Stoichiometry of sulphide oxidation with nitrate as electron acceptor
F.Giaccherini*, F. Spennati*, A.Mannucci*, F. Alatriste-Mondragon**, G.Mori*** and G.Munz*
* Dept of Civil and Environmental Engineering, University of Florence, Via S. Marta 3, 50139 Florence,
Italy
** IPICYT, Camino a la Presa San José 2055, 78216 San Luis Potosí, Mexico
2
*** Cer co, Consorzio Cuoiodepur, Via Arginale Ovest 56020 San Miniato (PISA), Italy
Abstract
Authotrophic denitrification with sulphide using nitrate as electron acceptor were investigated at bench
scale at different HRT (5 and 20 d). The results indicated that the process allows to achieve complete
sulphide removal in all testes conditions. Tested sulphide loads were estimated from the H2S produced in
a pilot scale anaerobic digester treating vegetable tannery primary sludge; nitrogen loads originated from
the nitrification of the supernatant. Average nitrogen removal efficiencies higher than 80% were observed
in all the tested conditions once steady state was reached.
Due to sulphide limitation, an incomplete denitrification have been observed and nitrite and thiosulphate
tend to accumulate especially in the presence of variable environmental conditions.
Autotrophic denitrification efficiency and intermediate sulphur compounds production have been
simulated using a two step denitrification and three step sulphide oxidation model. The model allowed to
estimate those parameters that can affect the production of biomass in a hypotetic real scale Biotrickling
-1
filter for biogas desulphurization, as the SOB yield factor (0.6 mgX SOB(COD) mgS ) and decay coefficients
-1
(0.06 d ).
Keywords
Activated sludge modeling, denitrification, nitrate, nitrogen removal, sulphide biological oxidation.
INTRODUCTION
The removal of nitrogen by using sulphide as electron donor is a potentially interesting solution when
sulphate and nitrogen are abundant in an effluent (such as for tannery industry) that can be
advantageously treated through anaerobic digestion or for the simultaneous removal of nitrogen, carbon
and sulfur from effluents of the petroleum industry (Reyes-Avila et al., 2014).
Generally, tannery effluents are rich in nitrogen, especially organic nitrogen (Lofrano et al. 2013). Despite
the use of anaerobic processes to treat tannery wastewater and primary tannery sludge (Giaccherini et
al., 2014) can be seen as an interesting alternative to conventional aerobic treatment, its application has
several drawbacks, among which consistent production of sulfide due to the reduction of sulfate which
occurs in the absence of alternative electron acceptors such as oxygen and nitrate; as a consequence the
implementation of adequate technology for H2S desorption and treatment is required (Mannucci et al.
2012).
Anaerobic denitrification using sulphide as electron acceptor was already investigated with reference both
to liquid (Lu et al., 2009) and gaseous streams (Kleerebezem and Mendez, 2002; Mora et al. 2014).
Static biotrickling filters (BTF) are often used for the application at real scale for the biogas
desulphurization through biological processes (Rodriguez et al., 2014) but when biogas is treated in a
static BTF, at high volumetric load, biomass accumulation may lead to the clogging of the bed (Mannucci
et al., 2012), and consequently to the increase of head loss and operational costs.
Within the Biosur Life+ project a moving bed biofilm reactor for gaseous effluent treatment, with the aim of
allowing biomass removal and solids retention time control have been developed; in this context,
however, it is still important to quantify the kinetics and the stoichiometry of the processes involved.
Most of previous work on the kinetics and stoichiometry of sulphur compounds oxidation through
anaerobic processes were carried out with thiosulphate as electron donor (Artiga et al., 2005; Jing et al.,
2010; Mora et al., 2014), while very few tests (Munz et al., 2009; Mora et al. 2014) have been carried out
with sulphide due to the difficulty of using a strongly volatile and, at the same time inhibiting, substrate.
Kinetics and physiology of autotrophic denitrifying sulfur bacteria were studied by Oh et al., (2010)
through a steady-state anaerobic master culture reactor (MCR) operated for over six months under a
semi-continuous mode and nitrate limiting conditions using nutrient/mineral/buffer (NMB) medium
containing thiosulfate and nitrate. Results shows a biokinetic parameter estimation in the following ranges
-1
-1
-1
for the parameter values: μmax =0.12-0.2 h ; kd=0.3-0.4 h ; Ks=3-10mg L ; YNO3=0.4-0.5mg Biomass mg
—
-1
2(NO3 N) . Denitrification inhibition occurred when the concentrations of NO 3 -N, and SO4 reached about
-1
-1
660 mg L and 2000 mg L , respectively. The autotrophic denitrifying sulfur bacteria were observed to be
very sensitive to nitrite but relatively tolerant of nitrate, sulfate, and thiosulfate.
Reyes-Avila et al (2014) studied the biological denitrification to eliminate carbon, nitrogen and sulfur in a
lab scale anaerobic continuous stirred tank reactor. Acetate and nitrate at a C/N ratio of 1.45 were fed at
-3
-3
loading rates of 0.29 kg C m d and 0.2 kg N m d, respectively. Under steady-state denitrifying
conditions, the carbon and nitrogen removal efficiencies were higher than 90%. Moreover, under these
22- -3 -1
conditions, sulfide (S ) was fed to the reactor at several sulfide loading rates (0.042–0.294 kg S m d ).
The high nitrate removal efficiency was maintained along the whole process, whereas the carbon removal
2- -3 -1
was 65% even at sulfide loading rates of 0.294 kg S m d . The sulfide removal increased up to 99% via
0
partial oxidation to insoluble elemental sulfur (S ) that accumulated inside the reactor.
The effect of both nitrite and nitrate on the performance of anaerobic sulphide oxidation process (ASO
process) have been studied by Jing et al. (2010).
In the nitrate–ASO process, the maximum influent sulfide concentration and volumetric removal rate
resulted higher than those in nitrite–ASO process. In nitrate–ASO process, the half-saturation
-1
-1
concentrations were 0.35 ± 0.09 mg L (nitrate) and 1.36 ± 0.36 mg L (sulfide), while for nitrite–ASO
-1
-1
process, the half-saturation concentrations were 0.26 ± 0.08 mg L (nitrite) and 1.80 ± 0.30 mg L
(sulfide).
Chung et al. (2014) carried out a study to determine the possibility of autotrophic denitritation using
thiosulfate as an electron donor, to compare the kinetics of autotrophic denitrification and denitritation,
and to study the effects of pH and sulfur/nitrogen (S/N) ratio on the denitrification rate. Results showed
that both nitrate and nitrite were removed by autotrophic denitrification using thiosulfate as an electron
-1
donor at concentrations up to 800 mg-N L . Denitrification required a S/N ratio of 5.1 for complete
denitrification, but denitritation was complete at a S/N ratio of 2.5, which indicated an electron donor cost
savings of 50%. Moreover biomass yields of 0.53 mg-VSS/mg-N for denitrification and 0.34 mg-VSS/mgN for denitritation were estimated, indicating less sludge production during denitritation, resulting in
reduced sludge handling for denitritation compared to denitrification.
Anoxic respirometry have been used by Mora et al. (2014) to characterize a sulfide-oxidizing nitratereducing (SO-NR) culture obtained from an anoxic biogas desulfurizing biotrickling filter treating high H2S
loads. Respirometric profiles showed that nitrite was formed as intermediate during nitrate reduction and
revealed that no competitive inhibition appeared when both electron acceptors were present in the
2medium. Although final bioreaction products depended on the initial S2O3 /NO3 ratio, such ratio did not
affect thiosulfate oxidation or denitrification rates. Moreover, respirometric profiles showed that the
specific nitrite uptake rate depended on the biomass characteristics being that of a SO-NR mixed culture
-1
-1
-1
(39.8 mg N g VSS h ) higher than that obtained from a pure culture of T. denitrificans (19.7 mg N g
-1
VSS h ). Taking into account those results in 2015 they propose a kinetic model describing the two-step
denitrification (NO3→NO2→N2) for the autotrophic denitrification with thiosulfate.
This work is aimed at investigating denitrification with sulphide, in the presence of typical S/N ratio that
can derive from the anaerobic digestion of primary sludge of tannery wastewater.
The supernatant originated from the anaerobic digestion is characterized by high ammonia concentration;
its nitrification in a side stream separated from the main treatment train will allow the production of liquid
streams with high nitrate and/or nitrite concentration that could be used as electron acceptor in
autotrophic denitrification process.
A two step denitrification and three step sulphide oxidation model based on ASM models have been used
to estimate those parameters that can affect the production of biomass in a BTF, that is mainly, the
stoichiometric coefficients of sulphide oxidizing bacteria (SOB) growth and the decay kinetics.
METHODS
2
The present experimentation have been conducted CER CO Lab located at the Cuoiodepur WWTP (San
Miniato, Pisa – Italy). A bench scale sequencing batch reactor (SBR) with working volume of 3.2 L and
HRT of 24 h, was used (Figure 1). Influent nitrogen and sulphide loads were chosen on the basis of the
results of an experimentation (data not shown) conducted using a pilot scale anaerobic digester (volume
2= 150 L, SRT = 25 d) fed with sludge from Cuoiodepur WWTP primary settler. The S-SO4 concentration
th
in the primary sludge have been monitored daily for more than 250 days (from January 28 to November
th
2-1
217 2013). The average sulphate concentration and the S-SO4 /Stot were 740 ± 200 mg L S-SO4 and
22-1
0.96, respectively. Anaerobic digester effluent S-SO4 and effluent S-S were 15 mg S L and 13 mg S L
1
, respectively. Experimental data confirmed that 96 % of the influent sulphur (as sulphate) was reduced
to H2S that exits the digester as the biogas. Ammonia concentration in the supernatant resulted 722 ± 60
-1
mg N L .
-1
Average S/N ratio for the autotrophic denitrification of 1.02 gS gN have been calculated.
An S/N ratio of 1.4 constitutes the worst estimated conditions (maximum sulphide and minimum
nitrate/nitrite) to the main objective of complete sulphide removal. S/N ratios similar to those obtained
during anaerobic digestion experimentation where tested as influent conditions causing sulphide limiting
conditions respect to stoichiometry proposed by Fajardo et al. (2012).
Figure 1.1 Schematic of the used bench scale SBR
-1
The reactor was fed filled with 1 L (SST = 8000 mg SST L ) of activated sludge collected from
Cuoiodepur WWTP biological section and diluted with tap water. The reactor was fed with two distinct
peristaltic pumps with sulphide (Solution A) and with a solution of micronutrients and either nitrate or
nitrite (Solution B and C, respectively). To obtain the sludge capable of autotrophic denitrification,
activated sludge was acclimatized by feeding synthetic wastewater (Solution A and Solution B) under
anoxic condition for more than 40 days.
R1 was operated for 160 days and after the start-up, the experimentation was divided into two phases
with different operational conditions as reported in Table 1.
Table 1.1 Operational conditions
Parameter
Value
Units
Sample
Phase I Phase II
SRT
d
5
20
S
mg S-HS- L-1
375
340
Inflow
N
mg N-NO3- L-1
220
220
Inflow
S/N
g S (g N)-1
1.70
1.55
Inflow
Solution A was stored in the absence of headspace, that is, in a storage tank with a variable volume, and
-1
at pH =10 to minimize the desorption of hydrogen sulphide. For Solution A NaHCO 3 (1.24 g L ) and
-1
Na2S3H2O (0.41 g L ) diluted in demineralised water were used; for Solution B and C Na 2HPO4  2H2O
-1
-1
-1
-1
-1
(0.66 g L ); KH2PO4 (0.52 g L ); NH4Cl (0.05 g L ); MgSO4  7H2O (0.063 g L ); KNO3 (1.63 g L ; B only)
-1
and NaNO2 (0.96 g L C only) were diluted in tap water.
The cycle phases for both reactors were as follow: feeding (30 min); mixing (60 min); settling (240 min);
decant (30 min). The excess biomass was removed during the last 2 min of the mixing phase. The
reactors were maintained at pH between 7 and 8 through the dosage of an HCl solution and at room
temperature (between 18 and 28 °C).
Samples were collected three times a week and the following parameters were monitored: COD (soluble
-,
2and total), VSS (Volatile Suspended Solids), TSS (Total Suspended Solids), N-NO2 N-NO3 , S-SO4 ,
Sulphide, Total Sulphur, T, pH. Thiosulphate and elemental sulphur, were estimated indirectly. Soluble
and total COD, TSS and VSS have been analyzed according to IRSA-CNR methods (Metodi analitici per
2le acque, 2003). S-SO4 , Sulphide, Nitrite and Nitrate have been measured through ionic chromatography
(ICS1000, Dionex, Sunnyvale - U.S.A.) while total sulphur have been measured using plasma
spectrophotometry (ICP-OES, Agilent Technology, Santa Clara – U.S.A.). A portable Hach-Lange (Berlin,
Germany) probe was used to measure the pH twice a day in the effluent. Elemental sulphur and
thiosulphate have been estimated on the results of total and soluble COD in the wasted mixed liquor.
Thiosulphate remains in soluble form and contributes to the soluble COD while elemental sulphur remains
in particulate and colloidal form and its contribution to the total COD have been estimated on the basis of
the difference between total and soluble COD. Thiosulphate and elemental sulphur concentration in the
reactors have been estimated according to Eq. 1 and Eq. 2:

gCOD gCOD  MWsulphur
S thiosulphate   CODS  N nitrite

gN nitrite gS thiosulphate  MWthiosulphate



VSS gCOD
gCOD
S elementalsulphur   CODP 
TSS 
TSS gSSV

 gS elemental sulphur
Eq. 1
Eq. 2
Where MW = molecular weight.
The stoichiometry of the process and the decay coefficient where estimated by using an activated sludge
model that included a two-step denitrification and three steps sulphide oxidation (to elemental sulphur,
thiosulphate and sulphate). The model was calibrated with the experimental data of two subperiods of
phase I (SRT=5d) and phase II (SRT=20 d) during which the process was stable.
The number of parameters to be calibrated was reduced with the following assumptions:
1) a single SOB population according to results reported by Mora et al. (2015);
2) the yield coefficient of SOB growing with nitrite and nitrate was identical according to Kaelin et al.
(2009);
3) the yield coefficient associate to the growth of SOB was proportional to the increase of the oxidation
state of sulphur in the reaction and expressed as a fraction of a total yield coefficient referring to the
20
22oxidation of HS to SO4 (YSOB). The oxidation state for HS is -2, for S is 0, for S2O3 is +2, for SO4 is +
24 and for SO4 is + 8. Following the previously hypotesis and the results of Mora, we calculate:
0
0
2220.145YSOB from HS to S ; 0.145YSOB from S to S2O3 and 0.29YSOB from S2O3 to SO4
4) the decay process was represented with the death regeneration model and heterotrophic biomass was
considered as growing on the SOB decay products.
RESULTS AND DISCUSSIONS
During the experiment sulphide removal efficiency (RE) in the SBR was always higher than 99%. This
results are the same of those reported by Fajardo et al. (2012) in SBR reactors treating sulphide and
nitrate in sulphide limiting conditions. Sulphide removal kinetics resulted extremely high and only few
-1
minutes were needed for sulphide concentration to decrease below 1 mg L during each cycle.
Nitrate removal was not complete and nitrite accumulated in the reactor and was present in the effluent
23(Table 1).Nitrate removal efficiency and the ratio between accumulated N-NO and reduced N-NO
(Figure 2) indicates an incomplete denitrification.
Phase I
Phase II
31
Phase III
90
29
80
27
70
60
25
50
23
40
21
30
19
20
Temperature[°C]
Nitrate RE,
Accumulated nitrite/Reduced Nitrate (%)
100
17
10
0
15
20
40
60
80
100
120
140
160
Time (d)
Nitrate RE
Accumulated Nitrite / Reduced Nitrate
-
Temperature
Figure 1.2 Nitrate removal efficiency and ratio between accumulated N-NO2 and reduced N-NO3
-
The analysis of the SBR effluent confirmed the presence of intermediate sulphur compounds (Table 1).
Table 1.2 Sulphur mass balance in the SBR in quasi-steady state conditions
Phase I
Parameter
Phase II
Unit
Experimental
Models
phase
estimation
Experimental
phase
Models
estimation
HS-
mg S L-1
0±0
0
0±0
0
S0
mg S L-1
2
0
0
0
S2O3-
mg S L-1
20.0
23
28.9
14
SO42-
mg S L-1
180±25
164.1
175±17
156
NO2-
mg N L-1
12±4
8
17± 6
9
NO3-
mg N L-1
10±3
7
13±6
10
In Table 1, the results of model simulation and experimental data are reported. The hypotheses at the
basis of the model where verified to be compatible with a representation of the phenomena related to
biomass production at SRT=5 d and SRT=20 d. The estimated values for the decay coefficient was 0.06
-1
d , quite lower than the values previously estimated in aerobic conditions (Munz et al. 2009), while the
-1
yield factor resulted YSOB = 0.6 mgXSOB(COD) mgS . A ratio between ΔStot, measured/ΔStot,theoretical equal to 1.06
and 0.94 resulted in steady state conditions in phase I and II, respectively. From an application
-1
perspective, this imply that for the load expected at Cuoiodepur WWTP (0.95 gS gN when anaerobic
digestion of the primary sludge would be applied), the process would allow to completely remove sulphide
and a large fraction of nitrogen from the supernatant.
CONCLUSIONS
The results of this study indicate that denitrification with sulphide as electron donor is effective for the
removal of sulphide in the tested conditions with different HRT. Sulphide removal efficiency was higher
than 99% in all the tested conditions and nitrogen removal efficiencies higher than 80% were obtained.
The loads of nitrogen present in tannery wastewater resulted as compatible with the application of biogas
biological treatment through autotrophic denitrification.
The proposed model characterized by two step denitrification and three step of sulphide oxidation
effectively predict intermediate sulphur compounds formation and incomplete denitrification in all tested
conditions.
ACNOWLEDGMENTS
The authors acknowledge the Life+ program (Life Env/IT/075 Biosur), the Toscana Region (Meta POR
2007-2013) activities and the Marie Curie program (Irses 295176).
REFERENCES
Artiga P., González F., Mosquera-Corral A., Campos J. L., Garrido J. M., Ficara E. and Méndez R. 2005 Multiple
analysis reprogrammable titration analyser for the kinetic characterization of nitrifying and autotrophic denitrifying
biomass. Biochemical Engineering Journal, 26, 176–183.
Chung J., Amin K., Kim S., Yoon S., Kwon K. and Bae W. 2014 Autotrophic denitrification of nitrate and nitrite using
thiosulfate as an electron donor. Water Research, 58, 169-178.
Fajardo C., Mosquera-Corral A., Campos J. L. and Mendez R. 2012 Autotrophic denitrification with sulphide in a
sequencing batch reactor. Journal of Environmental Management, 113, 552-556.
Giaccherini F., Mannucci A., Munz G., Mori G., Alatriste-Mondragon F. and Lubello C. 2014 Anaerobic digestion of
vegetable tannery and municipal sludge. Proceedings of XI Latinoamerican Symposium of Anaerobic Digestion, La
Habana, Cuba, November 2014.
Jing C., Ping Z. and Mahmood Q. 2010 Influence of various nitrogenous electron acceptors on the anaerobic sulfide
oxidation. Bioresource Technology, 101(9), 2931-2937.
Kaelin D., Manser R., Rieger L., EugsterJ., Rottermann K. and Siegrist H. 2009 Extension of ASM3 for two-step
nitrification and denitrification and its calibration and validation with batch tests and pilot scale data. Water Research,
43, 1680-1692.
Kleerebezem R. and Mendez R. 2002 Autotrophic denitrification for combined hydrogen sulfide removal from biogas
and post-denitrification. Anaerobic Digestion IX, 45(10), 349-356.
Lofrano G., Meriç S., Zengin G. E. and Orhon D. 2013 Chemical and biological treatment technologies for leather
tannery chemicals and wastewaters: a review. Science of the Total Environment, 461, 265-281.
Lu H., Wang J., Li S., Chen G. H., van Loosdrecht M. C. and Ekama G. A. 2009 Steady-state model-based evaluation
of sulfate reduction, autotrophic denitrification and nitrification integrated (SANI) process. Water Research, 43(14),
3613-3621.
Mannucci A., Munz G., Mori G., Lubello C. 2012 Biomass accumulation modelling in a highly loaded biotrickling filter
for hydrogen sulphide removal. Chemosphere, 88(6), 712-717.
Metodi analitici per le acque 2003, Agenzia per la protezione dell’ambiente e per i servizi tecnici/ Istituto di Ricerca
sulle Acque Consiglio Nazionale delle Ricerche, Rome, Italy.
Mora M., Lopez L. R., Gamisans X. and Gabriel D. 2014 Coupling respirometry and titrimetry for the characterization
of the biological activity of a SO-NR consortium. Chemical Engineering Journal, 251, 111-115.
Mora M., Dorado A. D., Gamisans X. and Gabriel D. 2015 Investigating the kinetics of autotrophic denitrification with
thiosulfate: Modeling the denitritation mechanisms and the effect of the acclimation of SO-NR cultures to nitrite.
Chemical Engineering Journal, 262, 235-241.
Munz G., Gori R., Mori G., ,Lubello C. 2009 Monitoring biological sulphide oxidation processes with combined
respirometric and titrimetric techniques. Chemosphere, 76, 644-650.
Oh S. E., Kim K. S., Choi H. C., Cho J. and Kim I. S. 2000 Kinetics and physiological characteristics of autotrophic
dentrification by denitrifying sulfur bacteria. Water Science and Technology, 42(3-4), 59-68.
Reyes-Avila J., Razo-Flores E. and Gomez J. 2014 Simultaneous biological removal of nitrogen, carbon and sulfur by
denitrification, Water Research, 38(14-15), 3313-3321.
Rodriguez G., Dorado A. D., Fortuny M., Gabriel D. and Gamisans X. 2014 Biotrickling filters for biogas seeetening:
Oxygen tranfer improvement for a reliable operation. Process safety and environmental Protection, 92(3), 261-268.