Effect of Carbon / Nitrogen Ratio , Temperature , and Inoculum Source on Hydrogen Production from Dark Codigestion of Fruit Peels and Sewage Sludge

This paper studies the use of fruit peel biomass and waste sludge from municipal wastewater treatment plants in the metropolitan area of Monterrey, Mexico as an alternative way of generating renewable energy. Using a Plackett–Burman experimental design, we investigated the effects of temperature, inoculum source, and the C/N (Carbon/Nitrogen) ratio on dark fermentation (DF). The results indicate that it is possible to produce hydrogen using fruit peels codigested with sewage sludge. By adjusting the C/N ratio in response to the physicochemical characterization of the substrates, it was revealed that the quantities of carbohydrates and nitrogen were sufficient for the occurrence of the fermentation process with biogas production greater than 2221 ± 5.8 mL LReactor and hydrogen selectivity of 23% (366 ± 1 mL H2·LReactor) at the central point. The kinetic parameters (Hmax= 86.6 mL·L−1, Rm = 2.6 mL L−1 h−1, and λ = 1.95 h) were calculated using the modified Gompertz model. The quantification of soluble metabolites, such as acetic acid (3600 mg L−1) and ethyl alcohol (3.4 ± 0.25% v/v), confirmed the presence of acetogenesis in the generation of hydrogen.


Introduction
The energy needs of economic activities have increased exponentially in the last few years, whereas fossil fuel reserves have diminished.According to the predictions from the U.S. Energy Information Administration (EIA), twice the current amount of energy will be needed by 2040 if we maintain the current annual growth rate of the U.S. Gross Domestic Product (GDP) of 2.6% [1].Therefore, to halt the negative effects on the environment due to carbon dioxide (CO 2 ) emissions and contribute to the generation of clean energy, it is necessary to develop renewable sources [2,3].Presently, various research groups throughout the world are exploring sustainable resources for obtaining dihydrogen (H 2 ) as an energy carrier [2,4,5].
Since 2000, H 2 has been considered one of the most viable alternative energy carriers for the future [6] due to its inherent characteristics, such as a higher energy content of 122 MJ kg −1 , which is approximately 2.75 times greater than that of its closest competitor, gasoline [2,7].Dihydrogen can be used directly in fuel cells for electricity generation [8,9] and produces only water as a byproduct from its combustion [10].From an environmental standpoint, the production of H 2 from biological processes that involve organic waste is an important area of opportunity in the production of bioenergy [11,12].In such cases, the H 2 produced is also called biohydrogen.Although biohydrogen can be generated through various processes, a number of investigators have described the fundamentals of the biological production of H 2 using photolysis or dark fermentation (DF).The latter is considered a simpler method and has the advantage of using municipal, agricultural, and industrial wastes as starting materials [13][14][15][16].
Gómez-Romero has described the use of different municipal wastes or industrial wastewater as substrates for DF [17].Data obtained in these studies show that the production of biohydrogen could be reduced or even inhibited due to some specific characteristics of wastewater or organic wastes, such as deficient buffer capacity, nutrient imbalance, and the presence of microbial populations that may consume biohydrogen or produce methane.One strategy to overcome these problems is the use of codigestion processes, due to the positive synergistic effect of the mixed material with complementary optimizing conditions for hydrogen production: controlling pH, and optimizing the C/N ratio [18].In fact, the C/N ratio (20 to 40) is used as a reference parameter associated with the enhancement of the buffer capacity of the system [17,[19][20][21].Currently, some geographic regions in the world have been using biowaste as a resource in anaerobic codigestion processes for H 2 production.This includes cheese whey-vegetables or vinasse-nejayote in Mexico [17,22], fruit-vegetable-fish waste in Tunisia [23], coffee residues-sugarcane-vinasse in Brazil [24], or mozzarella cheese whey and buttermilk in Italy [25].
The most recent publications have proposed the use of fruit wastes as substrates that are rich in carbohydrates, which provide a nutritional balance with other substrates that are rich in polysaccharides and proteins [19,26], in addition to providing a convenient habitat for the proliferation of H 2 , producing its natural microflora [27,28].Saidi et al. were able to duplicate hydrogen production (28 mmol h −1 L −1 ) in a batch reactor in the presence of Thermotoga maritima with a C/N ratio of 22, which was increased three-fold due to the codigestion of fruit-vegetable-fish waste [23].On the other hand, Gómez-Romero demonstrated that the adaptation of the predominant microbial community of Bifidobacterium was 85.4%, followed by Klebsiella (9.1%), Lactobacillus (0.97%), Citrobacter (0.21%), Enterobacter (0.27%), and Clostridium (0.18%) in a reactor fed with milk serum and fruit wastes at different C/N ratios (5 < C/N < 46).The highest specific biohydrogen production rate was 10.68 mmol h −1 L −1 for a C/N ratio of 21 [17].Furthermore, Bikram also studied the codigestion of fruit and vegetables mixed with cottage cheese serum, which obtained a maximal hydrogen productivity of 118.12 ± 1.05 mmol L −1 with a C/N ratio of 68 [21].In contrast with the results reported by Gómez-Romero, the predominant species were Escherichia coli, Clostridium butyricum, and Streptococcus henryi.
In developing countries, the majority of municipal solid wastes comprise fruit peels and vegetable wastes according to the literature [29,30].The metropolitan area of Monterrey (MAM), generated approximately 800 thousand tons of organic waste in 2017 [31].This amount of biomass with a high content of hydrolysable carbohydrates was an attractive criterion for selecting fruit peels as strong substrates for the production of biofuels, in combination with sewage sludge produced from the biological process.Generally, sewage sludge is treated by an anaerobic digestion process to produce methane; however, hydrogen is an important intermediate product in this process.Thence, both wastes can be conveniently processed and valorized in a dark codigestion process, because the raw materials are abundant throughout the year and there is zero cost.
Based on the above, this paper describes an exploratory study of dark codigestion using fruit peels (FP) with sewage sludge (SS) generated from a municipal wastewater-treatment plant.For this, three variables-temperature, two sources of inoculum, and the C/N ratio-were studied, employing a Plackett-Burman (PB) experimental design to identify the influence of these parameters on the degradation of the organic wastes for producing H 2 as a renewable energy source.

Substrate Preparation
The DF process was developed using a codigestion procedure on a mixture of different FP, such as melon (Cucumis spp.), papaya (Papaya spp.), and pineapple (Ananas spp.), which were obtained from a vegetarian restaurant in the locality, and SS from a municipal wastewater-treatment plant in the city of Monterrey, Nuevo León.This allowed us to meet the nutritional contribution of the substrate.Fruit peels were separated and ground using an industrial blender (Waring Laboratory Science, BLENDOR model).The SS was thermally treated at 100 • C for 1 h [32]; after this, the FP and SS were combined and the proportions were adjusted to obtain C/N ratios of 20, 30, and 40.

Inoculum Source
Two inoculum sources were used for the production of biohydrogen.The first source derived from the anaerobic sludge (AS) obtained from the wastewater-treatment plant of Monterrey Water and Drainage Services (Servicios de Agua y Drenaje de Monterrey-SADM), whereas the second source, granular anaerobic sludge (GAS), was recovered from the wastewater-treatment plant of a beverage company found in the locality.The physicochemical characteristics are listed in Table 1, revealing that GAS presents a major percentage of biomass.The inocula were treated separately in a water bath at 100 • C for 1 h to inhibit methanogenic microorganisms and guarantee the selective production of biohydrogen [32].

Batch Bioreactors
A PB experimental design was used to evaluate the productivity of the codigestion process in the production of hydrogen.The different combinations of the three factors: temperature (25, 37, and 55 • C), the C/N ratio (20, 30, and 40), and inoculum source (AS and GAS), are summarized in Table 2.For each of the 18 combinations, vessels with 124 mL of total volume were used as bioreactors, with a working volume of 80 mL.

Analytical Techniques
The following parameters were calculated during the process of characterizing the substrates: Total Solids (TS); Volatile Solids (VS); Total Carbohydrates (TCH); and pH.These parameters were analyzed according to standard methods [35].Chemical Oxygen Demand (COD) was measured using the Hach standard method, which was based on the digestion of the sample with potassium dichromate at a temperature of 150 • C under acidic conditions [36].The phenol-sulfuric method, which involves the hydrolyzing of sugars by the addition of acid to form colored complexes with phenol [37], and the Lowry method, which adds reactants containing Cu 2+ to form blue complexes with proteins [38], were used to analyze the composition of carbohydrates and proteins, respectively, in the FP.
The total volume of biogas was measured at each time interval by the plunger displacement method utilizing syringes, ranging from 10 to 50 mL [39].The amount of hydrogen in the gas was quantified in a gas chromatograph (SRI 8610C) with a Thermal Conductivity Detector (TCD) and two packed columns, a silica gel and a molecular sieve, both of which had dimensions of 6 × 1/8 .The temperatures of the injector and detector were 90 • C and 150 • C, respectively.The gas carrier was nitrogen at a flow rate of 20 mL min −1 .Acetic acid was monitored during DF using distillation-titration [40,41].The other parameters analyzed were pH, TCH, COD [35][36][37], and ethanol, which was measured with a digital refractometer (Wago, Model PET109, Monterrey, Mexico).

Data Analysis
The PB design was used to explore key factors in the production of biogas and hydrogen by the codigestion of FP and SS.Three variables were studied (C/N ratio, temperature, and inoculum source) and the biogas and H 2 cumulative volumes were selected as the dependent output across 18 runs.The software used was MINITAB ver.15.The design matrix is presented in Table 2.
Cumulative biogas production curves were obtained from the experimental design and analyzed employing the modified Gompertz model to acquire the hydrogen production potential (Hmax), the hydrogen production rate, and the lag phase [42,43].
where H is the cumulative hydrogen production (mL L −1 ), H max is the hydrogen production potential (mL L −1 ), Rm is the maximal hydrogen production rate (mL L −1 h −1 ), λ is the lag phase time (h), t is the time (h), and e is the Euler constant (2.71821).The kinetic parameters were determined by fitting the cumulative hydrogen production curve by minimizing the ratio of the sum of the square error to the correlation coefficient (r 2 ) using the Solver function in Microsoft Excel ver.10.

Characterization of Substrates
The physical and chemical characteristics of the FP are essential for the design and operation of anaerobic digesters, because they will affect the production of biogas and the stability of the DF process.Table 3 shows the results of the physicochemical characterization of the FP and the SS.Both substrates had a high humidity (ranging from 89.61 to 92.43%), which can facilitate the degradation process [34].The amount of nitrogen increased when SS was used.The C/N ratio is important, as it is closely related to the ability to produce hydrogen [40].Furthermore, nitrogen is the most essential nutrient, in that a proper supply of nitrogen could maximize microbial growth but excessive addition could cause ammonia inhibition [44].

Measuring Biogas
The progress of the 18 bioreactors was closely monitored.The microbial activity in the production of biogas was greater at 25 • C than at 55 • C. The cumulative volume of biogas after 140 h is presented in Table 4.The greatest cumulative gas volumes correspond to the central points of the PB matrix, with an average value of 2221 ± 5.8 mL•L −1 Reactor for the R6, R14, and R16 reactors.The most effective reactors are located at the central points of the experimental design with the following parameters: C/N ratio of 30, temperature of 37 • C, and GAS inoculum.This C/N ratio is consistent with ratios found by other authors in anaerobic processes for methane production, which permits us to theorize that DF and anaerobic digestion can be performed utilizing the same C/N ratio [45].When the same C/N ratio and temperature were used with AS, biogas production decreased by 30%.This result is consistent with the least productive series of reactors (<500 mL L −1 Reactor ), which was also inoculated with AS and incubated at 55 • C.
The Pareto charts of the standardized effects in Figure 1a reveal that the type of inoculum is the most significant factor determining the volume of gas produced.The efficiency of GAS can likely be attributed to the spherical form of the granules (<2 mm), which distributes the different microorganisms in layers [46].According to Hernández et al., the fermentative bacteria are distributed in the outer layers of the granule, facilitating direct contact with the substrates.In contrast, the intermediate layers contain acetogenic bacteria, including those that produce hydrogen, where the population dynamics aid in the oxidation-reduction processes [47].
Figure 1b shows the Pareto chart for H 2 production.The inoculum source used was again the most important factor, followed by temperature.A C/N ratio of 30 promoted the greatest production of H 2 in the codigestion, which was attributed to the synergistic effects of the nutritive balance and the buffer capacity that the sludge contributes to the mix to control pH naturally.Thus, this avoids acidification of the reactors and the inhibition of hydrogen production [47].

Biohydrogen Production
The chromatographic analysis of biogas samples showed that H2 and CO2 were reaction products and methane was not detected in any of the samples.The production of H2 exhibited trends similar to those of the generation of biogas, with a selectivity of 20-23% in tests of R2, R17, R18, R6, R14, and R16.Other reactors, including R4, R9, and R13, which had lower (<8%) selectivity in the production of H2, were activated with the same inoculum source but required a C/N ratio of 40 and temperature of 25 ± 2 °C.These results are no better than those in the literature or in comparison to the stoichiometry of the reaction, but they demonstrate that it is possible to use SS to generate renewable energy.The conventional management and treatment of SS implies a worldwide problem in finding locations for its ultimate disposal and a risk of groundwater pollution [48]; therefore, this is a promising alternative treatment [49].
The kinetic parameters obtained from the experimental data of the reactors are described in Table 5 using the Gompertz model (Equation (1)).

Biohydrogen Production
The chromatographic analysis of biogas samples showed that H 2 and CO 2 were reaction products and methane was not detected in any of the samples.The production of H 2 exhibited trends similar to those of the generation of biogas, with a selectivity of 20-23% in tests of R2, R17, R18, R6, R14, and R16.Other reactors, including R4, R9, and R13, which had lower (<8%) selectivity in the production of H 2 , were activated with the same inoculum source but required a C/N ratio of 40 and temperature of 25 ± 2 • C.These results are no better than those in the literature or in comparison to the stoichiometry of the reaction, but they demonstrate that it is possible to use SS to generate renewable energy.The conventional management and treatment of SS implies a worldwide problem in finding locations for its ultimate disposal and a risk of groundwater pollution [48]; therefore, this is a promising alternative treatment [49].
The kinetic parameters obtained from the experimental data of the reactors are described in Table 5 using the Gompertz model (Equation (1)).The behavior of the H 2 production for all treatments was adequately described by the model, with a correlation coefficient (r 2 ) of around 1.0.Lag time ranged from 1.95 to 22 h depending on the C/N ratio.For the R14 studied at 37 • C (C/N = 30), H max was 86.6 mL L −1 .For R7 and R13 with C/N ratios of 20 and 40 analyzed at 25 • C, H max values were 33 and 17.5 mL L −1 , respectively.Finally, for R15 at 55 • C, H max was 26.9 mL L −1 h −1 .The highest hydrogen production rate was 2.6 mL L −1 h −1 from R14.This result could be explained by the positive synergism established between the mixture of organic substrates (FP and SS) and the inoculum.However, the hydrolysis of fruit peels, such as melon or pineapple, could be a limiting step in anaerobic codigestion, which should be overcome in order to improve its performance [17].

Monitoring Parameters during DF
As part of the study, the development of the reactors inoculated with GAS, pH, TCH, and COD was monitored (see Table 6).First, pH is one of the most important environmental factors that can affect the metabolic pathway and the efficiency of the production of H 2 .In this study, the pH of all of the reactors was adjusted to 5.5 and, at the end of the reaction, the trials with the greatest H2 production (R6, R14, and R16) had a final pH of 5.1 (Figure 2). Figure 2 shows that the pH range for maximal H 2 production is 5.2-6.0,whereas the reactors showing lower biogas and hydrogen production had final pH values as low as 4.3.These values represent the acidification of the reactors as a result of the generation of other soluble metabolites, such as acetic acid (Figure 3) and ethanol.The production of hydrogen occurs primarily during the exponential growth phase of Clostridium, whereas during the stationary phase, there is a change from acidogenesis to solvent production [50].These changes in microbial metabolism occur at pH values below 4.5 [51], which are the same pH values observed in the reactors at the end of the DF process.All of the reactions consumed 45-60% of the total carbohydrates during the 140-h incubation.The greatest change occurred during the first 24 h, especially in reactors with a C/N ratio of 40.Despite a link between carbohydrate consumption and energy transformation, there was no evidence of this effect in these results.
Soluble COD represents the organic content of biodegradable substrates.Therefore, its removal is an indicator of the degree of hydrolysis achieved by the acidogenic bacteria [40].In this study, it was evident that the concentrations of COD exhibited the same trends as the carbohydrates due to their average decrease of 60%.The greatest COD decrease was observed in the first 40 h of codigestion and coincided with the adaptation and exponential growth stages of the microorganisms.
Table 7 depicts the optimal operation condition and hydrogen yield (HY) from different codigestion studies using different organic substrates.As shown in Table 7, the HY obtained in this study is slightly lower than that reported in other papers using mixtures of codigestion in batch reactors; however, the removal percentage of COD (50%) was the highest.It is noteworthy that HY depends not only on the rate of existing carbohydrates but also on the group of operation conditions of the reactor, such as pH, COD, and inoculum.
In fact, pH control within an optimal interval is decisive for the growth of biohydrogenproducing bacteria [17,28,34,52].Even taking this into account, the development of this proposal in Monterrey, Mexico comprises a preceding action for the achievement of wastes as valuable resources   All of the reactions consumed 45-60% of the total carbohydrates during the 140-h incubation.The greatest change occurred during the first 24 h, especially in reactors with a C/N ratio of 40.Despite a link between carbohydrate consumption and energy transformation, there was no evidence of this effect in these results.
Soluble COD represents the organic content of biodegradable substrates.Therefore, its removal is an indicator of the degree of hydrolysis achieved by the acidogenic bacteria [40].In this study, it was evident that the concentrations of COD exhibited the same trends as the carbohydrates due to their average decrease of 60%.The greatest COD decrease was observed in the first 40 h of codigestion and coincided with the adaptation and exponential growth stages of the microorganisms.
Table 7 depicts the optimal operation condition and hydrogen yield (HY) from different codigestion studies using different organic substrates.As shown in Table 7, the HY obtained in this study is slightly lower than that reported in other papers using mixtures of codigestion in batch reactors; however, the removal percentage of COD (50%) was the highest.It is noteworthy that HY depends not only on the rate of existing carbohydrates but also on the group of operation conditions of the reactor, such as pH, COD, and inoculum.
In fact, pH control within an optimal interval is decisive for the growth of biohydrogenproducing bacteria [17,28,34,52].Even taking this into account, the development of this proposal in Monterrey, Mexico comprises a preceding action for the achievement of wastes as valuable resources All of the reactions consumed 45-60% of the total carbohydrates during the 140-h incubation.The greatest change occurred during the first 24 h, especially in reactors with a C/N ratio of 40.Despite a link between carbohydrate consumption and energy transformation, there was no evidence of this effect in these results.
Soluble COD represents the organic content of biodegradable substrates.Therefore, its removal is an indicator of the degree of hydrolysis achieved by the acidogenic bacteria [40].In this study, it was evident that the concentrations of COD exhibited the same trends as the carbohydrates due to their average decrease of 60%.The greatest COD decrease was observed in the first 40 h of codigestion and coincided with the adaptation and exponential growth stages of the microorganisms.
Table 7 depicts the optimal operation condition and hydrogen yield (HY) from different codigestion studies using different organic substrates.As shown in Table 7, the HY obtained in this study is slightly lower than that reported in other papers using mixtures of codigestion in batch reactors; however, the removal percentage of COD (50%) was the highest.It is noteworthy that HY depends not only on the rate of existing carbohydrates but also on the group of operation conditions of the reactor, such as pH, COD, and inoculum.In fact, pH control within an optimal interval is decisive for the growth of biohydrogen-producing bacteria [17,28,34,52].Even taking this into account, the development of this proposal in Monterrey, Mexico comprises a preceding action for the achievement of wastes as valuable resources for the generation of economical clean energies.In addition, the findings related to the operational conditions in this work can apply for the production of other bioenergetics, such as methane.

Production of Soluble Metabolites
As reported in the literature, fermentation can produce solvents such as ethanol, and volatile fatty acids such as acetic acid, butyric acid, and propionic acid.According to Equation (2), 4 moles of hydrogen can be obtained together with acetic acid, whereas only 2 moles are obtained together with butyric acid (Equation ( 3)).
Evidence of hydrogen production during DF was obtained in part by quantifying acetic acid and ethanol in the reactors.The entire series of experiments (Table 6) demonstrated an increase in the production of acetic acid, primarily in reactor R6 (3600 mg L −1 ), followed by reactors R4, R9, and R14, in which the production was 45-60% lower.This is consistent with an analysis reported in the literature, in which the production of hydrogen was shown to be linked with that of volatile fatty acids, specifically acetic acid [53].In contrast, the presence of ethanol is indicative of the dominance of the acidogenesis process in the reaction.Reactors R6, R14, and R16 maintained a constant concentration of ethanol (3.25 ± 0.25% v/v) throughout the entire experiment.Reactors run at 55 • C did not show any clear trends.

Practical Application and Future Perspectives
The codigestion of FP biomass and waste sludge from municipal treatment plants for producing biofuels involves different biochemical reactions during anaerobic degradation.From this derives the importance of balancing the nutritional support of the substrates in the feeding process, because these latter exert an influence on the nature of the microbial communities that prosper and that direct the selectivity of the biological processes.In this exploratory work, we studied the interaction of the mixed microbial consortium (GAS) and the operational variables of the reactor (C/N ratio, temperature) as determining factors in the production of hydrogen.However, the work also allowed us to determine the presence of other parallel reactions with sub-products with a different commercial value and industrial use.Therefore, the results of this study can also be applied successfully as a useful tool for achieving total control of the physical, chemical, and anaerobic process in order to increase selectivity in terms of the production of a specific, final metabolite.

Conclusions
From the results from the PB experimental design, we concluded that the factors of greatest significance for the codigestion of fruit peel biomass and waste sludge from municipal treatment plants were the inoculum source for biogas production as well for hydrogen production, followed by a C/N ratio for the latter.
The codigestion of the reactor batch with the C/N = 30 ratio and using GAS inoculum produced the greatest accumulative volume of biogas of 2221 ± 5.8 mL L −1 reactor, with selectivity of hydrogen of 23% and H 2 yield of 27 mL H 2 gVS −1 .
The production of acetic acid increased during the development of codigestion as probable evidence that the mixture follows, among others, the metabolic route of acetogenesis where hydrogen production is favored.
The kinetic model of dark fermentation was adjusted to the Gompertz model, in particular the microreactors (R6, R14, and R16), with a C/N ratio = 30, GAS inoculum with yield of H 2 max = 86.6 mL L −1 , Rm = 2.6 mL L −1 h −1 , and λ = 1.95 h.

Figure 1 .
Figure 1.(a) Pareto chart of standardized effects for Biogas production, and (b) Pareto chart of standardized effects for H2 production.

Figure 1 .
Figure 1.(a) Pareto chart of standardized effects for Biogas production, and (b) Pareto chart of standardized effects for H 2 production.

Figure 2 .
Figure 2. pH measurements at central point reactors during DF.

Figure 3 .
Figure 3. Acetic acid production at central point reactors during DF.

Figure 2 .
Figure 2. pH measurements at central point reactors during DF.

Figure 2 .
Figure 2. pH measurements at central point reactors during DF.

Figure 3 .
Figure 3. Acetic acid production at central point reactors during DF.

Figure 3 .
Figure 3. Acetic acid production at central point reactors during DF.

Table 1 .
Characteristics of inoculum sources.

Table 4 .
Biogas and per H 2 production bioreactor.

Table 5 .
Gompertz equation coefficients for hydrogen production at different C/N ratio evaluations.
I 0 = Initial value; F = Final value.

Table 7 .
Comparison of the hydrogen yield (HY) from different codigestion studies.