Hydrogenotrophic denitri ﬁ cation for treating nitrate contaminated without/with reactive black 5 dye

NO 3 -N and dye colors discharged from textile wastewater pose environmental problems in Thailand. This study aimed to observe the nitrogen removal rate (NRR) with and without RB-5 color contamination via hydrogenotrophic denitri ﬁ cation (HD) processing, which uses H 2 gas as electron donor to reduce NO 3 -N and NO 2 -N; comparing with bioreactors treatment to evaluate systems that can simultaneously remove NO 3 -N and dye color. Five reactors under different operation and gas supply conditions were set-up under HRT of 24 h, including an aerobic reactor using air, two anaerobic reactors using argon and H 2, and a combined process using intermittent air/argon and air/ H 2 . NRR without dye varied between 45 and 90% for H 2 and air/H 2 by HD processing, while it was completely removed when adding color. H 2 and air/H 2 reactors experienced partial decolorization of approximately 20 – 30%, whereas the other three reactors remained unchanged. Ef ﬂ uent of NO 3 -N were close to wastewater standards, but the color was still easy to detect, which indicated that the treatment time needs to be suf ﬁ cient. In conclusion, HD and intermittent air/H 2 processing can completely remove NO 3 -N and NO 2 -N when contaminated with RB-5 color. Furthermore, RB-5 did not affect the NRR, whereas some particles of dye color can also reduce in these processes.


INTRODUCTION
Thailand is a newly industrialized and developing country, and textiles are one of its most important export industries, with 14,400 tons/yr of leather and textile fibers being produced and exported. Various processing stages such as bleaching, dyeing, printing, finishing, sizing and stiffening use chemical reagents and large volumes of water, thereby generating huge quantities of effluent wastewater that require discharging. Presently, the textile industry in Thailand generates approximately 2.55 × 10 6 m 3 /yr of wastewater contaminated with various toxic pollutants, in particular, nitrate (NO 3 -N) (Sahinkaya et al. ; Sairiam et al. ). NO 3 -N is a common pollutant that can be converted into nitrite , which is another toxic pollutant generally found in groundwater, surface water, and wastewater. Normally, NO 3 -N from textile wastewater is produced from salts such as sodium nitrate used in dye baths to improve textile fibers. The concentration of NO 3 -N can be in the range of 40-100 g/L in textile effluents (Cirik et al. ). These levels are problematic and have a harmful impact on groundwater and human health, for example, they cause methemoglobin in children (Rahman et al. ). The World Health Organization (WHO) recommends NO 3 -N and NO 2 -N concentration to be less than 11.3 and 0.91 mg/L in drinking water, respectively (WHO ). Many techniques have been developed to treat NO 3 -N pollution, including chemical-physical processes such as ion exchanger and ultra-filtration reverse osmosis, and biological denitrification. The biological denitrification system involves both heterotrophic and autotrophic processes and is commonly applied for treating NO 3 -N and NO 2 -N. The system is based on the respiration of facultative bacteria as heterotrophic denitrification bacteria (HDB) and autotrophic denitrification bacteria (ADB), with NO 3 -N used as an electron acceptor under anaerobic and anoxic conditions. Hydrogenotrophic denitrification (HD) or hydrogenbased autotrophic denitrification is a type of treatment processing that is used to reduce inorganic compounds such as NO 3 -N and NO 2 -N using H 2 as electron donors, and bicarbonate or carbon dioxide for biosynthesis. Owing to its low biomass, low residual organic compounds, low system costs and high nitrogen removal efficiency compared with heterotrophic denitrification, HD is most commonly used to treat groundwater (Eamrat ). Theoretical equations for heterotrophic (Equation (1)) and hydrogenotrophic (Equation (2)) denitrification are given below.
Heterotrophic denitrification using methanol as a carbon source (Rezvani et  In the aerobic system, air was constantly supplied to a reactor to observe the effect of all parameters when mixed with oxygen. Two anaerobic reactors were operated using various types of biological denitrification: the first was heterotrophic and operated using argon supplies, while the other used HD processing and was supplied with H 2 gas, which acted as the electron donor. The combined system (anaerobic/aerobic), which is widely used in the removal of dye color and organic matter, was supplied by an intermittent feeding of air/argon and an air/H 2 supply. Reactive Black 5 (RB-5), which is a type of azo dye color and is most commonly used in dyeing processes, was selected as the model dye color combined with NO 3 -N contamination.

Experimental configuration
Five reactors under various operating conditions were set up including an aerobic (with air), two anaerobic (one with argon and another with H 2 ), and a combined system reactor (one with air/argon and another with air/H 2 ); these reactors were used to observe the treatment performance for the removal of NO 3 -N, NO 2 -N, and dye color. All the reactors that were made from plastic were fabricated in a cylindrical form; 3.5 cm in diameter and 28 cm in height. The working volume of each of the reactors was approximately 2 L. All the reactors were fixed inside a water tub to control their temperatures in the range of 30-33 C. Furthermore, the reactors were connected with an influent synthetic wastewater tank, and for gas feeding, the pipe lines were fixed on the operating conditions of the reactors. The schematic of the experimental configuration of the reactors is depicted in Figure 1.

Preparation of synthetic textile wastewater
Synthetic wastewater was prepared using tap water based on the actual concentrations and the chemical reagents (g/L) adapted from a previous study (Panswad & Luangdilok ). The composition of wastewater is presented in Table 2. The influent NO 3 -N concentrations considered in this study ranged from 40, 50 and reached 80 mg NO 3 -N/L at the end of the experiment. In the second phase, RB-5 dye color was added into these reactors. The dye concentration

Influent loading rate and gas supply conditions
To examine the influent loading rates, experiments were conducted in two phases: NO 3 -N without RB-5 contamination, and NO 3 -N with RB-5 contamination. The treatment performance of each reactor was tested for 24 h, in a continuous feeding mode. At the beginning, the enhanced sludge from the HD reactor, which was operating for more than 400 d, was added to each reactor. This reactor was operated at an HRT of 8 h at a controlled temperature of 30 C and fed with influent water at approximately 8.7 mL/min, with an H 2 supply of 30 mL/min. The nitrogen loading rate (NLR) was 313 g-N/m 3 /d. The HD reactor showed the best performance, with the removal rates of NO 3 -N and NO 2 -N close to 85-95%. The enriched sludge was added between 0.2 and 0.4 g of volatile suspended solids (VSS)/L to each reactor. In the first phase in days 1-7, the air, argon, air/argon, and air/H 2 gas reactors were set up to maintain an NLR of 40 g-N/m 3 /d. On day 8, the H 2 reactor was constructed, and an NLR of 50 g-N/m 3 /d was maintained in all the reactors. On day 18, an NLR of 80 g-N/m 3 /d was maintained in the air, argon, air/argon, and air/H 2 reactors; this change was also done on day 11 for the H 2 reactor. These changes were made to observe the effect on the nitrogen removal rate (NRR) when the influent loading is increased. In the second phase, NO 3 -N with RB-5 was introduced on day 33 for all the reactors except for the H 2 reactor, where the pollutant was introduced on day 27, at 80 mg/L of RB-5 (amount equivalent to NLR). On day 48, for the four reactors, and on day 42, for the H 2 reactor, the RB-5 dose was reduced to 20 mg/L to isolate the effect of NO 3 -N removal on the dye.
For gas supply feeding, the pipelines were fixed depending on the operating conditions of each reactor. One of the aerobic reactors was continuously supplied with air to grow aerobic bacteria, while the two anaerobic reactors were continuously supplied with argon and H 2 gas for bacterial growth and for the comparison of treatment performances under the presence of heterotrophic and hydrogenotrophic denitrifying bacteria, respectively. Of the other two reactors, one was intermittently supplied with air/argon and the other with air/H 2 , and controlled under the combined anaerobic/ aerobic system. Air was released from an air pump to each reactor, while argon was discharged from a cylinder tank and H 2 was supplied from an H 2 generator, via an air stone diffuser. Initially, the volume of the continuous supply of air and argon was controlled at 300 mL/min, and H 2 at 100 mL/min. After a supply of 18 days of air and argon supply, and 12 days of H 2 , it was reduced to   Table 4.

Analytical methods
The influent and effluent from all the reactors were centrifuged (Hitachi CF 16RXII, Japan) at 10,000 rpm for 10 min, and stored in a freezer (-18 C) until the analyses. pH, dis-  (6):

Statistical analysis
To test the differences between the treatments performance of decolorization, all the reactors during the second phase were analyzed by the analysis of variance (ANOVA) and least significant difference (LSD) using the SPSS statistical software tool version 25. Statistical significance was kept at p < 0.05.

Effluent NO 3 -N and NO 2 -N concentrations
In this study, the reactors were continuously fed with synthetic wastewater at an HRT of 24 h to observe the    for NO 3 -N reduced to N 2 gas, which converts to an alkalinity rate of 3.579 g CaCO -3 per 1 g of NO 3 -N reduced. This concentration can lead to an increase in the pH values in these reactors. In contrast, the pH concentrations from the air, argon and air/argon reactors were the same as those in influent feeding, thereby proving that no reaction occurred in these reactors. According to the previous reports, the optimum pH conditions for the HD process ranged from 7.6 to 8.6 (Eamrat ). Similarly, pH levels from 9.13 to 9.93 can lead to high denitrification rates (Li et al. ). Conversely, owing to the lack of carbon sources in air, argon and air/argon can lead to a slower NO 3 -N reduction and bacterial growth compared to the heterotrophic process using organic matter.
In conclusion, HD processing can completely reduce NO 3 -N and NO 2 -N when contaminated with RB-5 dye; in contrast, this dye color did not affect the combined NO 3 -N and NO 2 -N removal. The dye color also did not significantly affect NO 3 -N removal because NO 3 -N had a faster oxidation-reduction reaction than the dye color (Panswad & Luangdilok ; Cirik et al. ). Furthermore, dye color can lead to an increase in the reaction rates of NO 3 -N and NO 2 -N, thus reducing the time required for treatment.

Effluent dye color concentrations
The biological degradation of a dye mostly occurs in anaerobic processes by the reductive cleavage of dye bonds and through an oxidation-reduction reaction that removes aromatic amines in the aerobic process. Conversely, the dye is an electron acceptor with limited reduction when it exists alongside more effective electron acceptors (Assadi et al. color amounts in these reactors decreased at a slightly faster rate than in the other reactors. The dye concentrations in the air, argon and air/argon reactors accumulated at a constant influent feeding rate, as presented in Figure 3(a), 3(b) and 3(d). A lack of bacterial growth incompletely influenced the biodegradation processing. The results show that the decolorization rate was approximately 20-35% in H 2 and 10-20% in air/H 2 reactors, respectively. The average effluent concentrations in the H 2 and air/H 2 reactors were approximately 58 and 67 mg /L at a feeding rate of 80 mg/L of RB-5, and 16 and 18 at a feeding rate of 20 mg/L of RB-5, respectively. However, the decolorization rates remained unchanged when the dye concentration was reduced from 80 to 20 mg/L. Hence, the RB-5 dye color can cause a slight reduction together with NO 3 -N and NO 2 -N during HD processing.
Dissolved organic carbon (DOC) was measured for the influent and effluent to check the bacterial activities in both reactors. The results of DOC increased by 23.4 and 13.7 mg/L for the H 2 and air/H 2 reactors, respectively, while the influent feeding DOC result was 8.5 mg/L. The DOC value may have increased from ADB activity that produced extracellular polymeric substances (ESP), which are biopolymers used to protect cells from external environmental or strength conditions. Bacterial activities in these reactors can help in the removal of NO 3 -N along with the dye color by the process of denitrification. Hence, the metabolism of ADB under HD processing can also generate NADH, which is another factor that facilitates dye color removal using azo-reductases to break azo bonds in the dye molecule, which then act as electron acceptors to produce colorless aromatic amines. One of the advantages of using H 2 gas for decolorization is that its electron-donating substrates can reduce and charge form of azo dye to aromatic amines (Saroyan et al. ). However, the reduction in dye concentrations was very minimal in the H 2 reactor and, to a slightly larger extent, in the air/H 2 reactor in the presence of NO 3 -N. The dye color was removed faster in the presence of lower NO 3 -N concentration (Lourenço et al. ). Conversely, oxygen and NO 3 -N are more active electron acceptors than a dye, and lead to incomplete decolorization (Lee et al. ). Table 1 presents that the actual influent and effluent concentrations of dye color under anaerobic and combined anaerobic/aerobic treatment processing in Thailand were found in the range of 40-500 mg/L reducing to 2-112 mg/L due to effective treatment processing and long-term operation. Hence, HRT is a crucial factor that can improve the treatment efficiency of decolorization using HD and the combined system in this study.
Relationship between nitrogen removal and decolorization rate Table 5 presents the relationship between NRR compared with the decolorization rate from all the reactors under different influent feeding regimes, including actual NO 3 -N and NO 3 -N contaminated with RB-5 dye. NRR and dye were mostly removed in the anaerobic and combined systems when H 2 gas was used during HD processing, whereas the aerobic, anaerobic, and combined systems using air and argon gas remained unchanged.
The average NRR results without and with RB-5 dye were reduced in the H 2 and air/H 2 reactors, at close to 95% in the H 2 reactor without RB-5 and close to 100% with RB-5 dye, although the latter dropped slightly to 90-100% after the azo dye concentration was reduced from 80 to 20 mg/L. Similarly, there were decreases in the nitrogen removal efficiency of the air/H 2 reactor after the RB-5 dye concentration was reduced from 80 to 20 mg/L, at 80 and 42%, respectively.
Hence, the HD process was accelerated when contaminated with dye color while it was delayed when the dye concentrations decreased at this stage. Previous studies suggested that NO 3 -N and NO 2 -N is an electron acceptor similar to dye color but it is highly effective; therefore these substances can compete for electrons with dye color leading to an According to previous studies that evaluated advanced treatment processing to simultaneously remove nitrogen and dye color in a single reactor, the treatment efficiency of both the parameters was varied based on the operation time (i.e. HRT) and the treatment process, such as anaerobic-biofilm anoxic-aerobic membrane bioreactor, sequencing batch reactor, anaerobic reactor, up-flow anaerobic filter, and hydrolysis/acidification, or multiple anoxic/aerobic process.
The results obtained in most of these studies showed the complete removal of nitrogen, while the decolorization rate was found to be in the range of 20-100%, except in the case of the membrane bioreactor and up-flow anaerobic filter, which exhibited high decolorization rates. However, the combined system is generally used to improve the quality of textile wastewater.

CONCLUSIONS
HD processing was performed under different operating and gas supply conditions; the reactors included an aerobic, two anaerobic (one supplied with argon, and the other with H 2 ) and two combined systems (one with intermittent air/argon, and the other with air/H 2 ). The results of NRR and decolorization rate showed high treatment efficiencies of the anaerobic processes and combined systems that used H 2 and air/H 2 as electron donors. The concentrations of NO 3 -N and dye color were significantly and simultaneously reduced when HD processing was performed in a single reactor. However, in the air, argon and air/argon reactors, no oxidation-reduction reactions occurred owing to a lack of bacterial growth. Furthermore, H 2 reactors showed higher NRR than the air/H 2 reactors because oxygen negatively affected the HD process. The NRR and decolorization rates were high in H 2 reactors when both NO 3 -N and RB-5 dye color were presented, thereby providing optimal conditions for H 2 to donate electrons. The performance of the H 2 reactors in terms of nitrogen removal was 90-100% and the decolorization ranged from 20 to 35%, whereas the air/H 2 reactor showed a 40-100% NRR and a 10-20% decolorization rate after an HRT of 24 h. This is because NO 3 -N is a better electron acceptor compared to the azo dye, and therefore, the decolorization rate was low. However, there was a reduction in the nitrogen removal efficiency after the concentration of RB-5 in the influent was reduced. The dye color may compete with NO 3 -N and NO 2 -N to improve the oxidation-reduction reaction rates of nitrogen. Hence, HD processing can still completely remove NO 3 -N and NO 2 -N when contaminated with RB-5 dye color, and RB-5 did not affect the NO 3 -N and NO 2 -N removal rates. However, the concentration of NO 3 -N after processing met the effluent wastewater standard in Thailand, and the concentration of color was easy to detect even in contaminated wastewater. In the future, HD and combined systems should use appropriate H 2 levels and sufficient treatment time to help facilitate the reduction of the concentration of dye color in a single reactor. Furthermore, these systems should be applied to textile wastewater to observe their performance in a real situation. This method provides an opportunity to improve and develop advanced technologies that can lower treatment costs, ensure sustainability and improve wastewater quality before being discharged, as well as reduce freshwater consumption in developing countries like Thailand.
for this research through the Science and Technology Research Partnership for Sustainable Development (SATREPS) program of JST and JICA.

DATA AVAILABILITY STATEMENT
Data cannot be made publicly available; readers should contact the corresponding author for details.