This is the accepted version of the journal article: Mora Garrido, Mabel; Lafuente Sancho, Francisco Javier; Gabriel, David. «In- fluence of crude glycerol load and pH shocks on the granulation and microbial diversity of a sulfidogenic Upflow Anaerobic Sludge Blanket reactor». Process Safety and Environmental Protection, Vol. 133 (January 2020), p. 159-168. DOI 10.1016/j.psep.2019.11.005 This version is available at https://ddd.uab.cat/record/266961 under the terms of the license 1 Influence of crude glycerol load and pH shocks on the 2 granulation and microbial diversity of a sulfidogenic Upflow 3 Anaerobic Sludge Blanket reactor M. Mora, J. Lafuente, D. Gabriel* GENOCOV Research Group, Department of Chemical, Biological and Environmental Engineering, Escola d’enginyeria, Universitat Autònoma de Barcelona, 08193 Bellaterra, Spain (*corresponding author: David.Gabriel@uab.cat) 4 5 ABSTRACT 6 Bioscrubbers are an environmental-friendly alternative to valorize SO2 contained in flue 7 gases to obtain elemental sulfur as final value-added product. The bottleneck of a SO2 8 bioscrubber relies on the heterotrophic reduction of the absorbed SO2 to obtain sulfide. 9 In this study, the performance and stability of a sulfidogenic Upflow Anaerobic Sludge 10 Blanket reactor (UASB) using crude glycerol was investigated during 6 months of 11 operation under variable organic loading rates. The UASB presented a maximum 12 elimination capacity and a sulfate removal efficiency of 110 mg S-SO42- L-1 h-1 and 100%, 13 respectively, when the chemical oxygen demand to sulfate ratio (COD/S-Sulfate) was 8.5 14 g O2 g-1 S-SO42-. The intermediate compounds identified from crude glycerol degradation 15 were mainly propionic and acetic acid, which varied along the sludge bed together with 16 the pH. Microbial diversity analyses of the sulfidogenic granular sludge showed that the 17 most abundant sulfate reducing genera were Desulfovibrio spp. and Desulfobulbus spp. 18 Methanosaeta and other fermentative/acidogenic microorganisms were also found in 19 significant amounts. Particle size distribution analyses showed that biogas production 20 allowed the granulation of the sulfidogenic sludge, which had an average particle size 21 ranging from 729.3 µm (lower part of the bed) to 391.7 µm (upper part of the test). A 22 short-term pH shock caused a detrimental effect over the system performance due to 23 degranulation. Concomitantly, biogas production was interrupted and acetic acid was 24 accumulated also causing a significant impact on microbial diversity. Unclassified 25 Clostridiales, Desulfovibrio spp. and Desulfobulbus spp. showed a higher resistance to 26 pH shocks. 27 28 Keywords: Crude glycerol, sludge bed stratification, SO2 valorization, sulfidogenesis, 29 SRB, UASB 30 31 32 1 33 1. INTRODUCTION 34 Combustion of sulfur-containing fuels results in SO2 formation, which cause detrimental 35 impacts on human health and the environment such as respiratory illness and acid rain, 36 respectively, among others (Srivastava and Jozewicz, 2001). During the last decades, the 37 SO2 release to the atmosphere in Europe has been limited by different directives and 38 decreased due to the implementation of treatment systems. However, anthropogenic 39 worldwide emissions of SO2 in 2011 were around 100 Tg mainly generated in the 40 energetic and industrial sectors (Klimont et al., 2013). The treatment of SO2 emissions 41 from flue gases performed through physical-chemical treatments such as absorption are 42 expensive and generate effluents that require further processing as limestone forced 43 oxidation or furnace sorbent injection (Srivastava and Jozewicz, 2001; Philip and 44 Deshusses, 2003). 45 Biological processes offer advantages with respect to physical-chemical ones such as 46 reduced operational costs and wastes generation. The circular economy culture is also 47 driving biological processes development towards waste valorization (Scherson and 48 Criddle, 2014). Some authors have proposed alternatives for the valorization of SO2 from 49 flue gases. However, many of them pose significant drawbacks such as the use of 50 expensive carbon sources for sulfate reduction (such as glucose, volatile fatty acids or 51 alcohols), the use of immobilized biomass for sulfur production (hindering elemental 52 sulfur recovery) or the use of iron or nitrate for sulfide oxidation (increased operational 53 costs) (Gasiorek, 1994; Philip and Deshusses, 2003; Qian et al., 2015). Considering that 54 elemental sulfur is a valuable compound actually obtained from non-renewable sources, 55 the two-step bioscrubbing process was envisaged as a promising alternative to valorize 56 SO2 from flue gases as elemental sulfur if a valuable waste organic source such as crude 57 glycerol is used (Mora et al., 2016). The two-step bioscrubbing process consists of a first 58 absorption stage with a slightly alkaline solution followed by a heterotrophic Upflow 59 Anaerobic Sludge Blanket (UASB) reactor for the reduction of sulfate to sulfide in series 60 with downstream bioreactor for partial oxidation of sulfide to elemental sulfur (see Figure 61 S1 in Supplementary Material). 62 Crude glycerol is a chemical oxygen demand (COD) rich by-product produced in large 63 quantities from the soap and biodiesel manufacturing processes (Dinkel et al., 2010a). In 64 the literature there are some studies assessing the chemical and/or biological conversion 65 of crude glycerol into more valuable products (Yazdani and Gonzalez, 2007). Anaerobic 2 66 digestion of crude glycerol to produce biogas is the most popular or well-known 67 alternative to obtain energy from this attractive waste (Lopez et al., 2009). Santos et al. 68 (2018) recently reported that the use of crude glycerol promotes sulfate-reducing 69 communities growth while Mora et al. (2018) also demonstrated that crude glycerol is a 70 powerful organic source to promote sulfidogenesis under certain conditions. Crude 71 glycerol use in sulfidogenic reactors has been described for recovering precious or 72 valuable metals, for immobilizing heavy/toxic metals or simply to study and get 73 knowledge about the competition between methanogens and sulfate reducers (Dinkel et 74 al., 2010b; Bertolino et al., 2014; Santos et al., 2017). On the contrary, the use of crude 75 glycerol in sulfidogenic UASB reactors towards elemental sulfur recovery is scarce. 76 The limiting step in SO2 bioscrubbing is the heterotrophic reduction of sulfate to sulfide 77 (Mora et al., 2016). The presence of organic matter promotes the methanogenic activity 78 and the production of biogas (Thirugnanasambandham et al., 2014; Sridhar et al., 2016; 79 Thirugnanasambandham et al., 2016; Thirugnanasambandham 2017), which is really 80 positive to enhance and maintain the granulation of the sludge due to the shear stress that 81 biogas circulation generates (Liu and Tay, 2002). However, the generation of biogas in 82 sulfidogenic reactors is not advisable since sulfide strips out from the liquid to the gas 83 phase resulting in generation of a highly corrosive biogas. Then, the production of biogas 84 has to be minimized by increasing the COD selectivity towards sulfate reduction. In 85 general, literature focuses on the enhancement of methane production avoiding 86 sulfidogenesis or at least the presence of the produced sulfide in the biogas. Many studies 87 have reported the start-up and operation of sulfidogenic UASB reactors (Alphenaar et al., 88 1993; Jing et al., 2013). Nonetheless, there is a deficiency of studies related with 89 sulfidogenesis promotion against methane production and a few target the reduction of 90 sulfate in UASB reactors with crude glycerol under switching conditions such as variable 91 carbon loads and pH shocks. There is even less information about the bed stratification 92 in terms of bioreactions progress, granulation and microbial diversity. This sort of 93 research is particularly needed since granular sludge is often used, not only because it 94 allows enhancing the distribution of the liquid phase and the settling capability of sludge 95 but because the granulation of sulfate reducing bacteria at low up-flow velocities is not 96 feasible. 97 The aim of the present study was to assess the performance and stability of a sulfidogenic 98 UASB, inoculated with methanogenic granular sludge, under different conditions in terms 99 of sludge and liquid phases stratification in the sludge blanket, microbial diversity 3 100 distribution and sulfur mass balance. The SO2 absorption stage and the partial oxidation 101 of sulfide into elemental sulfur, which are part of the two-step bioscrubber concept, were 102 out of the scope of the present study. 103 104 2. MATERIALS AND METHODS 105 2.1 Experimental setup for sulfidogenesis promotion in a UASB reactor 106 The experimental setup used in this study to promote sulfidogenesis from sulfate rich 107 effluents is presented in Figure 1. It consisted of a Upflow Anaerobic Sludge Blanket 108 (UASB) type reactor made of glass with a reaction volume of 1L and a total volume of 109 2L. An UASB reactor configuration was selected since UASBs are well-known, versatile, 110 work with high biomass densities and face well product inhibitions, overall leading to 111 larger volumetric productivities compared to other reactor configurations. The reactor 112 was divided in two different sections, i.e. a lower section (riser) containing the sludge 113 blanket or biological bed, with an internal diameter (D) of 0.05 m, and the upper section 114 acting as biomass, liquid and gas separator (D=0.11 m). Mineral medium was pumped 115 from bottom to top of the reactor. The composition of the mineral medium was (g L-1): 116 NH4Cl (0.5), K2HPO4 (3.5) and Na2SO4 (1) dissolved in tap water and adjusted to 117 pH=8.8-9.0 with NaOH (2 M). A total flow rate of 0.5 L h-1 was pumped into the system 118 (up-flow velocity of 0.25 m h-1), which consisted of a mixture of the mineral medium and 119 crude glycerol (ecoMotion S.A., Spain). The flow rate of the influent was set at 1.2 mL 120 h-1. Hence, crude glycerol was diluted to adjust the inlet COD concentration. The 121 hydraulic residence time (HRT) considering the reaction volume (riser) was 2h. T and pH 122 in the UASB reactor were not controlled but monitored. Inlet pH ranged from 8.8 to 9.0. 123 A 5 L Tedlar bag (SKC Inc.) was used to collect and analyze the composition and flow 124 of the biogas produced. Inlet and outlet flows were also sampled every two to three days 125 to analyze COD, sulfur species and volatile fatty acids (VFA). 126 127 2.2 Start-up and operating conditions of the process 128 The UASB was inoculated with 50 g of settled granular sludge from an anaerobic digester 129 treating wastewater in a paper recycling company. The operation of the UASB reactor 130 lasted for more than 6 months. Operating conditions at start-up were set at a COD/S- 131 Sulfate ratio of 2.5 g O2 g-1 S-SO42- since it has been proven that a COD limitation allows 132 the development of sulfidogenic bacteria while higher COD/S-Sulfate ratios enhance a 133 fast development of fermentative, acidogenic and methanogenic microorganisms that 4 134 compete with sulfidogenic heterotrophic bacteria (Choi and Rim, 1991; Hu et al., 2015). 135 Sulfate inlet concentration was set at 220 mg S-SO42- L-1 while crude glycerol was fed at 136 four different organic loading rates (OLR), which led to operate under five different 137 COD/S-Sulfate ratios (Table 1). A 24h pH shock was provoked on day 120. After 138 resumption, a lower HRT was obtained during the last operation period (140 d to 200 d) 139 since part of the sludge was progressively lost during the previous period (days 120 to 140 140) obtaining a lower reaction volume (blanket) in the UASB. A minimum of 30 days 141 between each load change were established. 142 143 2.3 Analytical methods 144 Sulfate (SO42-) and thiosulfate (S2O32-) concentrations were analyzed off-line by ion 145 chromatography using suppressed conductivity detection (ICS-2000, Dionex 146 Corporation). Chemical Oxygen Demand (COD) was analyzed off-line using COD 147 commercial kits (LCK314 and LCK1414, Hach LTD) and a spectrophotometer (DR2800, 148 Hach). A SenTix® 82 probe (WTW) connected to a benchtop meter (Inolab® Multi 740, 149 WTW) was used for routine pH measurements. Sulfide concentration was analyzed off- 150 line with a sulfide selective electrode connected to a benchtop meter (Symphony, VWR) 151 after conditioning the samples with sulfide antioxidant buffer (SAOB). The SAOB 152 composition was (g L-1): ascorbic acid (35) and EDTA (67) dissolved in NaOH (2M). 153 VFA were analyzed by gas chromatography (7820A, Agilent Technologies) equipped 154 with a DB-FFA column and using a flame ionization detector (FID) with helium as carrier 155 gas. Prior to VFA analyses, samples were prepared following the procedure described in 156 Baeza et al. (2017). 157 CH4, CO2 and H2S were analyzed by gas chromatography (HP 5890, Hewlett Packard) 158 equipped with a Porapack Q column and a thermal conductivity detector (TCD) with 159 helium as carrier gas. The volume of the gas produced in the UASB reactor was analyzed 160 following the Gas Bag Method (GBM) presented in Ambler and Logan (2011) using gas 161 chromatography (7820A, Agilent Technologies) equipped with an HP-Mole Sieve 162 column, a thermal conductivity detector (TCD) and using argon as the carrier gas. The 163 GBM procedure consisted basically of measuring the initial composition of the collected 164 gas in the bag, adding a known volume of tracer gas (nitrogen gas in this case) and 165 analyzing the new composition. From these two analyses (before and after the tracer 166 injection) the initial volume of gas could be calculated from mass balances (Baeza et al., 167 2017). 5 168 169 2.4 Granular sludge bed stratification and sequencing analysis 170 The UASB performance was also assessed at different bed heights (5, 25 and 44 cm). 171 Granulation of the anaerobic granular from the liquid phase was studied on days 0, 10, 172 60, 110 and 200 after startup while the supernatant of the samples was also analyzed for 173 sulfide, sulfate, COD and VFAs on days 110 and 200. To this aim, samples were 174 centrifuged at 8000 rpm during 2 min. The particle size distribution (PSD) was measured 175 by a laser particle size analysis system (Malvern MasterSizer Series 2600, Malvern 176 instruments Ltd., UK). Analysis of inorganic elements in the granules was also performed 177 to study the composition of the granular sludge and the formation of inorganic 178 precipitates. An Inductively Coupled Plasma-Mass Spectrometry (ICP-MS 7500 CE, 179 Agilent) was used. Previously, samples were washed and centrifuged twice (8000 rpm, 2 180 mins), dried at 105ºC during 24h and digested with nitrohydrochloric acid in a microwave 181 digester (Ultrawave, Milestone). Total solids and volatile solids concentrations in the bed 182 were also analyzed at different heights following the standard method of analysis (APHA, 183 2005). Microbial diversity was assessed on days 0, 110 and 200 by extracting and 184 sequencing the deoxyribonucleic acid (DNA) following the procedure described in Mora 185 et al. (2018). 186 187 3. RESULTS 188 3.1 Performance of the sulfidogenic UASB 189 Figure 2 shows the profiles obtained during the monitoring of the main species involved 190 in the process. As can be observed in Figure 2A, after 15 days of operation the removal 191 efficiency (RE) of sulfate (S-RE) was 42.1%, which corresponded to an elimination 192 capacity of 55.8 mg S-SO42- L-1 h-1. From this point until day 50, the average S-RE was 193 34.3±3.5% while the sulfur imbalance was around 10% due to the incomplete recovery 194 of sulfate as sulfide (Figure 2D). Regarding crude glycerol, the average RE of COD 195 (COD-RE) along this first period of 50 days was 76.9±7.0%. Another important aspect 196 observed from the first performance period was that the organic matter degraded was 197 mainly destined to reduce sulfate, nevertheless, almost 25% of the organic matter was 198 converted into methane (a methane production of 21.2±6.7 mL CH4 h-1 was obtained with 199 a purity up to 95-99%). As can be observed in Figure 2C, VFA were not present in the 200 effluent indicating that VFA were produced and completely utilized to reduce sulfate and 6 201 to produce methane. In Table 2 the summary of S-RE and COD-RE resulting from all the 202 UASB operation periods is presented. 203 During the second period (50 – 90 days) COD inlet concentration was increased (Table 204 1) to reach a theoretical COD/S-Sulfate ratio of 7.0 g O2 g-1 S-SO42-. The previous ratio 205 tested (2.5 g O2 g-1 S-SO42-) was not high enough to completely reduce sulfate to sulfide 206 although was adequate to avoid a high methanogenic activity. The sulfur imbalance in 207 this second operation period was more than 20% (Figure 2D), while the H2S content in 208 the gas was less than 5% of the inlet sulfate (data not shown). The OLR set during this 209 second period of operation (804±134 mg O2 L-1 h-1) allowed reaching S-RE and COD- 210 RE of 83.0±12.8% and 80.8±5.2%. Methane production increased up to 100 mL h-1, 211 corresponding to 38.6% of the degraded COD, and the biogas produced had a CH4 content 212 of 89.8±4.0% (data not shown). As can be also observed in Figure 2C (second operation 213 period from day 50 to 90), there was an instability of biogas production. At this point, 214 biomass accumulated into the gas collector did not allow the gas produced to flow to the 215 Tedlar bag, thus causing an inefficient separation of the gas phase in the upper part of the 216 reactor. 217 During the following operating period (90 - 110 days) the COD/S-Sulfate ratio set (8.5 g 218 O2 g-1 S-SO42-) allowed obtaining a successful S-RE. An average S-RE of 97.1±2.9% was 219 reached (Figure 2A and Table 2). Compared to the previous operating period, the COD- 220 RE and the COD elimination capacity (COD-EC) dropped drastically (from around 645 221 mg O2 L-1 h-1 to around 415 mg O2 L-1 h-1) which caused an increase of the COD 222 concentration in the effluent. A better collection of the biogas was obtained through the 223 addition of nitrogen pulses in the mineral medium inlet. At the end of this period, the 224 stratification study of the bed was also performed (day 110) to elucidate the processes 225 occurring along the sludge bed. 226 A COD/S-Sulfate ratio of 4.5 g O2 g-1 S-SO42- was set from day 110 until day 140 to 227 reduce the excess of organic matter in the effluent of the reactor. The inlet glycerol 228 concentration was reduced from 1891±370 mg O2 L-1 to 1026±90 mg O2 L-1 to obtain a 229 COD load (524±47 mg O2 L-1 h-1) slightly higher than the COD-EC of the previous period 230 (415 mg O2 L-1 h-1). The reduction of the inlet COD/S-Sulfate affected negatively the 231 UASB performance and caused the increase of sulfate and COD concentrations in the 232 effluent up to 55 mg S-SO42- L-1 and 636 mg O2 L-1, respectively. Moreover, from day 233 120 to day 121 a pH shock of 24h was provoked. The value of the pH in the mineral 234 medium was set to pH=12.5. Consequently, the S-RE and COD-RE decreased down to 7 235 57.6±13.5% and 35.3±6.6%, respectively. The pH shock caused the sludge blanket 236 migration and rupture due to the sludge degranulation (Figure S2 in Supplementary 237 Material). Consequently, the biomass rising caused a loss of 30% of sludge bed volume 238 during this operating period. Because of the bed instability, 100 mL of activated carbon 239 were added on day 140 in the lower part of the UASB to procure a better liquid 240 distribution. After the addition of activated carbon, S-RE was recovered and sulfide was 241 produced with a minimal percentage in the S imbalance (Figure 2D). However, 242 preferential paths were created again in the upper part of the bed leading to a S-RE 243 increase the last 25 days of operation. 244 245 3.2 Assessment of the sludge bed stratification 246 Operating conditions set in the reactor (particularly the up-flow velocity) caused the 247 stratification of the sludge bed, which naturally happens in plug-flow reactors such as 248 UASB reactors. This phenomenon creates a set of layers with different particle size, 249 microbial diversity and chemical compounds distribution along the sludge bed that were 250 analyzed to assess the process performance at different reactor heights. 251 252 3.2.1 Granular size distribution along the bed height 253 Figure 3 shows the particle size distribution at different bed heights at each sampling time 254 compared to that of the inoculum. The PSD 10 days after the start-up of the process 255 revealed that the granules size of the inoculum was maintained in the first part of the bed 256 (5 cm). However, granulation was almost lost at H=25 (half reactor, Figure S2A) since 257 half of the particles had an average diameter lower than 215 µm (i.e. D(0.5)=215 µm), 258 which is around the limit to consider a biomass aggregate as a granule (200 µm). 259 Granulation recovered in all bed heights between days 60 and 110 coincident with the 260 increase of the OLR and subsequent production of biogas (Figure 2C). The PSD revealed 261 that half of the granules measured at H=5 cm had a diameter lower than 729 µm while at 262 H=25 cm and H=44 cm the D(0.5) was below 392 µm. However, after the pH shock (days 263 120 and 121) the granulation was completely lost and many operational problems 264 occurred until the end of the operation, as migration events of the bed (Figure S2B) or the 265 creation of preferential paths due to the rupture of the bed (Figure S2C). 266 267 3.2.2 Stratification of chemical species along the sludge blanket 8 268 Figure 4 shows the profiles of sulfate, sulfide, VFA and pH along the sludge bed on days 269 110 (Figure 4A) and 200 (Figure 4B). After 110 days of operation, when S-RE was close 270 to 100%, sulfate and sulfide profiles indicated that sulfidogenesis was not occurring in 271 the lower part of the UASB since sulfate was not degraded at H<5cm. In contrast, crude 272 glycerol was rapidly hydrolyzed and mainly converted into propionic acid and acetic acid. 273 The pH also decreased concomitantly with VFA production. From H=25cm upward, 274 acetic acid was consumed while the propionic acid concentration was maintained. The 275 concentration of VFA in the effluent of the UASB was over 700 mg O2 L-1 and sulfate 276 was not completely recovered as sulfide. At this point the sulfur imbalance was 38%. 277 After 200 days of operation (Figure 4B), stratification of the bed had a completely 278 different pattern from that obtained when the operation of the UASB did not present any 279 problem (Figure 4A). In this case, an incomplete reduction of sulfate was observed while 280 the S imbalance obtained was lower than 10%. Interestingly, the VFAs accumulated 281 switched with respect to the previous operation period from propionic acid to acetic 282 (predominant) and butyric acid. 283 284 3.2.3 Microbial diversity developed in the sufidogenic UASB 285 The inoculum, the sludge bed after 110 days of operation at 3 bed heights and a single 286 sample at the end of the operation were analyzed for their microbial diversity. Figure 5 287 shows the percentages of main genera identified in the samples. Most of the genera 288 identified in the inoculum were mainly fermentative bacteria and acidogenic bacteria such 289 as Syntrophobacter (9.6%), Clostridium (3.3%), Smithella (3.0%), Bacteroides (2.7%), 290 Pseudomonas (1.1%), Saccharofermentans (1.0%) besides Unclassified Clostridia 291 (6.6%), Unclassified Anaeroliniales (2.0%) or the phylum Unclassified Firmicutes 292 (1.8%). Methanosaeta (15.9%) and Methanobacterium (1.0%) were the main 293 methanogenic microorganisms identified. These genera of archaea and bacteria are 294 typically found in sludge from anaerobic digesters (Ariesyady et al., 2007). On the 295 contrary, the SRB identified genus Desulfobulbus represented only 1.1% of the culture, 296 which can grow using organic matter to produce acetate and/or using H2 as electron donor. 297 It must be mentioned that also Unclassified Syntrophobacterales (4.3%) have been 298 described to reduce sulfate (Liu et al., 1999), which would enlarge the relative percentage 299 of SRB in the inoculum. 300 After 110 days of operation, the microbial diversity changed completely. A decrease in 301 the diversity from 22 to 16 identified genera compared to the inoculum was found. The 9 302 genus Anaerosinus, Citrobacter, Eubacterium, Klebsiella, Dysgonomonas and 303 Trichococcus were the most abundant genera along the sludge bed of the UASB on day 304 110. The abundances of Unclassified Clostridiales and also Methanosaeta were higher 305 than 5% in all bed heights. With regard to sulfate reducers, Desulfovibrio was the most 306 abundant genus ranging from 6.4% (upper part of the sludge bed) to 2.9% (lower part of 307 the sludge bed). Desulfobulbus was also found with abundances lower than 1.5%. After 308 the reduction of the organic load and the pH shock (120 days), the genera found in the 309 UASB were mainly Desulfobulbus (19.8%), Desulfovibrio (8.5%) and Unclassified 310 Clostridiales (45.2%). 311 312 4. DISCUSSION 313 4.1 Assessment of the sulfidogenic UASB performance 314 Figure 2A shows that complete sulfate reduction was only achieved in the third operating 315 period (days 90 to 110 of operation) when the inlet COD/S-Sulfate ratio was 8.5 g O2 g-1 316 S-SO42-. Incomplete sulfate removal was found at inlet COD/S-Sulfate ratios below 8.5 g 317 O2 g-1 S-SO42- despite COD was not completely removed. In fact, complete mineralization 318 of glycerol was never reached. As reported by Hansen et al. (2009) and Hu et al. (2012) 319 crude glycerol can contain more than 50% of organic matter different from glycerol and 320 methanol such as fatty acid methyl esters, free or long chain fatty acids and glycerides, 321 which are complex compounds difficult to be hydrolyzed and even inhibitory (Chen et 322 al., 2008). Thus, the COD consumed probably corresponded to the hydrolysable and 323 readily biodegradable fraction of crude glycerol (Figure 2B), that limited sulfate removal 324 at inlet COD/S-Sulfate ratios below 8.5 g O2 g-1 S-SO42-. The low hydraulic residence 325 time set in this study (HRT=2h) also contributed to reduce the capacity of the reactor to 326 hydrolyze the slowly biodegradable fraction of crude glycerol. Based on the COD fed and 327 the remaining COD in the first and second operating periods, the slowly biodegradable 328 fraction of the crude glycerol used herein was estimated to be around 15 to 20%. 329 When COD was not limiting, SRB activity resulted in a maximum S-EC of 110 mg S- 330 SO42- L-1 h-1, which is 2.2-fold that obtained by Bertolino et al. (2014) using glycerol as 331 the carbon source and comparable to that obtained by Celis-García et al. (2007) using 332 lactate and VFAs as carbon sources, which are readily biodegradable organic compounds. 333 However, complete sulfate removal was achieved at expenses of producing an effluent 334 containing an excess COD of around 1100 mg COD L-1 (Figure 2B). Results indicate that 10 335 improvement of the hydrolytic capacity of crude glycerol is warranted to increase the 336 sulfate reduction capacity of the UASB and, concomitantly, to reduce the electron 337 acceptor consumption (generally oxygen) in the downstream processing of the UASB 338 effluent. 339 Biogas was always produced until the pH shock occurred. The high content of methane 340 in the biogas produced before the pH shock was obtained not only due to the high pH set 341 in the UASB but to the presence of both acetoclastic and hydrogenotrophic methanogens 342 capable of converting acetate and hydrogen and CO2, respectively, into methane. As can 343 be also observed from VFA and methane profiles (Figure 2C), the pH shock certainly 344 caused a severe effect on methanogenic activity since methane production was 345 completely interrupted. Acetate accumulation coincided with methanogenic activity 346 cease, which caused acetate concentration peaks up to 650 mg O2 L-1. This result indicated 347 that acetate was consumed by acetoclastic methanogens and that heterotrophic sulfate 348 reducers were probably not consuming acetate to reduce sulfate as reported by other 349 authors (Qatibi et al., 1998; Dinkel et al., 2010b; Postgate, 2013; Santos et al., 2017). In 350 addition, and considering that around 15-20% of crude glycerol was slowly 351 biodegradable, an average outlet acetate concentration of 510 mg COD·L-1 (Figure 2C) 352 coupled to an outlet total COD concentration of around 700 mg COD·L-1 at the end of the 353 UASB operation (Figure 2B) indicated that acidogenesis and acetogenesis were not 354 severely affected by the pH shock. 355 During the UASB performance many operational problems occurred due to the granular 356 sludge rising, before and after the pH shock. It was suspected that crude glycerol 357 diminished the density of the biomass. Viana et al. (2012) also observed severe sludge 358 flotation in anaerobic reactors fed with crude glycerol due to the presence of long chain 359 fatty acids. Moreover, after the pH shock, biogas production disappeared, and so the shear 360 stress that contributes to granulation and, consequently, to the settling capability of the 361 sludge. Overall, the pH shock had detrimental effects on the UASB performance that 362 could not be reversed 80 days after the pH shock. The addition of active carbon or the 363 injection of nitrogen pulses were only temporarily effective. The loss of part of the bed 364 and the rupture of the bed caused a sharp decrease of the contact time of the liquid phase 365 in the sludge bed that led to the incomplete reduction of sulfate. In Figure 4B the effect 366 of the preferential paths formed from the middle of the bed (Figure S2C) are clearly 367 represented and also how the lower part of the bed had a softer effect due to the use of 368 activated carbon in the inlet of the UASB. The complete degranulation of the sludge and 11 369 methanogenic activity disappearance also affected the S-RE and COD-RE that decreased 370 down to 60.4% and 25.1%, respectively, at the end of the UASB operation (Table 2). 371 Other authors have also reported the operation difficulties of sulfidogenic reactors with 372 low methane production, which may cause sludge bed clogging due to the lack of shear 373 stress (Tsui et al., 2016). Results confirmed the beneficial effects of using granular 374 biomass in sulfidogenic UASB reactors and the benefits of a minimum production of 375 biogas. 376 377 4.2 Sulfur mass balance 378 The recovery of sulfide from sulfate in the sulfidogenic UASB was of high importance 379 since the aim of this study was the valorization of sulfate rich effluents to obtain elemental 380 sulfur in a microaerobic bioreactor located downstream of the UASB. If sulfate is 381 converted into other compounds different from sulfide, the productivity decreases 382 compromising the technical and economic feasibility of the process. 383 After avoiding a slight air diffusion in the first operating period on top of the UASB 384 headspace, still the sulfur imbalance was over 20% for the second and third operating 385 periods, when the OLR was higher compared to that of the first period. In the third 386 operating period, crude glycerol and biomass samples at three different heights of the 387 UASB (lower, medium and upper part of the riser) were analyzed for metals in order to 388 check if there were metallic sulfides precipitating in the reactor. Results did not reveal 389 any compound in crude glycerol that could precipitate with sulfide to result in such a high 390 percentage of imbalance. Also, biomass samples had less than 4 mg S g-1 solid. Moreover, 391 crude glycerol used in this study had large quantities of S, P, K, Na and Mg. These 392 findings all together indicated that organic sulfur compounds were probably produced as 393 intermediate compounds of the biodegradation process leading to higher sulfur 394 imbalances at higher COD loads. As reported by Andersson et al. (2004) volatile organic 395 sulfur compounds are typically produced from the anaerobic digestion of sulfur 396 containing organic wastes, which is the case of crude glycerol and the bioprocess studied 397 in this work. It would also be in concordance with VFAs concentration in the effluent, 398 which represented around 75-80% of its COD content (Figure 2B and 2C). Then, other 399 organic compounds (with or without sulfur) could represent the remaining 20-25% of the 400 COD in the effluent. Further analyses are warranted to assess the organosulfur 401 compounds formed and their potential impact on biogas composition as well as on the 402 emissions in downstream reactors. 12 403 404 4.3 Assessment of the sludge bed stratification 405 Results obtained from the monitoring of liquid and gas phases presented above allowed 406 generating some hypothesis about the processes that could be ongoing in the sludge bed. 407 However, the analysis of the microbial diversity was essential to determine which of the 408 hypothesis were unlikely. Assessment of the sludge bed stratification also helped linking 409 sludge granulation, microbial diversity and the presence of different compounds (such as 410 sulfide, sulfate, VFA and pH) to identify the main mechanisms occurring along the sludge 411 bed in each operation stage analyzed. 412 In terms of granulation, the performance of the sulfidogenic UASB was complex since 413 the up-flow velocity set was lower (0.25 m h-1) than the minimal velocity of 0.5 to 1 m h- 1 414 reported to keep the sludge granulated (Pol et al., 1983). The rationale behind starting 415 up the UASB with methanogenic granular sludge in this study was the low availability of 416 granular sulfidogenic sludge in full-scale systems. Thus, the closer type of granular sludge 417 for sulfidogenic UASB inoculation in practice is that from granular methanogenic 418 UASBs. Figure 3 shows the fast degranulation of part of the bed that occurred during the 419 first 10 days of the UASB operation indicating that the development of sulfate reducing 420 bacteria and the presence of sulfide besides the low up-flow velocity set in the UASB 421 interfered in granules preservation. After 110 days of operation, the high OLR set allowed 422 reaching maximum D(0.5) in each bed height, which indicated that methanogenic activity 423 was present along the whole bed even at sulfide concentrations above 100 mg S L-1, 424 contributing to sulfidogenic sludge granulation. As reported by Liu and Tay (2004), high 425 OLR sustains fast microbial growth and drives to a rapid granulation and stable treatment 426 process. Microbial diversity analysis on day 110 showed that the most abundant genera 427 in the sludge bed (Anaerosinus, Citrobacter, Eubacterium, Klebsiella, Dysgonomonas 428 and Trichococcus) were able to convert glycerol, or its fermentation products, into 429 metabolites such as 3-hydroxypropionaldehyde, 1,3-propanediol, ethanol, lactate, 430 hydrogen and VFAs (mainly acetate and propionate) among others (Schauder and Schink, 431 1989; Homann et al., 1990; Yazdani and Gonzalez, 2007; van Gelder et al., 2012). These 432 compounds could be easily consumed by SRB to produce sulfide. A larger abundance of 433 Desulfovibrio and Desulfobulbus were identified in the upper part of the UASB rather 434 than in the lower part. This result was in concordance with the analysis of sulfate and 435 sulfide along the bed (Figure 4A), which revealed that the higher sulfate reducing 436 activities were obtained in the medium (H=25cm) and upper (H=44cm) parts of the bed. 13 437 On the other hand, the main ongoing process in the lower part of the bed (H=5cm) was 438 the hydrolysis and acidogenesis of glycerol instead of the reduction of sulfate, which is 439 in agreement with the main genera found at this bed height: Citrobacter (9.47%), 440 Klebsiella (11.1%) and Trichococcus (31.0%). Acetic and propionic acids were produced 441 in large amounts as result of their activity. Then, acetic acid was progressively consumed 442 along the bed while propionic acid was still increasing, even if could not be ensured that 443 sulfate reduction was occurring using whether VFA (mainly propionic), hydrogen (an 444 intermediate product obtained from glycerol fermentation) or both. Eubacterium, a 445 hydrogenotrophic methanogenic bacterium, was also found in high abundance (11.6%). 446 The presence of Eubacterium, together with the presence of the methanogenic archaea 447 Methanosaeta (5.55%), explained the high percentage of methane in the biogas produced 448 (above 85%). Despite reduced HRTs of 2h-4h enhanced hydrogen production (Si et al., 449 2015), hydrogen was not always detected, indicating that H2 was probably produced and 450 consumed along the granular sludge bed. 451 A completely different performance was found after the reduction of the organic load and 452 the pH shock (day 120 on). The UASB operation became completely unstable. 453 Granulation was lost (Figure 3) and operational problems such as the sharp decrease of 454 the HRT in the blanket section of the bed, caused by the sludge bed breakage and the 455 formation of preferential paths (Figure S2, Supplementary Material), emerged. However, 456 it was observed that, even with such a malfunctioning of the reactor, the S-RE was 60.4% 457 with a low COD/S-Sulfate consumed (around 2.5 g O2 g-1 S). Then, bacteria overcoming 458 the pH shock were capable of hydrolyzing glycerol and to produce VFA and reduce 459 sulfate to obtain acetic acid and CO2 (such as Desulfobacterium sp. or some Desulfovibrio 460 species) as reported by other authors (Qatibi et al., 1991a; Qatibi et al., 1991b). However, 461 the analysis of microbial diversity revealed that the main genera that remained in the 462 UASB after the pH shock were not only Desulfobulbus (19.8%) and Desulfovibrio (8.5%) 463 but also Unclassified Clostridiales (45.2%). The theoretical COD/S-Sulfate required to 464 completely reduce sulfate to sulfide with the complete mineralization of glycerol is 2 g 465 O2 g-1 S (Eqs. 1 and 2) which is 25% lower than that obtained at the end of the UASB 466 operation. However, it was observed that the remaining COD in the effluent was mostly 467 acetate. The acetogenesis from glycerol without the reduction of sulfate implies the 468 consumption of 0.43 g O2 g-1 O2-glycerol (Eq. 3). This means that the conversion of 1100 469 mg O2-glycerol L-1 (COD concentration in the influent at the end of the UASB operation) 470 into acetate requires a consumption of 470 mg O2 L-1, which is similar to the COD 14 471 consumption obtained at the end of the UASB operation (Figure 2). A production of 58.9 472 mg H2 L-1 could be also produced from the acetogenesis (Eq. 3), which is high enough to 473 reduce 140 mg S L-1 (sulfate reduced at the end of the UASB operation) since the 474 theoretical ratio is 0.25 g H2 g-1 S-Sulfate (Eq. 4). The high abundance of Unclassified 475 Clostridiales, together with the COD/S-Sulfate consumed and the VFAs composition in 476 the effluent are in agreement with the hypothesis that acetogenesis and hydrogenotrophic 477 sulfate reduction were the main processes occurring in the sludge bed at the end of the 478 UASB operation. Complementary experiments such as microcosms are warranted to 479 define clearly the bioprocesses occurring in each height of the sulfidogenic glycerol- 480 degrading sludge bed. 481 C3H8O3 + 0.75 SO42-  CH3COOH + 0.75HS- + 0.75 OH- + H2CO3 + 0.25 H2O (1) CH3COOH + SO42-  HS- + HCO3- + H2CO3 (2) C3H8O3 + 2H2O  CH3COOH + 3H2 + H2CO3 (3) 4H2 + SO42-  HS- + 3H2O + OH- (4) 482 483 5. CONCLUSIONS 484 This work demonstrates that sulfate can be successfully treated in UASB reactors using 485 crude glycerol as carbon source at low up-flow velocities and short HRTs. An excess of 486 crude glycerol was needed since 15-20% of the crude glycerol COD was not hydrolyzed 487 and, consequently, not utilized. The study of the sludge bed stratification performed 488 before the system failure revealed that crude glycerol hydrolysis and acidogenesis 489 happened in the first half part of the reactor while methanogenic activity was present 490 along the whole bed even at sulfide concentrations above 100 mg S L-1. Sulfate was 491 reduced once the crude glycerol was converted into VFAs or hydrogen to become the 492 main process concomitantly with hydrogenotrophic and acetoclastic methanogenesis. The 493 lack of a minimum methane production leads to poor sludge granulation, thus driving the 494 UASB performance to an unstable operation. Also, a short, temporary pH shock up to 495 pH=12 dramatically affected the UASB performance by ceasing methanogenic activity 496 and leading to biomass degranulation. SRB revealed as more resistant and resilient to pH 497 shocks compared to methanogenic biomass. Further research is needed to enhance the 498 identification of the biodegradation mechanisms and intermediate compounds such 15 499 organosulfur compounds and VFA produced from the biodegradation of crude glycerol 500 in sulfidogenic UASBs. 501 502 6. ACKNOWLEDGMENTS 503 Authors are members of the GENOCOV research group from the Department of 504 Chemical, Biological and Environmental Engineering at UAB (Universitat Autònoma de 505 Barcelona), a unit of Biochemical Engineering of Xarxa de Referència en Biotecnologia 506 de Catalunya (XRB), Generalitat de Catalunya. Authors acknowledge the Spanish 507 Government, through the project RTI2018-099362-B-C21 MINECO/FEDER, EU, for the 508 financial support provided to perform this research. The authors also acknowledge UIPSA 509 and ecoMotion for their kind collaboration in this work. 510 511 7. REFERENCES 512 Alphenaar, P.A., Visser, A., Lettinga, G., 1993. The Effect of Liquid Upward Velocity 513 and Hydraulic Retention Time on Granulation in Uasb Reactors Treating Waste- 514 Water with a High Sulfate Content. Bioresource Technol 43, 249-258. 515 https://doi.org/10.1016/0960-8524(93)90038-D 516 Ambler, J.R., Logan, B.E., 2011. Evaluation of stainless steel cathodes and a bicarbonate 517 buffer for hydrogen production in microbial electrolysis cells using a new method for 518 measuring gas production. Int J Hydrogen Energ 36, 160-166. 519 https://doi.org/10.1016/j.ijhydene.2010.09.044 520 Andersson, F.A., Karlsson, A., Svensson, B.H., Ejlertsson, J., 2004. Occurrence and 521 abatement of volatile sulfur compounds during biogas production. J Air Waste 522 Manag Assoc. 54, 855-861. https://doi.org/10.1080/10473289.2004.10470953 523 APHA, 2005. Standard methods for the examination of water and wastewater. American 524 Public Health Association (APHA): Washington, DC, USA. 525 Ariesyady, H.D., Ito, T., Okabe, S., 2007. Functional bacterial and archaeal community 526 structures of major trophic groups in a full-scale anaerobic sludge digester. Water 527 Res 41, 1554-1568. https://doi.org/10.1016/j.watres.2006.12.036 528 Baeza, J.A., Martinez-Miro, A., Guerrero, J., Ruiz, Y., Guisasola, A., 2017. 529 Bioelectrochemical hydrogen production from urban wastewater on a pilot scale. J 530 Power Sources 356, 500-509. https://doi.org/10.1016/j.jpowsour.2017.02.087 531 Bertolino, S.M., Melgaco, L.A., Sa, R.G., Leao, V.A., 2014. Comparing lactate and 532 glycerol as a single-electron donor for sulfate reduction in fluidized bed reactors. 533 Biodegradation 25, 719-733. https://doi.org/10.1007/s10532-014-9694-1 16 534 Celis-Garcia, L.B., Razo-Flores, E., Monroy, O., 2007. Performance of a down-flow 535 fluidized-bed reactor under sulfate reduction conditions using volatile fatty acids as 536 electron donors. Biotechnol Bioeng 97, 771-779. https://doi.org/10.1002/bit.21288 537 Chen, Y., Cheng, J.J., Creamer, K.S., 2008. Inhibition of anaerobic digestion process: A 538 review. Bioresource Technol 99, 4044-4064. 539 https://doi.org/10.1016/j.biortech.2007.01.057 540 Choi, E., Rim, J.M., 1991. Competition and Inhibition of Sulfate Reducers and Methane 541 Producers in Anaerobic Treatment. Water Sci Technol 23, 1259-1264. 542 https://doi.org/10.2166/wst.1991.0577 543 Dinkel, V.G., Frechen, F.B., Dinkel, A.V., Smirnov, Y.Y., Kalyuzhnyi, S.V., 2010a. 544 Kinetics of Anaerobic Biodegradation of Glycerol by Sulfate-Reducing Bacteria. 545 Appl Biochem Micro+ 46, 712-718. https://doi.org/10.1134/S0003683810070069 546 Dinkel, W., Frechen, F.B., Kljavlin, M., 2010b. Kinetics of Anaerobic Biodegradation of 547 Glycerol by Sulfate-Reducing Bacteria. Chem-Ing-Tech 82, 1771-1780. 548 https://doi.org/10.1134/S0003683810070069 549 Elferink, S.J.W.H., Visser, A., Pol, L.W.H., Stams, A.J.M., 1994. Sulfate Reduction in 550 Methanogenic Bioreactors. Fems Microbiol Rev 15, 119-136. 551 https://doi.org/10.1111/j.1574-6976.1994.tb00130.x 552 Gasiorek, J. 1994. Microbial removal of sulfur dioxide from a gas stream. Fuel Processing 553 Technol 40, 129-138. https://doi.org/10.1016/0378-3820(94)90137-6 554 Hansen, C.F., Hernandez, A., Mullan, B.P., Moore, K., Trezona-Murray, M., King, R.H., 555 Pluske, J.R., 2009. A chemical analysis of samples of crude glycerol from the 556 production of biodiesel in Australia, and the effects of feeding crude glycerol to 557 growing-finishing pigs on performance, plasma metabolites and meat quality at 558 slaughter. Anim Prod Sci 49, 154-161. https://doi.org/10.1071/EA08210 559 Homann, T., Tag, C., Biebl, H., Deckwer, W.D., Schink, B., 1990. Fermentation of 560 Glycerol to 1,3-Propanediol by Klebsiella and Citrobacter Strains. Appl Microbiol 561 Biot 33, 121-126. https://dx.doi.org/10.1007/BF00176511 562 Hu, S.J., Luo, X.L., Wan, C.X., Li, Y.B., 2012. Characterization of Crude Glycerol from 563 Biodiesel Plants. J Agr Food Chem 60, 5915-5921. 564 https://doi.org/10.1021/jf3008629 565 Hu, Y., Jing, Z.Q., Sudo, Y., Niu, Q.G., Du, J.R., Wu, J., Li, Y.Y., 2015. Effect of influent 566 COD/SO42- ratios on UASB treatment of a synthetic sulfate-containing wastewater. 567 Chemosphere 130, 24-33. https://doi.org/10.1016/j.chemosphere.2015.02.019 568 Jing, Z.Q., Hu, Y., Niu, Q.G., Liu, Y.Y., Li, Y.Y., Wang, X.C.C., 2013. UASB 569 performance and electron competition between methane-producing archaea and 570 sulfate-reducing bacteria in treating sulfate-rich wastewater containing ethanol and 571 acetate. Bioresource Technol 137, 349-357. 572 https://doi.org/10.1016/j.biortech.2013.03.137 17 573 Klimont, Z., Smith, S.J., Cofala, J., 2013. The last decade of global anthropogenic sulfur 574 dioxide: 2000-2011 emissions. Environ. Res. Lett. 8. https://doi.org/10.1088/1748- 575 9326/8/1/014003. 576 Liu, Y., Tay, J.H., 2002. The essential role of hydrodynamic shear force in the formation 577 of biofilm and granular sludge. Water Res 36, 1653-1665. 578 https://doi.org/10.1016/S0043-1354(01)00379-7 579 Liu, Y., Tay, J.H., 2004. State of the art of biogranulation technology for wastewater 580 treatment. Biotechnol Adv 22, 533-563. 581 https://doi.org/10.1016/j.biotechadv.2004.05.001 582 Liu, Y.T., Balkwill, D.L., Aldrich, H.C., Drake, G.R., Boone, D.R., 1999. 583 Characterization of the anaerobic propioate-degrading syntrophs Smithella 584 propionica gen. nov., sp. nov. and Syntrophobacter wolinii. Int J Syst Bacteriol 49, 585 545-556. https://doi.org/10.1099/00207713-49-2-545 586 Lopez, J.A.S., Santos, M.D.M., Perez, A.F.C., Martin, A.M., 2009. Anaerobic digestion 587 of glycerol derived from biodiesel manufacturing. Bioresource Technol 100, 5609- 588 5615. https://doi.org/10.1016/j.biortech.2009.06.017 589 Mora, M., Lafuente, J., Fernández, J., March, R., Gabriel, D., 2016. Crude glycerol use 590 as carbon source in a sulfate-reducing UASB for sulfur recovery from S-rich 591 effluents. In: Proceedings of the 1st International Conference on Bioenergy and 592 Climate Change: Towards a sustainable development, June 6-7 2016, Soria, Spain. 593 Mora, M., Lafuente, J., Gabriel, D., 2018. Screening of biological sulfate reduction 594 conditions for sulfidogenesis promotion using a methanogenic granular sludge. 595 Chemosphere 210, 557–566. https://doi.org/10.1016/j.chemosphere.2018.07.025 596 Mora, M., Lopez, L.R., Lafuente, J., Perez, J., Kleerebezem, R., van Loosdrecht, M.C.M., 597 Gamisans, X., Gabriel, D., 2016. Respirometric characterization of aerobic sulfide, 598 thiosulfate and elemental sulfur oxidation by S-oxidizing biomass. Water Res 89, 599 282-292. https://doi.org/10.1016/j.watres.2015.11.061 600 Pagani, I., Lapidus, A., Nolan, M., Lucas, S., Hammon, N., Deshpande, S., Cheng, J.F., 601 Chertkov, O., Davenport, K., Tapia, R., Han, C., Goodwin, L., Pitluck, S., Liolios, 602 K., Mavromatis, K., Ivanova, N., Mikhailova, N., Pati, A., Chen, A., Palaniappan, 603 K., Land, M., Hauser, L., Chang, Y.J., Jeffries, C.D., Detter, J.C., Brambilla, E., 604 Kannan, K.P., Djao, O.D.N., Rohde, M., Pukall, R., Spring, S., Goker, M., Sikorski, 605 J., Woyke, T., Bristow, J., Eisen, J.A., Markowitz, V., Hugenholtz, P., Kyrpides, 606 N.C., Klenk, H.P., 2011. Complete genome sequence of Desulfobulbus propionicus 607 type strain (1pr3(T)). Stand Genomic Sci 4, 100-110. 608 https://doi.org/10.4056/sigs.1613929 609 Philip, L., Deshusses, M.A., 2003. Sulfur dioxide treatment from flue gases using a 610 biotrickling filter - Bioreactor system. Environ. Sci. Technol. 37, 1978–1982. 611 https://doi.org/10.1021/es026009d 18 612 Pol, L.W.H., Dezeeuw, W.J., Velzeboer, C.T.M., Lettinga, G., 1983. Granulation in 613 Uasb-Reactors. Water Sci Technol 15, 291-304. 614 https://doi.org/10.2166/wst.1983.0172 615 Pol, L.W.H., Lopes, S.I.D., Lettinga, G., Lens, P.N.L., 2004. Anaerobic sludge 616 granulation. Water Res 38, 1376-1389. https://doi.org/10.1016/j.watres.2003.12.002 617 Postgate, J., 2013. The sulfate-reducing bacteria: contemporary perspectives. Springer 618 Science & Business Media. 619 Qatibi, A.I., Bennisse, R., Jana, M., Garcia, J.L., 1998. Anaerobic degradation of glycerol 620 by Desulfovibrio fructosovorans and D-carbinolicus and evidence for glycerol- 621 dependent utilization of 1,2-propanediol. Curr Microbiol 36, 283-290. 622 https://doi.org/10.1007/s002849900311 623 Qatibi, A.I., Bories, A., Garcia, J.L., 1991a. Sulfate Reduction and Anaerobic Glycerol 624 Degradation by a Mixed Microbial Culture. Curr Microbiol 22, 47-52. 625 https://doi.org/10.1007/BF02106212 626 Qatibi, A.I., Niviere, V., Garcia, J.L., 1991b. Desulfovibrio-Alcoholovorans Sp-Nov, a 627 Sulfate-Reducing Bacterium Able to Grow on Glycerol, 1,2-Propanediol and 1,3- 628 Propanediol. Arch Microbiol 155, 143-148. https://doi.org/10.1007/BF00248608 629 Qian, J., Lu, H., Cui, Y., Wei, L., Liu, R., Chen, G.H., 2015. Investigation on thiosulfate- 630 involved organics and nitrogen removal by a sulfur cycle-based biological 631 wastewater treatment process. Water Res 69, 295-306. 632 https://doi.org/10.1016/j.watres.2014.11.038 633 Santos, S.C., Liebensteiner, M.G., van Gelder, A.H., Dimitrov, M.R., Almeida, P.F., 634 Quintella, C.M., Stams, A.J., Sánchez-Andrea, I., 2017. Bacterial glycerol oxidation 635 coupled to sulfate reduction at neutral and acidic pH. The Journal of general and 636 applied microbiology, 64, 1-8. https://doi.org/10.2323/jgam.2017.02.009 637 Schauder, R., Schink, B., 1989. Anaerovibrio-Glycerini Sp-Nov, an Anaerobic Bacterium 638 Fermenting Glycerol to Propionate, Cell Matter, and Hydrogen. Arch Microbiol 152, 639 473-478. https://dx.doi.org/10.1007/BF00446932 640 Scherson, Y.D., Criddle, C.S., 2014. Recovery of freshwater from wastewater: Upgrading 641 Process configurations to maximize energy recovery and minimize residuals. 642 Environ. Sci Technol 48, 8420–8432. https://doi.org/10.1021/es501701s 643 Si, B.C., Li, J.M., Li, B.M., Zhu, Z.B., Shen, R.X., Zhang, Y.H., Liu, Z.D., 2015. The 644 role of hydraulic retention time on controlling methanogenesis and 645 homoacetogenesis in biohydrogen production using upflow anaerobic sludge blanket 646 (UASB) reactor and packed bed reactor (PBR). Int J Hydrogen Energ 40, 11414- 647 11421. https://doi.org/10.1016/j.ijhydene.2015.04.035 648 Sridhar, R., Sivakumar, V., Thirugnanasambandham, K., 2016. Response surface 649 modeling and optimization of upflow anaerobic sludge blanket reactor process 650 parameters for the treatment of bagasse based pulp and paper industry wastewater. 19 651 Desalination and Water Treatment, 57(10), 4345-4356. 652 https://doi.org/10.1080/19443994.2014.999712 653 Srivastava, R.K., Jozewicz, W., 2001. Flue Gas Desulfurization : The State of the Art. Air 654 Waste Manag. 51, 1676–1688. https://doi.org/10.1080/10473289.2001.10464387 655 Thirugnanasambandham, K., Sivakumar, V., 2014. Investigation on Fluidized Bed 656 Bioreactor Treating Ice Cream Wastewater Using Response Surface Methodology 657 and Artificial Neural Network. International Journal of Chemical Reactor 658 Engineering, 12(1), 563-573. https://doi.org/10.1515/ijcre-2014-0112 659 Thirugnanasambandham, K., Sivakumar, V., Sruthi, B., 2016. Recovery of biogas from 660 meat industry wastewater using continuously stirred tank reactor (CSTR): modeling 661 and optimization. International Journal of Chemical Reactor Engineering, 14(1), 125- 662 132. https://doi.org/10.1515/ijcre-2014-0143 663 Thirugnanasambandham, K., 2017. Enhancement of biogas production from wastewater 664 using a batch anaerobic process. Energy Sources, Part A: Recovery, Utilization, and 665 Environmental Effects, 39(14), 1484-1490. 666 https://doi.org/10.1016/j.eng.2018.11.036 667 Tsui, T.H., Chen, L., Hao, T.W., Chen, G.H., 2016. A super high-rate sulfidogenic system 668 for saline sewage treatment. Water Res 104, 147-155. 669 https://doi.org/10.1016/j.watres.2016.08.013 670 van Gelder, A.H., Aydin, R., Alves, M.M., Stams, A.J.M., 2012. 1,3-Propanediol 671 production from glycerol by a newly isolated Trichococcus strain. Microb Biotechnol 672 5, 573-578. https://doi.org/10.1111/j.1751-7915.2011.00318.x 673 Viana, M.B., Freitas, A.V., Leitão, R.C., Pinto, G.A.S., Santaella, S.T., 2012. Anaerobic 674 digestion of crude glycerol: a review. Environmental Technology Reviews 1, 81-92. 675 https://doi.org/10.1080/09593330.2012.692723 676 Yazdani, S.S., Gonzalez, R., 2007. Anaerobic fermentation of glycerol: a path to 677 economic viability for the biofuels industry. Curr Opin Biotech 18, 213-219. 678 https://doi.org/10.1016/j.copbio.2007.05.002 679 680 681 20