Effect of pH on the Anaerobic Fermentation of Fruit/Vegetables and Disposable Nappies Hydrolysate for Bio-hydrogen Production

The objective of this work was to optimize the anaerobic fermentation of a mixed waste stream, consisted of fruit and vegetables that have lost their marketing value and a disposable nappies’ hydrolysate. More specifically, the aim was to identify the optimal pH value for maximum hydrogen production and valuable metabolites such as volatile fatty acids and ethanol. A wide range of pH values was tested (from 4.5 to 7.5 with 0.5 increment) using an automatic controller system, in batch fermentations that took place in mesophilic temperature conditions (37 °C). The first set of experiments was carried out with the fruit and vegetables mixture, diluted with water (2:3 v/v) and subsequent trials followed using the fruit and vegetable mixture with the disposable nappies’ hydrolysate at the same ratio (2:3 v/v). The maximum hydrogen volume was produced at pH 6.0 (1.34 L H2/LReactor) for the fruit/vegetable stream whereas, the maximum concentration of ethanol and volatile fatty acids (15.60 g/L) was reached at pH 6.5 for the same substrate. Regarding the mixed waste stream, both hydrogen production and metabolites concentration reached a maximum at pH 7.5 with 4.09 L H2/LReactor and 17.16 g/L respectively. Different optimum pH value for bio-hydrogen production was observed between the anaerobic fermentation of the two substrates (fruit/vegetables waste and mixed waste stream). Higher overall yields and concentrations of the metabolic products were obtained with the fermentation of the mixed substrate.


Introduction
Energy crisis as a result of fossil fuel stocks decline, rising fuel prices due to high demand and the dependence of all production processes on fuels, in combination with climate change, have incited researchers globally for alternative sustainable energy sources that could offer a solution to these concerns. Among the eco-friendly biofuels, hydrogen (H 2 ) is the most promising due to its high energy yield (122 kJ/g) on mass basis, which is almost three times greater than that from hydrocarbon fuels, and high exhaust gases purity, since only water vapor is produced from its combustion [1][2][3]. H 2 is currently produced via conventional water electrolysis and thermo-catalytic reformation using natural gas and oil. Albeit the scope of these processes is the production of a cleaner fuel, their final greenhouse-gas footprint is high which is a contradiction from an environmental point of view [4,5].
Dark or anaerobic fermentation (AF) is a practical method for bio-H 2 production, consuming less energy than physicochemical ones. Metabolic pathways of facultative or obligate anaerobic bacteria, lead to bio-H 2 production through the decomposition of organic carbon from various feedstock substrates. Besides bio-H 2 , volatile fatty acids (VFAs) and ethanol can be also produced, as by-products of the process, which may be further utilized for methane, biodiesel and bio-plastics synthesis [2], among other uses. Anaerobic fermentation can be also considered as a waste management method, since organic wastes and wastewaters are treated and biologically processed in this way, tackling the environmental burden of waste disposal [6].
Promising results of biological H 2 production have been obtained using different substrates. Municipal solid wastes [7], cellulose containing wastes [8], agroindustrial wastewaters [9], cheese whey [10], food waste [11] as well as used nappies [12], have been recently reported as potential substrates for AF, taking into account that the type of substrate is of crucial importance for the effective Η 2 production. Almost 1.5 billion tons of food waste is being disposed annually around the globe, which accounts for the 33% of the food production for human consumption [6]. In Greece, large Super Market chains, return most of the expired food products or products of low quality (i.e. meat, fish, pasta, milk etc.) to the production companies for waste management or lead it to animal food companies and social actions against poverty. Fruit and vegetables, due to their vulnerability, are being thrown away and led to landfilling. Interestingly, food waste is rich in carbohydrates and other nutrients and therefore could be utilized in fermentation-based bio-refining processes for the production of biofuels and/or added value products [5].
Disposable nappies (DN) are primarily composed of cellulose and synthetic fibres that can be treated by biological methods. According to EDANA [13] a typical composition of a disposable baby nappy is 36.6% cellulose pulp, 30.7% sodium polyacrylate (SAP), 16% polypropylene, 6.2% lowdensity polyethylene, and 10.5% of elastic and adhesive tapes. Cellulosic content is not the only biodegradable material that could be valorized from a used disposable nappy. The percentage of organic content can reach almost 87% if urine and excreta are also added [14]. To date, the aforementioned recyclable materials and potential energy sources of this waste stream, are disposed of to landfills or led to incineration, due to collection of nappies with unsorted Municipal Solid Wastes (MSW) [15][16][17]. In the USA, 3.4 million tonnes of nappies are produced each year and more than 90% is sent to landfills [17].
Increasing living standards, consumerism and throw-away mentality, combined with unsustainable waste disposal practices have led to increased waste amounts in landfills. Biogas and leachate migrate away from the landfill boundaries and are released into the surrounding environment causing numerous adverse effects such as, health hazards, fires and explosions, vegetation damage, unpleasant odors, landfill settlement, ground water pollution, air pollution, and global warming [18,19]. Development of sustainable and effective waste valorization is mandatory to circumvent waste disposal and environmental problems.
This work aimed at assessing the effect of pH on the anaerobic fermentation of a mixed waste stream, consisted of fruit and vegetables (FVW), that cannot be consumed or are considered of low quality from food suppliers, and a DN hydrolysate. More specifically, the purpose was to identify the optimal pH value for maximum bio-H 2 production and valuable metabolites such as VFAs and ethanol. According 1 3 to literature, pH regulation plays an important role on the final performance of AF. Even narrow changes of pH in the bioreactor, affect significantly the microbial balance, metabolic pathways and consequently the metabolic products. In general, acidic pH < 4 favors the accumulation of acidic metabolites and ethanol and inhibits H 2 production [20], while at pH > 7 propionate is predominant and H 2 synthesis is limited. Thus, it is crucial to maintain pH constant at an optimum value to maximize H 2 productivity [2,9]. It should be noted that optimum pH for H 2 production, varies depending on the fermented substrate and therefore must be determined accordingly [21]. In the present study, constant pH, varying from 4.5 to 7.5, was studied in batch experiments at mesophilic conditions, with FVW and a mixture of FVW and DN hydrolysate as fermentation feedstocks.

Anaerobic Sludge
Anaerobic sludge, obtained from a municipal wastewater treatment plant in Metamorphosis (Attica, Greece), was used as inoculum at 15% v/v for all batch experiments. The sludge was boiled at 100 °C for 20 min, prior to experiments, in order to enrich H 2 -producing bacteria via the deactivation of methanogens and H 2 consumers [22]. The total solids (TS) content of the sludge was 15.6 ± 3.6 g/L, with 70 ± 2% of it being volatile solids (VS).

Fermentation Substrates
Fruit and vegetables, that did not meet quality standards for the supply chain, were collected from a Super Market in Patras (Achaia, Greece). Hard pieces like stalks, were removed and the rest were pulped, mixed and homogenized with an analytical mill (Sigma-Aldrich, IKA A11).
Used DN were collected from a private nursery in Chalandri (Attica, Greece). The DN were cut manually with scissors and the process described by Conway et al. [23] was followed in order to obtain a hydrolysate with the containing cellulose and excreta.
Both FVW and DN hydrolysate, were stored separately at − 18 °C until further processing.

Experimental Set-Up and Procedure
All experiments were performed in 1-L double wall, cylindrical, stainless steel (INOX 316) bioreactors with a working volume of 750 mL at constant conditions. Temperature was controlled via a thermocouple and was maintained at 37 ± 0.2 °C using hot water circulation. A motor drive unit, on the top of the bioreactor, ensured continuous stirring. pH was automatically controlled (HACH controller, SC200) and kept constant with the addition of 0.5 N NaOH solution through a peristaltic pump. The range of pH values tested was 4.5-7.5, using a step of 0.5 for the FVW substrate (diluted with tap water at 2:3 v/v ratio). Subsequently, batch experiments with FVW and DN hydrolysate at a 2:3 v/v ratio were conducted, using 15% (v/v) of the boiled sludge as inoculum. The second set of experiments tested only the pH values that considered to produce adequate amounts of bio-H 2 and VFAs. A total of four (4) pH values were tested in the mixture of FVW and DN hydrolysate, namely 6.0, 6.5, 7.0 and 7.5. A mixture of N 2 /CO 2 gases at 80/20% v/v ratio was used to purge the headspace of bioreactors, to ensure an initial anaerobic environment. In Fig. 1 the schematic diagram of experimental methodology followed for the batch tests using FVW and DN hydrolysate, is presented. It should be noted that each experiment lasted until stable metabolites concentration was reached in the cultivation medium (96-140 h).

Analytical Methods
Physicochemical characterization of the used substrates, as well as chemical analysis of the effluents from the bioreactor, was performed. During the experiments, samples were taken in order to determine the composition of the produced biogas, other metabolites such as VFAS, lactic acid and ethanol as well as total and dissolved COD (t-COD and d-COD), total and dissolved carbohydrates, TSS and VSS. Samples were taken every 4-6 h according to the evolution of the experiment. Οff-line pH measurements were conducted using an electrode (Thermo Scientific, Orion ROSS Ultra Refillable pH/ATC Triode), while alkalinity, TS, VS, TSS, VSS, t-COD, d-COD, TKN, ammonium nitrogen, total and dissolved phosphorus were determined according to Standard Methods [24]. Total and dissolved carbohydrates were measured according to Joseffson [25], phenolic compounds were determined according to Waterman and Mole [26], and VFAs as well as ethanol were analyzed on a gas chromatograph (Agilent Technologies, 7890A) equipped with a flame ionization detector, as described by Dareioti et al. [9]. The oven temperature was gradually increased from 110 °C (held for 5 min) to 250 °C (held for 6 min) at a rate of 15 °C/min. The carrier gas was Helium at a flow rate of 15 mL/min. The injector temperature was set at 175 °C and the detector at 300 °C. A capillary column (DB-FFAP, 30 m in length, 0.25 mm I.D. and 0.25 μm packing film) was used for determining the concentration of the individual VFAs, i.e. acetic, propionic, isobutyric, butyric, isovaleric, valeric and caproic acid. Lactic acid was measured with a DIONEX IC300 ion chromatography system using a thermostated (30 °C) Dionex IonPac analytical column (AS19 length 4 × 250 mm and 7.5 mm I.D), a guard column (4 × 50 mm length and 1 3 12 mm I.D) and an electron conductivity detector (Dionex). Fats and Oils were measured after extraction with hexane using a Soxhlet extractor (Velp Scientifica, SER 148). All experiments were carried out in triplicate. A custom-made equipment was used for the total biogas and bio-H 2 volume quantification. The apparatus is composed of a U-tube filled with engine oil, an electron valve and a counter. The volume of gases is measured by counting the number of displacements of constant oil volume. Biogas composition analysis was performed by gas chromatography with a capillary column (HP-PLOT/Q, 30 m in length, 0.53 mm I.D. and 40 μm packing film), a thermal conductivity detector (TCD) and nitrogen as carrier gas. Biogas production was converted to standard conditions (i.e. STP = 0 °C and 1 atm).

Physicochemical Characterization of Fermentation Substrates
Samples of the fermentation substrates were taken throughout the experimentation period, and measured in triplicate, in order to determine their characteristics and ensure that the batch tests are performed with constant feedstock. As shown in Table 1, DN hydrolysate is characterized by a neutral to alkaline pH (7.73 ± 0.04), which can be attributed to water and the presence of urine that has a high pH value when stored, due to the decomposition of urea and urate [27]. In general, pH of food waste is characterized as acidic with values close to 5.0 [28]; in this case pH of FVW is more acidic (3.48 ± 0.30) and therefore the resulting mixture of FVW/DN hydrolysate ends up with pH 4.67 ± 0.18. Furthermore, FVW presents high organic content, which is mainly attributed to the total carbohydrates' concentration of this waste stream (64.53 ± 7.46 g/kg ww). Total solids (TS) of FVW is also high (111.55 ± 3.99 g/kg ww) with 90.5% of it being volatile solids (VS), and COD:N:P ratio is 303:4.5:1. Regarding DN hydrolysate, the absence of soluble carbohydrates leads to the conclusion that cellulosic fibers from the nappies are the main source of carbohydrates. Although the biodegradable content of a used nappy is considerable (cellulose fibers plus excreta), yet DN hydrolysate presents rather low concentrations of nutrients and organic carbon. This result is due to the high dilution ratio, that is needed for the pre-treatment. The latter has an impact on the TS as well, which is rather low (8.07 ± 0.97 g/L).
Finally, the mixture FVW/DN hydrolysate has lower organic load than FVW, due to its dilution with the hydrolysate and its COD:N:P ratio is 317.5:6:1, which is almost ideal considering that for anaerobic processes the optimum operational ratio is 350:7:1 [29].

Effect of pH on FVW Anaerobic Fermentation
Batch experiments with FVW diluted with tap water at 2:3 v/v ratio were conducted. The range of pH values tested was 4.5-7.5, using a step of 0.5. The volume of produced biogas as well as its composition was analyzed throughout the experiments. Other parameters such as, VFAs and ethanol, lactic acid, t-Carbohydrates, d-Carbohydrates, t-COD, d-COD, TSS and VSS were determined at regular basis, in order to monitor the process.
Biogas production at STP conditions, is depicted in Fig. 2a. In all cases, biogas was composed of H 2 and CO 2 , while CH 4 was not detected. At pH 4.5 and 5.0 there was practically no H 2 , since acidic conditions inhibit metabolic activities related with H 2 production [2,30,31]. Maximum H 2 volume and yield (1008.1 mL, 1.345 L H 2 / L Reactor ) was obtained at pH 6.0, in accordance with researchers that investigated the effect of pH on H 2 production with different substrates in batch experiments [9,32]. At pH 7.0 and 7.5 though, H 2 volume prevailed CO 2, constituting 56.2% and 78.1% respectively of the total biogas volume obtained. VFAs and ethanol production (Fig. 2b) is a useful tool for the assessment of the process and information can be extracted regarding the H 2 yields, since acidogenesis and H 2 generation are closely related. At pH 4.5, acetic acid and ethanol were mainly produced, and their concentrations reached 1683.77 and 1380.94 mg/L respectively, with the rest of the VFAs concentrations being less than 40 mg/L. Lactic acid appeared at pH 5.0, having a high concentration (4681.94 mg/L), even though it was expected at lower pH values as well, according to literature [32,33]. Lactate concentration was followed by acetic acid (1468.06 mg/L) and ethanol (1093.75 mg/L). Various intermediate metabolites were produced at pH 5.5. Acetic acid (2135.90 mg/L) and ethanol (2197.59 mg/L) were predominant at the end of the fermentation, while lactic acid was also produced but was not detected after 39.2 h. Caproic acid appeared at 54.3 h of fermentation and butyrate at 68 h. Their concentrations reached 340.63 and 732.84 mg/L respectively. Isobutyrate and isovalerate were less than 85 mg/L, while propionate concentration was 684.40 mg/L. Total VFAs and ethanol production of 10,199.47 mg/L was obtained at pH 6.0, where the highest H 2 production was also observed, as stated previously. At the same pH value, butyrate had the highest concentration (6262.82 mg/L), which is in agreement with  literature [34] where it is mentioned that butyric and acetic acid type fermentations result in higher H 2 production. Zhang et al. [35] also reported butyric acid as the predominant VFA at pH 6.0, in kitchen waste anaerobic fermentation experiments. Acetic acid reached its maximum concentration (12,216.46 mg/L) at pH 6.5 followed by propionic and butyric acid with 1210.98 and 1500.01 mg/L respectively. Lower concentrations of acetic acid and butyric acid are noticed at pH 7.0, whereas at pH 7.5 the same pattern as at pH 6.5 is observed. Acetic acid reached 11,856.98 mg/L and butyric acid 1516.42 mg/L. Total carbohydrates removal, expressed in glucose equivalents, ranged from 70.4 to 88.3%. The highest degradation was observed at pH 6.5, as shown in Fig. 2c. The increment of pH values was followed by an increment on the carbohydrates consumption as well. Dissolved carbohydrates consumption ranged from 88.6 to 96.7%. Maximum consumption was observed at pH 7.5. It is also worth mentioning that the H 2 production yield (mole of H 2 produced per mole of consumed carbohydrates) reached maximum efficiency at pH 5.5, which was equal to 0.55 mol H 2 /mol equivalent glucose consumed, while the optimum theoretical yield is 4 mol H 2 /mol equivalent glucose with acetic acid as the main end-product. Dareioti et al. [9] and Thauer et al. [36] observed low yields (0.642 mol H 2 /mol equivalent glucose consumed) as well, at pH 6.0. Dareioti et al. [9] reported that acetic acid was not detected at pH 6.0, in a study on the effect of pH regarding the optimization of H 2 production from agroindustrial wastewaters. Likewise, Stavropoulos et al. [37] reported a maximum H 2 yield at pH 5.0, reaching 0.84 mol H 2 /mol equivalent glucose consumed, with dairy products as an easily fermentable substrate. Figure 3a depicts the concentrations of main end-products as a function of time, regarding the batch experiment that took place at pH 6.0, which resulted in the maximum H 2 production. Fermentation lasted 118 h and it is evident that acetic and ethanol type fermentations were predominant until the 70th h, with butyric type fermentation starting to evolve from that point on and then prevail by the end of experiment. Acetic acid reached a maximum of 4140 mg/L at the 67th h of fermentation and then followed a downward trend. Ethanol concentration appears to follow the same pattern as well, while butyric acid concentration reached a maximum (6016 mg/L) and appears to be stable until the end of the fermentation. Both acetic and butyric acid are usually produced during fermentation processes, but their dynamics vary due to alterations of environmental factors like pH, H 2 pressure and concentrations of metabolic products [38]. H 2 production (Fig. 3b) follows acetic acid production trend, which is in accordance with literature [36,38]. The maximum H 2 production is observed at 30th h which remained stable during fermentation with a minor increase (1%) after the 90th h. Regarding total and dissolved carbohydrates (Fig. 3c), their consumption is rapid until the 30th h, whereas a lower further reduction is observed, which is not in line with the vast increase of the butyric acid concentration. High concentration of butyric acid and simultaneous reduction of acetic acid, implies that the later may be enzymatically converted to butyric acid and AcCoA by butyric acid bacteria through the simultaneous consumption of butyryl-CoA [39].

Effect of pH on FVW/DN Hydrolysate Anaerobic Fermentation
Subsequent experiments, regarding the optimum pH value for maximum bio-H 2 production from FVW and DN hydrolysate were conducted. The feedstock in these second series of experiments consisted of FVW and DN hydrolysate mixture at a 2:3 v/v ratio. A total of four (4) pH values were tested in the mixture, namely 6.0, 6.5, 7.0 and 7.5 due to previous experiment series' findings. H 2 production at pH values below 5.5 was considered inadequate. Table 2 presents comparative detailed information on the produced bio-H 2 and main end-products concentrations at the end of each experiment, as well as on the carbohydrates' removal from both used substrates at all tested pH values.
Results shown in Fig. 4a indicate that the highest volume of H 2 was produced at pH 7.5 where it reached 3021.91 mL and a yield of 4.02 L H 2 /L Reactor . The change of the feedstock nature led to a threefold increment in H 2 volume, compared to FVW fermentation at a different pH value, probably due to microbial consortium enhancement from DN hydrolysate. This is evident also from the presence of 5% CH 4 (data not shown) at pH 7.5, which is an optimum value for methanogenesis, as methanogenic bacteria from the contained excreta of the used nappies could have been added in the mixture. Figure 4c presents the extent of carbohydrates degradation. Total carbohydrates consumption ranged from 88 to 90.5%, whereas dissolved carbohydrates followed the same pattern of removal and their percentage of consumption ranged from 97.4 to 98.2%. Besides the microbial enhancement from the DN hydrolysate, cellulose, which is part of the total carbohydrates, was also added in the mixture increasing thus the carbohydrates content of the feedstock. Concerning the main end-products formation (Fig. 4b), results denote that the main metabolic pathways followed by the anaerobic microbiota were butyric and acetic acid fermentation. Butyric acid reached a maximum (8189.72 mg/L) at pH 6.0 and acetic (7359.35 mg/L) at pH 7.5. Ethanol was again present in all batches but seemed to decrease with the increment of pH value. Lactic acid was detected at pH 6.0 and 6.5 while caproic acid showed a maximum concentration at pH 7.0 (1009.66 mg/L). Results were in accordance with other researchers as well [32,35]. At pH 7.5, total VFAs and ethanol concentration show the highest value among the experiments (17,162.19 mg/L). According to literature [40], pH values close to neutral seem to enhance H 2 -consuming than H 2 -producing bacteria. However, low concentration of propionic acid, along with high acetic acid and butyric acid concentrations, indicate that butyric acid-type fermentation was predominant among the metabolic pathways followed by the microbial consortium. According to recent publications [41,42] this type of fermentation presents 4.5 times greater H 2 production capacity compared with the propionic acid-type fermentation. More details about the batch experiment (pH 7.5), where maximum H 2 yield was obtained, are given in Fig. 5. The alteration on the fermentation substrate due to DN hydrolysate addition on FVW, resulted not only at a different optimum pH value for H 2 maximization, but also at a difference on the metabolic pathways followed by the microbial consortia and the final concentrations of metabolic products. The fermentation of FVW/DN hydrolysate lasted 137.65 h and Fig. 5a presents the kinetics of main end-products formation. Acetic and butyric acid are produced from the beginning of the fermentation following the same trend and reaching both a concentration of 7000 mg/L after 89 h. Equal abundance of these two acids has been previously reported at a pH range of 6.5-7.0 in a glucose fermentation by a mixed culture [43]. This mixed-acid metabolic pathway is common, in similar waste streams fermentations i.e. food waste [44]. In the present study, VFAs with high molecular weight, such as caproic, and ethanol were detected at low concentrations. Propionic acid reached 180 mg/L after the 64th h, probably due to the presence of acidogenic bacteria such as Corynebacteria, Propionibacterium and Bifidobacterium, which are responsible for this acid's production via transcarboxylase cycle [45], without excluding other species as well [46,47]. An alternative metabolic pathway for propionic acid formation involves the presence of lactate [43], which in our case is undetectable, and is thus most probably not followed.
Carbohydrates (Fig. 5c) present a reduction corresponding to the rate of main end-products formation, with 60.9% of them being consumed until the 20th h of fermentation, while the rest followed a slow pace reduction. H 2 production reached a maximum at the 32.7th h and remained stable from then onwards till the completion of fermentation (Fig. 5b).

Conclusions
The effect of pH on the anaerobic fermentation of FVW and DN hydrolysate for bio-H 2 production was investigated in this study. Bio-H 2 was produced efficiently and maximized by properly adjusting the pH values according to the utilized fermentation substrate. For FVW, as single feedstock, the maximum H 2 yield was observed at pH 6.0 reaching 0.52 mol H 2 /mol equivalent glucose whereas, the maximum concentration of ethanol and organic acids (15,600 mg/L) was reached at pH 6.5 for the same substrate. Regarding the mixed waste stream (FVW/DN hydrolysate), both H 2 production and metabolites concentration reached a maximum at pH 7.5 with 1.12 mol H 2 /mol equivalent glucose   and 17,160 mg/L respectively. The use of FVW/DN hydrolysate instead of FVW alone, led to higher yields of H 2 and main end-products of the anaerobic fermentation, as well as greater carbohydrates consumption. Although the highest H 2 production was observed at pH 7.5, such pH value is not suggested for continuous operation of a fermentation system due to increased possibility of converting the reactor to a methanogenic one. Moreover, the need to operate continuously at such high pH value and sustain it constant, requires continuous addition of large quantities of alkaline solution, resulting thus to higher operational cost than operating at suboptimal pH levels.