Discontinuous biomass recycling as a successful strategy to enhance continuous hydrogen production at high organic loading rates

A novel strategy to discontinuously increase the biomass concentration in a continuous stirred-tank reactor was evaluated to enhance the performance of dark fermentation. Different concentrations of biomass were evaluated at organic loading rates (OLR) ranging from 90 to 160 g lactose/L-d with a hydraulic retention time (HRT) of 6 h. The study revealed that the discontinuous increase of biomass enhanced the hydrogen (H₂) production rates and carboxylic acids concentrations by 19–25% and 8–23%, respectively. In particular, a maximum H₂ production rate of 30.8 L H₂/L-d with carboxylic acids concentration of 20 g/L was reached at an OLR of 138 g lactose/L-d with a biomass concentration of 15 g volatile suspended solids/L. The analysis of microbial communities showed the co-dominance of Clostridium and lactic acid bacteria. Overall, the discontinuous increase of biomass was an effective strategy to improve the performance of suspended-biomass reactors operated at high OLR and low HRT.     

Converting waste molasses liquor into biohydrogen via dark fermentation using a continuous bioreactor

This study investigated the effects of substrate concentration, HRT (hydraulic retention time), and pre-treatment of the substrate molasses on biohydrogen production from waste molasses (condensed molasses fermentation solubles, CMS) with a CSTR (continuously-stirred tank reactor). First, the hydrogen production was performed with various CMS concentrations (40–90 g COD/L, total sugar 8.7–22.6 g/L) with 6 h HRT. The results show that the maximal hydrogen production rate (HPR) occurred at 80 g COD/L substrate (19.8 g ToSu/L, ToSu: Total Sugar), obtaining an HPR of 0.417 mol/L/d. However, maximum hydrogen yield (HY) of 1.44 mol H₂/mol hexose and overall hydrogen production efficiency (HPE) of 25.6% were achieved with a CMS concentration of 70 g COD/L (17.3 g ToSu/L). The substrate inhibition occurred when CMS concentration was increased to 90 g COD/L (22.6 g ToSu/L). Furthermore, it was observed that the optimal HPR, HY, and HPE all occurred at HRT 6 h. Operating at a lower HRT of 4 h decreased the hydrogen production performance because of lower substrate utilization efficiency. The employment of pre-heating treatment (60 °C for 1 h) of the substrate could markedly enhance the fermentation performance. With 6 h HRT and substrate pre-heating treatment, the HPE raised to 29.9%, which is 18% higher than that obtained without thermal pretreatment.     

Microbial ecology of a lactate-driven dark fermentation process producing hydrogen under carbohydrate-limiting conditions

Despite the high prevalence of lactic acid bacteria in dark fermentation (DF) processes, their ecological role is not yet completely elucidated, preventing their systematic use as “helpers” for hydrogen production. The aim of this study was to investigate the microbial community structure of a lactate-driven DF process that successfully produced hydrogen under carbohydrate-limiting conditions using tequila vinasse as a substrate. Microbial responses to stepwise decreases in hydraulic retention time (HRT) from 24 to 4 h were assessed by using Illumina MiSeq sequencing. HRTs above 12 h and below 6 h led to a lower hydrogen production rate (HPR; 0.2–3.3 L/L-d) and process stability (HPR variations within 25–65%), which were associated with the presence of Acetobacter lovaniensis, Clostridium luticellari, Blautia coccoides, and the high abundance of propionate and lactate. Interestingly, transient conditions from unsteady-to-steady state occurred at an HRT of 12 h, where species richness and evenness decreased remarkably. Accordingly, HRTs between 12 and 6 h resulted in higher HPRs of up to 11.7 ± 0.7 L/L-d with HPR variations of less than 10%, which closely matched with the dominance of Clostridium sp., and butyrate and acetate as the main aqueous products. Overall, the results indicate that the successfulness of exploiting the ‘unwanted’ LAB proliferation through lactate-driven DF processes requires the enrichment of lactate-consuming and hydrogen-producing bacteria, which entails the selection of proper biocatalysts and operating conditions/strategies such as the operation of DF reactors under carbohydrate-limiting conditions and low HRTs.     

Optimal operational conditions for biohydrogen production from sugar refinery wastewater in an ASBR

The performance of biohydrogen production in an anaerobic sequencing batch reactor (ASBR) was evaluated with respect to variations in the key operational parameters – pH, hydraulic retention time HRT, and organic loading rate OLR using sugar refinery wastewater as substrate. Analysis of variance (ANOVA) indicated HRT had less significant influence on hydrogen content and yield in comparison to pH and OLR, whereas OLR has much impact on hydrogen production rate. Taxonomic analysis results showed that diverse bacterial species contributed to hydrogen production and the dominant species in the bioreactor were governed by all operational parameters. Even without pretreatment of the seed sludge, a high proportion of Clostridium spp. over the other bacterial species was observed at pH 5.5, and this is compatible with the high hydrogen productivity. Consequently, pH 5.5, HRT 10 h, and OLR 15 kg/m³ d were delineated as the optimal operational conditions for an ASBR fed with sugar refinery wastewater.     

Biological stoichiometric analysis of nutrition and ammonia toxicity in thermophilic anaerobic co-digestion of organic substrates under different organic loading rates

This study used chemostats under different Organic Loading Rates (OLRs) to investigate the co-digestion of kitchen waste, swine wastewater sludge, and fruit and vegetable waste. In kitchen waste, mean Total Chemical Oxygen Demand (TCOD), Oil and Grease (O&G), and moisture content (MC) was 101.5 g/L, 33.0 g/L, and 82.7%, respectively. The TCOD/Total Kjeldahl Nitrogen (TKN) and TCOD/Total Phosphorus (TP) of kitchen waste were 319.5 and 230, respectively. In swine wastewater sludge, TCOD/TKN was 4.56–43.9 and TCOD/TP was 2.02–31.8. Biodegradability tests of fruit and vegetable waste showed that COD removal exceeded 56%, and methane recovery exceeded 80%. Co-digestion of these three organic wastes in chemostats suggests that the system functioned stably up to an OLR of 9.52 g COD/L-d at a Hydraulic Retention Time (HRT) of 5 days. When the OLR increased to 12.54 g COD/L-d, average COD and Volatile Suspended Solids (VSS) removal efficiencies decreased from 90% to 76.5% and from 93% to 76.5%, respectively. The analyzed NH₃–N concentration is 28% less than the stoichiometry-predicted concentration. The discrepancy may be due to differences in substrate biodegradabilities, TKN sampling and analysis procedures, and parameters used for stoichiometry calculations.     

Hydrogen and methane production in extreme thermophilic conditions in two-stage (upflow anaerobic sludge bed) UASB reactor system

Two-stage hydrogen and methane production in extreme thermophilic (70 °C) conditions was demonstrated for the first time in UASB-reactor system. Inoculum used in hydrogen and methane reactors was granular sludge from mesophilic internal circulation reactor and was first acclimated for extreme thermophilic conditions. In hydrogen reactor, operated with hydraulic retention time (HRT) of 5 h and organic loading rate (OLR) of 25.1 kg COD/m³/d, hydrogen yield was 0.73 mol/mol glucose_added. Methane was produced in second stage from hydrogen reactor effluent. In methane reactor operated with HRT of 13 h and OLR of 7.8 kg COD/m³/d, methane yield was 117.5 ml/g COD_added. These results prove that hydrogen and methane can be produced in extreme thermophilic temperatures, but as batch experiments confirmed, for methane production lower temperature would be more efficient.     

Efficient hydrogen gas production from molasses in hybrid anaerobic-activated sludge-rotating biological contactor

A hybrid anaerobic activated sludge and rotating biological contactor (AnAS-RBC) reactor with sludge return was investigated for biological hydrogen production from molasses by a mixed culture of hydrogen-producing bacteria via a dark fermentation process. The pH was not controlled during the experiments. The effects of organic loading rate (OLR) and disk rotational speed (DRS) on process performance were studied using response surface methodology. Both OLR and DRS showed significant effects on all process responses. High levels of DRS were found favorable for efficient process. Optimal mixing of the bacterial culture contributed towards higher acetate production resulting in higher hydrogen production. The highest hydrogen production (168 mL/g COD) was achieved by applying OLR and DRS of 23 g/L.d and 30 rpm, respectively. In this study, the highest percentage of hydrogen in biogas produced was 87% in hydraulic retention time of 12 h.     

Clostridium strain co-cultures for biohydrogen production enhancement from condensed molasses fermentation solubles

An anaerobic continuous-flow hydrogen fermentor was operated at a hydraulic retention time of 8 h using condensed molasses fermentation solubles (CMS) substrate of 40 g-COD/L. Serum bottles were used for seed micro-flora cultivation and batch hydrogen fermentation tests (CMS substrate concentrations of 10–160 g-COD/L). Three hydrogen-producing bacterial strains Clostridium sporosphaeroides F52, Clostridium tyrobutyricum F4 and Clostridium pasteurianum F40 were isolated from the seed fermentor and used as the seeding microbes in single and mixed-culture cultivations for determining their hydrogen productivity. These strains possessed specific hydrogenase genes that could be detected from CMS-fed hydrogen fermentors and were major hydrogen producers. C. pasteurianum F40 was the dominant strain with a high hydrogen production rate while C. sporosphaeroides F52 may play a main role in degrading carbohydrate and glutamate. These strains could be co-cultivated as a symbiotic mixed-culture process to enhance hydrogen productivity. C. pasteurianum F40 or C. tyrobutyricum F4 co-culture with the glutamate-utilizing strain C. sporosphaeroides F52 efficiently enhanced hydrogen production by 12–220% depending on the substrate CMS concentrations.     

Bio-hydrogen production from tequila vinasses: Effect of detoxification with activated charcoal on dark fermentation performance

Hydrogen production from dark fermentation is a potential source of sustainable fuel when it is generated from waste. This study compared hydrogen production resulting from fermentation using raw and detoxified tequila vinasse. Vinasse was detoxified with granular activated charcoal, which was used to adsorb compounds that could inhibit the production of hydrogen by dark fermentation. In batch cultures detoxification of vinasse led to up to 20% higher maximum velocities of hydrogen production, a 5.4 h reduction in the lag phase and an 11% higher molar yield, compared to results obtained with raw vinasse. Losses of sugars after detoxification provoked that the specific hydrogen volumetric yields obtained with detoxified vinasse were 30–40% lower with 5 g COD/L and 15 g COD/L initial concentrations, compared to the ones obtained with raw vinasse. For an initial 30 g COD/L no differences in specific hydrogen yields were observed between raw or detoxified vinasse in batch fermentation. Continuous culture fermentation of vinasse showed hydrogen production rates between 1.32 ± 0.07 to 1.39 ± 0.14 NL H₂/L-d when extra nutrients were added, while a stable production of hydrogen through fermentation of detoxified vinasse could not be maintained despite nutrient addition. Production of hydrogen from vinasse diluted with water with no additional nutrients was assessed and rates close to 0.42 ± 0.02 NL H₂/L-d and hydrogen content close to 37% were obtained. Accumulation of lactic acid and a predominant production of butyric acid over acetic acid suggested that the fermentation dynamics of vinasse with no supplementary nutrients were especially susceptible to high substrate loading rates and prolonged hydraulic retention times.     

Biohydrogen production from immobilized cells and suspended sludge systems with condensed molasses fermentation solubles

The anaerobic fermentation using the condensed molasses fermentation solubles (CMS) as substrate in a continuously stirred anaerobic bioreactor (CSABR) was carried out for optimal hydrogen production performance of biohydrogen production rate and yield, where as two kinds of bioreactors used. One is a suspended sludge bioreactor (SSB) which used suspended seed sludge. The other bioreactor is an immobilized cell bioreactor (ICB) which used immobilized cells and mix the same seed sludge in the SSB as the source of the bacteria. It was found that the hydrogen production rate increased with a decrease in the hydraulic retention time (HRT), when substrate concentration was 40 g COD/L in an SSB as well as maximum hydrogen production rate of 14.04 ± 2.08 L/d/L obtained at HRT 0.5 h (ca. 5.78 times value of HRT 4 h) in the SSB system. The hydrogen production rate at low dilution rate (HRT > 4 h), in the ICB is better than SSB, meanwhile at a high dilution rate (HRT < 4 h), due to the presence of enriched granules in the SSB (12.30 g VSS/L), absent in the ICB (9.89 g VSS/L), the hydrogen production rate was 7.60 ± 1.05 L/d/L (ca. 1.23 times higher than HRT 4 h), which was lower than the rate in the SSB. Eventually, the hydrogen production rate increased by increasing the substrate concentrations from 40 to 60 g COD/L within the HRT range of 2–4 h in both the SSB as well as in ICB systems.     

Evaluation of hidden H2-consuming pathways using metabolic flux-based analysis for a fermentative side-stream dynamic membrane bioreactor using untreated seed sludge

This study investigates the performance and hidden hydrogen consuming metabolic pathways of a fermentative side stream dynamic membrane (DM) bioreactor using flux balance analysis (FBA). The bioreactor was inoculated with untreated methanogenic seed sludge. It was found that fouling rate aggravated with increasing COD concentration (10–30 g/L) and was positively correlated to it rather than to the applied solid flux on the DM module. Due to increased fouling rate the hydraulic retention time (HRT) could not be reduced less than 0.82 ± 0.02 d. An increase in the organic loading rate (OLR) led to an increase in H₂ yield from 0.01 to 0.76 mol H₂/mol of sucrose. FBA revealed that homoacetogenesis was the main H₂-consuming pathway at lower OLRs (corresponding to 10 and 15 g COD/L), while for the OLR corresponding to 30 g COD/L, homoacetogens were suppressed. More importantly, caproic acid production pathway was identified for the first time as another H₂-consuming pathway at high OLR which was not significant at lower OLRs during fermentative dynamic membrane bioreactor operations.     

Effect of organic loading on a novel hydrogen bioreactor

This study investigated the impact of six organic loading rates (OLR) ranging from 6.5 gCOD/L-d to 206 gCOD/L-d on the performance of a novel integrated biohydrogen reactor clarifier systems (IBRCSs) comprised a continuously stirred reactor (CSTR) for biological hydrogen production, followed by an uncovered gravity settler for decoupling of solids retention time (SRT) from hydraulic retention time (HRT). The system was able to maintain a high molar hydrogen yield of 2.8 mol H₂/mol glucose at OLR ranging from 6.5 to 103 gCOD/L-d, but dropped precipitously to approximately 1.2 and 1.1 mol H₂/mol glucose for the OLRs of 154 and 206 gCOD/L-d, respectively. The optimum OLR at HRT of 8 h for maximizing both hydrogen molar yield and volumetric hydrogen production was 103 gCOD/L-d. A positive statistical correlation was observed between the molar hydrogen production and the molar acetate-to-butyrate ratio. Biomass yield correlated negatively with hydrogen yield, although not linearly. Analyzing the food-to-microorganisms (F/M) data in this study and others revealed that, both molar hydrogen yields and biomass specific hydrogen rates peaked at 2.8 mol H₂/mol glucose and 2.3 L/gVSS-d at F/M ratios ranging from 4.4 to 6.4 gCOD/gVSS-d. Microbial community analysis for OLRs of 6.5 and 25.7 gCOD/L-d showed the predominance of hydrogen producers such as Clostridium acetobutyricum, Klebsiella pneumonia, Clostridium butyricum, Clostridium pasteurianum. While at extremely high OLRs of 154 and 206 gCOD/L-d, a microbial shift was clearly evident due to the coexistence of the non-hydrogen producers such as Lactococcus sp. and Pseudomonas sp.     

Continuous fermentative hydrogen production from cheese whey wastewater under thermophilic anaerobic conditions

Hydrogen (H₂) production from cheese processing wastewater via dark anaerobic fermentation was conducted using mixed microbial communities under thermophilic conditions. The effects of varying hydraulic retention time (HRT: 1, 2 and 3.5 days) and especially high organic load rates (OLR: 21, 35 and 47 g chemical oxygen demand (COD)/l/day) on biohydrogen production in a continuous stirred tank reactor were investigated. The biogas contained 5–82% (45% on average) hydrogen and the hydrogen production rate ranged from 0.3 to 7.9 l H₂/l/day (2.5 l/l/day on average). H₂ yields of 22, 15 and 5 mmol/g COD (at a constant influent COD of 40 g/l) were achieved at HRT values of 3.5, 2, and 1 days, respectively. On the other hand, H₂ yields were monitored to be 3, 9 and 6 mmol/g COD, for OLR values of 47, 35 and 21 g COD/l/day, when HRT was kept constant at 1 day. The total measurable volatile fatty acid concentration in the effluent (as a function of influent COD) ranged between 118 and 27,012 mg/l, which was mainly composed of acetic acid, iso-butyric acid, butyric acid, propionic acid, formate and lactate. Ethanol and acetone production was also monitored from time to time.To characterize the microbial community in the bioreactor at different HRTs, DNA in mixed liquor samples was extracted immediately for PCR amplification of 16S RNA gene using eubacterial primers corresponding to 8F and 518R. The PCR product was cloned and subjected to DNA sequencing. The sequencing results were analyzed by using MegaBlast available on NCBI website which showed 99% identity to uncultured Thermoanaerobacteriaceae bacterium.     

Characteristics of biohydrogen fermentation from various substrates

The characteristics of biohydrogen production from sucrose, slurry-type piggery waste and food waste under the effects of the reactor configurations and operational pHs (6 and 9) were examined by using heat-treated anaerobic sludge as a seed biomass. When sucrose was used in the batch test, the maximum hydrogen yield was 0.12–0.13 g COD (as H₂)/g COD (1.41–1.43 mol/mol hexose) at pH 6. In contrast, 0.10–0.11 g COD (as H₂)/g COD (1.12–1.21 mol/mol hexose) hydrogen yield was achieved from the reactor at pH 9. On the other hand, hydrogen production was not observed in the continuous sequencing batch mode fermenters fed with sucrose. Profile analysis at each cycle revealed hydrogen production at the initial operation periods but eventually only methane at 36 days. When slurry-type piggery waste was used as the substrate, the upflow elutriation-type fermenters produced methane but not hydrogen after 30 days operation. The fermentation intermediate profile showed that the hydrogen produced might have been consumed by homoacetogenic or propionate producing reactions, and eventually converted into methane by acetoclastic methanogens. The downflow leaching bed fermenters using food waste produced 0.013 L H₂/g volatile solids (VS) (0.0061 g COD (as H₂)/g COD) at pH 6 with 54% VS reduction whereas 0.0041 L H₂/g VS (0.0020 g COD (as H₂)/g COD) was produced at pH 9 with 86% VS reduction. The results show that the hydrogen produced should be released rapidly from the reactor before it can be consumed in other biochemical reactions, and substrates with high pH level (>9.0) can be used directly to produce hydrogen without needing to adjust the pH.     

Decreasing methane production in hydrogenogenic UASB reactors fed with cheese whey

One of the problems in fermentative hydrogen producing reactors, inoculated with pre-treated anaerobic granular sludge, is the eventual methane production by hydrogen-consuming methanogens. In this study, strategies such as reduction of pH and HRT, organic shock loads and repeated biomass heat treatment were applied to hydrogenogenic UASB reactors fed with cheese whey, that showed methane production after certain time of continuous operation (between 10 and 60 days). The reduction of pH to 4.5 not only decreased methane production but also hydrogen production. Organic shock load (from 20 to 30 g COD/L-d) was the more effective strategy to decrease the methane production rate (75%) and to increase the hydrogen production rate (172%), without stopping reactor operation. Repeated heat treatment of the granular sludge was the only strategy that inhibited completely methane production, leading to high volumetric hydrogen production rates (1.67 L H₂/L-d), however this strategy required stopping reactor operation; in addition homoacetogenesis, another hydrogen-consuming pathway, was not completely inhibited. This work demonstrated that it was possible to control the methane activity in hydrogen producing reactors using operational strategies.     

Bioproduction of hydrogen from food waste by pilot-scale combined hydrogen/methane fermentation

A pilot-scale two-phase hydrogen/methane fermentation system generated 3.9 L biogas per unit time and reactor volume from food waste, of which the fraction of H₂ was approximately 60% at a hydraulic retention time (HRT) of 21 h. As substrate, 90% of the carbohydrates in the organic compounds were consumed, based on COD removal efficiency, and the hydrogen yield was approximately 1.82 (H₂-mol/glucose-mol). The maximum decomposition rate coefficient of hydrogen fermentation was observed at an HRT of 21 h, indicating that reducing HRTs improves hydrogen production. Over 80% of the methane was produced in the methane fermentation tank and the predominant fraction of organic acids after methane fermentation comprised acetic acid. Based on our economic evaluation, two-phase hydrogen/methane fermentation has greater potential for recovering energy than methane-only fermentation.     

Effect of hydraulic retention time on the hydrogen production in a horizontal and vertical continuous stirred-tank reactor

Hydraulic retention time (HRT) is the main process parameter for biohydrogen production by anaerobic fermentation. This paper investigated the effect of the different HRT on the hydrogen production of the ethanol-type fermentation process in two kinds of CSTR reactors (horizontal continuous stirred-tank reactor and vertical continuous stirred-tank reactor) with molasses as a substrate. Two kinds of CSTR reactors operated with the organic loading rates (OLR) of 12kgCOD/m3•d under the initial HRT of the 8 h condition, and then OLR was adjusted as 6kgCOD/m3•d when the pH drops rapidly. The VCSTR and HCSTR have reached the stable ethanol-type fermentation process within 21 days and 24 days respectively. Among the five HRTs settled in the range of 2–8 h, the maximum hydrogen production rate of 3.7LH₂/Ld and 5.1LH₂/Ld were investigated respectively in the VCSTR and HCSTR. At that time the COD concentration and HRT were 8000 mg/L and 5 h for VCSTR, while 10000 mg/L and 4 h for HCSTR respectively.Through the analysis on the composition of the liquid fermentation product and biomass under the different HRT condition in the two kinds of CSTR, it can found that the ethanol-type fermentation process in the HCSTR is more stable than VCSTR due to enhancing biomass retention of HCSTR at the short HTR.     

Non-sterile bio-hydrogen fermentation from food waste in a continuous stirred tank reactor (CSTR): Performance and population analysis

Bio-hydrogen production from food waste by anaerobic mixed cultures was conducted in a continuous stirred tank reactor (CSTR). The hydraulic retention time (HRT) was optimized in order to maximize hydrogen yield (HY) and hydrogen production rate (HPR). The maximum hydrogen content (38.6%), HPR (379 mL H₂/L. d) and HY (261 mL H₂/g-VS_added) were achieved at the optimum HRT of 60 h. The major soluble metabolite products were butyric and acetic acids which indicated a butyrate-acetate type fermentation. Operation of CSTR at HRT 60 h could select hydrogen producing bacteria and eliminate lactic acid bacteria and acetogenic bacteria. The microbial community analyzed by polymerase chain reaction-denaturing gradient gel electrophoresis (PCR-DGGE) revealed that the predominant hydrogen producer was Clostridium sp.     

Steady-state and dynamic modeling of biohydrogen production in an integrated biohydrogen reactor clarifier system

Steady-state operational data from the integrated biohydrogen reactor clarifier system (IBRCS) during anaerobic treatment of glucose-based synthetic wastewater at HRT of 8 h and SRT ranging from 26 to 50 h and organic loading rates of 6.5–206 gCOD/L-d were used to calibrate and verify a process model of the system developed using BioWin. The model accurately predicted biomass concentrations in both the bioreactor and the clarifier supernatant with average percentage errors (APEs) of 4.6% and 10%, respectively. Hydrogen production rates and hydrogen yields predicted by the model were in close agreement with the observed experimental results as reflected by an APE of less than 4%, while the hydrogen content was well correlated with an APE of 10%. The successful modeling culminated in the accurate prediction of soluble metabolites, i.e. volatile fatty acids in the reactor with an APE of 14%. The calibrated model confirmed the advantages of decoupling of the solids retention time (SRT) from the hydraulic retention time (HRT) in biohydrogen production, with the average hydrogen yield decreasing from 3.0 mol H₂/mol glucose to 0.8 mol H₂/mol glucose upon elimination of the clarifier. Dynamic modeling showed that the system responds favorably to short-term hydraulic and organic surges, recovering back to the original condition. Furthermore, the dynamic simulation revealed that with a prolonged startup periods of 10 and 30 days, the IBRCS can be operated at an HRT of 4 h and OLR as high as 206 gCOD/L-d without inhibition and/or marked performance deterioration.     

Pilot-scale hydrogen fermentation system start-up performance

A high-rate hydrogen production process able to produce H₂ at a maximum rate of 15 L/L/h was successfully developed by the Feng Chia University (FCU) biohydrogen research team. This highly efficient hydrogen fermentation system includes a 400 L pilot-scale system constructed for determining scale-up operation parameters for commercializing the bioH₂ production technology. The pilot-scale system is composed of a feedstock tank, mixing system, fermentor, gas/liquid separator and automatic control system. The fermentor is fed with sucrose (20 g COD/L) and operated at 35 °C. A batch strategy is used for system start-up. The fermentor was first operated in a batch mode for two days and then switched to a continuous-feeding mode (HRT 12 h) for one month. During the continuous operation, pH notably affected H₂ production efficiency and bacterial community. For the first 14-day operation, the H₂ production rate increased from 0.017 to 0.256 L/L/h with a pH variation from 5.0 to 7.0. The DGGE results indicate the presence of two Clostridium species (namely, Clostridium butyricum and Clostridium pasteurianum) in the fermenter. Stable hydrogen production rate was obtained at pH 5.5–6.0 when C. pasteurianum became dominant in the mixed culture.