Hydrogen and Methane Production from Food Residue Biomass Product 1 (FORBI)

10 This study concerns the production of hydrogen and methane from a Food Residue Biomass (FORBI) product [1], 11 generated from pre-sorted HFW in a CSTR and in a PABR respectively. FORBI is generated by drying and shredding 12 the fermentable fraction of household food waste collected door-to-door in the Municipality of Halandri, Greece. 13 Hydrogen production from FORBI through anaerobic fermentation under acidogenic mesophilic conditions was 14 carried out using a 4L CSTR, operated at 12 h HRT under an organic loading of 15 g TS·L -1 . The H 2 -CSTR was 15 operated for 40 days. During the operation of H 2 -CSTR the production of biogas reached up to 0,1026 L biogas ·g FORBI-1 16 and the percentage of hydrogen in the gas up to 48,2 %. 17 The conversion of FORBI into methane was carried out through the operation of a 77L PABR operated under 18 mesophilic methanogenic conditions at various operating parameters (OLR, HRT, T). Two different approaches were 19 adopted for the pre-treatment of the feedstock. For the two first phases of the experimental procedure, a liquid 20 extraction step was carried out before feeding the bioreactor with the separated liquid fraction, while in the subsequent 21 three phases, a whole suspension of FORBI was used as feed. The mean biogas production rate was 0,158±0,02 22 L biogas ·g FORBI -1 and the mean methane percentage in the biogas was 67,5±2,1%, in the first two phases. The mean biogas 23 production rate was 0,519±0,03 L biogas ·g FORBI -1 and the mean methane percentage in the biogas was 66±2,8%, when a 24 whole suspension of FORBI was fed to the PABR. The present work, pursued under the framework of the Horizon2020 project WASTE4think, focuses on investigating 27 the anaerobic digestion and dark fermentation potential of a biomass product (named FORBI: Food Residue Biomass) 28 for the production of biofuels (biomethane and biohydrogen respectively). The FORBI production process includes 29 drying and shredding of Household Fermentable Waste (HFW) and it is a promising idea since it significantly reduces 30 the mass and volume of the HFW, it can be stored for prolonged periods of time without deteriorating, odors are 31 eliminated and the product is homogenized. 32


37
The generation and disposal of MSW is dramatically increasing in the recent years due to the rapid population growth 38 (over 9 billion by 2050) [2] and modernization throughout the world. According to a recent report, published by the 39 World Bank, currently more than 1,3 billion tons of municipal solid waste are generated annually worldwide, while 40 waste generation is expected to exceed 2,2 billion tons in the next decade [3]. Breaking down the solid waste 41 generation data, it has been proved that 30-50% of the MSW is food waste, out of which 95% is ultimately landfilled. 42 These quantities of landfilled food waste reflect a great wastage of nutrient content and energy recovery potential, 43 while at the same time landfilled food waste is a major contributor to landfill methane and other GHG emissions. 44 Moreover, the Urban Development Series of World Bank suggests that "poorly managed waste has an enormous 45 impact on health, local and global environment and the economy. In addition, improperly managed waste usually 46 results in down-stream costs higher than what it would have cost to manage the waste properly in the first place" [3]. 47 Hence, policies have been developed recently, aiming at minimizing the amounts of fermentable waste ending up in 48 landfills 49 In Europe alone, 88 million tons of food is wasted, with an overall cost estimated at 143 billion Euros according to 50 the literature [4][5][6][7]. Scientific and technological developments offer a variety of valorization options and technologies 51 for the production of valuable chemicals, products and biofuels [8][9][10]. The exploitation of these -or some of these-52 options could drastically minimize the amount of food waste ending up in the landfills. 53 The current research work, pursued within the framework of WASTE4think, a Horizon2020 project, is based on work 54 previously done extensively studying anaerobic digestion and fermentation of food waste [8][9][10][11][12][13][14][15]. More specifically, 55 this paper focuses on the effectiveness and benefits of using pretreated food waste (dried and shredded) as a feedstock 56 for methane (CH4) production via anaerobic digestion and hydrogen (H2) production via dark fermentation. 57 WASTE4think proposes an innovative management approach of the HFW that includes source separation and separate 58 collection of this fraction in the Municipality of Halandri, followed by drying and shredding of the collected waste at 59 the Municipality level. The scope of the project is to evaluate the generated product, named FORBI (Food Residue  60 Biomass) as a potential feedstock for the production of biofuels, among various valorization alternatives. FORBI is a 61 high quality homogenized and dry biomass product weighing approximately 25% of the original food waste 62 collected[1], which may be stored for prolonged periods of time without deterioration. 63 The idea of implementing a pre-treatment step to improve the characteristics of the organic waste before using it as 64 an anaerobic digestion feedstock to enhance methane production is not new. There are a lot of research papers covering 65 the specific subject both for organic waste generally and food waste more specifically for methane [16][17][18] and 66 hydrogen [19] production. Moreover, the combination of thermal and mechanical pre-treatment has been evaluated in 67 the past [18], however not systematically and not in a homogenized food waste product like FORBI. 68 The scope of the current research paper is to investigate the effect of drying/shredding pre-treatment and pursue an 69 initial evaluation of FORBI as feedstock for the production of biomethane and biohydrogen, respectively. 70 Methane production was carried out using a Periodic Anaerobic Baffled Reactor, a novel high-rate bioreactor designed 71 to operate at high organic loading rates. The PABR resembles a simple ABR with the compartments configured 72 circularly. Variation of the switching frequency (or equivalently the switching period (T), i.e. the time for switching 73 the feed to all compartments) allows flexibility in the operation of the PABR. The PABR can be operated as a simple 74 ABR, if the switching frequency is set to zero, and, at relatively high switching frequency, as a single-compartment 75 upflow bioreactor [20][21][22][23]. 76

77
The measurements of tCOD and sCOD, TSS and VSS, total alkalinity and temperature were carried out according to 78 Standard Methods [24]. The pH and conductivity were measured using a digital pH-meter (WTW INOLAB PH720) 79 and conductivity meter (WTW INOLAB), respectively. For the quantification of VFAs, 1ml of sample acidified with 80 30μL of 20% H2SO4 was analyzed on a gas chromatograph (SHIMADZU GC-2010 plus) equipped with a flame 81 ionization detector and a capillary column (Agilent technologies, 30m x 0,53mm ID x 1μm film, HP-FFAP) and an 82 autosampler (SHIMADZU AOC-20s). The oven was programmed from 105 o C to 160 o C at a rate of 15 o C·min -1 and 83 subsequently to 225 o C (held for 3min) at a rate of 20 o C·min -1 . Helium was used as the carrier gas at 30ml·min -1 , the 84 injector temperature was set at 230 o C and the detector at 230 o C. For the quantification of the methane content of the 85 biogas, a GC-TCD was used (Shimadzy GC-2014). The separation column's (Carboxen 1000) length was 5 m and the 86 interior diameter 2,1 mm. The initial temperature of the GC-TCD was 40 o C. For the estimation of the methane content 87 a temperature programme was used (duration: 25 mins) during which the temperature was increasing 10 o C·min -1 until 88 reaching 185 o C and staying stable at this temperature for 5 minutes. The methane content then was calculated using 89 a standard calibration curve. The biogas production rate was measured using an oil displacement technique [20,21]. 90

91
FORBI is a biomass product generated by an alternative food waste management scheme. A GAIA  Waste Drying Machine developed by GAIA Corp. is fed with the raw food waste collected daily and after a 9-hour 93 drying/shredding process FORBI is collected from the dryer tank. 94 The proposed HFW management scheme offers a variety of benefits, since FORBI exhibits numerous advantages: 95 1. The volume and weight of the processed solid feedstock is reduced by as high as 75-80% compared to the 96 initial raw food waste. 97 2. A homogenized product is generated, with highly repeatable physicochemical characteristics (they don'tvary 98 appreciably) 99 3. No odors are emitted. 100 4. FORBI may be stored for prolonged periods without deteriorating. 101 A series of analyses had been conducted, in a previous study, to develop an in-depth characterization of FORBI [1]. 102 The results show,that FORBI exhibits high homogeneity. Moreover, the bulk density of FORBI was found to be as 103 high as 690 kg·m -3 . The basic results of the characterization of FORBI are shown in Table 2-1. 104 Moreover, an elemental analysis as well as an experimental determination of the NCV and GCV of FORBI are given 106 in Table 2-2 107 Gas and liquid samples were taken routinely and analyzed for hydrogen and methane content using the same method 115 as explained before. Total Suspended Solids (TSS), Volatile Suspended Solids (VSS), and Volatile Fatty Acids 116 (VFAs) were also estimated on a regular basis. The production of biogas was also measured. 117 During start-up, the CSTR was inoculated with 1 L of thermally treated (95 o C for 15 minutes) activated sludge and 118 was fed with an aquatic suspension of FORBI (15g FORBI·L -1 ). The bioreactor was started up in batch mode for 48 119 hours (data not shown). It was then operated in a continuous mode under an HRT (hydraulic retention time) of 12 120 hours using the same aquatic suspension of FORBI. The CSTR operated without pH regulation 121

122
The scope of the experimental procedure was to evaluate FORBI as an anaerobic digestion feedstock for methane 123 production using a pilot-scale PABR, of operating volume of 77L. Therefore, the bioreactor was operated at various 124 OLR and HRT outlined in Table 1. The PABR was equipped with sampling valves in every compartment, placed in 125 the middle-height of the compartment. The PABR consists of two concentric cylinders, the interior of which is filled 126 with water maintained at 35 o C through temperature control. 127 The PABR was initially operated with a HRT of 12,2d and a T of 2d, with an influent tCOD of 7250mg·L -1 for 86d 128 (phase #1). After a steady periodic state was reached, the HRT was decreased to 10d and the mean influent tCOD was 129 increased to 11690mg·L -1 for an operation period of 43d (phase #2). A liquid extraction step was used as pretreatment 130 during phases #1 and #2, feeding the bioreactor only with FORBI extract so as to keep the solids content of the feed 131 low, a general requirement for high-rate bioreactors. Initially, FORBI was suspended in water, in a proportion of 18 132 gFORBI·L -1 water and was vigorously stirred for 30 minutes. Then the slurry was filtered under pressure using a cloth 133 filter. The liquid phase (filtrate) retained 36,4% of the organic content of the waste during phase #1 and 29,3% during 134 phase #2. The solid phase collected was valorized for the production of compost. 135 After a steady periodic state was reached, the HRT was decreased to 8,7d and the PABR was fed with a mean influent 136 tCOD of 10760mg·L -1 (phase #3). During phase #3, a whole suspension of FORBI was used instead of the liquid 137 fraction, in order to see if FORBI hydrolysis is fast enough to sustain a high-rate anaerobic digestion without solids 138 separation. As the PABR proved to handle satisfactorily a whole FORBI suspension as a feed, for the next 139 experimental phase (phase#4) the HRT was decreased to 5d without changing the influent tCOD (tCOD =10830 mg·L -140 1 ). In phase #5, the HRT was maintained at 5d, while the tCOD in the feed was approximately doubled (22630 mg·L -141 1 ). The fifth phase (phase#5) was split into two sub-phases, investigating the response of the PABR when alternating 142 the switching frequency (T). More specifically, during phase#5.1 a sharp increase of the tCOD was implemented 143 leading to an influent tCOD of 22630mg·L -1 , while the HRT was not changed. Then, after 18 days of operation the 144 switching period was decreased from 2d to 1d without changing any other parameters (phase#5.2). 145

148
The CSTR was operated for 30 days. During the operation the production of biogas reached up to 0.1026 Lbiogas· 149 gFORBI -1 and the maximum percentage of hydrogen was 48,2%. 150 During the operation of the bioreactor, the biogas production rate was not constant Figure 3-1 (b), despite the fact that 151 the pH of the bioreactor remained 4,2-4,6 throughout the experiment Figure 3-1 (a). During the first 7 days, the 152 production rate of biogas increased from 0,65 L*Lbioreactor -1 *d -1 to 3,07 L*Lbioreactor -1 *d -1 . Afterwards, it decreased 153 significantly and was 1,37 L·Lbioreactor -1 ·d -1 on the average for the rest of the period. The concentrations of the main 154 metabolic products measured during the operation of the hydrogen producing reactor are presented in Figure 3-2. The 155 dominant metabolic products, measured, were acetic and butyric acids, which are common for biohydrogen producing 156 bioreactors [25]. The low concentrations of propionic acid indicate an efficient hydrogen production process, since 157 during the formation of propionate hydrogen is consumed [14,26]. 158 The decrease of the biogas production rate after the 7 th day of operation could be attributed to the consumption of 159 hydrogen by hydrogen consuming microorganisms which consume H2 and CO2 to produce acetate [25]. In order to 160 eliminate these hydrogen consuming microorganisms, the bioreactor was purged with air for one hour using an air 161 pump (arrows in figures 1a and 1b). After each purging, the biogas production rate increased significantly but not to 162 the level observed in the beginning of the operation. 163 164 A concise literature review was carried out to compare the obtained hydrogen production with others based on food 165 waste (Table 3-1). 166   The PABR exhibited great stability during all five phases of the process. During the stable periodic state of phase#1, 169 the mean tCOD removal rate was 89% (Figure 3-3 b) with a mean effluent tCOD concentration of 872 mg·L -1 . The 170 VSS remained below 0,5g L -1 (Figure 3-3 a) in all four compartments of the PABR during the whole phase #1. The 171 mean biogas production rate was 0,158 Lbiogas·gFORBI -1 and the mean methane composition of the biogas was 65-70%. 172 During phase #2, the OLR almost doubled from 0,59 gCOD·Lbioreactor -1 ·d -1 to 1,17 gCOD·Lbioreactor -1 ·d -1 , by decreasing 173 the HRT from 12,2d to 10d and by increasing the mean influent tCOD from 7320mg·L -1 to 11690mg·L -1 . The mean 174 tCOD removal rate was 93,5% ( Figure 3-3 b) and the mean biogas production rate was 0,11Lbiogas·gFORBI -1 . The pH 175 remained at optimum levels for methane production in all four compartments of the reactor during phases #1 and #2. 176 The mean methane composition of the biogas was 65-70%. 177 Subsequently, the reactor was fed with a whole suspension of FORBI and operated at an HRT of 8,7d (phase#3). The 178 feed TSS concentration was 10g·L -1 . The PABR responded efficiently to the change, while no problems were observed 179 by the high solids content, that is contained in FORBI's suspension (approximately 15g·L -1 ). The mean tCOD removal 180 rate was 86,4%. Feeding with a whole suspension of FORBI, with all its solids content, resulted in a substantial 181 increase of the mean biogas production rate to 0,56Lbiogas·gFORBI -1 . The mean methane composition of the biogas 182 remained in the range 65-70%. 183 In the next experimental phase (phase#4) the HRT was further decreased to 5 days, while the average tCOD of the 184 influent remained the same, leading to an organic loading rate of 2,14 gCOD·Lbioreactor -1 ·d -1 . The PABR handled very 185 well the decrease of the HRT. The average tCOD removal rate observed was 80,5% while the biogas productivity was 186 0,5 Lbiogas·gFORBI -1 , once a stable periodic operation was established. The average methane composition of the biogas 187 was 69,6%. 188 Finally, in the fifth phase of the experimental procedure (phase#5), the average feed tCOD was increased to 189 22630mg·L -1 leading to an organic loading rate of 4,53 gCOD·Lbioreactor -1 ·d -1 . Again, the bioreactor's stability was very 190 good, achieving an average tCOD removal of 85,6%, while slightly increasing the biogas productivity to 0,531 191 Lbiogas·gFORBI -1 . The methane content was stable, giving an average value of 62,7%. In this phase, the VSS (Figure 3-192 3 a) and the VFAs (Figure 3-5) of the bioreactor were slightly increased showing kinetic limitation at this loading rate. 193 Hence, the switching period was decreased from 2 days to 1 day to investigate the system's response to this change 194 to the operational parameters. The result of this change was that both the VSS and the VFAs showed a slight decrease, 195 without, however, reaching the very low levels of the previous phases. 196 197 198 (a) (b) 199

217
The present work concerns the production of Hydrogen and Methane from a Food Residue Biomass (FORBI) product, 218 generated from pre-sorted fermentable household waste in a CSTR and in a Periodic Anaerobic Baffled Reactor 219 (PABR) respectively. FORBI is generated by drying and shredding the fermentable fraction of household food waste 220 collected door-to-door in the Municipality of Halandri. 221 Operating at an HRT of 12 hours, the maximum production of hydrogen from FORBI was 0,1026Lbiogas·gFORBI -1 with 222 a hydrogen concentration of 48,2%. Homoacetogenesis seems to have reduced the amount of hydrogen produced. 223 Short aeration of the reactor has allowed temporary increase in the biogas production rate, but it is clear that further 224 work is necessary to secure a stable and optimal hydrogen production. 225 The PABR proved an excellent high rate reactor for methane production from FORBI. The reactor was capable of 226 handling successfully a suspension of FORBI, without the need for solids removal, yielding on the average 227 0,519Lbiogas·gFORBI -1 and 66% methane content. Some kinetic limitation has appeared at an organic loading of 4,53 228 gCOD·Lbioreactor -1 ·d -1 . 229 Compared to other published works, drying and shredding food waste has proved to be a promising method in terms 230 of maximizing biogas productivity and especially when it comes to biomethane production, through anaerobic 231 digestion. Hence, FORBI apart from its benefits regarding mass and volume reduction, homogeneity and stability, is 232 an excellent feedstock for either a dark fermentation or an anaerobic digestion process. 233