Bioresource Technology 182 (2015) 239–244
Contents lists available at ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
Multi-phased anaerobic baffled reactor treating food waste A. Ahamed a,1, C.-L. Chen a, R. Rajagopal a, D. Wu a, Y. Mao a, I.J.R. Ho a, J.W. Lim a, J.-Y. Wang a,b,⇑ a Residues and Resource Reclamation Centre, Nanyang Environment and Water Research Institute, Nanyang Technological University, #06-08 CleanTech One, 1 Cleantech Loop, Singapore 637141, Singapore b School of Civil and Environmental Engineering, Nanyang Technological University, 50 Nanyang Avenue, Singapore 639798, Singapore
h i g h l i g h t s Study on isolation of different phases of an anaerobic digestion process. First reference for biogas yield from food waste using anaerobic baffled reactor. Effluent quality shows this as an efficient treatment for food waste.
a r t i c l e
i n f o
Article history: Received 12 November 2014 Received in revised form 26 January 2015 Accepted 28 January 2015 Available online 4 February 2015 Keywords: Food waste Anaerobic baffled reactor Phase separation Biogas Anaerobic digestion
a b s t r a c t This study was conducted to identify the performance of a multi-phased anaerobic baffled reactor (MP-ABR) with food waste (FW) as the substrate for biogas production and thereby to promote an efficient energy recovery and treatment method for the wastes with high organic solid content through phase separation. A four-chambered ABR was operated at an HRT of 30 days with an OLR of 0.5–1.0 g-VS/L d for a period of 175 days at 35 ± 1 °C. Consistent overall removal efficiencies of 85.3% (CODt), 94.5% (CODs), 89.6% (VFA) and 86.4% (VS) were observed throughout the experiment displaying a great potential to treat FW. Biogas generated was 215.57 mL/g-VSremoved d. Phase separation was observed and supported by the COD and VFA trends, and an efficient recovery of bioenergy from FW was achieved. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Approximately one third of the food produced in the world for human consumption every year, which is approximately 1.3 billion tonnes, gets wasted globally. The wastage of food can be attributed into loss of resources in food production, and the unnecessary production of greenhouse gas emissions (FAO, 2013). In a world with depleting energy resources, the amount of food waste (FW) generated everyday needs an utmost use to move towards sustainable development. FW, rich in organic acids, constitutes an ideal source for bioenergy recovery. On the other hand, FW disposal has been a
⇑ Corresponding author at: Residues and Resource Reclamation Centre, Nanyang Environment and Water Research Institute, Nanyang Technological University, #0608 CleanTech One, 1 Cleantech Loop, Singapore 637141, Singapore. Tel.: +65 67904100; fax: +65 67927319. E-mail addresses: [email protected], [email protected] (A. Ahamed), jywang@ ntu.edu.sg (J.-Y. Wang). 1 Tel.: +65 67904102; fax: +65 67927319. http://dx.doi.org/10.1016/j.biortech.2015.01.117 0960-8524/Ó 2015 Elsevier Ltd. All rights reserved.
growing problem in highly populated cities such as Singapore comprising 21.1% of the municipal solid waste that is disposed (NEA, 2013). Wang et al. (2003) reported that FW has been main source of odour and leachate in collection and transportation of municipal solid waste due to their high volatile components and moisture content, and has been causing significant environmental impact. Anaerobic digestion (AD), a biological conversion of organic matter into methane, carbon dioxide, inorganic nutrients and humus-like matter, appears to be the most promising method for FW treatment (Liu et al., 2008). As compared to the aerobic method, the use of anaerobic processes for the waste streams provides greater economic and environmental benefits and advantages (Mohan and Bindhu, 2008). AD technology is a well known method for waste utilization, and various configurations of the reactor type have been developed thus far. The most prominent and simple design is the single stage process and it has been widely used in various applications. However, its low efficiency has been highlighted (Ke et al., 2005). Besides single stage system, two-phase reactors have gained attraction and relevant research works has been reviewed by Ke et al. (2005). The two-phase
240
A. Ahamed et al. / Bioresource Technology 182 (2015) 239–244
system attempts to separate the acidogenic and methanogenic phases of AD process. The goal is to improve the performance of the reactor by allowing enough buffering capacity and space for respective microbial communities to flourish under suitable conditions. Adding on to this, a two-phase hybrid anaerobic solid–liquid system (HASL) by Wang et al. (2002, 2003) and Liu et al. (2008), was studied in detail for FW digestion and has shown promising results. Apart from these reactor types, an up-flow anaerobic sludge blanket process elaborately reviewed by Chong et al. (2012) is extensively applied for wastewater treatment with relatively high removal efficiency at short hydraulic retention times (HRTs), but is also considered not suitable for wastes with high solid content such as municipal FW (Tawfik et al., 2011b). Improving the soluble component removal efficiency, longer biomass retention, enhanced bacterial activity and high rate of contact between the cells and the substrate are important attributes for an efficient anaerobic digestion system (Grobicki and Stuckey, 1991). The multi-phased anaerobic baffled reactor (MP-ABR) possess a great potential in improving the afore mentioned criterion, and also provides better resilience to hydraulic and organic shock loadings (Barber and Stuckey, 1999). The reactor design allows adequate space for the microbial communities to sustain without interferences. The significant advantage of the reactor type is its ability to separate different phases of the AD process longitudinally, allowing the reactor to behave as a ‘phase separated’ system, which may offer the best choice for high efficiency removal rates (Mohan and Bindhu, 2008). FW, which possesses about 75% moisture content and 24% volatile solids (VS) (97% VS/total solids (TS) ratio), qualifies as a perfect substrate type for this type of reactor design due to its high solids retention capacity. Although extensive studies have been conducted on anaerobic baffled reactor (ABR) as mentioned in the review by Barber and Stuckey (1999), little is known about biogas production from FW, except for the work of Tawfik et al. (2011a,b), using a laboratory scale five-chambered ABR for biological hydrogen production from diluted kitchen waste. The main objective of this study was to assess the performance of the MP-ABR system treating FW for biogas production, to isolate the different phases of the AD process and to promote an efficient energy recovery and treatment method for the wastes with high organic content. Various performance indicators including chemical oxygen demand (COD), volatile fatty
acid (VFA), TS and VS were also investigated to evaluate the efficiency of reactor performance. 2. Methods 2.1. Feed and inocula characteristics FW was collected from a canteen inside the university campus, which mainly composed of rice, noodles, meat, vegetables and condiments. After removing the bones and inorganic materials, the waste was homogenized using a kitchen blender to disintegrate the particulate organics into size < 2 mm. The feed was diluted with tap water according to the influent organic loading rate (OLR) and flow rate before being fed to the reactor. The characteristics of the feed were pH (5.36 ± 0.82); total COD (CODt) (17.31 ± 5.05 g/L); soluble COD (CODs) (4.98 ± 2.03 g/L); VS (14.78 g/L); total VFA (VFA) (1.214 g-COD/L) on an average. Inoculum was obtained from a local wastewater treatment plant (Ulu Pandan Wastewater Treatment Plant) that uses AD technology for sewage sludge treatment. The reactor was filled up with 70% of the seed sludge and 30% with tap water. The seed sludge contained TS of 1.77 g/kg, of which VS was 70%. The pH and alkalinity were about 6.9 and 1916 mg-CaCO3/L respectively. 2.2. MP-ABR system A rectangular acrylic container with internal vertical baffles hanging and standing alternately embodied the basic design of the reactor (Fig. 1). To allow the number of compartments and width of each to be adjustable, the baffles were designed to be movable. The design of the reactor was slightly different from the conventional ABR. As the FW contain very high solid content, the widths of the downflow segment were increased for this study to allow the solids to flow over smoothly without any blockage. The reactor was operated in two operational periods with a modification to the reactor configuration in the second period. A four compartment (N1, N2, N3 and N4) reactor model was selected to isolate the different phases of the AD process, which is also considered to perform better (Boopathy, 1998). The individual working volumes were 10.7 L each for the first three and 21.4 L for the last compartment during the first period of operation. Subsequently, the volumes of N3 and N4 were adjusted to be 16.05 L each in
Biogas collection
Effluent
Feed-in
N4
N3
N2
N1
Fig. 1. Flow scheme of the MP-ABR with the direction of sludge flow.
241
A. Ahamed et al. / Bioresource Technology 182 (2015) 239–244
the second period. Each compartment was provided with an overhead mixer rotating at a speed of 100–150 rpm to minimize dead zones and short circuiting. The sampling points were at the mid-level of each compartment and the draining ports were at the bottom level. The reactor was connected to a gas flow meter and data logger to measure the gas volume. The gas composition was measured by subsequently collecting the gas in a gas bag and followed by gas chromatograph (GC) analysis. 2.3. Operating strategy The reactor was fed with FW at an OLR of 0.5–1.0 g/L once a day. As the objective of the study was to isolate different phases of the process, the OLR was maintained low to facilitate the growth of the slow growing micro-organisms as recommended by Barber and Stuckey (1999). HRT of 30 days was chosen throughout the experiment to maintain constant OLR while the individual HRTs of each compartment were 6, 6, 6 and 12 days, respectively. The retention times slightly vary from the theoretical value as the reactor was slanted towards the N4 end at an angle of about 5° in order to prevent the backflow of the components. The final effluent collection was from the N4 after being allowed to settle for an hour to prevent the biomass washout. Samples for analysis were collected from respective sampling points whenever needed. The temperature of the reactor was maintained at 35 ± 1 °C in a temperature controlled room, relating to the tropical and subtropical climatic conditions. The reactor was operated for 175 days in two periods of operation which constituted 100 days in the first period and after certain modifications as mentioned in section 2.2 to further enhance and stabilize the performance, the reactor operation was continued for another 75 days.
that the acidification took place along the length of the reactor over the course of time as the operation continued. A gradual pH drop from around 7.5–3.5 was observed in N1 between days 0–30 with corresponding COD accumulation. A similar trend was noted for N2 and N3 between days 30–55 and 55–75, respectively. Furthermore, the system went into a steady state maintaining the pH and COD as shown in Fig. 2a and b. Gradual acidification resulted in the isolation of different phases at the respective compartments. The average COD removal efficiencies of CODt (81.7%) and CODs (95.1%) were high throughout the whole period. TOC and VFA removal efficiencies were also 87.6% and 92.7% on an average, respectively. As the actual CODs, TOC and VFA values reaches the maximum after the liquefaction process, in this study the removal rates of the same were measured from their respective highest values. Though the first period of operation showed excellent removal efficiencies, a rapid accumulation of VFA at N3 (between days 55 and 75) was observed which was increasing to be in-line with N1 and N2. It had to be rectified as a precaution to prevent reactor failure as the activity of fast growing acid producing bacteria increases rapidly, and overloads the capacity of slow growing syntrophs and methanogens, which finally leads to accumulation of reduced intermediates as reported by Nachaiyasit and Stuckey (1995). Another possible reason was considered to be due to insufficient HRT for the compartment N3 as the acetogenic communities are considered sensitive and slow growing. Recirculation of the sludge (1 L/d) from N4 to N3 was trialed as an immediate solution for 15 days (days 85–100) to control the pH drop and VFA concentration, which actually improved the condition (Fig. 2b). But on the other hand, this process also increased the loading to N4 while simultaneously diluting the biomass concentration of the same, 8
2.4. Analytical methods
3. Results and discussion 3.1. Reactor performance in period I Measurement of pH regularly provided ideal information to monitor the state of the reactor at different compartments. COD, TOC and VFA accumulation were observed throughout the operation period. The trend of the above mentioned parameters verified
pH
6
4
(a)
2
0
20
40
60
80 100 120 140 160 180 200 Time (d)
14
(b)
12 10 CODs (g/L)
Regular samples of feed, N1, N2, N3 and N4 were collected for analysis. pH was measured using a compact titrator (Mettler Toledo) equipped with a pH probe (Mettler Toledo DGi 115-SC). TS, VS, total suspended solids (TSS) and volatile suspended solids (VSS) were analysed according to the Standard Methods of U.S. EPA (2001). CODt and CODs measurements were done using COD digestion vials (Hach Chemical) and a spectrophotometer (DR/2800, Hach). Aqueous total organic carbon (TOC) and inorganic carbon was measured using a TOC analyzer (TOC-V CSH, Shimadzu, Japan). The determination of VFA was carried out using a GC (Agilent Technologies 7890A, USA), equipped with a flame ionization detector (FID) and a DB-FFAP (Agilent Technologies, USA) column (30 m 0.32 mm 0.50 lm). Supernatant of samples after centrifugation (KUBOTA 3700, Japan) at 12,000 rpm for 20 min were filtered through 0.45 lm cellulose acetate/nylon membrane filters (Membrane Solutions) and used for CODs, TOC and VFA measurements. Total biogas production was monitored daily using a gas flow meter (Model 50D-3E, Mcmillan Company, USA) with 0–50 mL/min measuring range connected to a data logger, while the biogas composition (methane, carbon dioxide, hydrogen and nitrogen contents) was analyzed by GC (Agilent Technologies 7890A, USA) equipped with a thermal conductivity detector (TCD).
8 6 4 2 0 0
20
40
60
80 100 120 140 160 180 200 Time (d)
Fig. 2. Temporal evaluation of (a) pH and (b) CODs in MP-ABR reactor in two periods of operation. d Influent, s N1, . N2, 4 N3, j effluent.
242
A. Ahamed et al. / Bioresource Technology 182 (2015) 239–244
hence it was discontinued. Therefore, a modification was proposed and carried out to the reactor to address this issue that was followed with the second period of operation. 3.2. Modifications to the reactor The retention time of N3 was increased to be 50% longer by increasing the volume of N3 to 16.05 L capacity. The step was taken as a measure to provide enough buffer from the pH and COD shock of the incoming partially digested and highly acidogenic substrate. As a result, the volume of N4 was compromised by 25% to 16.05 L capacity. Therefore, in the operation period II, the HRTs of each compartment of the reactor were 6, 6, 9 and 9 days, respectively. 3.3. Reactor performance in period II Feeding was continued with the same concentration after the alterations. The pH of N3 stabilized and remained at 4.5. A trend in the COD and VFA concentrations along the length of the reactor was observed. The concentration of COD was the highest at N1 as shown in Figs. 2b and 3a, subsequently decreasing along the length of the reactor, eventually resulting in around 90% removal efficiency. This implies that solubilization occurs at the first compartment of the reactor suggesting to be the hydrolysis phase of the system. Furthermore, the concentration of VFA reaches maximum at N2 as shown in Fig. 3b, and gradually decreasing towards the end. The inference could be attributed to the breakdown of complex fatty acids after liquefaction from N1 into simpler short chain VFAs. The three major VFAs observed in N2 were acetate (0.926 g-COD/L),
70 60
(a)
3.4. Overall removal efficiencies
CODt (g/L)
50 40 30 20 10 0
80
100
120
140
160
180
160
180
Time (d) 8
(b)
VFAs (gCOD/L)
6
4
2
0
80
butyrate (1.352 g-COD/L) and propionate (0.787 g-COD/L), that acted as a food source for the subsequent microbial communities. This finding supports that the acidogenesis took place at the second compartment. The conversion of the available butyrate and propionate from acidogenesis to acetate and hydrogen is the major role of acetogenic microorganisms. To verify this, butyrate to acetate ratio (B/A) and propionate to acetate ratio (P/A) were investigated. The ratios indirectly indicate the stability of the reactor performance according to previous reports (Gourdon and Vermande, 1987; Marchaim and Krause, 1993) that studied the effects of propionate in AD systems. In this study, the B/A ratio was 0.658, 1.461, 2.162 and 1.874 and P/A ratio was 0.539, 0.849, 1.179 and 1.994 for N1, N2, N3 and N4, respectively. A rise and fall in the ratio was observed between N1 and N4 for B/A and a gradual rise was observed for P/A. Though P/A ratio was increasing, the actual concentration of propionate have risen and fallen. The possible reason for the trend of B/A and P/A ratios increase in N2 and N3 could be due to the slow breakdown of complex substrates such as fats, proteins and complex sugars that resulted in increase in the concentrations of butyrate and propionate. Even though the ratios are higher towards the end, the final effluent concentrations of acetate (0.093 g-COD/L), butyrate (0.174 g-COD/L) and propionate (0.186 g-COD/L) were low. As there is no clear cut evidence from the trends of B/A and P/A ratios supporting the conversion of butyrate and propionate to acetate in N3 before being consumed by methanogens, a detailed microbial study is recommended in the near future to verify the presence of acetogens in N3. Moreover, most of the biogas mainly produced was from the N4 that used up around 90% of the VFAs from N3 as substrate for methanogens. This showed that methanogenesis eventuated at the final compartment.
100
120 140 Time (d)
Fig. 3. Temporal evaluation of (a) CODt and (b) VFA concentration in MP-ABR reactor in the second period of operation. d Influent, s N1, . N2, 4 N3, j effluent.
The overall removal rates of CODt and CODs were 85.3% and 94.5% on average. A study by Bachmann et al. (1985), using a five-chambered ABR to treat soluble organic wastes (carbohydrate–protein) with 6.3 L reactor volume at an inlet COD of 8 g-COD/L achieved 55–93% COD removal efficiency (operated under HRT of 4.8–71 h at 35 °C). The results achieved in both studies were similar and complimentary considering the similar substrate types. The VFA removal efficiency of 89.6% was observed on an average throughout the experiment. The trend of VFA removal efficiency indicated slightly better removal during the first period of operation than the second period. The main reason was due to the accumulation of VFA during the initial stabilization period that resulted in the acidification of the N1 and N2 over the period of time. Higher retention time for methanogenesis during the period I and the freshness of the seed sludge that rapidly consume the acetates could also be partially attributed to the better results. VS removal of the system was 86.4% on an average, which reached around 90% towards the latter stages of the experiment as shown in Fig. 4. The gradual increase in VS removal rate was associated with the growth of methanogenic microorganisms over the period of time that facilitated better VS removal efficiency. Relatively low VS removal (57–60%) in the HASL system was reported by Wang et al. (2002). Fig. 4 shows a sharp drop in solids removal between days 63 and 84 and towards the end of operation period. The drop was due to the biomass washout that resulted from increased biomass growth in N4. VSS concentration in N1 was around 30 g/L whereas in the rest of the compartments the concentration steadied around 5 g/L, this could be correlated to the concentration of biomass as represented in the previous studies by Shanmugam and Horan (2009) and Sreela-or et al. (2011).
A. Ahamed et al. / Bioresource Technology 182 (2015) 239–244
120
Solid removal %
100
80
60 TS removal VS removal TSS removal VSS removal TS removal regression VS removal regression TSS removal regression VSS removal regression
40
20
0 0
50
100
150
200
Time (d) Fig. 4. The trend and temporal evaluation of solid removal rate in MP-ABR reactor in the second period of operation.
One possible reason for higher biomass concentration in N1 is because only the soluble part gets transferred to the subsequent compartments and more than 90% of the solids are retained back at the N1. Another reason could be the slow growth rate of the acetogenic and methanogenic microbes as compared to the hydrolytic and acidogenic populations. The CODt plot also verifies the same inferences (Fig. 3a). The reduction in the trendline of TSS and VSS as shown in Fig. 4 was correlated to the increase in the concentration of the biomass inside the reactor that resulted in slightly higher biomass washout towards the latter stages of operation. The optimal biomass concentration for an ABR system at an OLR of 20 kg-COD/m3 was found to be 7 g/L (Grobicki and Stuckey, 1991). The authors also concluded that the reactor design facilitates low biomass washout during shock loads and offers very high specific reaction rate. To prevent the loss of biomass an additional settling tank can be added which could help to recycle the lost biomass subsequently increasing the performance.
3.5. Biogas production The biogas yield was low and inconsistent during the start up period owing to the VFA accumulation during the period I. The average biogas yield stabilized around 4.8 L/d in period II, which constituted of 50–60% methane. The VS loading was 25.78 g/d on an average with 86.4% removal efficiency. Hence, the biogas generated was 215.57 mL/g-VSremoved d, which is substantially high and proves high conversion efficiency considering the feasibility study by Tawfik et al. (2011a) that generated hydrogen production potential of 41.2 mL/g-VSremoved d at an HRT of 2.2 days and 35.3 mL/g-VSremoved d at an HRT of 1.1 day. Apart from that, a two-stage ABR for biohydrogen production from municipal FW investigated by Tawfik et al. (2011b), yielded 620 mL-H2/ g-VSremoved d in total, at a total HRT of 2.2 days and OLR of 29 kg-COD/m3 d. Though the conversion efficiency obtained by the experiment is high, H2 production requires strict pH range of 5–6 whereas controlling the pH of municipal FW is challenging and tends to reach a pH of as low as 3 as observed in our study. Also the VS removal efficiency was 68% (final VS 18 ± 6 g/L) whereas the MP-ABR system achieved around 90% (final VS 1.95 ± 1.2 g/L). Recent studies have reported on co-digestion of FW as the studies involving only FW as substrate have failed at high OLRs and eventually not feasible due to VFA accumulation (Han and Shin, 2004; Lin et al., 2011; Shen et al., 2013). Generally, the systems involving co-digestion of FW have performed better in terms of methane yield than single substrate. A study comparing single
243
and two phase systems treating fruit and vegetable waste and food waste co-digestion by Shen et al. (2013) reported biogas yield of approximately 4 L/d (0.5 g-VS/L), 6 L/d (1 g-VS/L), 11 L/d (2 g-VS/ L), 17 L/d (3 g-VS/L), 18.5 L/d (4 g-VS/L) and 12 L/d (5 g-VS/L) at their respective OLRs. Very little improvement in biogas yield was observed for increasing the OLR from 3 to 4 g-VS/L and faced a 33% drop in the biogas yield for OLRs 4–5 g-VS/L resulting in the accumulation of propionic and pentanoic acids. Though the high OLR affected the performance of two phase system, system failure was not observed showing good buffering capacity compared to single phase system that failed immediately due to VFA accumulation. In a similar case, MP-ABR could provide even better buffering capacity compared to two-phase system. The effects of mixture ratio on anaerobic co-digestion studied by Lin et al. (2011) reported biogas yield of 0.49 m3-CH4/kg-VS at an optimal mixing ratio of 1:1 for fruit and vegetable waste and food waste co-digestion with VS and CODs removal efficiencies of 74.9% and 96.1%. In the same study, when trialed with FW alone the reactor failed due to VFA accumulation at an OLR of 3 kg-VS/m3 with methane production dropping to 0.06 m3-CH4/kg-VS and CODs removal to 73.3%. The significant improvement in the performance is mainly due to the adjustments made in C/N ratio attained by mixing two or more substrates. Iacovidou et al. (2012) has reported the reason for this could be due to the improvements in kinetic coefficients of hydrolysis that increases the overall performance. In each case finding the optimal mixing ratio is also necessary. A study on FW and cattle manure (CM) co-digestion by Zhang et al. (2013), have discussed about the importance of carbon to nitrogen (C/N) ratio in AD systems. The optimum C/N ratio was 15.8, corresponding to the FW/CM ratio of 2. The total methane production at the optimum FW/CM ratio was enhanced by 41.1%, corresponding to the methane yield of 388 mL/g-VS. The appropriate C/N ratio and the higher biodegradation of lipids might be the main reason for the improved biogas production and methane yield during co-digestion. Whereas in this study, the C/N ratio of FW used was 26.28 which could be the reason for low biogas yield compared to co-digestion. A report by Boopathy (1998) concluded that the biogas production of the ABR was equal to or greater than conventional, upflow anaerobic filtered and attached growth anaerobic digester designs. This is attributed to the ability of the baffled reactor to effectively trap the small-diameter methane containing sludge particles and maintain long solids retention time. But considering the total reactor volume of MP-ABR in this study, the biogas production is relatively low which is 88.45 mL/L of the reactor volume. 4. Conclusion Overall, the MP-ABR system has achieved an efficient recovery of bioenergy from FW and has shown a great potential to treat the municipal FW considering the high removal rates achieved in this study. Though, the objective of separating different phases of an AD reactor has been successfully achieved in this study, a further detailed microbial analysis is recommended to better understand the microbial communities in different compartments. Additionally, future works could involve operating the reactor at higher OLR, recycling the effluent, and looking into the prospects of co-digestion towards a more sustainable operation model. Acknowledgements The authors are grateful to the National Research Foundation (NRF), Singapore for financial support (NRF-CRP5-2009-02). We appreciate Mr. Bernard Ng and NEWRI-R3C/NTU family for their contributions to this research program.
244
A. Ahamed et al. / Bioresource Technology 182 (2015) 239–244
References Bachmann, A., Beard, V.L., McCarty, P.L., 1985. Performance characteristics of the anaerobic baffled reactor. Water Res. 19 (1), 99–106. Barber, W.P., Stuckey, D.C., 1999. The use of the anaerobic baffled reactor (ABR) for wastewater treatment: a review. Water Res. 33 (7), 1559–1578. Boopathy, R., 1998. Biological treatment of swine waste using anaerobic baffled reactors. Bioresour. Technol. 64 (1), 1–6. Chong, S., Sen, T.K., Kayaalp, A., Ang, H.M., 2012. The performance enhancements of upflow anaerobic sludge blanket (UASB) reactors for domestic sludge treatment – a state-of-the-art review. Water Res. 46 (11), 3434–3470. FAO, 2013. Food and Agricultural Organization of the United Nations: energy-smart food for people and climate. Available from: http://www.fao.org/energy/81344/ en/ [cited: 17 April]. Gourdon, R., Vermande, P., 1987. Effects of propionic acid concentration on anaerobic digestion of pig manure. Biomass 13, 1–12. Grobicki, A., Stuckey, D.C., 1991. Performance of the Anaerobic Baffled Reactor under steady-state and shock loading conditions. Biotechnol. Bioeng. 37, 344–355. Han, S.-K., Shin, H.-S., 2004. Biohydrogen production by anaerobic fermentation of food waste. Int. J. Hydrogen Energ. 29, 569–577. Iacovidou, E., Ohandja, D.-G., Voulvoulis, N., 2012. Food waste co-digestion with sewage sludge – realising its potential in the UK. J. Environ. Manage. 112, 267– 274. Ke, S., Shi, Z., Fang, H.H.P., 2005. Applications of two-phase anaerobic degradation in industrial wastewater treatment. Int. J. Environ. Pollut. 23 (1), 65–80. Lin, J., Zuo, J., Gan, L., Li, P., Liu, F., Wang, K., Chen, L., Gan, H., 2011. Effects of mixture ratio on anaerobic co-digestion with fruit and vegetable waste and food waste of China. J. Environ. Sci. 23 (8), 1403–1408. Liu, X.Y., Ding, H.B., Sreeramachandran, S., Stabnikova, O., Wang, J.-Y., 2008. Enhancement of food waste digestion in the hybrid anaerobic solid–liquid system. Water Sci. Technol. 57 (9), 1369–1373. Marchaim, U., Krause, C., 1993. Propionic to acetic ratios in overloaded anaerobic digestion. Bioresour. Technol. 43, 195–203.
Mohan, S., Bindhu, B.K., 2008. Effect of phase separation on anaerobic digestion of kitchen waste. J. Environ. Eng. Sci. 7 (2), 91–103. Nachaiyasit, S., Stuckey, D.C., 1995. Microbial response to environmental changes in an Anaerobic Baffled Reactor (ABR). Anton. van Leeuw. 67, 111–123. NEA, 2013. National Environment Agency, Singapore. Available from: http://app2. nea.gov.sg/energy-waste/waste-management/waste-statistics-and-overallrecycling [cited: 15 November]. Shanmugam, P., Horan, N.J., 2009. Simple and rapid methods to evaluate methane potential and biomass yield for a range of mixed solid wastes. Bioresour. Technol. 100, 471–474. Shen, F., Yuan, H., Pang, Y., Chen, S., Zhu, B., Zou, D., Liu, Y., Ma, J., Yu, L., Li, X., 2013. Performances of anaerobic co-digestion of fruit & vegetable waste (FVW) and food waste (FW): single-phase vs. two-phase. Bioresour. Technol. 144, 80–85. Sreela-or, C., Imai, T., Plangklang, P., Reungsang, A., 2011. Optimization of key factors affecting hydrogen production from food waste by anaerobic mixed cultures. Int. J. Hydrogen Energ. 36, 14120–14133. Tawfik, A., Salem, A., El-Qelish, M., Abdullah, A.M., Abou Taleb, E., 2011a. Feasibility of biological hydrogen production from kitchen waste via anaerobic baffled reactor (ABR). Int. J. Sust. Water Environ. Syst. 2 (2), 117–122. Tawfik, A., Salem, A., El-Qelish, M., 2011b. Two stage anaerobic baffled reactors for bio-hydrogen production from municipal food waste. Bioresour. Technol. 102, 8723–8726. U.S. EPA, 2001. Method 1684: total, fixed, and volatile solids in water, solids, and biosolids. EPA-821-R-01-015. Office of Water, Office of Science and Technology, Engineering and Analysis Division, U.S. Environmental Protection Agency, Washington, DC. Wang, J.-Y., Xu, H.L., Tay, J.-H., 2002. A hybrid two-phase system for anaerobic digestion of food waste. Water Sci. Technol. 45 (12), 159–165. Wang, J.-Y., Zhang, H., Stabnikova, O., Tay, J.-H., 2003. Removal of ammonia in a modified two-phase food waste anaerobic digestion system coupled with an aerated submerged biofilter. Waste Manage. Res. 21, 527–534. Zhang, C., Xiao, G., Peng, L., Su, H., Tan, T., 2013. The anaerobic co-digestion of food waste and cattle manure. Bioresour. Technol. 129, 170–176.
Contents lists available at ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
Multi-phased anaerobic baffled reactor treating food waste A. Ahamed a,1, C.-L. Chen a, R. Rajagopal a, D. Wu a, Y. Mao a, I.J.R. Ho a, J.W. Lim a, J.-Y. Wang a,b,⇑ a Residues and Resource Reclamation Centre, Nanyang Environment and Water Research Institute, Nanyang Technological University, #06-08 CleanTech One, 1 Cleantech Loop, Singapore 637141, Singapore b School of Civil and Environmental Engineering, Nanyang Technological University, 50 Nanyang Avenue, Singapore 639798, Singapore
h i g h l i g h t s Study on isolation of different phases of an anaerobic digestion process. First reference for biogas yield from food waste using anaerobic baffled reactor. Effluent quality shows this as an efficient treatment for food waste.
a r t i c l e
i n f o
Article history: Received 12 November 2014 Received in revised form 26 January 2015 Accepted 28 January 2015 Available online 4 February 2015 Keywords: Food waste Anaerobic baffled reactor Phase separation Biogas Anaerobic digestion
a b s t r a c t This study was conducted to identify the performance of a multi-phased anaerobic baffled reactor (MP-ABR) with food waste (FW) as the substrate for biogas production and thereby to promote an efficient energy recovery and treatment method for the wastes with high organic solid content through phase separation. A four-chambered ABR was operated at an HRT of 30 days with an OLR of 0.5–1.0 g-VS/L d for a period of 175 days at 35 ± 1 °C. Consistent overall removal efficiencies of 85.3% (CODt), 94.5% (CODs), 89.6% (VFA) and 86.4% (VS) were observed throughout the experiment displaying a great potential to treat FW. Biogas generated was 215.57 mL/g-VSremoved d. Phase separation was observed and supported by the COD and VFA trends, and an efficient recovery of bioenergy from FW was achieved. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Approximately one third of the food produced in the world for human consumption every year, which is approximately 1.3 billion tonnes, gets wasted globally. The wastage of food can be attributed into loss of resources in food production, and the unnecessary production of greenhouse gas emissions (FAO, 2013). In a world with depleting energy resources, the amount of food waste (FW) generated everyday needs an utmost use to move towards sustainable development. FW, rich in organic acids, constitutes an ideal source for bioenergy recovery. On the other hand, FW disposal has been a
⇑ Corresponding author at: Residues and Resource Reclamation Centre, Nanyang Environment and Water Research Institute, Nanyang Technological University, #0608 CleanTech One, 1 Cleantech Loop, Singapore 637141, Singapore. Tel.: +65 67904100; fax: +65 67927319. E-mail addresses: [email protected], [email protected] (A. Ahamed), jywang@ ntu.edu.sg (J.-Y. Wang). 1 Tel.: +65 67904102; fax: +65 67927319. http://dx.doi.org/10.1016/j.biortech.2015.01.117 0960-8524/Ó 2015 Elsevier Ltd. All rights reserved.
growing problem in highly populated cities such as Singapore comprising 21.1% of the municipal solid waste that is disposed (NEA, 2013). Wang et al. (2003) reported that FW has been main source of odour and leachate in collection and transportation of municipal solid waste due to their high volatile components and moisture content, and has been causing significant environmental impact. Anaerobic digestion (AD), a biological conversion of organic matter into methane, carbon dioxide, inorganic nutrients and humus-like matter, appears to be the most promising method for FW treatment (Liu et al., 2008). As compared to the aerobic method, the use of anaerobic processes for the waste streams provides greater economic and environmental benefits and advantages (Mohan and Bindhu, 2008). AD technology is a well known method for waste utilization, and various configurations of the reactor type have been developed thus far. The most prominent and simple design is the single stage process and it has been widely used in various applications. However, its low efficiency has been highlighted (Ke et al., 2005). Besides single stage system, two-phase reactors have gained attraction and relevant research works has been reviewed by Ke et al. (2005). The two-phase
240
A. Ahamed et al. / Bioresource Technology 182 (2015) 239–244
system attempts to separate the acidogenic and methanogenic phases of AD process. The goal is to improve the performance of the reactor by allowing enough buffering capacity and space for respective microbial communities to flourish under suitable conditions. Adding on to this, a two-phase hybrid anaerobic solid–liquid system (HASL) by Wang et al. (2002, 2003) and Liu et al. (2008), was studied in detail for FW digestion and has shown promising results. Apart from these reactor types, an up-flow anaerobic sludge blanket process elaborately reviewed by Chong et al. (2012) is extensively applied for wastewater treatment with relatively high removal efficiency at short hydraulic retention times (HRTs), but is also considered not suitable for wastes with high solid content such as municipal FW (Tawfik et al., 2011b). Improving the soluble component removal efficiency, longer biomass retention, enhanced bacterial activity and high rate of contact between the cells and the substrate are important attributes for an efficient anaerobic digestion system (Grobicki and Stuckey, 1991). The multi-phased anaerobic baffled reactor (MP-ABR) possess a great potential in improving the afore mentioned criterion, and also provides better resilience to hydraulic and organic shock loadings (Barber and Stuckey, 1999). The reactor design allows adequate space for the microbial communities to sustain without interferences. The significant advantage of the reactor type is its ability to separate different phases of the AD process longitudinally, allowing the reactor to behave as a ‘phase separated’ system, which may offer the best choice for high efficiency removal rates (Mohan and Bindhu, 2008). FW, which possesses about 75% moisture content and 24% volatile solids (VS) (97% VS/total solids (TS) ratio), qualifies as a perfect substrate type for this type of reactor design due to its high solids retention capacity. Although extensive studies have been conducted on anaerobic baffled reactor (ABR) as mentioned in the review by Barber and Stuckey (1999), little is known about biogas production from FW, except for the work of Tawfik et al. (2011a,b), using a laboratory scale five-chambered ABR for biological hydrogen production from diluted kitchen waste. The main objective of this study was to assess the performance of the MP-ABR system treating FW for biogas production, to isolate the different phases of the AD process and to promote an efficient energy recovery and treatment method for the wastes with high organic content. Various performance indicators including chemical oxygen demand (COD), volatile fatty
acid (VFA), TS and VS were also investigated to evaluate the efficiency of reactor performance. 2. Methods 2.1. Feed and inocula characteristics FW was collected from a canteen inside the university campus, which mainly composed of rice, noodles, meat, vegetables and condiments. After removing the bones and inorganic materials, the waste was homogenized using a kitchen blender to disintegrate the particulate organics into size < 2 mm. The feed was diluted with tap water according to the influent organic loading rate (OLR) and flow rate before being fed to the reactor. The characteristics of the feed were pH (5.36 ± 0.82); total COD (CODt) (17.31 ± 5.05 g/L); soluble COD (CODs) (4.98 ± 2.03 g/L); VS (14.78 g/L); total VFA (VFA) (1.214 g-COD/L) on an average. Inoculum was obtained from a local wastewater treatment plant (Ulu Pandan Wastewater Treatment Plant) that uses AD technology for sewage sludge treatment. The reactor was filled up with 70% of the seed sludge and 30% with tap water. The seed sludge contained TS of 1.77 g/kg, of which VS was 70%. The pH and alkalinity were about 6.9 and 1916 mg-CaCO3/L respectively. 2.2. MP-ABR system A rectangular acrylic container with internal vertical baffles hanging and standing alternately embodied the basic design of the reactor (Fig. 1). To allow the number of compartments and width of each to be adjustable, the baffles were designed to be movable. The design of the reactor was slightly different from the conventional ABR. As the FW contain very high solid content, the widths of the downflow segment were increased for this study to allow the solids to flow over smoothly without any blockage. The reactor was operated in two operational periods with a modification to the reactor configuration in the second period. A four compartment (N1, N2, N3 and N4) reactor model was selected to isolate the different phases of the AD process, which is also considered to perform better (Boopathy, 1998). The individual working volumes were 10.7 L each for the first three and 21.4 L for the last compartment during the first period of operation. Subsequently, the volumes of N3 and N4 were adjusted to be 16.05 L each in
Biogas collection
Effluent
Feed-in
N4
N3
N2
N1
Fig. 1. Flow scheme of the MP-ABR with the direction of sludge flow.
241
A. Ahamed et al. / Bioresource Technology 182 (2015) 239–244
the second period. Each compartment was provided with an overhead mixer rotating at a speed of 100–150 rpm to minimize dead zones and short circuiting. The sampling points were at the mid-level of each compartment and the draining ports were at the bottom level. The reactor was connected to a gas flow meter and data logger to measure the gas volume. The gas composition was measured by subsequently collecting the gas in a gas bag and followed by gas chromatograph (GC) analysis. 2.3. Operating strategy The reactor was fed with FW at an OLR of 0.5–1.0 g/L once a day. As the objective of the study was to isolate different phases of the process, the OLR was maintained low to facilitate the growth of the slow growing micro-organisms as recommended by Barber and Stuckey (1999). HRT of 30 days was chosen throughout the experiment to maintain constant OLR while the individual HRTs of each compartment were 6, 6, 6 and 12 days, respectively. The retention times slightly vary from the theoretical value as the reactor was slanted towards the N4 end at an angle of about 5° in order to prevent the backflow of the components. The final effluent collection was from the N4 after being allowed to settle for an hour to prevent the biomass washout. Samples for analysis were collected from respective sampling points whenever needed. The temperature of the reactor was maintained at 35 ± 1 °C in a temperature controlled room, relating to the tropical and subtropical climatic conditions. The reactor was operated for 175 days in two periods of operation which constituted 100 days in the first period and after certain modifications as mentioned in section 2.2 to further enhance and stabilize the performance, the reactor operation was continued for another 75 days.
that the acidification took place along the length of the reactor over the course of time as the operation continued. A gradual pH drop from around 7.5–3.5 was observed in N1 between days 0–30 with corresponding COD accumulation. A similar trend was noted for N2 and N3 between days 30–55 and 55–75, respectively. Furthermore, the system went into a steady state maintaining the pH and COD as shown in Fig. 2a and b. Gradual acidification resulted in the isolation of different phases at the respective compartments. The average COD removal efficiencies of CODt (81.7%) and CODs (95.1%) were high throughout the whole period. TOC and VFA removal efficiencies were also 87.6% and 92.7% on an average, respectively. As the actual CODs, TOC and VFA values reaches the maximum after the liquefaction process, in this study the removal rates of the same were measured from their respective highest values. Though the first period of operation showed excellent removal efficiencies, a rapid accumulation of VFA at N3 (between days 55 and 75) was observed which was increasing to be in-line with N1 and N2. It had to be rectified as a precaution to prevent reactor failure as the activity of fast growing acid producing bacteria increases rapidly, and overloads the capacity of slow growing syntrophs and methanogens, which finally leads to accumulation of reduced intermediates as reported by Nachaiyasit and Stuckey (1995). Another possible reason was considered to be due to insufficient HRT for the compartment N3 as the acetogenic communities are considered sensitive and slow growing. Recirculation of the sludge (1 L/d) from N4 to N3 was trialed as an immediate solution for 15 days (days 85–100) to control the pH drop and VFA concentration, which actually improved the condition (Fig. 2b). But on the other hand, this process also increased the loading to N4 while simultaneously diluting the biomass concentration of the same, 8
2.4. Analytical methods
3. Results and discussion 3.1. Reactor performance in period I Measurement of pH regularly provided ideal information to monitor the state of the reactor at different compartments. COD, TOC and VFA accumulation were observed throughout the operation period. The trend of the above mentioned parameters verified
pH
6
4
(a)
2
0
20
40
60
80 100 120 140 160 180 200 Time (d)
14
(b)
12 10 CODs (g/L)
Regular samples of feed, N1, N2, N3 and N4 were collected for analysis. pH was measured using a compact titrator (Mettler Toledo) equipped with a pH probe (Mettler Toledo DGi 115-SC). TS, VS, total suspended solids (TSS) and volatile suspended solids (VSS) were analysed according to the Standard Methods of U.S. EPA (2001). CODt and CODs measurements were done using COD digestion vials (Hach Chemical) and a spectrophotometer (DR/2800, Hach). Aqueous total organic carbon (TOC) and inorganic carbon was measured using a TOC analyzer (TOC-V CSH, Shimadzu, Japan). The determination of VFA was carried out using a GC (Agilent Technologies 7890A, USA), equipped with a flame ionization detector (FID) and a DB-FFAP (Agilent Technologies, USA) column (30 m 0.32 mm 0.50 lm). Supernatant of samples after centrifugation (KUBOTA 3700, Japan) at 12,000 rpm for 20 min were filtered through 0.45 lm cellulose acetate/nylon membrane filters (Membrane Solutions) and used for CODs, TOC and VFA measurements. Total biogas production was monitored daily using a gas flow meter (Model 50D-3E, Mcmillan Company, USA) with 0–50 mL/min measuring range connected to a data logger, while the biogas composition (methane, carbon dioxide, hydrogen and nitrogen contents) was analyzed by GC (Agilent Technologies 7890A, USA) equipped with a thermal conductivity detector (TCD).
8 6 4 2 0 0
20
40
60
80 100 120 140 160 180 200 Time (d)
Fig. 2. Temporal evaluation of (a) pH and (b) CODs in MP-ABR reactor in two periods of operation. d Influent, s N1, . N2, 4 N3, j effluent.
242
A. Ahamed et al. / Bioresource Technology 182 (2015) 239–244
hence it was discontinued. Therefore, a modification was proposed and carried out to the reactor to address this issue that was followed with the second period of operation. 3.2. Modifications to the reactor The retention time of N3 was increased to be 50% longer by increasing the volume of N3 to 16.05 L capacity. The step was taken as a measure to provide enough buffer from the pH and COD shock of the incoming partially digested and highly acidogenic substrate. As a result, the volume of N4 was compromised by 25% to 16.05 L capacity. Therefore, in the operation period II, the HRTs of each compartment of the reactor were 6, 6, 9 and 9 days, respectively. 3.3. Reactor performance in period II Feeding was continued with the same concentration after the alterations. The pH of N3 stabilized and remained at 4.5. A trend in the COD and VFA concentrations along the length of the reactor was observed. The concentration of COD was the highest at N1 as shown in Figs. 2b and 3a, subsequently decreasing along the length of the reactor, eventually resulting in around 90% removal efficiency. This implies that solubilization occurs at the first compartment of the reactor suggesting to be the hydrolysis phase of the system. Furthermore, the concentration of VFA reaches maximum at N2 as shown in Fig. 3b, and gradually decreasing towards the end. The inference could be attributed to the breakdown of complex fatty acids after liquefaction from N1 into simpler short chain VFAs. The three major VFAs observed in N2 were acetate (0.926 g-COD/L),
70 60
(a)
3.4. Overall removal efficiencies
CODt (g/L)
50 40 30 20 10 0
80
100
120
140
160
180
160
180
Time (d) 8
(b)
VFAs (gCOD/L)
6
4
2
0
80
butyrate (1.352 g-COD/L) and propionate (0.787 g-COD/L), that acted as a food source for the subsequent microbial communities. This finding supports that the acidogenesis took place at the second compartment. The conversion of the available butyrate and propionate from acidogenesis to acetate and hydrogen is the major role of acetogenic microorganisms. To verify this, butyrate to acetate ratio (B/A) and propionate to acetate ratio (P/A) were investigated. The ratios indirectly indicate the stability of the reactor performance according to previous reports (Gourdon and Vermande, 1987; Marchaim and Krause, 1993) that studied the effects of propionate in AD systems. In this study, the B/A ratio was 0.658, 1.461, 2.162 and 1.874 and P/A ratio was 0.539, 0.849, 1.179 and 1.994 for N1, N2, N3 and N4, respectively. A rise and fall in the ratio was observed between N1 and N4 for B/A and a gradual rise was observed for P/A. Though P/A ratio was increasing, the actual concentration of propionate have risen and fallen. The possible reason for the trend of B/A and P/A ratios increase in N2 and N3 could be due to the slow breakdown of complex substrates such as fats, proteins and complex sugars that resulted in increase in the concentrations of butyrate and propionate. Even though the ratios are higher towards the end, the final effluent concentrations of acetate (0.093 g-COD/L), butyrate (0.174 g-COD/L) and propionate (0.186 g-COD/L) were low. As there is no clear cut evidence from the trends of B/A and P/A ratios supporting the conversion of butyrate and propionate to acetate in N3 before being consumed by methanogens, a detailed microbial study is recommended in the near future to verify the presence of acetogens in N3. Moreover, most of the biogas mainly produced was from the N4 that used up around 90% of the VFAs from N3 as substrate for methanogens. This showed that methanogenesis eventuated at the final compartment.
100
120 140 Time (d)
Fig. 3. Temporal evaluation of (a) CODt and (b) VFA concentration in MP-ABR reactor in the second period of operation. d Influent, s N1, . N2, 4 N3, j effluent.
The overall removal rates of CODt and CODs were 85.3% and 94.5% on average. A study by Bachmann et al. (1985), using a five-chambered ABR to treat soluble organic wastes (carbohydrate–protein) with 6.3 L reactor volume at an inlet COD of 8 g-COD/L achieved 55–93% COD removal efficiency (operated under HRT of 4.8–71 h at 35 °C). The results achieved in both studies were similar and complimentary considering the similar substrate types. The VFA removal efficiency of 89.6% was observed on an average throughout the experiment. The trend of VFA removal efficiency indicated slightly better removal during the first period of operation than the second period. The main reason was due to the accumulation of VFA during the initial stabilization period that resulted in the acidification of the N1 and N2 over the period of time. Higher retention time for methanogenesis during the period I and the freshness of the seed sludge that rapidly consume the acetates could also be partially attributed to the better results. VS removal of the system was 86.4% on an average, which reached around 90% towards the latter stages of the experiment as shown in Fig. 4. The gradual increase in VS removal rate was associated with the growth of methanogenic microorganisms over the period of time that facilitated better VS removal efficiency. Relatively low VS removal (57–60%) in the HASL system was reported by Wang et al. (2002). Fig. 4 shows a sharp drop in solids removal between days 63 and 84 and towards the end of operation period. The drop was due to the biomass washout that resulted from increased biomass growth in N4. VSS concentration in N1 was around 30 g/L whereas in the rest of the compartments the concentration steadied around 5 g/L, this could be correlated to the concentration of biomass as represented in the previous studies by Shanmugam and Horan (2009) and Sreela-or et al. (2011).
A. Ahamed et al. / Bioresource Technology 182 (2015) 239–244
120
Solid removal %
100
80
60 TS removal VS removal TSS removal VSS removal TS removal regression VS removal regression TSS removal regression VSS removal regression
40
20
0 0
50
100
150
200
Time (d) Fig. 4. The trend and temporal evaluation of solid removal rate in MP-ABR reactor in the second period of operation.
One possible reason for higher biomass concentration in N1 is because only the soluble part gets transferred to the subsequent compartments and more than 90% of the solids are retained back at the N1. Another reason could be the slow growth rate of the acetogenic and methanogenic microbes as compared to the hydrolytic and acidogenic populations. The CODt plot also verifies the same inferences (Fig. 3a). The reduction in the trendline of TSS and VSS as shown in Fig. 4 was correlated to the increase in the concentration of the biomass inside the reactor that resulted in slightly higher biomass washout towards the latter stages of operation. The optimal biomass concentration for an ABR system at an OLR of 20 kg-COD/m3 was found to be 7 g/L (Grobicki and Stuckey, 1991). The authors also concluded that the reactor design facilitates low biomass washout during shock loads and offers very high specific reaction rate. To prevent the loss of biomass an additional settling tank can be added which could help to recycle the lost biomass subsequently increasing the performance.
3.5. Biogas production The biogas yield was low and inconsistent during the start up period owing to the VFA accumulation during the period I. The average biogas yield stabilized around 4.8 L/d in period II, which constituted of 50–60% methane. The VS loading was 25.78 g/d on an average with 86.4% removal efficiency. Hence, the biogas generated was 215.57 mL/g-VSremoved d, which is substantially high and proves high conversion efficiency considering the feasibility study by Tawfik et al. (2011a) that generated hydrogen production potential of 41.2 mL/g-VSremoved d at an HRT of 2.2 days and 35.3 mL/g-VSremoved d at an HRT of 1.1 day. Apart from that, a two-stage ABR for biohydrogen production from municipal FW investigated by Tawfik et al. (2011b), yielded 620 mL-H2/ g-VSremoved d in total, at a total HRT of 2.2 days and OLR of 29 kg-COD/m3 d. Though the conversion efficiency obtained by the experiment is high, H2 production requires strict pH range of 5–6 whereas controlling the pH of municipal FW is challenging and tends to reach a pH of as low as 3 as observed in our study. Also the VS removal efficiency was 68% (final VS 18 ± 6 g/L) whereas the MP-ABR system achieved around 90% (final VS 1.95 ± 1.2 g/L). Recent studies have reported on co-digestion of FW as the studies involving only FW as substrate have failed at high OLRs and eventually not feasible due to VFA accumulation (Han and Shin, 2004; Lin et al., 2011; Shen et al., 2013). Generally, the systems involving co-digestion of FW have performed better in terms of methane yield than single substrate. A study comparing single
243
and two phase systems treating fruit and vegetable waste and food waste co-digestion by Shen et al. (2013) reported biogas yield of approximately 4 L/d (0.5 g-VS/L), 6 L/d (1 g-VS/L), 11 L/d (2 g-VS/ L), 17 L/d (3 g-VS/L), 18.5 L/d (4 g-VS/L) and 12 L/d (5 g-VS/L) at their respective OLRs. Very little improvement in biogas yield was observed for increasing the OLR from 3 to 4 g-VS/L and faced a 33% drop in the biogas yield for OLRs 4–5 g-VS/L resulting in the accumulation of propionic and pentanoic acids. Though the high OLR affected the performance of two phase system, system failure was not observed showing good buffering capacity compared to single phase system that failed immediately due to VFA accumulation. In a similar case, MP-ABR could provide even better buffering capacity compared to two-phase system. The effects of mixture ratio on anaerobic co-digestion studied by Lin et al. (2011) reported biogas yield of 0.49 m3-CH4/kg-VS at an optimal mixing ratio of 1:1 for fruit and vegetable waste and food waste co-digestion with VS and CODs removal efficiencies of 74.9% and 96.1%. In the same study, when trialed with FW alone the reactor failed due to VFA accumulation at an OLR of 3 kg-VS/m3 with methane production dropping to 0.06 m3-CH4/kg-VS and CODs removal to 73.3%. The significant improvement in the performance is mainly due to the adjustments made in C/N ratio attained by mixing two or more substrates. Iacovidou et al. (2012) has reported the reason for this could be due to the improvements in kinetic coefficients of hydrolysis that increases the overall performance. In each case finding the optimal mixing ratio is also necessary. A study on FW and cattle manure (CM) co-digestion by Zhang et al. (2013), have discussed about the importance of carbon to nitrogen (C/N) ratio in AD systems. The optimum C/N ratio was 15.8, corresponding to the FW/CM ratio of 2. The total methane production at the optimum FW/CM ratio was enhanced by 41.1%, corresponding to the methane yield of 388 mL/g-VS. The appropriate C/N ratio and the higher biodegradation of lipids might be the main reason for the improved biogas production and methane yield during co-digestion. Whereas in this study, the C/N ratio of FW used was 26.28 which could be the reason for low biogas yield compared to co-digestion. A report by Boopathy (1998) concluded that the biogas production of the ABR was equal to or greater than conventional, upflow anaerobic filtered and attached growth anaerobic digester designs. This is attributed to the ability of the baffled reactor to effectively trap the small-diameter methane containing sludge particles and maintain long solids retention time. But considering the total reactor volume of MP-ABR in this study, the biogas production is relatively low which is 88.45 mL/L of the reactor volume. 4. Conclusion Overall, the MP-ABR system has achieved an efficient recovery of bioenergy from FW and has shown a great potential to treat the municipal FW considering the high removal rates achieved in this study. Though, the objective of separating different phases of an AD reactor has been successfully achieved in this study, a further detailed microbial analysis is recommended to better understand the microbial communities in different compartments. Additionally, future works could involve operating the reactor at higher OLR, recycling the effluent, and looking into the prospects of co-digestion towards a more sustainable operation model. Acknowledgements The authors are grateful to the National Research Foundation (NRF), Singapore for financial support (NRF-CRP5-2009-02). We appreciate Mr. Bernard Ng and NEWRI-R3C/NTU family for their contributions to this research program.
244
A. Ahamed et al. / Bioresource Technology 182 (2015) 239–244
References Bachmann, A., Beard, V.L., McCarty, P.L., 1985. Performance characteristics of the anaerobic baffled reactor. Water Res. 19 (1), 99–106. Barber, W.P., Stuckey, D.C., 1999. The use of the anaerobic baffled reactor (ABR) for wastewater treatment: a review. Water Res. 33 (7), 1559–1578. Boopathy, R., 1998. Biological treatment of swine waste using anaerobic baffled reactors. Bioresour. Technol. 64 (1), 1–6. Chong, S., Sen, T.K., Kayaalp, A., Ang, H.M., 2012. The performance enhancements of upflow anaerobic sludge blanket (UASB) reactors for domestic sludge treatment – a state-of-the-art review. Water Res. 46 (11), 3434–3470. FAO, 2013. Food and Agricultural Organization of the United Nations: energy-smart food for people and climate. Available from: http://www.fao.org/energy/81344/ en/ [cited: 17 April]. Gourdon, R., Vermande, P., 1987. Effects of propionic acid concentration on anaerobic digestion of pig manure. Biomass 13, 1–12. Grobicki, A., Stuckey, D.C., 1991. Performance of the Anaerobic Baffled Reactor under steady-state and shock loading conditions. Biotechnol. Bioeng. 37, 344–355. Han, S.-K., Shin, H.-S., 2004. Biohydrogen production by anaerobic fermentation of food waste. Int. J. Hydrogen Energ. 29, 569–577. Iacovidou, E., Ohandja, D.-G., Voulvoulis, N., 2012. Food waste co-digestion with sewage sludge – realising its potential in the UK. J. Environ. Manage. 112, 267– 274. Ke, S., Shi, Z., Fang, H.H.P., 2005. Applications of two-phase anaerobic degradation in industrial wastewater treatment. Int. J. Environ. Pollut. 23 (1), 65–80. Lin, J., Zuo, J., Gan, L., Li, P., Liu, F., Wang, K., Chen, L., Gan, H., 2011. Effects of mixture ratio on anaerobic co-digestion with fruit and vegetable waste and food waste of China. J. Environ. Sci. 23 (8), 1403–1408. Liu, X.Y., Ding, H.B., Sreeramachandran, S., Stabnikova, O., Wang, J.-Y., 2008. Enhancement of food waste digestion in the hybrid anaerobic solid–liquid system. Water Sci. Technol. 57 (9), 1369–1373. Marchaim, U., Krause, C., 1993. Propionic to acetic ratios in overloaded anaerobic digestion. Bioresour. Technol. 43, 195–203.
Mohan, S., Bindhu, B.K., 2008. Effect of phase separation on anaerobic digestion of kitchen waste. J. Environ. Eng. Sci. 7 (2), 91–103. Nachaiyasit, S., Stuckey, D.C., 1995. Microbial response to environmental changes in an Anaerobic Baffled Reactor (ABR). Anton. van Leeuw. 67, 111–123. NEA, 2013. National Environment Agency, Singapore. Available from: http://app2. nea.gov.sg/energy-waste/waste-management/waste-statistics-and-overallrecycling [cited: 15 November]. Shanmugam, P., Horan, N.J., 2009. Simple and rapid methods to evaluate methane potential and biomass yield for a range of mixed solid wastes. Bioresour. Technol. 100, 471–474. Shen, F., Yuan, H., Pang, Y., Chen, S., Zhu, B., Zou, D., Liu, Y., Ma, J., Yu, L., Li, X., 2013. Performances of anaerobic co-digestion of fruit & vegetable waste (FVW) and food waste (FW): single-phase vs. two-phase. Bioresour. Technol. 144, 80–85. Sreela-or, C., Imai, T., Plangklang, P., Reungsang, A., 2011. Optimization of key factors affecting hydrogen production from food waste by anaerobic mixed cultures. Int. J. Hydrogen Energ. 36, 14120–14133. Tawfik, A., Salem, A., El-Qelish, M., Abdullah, A.M., Abou Taleb, E., 2011a. Feasibility of biological hydrogen production from kitchen waste via anaerobic baffled reactor (ABR). Int. J. Sust. Water Environ. Syst. 2 (2), 117–122. Tawfik, A., Salem, A., El-Qelish, M., 2011b. Two stage anaerobic baffled reactors for bio-hydrogen production from municipal food waste. Bioresour. Technol. 102, 8723–8726. U.S. EPA, 2001. Method 1684: total, fixed, and volatile solids in water, solids, and biosolids. EPA-821-R-01-015. Office of Water, Office of Science and Technology, Engineering and Analysis Division, U.S. Environmental Protection Agency, Washington, DC. Wang, J.-Y., Xu, H.L., Tay, J.-H., 2002. A hybrid two-phase system for anaerobic digestion of food waste. Water Sci. Technol. 45 (12), 159–165. Wang, J.-Y., Zhang, H., Stabnikova, O., Tay, J.-H., 2003. Removal of ammonia in a modified two-phase food waste anaerobic digestion system coupled with an aerated submerged biofilter. Waste Manage. Res. 21, 527–534. Zhang, C., Xiao, G., Peng, L., Su, H., Tan, T., 2013. The anaerobic co-digestion of food waste and cattle manure. Bioresour. Technol. 129, 170–176.
