Optimization of solid content,carbon/nitrogen ratio and food/inoculum ratio for biogas production from food waste

Abstract

Biogas production from food waste has been used as an efficient waste treatment option for years. The methane yields from decomposition of waste are, however, highly variable under different operating conditions. In this study, a statistical experimental design method (Taguchi OA₉) was implemented to investigate the effects of simultaneous variations of three parameters on methane production. The parameters investigated were solid content (SC), carbon/nitrogen ratio (C/N) and food/inoculum ratio (F/I). Two sets of experiments were conducted with nine anaerobic reactors operating under different conditions. Optimum conditions were determined using statistical analysis, such as analysis of variance (ANOVA). A confirmation experiment was carried out at optimum conditions to investigate the validity of the results. Statistical analysis showed that SC was the most important parameter for methane production with a 45% contribution, followed by F/I ratio with a 35% contribution. The optimum methane yield of 151 l kg⁻¹ volatile solids (VS) was achieved after 24 days of digestion when SC was 4%, C/N was 28 and F/I were 0.3. The confirmation experiment provided a methane yield of 167 l kg⁻¹ VS after 24 days. The analysis showed biogas production from food waste may be increased by optimization of operating conditions.

Keywords

Food waste methane potential optimization Taguchi methods anaerobic digestion

Introduction

Anaerobic digestion is an effective process for generation of methane, which involves complicated microbial processes on decomposition of organic waste and subsequent conversion of metabolic intermediate products to methane and other gases. Anaerobic digestion of food waste has become an important issue for Turkey in recent years. In Turkey, organic constituents account for more than 50% of municipal solid waste (Metin et al., 2003). About 70% of the total amount of municipal solid waste was disposed in landfills in Turkey in 2004 (Turan et al., 2009). European Union (EU) Waste Framework Directive requires a reduction in the amount of landfilled biodegradable materials. Being a candidate for the EU, Turkey has planned to reduce the amount of landfilled biodegradable materials by 50% of its level in 2005 by year 2018 and 65% by 2025. Examination of biogas potential of food waste and optimization of operating conditions for maximization of methane production from food waste have, therefore, become more important.

Many previous studies showed that food waste may be used to produce biogas. Regarding the energy potential of food waste, different values were reported in the literature. Cho et al. (1995) showed that the ultimate methane yields of cooked meat, boiled rice, fresh cabbage and mixed food waste were 482, 294, 277 and 472 l kg⁻¹ volatile solids (VS) added, respectively, under different days of operation time. Li et al. (2009) reported that ultimate methane yield of food waste (kitchen waste) was 362 l kg⁻¹ VS. El-Mashad and Zhang (2010) found methane yields of approximately 353 l kg⁻¹ VS after 30 days of digestion. Zhang et al. (2007) showed that methane yield can be as high as 348 and 435 l kg⁻¹ VS, respectively, after 10 and 28 days of digestion. Beck and Adolph (2010) reported methane yields of approximately 350 l kg⁻¹ VS, while their pilot study achieved 640 l kg⁻¹ VS. Grasmug and Braun (2002) reported biogas (including methane, carbon dioxide and some other gases) yields of 1100 l kg⁻¹ VS. Schott et al. (2013) reported up to 399 l kg⁻¹ VS ultimate methane yield from food waste. Browne and Murphy (2013) reported that methane potential of food waste was between 467 and 529 l kg⁻¹ VS. It is evident from the previous literature that methane yields from food waste are highly variable. This variability was not only related to composition of food waste but also operating conditions of anaerobic systems.

Anaerobic processes are controlled by complex interaction of several factors (Rittmann and McCarty, 2001). Factors such as substrate preparation, food/inoculum ratio (F/I), carbon/nitrogen ratio (C/N), liquid and headspace volumes, pH of substrate and inoculum, headspace pressure can affect methane yields (Browne and Murphy, 2013). Although anaerobic technology has matured significantly in recent decades, there is still uncertainty regarding how interaction of several factors affects a system’s performance. Some previous studies have investigated the effects of experimental conditions on biogas production from food waste. Lopes et al. (2004) and Forster-Carneiro et al. (2007) aimed to understand the influence of inoculum during anaerobic treatment of the organic fraction of municipal solid waste. Guendouz et al. (2012) investigated the effect of solid content (SC) on anaerobic digestion in batch reactors. Bouallagui et al. (2003) investigated the effect of fruit and vegetable composition on biogas production. However, most studies focused on understanding the role of single parameter on biogas production and studies that examined the relative importance of several parameters and their optimum levels were rare. Optimization of operating conditions is necessary for maximum recovery of methane from organic waste (Khalida et al., 2011).

The main objectives of this study were: (1) to optimize the conditions for methane production from food waste, and (2) to examine the relative importance of three parameters on methane production from food waste. The parameters selected were SC, C/N and F/I. These parameters are selected as they are directly related to the characteristics of the food waste (Zhang et al., 2007). A statistical experimental design method (Taguchi OA₉) was implemented to examine the contribution of each variable on system’s performance. Nine anaerobic reactors having different characteristics were operated under mesophilic conditions (37 ± 1°C) at pH 7 for 24 days. The optimum conditions were determined by application of statistical tests such as analysis of variance (ANOVA). The validity of the results was evaluated using a confirmation experiment.

Materials and methods

Materials

To minimize the effects of variation in food composition, we preferred to prepare a mixture from a specific group of food waste. Food waste, consisting of rice, pasta, lettuce and meat, was obtained from the Erciyes University Cafeteria. The food was shredded into a slurry state with a food grinder. The prepared materials were stored in a freezer at −18^oC for later use. The materials were then defrosted at room temperature prior to experimental use. The inoculum was obtained from the methanogenic reactor of Pakmaya Yeast Baker Factory in Izmir.

Experimental and analytical setup

Total solids, VS and moisture content were determined according to standard methods (APHA, 2005). The pH was determined using a pH meter (Hach-Lange, HQ-40D). Elemental composition (C, H, N and S) of the food waste was attained by ultimate analysis using element analyser (TruSpec MICRO). Methane production was measured by a liquid displacement method by passing the total gas through distilled water containing 1% KOH (w/v) to remove CO₂ produced. All methane yields were reported at standard temperature and pressure. Total Kjeldahl nitrogen (TKN) was analysed by a TKN analyser (Buchi-b-324).

The experiments were performed in anaerobic glass reactors (batch reactors) at a mesophilic temperature of 37 ± 1°C. Nine lab-scale reactors were used in this work according to the reactor design shown in Figure 1. In this study, the selected parameters consisted of SC, C/N and F/I, which are directly related to the characteristics of food waste. Three levels were selected for each parameter. The levels for SC were 4%, 8%, 12%; they were 28, 32 and 36 for C/N, and 0.3, 0.5 and 1 on VS basis for F/I (Table 1). The levels for C/N and F/I were reached by mixing the food waste (rice, pasta, lettuce and meat) and inoculum at specific proportions. The waste composition of each reactor is presented in Table 2. After C/N values were set, we added water to reach the appropriate SC levels.

Table 1.

Parameters and their levels for Taguchi method.

Parameter	Level 1	Level 2	Level 3
A: Solid content (%)	4	8	12
B: Carbon/nitrogen ratio	28	32	36
C: Food/inoculum ratio	0.3	0.5	1

Table 2.

Composition of wastes added to each reactor. The percentages are given in weight basis.

Parameters/materials	R-1	R-2	R-3	R-4	R-5	R-6	R-7	R-8	R-9
Meat	67%	44%	26%	77%	33%	17%	80%	63%	38%
Pasta	17%	22%	20%	12%	22%	17%	11%	13%	10%
Lettuce	17%	22%	36%	8%	33%	11%	5%	13%	10%
Rice	0%	11%	18%	4%	11%	56%	4%	13%	43%

All nine reactors had 1.5 l capacity (working volume: 1.2 l) with an internal diameter of 11 cm and a height of 20 cm. Initially, inoculums and four kinds of food (rice, pasta, lettuce and meat) waste were mixed to obtain the proportions summarized in Table 1 before feeding into the reactors. The reactors were buffered with a buffer solution of pH 7. To ensure anaerobic conditions, the headspace was purged with an inert gas (N₂) for 5 min. Reactors were put into a drying oven that was connected to thermostatically controlled units kept at temperature of 37 ± 1°C. Each reactor was shake-mixed manually once a day. To remove CO₂ produced, 1% KOH solution was filled in the gas-collecting bottle. The methane produced displaced a measurable volume of KOH solution from the gas-collecting bottle, which was equivalent to the methane volume. The volume of methane from each reactor was determined by a measuring cylinder, which was connected to the gas-collecting bottle (Kawai et al., 2014; Lin et al., 2013).

The reactors were run during a 24-day processing period twice (defined as first and second experimental series). Upon completion of the 24-day period in the first series of experiments, under the same conditions second series of experiments were started. At least two experimental series were required to reach conclusions about the influence of parameters investigated.

The theoretical methane potential was calculated to determine the theoretical methane yields for each reactor. The USDA (2015) National Nutrient Database for Standard Reference was the major source of food composition data. This database provides the foundation for most food composition databases in the public and private sectors. Average nutrient values (expressed in terms of lipids, proteins and carbohydrates) per 100 g for the foods used in this study (meat, pasta, lettuce and rice) were taken from this database (e.g. nutrient values for rice: 2.02 g protein, 0.19 g lipid, 21.09 g carbohydrate). Theoretical methane yields of carbohydrates, proteins and lipids were obtained from Neves et al. (2008) (Table 3). Based on these data, theoretical methane potential was calculated for each reactor.

Table 3.

Theoretical methane potential of carbohydrates, proteins and lipids (Neves et al., 2008).

Organic compound	Final methane yield (l kg⁻¹ VS added)
Carbohydrates	370
Protein	390
Lipids	430

VS, volatile solids.

Experimental design

In this study, the Taguchi method was used for experimental design. The Taguchi method (Taguchi, 1986) is a multi-parameter optimization procedure, which is very useful in identifying and optimizing dominant process parameters with a minimum number of experiments (Roozbehani et al., 2015).

The method is based on an orthogonal array of experiments. An orthogonal array is a minimal set of experiments with various combinations of parameter levels. Output of the orthogonal array, which indicates the relative influences of various parameters on the formation of the desired product, is used to optimize an objective function (Taguchi and Konishi, 1987). There are three types of objective functions: larger-the-better, smaller-the-better and nominal-the-best. The larger-the-better function aims to maximize the response from a system, while the smaller-the-better function aims to minimize it. Nominal-the-best targets a nominal value and aims to minimize the variability around it. The influences are commonly referred in terms of signal-to-noise (S/N) ratio. In this study, for optimization of methane production, the larger-the-better category was used. The exact relation between S/N ratio and the signal is given as in Eq. 1.

S / N = - 10 L o g_{10} [\frac{1}{n} \sum_{i = 1}^{n} \frac{1}{Y_{i}^{2}}]

(1)

In Eq. 1, Y_i is the signal (methane production level) measured in each experiment averaged over n repetitions. The performance value corresponding to the confirmation experiment conditions may be predicted by utilizing the balanced characteristic of the orthogonal array. Therefore, the following extra model, given in Eq. 2, may be used.

Y_{i} = μ + X_{i} + e_{i}

(2)

In Eq. 2, X_i is the fixed effect of the parameter level combination used in the ith experiment, µ is the overall mean of performance value and e_i is the random error in the ith experiment. When the experimental results are given in percentage (%), before applying Eq. 2, the Ω transformation of percentage values is performed using Eq. 3 (Taguchi and Konishi, 1987).

Ω (d b) = - 10 L o g (\frac{1}{P} - 1)

(3)

In Eq. 3, P is the percentage of the methane production level and Ω is the decibel value of percentage value for omega transformation. The prediction error is calculated by subtracting the observed Y_i from the predicted Y_i. The confidence intervals for the prediction error may be calculated as in Eqs 4, 5 and 6. In these equations S_e is the standard deviation confidence interval, n is the number of rows in the matrix experiment, n_r is the repetition number of confirmation experiments and n_Ai, n_Bi, n_Ci,… are the replication numbers for the parameter levels A_i, B_i, C_i, …

S_{e} = \pm 2 \sqrt{(\frac{1}{n_{0}}) σ_{e}^{2} + (\frac{1}{n_{r}}) σ_{e}^{2}}

(4)

σ e^{2} = (s u m o f s q u a r e s d u e t o e r o r) / (d e g r e e s o f f r e e d o m f o r e r o r)

(5)

\frac{1}{n_{o}} = \frac{1}{n} + [\frac{1}{n_{A i}} - \frac{1}{n}] + [\frac{1}{n_{B i}} - \frac{1}{n}] + [\frac{1}{n_{b C}} - \frac{1}{n}] + \dots .

(6)

If the prediction value exceeds the confidence interval, there is a possibility that the extra model does not give accurate results.

In this study, Taguchi method has been adopted for three parameters in three different levels. The selected parameters were SC, C/N and F/I (Table 1). This is a three-parameter-three-level design. Therefore, L-9 orthogonal array (Taguchi and Konishi, 1987) was chosen as per design as presented in Table 4.

Table 4.

Experimental design matrix (L-9 orthogonal array) and results of experiments.

Experiment no.	Variables and their levels			Methane production (l kg⁻¹ VS)
Experiment no.	A	B	C	First series	Second series	Average
1	1	1	1	170	132	151
2	1	2	2	194	122	158
3	1	3	3	143	95	119
4	2	1	2	159	137	148
5	2	2	3	14	7	11
6	2	3	1	97	99	98
7	3	1	3	40	36	38
8	3	2	1	163	113	138
9	3	3	2	13	5	9

VS, volatile solids.

Statistical analysis

To identify which parameters are effective in the methane production process, a statistical ANOVA was performed. The F-test is a tool to see which process parameters have a significant effect on the methane production. The F-value for each process parameter is simply a ratio of mean of the squared deviations to the mean of squared error. When F-value is larger for a parameter, its effect on the performance criteria value is also larger. We also calculated the contribution values for each parameter and p-values. Contribution is defined as the ratio of pooled sum of squares to the total sum of squares. p-values provided information about the statistical significance of the results.

Results and discussion

Characteristics of the waste materials and inoculum

The average values for total solids, VS, moisture content, carbon content and C/N for food waste samples and inoculum are shown in Table 5. All values for total solids, VS and moisture content were calculated on wet weight basis. As seen, C/N ratios for meat, pasta, lettuce and rice, used in this study, changed between 17.8 and 127.7, and moisture contents changed between 57.8% and 88.6%. The inoculum had C/N ratio of 29.3 and moisture content of 91.4%.

Table 5.

Physicochemical characteristics of food waste and inoculum.

Parameters/materials	Total solids (%)	Volatile solids (%)	Moisture content (%)	TKN (%)	C (%)	C/N ratio
Meat	35.6	29.7	64.4	3.1	54.5	17.8
Pasta	18.3	11.9	81.7	1.9	55.4	29.0
Lettuce	11.4	10.5	88.6	1.0	55.1	55.9
Rice	42.2	41.3	57.8	0.4	55.1	127.7
Inoculum	8.6	2.9	91.4	1.8	52.5	29.3

TKN, total Kjeldahl nitrogen.

This food waste and inoculum were mixed in a way to obtain the conditions given in Table 1. For example, for the first reactor, SC, C/N and F/I were set to their first parameter levels given in Table 1 (4%, 28 and 0.3, respectively). As explained, same conditions with the same proportions of food and inoculum were used for the first and second experimental series.

Methane production

Both series of experiments were continued for 24 days. In Figure 2 and Table 4, cumulative methane production values obtained from nine reactors during the 24-day period were shown.

Reactor-2 and Reactor-1 provided the highest two cumulative methane production (194 and 170 l kg⁻¹ VS, respectively) after 24 days digestion during the first experimental series (Table 4 and Figure 2). For the second experimental series, the highest cumulative methane production was observed at Reactor 4 with 136 l kg⁻¹ VS. Reactor 1, with cumulative methane production of 130 l kg⁻¹ VS, had the second highest methane yield. In both series, the lowest two cumulative biogas yields were observed at Reactor-5 and Reactor-9. In general, the methane yields at all reactors in the first experimental series were higher than that of the second experimental series. However, the differences between two experimental series were minor and all reactors showed an almost similar amount of change. The change between two series is most likely to be related to the differences in inoculum activity.

We calculated theoretical methane potentials for nine reactors to be able to compare the theoretical values with the experimental ones (Table 6). After 24 days of digestion in the first experimental series, Reactor-2 was able to produce 50% of its theoretical methane potential and Reactor-1 was able to produce 43% of its theoretical methane potential. For the second series, Reactor-2 and Reactor-1 reached 38% and 41% of their theoretical methane potentials.

Table 6.

Theoretical methane potential and final methane production.

Reactors	CH₄, l kg⁻¹ VS	First series		Second series		Average (first and second series)		Confirmation experiment
	Theoretical CH₄ production	CH₄, l kg⁻¹ VS	(%)	CH₄, l kg⁻¹ VS	(%)	CH₄, l kg⁻¹ VS	(%)	CH₄, l kg⁻¹ VS	(%)
	Theoretical CH₄ production	Methane production 24-day	Achieved methane potential	Methane production 24-day	Achieved methane potential	Methane production 24-day	Achieved methane potential	Methane production 24-day	Achieved methane potential
R-1	396	170	43	132	33	151	38	167	42
R-2	389	194	50	122	31	158	41	–	–
R-3	384	143	37	95	25	119	31	–	–
R-4	396	159	40	137	34	148	37	–	–
R-5	387	14	4	7	2	11	3	–	–
R-6	377	96	26	99	26	98	26	–	–
R-7	396	40	10	36	9	38	10	–	–
R-8	392	163	41	113	29	138	35	–	–
R-9	384	13	3	5	1	9	2	–	–

VS, volatile solids.

In general, the methane yields obtained for the nine reactors in this study were lower than 200 l kg⁻¹ VS. These values seem to be lower than the values reported in the literature. Previous studies estimated ultimate methane yields on the order of 200–400 l kg⁻¹ VS from food waste under different operating conditions (Cho et al., 1995; Grasmug and Braun, 2002). However, the characteristics of food waste included in the previous studies are very different from the characteristics of food waste in this study. In this study, our objective was to obtain highest yields possible from a specific mixture of waste. Therefore, the food waste used in this study included very basic foods. In other studies, complex mixtures were used. Therefore, the difference in methane yields is an expected finding. Here, we also have to note that the experiments in this study were continued for 24 days and the methane yields reported are not ultimate yields. The ultimate methane yields from our experiments would probably be higher.

Optimum operating conditions

To identify which parameters are effective in the methane production process, a statistical ANOVA was performed. F-values and contribution ratios were calculated. With the performance characteristics and ANOVA analyses, the optimal combination of process parameters was predicted. The results of variance analysis are given in Table 7.

Table 7.

The results of variance analysis.

Parameter	Sum of squares	Degrees of freedom	Mean of squares	F-value	Contribution	p-value
Solid content	20 847	2	10 423	4.71	45%	0.02
Carbon/nitrogen	4393	2	2196	0.99	9.5%	0.39
Food/inoculum	16 583	2	8292	3.75	36%	0.04
Error	2348	11	2213		9.5%
Total	66 171	17	23 124

According to F-values and contribution ratios given in Table 7, SC is the parameter that had the largest effect on methane production from food waste. The contribution of SC on the methane production was 45% (p=0.02). This parameter is followed by F/I and C/N. F/I had contribution of 36% (p=0.04). The contribution of C/N was much lower than other two (9.5% contribution, p=0.39). The contribution ratios of three parameters sum up to 90.5%. This shows that there may be some uncertainties in predictions or there may be some other factors that control methane yields. Nevertheless, the results are still very informative. This analysis showed by controlling SC and F/I parameters, we can control 81% of the variation in methane yields.

The optimal level of a parameter is the level with the highest S/N value. Figure 3 shows the variation of the performance characteristics with the variables. To determine the experimental conditions for the first data point for the first parameter, which is level 1 or 4% for SC parameter, the experiments with this condition were determined. These experiments were experiments 1, 2 and 3. The performance characteristics value of the first data point was then calculated as the average of those obtained from experiments 1, 2 and 3. Experimental conditions for the second data point are the conditions of the experiments for which SC is level 2 or 8% (experiments 4, 5 and 6). Similarly, experimental conditions for the third data point are the conditions of the experiments for which SC is level 3 or 12% (experiments 7, 8 and 9). The numerical value of the maximum point in each graph provides the best condition for that parameter. Based on Figure 3, the best value for each parameter was found as A1 (SC of 4%), B1 (C/N of 28) and C1 (F/I of 0.3). These parameter values provide the optimum conditions.

In this study, we found that SC is the most critical parameter. The importance of SC on biogas yields was demonstrated by some previous studies. Kossmann and Pönitz (1999) stated that the mobility of the methanogens within the substrate is gradually impaired by increasing SC, which lowers the methane production. Using Anaerobic Digestion Model No. 1 and experimental data, Guendouz (2012) showed that mass transfer limitation caused low methane production at high SC, and that hydrolysis rate constants decrease with increasing SC. Desai et al. (1994) mentioned that biogas yield was high for SC values between 6% and 10% and gas production slowed when SC exceeds 12%. In this study, we found that the highest methane yields were possible at SC of 4%.

F/I ratio was the second most important parameter in our experiments. The effect of F/I on biogas production was shown to be significant by some studies. (Hashimoto, 1989; Liu et al., 2009). F/I affects the amount of active inoculum requirements (Cheng and Zhong, 2014). Decreases in methane yields were reported when F/I value is higher. Schievano et al. (2010) applied different F/I (1, 0.5 and 0.33) to predict biogas production from food waste and inhibition rate. The results showed that the highest biogas production was achieved for F/I of 0.33. Similar to these studies, we found that optimum value for F/I parameter was 0.3.

Sosnowski et al. (2003), Amon et al. (2007) and Eltawil and Belal (2009) investigated the effects of different C/N on methane yields. Experimental results indicated that for three studies, the highest methane yield was achieved at C/N of 25–28. In this study, we obtained the highest methane yields at C/N of 28, with 24 days of digestion. As can be seen from the results, optimum C/N ratio obtained in this study demonstrated compliance with the literature values.

In order to test the validity of predicted results, a confirmation experiment was carried out at the optimum working conditions. A reactor was set with SC of 4%, C/N of 28% and F/I of 0.3, and followed for methane production for another 24 days. Cumulative methane productions obtained from the confirmation experiment were shown graphically on Figure 2 together with the results obtained from the Reactor-1 in the first and second experimental series, respectively. The comparison of the theoretical methane potential and methane production in the confirmation experiment was also shown in Table 6. After 24 days, the methane production during the confirmation experiment reached 42% of the theoretical methane potential with 167 l kg⁻¹ VS methane production. The cumulative methane production at 30 days was 210 l kg⁻¹ VS, which was 53% of the theoretical methane yield. The confirmation experiment showed that the results found in this study are repeatable.

Conclusions

Biogas production from food waste has become an important topic in recent years due to recent legislation in EU and, in Turkey, that requires reductions in the amount of landfilled biodegradable waste. In Turkey, the proportion of biodegradable waste in municipal solid waste is significant; therefore, anaerobic digestion provides a feasible and economical alternative. In this study, we aimed to examine the optimum levels of operating conditions to maximize methane production from food waste.

The experiments were designed according the Taguchi method. Taguchi’s L-9 orthogonal array design provided a quick and cost-effective method, as fewer experiments are required for determining optimum conditions. In this study, we focused on the effects of SC, C/N and F/I on methane yields. With the Taguchi experimental design, we were able to examine not only the individual effects of these parameters but also their interactions with each other. The results showed that SC was the most important parameter on methane yields, having 45% contribution on performance. F/I ratio was shown to be the second most important parameter. The highest average cumulative methane production (151 l kg⁻¹ VS) was obtained at SC of 4%, C/N of 28 and F/I of 0.3 with 24 days of digestion. This amount corresponds to 41% of theoretical methane potential. The results were confirmed with a confirmation experiment set at the optimum operating conditions.

This study showed that Taguchi methodology may be used for optimizing operating conditions for biogas production. In the future, more parameters and levels may be incorporated into the experimental design to obtain a complete picture of biogas production from food waste. The results from this study could also be used in practice for maximizing biogas production in biogas plants utilizing food or municipal waste and for planning new biogas plants for this purpose.

Footnotes

Acknowledgements

We would like to thank Emine IRBAS and Sibel SAYGI for providing assistance in the laboratory studies. We also would like to thank Erciyes University Student Cafeteria and Pakmaya for providing the food waste and inoculum.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by the Scientific and Technological Research Council of Turkey (TUBITAK), Project Number: CAYDAG 112Y240.

References

Amon

Kryvoruchko

Zollitsch

Mayer

Gruber

(2007) Biogas production from maize and dairy cattle manure – influence of biomass composition on the methane yield. Agriculture, Ecosystems and Environment 118: 173–182.

APHA (2005) Standard methods for the examination of water and wastewater. Washington, DC: American Public Health Association.

Beck

Adolph

(2010) Anaerobic usage of leftovers for energy production – experiences of laboratory and pilot scale plants (in German). University of Hohenheim.

Bouallagui

Cheikh

Marouani

Hamdi

(2003) Mesophilic biogas production from fruit and vegetable waste in a tubular digester. Bioresource Technology 86: 85–89.

Browne

Murphy

(2013) Assessment of the resource associated with biomethane from food waste. Applied Energy 104: 170–177.

Cheng

X-Y

Zhong

(2014) Effects of feed to inoculum ratio, co-digestion, and pretreatment on biogas production from anaerobic digestion of cotton stalk. Energy & Fuels 28: 3157–3166.

Cho

Park

Chang

(1995) Biochemical methane potential and solid state anaerobic digestion of Korean food wastes. Bioresource Technology 52: 245–253.

Desai

Patel

Madamvar

(1994) Effect of temperature and retention time on biomethanation of cheese whey–poultry waste–cattle dung. Environmental Pollution 83: 311–315.

El-Mashad

Zhang

(2010) Biogas production from co-digestion of dairy manure and food waste. Bioresource Technology 101: 4021–4028.

10.

Eltawil

Belal

EBA

(2009) Evaluation and scrubbing of biogas generation from agricultural wastes and water hyacinth. Misr Journal of Agricultural Engineering 26: 534–560.

11.

Forster-Carneiro

Pérez

Romero

Sales

(2007) Dry-thermophilic anaerobic digestion of organic fraction of the municipal solid waste: focusing on the inoculum sources. Bioresource Technology 98: 3195–3203.

12.

Grasmug

Braun

(2002) Anaerobic digestion of biological waste from Vienna – fermentability and usability of leftovers and market waste. Department for Agrobiotechnology, Tulln, Austria.

13.

Guendouz

Brockmann

Trably

Delgenès

Steyer

Escudié

(2012) Total solids content drives high solid anaerobic digestion via mass transfer limitation. Bioresource Technology 111: 55–61.

14.

Hashimoto

(1989) Effect of inoculum/substrate ratio on methane yield and production rate from straw. Biological Wastes 28: 247–255.

15.

Kawai

Nagao

Tajima

Niwa

Matsuyama

Toda

(2014) The effect of the labile organic fraction in food waste and the substrate/inoculum ratio on anaerobic digestion for a reliable methane yield. Bioresource Technology 157: 174–180.

16.

Khalida

Arshadb

Anjuma

Mahmood

Dawson

(2011) The anaerobic digestion of solid organic waste. Waste Management 31: 1737–1744.

17.

Kossmann

Pönitz

(1999) Biogas Digest. Biogas Basics Volume I, Information and Advisory Service on Appropriate Technology (ISAT), GATE in Deutsche Gesellschaft für Technische Zusammenarbeit (GTZ), GmbH, Eschborn.

18.

Chen

(2009) Anaerobic co-digestion of kitchen waste and cattle manure for methane production. Energy Sources, Part A: Recovery, Utilization, and Environmental Effects 31: 1848–1856.

19.

Lin

Wang

(2013) Hydrogen-methane production from pulp & paper sludge and food waste by mesophilic–thermophilic anaerobic co-digestion. International Journal of Hydrogen Energy 38: 15055–15062.

20.

Liu

Zhang

El-Mashad

Dong

(2009) Effect of feed to inoculum ratios on biogas yields of food and green wastes. Bioresource Technology 100: 5103–5108.

21.

Lopes

Leite

Prasad

(2004) Influence of inoculum on performance of anaerobic reactors for treating municipal solid waste. Bioresource Technology 94: 261–266.

22.

Metin

Eröztürk

Neyim

(2003) Solid waste management practices and review of recovery and recycling operations in Turkey. Waste Management 23: 425–432.

23.

Neves

Gonçalo

Oliviera

Alves

(2008) Influence of composition on the biomethanation potential of restaurant waste at mesophilic temperatures. Waste Management 28: 965–972.

24.

Rittmann

McCarty

(2001) Environmental Biotechnology. New York: McGraw-Hill.

25.

Roozbehani

Sakaki

Shishesaz

Abdollahkhani

Hamedifar

(2015) Taguchi method approach on catalytic degradation of polyethylene and polypropylene into gasoline. Clean Technologies and Environmental Policy 17: 1873–1882.

26.

Schievano

D’Imporzano

Malagutti

Fragali

Ruboni

Adani

(2010) Evaluating inhibition conditions in high-solids anaerobic digestion of organic fraction of municipal solid waste. Bioresource Technology 101: 5728–5732.

27.

Schott

Vukicevic

Bohn

Andersson

(2013) Potentials for food waste minimization and effects on potential biogas production. Waste Management & Research 31: 811–819. doi: 810.1177/0734242X13487584.

28.

Sosnowski

Wieczorek

Ledakowicz

(2003) Anaerobic co-digestion of sewage sludge and organic fraction of municipal solid wastes. Advances in Environmental Research 7: 609–616.

29.

Taguchi

(1986) Introduction to Quality Engineering: Designing Quality into Products and Processes. Asian Productivity Organization.

30.

Taguchi

Konishi

(1987) Orthogonal Arrays and Linear Graphs. Dearborn, MI: SI Press.

31.

Turan

Çoruh

Akdemir

Ergun

(2009) Municipal solid waste management strategies in Turkey. Waste Management 29: 465–469.

32.

USDA (US Department of Agriculture) (2015) US Department of Agriculture, Agricultural Research Service, Nutrient Data Laboratory. USDA National Nutrient Database for Standard Reference, Release 28. Version Current: September 2015. Available at: http://www.ars.usda.gov/nea/bhnrc/ndl

33.

Zhang

El-Mashad

Hartman

Wang

Liu

Choate

and Gamble

(2007) Characterization of food waste as feedstock for anaerobic digestion. Bioresource Technology 98: 929–935.