Hybrid PVA-alginate immobilization enables configuration-dependent biohydrogen production from Napier grass and oil palm fronds with superior multi-cycle stability Authors: Nantharat Wongfaed, Sureewan Sittijunda, Alissara Reungsang Published: 2026 ABSTRACT Dual-matrix approaches face stability limitations (83.8% retention, 4 cycles). We developed a simplified hybrid polyvinyl alcohol (PVA)-alginate immobilization system (4.50% PVA, 1.50% alginate, 0.25% silica, 0.50% chitosan) that independently immobilizes cellulolytic organisms (modified KKU-MC1) and hydrogen (H2)-producing bacteria (HPB). This platform enabled systematic evaluation of four process configurations combining simultaneous saccharification and fermentation (SSF) or separate hydrolysis and fermentation (SHF) with selective immobilization using Napier grass (NG) and oil palm fronds (OPF). Configuration profoundly influenced performance: SHF1 (free modified KKU-MC1 + immobilized HPB) achieved 81.5 ± 7.5 mL-H2/g-volatile solid (VS) from NG 2.2-fold higher than SSF while SSF2 (immobilized modified KKU-MC1 + free HPB) demonstrated extraordinary co-digestion synergy at 7.5:2.5 NG:OPF ratio (61.6 ± 7.4 mL-H2/g-VS, 12.5-fold the theoretical). Five-cycle testing revealed 84.0–89.1% H2 retention, 50.0–66.7% enzyme retention, and 50.0–80.0% biomass retention with intact beads, outperforming dual-matrix systems (5 vs. 4 cycles). Process configuration and substrate characteristics proved the primary determinants of H2 yield, while medium components represented 57.3% of costs. With five-cycle stability providing 5.0-fold economic advantages (0.00364 USD-H2/USD-total), this platform establishes configuration-substrate-immobilization as critical for agricultural bio-H2 production.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-026-53362-1. Introduction The global transition from fossil fuels to renewable energy systems has become increasingly urgent, with atmospheric CO₂ levels exceeding 425 parts per million as of 20241. Global energy demand continues to intensify, with hydrogen (H2) projected to constitute 12–22% of total final energy consumption by 2050 under net-zero scenarios2. Among emerging renewable energy carriers, bio-hydrogen (bio-H₂) offers singularly high gravimetric energy density of 142 MJ/kg—approximately three times that of gasoline—zero-emission combustion, and potential for carbon–neutral production3. Biological hydrogen production (HP) by dark fermentation (DF) operates continuously at ambient conditions without light dependence, offering advantages including waste valorization, diverse substrate acceptance, and relatively high production rates3,4. Agricultural residues represent a substantially underutilized feedstock source, with 5.5 billion tons produced annually worldwide5. Quantitative modeling suggests bio-HP potential from agricultural residues could reach approximately 300 PJ annually with proper pretreatment strategies, representing 191% improvement over non-pretreated biomass6. In Southeast Asia, Napier grass (NG) and oil palm fronds (OPF) are among the most abundant lignocellulosic materials. NG offers high yields of 200–300 tons fresh matter per hectare annually, contributing to Thailand’s biomass potential exceeding 190 million tons per year7. NG contains cellulose (35–40%), hemicellulose (25–30%), and lignin (8–15%)8. OPF accumulate at 10–15 tons dry matter per hectare per year across Indonesia, Malaysia, and Thailand9,10, containing approximately 35–40% cellulose, 20–38% hemicellulose, and 15–32% lignin with rich macro- and micronutrients9,11. However, lignocellulosic biomass recalcitrance necessitates effective process optimization for economically viable hydrogen yields (HYs)12,13. Mixed microbial consortia offer substantial operational advantages over pure cultures for lignocellulose-to-H2 bioconversion, as pure cultures possess limited hydrolytic activities, persistent contamination susceptibility, and require stringent sterile conditions14. Mixed cultures demonstrate HY improvements through metabolic diversity, process robustness, and elimination of sterility requirements15,16. Mixed culture biotechnology is increasingly recognized as a robust alternative with efficient waste treatment capacity and bio-H2 co-generation17,18. The modified KKU-MC1 consortium is a cellulose-degrading microbial community derived from rice straw compost, soil from goat and sheep stalls, and termite guts. This consortium exhibits mesophilic growth, broad substrate utilization, and steady bio-HP without waste sterilization. Methanogenic activity was eliminated through 2-bromoethanesulfonate (BES) treatment for bio-H2 studies16. The resulting modified consortium comprises predominantly Pseudomonadota (formerly Proteobacteria) (46%), Bacillota (formerly Firmicutes) (28%), and Bacteroidota (formerly Bacteroidetes) (7%), with minor contributions from Actinomycetota (formerly Actinobacteria) and Mycoplasmatota (formerly Tenericutes/Mollicutes), providing complementary hydrolytic and fermentative capabilities (Wongfaed et al., submitted). Cell immobilization technology addresses operational challenges in continuous and repeated-batch systems through improved cell density, enhanced substrate tolerance, cell washout prevention, pH stability, and tolerance against toxic components4,19–21. Immobilization emerges as critical strategy to overcome suspended culture limitations, particularly where biomass retention is essential22. Previous work by Wongfaed et al.23 developed a dual-matrix core–shell entrapment system reporting 119.9 ± 10.5 mL-H₂/g-volatile solid (VS) from NG with 83.8% retention over four cycles. Despite this advancement, critical limitations remained: (1) immobilization applied only to cellulolytic modified KKU-MC1 while H2-producing bacteria (HPB) remained free; (2) complex sequential core–shell preparation; (3) structural degradation after approximately 30 days; and (4) absence of comprehensive economic assessment. These limitations reveal three critical knowledge gaps. First, whether to immobilize cellulolytic organisms or HPB, and how this interacts with process architecture—Simultaneous Saccharification and Fermentation (SSF) versus Separate Hydrolysis and Fermentation (SHF)—requires clarification. While commercial enzyme cocktails enable environmentally friendly saccharification, their industrial viability is limited by high costs24. Second, although co-digestion strategies have been investigated separately25,26, systematic assessment of synergistic improvements between NG and OPF with hybrid immobilization remains unexplored. Co-digestion offers complementary nutrient profiles, balanced carbon to nitrogen (C/N) ratios, and diverse carbohydrate compositions8,27. Third, material selection requires optimization balancing performance, cost, and stability. PVA-alginate carriers offer good mechanical properties through dual gelation mechanisms19,20,28,29. Silica reinforcement provides structural enhancement30, while chitosan enhances cell adhesion19. To address these gaps, this study developed a hybrid PVA-alginate-silica-chitosan immobilization system that independently immobilizes cellulolytic organisms (modified KKU-MC1) and HPB, enabling systematic comparison of four process configurations (SSF1, SSF2, SHF1, SHF2) across mono- and co-digestion of NG and OPF, assessing configuration-dependent HP performance, co-digestion synergy, multi-cycle operational stability, and economic viability. Materials and methods Substrates preparation and characterization NG (Pak Chong 1 variety) was harvested at 45–50 days maturity. OPF from Tenera variety oil palms (5–7 years old) were collected during routine fruit harvesting operations; fresh OPF are typically removed to facilitate harvesting and left beneath trees for natural decomposition. Both biomass materials were kindly provided by Sahaviriya Steel Industries Public Company Limited from their plantations in Prachuap Khiri Khan province, Thailand, with permission for use in this research. To minimize compositional variations from seasonal and maturity factors, all materials were collected from a single location within a defined period. Substrates were cut into 2 cm segments and dried at 60 °C for 72–96 h until moisture content fell below 10%. Materials were further reduced to 0.2–0.4 cm using a food blender, sieved through No.8 mesh, and thoroughly mixed to ensure uniform particle size distribution for optimal enzyme–substrate contact23. Samples were stored in sealed containers at room temperature (25 ± 2 °C) until use. Substrate composition was analyzed for structural carbohydrates and lignin following NREL protocols31, and for total solids and volatile solids following APHA32, with physicochemical characteristics presented in Supplementary Table S1. Microorganisms and inoculum preparation The original KKU-MC1 was a cellulose-degrading microbial community derived from naturally rich lignocellulose-degrading enzymes sources (rice straw compost, soil from goat and sheep stalls, and termite guts), comprising predominantly Bacteroidota (51%), Pseudomonadota (29%), and Bacillota (10%), with the remaining 10% consisting of unknown bacteria (8%), unidentified organisms (0.4%), Archaea (0.6%), and other minor bacterial species (1%)16. To eliminate methanogenic activity, the original consortium was modified through supplementation with 10 mM BES at 37 °C for 72 h, followed by more than 10 successive subcultures in peptone cellulose solution (PCS) medium to ensure stable methanogen suppression. The resulting modified KKU-MC1 comprised predominantly Pseudomonadota (46%), Bacillota (28%), and Bacteroidota (7%), with minor contributions from Actinomycetota and Mycoplasmatota (Wongfaed et al., submitted). Bacillota fraction includes obligate anaerobes such as Clostridium spp., while the Pseudomonadota fraction includes facultative anaerobes such as Enterobacter spp., both of which are well-established dark fermentative H₂ producers that thrive under anaerobic conditions (Wongfaed et al., submitted). The HPB were obtained from an upflow anaerobic sludge blanket (UASB) reactor treating sugarcane leaves. To suppress methanogenic activity, the activated sludge was heat-treated at 95 °C for 30 min prior to enrichment. The community was stabilized through more than 10 successive subcultures in BM7 medium (10 g/L glucose, 5 g/L xylose) at 37 °C and 150 rpm. The enriched community was dominated by Bacillota (72%) and Pseudomonadota (11%), with no detectable methanogenic archaea confirmed by 16S rRNA sequencing (Wongfaed et al., submitted). Cultures were grown aerobically under the same conditions for 18–24 h before use in HP experiments to promote rapid biomass accumulation; upon transfer to sealed, N₂-flushed serum bottles, the facultatively anaerobic community shifts to anaerobic DF for HP (Wongfaed et al., submitted). Hybrid immobilization by entrapment-encapsulation technique Cell preparation Modified KKU-MC1 was cultivated in PCS medium containing Whatman No. 1 filter paper (1.0 × 6.0 cm) as the cellulose source. For both immobilization and free cell experiments, cultures were harvested by centrifugation (5,000 rpm, 5 min) and rinsed three times with sterile 0.85% (w/v) NaCl solution to remove residual media components. For modified KKU-MC1 cultures grown in PCS medium, residual Whatman No. 1 filter paper pieces were first removed by decanting the culture suspension into a clean vessel prior to centrifugation, as the macroscopic paper fragments (1.0 × 6.0 cm) are physically distinguishable from microbial cells and sediment independently upon standing. Cells were resuspended in sterile 0.85% NaCl to OD600 = 1.8 ± 0.1 (corresponding to VSS 2.10 ± 0.50% w/v). HPB were prepared by the same centrifugation and rinsing procedure, resuspended to OD600 = 1.5 ± 0.2 (corresponding to VSS 2.00 ± 0.75% w/v). These stock cultures were used for both free cell experiments and immobilization procedures. For free cell experiments, stock cultures were inoculated at organism-specific volumes into experimental bottles: modified KKU-MC1 at 20% (v/v) and HPB at 20% (v/v), achieving target initial VSS concentrations of 0.42% and 0.40% (w/v), respectively, in working volume, resulting in comparable total biomass loading across all configurations (Table S2). Filter paper from the pre-culture medium was not transferred into any experimental bottle; the sole cellulosic carbon source in all SSF and SHF experiments was the added NG and/or OPF substrate. Bead matrix preparation A polymer solution was prepared by heating 50 mL sterile distilled water to 95 °C on a hotplate stirrer, followed by sequential addition of matrix components: 4.50% (w/v) PVA (degree of polymerization = 1750, molecular weight 85,000–124,000, 99% hydrolyzed), 1.50% (w/v) sodium alginate, 0.25% (w/v) silica, and 0.50% (w/v) chitosan. The mixture was stirred continuously at 60 rpm for 1 h to ensure complete dissolution. Subsequently, 42 mL sterile distilled water was added, and stirring continued until the solution cooled to ambient temperature and achieved homogeneity. Cell immobilization procedure The 0.5 g wet cell pellet (modified KKU-MC1 or HPB) was directly incorporated into the cooled matrix solution and mixed at 60 rpm for 1 h to ensure homogeneous cell distribution. The cell–matrix mixture was extruded dropwise through a 10 mL syringe (needle diameter: 0.8 mm) into gelation medium containing 3% (w/v) calcium chloride (CaCl2) in saturated boric acid solution. Forming beads were maintained in gelation solution for 24 h at ambient temperature under gentle agitation to ensure complete cross-linking and matrix stabilization through dual mechanisms: (1) borate ester formation with PVA hydroxyl groups, and (2) ionic alginate gelation by calcium ions. Solidified beads were washed three times with sterile distilled water to remove excess CaCl2 and stored at 4 °C until use. Bead characterization and loading standardization Immobilized beads exhibited uniform morphology with average diameter 2.5 ± 0.2 mm. Due to differing cell densities in original cultures, VSS content per bead varied between strains: modified KKU-MC1 beads contained approximately 1.2 ± 0.2 mg-VSS/bead, while HPB beads contained approximately 1.8 ± 0.3 mg-VSS/bead. Both cell types were immobilized using identical protocol. To minimize confounding effects from different initial biomass concentrations between configurations, inoculum volumes were adjusted to achieve approximately equal total VSS across all experimental groups (see detailed Table S2). Both modified KKU-MC1 and HPB free cells were inoculated at 20% (v/v), with stock concentrations of 2.10 ± 0.50% and 2.00 ± 0.75% w/v, respectively, resulting in comparable initial free-cell VSS. This standardization, combined with standardized immobilized bead loading (0.63 ± 0.1 g for HPB beads, 0.89 ± 0.1 g for KKU-MC1 beads), resulted in comparable total initial VSS across configurations: 0.70–0.72% total VSS, enabling direct comparison of immobilization strategy effects independent of absolute biomass loading. Final VSS concentrations in working volumes were: (1) Free modified KKU-MC1: 0.42% (from 20% v/v inoculation), (2) Free HPB: 0.40% (from 20% v/v inoculation), and (3) Immobilized cells: 0.30% for both organisms (based on bead VSS content). The complete hybrid immobilization procedure is illustrated in Fig. 1.Fig. 1Schematic illustration of hybrid immobilization procedure by entrapment-encapsulation technique. The process comprises five main steps: (1) Cell harvest and washing through centrifugation and rinsing with 0.85% NaCl; (2) Polymer matrix preparation by dissolving 4.50% PVA, 1.50% sodium alginate, 0.25% silica, and 0.50% chitosan in hot water; (3) Cell–matrix mixing to achieve homogeneous cell distribution; (4) Bead formation through dropwise extrusion into calcium chloride-boric acid gelation medium enabling dual cross-linking mechanisms; (5) Curing (24 h), washing, and storage at 4 °C. The final product consists of uniform beads with average diameter 2.5 ± 0.2 mm containing immobilized microbial cells. Schematic illustration of hybrid immobilization procedure by entrapment-encapsulation technique. The process comprises five main steps: (1) Cell harvest and washing through centrifugation and rinsing with 0.85% NaCl; (2) Polymer matrix preparation by dissolving 4.50% PVA, 1.50% sodium alginate, 0.25% silica, and 0.50% chitosan in hot water; (3) Cell–matrix mixing to achieve homogeneous cell distribution; (4) Bead formation through dropwise extrusion into calcium chloride-boric acid gelation medium enabling dual cross-linking mechanisms; (5) Curing (24 h), washing, and storage at 4 °C. The final product consists of uniform beads with average diameter 2.5 ± 0.2 mm containing immobilized microbial cells. Hydrogen production by simultaneous saccharification and fermentation (SSF) Bio-HP from agricultural residues (NG and OPF) was investigated using SSF process, wherein enzymatic saccharification and fermentative HP occur concurrently. Two configurations (SSF1 and SSF2) were evaluated as detailed in Table S2. Biomass substrates at loadings of 10, 30, or 50 g-VS/L were added to 100 mL serum bottles. Configuration-specific inoculation protocols were: (1) SSF1: free modified KKU-MC1 (20% v/v, 10 mL) with immobilized HPB (0.63 ± 0.1 g wet weight); (2) SSF2: immobilized modified KKU-MC1 (0.89 ± 0.1 g wet weight) with free HPB (20% v/v, 10 mL). Working volume was adjusted to 50 mL with PCS medium. Initial pH was adjusted to 7.5 using 5 M NaOH or HCl. Bottles were flushed with nitrogen gas for 5 min to establish anaerobic conditions, sealed with butyl rubber stoppers and aluminum crimp caps, and incubated at 37 °C with agitation at 150 rpm for 10 days or until HP ceased. Biogas volume was measured daily using 50 mL gas-tight glass syringes, and composition was analyzed by Gas Chromatography-Thermal Conductivity Detector (GC-TCD). All experiments were conducted in triplicate. Data are presented as mean ± SD (n = 3). Hydrogen production by separate hydrolysis and fermentation (SHF) HP was investigated using a two-stage SHF process comprising sequential enzymatic pre-hydrolysis and fermentation (Table S2). Stage 1: Pre-hydrolysis (5 days) Biomass substrates at loadings of 10, 30, or 50 g-VS/L were added to 100 mL serum bottles. Modified KKU-MC1 was added in configuration-specific formats: (1) free cells at 20% (v/v) (8 mL per bottle), or (2) hybrid-immobilized cells (0.89 ± 0.1 g wet weight). Working volume was adjusted to 40 mL with PCS medium. Initial pH was adjusted to 7.0 using 5 M NaOH or HCl. Bottles were flushed with N₂ for 5 min, sealed with butyl rubber stoppers and aluminum crimp caps, and incubated at 37 °C with 150 rpm agitation for 5 days. Biogas volume was measured daily using gas-tight glass syringes and H₂ content analyzed by GC-TCD; cumulative HY from Stage 1 is reported separately as 1st Pre-hydrolysis HY (mL-H₂/g-VS) in Table 2. Stage 2: Hydrogen fermentation Complete hydrolysate (40 mL, containing both solid and liquid fractions) was inoculated with HPB in configuration-specific formats: (1) free cells at 20% (v/v) (10 mL per bottle), or (2) hybrid-immobilized cells (0.63 ± 0.1 g wet weight). Working volume was adjusted to 50 mL with PCS medium, pH was maintained at 7.0, and bottles were incubated at 37 °C with 150 rpm agitation until biogas production ceased. Biogas volume was measured daily using gas-tight glass syringes and H₂ content analyzed by GC-TCD; cumulative HY from Stage 2 is reported as 2nd Fermentation HY (mL-H₂/g-VS) in Table 2. Total HY represents the sum of both stage contributions. All experiments were conducted in triplicate. Reusability evaluation Three optimal configurations were selected based on results from Sects. "Configuration-dependent performance in mono-digestion" and "Extraordinary synergy in co-digestion systems": SHF1 (10 g-VS/L NG, optimal mono-digestion condition), SHF2 (10 g-VS/L OPF, optimal mono-digestion condition), and SSF2 (7.5:2.5 NG: OPF, optimal co-digestion ratio), and evaluated through five consecutive batch cycles. Following each cycle, spent medium and undigested substrate residues were decanted together while retaining immobilized beads. Beads were rinsed once with sterile distilled water to remove loosely attached debris, then fresh substrate and PCS medium were added to restore original working volume (40–50 mL). Fresh free-cell inoculum was added for the non-immobilized component: HPB at 20% (v/v) for SHF2 and SSF2, or modified KKU-MC1 at 20% (v/v) for SHF1. Initial pH was adjusted to 7.0–7.5, bottles were flushed with N₂, sealed, and incubated as described in Sects. "Hydrogen production by simultaneous saccharification and fermentation (SSF)" and "Hydrogen production by separate hydrolysis and fermentation (SHF)". Key performance indicators monitored included: (1) HY as primary productivity metric, (2) Filter Paperase (FPase) activity using dinitrosalicylic acid (DNS) method (Sect. "Physical and chemical analyses"), and (3) VSS in immobilized beads by drying-combustion method. Physical bead integrity was assessed visually and through diameter measurements. Retention percentages were calculated as ratio of cycle 5 to initial values; decline rates as average percentage decrease per cycle. Statistical analysis used one-way ANOVA with Tukey’s HSD post-hoc test (p < 0.05). Analytical methods Physical and chemical analyses Total solid (TS) and VS of substrates were determined according to APHA32. Cellulose, hemicellulose, and lignin content were analyzed following NREL protocols31. pH was measured using a pH meter (pH-500, Queen, USA). Cell density was determined spectrophotometrically at OD₆₀₀ (UV-1800, Shimadzu, Japan) with 0.85% NaCl as blank. VSS concentration was assessed following APHA32. For immobilized beads, individual beads were dried at 90 °C for 1 h, weighed, combusted at 550 °C for 4 h, washed, and the drying-combustion cycle repeated three times, modified from Jamali et al.33. FPase activity was determined using DNS method34. Whatman No. 1 filter paper (1.0 × 6.0 cm, ~ 50 mg) was incubated with enzyme solution in 0.05 M sodium citrate buffer (pH 4.8) at 50 °C for 60 min. Reducing sugars were quantified at 540 nm using glucose standard. One International Unit (IU) was defined as enzyme releasing 1 μmol reducing sugar per minute. Volatile fatty acids (VFAs), including acetic acid (HAc), propionic acid (HPr), and butyric acid (HBu), as well as lactic acid (HLa), were quantified using High-Performance Liquid Chromatography (HPLC) (Shimadzu LC-20AD, Japan) with refractive index detector and Aminex HPX-87H column (Bio-Rad, USA) at 65 °C with 5 mM H₂SO₄ at 0.6 mL/min23. Biogas volume was measured daily using gas-tight syringes. H2 content was determined by GC-TCD with operating conditions (Shimadzu GC-2014, Japan): Shin Carbon ST 80/100 packed column (2 m × 3 mm i.d.); carrier gas argon at 30 mL/min; injector temperature 100 °C; column temperature 45 °C (isothermal); detector temperature 200 °C; injection volume 0.1 mL via gas-tight syringe23. HY was expressed as mL-H₂/g-VS. Scanning Electron Microscope (SEM) (JSM-6460LV, JEOL, Japan) was performed at 15 kV,samples were gold-coated (~ 15 nm) and examined at 500–15,000 × magnification. Kinetic modeling HP kinetics were analyzed using the modified Gompertz equation35:1\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathrm{H}(\mathrm{t}) = \mathrm{P}\cdot \mathrm{e}\mathrm{x}\mathrm{p}\left\{-\mathrm{e}\mathrm{x}\mathrm{p}\left[\frac{\mathrm{R}\mathrm{m}\mathrm{a}\mathrm{x} \cdot \mathrm{e}}{\mathrm{P}}\left(\uplambda -\mathrm{t}\right)+1\right]\right\}$$\end{document}where H(t) = cumulative HP (mL) at time t (h); P = maximum HP potential (mL); Rmax = maximum HP rate (mL/h); λ = lag phase duration (h); e = Euler’s number (2.718). Daily biogas measurements were used to construct cumulative production curves. Gompertz parameters (P, Rmax, λ) were estimated by non-linear regression using Solver add-in (GRG Nonlinear method) in Microsoft Excel 2019 (Microsoft Corporation, USA). The objective function minimizes sum of squared residuals between observed and predicted cumulative HP values. Model fit quality was assessed by coefficient of determination (R2). HY was calculated by dividing maximum production potential by mass of VS added:2\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathrm{H}\mathrm{Y} (\mathrm{m}\mathrm{L}-\mathrm{H}_2/\mathrm{g}-\mathrm{V}\mathrm{S}) =\frac{\mathrm{P} (\mathrm{m}\mathrm{L}) }{\mathrm{S}\mathrm{u}\mathrm{b}\mathrm{s}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e} \mathrm{V}\mathrm{S} (\mathrm{g})}$$\end{document} Both experimentally observed final yields (HYexp) and Gompertz-predicted yields (calculated from P parameter) are reported for comparison. Economic assessment Energy yield (EY) was calculated as:3\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathrm{E}\mathrm{Y} = \frac{(\mathrm{H}\mathrm{Y}\times \mathrm{H}\mathrm{H}\mathrm{V})}{\mathrm{M}\mathrm{o}\mathrm{l}\mathrm{a}\mathrm{r} \, \mathrm{v}\mathrm{o}\mathrm{l}}$$\end{document} Production cost per unit H2 was calculated based on medium and immobilization material costs, excluding biomass feedstock (valued at zero as agricultural waste), size reduction energy, and facility depreciation8,36:4\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathrm{C}\mathrm{o}\mathrm{s}\mathrm{t} =\frac{\sum (\mathrm{C}\mathrm{N})}{\mathrm{H}}$$\end{document}where C is price of nutrients/raw materials (USD/kg), N is quantity (kg), and H is H2 produced (kg). Medium-based economic yield (MEY) was calculated as ratio of H2 value to input costs:5\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathrm{M}\mathrm{E}\mathrm{Y} =\frac{\mathrm{H}\times \mathrm{I}}{\sum (\mathrm{C}\mathrm{N})}$$\end{document}where I is H2 selling price (USD/kg). H was valued at 6 USD/kg representing conservative mid-to-upper range estimate for renewable H2 based on current market prices (typical range: 4–8 USD/kg;2,37). Chemical prices (industrial grade) were obtained from verified suppliers on www.alibaba.com (accessed November 2024). MEY > 1 indicates economic feasibility. Co-digestion synergy evaluation Co-digestion synergy was evaluated by comparing actual HY from mixed substrates against theoretical additive yields calculated from mono-digestion performance. Theoretical additive yield (TAY) assumes no interaction between substrates:6\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mathrm{TAY}} = \Sigma \left( {{\mathrm{HY}}_{{\mathrm{i}}} \times {\mathrm{w}}_{{\mathrm{i}}} } \right)$$\end{document}where TAY = theoretical additive yield (mL-H₂/g-VS); HYi = hydrogen yield from mono-digestion of substrate i (mL-H₂/g-VS); wi = mass fraction of substrate i in mixture (dimensionless). Synergistic enhancement (SE) quantifies deviation from additive behavior:7\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mathrm{SE}}\left( \% \right) = \left[ {\left( {{\mathrm{HY}}_{{{\mathrm{actual}}}} - {\mathrm{TAY}}} \right)/{\mathrm{TAY}}} \right] \times {1}00$$\end{document}where SE = synergistic enhancement (%); HY_actual = experimentally measured HY from co-digestion (mL-H₂/g-VS). Positive SE values indicate synergistic effects (actual yield exceeds theoretical prediction), while negative values indicate antagonistic effects. Statistical significance of synergy was assessed using one-way ANOVA with p < 0.05 as threshold for significance. For TAY calculations, mono-digestion HY values were selected from the same immobilization configuration and process type (SSF or SHF) at comparable substrate concentrations to ensure meaningful comparison. Statistical analysis Statistical analysis was performed using SPSS Statistics 26.0 (IBM Corporation, USA). All data are presented as mean ± SD from at least three independent biological replicates. Data normality was assessed using Shapiro–Wilk test. One-way ANOVA was employed to evaluate differences between experimental configurations and substrate concentrations. When significant main effects were detected (p < 0.05), Tukey’s HSD post-hoc test was applied for pairwise comparisons. For reusability studies, one-way ANOVA with Tukey’s HSD was used to compare performance across consecutive cycles (initial, Cycles 1–5) for each immobilized system. Different superscript letters in tables indicate statistically significant differences (p < 0.05). Pearson’s correlation coefficient (r) was calculated to assess linear relationships between variables, including HY, enzyme activity, VSS concentration, metabolite composition, and kinetic parameters. Correlation strength was interpreted as: |r|< 0.3 (weak), 0.3–0.7 (moderate), > 0.7 (strong). Statistical significance of correlations was determined at p < 0.05. Results and discussion Performance benchmarking and system characterization The hybrid immobilization system was developed through systematic screening of seven preliminary matrix formulations (Supplementary Table S3). The optimized formulation (4.50% PVA, 1.50% alginate, 0.50% chitosan, 0.25% silica with saturated boric acid + 3% CaCl2 dual gelation) produced consistently spherical beads with mean diameter 2.5 ± 0.2 mm and coefficient of variation 8.0% (n = 50). The dual gelation mechanism—boric acid-mediated PVA network formation via di-diol borate ester complexes and divalent calcium cross-linking of alginate chains into egg-box structures—confers mechanical robustness to the resulting interpenetrating polymer networks20,29. The finalized matrix showed excellent resistance to dissolution, swelling, and mechanical degradation during repeated fermentation over five consecutive 10-day cycles, confirming successful optimization. Cell loading was 1.2 ± 0.2 mg-VSS/bead for modified KKU-MC1 and 1.8 ± 0.3 mg-VSS/bead for HPB—a 1.5-fold difference reflecting inherent biomass yield characteristics. Low relative standard deviations (< 15%) confirmed excellent protocol reproducibility for scale-up considerations. SEM revealed the microstructural architecture of both bead types (Fig. 2). Modified KKU-MC1 beads before fermentation showed a three-dimensional porous network structure conducive to substrate penetration and product diffusion (Fig. 2A). After fermentation, beads exhibited dense microbial colonization with spherical cell aggregates distributed throughout the matrix interior (Fig. 2B, C), confirming successful cell retention and maintained metabolic activity. HPB beads displayed interconnected fibrous network structures before fermentation (Fig. 2D), with roughened surface topology and biofilm formation after fermentation (Fig. 2E-F). These morphological transitions indicate active bio-HP, where microbial proliferation and the secretion of extracellular polymeric substances (EPS) enhance matrix-cell interactions. Importantly, bead integrity was maintained despite extensive colonization, demonstrating well-balanced mechanical properties accommodating cellular activities without compromising structural stability19. No discernible matrix dissolution or structural degradation occurred during five reusability cycles, in contrast to preliminary formulations with 5–10-day stability limitations. Maintaining structural integrity under operational conditions (37 °C, 150 rpm, pH 7.0–7.5) establishes that the hybrid matrix provides sufficient mechanical stability with biological compatibility for HP.Fig. 2Scanning electron microscopy characterization of hybrid immobilized cell beads. SEM images demonstrate morphology and structural characteristics of immobilized modified KKU-MC1 consortium (A–C) and HPB (D–F). Modified KKU-MC1 beads showing matrix structure before fermentation (A) and dense cell colonization after fermentation (B, C). HPB beads display fibrous network structure (D) and biofilm formation after fermentation (E, F). Images captured at 7500 × − 15000 × magnification. Scanning electron microscopy characterization of hybrid immobilized cell beads. SEM images demonstrate morphology and structural characteristics of immobilized modified KKU-MC1 consortium (A–C) and HPB (D–F). Modified KKU-MC1 beads showing matrix structure before fermentation (A) and dense cell colonization after fermentation (B, C). HPB beads display fibrous network structure (D) and biofilm formation after fermentation (E, F). Images captured at 7500 × − 15000 × magnification. Configuration-dependent performance in mono-digestion Cumulative HP profiles for mono-digestion of NG and OPF at 10, 30, and 50 g-VS/L are presented in Fig. 3A–D. Cumulative HP characterizes absolute volumetric production capacity, while HY and kinetics from Tables 1 and 2 serves as the primary substrate-normalized performance metric.Fig. 3Cumulative hydrogen production from mono-digestion of Napier grass (NG) and oil palm fronds (OPF) under SSF configurations (A: SSF1; B: SSF2) and SHF configurations (C: SHF1; D: SHF2) using free cell or hybrid immobilized modified KKU-MC1 or HPB at substrate loadings of 10, 30, and 50 g-VS/L. For SHF configurations (C, D), cumulative H₂ production represents combined daily measurements from Stage 1 (pre-hydrolysis, days 1–5) and Stage 2 (fermentation); individual stage contributions are reported separately in Table 2 as 1st Pre-hydrolysis HY and 2nd Fermentation HY (mL-H₂/g-VS). Error bars represent SD (n = 3).Table 1Hydrogen production performance from mono-digestion substrates using free—and immobilized- modified KKU-MC1 or HPB under SSF configurations.Substrate (g/L)HY_exp (mL-H2/g-VS)HY (mL-H2/g-VS)Rmax (mL/L.h)λ (h)R2SSF1 (Free-modified KKU-MC1 + Hybrid immobilized-HPB)10-NG37.0 ± 6.8ᵃ37.50.3ᵃ8.4ᵃ0.97130-NG30.0 ± 8.0ᵃᵇ28.10.3ᵃ0.0ᵇ0.93750-NG13.8 ± 2.5ᶜ13.00.1ᵇ0.0ᵇ0.84110-OPF6.9 ± 1.0ᵃ7.00.1ᵇ0.0ᵇ0.96330-OPF6.0 ± 0.0ᵃ3.00.1ᵇ21.2ᵃ0.98250-OPF6.7 ± 1.2ᵃ6.70.1ᵇ0.0ᵇ0.969SSF2 (Hybrid immobilized-modified KKU-MC1 + Free-HPB)10-NG5.9 ± 0.5ᶜ5.91.2ᵃ78.6ᵃ0.99930-NG21.3 ± 1.2ᵃ21.10.5ᵇ0.0ᵇ0.96650-NG13.7 ± 1.0ᵇ13.20.6ᵇ3.4ᵇ0.84810-OPF2.0 ± 0.0ᵇ2.00.1c31.7ᵃ0.99930-OPF7.3 ± 1.3ᵃ7.10.3c1.3ᵇ0.91950-OPF7.8 ± 1.5ᵃ7.70.7ᵃ7.4ᵇ0.996NG, Napier grass; OPF, Oil palm fronds. Values represent mean ± SD (n = 3). Different superscript letters within the same substrate type and configuration indicate significant differences (Tukey’s HSD, p < 0.05).Table 2Hydrogen production performance from mono-digestion substrates using free cell and hybrid immobilized cell of modified KKU-MC1 or HPB under SHF configurations.Substrate (g/L)1st Pre-hydrolysis HY_exp (mL-H2/g-VS)2nd Fermentation HY_exp (mL-H2/g-VS)Total HY (mL-H2/g-VS)Gompertz Total HYRmax (mL/L.h)λ (h)R2SHF1 (Free-modified KKU-MC1 + Hybrid immobilized-HPB)10-NG43.0 ± 3.538.5 ± 4.081.5 ± 7.5ᵃ83.80.7ᵃ30.8ᵇ0.96530-NG24.7 ± 2.023.9 ± 2.448.6 ± 4.4ᵇ48.90.3ᵇ0.0ᶜ0.89650-NG15.6 ± 4.016.9 ± 1.532.5 ± 5.5ᶜ33.60.2b0.0ᶜ0.90810-OPF6.0 ± 0.523.2 ± 2.029.2 ± 2.5ᵇ30.90.2ᵇ74.7ᵃ0.98530-OPF20.1 ± 4.020.6 ± 0.040.7 ± 4.0ᵃ41.40.2ᵇ11.6ᵇ0.97450-OPF4.1 ± 0.411.6 ± 1.415.7 ± 1.8ᶜ18.60.1ᶜ40.6ᵇ0.965SHF2 (Hybrid immobilized-modified KKU-MC1 + Free-HPB)10-NG8.1 ± 1.012.0 ± 1.520.1 ± 2.5ᵇ20.90.2ᵃ57.9ᵃ0.96130-NG14.1 ± 2.015.5 ± 1.029.5 ± 3.0ᵃ30.80.2ᵃ13.2ᵇ0.94450-NG8.8 ± 0.013.8 ± 1.422.7 ± 1.4ᵇ24.80.1ᵇ0.0ᶜ0.92810-OPF17.0 ± 1.540.3 ± 2.857.3 ± 4.3ᵃ60.20.4ᵃ61.9ᵃ0.98630-OPF7.4 ± 1.019.5 ± 2.026.9 ± 3.0ᵇ29.80.1ᵇ53.6ᵃ0.98350-OPF4.9 ± 0.45.3 ± 1.510.2 ± 1.9ᶜ10.50.1ᵇ49.8ᵃ0.970NG, Napier grass; OPF, Oil palm fronds. Values represent mean ± SD (n = 3). Different superscript letters within the same substrate type and configuration indicate significant differences (Tukey’s HSD, p < 0.05). Cumulative hydrogen production from mono-digestion of Napier grass (NG) and oil palm fronds (OPF) under SSF configurations (A: SSF1; B: SSF2) and SHF configurations (C: SHF1; D: SHF2) using free cell or hybrid immobilized modified KKU-MC1 or HPB at substrate loadings of 10, 30, and 50 g-VS/L. For SHF configurations (C, D), cumulative H₂ production represents combined daily measurements from Stage 1 (pre-hydrolysis, days 1–5) and Stage 2 (fermentation); individual stage contributions are reported separately in Table 2 as 1st Pre-hydrolysis HY and 2nd Fermentation HY (mL-H₂/g-VS). Error bars represent SD (n = 3). Table 1Hydrogen production performance from mono-digestion substrates using free—and immobilized- modified KKU-MC1 or HPB under SSF configurations.Substrate (g/L)HY_exp (mL-H2/g-VS)HY (mL-H2/g-VS)Rmax (mL/L.h)λ (h)R2SSF1 (Free-modified KKU-MC1 + Hybrid immobilized-HPB)10-NG37.0 ± 6.8ᵃ37.50.3ᵃ8.4ᵃ0.97130-NG30.0 ± 8.0ᵃᵇ28.10.3ᵃ0.0ᵇ0.93750-NG13.8 ± 2.5ᶜ13.00.1ᵇ0.0ᵇ0.84110-OPF6.9 ± 1.0ᵃ7.00.1ᵇ0.0ᵇ0.96330-OPF6.0 ± 0.0ᵃ3.00.1ᵇ21.2ᵃ0.98250-OPF6.7 ± 1.2ᵃ6.70.1ᵇ0.0ᵇ0.969SSF2 (Hybrid immobilized-modified KKU-MC1 + Free-HPB)10-NG5.9 ± 0.5ᶜ5.91.2ᵃ78.6ᵃ0.99930-NG21.3 ± 1.2ᵃ21.10.5ᵇ0.0ᵇ0.96650-NG13.7 ± 1.0ᵇ13.20.6ᵇ3.4ᵇ0.84810-OPF2.0 ± 0.0ᵇ2.00.1c31.7ᵃ0.99930-OPF7.3 ± 1.3ᵃ7.10.3c1.3ᵇ0.91950-OPF7.8 ± 1.5ᵃ7.70.7ᵃ7.4ᵇ0.996NG, Napier grass; OPF, Oil palm fronds. Values represent mean ± SD (n = 3). Different superscript letters within the same substrate type and configuration indicate significant differences (Tukey’s HSD, p < 0.05). Hydrogen production performance from mono-digestion substrates using free—and immobilized- modified KKU-MC1 or HPB under SSF configurations. NG, Napier grass; OPF, Oil palm fronds. Values represent mean ± SD (n = 3). Different superscript letters within the same substrate type and configuration indicate significant differences (Tukey’s HSD, p < 0.05). Table 2Hydrogen production performance from mono-digestion substrates using free cell and hybrid immobilized cell of modified KKU-MC1 or HPB under SHF configurations.Substrate (g/L)1st Pre-hydrolysis HY_exp (mL-H2/g-VS)2nd Fermentation HY_exp (mL-H2/g-VS)Total HY (mL-H2/g-VS)Gompertz Total HYRmax (mL/L.h)λ (h)R2SHF1 (Free-modified KKU-MC1 + Hybrid immobilized-HPB)10-NG43.0 ± 3.538.5 ± 4.081.5 ± 7.5ᵃ83.80.7ᵃ30.8ᵇ0.96530-NG24.7 ± 2.023.9 ± 2.448.6 ± 4.4ᵇ48.90.3ᵇ0.0ᶜ0.89650-NG15.6 ± 4.016.9 ± 1.532.5 ± 5.5ᶜ33.60.2b0.0ᶜ0.90810-OPF6.0 ± 0.523.2 ± 2.029.2 ± 2.5ᵇ30.90.2ᵇ74.7ᵃ0.98530-OPF20.1 ± 4.020.6 ± 0.040.7 ± 4.0ᵃ41.40.2ᵇ11.6ᵇ0.97450-OPF4.1 ± 0.411.6 ± 1.415.7 ± 1.8ᶜ18.60.1ᶜ40.6ᵇ0.965SHF2 (Hybrid immobilized-modified KKU-MC1 + Free-HPB)10-NG8.1 ± 1.012.0 ± 1.520.1 ± 2.5ᵇ20.90.2ᵃ57.9ᵃ0.96130-NG14.1 ± 2.015.5 ± 1.029.5 ± 3.0ᵃ30.80.2ᵃ13.2ᵇ0.94450-NG8.8 ± 0.013.8 ± 1.422.7 ± 1.4ᵇ24.80.1ᵇ0.0ᶜ0.92810-OPF17.0 ± 1.540.3 ± 2.857.3 ± 4.3ᵃ60.20.4ᵃ61.9ᵃ0.98630-OPF7.4 ± 1.019.5 ± 2.026.9 ± 3.0ᵇ29.80.1ᵇ53.6ᵃ0.98350-OPF4.9 ± 0.45.3 ± 1.510.2 ± 1.9ᶜ10.50.1ᵇ49.8ᵃ0.970NG, Napier grass; OPF, Oil palm fronds. Values represent mean ± SD (n = 3). Different superscript letters within the same substrate type and configuration indicate significant differences (Tukey’s HSD, p < 0.05). Hydrogen production performance from mono-digestion substrates using free cell and hybrid immobilized cell of modified KKU-MC1 or HPB under SHF configurations. NG, Napier grass; OPF, Oil palm fronds. Values represent mean ± SD (n = 3). Different superscript letters within the same substrate type and configuration indicate significant differences (Tukey’s HSD, p < 0.05). Hydrogen production performance under SSF configurations HP from mono-digestion under SSF configurations exhibited significant disparities among substrates, cell immobilization strategies, and substrate loading rates (Fig. 3A, B). Statistical analysis indicated substrate type and cell immobilization format significantly impacted HYs (p < 0.001), while kinetic parameters Rmax and λ varied with substrate (Table 1). SSF1 configuration (free modified KKU-MC1 with immobilized HPB) NG showed substrate loading-dependent performance. Highest yield at 10 g-VS/L was 37.0 ± 6.8 mL-H₂/g-VS. This significantly decreased with increasing substrate concentration to 30.0 ± 8.0 mL-H₂/g-VS at 30 g-VS/L (p < 0.05) and 13.8 ± 2.5 mL-H₂/g-VS at 50 g-VS/L (p < 0.001). Rmax for NG decreased from 0.3 mL/L·h at 10 and 30 g-VS/L to 0.1 mL/L·h at 50 g-VS/L (p < 0.05), while λ was shortest at 10 g-VS/L (8.4 h) compared to minimal or zero λ at higher loadings (p < 0.05). This substrate loading effect agrees with observations that increasing substrate concentrations cause inhibition by accumulated VFAs (HAc, HPr, HBu and HLa) and elevated H₂ partial pressure, alongside mass transfer limitations in immobilized systems. OPF consistently produced lower yields across the concentration range (6.0–6.9 mL-H2/g-VS), with no significant differences observed among loading rates (p > 0.05). This reflects OPF’s intrinsic recalcitrance, which stems not solely from absolute lignin content (23.0 ± 7.2%, comparable to NG at 14.6 ± 8.0%, Table S1), but primarily from its distinct structural heterogeneity between rachis and leaflet fractions9,11. OPF consists of two main components—the rachis (woody petiole) and the leaflets—which differ significantly in their physical structure, composition, and recalcitrance. The rachis represents the woody fraction with higher recalcitrance, making it more difficult to decompose than the leaflets, which contain higher protein and lower fiber content38. This compositional heterogeneity, combined with variations attributable to biomass origin, tree age, and geographical factors39, contributes to the wide range of lignin values reported in literature (15–30%). However, Rmax showed significant differences: 10–50 g-VS/L remained at 0.1 mL/L·h. For OPF, λ significantly varied–30 g-VS/L resulted in the longest λ (21.2 h, p < 0.05), whereas 10 and 50 g-VS/L showed minimal λ. SSF2 configuration (immobilized modified KKU-MC1 with free HPB) The SSF2 configuration, comprising immobilized modified KKU-MC1 and free HPB, exhibited distinct performance patterns compared to SSF1 (Fig. 3B). For NG, HYs increased significantly from 5.9 ± 0.5 mL-H2/g-VS at 10 g-VS/L to 21.3 ± 1.2 mL-H2/g-VS at 30 g-VS/L (p < 0.001), before declining to 13.7 ± 1.0 mL-H2/g-VS at 50 g-VS/L. This concentration-dependent behavior was notably opposite to the trends observed in SSF1 (Table 1). Kinetic parameters showed highest Rmax at 10 g-VS/L (1.2 mL/L·h) but exceptionally long λ (78.6 h), while higher loadings showed lower Rmax (0.5–0.6 mL/L·h) but minimal λ—suggesting immobilized cells require extended adaptation at low concentrations20. In contrast, the performance of OPF in the SSF2 configuration showed minimal variation, with yields ranging from 2.0 to 7.8 mL-H2/g-VS. However, Rmax increased significantly with substrate loading, rising from 0.1 mL/L⋅h at 10 g-VS/L to 0.7 mL/L⋅h at 50 g-VS/L (p < 0.05). Simultaneously, λ decreased from 31.7 h at 10 g-VS/L to 1.3–7.4 h at higher concentrations (p < 0.05). These results suggest that when cellulolytic organisms are immobilized, the system achieves better kinetic efficiency and faster initiation at higher OPF concentrations. Gompertz kinetic modeling provided an excellent fit to experimental data (R2 > 0.84 for all conditions), confirming model appropriateness for describing HP kinetics from lignocellulosic substrates. Kinetic parameters Rmax and λ varied significantly by configuration and substrate. SSF1 generally exhibited shorter λ with substrate loading-dependent production rates, while SSF2 showed longer adaptation periods particularly at low substrate concentrations but concentration-dependent improvements in kinetic parameters for OPF. Statistical comparison between SSF1 and SSF2 revealed significant configuration-substrate interactions (p < 0.001). SSF1 performed better for NG at optimal loading (10 g-VS/L: 37.0 vs. 5.9 mL-H₂/g-VS), while both configurations showed relatively poor OPF utilization, although SSF2 performed better at higher OPF loadings. These findings highlight necessity for selecting appropriate process configurations matching substrate characteristics and immobilization strategy for bio-HP system optimization. Hydrogen production performance under SHF configurations SHF exhibited different performance profiles compared with SSF, with significant impacts of process configuration on HP and HY (Table 2, Fig. 3C, D). The two-stage process allowed optimization of enzymatic hydrolysis and H₂ fermentation in stages, although effectiveness varied among substrate types and immobilization strategies. SHF1 configuration (free modified KKU-MC1 with immobilized HPB) Cumulative HP profiles across substrate types and loadings are presented in Fig. 3C. The SHF1 configuration achieved a peak HY of 81.5 ± 7.5 mL-H₂/g-VS using NG at 10 g-VS/L. This performance was characterized by balanced metabolic contributions from the pre-hydrolysis (43.0 ± 3.5 mL-H₂/g-VS) and fermentation stages (38.5 ± 4.0 mL-H₂/g-VS), supported by an Rmax of 0.7 mL/L⋅h (p < 0.001). However, increasing the substrate loading to 50 g-VS/L induced significant substrate inhibition, as evidenced by a 60% reduction in total yield and a decrease in Rmax to 0.2 mL/L⋅h (p < 0.001). Notably, the 2.2-fold yield increase in SHF1 compared to the matched SSF1 configuration suggests that decoupling the stages alleviates metabolic crosstalk and feedback inhibition. This separation allows the cellulolytic enzymes to act on the porous NG fibers without interference from high H2 partial pressures or inhibitory VFAs generated during active fermentation. In contrast, OPF reached maximum yield (40.7 ± 4.0 mL-H₂/g-VS) (Rmax 0.2 mL/L·h, λ 11.6 h, p < 0.05) at 30 g-VS/L, where increased microbe-to-fiber contact compensated for substrate recalcitrance. The fermentation stage dominated over pre-hydrolysis, suggesting the initial phase functioned primarily as bioliquefaction facilitating subsequent oligosaccharide conversion. SHF2 configuration (immobilized modified KKU-MC1 with free HPB) The SHF2 configuration (immobilized KKU-MC1 and free HPB) exhibited substrate-specific advantages, particularly for the more recalcitrant OPF (Fig. 3D). SHF2-OPF achieved a maximum yield of 57.3 ± 4.3 mL-H2/g-VS at 10 g-VS/L, the highest recorded performance for OPF across all experimental configurations. This yield nearly doubled that of the SHF1-OPF trial (29.2 mL-H2/g-VS, p < 0.001), primarily driven by exceptional productivity during the fermentation stage (40.3 ± 2.8 mL-H2/g-VS). This suggests that the immobilized cellulolytic consortium facilitated sustained enzymatic release during pre-hydrolysis, effectively liberating fermentable sugars for subsequent conversion by free-cell HPB. These findings align with observations that immobilization matrices protect cellulolytic activity21. However, Rmax significantly decreased at higher OPF loadings (0.1 mL/L⋅h at 30–50 g-VS/L, p < 0.05), with λ remaining elevated between 49.8 and 61.9 h, indicating persistent mass-transfer limitations at higher solid concentrations. Conversely, the SHF2 configuration was less effective for NG substrates. At 10 g-VS/L, NG yielded only 20.1 ± 2.5 mL-H2/g-VS, a significant decrease compared to the SHF1-NG trial (p < 0.001). Performance peaked at 30 g-VS/L (29.5 ± 3.0 mL-H2/g-VS) with a notable reduction in λ from 57.9 h to 13.2 h (p < 0.05). This kinetic shift suggests that moderate substrate loadings may be required to maintain optimal enzyme induction levels within the immobilization matrix when processing highly accessible substrates like NG. SHF configurations generally exhibited longer λ values (11.6 to 74.7 h) than SSF configurations, reflecting the time-intensive nature of sequential hydrolysis and the inherent recalcitrance of the biomass. Rmax values varied between 0.1 and 0.7 mL/L⋅h, with the highest rate observed for SHF1-NG at 10 g-VS/L. Statistical analysis confirmed that both Rmax and λ were significantly influenced by substrate loading (p < 0.05; Table 2). Across all trials, the modified Gompertz equation showed excellent fit (R2 > 0.90), validating the derived kinetic parameters as reliable indicators of system performance. Among all SHF conditions evaluated, SHF1 at 10 g-VS/L NG achieved the highest total HY of 81.5 ± 7.5 mL-H₂/g-VS, representing the peak HY across all mono-digestion configurations tested in this study and 2.2-fold higher than the best SSF configuration (SSF1 at 10 g-VS/L NG: 37.0 ± 6.8 mL-H₂/g-VS). The two-stage SHF process provided superior performance for the more recalcitrant OPF substrate, as demonstrated by the 7.4-fold higher HY achieved by SHF2-OPF (57.3 mL-H₂/g-VS) compared to SSF2-OPF (7.8 mL-H₂/g-VS) in the present study, consistent with previous reports that temporal separation of enzymatic hydrolysis and fermentation alleviates product inhibition and allows independent optimization of each stage8,13,23. A configuration with both modified KKU-MC1 and HPB simultaneously immobilized was not evaluated in this study, as the primary objective was to systematically determine whether the location of immobilization of cellulolytic versus fermentative organisms governs performance,dual-immobilization of both organisms is identified as an important direction for future investigation. Extraordinary synergy in co-digestion systems Co-digestion performance was evaluated at NG:OPF ratios of 2.5:7.5, 5.0:5.0, and 7.5:2.5 (g-VS/L:g-VS/L) at a total loading of 10 g-VS/L across all four configurations. Cumulative HP from Fig. 4A–D characterizes absolute volumetric production capacity, while HY and kinetics from Tables 3 and 4 serves as the primary substrate-normalized performance metric. Synergistic enhancement (SE, %) quantifies the deviation of actual HY from the TAY calculated from mono-digestion data at equivalent loadings (Sect. "Co-digestion synergy evaluation", Fig. 5).Fig. 4Cumulative hydrogen production from co-digestion of NG and OPF at ratios of 2.5:7.5, 5.0:5.0, and 7.5:2.5 (g-VS/L:g-VS/L) under SSF configurations (A: SSF1; B: SSF2) and SHF configurations (C: SHF1; D: SHF2) using free cell or hybrid immobilized modified KKU-MC1 or HPB. For SHF configurations (C, D), cumulative H₂ production represents combined daily measurements from Stage 1 (pre-hydrolysis, days 1–5) and Stage 2 (fermentation); individual stage contributions are reported separately in Table 4 as 1st Pre-hydrolysis HY and 2nd Fermentation HY (mL-H₂/g-VS). Error bars represent SD (n = 3).Table 3Hydrogen production performance from co-digestion substrates using free cell and hybrid immobilized cell of modified KKU-MC1 or HPB under SSF configurations.Substrate ratio (g-VS/L:g-VS/L)HY_exp (mL-H2/g-VS)Gompertz HY (mL-H2/g-VS)Rmax (mL/L.h)λ (h)R2SSF1 (Free-modified KKU-MC1 + Hybrid immobilized-HPB)2.5NG + 7.5OPF19.6 ± 5.0ᵇ20.30.2ᵇ32.2ᵃ0.9945.0NG + 5.0OPF38.9 ± 9.5ᵃ40.10.4ᵃ28.0ᵃ0.9947.5NG + 2.5OPF33.6 ± 7.1ᵃᵇ34.80.3ᵃᵇ32.3ᵃ0.994SSF2 (Hybrid immobilized-modified KKU-MC1 + Free-HPB)2.5NG + 7.5OPF43.7 ± 4.5ᵇ44.60.5ᵇ15.4ᵇ0.9925.0NG + 5.0OPF58.6 ± 6.4ᵃ59.50.9ᵃ23.6ᵃ0.9917.5NG + 2.5OPF61.6 ± 7.4ᵃ62.30.9ᵃ16.7ᵇ0.990NG, Napier grass; OPF, Oil palm fronds. Values represent mean ± SD (n = 3). Different superscript letters within the same substrate type and configuration indicate significant differences (Tukey’s HSD, p < 0.05).Table 4Hydrogen production performance from co-digestion substrates using free cell and hybrid immobilized modified KKU-MC1 or HPB under SHF configurations.Substrate ratio1st Pre-hydrolysis (mL-H2/g-VS)2nd Fermentation (mL-H2/g-VS)Total HY (mL-H2/g-VS)Gompertz Total HY (mL-H2/g-VS)Rmax (mL/L.h)λ (h)R2SHF1 (Free-modified KKU-MC1 + Hybrid immobilized-HPB)2.5NG + 7.5OPF20.5 ± 2.528.3 ± 2.148.8 ± 4.6ᵃ47.70.3ᵃ33.8ᵇ0.9815.0NG + 5.0OPF12.6 ± 1.228.9 ± 2.541.4 ± 3.7ᵇ39.50.3ᵃ61.3ᵃ0.9787.5NG + 2.5OPF2.5 ± 0.01.1 ± 0.03.6 ± 0.0ᶜ3.60.1b1.6ᶜ0.986SHF2 (Hybrid immobilized-modified KKU-MC1 + Free-HPB)2.5NG + 7.5OPF3.6 ± 1.12.5 ± 0.16.2 ± 1.2ᵇ5.80.1ᵃ42.2ᵇ0.9925.0NG + 5.0OPF2.4 ± 0.016.5 ± 1.118.9 ± 1.1ᵃ19.80.1ᵃ70.4ᵃ0.9957.5NG + 2.5OPF4.2 ± 1.20.9 ± 0.15.1 ± 1.3ᵇ4.90.1ᵃ36.6ᵇ0.993NG, Napier grass; OPF, Oil palm fronds. Values represent mean ± SD (n = 3). Different superscript letters within the same substrate type and configuration indicate significant differences (Tukey’s HSD, p < 0.05).Fig. 5 Synergy analysis of co-digestion at different NG:OPF ratios across immobilization configurations. (A) SSF1: free modified KKU-MC1 with hybrid immobilized HPB; (B) SSF2: hybrid immobilized modified KKU-MC1 with free HPB; (C) SHF1: free modified KKU-MC1 with hybrid immobilized HPB; (D) SHF2: hybrid immobilized modified KKU-MC1 with free HPB. Gray bars represent theoretical additive yield (TAY) calculated from mono-digestion baselines using Eq. (6); colored bars represent actual HY (mean ± SD, n = 3); red line with markers indicates synergistic enhancement (SE, %) calculated using Eq. (7). Positive SE values indicate synergistic effects; negative values indicate antagonistic effects. Cumulative hydrogen production from co-digestion of NG and OPF at ratios of 2.5:7.5, 5.0:5.0, and 7.5:2.5 (g-VS/L:g-VS/L) under SSF configurations (A: SSF1; B: SSF2) and SHF configurations (C: SHF1; D: SHF2) using free cell or hybrid immobilized modified KKU-MC1 or HPB. For SHF configurations (C, D), cumulative H₂ production represents combined daily measurements from Stage 1 (pre-hydrolysis, days 1–5) and Stage 2 (fermentation); individual stage contributions are reported separately in Table 4 as 1st Pre-hydrolysis HY and 2nd Fermentation HY (mL-H₂/g-VS). Error bars represent SD (n = 3). Table 3Hydrogen production performance from co-digestion substrates using free cell and hybrid immobilized cell of modified KKU-MC1 or HPB under SSF configurations.Substrate ratio (g-VS/L:g-VS/L)HY_exp (mL-H2/g-VS)Gompertz HY (mL-H2/g-VS)Rmax (mL/L.h)λ (h)R2SSF1 (Free-modified KKU-MC1 + Hybrid immobilized-HPB)2.5NG + 7.5OPF19.6 ± 5.0ᵇ20.30.2ᵇ32.2ᵃ0.9945.0NG + 5.0OPF38.9 ± 9.5ᵃ40.10.4ᵃ28.0ᵃ0.9947.5NG + 2.5OPF33.6 ± 7.1ᵃᵇ34.80.3ᵃᵇ32.3ᵃ0.994SSF2 (Hybrid immobilized-modified KKU-MC1 + Free-HPB)2.5NG + 7.5OPF43.7 ± 4.5ᵇ44.60.5ᵇ15.4ᵇ0.9925.0NG + 5.0OPF58.6 ± 6.4ᵃ59.50.9ᵃ23.6ᵃ0.9917.5NG + 2.5OPF61.6 ± 7.4ᵃ62.30.9ᵃ16.7ᵇ0.990NG, Napier grass; OPF, Oil palm fronds. Values represent mean ± SD (n = 3). Different superscript letters within the same substrate type and configuration indicate significant differences (Tukey’s HSD, p < 0.05). Hydrogen production performance from co-digestion substrates using free cell and hybrid immobilized cell of modified KKU-MC1 or HPB under SSF configurations. NG, Napier grass; OPF, Oil palm fronds. Values represent mean ± SD (n = 3). Different superscript letters within the same substrate type and configuration indicate significant differences (Tukey’s HSD, p < 0.05). Table 4Hydrogen production performance from co-digestion substrates using free cell and hybrid immobilized modified KKU-MC1 or HPB under SHF configurations.Substrate ratio1st Pre-hydrolysis (mL-H2/g-VS)2nd Fermentation (mL-H2/g-VS)Total HY (mL-H2/g-VS)Gompertz Total HY (mL-H2/g-VS)Rmax (mL/L.h)λ (h)R2SHF1 (Free-modified KKU-MC1 + Hybrid immobilized-HPB)2.5NG + 7.5OPF20.5 ± 2.528.3 ± 2.148.8 ± 4.6ᵃ47.70.3ᵃ33.8ᵇ0.9815.0NG + 5.0OPF12.6 ± 1.228.9 ± 2.541.4 ± 3.7ᵇ39.50.3ᵃ61.3ᵃ0.9787.5NG + 2.5OPF2.5 ± 0.01.1 ± 0.03.6 ± 0.0ᶜ3.60.1b1.6ᶜ0.986SHF2 (Hybrid immobilized-modified KKU-MC1 + Free-HPB)2.5NG + 7.5OPF3.6 ± 1.12.5 ± 0.16.2 ± 1.2ᵇ5.80.1ᵃ42.2ᵇ0.9925.0NG + 5.0OPF2.4 ± 0.016.5 ± 1.118.9 ± 1.1ᵃ19.80.1ᵃ70.4ᵃ0.9957.5NG + 2.5OPF4.2 ± 1.20.9 ± 0.15.1 ± 1.3ᵇ4.90.1ᵃ36.6ᵇ0.993NG, Napier grass; OPF, Oil palm fronds. Values represent mean ± SD (n = 3). Different superscript letters within the same substrate type and configuration indicate significant differences (Tukey’s HSD, p < 0.05). Hydrogen production performance from co-digestion substrates using free cell and hybrid immobilized modified KKU-MC1 or HPB under SHF configurations. NG, Napier grass; OPF, Oil palm fronds. Values represent mean ± SD (n = 3). Different superscript letters within the same substrate type and configuration indicate significant differences (Tukey’s HSD, p < 0.05). Synergy analysis of co-digestion at different NG:OPF ratios across immobilization configurations. (A) SSF1: free modified KKU-MC1 with hybrid immobilized HPB; (B) SSF2: hybrid immobilized modified KKU-MC1 with free HPB; (C) SHF1: free modified KKU-MC1 with hybrid immobilized HPB; (D) SHF2: hybrid immobilized modified KKU-MC1 with free HPB. Gray bars represent theoretical additive yield (TAY) calculated from mono-digestion baselines using Eq. (6); colored bars represent actual HY (mean ± SD, n = 3); red line with markers indicates synergistic enhancement (SE, %) calculated using Eq. (7). Positive SE values indicate synergistic effects; negative values indicate antagonistic effects. Hydrogen production performance under SSF configurations HP from NG:OPF co-digestion under SSF configurations showed clear ratio-dependent performance and significant differences configurations (Fig. 4A, B, Table 3). The substrate ratio and immobilization strategy both significantly impacted HP in co-digestion with configuration-ratio interaction critically in determining optimal mixing proportion. SSF1 configuration performance (free modified KKU-MC1 with immobilized HPB) The 5.0:5.0 NG:OPF ratio achieved the highest yield of 38.9 ± 9.5 mL-H₂/g-VS with the fastest production rate (Rmax 0.4 mL/L·h) and moderate λ (28.0 h), significantly outperforming the 2.5:7.5 ratio (19.6 ± 5.0 mL-H₂/g-VS, Rmax 0.2 mL/L·h, λ 32.2 h, p < 0.01) (Fig. 4A). Although the 7.5:2.5 ratio achieved a comparable yield (33.6 ± 7.1 mL-H₂/g-VS, Rmax 0.3 mL/L·h, λ 32.3 h, p > 0.05 vs. 5.0:5.0), the balanced mixture demonstrated superior overall performance. The λ values showed no significant differences among ratios (28.0–32.3 h, p > 0.05), suggesting similar adaptation requirements for untreated co-substrates. The superior performance at the 5.0:5.0 ratio reflects optimal C/N balance; NG has a C/N ratio of approximately 10040 and OPF approximately 40–1209,41, yielding a blended C/N of approximately 70–110 at the 5.0:5.0 ratio, within the range considered favorable for fermentative HP. NG typically contains approximately 45% C, 5.9% H, and 0.45% N (dry basis)40, resulting in a high C/N ratio favorable for fermentation but potentially N-limiting mono-digestion. In contrast, OPF exhibits the highest N content among oil palm biomass fractions with approximately 43–48% C, 5–10% H, and 0.4–12% N (dry basis)9,41. The complementary C/N ratios between NG (high C/N and OPF (lower C/N likely optimize microbial nitrogen nutrition during co-digestion, supporting enhanced cellulolytic and fermentative activities. Furthermore, the trace element complementarity provided by NG (382.0 mg/kg Fe and OPF (238.0 mg/kg Fe which could support robust [FeFe]-hydrogenase activity8. Gompertz kinetic modeling showed excellent fit (R2 > 0.990) with production rates of 0.2–0.4 mL/L·h and moderate λ (28.0–32.3 h), indicating faster adaptation to mixed substrate conditions compared to some mono-digestion scenarios. The 5.0:5.0 ratio demonstrated the highest synergistic enhancement (SE = + 77%,p < 0.01), with actual yields substantially exceeding the theoretical additive predictions (TAY = 22.0 mL-H2/g-VS) (Fig. 5A). Positive synergy was also observed at the 2.5:7.5 (SE = + 36%) and 7.5:2.5 ratios (SE = + 14%). This synergy arises from the integration of NG’s readily fermentable carbohydrates with the structural diversity and mineral supplementation of OPF. An equal proportion appears to optimize microbial community dynamics by preventing substrate-induced inhibition while maintaining sufficient carbon flux for sustained hydrogen evolution. SSF2 configuration performance (immobilized modified KKU-MC1 with free HPB) This configuration showed distinctly different co-digestion patterns. The 7.5:2.5 NG:OPF ratio reached highest yield of 61.6 ± 7.4 mL-H2/g-VS (Rmax 0.9 mL/L·h, λ 16.7 h), significantly higher than both 5.0:5.0 (58.6 ± 6.4 mL-H2/g-VS, Rmax 0.9 mL/L·h, λ 23.6 h, p > 0.05) and 2.5:7.5 ratios (43.7 ± 4.5 mL-H₂/g-VS, Rmax 0.5 mL/L·h, λ 15.4 h, p < 0.01) (Fig. 4B). The 5.0:5.0 and 7.5:2.5 ratios exhibited comparably high production rates (0.9 mL/L·h, p > 0.05), significantly higher than 2.5:7.5 ratio (0.5 mL/L·h, p < 0.05). Notably, 7.5:2.5 ratio exhibited shortest λ among high-performing ratios (16.7 h vs. 23.6 h for 5.0:5.0, p < 0.05). This shift in optimal ratio compared to SSF1 (p < 0.001) suggests that immobilizing cellulolytic organisms favors NG-dominant compositions through enhanced enzymatic accessibility within the matrix. Synergy evaluation for SSF2: All three co-digestion ratios exhibited extraordinary synergistic enhancement in SSF2 (Fig. 5B). The 5.0:5.0 ratio achieved the highest SE (+ 1,384%), followed by the 2.5:7.5 ratio (SE = + 1,369%) and the 7.5:2.5 ratio (SE = + 1,150%). These remarkable values reflect configuration-specific responses rather than artifacts, as SSF2 mono-digestion at equivalent concentrations showed constrained performance (TAY = 3.0–4.9 mL-H₂/g-VS) due to mass transfer limitations of immobilized cellulolytic cells. Co-digestion effectively overcame these constraints, with actual yields reaching 43.7–61.6 mL-H₂/g-VS through enhanced substrate accessibility. Co-digestion overcomes mass transfer limitations in immobilized systems through three specific mechanisms: (1) OPF’s different fiber structure disrupts dense substrate packing around beads, creating additional diffusion channels for enzymes and soluble sugars; (2) complementary mineral profiles of NG (Fe 382 mg/kg) and OPF (Fe 238 mg/kg) reduce nutrient depletion gradients near immobilized cells; and (3) the mixed carbohydrate pool generates a broader spectrum of short-chain oligosaccharides that diffuse through the polymer matrix more readily than mono-substrate products. Although the 7.5:2.5 ratio achieved the highest absolute yield (61.6 mL-H₂/g-VS), the 5.0:5.0 ratio showed the greatest synergistic enhancement relative to theoretical predictions. This finding demonstrates that immobilization does not inherently diminish co-digestion benefits; rather, appropriate substrate formulation can amplify synergistic effects through improved spatial coordination of hydrolytic and fermentative activities15. Configuration-specific synergy mechanisms: Contrasting optimal ratios between SSF1 (5.0:5.0) and SSF2 (7.5:2.5) reveal that synergy magnitude and substrate preference are strongly dependent on immobilization configuration (p < 0.001). When cellulolytic organisms remain free (SSF1), balanced mixing optimizes enzyme–substrate contact and nutrient distribution. Conversely, immobilization of KKU-MC1 (SSF2) shifts optimum toward NG-dominance, likely because easily hydrolysable substrates compensate for diffusion limitations within immobilization matrix while maintaining sufficient metabolic activity for HP. The superior co-digestion performance of SSF2 over SSF1 contrasts with the mono-digestion results, reflecting that the immobilization matrix creates protective microenvironments that shield cellulolytic cells from product inhibition—an advantage that becomes significant when diverse, complementary substrates are simultaneously available. Hydrogen production performance under SHF configurations SHF of co-digested substrates showed markedly different performance characteristics than SSF, with significant process-ratio interactions (Fig. 4C, D, Table 4). The two-step approach enabled separate evaluation of pre-hydrolysis and fermentation contributions, providing insights into synergistic mechanisms and configuration-specific limitations. SHF1 configuration performance (free modified KKU-MC1 with immobilized HPB) This configuration showed lowest HP among co-digestion patterns (Fig. 4C). Highest total yield of 48.8 ± 4.6 mL-H₂/g-VS (Rmax 0.3 mL/L·h, λ 33.8 h) was achieved by 2.5:7.5 NG:OPF ratio, with pre-hydrolysis and fermentation stages contributing balanced amounts of 20.5 ± 2.5 mL-H₂/g-VS and 28.3 ± 2.1 mL-H₂/g-VS, respectively, showing significant improvement over 5.0:5.0 at 41.4 ± 3.7 mL-H₂/g-VS (Rmax 0.3 mL/L·h, λ 61.3 h), and far above 7.5:2.5 ratio which resulted in severely impaired performance of 3.6 ± 0.0 mL-H₂/g-VS (Rmax 0.1 mL/L·h, λ 1.6 h). Rates for 2.5:7.5 and 5.0:5.0 were identical (0.3 mL/L·h); however, at 5.0:5.0 ratio, λ was significantly longer (61.3 h vs. 33.8 h), revealing poor adaptation at balanced proportions. Conversely, SHF1 declined sharply at high NG proportions, supported by near-zero Rmax in 7.5:2.5 (p < 0.001) with λ disappearing within 1.6 h, suggesting excessively high cellulosic content surpasses pre-hydrolysis stage capacity, causing incomplete saccharification followed by fermentation failure. This evidence contrasts with SSF performance, underlining sensitivity of sequential processing to substrate composition. The OPF-dominant ratio (2.5:7.5) was the only co-digestion condition achieving positive synergy in SHF1 (SE = + 15%, p < 0.05), with actual yields (48.8 mL-H₂/g-VS) modestly exceeding theoretical predictions (TAY = 42.3 mL-H₂/g-VS) (Fig. 5C). In contrast, the 5.0:5.0 ratio showed antagonistic effects (SE = − 25%), with actual yields (41.4 mL-H₂/g-VS) falling below theoretical expectations (TAY = 55.4 mL-H₂/g-VS). The 7.5:2.5 ratio exhibited severely antagonistic effects (SE = − 95%), with actual yields (3.6 mL-H₂/g-VS) dramatically lower than theoretical predictions (TAY = 68.4 mL-H₂/g-VS). This pattern indicates that sequential processing with free cellulolytic organisms cannot efficiently handle NG-dominant mixtures; excessively high cellulosic content surpasses pre-hydrolysis stage capacity, causing incomplete saccharification followed by fermentation failure. OPF-dominant mixtures provided sufficient hydrolysis products during pre-treatment while avoiding the bottlenecks observed with higher NG proportions. SHF2 configuration performance (immobilized modified KKU-MC1 with free HPB) This configuration showed limited performance across all ratios (Fig. 4D). The 5.0:5.0 ratio achieved highest yield of 18.9 ± 1.1 mL-H₂/g-VS (Rmax 0.1 mL/L·h, λ 70.4 h), significantly higher than 2.5:7.5 (6.2 ± 1.2 mL-H₂/g-VS, Rmax 0.1 mL/L·h, λ 42.2 h) and 7.5:2.5 ratios (5.1 ± 1.3 mL-H₂/g-VS, Rmax 0.1 mL/L·h, λ 36.6 h). No positive synergy was observed in SHF2; actual yields consistently fell below theoretical predictions across all ratios tested, indicating antagonistic interactions (Fig. 5D). The 5.0:5.0 ratio showed the least negative effect (SE = − 23%; actual 18.9 vs. TAY 24.5 mL-H₂/g-VS), while the 2.5:7.5 ratio (SE = − 59%; actual 6.2 vs. TAY 15.3 mL-H₂/g-VS) and 7.5:2.5 ratio (SE = − 85%; actual 5.1 vs. TAY 33.8 mL-H₂/g-VS) exhibited more severe antagonism. This negative outcome stems from fundamental incompatibility between immobilized cellulolytic cells and sequential processing architecture. During separate hydrolysis phase, enzyme release from immobilized KKU-MC1 proved insufficient for adequate saccharification, resulting in limited fermentable sugar availability regardless of substrate ratio. The antagonistic effect underscores that successful co-digestion synergy requires not only complementary substrates but also compatible process-configuration matching. A comprehensive comparison of HYs across different immobilization or free cell systems, substrates, and pretreatment methods is presented in Supplementary Table S4, demonstrating that this simplified hybrid system achieves competitive performance among immobilized cell technologies for lignocellulosic bio-HP. Mechanistic basis of synergy Collectively, these findings establish a hierarchy of synergy potential: SSF2 > SSF1 > SHF1 > SHF2, demonstrating that process architecture fundamentally governs co-digestion outcomes (Fig. 5). SSF2 achieved extraordinary enhancement (+ 1,150 to + 1,384%), SSF1 showed moderate positive synergy (+ 14 to + 77%), SHF1 exhibited ratio-dependent responses ranging from synergistic (+ 15%) to severely antagonistic (-95%), while SHF2 demonstrated consistent antagonism (− 23 to − 85%). The data supports a mechanistic model wherein synergy emerges from: (i) complementary nutrient profiles ensuring balanced microbial nutrition; (ii) diverse carbohydrate compositions enabling broader enzymatic utilization; (iii) metabolic buffering through distributed VFA production; and (iv) potential cross-feeding between populations specialized for different substrate components. Critically, these synergistic mechanisms require appropriate spatial and temporal organization—simultaneous processing (SSF) better preserves substrate interactions than sequential approaches (SHF), while immobilization configuration determines optimal substrate ratios. Metabolic pathway analysis The metabolite profiles revealed pathway-performance relationships governing system efficiency (Figs. 6, 7A–D). VFA composition—particularly HAc, HBu, HPr, and HLa distribution—provided mechanistic insights into fermentation pathways across substrate-configuration combinations.Fig. 6Volatile fatty acid profiles and final pH from mono-digestion systems. VFA composition (HLa = lactic acid, HAc = acetic acid, HPr = propionic acid, HBu = butyric acid) and final pH under (A) SSF1, (B) SSF2, (C) SHF1, and (D) SHF2 configurations. Substrates: NG = Napier grass; OPF = oil palm fronds at 10, 30, and 50 g-VS/L loadings. Error bars represent SD (n = 3).Fig. 7Volatile fatty acid profiles and final pH from co-digestion systems. VFA composition (HLa = lactic acid, HAc = acetic acid, HPr = propionic acid, HBu = butyric acid) and final pH under (A) SSF1, (B) SSF2, (C) SHF1, and (D) SHF2 configurations. Co-digestion of NG:OPF at ratios of 2.5:7.5, 5.0:5.0, and 7.5:2.5 (g-VS/L:g-VS/L), total loading 10 g-VS/L. Error bars represent SD (n = 3). Volatile fatty acid profiles and final pH from mono-digestion systems. VFA composition (HLa = lactic acid, HAc = acetic acid, HPr = propionic acid, HBu = butyric acid) and final pH under (A) SSF1, (B) SSF2, (C) SHF1, and (D) SHF2 configurations. Substrates: NG = Napier grass; OPF = oil palm fronds at 10, 30, and 50 g-VS/L loadings. Error bars represent SD (n = 3). Volatile fatty acid profiles and final pH from co-digestion systems. VFA composition (HLa = lactic acid, HAc = acetic acid, HPr = propionic acid, HBu = butyric acid) and final pH under (A) SSF1, (B) SSF2, (C) SHF1, and (D) SHF2 configurations. Co-digestion of NG:OPF at ratios of 2.5:7.5, 5.0:5.0, and 7.5:2.5 (g-VS/L:g-VS/L), total loading 10 g-VS/L. Error bars represent SD (n = 3). Mono-digestion patterns Mono-digestion conditions produced predominantly acetate-dominant VFA profiles across all configurations, with HAc accounting for 33–64% of total VFA (Fig. 6A–D). High-performing configurations, including SHF1 with NG (81.5 mL-H₂/g-VS) and SHF2 with OPF (57.3 mL-H₂/g-VS), maintained HAc proportions of 54% and 57%, respectively, consistent with acetate- and butyrate-producing pathways associated with efficient H₂ recovery42. However, no statistically significant linear correlation between HAc proportion and HY was observed across mono-digestion conditions (r = − 0.03, p = 0.87, n = 24), indicating that process configuration and substrate characteristics were stronger determinants of yield than VFA pathway distribution alone. Notably, HLa accumulation was minimal across most high-performing conditions, while HBu proportions increased with substrate loading, with SSF2 exhibiting elevated HBu proportions compared to SSF1, suggesting that immobilization of cellulolytic organisms shifts electron flux toward butyrate pathways. High-performing systems maintained final pH within 5.5–6.5, the optimal range for HP (Fig. 6A–D). Co-digestion metabolic shifts SSF2 co-digestion produced substantially higher total VFA (Fig. 7B) with a distinctive shift toward HPr-dominant fermentation. The best-performing 7.5:2.5 ratio achieved total VFA of 12.05 g/L with HPr (6.65 g/L) as dominant acid—approximately 3.4-fold higher than SSF1 (2.83–3.50 g/L). Despite elevated HPr production—typically considered unfavorable for HP—SSF2 achieved extraordinary yields (43.7–61.6 mL-H₂/g-VS) with synergistic enhancement (+ 1,150 to + 1,384%). This demonstrates that total metabolic flux and conversion efficiency can override pathway-specific limitations in immobilized co-digestion systems. Final pH remained moderate (6.24–6.29) despite high VFA production, suggesting matrix-contributed buffering effects (Fig. 7A–D). The HPr paradox observed in SSF2 challenges classical dark fermentation theory, which posits that propionate pathways act as competitive electron sinks43,44. The propionate-dominant metabolic profile in SSF2 co-digestion may attribute to elevated local H₂ partial pressure within the PVA-alginate matrix, which shifts thermodynamic equilibrium against further HP via acetate pathways, making the NADH-linked succinate-propionate route thermodynamically favorable as an alternative electron sink43,44. The substantially higher total carbon flux in SSF2 co-digestion (total VFA 12.05 g/L vs. 2.83–3.50 g/L in other configurations) channels excess electron equivalents through propionate despite overall high HYs. However, the exact mechanism underlying this phenomenon remains to be directly confirmed through in situ measurements of H₂ partial pressure and community profiling within the bead matrix. The high HY recorded here suggests alternative electron routing, likely facilitated by spatial compartmentalization within the immobilization matrix. Such microenvironments may support the coexistence of distinct microbial subpopulations, enabling simultaneous HPr synthesis and H₂ evolution. Moreover, matrix-mediated buffering maintains optimal hydrogenase activity, circumventing the kinetic inhibition typically observed in free-cell systems at high acid concentrations. In contrast, SHF configurations aligned with classical theory: SHF1 co-digestion showed elevated HLa under OPF-dominant conditions, while SHF2 maintained low-VFA, acetate-dominant profiles despite inferior yields. Collectively, these findings establish that immobilization-mediated metabolic flux, rather than strict adherence to traditional pathway hierarchies, dictates the performance limits of bio-H₂ systems. Long-term stability and economic viability Multi-cycle operational stability Three optimal configurations were evaluated over five cycles (Table 5). All showed gradual HY decline (p < 0.001), achieving 84.0–89.1% retention by cycle 5, with SSF2 demonstrating superior stability (89.1%). Notably, the hybrid PVA-alginate system achieved 84.0–89.1% retention compared to 83.8% reported for dual-matrix systems at four cycles23. SSF2’s superior retention (89.1% vs. 83.8%) may result from elimination of delamination issues associated with core–shell architectures. FPase activity decreased from 0.08–0.09 to 0.04–0.06 IU/mL by cycle 5, achieving 50.0–66.7% retention (p < 0.001). The moderate enzyme retention compared to HY retention (84.0–89.1%) suggests immobilized systems maintain sufficient cellulolytic capacity through continued enzyme production within beads21. VSS in immobilized beads declined from 0.30 to 0.15–0.24% (w/v), achieving 50.0–80.0% retention with configuration-dependent variation (p < 0.05). SSF configurations showed superior cell retention compared to SHF, possibly due to less severe environmental transitions in single-stage processing. Physical assessment throughout five cycles revealed maintained bead integrity across all configurations (Table 5). Beads remained intact with no visible matrix dissolution, fragmentation, or structural collapse under operational conditions (37 °C, 150 rpm, pH 7.0–7.5). Surface roughness increased from smooth to rough by cycle 5 due to biofilm accumulation, indicating active microbial colonization without compromising matrix integrity. Bead diameter showed minimal variation (CV < 10% relative to initial 2.5 ± 0.2 mm), confirming absence of significant swelling or shrinkage—contrasting with preliminary formulations that exhibited dimensional instability. Although VSS measurements quantified overall biomass retention within beads, they cannot distinguish between cell death and cell washout into the surrounding medium. Future studies should employ 16S rRNA amplicon sequencing to characterize microbial community dynamics within immobilized beads across operational cycles and between mono- and co-digestion conditions, which would allow observation of community shifts within the immobilization matrix. Additionally, quantification of cells in the spent medium supernatant would complement in-bead VSS measurements by directly assessing cell washout rates, providing a more complete picture of microbial retention and succession within the PVA-alginate matrix over extended operation.Table 5Reusability and stability of hybrid immobilized cells over five consecutive operational cycles.ParameterProcess-SubstrateInitial12345Retention (%)Decline rate (%/cycle)P-value†A. Hydrogen Production PerformanceHY (mL-H2/g-VS)SHF1 (10-NG)81.5 ± 7.5a78.2 ± 7.1ab75.8 ± 6.8bc73.1 ± 6.2 cd70.8 ± 5.9de68.5 ± 5.5e84.0c3.20c < 0.001***SHF2 (10-OPF)57.3 ± 4.3a55.1 ± 4.0ab53.4 ± 3.8bc51.6 ± 3.6 cd50.1 ± 3.4de48.7 ± 3.2e85.0b3.00b < 0.001***SSF2 (7.5NG + 2.5OPF)61.6 ± 7.4a60.5 ± 7.0ab59.1 ± 6.6bc57.5 ± 6.2 cd56.2 ± 5.9de54.9 ± 5.6e89.1a2.18a < 0.001***Energy yield (MJ/ton-FW)SHF1 (10-NG)177.7 ± 16.4a170.5 ± 15.5ab165.3 ± 14.8bc159.4 ± 13.5 cd154.4 ± 12.9de149.4 ± 12.0e84.1c3.18c < 0.001***SHF2 (10-OPF)79.4 ± 6.0a76.4 ± 5.5ab74.0 ± 5.3bc71.5 ± 5.0 cd69.4 ± 4.7de67.5 ± 4.4e85.0b3.00b < 0.001***SSF2 (7.5NG + 2.5OPF)122.5 ± 14.7a120.3 ± 13.9ab117.5 ± 13.1bc114.3 ± 12.3 cd111.7 ± 11.7de109.1 ± 11.1e89.1a2.17a < 0.001***B. Enzyme Activity: FPase (IU/mL)SHF1 (10-NG)0.09 ± 0.01a0.08 ± 0.02ab0.08 ± 0.00bc0.07 ± 0.01 cd0.06 ± 0.02de0.05 ± 0.02e55.6c8.88c < 0.001***SHF2 (10-OPF)0.08 ± 0.01a0.06 ± 0.01ab0.05 ± 0.02bc0.06 ± 0.01 cd0.05 ± 0.03de0.04 ± 0.02e50.0b10.00b < 0.001***SSF2 (7.5NG + 2.5OPF)0.09 ± 0.02a0.07 ± 0.02ab0.06 ± 0.02bc0.06 ± 0.01 cd0.05 ± 0.03de0.06 ± 0.02e66.7a6.67a < 0.001***C. Cell Viability and Biomass: VSS in immobilized beads (% w/v)SHF1 (10-NG)0.30 ± 0.09a0.28 ± 0.02ab0.22 ± 0.04bc0.20 ± 0.12 cd0.19 ± 0.11de0.15 ± 0.04e50.0c10.00c < 0.001***SHF2 (10-OPF)0.30 ± 0.10a0.29 ± 0.04ab0.28 ± 0.05bc0.28 ± 0.10 cd0.27 ± 0.10de0.22 ± 0.10e73.3b5.33b < 0.001***SSF2 (7.5NG + 2.5OPF)0.30 ± 0.10a0.29 ± 0.01ab0.28 ± 0.10bc0.26 ± 0.15 cd0.25 ± 0.08de0.24 ± 0.10e80.0a4.00a < 0.001***D. Physical Bead IntegrityStructural integrityAll conditionsIntactIntactIntactIntactIntactGoodMaintained––Surface roughnessAll conditionsSmoothSmoothModerateModerateRoughRough–––Values represent mean ± SD (n = 3). Different superscript letters within the same substrate type and configuration indicate significant differences (Tukey’s HSD, p < 0.05). Retention % and Decline rate superscripts (a,b,c): Significant differences between systems. †P-values from one-way ANOVA testing differences across all six time points for each system.Energy Yield Calculation EY (MJ/ton-FW) = HY × (MW/Vm) × HHV × VS% × DM% × 10⁶, where MW = 2.016 g/mol, Vm = 22,400 mL/mol, HHV = 142 MJ/kg.Conversion factors NG = 2.180 (VS = 88.4%, Moisture = 80.7%); OPF = 1.386 (VS = 91.9%, Moisture = 88.2%); Mix = 1.988 (weighted average). Table 5Reusability and stability of hybrid immobilized cells over five consecutive operational cycles.ParameterProcess-SubstrateInitial12345Retention (%)Decline rate (%/cycle)P-value†A. Hydrogen Production PerformanceHY (mL-H2/g-VS)SHF1 (10-NG)81.5 ± 7.5a78.2 ± 7.1ab75.8 ± 6.8bc73.1 ± 6.2 cd70.8 ± 5.9de68.5 ± 5.5e84.0c3.20c < 0.001***SHF2 (10-OPF)57.3 ± 4.3a55.1 ± 4.0ab53.4 ± 3.8bc51.6 ± 3.6 cd50.1 ± 3.4de48.7 ± 3.2e85.0b3.00b < 0.001***SSF2 (7.5NG + 2.5OPF)61.6 ± 7.4a60.5 ± 7.0ab59.1 ± 6.6bc57.5 ± 6.2 cd56.2 ± 5.9de54.9 ± 5.6e89.1a2.18a < 0.001***Energy yield (MJ/ton-FW)SHF1 (10-NG)177.7 ± 16.4a170.5 ± 15.5ab165.3 ± 14.8bc159.4 ± 13.5 cd154.4 ± 12.9de149.4 ± 12.0e84.1c3.18c < 0.001***SHF2 (10-OPF)79.4 ± 6.0a76.4 ± 5.5ab74.0 ± 5.3bc71.5 ± 5.0 cd69.4 ± 4.7de67.5 ± 4.4e85.0b3.00b < 0.001***SSF2 (7.5NG + 2.5OPF)122.5 ± 14.7a120.3 ± 13.9ab117.5 ± 13.1bc114.3 ± 12.3 cd111.7 ± 11.7de109.1 ± 11.1e89.1a2.17a < 0.001***B. Enzyme Activity: FPase (IU/mL)SHF1 (10-NG)0.09 ± 0.01a0.08 ± 0.02ab0.08 ± 0.00bc0.07 ± 0.01 cd0.06 ± 0.02de0.05 ± 0.02e55.6c8.88c < 0.001***SHF2 (10-OPF)0.08 ± 0.01a0.06 ± 0.01ab0.05 ± 0.02bc0.06 ± 0.01 cd0.05 ± 0.03de0.04 ± 0.02e50.0b10.00b < 0.001***SSF2 (7.5NG + 2.5OPF)0.09 ± 0.02a0.07 ± 0.02ab0.06 ± 0.02bc0.06 ± 0.01 cd0.05 ± 0.03de0.06 ± 0.02e66.7a6.67a < 0.001***C. Cell Viability and Biomass: VSS in immobilized beads (% w/v)SHF1 (10-NG)0.30 ± 0.09a0.28 ± 0.02ab0.22 ± 0.04bc0.20 ± 0.12 cd0.19 ± 0.11de0.15 ± 0.04e50.0c10.00c < 0.001***SHF2 (10-OPF)0.30 ± 0.10a0.29 ± 0.04ab0.28 ± 0.05bc0.28 ± 0.10 cd0.27 ± 0.10de0.22 ± 0.10e73.3b5.33b < 0.001***SSF2 (7.5NG + 2.5OPF)0.30 ± 0.10a0.29 ± 0.01ab0.28 ± 0.10bc0.26 ± 0.15 cd0.25 ± 0.08de0.24 ± 0.10e80.0a4.00a < 0.001***D. Physical Bead IntegrityStructural integrityAll conditionsIntactIntactIntactIntactIntactGoodMaintained––Surface roughnessAll conditionsSmoothSmoothModerateModerateRoughRough–––Values represent mean ± SD (n = 3). Different superscript letters within the same substrate type and configuration indicate significant differences (Tukey’s HSD, p < 0.05). Retention % and Decline rate superscripts (a,b,c): Significant differences between systems. †P-values from one-way ANOVA testing differences across all six time points for each system.Energy Yield Calculation EY (MJ/ton-FW) = HY × (MW/Vm) × HHV × VS% × DM% × 10⁶, where MW = 2.016 g/mol, Vm = 22,400 mL/mol, HHV = 142 MJ/kg.Conversion factors NG = 2.180 (VS = 88.4%, Moisture = 80.7%); OPF = 1.386 (VS = 91.9%, Moisture = 88.2%); Mix = 1.988 (weighted average). Reusability and stability of hybrid immobilized cells over five consecutive operational cycles. Values represent mean ± SD (n = 3). Different superscript letters within the same substrate type and configuration indicate significant differences (Tukey’s HSD, p < 0.05). Retention % and Decline rate superscripts (a,b,c): Significant differences between systems. †P-values from one-way ANOVA testing differences across all six time points for each system. Energy Yield Calculation EY (MJ/ton-FW) = HY × (MW/Vm) × HHV × VS% × DM% × 10⁶, where MW = 2.016 g/mol, Vm = 22,400 mL/mol, HHV = 142 MJ/kg. Conversion factors NG = 2.180 (VS = 88.4%, Moisture = 80.7%); OPF = 1.386 (VS = 91.9%, Moisture = 88.2%); Mix = 1.988 (weighted average). Economic assessment Production costs for a representative 50 mL batch (SHF1, 10 g-VS/L NG) totaled 0.0060 USD, comprising medium components (0.0035 USD, 57.3%) and immobilization materials (0.0026 USD, 42.7%). Peptone alone accounts for 41.5% of total costs. HP of 40.75 mL at 6 USD/kg selling price2,37 generated 0.000022 USD revenue, yielding MEY of 0.00364 USD-H₂/USD-total cost. Five-cycle operational stability provides a 5.0-fold cost advantage over single-use systems through amortized immobilization costs (Supplementary Table S5). For comparison, Sani et al.8 reported economic yields of 0.030–0.035 USD-H₂/USD-medium for pilot-scale SHF processes using co-digestion NG and OPF (5.0:5.0 g-VS/L), with trace element supplementation improving yields by 16.6%. While the current lab-scale MEY appears lower, this difference reflects both scale effects and different system boundaries: the present study includes immobilization material costs (42.7% of total), whereas pilot-scale operations benefit from economies of scale and optimized medium utilization. Similarly, Rodríguez-Valderrama et al.45 reported a net benefit of 0.009 USD/kg co-substrates (wet basis) for bio-H2 co-digestion of dilute acid hydrolysates from fruits-and-vegetables wastes with corn stover, reinforcing that co-digestion of complementary agricultural residues enhances both yield and economic performance. This lab-scale assessment excludes capital costs, labor, and scale-up factors. Pilot-scale validation is necessary to confirm economic viability, though co-digestion studies45 and our previous work8 suggest favorable economics persist at larger scales. Primary improvement pathways include: (1) medium cost reduction through waste-derived nitrogen sources (57.3% of current expenses); (2) substrate loading optimization beyond 10 g-VS/L; and (3) industrial-scale material procurement. Energy yield retention paralleled HY trends: 84.1% for SHF1-NG (177.7 → 149.4 MJ/ton-FW), 85.0% for SHF2-OPF, and 89.1% for SSF2 co-digestion (Supplementary Table S6), demonstrating substantial energy recovery potential after five operational cycles. Morphological evidence of system performance SEM analysis provided morphological insights into substrate transformation and bead integrity throughout the bio-HP process (Fig. 8), revealing structure-performance relationships that explain configuration-dependent yields. NG’s irregular topography implies high natural accessibility (Fig. 8A), unlike OPF’s smooth, intact matrix which restricts access (Fig. 8B).Fig. 8Scanning electron microscopy (SEM) characterization of substrates and hybrid immobilized cell beads. (A, B) Untreated substrates before fermentation: (A) Napier grass (NG); (B) oil palm fronds (OPF). (C–F) Substrate residues after fermentation at optimal conditions: (C) SHF1 (10 g-VS/L NG); (D) SHF2 (10 g-VS/L OPF); (E) SSF1 and (F) SSF2 co-digestion residues (7.5:2.5 NG:OPF). (G–J) Hybrid immobilized beads after five consecutive reusability cycles: (G–H) SHF1-HPB beads; (I–J) SSF2-Modified KKU-MC1 beads. No visible matrix dissolution or structural collapse was observed in either bead type. Substrate images obtained at 500 × ; immobilized bead images at 1,500 × (G, I) and 7,500 × (H, J). Scanning electron microscopy (SEM) characterization of substrates and hybrid immobilized cell beads. (A, B) Untreated substrates before fermentation: (A) Napier grass (NG); (B) oil palm fronds (OPF). (C–F) Substrate residues after fermentation at optimal conditions: (C) SHF1 (10 g-VS/L NG); (D) SHF2 (10 g-VS/L OPF); (E) SSF1 and (F) SSF2 co-digestion residues (7.5:2.5 NG:OPF). (G–J) Hybrid immobilized beads after five consecutive reusability cycles: (G–H) SHF1-HPB beads; (I–J) SSF2-Modified KKU-MC1 beads. No visible matrix dissolution or structural collapse was observed in either bead type. Substrate images obtained at 500 × ; immobilized bead images at 1,500 × (G, I) and 7,500 × (H, J). Substrate degradation patterns NG residues from SHF1 (81.5 ± 7.5 mL-H2/g-VS) presented extensive degradation (Fig. 8C): severely disrupted fiber structures, considerable surface erosion, fragmentation into 10–50 μm segments, and deeply pitted surfaces indicating thorough enzymatic attack. Residual cellulose appeared partially amorphized while hemicellulose was almost completely removed. This morphological evidence supports the 2.2-fold yield improvement of SHF1 over SSF1 for NG, where temporal separation allowed optimal enzymatic hydrolysis conditions23. In contrast, OPF residues from SHF2 (57.3 ± 4.3 mL-H2/g-VS) showed heterogeneous degradation patterns (Fig. 8D) with areas of extensive fiber disruption alongside relatively intact regions. Despite incomplete degradation, this represented the highest OPF performance across all configurations, with residual undegraded material representing recalcitrant lignin-rich fractions. Co-digestion residue morphology SSF2 residues (7.5:2.5 NG:OPF, Fig. 8F) showed preferential NG degradation with deep pitting alongside moderate OPF erosion—explaining the optimal ratio where rapid NG hydrolysis maintains sugar availability while OPF extends substrate duration, contributing to + 1,150% synergy. Immobilized bead characterization Post-fermentation beads from all configurations exhibited preserved structural integrity despite heavy microbial colonization (Fig. 8G–J). SHF1 HPB beads displayed a fibrous interconnected polygon network structure with cells distributed throughout the polymer matrix connected through extracellular polymeric substances (EPS) (Fig. 8G–H), indicating active metabolic function and strong matrix-cell adhesion19. Cross-sectional examination revealed uniform cell distribution from surface to interior, suggesting good substrate and product diffusivities. Maintained porosity despite cell loading explains the 84.0% HY retention over 5 cycles in SHF1-NG. SSF2 Modified KKU-MC1 beads showed dense spherical cell aggregates distributed throughout the porous matrix interior with biofilm formation and roughened surface topography (Fig. 8I–J). The three-dimensional network structure remained intact with bacterial cells embedded within the polymer matrix connected through EPS, indicating active metabolic function and sustained enzyme production. This extensive interior colonization maintains continuous enzyme secretion, explaining the 66.7% FPase retention during reusability studies. Interior examination revealed that in the SSF configuration, cellulolytic organisms distribute uniformly as spherical aggregates throughout the matrix interior, maintaining enzymatic activity throughout the bead volume, while HPB in SHF1 develop fibrous network structures that facilitate access to soluble fermentation products diffusing through the polymer matrix. This spatial organization explains the 89.1% HY retention and 80.0% VSS retention achieved in SSF2 co-digestion. No visible matrix dissolution, bead fragmentation, or structural degradation was observed throughout the five-cycle study, confirming that the simplified hybrid PVA-alginate system provides adequate mechanical stability under operational conditions (37 °C, 150 rpm, pH 5.2–7.5) while supporting high cell densities and active metabolism—a significant improvement over previous formulations with limited operational lifetime. Conclusions This study establishes that simplified hybrid immobilization provides a practical alternative to complex dual-matrix systems for bio-HP from agricultural residues. The system achieved 89.1% HY retention over five cycles—outperforming dual-matrix approaches (83.8%, four cycles). The system achieved competitive yields while providing 5.0-fold cost advantages through multi-cycle operation. Process configuration profoundly influenced performance, with SSF2 co-digestion achieving high synergy for immobilized systems, while SHF provided 2.2-fold advantages for recalcitrant substrates. Process configuration proved the primary determinant of yield, though immobilization location fundamentally altered metabolic flux. Economic analysis identified medium components (57.3% of costs) as the primary optimization target, with waste-derived nitrogen substitution offering immediate cost reduction potential. The simplified platform extends beyond immediate applications to broader sustainable energy systems. Integration with biomethane production could improve energy recovery by 40–60%, while configuration-substrate matching principles provide general guidelines for diverse agricultural residues. This work is bio-HP from laboratory demonstration toward practical renewable energy technology for distributing HP from agricultural wastes. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Information 1 Supplementary Information 1