Hydrogen Production Through Anaerobic Co-Digestion of Different Agroindustrial Waste and Food Waste at Mesophilic Conditions

Abstract

Mesophilic anaerobic co-digestion of eight distinct substrate mixtures of agroindustrial and food wastes was assessed to determine the most efficient waste mixture for maximizing hydrogen production. To evaluate the impact of adding various mixtures on dark fermentation (DF), batch tests were conducted for 250 h at 37 °C and a pH range between 5.0 and 5.9. Ethanol, butyric, propionic, acetic, and isobutyric acids were identified as the principal fermentation end products. The hydrogen production rate reached in a decreasing order from a mixture comprising 55% Olive Mill Wastewater (OMW), 40% Cheese Whey (CW), and 5% Sewage Sludge (SS) or Liquid Pig Manure (LPM) (38 NmL/gVS) to 55% OMW, 40% CW and 5% diluted Food Waste (FW_dil) (30 NmL/gVS), 60% CW and 40% Grape Residues (GR) (27 NmL/gVS), 80% CW and 20% LPM (13 NmL/gVS), 60% OMW and 40% FW_dil. (10 NmL/gVS), 60% CW and 40% FW_dil, (8 NmL/gVS) and 70% OMW and 30% SS (5 NmL/gVS). These results indicated that H₂ was generated through mixed fermentation pathways, while the addition of OMW > 55% inhibited microbial activity and reduced hydrogen production. The highest hydrogen yield (38 NmL/gVS), accompanied by 27.6%, Volatile Solids (VS) reduction and the highest Volatile Fatty Acids (VFAs) concentration (6.1 g/L). The same substrate mixture resulted in the highest accumulation of acetic and butyric acid in the acidified effluent, indicating the dominance of hydrogen-producing metabolic routes. The data suggest that co-fermentation of the selected residues not only enhances hydrogen production but also creates more stable operational conditions -including improved pH regulation, increased carbohydrate conversion, and greater VFAs accumulation- making the process more robust and viable for practical application.

1. Introduction

The world’s primary energy source is fossil fuels, but using them contributes to climate change, pollution, and the greenhouse effect [1]. Switching to carbon dioxide-neutral and environmentally benign renewable energy sources could solve these issues. It is anticipated that fossil fuels will eventually be replaced by wind, solar, hydro, and biomass energy [2]. Since hydrogen has the highest energy density of any fuel and produces only water as a byproduct, it is considered a highly desirable energy transporter [3]. Fossil fuels (such as coal and natural gas) and organic wastes (including plant and animal wastes, fruit and vegetable wastes, and agricultural wastes) can all be used to produce hydrogen using a variety of biological technologies [4]. To protect the integrity of biological systems, the environment must be cleaned up, and industrial effluents must be excluded [5]. Interestingly, the high organic content of industrial wastewater offers the possibility of producing hydrogen through biological processes. However, increasing attention is also being directed toward hydrogen production from renewable sources to support the sustainable development of a hydrogen-based society [6]. Therefore, the growing interest in sustainable and low-emission technologies suggests that hydrogen production as an alternative fuel source may become a viable and widely adopted solution in the near future. In fact, from 2024 to 2030, the global hydrogen generating market is projected to expand at a compound annual growth rate (CAGR) of 9.3%, from its estimated USD 170.14 billion in 2023.

Dark fermentation (DF) is one of the most commonly employed methods in biological hydrogen production processes largely due to its ability to reduce waste, generate energy, and enhance the economic feasibility of the process. The metabolic mechanics of DF, an anaerobic process, are very similar to those of biomethane synthesis [7]. Nevertheless, effective DF requires substrates rich in carbohydrates, which severely restricts the range of suitable feedstocks. Additionally, optimizing and maintaining optimal pH levels presents a major challenge for continuous biohydrogen generation. The ideal pH range for DF is 5.5 ± 1, which provides a broad window for biological activity, however without proper control, the accumulation of VFAs, during DF can lead to acidification, ultimately inhibiting microbial performance [7]. Some of the main challenges that still prevent the commercial development of biohydrogen production are low yields and rates of H₂ production, poor substrate conversion efficiency, incomplete substrate degradation and partial substrate conversion into organic acids and CO₂ [8]. Agro-industries such as pig farms, cheese factories, vineyards, and olive mills account for a sizeable portion of the global economy, particularly in the Mediterranean region. Millions of tons of wastewater and numerous byproducts are produced by these sectors; many of which remain completely untapped and pose a threat to the environment [9]. In recent years, several research studies have focused on the use of agro-industrial wastes as substrates in fermentative processes to produce molecules of commercial relevance, like VFAs [10]. The most frequently studied VFAs are lactic, propionic, butyric, acetic, and formic acids. VFAs have found important applications in industries such as biofuels, food preservation, pharmaceuticals, chemicals, and agriculture. Importantly, VFAs can also serve as precursors in the production of biopolymers such as polyhydroxyalkanoates (PHAs) which are used in the manufacture of biodegradable plastics [11].

VFAs are of paramount importance in the context of DF as they serve as crucial intermediates within the metabolic pathways that culminate in hydrogen production. Throughout the DF process, carbohydrates are converted through glycolysis into pyruvate, which is subsequently oxidized to form VFAs such as acetic, butyric, and propionic acids, accompanied by the release of H₂ and CO₂. The proportional allocation of these VFAs significantly impacts hydrogen output, as fermentations of the acetic and butyric varieties commonly exhibit heightened hydrogen generation rates, while the accrual of propionic acid signifies pathways that consume hydrogen [12,13]. Consequently, comprehensive monitoring and regulation of VFA synthesis are indispensable for the enhancement of hydrogen synthesis efficiency and the preservation of steady anaerobic functionality.

Microorganisms hydrolyze complex carbohydrates into glucose and other monosaccharides, which are subsequently metabolized via glycolysis to form pyruvate. Pyruvate is then further converted into VFAs, hydrogen, and other fermentation products [14]. Co-fermentation of multiple substrate types could offer an attractive alternative strategy to maintain the pH of the DF process within the optimal range. Previous studies have assessed how co-fermentation of two components affects the production of hydrogen [15]. To optimize the response variables, a thorough investigation of the impact of substrate composition on the DF process needs to be performed as relevant studies are limited. To address this gap, mixture design -a unique experiment design that focuses on composition optimization for multi-component systems- can be employed. Experimentally, in mixture design the relative proportions of the components in the mixture determine the response, but the overall number of substrates is kept constant. It has a significant benefit over traditional one-factor-at-a-time techniques and is particularly valuable in assessing the interacting effects of substrates [15].

The current yield of biohydrogen produced through DF ranges from 1.2 to 2.3 moles of hydrogen per mole of hexose. This corresponds to approximately 30–50% of the theoretical maximum yield, which is four moles of hydrogen per mole of glucose [16]. To quickly, affordably, and effectively evaluate the specific hydrogen production (SHP) of a substrate, biochemical hydrogen potential (BHP) tests are commonly employed in the literature. BHP assays determine the potential amount of hydrogen that can be generated when a specific substrate or a mixture of waste undergoes biodegradation under fermentative conditions. These tests are especially important as initial experimental evaluations to assess the viability of applying DF treatment to specific wastes [17].

The intricate nature of multi-substrate co-fermentation systems necessitates the utilization of the mixture design methodology as a comprehensive and structured statistical tool to enhance the efficiency of substrate composition optimization. This approach facilitates the examination of cooperative and opposing relationships between distinct feedstocks, leading to the determination of their most advantageous ratios to amplify hydrogen yield. In contrast to traditional sequential experimentation techniques, the mixture design strategy adeptly accounts for the interdependent influences among constituents, proving especially beneficial for heterogeneous agro-industrial residues with varying nutrient profiles and biodegradation attributes. In this particular context, our hypothesis posits that the amalgamation of organic waste materials possessing complementary attributes—such as the carbohydrate-rich CW, the nutrient-balanced SS and LPM, along with the phenolic compound-containing OMW—is likely to augment the efficiency of hydrogen production. This enhancement is expected to occur through the facilitation of advantageous metabolic pathways and the maintenance of pH levels within the optimal spectrum for DF. The integration of GR and FW serves to broaden the array of carbon and nitrogen inputs, potentially enhancing microbial functionality and operational steadiness. The principal objective of this investigation, therefore, is to determine the most efficacious combinations of substrates conducive to biohydrogen production, employing a mixture design strategy.

Along these lines, the primary objective of the present study was to develop an innovative strategy for optimizing substrate composition to enhance biohydrogen production, using a statistical mixture design approach. This research uniquely explores the synergistic interactions among a diverse range of readily available agro-industrial and municipal wastes, including food waste (FW), pig manure (PM), cheese whey (CW), grape residues (GR), olive mill wastewater (OMW), and sewage sludge (SS). A key novelty of this study lies in the strategic blending of multiple waste streams, which not only enhances hydrogen yields but also naturally maintains the pH below 6.0, eliminating the need for chemical adjustments. In addition, the study develops substrate mixtures with stable composition and an optimized balance of VFAs, crucial intermediates for efficient dark fermentation. By integrating these approaches, the present work advances previous research by providing a resource-efficient, holistic framework that maximizes biohydrogen production while simultaneously improving process stability through the coordinated utilization of diverse agro-industrial and municipal wastes.

2. Materials and Methods

2.1. Feedstock Used in Fermentation Experiments

SS was the main sludge derived from the Municipal Sewage Treatment Plant (MSTP) in Heraklion, Crete, which has a population of around 175,000. Before being used, the sludge was kept in the refrigerator at 4 °C for less than 24 h to minimize compositional changes. The fresh OMW used in this investigation came from a Heraklion-based olive oil production facility that extracts olive oil using a three-phase decanter centrifugation technique. The OMW sample was stored in the freezer at −18 °C to ensure a consistent feed composition during the experimentation period due to its seasonal production and fermentation tendency. The CW was acquired from a nearby cheese factory that makes cheese using conventional methods. The Hellenic Mediterranean University (HMU) student restaurant in Heraklion provided the Food Waste (FW) used in this investigation. On a wet-weight basis, the FW composition was 10% bread, 20% raw-fresh food (vegetables), and 70% cooked food. For the purpose of pumping and stirring in the laboratory-scale experimental apparatus, FW was homogenized using a mechanical mixer (Rohnson, Athens, Greece) (about 4.0 mm in size) and diluted with tap water to about 3% Total Solids (TS). To prevent any kind of obstruction during the investigation, the collected FW was combined and blended into a paste using an electrical mixer (Rohnson, Athens, Greece) before being filtered through a sieve. Τhe filtered material was kept in the refrigerator at 4 °C for less than 24 h. Pig manure (PM) was collected from a nearby pig farm that feeds 70 sows. PM underwent a 24-h thermal pretreatment at 75 °C, a 1:4 water dilution, and a 4-mm-diameter mesh screen before being stored at 4 °C. Grape residues (GR) were acquired from Alexakis SA winery in Heraklion, Crete. For GR, thermal pretreatment was performed at 60 °C for 60 min and the mixture was diluted 1:4 with water.

Table 1. Composition of Sewage Sludge (SS), Food Waste (FW), Grape Residues (GR), Cheese Whey (CW), Olive Mill Waste Water (OMW) and Liquid Pig Manure (LPM).

Parameters	SS	FWdil	GR	CW	OMW	LPM
pH	6.3 ± 0.3	4.5 ± 0.1	3.5 ± 0.1	5.2 ± 0.2	4.4 ± 0.1	5.8 ± 0.1
TS (g/L)	38.2 ± 0.7	54.7 ± 0.6	12.7 ± 0.8	100.6 ± 3.9	79.1 ± 0.3	26.3 ± 0.9
VS (g/L)	26.1 ± 0.5	51.2 ± 0.7	8.7 ± 0.7	90.5 ± 4.0	71.5 ± 0.3	21.7 ± 0.4
TCOD (g/L)	45.3 ± 7.6	74.4 ± 10.0	19.8 ± 7.9	134 ± 2.4	185.3 ± 22.4	35.8 ± 6.2
sCOD (g/L)	1.4 ± 1.1	1.9 ± 0.5	16.1 ± 1.4	75.4 ± 3.2	11.4 ± 0.7	10.6 ± 3.1
BOD5 (g/L)	8.0 ± 0.0	27.0 ± 0.0	6.0 ± 0.0	52.00 ± 0.0	20.0 ± 0.0	11.0 ± 0.0
TN (g/L)	1.4 ± 0.0	1.5 ± 0	0.01 ± 0.0	2.1 ± 0.0	1.3 ± 0.0	1.3 ± 0.1
Proteins (%)	0.9 ± 0.0	0.9 ± 0.0	0.6 ± 0.0	1.3 ± 0.0	0.8 ± 0.0	0.8 ± 0.0
TP (mg/L)	137.1 ± 0.1	25.8 ± 0.2	193.0 ± 0.5	201 ± 1.00	236 ± 2.05	74 ± 2.05
Total carbohydrates (g/L)	3.6 ± 0.1	6.7 ± 0.1	4.6 ± 0.1	17.7 ± 3.4	7.6 ± 0.95	2.7 ± 0.1
TOC (%)	38.7 ± 0.1	52.0 ± 0.1	45.3 ± 0.6	49.9 ± 0.2	50.6 ± 0.0	46.6 ± 1.1
C/N	10.6	18.9	-	23.7	31.7	9.5

summarizes the mean composition of raw SS, FW_dil, GR, CW, LPM, and OMW. Thermal pretreatment was applied to the solid wastes (PM and GR) due to their high solids content and complex organic structures, which require more intensive treatment to enhance hydrolysis and improve biodegradability. In contrast, FW contains a higher moisture content and lower structural complexity, making it more suitable for mechanical pretreatment. This approach aimed to optimize substrate accessibility for microbial activity based on the specific physical characteristics of each waste type. A mixture design within Microsoft Excel was formulated, considering the physicochemical characteristics of distinct waste substrates. This approach aimed to attain a well-rounded nutrient profile and optimal substrate proportions conducive to dark fermentation processes.

The characteristics of the mixed substrate used as feedstock are summarized in

Table 2. Characteristics of experimental materials as feedstock.

Parameters	M1	M2	M3	M4	M5	M6	M7	M8
pH	5.3 ± 0.0	5.9 ± 0.0	5.6 ± 0.0	5.5 ± 0.0	5.3 ± 0.0	5.7 ± 0.0	5.0 ± 0.0	5.4 ± 0.0
TS (g/L)	44.9 ± 0.3	36.6 ± 0.9	42.7 ± 0.6	41.9 ± 0.7	37.9 ± 0.6	31.0 ± 0.4	32.0 ± 0.6	36.0 ± 0.4
VS (g/L)	37.8 ± 0.6	30.4 ± 0.2	35.9 ± 0.4	35.0 ± 0.5	30.3 ± 0.6	24.4 ± 0.8	25.8 ± 0.4	30.3 ± 0.7
sCOD (g/L)	20.3± 4.1	10.8 ± 0.5	18.1 ± 1.2	18.2 ± 3.1	21.9 ± 0.3	16.1 ± 3.2	16.5 ± 2.1	12.2 ± 1.1
TN (g/L)	2.1 ± 0.0	2.0 ± 0	1.6± 0.0	1.3 ± 0.0	2.0 ± 0.0	1.8 ± 0.1	1.8 ± 0.1	1.3 ± 0.1
Proteins (%)	1.3 ± 0.0	1.3 ± 0.0	1.0 ± 0.0	0.8 ± 0.0	1.3 ± 0.0	1.1 ± 0.0	1.1 ± 0.0	0.8 ± 0.0
Soluble carbohydrates (g/L)	5.4 ± 0.5	6.8 ± 0.4	6.4 ± 0.4	6.7 ± 0.4	3.9 ± 0.2	6.7 ± 0.2	5.0 ± 0.5	4.8 ± 0.6

M1: 55% OMW & 40% CW &5% FW_dil, M2: 70% OMW & 30% SS, M3: 55% OMW & 40% CW & 5% SS, M4: 55% OMW & 40% CW & 5% LPM, M5: 80% CW & 20% LPM, M6: 60% CW & 40% GR, M7: 60% CW & 40% FW_dil, M8: 60% OMW & 40% FW_dil.

.

2.2. Inoculum

The inoculum for this investigation was anaerobic sludge sourced from a municipal sewage treatment plant in Crete, Greece. The anaerobic sludge had a VS content of roughly 13.8 g/L and the seed biomass exhibited a nearly neutral pH of about 7.0. In order to deactivate methanogenic bacteria and promote biohydrogen production without methane formation, the sludge underwent thermal pretreatment, which involved heating and stirring it at 100 °C for 20 min.

2.3. Experimental Procedures

The BHP of the substrates was assessed under mesophilic conditions (37 ± 1 °C) using laboratory-scale fermentation procedures. 120 mL glass bottles that were hermetically sealed with nylon tops were used for batch testing. The schematic diagram of the batch BHP reactor is presented in Figure 1. Eighty milliliters was the working volume used in the bottles; the remaining forty milliliters was headspace volume. Thirty to thirty five milliliters of substrate and forty five to fifty milliliters of inoculum were placed into the bottles depending on the mixture. Eight distinct substrate mixtures (M1–M8) were designed to investigate their effectiveness in biohydrogen production via DF, using various combinations of agro-industrial wastes. Mixture M1 comprised 55% OMW, 40% CW and 5% diluted FW_dil, M2 included 70% OMW and 30% SS. M3 consisted of 55% OMW, 40% CW, and 5% SS, while M4 was composed of 55% OMW, 40% CW, and 5% LPM. In M5, 80% CW was combined with 20% LPM. M6 mixed 60% CW with 40% GR, and M7 blended 60% CW with 40% FW_dil. Finally, M8 contained 60% OMW and 40% FW_dil. These mixtures were formulated to explore the synergistic effects of co-substrates with varying organic compositions and nutrient contents on hydrogen production efficiency. On a VS basis, the substrate/inoculum (S/I) ratio was established at 4.7 to inherently regulate pH to values below 6.0 eliminating the need for chemical adjustment. After two minutes of N₂ gas purging to remove air, each container was placed in an orbital shaker set to 150 rpm for 250 h. The hydrogen generation of every flask achieved a peak after 140 h of the experiment.

2.4. Analytical Methods

The pH was measured using a pH-meter (model GLP21, Crison, Alella, Spain), while BOD, total nitrogen (TN), total phosphorus (TP), and total (TCOD) and soluble chemical oxygen demand (sCOD) were determined using standard test kits (Hach, Düsseldorf, Germany) and spectrophotometric methods, all in accordance with APHA (2005) [18]. The crude protein (CP) content was calculated by multiplying total nitrogen by 6.25, as most proteins contain 16% nitrogen [19]. L-tryptophan, sulfuric, and boric acids were added to assess the amount of total and soluble carbohydrates [20]. This resulted in a colored sugar derivative that was then measured colorimetrically at 520 nm. Acetic, propionic, iso-butyric, butyric, iso-valeric, and valeric acid VFA concentrations were assessed. After collecting gas samples in gas-tight syringes, the needle was sealed with a butyl rubber stopper before being transferred to the gas chromatograph. To analyze hydrogen, twenty microliters were injected into an Agilent 6890 N GC System gas chromatograph (Agilent Technologies, Inc., Santa Clara, CA, USA). A capillary column (Agilent, Santa Clara, CA, USA, GS Carbonplot, 30 m × 0.32 mm, 3 μm) and a thermal conductivity detector were employed. The detector port was maintained at 150 °C, and the column operated isothermally at 80 °C. Helium was employed as a carrier gas at a flow rate of 15 mL/min. The test was continued until the rate of hydrogen production declined. All experiments were performed in triplicate to ensure reproducibility and statistical reliability of the results Using Origin 9.0, the study’s data and findings were statistically analyzed (average values, variance, and standards deviation were examined) (OriginLab, Northampton, MA, USA).

3. Results and Discussion

3.1. Biohydrogen Production

The type and concentration of the substrate, the pretreatment techniques, the source of the inoculum, the fermentation temperature, the initial pH, and other variables all have an impact on the production of hydrogen through fermentation. Figure 2 illustrates the variation of cumulative biogas production over time for fermentations with different substrate mixtures. Data points represent the average of the triplicate flasks with deviations from the mean was less than 10%. After 220 h, cumulative biogas production was recorded as 219, 217, 248, 245, 103, 171, 66 and 77 mL at the end of 220 h for mixtures M1 through M8, respectively. At a fixed substrate-to-inoculum (S/I) ratio, the cumulative biogas production was highest for M3 followed by M4, M1 and M6. The lowest biogas yield was obtained for mixture M2. Additionally, the amount of hydrogen and biogas generated at STP conditions is shown in Figure 3 as a function of various substrates. Hydrogen content ranged from 25% (M2) to 36% (M4). In all tested substrates, the mixed liquor contained ethanol and VFAs, while the biogas primarily consisted of carbon dioxide and hydrogen.

The highest hydrogen yields (38 mL/gVS) were obtained from substrates M3 and M4 with corresponding pH values of 5.6 and 5.5 respectively. In contrast, M2 resulted in the lowest yield (5 mL/gVS) at a pH of 5.9. Methane was not detected in any of the fermentation tests, indicating that only the acidogenic phase was active. A mass balance for M3 was conducted to evaluate the anaerobic digestion performance in terms of VS reduction and biogas yield. The total VS input was 35.9 g/L, while the remaining VS in the digestate was 25.9 g/L, resulting in a VS removal efficiency of 27.6% (Figure 4). This indicates effective biodegradation of organic matter during the fermentation process. Correspondingly, the specific biogas yield achieved was 38 mL/gVS, reflecting the volume of biogas produced per gram of VS added These results demonstrate that a substantial portion of the organic material was converted into biogas, supporting the effectiveness of the co-digestion strategy applied in this study. Figure 3 presents the biogas yield of the substrates used.

The cumulative hydrogen production profiles were successfully fitted to the modified Gompertz model (R² = 0.875–0.995) (

Table 3. Biogas and Biohydrogen production, biogas composition and yield, TVFAs, sCOD, pH_in-out, VS, sCarbohydrate removal and the modified Gompertz model for biohydrogen production kinetics. for different co-digestions.

Parameters		M1	M2	M3	M4	M5	M6	M7	M8
Total Biogas production (NmL)		219.0	42.0	248.0	245.0	103.0	171.0	66.0	77.0
Biogas composition (%)–H2		32.0	25.0	35.0	36.0	29.0	32.0	26.0	30.0
Biohydrogen production (NmL)		70.1	10.5	86.8	88.2	29.9	54.7	17.2	23.1
Biogas yield (NmL/gVS)		93.0	21.0	108.0	107.0	45.0	85.0	29.0	35.0
H2 yield (NmL/gVS)		30.0	5.0	38.0	38.0	13.0	27.0	8.0	10.0
sCOD removal (%)		6.4	4.5	9.6	12.3	11.2	6.8	3.9	4.6
VS removal (%)		23.8	16.2	27.6	20.7	17.2	24.0	1.9	13.6
sCarbohydrate removal (%)		22.1	22.7	45.0	38.9	1.3	23.2	36.7	1.0
TVFAs (g/L)		3.9	3.2	5.1	6.1	2.3	2.7	1.2	4.9
pHin		5.3	5.9	5.6	5.5	5.3	5.7	5.0	5.4
pHout		5.4	5.4	5.4	5.5	4.5	4.9	4.9	5.2
Modified Gompertz model	Hmax (NmL H2)	70.1	10.5	86.8	88.2	29.9	54.7	17.2	23.1
	Rmax (NmL H2/h)	0.36	0.11	1.15	0.57	0.6	0.45	0.35	0.34
	λ (hours)	3.2	3.2	3.2	3.2	4.2	4.2	4.2	3.2
	R2	0.968	0.875	0.995	0.952	0.921	0.879	0.947	0.973

M1: 55% OMW & 40% CW &5% FW_dil, M2: 70% OMW & 30% SS, M3: 55% OMW & 40% CW & 5% SS, M4: 55% OMW & 40% CW & 5% LPM, M5: 80% CW & 20% LPM, M6: 60% CW & 40% GR, M7: 60% CW & 40% FW_dil, M8: 60% OMW & 40% FW_dil.

), confirming that this model accurately described the experimental data. According to Guo and Wang, 2024 [21] these values affirm this mode’s efficiency in accurately describing the kinetics of the process. Among the tested mixtures, the M3 and M4 mixture exhibited the highest maximum hydrogen potential (H_max = 86.8 NmL H₂ and 88.2, respectively) and the highest hydrogen production rate (R_max = 1.15 NmL H₂/h and 0.57, respectively), along with a short lag phase (λ = 3.2 h). This behavior indicates improved substrate biodegradability and rapid microbial adaptation when co-fermenting multiple waste types. In contrast, M5, M6, and M7 exhibited prolonged lag phases and reduced hydrogen yields, suggesting that inadequate nutrient balance and buffering capacity constrained microbial activity.

According to published research [12], acetic acid is considered the optimal fermentation end-product in order to obtain the highest possible hydrogen output. However, in our study, the production of both acetic and butyric acid in M1, M3, and M4 appeared to be directly correlated with increased hydrogen productivity (Figure 5). This observation is consistent with the findings of Dareioti and collaborators [22] who also reported that butyric acid production seems to be tightly linked to hydrogen formation. Furthermore, Ghimire et al. [23], reported a maximum H₂ production for OMW under mesophilic batch tests of 33.8 mL/gVS, whereas in our work, the corresponding yields were calculated to be significantly lower (5 and 10 mL/gVS) for a mixture of OMW (70%) and SS (30%) and OMW (60%) and FW_dil (40%), respectively. The decreased hydrogen yield may be attributed to the presence of phenolic compounds in OMW which are known to inhibit microbial activity. Conversely, when the OMW content was adjusted to 55%, hydrogen yields increased substantially, ranging from 30 to 38 mL/gVS.

The pH values remained relatively stable (4.5–5.5), Table 3, due to the inherent buffering capacity of the substrate mixtures, confirming that the co-fermentation strategy provided natural pH control without external adjustment. Numerous studies in literature have investigated the influence of pH on hydrogen production using various substrates, including those examined in this study. These findings largely corroborate the results of the current study, indicating that an optimal pH for maximizing hydrogen yield is around 5.5. For example, Hernández and Rodríguez [24] reported a maximum hydrogen content of 26.9% and a production rate of 31.8 NmL H₂/h at an optimal pH of 5.5, supporting the findings of the present study.

According to Elbeshbishy et al. [25], pH is one of the most crucial factors in dark fermentation, as it significantly influences the formation of end-products. Maintaining pH within an optimal range is essential to enhance hydrogen production and prevent process inhibition. Additionally, since various inhibitors in DF such as nutrient imbalances, high organic loading, or toxic compounds have been identified, co-fermentation of multiple feedstocks offers a practical approach to mitigate or even eliminate these inhibitory effects. Moreover, operating at pH 5.5 reduces the need for alkaline solutions for pH adjustment, which in turn lowers the overall operational costs. This study introduces a strategic waste blending approach that significantly enhances hydrogen yields while inherently stabilizing pH around 5.5, eliminating the need for external chemical pH control [26,27].

3.2. Organic Removal

After 220 h of anaerobic fermentation, the carbohydrate removal efficiencies of the various substrate mixtures ranged from 1% to 45% (Figure 4). The carbohydrate breakdown efficiency might have been affected by the intrinsic complexity and heterogeneity of the substrate compositions. These findings are also consistent with the synthesis of organic acids, which increased in quantity at substrates M3 and M4 (Figure 5). In anaerobic digesters, carbohydrates are known to yield more hydrogen than other carbonaceous substances (such as lipids or protein-rich materials) [28]. The elevated content of butyric acid observed in these mixtures are indicative of the breakdown of carbohydrates (Figure 4).

All samples exhibited sCOD removal efficiencies below 15%, which is typical for the acidogenesis stage of anaerobic digestion (Figure 4), where complex organic molecules are primarily converted into soluble metabolic products (alcohols and VFAs) [28]. As a result, the minimal COD abatement observed hydrogen production was anticipated and consistent with the findings of earlier research. Chu and collaborators [29] reported a 9.3% COD removal during the acidogenic fermentation of food residues under thermophilic conditions. Zhou and collaborators [30] found that when FW and sewage sludge were co-digested under mesophilic conditions, the overall COD decreased by 6 to 12%. An additional methanogenic phase may be used to further minimize residual organic load and optimize the production of biogas in order to utilize the energy potential of the non-converted organic substrate [28].

The overall COD recovery ranged between 99–110% (

Table 4. COD balances, % of input COD for VFAs, Ethanol, H₂, biomass and Residual COD.

Parameters	M1	M2	M3	M4	M5	M6	M7	M8
Net VFAs	8.5	7.3	10.2	11.7	2.1	2.9	3.0	11.1
Net Ethanol	3.7	2.9	3.7	4.2	5.2	7.8	2.0	0.2
H2	0.2	0.1	0.3	0.3	0.1	0.2	0.1	0.1
Biomass	2.0	5.0	2.0	3.0	3.0	4.0	4.0	5.0
Residual COD	93.6	95.5	90.4	87.7	88.8	93.2	96.1	95.4
Total	108.0	110.8	106.7	107.0	99.3	108.1	105.1	111.8

M1: 55% OMW & 40% CW &5% FW_dil, M2: 70% OMW & 30% SS, M3: 55% OMW & 40% CW & 5% SS, M4: 55% OMW & 40% CW & 5% LPM, M5: 80% CW & 20% LPM, M6: 60% CW & 40% GR, M7: 60% CW & 40% FW_dil, M8: 60% OMW & 40% FW_dil.

), confirming that the analytical determinations of VFAs, ethanol, and hydrogen accounted for the main carbon sinks in the system. COD recoveries slightly above 100% may result from cumulative analytical errors or overestimation of certain metabolites. These results indicate that the COD balance was satisfactorily closed, validating the reliability of the experimental measurements.

3.3. VFA and Ethanol Profiles

Since hydrogen gas is a byproduct of VFA biosynthesis, its production during DF is proportional to VFAs formation. However, there is also a close correlation between the type of VFAs produced and the amount of biohydrogen produced. Depending on the microbial species and environmental factors, the breakdown of glucose in anaerobic environments results in the production of hydrogen and a number of metabolic products, primarily VFAs (acetic, propionic, and butyric acids), lactic acid, and alcohols (ethanol). Therefore, the analysis of metabolic products can offer valuable insights to interpret the observed hydrogen yields of generation and offers valuable insights into the evolution of the process.

In the current investigation, soluble metabolites at the end of the fermentation process were analyzed. Ethanol, butyric, propionic, acetic, and isobutyric acids were detected across all samples. The highest total VFAs concentration was detected at substrate M4 (6.1 g/L) primarily due to an elevated concentration of butyric acid (Figure 5), which was observed to increase when replacing SS with LPM. The correlation analysis between individual VFAs and hydrogen yield demonstrated distinct relationships among the fermentation intermediates. Among the VFAs analyzed, isobutyrate and butyrate exhibited the highest coefficients of determination with H₂ yield (r² ≈ 0.61), indicating a strong association between their accumulation and enhanced hydrogen production. Propionate displayed a moderate correlation (r² ≈ 0.35), while acetate and ethanol showed negligible linear relationships with H₂ yield (r² < 0.05). These findings suggest that metabolic pathways favoring butyrate and isobutyrate formation are more conducive to hydrogenogenesis, whereas acetate- and ethanol-type fermentations are less efficient in promoting hydrogen evolution under experimental conditions. The concentrations of acetic, propionic and isobutyric acid were practically unchanged between substrates. One of the more efficient methods for producing H₂ is butyric acid-type fermentation, particularly when facilitated by spore-forming Clostridium species. In fact, butyric acid is frequently seen to dominate the VFA profile in the anaerobic digestion of a variety of substrate types [28].

It was previously confirmed that the accumulation of butyric acid is positively correlated with biohydrogen production [31]. In contrast, several other studies have reported that biohydrogen production during dark fermentation is often associated with elevated acetic acid levels [13,32]. These differences are primarily due to the varying metabolic pathways of hydrogen-producing microorganisms, which result in different patterns of VFA accumulation [31]. In this study, biohydrogen production was enhanced in the test group using 55% OMW, 40% CW and 5% SS or LPM, primarily driven by the generation of butyric acid. pH is a key factor influencing metabolic pathways: under acidic conditions (<5.5), microorganisms tend to shift toward solventogenesis, producing compounds like ethanol and butanol. In contrast, maintaining a neutral to mildly acidic pH range (5.5–6.5) promotes the formation of acetate and butyrate, thereby maximizing hydrogen production [33].

During DF, a range of intermediates and by-products are produced alongside the target product, biohydrogen. Typically, the energy conversion ratio to hydrogen is below 40% with the remaining 60% primarily diverted toward microbial growth and the production of VFAs. Downstream procedures are advised to recover the residual energy and reduce waste. On the other hand, an essential step in the fermentation of biomass to produce hydrogen is pretreatment as it may have an impact on the organic matter composition in the feedstock in addition to the subsequent capacity of the biomass to decompose [6]

4. Conclusions

In this study, the synergic and antagonistic effects of OMW, FW, LPM, CW, GR and SS on hydrogen production via DF were evaluated using a mixture design approach. The experimental findings highlighted the importance of the acetic-butyric metabolic route for hydrogen production, with OMW, CW, SS or LPM exhibiting the strongest synergistic effect. Particularly, using mixture ratio of OMW:CW:SS or LPM = 55:40:5, promoted favorable fermentative pathways dominated by acetic and butyric acid formation. The ideal hydrogen yield of 38 mL/g VS was obtained at these combinations and was accompanied by VS reduction of 27.6% and pH levels (pH 5.5) favorable for microbial hydrogenogenesis.

The study highlights several important conclusions:

• Synergistic substrate interactions: Combining carbohydrate-rich CW with nutrient-balanced SS or LPM, along with moderate amounts of OMW, enhanced metabolic pathways favorable for hydrogen production while reducing the inhibitory impact of phenolic compounds present in OMW.
• Operational stability: The process of co-fermentation naturally regulated pH levels to support dark fermentation, stimulating microbial function and carbohydrate transformation without requiring chemical interventions.
• Process efficiency and feasibility: The strategy successfully combined efficient hydrogen generation and significant VFA accumulation, showcasing the feasibility of merging bioenergy production with waste utilization.

The findings provide a strong foundation for the advancement of co-digestion systems and the incorporation of additional processes such as VFAs recovery or biohydrogen enrichment to maximize energy efficiency. Future studies should focus on sustained operational performance, scale-up, integration with methanogenic stages to emphasize the potential applicability of the process in two-stage anaerobic systems, techno-economic assessments, and the analysis of microbial community dynamics to improve hydrogen generation and system stability. Ultimately, this study validates the efficacy of intentional co-digestion strategies in improving sustainable waste management practices and promoting the development of renewable energy sources.