Abstract

To investigate the advantages of mixed carbon source over a single one in deep denitrification, sodium acetate, glucose and their mixture were used as carbon sources in present study. Denitrification performance, effluent pH, microbial community and carbon source cost were taken into account. With the same influent NO3-N concentration of 50 mg/L and the same C/N ratio of 1.5, the NO3-N removal rate with the mixed carbon source (96.53%) was slightly lower than that with sodium acetate (98.15%), but significantly higher than that with glucose (74.69%). The specific denitrification rates of the sodium acetate, glucose and sodium acetate/glucose reactor were 47.7, 29.7 and 45.4 mg N/g VSS d, respectively. The effluent pH with sodium acetate varied in the range of 9.13–9.60, exceeding the discharge standard limit of 9.0, whereas the sodium acetate/glucose reactor could keep pH in the range of 7.80–8.23. The 16S rRNA gene-based high-throughput sequencing revealed that carbon sources determined the microbial community structure and the sludge Shannon index with the mixed carbon source was the highest. Furthermore, cost estimation indicated that the mixed carbon source was the cheapest. This study is significant as it tests reasonable selection of carbon sources for deep denitrification in practice.

INTRODUCTION

The anoxic/oxic (A/O) process is widely used to treat the wastewater containing nitrogenous components. Although fairly good removal efficiency of ammonia nitrogen and total nitrogen can be achieved for A/O process, nitrate concentration in its effluent usually maintains at a high level when the raw wastewater contains high concentrations of nitrogen. In this case, the increasingly stringent discharge standards cannot be met. To remove residual nitrate, deep denitrification is commonly applied. Denitrification is the reduction of nitrate to N2 carried out by denitrifying bacteria (Xu et al. 2018). Generally, most denitrifying bacteria are heterotrophic and therefore require organic carbon sources which have an important effect on the denitrification process for cell growth and nitrate reduction (Zhang et al. 2018). However, the organic carbon in the raw wastewater has been mostly consumed during the A/O process, resulting in C/N ratio of the secondary effluent insufficient for further denitrification. Hence, the addition of an external carbon source is needed.

Traditionally, liquid organic substances such as methanol, ethanol, glucose and acetate serve as external carbon sources for nitrate removal (Calderer et al. 2010; Ribera-Guardia et al. 2014). Methanol was once the most commonly employed external carbon source due to being easily assimilated by denitrifying bacteria and its low cost (Rabah & Dahab 2004). Ethanol was found to be considerably more readily available as a carbon source for denitrification than was methanol (Christensson et al. 1994). However, both methanol and ethanol are flammable carbon sources, presenting security hazards for transportation, storage and operation (Li et al. 2016). Thus, glucose and acetate received more attention in recent years.

Glucose and acetate proved to be beneficial for denitrication (Calderer et al. 2010; Chen et al. 2015; Xu et al. 2018). It was reported that high denitrification rate could be obtained with acetate as carbon source instead of glucose (Yang et al. 2012). However, in the acetate system, pH often increased significantly to exceed the limit of 9.0 (Yang et al. 2012; Ma et al. 2015). As regards glucose used as carbon source, relatively low denitrification rate and undesirable accumulation of NO2-N and NH4+-N hindered its application (Srinandan et al. 2012). Although utilizing sodium acetate or glucose as the single carbon source has its own limitations, employing their mixture as carbon source may be a reasonable alternative. Unfortunately, the denitrification performance of using binary mixture of sodium acetate and glucose as carbon source remained unclear.

In this study, sodium acetate, glucose and their mixture were adopted as external carbon source to promote biological denitrification, respectively. Denitrification performance with different carbon sources was explored. Special attention has been paid to effluent pH. Meanwhile, microbial community was investigated to shed light on the microbial utilization mechanism of different carbon sources. In addition, cost estimation with different carbon sources was carried out.

MATERIALS AND METHODS

Experimental materials

The experiment set-up was shown in Figure 1. Three identical up-flow anaerobic sludge bed (UASB) reactors with an internal diameter of 70 mm and a height of 280 mm were designed. The working volume of each reactor was 1 L. The synthetic influent for the three reactors was mainly composed of NaNO3, different carbon source, KH2PO4 and trace element solution. Sodium acetate, glucose and their mixture were used as carbon sources in the influent, respectively. To simulate the condition of low C/N ratio, the influent NO3-N concentration was kept at 50 mg/L and the C/N ratio was set as 1.5 during the whole experiment. The detailed characteristics of the influent were shown in Table 1.

Table 1

Characteristics of synthetic influent

Influent characteristicsSodium acetate reactorSodium acetate/glucose reactorGlucose reactor
NaNO3 (mg/L) 303.57 303.57 303.57 
Sodium acetate (mg/L) 256.25 64.06 
C6H12O6 (mg/L) 140.75 187.67 
KH2PO4 (mg/L) 
pH 7.3 ± 0.2 7.3 ± 0.2 7.3 ± 0.2 
Influent characteristicsSodium acetate reactorSodium acetate/glucose reactorGlucose reactor
NaNO3 (mg/L) 303.57 303.57 303.57 
Sodium acetate (mg/L) 256.25 64.06 
C6H12O6 (mg/L) 140.75 187.67 
KH2PO4 (mg/L) 
pH 7.3 ± 0.2 7.3 ± 0.2 7.3 ± 0.2 
Figure 1

Schematic diagram of the experimental set-up.

Figure 1

Schematic diagram of the experimental set-up.

The inoculated sludge was obtained from an anoxic tank of an industrial wastewater treatment plant in Chengdu, China. The ratio of volatile suspended sludge (MLVSS) to total suspended sludge (TSS) was 0.61 with an initial MLVSS of 5,730 mg/L.

Experimental procedures

All reactors were operated at 30 ± 2 °C. The synthetic wastewater was pumped into the bottom of the reactor with a peristaltic pump. Three reactors (i.e. the glucose, sodium acetate and sodium acetate/glucose reactor) were operated under a continuous mode with a hydraulic retention time (HRT) of 4 h. The steady-operation phase lasted for 48 days.

Analytical methods

Prior to analysis, samples were filtered through a 0.45 μm polyethersulfone membrane. NO3-N was analyzed by the phenol disulfonic acid method. NO2-N was measured by the N-(1-naphthalene)-diaminoethane photometry method. pH was measured with a pH electrode (Thermo, 3-star 310p-02, USA). MLSS and MLVSS were measured according to the standard methods (APHA et al. 2005). All tests were performed in triplicate.

The specific denitrification rate (SDR, mg N/g VSS d) was defined as follows:  
formula
(1)
where Cin is the initial NO3-N + NO2-N concentration (mg/L); Cef is the effluent NO3-N + NO2-N concentration (mg/L); Q is the flow rate (L/d); V is the reactor volume (L); and XV is the MLVSS concentration (g/L).

Microbial community analysis

DNA extraction and PCR

Four sludge samples (R0, R1, R2 and R3) were collected. R0 represented the inoculated sludge and R1, R2 and R3 represented the sludge sample collected from the sodium acetate, sodium acetate/glucose and glucose reactor on the end of experiments, respectively. All samples were freeze-dried by a lyophilizer (LABCONCO Co., Free Zone, USA). DNA was extracted from dried sludge using the E.Z.N.ATM Mag-Bind Soil DNA Kit (Omega Bio-tek, Norcross, GA, USA) following the manufacturer's protocol. DNA concentration and purity of the extracted DNA were evaluated by using a spectrophotometer (UV2700, Shimadzu, Japan) and checked by agarose gel (1%) electrophoresis.

The bacterial 16S rRNA genes were PCR-amplified with forward primer 341F (CCTACGGGNGGCWGCAG) and reverse primer 805R (GACTACHVGGGTATCTAATCC) for the V3 and V4 region. The PCR reaction mixture contained 15 µL 2 × Taq master mix, 1 µL forward primer (10 µM), 1 µL reverse primer (5 µmol/L), 10 ng genomic DNA and added H2O to 30 µL. The PCR thermal programs consisted of an initial denaturation at 94 °C for 3 min, then followed by the following conditions: five cycles of denaturing at 94 °C for 30 s, 45 °C annealing at for 20 s, and extension 65 °C for 30 s; 20 cycles of denaturing at 94 °C for 20 s, 55 °C annealing at for 20 s, and extension72 °C for 30 s, and a final extension at 72 °C for 5 min.

High-throughput sequencing and data analysis

Sequencing of amplicons from all samples was carried out on Illumina MiSeq PE300 platform. The efficient readings were selected based on the barcode and primers at the beginning and end. After removing low quality sequences and chimeras, the effective sequences of R0, R1, R2 and R3 were 63,578, 69,886, 76,645 and 81,301, the mean lengths were 419.60, 422.61, 417.83 and 420.07 bp, respectively.

Operational taxonomic units (OTUs) used 97% identity thresholds (i.e. 3% dissimilarity levels) in sequences by Usearch (Usearch 5.2.236). The species richness, rarefaction and Shannon diversity index were generated by Mothur (Mothur 1.30.1) and the taxonomy was assigned via the RDP classifier with the Silva databases.

Statistical methods

Data analyses were performed for all samples and the mean values with standard deviation are presented. One-way-analysis of variance (ANOVA) was conducted using the statistical software package SAS 9.0. For values of p < 0.05, the statistical analysis was considered significant.

RESULTS AND DISCUSSION

Denitrification performance

As is shown in Figure 2, the effluent NO3-N concentrations of the sodium acetate, glucose and sodium acetate/glucose reactor were 0.93, 12.66 and 1.74 mg/L, respectively. The corresponding NO3-N removal rates were 98.15%, 74.69% and 96.53%, respectively.

Figure 2

NO3-N removal performance with different carbon sources. Data are the mean values of different measurements. Error bars represent standard deviations of statistical analysis. n means measurement times. The asterisk (*) indicates statistical differences (p < 0.05) from the sodium acetate reactor.

Figure 2

NO3-N removal performance with different carbon sources. Data are the mean values of different measurements. Error bars represent standard deviations of statistical analysis. n means measurement times. The asterisk (*) indicates statistical differences (p < 0.05) from the sodium acetate reactor.

Only taking nitrate removal into account, sodium acetate was evidently the best and glucose the worst carbon source. Zhang et al. (2016a) reported a similar conclusion that when the initial NO3-N concentration was 40 mg/L and the HRT was 6 h, the effluent NO3-N concentration was 22.2 mg/L with glucose as carbon source and less than 1 mg/L with sodium acetate as carbon source. This can be explained as follows. Firstly, compared to sodium acetate, less NO3-N could be reduced by glucose under the same C/N ratio. In the case of sodium acetate as carbon source, the C/N ratio required for the complete reduction of NO3-N to N2 by heterotrophic denitrification is 1.5, as calculated in Equation (2) (Huang et al. 2013):  
formula
(2)
However, the C/N ratio increased to 1.56 if glucose was used as a carbon source, as calculated in Equation (3) (Henze 1991):  
formula
(3)

Secondly, acetic acid, as the hydrolysate of acetate, can be directly inserted into the metabolic process without modification (Elefsiniotis et al. 2004), whereas glucose requires some enzymatic conversion before it can enter the metabolism of most denitrifying bacteria. As is known, propionic acid fermentation and butyric acid fermentation are two main fermentation types. Whichever fermentation type it is, the end product of fermentation has other ingredients besides acetic acid. In addition, glucose could not be fully fermented in time due to the relatively short HRT (4 h) used in this study.

The NO2-N and NH4+-N concentration profiles of different carbon sources were depicted in Figure 3.

Figure 3

NO2-N and NH4+-N concentration profiles with different carbon sources. Data are the mean values of different measurements. Error bars represent standard deviations of statistical analysis. n means measurement times. The asterisk (*) indicates statistical differences (p < 0.05) from the sodium acetate reactor.

Figure 3

NO2-N and NH4+-N concentration profiles with different carbon sources. Data are the mean values of different measurements. Error bars represent standard deviations of statistical analysis. n means measurement times. The asterisk (*) indicates statistical differences (p < 0.05) from the sodium acetate reactor.

The effluent NO2-N concentrations of the sodium acetate, glucose and sodium acetate/glucose reactor were 3.50, 9.00 and 4.93 mg/L, respectively. Obviously, NO2-N accumulated in all reactors. It could be ascribed to insufficient carbon source in the influent with C/N ratio as low as 1.5. The NO2-N accumulation with glucose was the most significant. High nitrate reduction rate than nitrite reduction of the selected species in the glucose-fed reactor could be the possible reason for nitrite accumulation (Chakravarthy et al. 2011). Similar observations had been obtained by Ge et al. (2012) and Yang et al. (2012). However, when mixed carbon source (sodium acetate/glucose) was added, the NO2-N accumulation was greatly alleviated. For NH4+-N, although not contained in the influent, its concentration in the effluent of the sodium acetate, glucose and sodium acetate/glucose reactor reached 0.13, 1.56 and 0.95 mg/L, respectively. The dissimilatory nitrate reduction to ammonium (DNRA) was assumed to be responsible for the NH4+-N production (Hardison et al. 2015). In general, denitrification process is the main competitor of DNRA process, and DNRA phenomenon often appears at higher C/N ratio. However, present experimental results showed that DNRA could happen when C/N ratio was as low as 1.5. Akunna et al. (1993) pointed out that using glucose as carbon source may foster the reduction of NO3-N to NH4+-N whereas using acetic acid as the carbon source was beneficial to the conversion of NO3-N to N2. A similar result was reported by Srinandan et al. (2012). These are in agreement with the results in the present study that effluent NH4+-N concentration of glucose reactor was higher than that of sodium acetate reactor. Herein, it is worth noting that using binary mixtures of sodium acetate and glucose produced less NO2-N and NH4+-N, compared with using glucose alone.

Comparison of denitrification performance

Denitrification performance observed in this study was compared with other published studies using acetate or glucose for denitrification (Table 2).

Table 2

Comparison of SDRs using acetate/glucose from other literature and this study

Carbon sourceOperational modeInitial nitrate concentration (mg/L)SDR (mg N/g VSS d)C/NHRT (h)T (°C)Source of inoculaReference
Acetate Batch 1,000 220a 2.0 15.8 37 An up-flow denitrification sludge blanket reactor Sánchez et al. (2000)  
Acetate Batch 1,000 350a 1.3 15.8 37   
Acetate Batch 750 238a 1.94 23 A wastewater treatment plant Bilanovic et al. (1999)  
Glucose Batch 200 64.8b 5.4 96 30 A pilot anaerobic digester Akunna et al. (1993)  
Acetate Continuous 50 47.7b 1.5 30 An anoxic tank of an industrial wastewater treatment plant The present study 
Glucose Continuous 50 29.7b 1.5 30   
Acetate + glucose Continuous 50 45.4b 1.5 30   
Carbon sourceOperational modeInitial nitrate concentration (mg/L)SDR (mg N/g VSS d)C/NHRT (h)T (°C)Source of inoculaReference
Acetate Batch 1,000 220a 2.0 15.8 37 An up-flow denitrification sludge blanket reactor Sánchez et al. (2000)  
Acetate Batch 1,000 350a 1.3 15.8 37   
Acetate Batch 750 238a 1.94 23 A wastewater treatment plant Bilanovic et al. (1999)  
Glucose Batch 200 64.8b 5.4 96 30 A pilot anaerobic digester Akunna et al. (1993)  
Acetate Continuous 50 47.7b 1.5 30 An anoxic tank of an industrial wastewater treatment plant The present study 
Glucose Continuous 50 29.7b 1.5 30   
Acetate + glucose Continuous 50 45.4b 1.5 30   

aMaximum value.

bAverage value.

By a simple comparison, the SDR values obtained in present study were lower than those in the references listed. This can be largely ascribed to lower nitrate concentrations or lower C/N ratios in this study compared to other studies. In fact, SDR values are correlated with many experimental factors, such as carbon source, influent nitrate concentration, C/N ratio, HRT, temperature, inoculum, reactor's operational mode and so on; thereby the reported SDR data in different literature are usually scattered so that an in-depth comparison becomes difficult.

Among various factors influencing the denitrification performance, C/N ratio and HRT are of great concern. Thereupon, the following discussion focused on these two factors.

In terms of C/N ratio, it has been referred as a key factor to influence denitrification process. Akunna et al. (1992) reported that in a continuous flow anaerobic reactor fed with nitrate and glucose, methane production without denitrification took place and DNRA was the main pathway for nitrate reduction at C/N > 19.9, denitrification and methane production occurred simultaneously at 3.3 ≤ C/N ≤ 19.9 and denitrification without methane production happened at C/N < 3.3. Similarly, Ruiz et al. (2006) found that C/N ratio showed a strong influence on the biomass activity, and therefore on the metabolic pathways of nitrate and organic matter utilization in UASB reactors fed with nitrate and sodium acetate. At C/N = 0.375, nitrate removal was poor and serious nitrite accumulation occurred; whereas at C/N > 1.875, nitrate was practically 100% eliminated. At C/N ≤ 3.75, denitrification represented by far the main rout of organic matter consumption. At C/N = 37.5, over 97% of the organic carbon source was used to produce methane. Chiu & Chung (2003) investigated the optimal C/N ratio for complete denitrifications in a semi-continuous flow reactor with sodium acetate as carbon source. They concluded that the lower the initial nitrate concentration, the higher the corresponding optimal C/N ratio. Specifically, the optimal C/N ratios for initial nitrate concentrations of 25, 50, 100, and 200 mg/L were 5.5, 4.5, 4.0 and 2.6, respectively. In addition, one of our previous works demonstrated that 99.5% nitrate removal could be achieved at a C/N ratio as low as 1.5 in a continuous-operation biofilm-electrode reactor with sodium acetate as the carbon source (Peng et al. 2015). In that case, the heterotrophic denitrification accounted for 92% of nitrate removal. Therefore, the C/N ratio was set as 1.5 in the present study and pretty good nitrate removal efficiency (96.53%) was obtained. In theory, the minimum C/N ratio required for complete heterotrophic denitrification is 1.5 and 1.56, corresponding to utilizing acetate and glucose as carbon source, respectively, hence a C/N ratio lower than 1.5 is unreasonable. It is very likely that denitrification will be impeded due to insufficient carbon source, which was verified in our previous work (Peng et al. 2015). A sharp drop of nitrate removal rate from 99.5% to 64.1% was observed as the C/N ratio decreased from 1.5 to 0.8. On the other hand, from the perspective of further optimization, a somewhat larger C/N ratio over 1.5 would probably be better. In practice, more carbon will be demanded as it is not only used for respiration, but also for cell growth and maintenance (Constantin & Fick 1997).

As regards HRT, its effects on denitrification were investigated in our earlier work (Wu et al. 2018). In that work, the influent nitrate concentration of 50 mg/L and the C/N ratio of 1.5 were kept unchanged in a UASB reactor and different HRTs (8, 6, 4, 2 h) were tested in order. To attain the highest removal rates of NO3-N and total nitrogen, the optimal HRT was found to be 6 h with glucose as the carbon source while it became 4 h as the carbon source was converted to sodium acetate. A conclusion was drawn that HRT, combined with carbon source types, had a significant effect on denitrification and HRT affected the balance among denitrifying bacteria, bacteria responsible for DNRA and other heterotrophic bacteria. In the present study, HRT was set as 4 h considering the mixed carbon source employed was composed of sodium acetate and glucose. What can be expected is further reduction of HRT is inadvisable. If HRT gets too short, a portion of glucose in the form of intermediate metabolites will be washed out of the reactor at the stage of acidification fermentation, as a result, they cannot provide electrons for denitrification in time. On the other hand, appropriate extension of HRT in the range of 4–6 h may be desirable and worth further research.

Overall, it must be noted that the optimal C/N ratio and HRT for a biological denitrification system treating a specific wastewater under special conditions should be determined experimentally.

Effluent pH

pH is a very important index related to denitrification performance (Ge et al. 2012). The effluent pH values of the sodium acetate, glucose and sodium acetate/glucose reactor varied in the range of 9.13–9.60, 7.32–7.60 and 7.80–8.23, respectively. This observation is in agreement with previous studies (Yang et al. 2012; Ma et al. 2015). In theory, regardless of what type of carbon source used, the total amount of OH produced by denitrification to remove the same NO3-N is equal. However, the difference between the effluent pH values of the sodium acetate and glucose reactor was significant. With the same pH (7.3 ± 0.2) in the influent, the effluent pH of the sodium acetate reactor increased to exceed the discharge standard limit of 9.0, whereas that of the glucose reactor changed little. This difference could be explained from two aspects. Firstly, sodium acetate, as a strong base weak acid salt, once dissolved in water, produced OH. Secondly, CO2 was produced in the fermentation of glucose and the resulting CO32− or HCO3 could buffer the system's pH. The experimental results demonstrated that keeping pH in a reasonable range was one of the advantages of mixed carbon sources.

Microbial community

Diversity of microbial communities

To explore the microbial utilization mechanisms of different carbon sources, the 16S rRNA gene based high-throughput sequencing was used to analyze sludge samples. The Shannon indexes of R0, R1, R2, and R3 were 6.41, 5.53, 5.74 and 5.34, respectively. The larger the Shannon value, the higher the community diversity. Obviously, the Shannon index of the mixed carbon sources was increased compared to that of a single carbon source. This result hinted that the mixed carbon source was beneficial to the improvement of microbial community diversity, which was linked to the stability of the microbial system (Hagman et al. 2008; Zhang et al. 2016b).

In order to deeply explore the effect of carbon sources on microbial community, sludge samples were analyzed at phylum level (Figure 4).

Figure 4

Distribution of bacterial community with different carbon sources at phylum level.

Figure 4

Distribution of bacterial community with different carbon sources at phylum level.

Compared with R0, Proteobacteria and Firmicutes were enriched in R1, R2 and R3. The relative abundance of Proteobacteria and Firmicutes in R0 were 41.8% and 1.54%, respectively; whereas the relative abundance of Proteobacteria and Firmicutes were 61.13% and 4.85% in R1, 60.39% and 2.53% in R2, and 68.69% and 2.24% in R3, respectively. The microbial community of denitrification was mainly distributed in Proteobacteria (Zhang et al. 2012; Ma et al. 2015), and Firmicutes was found to be responsible for denitrification process under anaerobic conditions and could survive under the conditions lacking organic substrates (Wang et al. 2009). By contrast, the relative abundance of Bacteroidetes, Planctomycetes and Acidobacteria decreased from 19.54%, 8.16% and 8.27% in R0 to 10%, 3.36% and 3.52% in R1, 14.22%, 3.01% and 2.67% in R2, and 13.39%, 2.32% and 2.14% in R3, respectively. Bacteroidetes, Planctomycetes and Acidobacteria were found in granular sludge bed reactor via autotrophic nitrogen removal process (Wang et al. 2012). However, the difference of above-mentioned phyla among R1, R2 and R3 was not significant.

The heatmap of genera with different carbon sources is shown in Figure 5. For the sodium acetate reactor (R1), Thauera (15.03%) and Comamonas (12.30%) were the two dominant genera. For the glucose reactor (R3), Rhizobium (12.23%), Comamonas (5.95%), Aeromonas (4.71%), Acidovorax (4.39%) and Elstera (3.53%) were the dominant genera. As regards the sodium acetate/glucose reactor (R2), Rhizobium (13.15%), Elstera (5.05%), Gemmobacter (3.87%) and Comamonas (3.83%) were the dominant genera. Among the abovementioned genera, only Thauera (2.22%) and Comamonas (3.84%) were the dominant genera for R0. Thauera was reported to be the dominant denitrifying bacteria in acetate supported denitrification system (Ginige et al. 2005; Osaka et al. 2006). Su et al. (2020) discovered that Comamonas sp. exhibited excellent nitrogen removal ability under low C/N conditions (C/N = 2.5), with no accumulation of nitrite. The important role of Comamonas in the nitrogen removal was also confirmed in present study. Under a C/N ratio as low as 1.5, the relative abundance of Comamonas in R1 was the highest (12.30%), meanwhile the NO3-N removal rate of R1 was the highest (98.15%) with the lowest effluent nitrite concentration. Compared with glucose, sodium acetate proved to be more favorable for enriching Comamonas. Considering the relative abundance of Rhizobium in R2 (13.15%) and R3 (12.23%), which were markedly higher than that in R1 (0.37%) and R0 (0.01%), it is inferred that the enrichment of Rhizobium was closely correlated with glucose. According to Gonçalves Pessoa et al. (2019), the genus Aeromonas comprises D-glucose fermentative facultative anaerobic bacteria that are D-glucose fermenters with or without gas production. In addition, they are catalase and oxidase positive, reduce nitrate to nitrite, produce several enzymes and are capable of using other carbohydrates besides glucose. For Acidovorax, it belongs to the class Betaproteobacteria, and was reported to be the most dominant genus at a low COD to nitrate ratio of 1:1 (Rungkitwatananukul et al. 2016). In the present study, more Acidovorax was enriched with glucose than with sodium acetate. With regard to Elstera, it belongs to the family Rhodospirillaceae and class α-proteobacteria, comprising just two species so far, i.e. Elstera cyanobacteriorum and Elstera litoralis. E. cyanobacteriorum was found to be positive for nitrate reduction and denitrification (Cai et al. 2018). Besides, Gemmobacter exhibited a relative abundance of 3.87% in R2 and 1.35% in R3. Gemmobacter belonged to the class α-Proteobacteria (Liu et al. 2014), and had been identified as denitrifying bacteria (Wei et al. 2017). These results mentioned above indicated that different carbon sources led to a different microbial community structure at the genus level.

Figure 5

Heatmap of genera with different carbon sources.

Figure 5

Heatmap of genera with different carbon sources.

Cost estimation

Actually, the choice of carbon sources in sewage treatment plants is often affected by economic factors. Based on the denitrification performance of sodium acetate, sodium acetate/glucose and glucose system, the economic analysis of different carbon sources was conducted. It is assumed that the treatment scale of denitrification system is 1,000 m3/d and the characteristics of the influent are the same as presented in this study, aiming to obtain an effluent nitrate concentration of 5 mg/L. Sodium acetate, sodium acetate/glucose and glucose are respectively used as the external carbon sources for denitrification. The estimated cost with different carbon sources was presented in Table 3.

Table 3

Carbon source cost with different carbon sources

ItemValueUnitRemark
Operation parameter 
 Wastewater treatment capacity 1,000 m3/d  
 Initial concentration 50 mg/L  
 Effluent concentration mg/L  
Price of carbon source 
 Sodium acetate 2,500 RMB/ton Purity (58%) 
 Glucose 2,400 RMB/ton Purity (99%) 
Cost of carbon source 
 Sodium acetate 1,005 RMB/d Sodium acetate (402 kg/d) 
 Sodium acetate/glucose 642.7 RMB/d Sodium acetate (115 kg/d)/glucose (148 kg/d) 
 Glucose 669.6 RMB/d Glucose (279 kg/d) 
ItemValueUnitRemark
Operation parameter 
 Wastewater treatment capacity 1,000 m3/d  
 Initial concentration 50 mg/L  
 Effluent concentration mg/L  
Price of carbon source 
 Sodium acetate 2,500 RMB/ton Purity (58%) 
 Glucose 2,400 RMB/ton Purity (99%) 
Cost of carbon source 
 Sodium acetate 1,005 RMB/d Sodium acetate (402 kg/d) 
 Sodium acetate/glucose 642.7 RMB/d Sodium acetate (115 kg/d)/glucose (148 kg/d) 
 Glucose 669.6 RMB/d Glucose (279 kg/d) 

Clearly, the cost of sodium acetate/glucose was the lowest, which was about 0.64 RMB/m3. Compared with sodium acetate and glucose, sodium acetate/glucose could reduce the carbon source cost by 0.36 and 0.03 RMB/m3, respectively. In addition, compared with sodium acetate, sodium acetate/glucose could save the chemical cost for pH adjustment.

CONCLUSIONS

Glucose and acetate are traditional external carbon sources for nitrate removal. However, they both have inherent flaws, making a preferable alternative to be expected. This study aimed to investigate the comprehensive performance of the binary mixture of sodium acetate and glucose for deep denitrification. The experimental results revealed that the mixed carbon source was advantageous over each single component. Compared with glucose, the mixture exhibited higher denitrification rate and produced less undesirable NO2-N and NH4+-N. Compared with sodium acetate, the mixture afforded an effluent pH in compliance with emission standards. In addition, the biodiversity with the mixed carbon source was the highest, whereas its cost was the lowest. These findings contribute to a reasonable selection of carbon sources for deep denitrification in practice. However, the optimizations of C/N ratio and HRT were not included in the present study. These issues deserve to be addressed in further research.

ACKNOWLEDGEMENT

This work was funded by the Education Department of Sichuan Province (Project No. 16ZB0051).

REFERENCES

REFERENCES
Akunna
J. C.
Bizeau
C.
Moletta
R.
1992
Denitrification in anaerobic digesters: possibilities and influence of wastewater COD/N-NOx ratio
.
Environmental Technology
13
(
9
),
825
836
.
APHA, AWWA & WEF
2005
Standard Methods for the Examination of Water and Wastewater
, 21st edn.
American Public Health Association/American Water Works Association/Water Environment Federation
,
Washington, DC
,
USA
.
Bilanovic
D.
Battistoni
P.
Cecchi
F.
Pavan
P.
Mata-Alvarez
J.
1999
Denitrification under high nitrate concentration and alternating anoxic conditions
.
Water Research
33
(
15
),
3311
3320
.
Cai
H. Y.
Cui
H. L.
Zeng
Y. H.
An
M. L.
Jiang
H. L.
2018
Sandarakinorhabdus cyanobacteriorum sp. nov., a novel bacterium isolated from cyanobacterial aggregates in a eutrophic lake
.
International Journal of Systematic and Evolutionary Microbiology
68
(
3
),
730
735
.
Calderer
M.
Gibert
O.
Marti
V.
Rovira
M.
de Pablo
J.
Jordana
S.
Duro
L.
Guimera
J.
Bruno
J.
2010
Denitrification in presence of acetate and glucose for bioremediation of nitrate-contaminated groundwater
.
Environmental Technology
31
(
7
),
799
814
.
Chakravarthy
S. S.
Pande
S.
Kapoor
A.
Nerurkar
A. S.
2011
Comparison of denitrification between Paracoccus sp. and Diaphorobacter sp.
Applied Biochemistry and Biotechnology
165
(
1
),
260
269
.
Chen
H. B.
Wang
D. B.
Li
X. M.
Yang
Q.
Zeng
G. M.
2015
Enhancement of post-anoxic denitrification for biological nutrient removal: effect of different carbon sources
.
Environmental Science and Pollution Research
22
(
8
),
5887
5894
.
Chiu
Y.-C.
Chung
M.-S.
2003
Determination of optimal COD/nitrate ratio for biological denitrification
.
International Biodeterioration & Biodegradation
51
(
1
),
43
49
.
Christensson
M.
Lie
E.
Welander
T.
1994
A comparison between ethanol and methanol as carbon sources for denitrification
.
Water Science and Technology
30
(
6
),
83
90
.
Elefsiniotis
P.
Wareham
D. G.
Smith
M. O.
2004
Use of volatile fatty acids from an acid-phase digester for denitrification
.
Journal of Biotechnology
114
(
3
),
289
297
.
Gonçalves Pessoa
R. B.
de Oliveira
W. F.
Marques
D. S. C.
dos Santos Correia
M. T.
de Carvalho
E. V. M. M.
Coelho
L. C. B. B.
2019
The genus Aeromonas: a general approach
.
Microbial Pathogenesis
130
,
81
94
.
Henze
M.
1991
Capabilities of biological nitrogen removal process from wastewater
.
Water Science and Technology
23
(
4
),
669
679
.
Huang
B. C.
Feng
H. J.
Wang
M. Z.
Li
N.
Cong
Y. Q.
Shen
D. S.
2013
The effect of C/N ratio on nitrogen removal in a bioelectrochemical system
.
Bioresource Technology
132
,
91
98
.
Liu
J.
Zhang
X.
Chi
F.
Pan
J.
Sun
C.
Wu
M.
2014
Gemmobacter megaterium sp. nov., isolated from coastal planktonic seaweeds
.
International Journal of Systematic and Evolutionary Microbiology
64
,
66
71
.
Osaka
T.
Yoshie
S.
Tsuneda
S.
Hirata
A.
Iwami
N.
Inamori
Y.
2006
Identification of acetate- or methanol-assimilating bacteria under nitrate-reducing conditions by stable-isotope probing
.
Microbial Ecology
52
(
2
),
253
266
.
Peng
H.
Liu
W. Y.
Liu
S. X.
Chen
Y. H.
Xiao
H.
2015
Effect of low C/N ratios on denitrification in a biofilm-electrode reactor
.
Chinese Journal of Environmental Engineering
9
(
7
),
3263
3268
(in Chinese)
.
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
.
Rungkitwatananukul
P.
Nomai
S.
Hirakata
Y.
Pungrasmi
W.
Puprasert
C.
Hatamoto
M.
Yamaguchi
T.
2016
Microbial community analysis using MiSeq sequencing in a novel configuration fluidized bed reactor for effective denitrification
.
Bioresource Technology
221
,
677
681
.
Sánchez
M.
Mosquera-Corral
A.
Méndez
R.
Lema
J. M.
2000
Simple methods for the determination of the denitrifying activity of sludges
.
Bioresource Technology
75
(
1
),
1
6
.
Srinandan
C. S.
D'Souza
G.
Srivastava
N.
Nayak
B. B.
Nerurkar
A. S.
2012
Carbon sources influence the nitrate removal activity, community structure and biofilm architecture
.
Bioresource Technology
117
,
292
299
.
Su
J. F.
Yang
S.
Huang
T. L.
Li
M.
Liu
J. R.
Yao
Y. X.
2020
Enhancement of the denitrification in low C/N condition and its mechanism by a novel isolated Comamonas sp. YSF15
.
Environmental Pollution
256
,
113294
.
Wang
L.
Zheng
P.
Chen
T. T.
Chen
J. W.
Xing
Y. J.
Ji
Q. X.
Zhang
M.
Zhang
J. Q.
2012
Performance of autotrophic nitrogen removal in the granular sludge bed reactor
.
Bioresource Technology
123
,
78
85
.
Wu
J. J.
Li
Y. M.
Tan
D. M.
Peng
H.
Xiao
H.
2018
Effects of HRT on anaerobic denitrification in UASB
.
Chinese Journal of Environmental Engineering
12
(
5
),
1510
1516
(in Chinese)
.
Zhang
Y. M.
Wang
X. C. C.
Cheng
Z.
Li
Y. Y.
Tang
J. L.
2016b
Effects of additional fermented food wastes on nitrogen removal enhancement and sludge characteristics in a sequential batch reactor for wastewater treatment
.
Environmental Science and Pollution Research
23
(
13
),
12890
12899
.
Zhang
R.-C.
Xu
X.-J.
Chen
C.
Xing
D.-F.
Shao
B.
Liu
W.-Z.
Wang
A.-J.
Lee
D.-J.
Ren
N.-Q.
2018
Interactions of functional bacteria and their contributions to the performance in integrated autotrophic and heterotrophic denitrification
.
Water Research
143
,
355
366
.