Volatile fatty acid production from saline cooked mussel processing wastewater at low pH

1 https://doi.org/10.1016/j.scitotenv.2020.139337 1 Volatile fatty acid production from saline 2 cooked mussel processing wastewater at low 3 pH 4 5 Andrea Fra-Vázquez, Alba Pedrouso*, Angeles Val del Río and Anuska Mosquera6 Corral 7 CRETUS Institute, Department of Chemical Engineering, Universidade de Santiago de 8 Compostela, Rua Lope Gomez de Marzoa, s/n, Campus Vida, E-15782, Santiago de 9 Compostela, Galicia, Spain. 10 11 * Corresponding author: [email protected], Tel.: +34 881816779. 12 13 ABSTRACT 14 The production of VFA using as substrate the wastewater produced in a cooked 15 mussel processing factory, containing large COD (13.7 ± 3.2 g COD/L) and salt 16 concentrations (21.8 ± 2.8 g NaCl/L) and characterized by low pH (4.6 ± 0.6) was 17 evaluated. This wastewater was fed to a 5-L completely stirred tank reactor operated 18 in continuous mode. The conversion efficiency of its COD content into volatile fatty 19 acids (VFA) was evaluated. The maximum acidification of 43 % (total VFA on 20 soluble COD basis) was obtained when an organic loading rate of 2.5 ± 0.4 g 21 COD/(L·d) was applied to the reactor and corresponded to a VFA volumetric 22 2 productivity of 0.72 ± 0.07 g CODVFA/(L·d). Under steady-state conditions, the 23 obtained mixture of VFA was composed by 80:18:2 as acetic:propionic:butyric acids 24 (percentage of VFA on soluble COD basis). Carbohydrates were degraded up to 96 % 25 while protein fermentation did not take place, probably due to the low pH value, 26 limiting the maximum acidification of the wastewater. Batch experiments showed that 27 the increase of the pH from 4.2 to 4.9 by the addition of NaHCO3 resulted in the 28 improvement of the acidification and changed the VFA mixture composition. Thus, 29 this study demonstrates the opportunity of using complex substrates, as cooked mussel 30 processing wastewater, to produce rich-VFA streams under unfavorable operational 31 conditions, such as high salinity and low pH. 32 Keywords: Anaerobic fermentation; Biorefinery; Industrial wastewater; Protein 33 degradation; Salinity; VFA. 34 35 36 3 1. INTRODUCTION 37 The fish and seafood canning industry is a crucial economic sector in Galicia (North-West of 38 Spain) which nowadays amounts to 67 % and 80 % of the European and Spanish production, 39 respectively (FAO, 2019). Indeed, Galicia is the third producer worldwide just after Thailand 40 and China. More specifically, mussels are one of the most popularly consumed seafood, and 41 Galicia represents 50 % of the worldwide production (OPMEGA, 2020). This industrial sector 42 consumes an enormous amount of water, either freshwater and/or seawater, which on average is 43 above 10 m3/tonne of raw mussel processed. As a consequence similar large volumes of highly 44 polluted wastewater are generated (Bello Bugallo et al., 2012). The main environmental 45 problem associated with this produced wastewater relates to the high organic matter (up to 42 46 g/L), comprising proteins (15 - 20 % of wet weight) (Tay et al., 2005), and salt concentrations 47 that could reach values over 20 g NaCl/L (Méndez et al., 1992). The discharge to the 48 environment of these streams without appropriate treatment could provoke continual oxygen 49 depletion, due to the contained organic matter, which causes the death of the aquatic life. 50 Furthermore, the discharge of nitrogen from proteins favours algae overgrowth leading to the 51 eutrophication of the receiving water body. In addition, if salty wastewater is not withdrawn 52 directly into the sea but to freshwater water bodies is responsible for the increase of salinity of 53 these ecosystems, similarly if it is treated in municipal wastewater treatment plants that 54 discharge in interior areas. 55 The treatment of the fish and seafood processing wastewater is particularly challenging due to 56 its complex characteristics (high organic matter and salt concentrations). In addition, the 57 seasonal activity of the factories and the fact that they commonly process different products 58 within one single week involves the generation of wastewater streams with different 59 composition in the same facility. The wastewater characteristics depend on the processing step 60 where it was generated: preliminary operations (reception, washing, brining, cutting…), 61 processing (cooking, canning and trimming), final operations (sealing and sterilization) or 62 auxiliary operations such as steam generation (Carrera et al., 2019; Cristóvão et al., 2016; 63 Méndez et al., 1992). For example, a high volume of diluted washing wastewater is generated 64 4 while the volume of cooking process wastewater is highly polluted is low. Nevertheless, the 65 different generated wastewater types are usually treated together after being homogenized in a 66 tank (Cristóvão et al., 2016). The most common technologies applied for the treatment of fish 67 and seafood processing wastewater are based on physical-chemical (membrane separation, 68 chemical destabilization and electrochemical methods) and biological (anaerobic and aerobic) 69 processes (Carrera et al., 2019; Cristóvão et al., 2012). Biological processes enable the recovery 70 of resources from wastewater especially when the valorized stream contains large 71 concentrations of organic matter, as it is the case of the fish and seafood canning processing 72 wastewater attracting great interest. 73 Anaerobic digestion is suggested as a suitable treatment for seafood wastewater due to its high 74 organic matter removal capacity, low energy consumption, low sludge production and energy 75 production as biogas (mainly CH4 and CO2) (Chowdhury et al., 2010). Anaerobic processes 76 with high removal efficiencies (55 - 97 %) and treating organic loads of 1 - 4 kg COD/(m3·d) 77 have been applied to treat these effluents (Méndez et al., 1992; Panpong et al., 2014; Prasertsan 78 et al., 1994; Sillapacharoenkul and Sinbuathong, 2020). In the frame of the circular economy, 79 the waste conversion into volatile fatty acids (VFA), which are short-chain fatty acids obtained 80 as metabolic intermediates in the anaerobic digestion, has recently gained attention due to their 81 wide variety of applications (Kleerebezem et al., 2015). VFA are intermediate products of the 82 anaerobic digestion process. Thus, VFA-rich streams are produced in fermentation processes 83 where the methanogenic step is suppressed (Wainaina et al., 2019). Application alternatives of 84 the waste-derived VFA are the generation of biofuels, bulk chemicals, the biological removal of 85 nutrients from wastewater and the production of bioplastics or food additives. For example, 86 VFA can be used as a carbon source during the denitrification or the biological phosphorus 87 removal processes. VFA act also as substrate in the production of polyhydroxyalkanoates 88 (PHA), a type of bioplastic, by mixed microbial cultures (Atasoy et al., 2018; Wainaina et al., 89 2019). 90 Operational conditions of the anaerobic systems significantly influence the concentration, yield 91 and composition of the VFA produced from wastes. Organic acid production is strongly 92 5 affected by the pH of the reaction media since it has a great influence on the growth rate of the 93 microorganisms involved in the anaerobic digestion (Wainaina et al., 2019). Indeed, methane 94 production is barely observed out of its optimal pH range (6.5 - 8.5). Nevertheless, hydrolytic 95 and acidogenic microorganisms operate at an optimal pH range of 5 - 11 and cannot survive in 96 extremely acidic (pH 3) or alkaline (pH 12) conditions (Jankowska et al., 2015; Wainaina et al., 97 2019). The optimal pH to maximize the acidification efficiency varies according to the waste 98 characteristics and the operational conditions. Jankowska et al. (2015) observed that, in 99 unbuffered systems, acidic pH promoted the VFA production at short retention time (5 days) 100 while alkaline pH (10 - 11) maximized VFA accumulation at longer retention times (15 days). 101 Wainaina et al. (2019) stated that acidic pH is suitable to produce VFA from a variety of easily 102 degradable wastes while alkaline pH values are recommended when complex substrates are 103 used. For example, different optimal pH values to obtain VFA were reported: from cheese whey 104 is 5.2 - 5.5 (Bengtsson et al., 2008), from food waste and the organic fraction of municipal solid 105 waste is 9.0 (Cheah et al., 2019; Moretto et al., 2019), from kitchen waste is 7.0 (Zhang et al., 106 2005) and from wasted activated sludge ranges from 9.5 to 11.0 (Chen et al., 2007; Liu et al., 107 2020). 108 Since most cooked mussel processing factories use seawater in their processes, another primary 109 concern with the produced wastewater is its high salinity (Xiao and Roberts, 2010). Significant 110 salt concentrations can inhibit the anaerobic processes, especially the methanogenesis step at 111 concentrations above 10 g NaCl/L (Panpong et al., 2015). However, the adaptation of the 112 anaerobic biomass to high salt concentrations (Artiga et al., 2008; Sudmalis et al., 2018; Zhang 113 et al., 2017), or the use of halophilic inoculum (Aspé et al., 1997; Scoma et al., 2017; Tan et al., 114 2019) are suitable strategies to develop an anaerobic treatment process for saline wastewater. 115 The purpose of this study was to evaluate the suitability of the wastewater generated in a cooked 116 mussel processing factory as feedstock to produce a VFA-rich effluent, with the novelty of 117 operating the continuous acidifying reactor at very low pH and high salt concentration. Batch 118 experiments were also performed to investigate the effect of the pH on the productivity and 119 composition of the produced VFA. 120 6 2. MATERIALS AND METHODS 121 2.1 Cooked mussel processing wastewater characterization 122 The wastewater used in the present study was taken directly from the cookers of a mussel 123 processing factory (Cocedero Suárez, Vilanova de Arousa, Spain). The pH of the mussel 124 cooking wastewater at the time of the collection was approximately 7 but it dropped to 4 - 5 125 (Table 1) after a couple of days stored at 4 ºC. Wastewater was stored at low temperature to 126 prevent the degradation of the organic matter. Carbohydrates were the predominant organic 127 compounds (50 % of the soluble COD), followed by proteins (30 % of the soluble COD). The 128 concentration of proteins and carbohydrates as chemical oxygen demand (COD) was calculated 129 using the following factors: 1.5 g CODprotein/g protein and 1.1 g CODcarbohydrates/g carbohydrate 130 (Mahmoud et al., 2004). The lipid concentration was not significant. The wastewater 131 composition fluctuated due to changes in the factory process, and its variability defined the 132 three operational stages carried out in the acidification reactor, as indicated in Table 1. 133 134 Table 1. Average values of the main characteristic parameters of the wastewater treated and 135 reactor operational conditions. 136 Parameters Units Stage I Stage II Stage III 0 - 59 days 60 – 279 days 280 - 400 days OLR g COD/(L·d) 7.3 ± 0.5 2.6 ± 0.4 2.2 ± 0.2 HRT d 3.1 6.3 6.3 pH -- 4.7 ± 0.4 4.4 ± 0.5 5.1 ± 0.7 sCOD g/L 18.3 ± 1.3 13.1 ± 0.4 11.1 ± 1.1 Carbohydrates g/L ND 5.5 ± 1.6 5.3 ± 1.3 Proteins g/L ND 2.8 ± 0.4 1.7 ± 0.2 VFA g CODVFA/L 0.7 ± 0.3 2.2 ± 1.3 1.7 ± 0.7 Ammonium g NH4+-N/L 0.09 ± 0.02 0.19 ± 0.06 0.19 ± 0.05 NaCl g/L 19.1 ± 2.1 22.7 ± 2.4 22.1 ± 3.0 OLR: organic loading rate; HRT: hydraulic retention time; COD: chemical oxygen demand; VFA: volatile 137 fatty acids; ND: Not determined. 138 139 7 2.2 Experimental set-up 140 2.2.1 Continuous reactor for VFA production 141 A continuous stirred tank reactor with a working volume of 5 L was used to produce VFA. It 142 was directly fed with raw cooked mussel processing wastewater (Table 1). The temperature was 143 maintained in the mesophilic range (37 ± 1 °C) using a thermostatic bath (Techne Inc., USA). 144 The reactor was inoculated with anaerobic granular sludge from a pilot-scale up-flow anaerobic 145 sludge blanket (UASB) reactor that treated mimicked municipal wastewater (Silva-Teira et al., 146 2017). Short solid retention times (SRT) were imposed to washout the methanogenic 147 microorganisms from the anaerobic mixed culture as they present growth rates lower than the 148 acidogenic bacteria (Khan et al., 2016). The gas-phase composition was measured during the 149 first days of Stage I to check the absence of methane production due to the inhibition of 150 methanogenic microorganisms. 151 The operation of the reactor lasted 400 days, divided into three different stages (Table 1). 152 During Stage I (the first 59 days) an organic loading rate (OLR) of 7.3 ± 0.5 g COD/(L·d) was 153 applied, with a hydraulic retention time (HRT) of 3.1 days. Then in Stage II (days 60 to 279), 154 the OLR was diminished to 2.6 ± 0.4 g COD/(L·d) by increasing the HRT to 6.3 days. Finally, 155 the OLR was further decreased in Stage III (days 280-400) to 2.2 ± 0.2 g COD/(L·d) while HRT 156 was maintained. The SBR operated under complete mixing conditions by means of the action of 157 a mechanical stirrer at 120 rpm (Heidolph, Germany); therefore, the SRT was equal to the HRT. 158 The pH of the media was not controlled. 159 160 2.2.2 Acidification batch tests 161 The acidification batch assays were carried out in 500 mL Pirex-glass bottles (400 mL of 162 working volume), following the methodology described by Silva et al. (2013). The bottles were 163 filled in with the corresponding volumes of substrate, macroand micro-nutrients solutions and 164 acidifying biomass from the continuous acidification reactor (Table 2). The substrate 165 composition corresponded to Stage I of Table 1. The substrate to biomass ratio was set at 3 g 166 COD/g VSS. 167 8 In total, 6 bottles were prepared with 3 different conditions (duplicates): two as control 168 experiments without inoculum addition for measuring the abiotic disappearance of the substrate 169 (E1); two without alkalinity addition (E2) and two containing NaHCO3 in a ratio of 1:1 with 170 respect to VSS (E3). After the addition of the substrate, biomass and medium, the headspace of 171 each vial was bubbled with N2 and the bottles were sealed with rubber stoppers and capped with 172 plastic seals. Then, bottles were incubated in a shaker (120 rpm) at 37 °C. VFA production was 173 monitored throughout time by the analysis of the periodically collected samples from the liquid 174 phase of each bottle. Before collecting these liquid samples, 1 mL-gas sample was taken and 175 measured by gas chromatography (Hewlett Packard 5890 Series II instrument) to assess the 176 occurrence of methane production. The evolution of the concentration of VFA (expressed as g 177 CODVFA/L) versus time was plotted. The specific acidogenic activity (g CODVFA/(g VSS·d)) 178 was estimated as the ratio between the maximum slope of the appearance of VFA (g 179 CODVFA/(L·d)) and the concentration of biomass present in the bottles (g VSS/L). 180 Table 2. Initial operational conditions of the acidification batch experiments. 181 Volumes added of different compounds Experiment Control (E1) No alkalinity (E2) Alkalinity (E3) Acidifying sludge (mL) 0 23 23 Wastewater (mL) 61.2 61.2 61.2 Macronutrients solution (mL)* 66 66 66 Micronutrients solution (mL)* 13 13 13 10 g NaHCO3/L solution (mL) 0 0 28 *Compositions of macroand micronutrient solutions described in Silva et al. (2013). 182 183 2.3 Analytical methods 184 Total suspended solids (TSS), volatile suspended solids (VSS), alkalinity and COD were 185 analysed according to Standard Methods for the Examination of Water and Wastewater 186 (APHA-AWWA-WEF, 2017). Liquid samples were filtered through a cellulose-ester filter of 187 0.45 µm of pore size (Advantec, Japan) for the quantification of total organic carbon (TOC), 188 9 ammonium (NH4+), soluble chemical oxygen demand (sCOD), proteins, carbohydrates, VFA 189 and other ions to determine salt concentration. Ammonium concentration was determined 190 according to the Bower and Holm-Hansen method (Bower and Holm-Hansen, 1980). TOC 191 concentration was determined by catalytic combustion (Analyser model TOC-L CSN, 192 Shimadzu, Japan). VFA concentration was determined by gas chromatography (GC) (Hewlett 193 Packard, USA). Protein and carbohydrate concentrations were measured according to Lowry et 194 al. (1951) and Loewus (1952) methods, respectively. Anions (e.g. Cl-) and cations (e.g. Na+) 195 were determined by ion chromatography (861 Advanced Compact IC system, Metrohm, 196 Switzerland). 197 198 2.4 Calculations 199 The individual acid concentrations for acetic acid (HAc), propionic acid (HPr), butyric acid 200 (HBu) and valeric acid (HVa) were converted to COD units by the application of corresponding 201 coefficients: 1.07 g CODHAc/g HAc, 1.51 g CODHPr/g HPr, 1.82 g CODHBu/g HBu and 2.04 g 202 CODHVa/g HVa. The acidification percentage was calculated as the sum of the individual VFA 203 measured by GC, converted to COD units (g CODVFA), and divided by the amount of COD at 204 the beginning of the experiment (CODi), as indicated in the following equation: 205 Acidification (%) = g CODVFA g CODi · 100 Statistical analysis of data was carried out using the software R version 3.5.1. The normality and 206 homogeneity of variance were evaluated by means of the Shapiro-Wilk and Levene tests, 207 respectively. ANOVA parametric test was used when both tests could be confirmed, and if not, 208 non-parametric Kruskal-Wallis test was applied. Differences in the experimental values of the 209 pH, acidification percentage and VFA concentration obtained in the acidification batch tests 210 were compared with the calculation of the area under the curve (AUC) using the package PK. 211 212 213 214 16 changes in the composition of the cooked mussel processing wastewater during the whole 344 operation, it showed more stability during the Stage III when the acidification degree and 345 composition of the mixture of VFA remained relatively constant. 346 347 3.2 Alkalinity effect on VFA production: proteins degradation 348 Batch tests were performed to evaluate the influence of the pH value on the VFA production 349 from cooked mussel processing wastewater (Figure 4 and Table S1 in Supporting Material). 350 Acidifying biomass from the reactor was collected on day 76 and used as inoculum. An 351 experiment without acidifying sludge or alkalinity addition was carried out as control (E1). 352 Then, the effect of the alkalinity was studied without (E2) and with (E3) the external addition of 353 NaHCO3, in batch experiment that already contained the same inoculum and substrate 354 concentrations. The initial VFA concentration in all bottles (E1, E2 and E3) was approximately 355 900 mg CODVFA/L. Even though the biomass collected from the acidification reactor was 356 washed before the experiment, the inoculum media contained a remaining amount of VFA (< 357 0.1 g CODVFA/L). In all the bottles, no methane production was observed during the tests. 358 In the control flasks (E1), where only substrate was added, no differences in the VFA 359 concentrations were observed throughout the batch test. Experiments E2 and E3 with substrate 360 and acidifying biomass showed an increase of the VFA concentration during the first days of the 361 batch experiment (Figure 4). However, the increase of the acidification in experiment E2 was 362 lower than in E3 and the acidification values on day 2 were 48.2 % and 61.6 %, respectively. 363 From that day onwards the VFA concentration remained at approximately 1.4 ± 0.2 g 364 CODVFA/L in E2, whereas in E3 reached a value of 2.5 ± 0.1 g CODVFA/L after 23 days of 365 experiment. This latter value corresponded to an acidification degree of 70 % of initial COD. 366 Statistical analysis was applied by comparing the area under the curve (AUC) described by the 367 VFA produced throughout the batch test and showed significant differences in the acidification 368 percentage between the flasks without (E2) and with (E3) alkalinity (p = 0.061), with 90 % 369 confidence. 370 371 17 Figure 4. Evolution of the VFA concentrations (columns), percentage of acidification 372 (continuous lines) and pH value of the liquid media (discontinuous lines) in the acidification batch 373 experiments using cooked mussel processing wastewater. Grey colour corresponds to experiment 374 E2 and black colour to experiment E3. The error bars of the columns represent the standard 375 deviation of the point. 376 377 The specific acidogenic activity of 0.79 g CODVFA/(g VSS·d) in E3, was almost three times 378 higher than in E2 (0.27 g CODVFA/(g VSS·d)). The main difference in both experiments was the 379 pH value. Without alkalinity (E2) the pH value was 4.2 ± 0.3, whereas in E3 the addition of 380 NaHCO3 promoted the maintenance of higher pH (4.9 ± 0.1). These results were in accordance 381 with the specific activities estimated for the acidifying reactor. During Stage I the acidogenic 382 activity was 0.24 ± 0.11 g CODVFA/(g VSS·d), which was very similar to the value obtained in 383 E2. This value increased during Stage III when a higher pH was measured in the reactor and 384 correlated with an increase of the acidogenic activity, being the average value of 0.33 ± 0.03 g 385 CODVFA/(g VSS·d). Therefore, batch results indicated that the increase in the pH, by addition of 386 alkalinity, had a positive effect in terms of conversion of VFA from the cooked mussel 387 processing wastewater. Yu and Fang (2002) also observed changes in the VFA production from 388 0 20 40 60 80 0 2 4 6 8 0 1 2 3 4 7 9 11 14 16 18 23 Acidfication (%) VFA (g CODVFA/L); pH Time (d) 18 dairy wastewater at variable pH and obtained an increase of the microbial activity from 0.146 g 389 COD/(g VSS·d) at pH 4 to 0.320 g COD/(g VSS·d) at pH 5.5. 390 A shift of the VFA distribution was observed in experiments at different operational pH (Figure 391 5 and Table S1 in Supporting Material). During the acidification experiments without (E2) and 392 with (E3) alkalinity, acetic acid was the dominant organic acid, whereas valeric acid was 393 produced at the lowest concentration. However, propionic and butyric acids showed inverse 394 behaviour in the two experimental conditions (Figure 5). In E2 (lower pH), the butyric acid 395 concentration increased and reached the same value as acetic acid from day 7 onwards (Figure 396 5A). Propionic acid concentration slightly increased during the first days, and it remained stable 397 during most part of the experiment. In E3 (higher pH), butyric acid concentration did not 398 experience the same evolution as in E2 and approximately the same concentration was 399 maintained until the end of the experiment. However, propionic acid production increased at the 400 beginning of the assay and became the second most-produced acid after acetic (Figure 5B). 401 Previous studies have reported the influence of the pH not only on the concentration of VFA 402 produced but also on the metabolic pathways in acidogenic fermentation and, therefore, of the 403 product distribution. However, there are no consistent conclusions on the influence of pH on the 404 composition of VFA (Zhou et al., 2018). In the batch experiments of the present research work, 405 butyric acid production was improved under low pH conditions. These results agreed with 406 previous studies that reported that the butyrate metabolic pathway was enhanced under acidic 407 conditions (González-Cabaleiro et al., 2015; Jankowska et al., 2017; Temudo et al., 2007). 408 A positive effect of the acidic pH was observed in the reactor to select acidifying bacteria and 409 wash out methanogenic microorganisms from the anaerobic mixed culture used as inoculum. 410 However, the acidogenic activity was limited by the low pH values (below 4 during most of the 411 operational time). Results of the batch experiments showed that the addition of alkalinity 412 improved the VFA production and modified the obtained products, with respect to the 413 experiments without NaHCO3 addition. However, the increase of pH up to 5 was insufficient to 414 achieve complete acidification of the substrate. Even though the protein concentration was not 415 19 measured during the batch experiment, the 30 % of non-acidified COD probably corresponded 416 mainly to the protein content of the substrate. 417 A B Figure 5. Concentrations of VFA produced in the batch assays without-E2 (A) and with-E3 (B) 418 alkalinity. Acetic acid (), propionic acid (), butyric acid () and valeric acid (). The error 419 bars represent the standard deviation of the point. 420 421 Residual carbohydrate concentration is also expected due to kinetic and energetic or 422 thermodynamic conversion limitations (González-Cabaleiro et al., 2015). Considering the 423 carbohydrate affinity for the process of 1 mM (expressed as glucose) (González-Cabaleiro et al., 424 0 200 400 600 800 1000 1200 0 5 10 15 20 25 VFA (mg CODVFA/L) Time (d) 0 200 400 600 800 1000 1200 0 5 10 15 20 25 VFA (mg CODVFA/L) Time (d) 20 2015), 0.2 g COD/L would remain as carbohydrates. Thus, protein partial degradation is 425 suggested to contribute to the achievement of the 70 % of acidification in E3. If only 426 carbohydrate were degraded, the acidification efficiency would be limited to 64 %. A more 427 detailed study is required to optimise the pH via the long-term addition of NaHCO3 to the 428 feeding of the reactor and to evaluate is effect on protein degradation and, eventually, on the 429 amount of VFA produced. 430 To sum up, obtained results suggested that an increase in the pH of the reactor media could 431 promote protein degradation fostering VFA production. However, a techno-economical study 432 would be required to evaluate the process benefits in terms of acidification efficiency and 433 increase of operational costs due to the addition of chemicals to adjust the pH value. Other 434 factors like HRT and OLR should be considered to define the best operational strategy and set 435 the optimal pH value. The obtained VFA-rich stream could be used to produce PHA, as carbon 436 source for nutrient removal or purified to use the VFA as platform chemicals, among other 437 applications. Depending on the final use, the composition of the VFA mixture will be relevant 438 (as platform chemical or affecting the obtained PHA properties) or not (for nutrient removal) 439 (Atasoy et al., 2018). The final application will also determine the downstream processes 440 required to obtain the final product and a clean effluent for discharge. In the present study, it 441 was demonstrated that mussel cooking wastewater is a good candidate to produce VFA-rich 442 streams. Thus, this wastewater could be valorised, under uncontrolled pH conditions, instead of 443 being just treated consuming resources like energy or chemicals. As in the present study the aim 444 is to produce VFA subsequent treatment/processing steps are required to produce an effluent 445 with the required composition to be discharged to the environment. 446 447 4. Conclusions 448 Acidogenic fermentation of cooked mussel processing wastewater resulted in a significant VFA 449 productivity of 0.72 ± 0.07 g CODVFA/(L·d), considering the complex composition of the 450 substrate, mainly characterized by high organic matter content (13.8 ± 3.2 g COD/L), high 451 salinity (21.8 ± 2.8 g NaCl/L) and low pH (4.6 ± 0.6). The maximum acidification percentage 452 21 obtained was 43 % and the composition of the VFA mixture obtained was of 80:18:2 as 453 HAc:HPr:HBu. Carbohydrate conversion reached up to 96 % and contributed to the production 454 of VFA. However, the acidification efficiency was hindered by a deficient protein degradation, 455 probably associated to the acidic conditions inside the reactor. 456 Batch experiments showed that the increase of the pH from 4.2 to 4.9 by the addition of 457 NaHCO3 resulted in a higher acidification efficiency. In addition to increasing VFA production, 458 the composition of the mixture switched from containing mostly acetic and propionic acids to 459 containing mostly acetic and butyric. Nevertheless, part of the COD remained as non-acidified 460 COD even at pH 5, probably due to the slight degradation of proteins. 461 462 Acknowledgements 463 This research was supported by the Spanish Government (AEI) through the FISHPOL 464 (CTQ2014-55021-R) and TREASURE (CTQ2017-83225-C2-1-R) projects. The authors belong 465 to the Galician Competitive Research Group GRC ED431C 2017/29 and to the CRETUS 466 Strategic Partnership (ED431E 2018/01). All these programs are co-funded by the FEDER 467 (EU). Special thanks to Dr. Thelmo A. Lú-Chau for his contribution to the statistical analysis of 468 data. 469 470 References 471 APHA-AWWA-WEF. Standard methods for the examination of water and wastewater. 472 Washington DC, USA: American Public Health Association/American Water Works 473 Association/Water Environment Federation, 2017. 474 Artiga P, García-Toriello G, Méndez R, Garrido JM. Use of a hybrid membrane bioreactor for 475 the treatment of saline wastewater from a fish canning factory. Desalination 2008; 221: 476 518-525. doi: 10.1016/j.desal.2007.01.112. 477 22 Aspé E, Marti MC, Roeckel M. Anaerobic treatment of fishery wastewater using a marine 478 sediment inoculum. Water Research 1997; 31: 2147-2160. doi: 10.1016/S0043479 1354(97)00051-1. 480 Atasoy M, Eyice O, Schnürer A, Cetecioglu Z. Volatile fatty acids production via mixed culture 481 fermentation: Revealing the link between pH, inoculum type and bacterial composition. 482 Bioresource Technology 2019; 292: 121889. doi: 10.1016/j.biortech.2019.121889. 483 Atasoy M, Owusu-Agyeman I, Plaza E, Cetecioglu Z. Bio-based volatile fatty acid production 484 and recovery from waste streams: Current status and future challenges. Bioresource 485 Technology 2018; 268: 773-786. https://doi.org/10.1016/j.biortech.2018.07.042. 486 Bello Bugallo PM, Stupak A, Cristóbal Andrade L, Torres López R. Material Flow Analysis in 487 a cooked mussel processing industry. Journal of Food Engineering 2012; 113: 100-117. 488 doi: 10.1016/j.jfoodeng.2012.05.014. 489 Bengtsson S, Hallquist J, Werker A, Welander T. Acidogenic fermentation of industrial 490 wastewaters: Effects of chemostat retention time and pH on volatile fatty acids 491 production. Biochemical Engineering Journal 2008; 40: 492-499. doi: 492 10.1016/j.bej.2008.02.004. 493 Bermudez-Penabad N, Kennes C, Veiga MC. Anaerobic digestion of tuna waste for the 494 production of volatile fatty acids. Waste Management 2017; 68: 96-102. doi: 495 10.1016/j.wasman.2017.06.010. 496 Bower CE, Holm-Hansen T. A Salicylate–Hypochlorite Method for Determining Ammonia in 497 Seawater. Canadian Journal of Fisheries and Aquatic Sciences 1980; 37: 794-798. doi: 498 10.1139/f80-106. 499 Carrera P, Campo R, Méndez R, Di Bella G, Campos JL, Mosquera-Corral A, et al. Does the 500 feeding strategy enhance the aerobic granular sludge stability treating saline effluents? 501 Chemosphere 2019; 226: 865-873. https://doi.org/10.1016/j.chemosphere.2019.03.127. 502 Cristóvão RO, Botelho CM, Martins RJE, Boaventura RAR. Chemical and biological treatment 503 of fish canning wastewaters. international Journal of Bioscience and Bioinformatics 504 2012; 2: 237-242. doi: 10.7763/IJBBB.2012.V2.108. 505 23 Cristóvão RO, Pinto VMS, Gonçalves A, Martins RJE, Loureiro JM, Boaventura RAR. Fish 506 canning industry wastewater variability assessment using multivariate statistical 507 methods. Process Safety and Environmental Protection 2016; 102: 263-276. 508 https://doi.org/10.1016/j.psep.2016.03.016. 509 Cheah Y-K, Vidal-Antich C, Dosta J, Mata-Álvarez J. Volatile fatty acid production from 510 mesophilic acidogenic fermentation of organic fraction of municipal solid waste and 511 food waste under acidic and alkaline pH. Environmental Science and Pollution 512 Research 2019; 26: 35509-35522. doi: 10.1007/s11356-019-05394-6. 513 Chen Y, Jiang S, Yuan H, Zhou Q, Gu G. Hydrolysis and acidification of waste activated sludge 514 at different pHs. Water Research 2007; 41: 683-689. doi: 10.1016/j.watres.2006.07.030. 515 Chowdhury P, Viraraghavan T, Srinivasan A. Biological treatment processes for fish processing 516 wastewater - A review. Bioresource Technology 2010; 101: 439-449. doi: 517 10.1016/j.biortech.2009.08.065. 518 Duong TH, Grolle K, Nga TTV, Zeeman G, Temmink H, van Eekert M. Protein hydrolysis and 519 fermentation under methanogenic and acidifying conditions. Biotechnology for Biofuels 520 2019; 12: 254. doi: 10.1186/s13068-019-1592-7. 521 FAO. The canned seafood sector in Spain. Globefish. FAO. Food and Agriculture Organization 522 of the United Nations, 2019. 523 González-Cabaleiro R, Lema JM, Rodríguez J. Metabolic Energy-Based Modelling Explains 524 Product Yielding in Anaerobic Mixed Culture Fermentations. PLOS ONE 2015; 10: 525 e0126739. 10.1371/journal.pone.0126739. 526 Gouveia AR, Freitas EB, Galinha CF, Carvalho G, Duque AF, Reis MAM. Dynamic change of 527 pH in acidogenic fermentation of cheese whey towards polyhydroxyalkanoates 528 production: Impact on performance and microbial population. New Biotechnology 529 2017; 37: 108-116. doi: 10.1016/j.nbt.2016.07.001. 530 Jankowska E, Chwialkowska J, Stodolny M, Oleskowicz-Popiel P. Volatile fatty acids 531 production during mixed culture fermentation – The impact of substrate complexity and 532 pH. Chemical Engineering Journal 2017; 326: 901-910. doi: 10.1016/j.cej.2017.06.021. 533 24 Jankowska E, Chwiałkowska J, Stodolny M, Oleskowicz-Popiel P. Effect of pH and retention 534 time on volatile fatty acids production during mixed culture fermentation. Bioresource 535 Technology 2015; 190: 274-280. doi: 10.1016/j.biortech.2015.04.096. 536 Khan MA, Ngo HH, Guo WS, Liu Y, Nghiem LD, Hai FI, et al. Optimization of process 537 parameters for production of volatile fatty acid, biohydrogen and methane from 538 anaerobic digestion. Bioresource Technology 2016; 219: 738-748. doi: 539 10.1016/j.biortech.2016.08.073. 540 Kleerebezem R, Joosse B, Rozendal R, Van Loosdrecht MCM. Anaerobic digestion without 541 biogas? Reviews in Environmental Science and Bio/Technology 2015; 14: 787-801. 542 doi: 10.1007/s11157-015-9374-6. 543 Liu H, Wang J, Liu X, Fu B, Chen J, Yu H-Q. Acidogenic fermentation of proteinaceous 544 sewage sludge: Effect of pH. Water Research 2012; 46: 799-807. doi: 545 10.1016/j.watres.2011.11.047. 546 Liu X, Du M, Yang J, Wu Y, Xu Q, Wang D, et al. Sulfite serving as a pretreatment method for 547 alkaline fermentation to enhance short-chain fatty acid production from waste activated 548 sludge. Chemical Engineering Journal 2020; 385: 123991. 549 https://doi.org/10.1016/j.cej.2019.123991. 550 Loewus FA. Improvement in anthrone method for determination of carbohydrates. Analytical 551 Chemistry 1952; 24: 1. doi: 10.1021/ac60061a050. 552 Lowry OH, Rosebrough NJ, Farr AL, Randall RJ. Protein measurement with the Folin phenol 553 reagent. Journal of Biological Chemistry 1951; 193: 265-75. 554 Mahmoud N, Zeeman G, Gijzen H, Lettinga G. Anaerobic stabilisation and conversion of 555 biopolymers in primary sludge—effect of temperature and sludge retention time. Water 556 Research 2004; 38: 983-991. 557 Méndez R, Omil F, Soto M, Lema JM. Pilot plant studies on the anaerobic treatment of different 558 wastewaters from a fish-canning factory. Water Science and Technology 1992; 25: 37559 44. 560 25 Moretto G, Valentino F, Pavan P, Majone M, Bolzonella D. Optimization of urban waste 561 fermentation for volatile fatty acids production. Waste Management 2019; 92: 21-29. 562 doi: 10.1016/j.wasman.2019.05.010. 563 OPMEGA. The mussel cultivation. OPMEGA. Producers' Organization of Mussel from Galicia, 564 2020. 565 Panpong K, Nuithitikul K, O-thong S, Kongjan P. Anaerobic Co-Digestion Biomethanation of 566 Cannery Seafood Wastewater with Microcystis SP; Blue Green Algae with/without 567 Glycerol Waste. Energy Procedia 2015; 79: 103-110. doi: 568 10.1016/j.egypro.2015.11.487. 569 Panpong K, Srisuwan G, O-Thong S, Kongjan P. Anaerobic co-digestion of canned seafood 570 wastewater with glycerol waste for enhanced biogas production. Energy Procedia 2014; 571 52: 328-336. doi: 10.1016/j.egypro.2014.07.084. 572 Parawira W, Murto M, Read JS, Mattiasson B. Profile of hydrolases and biogas production 573 during two-stage mesophilic anaerobic digestion of solid potato waste. Process 574 Biochemistry 2005; 40: 2945-2952. doi: 10.1016/j.procbio.2005.01.010. 575 Prasertsan P, Jung S, Buckle KA. Anaerobic filter treatment of fishery wastewater. World 576 Journal of Microbiology and Biotechnology 1994; 10: 11-13. doi: 10.1007/bf00357553. 577 Scoma A, Coma M, Kerckhof FM, Boon N, Rabaey K. Efficient molasses fermentation under 578 high salinity by inocula of marine and terrestrial origin. Biotechnology for Biofuels 579 2017; 10. 10.1186/s13068-017-0701-8. 580 Silva-Teira A, Sánchez A, Buntner D, Rodríguez-Hernández L, Garrido JM. Removal of 581 dissolved methane and nitrogen from anaerobically treated effluents at low temperature 582 by MBR post-treatment. Chemical Engineering Journal 2017; 326: 970-979. 583 https://doi.org/10.1016/j.cej.2017.06.047. 584 Silva FC, Serafim LS, Nadais H, Arroja L, Capela I. Acidogenic fermentation towards 585 valorisation of organic waste streams into volatile fatty acids. Chemical and 586 Biochemical Engineering Quarterly 2013; 27: 467-476. 587