The anaerobic digestion of substrates with high ammonia content has always been a bottleneck in the methanisation process of biomasses. Since microbial communities in anaerobic digesters are sensitive to free ammonia at certain conditions, the digestion of nitrogen-rich substrates such as livestock wastes may result in inhibition/toxicity eventually leading to process failures, unless appropriate engineering precautions are taken. There are many different options reported in literature to remove ammonia from anaerobic digesters to achieve a safe and stable process so that along with high methane yields, a good quality of effluents can also be obtained. Conventional techniques to remove ammonia include physical/chemical methods, immobilization and adaptation of microorganisms, while novel methods include ultrasonication, microwave, hollow fiber membranes and microbial fuel cell applications. This paper discusses conventional and novel methods of ammonia removal from anaerobic digesters using nitrogen-rich substrates, with particular focus on recent literature available about this topic.
Anaerobic digestion (AD) is an effective biotechnology in waste and wastewater treatment processes to reduce waste volume and to produce agricultural fertilizers as well as biomethane, a source of renewable energy. AD processes require a low energy input and have a relatively lower amount of sludge production in comparison to aerobic waste treatment systems. The biogas process is divided into four phases: hydrolysis, acidogenesis, acetogenesis and methane formation, which are carried out by different groups of microorganisms. However, the conflicting needs between microorganisms regarding nutrients, environmental conditions and growth kinetics, especially between the fast growing acid formers (Bacteria) and the slow growing methane formers (Archaea), often cause process instability. Correspondingly, the microorganisms react in a varying degree to the presence of inhibitory substances in wastes and wastewater. The dysfunction of one single microbial group participating in the anaerobic food chain affect the whole digestion process.
Inhibitors in AD processes include organics or inorganics in nature or a combination thereof. Among these compounds, chlorophenols, halogenated aliphatics, N-substituted aromatics, long chain fatty acids, lignins and lignin related compounds are potential organic inhibitors. Inorganic inhibitors are ammonia, sulfide, light metal ions (Na, K, Mg, Ca, Al) and heavy metals (Anjum et al. 2016). However, the limit of toxicity concentrations vary considerably, since AD processes are very complex. Particularly, the interactions of microbial communities and their ability to adapt to substrate and process conditions are often unpredictable. Therefore, the inhibition of AD systems due to excess ammonia levels have often been reported and studied in literature. This paper provides an overview of research activities in AD conducted on ammonia mitigation from animal manure, with particular attention to recent methodologies.
Nitrogen, at an optimal level, is a nutrient source for microorganisms and ensures the buffering capacity in AD processes. However, substrates with high nitrogen fractions inhibit the degradation, which may lead to reduced biogas quality and quantity. Hereof, a wide range of inhibiting ammonia concentrations were reported in the literature, ranging from 1,700 mg/L to 14,000 mg/L (Anjum & Krakat 2016). But also, ammonium has been known for having a toxic impact on the microbial biocoenosis. Concentrations greater than 3,000 mg/L showed an inhibitory effect by 40% on the biogas process (Braun 1982).
During the AD process, the organic nitrogen in the form of proteins, amino acids and uric acids is hydrolyzed to inorganic ammonia. The protein degradation process is very slow and the released ammonia tends to accumulate (Zeshan et al. 2012). Substrates known to have high nitrogen contents are animal wastes, municipal wastes (bio-wastes), meat processing wastes and dairy wastes. Besides being an inhibitor in the AD process, free ammonia is also an environmental pollutant. Free ammonia and the increased eutrophication, respectively, is toxic for most fish species, it decreases dissolved oxygen, causes corrosion and reduces disinfection efficiencies (Lauterböck et al. 2012).
As the human population grows, the demand of meat and dairy products also increases. Animal by-products, including chicken poultry, swine and cattle manure grows as well. Particularly, the global population of poultry industries has been growing rapidly. The average stock of poultry is almost 23 billion worldwide, producing about 587 billion tons of ammonium-rich excreta per year (Anjum et al. 2016). A part of this manure is used as fertilizer, but the major portion must be disposed of properly. Untreated, the animal wastes cause environmental problems, such as malodor, ammonia volatilization and groundwater contamination.
Slaughterhouse wastes are by-products of the meat processing industries. They contain the inedible parts of animals, such as bones, blood, viscera and feathers. Due to their high lipid and protein content, slaughterhouse wastes offer a sustainable treatment method to generate energy in combination with the use of residues as fertilizers (Edström et al. 2003). However, to meet the current European regulations for the disposal or use of animal by-products, slaughterhouse wastes must be pretreated. Other compounds with high nitrogenous content are dairy effluents, wastewater from food industry processing whey, cheese and casein (Kovács et al. 2013). The composition and properties of feedstock is consequential to evaluate further treatment. Although AD offers a solution to the waste management with many advantages, the nitrogen problem still persists.
Many possibilities to control ammonia inhibition have been studied and reported in literature. Some methods were practical and applicable in large scales, while some are still in research at laboratory scale. Nevertheless, each method has its pros and cons, depending on the inoculum, type and characteristics of the substrate, reactor configuration, and environmental and operational conditions.
AMMONIA INHIBITION MECHANISM
Only a proportional fraction of the organic nitrogen is biologically degraded to inorganic ammonia (-N/NH3-N). Gallert & Winter (1997) reported that only about 1/3 and 1/2 of the total Kjeldahl nitrogen was converted to ammonia during mesophilic and thermophilic degradation. Bujoczek et al. (2000) referred to an organic nitrogen conversion rate ranging between 62.6% and 80.3%. Yabu et al. (2011) reported that only 34–38% of the total nitrogen of the garbage was converted to ammonia.
Among the anaerobic degrading microorganisms, methanogens (Euryarchaeota) are reported to be the most affected groups by elevated ammonia levels (>1,800 mg/L) and the first to be inhibited (Kayhanian 1999; Chen et al. 2008; Niu et al. 2014). However, hydrogenotrophic methanogens were reported to be dominant in engineered habitats with low ammonium concentrations (Krakat et al. 2010a, 2010b). As for the bacterial consortium, mostly the dominance of the Firmicutes phylum over the Proteobacteria phylum was reported in stable AD processes, although the dominance of Firmicutes, Bacteroidetes, and Proteobacteria was also reported (De Vrieze et al. 2015). Furthermore, the hydrolysis and acidogenesis conversion ratio declined when the total nitrogen concentration was higher 5,000 mg/L (Niu et al. 2014).
Several effect mechanisms of ammonia inhibition, mainly studied with pure cultures of methanogens, were proposed: (1) free ammonia inhibits the methane synthesizing system directly, (2) free ammonia causes proton imbalance, because ammonia diffuse passively into the cells of methanogens, (3) free ammonia changes the intracellular pH, (4) free ammonia increases the maintenance energy requirement, and (5) free ammonia inhibits specific enzymatic reactions (Wittmann et al. 1995; Kayhanian 1999; De Vrieze et al. 2015).
When the methanization is interrupted, the concentration of volatile fatty acids (VFA) accumulates. In turn, the pH remains stable unless there is enough buffering capacity (alkalinity) available in the system and, thus, the concentration of free ammonia decreases. That results in an ‘inhibited steady-state’, a condition with significant losses of the biogas and methane production yields (Fotidis et al. 2014). It was reported by Yenigün & Demirel (2013), that the ammonia inhibition was reversible for the methanogenesis phase. The microbial community even increases its tolerance towards ammonium after recovery. However, in contrary, Niu et al. (2014) reported that the former steady-state could not be recovered after a thermophilic reactor was loaded with a very high amount of nitrogen.
Methanogens are primarily classified as acetoclastic, hydrogenotropic and methylotrophic organisms (Liu & Whitman 2008). There is contradictory information reported in literature regarding the sensitivity of the methanogens to environmental conditions. Acetoclastic methanogens are acetate consumers, while hydrogenotropic methanogens are capable of utilizing H2, CO2 and formate to produce methane. Some studies in literature observed that acetoclastic methanogens were more sensitive to ammonia inhibition (Borja et al. 1996; Calli et al. 2005; Niu et al. 2013), while others observed a lower tolerance by hydrogenotropic methanogens (Karakashev et al. 2005; Chen et al. 2008; Song et al. 2010). At elevated ammonia levels, the microbial community shifts from the acetoclastic methanogenesis to syntrophic acetate oxidation (Schnürer & Nordberg 2008; Westerholm et al. 2012a). The syntrophic acetate oxidation involves a two-step reaction consisting of acetate oxidation to hydrogen and carbon dioxide by syntrophic acetate-oxidizing bacteria, followed by conversion of these products to methane by hydrogenotropic methanogens (Westerholm et al. 2012a).
Some studies reported that thermophilic microbial consortia could tolerate double the amount of free ammonia compared to mesophilic ones (Gallert & Winter 1997; Gallert et al. 1998), while others studies revealed that thermophilic conditions with high free ammonia concentrations were unstable and could be more easily inhibited (Braun et al. 1981; Hashimoto 1986; Angelidaki & Ahring 1994).
The toxicity limits of free ammonia described in literature differed significantly, with concentration ranging between 50 to 1,500 mg NH3-N/L (Braun et al. 1981; Koster & Lettinga 1984; Angelidaki & Ahring 1993, 1994; Gallert & Winter 1997; Hansen et al. 1998; Kayhanian 1999; Bujoczek et al. 2000; Siles et al. 2010). Nevertheless, ammonium thresholds greater than 12 g/L were also reported as tolerable concentrations for microorganisms detected in thermophilic fermented chicken dung (Anjum & Krakat 2016). The degradation process was either partly or totally inhibited at given concentrations and conditions. Many research groups have published results of ammonia thresholds. A summary of these results was provided in different reviews (Rajagopal et al. 2013; Yenigün & Demirel 2013). The varying results about methanogens sensitivity and ammonia toxicity limits may be attributed to (1) differences in the types of the substrates used, (2) varied inoculum sources, (3) environmental conditions such as pH, temperature, and (4) operational conditions such as substrate loading rate, retention time and acclimation periods microorganisms (De Baere et al. 1984; Angelidaki & Ahring 1994; Liu & Sung 2002).
STRATEGIES TO OVERCOME AMMONIA INHIBITION
The conventional biological method of ammonia elimination is nitrification and denitrification. These processes require a significant energy input and carbon sources. Therefore, they are not suitable in combination with AD processes. Moreover, the process is generally performed on wastewater containing low nitrogen concentration (Ahn 2006; Lin et al. 2009b). But also less common approaches such as break point chlorination and membrane based technologies are used.
Some studies focused on the adjustment of feedstock C/N ratios to guarantee an optimal microbial growth (Kayhanian 1999; Siles et al. 2010; Zeshan et al. 2012). Kayhanian (1999) reported a C/N ratio of 27 to 32 as sufficient for the degradation of biodegradable organic fraction of municipal solid waste. Zeshan et al. (2012), however, prepared simulations using bio-degradable feedstocks and identified a C/N ratio 32 as appropriate to minimize or to avoid ammonia inhibition at dry thermophilic ADs, while a C/N ratio 27 was less adequate. Siles et al. (2010) mentioned, that the biogas production from the degradation of glucose-containing synthetic solutions decreased significantly when the C/N ratio was below 4.42.
A common approach is to dilute the substrate to a total solids (TS) level of 0.5%–3% (Chen et al. 2008). Even though no additional equipment was required, this approach proved to be economically unattractive, since dilution seriously decreased the digester's gas production level and also resulted in an increased waste volume and increased dewatering costs, respectively (Kayhanian 1999).
The anaerobic population of wastewater reactors and biogas plants may overcome the inhibitory effects of ammonia by acclimation or long-term adaptation (Angelidaki & Ahring 1993; Beinersdorf et al. 2015). Yet adaptation processes may take a long time and, thus, they are not suitable for a fermentations in which ammonia concentrations vary strongly and shock loadings may occur.
Other processes based on a physical-chemical reaction are namely stripping and chemical precipitation. The ammonia can either be reduced in a pretreatment step, during AD or as a post-treatment of the AD effluent (Serna-Maza et al. 2015). For instance, stripping is a conventional method, which is applicable as a pre-treatment step (Zhang et al. 2012; Markou 2015), during the digestion (Yuan et al. 2016) and as post-treatment (Lei et al. 2007). A successful air stripping as pretreatment step was performed by Zhang et al. (2012) to increase the methane yield from AD of piggery wastewater. The ammonia stripping pre-treatment showed that the removal of ammonia seemed to be dependent on pH and aeration rate, and particularly, it could be an option at alkaline pH levels to prevent digester failure during anaerobic treatment of raw piggery wastewater. Bousek et al. (2016) recently investigated effects of ammonia stripping in anaerobic digesters with high ammonia levels. The authors concluded, that the side stream air stripping could be a promising tool to decrease high ammonia levels in anaerobic digesters. However, this approach might pose adverse impact on the microbial community due to inhibition through oxygen exposure. Abouelenien et al. (2010) reported that nitrogen can successfully be eliminated by air stripping combined with gas washing in sulfuric acid to capture ammonia when chicken dung manure was digested. Thereby, a recirculation of the stripped liquid fraction to the fermenter could be a solution in large-scale applications to control the ammonia concentration (Nie et al. 2015). However, other carrier gases such as biogas and hot steam can also be used for elimination processes.
Numerous studies have been conducted to find practical solutions to overcome ammonia inhibition. These strategies are discussed in the following parts.
The addition of ‘stress-resistant’ – or ‘efficient biogas producing’ microorganisms in bioreactors to improve the AD process, known as bioaugmentation, has been used for the last 15 years (Nzila 2016). One of the advantages of bioaugmentation is the decreasing of the adaptation period of bioreactors. The challenge for the bioaugmented culture is, however, to adapt and to survive under reactor conditions, which may be not similar to their optimal growth conditions. Moreover, added cultures must be able to assert themselves to compete with the established microbial populations.
Lü et al. (2013) studied the combined effect of acids and ammonia on acetoclastic methanogens, hydrogenotrophic methanogenes and syntrophic acetate oxidizers (SAO) during a thermophilic AD process. It was reported that, foremost, the stress of acetate and ammonia was synergetic. Methanogenic pathways were regulated by acetate and ammonia, jointly. Therefore, the microbial population structure in a thermophilic anaerobic culture shifted, depending on the combined effect of acetate and ammonia. At lower acetate concentrations (50 mmol/L), acetoclastic methanogens were dominant, regardless of the ammonia concentration. When the acetate concentration was high (150 mmol/L and 250 mmol/L) and ammonia levels were moderate (1–4 g/L), acetate was mainly degraded by the acetoclastic methanogens.
Generally, the efficiency of acetate degradation by both SAO and hydrogenotrophic methanogens were lower than that of the acetoclastic methanogens. However, at elevated ammonia concentrations, SAO and hydrogenotropic methanogens were more robust and could continue to function.
Nielsen et al. (2007) studied the effect of bioaugmentation during a two-stage thermophilic AD of cattle manure containing large fraction of biofibers with two different strains of hydrolytic hyperthermophilic bacteria, Caldicellulosiruptor lactoaceticus and Dictyoglomus. Even though the free ammonia concentration was lower than the reported inhibition concentration, the improvement of the methane yield was modest, only about 9–10%, compared to that in the control reactors without bioaugmentation.
Westerholm et al. (2012b) explored the enhancement of the biogas yield by bioaugmentation with SAO cultures containing Clostridium ultunense sp., Syntrophaceticus schinkii, Tepidanaerobacter acetatoxydans and Methanoculleus sp. The cultures were added daily to stirred tank reactors fed with a mixture of cattle manure and whole stillage. To achieve elevated ammonia concentrations, the whole stillage was substituted by egg albumin powder. Surprisingly, SAO activities were also detected at low acetate and ammonia concentrations. This was attributed to the long hydraulic retention time adjusted. At higher dilution rates, the acetoclastic methanogens dominated. Overall, the results of reference and bioaugmented reactors were similar, which indicated that the bioaugmentation with SAO and hydrogenotropic cultures did not have any dominant influence on the acetate degradation pathway. Moreover, the bioaugmentation did not improve the reactors’ operation during periods of increasing ammonia levels.
A similar study was conducted by Fotidis et al. (2013) in mesophilic up-flow anaerobic sludge blanket reactors fed with synthetic medium containing glucose and NH4Cl as carbon and ammonium nitrogen source, respectively. A SAO co-culture (Clostridium ultunense spp. in living association with Methanoculleus spp.) was used for bioaugmentation. Again, an identical performance of reference and bioaugmented reactors was observed indicating that the bioaugmentation process did not affect the methane production at high ammonia levels. This could be attributed to very slow growth rates of SAO co-cultures and to the challenge to immobilize used co-cultures. To elucidate the reason for the unsuccessful bioaugmentation, the SAO co-culture was co-cultivated with hydrogenotrophic Methanoculleus bourgensis in fed-batch reactors. The total incubation period of SAO co-culture was lowered by 33% with M. bourgensis and increased the growth rate by 42%. This indicated that Methanoculleus spp. as partner in the SAO co-culture was the rate limiting factor in the consortium. Presumably, Methanoculleus spp. reduced the hydrogen partial pressure less effectively as Methanoculleus bourgensis and, therefore, did not bring any advantage for the biomethanization process.
An improved bioaugmentation was conducted with Methanoculleus bourgensis by Fotidis et al. (2014) in continuous stirred tank reactors (CSTR) fed with pig and cattle manure (70–90%) and organic waste (10–30%). Methanoculleus bourgensis was a fast growing hydrogenotrophic methanogen that could produce methane at elevated ammonia concentration (4 g ). In comparison to the reference reactors, the methane yield of the bioaugmented reactors showed an immediate improvement by an increase of 31.8%. The bioaugmentation was successfully performed without interrupting the continuous operation of the reactor and without changing the ammonia-rich feedstock. It was assumed that a ‘critical biomass’ of ammonia tolerant methanogens was necessary, which was a minimum amount of microorganisms needed to promote the desired microbial activity in the reactor. The advantages of bioaugmentation include no changes in reactor operational parameters such as pH and temperature, and a low cost. However, choosing the suitable microorganisms and wash-out occurrences are potential disadvantages of bioaugmentation.
Anaerobic ammonia oxidation
A further strategy to remove nitrogen form process liquids is the anaerobic ammonium oxidation (Anammox) process catalyzed by chemolitoautotrophic Planctomycetes. The process deals in the oxidation of ammonia to di-nitrogen gas with nitrite as an electron acceptor (Kimura et al. 2010). Commonly, is removed by nitrification to nitrate, followed by denitrification. In Anammox-based processes, only about 50% of ammonium is oxidized to nitrite, when oxygen is present as terminal election acceptor, after which, in oxygen free environments, the produced nitrite reacts with the remaining ammonium to form nitrogen gas (N2). This approach is preferably used for the elimination of nitrogen from wastewaters with low C/N ratios and high ammonium levels with low amounts of organics, respectively (Zhang et al. 2016a). Therefore, aerobic single reactor systems for high ammonia removal over nitrite treatment processes (SHARON) are frequently used which involve part conversion of ammonium to nitrite often coupled with a second reactor in which Anammox occurs. Alternatively, the completely autographic nitrogen removal over nitrite (CANON) process may be used, which involves nitrogen removal within one single reactor under oxygen limited conditions (Khin & Annachhatre 2004).
The SHARON-Anammox process does not require the addition of chemical oxygen demand (COD) and, thus, the combined system saves 50% of oxygen and 100% on the external carbon source, reducing CO2 emissions by about 100%. Overall, this approach is 90% less expensive than conventional processes (Dijkman & Strous 1999).
The CANON process is autotrophic and, therefore, requires no added COD. The entire nitrogen removal can be achieved in a single reactor with little aeration reducing space and energy requirements. The CANON process consumes 65% less oxygen and 100% less reducing agents compared to conventional removal process (Khin & Annachhatre 2004). However, the Anammox process offers many advantages and disadvantages for the ammonium removal (van der Star 2008).
Anammox – advantages:
No requirement for external oxidants
No CO2 emission (CO2 is consumed by autotrophic Anammox-bacteria instead of produced)
Lower energy costs due to minimized aeration costs.
Anammox – disadvantages:
Anammox organisms grow notoriously slowly (doubling time is approximately 10–15 days under known optimal conditions).
Anammox bacteria react very sensitively to oxygen, even at low concentrations, with growth inhibition and cell death as a consequence.
High sensitivity to changing environmental conditions.
Natural zeolites crystalline, hydrated alumina silicates of alkali and alkaline earth cations, are large mineable deposits in many parts of the world. Especially in AD processes, the porous materials of natural zeolites provided a support and immobilization material enabling the anaerobic reactor to retain high biomass concentrations and significantly reduced retention times. Moreover, the properties of zeolite also showed a great capacity for metal adsorption, which was useful for removing toxic materials causing inhibition in AD processes (Zhang et al. 2016b). So, the high ion exchange capacity of zeolites, their large reserves, the shortage of competing minerals and the relatively low market price make zeolite an attractive mineral to be applied in large scales (Lin et al. 2013).
Two mechanisms of zeolites were advantageous for the AD process: as an ion exchange element for NH4-N and the adsorption of NH3. Due to the presence of Na+, K+, Ca2+ and Mg2+ cations, zeolite has been used as an ion exchange element for NH4-N- removal in AD (Montalvo et al. 2012). Since the cations are loosely bound, they can be easily exchanged by other surrounding cations, for example by ammonium ions. These cations stimulated the microbial growth and they were antagonistic to ammonia inhibition. However, when high levels of these cations were exchanged and released at the adsorption process, they were toxic for the microorganisms (Lin et al. 2013). Another mechanism proposed that the zeolite adsorbed NH3 on active areas of the material (Milán et al. 2001).
Zeolites can either be used in their natural form or modified. Since the natural zeolites have a lower adsorption capacity, they usually need to be modified to improve their adsorption capacity and purity (Huang et al. 2014). One of the possible modification methods is a sodium chloride (NaCl) solution treatment. The surface of modified zeolite became rougher and more irregular, compared with the natural zeolite that, in turn, increased the surface area, total pore volume and average pore diameter significantly (Lin et al. 2013). After that modification, more Na+ ions were available and they were readily exchanged by ammonium ions on zeolite surfaces. NaCl modification also replaced Ca2+ and Mg2+ ions on zeolite, resulting in the production of large pores and cavities in the zeolite. In this study, the modified zeolite exhibited 58% higher ammonium adsorption capacity and faster adsorption rates.
The doses and types of zeolites to be employed is an important parameter. When the amount of zeolite is too high, the apparent viscosity will increase and thereby will hinder the mass transfer between the substrate and microorganisms (Milán et al. 2001).
Milan et al. (1997) compared the effect of different homoionic zeolites on piggery wastewater by anaerobic fixed bed reactors (Na, Ca, K and Mg). From all these ions, homoionic sodic zeolite showed the best performance, with 91% and 58% ammonia removal capacity after 10 h and 30 h of operation, respectively. According to the experimental results obtained, the exchange capacity was in the order Na-Zeo > Ca-Zeo > K-Zeo > Mg-Zeo. This order of exchange capacity was also confirmed by Lin et al. (2013). Especially Ca-Zeo and Mg-Zeo were affected by the high concentration of suspended solids and the viscosity of the influent, which led to a low mobility of these cations in the liquid phase. Moreover, Lin et al. (2013) stated that Na+ was only dominant when the ammonium concentration was low (under 500 mg N/L). At high ammonium ion concentration (>1,000 mg N/L), Ca2+ replaced Na+ as dominant ions for ammonium adsorption. Depending on the pH, the adsorption mechanism shifted from ion exchange to molecular adsorption. The significance of molecular adsorption was negligible at pH < 8. At a higher pH, the molecular adsorption contributed more to the ammonia removal, and became significant at pH 11, owing to the absence of ionized ammonium in the solution.
A series of experiments treating swine manure with zeolite doses between 0.2 and 10 g/L showed that the AD process was favored by the addition of natural zeolite at doses between 2 and 4 g/L and increasingly inhibited at doses beyond 6 g/L (Milán et al. 2001). A digestion failure was observed at zeolite doses of 10 g/L. This was explained by the decrease of total volatile fatty acid/alkalinity ratio and the simultaneously increasing pH value.
Since the thermophilic AD of animal wastes or manure was very challenging, mainly due to the ammonia inhibition, relatively fewer studies were conducted at this temperature. By increasing the ammonia concentration gradually up to 5 g/L, Angelidaki & Ahring (1992) studied the effect of addition of bentonite and the waste product bentonite-bound oil (BBO), each with 1.1% bentonite, on the AD of cattle manure under thermophilic conditions. Bentonite is a clay mineral with a characteristically layered structure, allowing water, numerous elements and organic matter to enter the layer spaces. In a continuous process, as the ammonia concentration was gradually increased, the methane production decreased in all reactors. However, bentonite and BBO added reactors recovered and reached the same level as before the inhibition. In batch culture experiments, the addition of bentonite shortened the lag-phase.
Kotsopoulos et al. (2008) studied the effect of zeolite addition on the thermophilic AD of pig wastes. The preliminary experiments showed that the obvious differences in biogas production were observed at zeolite doses of 0, 4, 8, and 12 g/L, where the degradation process also ceased after 15 days. Especially, the addition of 8 g/L and 12 g/L of zeolites increased the methane production significantly. The total ammonia concentrations only slightly decreased when increased zeolite doses were added. However, the free ammonia concentration increased in the treatments with zeolite. The pH value remained, in any case, within the optimal range for the AD (6.8–7.5). Thus, it was also concluded in this study that the positive effect of zeolite addition was not only the result of the ion exchange capacity of zeolite, but also due to the immobilization surface for the microorganisms.
The growth rate of microorganisms immobilized on zeolites was higher compared to those in suspension, as calculated by Borja et al. (1994). The kinetic constant and maximum growth rate in the zeolite loaded reactor were 35% and 59% higher compared to the control reactor without zeolite. The maximum volumetric methane production rate was also increased by approximately 62% when biomass was added. This result was confirmed by Nikolaeva et al. (2009) using up-flow anaerobic fixed bed digesters packed with waste tire rubber. Particularly the combination of waste tire rubber and zeolite improved the values of the maximum methane yield and kinetic constants by 11.1% and 29.4%, respectively.
Different removal strategies also were investigated by Ho & Ho (2012) to mitigate ammonia inhibition on thermophilic AD of piggery wastewater. The pH was adjusted and the effects of different additives (biomass, natural zeolite and humic acid) were investigated, individually or in combination. The enhancement with biomass and humic acid did not show any desirable result, with or without pH adjustment. On the other hand, the addition of 10–20 g/L of zeolite showed an increase in methane production. The enhancement effect of zeolite was greater on the pH-unadjusted piggery wastewater (pH 8.1) than on the pH-reduced wastewater (pH 6.5). The high ammonia concentration of piggery wastewater used in this study, about 1,740 mg NH4-N/L, presumably also required a high amount of zeolite addition. However, the ammonium-nitrogen concentration remained unchanged after 10 days of batch digestion, which suggested that the applied zeolite concentrations were inadequate in reducing the high ammonia-nitrogen concentrations. Since the addition of zeolite was nonetheless effective in increasing methane production, it was hypothesized that a combination of microbial immobilization and stimulation by unknown exchanged cations released from the zeolite were likely factors that contributed to the beneficial effects.
The effect of different amounts of zeolites coupled with acclimatized inoculum was studied by Kougias et al. (2013) in mesophilic batch reactors. Using well digested swine manure as inoculum, poultry manure was anaerobically digested with 5 and 10 g/L of natural zeolites. The results showed that a significant increase of methane production was observed when zeolite was added to the system. The best results were obtained by the addition of 10 g/L zeolite, with a 109.75% increase of methane production compared to the reference reactor without zeolite. The VFA concentrations were also noticeably lower in reactors loaded with zeolite, indicating a stable digestion process.
The effect of zeolite addition on the mesophilic AD of swine manure with 10% TS was investigated (Lin et al. 2013). Based on the assumption that the ammonium adsorption capacity of zeolite was 13.3 mg N/g zeolite, 60 g/L of zeolite was added to the reactor. The results showed that the zeolite addition led to a faster start-up and a better performance in term of biogas production and methane yield, especially at high ammonium concentrations. The total ammonia nitrogen (TAN) was rapidly absorbed by zeolite in the first 5 days. However, the calculated zeolite adsorption capacity was lower than the maximal capacity. This might be due to two reasons: the competition with Na+ and K+ ions co-existing in the digester and the ammonium mass transfer might be hindered and retarded by the high TS concentration and the sedimentation of zeolite particles covered by solid substrates. The zeolite addition did not have any significant influence on the microbial biodiversity.
In an another study, Huang et al. (2014) proposed a process to simultaneously remove ammonium nitrogen and phosphate from simulated swine wastewater by modified zeolite and struvite crystallization. Recently, Wang et al. (2016) used struvite precipitation (SP) to combat ammonia toxicity in a two-stage anaerobic digester fed with protein-rich feedstocks. The authors observed that the SP treatment was successful in removal of TAN along with better methane yields (Wang et al. 2016).
To minimize the operational costs, an effective way to regenerate exhausted zeolite, especially one that can adapt to condition of higher ammonia loading from manure wastewater, is of great significance. Zeolite regeneration is generally realized using brine solutions with different brine compositions or by a biological process, where ammonium adsorbed on the zeolite is transferred to gaseous nitrogen species by nitrification-denitrification (Guo et al. 2013). However, if organic matter adsorbs proportionately, the effluent may contain high concentrations of salt and COD.
A study was conducted by Guo et al. (2013) using different combinations of alkaline (NaOH) and salt (NaCl) concentrations from batch desorption tests to expanded fixed-bed column. The regeneration efficiency was dependent on alkaline concentrations (pH), salt strength and flow rate. The desorption rate increased significantly from 70% to 90–95% with increasing alkaline concentration between 0.032 and 0.1 mol/L. At this lower alkaline concentration, the desorption rate also increased with increasing salt concentration (from 10 to 58.5 g/L). No influence on ammonium desorption was observed, when the salt concentration was further increased for more than 20 g/L at alkaline concentration greater than 0.1 mol/L. The elution curves showed best results at flow rates in the range of 2.5–3.0 bed volumes per hour. For most cases in this study, complete ammonia elution was achieved in 10 bed volumes. Later, the ammonia retained in the regenerant can be recovered by air or steam stripping and, consequently, the regenerant can be recycled. This work also provides a comprehensive table presenting a comparison of regeneration parameters and corresponding performance between different literatures.
Deng et al. (2014) described an effective and economic method to regenerate exhausted zeolite. The study included the assessment of ammonia exchange capacity, zeolite regeneration efficiency and economic analysis. The ammonia exchange capacity of natural zeolite increased initially with increasing ammonium ion and then reached a plateau. This exchange, however, was also influenced by pH. At high pH-values (9.5 and 10.5), a significant decrease in ammonia exchange capacity was observed, probably due to the transformation of ammonium ion to unionized ammonia that was then inaccessible for the ion exchange. The regeneration efficiency was strongly affected by the mass ratio of Na+ ions and zeolite-NH4+-N. The alkaline regeneration at pH 12 decreased the mass ratio to 4.2, and only 10 g/L NaCl was needed for the recovery (85% in 2 h in continuous columns tests). The economic analysis showed that this alkaline regeneration saved chemical costs over 10 times as compared with a conventional regeneration method at pH 9.
A combination of ammonia and phosphate removal was investigated by Huang et al. (2014). Using modified zeolite, which was natural zeolite modified with magnesium salts, as adsorbent for ammonium ions, the magnesium ions released from zeolite served as the source of SP for the removal of phosphate. Both TAN and phosphate were effectively removed. The removal efficiencies were maximum at pH-values 8.5–9, where 82% total ammonium and 98% phosphate were removed with 110 g/L of modified zeolite and 40 minutes reaction time from simulated swine wastewater. However, the presence of various cations (K2+, Ca2+, Na2+ and Mg2+) had a significantly negative effect on the removal of ammonia-nitrogen. This method has yet to be tested using real animal wastes.
In a relatively recent work, Zheng et al. (2015) developed a bio-zeolite fixed-bed reactor using different materials to mitigate ammonia inhibition during AD of livestock wastes with high ammonium contents. The novel fixed-bed reactor designed using chlorinated polyethylene (CPE) material was eventually recommended by the authors to improve AD of livestock wastes of high ammonium content, since CPE fixed zeolite favored the adsorption of ammonia and immobilization of microorganisms.
Zeolite – advantages:
Relative easy to operate
Immobilization surface, enhance growth
Adsorption of ammonia (not in every study)
Somehow counteract ammonia inhibition, without reducing ammonia concentration
Ability to recover inhibited process
Enhance methane production (not in every study).
Zeolite – disadvantages:
Adsorption capacity in some studies was quite low, i.e. application for high strength NH4-N possible?
In some publications, no reduction of ammonia was observed
Regeneration (high dose chemicals), i.e. high costs
Presence of other cations can decrease ammonium exchange capacity.
Ultrasonication has been widely used to degrade organic compounds from wastewater (Wang et al. 2008). It evokes cavitation in the organic material by rapid bubble formation in the liquid phase. When the bubbles collapsed, instant high temperature and pressure were produced, leading to the formation of radicals and degradation of the particulate matters (Luste & Luostarinen 2011). In the removal of ammonia nitrogen, most of the free ammonia molecules enter into cavitation bubbles in which they are transformed into nitrogen and hydrogen molecules via pyrolysis under instant high temperature and high pressure in the cavitation bubbles (Wang et al. 2008). Luste & Luostarinen (2011) studied the effect of ultrasound and thermal pretreatments on dairy cattle slurry. Both the CH4-yield and concentration of ultrasound tested (6,000 kJ/kg TS) cattle slurry was about 20% higher compared to the untreated slurry. The ammonium content increased afterwards due to the increased amount of soluble organic materials. A decrease in ammonia concentration was reported by Cho et al. (2014). Applying ultrasonication technique at two different frequencies (28 and 40 kHz) to simulated and real livestock wastes, the ammonia removal efficiency was tested at diverse pH-values (10–12), duration (5–60 min) and temperature (30–72 °C) ranges. The elevated pH was necessary to shift the equilibrium to NH3, since only NH3 could be stripped. Interestingly, the temperature increase of the samples treated in the same duration of time was similar at both ultrasonication intensities. A high removal rate was obtained at lower frequency of 28 kHz and pH 11 for 15 min. The higher ammonia removal efficiency at lower frequency was assumed due to the non-thermal effect, which contributed to different convective phenomena (micro-streaming, micro-convection, shock waves, micro-jets). The non-thermal effects were the dominant mechanism in the ultrasound assisted ammonia stripping, accounting for 64% of the total ammonia removal rate. The experiments with the livestock waste showed an ammonia removal rate of up to 55% and an increase in the methane yield by 58%. An additional aeration did not bring any positive increase of results.
Ultrasonication – advantages:
Enhance solubilization, thereby increasing hydrolysis rate and simultaneously reducing ammonia.
Ultrasonication – disadvantages:
No/low reduction of ammonia concentration
Rarely applied in actual plant
Released other toxic compounds when too intense
Destruction of microorganisms.
Similar to the ultrasonication technique, the ammonia removal by microwave radiation is also based on the shift of the ammonia-nitrogen equilibrium to molecular ammonia (NH3) and its subsequent evaporation by microwave radiation. In order to treat coke-plant wastewater containing high level of ammonia (5,000 mg/L at pH 11), Lin et al. (2009a) developed an effective method to remove the ammonia by microwave radiation. To find an optimal operating condition, the experiments were conducted under different pH levels, radiation time, with and without aeration, and initial ammonia concentration. The aeration did not show any significant influence on the ammonia removal. However, a fundamental difference between conventional heating and microwave radiation could be observed according to the NH3 transfer into the gas phase. Throughout, the elimination efficiency was increased when microwave radiation was applied.
MW radiation seems to be an effective strategy for the removal of high ammonia concentrations from wastewater in a short time. 10 min MW radiation (750 W) led to a decreasing ammonia concentration from 5,000 mg/L to 350 mg/L at a pH-value of 11. However, both biological and economic disadvantages may be considered, as microorganisms can be destroyed and costs for high power consumption can rise.
Hollow fiber membrane
Using hollow membrane contactor, Lauterböck et al. (2012) tried to counteract ammonia inhibition in slaughterhouse wastes with ammonium concentration ranging from 6 to 7.4 g/L. A module consisted of seven hollow fibers, with the microporous hydrophobic membrane material made from polypropylene without outer casing, were directly submerged into the fermentation broth. Simultaneously, sulfuric acids were circulated through the lumen of the fibers and the nitrogen was recovered in the form of an ammonium sulfate solution. This technology allows a gaseous transfer between two liquid phases (NH3 rich feed and the acidic absorption solution), with the gas filled pores of the membrane as the actual transfer area. The driving force for the mass transfer is the difference in NH3 partial pressure between the two liquid phases (Lauterböck et al. 2014). The ammonia was continuously extracted by the membrane contractor. The results showed that was significantly reduced by the hollow fibers contactor. Moreover, the pH was decreased as was removed and thus lowering the free ammonia concentration by 70%. It was also observed that the VFA concentrations were lower in the membrane reactor and the biogas yield was higher.
In addition to these methods, Kim et al. (2015) investigated microbial fuel cells (MFCs) in order to remove TAN and residual COD from the effluent of an anaerobic reactor fed with swine wastewater as substrate in batch mode. The authors observed that MFCs could be useful to remove TAN. Cerrillo et al. (2016) employed a combination of an AD system with a microbial electrolysis cell (MEC) coupled with an ammonia stripping unit to improve the quality of the effluent. It was concluded by the authors that this combined AD-MEC system could be used to prevent nitrogen overloads and to achieve a high quality of effluents along with the possibility of nutrient recovery. Zhang & Angelidaki (2015) employed a submersible microbial desalination cell in a CSTR to reduce ammonia levels. As a result of batch experiments for 30 days, the authors found out that the concentration of ammonia in the CSTR decreased from 6 to 0.7 g N/L.
SP has been widely applied to remove ammonium-nitrogen from wastewaters (Darwish et al. 2016). The ammonium removal by adding magnesium salt and phosphate to form equal molar concentrations of MgNH4PO4•6H2O (MAP) under alkaline conditions is a useful method to produce struvite. Magnesium, ammonium and phosphate are required in equimolar quantities (1:1:1) to form MAP. For optimum recovery of ammonia; however, slight excess of M and P are required (Çelen & Türker 2001). MAP has a comparable composition of Mg, P and N as commercial fertilizers in soil. Those MAP fertilizers are most commonly applied in granular form in the field (Degryse et al. 2016). However, wastewaters also contain large amounts of phosphorus, and there are two reasons to remove P from wastewater. Firstly, the removal of P is necessary to avoid and prevent eutrophication in the environment. Secondly, phosphate is a limited source, and thus recycling P is vital for the sustainable production of food in future (Cordell et al. 2011).
Different processes are used for the precipitation of MAP. According to reagents used, most commonly MgCl2, is applied while the pH is adjusted with NaOH. These Mg salts react quickly with phosphorus during crystallization process so that high quality struvite is formed (Britton et al. 2005). Alternatively, MgO or Mg(OH)2 can be used, which is cheaper than Mg salts and, furthermore, the need for NaOH is less during the process. However, the reaction time is slower than those of Mg salts resulting in a product containing excess MgO of Mg(OH)2 (Capdevielle et al. 2013). The solubility of MAP crystals in water decreases with increasing pH below 9 (Bhuiyan et al. 2007).
To prevent process failures due to ammonia toxicity, different conventional strategies such as stripping, chemical precipitation, adjustment of C:N-ratios, immobilization and adaption of microorganisms, bioaugmentation, dilution of substrates and/or co-digestion of nitrogen rich wastes have been suggested.
Many studies also reported about using zeolite. Benefits such as the ability to recover process failure and easy operation were discussed, but the possibility of low adsorption in some cases and high costs for regeneration were the pronounced drawbacks. Among the novel methods, ultrasonication and microwave techniques were reported to provide ammonia reduction. Also MFC/microbial desalination cell seemed to provide high TAN-removals from AD systems along with good effluent qualities.
N. K., D. D. and R. A. are thankful for grants provided by the German Federal Ministry of Food and Agriculture (BMEL, grant no. 22026411) with Fachagentur Nachwachsende Rohstoffe e. V. (FNR) as promotor.