Introduction

Recently, there has been considerable interest concerning the use of livestock manure as an alternative renewable source of energy. This is due in part to the continuous economic and environmental concerns facing farmers and governments [1]. Anaerobic digestion (AD) is considered to be an important and advantageous process in livestock manure waste treatment. It converts plant biomass, crop residues, animal manures, and other organic wastes into methane-rich biogas, a widely used source of renewable energy, while reducing the environmental hazard of livestock waste [2].

AD is usually classified as wet, semi-solid, or solid state, referring to total solids (TS) of substrate percentages of <10, 10–15, or >15 %, respectively [3]. Of these, dry anaerobic digestion has several benefits over conventional wet anaerobic digestion [4], including reduction of the reactor volume, high volumetric methane productivity, low energy requirements for heating, positive energy balance, and less wastewater, as well as an end product of more easily transportable fertilizer [5, 6]. Thus, dry anaerobic digestion of high solid content feedstock has shown continuous growth [7–9]. However, while the wet or semi-dry digestion process of animal manure is an established method and has been widely studied [10], very few studies have been conducted on the AD of poultry manure (PM), especially in a solid state or dry condition [11–13].

An excess of either nitrogen or carbon content could lead to process inhibition [14]. High ammonia content derived from organic nitrogen in PM under dry conditions makes this substrate difficult to digest anaerobically [15]. Several researchers have investigated the inhibitory effect of ammonia during anaerobic digestion of animal wastes [16, 17]. Many studies have focused on reducing the inhibitory effect of ammonia and volatile fatty acids (VFA) to improve the fermentation of PM for the production of methane. In one such study, PM was diluted with water to decrease the total percentage of solid [18]. Ammonia removal has also been attempted by different researchers in the effort to enhance methane fermentation [19]. However, the reported removal methods also necessitated decreasing the TS% through dilution with water, resulting in an increase in waste volume. Chemicals were also needed to remove ammonia, thus increasing cost.

An alternative method for dry AD of PM is co-digestion. It offers several benefits including increased biogas production resulting from increased loading of readily biodegradable organics, improved balance of nutrients and C/N ratio, dilution of toxic substances including ammonia without the addition of water and expensive chemicals, a better quality of a digested product, and reduced costs from processing several substrates in one installation [2, 20, 21]. Actually, Borowski and Weatherley [15] showed that 30 % addition of PM to sewage sludge increased total gas production by 50 %. There are many examples of co-digestion of PM, and other research has shown positive results with other types of livestock waste, including cattle manure [13], hog wastes [22], anaerobically digested sludge (ADS) [23], fruit and vegetable wastes (FVW), cattle slurries [24], buffalo manure with OFMSW [25], and a mixture containing 40 % dairy manure, 40 % PM and 20 % wheat straw [2]. It may therefore be said that co-digestion of PM with different waste materials is an effective way to reduce the ammonia derived from PM and increase biogas production. However, feedstocks co-digested with PM need to be restricted based on their C/N ratio, and the supplied amount balanced with that of PM. Agriculture wastes with low C/N ratio are thus promising candidates, because increased methane yield and biodegradation rate by co-digestion of PM and agriculture wastes will result in more methane generation per unit fed mass and reduce the size of the reactor required, which makes the process more attractive economically [26].

Coffee is the second largest traded commodity in the world, and its processing generates large amounts of by-products and residues. Although wastes and by-products produced during coffee berry processing constitute a source of severe contamination and pose serious environmental threats in coffee-producing countries, reports on anaerobic digestion of coffee wastes are limited [27]. Coconut and cassava wastes have a high energy potential and a large amount of it is exhausted. They could be used as a co-substrate for anaerobic digestion [28, 29]. However, previously reported experiments on co-digestion of PM and other feedstocks have been carried out under diluted conditions with high water content and low ammonia concentration.

To the best of our knowledge, this study is the first to investigate the mixture of agriculture wastes (AWS) consisting of coconut, cassava wastes, and coffee ground as a co-substrate with PM. One previous study attempted co-digestion of the mixture described above under semi-solid condition with TS of 10 % and obtained 93 % increase in methane production [30]. The main goal of this study was to evaluate the effect of agricultural wastes added as a co-substrate on performance of anaerobic digestion under dry conditions (TS 20 %). Both mesophilic (35 ± 2 °C) and thermophilic (55 ± 2 °C) fermentation condition were used and compared. Methane production, ammonia accumulation, and acetate degradation were used as evaluation parameters.

Materials and Methods

Substrates and Seed Sludge Used for Anaerobic Co-Digestion Processes

Fresh poultry manure (PM) from Hiroshima University Poultry farm (cage layer system) was collected from deposits directly under chicken cages. PM had the following characteristics: 25 % (w/w) TS, 69.8 % TS of VS, 380 g C kg TS−1 total organic carbon (TOC), 87 g N kg TS−1 total Kjeldahl nitrogen (TKN), and 26.5 g N kg TS−1 total ammonia nitrogen (TAN). Treated PM (TPM) was produced by stripping ammonia from PM using the technique described in [19]. TPM was characterized as 20 % (w/w) TS, 70.3 % TS of VS, 85.5 g N kg TS−1 of TKN, and 10.5 g N kg TS−1 of TAN. Agriculture wastes (AWS), consisting of a mixture of cassava (root residue wet cake), coconut (wet cake), and coffee ground waste, was shipped frozen from Thailand and received by Hitachi Engineering and Services, Japan (HES-TH). Characteristics of the AWS are presented in Table 1. The seed sludge collected from a Wastewater Treatment Center in Hiroshima, Japan, was anaerobically incubated at 55 °C for 60 days in the laboratory in order to achieve complete consumption of the substrate. The seed sludge was characterized as 20 % (w/w) TS, 53 % TS of VS, 268 g C kg TS−1 of TOC, 3.2 g N kg TS−1 of TKN, and 3.2 g N kg TS−1 of TAN. All samples were collected in triplicate, and the averages of the three measurements are presented.

Table 1 Characteristics of agriculture wastes (AWS)
Full size table

Experimental Setup and Procedures

Experiment 1 (Exp 1) Dry (20 %) Anaerobic Co-Digestion of PM and AWS with Repeated Batch Culture

The substrates consisted of PM, mixed with AWS, and inoculated with seed sludge. Water was added to adjust TS to 20 %, so that the ratio of PM to AWS was 7:3 (not controlled). Ratio of inoculum (Ozu sludge) to substrate (PM + AWS) was kept 3:1 (v/v). The substrate were placed in a set of 500-mL capacity anaerobic vials, 200 g of each (mixture 1 (M1)). As control, vials without AWS supplement were used (PM control). The headspace in the vials was purged with N2 gas and sealed with rubber stoppers in crimped aluminum caps. These bottles were incubated anaerobically at 35 ± 2 °C (mesophilic condition) or 55 ± 2 °C (thermophilic condition). Semi-continuous batch culture was used for this co-digestion. Triplicate vials for each condition were used. The amounts, composition, and conditions of anaerobic digestion of these substrates and co-substrate are illustrated in Table 2.

Table 2 Experiment 1: dry anaerobic co-digestion of PM and AWs with 20 % TS by repeated batch culture
Full size table

Experiment 2 (Exp 2) Dry (20 %) Anaerobic Co-Digestion of TPM and AWS with Repeated Batch Culture

The treatment procedures of TPM and PM (mixture 2 (M2)) were carried out identically, which was described in “Experiment 1 (Exp 1) Dry (20 %) Anaerobic Co-Digestion of PM and AWS with Repeated Batch Culture” section.

The amounts, composition, and conditions of anaerobic digestion of these substrates and co-substrate are illustrated in Table 3.

Table 3 Experiment 2: dry anaerobic co-digestion of TPM and AWs with 20 % TS by repeated batch culture
Full size table

Analytical Methods

Volumes of gases and their composition were monitored every day. When gas production stopped, vials were opened and samples were taken to measure the produced ammonia, volatile fatty acids (VFAs), and pH. For the next step of the new batch culture, half of the vial contents were removed, and the vials were replenished with the same amount of substrate. This procedure for batch culture was repeated for three batches and was conducted over a total duration of 114 days. Duration of the first batch was 40 days, the second batch was 35 days, and the third batch was 39 days.

Fermentation samples (ca. 0.3 g wet weight) were withdrawn into 2-mL plastic tubes and suspended with 1.2 mL deionized water. The suspension was centrifuged at 3000 rpm for 10 min at 4 °C, and the clear supernatant was used for measurement of pH, ammonia, and VFAs. VFAs were measured using a High Performance Liquid Chromatograph (Shimadzu, Kyoto, Japan) equipped with Aminex HPX-87H Column, 300 mm × 7.8 mm (Bio-Rad, Tokyo, Japan). The column temperature was 65 °C. The flow rate was 0.8 mL min−1 for 0.005 M H2SO4 solution, used as a mobile phase. Ammonia was measured using a commercially available ammonia testing kit (Wako Ltd., Osaka, Japan). TOC was determined with a TOC analyzer (TOC-5000, Shimadzu). TS, VS, TKN, and pH were measured in accordance with the standard methods [31]. Soluble carbohydrate concentration was measured by the phenol sulfuric acid method, with glucose as standard [32]. Soluble protein was determined by the Lowry–Folin method, with bovine serum albumin as the standard [33]. The lipid content was analyzed by Soxhlet with petroleum ether extraction according to AOAC methods [34].

Gas production was measured periodically by displacement of saturated aqueous NaCl in a graduated cylinder. The composition of CH4, H2, and CO2 was determined by a gas chromatograph (GC-8A, Shimadzu) with a thermal conductivity detector equipped with a glass column (2 m × 3 mm) packed with unibeads C 60/80 (Shimadzu) at 140 °C. Argon was used as the carrier gas at a pressure of 100 kPa. All samples were collected in triplicate, and the averages of the three measurements are presented.

Results

Characteristics of Substrate

As mentioned above, we conducted two experiments (Exp 1 and Exp 2); substrate characteristics, components, and contents/concentrations of PM, TPM, and AWS for both are presented in Tables 1 and 2. On a dry matter (TS) basis, AWS contained higher volatile solids (VS) but lower fixed solids (FS) than PM. The higher VS content of AWS (98.7 %) compared with that of manure (70 %) means higher energy content, which is desirable economically with regard to both transportation and biogas energy production. The initial pH of the AWS was 4.85, while it was 8.5 and 7.75 for PM and TPM, respectively.

Methane Production

The cumulative methane profile of dry anaerobic co-digestion from Exps 1 and 2 is shown in Fig. 1. The substrate was composed of PM or TPM mixed with AWs and inoculated with seed sludge at a ratio of 1:3; control vials of PM or TPM with no added AWS were also used. Semi-continuous batch culture was used for the co-digestion. As already mentioned, three batches were carried out, the first for 40 days, the second for 35 days, and the third for 39 days for a total length of each experiment of 114 days. Thermophilic (55 ± 2 °C) and mesophilic (35 ± 2 °C) conditions were used in both experiments.

Fig. 1
figure 1

Methane production from dry anaerobic co-digestion (20 % TS) by repeated batch culture of a Exp 1 (PM only (PM control) or co-digested with AWs (M1)) and b Exp 2 (digestion of treated poultry manure (TPM) only (TPM control) or mixed with AWs (M2)). Seed sludge was used as inoculum. Fermentation temperatures were 35 °C (mesophilic) and 55 °C (thermophilic). Duration of batch I, 40 days; batch II, 35 days; batch III, 39 days

Full size image

Figure 1a shows that in Exp 1 under thermophilic conditions (55 ± 2 °C), co-digestion of PM with AWS (M1) increased methane production by 150 % over the control (maximum production of 323.4 vs. 129 mL g VS−1). Complete inhibition of methane production was reported in PM control at the second and third batches. A maximum of 195 % increase in methane production resulted from M1 under the mesophilic (35 ± 2 °C) condition (highest amount of methane produced was 406 vs. 136.9 mL g VS−1 from PM control). Complete inhibition of methane production was observed in the third batch with PM control.

Figure 1b shows the methane profile of Exp 2 under thermophilic conditions; there was a 70 % increase in methane production from M2 compared with TPM control (297.6 vs. 175.4 mL g VS−1). In the third batch, methane production was totally inhibited in TPM control. On the other hand, co-digestion of M2 under mesophilic conditions (35 ± 2 °C) resulted in 63 % as a maximum increase in methane production from M2 over TPM control. Maximum methane produced was 562 mL g VS−1 from M2 vs. 344 mL g VS−1 from TPM control.

VFAs (Acetate) Production/Degradation, pH, and Ammonia

Table 4 illustrates the course of acetate production/degradation under both thermophilic and mesophilic conditions in Exps 1 and 2. In Exp 1, under thermophilic conditions, acetate was accumulated in both PM control and M1 (second and third batches) with a maximum of 286 mM kg−1 bed accumulated in PM control. Meanwhile, 77.8 % of the resulting acetate was degraded with M1 in the first batch (Table 4). In Exp 2, TPM control and M2 (second and third batches), at 55 °C, acetate accumulated to reach a maximum of 290 mM kg−1 bed, while 31 % of acetate produced was degraded in the first batch of M2. In contrast, in Exp 1, M1 under 35 °C acetate was degraded by 86, 100, and 87 % in the 1st, 2nd and 3rd batches respectively (Table 4). Mesophilic conditions of Exp 2 TPM control showed 97, 67, and 5.2 % degradation of acetate in the first, second, and third batches. Finally, the best acetate degradation was observed with M2 under the mesophilic condition, which showed almost 100 % acetate degradation in all batches (Table 4).

Table 4 Acetate (mM kg−1 bed) production and degradation during dry anaerobic co-digestion (20 % TS) by repeated batch culture
Full size table

Table 5 shows that in both Exps 1 and 2, pH changed according to the degree of acetate degradation. Initial pH for all mixtures ranged from 7.3 to 8; then, pH was increased by degradation of acetate and production of methane. In both experiments, pH increased to a range of 8.1–8.6 by the end of the experiments. In addition, a clear drop in pH values was observed under thermophilic conditions in both experiments.

Table 5 Initial and final changes in pH during dry anaerobic co-digestion (20 % TS) by repeated batch culture
Full size table

Table 6 illustrates that ammonia production increased gradually from the first batch to the end of the third batch under all conditions in both experiments. Table 6 shows that initial ammonia values were 1.8 and 1.3 g N kg−1 bed in the Exp 1 PM control and M1, respectively, under both mesophilic and thermophilic conditions. Under 55 °C, ammonia was accumulated until it reached 12.3 and 7.4 g N kg−1 bed at the end of the third batch of the Exp 1 PM control and M1, respectively, while under 35 °C, ammonia values were 8.2 and 5.4 g N kg−1 bed, respectively. Under the thermophilic condition, a 39 % decrease in ammonia accumulation was found with M1 compared to PM control, while a 34 % decrease was achieved in M1 under the mesophilic condition. The ammonia profile of Exp 2 illustrated in Table 6 shows that initial ammonia values in the bed were 1.04 and 1.2 g N kg−1 bed in the Exp 2 TPM control and M2, respectively, under both mesophilic and thermophilic conditions. In TPM control, ammonia was accumulated until it reached 7.18 and 10.2 g N/kg bed at the end of the third batch, at 35 and 55 °C, respectively, while a 6.06 and 8.5 g N/kg bed was accumulated at the end of the third batch from M2 under 35 and 55 °C, respectively. Addition of AWS resulted in decreased ammonia accumulation by 16 and 24 % at 35 and 55 °C, respectively.

Table 6 Ammonia (NH3-N g kg−1 bed) production during dry anaerobic co-digestion (20 % TS) by repeated batch culture
Full size table

Discussion

The aim of these experiments was to monitor biogas production and composition and to measure the processing parameters that characterize the anaerobic process (pH, VFA, and NH3-N). For anaerobic systems, these parameters are indicative and a change in one parameter will cause change in others [22, 35].

To evaluate biogas production, it is necessary to know the concrete content of organic matter and its composition. This can be determined by several methods, but the most used is measurement of VS that give the content of the organic matter. The theoretical maximum yield per gram of VS depends on the type of organic matter, carbohydrate, proteins, lipids, and VFA [36]. Table 1 illustrate the contents of VS, carbohydrate, proteins, and lipids of AWS. The results clearly indicate that AWS have higher energy content, which makes them suitable as a co-substrate; this result was compatible with that obtained in previous studies [37]. Both manure and AWS contained well-balanced nutrients for anaerobic microorganisms. Results of pH tests showed that initial pH value of AWS is 4.85, while by mixing PM and AWS, the resulting mixture pH (ranged from 7.3 to 8) becomes more suitable for fermentation than with either substrate alone.

Manure is excellent as a “carrier” substrate to allow anaerobic digestion of other kinds of wastes, which would be difficult to treat separately. There are several reasons for this suitability, including the fact that the high buffering capacity contained in manure protects the process against failure owing to drop in pH in case of temporary VFA accumulation. So, the mixing of wastes produces a more suitable pH for anaerobic digestion.

Dry anaerobic fermentation of PM has usually encountered problems with methane inhibition as a result of ammonia accumulation. Improvement of dry anaerobic fermentation of PM had previously studied [11, 12, 19] through either acclimatization or ammonia stripping technology. Here, we investigated improvement of dry fermentation of PM through co-digestion with AWS.

Methane production was a crucial indicator of the success or fail of the process. In Exp 1, highest methane production was 406 mL g V−1 obtained from M1 at 35 °C. This higher methane was attributed mainly to the addition of AWS, which resulted in a 195 % increase in methane production compared with PM control (Fig. 1a). The enhanced methane production resulted from a decrease in ammonia accumulation—by about 52 % compared with PM control (Table 6). In turn, the decrease in ammonia accumulation resulted in complete degradation of VFA (100 % of acetate produced was degraded to methane), keeping pH within the range of 8.1–8.8 (Table 5), compatible with the normal growth of anaerobic microorganisms [38]. Methane production from M1 at 55 °C showed lower methane production of 324 mL g VS−1 than under similar mesophilic condition (Fig. 1a). However, even this amount was 151 % higher than the PM control at the same temperature; there was a 43.7 % decrease in ammonia accumulation in M1 at 55 °C compared to PM control. It was observed that acetate was accumulated in the third batch with both M1 and PM control at 55 °C, and with PM control at 35 °C; this was then responsible for the inhibition of methane production (Table 5). In spite of acetate accumulation, no drop in pH was observed; this may be explained by the buffering action of high ammonia accumulated under PM control (Table 5).

In Exp 2, highest methane production was 562 mL g VS−1 obtained from M2 under the mesophilic condition. This higher methane was attributed mainly to the addition of AWS, which resulted in a 63 % increase in methane production compared with TPM control at 35 °C Table 5. Enhanced methane production resulted from lower ammonia accumulation than control (by 35 % compared to TPM control; Table 6). The amount of methane produced from M2 under the mesophilic condition was 38.4 % higher than that obtained from M1 under the same condition. This may be attributed to lower ammonia accumulation with M2 (3.9, compared with 4.1 g N kg−1 bed from M1).

At 55 °C, M2 showed lower methane production of 298 mL g VS−1 than M2 at 35 °C but was higher than that produced from the M2 control (Fig. 1b). Complete inhibition of methane production was observed in the third batch of TPM control and M2 at 55 °C (Fig. 1b); this can be explained by the acetate accumulation that occurred concurrently with ammonia accumulation (Tables 4 and 6).

In all experiments, methane percentage ranged from 58 to 70 % of biogas produced, which is nearly the same as previously reported [23, 39, 40]. The addition of AWS to PM resulted in an increase in the methane fraction of biogas compared with control with no AWS added (data not shown). This result is incompatible with that obtained by Borowski and Weatherley [15], who reported that addition of poultry manure to the sludge had no impact either on methane content or on the amount of trace gases, even though they reported nearly the same range of 66–68 %.

The superior biogas production and methane yields of M1 and M2 compared to the control demonstrate that co-digestion of PM or TPM and AWS using high TS% (20 %) is available and results in a synergistic effect [2, 22, 42]. Our results proved better performance of co-digestion under mesophilic conditions than under thermophilic conditions, which may be attributed to less ammonia accumulation under the mesophilic (about a 29 % decrease in ammonia accumulation in both M1 and M2 test) (Table 6) [11, 42]. Accumulation of acetate was under the thermophilic condition, resulting in complete cessation of methane production. Acetate accumulation under the thermophilic condition was the limiting factor leading to both a decrease in pH and accumulation of ammonia, which combined to make ammonium acetate that consequently inhibit methane production [19]. In the second batch, the PM control showed accumulation of VFA (especially acetate) above the threshold limit without cessation in methane production, as pH, at 8.7, was high enough to inhibit the toxicity of VFA [15, 23, 39].

Our present result was better than that obtained by Ahn et al. [41], who found that co-digestion of switch grass-poultry manure mixture resulted in slow production of VFA, as well as the accumulation of it, which resulted from minimal conversion of VFAs to CH4 by methanogens. They added that the accumulated VFAs resulted in irreversible digester acidification and the low pH inhibited methanogens from converting VFAs to methane. As no significant drop in pH was observed in our results, this proved that ammonia and acetate accumulation together were likely the limiting factor, rather than the pH drop.

Magbanua et al. [22] obtained similar results from co-digestion of hog and poultry waste They obtained a fixed pH range of 7 ± 0.5 to 7.5 ± 0.5, which is optimum for methanogenesis. But, in contrast to our results, when they used poultry manure alone, pH dropped over the first 15–20 days, to as low as 4.5 ± 0.5.

Dry anaerobic fermentation of PM under thermophilic conditions often encounters very high concentrations of ammonia due to extensive hydrolysis of nitrogen compounds, and the inhibitory effect of ammonia on thermophilic methanogens [41] becomes significant. We obtained higher ammonia accumulation at PM and TPM control at 55 °C (12.28 and 10.2 g N kg−1 bed; Table 6), which caused inhibition of methane production. Mixing of AWS with PM or TPM resulted in lower accumulated ammonia (ranging from 5.35 to 8.55 g N kg−1 bed), which allowed methane production despite an ammonia level higher than the threshold one. This result is different from that obtained by Sung and Liu [42], who observed that total ammonia nitrogen (TAN) concentration of 4.92 and 5.77 g L−1 decreased the production of methane by 39 and 64 %, respectively. Moreover, Sung and Liu [42] reported 100 % inhibition in the range of 8–13 g L−1, depending on the condition of acclimatization and the pH of the system.

The amount of methane produced in these experiments was higher than that obtained in previous studies [11, 12, 19]. Levels of 562 and 406 mL g VS−1 were the highest amount of methane produced from M1 and M2 under 35 °C, respectively. Abouelenien et al. [11], studying dry anaerobic fermentation of PM (25 % TS) under mesophilic conditions, obtained 4.4 l kg−1h6PM-DW (31 mL g VS−1) after 254 days. In another study, they reported a total amount of 103.5 mL g VS−1 methane was produced from twice ammonia stripped PM after 55 days of fermentation under 55 °C [12]. Simultaneous ammonia striping and methane fermentation of PM using a bench-scale reactor under thermophilic conditions resulted in 194.5 and 157.2 mL g VS−1 of methane obtained from TPM and mixture of TPM/PM (1:1), respectively [19].

Additionally, we obtained higher methane production than that obtained by Chen et al. [43], who reported 107.25 mL g TS−1 of biogas (76.92 % methane) was produced from co-digestion of PM with spartina alterniflora residues (SAR) at 35 °C with initial TS of 8 %. Ahn et al. [41] found that anaerobic digestion of poultry manure-switch grass mixture under 15 % TS and thermophilic conditions (55 °C) resulted in very poor methane yield of 2 mL g VS−1 after 62-day digestion. They attributed this poor methane yield to VFA accumulation and pH drop.

Magbanua et al. [22] tried co-digestion of hog and poultry waste in various proportions (poultry manure TS was 1.74 %; digestion temperature was 35 ± 2). They found that treatments that received both wastes produced higher yields of biogas; however, they obtained lower methane production of 130 ± 20 mL g VS−1 than that found in the current experiment. The co-digestion process of PM (PM and TPM) and AWS under dry condition (TS 20 %) resulted in great enhancement of methane production. Results of ammonia accumulation (5.35–12.28 g N kg−1 bed) indicated that co-digestion of PM with AWS increased ammonia tolerance of methanogenesis and allowed methane to be successfully produced.

Conclusion

Dry co-digestion of PM or TPM and AWS (TS% of 20) performed better than individual digestion because of lower ammonia accumulation and better degradation of VFA. Mixtures of PM or TPM and AWS in co-digestion presented higher synergetic effect than digestion of single manure. Comparing thermophilic and mesophilic co-digestion conditions, better performance was achieved under the mesophilic condition. Optimum methane production of 562 mL g VS−1 was obtained under the mesophilic, and this amount is the highest volume achieved so far. Ammonia accumulation is reduced by 43.7 % during the test experiments compare to the control.