In situ buffering and enzymatic regulation of acidogenesis in two-stage anaerobic digestion of composite vegetable waste Authors: Debkumar Chakraborty, Abhishek Pitta, Makarand Ghangrekar, Gorakhanath Jadhav, Anil Dhanda, Paula Sfirloaga, Narcis Duteanu Published: 2026 ABSTRACT Peels of vegetable waste (VW), rich in cellulose, lignin, and hemicellulose, can impede hydrolysis due to reduced enzymatic accessibility and activity. This research comparatively evaluated the effects of composite vegetable waste (CVW) comprising of 2:3 food waste (FW):VW on biological buffering potential and enzymatic activity during acidogenesis in a two-stage (Leached Bed Reactor-airlift reactor) system. Results demonstrated that maximum buffering and enzymatic activity occurred within the first 3–5 days. The system maintained an in situ maximum buffering capacity of 0.145 mol/L (FW) and 0.089 mol/L (CVW) without chemical additives. Peak α-amylase activities reached 72.2 U/mL (FW) and 64 U/mL (CVW), while protease reached 34.94 U/mL (CVW). The ratio of 2:3 for FW:VW moderated enzymatic hydrolysis rates through its lignocellulosic content and stabilized the acidogenic environment, which in turn regulated VFA dynamics, resulting in a 1.8-fold increase in acetate yield (18.19 g/L). Conversely, FW achieved a higher biohydrogen yield of 121.3 mL/g VS (43% composition) compared to 110 mL/g VS (36.4% composition) for CVW. The FW also produced 4.2 times higher propionic acid (95.3 g/L) than CVW (22.64 g/L). Life Cycle Assessment (LCA) confirmed the sustainability of this minimized-discharge model, with H2 production from FW showing a lower Global Warming Potential (0.056 kg CO2 eq/L) than CVW (0.269 kg CO2 eq/L) due to higher gas yields. The system maintained in situ buffering without chemical additives, thus advancing the development of cost-effective routes for bioresource recovery from complex organic waste. It has also been emphasized that the sustainability of process is bolstered by the recirculation of airlift reactor effluent, which mitigates the need for chemical additives.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-026-57332-5. Introduction While fruits and vegetables, plus roots and tubers, are among the highest of food waste (FW), around 21 million tonnes (45%) of the world’s fruit and vegetables are lost each year1,2. For example, in India, 50 MT of vegetable waste (VW) is generated along the food supply chain, accounting for 30% of total production and 40.15% of municipal solid waste (MSW) generation3–6. Such vegetable-based composite solid waste can be used as a substrate to generate hydrogen (H2) and short-chain fatty acids via acidogenesis, due to the high organic content7,8. However, along with fermentable sugar fraction, the vegetable-based waste also contains complex polymers, such as cellulose, hemicellulose, and lignin, which hinder microbial degradation9. Hydrolysis of these complex polymers into simple polymers and monomers is the rate-limiting step in the anaerobic digestion (AD) process. Regulation of hydrolysis could control the release of sugars, which in turn shapes the production dynamics of volatile fatty acids (VFAs) and H2. Hence, in the past, researchers focused on physiochemical (acid and heat treatment) and enzyme treatment processes that remove the lignin barrier and recover a stream of sugars from VW, but these processes are complex and expensive10–12. An alternate, eco-friendly, and cost-effective approach would be using a hydrolytic-enzyme-enriched inoculum for the acidogenesis of VW, which circumvents the requirement of pre-treatment. Unlike lignocellulosic vegetable waste, FW degradation is faster, and regulation of the pH to balance the proton gradient due to accumulated VFA is critical to improve acidogenesis. Regulation of the pH by adding bases (sodium hydroxide) or buffering agents (phosphate and bicarbonate buffers) during leachate recirculation is a common practice in leached-bed reactor (LBR) operation13,14. However, excessive use of chemical buffering agents can increase costs, plasmolyze cells, and interfere with the enzymatic activities of acidogens and methanogens, thereby hindering the production of VFA, H2, and methane15,16. To overcome the limitations of chemical additives, researchers have explored alternative biological buffering strategies, such as co-digestion with complex substrates17,18 and addition of salt16 during the anaerobic fermentation process. Nevertheless, a significant research gap remains in harnessing lignocellulosic vegetable waste as a functional substrate for intrinsic buffering. Furthermore, the recirculation of alkaline effluents from downstream methanogenic stage to biologically regulate the acidogenic environment and enzymatic activity remains insufficiently investigated, despite its economic feasibility and compatibility with microbial systems. In our earlier investigation, it was concluded that co-digestion of complex vegetable waste with easily degradable FW regulated pH, VFA dynamics, and methane recovery during the two-stage AD (LBR-Airlift) process17,18. However, a critical research gap remains regarding the mechanistic relationship between in situ biological buffering and hydrolytic enzyme functionality. Elucidating this link is critical for advancing from empirical observations to a mechanistic understanding of microbial homeostasis during acidogenesis. The acidogenic environment does not just influence survival; it dictates the transmembrane pH gradient and the thermodynamic feasibility of metabolic pathways, which in turn, regulate the secretion and catalytic efficiency of extracellular enzymes. Understanding this synergy is essential for defining how biological systems can self-regulate during acidogenesis. In this context, the present study correlates specific buffering regimes (pKa zones) with the activities of key hydrolytic enzymes α-amylase, β-galactosidase, and protease to develop a mechanistic framework describing lignocellulosic vegetable waste as a biological regulator of the acidogenic environment. This approach aims to optimize VFA and H2 production without external chemical inputs. Additionally, life cycle assessment (LCA) is employed to evaluate the environmental sustainability of the proposed system by considering the entire process chain from resource input to end-product generation and return to the environment. Specifically, this investigation aims to: (i) evaluate the buffering potential of composite vegetable waste (CVW) relative to FW; (ii) investigate the influence of in situ buffering on hydrolytic enzyme activities (α-amylase, β-galactosidase, and protease); and (iii) assess the environmental performance of the self-regulating system through LCA. Materials and methods Substrate collection and preparation of composite vegetable waste Vegetable waste and FW were collected from the institute’s canteen, and after separation of the individual components, the proportions of each component were measured. During sample collection, care was taken to avoid the inclusion of unwanted materials. The vegetable waste was prepared using different kinds of vegetables viz., bottle gourd (21%), brinjal (20%), drum stick (19%), carrot (15%), coriander stem and leaves (11%), tomatoes (10%), muskmelon (4%), and cabbage (1%) to maintain uniformity with previous optimized investigation18. The VW was chopped by a kitchen knife into particles of about 2 mm in size. The FW, which was dominated by boiled rice, sambar, egg yolk, etc., was also collected from the institute’s canteen, and both the wastes were stored at − 4 °C overnight. The choice of a 2:3 (FW:VW) ratio for the CVW was a strategic decision based on prior optimization studies17,18, which found this specific blend highly effective at regulating pH, VFA dynamics, and methane recovery in two-stage systems. This ratio was selected to balance the disparate degradation kinetics of the two streams: the rapidly fermentable FW provides immediate substrate for acidogenesis, while the slower-degrading lignocellulosic VW (containing 20% lignin and 17% hemicellulose) provides long-term stability and acts as a biological buffer. This synergy prevents the rapid acidification typical of 100% FW systems, thereby maintaining a acidogenic environment (pH ~ 4.5) conducive to sustained enzymatic activity without the need for external chemical buffers. The physicochemical characteristics of the substrates were determined prior to reactor loading. The FW exhibited total solid (TS) and volatile solid (VS) contents of 19.46 ± 0.66% and 96.83 ± 0.54%, respectively, whereas VW showed TS and VS values of 20.00 ± 1.51% and 94.00 ± 0.46%, respectively. Based on these measurements, the CVW system (2:3 FW:VW) was prepared by mixing 440 g of FW and 660 g of VW, resulting in a total substrate loading of 1100 g and an overall VS content of approximately 207 g. For the mono-substrate FW system, 1100 g of FW was loaded, corresponding to a total VS content of 207.3 g, thereby ensuring a uniform initial VS loading between both experimental systems. In addition, 100 mL of enriched inoculum (TS: 9.8 ± 1.63%; VS: 92.93 ± 1.73%) was added to each reactor to initiate fermentation (See supplementary information). Two-stage leached bed reactor-airlift reactor design The two-stage high-solid systems, comprising a Leached Bed Reactor (LBR) for hydrolysis and acidogenesis and an air-lift reactor for methanogenesis, were coupled (See supplementary information) as described by Chakraborty and Venkata Mohan17. Two LBRs with a total volume of 1.8 L and a working volume of 1.2 L were used, with a nylon mesh (pore diameter < 1 mm) inside. Two silicone tubes were connected to LBR. Bottom one was used to collect leachate while the top tube was used to collect acidogenic gas into tedlar gas sampling bags for analysis. Enriched, activated acidogenic leachate from the previous experiment was used as the seed inoculum17. One LBR was fed with CVW (2:3 FW:VW ratio), and the other was fed FW alone. Two cylindrical airlift reactors with a total volume of 2.5 L and a working volume of 1.8 L were used in the experiment. The enriched effluent from air-lift reactors used in the previous experiment was used as the seed inoculum. Finally, these two coupled reactors (LBR-Airlift) were used for the batch experiment17. One of the airlift reactors was fed with the leachate from the LBR fed with CVW, and the other with leachate from FW to LBR. Reactor operation The LBR-Airlift systems were operated in duplicate (n = 2) to ensure data reproducibility. After thorough mixing at room temperature, the CVW and FW were fed to the CVW-LBRs (test) and FW-LBRs (control), respectively, and the LBRs were operated for 21 days. Finally, 1100 mL of tap water was added to each LBR to maintain a 1:1 solid-to-water ratio and enhance hydrolysis. From the first day, leachate from the LBRs was collected every other day for a total of 21 days. To maintain a consistent hydraulic regime and prevent washout, a precise recirculation strategy was employed, with 50% of the leachate collected from LBRs transferred to the airlift reactors via a peristaltic pump. The resulting effluent from the airlift reactor was then blended with the remaining 50% of the LBR leachate. This mixture was standardized to a final volume of 1 L with tap water to compensate for any moisture loss, then recirculated back to the top of the LBR. This 24-h recirculation cycle enhanced convective mass transfer and mixing, thereby promoting uniform distribution of enzymes and microbial consortia across the substrate matrix, compensating for the lack of internal agitation in the high-solid LBR bed. The same operational strategy was also followed in the control LBR (fed with FW). Analytical methods Physicochemical analysis The food and vegetable wastes were made into a slurry, and key parameters, including TS, VS, carbohydrate, and protein content, were analyzed at the start (1st day) and end (21st day) of acidogenesis for the fresh CVW and FW samples. The leachate volume, pH, buffering capacity, hydrolytic enzyme (α-amylase, β-galactosidase, protease) activities, ammonium nitrogen (NH4+-N), and distribution of reducing sugars, VFA, organic acids, and alcohols of LBR leachate samples were analyzed on alternate days starting from the first day till end of the experiment, according to the Standard Methods19. The composition of acidogenic gas generated from LBR was measured using gas chromatography (GC; NUCON 5765) coupled with a thermal conductivity detector, with a 1/8″ × 2 m Heysep Q column employing nitrogen as carrier gas, and the total gas volume in a Tedlar bag was measured using a water displacement facility. The injector and detector temperatures were both maintained at 60 °C, while the oven was operated isothermally at 40 °C17. Total soluble products (TSP) and their compositional profiles including alcohols, VFA, and lactic acid, as well as sugars (glucose, ribose, arabinose, fructose, xylose, galactose, mannose, lactose, and maltose) were quantified using high-performance liquid chromatography (HPLC; Shimadzu Corporation LC-20A) equipped with a refractive index detector (RID-20A) and a C18 reverse-phase column (250 × 4.6 mm, 5 µm particle size). Filtered leachate samples (20 µL) were injected and separated under isocratic conditions using water as the mobile phase at a flow rate of 0.7 mL min−1. The column temperature was maintained at 80 °C in a column oven (RA-20A)20. Standard graphs were plotted for each VFA, lactic acid, alcohol, and sugar to measure their concentrations. Buffering capacity estimation Buffering capacity (β, mol/L) was estimated as the moles of acid or base required to change the pH of 1 L of solution by 1 unit, using the Eq. 1.1\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\frac{{db}}{{dpH}} = \frac{{\left( {{\mathrm{Volume}}\,{\mathrm{of}}\,{\mathrm{acid}}\,{\mathrm{and}}\,{\mathrm{base}}\,{\mathrm{added}}} \right)*({\mathrm{Normality}}\,{\mathrm{of}}\,{\mathrm{acid}}\,{\mathrm{and}}\,{\mathrm{base}})}}{{\left( {Volume\,of\,sample} \right)*(pH\,change\,produced)}}$$\end{document} Samples were prepared by centrifuging crude leachate at 6000 rpm for 10 min, with 3 mL of supernatant diluted to 40 mL using tap water. Titration was performed against 0.1 N HCl or 0.1 N NaOH using an auto-titrator21. Determination of biochemical rate constant and enzyme activity Extracellular hydrolytic enzymes secreted by anaerobic microflora play a key role during acidogenesis22. The solubilization of organic waste was described using a first-order kinetic model to estimate rate constants (k, expressed in d−1) for hydrolysis and the subsequent conversion into intermediate products during acidogenesis and acetogenesis. This model assumes that the rate of substrate degradation is directly proportional to the concentration of available organic matter, governed by the collective activity of extracellular hydrolytic enzymes (α-amylase, β-galactosidase, and protease). The hydrolysis, acidogenesis (mixed-acid production), and acetate production rates were determined based on the ratio differences between tCOD and sCOD, sCOD and total VFAs, and total VFAs and acetate fraction, respectively, using previous methodology17. The rate constant for each phase was estimated based on the cumulative increase in Total Soluble Products (TSP), specifically VFA and reducing sugars, following the methodology described previously17,23. The rate constants for hydrolysis, mixed-acid production, and acetate formation were determined assuming first-order kinetics for substrate biodegradation (dS/dt = kS), where S represents VS concentration, p denotes product concentration, and k is the first-order rate constant. Under the assumption that total product formation is proportional to substrate utilization (dS≈dP), the rate constants were estimated using Eq. (2)2\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\frac{dp}{dt}=rate\; constant*total\; substrate\; concentration$$\end{document}where, dP corresponds to the net product formation for each process, estimated as follows: hydrolysis was represented by the increase in soluble chemical oxygen demand (sCOD) (cumulative sCOD—initial sCOD), mixed volatile acid production by the accumulation of total VFA (cumulative VFA—initial VFA), and acetate production by the net acetate concentration (cumulative acetate-initial acetate). It should be noted that potential sCOD losses in the form of gaseous products were not accounted for in this analysis. The specific hydrolytic enzyme activities, e.g., α-amylase, β-galactosidase, and proteases, were measured by spectroscopic analysis on alternate days using the following detailed procedures. Filtered leachate samples were used as the crude enzyme source to measure enzyme activity. To ensure that measured signals were strictly enzymatic and not influenced by matrix pigments or reducing sugars, each assay included a negative control consisting of thermally inactivated (boiled) leachate, which served as the background blank for subtraction. Positive controls using commercially available purified enzymes were utilized to calibrate the spectroscopic readings and verify the activity of the reagents as described by Chakraborty and Venkata Mohan17; Chakraborty et al.23. α-amylase activity The α-amylase activity was determined by measuring the starch to maltose formation rate (Eq. 3). Reducing activity of maltose converts an alkaline solution of 3,5-dinitrosalicylic acid into orange colour, and 3-amino-5-nitrosalicylic acid was measured at OD540. Enzyme activity was determined using Eq. 4, where one enzyme unit is the concentration of enzyme required to release 1 mg of α-maltose per minute from starch at 37 °C and pH of 7.0. In this equation, C represents the concentration of maltose released (mg/mL), V total is the total reaction volume (mL), Venzyme is the volume of enzyme used (mL) and t is the incubation time24.3\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathrm{S}\mathrm{t}\mathrm{a}\mathrm{r}\mathrm{c}\mathrm{h} + \mathrm{H}2\mathrm{O} + (\upalpha -\mathrm{a}\mathrm{m}\mathrm{y}\mathrm{l}\mathrm{a}\mathrm{s}\mathrm{e})=\mathrm{R}\mathrm{e}\mathrm{d}\mathrm{u}\mathrm{c}\mathrm{i}\mathrm{n}\mathrm{g}\; \mathrm{G}\mathrm{r}\mathrm{o}\mathrm{u}\mathrm{p}\mathrm{s} (\mathrm{M}\mathrm{a}\mathrm{l}\mathrm{t}\mathrm{o}\mathrm{s}\mathrm{e})$$\end{document}4\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\upalpha -\mathrm{a}\mathrm{m}\mathrm{y}\mathrm{l}\mathrm{a}\mathrm{s}\mathrm{e}\; activity \left(\frac{U}{mL/min}\right)=\frac{C*V\; total}{Venzyme*t}$$\end{document} β-galactosidase assay Β galactosidase releases o-Nitrophenol from o-Nitrophenyl ß-D-Galactopyranoside (ONPG), which can be measured at OD405 nm (Eq. 5). Enzyme activity was estimated using Eq. 6, where one unit of β galactosidase was determined, which depends on the hydrolysis of 1.0 µmole of o-Nitrophenyl ß-D-Galactoside to o-Nitrophenol and D-galactose per minute at pH 6.0 and 37 °C25. In this equation, C represents the concentration of o-nitrophenol released (µmol/mL), V total is the total reaction volume (mL), Venzyme is the volume of enzyme used (mL) and t is the incubation time.5\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mathrm{ONPG}} + \left( {\beta \;{\mathrm{galactosidase}}} \right) = {\mathrm{o}} - {\mathrm{Nitrophenol}} + {{{\ss}}} - {\mathrm{D}} - {\mathrm{Galactose}}$$\end{document}6\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\upbeta\; \mathrm{g}\mathrm{a}\mathrm{l}\mathrm{a}\mathrm{c}\mathrm{t}\mathrm{o}\mathrm{s}\mathrm{i}\mathrm{d}\mathrm{a}\mathrm{s}\mathrm{e}\; activity \left(\frac{U}{mL/min}\right)=\frac{C*V total}{Venzyme*t}$$\end{document} Protease activity In the protease activity assay, Na-p-tosyl-L-arginine methyl ester (TAME) was used as the substrate. Enzyme activity was measured by the release of Na-p-Tosyl-L-Arginine from solution, as described in Eq. (7), at 247 nm. One unit (U) of protease activity was determined as the amount of enzyme required to hydrolyze 1.0 µmol of TAME per minute at 25 °C and pH of 8.026 (Eq. 8). In this equation, ΔA₂₄₇ is the change in absorbance per minute, ε is the molar extinction coefficient, l is the path length of the cuvette (cm), Vtotal is the total reaction volume (mL), and Venzyme is the volume of enzyme used (mL).7\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathrm{T}\mathrm{A}\mathrm{M}\mathrm{E} + \mathrm{H}2\mathrm{O}=\mathrm{N}\mathrm{a}-\mathrm{p}-\mathrm{T}\mathrm{o}\mathrm{s}\mathrm{y}\mathrm{l}-\mathrm{L}-\mathrm{A}\mathrm{r}\mathrm{g}\mathrm{i}\mathrm{n}\mathrm{i}\mathrm{n}\mathrm{e}+\mathrm{M}\mathrm{e}\mathrm{t}\mathrm{h}\mathrm{a}\mathrm{n}\mathrm{o}\mathrm{l}$$\end{document}8\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathrm{P}\mathrm{r}\mathrm{o}\mathrm{t}\mathrm{e}\mathrm{a}\mathrm{s}\mathrm{e}\; \mathrm{a}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{i}\mathrm{t}\mathrm{y} \left(\frac{\mathrm{U}}{mL/min}\right)=\frac{\Delta \mathrm{A}247*V\; total}{\upvarepsilon *l*Venzyme*t}$$\end{document} Sustainability analysis and life cycle assessment The LCA for the inventory analysis of H2 production during two-stage anaerobic digestion of FW and CVW, was conducted using the SimaPro 8.5.2.0 tool following the CML-IA baseline and IPCC GWP 100a methods according to the model developed by Katakojwala and Venkata Mohan27. This model evaluates and compares the process flows between two treatments, considering system boundaries, raw material consumption, and product generation using a mixture of attributional and consequential LCA. The system boundary was defined as cradle-to-gate, encompassing substrate preparation, reactor operation, biogeochemical field emissions, and the recovery of dominant metabolites (1 L of H2, 1 kg lactic acid, 1 kg propionic acid, and 1 kg acetic acid). The functional unit was 1 L of H2 produced. The inventory assumptions and upstream impacts were measured based on experimental data: 1.1 L of process water for moisture maintenance and 350 Wh of electricity for the peristaltic pump over the 21-day period. Furthermore, the model includes biogeochemical field emissions resulting from the fermentation of the CVW and FW. Statistical analysis All experimental runs for both CVW and FW systems were conducted in duplicate (n = 2). However, to ensure analytical precision, measurements for all parameters, including pH, VFA concentrations, and enzymatic activities, were performed in triplicate for each reactor. Statistical significance of temporal variations in pH and enzyme activities was determined using One-way ANOVA to assess differences between operational phases. Results presented in text, tables, and figures represent the mean values ± the standard deviation (SD). Results and discussion The TS, carbohydrate, protein, and cellulose contents were 14.5%, 50.9%, 15.41%, 3 g/L in CVW and 19.46%, 65.35%, 9.56%, 3.4 g/L in FW. Additionally, the CVW had 12 g/L lignin and 5 g/L hemicelluloses. Regulation of acidogenic environment and buffering capacity of the system pH and leachate production profile of LBR Rapid drop in pH was observed in both experiments; for CVW-LBR the pH dropped from 6.51 to pH 3.67 by 48 h and for FW-LBR from 6.21 to 3.2 by 72 h from the start of the experiments (Fig. 1a). However, in both reactors, after 5th day the pH started rising (p < 0.05) and on the 9th day, it was above 4.5 and this upward trend continued significantly until Day 21 (p < 0.01). The pH of CVW-LBR leachate always remained above the pH of FW-LBR leachate. The difference in pH is attributed to variations in acid production during acidogenesis in both experiments, as discussed later in "VFA distribution" section.Fig. 1Acidogenic environment during LBR operation, showing (a) pH, (b) daily leachate production, (c) Buffering activity of CVW, and (d) Buffering activity of FW. Acidogenic environment during LBR operation, showing (a) pH, (b) daily leachate production, (c) Buffering activity of CVW, and (d) Buffering activity of FW. The rate of pH decrease was lower in CVW than in FW due to reduced hydrolysis. Generally, during the hydrolysis of FW, a rapid drop in pH is common, which decreases the specific growth rates of acidogenic microorganisms, as they cannot maintain homeostatic intracellular pH beyond threshold limits16,28. At low pH, the production of VFAs further acidifies the medium, allowing undissociated species to penetrate the microbial cell and elicit intracellular H+ release29. At acidic pH, VFA was highly toxic to microorganisms in their free acid form, which is more dominant at pH below the pKa30. The system pH plays a crucial role in substrate fermentation by influencing organic matter solubilization, microbial activity, and community structure31,32. The average leachate production rate of CVW-LBR and FW-LBR was 915.5 and 904 mL/day, respectively (Fig. 1b). However, the volume of leachate production varied from the first day onwards, mainly due to differences in the percolating property of the substrates. A higher hydrolysis rate in CVW-LBR influences percolation properties, leading to increased leachate volume17. In the FW-LBR, leachate production increased until the 7th day; from the 8th day to the 15th day, it decreased, but it started increasing again after the 15th day and continued until the end of the experiment. The volume of leachate production in CVW-LBR was almost constant and steadily increasing from the first day till the end of the experiment, except for a slight decrease on the 15th day. LBR has limitations of channeling effects and lack of agitation which can hinder effective contact between microorganisms and substrates, thereby reducing hydrolysis rates15. However, leachate recirculation maintains equilibrium between biological degradation processes and the hydrological capacity of the waste33. Recirculation of LBR-leachate and airlift reactor-effluent at 24 h intervals had many advantages, including better percolation of leachate through the substrate, resist sudden pH drop in LBRs by avoiding the inhibitory effects of VFAs and other acid metabolites, redistributing enzymes and microbes, minimizing nutrient deficits (nitrogen, phosphorus, and trace elements), promoting mass transfer and retaining microorganisms, which enabled rapid degradation of the substrates and stabilized the leachate production33–35. Eventually, acidogenesis and methanogenesis occur faster in leachate-recycled systems than in systems without recirculation36,37. Additionally, water is essential to AD process because it dissolves nutrients and substrates and promotes their diffusion to bacterial cells38,39. Increasing the water recirculation rate (WRR) from low to high levels (e.g., 5% to 100%) has been shown to enhance solubilization of FW, which was reflected in improved VS removal efficiencies. Although higher WRR may lead to enzyme washout, this is compensated for by increased extracellular enzyme regeneration. Furthermore, dilution at higher WRR helps to mitigate the accumulation of inhibitory compounds, thereby improving overall process performance. Buffering activity A titration curve shows the buffering capacity and pKa of the weak acid/weak base in leachate at regular intervals. The buffering capacity of leachate from CVW and FW fed LBRs varied during acid and base titrations for a given operation time (Fig. 1c & d), which was further correlated with pH, enzyme activity, and H2 production. The titration curves for both CVW and FW showed three distinct buffering zones, providing a chemical fingerprint of the acidogenic process. Buffering Zone-I (pH 2.0–4.0). Buffering Zone-I (pH 2.0–4.0). Buffering Zone-I (pH 2.0–4.0). Zone I (pH 2.0–4.0, pKa ~ 3) is directly attributed to the rapid accumulation of lactic and short-chain fatty acids; the maximum β value for FW (0.019 ± 0.001) was significantly higher than CVW, correlating with the higher VFA and lactic acid yields measured in the FW system. In both experiments, and the buffering capacity (β) increased with time.b) Buffering Zone-II (pH 6.0–9.0). b) Buffering Zone-II (pH 6.0–9.0). Buffering Zone-II (pH 6.0–9.0). Zone II (pH 6.0–9.0, pKa ~ 7.5) represents the system’s primary defensive mechanism against acidification. Both CVW and FW showed maximum buffering capacity in this buffering zone, with specific bicarbonate and acetate buffering peaks at pKa 7.8 and 8.2, respectively. The fact that CVW reached its maximum buffering capacity (0.089 ± 0.001) on the third day, coinciding with its peak protease activity (34.94 U/mL), suggests that the degradation of proteinaceous and lignocellulosic components in vegetable waste contributes to this in situ stabilization (Fig. 1c & d). Whereas, the buffering characteristics of FW initially decreased till the third day, then increased to a maximum of (0.145 ± 0.001) on the fifth day, then decreased with time as the pH increased (Fig. 1c & d).c) Buffering Zone-III (pH 9.0–11.0). c) Buffering Zone-III (pH 9.0–11.0). Buffering Zone-III (pH 9.0–11.0). Finally, the alkaline buffering in Zone III (pH 9.0–11.0, pKa ~ 10) was observed in the leachates from CVW and FW LBRs, increasing initially until the 3rd day (CVW) and the 5th day (FW), then decreasing. This zone is associated with the presence of ammonium ions (NH4+-N), which reached a maximum of 57.53 mg/L in CVW due to sustained protease activity. Cumulative buffer intensity was minimum and maximum in the buffering zones I and II, respectively, while in zone III it was moderate (Fig. 1c & d). Overall, the buffering capacity of the leachate varied markedly between the two systems. The CVW-LBR exhibited a relatively low buffering capacity (0.089 ± 0.001), indicating that a smaller amount of base was required to neutralize the acids formed. In contrast, the FW-LBR showed a higher buffering capacity (0.145 ± 0.001), resulting in rapid accumulation of VFAs at significantly higher concentrations than in the CVW system (Fig. 1c & d). This elevated buffering capacity in FW is primarily attributed to its 4.2-fold higher propionic acid production (95.3 g/L) compared to CVW. Conversely, the comparatively lower but more stable buffering observed in the CVW system can be linked to the gradual release of fermentable sugars from its lignocellulosic fraction, which mitigated abrupt acidification and contributed to a more controlled acidogenic environment. Higher concentrations of VFA and their salts resulted in higher buffering capacity16, and continuous leachate and effluent recirculation helped maintain uniform buffering strength in both cases. Notably, stable buffering and pH regulation were achieved without the addition of external chemical buffers, likely due to the dynamic equilibrium between VFAs and dissolved CO₂ during fermentation40. In this experiments, buffering regions corresponding to bicarbonate (HCO₃–) and acetate (CH₃COO–) systems were identified at pKa values of 7.8 and 8.2, respectively. A pKa near ~ 4.8 was attributed to the presence of short-chain fatty acids, whereas the pKa around ~ 10 was associated with the formation of conjugate bases, including the conversion of bicarbonate to carbonate, phenols to phenoxide, and ammonium ions to ammonia, as well as the accumulation of alkaline species during anaerobic digestion (Fig. 1c and d). The variation in buffering capacity (β) peak intensities closely correlated with fluctuations in VFA concentrations and buffering potential16, demonstrating the linkage between metabolic intermediates and system stability. Importantly, these buffering peaks coincided with the maximum enzymatic activities of α-amylase and protease observed between days 3 and 5, indicating that the inherent ‘biological buffering’ suggests a homeostatic proton-mediated environment conducive to efficient hydrolysis. By returning the methanogenic effluent to the LBR, the system leveraged biological byproducts (e.g., bicarbonate and ammonium) to counteract acid accumulation, successfully maintaining the pH within a range suitable for sustained enzymatic activity throughout the 21-day period. Hydrolytic enzyme activities in anaerobic digestion α amylase activity In the CVW-LBR system, a rapid increase in extracellular α-amylase activity was observed within the first 24 h, reaching a maximum of 64 U/mL. Subsequently, this activity decreased until the fifth day, then increased slowly toward the end of the experiment. In contrast, in the FW-LBR, significant enzyme activity was only observed after 24 h, followed by a rapid increase until the fifth day, where it peaked at 72.2 U/mL. This was followed by a sharp decrease until the 12th day, a marginal improvement on the 15th day, and a final decline during the end of the operation (Fig. 2a). The overall higher α-amylase activity in the FW-LBR is attributed to the predominance of starch-rich rice waste in the FW substrate.Fig. 2Hydrolytic enzyme (a) α-amylase, (b) β-galactosidase, and (c) protease activities during anaerobic co-digestion. Hydrolytic enzyme (a) α-amylase, (b) β-galactosidase, and (c) protease activities during anaerobic co-digestion. β-galactosidase activity This hydrolytic enzyme cleaves glycosidic bonds and releases galactose from polysaccharides (e.g., the pectin moiety in fruits and vegetables). Though the overall β galactosidase activity was lesser than α amylase and protease activities, it was higher in CVW-LBR (~ 0.025 U/ mL) due to its more polysaccharide content than in the FW-LBR (~ 0.015 U/ mL) (Fig. 2b). After the initial increment, the β galactosidase activity almost remained steady throughout the experiment in both CVW and FW LBRs. Protease activity The peak protease activity occurred significantly earlier in CVW (Day 3; 34.94 U/mL) compared to FW (Day 7; 23.17 U/mL) (Fig. 2c). This discrepancy is attributed to the higher initial protein load in the CVW substrate (15.41%), which facilitated earlier enzymatic induction. Furthermore, the in situ biological buffering provided by the CVW system reached its maximum potential on Day 3, maintaining a more favorable acidogenic environment for protease enzymes during the initial stages. In contrast, the rapid acidification of the FW-LBR during the first three days (pH ~ 3.2) likely disrupted the transmembrane pH gradient of the acidogenic microflora, thereby suppressing protease secretion until the pH was regulated above 4.5 through leachate recirculation later in the first week. These results suggest that substrate-induced regulation and early pH stability are critical for maximizing protein hydrolysis in complex waste streams. The temporal dynamics of hydrolytic enzyme activity revealed that the initial 3–5 days constituted a critical phase for substrate solubilization. During this period, α-amylase and protease activities reached their maximum, coinciding with peak buffering capacity in both CVW and FW systems (Figs. 1c and 2a-c). In the CVW-LBR, maximum α-amylase activity was observed on day one at a relatively low pH (~ 3.7), followed by a decline until day five, after which enzyme activity gradually increased in parallel with rising pH. In contrast, the FW-LBR exhibited peak α-amylase activity on 5th day at a comparatively higher pH (~ 4.2), indicating a delayed but more sustained enzymatic response. The restricted hydrolysis in CVW shortened the duration of hydrolytic enzyme activity and reduced both the buffering potential and its persistence, whereas FW demonstrated a progressive increase in buffering capacity from day 5 onward, corresponding with enhanced amylase activity during the first week. From 12th day onwards (pH > 5.2), significant accumulation of VFAs in the FW-LBR further increased the buffering potential (Figs. 1d and 2a-c). From the seventh day onwards, the hydrolytic enzymatic activity decreased in both LBRs due to changes in redox and buffering environments and in VFA dynamics28,41. The polysaccharide degradation rate was lower for CVW due to decreased enzyme activity on lignocellulosic biomass, leading to incomplete degradation over time42. Hence, for lignocellulosic biomass, enzyme addition was considered the best strategy to enhance biogas formation43–45. However, enzyme addition increases the operational cost and complexity of the anaerobic digestion process. The present investigation continuous recirculation of LBR leachate and methanogenic effluent helped regulate pH and enhance the bioavailability of enriched acidogenic microorganisms, thereby increasing bioactivity in the LBRs. Rate constant analysis To understand the effects of enzyme characteristics on acidogenic environment, it is important to measure hydrolysis, acidogenesis, and acetogenesis rates and rate constants in each phase of FW-LBR and CVW-LBR operation. Organic waste solubilization inside the LBR, driven by the cumulative action of hydrolytic enzymes, was modeled using first-order kinetics (dC/dt = -kC), where k represents the rate constant for each biochemical phase. This model effectively describes the spontaneous increase in VFA and sugar concentrations observed during the first half of the digestion period. Hydrolysis, acidogenesis, and acetogenesis rates and rate constants (d−1) were observed as (78%, 40.31%, 35.78% and 0.037, 0.019, 0.017) for CVW and (84%, 46.37%, 12% and 0.040, 0.022, 0.0057) for FW (Table 1). The analysis shows that hydrolysis and acidogenesis rates were higher for FW, whereas the acetogenesis rate was approximately 3 times higher for CVW (Table 1).Table 1Summary of acidogenesis performance in the leach bed reactor.Parameters2:3 FW:VWFood waste0 day21st day0 day21st dayTS (%)14.5 ± 1.66.73 ± 1.2119.46 ± 0.349.43 ± 1.41VS (%)94.65 ± 7.4138.65 ± 12.5196.83 ± 0.4556.05 ± 0.97Carbohydrate (%)50.90 ± 8.216.69 ± 0.2165.35 ± 8.528.77 ± 1.21Protein (%)15.41 ± 3.2712.39 ± 2.59.56 ± 2.322.65 ± 1.33Lignin (%)20 ± 3.512 ± 3.22tracesTracesHemicellulose (%)17 ± 4.15 ± 2.1tracesTracesCellulose (%)38 ± 6.323 ± 0.43.4 ± 0.211.2 ± 0.3COD removal (%)80.6785.24VS removal (%)59.1642.01Hydrolysis rate (%)7884Hydrolysis rate constant (d-1)0.0370.040Acidogenesis rate (%)40.3146.37Acidogenesis rate constant (d-1)0.0190.022Acetogenesis rate (%)35.7812Acetogenesis rate constant (d-1)0.0170.0057Peak sugar (day 9) (g/L)2.632.56Lactic acid (g/L)60.90 g/L (642.46 g/kg VS removed)73.58 g/L (585.28 g/kg VS removed)Acetic acid (g/L)18.19 g/L (191.92 g/kg VS removed)10.16 g/L (80.83 g/kg VS removed)Propionic acid (g/L)22.64 g/L (238.84 g/kg VS removed)95.29 g/L (758.01 g/kg VS removed)Butyric acid (g/L)4.56 g/L (48.12 g/kg VS removed)5.54 g/L (44.06 g/kg VS removed)Total H2 Yield (mL/g VS removed)110 mL/g VS removed (36.4% of biogas composition)121.3 mL/g VS removed (43% of biogas composition) Table 1Summary of acidogenesis performance in the leach bed reactor.Parameters2:3 FW:VWFood waste0 day21st day0 day21st dayTS (%)14.5 ± 1.66.73 ± 1.2119.46 ± 0.349.43 ± 1.41VS (%)94.65 ± 7.4138.65 ± 12.5196.83 ± 0.4556.05 ± 0.97Carbohydrate (%)50.90 ± 8.216.69 ± 0.2165.35 ± 8.528.77 ± 1.21Protein (%)15.41 ± 3.2712.39 ± 2.59.56 ± 2.322.65 ± 1.33Lignin (%)20 ± 3.512 ± 3.22tracesTracesHemicellulose (%)17 ± 4.15 ± 2.1tracesTracesCellulose (%)38 ± 6.323 ± 0.43.4 ± 0.211.2 ± 0.3COD removal (%)80.6785.24VS removal (%)59.1642.01Hydrolysis rate (%)7884Hydrolysis rate constant (d-1)0.0370.040Acidogenesis rate (%)40.3146.37Acidogenesis rate constant (d-1)0.0190.022Acetogenesis rate (%)35.7812Acetogenesis rate constant (d-1)0.0170.0057Peak sugar (day 9) (g/L)2.632.56Lactic acid (g/L)60.90 g/L (642.46 g/kg VS removed)73.58 g/L (585.28 g/kg VS removed)Acetic acid (g/L)18.19 g/L (191.92 g/kg VS removed)10.16 g/L (80.83 g/kg VS removed)Propionic acid (g/L)22.64 g/L (238.84 g/kg VS removed)95.29 g/L (758.01 g/kg VS removed)Butyric acid (g/L)4.56 g/L (48.12 g/kg VS removed)5.54 g/L (44.06 g/kg VS removed)Total H2 Yield (mL/g VS removed)110 mL/g VS removed (36.4% of biogas composition)121.3 mL/g VS removed (43% of biogas composition) Summary of acidogenesis performance in the leach bed reactor. While acetogenesis is a H2-producing step, the higher total H2​ yield in FW (121.3 mL/gVS removed) is attributed to its higher overall acidogenesis rate constant (0.022 d−1), which reflects a more robust metabolic flux through primary H2-producing pathways. In contrast, the higher acetogenesis rate constant in CVW (0.017 d−1) resulted in superior acetic acid recovery (18.19 g/L) but did not translate to higher gas yields because the system was limited by the slower hydrolysis of its lignocellulosic fractions in CVW (0.037 d−1) compared to FW (0.040 d−1). This confirms that H2​ production is more closely linked to the total vigor of the initial acidogenic phase rather than the relative efficiency of the final acetogenic conversion. For CVW, a minor lag in the first 48 h was attributed to the complex lignocellulosic polymers (20% lignin), after which the system strictly obeyed first-order kinetics as the biological buffering stabilized the environment. Enzyme action varied depending upon substrate types, and retarded activity was observed during hydrolysis of lignocellulosic waste33,46. Total soluble products (TSP) analysis Reducing sugar In this investigation, the formation of reduced sugars from different feedstocks and the relationships with enzymatic activities were studied. Reducing sugar dynamics differed markedly between the systems, reflecting differences in substrate characteristics. In FW-LBR, concentrations increased rapidly from 0.365 to 2.56 g/L by day 9 (Fig. 3a), followed by a sharp decline, indicating rapid hydrolysis of readily biodegradable substrates and subsequent consumption. In contrast, CVW-LBR showed an initial decrease within 24 h due to the lignocellulosic barrier, followed by a steady increase from 0.752 to 2.63 g/L until day 9 and a gradual decline thereafter (Fig. 3a). This pattern suggests regulated hydrolysis mediated by the lignocellulosic matrix, resulting in sustained sugar release. Both systems peaked on day 9, but sugar composition shifted from sucrose and maltose in the early phase to pentoses (ribose in FW and arabinose in CVW) at later stages, reflecting progressive degradation of complex polysaccharides. The NH₄⁺–N concentration in the CVW system increased progressively, reaching a maximum of 57.53 mg/L by the third week, which was likely due to continuous protein degradation facilitated by higher protease activity (Fig. 3b). In contrast, NH₄⁺–N levels in the FW system remained negligible, thus consistent with its lower initial protein content and comparatively lower protease activity (23.17 U/mL vs. 34.94 U/mL in CVW). Additionally, rapid acidification in the FW reactor (pH 3.2 within 72 h) likely inhibited proteolytic microbial activity and enzyme secretion, further limiting ammonium production. Notably, NH₄⁺–N concentrations in both systems remained below inhibitory levels17.Fig. 3Comparative production of (a) reducing sugar, (b) NH4+-N and sugar distribution, (c) for composite vegetable waste (2:3 FW: VW), and (d) Food waste in LBR leachate. Comparative production of (a) reducing sugar, (b) NH4+-N and sugar distribution, (c) for composite vegetable waste (2:3 FW: VW), and (d) Food waste in LBR leachate. The distribution of reducing sugars differed between CVW and FW fed LBR leachates. In CVW, the maltose concentration and α-amylase activity were directly proportional; they were higher on the first day and decreased on the third day. Later, the maltose production started steadily increasing and reached a maximum on the 12th day, when the glucose concentration was also maximum. After 12th day, the arabinose concentration started increasing and was dominating till the end of the experiment (Fig. 3c). Release of galactose on the third day indicated β galactosidase activity in the medium. Whereas, for FW on the first day, sucrose concentration was higher than maltose which remained steady till the 15th day. From the third day onwards, the maltose concentration decreased, but from the fifth day onwards, the ribose concentration increased and remained dominant until the end of the experiment (Fig. 3d). Higher α-amylase activity in FW is justified by higher production of reducing sugars and VFA. The peak α-amylase activity (72.2 U/mL) and protease activity (34.94 U/mL) observed in our reactors are significantly higher than values typically reported for non-enriched anaerobic systems12,47. Whereas β-galactosidase and protease activity were prominent for CVW due to the presence of pectin content (poly galacturonic acid) and polyphenol contents in the cell walls of vegetable wastes23,47,48. While some researchers utilize commercial enzyme cocktails to enhance the hydrolysis of lignocellulosic waste12, our study achieves comparable or superior hydrolytic vigor through in situ biological regulation and leachate recirculation. VFA distribution Quantification of VFAs at regular intervals assists in understanding the metabolic shifts. Production of mixed VFAs was higher in FW over CVW attributing to the mixed fermentation pathways generating VFA along with H2. Lactic acid, acetic acid, propionic acid, butyric acid, and isovaleric acid were dominant in both FW and CVW. For CVW, the cumulative production of acetic acid was 18.19 g/L (191.92 g/kg VS removed), propionic acid was 22.64 g/L (238.84 g/kg VS removed), butyric acid was 4.56 g/L (48.12 g/kg VS removed), and lactic acid was 60.90 g/L (642.46 g/kg VS removed). In comparison, FW showed a higher accumulation of propionic acid at 95.29 g/L (758.01 g/kg VS removed) and lactic acid at 73.58 g/L (585.28 g/kg VS removed), but a lower acetic acid of 10.16 g/L (80.83 g/kg VS removed) and butyric acid of 5.54 g/L (44.06 g/kg VS removed) production. This standardized reporting confirms that while FW yields higher total organic acids, CVW is significantly more efficient for acetate-specific recovery, with a 1.8-times higher volumetric concentration. However, the productivity of other short and long chain fatty acids, e.g., propionic acid, butyric acid, isovaleric acid, and lactic acid, was 4.2, 1.2, 2.06, and 1.21 times higher for FW, respectively (Table 1). Lactate, being a highly reduced and electron-rich substrate, is preferentially converted to propionate rather than acetate or butyrate, under reducing conditions or when H2 removal is limited. Consequently, the higher lactate accumulation in FW-LBR favoured its spontaneous conversion to propionate under reduced conditions compared to CVW-LBR49. For CVW, the butyric acid/acetic acid ratio of 0.37–0.54 was less than FW (0.37–1.0), indicating that the anaerobic co-digestion conditions were stable for H2 production23,46. The system’s pH directly affects bacterial growth and metabolism by influencing the thermodynamics of metabolic pathways and the products formed, thereby improving the efficiency of VFA and H2 production29. The VFA profiles demonstrated clear substrate-dependent metabolic differentiation between FW and CVW systems. While FW produced substantially higher propionic acid concentrations (95.3 g/L), CVW exhibited 1.8-times higher acetic acid productivity (18.2 g/L), indicating distinct acidogenic pathways. During the initial phase (days 1–9), lactic acid dominated the system, accounting for up to 99% in FW, coinciding with elevated hydrolytic enzyme activity at acidic pH (< 4.5). Following this phase, a metabolic transition was observed. After day 7, acidogenesis intensified, and pH stabilization above 5.0 via in situ buffering led to a decline in lactic acid and a concomitant increase in VFA. In the CVW system, this transition was characterized by a progressive shift toward acetic acid, which reached 42% by day 15, alongside a gradual increase in propionic acid, which ultimately reached 64% by day 18. Additionally, butyric acid formation in CVW (14–23%) was observed predominantly during the final stages of the experiment (Fig. 4a). In contrast, FW exhibited the dominance of propionic acid (40–86%), with comparatively lower acetic acid contribution (13.6%), along with the production of other VFAs such as butyric (12%), isovaleric (13.31%), and valeric acids (11%) (Fig. 4b).Fig. 4Volatile fatty acid distribution for (a) composite vegetable waste, (b) food waste, and (c) comparative H2 gas production analysis from (CVW) and food waste. Volatile fatty acid distribution for (a) composite vegetable waste, (b) food waste, and (c) comparative H2 gas production analysis from (CVW) and food waste. These shifts in VFA distribution are mechanistically correlated with pH-dependent enzymatic profiles. The alignment of VFA transitions with specific pKa buffering zones (7.8 and 8.2) and peak enzymatic activities suggests a biochemical adaptation of the acidogenic environment. The measured rate constants (k) further support a robust metabolic flux shift that explains the 4.2-fold higher propionic acid yield in the FW system. Maintaining the system pH above the pKa of specific VFAs is essential for sustained production, as a decrease below pKa leads to intracellular accumulation of undissociated acids, inhibiting microbial growth and reducing VFA formation29,50. Increased ionic strength and inhibition by undissociated acids (pKa ≈ 4.75–4.8 for acetic and butyric acids) further suppress metabolic activity under acidic conditions (pH 4.0—5.0)29. Moreover, during the redox phase of AD, distinct groups of acidogenic microorganisms proliferate within specific pH niches, influencing substrate utilization and overall metabolic pathways31,32,51. The observed VFA distribution thus proposed biochemical framework to maintain intracellular pH homeostasis through selective regulation of metabolic pathways28. Notably, the duration of acidogenesis was closely linked to buffering capacity; in FW, prolonged acidogenesis during the first 12 days corresponded with increased buffering activity, driven by VFA accumulation production52,53 (Fig. 1c & d). Biohydrogen production Hydrogen offers a clean and efficient energy alternative with the highest gravimetric energy density (122 kJ/g) and can be produced from renewable resources such as food and vegetable wastes8,20. Hydrogen yield from FW was higher (121.3 mL/g VS removed; 43% of biogas) compared to CVW (110 mL/g VS removed; 36.4% of biogas) (Fig. 4c). Hydrogen production was observed till the 18th day for FW, during which the VFA accumulation was higher. In the case of CVW, the H2 production was observed till the 21st day of the experiment and decreased after that, and the cumulative volume of H2 production was less in CVW than in FW. These yields are comparable to the previous reports15,17,52, who achieved similar H2 production from FW via acidogenic fermentation. Notably, our results were obtained without the external addition of chemical buffers (e.g., NaHCO3​), which are typically required in literature to prevent acidification. Steady production of undissociated organic acids, detected after the 12th day of fermentation in FW, is linked to increased H2 production. With increasing pH, the VFA concentration increased because mass transfer and substrate degradation rates increased, while the diffusion-limited dissolution rate decreased54. At the highest undissociated acid concentrations, the buffer capacity decreased, and H2 production decreased in both FW and CVW. The VFA composition reflects the potential differences in H2 composition and yields from FW and CVW. A decrement in VFA due to slower hydrolysis was observed until the end of the experiment for CVW, which led to cessation of H2 production. Whereas, for FW from the third day onwards, propionic acid increased rapidly from 40 to 86%, and the total VFA production was also higher, leading to higher H2 yield55. Although propionate-forming pathways are typically H2-consuming, the high yields of both H2​ and propionic acid in the FW system are attributed to the exceptionally high acidogenic flux (46.37% rate) of the starch-rich FW (65.35% carbohydrate). This rapid fermentation generated high amount of total soluble products created an environment where H2-producing pathways (yielding acetate and butyrate) were highly active, generating enough H2​ to both satisfy the requirements of propionate-producing microflora and maintain a high cumulative gas yield. Higher butyric acid/acetic acid ratio of FW-LBR (0.37 to 1) than CVW-LBR (0.37–0.54) indicates more H2 yield in FW23,46. Furthermore, the metabolic shift from lactic acid to propionic acid after 9th day coincided with an increase in pH and improved mass transfer. As the concentration of undissociated organic acids was regulated through in situ buffering, the inhibition of hydrogenase enzymes was mitigated, allowing for sustained H2​ production even during the phase of high propionate accumulation. Sustainability, LCA analysis, study limitations and future scope Sustainability analysis was compared between H2 production via CVW and FW using the CML-IA baseline V3.05 / World 2000- and single-issue IPCC GWP 100a methods in SimaPro 8.5.2.0. The sustainability analysis (Table 2) shows that the process has a negligible impact on most environmental factors, as it lacks chemical pretreatment or external buffers. This process emphasizes reuse, minimal water use, and wastewater treatment. A total of 1.1 L of water and 0.35 kWh of electricity were consumed by the peristaltic pump to transfer leachate from LBR to the airlift reactor. While the process avoids chemical-intensive treatments, a critical assessment of the sustainability parameters reveals that the overall environmental footprint is highly sensitive to the electricity demand for recirculation. This low energy demand confirms that the role of electricity in influencing environmental parameters is minimal when compared to chemical-intensive treatment methods. Abiotic Depletion (Fossil Fuels) and Marine Aquatic Ecotoxicity were identified as the primary environmental hotspots. Specifically, CVW showed significantly higher abiotic depletion (2.71 MJ) and marine ecotoxicity (477.07 kg 1,4-DB eq) for H2 production compared to FW (0.568 MJ and 99.93 kg 1,4-DB eq, respectively) (Table 2). The difference in high Abiotic Depletion of fossil fuels is largely due to the electricity demand of the recirculation pumps relative to the final product yield. This is also attributed to the higher energy intensity required to process the lower-yielding lignocellulosic waste.Table 2Comparative environmental impacts of acidogenesis products and biohydrogen production from FW and CVW.Impact categoryUnitLactic acidPropionic acidAcetic acidHydrogenFood waste2:3 FW:VWFood waste2:3 FW:VWFood waste2:3 FW:VWFood waste2:3 FW:VWAbiotic depletionkg Sb eq1.20E-074.97E-079.31E-081.34E-068.69E-071.66E-062.53E-051.2E-3Abiotic depletion (fossil fuels)MJ2.7E-30.0112.1E-30.0290.0190.0370.572.71Global warming (GWP100a)kg CO2 eq0.27E-31.1E-30.20E-32.97E-31.93E-33.7E-30.0560.269Ozone layer depletion (ODP)kg CFC-11 eq2.54E-091.05E-081.96E-092.81E-081.83E-083.50E-085.34E-072.55E-06Human toxicitykg 1,4-DB eq0.110.430.081.20.7581.44922.110105.55Fresh water aquatic ecotoxkg 1,4-DB eq0.13E-30.56E-30.1E-31.5E-30.97E-31.86E-30.0280.14Marine aquatic ecotoxicitykg 1,4-DB eq0.481.960.355.273.426.5599.93477.07Terrestrial ecotoxicitykg 1,4-DB eq0.00080.00330.00060.0080.0060.0110.170.81Photochemical oxidationkg C2H4 eq4.30E-050.00023.32E-050.00040.00030.00060.0090.043Acidificationkg SO2 eq0.00090.00400.00070.0110.0070.0130.2050.98Eutrophicationkg PO4–- eq0.00050.0020.00040.0050.0040.0070.110.52Significant values are in bold. Table 2Comparative environmental impacts of acidogenesis products and biohydrogen production from FW and CVW.Impact categoryUnitLactic acidPropionic acidAcetic acidHydrogenFood waste2:3 FW:VWFood waste2:3 FW:VWFood waste2:3 FW:VWFood waste2:3 FW:VWAbiotic depletionkg Sb eq1.20E-074.97E-079.31E-081.34E-068.69E-071.66E-062.53E-051.2E-3Abiotic depletion (fossil fuels)MJ2.7E-30.0112.1E-30.0290.0190.0370.572.71Global warming (GWP100a)kg CO2 eq0.27E-31.1E-30.20E-32.97E-31.93E-33.7E-30.0560.269Ozone layer depletion (ODP)kg CFC-11 eq2.54E-091.05E-081.96E-092.81E-081.83E-083.50E-085.34E-072.55E-06Human toxicitykg 1,4-DB eq0.110.430.081.20.7581.44922.110105.55Fresh water aquatic ecotoxkg 1,4-DB eq0.13E-30.56E-30.1E-31.5E-30.97E-31.86E-30.0280.14Marine aquatic ecotoxicitykg 1,4-DB eq0.481.960.355.273.426.5599.93477.07Terrestrial ecotoxicitykg 1,4-DB eq0.00080.00330.00060.0080.0060.0110.170.81Photochemical oxidationkg C2H4 eq4.30E-050.00023.32E-050.00040.00030.00060.0090.043Acidificationkg SO2 eq0.00090.00400.00070.0110.0070.0130.2050.98Eutrophicationkg PO4–- eq0.00050.0020.00040.0050.0040.0070.110.52Significant values are in bold. Comparative environmental impacts of acidogenesis products and biohydrogen production from FW and CVW. Significant values are in bold. Sensitivity analysis depicted that the present process for the production of H2 and biochemicals through two stages using FW has negligible influence on marine aquatic ecotoxicity, on global warming (GWP), ozone layer depletion (ODP), human toxicity, fresh water aquatic ecotoxicity, marine aquatic, terrestrial ecotoxicity, photochemical oxidation, acidification and eutrophication except on abiotic depletion (fossil fuels) over CVW (Table 2). A critical assessment of the sustainability parameters (Table 2) reveals that the environmental impacts of the CVW system are roughly 4.7 times higher than those of the FW system for H2 production. This discrepancy is primarily attributed to the inverse correlation between substrate complexity and product yield. The CVW is characterized by a significant lignocellulosic fraction (20% lignin and 38% cellulose), which retards the rate-limiting hydrolysis step (CVW 0.037 d−1 for CVW vs 0.040 d−1 for FW) and results in a lower cumulative H2 yield (110 mL/gVS removed) compared to the easily fermentable FW (121.3 mL/gVS removed). This slower solubilization means the system operates less efficiently in terms of rapid product conversion, leading to higher impacts per unit of bioresource recovered. Since the electricity demand for the peristaltic pump (350 Wh) was identical for both reactors, the lower H2 yield in the CVW system resulted in a significantly higher environmental impact per functional unit compared to FW. This identifies electricity consumption as the primary environmental hotspot. Consequently, while CVW is superior for maintaining the homeostatic acidogenic environment and avoiding chemical costs, it carries a higher ‘energy penalty’ due to its complex biochemical structure. While FW is more “efficient” per liter of H2, it is prone to rapid acidification and metabolic inhibition, which would require chemical intervention (e.g., NaOH or carbonate buffers) in a non-recirculating or non-co-digested system inputs that were intentionally excluded from this zero-chemical model. Future optimizations focusing on reducing the recirculation energy or enhancing the solubilization of lignin could bridge this sustainability gap between CVW and FW. Additionally, in LBR, biological hydrolysis and acidogenesis lead to the production of short chain fatty acids (SCFA), with lactate and propionate as the predominant products. This research mechanistically correlates an SCFA recovery model with lactate dominating the others during a two-stage AD without pretreatment. Lactate and propionate have huge market potential and are expected to reach USD 8.77 billion for lactate by 2025 at a CAGR of 18.7% and USD 363.3 million for calcium propionate by 2023 at a CAGR of 5.6%. Hence, recovery of these SCFA (mainly lactate and propionate) at pilot-scale will add value to the process and also serves as the future scope of this investigation. Despite the advancement of a zero-chemical-addition model, this investigation has limitations that should be addressed in future research. First, the LBR configuration is subject to channeling and a lack of agitation, which can limit the mass transfer of enzymes and nutrients. Second, while the co-digestion of CVW regulated the acidogenic environment, the lignin content hindered complete degradation of the substrate within the batch timeframe. Third, the sustainability of the process is currently sensitive to the electricity demand of operational equipment. Future research will focus on scaling this two-stage system to pilot-level continuous-flow regimes to optimize the recovery of high-value metabolites and integrating metagenomic and pathway flux analyses to validate the biochemical mechanistic correlations documented in this investigation. Conclusions This research indicates a novel, sustainable model for in situ biological buffering to regulate the pH during two-stage anaerobic digestion, achieved through co-digestion and leachate recirculation. The results demonstrate that using lignocellulosic vegetable waste in a co-digestion strategy effectively regulates the acidogenic environment and sustains endogenous enzymatic activities without the need for external chemical buffers. The defining contribution of this investigation is the correlation between biological buffering potential and a panel of three hydrolytic enzymes (α-amylase, β-galactosidase, and protease), which directly dictates the rate of hydrolysis and subsequent H2 yield. Key findings indicated that FW exhibited a higher buffering capacity (0.145 ± 0.001 mol/L) and H2 yield (121.3 mL/g VS removed). However, the CVW provided a more regulated enzyme-pH environment, resulting in 1.8 times higher acetic acid productivity (18.2 g/L) compared to FW. Moreover, methanogenic effluent recirculation significantly reduced freshwater intake, enhanced in situ buffering, contributing to a more resource-efficient process. Finally, LCA indicates that the process has a reduced environmental footprint compared to chemical-intensive methods, it also identifies electricity consumption for recirculation as a primary environmental hotspot. Despite limitations such as reactor channeling and incomplete lignocellulosic degradation, the model is confirmed as a zero-chemical-addition and minimized-discharge process. Overall, linking biological buffering with enzymatic performance provides a robust framework for optimizing H2 and VFA recovery from complex organic waste streams, with clear potential for future scale-up following energy and process optimization. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1