Optimization of Cattle Manure Liquid Fraction Anaerobic Digestion at Different Temperatures Modellin

2.1. Chemicals

In this study, nitrogen gas was used to provide an anaerobic environment in BMP bottles. NaHCO3 was used to prevent pH inhibition in the mixtures. The gas content produced under different operating conditions was determined according to the principle of displacement with water. 5% KOH was used to detect the methane portion of the produced gas, in other words, to occlude CO2 and H2S gas. The chemicals used are of high purity and obtained from Sigma Aldrich. In this study, no nutrients were added.

2.2. Characterization of Cattle Manure Liquid Fraction and Inoculation

The municipal sewage sludge as inoculation was used and it was taken from the anaerobic reactor in which domestic and textile wastewater treated by the National Membrane Technologies Application Center (Istanbul, Turkey). The sludge was precipitated and the lower phase was used in the study. pH, conductivity (EC), total solids (TS), and volatile suspended solids (VSS) analyzes were performed. pH, EC, TS, VSS and TS/VSS values were 5.24 ± 0.03, 1515 ± 3.54 μS/cm, 1.3 ± 0.03%, 91 ± 0.71% and 70, respectively. The addition of inoculation was preferred to shorten the commissioning of the system and as a source of essential nutrients/microorganisms. CMLF was taken from the facility where dairy cattle raising are performed by Kain Farm Agriculture and Livestock Inc. in Yalova province (Turkey) and stored at 4˚C during the study. As soon as the CMLF was provided, it could not be fed into the system. Before, it was carried out some analyses. The solid-liquid separation from the raw manure is made by separator. The CMLF provided was not subjected to any pre-treatment. This company evaluates SF by composting. Characterization analyzes of CMLF were carried out before the system was started. All analyzes were performed in 3 replications according to APHA standard methods [33] and the results have given in Table 1. Table 1. Characteristics of cattle manure liquid fraction Parameter pH EC (μS/cm) Salinity (‰) TS (%) VSS (%)

Parameter Alkalinity (mg/L) COD (mg/L) VFA (mg/L) Ammonia (mg/L) Orthophosphate (mg/L)

Value 2244.5±2.83 23400±141.42 1439.3±27.79 170.8±1.56 34±2.83

2.3. Lab-Scale System and Operation

Table 2 shows the experimental conditions created with Taguchi Method. CMLF was put into BMP bottles with a total volume of 250 mL. The inoculation was then added to BMP bottles according to different substrate/inoculation (S/I) ratio. 50 gr/L NaHCO3 solution was added to prevent pH inhibition and pH measurements were made. pH of all samples was measured between 7.1-7.2. The bottles were closed by gassing nitrogen gas for 5 minutes to remove oxygen. The prepared bottles were placed in the following BMP apparatus. Gas production was run until reduced (30 days). This set was carried out in two replicates and average results were given. In

anaerobic systems, the active inoculation concentration is recommended to be between 3-5 g VSS/L. In this study, by selecting it 3 gr VSS/L, the S/I ratio was calculated and added to the system [34]. The S/I ratio was also determined as 1.5 and 4.5. The system used in the study is given in Figure 1. While the system was started, the bottles were wrapped with aluminum foil to allow the reaction to take place in the dark. Moreover, the bottles were operated in a beaker filled with water to minimize the temperature fluctuation. Ghanimeh et al. [24] conducted on the effect of mixing on thermophilic AD. It was stated that mixing densities above 200 rpm may lead to the destruction of microbial populations; therefore slow mixing (100 rpm) provides better yields than vigorous mixing. According to this information, 90 and 120 rpm were selected as the mixing speed in this study. Table 2. Experiments conditions Sample 1 2 3 4 5 6 7 8

Figure 1. The system used in the study The measurement of gas was carried out according to the principle of displacement with water. KOH was added to the solution prepared to occlude CO2 gas produced. Thus, all of the gas produced was determined as CH4. The analyses in the digested samples were carried out in the upper phase after 1-hour precipitation. 1757

Optimization and statistical analysis by Taguchi method were performed by using Minitab 18 program. Data were transformed by Box-Cox transformation and one-way analysis of variance was performed. Taguchi method is one of the methods based on the principle of least experimentation to keep the costs to a minimum level developed by the Japanese Scientist Genichi Taguchi. It is an experience design that uses an orthogonal sequence [35]. The results obtained from the experiments have been converted to the signal/noise ratio (S/N). S and N represent the signal factor and noise factor, respectively. The signal factor refers to the actual value received from the system, and the noise factor refers to the factors that cannot participate in the design of the experiment but affect the result of the experiment. Noise sources are all variables that cause the performance characteristics to be obtained to deviate from the target value. Thus, the smaller the N value expressing the noise factors in the S/N ratio, the closer to the desired target value is. Therefore the purpose of this analysis is to maximize the S/N ratio and the “biggest best” (1) criterion was used for Nussetl [36]. 𝑆⁄ = −10 ∗ 𝑙𝑜𝑔 [1 ∑𝑛 1 ] 𝑁 𝑛 𝑓=1 𝑦𝑖2

where, yi denotes the performance characteristic value (Nusselt number), and n denotes the number of Y values. In this study, 3 parameters (all are 2 levels) were selected for BMP tests. In this study, L8 (2^7) orthogonal sequence and the largest is best approach are used. The temperature was chosen as 35-55˚C, S/I as 1.5-4.5, and mixing speed as 90-120 rpm to model by Taguchi method the results obtained in determining the methane production potential of CMLF in different operating parameters.

3.1. Biogas/Bio-methane Production

Different mixing speeds, temperatures, and S/I ratios from the important operating parameters in anaerobic digestion were studied. The results obtained with Taguchi method are given in Figure 2. 54

According to S/N ratios figure, it is seen that as mixing speed, temperature and S/I ratio increase, methane production increases. The thermophilic temperature has increased methane yield. However, it is stated that gas production decreases above 60 ˚C and stops above 70 ˚C [37]. Therefore, operating above 55±0.5 ˚C may not be suitable depending on the properties of the raw substrate. Suhartini et al. [38] argued that thermophilic conditions have more advantages with high methane production, increased buffering capacity, and decreased foaming and dewatering of increased digested products in the study in which they stabilized sugar pulp under mesophilic and thermophilic conditions. This was because there was no formation of extracellular polymeric substances (EPS) as a result of better substrate hydrolysis [38]. Labatut et al. [11] operated cattle manure and dog food by continuously mixing under mesophilic and thermophilic conditions. It was stated that the stability of the system under thermophilic conditions was highly dependent on the composition of the feedstock. It can be said that it is suitable for digestion of LF at mesophilic and thermophilic temperatures. It must be forgotten that the higher the temperature is, the higher cost is. For the cost, mesophilic temperatures may be preferred. In the study conducted by Tufaner and Avsar [39] in which biogas was produced from cattle manure at mesophilic temperatures, (effective reactor volume 6.15 L), 17 L of biogas was produced in 31 days. A laboratory-scale upstream anaerobic reactor was used in this study. Optimum biogas was produced with the addition of 20-30% different wastes in the digestion of different ratios of cattle manure and different wastes together [40]. In the literature, it has been stated that the use of inoculation at different rates has a positive effect since it allows different microorganism communities. As can be seen from Figure 2, increasing this ratio from 1.5 to 4.5 does not have the same effect as the mixing speed, but still has increased the gas formation. The reason why inoculation rates do not have a major effect may indicate that LF provides adequate properties. For a clearer result, both different rates should be studied and microorganism analyzes should be conducted. In the literature, the effects of different inoculation contents on anaerobic digestion of different waste and/or wastewater have been studied [41-44]. Inoculation characteristics are important for AD performance. The use of it has been reported to alter degradation rate, biogas composition, and reactor stability. It has observed that co-inoculation increased the methane yield more [43]. It may be studied the effect of coinoculation addition on methane yield. From operating parameters studied, it was observed that mixing has a higher effect on methane yield. Increasing the mixing speed from 90 rpm to 120 rpm increased methane yield. Mixing results in additional investment and operating costs for the AD system. However, efficient mixing will reduce the energy requirement by providing higher methane [26]. Mixing should prevent solid matters from sedimentation (provide homogeneous media), increase contact between nutrients and microorganisms, increase the activity of the microbial population, facilitate the separation of organics, reduce stratification, minimize temperature gradients and reduce particle size. It was stated that slow mixing improves fermentation and vigorous mixing delays fermentation and causes low methane production [23, 45]. A high mixing speed means more energy consumption. Therefore, high-speed mixing may not be preferred. Ghanimeh et al. [24] studied the effect of mixing on thermophilic (55˚C) AD of organic fraction of municipal solid waste and fresh manure as inoculation. An average of 2819 L and 2504 L methane per day were produced in the reactors (14 L volumes; 9 L working volumes) mixed continuous and intermittent mixing at 100 rpm. It was stated that mixing velocities above 100 rpm may lead to destruction of microbial communities, and therefore slow mixing (100 rpm) provides better efficiency than vigorous mixing. Lindmark et al. [25] reported that more biogas was produced at 25 rpm compared to 150 rpm. The vigorous mixing does not recommend as it can be harmful to microbial consortia [26] and reduces contact between the nutrient and microorganism. It was stated that mixing should be done on substrates with foaming and/or solid-liquid separation [46]. Optimum mixing is still a matter of debate. Because there is no clear information about which intermittent or continuous mixing is more convenient [47]. Intermittent mixing is suggested to be economical

for energy consumption [46]. It was stated that mixing on digesters operated at lower TS and longer hydraulic retention time (HRT) has no significant effect and that mixing on digesters operated at higher TS has a significant effect on biogas yield [47]. Karim et al. [22] studied the effect of mixing on AD of manure at mesophilic temperatures. In the study, it was stated that mixing in 5% TS content has no significant effect and increased methane content by 10-30% in digestion of substrate with 10% and 15% TS content. It is especially recommended for substrates with solid matter content higher than 5% [48, 49]. Slow mixing has been observed to increase the capacity of the digester, the stability of the system and the treatment efficiency. Mixing is an important parameter because it prevents hydrogen accumulation, but slow or intermittent mixing compared to no mixing and vigorous mixing has been suggested by various researchers [22-24]. The effect of intermittent mixing can be studied in the future. It can be said that the dead volume is reduced, sufficient contact between the microorganism and substrate is ensured, solid-liquid separation is minimized, the temperature distribution is ensured and gas outflow is facilitated in 90-120 rpm mixing speeds. Furthermore, the accumulation of VFA with continuous mixing may have been minimal. For this reason, 90-120 rpm mixing velocities have been a positive effect on methane yield. Regression analysis was performed according to Box-Cox transformation. The results obtained are given in Table 4. Table 4. Regression analysis Source Regression Mixing, rpm Temperature, ◦C S/I Error Total

Adj SS Adj MS 13.6899 4.5633 13.1289 13.1289 0.4659 0.4659 0.0951 0.0951 0.1489 0.0372 13.8388 *0.05

P values less than 0.05 indicate that coefficient estimates are significant. The fact that F values are high and P values are low shows that the regression coefficients are holistically significant. As a result of the regression analysis, R2 value was calculated as 98.92. R2 value is suitable as it is greater than 95. As a result of the calculations, the following regression equation has obtained. 𝑀𝑒𝑡ℎ𝑎𝑛𝑒 0.5 = 10.286 + 0.08540 × 𝑀𝑖𝑥𝑖𝑛𝑔 + 0.02413 × 𝑇𝑒𝑚𝑝𝑒𝑟𝑎𝑡𝑢𝑟𝑒 + 0.0363 ∗ 𝑆⁄𝐼 (2) This equation indicates that on methane yield, mixing has the highest effect and temperature has the lowest effect. Temperature is a parameter that affects the optimum reproduction of microorganisms and the decomposition rate/solubility of organic wastes. It can be said that mixing increased homogeneous distribution of total solids content, no dead zone formation, and microorganisms' access to food. Less effect of mesophilic and thermophilic conditions (temperature) may be due to the use of CMLF with more than 80% biodegradability [21]. It can also be related to both total solids content of this mixture and the type/number of microorganisms in the medium as a result of inoculation addition. The temperature may not have been had a major impact on the biogas yield, as it was studied at the optimum temperature values (35-55 C). Figure 3 shows the cumulative biogas and final bio-methane content. It is seen that higher biogas production and methane production in 5, 6, 7 and 8 samples than others. This result shows that mixing speed compared to S/I ratio and temperature is a more effective parameter on gas production under these conditions.

3.2. Properties of liquid products

pH of the samples was adjusted to 7.1 - 7.2 before starting the study, pH values of the liquid products were measured in the range of 7.73-8.18 after the digestion. When the results in Table 5 are examined, it can be said that the ratio of the S/I rather than the mixing speed and temperature affects pH more. Moreover, pH may have changed as a result of the conversion of nitrogenous compounds to ammonia [54] and the change in CO2 solubility [5]. It was observed that pH does not fluctuate much in systems where cattle manure is used [55]. In this study, it can be said that there is not a great fluctuation and that the observed change may be due to the CMLF use and S/I ratio. COD removal was achieved by over 81% in all samples. The highest removal efficiency was 88.8% in sample 8. It can be said that the mixing speed has a higher effect on COD removal efficiency. It was stated that mixing increases soluble COD removal [24]. The use of only the liquid fraction of the cattle manure after it has passed through the separator has resulted in a low number of colloidal substances. Moreover, the fact that the particle size of the municipal sewage sludge used was very small increased the decomposition efficiency. The low content of colloidal matter also indicates in the low level of insoluble/difficult to decompose components [6]. Table 5. Properties of the supernatant in liquid products after anaerobic digestion Sample Mixing speed Temperature (rpm) (℃) 1 2 3 4 5 6 7 8