Abstract

A novel Fe3O4@Cu/PVA biomaterial as a new adsorbent and bacterial cell immobilized carrier was synthesized in this work. The structure and morphology were characterized by scanning electron microscopy (SEM). Effects of factors on Mn(II)-based autotrophic denitrification were investigated in a moving bed biofilm reactor (MBBR). The results indicate that the highest nitrate removal and Mn(II) oxidation efficiency occurred under the conditions of initial Mn(II) concentration of 80 mg·L−1, hydraulic retention time (HRT) of 10 h and pH 7. Meteorological chromatography analysis showed that N2 was produced as an end-product, and that gas compositions were different depending on the concentration of Mn(II) in the MBBR. The community diversity in the MBBR was markedly influenced by the concentration of Mn(II) and Pseudomonas sp. H-117 played a primary role in the process of nitrate removal and Mn(II) oxidation.

INTRODUCTION

Nitrate-nitrogen contamination of groundwater has become an environmental and public health issue in developed and developing countries (Su et al. 2018). Several abiotic and biotic techniques for the removal of nitrates from water have been proposed in recent years (Wang et al. 2018). Among them the reduction of nitrate into harmless nitrogen gas (i.e., denitrification) by biological processes has become the most popular technology (Ghafari et al. 2008). However, low carbon-to-nitrogen ratios in many wastewaters require a supply of external carbon to meet complete heterotrophic denitrification (Wei et al. 2017). Autotrophic denitrification has received significant attention lately due to several inherent merits compared with conventional heterotrophic denitrification (Yang et al. 2016). Compared with heterotrophic denitrification, autotrophic denitrification has associated low operational costs and reduced sludge production autotrophic denitrifying (Yang et al. 2017). However, bacteria have a relatively low growth rate (Montalvo et al. 2016). To address these issues, immobilized cell technology can be applied in autotrophic denitrification.

Polymer brush-grafted magnetic nanoparticles for highly effective adsorption of heavy metal ions (Bae et al. 2016), and Fe3O4 nanoparticles also have potential for re-collection of metal ions. The objectives of this work were to: (1) determine nitrate removal and Mn(II) oxidation capability of the Fe3O4@Cu/PVA immobilized cells and removal mechanism; (2) evaluate the nitrate removal and Mn(II) oxidation capability under conditions of different pH, hydraulic retention time (HRT) and initial Mn(II) concentrations in the moving bed biofilm reactor (MBBR); (3) investigate the impacts of Mn(II) concentration on the microbial community and metabolic activity in the MBBR.

MATERIALS AND METHODS

Media and cultivation

Strain H-117 (Su et al. 2017) was isolated from the Shi Bian Yu reservoir, and grown in an anaerobic autotrophic medium (AM). The ingredients of the AM and trace element solution (TE) are presented in Table 1.

Table 1

The components of the media used in the autotrophic denitrification and operational conditions of the MBBR

TE g·L−1 AM g·L−1 
FeCl2·4H21.5 NaHCO3 0.5 
CuCl2·2H20.02 NaNO3 0.1 
MnSO4·7H20.1 KH2PO4 0.1 
Na2MoO4·2H20.024 MgSO4·7H20.05 
ZnCl2 0.006 CaCl2 0.05 
H3BO3 0.007 TE 2 mL 
Phase Initial Mn(II) (mg/L) Initial pH HRT (h) 
Phase I Period 1 80 10 
Period 2 80 
Period 3 80 
Period 4 80 10 
Period 5 80 10 
Phase II Period 6 60 10 
Period 7 60 
Period 8 60 
Period 9 60 10 
Period 10 60 10 
Phase III Period 11 40 10 
Period 12 40 
Period 13 40 
Period 14 40 10 
Period 15 40 10 
TE g·L−1 AM g·L−1 
FeCl2·4H21.5 NaHCO3 0.5 
CuCl2·2H20.02 NaNO3 0.1 
MnSO4·7H20.1 KH2PO4 0.1 
Na2MoO4·2H20.024 MgSO4·7H20.05 
ZnCl2 0.006 CaCl2 0.05 
H3BO3 0.007 TE 2 mL 
Phase Initial Mn(II) (mg/L) Initial pH HRT (h) 
Phase I Period 1 80 10 
Period 2 80 
Period 3 80 
Period 4 80 10 
Period 5 80 10 
Phase II Period 6 60 10 
Period 7 60 
Period 8 60 
Period 9 60 10 
Period 10 60 10 
Phase III Period 11 40 10 
Period 12 40 
Period 13 40 
Period 14 40 10 
Period 15 40 10 

Reactor operation and evaluation of the optimum conditions

The polyvinyl alcohol (PVA) sponge was purchased from Xiya Chemical Industry Co., Ltd (Shandong, China). The preparation of the Fe3O4@Cu/PVA functionalized biomaterials was according to Nodeh et al. (2016) and Azizi et al. (2014). The bioreactor system consisted of three modules: feed tank, pump, and MBBR. The MBBR was a 0.10 m organic glass column of 0.25 m inner diameter. The total volume was 4.5 L and the working volume was 4.0 L. The working volume was filled to a depth of 0.06 m with Fe3O4@Cu/PVA biomaterials (1 × 1 × 1 cm), and void ratio of 95%. The reactor was sealed by a gasket. The experiment involved running two parallel reactors, one reactor with no immobilization as a control, the other as the immobilized reactor. The MBBR was continuously and steadily operated for 115 days. As shown in Table 1, the operation of the MBBR was divided into 15 stages. Firstly, AM were prepared in the feed tank and the AM was pumped into the MBBR at the desired flow rate. The inorganic electron donor (Mn(II)) was prepared by diluting MnSO4·H2O in distilled water to the desired concentration. The reactor was operated at 30 ± 2 °C. The effluent port was set in the bottom of the MBBR and measured the nitrate-N, Mn(II), and nitrite-N concentrations each day.

Determination of nitrate removal and Mn(II) oxidation ability was performed by variation of HRT, pH and initial Mn(II) concentration, and the MBBR was employed to optimize the operating variables. The optimum conditions could be confirmed by the above 15 stages. After the preliminary study, the MBBR was operated under the optimum conditions.

Analytical methods

In all experiments, the effluent samples were filtered by a membrane (0.45 um pore-size) through a suction filter machine. The concentrations of nitrate, nitrite and Mn(II) were measured according to the standard methods. Water pH was measured by a bench-top pH meter (MM110, HACH, USA). The morphological properties of Fe3O4@Cu/PVA biomaterials and the bacteria were characterized by a scanning electron microscopy (SEM, JSM-5800, Japan JEOL) system, and the crystallographic structure of the Fe3O4@Cu/PVA biomaterials was determined by X-ray diffraction (XRD, Rigaku Ultima IV). The gases were analyzed immediately after collection using gas chromatography (Agilent6890, Japan; and PerkinElmer clarus 600, USA). Bacterial genomic DNA in the MBBR was extracted from the biofilm samples for bacterial community analysis.

RESULTS AND DISCUSSION

SEM photographs and XRD spectra of Fe3O4@Cu/PVA

The SEM images of Fe3O4@Cu/PVA and bacteria are shown in Figure 1. The Fe3O4@Cu/PVA exhibited a 3D interconnected network skeleton and a bumpy surface morphology (Figure 1(a)). After 10 days of incubation, strain H-117 was wrapped on the bumpy surface of the Fe3O4@Cu/PVA (Figure 1(c)). Comparing Figure 1(a) and 1(c), it is observed that strain H-117 was immobilized on the surface of the Fe3O4@Cu/PVA. The sediments of the experimental groups were also observed by SEM (Figure 1(b)).

Figure 1

SEM micrographs of (a) Fe3O4@Cu/PVA, (b) the sediment in the experimental groups, (c) Fe3O4@Cu/PVA immobilized cells, and (d) XRD patterns of the Fe3O4@Cu/PVA.

Figure 1

SEM micrographs of (a) Fe3O4@Cu/PVA, (b) the sediment in the experimental groups, (c) Fe3O4@Cu/PVA immobilized cells, and (d) XRD patterns of the Fe3O4@Cu/PVA.

In order to determine the crystal structure and the composition of the as-prepared samples, the PVA before and after the adsorption of Fe3O4@Cu was subjected to XRD analysis under a wide range of scans. Figure 1(d) shows the XRD patterns of PVA, Fe3O4@Cu, and Fe3O4@Cu/PVA, and these data indicate the presence of Fe3O4@Cu on the surface of the PVA. The Fe3O4@Cu/PVA shows four strong peaks at 30.19 (2.96 Å), 35.62 (2.52 Å), 56.98 (1.61 Å), and 62.78 (1.48 Å), which correspond to peaks [220], [311], [511], and [440] of Fe3O4@Cu.

Denitrification and Mn(II) oxidation in the MBBR

The experimental and control groups were operated in parallel under period 1–15 conditions for 90 days (Table 1). In the first three periods, the effect of decreasing HRT from 10 h (period 1) to 6 h (period 3) was tested. The nitrate removal ratio decreased from 100% (1.61 mg·L−1·h−1) in period 1 to 91% (2.31 mg·L−1·h−1) in period 3, and the nitrate removal ratio reached a maximum value of 100% when the HRT was 10 h (period 1). The data demonstrate that the denitrification efficiency was improved when the HRT increased. For Mn(II)-based autotrophic denitrification, Mn(II) was oxidized during nitrate removal. The Mn(II) oxidation ratio decreased from 74.83% (period 1) to 64.90% (period 2) and 60.32% (period 3) with the decrease in the HRT from 10 to 8 and 6 h, respectively (Figure 2(a)), which is in good agreement with the theoretical predictions. The Mn(II) could not be completely oxidized by microorganisms when the HRT was 6 h, resulting in a high concentration of Mn(II) in the effluent. The results suggest that good Mn(II) oxidation efficiency was obtained with long HRT in the MBBR. This may be because a long HRT provides bacteria with enough time to adapt to the new environment and Mn(II) had enough time to degrade. Bai et al. (2016) reported that a longer HRT may increase the reaction time, and thus improve the treatment effect. Compared with the removal of nitrate and Mn(II) oxidation, the accumulation of nitrite followed a different rule, and the concentration of nitrite remained below 2.5 mg·L−1 during the experiment.

Figure 2

Nitrate, nitrite, Mn(II) variations in the MBBR and control group: (a) Mn(II) = 80 mg·L−1, (b) Mn(II) = 60 mg·L−1, (c) Mn(II) = 40 mg·L−1.

Figure 2

Nitrate, nitrite, Mn(II) variations in the MBBR and control group: (a) Mn(II) = 80 mg·L−1, (b) Mn(II) = 60 mg·L−1, (c) Mn(II) = 40 mg·L−1.

In periods 4–5, pH in the feed was increased from 6 (period 4) to 8 (period 5). In order to recover the denitrification performance, the HRT was increased back to 10 h. As shown in Figure 2(a), the average nitrate removal efficiencies were 90.90% (1.42 mg·L−1·h−1), 100.00% (1.61 mg·L−1·h−1) and 89.86% (1.34 mg·L−1·h−1) at pH 6.0, 7.0, and 8.0, respectively. These data indicate that the Mn(II)-based autotrophic denitrification process was easily influenced by pH and that 16 mg·L−1 nitrate was almost completely degraded at pH 7.0 in the MBBR. Ghafari et al. (2010) also showed that the optimal pH value for the denitrification process was 7.5–8.0. The Mn(II) concentration is also shown in Figure 2(a). The Mn(II) oxidation ratios were 74.83% (pH 7.0), 64.90% (pH 6.0), and 60.32% (pH 8.0). The data can be explained by the fact that Mn(II) oxidation bacteria are more efficient in neutral pH conditions, which might be attributed to low or high pH values inhibiting Mn(II) oxidation microbial communities in the MBBR. Chen et al. (2015) also reported that low pH values (pH 6.0) inhibited the denitrification process due to the reactivity of hydrogenotrophic denitrifying bacteria being affected by the decomposition of carbonate ions in this environment. Nitrite concentration remained at very low levels (<2.5 mg·L−1).

From period 6 to period 10, the Mn(II) concentration was kept at 60 mg·L−1. As shown in Figure 2(b), the highest nitrate removal (1.44 mg·L−1·h−1) and Mn(II) oxidation ratio (4.20 mg·L−1·h−1) occurred during period 6 (HRT 10.0 h and pH 7.0). From period 11 to period 15, the Mn(II) concentration was kept at 40 mg·L−1. As shown in Figure 2(c), the maximum Mn(II) oxidation (1.26 mg·L−1·h−1) and denitrification efficiency (2.30 mg·L−1·h−1) were observed during period 11 (HRT 10 h and pH 7.0). Comparing period 1 (80 mg·L−1 Mn(II)), period 6 (60 mg·L−1 Mn(II)) and period 11 (40 mg·L−1 Mn(II)), the nitrate removal ratio increased when the Mn(II) concentration increased, and the maximum nitrate removal ratio of 100% (1.61 mg·L−1·h−1) was obtained at 80 mg·L−1 Mn(II). The results show that autotrophic denitrification ability is affected by Mn(II) concentration (Mn(II) as electron donor) and strain H-117 could tolerate and oxidize Mn(II) at concentrations as high as 80 mg·L−1.

Based on data in Figure 2, the strain H-117 may conduct the following sequential Mn(II)-based autotrophic denitrification reactions (Swathi et al. 2016): 
formula
(1)
 
formula
(2)
 
formula
(3)
 
formula
(4)
 
formula
(5)
 
formula

Detection of gaseous nitrogen compounds under different conditions

To confirm the nitrogen removal mechanism of the MBBR, gas chromatography (GC) was used to monitor the gas composition produced by the MBBR. Figure 3 shows that there was 99.86% nitrogen in the MBBR when the Mn(II) concentration was 80 mg·L−1 (period 1), 96.53% nitrogen when 60 mg·L−1 Mn(II) was used (period 6), and 88.76% nitrogen with 40 mg·L−1 Mn(II) (period 11). Meanwhile, 79% nitrogen was detected in the atmospheric background. It could be concluded that the gas compositions were different in the various Mn(II) concentrations. The maximum N2 emission (99.86%) was obtained when 80 mg·L−1 Mn (II) was used in the MBBR as the only electron donor. The result is in accordance with the nitrate removal ratio in the MBBR because N2 emission increased with the increased nitrate removal ratio. Zhou et al. (2017) also reported that a higher excess N2 concentration indicated a higher denitrification rate during NPKM treatment.

Figure 3

The gas composition produced by the MBBR: (a) Mn(II) = 80 mg·L−1, (b) Mn(II) = 60 mg·L−1, (c) Mn(II) = 40 mg·L−1, (d) feed and effluent nitrate, nitrite and Mn(II) variations in the MBBR under the optimum conditions.

Figure 3

The gas composition produced by the MBBR: (a) Mn(II) = 80 mg·L−1, (b) Mn(II) = 60 mg·L−1, (c) Mn(II) = 40 mg·L−1, (d) feed and effluent nitrate, nitrite and Mn(II) variations in the MBBR under the optimum conditions.

At the end of the experiment, N2O was not detected under anaerobic conditions when the concentration of the electron donor was 80 mg·L−1, which is in agreement with the results found by Ribera-Guardia et al. (2014). However, a low concentration of N2O was detected by the PE600 at 60 mg·L−1 Mn(II) (1.68 mg·L−1 N2O) and 40 mg·L−1 Mn(II) (3.09 mg·L−1 N2O). These results suggest that NO2 levels in the MBBR were slightly different at different Mn(II) concentrations, and the highest concentration of NO2 was found when the Mn(II) concentration was 60 mg·L−1. Itokawa et al. (2001) reported that limited availability of the electron donor could cause N2O accumulation. In this sense, there may be more chance of N2O accumulation in the MBBR with a low concentration of Mn(II). The results indicated that the MBBR shows good superiority in practical applications due to its negligible N2O emissions.

The MBBR performance under optimum conditions

As shown above, the optimum conditions for nitrate removal were initial Mn(II) concentration of 80 mg·L−1, HRT of 10 h, and pH of 7, and these conditions were used to carry out the following experiment. As shown in Figure 3(d), the average nitrate removal and Mn(II) oxidation efficiency of the MBBR were not stable during the first 4 days of the experiment due to the MBBR being in the start-up stage. After the fourth day, the nitrate removal and Mn(II) oxidation efficiency were 100.00% (1.59 mg·L−1·h−1) and 78.50% (6.30 mg·L−1·h−1), which are higher than in the other periods.

Pyrosequencing analysis of bacterial community structures

The bacterial community structure varied when different Mn(II) concentrations were applied to the MBBR. In this study, the samples were collected at initial Mn(II) concentrations of 80 mg·L−1 (FM1), 60 mg·L−1 (FM2), 40 mg·L−1 (FM3) and 80 mg·L−1 (control group (FM4)).

Altogether 19 bacterial phyla were detected in FM1, FM2 and FM3, and only 14 bacterial phyla were detected in FM4. These results indicate that bacteria could tolerate and grow at concentrations as high as 80 mg·L−1. The results in Figure 4(a) show that Proteobacteria, Firmicutes, Actinobacteria, Bacteroidetes, and Chloroflexi played a major role in the MBBR, and phylum Proteobacteria was the most dominant community in the MBBR. Comparing the relative abundance of Proteobacteria in FM1, FM2, and FM3 indicates that the Proteobacteria were accustomed to the high Mn(II) concentration (80 mg·L−1) and contributed to nitrate removal. Sun et al. (2017) also indicated that Proteobacteria were the main denitrification bacteria.

Figure 4

Taxonomic classification of bacterial 16S rRNA gene reads at (a) phylum level and (b) class level of samples; (c) genus levels FM1–FM4.

Figure 4

Taxonomic classification of bacterial 16S rRNA gene reads at (a) phylum level and (b) class level of samples; (c) genus levels FM1–FM4.

At the molecular level, 16S rRNA sequences belonging to 43 classes were identified in the biofilm of the MBBR. As shown in Figure 4(b), the major classes of the four samples were Gammaproteobacteria, Actinobacteria, Clostridia and Bacilli. According to our previous study, H-117 belongs to Gammaproteobacteria class, and the sum abundances of Gammaproteobacteria in samples FM1, FM2, FM3 and FM4 were 38.54%, 29.66%, 27.11%, and 18.46%, respectively. These results indicate that the relative abundance of Gammaproteobacteria increased as the Mn(II) concentration increased. It was shown above that the nitrate removal and Mn(II) oxidation ratio increased when the Mn(II) concentration increased, which suggests that class Gammaproteobacteria was important and the dominant community for nitrate removal and Mn(II) oxidation in the MBBR. Other studies have also indicated that Gammaproteobacteria play a vital role in the denitrification process and perform with nitrate removal abilities in a CBSAD reactor (Chen et al. 2018). Meanwhile, Pseudomonas was present in the Gammaproteobacteria, which is known to be an Mn(II) oxidizer and detected in oligotrophic environments (Cao et al. 2015).

The microbial community heatmap (Figure 4(c)) shows the similarity and abundances of the four samples; Pseudomonas (6.40%) was the most abundant genus in FM1, and FM2, FM3, FM4 showed different relative abundances of Pseudomonas (4.96% in FM2, 4.23% in FM3 and 1.57% in FM4). It could be concluded that Pseudomonas showed an increasing trend when the Mn(II) concentration was increased. Meanwhile, as the most dominant genus in the MBBR, Pseudomonas may make a significant contribution to nitrate removal and Mn(II) oxidation in the water treatment process. Villalobos et al. (2003) reported that Pseudomonas has the ability to oxidize Mn(II). Diep et al. (2009) reported that Pseudomonas can be used to remove nitrate from wastewater. Most of the iron-oxidizing bacteria detected in the biofilter, such as Pseudomonas, Crenothrix, Gallionella, and Bacillus, had the ability to oxidize manganese (Cheng et al. 2017). Acinetobacter, Chryseobacterium and Rhodococcus were also detected in the four samples and these bacteria also play important roles in the process of nitrate removal and Mn(II) oxidation.

CONCLUSIONS

The Fe3O4@Cu/PVA biomaterials inoculated with the nitrate-removing and Mn(II)-oxidizing bacteria Pseudomonas sp. H-117 effectively removed nitrate and oxidized Mn(II) under anaerobic conditions. Fe3O4@Cu/PVA-immobilized cells were used for nitrate removal in the MBBR, and the highest nitrate removal efficiency conditions were an initial Mn(II) concentration of 80 mg·L−1, HRT of 10 h, and pH 7. Meteorological chromatography analysis suggested that N2 and NO2 levels varied depending on the concentration of Mn(II). High-throughput sequencing suggested that Pseudomonas sp. H-117 was the dominant contributor for effective removal of nitrate and oxidization of Mn(II) in the MBBR.

ACKNOWLEDGEMENTS

This research work was partly supported by the National Natural Science Foundation of China (NSFC) (No. 51678471), State Key Laboratory of Urban Water Resource and Environment (Harbin Institute of Technology) (No. 2017TS06), the Foundation for Creative Research Groups of China (50821002), Key Scientific Technological Innovation Team Plan of Shanxi Province (No. 2017KCT-19-02), and Key Research and Development Plan in Shaanxi Province of China (2018ZDXM-SF-029).

REFERENCES

REFERENCES
Azizi
K.
,
Karimi
M.
,
Nikbakht
F.
&
Heydari
A.
2014
Direct oxidative amidation of benzyl alcohols using EDTA@Cu(II) functionalized superparamagnetic nanoparticles
.
Applied Catalysis A: General
482
,
336
343
.
Cao
L. T. T.
,
Kodera
H.
,
Abe
K.
,
Imachi
H.
,
Aoi
Y.
,
Kindaichi
T.
,
Ozaki
N.
&
Ohashi
A.
2015
Biological oxidation of Mn(II) coupled with nitrification for removal and recovery of minor metals by downflow hanging sponge reactor
.
Water Research
68
,
545
553
.
Chen
D.
,
Wang
H.
,
Ji
B.
,
Yang
K.
,
Wei
L.
&
Jiang
Y.
2015
A high-throughput sequencing study of bacterial communities in an autohydrogenotrophic denitrifying bio-ceramsite reactor
.
Process Biochemistry
50
(
11
),
1904
1910
.
Ghafari
S.
,
Hasan
M.
&
Aroua
M. K.
2008
Bio-electrochemical removal of nitrate from water and wastewater: a review
.
Bioresource Technology
99
,
3965
3974
.
Montalvo
S.
,
Huiliñir
C.
,
Gálvez
D.
,
Roca
N. A.
&
Guerrero
L.
2016
Autotrophic denitrification with sulfide as electron donor: effect of zeolite, organic matter and temperature in batch and continuous UASB reactors
.
International Biodeterioration and Biodegradation
108
,
158
165
.
Nodeh
H. R
,
Ibrahim
W. A. W.
,
Ali
I.
&
Sanagi
M. M.
2016
Development of magnetic graphene oxide adsorbent for the removal and preconcentration of As(III) and As(V) species from environmental water samples
.
Environmental Science and Pollution Research
23
(
10
),
9759
9773
.
Ribera-Guardia
A.
,
Kassotaki
E.
,
Gutierrez
O.
&
Pijuan
M.
2014
Effect of carbon source and competition for electrons on nitrous oxide reduction in a mixed denitrifying microbial community
.
Process Biochemistry
49
(
12
),
2228
2234
.
Su
J. F.
,
Ma
M.
,
Huang
T. L.
,
Ma
F.
,
Shao
S. C.
,
Lu
J. S.
&
Deng
L. Y.
2017
Characteristics of autotrophic and heterotrophic denitrification by the strain Pseudomonas sp. H117
.
Geomicrobiology
34
(
1
),
45
52
.
Sun
H.
,
Qiang
W.
,
Yu
P.
,
Zhang
L.
,
Ye
L.
,
Zhang
X.
&
Ren
H.
2017
Denitrification using excess activated sludge as carbon source: performance and the microbial community dynamics
.
Bioresource Technology
238
,
624
632
.
Swathi
D.
,
Sabumon
P. C.
&
Maliyekkal
S. M.
2016
Microbial mediated anoxic nitrification-denitrification in the presence of nanoscale oxides of manganese
.
International Biodeterioration and Biodegradation
119
,
499
510
.
Villalobos
M.
,
Toner
B.
,
Bargar
J.
&
Sposito
G.
2003
Characterization of the manganese oxide produced by Pseudomonas putida strain MbB1
.
Geochimica et Cosmochimica Acta
67
(
14
),
2649
2662
.
Wang
H. Y.
,
Song
Q.
,
Wang
J.
,
Zhang
H.
,
He
Q. L.
,
Zhang
W.
,
Song
J. Y.
,
Zhou
J. P.
&
Li
H.
2018
Simultaneous nitrification, denitrification and phosphorus removal in an aerobic granular sludge sequencing batch reactor with high dissolved oxygen: effects of carbon to nitrogen ratios
.
Science of the Total Environment
642
,
1145
1152
.
Yang
W. M.
,
Lu
H.
,
Khanal
S. K.
,
Zhao
Q.
,
Meng
L.
&
Chen
G. H.
2016
Granulation of sulfur-oxidizing bacteria for autotrophic denitrification
.
Water Research
104
,
507
519
.
Yang
Y.
,
Chen
T. H.
,
Morrison
L.
,
Gerrity
S.
,
Collins
G.
,
Porca
E.
,
Li
R. H.
&
Zhan
X. M.
2017
Nanostructured pyrrhotite supports autotrophic denitrification for simultaneous nitrogen and phosphorus removal from secondary effluents
.
Chemical Engineering Journal
328
,
511
518
.
Zhou
W.
,
Xia
L.
&
Yan
X.
2017
Vertical distribution of denitrification end-products in paddy soils
.
Science of the Total Environment
576
,
462
471
.