The Effect of Ammonia Toxicity on Methane Production of a Full-Scale Biogas Plant—An Estimation Method

: Ammonia accumulation in biogas plants reactors is becoming more frequently encountered, resulting in reduced methane (CH 4 ) production. Ammonia toxicity occurs when N-rich substrates represent a signiﬁcant part of the biogas plant’s feedstock. The aim of this study was to develop an estimation method for the effect of ammonia toxicity on the CH 4 production of biogas plants. Two periods where a biogas plant operated at 3200 mg · L − 1 (1st period) and 4400 mg · L − 1 (2nd period) of ammonium nitrogen (NH 4+ –N) were examined. Biomethane potentials (BMPs) of the individual substrates collected during these periods and of the mixture of substrates with the weight ratio used by the biogas plant under different ammonia levels (2000–5200 mg · L − 1 NH 4+ –N) were determined. CH 4 production calculated from the substrates’ BMPs and the quantities used of each substrate by the biogas plant was compared with actual CH 4 production on-site. Biogas plant’s CH 4 production was 9.9% lower in the 1st and 20.3% in the 2nd period in comparison with the BMP calculated CH 4 production, of which 3% and 14% was due to ammonia toxicity, respectively. BMPs of the mixtures showed that the actual CH 4 reduction rate of the biogas plant could be approximately estimated by the ammonia concentrations levels.


Introduction
Organic waste treatment technologies are becoming more and more inextricably linked to renewable energy production and they are an integral part of pollution reduction strategies and policies. The biogas sector contributes a significant part of the renewable energy production in EU, an industry under rapid expansion as more than 18,000 plants were operating in 2019 in contrast with approximately 6000 in 2009 [1]. Promotion of environmental awareness along with favourable legal framework and the high efficiency of the anaerobic digestion technology were the main reasons for the development of the biogas industry [2]. Biogas is a gaseous mixture consisted mainly by CH 4 and carbon dioxide (CO 2 ) and it derives from the degradation of organic raw materials by microorganisms under anaerobic conditions [3]. Biogas yield is directly depended on the energy content of the feedstock that can be transformed into CH 4 .
The availability of raw materials is directly related to the viability and profitability of biogas plants. Managers must establish strong financial interconnections with raw materials suppliers that are preferably located near to their plants to secure long term developed for the estimation of methane production linked to substrate ammonia content, which will enable a stable operation of the biogas plants, thus preventing economic losses and protecting the environment from residual emissions caused by inadequate anaerobic digestion process.
The aim of the present study is to develop a laboratory approach to evaluate the ammonia toxicity effect on CH 4 production in a full-scale biogas plant and help managers to achieve a cost-effective anaerobic reactor operation. A detailed valuation of the operating conditions of the biogas plant's anaerobic treatment process prevailing during two periods with NH 4 + -N concentrations of 3200 mg·L −1 (1st period) and 4400 mg·L −1 (2nd period), was performed. A comparison of the calculated total methane potential with the actual methane production of the biogas plant during these two periods and for a Hydraulic Retention Time (HRT) of 30 days was done. An estimation of the reduction rate of CH 4 production took place in the case of a sharp increase of the ammonia concentration in the content of the biogas reactor.

Substrates and Inoculum
The biogas plant is in the Region of Central Macedonia in Greece with an electric power capacity of 1MW, owned by Biogas Lagada S.A. (Lagadas, Central Macedonia, Greece). It is operated in the mesophilic temperature range and an HRT of about 30 days. The biogas plant consists of two anaerobic digestion reactors D1 and D2 that operate in series and each of them with a working volume of 3800 m 3 . The examination of the ammonia toxicity effect on CH 4 production was performed for two different periods. The ammonia concentration in the 1st period between June and July of 2018 was 3200 mg·L −1 NH 4 + -N and in the 2nd period in September of 2018 was 4400 mg·L −1 NH 4 + -N. The daily input of substrates in the biogas plant was 247 tons and their characteristics measured after laboratory analyses of samples obtained by each of the waste streams utilized by the biogas plant are presented in Table 1. The input substrates ratio for the two periods under examination are shown in Figure 1. The substrates used by the biogas plant in these two periods were also used in this study for the determination of CH 4 potential.  The inoculum was collected from the effluent of the D1 anaerobic reactor of the same biogas plant. At the day of collection, it was stored at 37 °C for 7 days to reduce the residual CH4 production. The same inoculum was used for all the Biomethane Potential (BMP) tests performed in this study. Inoculum characteristics are presented in Table 1 and some additional characteristics are reported in Table S8 (Supplementary Material).

Experimental Design
In order to calculate the actual impact of ammonia toxicity on the biogas plant, a BMP test of each different substrate used during these two periods was performed. The results of the substrates' BMP tests of each period were compared with the corresponding biogas plant's CH4 production. BMP is a laboratory scale test carried out on batch reactors where a portion of a homogenized sample is digested by a suitable inoculum [19]. It should be noted that BMP tests are used to determine the maximum CH4 content that could be recovered by the organic fraction of a substrate. Substrate's methane yield achieved at a fullscale reactor is expected to be lower than the CH4 potential determined by BMP tests as in the continuous operation of a full-scale reactors different parameters have an effect on methane production [19]. However, Mönch-Tegeder et al. reported negligible differences between the BMP tests of grinded substrates and the full-scale actual methane production of a digester equipped with a pre-treatment device [20]. Therefore, in this study the difference between BMP tests and actual CH4 production was further investigated. The total CH4 production of each period was calculated by multiplying the CH4 production per quantity used in the substrates BMP tests with the actual quantity that was introduced into the biogas plant reactor. When the actual CH4 production from the biogas plant is substantially lower than the calculated CH4 production by the substrates BMP tests, it is a first indication of an inhibition of the anaerobic digestion process. Due to the biogas plant's ammonia levels being relatively high during these periods, this reduction in CH4 production was possibly by ammonia toxicity. To clarify which percentage of the reduction was due to the ammonia toxicity, effluent samples from the biogas plant were obtained during the two periods of examination and their VFAs concentrations were measured. Moreover, an effluent sample from the 1st period and a 1:1 v/v diluted sample of the same effluent with water were used in triplicates and anaerobic conditions were created on site to measure their residual CH4 potential. The CH4 production measurements from the effluent samples were started on the same day they were collected and prepared.
In parallel, co-digestion BMP tests were performed by mixing the substrates of the 1st period to the exact weight ratio that was used in the biogas plant ( Figure 1) to accurately calculate the effect of ammonia toxicity on CH4 production under different ammonia concentrations. The mixing of the substrates resulted in 2060 mg·L −1 (2N) of ammonium nitrogen (NH4 + -N). Co-digestion BMP tests were performed at higher concentra- The inoculum was collected from the effluent of the D1 anaerobic reactor of the same biogas plant. At the day of collection, it was stored at 37 • C for 7 days to reduce the residual CH 4 production. The same inoculum was used for all the Biomethane Potential (BMP) tests performed in this study. Inoculum characteristics are presented in Table 1 and some additional characteristics are reported in Table S8 (Supplementary Material).

Experimental Design
In order to calculate the actual impact of ammonia toxicity on the biogas plant, a BMP test of each different substrate used during these two periods was performed. The results of the substrates' BMP tests of each period were compared with the corresponding biogas plant's CH 4 production. BMP is a laboratory scale test carried out on batch reactors where a portion of a homogenized sample is digested by a suitable inoculum [19]. It should be noted that BMP tests are used to determine the maximum CH 4 content that could be recovered by the organic fraction of a substrate. Substrate's methane yield achieved at a full-scale reactor is expected to be lower than the CH 4 potential determined by BMP tests as in the continuous operation of a full-scale reactors different parameters have an effect on methane production [19]. However, Mönch-Tegeder et al. reported negligible differences between the BMP tests of grinded substrates and the full-scale actual methane production of a digester equipped with a pre-treatment device [20]. Therefore, in this study the difference between BMP tests and actual CH 4 production was further investigated. The total CH 4 production of each period was calculated by multiplying the CH 4 production per quantity used in the substrates BMP tests with the actual quantity that was introduced into the biogas plant reactor. When the actual CH 4 production from the biogas plant is substantially lower than the calculated CH 4 production by the substrates BMP tests, it is a first indication of an inhibition of the anaerobic digestion process. Due to the biogas plant's ammonia levels being relatively high during these periods, this reduction in CH 4 production was possibly by ammonia toxicity. To clarify which percentage of the reduction was due to the ammonia toxicity, effluent samples from the biogas plant were obtained during the two periods of examination and their VFAs concentrations were measured. Moreover, an effluent sample from the 1st period and a 1:1 v/v diluted sample of the same effluent with water were used in triplicates and anaerobic conditions were created on site to measure their residual CH 4 potential. The CH 4 production measurements from the effluent samples were started on the same day they were collected and prepared.
In parallel, co-digestion BMP tests were performed by mixing the substrates of the 1st period to the exact weight ratio that was used in the biogas plant ( Figure 1) to accurately calculate the effect of ammonia toxicity on CH 4 production under different ammonia concentrations. The mixing of the substrates resulted in 2060 mg·L −1 (2N) of ammonium nitrogen (NH 4 + -N). Co-digestion BMP tests were performed at higher concentrations of ammonia by addition of an ammonium chloride solution to predict the CH 4 reduction Energies 2021, 14, 5031 5 of 13 rate of the biogas plant in the case of a sharp increase of reactor's ammonia concentration. The nominal concentrations used were 3000 mg·L −1 (3N), 4000 mg·L −1 (4N) and 5000 mg·L −1 (5N) of NH 4 + -N. These ammonia levels were previously observed at the biogas plant depending on the seasonality and availability of substrates.

BMP Tests
The CH 4 production from each substrate used during the two examination periods and from the co-digestion of substrates were determined by BMP tests. All substrates and created substrates mixtures were homogenized. The inoculum to single substrate and inoculum to substrates ratio were 1:1 in terms of Volatile Solids (VS). For the 1st period substrates, co-digestion experiment and the effluent residual CH 4 potential tests, batch glass reactors equipped with gas tight rubber stoppers were used and had a total and working volume of 297.66 mL and 100.00 mL, respectively. All the BMP tests were performed at the temperature of 37 ± 0.5 • C. Anaerobic conditions were created by flashing a mixture gas of nitrogen (80% N 2 ) and 20% CO 2 into reactors for 2 min. Single substrate and co-digestion of substrates BMP tests were done in triplicate, and they were vigorously mixed by hand once per day. Three reactors filled only with inoculum were used to monitor residual CH 4 production from the inoculum. For substrates of the 2nd period the batch reactors that were used had a total and working volume of 2300.00 mL and 2000.00 mL, respectively and stirring was performed by the Biogas Endeavour device.

Analytical Methods
The measurement of Total Solids (TS), Volatile Solids (VS), Total Kjeldahl Nitrogen (TKN) and Total Ammonium Nitrogen (TAN) was performed according to APHA's Standard Methods [21]. The pH value was measured with a digital pH meter, Jenway 3520 (Cole and Palmer, Staffordshire, UK).
Biogas quantity and composition from the mono-digestion and co-digestion of substrates used in the 1st period, as well as from the 1st period effluent residual CH 4 potential tests were determined daily by measuring the headspace pressure of the reactors with a gas chromatography analysis. The gas chromatograph-GC2010 plus AT (Shimadzu, Japan) for biogas analysis was equipped with a Thermal Conductivity Detector (TCD) and two connected columns Valco-PLOT Hayesep D and VP-Molesieve Column FS of the same length (15 m) and internal diameter (0.53 mm) [22]. The injection volume was 0.15 mL, and it was done with a gas-tight syringe. CH 4 potential from the substrates of the 2nd period was measured via the device Gas Endeavour (Bioprocess Control, Lund, Sweden) [23]. VFAs analysis was performed in a gas chromatograph GC−2010plusAT (Shimadzu, Japan) equipped with a Flame Ionization Detector (FID) and a high polarity column ZB-FFAP (Phenomenex, Torrance, Los Angeles, CA, USA) with 30 m length and 0.53 mm internal diameter. Samples were injected to the GC-FID with the help of an autosampler AOC−20s (Shimadzu, Japan) and the volume injected was 1 µL. Helium was the carrier gas for both gas chromatographers.
Statistical analysis of data was performed with the IBM SPSS Statistics, version 25. One-way analysis of variance was used for the comparison of the means for the co-digestion BMP experiments, and the pair-wise differences were assessed with the post hoc analysis (statistically significant difference if p < 0.05).

BMPs of Substrates and Comparison with the Actual Methane Production of the Biogas Plant-Preliminary Assessment of Ammonia Toxicity
The waste quantities introduced in the biogas reactor for each period, their CH 4 potential and the calculated CH 4 production based on the added quantity and the BMP potential of each substate are presented in Table 2. Water was added to reach the working volume of the reactors used for the mono-digestion BMP tests of substrates and the final ammonia concentration did not exceed 1150 m L −1 of NH 4 + -N for each of the reactors. The total CH 4 production (m 3 ) for each examination period (30 days) was calculated by adding the CH 4 production (m 3 ) of each substrate, which was calculated by multiplying the input quantity (t) with the CH 4 potential of each substrate (m 3 t −1 ), similar to a calculation performed by Hollinger et al. [24]. This total methane production indicates the maximum production that could occur by exploitation of the degradable matter of each individual substrate of the feedstock under anaerobic digestion in a period of 30 days. The highest CH 4 potential between wastes of both periods was obtained by glycerine (428.33 m 3 CH 4 t −1 ) and dough waste (268.58 m 3 CH 4 t −1 ). However, both wastes formed the smallest fractions of the total feedstock of the 2nd period, thus affecting less the total CH 4 production for this period. On the other hand, poultry manure with 79.38 m 3 CH 4 t −1 was accounted for most of the total CH 4 production in the 2nd period (36%), due to the large quantity (17%) that was introduced into the reactors. For the same reason, cattle manure with 15.68 m 3 CH 4 t −1 was responsible for the 44% of the total feedstock in the 1st period. Regarding the CH 4 yield of the substrates, similar results were reported by previous studies about the BMP of cattle manure [25], pig manure [26], poultry manure [27], glycerine [28], corn silage [29], cheese whey [30] and grape pomace [31].
Concentrations of NH 4 + -N, the CH 4 production of the biogas plant, and the total CH 4 production for each period are given in Table 3. The NH 4 + -N concentration was increased by 1200 mg·L −1 in the 2nd period in comparison with the 1st. This increase was due to the different substrates that were used in the 2nd period. Specifically, poultry manure, dry poultry manure and dried digestate which are characterized with a high nitrogen content (Table 1) were accounted for 24.14% of the total waste input in the 2nd period, where in the 1st period, poultry manure was the only high content nitrogen source which was representing a 12.74% of the total waste input. The biogas plant's CH 4 production for both periods was calculated based on the characteristics of the Combined Heat and Power (CHP) engine provided by the biogas plant. The actual CH4 production was the same during the two periods of examination as the roof of the biogas plant reactors was equipped with a single-layered gas storage system suitable to store the excess produced biogas and supply it to the CHP engine at a period of lower production. In combination with the steady stream of wastes to the reactor the CHP engine was running at a maximum electric capacity and the emergency biogas flare was never put into operation during these two periods. The CHP efficiency in electric power was 40.07%, the low heating value for CH 4 (65%) was 6.5 KWh, and the daily output was 23.35 MWh. As expected, the calculated CH 4 production from the substrates' mono-digestion BMP tests was greater than the biogas plant's in both periods. However, the biogas plant's CH 4 production in the 2nd period was 20.3% lower from the calculated total CH 4 production, in contrast with only a 9.9% reduction in the 1st period. In combination with the higher concentration of ammonia nitrogen and VFAs concentration as shown below during the 2nd period, it is strongly indicated that ammonia inhibited the methanogenic process in this period. The Volatile Fatty Acids (VFAs) concentrations act as a good indicator of process imbalance in anaerobic digestion [32,33]. The VFAs concentrations of the biogas plant effluent samples from the two periods are presented in Table 4. In the 1st study period, all the VFAs concentrations where lower than those of the 2nd period, thus revealing a toxicity effect but as not as strong as it was in the 2nd period. The acetic and propionic acid concentrations during the 2nd period of examination were very high, revealing the presence of toxicity in the anaerobic reactor of the biogas plant. It was reported that the increased concentration of acetic acid is associated to the absence or restricted growth of aceticlastic methanogens, which was probably due to high concentrations of ammonia [10]. Furthermore, Wang et al. reported that when the concentration of propionic acid exceeded the threshold of 900 mg·L −1 , the bacteria concentration was decreased with detrimental effects on their activity [34]. Inhibition of methanogens growth was observed in an anaerobic reactor at a similar propionic acid concentration of 950 mg·L −1 [35]. From the stoichiometry of the CH 4 biological production (Equation (1)) and based on the assumption that 1 mole of gas under standards conditions has 22.4 NL of volume, it turns out that from 1 mmol of acetic acid and 1 mmol of propionic acid could be produced 22.4 N mL and 39.2 N mL of CH 4 , respectively [36]. If the amounts of these two acids in the 2nd study period (Table 4) had been converted to CH 4 , an additional volume of 1817.5 mL CH 4 kg −1 or 1.8175 m 3 CH 4 t −1 of input substrates would be produced. The above calculations correspond to a net loss of 13,467 m 3 (30 days HRT * 247 t day −1 * 1.81745 m 3 CH 4 t −1 ) of CH 4 for the 2nd study period without considering the reduction in methane production due to the instability of the anaerobic process.
The reduction by 9.9% of the total calculated CH 4 production in the 1st period was probably due also to other factors, as the VFAs concentrations were not high enough as to be considered responsible for this reduction. To determine the reduction rate caused by toxicity of ammonia or to other factors, effluent from the biogas reactor was collected during the 1st period for further examination. A dilution with water of the biogas effluent was performed to minimize any toxicity effect and the CH 4 potential of the diluted and undiluted effluent sample was determined. The cumulative methane production in mL CH 4 g −1 of effluent for both samples is presented in Figure 2 and the VFAs concentrations at the end of BMP tests are depicted in Table S1 (Supplementary Material). The diluted sample had a 28.88% increase in CH 4 production in comparison with the undiluted sample. Furthermore, 60% of the total CH 4 production was produced in 12 days for the diluted effluent sample, where the undiluted reach the same level of production in 20 days. This difference in CH 4 production is caused by toxicity factors and probably by the inhibitory action of ammonia.
Energies 2021, 14, 5031 8 of 13 mL CH4 kg −1 or 1.8175 m 3 CH4 t −1 of input substrates would be produced. The above calculations correspond to a net loss of 13,467 m 3 (30 days HRT * 247 t day −1 * 1.81745 m 3 CH4 t −1 ) of CH4 for the 2nd study period without considering the reduction in methane production due to the instability of the anaerobic process.
The reduction by 9.9% of the total calculated CH4 production in the 1st period was probably due also to other factors, as the VFAs concentrations were not high enough as to be considered responsible for this reduction. To determine the reduction rate caused by toxicity of ammonia or to other factors, effluent from the biogas reactor was collected during the 1st period for further examination. A dilution with water of the biogas effluent was performed to minimize any toxicity effect and the CH4 potential of the diluted and undiluted effluent sample was determined. The cumulative methane production in mL CH4 g −1 of effluent for both samples is presented in Figure 2 and the VFAs concentrations at the end of BMP tests are depicted in Table S1 (Supplementary Material). The diluted sample had a 28.88% increase in CH4 production in comparison with the undiluted sample. Furthermore, 60% of the total CH4 production was produced in 12 days for the diluted effluent sample, where the undiluted reach the same level of production in 20 days. This difference in CH4 production is caused by toxicity factors and probably by the inhibitory action of ammonia. According to data provided by the biogas plant, the daily waste input in the reactor was approximately 247 tons with a simultaneous equal amount of effluent discharge. Therefore, it should be noted that according to the above results, during the 1st study According to data provided by the biogas plant, the daily waste input in the reactor was approximately 247 tons with a simultaneous equal amount of effluent discharge. Therefore, it should be noted that according to the above results, during the 1st study period, the biogas plant did not exploit a daily production capacity equal to 864.5 m 3 day −1 of CH 4 , which a fraction of 192.66 m 3 day −1 was probably due to ammonia toxicity, calculated from the difference of CH 4 production from the diluted and undiluted sample. This fraction corresponds to a reduction in total methane production of 3% (100 × 192.66 m 3 day −1 × 30 day/190,958.13 m 3 ). The remaining reduction (6.9%) is attributable to other factors, such as a longer HRT was required to anaerobically degrade the substrates. The measured CH 4 production from the undiluted and diluted effluent samples in combination with the volatile solids reduction of 31.7% and 40.5%, respectively, proves that the 6.9% decrease is due to other factors independent of toxicity. Similar results were reported by other researchers that studied the scale effect on calculated CH 4 production from BMPs of different substrates compared with the CH 4 production of a full-scale biogas plant, where the overestimation was on average 6.0 ± 6.8% and concluded that the reasons were the different TS solids content and the particle size of the substrates [24].

Comparison of Methane Production by Anaerobic Co-Digestion of Substrates with Different Ammonia Concentrations
Anaerobic co-digestion experiments were performed to find the effect of ammonia toxicity on methane production under different concentrations of ammonia that were previously reported at the biogas plant. For this reason, a mixture of substrates used at the 1st period was prepared and the mixing ratio was identical to the one used by the biogas plant. The final concentrations of N-NH 4 + after anaerobic digestion process in 2N (3157 mg·L −1 ) and 3N (4580 mg·L −1 ) were close to the ammonia concentrations during the biogas plant operation in the 1st and 2nd period, respectively. The increase of the ammonia concentration during anaerobic digestion was expected due to the biodegradation of proteins and urea hydrolysis.
The cumulative CH 4 production of the co-digestion batch reactors for 30 days (biogas plant's HRT operation), as also the total CH 4 productions after 60 days of anaerobic digestion are illustrated in Figure 3. Results from the statistical analysis of the data from the BMP co-digestion tests are presented in Tables S2-S7 (Supplementary Material). The total CH 4 production was strongly affected by the increase of the ammonia concentration in the batch reactors. The highest CH 4 production was observed by 2N, and its exponential increase was completed in a shorter time among all treatments. This was possibly because of the inoculum's microbial population acclimatization to the ammonia concentration range of 2060 to 3158 mg·L −1 which was previously maintained at the biogas plant during the 1st period (Table 1, inoculum's NH 4 + -N concentration). If the 2N CH 4 potential of 330.68 mL g −1 of VS (30 days) is multiplied by the 579.78 tons of VS of substrates that were used for a 30-day HRT in the 1st period results to 197,721.65 m 3 of CH 4 , which is almost the same (0.4% deviation) with the total CH 4 volume calculated by the mono-digestion BMP tests of substrates for the same period (Table 2). This result indicates that a co-digestion synergetic or antagonistic effect was not apparent and probably the inoculum's microbial population was efficient under the specific range of ammonia concentration. Similar results were reported when co-digestion of different organic substrates did improve the anaerobic digestion kinetics, but the ultimate methane potential was not affected [37]. An analytical presentation of the estimated rates of CH 4 potential losses when the biogas plant operates under different ammonia concentrations for an HRT of 30 days is shown in Table 5.  The calculated decrease in CH4 production derived from the substates mono-digestion BMP tests for the 2nd period is in line with the results of the 3N co-digestion treatment. The ammonia concentration in 3N treatment was at the same level with the biogas plant's reactor during the 2nd period of examination (Tables 3 and 5). The results from the comparison of the biogas plant's CH4 production with the total calculated CH4 production (substrates mono-digestion) of 2nd period exposed a reduction rate of 20.3% ( Table 3). Moreover, in the 1st period from the total 9.9% reduction rate, the 6.9% was allocated to other factors as the 3% was proved by the dilution of the effluent to be due to toxicity factors. The BMP co-digestion tests revealed a 14% CH4 reduction rate of 3N in comparison with 2N treatment which was exclusively caused by ammonia toxicity as all the other experimental conditions were the same for these two treatments (Table 5). Based on the assumption that the 6.9% CH4 reduction rate would be also present in the 2nd period as the feedstock's residual potential was not exploited due to short HRT or other operational factors of the biogas plant, the total CH4 reduction rate would be equivalent to around 20% as it was calculated before with the mono-digestion BMP tests. From these results, it is indicated that the co-digestion BMP tests are a sufficient method to predict the biogas plant's CH4 performance only by acknowledging the ammonia concentration.  The calculated decrease in CH 4 production derived from the substates mono-digestion BMP tests for the 2nd period is in line with the results of the 3N co-digestion treatment. The ammonia concentration in 3N treatment was at the same level with the biogas plant's reactor during the 2nd period of examination (Tables 3 and 5). The results from the comparison of the biogas plant's CH 4 production with the total calculated CH 4 production (substrates mono-digestion) of 2nd period exposed a reduction rate of 20.3% (Table 3). Moreover, in the 1st period from the total 9.9% reduction rate, the 6.9% was allocated to other factors as the 3% was proved by the dilution of the effluent to be due to toxicity factors. The BMP co-digestion tests revealed a 14% CH 4 reduction rate of 3N in comparison with 2N treatment which was exclusively caused by ammonia toxicity as all the other experimental conditions were the same for these two treatments (Table 5). Based on the assumption that the 6.9% CH 4 reduction rate would be also present in the 2nd period as the feedstock's residual potential was not exploited due to short HRT or other operational factors of the biogas plant, the total CH 4 reduction rate would be equivalent to around 20% as it was calculated before with the mono-digestion BMP tests. From these results, it is indicated that the co-digestion BMP tests are a sufficient method to predict the biogas plant's CH 4 performance only by acknowledging the ammonia concentration. The accumulative CH 4 production of 4N seems to increase after the thirtieth day and probably the microbial population was efficient in higher levels of ammonia. However, the total CH 4 production in the 4N treatment was statistically significantly lower by almost 20% compared to the 2N treatment after 60 days. The treatment with the highest ammonia concentration was the 5N and compared with the 2N, it had a less CH 4 production rate of 35% and greater than 25% in 30 days and 60 days of anaerobic digestion, respectively. Therefore, it is indicated that the microbial population of the biogas plant (represented by the inoculum in this study) could not be efficient after a sharp increase of ammonia, which could be materialized by the addition of a substrate with high nitrogen concentration into reactor, such as poultry manure. Previous researchers had reported that CH 4 production by anaerobic digestion would stop at a concentration of 4000 mg·L −1 of NH 4 + when non-acclimated microbial consortiums to high NH 4 + concentrations were used [38]. This was verified in this study by the direct exposure of the microbial population to high levels of ammonia, where an acclimatization was not observed even after 60 days of anaerobic digestion.
One significant difference between the substrates of the 1st and 2nd period was the utilization of wastes with high nitrogen content such as dry poultry manure and dried digestate. These two wastes represented almost 7% of the total quantity of substrates introduced in the biogas reactor in a period of 30 days. If the above percent were replaced by different substrates or with an increase of the substrates already in use with a lower nitrogen content, as it was done during the 1st study period, the specific reduction of 20.3% in the calculated CH 4 production potential would probably not be observed. Moreover, the additional transport costs for the biogas plant would have been probably avoided, as well as the residual CH 4 potential would not have been released into the environment, thus avoiding the cause of pollution.

Conclusions
The results of this study showed that is possible to estimate the CH 4 reduction of a biogas plant caused by high ammonia concentrations depending on the nitrogen composition of the raw influent wastes. The steps taken in this study can potentially be applied by biogas plants that need to predict the drop in methane production when N-rich substrates are used. The BMP co-digestion tests that were conducted within the range of the biogas plant's ammonia concentration almost predicted the reduction of the actual CH 4 production during the 2nd period. During the 1st period (3200 mg·L −1 NH 4 + -N) of examination only 3% reduction of the calculated CH 4 potential could correspond to ammonia toxicity, where 6.9% reduction was caused by other factors, such as a short HRT. Ammonia inhibition was strong during the 2nd period (4400 mg·L −1 NH 4 + -N) resulting in almost 14% reduction of the calculated CH 4 potential. It was estimated by this study approach that operation of the specific biogas plant under ammonia concentrations within the range of 5300 to 6437 mg·L −1 of NH 4 + -N will result in 35% reduction of CH 4 production, causing significant financial losses. The microbial population of the biogas plant during the period of examination could not efficiently withstand a sudden increase of ammonia concentration by utilization of high quantities of substrates with significant nitrogen concentrations. Therefore, the ammonia content of the anaerobic reactors of the biogas plants as also of the substrates that will be finally utilized should be monitored constantly to prevent the inhibitory effect of ammonia toxicity on CH 4 production.