Cost Reduction in the Process of Biological Denitrification by Choosing Traditional or Alternative Carbon Sources

Abstract

Biological denitrification is a generally economically viable and reliable technology for nitrate (NO₃⁻-N) and nitrite (NO₂⁻-N) removal from wastewater. Nitrate/nitrite reduction requires an electron donor, which can be supplied from influent wastewater organic compounds, added as an external carbon source, or by endogenous respiration. Decisions regarding the selection of an external carbon source should take into consideration its cost, sludge generation quantity, the rate and efficiency of denitrification, and chemical storage safety. The expense of waste management and external carbon sources accounts for up to 50% of the overall cost of wastewater treatment. The carbon source characteristics required for biological denitrification affect the microbial community structure, denitrification rate, and intermediate products. This review is based on a bibliometric analysis and a systematic literature review providing information and insight into the topic of the denitrification process using different carbon sources. In this review, the main focus was on discussing alternative carbon sources vs. traditional carbon sources in terms of the carbon source price, C/N ratio, denitrification efficiency and rate, operational parameters, and advantages and disadvantages, as well as the limitations in the denitrification process. Future perspectives may include the operating parameters influencing the stability of the removal performance; the maintenance and improvement of nitrate removal; a study of the diversity of the microbial community; research on the application of new alternative carbon sources in denitrification; and N₂O emission detection and minimisation.

1. Introduction

Denitrification is carried out by group of organisms called denitrification bacteria that can use nitrate (NO₃⁻-N) and/or nitrite (NO₂⁻-N) nitrogen as a terminal electron acceptor in the absence of dissolved oxygen (DO) [1]. The majority of denitrifying bacteria are heterotrophs that need organic carbon as an electron donor for the reduction of nitrate/nitrite, and for cell growth [2]. Biological denitrification includes the sequential reactions of nitrate reduction to the final product, gaseous nitrogen (N₂), driven by several reductases. Respiratory denitrification begins with the one-electron reduction of NO₃⁻ to NO₂⁻, catalysed by membrane-associated dissimilatory nitrate reductase enzymes, encoded by the nas, nar, or nap genes [3,4], with narG being commonly employed for nitrate reduction investigation [3]. Then, the one-electron reduction of NO₂⁻ to nitric oxide (NO) is catalysed by one of two non-homologous nitrite reductases, a copper-containing (nirK) or a cytochrome cd₁-containing (nirS) nitrite reductase [5]. Denitrifiers containing nirK are more varied and susceptible to variations in environment in comparison to nirS denitrifiers [6]. NO is further reduced to nitrous oxide (N₂O) by cytochrome bc nitric oxide reductase, which is encoded by the nor gene [3,4]. N₂O is reduced to N₂ by nitrous oxide reductase, encoded by the nosZ gene [3,4,7,8].

Carbon sources for biological denitrification include (i) organics present in wastewater [9,10,11,12,13,14,15]; (ii) external carbon sources [16,17,18,19,20,21,22]; and (iii) intracellular carbon storage [23,24,25,26,27,28,29,30,31,32]. Decisions regarding the selection of an external carbon source should take into consideration its cost, sludge generation quantity, the rate and efficiency of denitrification, and chemical storage safety [33]. The expense of waste management and external carbon source accounts for up to 50% of the overall cost of wastewater treatment [34,35]. Issues such as the handling and storage of carbon sources can be substantially lowered by disposing and reutilising the “wastes” [36,37]. The carbon source most often used is methanol because of the relatively lower required ratio of methanol to nitrate, the higher denitrification efficiency, wide accessibility on the market, and lower price [38]. However, methanol has many drawbacks. For example, (i) at lower temperatures, the denitrification efficiency is quite low with methylotrophs [39]; (ii) a long adaptation period of activated sludge microorganisms to methanol is required in order to enrich the biomass with methylotrophs [40]; (iii) since the methanol is toxic and reactive, safety measurements during its handling, transportation, and storage should be taken [38]; (iv) for methanol as a carbon source, and its storage, pumping and delivery systems, 25–31% more of the capital construction cost should be ensured to fulfil safety standards in comparison to other non-hazardous and non-flammable carbon sources [41]; and (v) the price of methanol can vary [42], and shortfalls have been recorded [38]. Other common carbon sources are acetic acids and ethanol; however, they also raise the operational costs [34,35]. Not only that, but during the transport and storage of ethanol, there is a possibility of severe safety risks [33].

The cost reduction and high denitrification efficiency and rate can be achieved with alternative carbon sources such as volatile fatty acids (VFAs) because they can be obtained on-site through the acid-phase anaerobic fermentation or thermal hydrolysis of sludge and wastewater [36,37]. VFA production is usually low because of their uptake by methanogens [43]. Also, a separation unit should be installed for the production of organic-rich fermented liquid [44]. Some authors proposed using sludge directly as an external carbon source since the fermentation of sludge and denitrification can be performed at the same time in the same bioreactor [28,33,45,46].

This review is based on bibliometric analysis and a systematic literature review, providing information and insight into the denitrification process using different carbon sources. The following phrase was used in the search engine: alternative and traditional carbon sources in denitrification. The review provides current knowledge about carbon sources and the efficiency of the denitrification process with different carbon sources from the aspect of cost benefits and reducing costs through the selection of carbon sources and, thus, the environmental footprint, as well as the recovery of waste as a source of carbon for denitrification. The types of functional microorganisms as well as the enzymes and functional denitrification genes responsible for the denitrification process are highlighted, along with the advantages, disadvantages, limitations, and future perspectives of the denitrification process.

2. Carbon Sources for Denitrification

Wastewater composition affects the final quality of the effluent. Organics are common compounds in wastewater, and can be used as source of electron donors for denitrification. Nevertheless, some wastewaters, such as municipal wastewater, are characterised by a low carbon/nitrogen (C/N) ratio, and the common challenge with such wastewater is the shortage of organics for complete nutrient removal. For the improvement of nutrient removal, external carbon sources are added [21,22]. However, the addition of an external carbon source represents a higher cost for the biological wastewater treatment plant, so research in the field of biological wastewater treatment is focused in the direction of alternative carbon sources for efficient nutrient removal, or on operational modes that enable satisfactory effluent quality at a low C/N ratio. External carbon sources can be divided into liquid and solid organic sources, and traditional and alternative carbon sources. Examples of liquid carbon sources, which are also traditional, are methanol (CH₃OH), ethanol (C₂H₆O), and glucose (C₆H₁₂O₆). Examples of solid carbon sources are biodegradable polymers such as polycaprolactone (PCL) [47,48]. Alternative carbon sources for biological denitrification have been the focus of researchers, such as digested sludge with the addition of different pure organics [49], pine bark, sawdust, bamboo biomass [50,51,52], and cattle manure [53]. Besides the organics present in the wastewater influent and external carbon sources, denitrifiers can use poly-β-hydroxybutyrate (PHB) ([C₄H₆O₂]_n), an internal carbon source, for denitrification as well [31,54]. The carbon source characteristics used for biological denitrification affect the microbial community structure, nitrate removal activity [55], denitrification rate, and intermediate products [56].

2.1. Liquid Carbon Sources for Denitrification

For heterotrophic denitrification, liquid organic sources are usually used. Among them, there are low-molecular-weight organic sources, such as sodium acetate (CH₃COONa) [16,20], methanol [20], ethanol [16,20]; saccharides, like glucose [16,19,20] and fructose [16,19,20]; ascorbic acid [16]; and a mixture of ethanol and glycerol (C₃H₅(OH)₃) [17]. In spite of the fact that denitrification driven by liquid carbon sources like sodium acetate and glucose is characterised by higher rate of denitrification, there is a possibility that in the case of a non-adequate amount of carbon source, the quality of the effluent can be worsened [57]. Liquid carbon sources are not appropriate for natural water bodies because their addition could lead to an increment in the organic pollution load [58]. In denitrification driven by glucose and with nitrate as an electron acceptor, the molar ratio of NO₃⁻ reduction to carbon dioxide (CO₂) production amounts to 0.8. This ratio depends on the carbon source and electron acceptor (nitrate or nitrite). When the labile compounds (for example, soluble sugars and unshielded cellulose, glucose, glucosamine, acetic acid) are used as a carbon source, the ratio amounts to ~0.7, and with non-labile compounds (for example, cellulose, stearic acid, vanillin, phytic acid, cattle slurry, pig slurry), the ratio is lower than 0.2 [53,59,60]. With low-molecular-weight organic carbon sources such as glucose, ethanol, sodium acetate, and methanol, a high nitrate reduction efficiency can be achieved because they provide enough electrons for denitrification, and microbes can easily use them [7,61]. The preferred nitrate reduction pathway depends on the carbon source used. For instance, glucose as a carbon source is connected to the dissimilatory nitrate reduction to ammonium, and volatile fatty acids (sodium acetate) [49], methanol, and ethanol [55] are connected to denitrification.

The majority of the microbes used for carbon utilisation use two metabolic pathways, the tricarboxylic acid (TCA) cycle and the glycolytic pathway [62]. Due to the distinctive metabolic features, sodium acetate is more effective as a carbon source for biological denitrification in comparison to methanol or glucose [62]. Sodium acetate forms acetyl Co-A, so it can be directly used for denitrification without modification, since acetyl Co-A is an essential compound of both metabolic pathways [62]. However, most denitrifiers have to conduct some enzymatic conversion of the glucose prior to metabolism [63]. With ethanol as a carbon source, higher denitrification efficiency can be achieved than with methanol because of the simple conversion of ethanol into acetyl Co-A [62]. For methanol as carbon source for denitrification, a specific group of denitrifiers is required, i.e., methylotrophs or methanotrophic bacteria, that uptake single carbon organic compounds (C₁) like methane or methanol [62]. Methanotrophs contribute to lowering methane emissions into the atmosphere [64]. They are usually anaerobic organisms, but there are some aerobic methane-oxidising bacteria (MOB) [64]. For methanotrophic bacteria, nitrite and ammonia (NH₃) can have an inhibitory effect [65]. In methanotrophs, nitrite inhibits the formate dehydrogenase enzyme that is crucial for formate’s oxidation to CO₂. Nitrite, which can be produced by methanotrophs during the oxidation of ammonia to nitrite, is a toxin that shows bacteriostatic characteristics. Ammonia shows an inhibitory effect towards the methane monooxygenase enzyme [66,67,68]. The aerobic methanotroph genome commonly contains cytochrome c-dependent nitric oxide reductase (norB) [69], probably to detoxify NO, which comes from aerobic ammonia oxidation, in such a way that NO is reduced to N₂O [70]. Organic carbon sources such as ethane (C₂H₆) and ethanol can be used as electron donors for denitrification for some obligate methanotrophs, including Methylomicrobium album strain BG, with the enzymes methanol dehydrogenase (MDH) and particulate methane monooxygenase (pMMO), although no growth has been observed [71,72,73]. Since numerous aerobic methanotrophs, including Methylomicrobium album strain BG8, have the ability to oxidise ammonia to nitrite because of the homologous inventory to ammonia-oxidising bacteria [69,74], ammonia can be used to support methanotrophic denitrification. This is enabled by enzyme copper-containing monooxygenase enzyme (CuMMO) and, for certain methanotrophs, by a hydroxylamine dehydrogenase homolog [75]. Kits et al. [64] demonstrated that Methylomicrobium album strain BG8 has the ability to use methane, methanol, formaldehyde (CH₂O), formate (HCO₂H), ethane, ethanol, and ammonia in denitrification with nitrite and under hypoxia. Hypoxia occurs at dissolved oxygen concentrations below 2 mg/L [76]. Kits et al. [64] point out that methanotrophic pathway intermediates, as well as co-substrates of pMMO, MDH, and probably hydroxylamine dehydrogenase, support the process of denitrification. Thus, methanotrophs such as M. album strain BG8 have the ability to utilise not only C₁ carbon sources for denitrification, but also C₂ carbon sources, as well as electrons derived from the NH₃ oxidation to NO₂⁻ that nitrite and nitric oxide reductases used. Therefore, methanotrophs have another pathway for the production of N₂O with the condition that methane monooxygenase can access endogenous reductant [65,69,77]. Kits et al. [64] emphasise that M. album strain BG8 does not have the ability to reduce NO₃⁻ to N₂O because the genome lacks dissimilatory nitrate reductase; however, it has the ability to respire NO₂⁻ due to the expression of the predicted denitrification genes nirS and norB1. Their experiments showed an increase in the transcript abundance of putative denitrification genes, nirS and one of two norB genes, in response to nitrite. In response to hypoxia and nitrite, the transcript abundance of pxmA, that encodes the alpha subunit of a putative copper-containing monooxygenase, also increased. The authors suggest that the genome of aerobic methanotrophs can express denitrification genes and, therefore, allow the coupling of substrate oxidation to the reduction of nitrogen oxide terminal electron acceptors under low-DO conditions [64].

Lee et al. [78] investigated the effects of carbon source (sodium acetate, sodium propionate (C₃H₅O₂Na), glucose, glycerol, ethanol, methanol), C/N ratio (0, 1.28, 2.57, 5.14, 12.85), and initial concentration of nitrate (75.9, 151.6, 303.2, 606.4, 1516 mg N/L) on functional denitrifying genes during heterotrophic denitrification. The experiments were performed with a bacterial-denitrifying consortium cultivated from humic soil as the inoculum. The achieved removal rates of NO₃⁻-N with carbon sources were as follows: ethanol (1695 ± 113.8 μg N/g VSS,d) > sodium propionate (1503 ± 454.4 μg N/g VSS,d) > glucose (1019 ± 689.8 μg N/g VSS,d) > glycerol (1005 ± 79.1 μg N/g VSS,d) > methanol (806 ± 33.6 μg N/g VSS,d) > sodium acetate (753 ± 94.1 μg N/g VSS,d). The following removal efficiencies with carbon sources were achieved: methanol (100%), ethanol (100%) > glycerol (80.8 ± 3.41%) > sodium propionate (59.5 ± 8.05%) > sodium acetate (47.9 ± 5.39%) > glucose (31.1 ± 6.75%). The complete removal efficiency of NO₃⁻-N was obtained with C/N ratios of 5.14 and 12.85 (with sodium acetate as a carbon source). The combination of low initial nitrate concentration, 75.9 and 151.6 mg N/L, and high ratios of C/N, 5.14 and 12.85, ensured complete denitrification (with sodium acetate as a carbon source). Nitrite accumulation was only observed with glycerol as the carbon source among the investigated ones. The calculated NO₂⁻-N production rates with carbon sources were sodium acetate (868 ± 1.0 μg N/g VSS,d) > glycerol (540 ± 94.7 μg N/g VSS,d) > sodium propionate (373 ± 198.9 μg N/g VSS,d) > glucose (328 ± 36.2 μg N/g VSS,d) > glycerol (130 ± 18.2 μg N/g VSS,d). The authors attribute the higher denitrification performance with methane as a carbon source in comparison to other investigated carbon sources to the methylotrophic denitrifying community existing in the microbial inoculum. The removal rates of NO₃⁻-N with sodium acetate as a carbon source were as follows: C/N 0 (59 ± 0.0 µg N/g VSS,d) < C/N 2.57 (1379 ± 0.0 µg N/g VSS,d) < C/N 1.28 (2186 ± 143.8 µg N/g VSS,d) < C/N 12.85 (2345 ± 680.6 µg N/g VSS,d) < C/N 5.14 (3017 ± 44.7 µg N/g VSS,d) [78].

2.2. Solid Carbon Sources for Denitrification

The benefit of a solid carbon source is that it can be used as a carbon source and, in solid-phase denitrification systems, as a physical support for biofilm formation [79]. The release of carbon is gradual, easy to manage, and can be used long term [58]. Solid carbon sources for denitrification can be (i) natural, such as wood chips, wheat straw, fruit shells [58], corn cobs, corn stover, cardboard fibres, leaf litter, tree bark, and fruit kernels [80,81,82,83]; (ii) synthetic, such as biodegradable polymers like poly(3-hydroxybutyrate-co-3-hydroxyvalerate) (PHBV)/starch [79], PCL [47,48], starch/PCL [84], starch/polyvinyl alcohol (PVA) [85], polyhydroxyalkanoates (PHAs) [86], PHBV/poly (lactic acid), (PLA) polymer [87,88], poly vinyl alcohol (PVA) [89], and polybutylene succinate (PBS) [90]; and other (iii), such as hydrolysed sludge, liquids from food waste, and other reprocessed organic materials [91,92]. The benefit of using natural solid carbon sources is their low cost and good availability. They are applied mostly in nearby denitrification systems. At the moment, the broader usage of natural carbon sources is limited because of the increased colour intensity in the effluent, excessive DOC release, unstable carbon release rates, and slow denitrification rates [58]. Solid carbon sources for denitrification are appropriate for water bodies such as wastewater, aquaculture wastewater, surface water, agriculture subsurface, and groundwater [58]. The prerequisite of a solid carbon source to be used by microorganisms is to be first converted to low-molecular-weight compounds, like acetic acid, formic acid, and methanol [58]. During the degradation of a polymer, large and complex molecules are broken down into smaller molecules [93]. When a biodegradable polymer is used as a carrier for microorganisms, firstly, on the surface of the polymer, the formation of biofilms through the attachment and growth of microorganisms occurs, followed by the cleavage of the polymer chains by extracellular enzymes, which leads to the hydrolysis of the polymer into soluble compounds of low molecular weight [94,95]. At that moment, microorganisms can uptake these low-molecular-weight compounds through semipermeable membranes and use them as electron donors and acceptors. Hence, the application of biodegradable polymers includes hydrolysis and denitrification, with hydrolysis as the rate-determining step [96]. During the microbial degradation of solid carbon sources, significant reductions in the polymer occur, coupled with chemical structure changes, mechanical deformation due to structural dissolution and breakdown, increased surface roughness, and the formation of perforations and pits [97]. All of these processes lead to changes in the carbon release rate of the material and to a decrease in the crystalline phase content, and, therefore, to the increase in readily hydrolysable amorphous content, which, in turn, changes the hydrophilicity [58].

The chemical structure and molecular weight of biodegradable polymers, commonly more complicated and higher in comparison to liquid carbon sources, considerably influence the mechanism of the utilisation of the carbon source, particularly the microbial community [87]. For instance, in denitrification experiments with ethanol and starch/PCL as carbon sources, the microbial diversity was much more abundant with the starch/PCL system, whereas the denitrification rate was higher with ethanol as the carbon source [98]. The advantage of PHBV/PLA polymer blends as a carbon source for denitrification is the low cost and high denitrification performance [87,99]. In denitrification with solid carbon sources, the concentration of dissolved organic carbon (DOC) in the effluent is linked to soluble organic substance production and consumption [100]. The degradation of the polymer results in soluble organic substances [87]. The decrease in hydraulic retention time (HRT) results in a decrease in soluble organic substance production, so the DOC in the effluent is lower with the lower HRT, and vice versa [87]. Xu et al. [87] investigated the effect of various carbon sources—glucose, CH₃COONa, and PHBV/PLA polymer—on denitrification performance and denitrification genes. The sludge inoculated to the bioreactor adapted to all three carbon sources quickly in spite of the different chemical structures and molecular weight. The experiments were conducted with volcanic rock as a biofilm carrier. The ratios of C/N for glucose and CH₃COONa were in the range of 3.6–6. The experiments with PHBV/PLA lasted 91 days and, with glucose and CH₃COONa, 49 days. The nitrate loading rates (NLRs) for PHBV/PLA were in the range of 0.15–0.78 gN/Ld, and for glucose and CH₃COONa were 0.19–0.61 gN/Ld. The influent NO₃⁻-N concentrations for glucose, CH₃COONa, and PHBV/PLA were in range of 15.49 ± 0.55–50.60 ± 0.27 mgN/L. The maximum denitrification rates were obtained with glucose (0.46 gN/Ld), followed by CH₃COONa (0.39 gN/Ld), and the lowest with PHBV/PLA (0.37 gN/Ld). The ratio C/N 6 enabled complete NO₃⁻-N removal with glucose and CH₃COONa carbon sources [87]. The authors point out that in denitrification with PHBV/PLA as a carbon source, the released energy from the degradation of the polymer was a lot more that the energy consumed for the growth of bacteria and denitrification [87]. When PHBV/starch was used as a carbon source for denitrification, a higher accumulation of DOC during the lag phase occurred [79]. Xu et al. [87] highlighted that in denitrification with PHBV/PLA as a carbon source, the degradation of PHBV/PLA polymer affects pH value and DOC concentration in the effluent. When the uptake of acidic intermediate products of the PHBV/PLA polymer increases, the concentration of DOC in the effluent decreases, and the value of pH increases [87]. Bacteria use depolymerase for the degradation of PHBV/PLA polymer into low-weight compounds [99,100] that denitrificants use as a carbon source [87]. The majority of denitrificants prefer sodium acetate or 3-hydroxybutyrate (3HB) over PHBV polymer in denitrification driven by PHBV granules [101].

2.3. Alternative Carbon Sources for Denitrification

As alternative carbon sources for biological denitrification, digested sludge with the addition of different pure organics [49] such as pine bark, sawdust, bamboo biomass [50,51,52], and cattle manure [53] can be used. The products of the decomposition of animal slurries are mostly volatile fatty acids [102,103], usually acetic, propionic, and butyric acids [104]. Grazed grassland soils are rich in a variety of organics that originate from animal excreta and plant material [105] such as stearic acid [106], glucosamine [107], benzoic acid [108], glucose, cellulose [105], phytic acid [109], and vanillin [110]. Pig slurry is mainly composed of readily decomposable organics, with a low proportion of lignin and cellulose [111]. Animal slurries contain commonly low amounts of nitrate because of the anaerobic conditions and nearly complete inorganic N in the ammonium form, and 40–60% of the total N in slurries is organic N [112]. The availability of organics for microbial processes depends on their ability to support microbial growth and to their decomposability [113]. Dlamini et al. [59] investigated the availability of eight different standard organic compounds found in manure (glucose, glucosamine, cellulose, stearic acid, benzoic acid, lignin, vanillin, phytic acid) and their role in promoting denitrification by establishing the reactivity of varying C compounds found in cattle and pig slurry through their availability for denitrification, and explored C-quality effects in promoting denitrification. They tested the C availability of pig and cattle slurry by comparing them to four standard organic compounds: glucose, acetic acid, vanillin, and cellulose. The experiments were performed in soil. For the evaluation of different carbon sources for the promotion of denitrification, they used the molar ratio of CO₂ production to NO₃⁻ reduction after incubation. The authors point out that the molar ratio of CO₂ evolution to NO₃⁻ reduction indicated that glucose and glucosamine are highly reactive organics. Also, they highlighted that pig slurry and acetic acid are carbon sources of good quality for the promotion of potential denitrification [59]. Cao et al. [33] used primary sludge as a carbon source directly in a denitrification bioreactor (an integrated denitrification system with primary sludge as a solid carbon source). At a dosage of 6.0 g VSS/g N, 100% denitrification was achieved, with no nitrite accumulation, and with a maximum specific nitrate reduction rate of 6.4 mg N/g VSS,h. Also, in this way, the reduction of primary sludge of 65.3–85.1% was achieved. However, the release of ammonium, phosphate, and recalcitrant matter occurred. During the denitrification process, reductions in ammonium and phosphate were recorded [33].

2.4. Endogenous Denitrification/Internal Carbon Sources

During the process of nitrogen removal in biological wastewater treatment plants, the combination of nitrogen assimilation and dissimilation occurs. Nitrogen assimilation includes nitrogen uptake by cell growth, and nitrogen dissimilation includes nitrification and denitrification. Sufficient nitrogen is ensured at a ratio of COD/N 20. At low COD/N ratios (<5–10), the carbon source is not enough for denitrification or cell growth. At high ratios of COD/N (>20–30), the nitrogen source is not enough for cell growth [15]. Besides intrinsic organic matter in wastewater and external carbon sources, the activated sludge biomass can also serve as an electron donor for denitrification [114,115]. At low concentrations of chemical oxygen demand (COD) in wastewater for denitrification, high sludge concentrations of MLSS > 4000–10,000 mg/L (mixed liquor suspended solids) should be ensured [116]. The requirement of external carbon source supplementation, the usual limitation of the nitrogen removal efficiency, can be overcome with the use of microbial biomass with traits of PHA storage capacity [117]. PHAs, as intracellular polyesters, include polymers such as PHB and PHBV, and can serve as electron donors under anaerobic and aerobic conditions [118]. PHAs act as channels for intracellular reducing power and provide the carbon for metabolic intermediates to balance the storage and use of carbon and energy [24,117]. When faced with stress conditions such as aerobic dynamic feeding (feast–famine regime), microorganisms can intracellularly store PHA [119]. In feast–famine strategies, microorganisms store PHA during the feast phase, and then consume PHA during the famine phase. Compared to the traditional heterotrophic oxidation of organics, the endogenous oxidation of PHA is six times slower [117]. For complete denitrification, without taking into consideration the biomass growth, for hydroxybutyrate and hydroxyvalerate, the required ratio amounts are 2.853 g COD_PHA/g N and 2.855 g COD_PHA/g N, respectively. The batch experiments optimised to determine the specific denitrifying activity of sludge using stored PHA as a carbon source showed the optimal conditions: a concentration of biomass 0.5–2.0 g VSS/L, a ratio of COD_PHA/N higher than 5.4 g/g, and a nitrate concentration in the range of 40–60 mg NO₃⁻-N/L. Also, the maximal specific endogenous denitrifying activity values were not obtained when the concentration of PHA in the biomass was lower than 5% because, at low PHA concentrations, the biomass tends to store it as a carbon source instead of consuming it for denitrification purposes [117].

3. Aerobic Denitrification

Besides anoxic denitrification, there is also denitrification under aerobic conditions, in which nitrates and/or nitrites are reduced to dinitrogen gas [120]. In that case, nitrification and denitrification are conducted in the same bioreactor, which is beneficial for the alkalinity level since the alkalinity is consumed during nitrification and supplemented during denitrification [121]. The process of aerobic denitrification is affected by several factors such as the C/N ratio [122], operational mode [123], DO concentration [122], carbon source [124], and temperature [125], and among them, the strongest influence is the C/N ratio and DO concentration [122]. The aerobic denitrification process offers higher total nitrogen (TN) removal efficiency in comparison to conventional processes for N removal, like membrane bioreactors (MBRs), sequencing batch bioreactors (SBRs), and sequencing batch biofilm bioreactor (SBBRs) [11]. Unlike conventional biological removal processes, aerobic denitrification requires a higher ratio of C/N, such as C/N 5–10 for the efficient effluent characteristics [126]. In cases where there is surplus reductant that has to be depleted, or in DO limitation where the respiration should be maximised, chemo-organoheterotrophs switch to aerobic denitrification and ammonia-oxidising bacteria to nitrifier-denitrification [32,127]. The operational mode anoxic/oxic (A/O) is more beneficial for aerobic denitrification than a completely oxic operational mode [26].

Aerobic denitrifying bacteria can be isolated from domestic wastewater treatment plants [11,128]. They belong to the phylum Proteobacteria and they are usually Gram-negative bacteria [126]. The carbon source influences the activity of periplasmic nitrate reductase (Nap). This reductase is responsible for aerobic denitrification [124]. In pure cultivation, the Nap activity was higher with a higher amount of the reduced carbon sources [124,129]. Among the carbon sources malate, succionate, butyrate, and caproate, Paracoccus pantotrophus exhibited the highest activity when growing on the last two carbon sources [124]. Later, it was shown that in Paracoccus pantotrophus, the highest activity of Nap was with the growth substrate butyrate [129]. Methylomicrobium album strain BG8 utilises NO₂⁻ either instead of DO or in combination with DO in order to reduce the overall demand of cellular O₂, which enables O₂ conservation for the additional oxidation of methane (CH₄) [64]. The authors suggest that the NO₂⁻ usage under the limitation of DO for M. album strain BG8 is the way to maximise total respiration [64].

PHB can also be used as a carbon source for aerobic denitrification [31]. PHB synthesis is affected by the microbial species, carbon source, and operational mode. Among the investigated carbon sources of sodium acetate, ethanol, and glucose, the storage of PHB is the only relevant mechanism for the carbon sources ethanol and sodium acetate [54]. The operational mode anaerobic–aerobic and feast–famine contributed to the maximum yield of 64 wt% PHB, in contrast to the completely aerobic and feast–famine mode, that contributed to the maximum yield of 53 wt% PHB [27]. The producing capacity of PHB, as well as the growth characteristics, are different for different microorganisms, such as Thauera selenatis and Plasticicumulans acidivorans [30]. Since aerobic denitrifying microorganisms cannot directly use reduced carbon sources, they have to be converted to PHB prior to uptake [32].

Wu et al. [28] conducted experiments with denitrifiers acclimated with different organic carbons (methanol or sodium acetate). The experiments were conducted in batch mode, with initial concentrations of 100 mg NO₃⁻-N/L, 50 mg NO₂⁻-N/L, and 100 mg NO₃⁻-N/L and 20 mg NO₂⁻-N/L. The biomass for the experiments was taken from two parental reactors acclimatised to sodium acetate or methane as a carbon source that operated six cycles per day as follows: fill (20 min), anoxic (150 min), aerobic (20 min), settlement (40 min), draw/idle (20 min). The rate of endogenous respiration for denitrifying microorganisms acclimatised to methanol or sodium acetate amounted to less than 2 mg/g VSS,h. A slightly higher reduction rate of NO₃⁻-N was observed with the biomass acclimatised to sodium acetate in comparison to the biomass acclimatised to methanol. Under endogenous conditions, diverse types of denitrifying organisms might have diverse denitrification modes. The authors point out that with denitrification driven by sodium acetate, the small amount of PHB remained inside the biomass and it was used for respiration slowly. Also, these experiments showed higher reduction rates for nitrate than nitrite. The authors highlighted that denitrifying organisms under endogenous respiration conditions prefer nitrate rather than nitrite when the amount of organics is limited. The denitrifying organisms acclimatised to sodium acetate as the carbon source showed higher potential of nitrite accumulation in comparison to methanol. The ratio of produced NO₂⁻-N to the reduced NO₃⁻-N was higher with sodium acetate as the carbon source (53%) in comparison to methanol (38%). However, with the mixture of electron acceptors (NO₃⁻-N + NO₂⁻-N), the same ratio amounted 66% in denitrification driven by sodium acetate and 28% in denitrification driven by methanol [28]. The partial PHB utilisation is linked to nitrite accumulation since the high accumulation of nitrite was observed in denitrification driven by PHB [90]. The authors suggest the application of high biomass concentration or a long reaction time in order to obtain the expected denitrification efficiency in the process with post-denitrification with endogenous respiration for nitrogen removal because of the low rate of endogenous denitrification, like for membrane bioreactors or biofilm systems [28].

Dyagelev et al. [25] conducted batch experiments and reported a denitrification rate for easily oxidisable substrate of 0.71 ± 0.04 mg NO₃⁻-N/mg,d, a denitrification rate via endogenous respiration of 0.12 ± 0.02 mg NO₃⁻-N/mg,d, and denitrification rate for difficult oxidisable substrate of 0.12 ± 0.01 mg NO₃⁻-N/mg,d. Hu et al. [26] investigated the effect of carbon sources and operation modes on the performance of the aerobic denitrification process. The synthesis of PHB, an essential component for the aerobic denitrification process, is affected by several factors, such as operation mode, carbon source, and microbial community composition. Experiments were performed with CH₃COONa and CH₃CH₂CH₂COONa in SBR under two modes: anoxic/oxic and entirely oxic mode. The experiments were performed with an initial 500 mg COD/L, 50 mg NH₄-N/L, and 20 mg PO₄-P/L, and with DO concentration of ~6.5 mg DO/L. The cycle consisted of feeding (10 min), mixing (120 min), aeration and mixing (540 min), settling (10 min), and resting (30 min) for anoxic/oxic mode, and of feeding (10 min), aeration and mixing (660 min), settling (10 min), and resting (30 min) for oxic mode. The cultivation lasted 40 days, and after that, the process of aerobic denitrification in the SBR reactor was well set up. The removal efficiency of the total nitrogen was in the range of 86.11–90.05%, and the COD removal efficiency was higher than 94%. The authors point out the relationship between TN removal efficiency and PHB content for the same operational mode. In both investigated operational modes (oxic and anoxic/oxic), the removal efficiency of TN and the maximum content of PHB were proportional, but for both carbon sources under oxic mode, no differences were observed for the removal efficiency of TN. The combination of anoxic/oxic mode and CH₃CH₂CH₂COONa as the carbon source resulted in the maximum PHB content and maximum removal efficiency of TN. The content of PHB was higher in the experiments with CH₃CH₂CH₂COONa than with CH₃COONa under the same operational mode because CH₃CH₂CH₂COONa is more reduced than CH₃COONa. Also, the highest PHB content was detected under A/O mode in comparison to the complete oxic mode [26] because the alternative anaerobic and oxic conditions enhance the synthesis of PHB [27]. The external carbon source is generally converted to PHB under anaerobic conditions [130]. For Paracoccus pantotrophus, CH₃COONa is more favourable than CH₃CH₂CH₂COONa because it can accelerate Nap during aerobic growth [131]. Using CH₃CH₂CH₂COONa as a carbon source could result in microorganisms with a saturation of reductant and, hence, to the saturation of adenosine triphosphate (ATP). When the microorganisms are faced with excessive ATP, it might lead to a low concentration of adenosine diphosphate (ADP) and the limitation of the respiratory rate [131]. Such conditions could result in the limitation of certain biochemical reactions in microorganisms, which leads to the accumulation of certain intermediates from those reactions [26]. In spite of the high removal efficiency of TN in aerobic denitrification with CH₃CH₂CH₂COONa, a part of the TN is removed over NO and N₂O. Thus, in spite of the fact that Nap activity is higher with CH₃CH₂CH₂COONa, CH₃COONa is a better choice of carbon source because more NO and N₂O is emitted with CH₃CH₂CH₂COONa [26]. These experiments were performed with excessive amounts of organics (at a ratio of C/N 10), and the amount of electron donors that flowed into the electron transport chain was large enough to stimulate the process of aerobic denitrification, which was the reason for the high removal efficiency of TN. Since the process of aerobic denitrification is favourable in alternate anoxic and oxic operation mode, the highest removal efficiency of TN is obtained under the anoxic/oxic operational mode [26].

4. The Role of Temperature and pH on Denitrification

Temperature, an environmental factor, and pH, an operational factor, affect the process of denitrification because they influence the stability and metabolism of both electron donor and electron acceptors [132,133].

Ortmeyer et al. [16] investigated biological denitrification on circulation columns with sediment, and with the carbon sources sodium acetate, glucose, ascorbic acid, and ethanol at room temperature (around 21.5 °C) and at 10 °C (typical groundwater temperature). The experiments were performed at an initial nitrate concentration of around 250 mg NO₃⁻/L. Among the investigated carbon sources (sodium acetate, glucose, ascorbic acid, and ethanol) and temperatures (around 21.5 °C and 10 °C), ethanol was the most effective carbon source at 10 °C. At room temperature, glucose showed a much higher denitrification rate in comparison to the temperature of 10 °C. Ascorbic acid was not found to be a convenient carbon source for denitrification due to low denitrification at both 10 °C and at room temperature, with biomass production that caused column clogging. In the denitrification at low temperature (10 °C), nitrite accumulation was recorded in much higher concentrations that at room temperature. The low concentration of electron donors led to nitrite accumulation [16].

The temperature of 25–35 °C did not have effect on the denitrification rate [78]. The removal efficiencies of nitrate were 25 °C (75.8 ± 2.40%) > 30 °C (75.7 ± 6.02%) > 35 °C (73.4 ± 5.73%) > 20 °C (68.5 ± 9.80%) > 15 °C (53.9 ± 8.08%). The removal rates of NO₃⁻-N were 35 °C (3011.9 ± 267.75 μg N/g VSS,d) > 25 °C (2907.1 ± 70.56 μg N/g VSS,d) > 30 °C (2784.4 ± 250.82 μg N/g VSS,d) > 20 °C (1909 ± 118.6 μg N/g VSS,d) > 15 °C (651 ± 42.3 μg N/g VSS,d) [78]. A lower temperature lowers the denitrification rate [132].

The value pH in the range of 7–8 is considered the optimal pH for the process of denitrification [134], and at low pH values, denitrification gene expression and denitrification rates have the tendency to decrease [132]. The pH value of 5–6 decreased the denitrification rate [78]. The removal efficiencies of NO₃⁻-N were pH 5 (47.5 ± 4.9 5%) < pH 6 (50.0 ± 5.71%) < pH 7 (62.4 ± 5.90%) < pH 8 (65.2 ± 2.34%) < pH 9 (68.7 ± 0.35%). The removal efficiencies of NO₃⁻-N were pH 5 (591 ± 61.0 µg N/g VSS,d) < pH 6 (1011 ± 19.2 µg N/g VSS,d) < pH 7 (1027 ± 137.2 µg N/g VSS,d) < pH 8 (1323 ± 35.8 µg N/g VSS,d) < pH 9 (1383 ± 86.9 µg N/g VSS,d) [78].

5. Microbial Community in Denitrification

The carbon source affects microbial community composition. For instance, with sodium acetate as the carbon source for denitrification, Halomonas were the dominant microbes [135], and Rhodobacteraceae from the Alpha-proteobacteria group and Rhodocyclaceae and Comamonadaceae from the Betaproteobacteria were found [136,137]. With methanol, the dominant microbes were Paracoccus and Hyphomicrobium [135], and Hyphomicrobiaceae and Methylaphilaceae [136]. Methylomicrobium album strain BG8 performs denitrification under hypoxia with nitrite as the electron acceptor and the electron donors methane, methanol, formaldehyde, formate, ethane, ethanol, and ammonia in denitrification with nitrite and under hypoxia [64]. Microbial analysis showed the shift in the microbial community dependening on the carbon source (sodium acetate, glucose, ascorbic acid, and ethanol) and temperature (around 21.5 °C and at 10 °C) [16]. In experiments with glucose at room temperature, the dominant populations were Proteobacteria (53–60% relative abundance) and Actinobacteria (40–47%), but at 10 °C, 99% of the communities belonged to Actinobacteria. With ethanol at room temperature, Proteobacteria were detected (98%), and at 10 °C, Actinobacteria were detected (89–93%). Sodium acetate as a carbon source at room temperature contributed to 87% of Actinobacteria and, at 10 °C, 100% of Proteobacteria. Ascorbic acid at room temperature was dominated by Actinobacteria (73–81%) and at 10 °C, also Actinobacteria (89–99%) [16].

An Illumina MiSeq sequencing analysis [87] showed that, in the denitrification system, the most dominant communities were Brevinema/Thauera/Dechloromonas with PHBV/PLA as a carbon source, Tolumonas/Thauera/Dechloromonas with glucose, and Thauera with CH₃COONa as a carbon source. A transcriptome-based analysis of the genes involved in the dissimilatory nitrate reduction process revealed that the highest FPKM values (the fragments per kilobase per million reads) were found in the denitrification system with glucose as the carbon source, which is in accordance with the highest NH₄⁺-N concentration in effluent [87].

With CH₃COONa as the carbon source for aerobic denitrification and the operation mode shifting from A/O to completely oxic operation mode, the quantity of aerobic denitrifiers and microorganisms that synthesise extracellular polymetric substances (EPSs) (Ohtaekwangia sp.) rises [26]. EPSs are important for the stabilisation of the structure of aerobic granules [138] and for the formation of the floccule of activated sludge by agglomerating activated sludge [139]. With CH₃CH₂CH₂COONa as the carbon source in completely oxic operational mode, the most dominant was Plasticicumulans sp. [26], known for the synthesis of PHB [30]. The conversion of the operational mode from completely oxic to A/O with CH₃CH₂CH₂COONa resulted in a shift of aerobic denitrifiers from 3.70% to 13.30% and to the decrease in microorganisms that produce PHB from 43.68% to 37.83% [26]. The abundance of the phylum Proteobacteria was higher with CH₃CH₂CH₂COONa than with CH₃COONa. The diversity of bacteria was higher with CH₃COONa than with CH₃CH₂CH₂COONa [26]. In all four investigated combinations, i.e., completely oxic, A/O, CH₃COONa, and CH₃CH₂CH₂COONa, the proportion of nitrifying bacteria was low [26]. With CH₃COONa as the carbon source, no genus was detected as dominant, but in the case of CH₃CH₂CH₂COONa, the dominant genus was Plasticicumulans sp. [26].

6. N₂O and NO Emission during Denitrification

Nitrous oxide, the leading ozone-depleting substance in the 21st century [140], is a potent greenhouse gas with a 114-year lifecycle [141]. N₂O exhibits 300 times higher global warming potential in comparison to carbon dioxide [142]. The anthropogenic emission of N₂O comes from agriculture, wastewater treatment processes, manure management, and fossil fuel combustion [14,142]. For the development of sustainable food production practices, the reduction of greenhouse gas emissions from agricultural systems should be employed [143]. In the soil respiratory system, denitrification contributes to around 37% of the produced CO₂ [144]. Regarding biological wastewater treatment, the major source of N₂O emission is incomplete heterotrophic denitrification [7,8,145]. Many factors affect the emission of N₂O from denitrification, like nitrite concentration, the concentration of DO, the ratio of C/N, the carbon source, temperature, the initial nitrate concentration, pH [7,14,118], and endogenous denitrification [28]. An insufficient ratio of C/N leads to N₂O emission because of the lack of an organic carbon source [145]. The ratio of C/N represents the demand of the mass of organics (electron donors) for the mass of reduced nitrogen compounds (electron acceptors) [146]. With sodium acetate as the carbon source, the theoretical C/N ratio for complete denitrification was 3.74, 2.92 for PHB, 2.68 for glucose, 1.90 for methanol, and 1.37 for ethanol. When biomass formation is taken into consideration, the required C/N ratios are 3.03 for PHB, 2.47 for methanol, and 2.01 for ethanol [8]. The theoretical C/N ratio is lower than the obtained C/N ratio from the experiments, such as 5.14 and 12.85 [78], because the theoretical C/N ratio did not take into consideration the other heterotrophic organisms. A low C/N ratio is common for municipal wastewater, and during the biological treatment of wastewater with a low C/N ratio, the accumulation of NO₂⁻-N and N₂O usually occurs due to an insufficient amount of electron donors [145,146]. In denitrification with a C/N ratio under 2.5, the competition for electron donors occurs between nitrate and nitrite reduction, so the NO₂⁻-N or N₂O accumulation can be observed [135,147]. In the process of heterotrophic denitrification, for the emission of N₂O, the C/N ratio was crucial [78]. In the denitrifying microbial community, the enhancement of the activity of N₂O reductase could be a potential strategy for reducing N₂O emission [148]. Nitrate was mainly denitrified with methanol and ethanol as the carbon source (among the tested carbon sources: sodium acetate, sodium propionate, glucose, glycerol, ethanol, methanol), with no nitrite and N₂O-N detected [78]. With CH₃COONa as a carbon source, the emissions were NO (~0.4 mg/L) > N₂O (~0.02 mg/L), and with CH₃CH₂CH₂COONa, the emissions were NO (~1.0 mg/L) > N₂O (~0.2 mg/L) [26]. The authors point out that CH₃COONa is a more convenient carbon source for aerobic denitrification in comparison to CH₃CH₂CH₂COONa due to the amount of emitted NO and N₂O [26].

Intracellular carbon sources for denitrification, like PHB, exhibit a high potential for N₂O emission because of its low degradation rate [90]. Wu et al. [28] speculate that higher N₂O emission under denitrification during endogenous conditions could be due to the low rate of denitrification.

The electron acceptor (nitrate and nitrite) affected the amount of N₂O emitted; a 10% emission rate was obtained with nitrite and a less than 1% rate with nitrate as the electron acceptor, for both carbon sources (methanol and sodium acetate) [28]. Since nitrite inhibits N₂O reductase during denitrification, nitrite also affects the emission of N₂O [149]. The authors point to the possible source of N₂O emission—for example, the settlement tank—in the short-cut nitrogen removal processes over nitrite [28]. The emission of N₂O in denitrification driven by sodium acetate was NO₂⁻-N (11.8%) > NO₃⁻-N + NO₂⁻-N (3.5%) > NO₃⁻-N (0.8%), and in denitrification driven by methanol was NO₂⁻-N (14.0%) > NO₃⁻-N + NO₂⁻-N (3.4%) > NO₃⁻-N (0.25%). Nitrite enhanced the emission of N₂O in both the nitrification and denitrification process. Unlike aerobic conditions, where stripping takes place, during denitrification, the emission of N₂O mostly occurred because of diffusion from the liquid to the gas phase [28].

The emission of NO and N₂O can be reduced by implementing aerobic denitrification [150]. The emission of N₂O during aerobic denitrification is connected to the type of microorganisms and their oxygen tolerance of nitrous oxide reductase [151]. In the denitrification conducted under pH values of 5 and 7, no NO₂⁻-N or N₂O-N were detected [78]. Denitrification conducted at 15 °C resulted in the conversion of nitrate to N₂ gas, with no nitrite and N₂O [78]. In denitrification conducted under the temperature range of 20–35 °C the nitrite and N₂O were detected. The intermediates of denitrification, NO₂⁻-N, and N₂O-N were detected in experiments conducted at 20, 25, 30, and 35 °C [78].

A comprehensive overview of the denitrification rate and effectiveness of the different carbon sources at C/N ratios, with advantages and disadvantages, is presented in

Table 1. An overview of denitrification rate and effectiveness of different carbon sources at C/N ratios, with advantages and disadvantages.

Process	C Source	C/N Ratio Expression	Experimental Conditions	Findings	rDNI	Effectiveness, % DNI	Advantages/ Disadvantages	Price Ref.
Batch test	Acetate, 30 °C	1998.4 mg C/L cca 250 mg NO3−-N/L		The C/N ratio was a key factor for N2O emission during the heterotrophic denitrification process		cca 53%	Inoculum: humic soil	1 [78]
Batch test	Methanol, 30 °C	1998.4 mg C/L cca 290 mg NO3−-N/L				100%	Inoculum: humic soil	2 [78]
Batch test	Ethanol, 30 °C	1998.4 mg C/L cca 270 mg NO3−-N/L				100%	Inoculum: humic soil	3 [78]
Batch test	Glucose, 30 °C	1998.4 mg C/L 275 mg NO3−-N/L				31%	Inoculum: humic soil	4 [78]
Batch test	Propionate, 30 °C	1998.4 mg C/L cca 250 mg NO3−-N/L				cca 66%	Inoculum: humic soil	5 [78]
Batch test	Glycerol, 30 °C	1998.4 mg C/L cca 250 mg NO3−-N/L				cca 83%	Inoculum: humic soil	6 [78]
Batch test, acclimatised sludge	MicroCTM h, 20 °C	COD/N 6.5 ± 3.7		The ability of a specific carbon-acclimated denitrifying population to instantly use other carbon sources also was investigated, and the chemical-structure-associated behaviour patterns observed suggested that the complex biochemical pathways/enzymes involved in the denitrification process depended on the carbon sources used	6.4 ± 3.6 mgN/g VSS,h		Presence of nitrite was minimal	7 [38]
Batch test, acclimatised sludge	MicroCTM h, 10 °C	COD/N 6.5 ± 3.7			2.5 mgN/g VSS,h		Presence of nitrite was minimal	7 [38]
Batch test, unacclimatised WWTP sludge	MicroCTM h, 20 °C	COD/N 4.0			4.3 mgN/g VSS,h			7 [38]
Batch test, acclimatised WWTP sludge	MicroCTM h, 20 °C	COD/N 7.0 ± 1.4			4.7 mgN/g VSS,h			7 [38]
Batch test, acclimatised sludge	Methanol, 10 °C	COD/N 4.8 ± 1.5			2.3 mgN/g VSS,h		No nitrite accumulation	2 [38]
Batch test, acclimatised sludge	Methanol, 20 °C	COD/N 4.8 ± 1.5			6.1 ± 0.7 mgN/g VSS,h		No nitrite accumulation	2 [38]
Batch test, acclimatised sludge	Acetate, 10 °C	COD/N 5.7 ± 1.3			3.6 mgN/g VSS,h		Significant nitrite accumulation	1 [38]
Batch test, acclimatised sludge	Acetate, 20 °C	COD/N 5.7 ± 1.3			13.6 ± 1.9 mgN/g VSS,h		Significant nitrite accumulation	1 [38]
Long-term SBR	Primary sludge, 30 °C	6.0 g VSS/g NO3−-N	Cycle 24 h–48 h HRT 8 d–2 d	Cycle study suggests that an appropriate denitrification cycle/duration time would largely lower the effluent organics concentration, which can be achieved by monitoring the pH turning point	6.4 mg N/g VSS,h	100%	Reduction in primary sludge of 65.3–85.1%	[33]
Batch test	Acetate, 10 °C	2.5, 5, and 10 mmol C 250 mg NO3−/L		Addition of organic substances and temperature strongly modify the denitrifying microbial community	0.01 mmol/L,d (2.5 mmol) 0.06 mmol/L,d (5 mmol) 0.09 mmol/L,d (10 mmol)	31% (2.5 mmol) 38% (5 mmol) 45% (10 mmol)	Inoculum: sediment, column experiments	1 [16]
Batch test	Acetate, 21.5 °C	2.5, 5, and 10 mmol C 250 mg NO3−/L			0.02 mmol/L,d (2.5 mmol) 0.08 mmol/L,d (5 mmol) 0.1 mmol/L,d (10 mmol)	60% (2.5 mmol) 59% (5 mmol) 45% (10 mmol)	Inoculum: sediment, column experiments	1 [16]
Batch test	Ethanol, 21.5 °C	2.5, 5, and 10 mmol C 250 mg NO3−/L			0.05 mmol/L,d (2.5 mmol) 0.11 mmol/L,d (5 mmol) 0.21 mmol/L,d (10 mmol)	54% (2.5 mmol) 55% (5 mmol) 45% (10 mmol)	Inoculum: sediment, column experiments	3 [16]
Batch test	Ethanol, 10 °C	2.5, 5, and 10 mmol C 250 mg NO3−/L			0.06 mmol/L,d (2.5 mmol) 0.17 mmol/L,d (5 mmol) 0.14 mmol/L,d (10 mmol)	66% (2.5 mmol) 97% (5 mmol) 44% (10 mmol)	Inoculum: sediment, column experiments	3 [16]
Batch test	Glucose, 21.5 °C	2.5, 5, and 10 mmol C 250 mg NO3−/L			0.08 mmol/L,d (2.5 mmol) 0.28 mmol/L,d (5 mmol) 0.64 mmol/L,d (10 mmol)	48% (2.5 mmol) 25% (5 mmol) 21% (10 mmol)	Inoculum: sediment, column experiments	4 [16]
Batch test	Glucose, 10 °C	2.5, 5, and 10 mmol C 250 mg NO3−/L			0.04 mmol/L,d (2.5 mmol) 0.05 mmol/L,d (5 mmol) 0.09 mmol/L,d (10 mmol)	41% (2.5 mmol) 19% (5 mmol) 22% (10 mmol)	Inoculum: sediment, column experiments	4 [16]
Batch test	Ascorbic acid, 21.5 °C	2.5, 5, and 10 mmol C 250 mg NO3−/L			<0.01 mmol/L,d (2.5 mmol) 0.05 mmol/L,d (5 mmol) 0.09 mmol/L,d (10 mmol)	3% (2.5 mmol) 18% (5 mmol) 19% (10 mmol)	Inoculum: sediment, column experiments	8 [16]
Batch test	Ascorbic acid, 10 °C	2.5, 5, and 10 mmol C 250 mg NO3−/L			0.01 mmol/L,d (2.5 mmol) 0.02 mmol/L,d (5 mmol and 10 mmol)	6% (2.5 mmol) 6% (5 mmol) 10% (10 mmol)	Inoculum: sediment, column experiments	8 [16]
Batch test	Easily oxidisable substrate, real wastewater of the sewage treatment facility	Not mentioned		The obtained denitrification rates allow calculation of the volumes of anoxic and aerobic structures. (When increasing the time of the sludge mixture in anoxic conditions, it becomes necessary to proportionally increase the volume of aerobic structures, in order to maintain the growth rate of autotrophs)	0.71 ± 0.04 mg NO3-N/mg VSS,d			[25]
Batch test		Not mentioned			0.12 ± 0.02 mg NO3-N/mg VSS,d		Endogenous respiration	[25]
Batch test	Difficult-to-oxidise substrate	Not mentioned			0.12 ± 0.01 mg NO3-N/mg VSS,d			[25]
Batch test	Methanol, 25 °C	MeOH/NO3−-N 3.5		After acclimation to nitrate, the dominant strains were Pseudomonas and Paracoccus spp.	21 mgNO3-N/g VSS,h	100%	Mixed bacterial culture, originated from two-stage anaerobic–aerobic industrial yeasts production wastewater treatment plant, low accumulation of nitrite-N (0.1 mg/L)	2 [152]
Continuous-flow stirred reactor	Methanol, 25 °C	MeOH/NO3−-N 3.0			142 mgNO3-N/g VSS,h	100%	Mixed bacterial culture, originated from two-stage anaerobic–aerobic industrial yeasts production wastewater treatment plant, HRT 51.6 h, DO 2.5 mg/L	2 [152]
Denitrification biofilter, gravel as matrix, batch mode, straw of 5 cm	Flower straw	57.0 ± 0.4 mg NO3−-N/L	Wastewater fed at the beginning of each batch within 30 min, and then drained with gravity after a retention time of 3 days (72 h)	Nitrate removal was efficiently enhanced by the addition of flower straws, but decreased gradually as the organic substances were consumed.	High nitrate removal rates	cca 35%	Optimisation of carbon source addition is required	9 [80]
SBR, feast/famine regime	PHA, 30 °C	CODPHA/N ratio higher than 5.4 g/g	HRT 24 h Allylthiourea added	PHA concentrations lower than 5% do not allow the obtainment of maximal specific endogenous denitrifying activity value	0.26–0.39 g N/g VSS,d	100%	No N2O was detected in the gas phase	[117]
Up-flow lab reactor, corn cob granules of 2 cm	Corn cobs, 27–33 °C	25.3 mg N/L	The weight and filling height of the substances were measured to determine the percentage loss of material during the experiment	A time-dependent decrease in nitrate removal efficiency was observed after 67 days of operation. The addition of fresh corn cobs brought about a rapid increase in nitrate removal efficiency	0.203 kg/m3,d	cca 100%	Carbon source and biofilm carrier, low nitrite accumulation	10 [153]
Pilot-scale drainage bioreactor	Wood chips, 10–15 °C	10.1 mg NO3−-N/L (mean)	Pilot-scale reactors with identical volumes (0.71 m3) and depths (0.6 m), and three cross-sectional geometries—channel, rectangular, and trapezoidal—were constructed with plywood	The percent reduction of the influent nitrate mass was linearly correlated to the theoretical HRT with 30 to 70% NO3−-N removals observed within the 4 to 8 h of retention time suggested for field installations	3.8–5.6 g N/m3,d	30–70%	Suitable for drained agriculture fields to surface water	[154]
Woodchip solid-phase denitrification bioreactor with gradient filling or uniform filling bioreactor	Wood chips	50.04 ± 0.81 mg NO3−-N/L	pH 7.89 ± 0.10 DO 7.21–8.89 mg/L Wood chips 1.0–5.0 mm size 150-day operation	Gradient filling improved the NO3− removal rate and reduced the bioreactor size. Better option than uniform filling bioreactor	35.66–174.55 mg N/L,d—gradient filling bioreactor 25.17–111.72 mg N/L,d—uniform filling bioreactor	100%—gradient filling bioreactor 100%—uniform filling bioreactor	Gradient filling bioreactor promoted complete denitrification with low NO2− accumulation	11 [155]
SBR	Industrial waste polyglycolic acid	C/N 3–5	12 h cycle, feeding 0, 2 h, anaerobic 5 h, aerobic 6 h, settling 0.6 h, drainage 0.2 h	An optimal denitrification performance in a methanol-fed activated sludge system was achieved with a polyglycolic acid dosage of 1.2 mL/L, pH 7–8, DO 3 ± 0.5 mg/L		47–89%	Polyglycolic acid upregulated the expression of nitrogen-metabolism-related genera, including amo, hao, nar, and nor, which improved the denitrification performance of the system	[156]
Flat biofilm reactor	Polycaprolactone, 24 ± 2 °C	NO3−-N ~20.0 mg/L	HRT: 2.5, 2.0, 1.5, 1.0, and 0.5 h	Coexisting ecological assemblages and coupled metabolic patterns of polymer degradation and denitrification in the system	12.58 mg/L,h	30–100%	Dominant denitrifying bacteria replaced by Acidovorax, which is capable of metabolising polyester	[157]

¹ https://www.laballey.com/products/sodium-acetate-anhydrous-powder-lab-grade?variant=7219061030971 (accessed on 3 June 2024). ² https://www.methanex.com/about-methanol/pricing/ (accessed on 3 June 2024). ³ https://grains.org/ethanol_report/ethanol-market-and-pricing-data-january-3-2024/ (accessed on 3 June 2024). ⁴ https://www.chemicalbook.com/SupplierPriceList_EN.aspx?cbn=CB2250047&c=100kg#price (accessed on 3 June 2024). ⁵ https://www.sigmaaldrich.com/HR/en/substance/sodiumpropionate9606137406 (accessed on 3 June 2024). ⁶ https://www.sigmaaldrich.com/HR/en/product/sigald/g7893 (accessed on 3 June 2024). ⁷ 0.48 USD/L, MicroCTM is a proprietary product, undisclosed composition, non-flammable, agriculturally derived carbon source [38]. ⁸ https://scottlabsltd.com/en-us/ascorbic-acid-500-g-30--15053j (accessed on 3 June 2024). ⁹ https://ahdb.org.uk/dairy/hay-and-straw-prices (accessed on 3 June 2024). ¹⁰ https://www.alibaba.com/showroom/corn-cob-price.html (accessed on 3 June 2024). ¹¹ https://www.paperindex.com/product-listings/wood-chips/18614/21 (accessed on 3 June 2024).

.

7. Limitations and Future Perspectives

In this review, the main focus was to discuss alternative carbon sources vs. traditional carbon sources in the context of the carbon source price, C/N ratio, denitrification efficiency and rate, operational parameters, and advantages and disadvantages, as well as the limitations in the denitrification process.

The advantages and disadvantages/limitations when used liquid external carbon sources are as follows:

i.

They are readily degradable, highly efficient, and have a high rate of denitrification [7,57,61].

ii.

They can be quickly decomposed during non-denitrification processes, which can often result in problems of low cost-effectiveness [33,34,35].

iii.

Some compounds in liquid form are highly combustible and must be strictly controlled during transport, storage, and operation [36,37,38].

iv.

The storage, pumping, and delivery systems for combustible carbon sources require higher capital construction costs to fulfil safety standards in comparison to other non-hazardous and non-flammable carbon sources [41].

v.

Overdose is a common problem that may result in significant secondary contamination [58].

vi.

There is a long adaptation period of activated sludge microorganisms to some carbon sources [40].

The advantages and disadvantages/limitations when used solid external carbon sources are as follows:

i.

The release of carbon is gradual, making it an easy-to-manage, long-term operation [58].

ii.

They are low in cost, easily available, easy to control, and can avoid the overdose problem [58].

iii.

It is easy to maintain a stable operation for the system [58].

iv.

The usage of natural carbon sources is limited because of increased colour intensity in the effluent, excessive dissolved organic carbon release, unstable carbon release rates, and slow denitrification rates [58].

v.

The chemical structure and molecular weight of biodegradable polymers considerably influence the mechanism of the utilisation of the carbon source, particularly the microbial community [87].

vi.

The amount of CO₂ released is lower than from liquid carbon sources [59].

vii.

The drawbacks include possible susceptibility to the operational conditions; low electron donor availability; the accumulation of products such as greenhouse gases, residual organic compounds, nitrite, nitrous oxide, and ammonium; and a low efficiency rate [62].

Future perspectives may include the operating parameters influencing the stability of the removal performance; the maintenance and improvement of nitrate removal; the study of the diversity of the microbial community; research into the application of new alternative carbon sources in denitrification; and N₂O emission detection and minimisation.

8. Economic and Environmental Benefits and Drawbacks of Some Carbon Sources for Denitrification

In selecting an appropriate substrate for denitrification, a number of factors should be considered, such as denitrification potential, yield, denitrification rate, and the required COD/N ratio. The reported COD/N ratios for some traditional carbon sources are 4.45 g COD/g NO₃⁻-N for methanol in the continuous experiments [40]; 4.0 g COD/g NO₃⁻-N for methanol at 15 °C and 4.16 g COD/g NO₃⁻-N for methanol at 25 °C in pure culture batch cultivations [40]; 4.16 g COD/g NO₃⁻-N corresponding to 2.0 g ethanol to denitrify 1 g of NO₃⁻-N [158]; and a COD/NO₃⁻-N ratio of 3.1 to 3.7 for denitrification with acetic acid as a carbon source [159]. The literature reports yields for methanol of 0.57–0.66 [160], 0.4 at a pH of 7.0 [161], 0.4 mg COD/mg COD [162], 0.45 ± 0.05 g COD/g COD at 13 °C; 0.53 ± 0.06 g COD/g COD for ethanol at 13 °C; and 0.66 ± 0.06 g COD/g COD for acetate at 13 °C [21]. The values of the specific denitrification rate (SDNR) for acetate (31.0 ± 4.6 mgNO₃⁻-N/g VSS,h) and ethanol (29.6 ± 5.6 mgNO₃⁻-N/g VSS,h) are higher than that for methanol (10.1 ± 2.5 mgNO₃⁻-N/g VSS,h) at 13 °C. Also, the operating cost for ethanol is more than twice that of methanol, and for acetate is over seven times that of methanol [21].

Applying food industry effluent and industrial wastewater as external carbon sources for denitrification has benefits such as availability in the required quantities, stability in terms of the composition and content of readily biodegradable organic compounds, and cleanliness, since they are relatively free of metals and other contaminants [37]. The fermentation products of anaerobic digestion from food waste, as an external carbon source, show a lack of practical application, such as the chemicals required, especially for pH adjustment; high time consumption; and the degradation of organic substrates—more than 50%—in the form of gas release [163,164,165]. By using the fermentation liquid from food wastes as a carbon source, produced by short-term free anaerobic fermentation, the maximum denitrification was achieved at COD/TN 6, with a denitrification rate of 12.89 mg/g VSS,h and a denitrification potential of 0.174 gN/gCOD, and value of yield of 0.5 g COD/gCOD. The productivity of carbon sources from food waste is 0.096 g SCOD/g FW (readily biodegradable chemical oxygen demand, SCOD), and the readily biodegradable fraction in the fermentation liquid from food waste was evaluated as 58.35%, which indicates that it is a high-quality carbon source for enhancing denitrification [37], while the readily biodegradable COD of domestic wastewater is usually 10–20% of total chemical oxygen demand [166]. The value of the yield coefficient of 0.44–0.65 was determined with agro-food wastewaters [167], and 0.66 was calculated under an assumption that the internal storage of soluble COD occurred with acetate [168]. It was calculated that 0.9–1.2 kg food waste would be needed for the production of fermentation liquid from food wastes for the treatment per m³ wastewater, which is equivalent to the food waste generated by two or three persons for complete N removal [169].

A promising alternative carbon source in biological nutrient removal is waste-activated sludge (WAS) fermentation liquid, generally rich in volatile fatty acids (VFAs), which are the preferred carbon source for denitrification [170,171,172,173,174]. There were more refractory and non-biodegradable organics released from the alkali-treated sludge, which were difficult to utilise with denitrifying bacteria. The major recalcitrant compounds included building blocks, high-molecular-weight protein, and high-molecular-weight polysaccharides. Also, the alkaline-fermented sludge showed very poor dewaterability, which, in turn, would cause difficulties and high operating costs in liquid and sludge separation [33].

Among natural materials, wood chips are the most widely and successfully used carbon source [175,176]. The benefits of their applications are their high permeability, low material cost, high C/N ratio, and robust durability [177], while the drawback is their low release of dissolved organic carbon, resulting in a lower denitrification capability than for easily released carbon sources [178]. Hence, to ensure complete denitrification, an excess amount of the solid carbon source is usually used, but this results in high amounts of dissolved organic carbon in the effluent [177].

9. Conclusions

The research progress of the application of various carbon sources in biological denitrification are summarised in this review, which points out the prospects and the challenges of denitrification with traditional and alternative carbon sources by estimating their advantages, disadvantages, and costs. The outcome of the denitrification process is affected by the type of selected carbon source, utilisation degree, produced metabolites, denitrification rate, and functional microorganisms. Many studies have shown that alternative carbon sources can replace traditional carbon sources. For the selection of external carbon sources, numerous factors should be taken into account, including the operation cost, denitrification efficiency, dosing conditions, and effluent quality. The evaluation of denitrification performance and potential cost reduction includes the calculation of the denitrification rate and efficiency, and the denitrification potential of different carbon sources. A potential contribution to cost reduction in denitrification is the use of alternative carbon sources. For example, the prices of some alternative carbon sources are 108.3 EUR/t (wheat straw), 0.1–0.2 EUR/t (corn cob), and 128.6 EUR/t (wood chips); and the prices of some traditional carbon sources are 14.2 EUR/kg (sodium acetate), 410.6 EUR/L (ethanol), and 535 EUR/t (methanol). Future research should focus on the investigation of the use of new alternative carbon sources for biological denitrification, the operating parameters influencing the stability of the removal performance, the functional microorganisms responsible, the maintenance and improvement of stable nitrate removal, and N₂O emission detection and minimisation. This review has the potential to influence the selection of carbon sources in wastewater treatment plants, which could lead to significant cost reductions and improvements in process efficiency.