was larger than the
size of the plotting symbol, the standard deviation is shown by
þ/ error bars.
(3) 3. Results and discussion
3.1. Denitrification and sulfate production in the column
reactor
The laboratory-scale glass column reactor was operated under
autotrophic (periods 1e5) and mixotrophic (periods 6e8)
conditions for around 200 days. In the first three periods, the
feed nitrate concentration was kept at 50 mg/L NO3eN and the
effect of decreasing HRT from 15 h (period 1) to 5.6 (period 3)
on the process performance was investigated (Table 1 and
Fig. 1A). Denitrification was almost complete in the first two
periods, but nitrite accumulated up to 10 mg/L NO2eN in the
3rd period with corresponding denitrification efficiency of
around 80% (Fig. 1A). Nitrite accumulation is a reliable marker
of over loading of sulfur-based denitrification process (Sierra-
Alvarez et al., 2007). The loading rate was increased from
0.080 g NO3eN/(L.d) in period 1 to 0.214 g NO3eN/(L.d) in period
3. The maximum denitrification rate was around 0.20 g N/(L.d)
observed in period 3. Similarly, Soares (2002) observed
6664 w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 6 6 6 1 e6 6 6 7

PERIODS concentrations (Table 1). The fraction of nitrate denitrified by

heterotrophs was determined indirectly from the production
1 2 3 4 5 6 7 8
of sulfate (Oh et al., 2001). Addition of 37 mg/L methanol
NO3-N or NO2-N (mg/L)

80
(14 mg/L as DOC) did not stimulated heterotrophic denitrifi-
60 Feed NO3-N
cation as the sulfate production was similar to the theoretical
Effluent NO3-N
autotrophic sulfate production calculated according to the
40 Effluent NO2-N
reaction (1) (Fig. 1B). Although methanol was almost
20 completely oxidized (Fig. 2), it is surprising not to observe any
A heterotrophic denitrification activity, which should decrease
0 the sulfate production due to the decreased fraction of nitrate
600
denitrified by sulfur-oxidizing autotrophs. This may be inter-
Sulfate (mg/L)

500 preted as the presence of facultative chemolithoautotrophic

400 denitrifiers, such as Thiobacillus versutus, Thiobacillus thyasiris,
Feed
300 Thiosphaera pantotropha and Paracoccus denitrificans as they are
Effluent
200 Theoretical not only able to grow autotrophically, by using reduced sulfur
100
0
B compounds as energy source, but are also capable of hetero-
trophic growth. Therefore, facultative chemolithoautotrophic
9
denitrifiers can apparently adapt to different environments,
8 such as autotrophic, heterotrophic or mixotrophic (Oh et al.,
2001). In period 7, methanol concentration in the feed was
pH

7 increased to 28 mg/L DOC. The increase in methanol

Feed
Effluent concentration neither started heterotrophic denitrification
6
C nor increased the process performance. On day 139, around
300 20% of the reactor content was replaced with fresh sulfur
Alkalinity (mg/CaCO3)

Feed
250 particles thinking that the sulfur dissolution limits the process
Effluent
performance, which resulted in accumulation of nitrate in the
200
effluent. After that, the reactor performance increased sharply
150
and denitrification efficiency approached 100% although
100 hetereotophic denitrification did not start. In period 8, with
50
D the increase of methanol concentration to 47 mg/L DOC,
0 sulfate concentration in the effluent decreased appreciably to
0 50 100 150 200 below the 250 mg/L, which is the maximum level set by US
Day EPA. Hence, the fraction of NO3eN denitrified by heterotrophs

Fig. 1 e Feed and effluent NO3eN, NO2eN, sulfate, alkalinity

and pH variations. The theoretical sulfate concentration PERIODS
was calculated according to reaction (1) assuming that all
NO3eN and NO2eN were autotrophically removed. A 60
1 2 3 4 5 6 7 8

50 Influent
Effluent
DOC (mg/L)

40
denitrification rate of up to 0.20 g N/(L.d) at a HRT of 1.0 h and
NO3eN loading of 0.24 g NO3eN/(L.d). Lee et al. (2001) obtained 30
much higher removal rate (around 5.0 g N/(L.d)) during 20
simultaneous heterotrophic and sulfur-utilizing autotrophic
10
denitrification for nitrified leachate containing 700e900 mg/L
NO3eN. The limited dissolution of sulfur and/or lime-stone 0
B 120
may limit the denitrification rate as will be discussed later.
Gas Production (mL/day)

100 Measured
In period 4, in order to recover the process performance Theorethical
(i.e., the process without nitrite accumulation); HRT was 80
increased back to 11.0 h. In period 5, the feed nitrate 60
concentration was increased to 75 mg/L NO3eN. In this period, 40
the effluent NO3eN and NO2eN increased over 15 mg/L and
20
the approximate denitrification efficiency was 75% (Fig. 1A).
Under autotrophic conditions (between periods 1 and 5), the 0
measured sulfate concentrations are in good agreement with
0 50 100 150 200
the theoretical sulfate production calculated according to
Time (day)
reaction (1) (Fig. 1B).
After period 5, the reactor was operated under mixotrophic Fig. 2 e The variations of feed and effluent DOC
conditions and heterotrophic denitrification was stimulated concentrations (A), and produced and theoretical gas
by the addition of methanol to the feed at various production rates (B).
 w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 6 6 6 1 e6 6 6 7 6665

was about 60% in period 8. According to reaction (2), the Due to the presence of nitrite in the middle of the reactor,
theoretical nitrate that can be removed by the addition of sulfide was only detected at the effluent of the reactor. The
47 mg/L DOC is around 50 mg/L NO3eN, corresponding to produced sulfide may be oxidized in the reactor with nitrate as
about 67% denitrification by heterotrophic denitrifies. Hence, the electron acceptor. Similarly, Moon et al. (2008) reported
the required amount of organic substances for heterotrophic that at lower part of the column Thiobacillus denitrificans,
denitrification is higher than the theoretical value, which may a sulfur-oxidizing bacterium, was dominant, whereas at the
be due to higher cell yield than the prediction, and/or the use higher part of the column T. denitrificans was disappeared and
some portion of methanol as a carbon source by facultative Chlorobium limicola, a sulfide oxidizing bacterium, became
chemolithoautotrophic denitrifiers as discussed before. dominant. This data shows that sulfate reduction may occur
Between days 141 and 162, samples from the middle of the at the higher part of the column, where nitrate and nitrite
reactor was withdrawn for the measurement of NO3eN and were consumed.
NO2eN (data not shown). The NO3eN and NO2eN concentra-
tions averaged 2  2 mg/L and 13  3 mg/L, respectively.
3.2. pH, alkalinity and hardness in the column reactor
Results indicated that the conversion rate of NO3eN to NO2eN
is faster than the rate of NO2eN conversion to N2, as also
The influent pH was around 8.0 throughout the study. The
illustrated in batch tests, which will be discussed later.
effluent pH (6.2e7.7) was lower than the influent pH until
Fig. 2B illustrates that the collected and the theoretical N2
period 8 due to acid production of sulfur-utilizing autotrophic
gas production rate are in close agreement as the N2 gas
denitrifiers according to reaction (1). The optimum pH for
recovery throughout the study averaged 103%. Observing high
sulfur-oxidizing denitrifiers is between 6 and 9 (Holt et al.,
oscillating results in collected N2 amount was due to the
1994). According to reaction (1), 1 mol nitrate removal under
entrapment and sudden release of the produced gas in the
autotrophic conditions produces 1.28 mol hydrogen ions,
column due to the use of small sulfur and lime-stone particles
corresponding to 4.57 g CaCO3 consumption per gram of
as the packing materials. Similarly, Moon et al. (2008) reported
nitrate removal. In period 1, the effluent alkalinity was higher
that a portion of the N2 gas produced during denitrification
than the influent alkalinity due to rapid dissolution of lime-
process was entrapped in the pores of the column support
stone. In period 2, the effluent alkalinity decreased with
materials especially at low up-flow velocities. This entrap-
increasing feed flow due to limited dissolution rate of lime-
ment of the gas in the column may decrease the denitrifica-
stone. With increasing feed flow and decreasing HRT to
tion rate due to mass transfer limitation.
5.6 h, the effluent alkalinity decreased appreciably as alka-
In the case of complete denitrification, sulfide
linity supplied by lime-stone dissolution cannot balance the
(0.02e0.1 mg/L) was detected in the effluent of the column
acid produced by autotrophic denitrifiers. Hence, the decrease
(data not shown) even in the absence of external-organic
of denitrification efficiency in the 3rd period may be due to
supplementation (periods 1e5). The reason of this observa-
decreased alkalinity in the reactor. Although the use of lime-
tion was due to the activity of sulfate reducing bacteria at high
stone seems to be an effective and economical way of alka-
sulfate concentrations. Biomass decay may supplement the
linity supplementation, its limited dissolution rate makes
organics required for the reduction of sulfate (Sahinkaya,
difficult to provide enough alkalinity at high nitrate ladings
2009). Sulfur disproportionation according to reaction (4)
(Oh et al., 2001).
may also be responsible for sulfide detection. Sulfur dispro-
In period 8, the effluent pH and alkalinity increased sharply
portionation develops under anaerobic conditions after
reaching to the values higher than those of influent. According
depletion of all electron acceptors (NO 
3 and NO2 ), which may
to reaction (2), 3.57 g alkalinity is produced per gram of NO3eN
indicate that the reactor was being operated at lower loadings
denitrified heterotrophically. Hence, the increase of pH and
than the maximum capacity (Luna-Velasco et al., 2010).
alkalinity in the 8th period was due to commencement of
4S0 þ 4H2 O/3H2 S þ SO2 þ heterotrophic denitrification as discussed above. The required
4 þ 2H (4)

PERIODS
NO3-N
1 2 3 4 5 6 7 8 Sulfate
8
NO3-N or NO2-N (mg/L)

NO2-N 1200
50
Effluent 1000
Sulfate (mg/L)

6 Influent 40
800
Ca (meq/L)

30 600
4 20 400
2+

10 200
2
0 0
0 10 20 30 40 50
0 Time (day)
0 40 80 120 160 200
Fig. 4 e The variations of NO3eN, NO2eN and sulfate
Time (day)
concentrations in activity test (Temperature: 30  C, initial
Fig. 3 e The variations of feed and effluent Ca2D nitrate concentration: 50 mg/L NO3eN, initial biomass
concentrations. concentration: 127 ± 18 mg VSS/L).
6666 w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 6 6 6 1 e6 6 6 7

methanol for heterotrophic denitrification of each gram
NO3eN was calculated as 2.72 g. Similarly, Liu et al. (2009)
reported that when the methanol/NO3eN ratio was 2.47 the
denitrification was not complete and methanol remained in
the effluent when the ratio was 3.0.
In addition to sulfate production and alkalinity consump-
tion, increase of hardness due to lime-stone dissolution is
another disadvantage of sulfur and limestone autotrophic
denitrification (SLAD) process. Although Fe2þ, Mn2þ, Mg2þ and
Naþ concentrations in the influent and effluent did not
change, the average Ca2þ concentration in the effluent
increased during autotrophic denitrification process due to
dissolution of lime-stone (Fig. 3). The effluent Ca2þ increased
up to 7.15 meq/L (or 358 mg/L CaCO3) after increasing feed
NO3eN concentration from 50 mg/L to 75 mg/L, which was due
to increased acid production according to reaction (1) and
subsequent increase in the limestone dissolution. After
methanol supplementation, effluent Ca2þ started to decrease,
which should be due to decreased acid production as
a consequence of heterotrophic denitrification activity and
decreased limestone dissolution. In period 8, where hetero-
trophic denitrification was maximum and effluent sulfate
concentration below 250 mg/L, effluent Ca2þ concentration
decreased to around 2 meq/L (or 100 mg/L CaCO3) due to
decreased limestone dissolution as a result of decreased acid
generation under mixotrophic conditions. Hence, mixotrophic
denitrification has four significant advantages over SLAD
process: increased process efficiency, decreased effluent
sulfate concentration, decreased alkalinity requirement and,
lastly decreased effluent hardness.
3.3. Activity and kinetic tests
During the batch assays, sulfate concentration increased as
the NO2eN and NO3eN were consumed due to the activity of
sulfur-oxidizing denitrifiers (Fig. 4). Nitrite was detected as an
intermediate during the conversion of the NO3eN and the
maximum NO2eN concentration was reached when the
NO3eN was completely depleted, similar to the results re-
ported by Sierra-Alvarez et al. (2007). Similar to the column
reactor, the NO2eN reduction is the limiting step in autotro-
phic denitrification activity tests. Although NO3eN was
completely removed within 10 days (Fig. 4), NO2eN concen-
tration did not change significantly for a further 10 days, then
NO2eN was consumed rapidly. The linear portion of the
nitrate or nitrite-time curve was used to calculate the removal
rate. The specific NO3eN and NO2eN reduction rates were
65 mg NO3eN/(gVSS.d) and 21 mg NO2eN/g VSS.d, respec-
tively. Therefore, NO3eN reduction was over three times
faster than NO2eN reduction which caused NO2eN accumu-
lation at high loadings in the column reactors. In the study of
Sierra-Alvarez et al. (2007), NO3eN and NO2eN reduction rates
increased with increasing initial NO3eN concentration and
the maximum specific reduction rates were 77 mg NO3eN/
(gVSS.d) and 60.2 mg NO2eN/(gVSS.d), respectively.
The reduction rates of NO3eN and NO2eN at various initial
NO3eN concentrations were studied in batch serum bottles,
inoculated with 5 mL mixed liquor of the serum bottles used in
activity tests. The results of the kinetic tests were illustrated
in Fig. 5. Lag phase, between 6 and 10 days, was observed in
the tests depending on the initial nitrate concentration, which
should be due to low initial biomass concentrations in the
tests. Similar to the activity tests, the reduction of nitrate
produced nitrite. When the initial NO3eN concentrations were
83 and 110 mg/L, NO2eN concentration reached plateau and
its reduction did not commence for further 16 days. In the
tests, the NO3eN reduction rate did not change significantly at
varying initial concentrations, and it averaged 4.0  1.0 mg
NO3eN/L.d. The NO2eN reduction rate for the initial NO3-N
concentrations of 30 and 56 mg/L was similar, averaging
1.74  0.06 mg NO2eN/(L.d). The NO3eN reduction rate was
around 2.3 times higher than NO2eN reduction rate, similar to

A 500 100
B 600
30 400 80 500
400
300 60
NO3-N or NO2-N (mg/L)

20 300
NO2-N or NO3-N (mg/L)

200 40
Sulfate (mg/L)

200
Sulfate (mg/L)

10
100 20 100
0 0 0 0

70
C 1000 D 400
60 800 120
50 100 300
40 600 80
30 60 200
400
20 40
200 100
10 20
0 0 0 0
0 10 20 30 40 0 10 20 30 40
NO3-N
Time (day) Sulfate
NO2-N

Fig. 5 e The variations of NO3eN, NO2eN and sulfate concentrations in kinetic tests at different initial NO3eN concentrations
(The serum bottles were inoculated with 5 mL suspension (127 ± 18 mg VSS/L medium) of serum bottles used for activity
tests. Temperature was 30  C).
 w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 6 6 6 1 e6 6 6 7
the activity tests. The observation of similar reduction rates at
varying initial NO3eN concentrations are in good agreement
with the half-saturation constant (Ks) of 0.4 mg/L NO3eN for
attached sulfur-oxidizing denitrifiers (Zeng and Zhang, 2005).
The obtained results are in contradiction with those of Sierra-
Alvarez et al. (2007) as they observed linear increase in the
NO3eN reduction rate with increasing initial nitrate concen-
tration. The reason for this increase in the study of Sierra-
Alvarez et al. (2007) should be due to use of granular consor-
tium (0.5e3 mm), which may result in diffusion limitation and
higher apparent Ks values.
4. Conclusions
Simultaneous sulfur-based autotrophic and heterotrophic
denitrification can be achieved in one-reactor for drinking
water treatment. Complete denitrification of 75 mg/L NO3eN
with effluent sulfate concentration of around 225 mg/L was
achieved when feed methanol/NO3eN ratio was 1.67 g/g,
which was much lower than the heterotrophic theoretical
value of 2.47 g/g. Stimulating mixotrophic denitrification also
decreased Caþ2 release and alkalinity requirement of SLAD
process due to the alkalinity production of heterotrophic
process. Batch experiments showed that sulfur-based auto-
trophic NO2eN reduction rate was around three times lower
than the NO3eN reduction rate.
Acknowledgments
_
This study was funded by TUBITAK (Project No: 110Y256) and
Harran University Research Fund (Project No: 1062).
references
APHA AWWA WEF, 2005. Standard Methods for the Examination
of Water and Wastewater, first ed. American Public Health
Association, American Water Works Association, Water
Environmental Federation, Washington DC, USA.
Cord-Ruwish, R., 1985. A quick method for the determination of
dissolved and precipitated sulfides in cultures of sulfate-
reducing bacteria. J. Microbiol. Methods 4, 33e36.
6667
Holt, J.G., Krieg, N.R., Sneath, P.H.A., Staley, J.T., Williams, S.T.,
1994. Bergey’s Manual of Determinative Bacteriology. Williams
and Wilkins, Baltimore, MD.
Kim, E.W., Bae, J.H., 2000. Alkalinity requirements and the
possibility of simultaneous heterotrophic denitrification
during sulfur-utilizing autotrophic denitrification. Water Sci.
Technol. 42, 233e238.
Lee, D.U., Lee, I.S., Chai, Y.D., Bae, J.H., 2001. Effects of external
carbon source and empty bed contact time on simultaneous
heterotrophic and sulfur-utilizing autotrophic denitrification.
Process Biochem. 36, 1215e1224.
Liu, L.H., Koenig, A., 2002. Use of limestone for pH control in
autotrophic denitrification: batch experiments. Process
Biochem. 37 (8), 885e893.
Liu, H., Jiang, W., Wan, D., Qu, J., 2009. Study of a combined
heterotrophic and sulfur autotrophic denitrification
technology for removal of nitrate in water. J. Haz. Mat 169,
23e28.
Luna-Velasco, A., Sierra-Alvarez, R., Castro, B., Field, J.A., 2010.
Removal of nitrate and hexavalent uranium from
groundwater by sequential treatment in bioreactors packed
with elemental sulfur and zero-valent iron. Biotechnol.
Bioeng. 107, 933e942.
Moon, H.S., Shin, D.Y., Nam, K., Kim, J.Y., 2008. A long-term
performance test on an autotrophic denitrification column for
application as a permeable reactive barrier. Chemosphere 73,
723e728.
Oh, S.E., Yoo, Y.B., Young, J.C., Kim, I.S., 2001. Effect of organics on
sulfur-utilizing autotrophic denitrification under mixotrophic
conditions. J. Biotechnol. 92, 1e8.
Park, J.H., Shin, H.S., Lee, I.S., Bae, J.H., 2002. Denitrification of
high NO3eN containing wastewater using elemental sulfur;
nitrogen loading rate and N2O production. Environ. Technol.
23, 53e65.
Rittmann, B.E., McCarty, P.L., 2001. Environmental Biotechnology:
Principles and Applications. McGraw-Hill Book Co, New York.
Sahinkaya, E., 2009. Microbial Sulfate reduction at low (8  C)
temperature using waste sludge as a carbon and seed source.
Int. Biodeter. Biodeg 63, 245e251.
Sierra-Alvarez, R., Beristan-Cardoso, R., Salazar, M., Gomez, J.,
Razo-Flores, E., Field, J.A., 2007. Chemolithotrophic
denitrification with elemental sulfur for groundwater
treatment. Water Res. 41, 1253e1262.
Soares, M.I.M., 2002. Denitrification of groundwater with
elemental sulfur. Water Res. 36, 1392e1395.
Yesilnacar, M.I., Sahinkaya, E., Naz, M., Ozkaya, B., 2008. Neural
network prediction of nitrate in groundwater of Harran Plain,
Turkey. Environ. Geol. 56, 19e25.
Zeng, H., Zhang, T.C., 2005. Evaluation of kinetic parameters of
a sulfurelimestone autotrophic denitrification biofilm
process. Water Res. 39, 4941e4952.