Bioresource Technology 156 (2014) 307–313

Contents lists available at ScienceDirect

Bioresource Technology

journal homepage: www.elsevier .com/locate /bior tech

Semi-continuous anaerobic co-digestion of dairy manure with three cropresidues for biogas production

http://dx.doi.org/10.1016/j.biortech.2014.01.0640960-8524/� 2014 Elsevier Ltd. All rights reserved.

⇑ Corresponding author. Address: Biogas Scientific Research Institute of theMinistry of Agriculture, No. 13, 4th Section, South Renmin Rd., Chengdu, Sichuan,China. Tel.: +86 28 85230701.

E-mail address: biomagongcheng2013@hotmail.com (G. Zhang).

Jiang Li, Luoyu Wei, Qiwu Duan, Guoquan Hu, Guozhi Zhang ⇑Biogas Scientific Research Institute of the Ministry of Agriculture, Chengdu 610041, China

h i g h l i g h t s

� Semi-continuous co-digestion of wastes under five mass ratios can be operated stably.� High biogas yields are achievable in mass ratio 5:5.� Four periods were formed for the digestion.� The N, P, S, Fe, Co and Ni improved gas production and kept the stability of AD.

a r t i c l e i n f o

Article history:Received 19 November 2013Received in revised form 13 January 2014Accepted 15 January 2014Available online 25 January 2014

Keywords:Semi-continuous anaerobic co-digestionCrop straw residueDairy manureMass mixing ratioBiogas

a b s t r a c t

The characteristics of anaerobic semi-continuous co-digestion of dairy manure (DM) with three cropstraw residues (SRs), rice straw, corn stalks and wheat straw under five mass mixing ratios (SRs/DM)were investigated. During the anaerobic digestion (AD) process, four periods were identified: startup, firststage of stabilization, second stage of stabilization, and suppression. Following the four periods, thebiogas production rate varied between 101 and 576 mL L�1 d�1. A high CH4 content and volatile solidreduction was maintained at the SRs/DM mass mixing ratio 1:9. The highest cumulative biogas produc-tion of more than 19 L was obtained at ratio 5:5. However, ratio 9:1 performed worst in the whole pro-cess. Systematic analysis of the elements revealed nitrogen, phosphorus, and trace elements contentswere important for the AD. Overall, the semi-continuous AD is efficient within a wide range of SRs/DMmass mixing ratios.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Anaerobic digestion (AD) is a biological process that producesbiogas from bio-degradable wastes by microorganism under pooror no oxygen conditions. AD is gaining more attention, not onlyas a solution to environmental concerns, but also as a potentialenergy resource for today’s energy-demanding life style. China isone of the largest agricultural countries, which produces over600 million ton of crop straw residues (SRs) every year, rankingfirst in the world (MOA, 2011). Rice straw (RS), corn stalks (CS)and wheat straw (WS) are the top three crop straw wastes in Chinaand account for 32.3%, 25.0% and 18.3% of the total crop straw out-put, respectively (MOA, 2011). Thus, making use of these wastesfor biogas generation can be quite significant. However, cropwastes cannot be effectively degraded due to an imbalance in

nutrients for microorganism and a lack of buffering capacity forthe chemical reaction (Babaee et al., 2013). This can often be over-come by co-digestion with livestock manure. Annual yield of live-stock manure in China is over 2.1 billion ton, including dairymanure (DM), swine manure, sheep manure and chicken manure(Zhang, 2010). The use of these wastes is a major component forproducing renewable energy, and it is suitable for narrowing thegap between the energy requirement of the industrialized worldand inability to replenish such needs from the limited sources ofenergy like fossil fuels.

Studies have been done to improve biogas production fromco-digestion of livestock manure and other wastes by AD (Astalset al., 2013; Chen et al., 2013; Saidu et al., 2013). Compared withthe digestion of single feedstock, co-digestion increases the biogasproduction rate because of the better nutrient balance andimprovement of AD efficiency. Generally, livestock manure con-tains a high total nitrogen (TN), which decreases the carbon-to-nitrogen (C/N) ratios of single SRs substrates and is beneficial toco-digestion with SRs. Livestock manure is also helpful to achievea suitable pH during anaerobic fermentation with the production

308 J. Li et al. / Bioresource Technology 156 (2014) 307–313

of ammonia (Ashekuzzaman and Poulsen, 2011). Hence, livestockmanure is excellent raw material for anaerobic co-digestion withSRs. As one of the main livestock manure in China, DM has buffer-ing capacity and excess nitrogen nutrients that can supportadditional carbon conversion of straws to methane gas. In fact,co-digestion, which utilizes various raw materials, such as agricul-tural waste, animal manure, sewage sludge and food waste, hasbeen extensively applied as an effective waste management andenergy production treatment (Ashekuzzaman and Poulsen, 2011;Curcio et al., 2010; De Vrieze et al., 2013; Larsen et al., 2013). How-ever, these studies have mostly been carried out in batch fermen-tation, and the suitable mixing ratios of multi-componentsubstrates such as SRs and DM are largely unknown. Most previousstudies focused only on the effect of C/N ratio in the AD, omittingthe analyses of the roles of other important elements, such asphosphorus (P), potassium (K), sulfur (S), iron (Fe), cobalt (Co),and nickel (Ni) in the degradation of substrates (De Vrieze et al.,2013; Podmirseg et al., 2013; Rajagopal et al., 2013; Wang et al.,2013). For example, methanogens need Fe, Co, and Ni to makethe methane production feasible (Gustavsson et al., 2011; Uemura,2010; Zhang and Jahng, 2012; Zitomer et al., 2008).

This study investigated how the different SRs and DM massmixing ratios affected the biogas-producing efficiency of anaerobicco-digestion in a semi-continuous style. The main strategy was todetermine the optimal ratio of SRs and DM, characterize the semi-continuous co-digestion of them, and evaluate the effect of theeight elements (C, N, P, K, S, Fe, Co and Ni) of the substrate onthe biogas production.

2. Methods

2.1. Collection and preparation of substrates

SRs and DM were obtained from a local farm in ShuangliuCounty, Chengdu, Sichuan, China. DM had a total solid (TS) of18.4%, and was stored at 4 �C. RS, CS and WS were prepared by cut-ting the residues into sections of 2–3 cm by using a grinder, with aTS of 90.2%, 89.6% and 90.0%, respectively. Inoculum was theanaerobic sludge, was obtained from an anaerobic digester of asewage plant in Chengdu City, and had a TS of 11.2%, and was alsostored at 4 �C. For preservation of more than one week, DM andsludge were stored at �17 �C.

2.2. Experimental design and set-up

The experiment was conducted by using lab-scale anaerobicdigesters fabricated from 1 L polycarbonate cups with whorl coverfor discharging and feeding flexibly. The cup had nozzle on the cov-er sealed with rubber stopper, where glass tube were inserted forgas flowing to the collection bottle full of water, forcing the waterto be pressed out. Gas volume was recorded by measuring the vol-ume of the drain. In this work, semi-continuous fermentation wasused to determine the co-digestion of DM mixed with the threetypes of SRs. The working volume of each digester was 800 mL,including 91.69 g inoculum and an appropriate mass ratio of SRsand DM. To obtain the best mixing ratio of the co-digestion ofDM and the three SRs, five different mass mixing ratios at 1:9,3:7, 5:5, 7:3 and 9:1 were tested under mesophilic conditions(35 �C) for 47 days. Tap water was added to digesters to maintaina TS content of 8.0%. After 7 days’ startup, anaerobic co-digestionwas then initiated in a semi-continuous style. All reactors weregently mixed manually for approximately 1 min prior to discharg-ing and feeding, with a conservative organic loading rate (OLR) of3.2 g L�1 every two days, as difficulties were experienced in achiev-ing steady state performance at an OLR of 1.6 g L�1 every day. Each

treatment was performed in triple replicate to investigate theeffect of different mixing ratios on biogas production.

2.3. Analysis and statistics

The amount of biogas produced from each digester was recordedevery day by using the water displacement method during thedigestion period. Biogas composition (CH4 and CO2 contents) wasmeasured using biogas 5000 (Geotech Inc., China). pH was mea-sured using an acidimeter (PHS-3C, DAPU, China). Chemical analy-sis tests were performed on the substrate of each reactor after theexperiment was completed (the same batch of DM, SRs and sludgewas used throughout testing). The TS, volatile solid (VS) were ana-lyzed by Hach Method 8271 and 8276. Total carbon (TC), TN, totalphosphorus (TP), and total potassium (TK) were determined inaccordance with the standard methods for the examination of or-ganic fertilizer of the Chinese agriculture industry standard (MOA,2012). Total iron (Tfe) was determined according to the microwavedigestion – atomic absorption spectrometry method (Xiao, 2006).Total sulfur (Tsu), total cobalt (Tco), total nickel (Tni) content ofthe samples were determined according to the microwave digestion– ICP-MS method (Wang et al., 2008). In the next step, principalcomponent analysis (PCA) was utilized for the determination ofthe effect of the eight elements. PCA was performed using theCANOCO for windows 4.5 software. Each semi-continuous experi-ment was deemed complete when a clearly stable trend in daily bio-gas volume produced was observed for at least 7 days. ANOVA wasperformed to determine the significant differences among eachtreatment by using SPSS version 17.0 (SPSS China Inc.).

3. Results and discussion

3.1. Biogas yields and production rates at different DM/SRs ratios

The daily biogas production by the co-digestion of SRs and DMduring 47 days was recorded under five mass mixing ratios (Fig. 1).From Fig. 1, it can be seen that, for all the five mass mixing ratios,the semi-continuous co-digestion of SRs and DM could be dividedinto four periods: startup, first stage of stabilization, second stageof stabilization and suppression. During the startup period, the bio-gas production increased rapidly, reaching about 400 mL d�1 onday 7, except for ratios of 7:3 and 9:1,which was only about100 mL d�1. It was suggested that recalcitrant polymers withinstraws limited their degradation, and the lower amounts of solublecarbohydrates in straws resulted in slow hydrolysis and fermenta-tion. Consequently, ratio 7:3 and 9:1 showed a low daily biogasproduction during the startup period.

After the feeding at OLR of 3.2 g L�1 every two days, the semi-continuous co-digestion of the five ratios went into the first stageof stabilization period, and the biogas production of ratio 1:9, 3:7,5:5, and 7:3 remained stable around 600 mL d�1 for 7 days, whichsuggested the destruction of substrates. However, after day 22, thegas production of the four ratios decreased to about 400 mL d�1 inthe second stage of stabilization. On day 32, further decrease wasobserved, with a gas production of about 200 mL d�1 for the fourratios, which indicated the suppression period came, and theexperiment was stopped on day 47. In fact, for all the SRs/DM,the mixing ratio 9:1 performed worst, especially in the startup per-iod, with a daily gas production of less than 100 mL d�1. In the firststage of stabilization period, daily gas production of SRs/DM 9:1 in-creased gradually, but still no more than 200 mL d�1. Interestingly,in the second stage of stabilization period, all the SRs/DM at ratio9:1 produced gas as much as other ratios, sometimes even more,and it remained in this state until suppression period. This couldbe explained that the substrate which was not used in the

Fig. 1. Daily biogas production from the semi-continuous co-digestion of SRs and DM under five mass mixing ratios. Each data point is the average of three independentreplications. Vertical bars represent standard deviations.

J. Li et al. / Bioresource Technology 156 (2014) 307–313 309

beginning was degraded later. Moreover, these results indicatedthat the co-digestion of SRs and DM could significantly delay theattainment of the highest daily gas production at the mass mixingratios of 9: 1.

In this work, feeding was carried every two day. It was observedthat, the daily biogas volume before feeding was always higherthan that after feeding, with a gap of more than 150 mL d�1 forall the five ratios, especially in the suppression period. This couldbe caused by imported fresh air during periodic feeding, whereoxygen might inhibit the activity of anaerobic microbiology. Fur-thermore, stirring during discharging and feeding might have hadan adverse effect on microbiology in digester with the destructionof localized pockets of high acetate concentration for the methaneproduction and the growth of methanogens (Sindall et al., 2013).

Among the three types of straw, WS/DM had the highest gasproduction on the first day, with more than 400 mL, which sug-gested it had intense microbial activity and rapid substrate decom-posing reaction at the beginning. Biogas from the RS/DM mixingratios of 1:9, 3:7, 5:5, and 7:3 reached their peak yield values at813, 788, 850 and 953 mL d�1 on day 18, 22, 22 and 18, respec-tively (Fig. 1, first stage of stabilization period). The digestion ofRS/DM 9:1 produced substantial biogas only after day 21 i.e., laterthan other RS/DM combinations, but, by the end, it had a peak ofabout the same size to other ratios (802 mL d�1 on day 32, Fig. 1second stage of stabilization period). This can also be attributedto a high amount of indigestible RS and slow decomposition. Com-pared to RS/DM, CS/DM had a relatively lower and later peak yield.WS/DM has its own daily biogas production characteristics, with apeak yield value range from 700 to 820 mL d�1 for the five ratios.But WS/DM 9:1 increased rapidly only after day 26 (Fig. 1). Liewet al. (2012) reported that WS contained more lignin, leading to amore recalcitrant structure than RS and CS, and thus limiting thedegradation of this lignocellulose.

The cumulative biogas productions by the co-digestion of SRsand DM at five mixing ratios are shown in Fig. 2. For all the mixingratios, CS/DM 5:5 had the highest biogas yield at 19,428 mL after47 days of digestion, but this was not significantly different fromWS/DM 5:5 (19,127 mL (p > 0.05)). In fact, in this work, ratio 5:5showed for the best results in the final cumulative biogas produc-

tion among the five ratios, no matter which SRs was co-digestedwith DM. However, CS/DM 9:1 had the lowest value at10,523 mL, nearly 46.2% lower than CS/DM 5:5. In addition, thecumulative biogas production for ratio 9:1 also was the worst inthe five ratios of all the SRs /DM. This was consistent with the dailybiogas production data (Fig. 1).

The biogas production rate of SRs and DM at five mass mixingratios in four periods are shown in Fig. 3. For RS/DM, ratio 1:9had the highest biogas production rate in the startup period, whichreached 236 mL L�1 d�1, and showed an increase of about 50.4%compared with other RS/DM ratios (101–123 mL L�1 d�1)(p < 0.05). This suggested that setting RS/DM mass mixing ratioat 1:9 could make the AD start smoothly. For CS/DM in the startupperiod, ratio 7:3 had the highest biogas production rate, suggestingthis ration improved the destruction of CS at the beginning of theAD. And for WS/DM, there was almost no difference among the fiveratios in the startup period. In the first stage of stabilization, for RS/DM, ratio 1:9 kept stay on the top, but ratio 3:7, 5:5, and 7:3 alsoreached a relatively high biogas production rate, which was 454,520 and 453 mL L�1 d�1, respectively, leaving RS/DM 9:1 in thebottom, which was drastically lower, averaging approximately121 mL L�1 d�1. These results verified slower biogas productionrate was accompanied by more straws in the substrates. CS/DMand WS/DM had a similar trend with RS/DM in the first stage ofstabilization, except for WS/DM 7:3, averaging 32.3% less thanWS/DM 1:9, 3:7, and 5:5, but 44.6% more than WS/DM 9:1, asshowed in Fig. 3 (p < 0.05). In the second stage, stabilization period,RS/DM 5:5 and 9:1 produced biogas faster than the other ratios,with rates of about 525 mL L�1 d�1. In the suppression period, ratio7:3 reached the highest biogas production rate in the five ratios ofSRs/DM, while ratio 1:9 reached the lowest (220–230 mL L�1 d�1),which indicated ratio 7:3 had significant substrate degradation,even there was some suppression for other ratios.

In the whole process, the biogas production rate reached high-est at different mass mixing ratios for different periods. But in thefirst stage of stabilization, the biogas production rates of all the ra-tios were always higher than that in the other periods, except forSRs /DM 9:1. According to the results from Fig. 2 and Fig. 3, themass mixing ratio 9:1 was not feasible for the semi-continuous

Fig. 2. Cumulative biogas production from the semi-continuous co-digestion of SRs and DM under five mass mixing ratios. Each data point is the average of threeindependent replications.

Fig. 3. Biogas production rate in four periods (startup, first stage of stabilization,second stage of stabilization and suppression) from the semi-continuous co-digestion of SRs and DM under five mass mixing ratios. Each data point is theaverage of three independent replications.

310 J. Li et al. / Bioresource Technology 156 (2014) 307–313

co-digestion of SRs and DM. Possible reasons for the low carbon tobiogas efficiency in the SRs/DM 9:1 reactors include the following:(1) Imbalance of C/N in the mass mixing ratio 9:1of SRs/DM; (2)Lack of trace nutrients through the presence of the ratio; (3) Organ-ic overloading, resulting in a reduction in the methanogenic activ-ity; and (4) High levels of carbohydrates in the straws thatconverted carbon to excessive acid and occurred before methano-genesis (Mata-Alvarez et al., 2000).

3.2. pH, biogas composition and organic solids degradation

pH is one of the key factors in AD, and the growth of methano-gens can be significantly influenced by the pH level (Rajagopalet al., 2013). The pH value reflected the changing processes inthe digesters with similar trends in all the mixtures during the47 days of semi-continuous co-digestion (Fig. 4). On day 2, thepH values decreased from 6.80 to 5.25 with SRs percentageincreasing in the five mixing ratios, and WS/DM 9:1 had the lowestpH value (5.25). But from day 13, the pH values increased in all themixtures, and then remained at approximately 7.00 until the endof the experiment. This stability confirmed that the biogas produc-tion of each mixture reached the methanogenesis stage. It furtherindicated that excess acid was not formed in the suppression per-iod, as moderate pH was recorded. However, the pH of SRs/DM 9:1valued from 5.50 to 5.25 before day 23, showing the bufferingcapacity of DM did not work, as SRs composed 90% of the total sub-strate mass. According to the results from both daily and cumula-tive biogas production, it could be concluded that the optimal pHvalues for the co-digestion of SRs and DM ranged from 6.80 to7.20, as ratio 5:5 showed. These results are consistent with Aboue-lenien et al. (2010) who found that with an ammonia-strippingunit (here similar to SRs), methane was successfully produced fromthe treated chicken manure at a pH of approximately 6.70.

The methane content of the biogas from all the SRs/DM ratiosvaried among the four periods, from 29.3% to 40.1% for the startuptime, 50.4–64.5% for the first and second stage, and 50.1–30.2% forthe suppression, which showed a relatively big deviation over theentire digestion (Fig. 4). This was consistent with the result fromdaily biogas production in Fig. 1. In the first 5 days, the methanecontent had a faster increase only for ratio 1:9 of RS/DM andWS/DM, from 40.6% to 55.3%. This increase could be observed inall the ratios of CS/DM, indicating that CS reached the methanogen-esis stage earliest among the three SRs. The methane content ofother four ratios of RS/DM and WS/DM were below 40% in the first5 days, some even had a value of 29.3%. Reason for this might bethat, compared to RS and WS, CS had more easily degradable or-ganic matter, which boosted the activity of the microbiology in

Fig. 4. Methane content (line and symbol) and pH values (symbol) from the semi-continuous co-digestion of SRs and DM under five mass mixing ratios. Each data point is theaverage of three independent replications.

J. Li et al. / Bioresource Technology 156 (2014) 307–313 311

the startup time (Liew et al., 2012). When the easily degraded mat-ter was used up, CS/DM got a similarly methane content as RS/DMand WS/DM, as Fig. 4 showed. However, unlike their methane con-tent, the difference for the three SRs was not evident in the biogasproduction rate or daily biogas production in the startup time.After day 35, the methane content decreased dramatically for allthe treatments, especially for WS/DM 1:9, which was only 30.7%on day 36. As pH was normal, this could be the result of inhibitorsgenerated with the digestion going on, thus affecting the growth ofmethanogens during the AD process (Chen et al., 2008; Madsenet al., 2011).

In order to determine the amount of co-substrate that had beendegraded in the co-digestion experiments, VS and TS analysis of allthe ratios were carried out at the end of the digestion. The calcu-lated reductions for each ratio are reported in Fig. 5A and B. TheANOVA results indicated that just as the VS reduction for differentratios of RS/DM, the population means of TS reduction for differentratios of RS/DM and WS/DM were not significantly different, rang-ing from 38.4% to 45.2%. For the VS reduction of WS/DM, ratio 1:9was 50.0% more than that of WS/DM 9:1(p < 0.05). As it is known,TS of lignocellulose is mainly composed of VS and ash (Soest et al.,1991). And the results indicated that there might be more ash con-tent in the WS, causing the TS reduction trend to disagree with thatof the VS. Unlike RS/DM and WS/DM, the TS reduction of CS/DM1:9 was significantly different from that of CS/DM 7:3, with an in-crease of 40.2% (p < 0.05). And for VS reduction, CS/DM 1:9 was sig-nificantly different from that of 7:3 and 9:1, increasing 35.0% and35.5%, respectively (p < 0.05). In sum, SRs/DM 1:9 performed bestin the organic solids destruction, and rather than those of RS/DMand WS/DM, different ratios of CS/DM led to different substratereduction results.

The ANOVA also showed that the means of three straws werenot significantly different from each other in the VS and TS reduc-tions. However, Fig. 5 C differed as the biogas productivity of WS/DM 3:7 reached 198 mL g�1 VS, 9.6% and 16.7% more than that ofRS/DM 3:7 and CS/DM 3:7, respectively (p < 0.05). Compared toRS/DM 5:5, the biogas productivity of WS/DM 5:5 was better,reaching 209 mL g�1 VS, which also was the best of all the SRs/DM ratios (p < 0.05). These results disagreed with those of Wuet al. (2010), who found that WS demonstrated a lower biogas pro-ductivity than CS and oat straw even it had a higher carbon content

than the latter two residues. But for the biogas production from VSmass in this work, WS had a better performance than RS and CS.

3.3. Elements analysis using a PCA

Generally speaking, the C/N ratios of different substrates mix-tures in AD greatly influence biogas production (Kayhanian,1999; Wang et al., 2012). Results of the element analysis showedthat the feedstock of SRs/DM in the five mixing ratios had a C/N ra-tio of about 7.8, 10.7, 15, 21 and 33, respectively. However, withthe increasing of the C/N, there was no obvious trend for the biogasproduction change in the five mixing ratios. And with the semi-continuous co-digestion’s going on, the biogas production cameto a stop.

To better understand the processes in play during the biogasproduction, this work tested not only C and N contents, but alsoP, K, S, Fe, Co, and Ni in all the feedstocks and effluent samples atthe end of the digestion. And PCA was used to determine theimportance of each element.

In this study, PCA showed a smaller partial correlation(KMO = 0.931) and a high dependence (P < 0.001) (Fig. 6). In gen-eral, the KMO value is higher than 0.9 and the P value is lower than0.001 can be viewed as very suitable for PCA. The PCA of all thesamples yielded two PC factors (PC1 and PC2) that accounted for84.69% of the variance. The plots of PCA factor scores vs. elementsprovided an indication of the element types of each PC and there-fore identified the elements that contributed to the variance in thespecific SRs/DM ratio, which may explain the biogas productiondifferences (Fig. 6). Eight elements could be identified by PCA.PC1 (64.51% of the variance) was mainly related with TN, TP, Tsu,Tfe, Tco and Tni, which were indicative of ammonia, phosphateand trace element. PC2 (20.18% of the variance) was associatedwith a primary element, carbon, which was indicative of carbohy-drate, including hemicellulose (five carbon polymers), cellulose(six carbon polymers) and lignin (phenol polymers), which weremainly in straws (Bauer et al., 2009). TK had comparative scoresin both of the two PC plots.

In Fig. 6, five regions could be drawn. Region I, including threeeffluent samples (outCS5:5, outWS5:5, and outWS1:9), clusteredwith higher second (>0.6) and moderate first (<0.4) PCA loading,indicated that element C and K were dominant in the three effluent

Fig. 5. Substrate utilization of the semi-continuous co-digestion of SRs and DM under five mass mixing ratios: (A) VS reduction (B) TS reduction (C) Biogas production. Eachdata point is the average of three independent replications. Vertical bars represent standard deviations. The ANOVA test was conducted to determine the differences betweenratios. Values with the same letters indicate no significant difference at p < 0.05.

Fig. 6. PCA of the semi-continuous co-digestion of SRs and DM under five massmixing ratios. Sample inCS, inRS and inWS stand for the feedstock of co-digestion ofSRs and DM. Sample outCS, outRS and outWS stand for the effluent of co-digestionof SRs and DM at the end of the digestion.

312 J. Li et al. / Bioresource Technology 156 (2014) 307–313

samples. As Fig. 5C showed, WS/DM 5:5 obtained the highest bio-gas productivity per gram of VS. PCA here may give reason for this,that was, compared with other ‘‘out’’ samples, outWS5:5 had thehighest TC and TK content, and contained moderate TN, TP, and

trace element amount, which was suitable for methanogenesis.However, biogas yield of WS5:5 was not significant different fromthat of CS/DM 5:5. This indicated that good biogas production alsoneeded suitable elements content as SRs/DM 5:5, not only focusingon the C/N ratio. Region III, including outCS7:3 and outWS3:7, ob-tained moderate PC2 (>0.1) and lower PC1 (<�0.1) loading, whichsuggested trace elements were lacking in the three samples. Thatmaybe why CS/DM 7:3 had the worst result for TS and VS removal,while CS/DM 1:9, RS/DM 3:7and WS/DM 1:9 had the best (outCS1:9 and outRS 3:7 gathered together in Fig. 6 Region II with PC 1loading >0.2). Similarly, region IV and V were also lack of trace ele-ments. What’s worse, region V had the least content of the eightelements (PC 1 loading <�0.15, PC 2 loading <�0.4). As a result,mixtures in this region (WS/DM 9:1 and CS/DM 9:1) had a bad per-formance in the gas production. Furthermore, the similarity of PCAloading of CS/DM 9:1, WS/DM 9:1, RS/DM 7:3and WS/DM 7:3 inFig. 6 region V was consistent with the pH trend of them in theco-digestion as shown in Fig. 4. Ratios 7:3 and 9:1 had more strawmass, which meant more carbon and potassium contents. This wasseemingly contradictory with the PCA loading in region V, and fur-ther investigation will be needed to better understand the results.Overall, from PCA and biogas production results, it showed thatcompared with region III, IV, and V, region I and II had more N, P,and trace elements contents, and got better AD performance. Itwas known that methanogens need trace elements, especially Fe,Co, Ni, to initiate methane production (Gustavsson et al., 2011;Uemura, 2010; Zhang and Jahng, 2012; Zitomer et al., 2008). As aresult, ratios in region I and II with more Fe, Co, Ni and appropriateelement composition obtained higher biogas production rate andyield than the others.

PCA also showed that with ratios changing from 1:9 to 9:1, i.e.straws became more and more in the feedstock, the PC1 loading of

J. Li et al. / Bioresource Technology 156 (2014) 307–313 313

the feedstock samples became lower and lower, but the PC2 load-ing varied little, indicating it were N, P, and trace elements in PC1that determined the differences of the element composition of thefeedstocks. In other words, if there was much more straw in thesubstrate, there was less N, P, and trace elements, which leadedto less biogas yield as this work showed. In general, these PCA re-sults suggested that co-digestion of SRs and DM with suitable N, P,and trace elements content as ratio 5:5 was an effective way toprolong the period of the highest gas production and improve bio-gas yield.

4. Conclusion

This study examined the mechanisms underlying the biogasproduction of anaerobic semi-continuous co-digestion using threetypes of straws and five mixing ratios. First, it was found the ADcould be divided into four periods, with the highest biogas produc-tion rate occurring in the first stage of stabilization, except for SRs/DM 9:1. CS/DM 5:5 showed the highest final cumulative biogasproduction. SRs/DM 1:9 increased fast in methane content. FromPCA, it was suggested good biogas production needed suitable N,P, and trace elements contents. Overall, except SRs/DM 9:1, theother four ratios had a high potential for anaerobic semi-continu-ous co-digestion.

Acknowledgements

The authors gratefully thank Dr. Liette Vasseur (Minjiang scho-lar at Fujian Agriculture and Forestry University) for the Englishscientific editing. The authors would like to acknowledge financialsupport from the National Science & Technology Pillar Programduring the Twelfth Five-year Plan Period (2011BAD15B03).

References

Abouelenien, F., Fujiwara, W., Namba, Y., Kosseva, M., Nishio, N., Nakashimada, Y.,2010. Improved methane fermentation of chicken manure via ammoniaremoval by biogas recycle. Bioresour. Technol. 101, 6368–6373.

Ashekuzzaman, S.M., Poulsen, T.G., 2011. Optimizing feed composition forimproved methane yield during anaerobic digestion of cow manure basedwaste mixtures. Bioresour. Technol. 102, 2213–2218.

Astals, S., Nolla-Ardevol, V., Mata-Alvarez, J., 2013. Thermophilic co-digestion of pigmanure and crude glycerol: process performance and digestate stability. J.Biotechnol. 166, 97–104.

Babaee, A., Shayegan, J., Roshani, A., 2013. Anaerobic slurry co-digestion of poultrymanure and straw: effect of organic loading and temperature. J. Environ. HealthSci. Eng. 11, 15.

Bauer, A., Bosch, P., Friedl, A., Amon, T., 2009. Analysis of methane potentials ofsteam-exploded wheat straw and estimation of energy yields of combinedethanol and methane production. J. Biotechnol. 142, 50–55.

Chen, J.H., Lin, C.C., Wang, K.S., 2013. Potential of methane production bythermophilic anaerobic co-digestion of pulp and paper sludge with pigmanure. J. Biobased Mater. Bioenergy 7, 300–304.

Chen, Y., Cheng, J.J., Creamer, K.S., 2008. Inhibition of anaerobic digestion process: areview. Bioresour. Technol. 99, 4044–4064.

Curcio, S., Calabro, V., Aversa, M., Ricca, E., Sansonetti, S., Iorio, G., 2010.Optimization of biogas production with bioconversion of organic solid wastes(manure) and food industry wastes. J. Biotechnol. 150, S165–S165.

De Vrieze, J., De Lathouwer, L., Verstraete, W., Boon, N., 2013. High-rate iron-richactivated sludge as stabilizing agent for the anaerobic digestion of kitchenwaste. Water Res. 47, 3732–3741.

Gustavsson, J., Svensson, B.H., Karlsson, A., 2011. The feasibility of trace elementsupplementation for stable operation of wheat stillage-fed biogas tank reactors.Water Sci. Technol. 64, 320–325.

Kayhanian, M., 1999. Ammonia inhibition in high-solids biogasification: anoverview and practical solutions. Environ. Technol. 20, 355–365.

Larsen, A.C., Gomes, B.M., Gomes, S.D., Zenatti, D.C., Torres, D.G.B., 2013. Anaerobicco-digestion of crude glycerin and starch industry effluent. Eng. Agric. 33, 341–352.

Liew, L.N., Shi, J., Li, Y.B., 2012. Methane production from solid-state anaerobicdigestion of lignocellulosic biomass. Biomass Bioenergy 46, 125–132.

Madsen, M., Holm-Nielsen, J.B., Esbensen, K.H., 2011. Monitoring of anaerobicdigestion processes: a review perspective. Renew. Sustainable Energy Rev. 15,3141–3155.

Mata-Alvarez, J., Mace, S., Llabres, P., 2000. Anaerobic digestion of organic solidwastes. An overview of research achievements and perspectives. Bioresour.Technol. 74, 3–16.

MOA, 2012. Chinese Agriculture Industry Standard NY 525–2012. Ministry ofAgriculture of China, Beijing.

MOA, 2011. National crop straw resources survey and evaluation reports. Agric. Eng.Technol. (Renew. Energy Ind.) 02, 2–5.

Podmirseg, S.M., Seewald, M.S.A., Knapp, B.A., Bouzid, O., Biderre-Petit, C., Peyret, P.,Insam, H., 2013. Wood ash amendment to biogas reactors as an alternative tolandfilling? A preliminary study on changes in process chemistry and biology.Waste Manage. Res. 31, 829–842.

Rajagopal, R., Masse, D.I., Singh, G., 2013. A critical review on inhibition of anaerobicdigestion process by excess ammonia. Bioresour. Technol. 143, 632–641.

Saidu, M., Yuzir, A., Salim, M.R., Salmiati, Azman, S., Abdullah, N., 2013. Influence ofpalm oil mill effluent as inoculum on anaerobic digestion of cattle manure forbiogas production. Bioresour. Technol. 141, 174–176.

Sindall, R., Bridgeman, J., Carliell-Marquet, C., 2013. Velocity gradient as a tool tocharacterise the link between mixing and biogas production in anaerobic wastedigesters. Water Sci. Technol. 67, 2800–2806.

Soest, P.J.v., Robertson, J.B., Lewis, B.A., Bridgeman, J., 1991. Methods of dietary fiber,neutral detergent fiber, and nonstarch polysaccharides in relation to animalnutrition. J. Dairy Sci. 74, 3583–3597.

Uemura, S., 2010. Mineral requirements for mesophilic and thermophilic anaerobicdigestion of organic solid waste. Int. J. Environ. Res. 4, 33–40.

Wang, Q.F., Peng, L.Y., Su, H.J., 2013. The effect of a buffer function on the semi-continuous anaerobic digestion. Bioresour. Technol. 139, 43–49.

Wang, X., Yang, G., Feng, Y., Ren, G., Han, X., 2012. Optimizing feeding compositionand carbon–nitrogen ratios for improved methane yield during anaerobic co-digestion of dairy, chicken manure and wheat straw. Bioresour. Technol. 120,78–83.

Wang, X., Zhang, Y., Liu, N., Wang, L., 2008. Determination of heavy metal ions insoil by ICP-MS with microwave digestion. Chin. J. Spectrosc. Lab. 25, 1183–1187.

Wu, X., Yao, W.Y., Zhu, J., Miller, C., 2010. Biogas and ch4 productivity by co-digesting swine manure with three crop residues as an external carbon source.Bioresour. Technol. 101, 4042–4047.

Xiao, G., 2006. Determination of microelements in tea and its cultivating soil bymicrowave digestion-atomic absorption spectrometry. Chin. J. Spectrosc. Lab.23, 493–496.

Zhang, L., Jahng, D., 2012. Long-term anaerobic digestion of food waste stabilized bytrace elements. Waste Manage. (Oxford) 32, 1509–1515.

Zhang, M., 2010. Recycling models and technologies for manure. Mod. Agric. Sci.Technol. 14, 280–283.

Zitomer, D.H., Johnson, C.C., Speece, R.E., 2008. Metal stimulation and municipaldigester thermophilic/mesophilic activity. J. Environ. Eng.-ASCE 134, 42–47.