Bioresource Technology 100 (2009) 5472–5477

Contents lists available at ScienceDirect

Bioresource Technology

journal homepage: www.elsevier .com/ locate/bior tech

Swine manure fermentation for hydrogen production

Jun Zhu a,*, Yecong Li a, Xiao Wu a, Curtis Miller a, Paul Chen b, Roger Ruan b

a Southern Research and Outreach Center, University of Minnesota, 35838 120th Street, Waseca, MN 56093, USAb Bioproducts and Biosystems Engineering Department, University of Minnesota, 1390 Eckles Avenue, St. Paul, MN 55108, USA

a r t i c l e i n f o

Article history:Received 5 June 2008Received in revised form 26 November 2008Accepted 27 November 2008Available online 20 January 2009

Keywords:Hydrogen productionFermentationSwine manureBioenergy

0960-8524/$ - see front matter � 2008 Elsevier Ltd. Adoi:10.1016/j.biortech.2008.11.045

* Corresponding author. Tel.: +1 507 837 5625; faxE-mail address: zhuxx034@umn.edu (J. Zhu).

a b s t r a c t

Biohydrogen fermentation using liquid swine manure as substrate supplemented with glucose was inves-tigated in this project. Experiments were conducted using a semi-continuously-fed fermenter (8 L in totalvolume and 4 L in working volume) with varying pHs from 4.7 through 5.9 under controlled temperature(35 ± 1 �C). The hydraulic retention time (HRT) tested include 16, 20, and 24 h; however, in two pH con-ditions (5.0 and 5.3), an additional HRT of 12 h was also tried. The experimental design combining HRTand pH provided insight on the fermenter performance in terms of hydrogen generation. The results indi-cated that both HRT and pH had profound influences on fermentative hydrogen productivity. A rising HRTwould lead to greater variation in hydrogen concentration in the offgas and the best HRT was found to be16 h for the fermenter in this study. The best pH value in correspondence to the highest hydrogen gen-eration was revealed to be 5.0 among all the pHs studied. There was no obvious inhibition on hydrogenproduction by methanogenesis when methane content in the offgas was lower than 2%. Otherwise, aninverse linear relationship between hydrogen and methane content was observed with a correlation coef-ficient of 0.9699. Therefore, to increase hydrogen content in the offgas, methane production has to be lim-ited to below 2%.

� 2008 Elsevier Ltd. All rights reserved.

1. Introduction

Hydrogen is a clean energy source that has the potential to re-place the fossil fuel in the global energy supply (Midilli and Dincer,2007). However, for this transition to happen, hydrogen must bemade from renewable resources rather than from fossil fuel basedsources such as natural gas (Hawkes et al., 2007). Currently,approximately 95% of the commercially produced hydrogen comesfrom carbon-containing raw materials, primarily fossil in origin(Elam et al., 2003). The nature of these processes has, however,ultimately defined that they cannot be sustainable because ofrequiring electricity derived from fossil fuel combustion to operate,thus by no means lessening our reliance on the consumption ofpetroleum-based energy sources. Therefore, it is ultimately impor-tant to find other avenues in producing renewable hydrogen.

One of the methods to produce hydrogen in a sustainable way isdark fermentation, which has been studied since 1960s (Nandi andSengupta, 1998). Reviewing literature indicates that most reportedstudies, especially the early ones conducted by microbiologists, arefocused on using pure cultures (Nandi and Sengupta, 1998). In thepast two decades, this research has made its way to many otherdisciplines, especially in the environmental engineering fieldwhere hydrogen production by mixed cultures has attracted more

ll rights reserved.

: +1 507 835 3622.

attention because the process will not only produce hydrogen butalso accomplish wastewater reuse and treatment (Valdez-Vazquezet al., 2005). The waste materials that have been investigated todate for hydrogen fermentation include municipal solid wastes(Lay et al., 1999), food wastes (a mixture of grains, vegetables,and meats) (Han and Shin, 2004), food processing wastewater froma tofu plant (Zhu et al., 2002), rice winery wastewater (Yu et al.,2002), sweet potato starch residues (Yokoi et al., 2001), paper millwastes (Valdez-Vazquez et al., 2005), and wheat starch co-product(Hussy et al., 2003). Nonetheless, none of the above studies are re-lated to hydrogen production using liquid swine manure as a sub-strate material. With the fast growth of animal industry in theUnited States and around the world, the increasing amount ofmanure produced from animal operations has provided an almostinexhaustible renewable source for potential hydrogen fermenta-tion. Therefore, research effort in this area is vital in building a sus-tainable economy and environment.

The objective of this project is to investigate the feasibility ofusing swine manure as substrate supplemented with glucose as asugar source to produce biohydrogen. The process used a semi-continuously-fed fermenter with four different hydraulic retentiontimes (HRT) and five pHs. The fermenter performance was exam-ined based on its efficiency in producing hydrogen under differentHRTs and pHs and the best condition combining HRT and pH wasidentified. Some inhibitory factors such as methane detected inthe offgas were also discussed.

J. Zhu et al. / Bioresource Technology 100 (2009) 5472–5477 5473

2. Methods

2.1. Fermenter configuration

A fermentation system (Fig. 1) was constructed for the experi-ment that included a fermenter body using a polyethylene jar pur-chased from the Fisher Scientific Company (catalog number: 11-823-18; Hanover Park, IL 60103), which had a working volume of4 L and a total volume of 8 L, peristaltic pumps that controlledthe influent and effluent of the fermenter (catalog number: 900-0857; Barnant Company, Butler, NJ 07405) and were operated bya programmable control module (catalog number: CR1000, Camp-bell Scientific, Inc., Logan, Utah 84321), a hot plate stirrer on whichthe fermenter was placed (catalog number 84303-20, Coleparmer,Vernon Hills, IL 60061) that not only provided mixing to but alsomaintained the temperature of the fermenter content, and apH/oxidation–reduction potential (ORP) controller with a self-cleaning pH electrode (catalog number: EW-27003-02, Coleparm-er, Vernon Hills, IL 60061) to keep the pH of the fermenter contentwithin the test range by adding either hydroxide (1.0 M NaOH) oracid (1.0 M HCl) to the liquid through a peristaltic pump duringmixing. The mixing speed was controlled at 200 rpm while thetemperature was maintained at 35 ± 1 �C.

2.2. Substrate preparation and fermenter startup

The substrate for fermentation consisted primarily of liquidswine manure supplemented with glucose and limited amountsof trace minerals to assist the growth of hydrogen-producing bac-teria, mainly Clostridium sp. The added nutrients in mg L-1 include:CoCl2 � 5H2O, 0.125; FeSO4, 55; Glucose, 10,000; K2HPO4, 125;MgCl2 � 6H2O, 100; NaHCO3, 600; and NH4HCO3, 500. Selection of10 g L�1 glucose concentration was based on findings reported byHawkes et al. (2002), in which the hydrogen productivity wasfound to be around 1.61–2.36 mol H2 mol�1 glucose for Clostridium

Fig. 1. Fermenter system setup

sp. The seed sludge was obtained from an operating dairy manureanaerobic digester located in St. Peter, MN, and a volume of 100 mLwas boiled for 15 min to inactivate hydrogenotrophic bacteria andto harvest anaerobic spore-forming bacteria such as Clostridium sp.Fresh manure was used as the feedstock for the fermenter and wasretrieved from a finishing building at the University of MinnesotaSouthern Research and Outreach Center at Waseca, which wasequipped with a pull-plug manure handling system. The collectedmanure underwent a total solids adjustment with tap water toachieve a solids content of around 1% (�0.8% volatile solids). Theprepared manure has a pH value of 7.56 and contains 1.05% totalsolids (TS), 0.78% total volatile solids (TVS), 0.65% total suspendedsolids (TSS), 1542 mg L�1 total Kjeldahl nitrogen (TKN), 788 mg L�1

ammonium nitrogen, 2438 mg L�1 5-day biochemical oxygen de-mand (BOD), 1720 mg L�1 volatile fatty acids (VFAs), and6540 mg L�1 chemical oxygen demand (COD).

For the fermenter startup, the heat-treated inoculum was firstplaced into the fermenter together with the prepared fresh manureinfluent to fill up to the 4 L working volume. The fermenter wasthen sealed airtight and flushed using argon gas for 1 min to re-move the oxygen left in the headspace and create an anaerobicenvironment. The startup period was considered complete whenthe gas from the fermenter was detected by the gas meter. Oncestarted, the fermenter was stirred constantly at 200 rpm, exceptin the settling and withdrawal phases (about 20 min total), to en-sure a thorough mixing and to facilitate rapid diffusion of hydro-gen. The length of test for each HRT was 14 days, which wasdetermined by preliminary trials (unpublished data) showing thata stable fermenter operation featuring constant gas productioncould normally be achieved within that period.

2.3. Experimental design and gas sampling

A 3 � 5 factorial design consisting of two variables (hydraulicretention time and pH) was adopted for the experiment. The three

for the biohydrogen study.

Fig. 2. Average daily biogas production by the fermenter at different pH values.

5474 J. Zhu et al. / Bioresource Technology 100 (2009) 5472–5477

HRTs studied were 16, 20, and 24 h and the five pH tested were 4.7,5.0, 5.3, 5.6, and 5.9. An additional HRT (12 h) was also experi-mented at a later stage at selected pHs showing the best hydrogenproductivity (pH 5.0 and 5.3). The goal of this design was aimed atproviding a holistic picture of the impact of combining HRT and pHon the fermenter performance in hydrogen production. Once thefermenter entered steady state, it was operated in a continuousmode and was fed four times during one HRT at every 4, 5, or6 h, which was in accordance with HRT 16, 20, and 24 h. The steadystate was considered established after the daily gas volume pro-duced reached a relatively constant value (within a variation of5–10%). For each feeding, one liter of the reactor content was re-moved followed by the same amount of influent added. The reactorwas stirred constantly, except during the settling and feeding peri-ods. As indicated early, the fermenter temperature was maintainedat 35 �C ± 1 �C for all the experiments in this study.

Gas sampling was conducted every other HRT at the biogas dis-charge port of the fermenter throughout the experiment and thegas volume was recorded daily by a wet gas meter (model: TG-05, Calibrated Instrument, Inc., New York). In addition, the biogasinside the fermenter was released continuously to a samplingbag to keep a low hydrogen partial pressure in the headspace,which was considered having an inhibitory effect on hydrogengeneration.

2.4. Analytical methods

Biogas composition was analyzed using a gas chromatograph(CP-4900 QUAD MICRO-GC, Varian Inc., Palo Alto, CA 94304) thatwas equipped with an ultra low volume thermal conductivitydetector (TCD) and two columns (molecular sieve 5A and PorapakQ), with the former for analyzing gaseous hydrogen (H2), oxygen(O2), methane (CH4), carbon monoxide (CO), and nitrogen (N2),and the latter for carbon dioxide (CO2). The operating conditionsfor the micro-GC were 10 s for stabilization, 100 ms for sampleinjection, 30 s for sampling, 120 s for running, and 8 s for backflu-shing. The temperatures for the sampling line, columns, and theinjector were set at 50, 80, and 110 �C, respectively. Argon at apressure of 4.2 kg/cm2 was used as the carrier gas and its flow ratewas automatically controlled by the micro-GC.

Total solids (TS), total volatile solids (TVS), total suspended sol-ids (TSS), 5-day biochemical oxygen demand (BOD5), ammonium,and chemical oxygen demand (COD) were analyzed according tothe American Public and Health Association Standard methods(APHA, 1998). Total Kjeldahl nitrogen (TKN) was measured usinga Foss Kjeldahl Analyser following digestion. The concentration ofvolatile fatty acids (VFAs) was measured according to Method8196 in the DR/3000 Spectrophotometer Manual (Hach, 1993).

Fig. 3. Average daily hydrogen concentration in the biogas at different pH values.

3. Results and discussion

3.1. Effect of hydraulic retention time and pH on biogas and hydrogenproduction

Figs. 2 and 3 present information about the changes in averagedaily biogas production, hydrogen concentration, and volume re-lated to hydraulic retention time (HRT) for different pHs. It is inter-esting to note that in most cases, HRT has more influence on thedaily volume of biogas production than pH for HRT equal to orgreater than 16 h (Fig. 2). Except for pH 4.7 and 5.3, roughly an in-verse linear relationship was observed between HRT and averagedaily biogas volume, indicating that the longer the HRT, the lowerthe biogas volume was produced per day. However, when HRT wasreduced to 12 h at pH 5.0 and 5.3, the volume of biogas was eithersignificantly reduced (pH 5.0) or virtually unchanged (pH 5.3), sug-

gesting that the best HRT for this particular fermenter using swinemanure as substrate to produce biohydrogen was 16 h. This resultevinces that although shorter HRT might be beneficial to enhancingthe productivity of hydrogen from the fermenter studied in thisproject, too short an HRT may be detrimental to the system possi-bly due to the potential for washout of hydrogen-producing bacte-ria, thus leading to the reduced volume of biogas generated.

Considering the data in Fig. 3, it is clear that HRT also has a pro-found influence on the hydrogen content in the offgas. The highesthydrogen concentration (35.8–37.6%) occurred when HRT fell be-tween 16 and 20 h at a pH value of 5.0. At all pH values except5.6, the 16 h HRT achieved the highest hydrogen concentrations.At pH 5.6, the highest percent hydrogen concentration was ob-served at HRT 20 h, which was around 22.4%. Combining data fromboth figures (Figs. 2 and 3) has evidenced that pH 5.0 is the mosteffective pH in assisting the hydrogen-producing consortium inthe fermenter system to produce biogas and biohydrogen. There-fore, in this particular study, the best operating condition wasfound to be at pH 5.0 and HRT 16 h when the highest biogas vol-

Fig. 5. The range of hydrogen content variation in daily biogas at different pHs andHRTs.

J. Zhu et al. / Bioresource Technology 100 (2009) 5472–5477 5475

ume production in conjunction with the highest hydrogen contentin the biogas was achieved.

Literature review shows that pH and HRT have been claimed tobe able to govern the process efficiency of continuous hydrogenproduction (Afschar and Schaller, 1991; Bahl et al., 1982; Holtet al., 1988; Ueno et al., 1996). However, research into the role thatHRT plays in enhancing the biohydrogen generation via fermenta-tion has not been extensive with relatively limited informationavailable in the literature. Zhu et al. (2002) reported that shortretention times (<4 h) were advantageous to produce hydrogen be-cause they made it difficult for methanogens to populate. By usingrice winery wastewater as substrate for an upflow anaerobic reac-tor, Yu et al. (2002) employed a range of HRT (from 2 to 24 h) toexamine its effect on hydrogen production. Their results indicatedthat the influence of HRT on hydrogen yield was limited but wassignificant on the specific hydrogen production rate, with shorterHRTs corresponding to higher production rates. Another study byChen et al. (2001) using sucrose as substrate for hydrogen produc-tion by anaerobic cultures also revealed that the hydrogen produc-tion rate decreased from 0.094 to 0.032 mol h�1 when the HRTincreased from 6 to 13.3 h. Although these researchers used differ-ent substrate materials other than liquid swine manure, the gen-eral conclusion that reducing HRT might be instrumental inimproving hydrogen productivity was consistent with the findingsobserved in this study.

3.2. Effect of hydraulic retention time on system stability and purehydrogen gas volume

The variation of biogas volume and its percent hydrogen con-tent are presented in Figs. 4 and 5. As discussed early, the best con-dition for the hydrogen fermenter studied here was at pH 5.0 andHRT 16 h. However, it was also a condition under which the largestvariation (21.4–38.3 L d�1) in biogas volume production was ob-served (Fig. 4). For other pHs under HRT 16 h, the variations in bio-gas volume are smaller (4.7–12.9 L d�1 at pH 4.7; 7.8–10.8 L d�1 atpH 5.3; 14.0–16.9 L d�1 at pH 5.6; 10.8–12.6 L d�1 at pH 5.9). De-spite the large biogas volume variation, it is recognized that eventhe minimal gas production at pH 5.0 and HRT 16 h (21.4 L d�1)is higher than the upper limits of all other volumes produced underall HRTs. Since pure hydrogen volume generated by the system isthe product of biogas volume and its percent hydrogen content,the high volume of biogas produced under this experimental con-dition is thus significant in achieving greater hydrogen productiv-

Fig. 4. The range of daily biogas volume variation under different pHs and HRTs.

ity. More importantly, if Fig. 5 is referenced, it is not difficult toidentify that under the operating condition of pH 5.0, HRT 16 hhas achieved the lowest variation in percent hydrogen content inthe biogas produced by the fermenter among all other HRTs, indi-cating a good condition for producing hydrogen with stable con-centration. In this case, the average pure hydrogen volumegenerated daily is 27.4 L d�1 � 35.8% = 9.81 L d�1 = 876 mg d-1, or18.7 mg g�1 – TVS. Considering the fermenter size, which has aworking volume of 4 L with an HRT of 16 h, the product/substrateratio has therefore reached 0.876 g H2/46.8 g TVS = 18.7 � 10�3 gH2 per g TVS. In other words, running 6 L of liquid swine manurethrough the fermenter can lead to 9.81 L of pure hydrogen beingproduced.

One caveat that needs to be brought up regarding the experi-mental results is the use of glucose as a supplemental nutrient toassist in the hydrogen-producing bacterial growth, which can bea major economic barrier with respect to any scale-up operationto produce biohydrogen in a commercial setting because of thecost associated with the glucose addition. Nonetheless, this doesnot mean that the process experimented and reported herein hasno sense of practical applications. It only warrants more researchwork that should look into co-fermenting liquid swine manurewith other sugar-rich waste streams such as sugar-processingwastewaters. A recent study done by Van Ginkel et al. (2005)showed that apple and potato processing wastewaters that wererich in sugar scored high in hydrogen production through fermen-tation. But they did acknowledge that external nitrogen and phos-phorus sources were added to overcome the shortage of thesenutrients in the wastewater used. Therefore, it is not unreasonableto infer that coupling liquid swine manure with food processingwastewater could bring together the fortes of both wastewaterstreams (one rich in nitrogen and phosphorus and the other in car-bohydrates) to make a perfect substrate for hydrogen fermenta-tion. An added benefit for such a treatment rests with itscapability of treating two wastewaters simultaneously withoutlosing the efficiency and economical feasibility of producing biohy-drogen from renewable resources. Further research effort certainlyshould follow in this direction.

3.3. Influence of methane content on hydrogen production

In a hydrogen generation fermenter, methanogenesis should beavoided because according to the methane molecular formula, two

Fig. 6. The methane percent content under different values of HRT and pH.

5476 J. Zhu et al. / Bioresource Technology 100 (2009) 5472–5477

moles of hydrogen need to be consumed for every mole of methaneproduced. As such, when using mixed cultures enriched from nat-ural environments as the fermentative working force, it is neces-sary to avoid the presence of organisms utilizing H2, such asmethanogens. Efforts have been exercised to prevent this conditionfrom happening by past researchers by operating the fermenter atlow pH for batch operation and short HRTs for continuous opera-tion because methanogens are more affected by lower pH andare slower growing than fermentative organisms (Hawkes et al.,2002). The results from this study generally are consistent withthose from previous reports in that the methane content in the bio-gas increases with the increase in fermenter pH (from 5.0 to 5.9)under the same HRT (Fig. 6). With respect to HRT, 16 and 20 hshowed the highest methane productivity at pH 5.9, while themethane productivity was low for all HRTs at pH 5.0, implying thatmethanogens could survive in higher pH and were less vulnerableto the changes in HRT (except under HRT 24 h and pH 5.9 for whichthe reason was unknown). Another phenomenon observed in thisstudy is somewhat revealing the relationship between methaneand hydrogen concentrations in the biogas (Fig. 7). When themethane content was kept below 2%, its presence had an insignif-icant impact on hydrogen concentration. However, when themethane concentration exceeded the 2% mark, an inverse linearrelationship between the concentrations of methane and hydrogenwas found with a correlation coefficient of 0.9699. This observationagain evidences that to maintain a healthy and efficient hydrogen

Fig. 7. The relationship between %CH4 and %H2.

fermentation, eliminating methane production is critical for theprocess. Although 2% is suggested as the maximum level of meth-ane that can be present in the fermenter based on this study, addi-tional information is still needed to verify the finding in order tobetter understand this limiting factor for hydrogen productionvia fermentation.

4. Conclusion

The HRT has profound impact on biogas production for all pHsexcept 4.7 and 5.3 and the best HRT is 16 h at pH 5.0 when thehighest biogas production (21.4–38.3 L d�1) and H2 concentrationin the biogas (35.8–37.6%) are obtained. The product/substrate ra-tio for the fermenter is 18.7 � 10�3 g H2 per g TVS, i.e., for every 6 Lof manure, 9.81 L of pure hydrogen gas can be generated. For effi-cient hydrogen gas production, the methane content in the fer-menter needs to be controlled at <2%. Otherwise, an inverselinear relationship (R = 0.9699) may exist.

Acknowledgements

The authors wish to thank the University of Minnesota Initia-tives for Renewable Energy and Environment for providing thefunding for this project.

References

Afschar, A.S., Schaller, K., 1991. Production of acetone and butanol from starch bycontinuous bioprocess. Journal of Biotechnology 18 (3), 255–263.

APHA, 1998. Standard Methods for the Examination of Water and Wastewater.American Public Health Association, Washington DC.

Bahl, H., Andersch, W., Braun, K., Gottschalk, G., 1982. Effect of pH and butyrateconcentration on the production of acetone and butanol by clostridiumacetobutylicum grown in continuous culture. European Journal of AppliedMicrobiology and Biotechnology 14, 17–20.

Chen, C.C., Lin, C.Y., Chang, J.S., 2001. Kinetics of hydrogen production withcontinuous anaerobic cultures utilizing sucrose as the limiting substrate.Applied Microbiology and Biotechnology 57 (1–2), 56–64.

Elam, C.C., Padro, C.E.G., Sandrock, G., Luzzi, A., Lindblad, P., Hagen, E.F., 2003.Realizing the hydrogen future: the international energy agency’s efforts toadvance hydrogen energy technologies. International Journal of HydrogenEnergy 28 (6), 601–607.

Hach, 1993. Procedure Manual. Loveland, CO 80539.Han, S.K., Shin, H.S., 2004. Biohydrogen production by anaerobic fermentation of

food waste. International Journal of Hydrogen Energy 29 (6), 569–577.Hawkes, F.R., Dinsdale, R., Hawkes, D.L., Hussy, I., 2002. Sustainable fermentative

hydrogen production: challenges for process optimisation. International Journalof Hydrogen Energy 27 (11–12), 1339–1347.

Hawkes, F.R., Hussy, I., Kyazze, G., Dinsdale, R., Hawkes, D.L., 2007. Continuousdark fermentative hydrogen production by mesophilic microflora:principles and progress. International Journal of Hydrogen Energy 32 (2),172–184.

Holt, R.A., Cairns, A.J., Morris, J.G., 1988. Production of butanol by Clostridiumpuniceum in batch and continuous culture. Applied Microbiology andBiotechnology 27 (4), 319–324.

Hussy, I., Hawkes, F.R., Dinsdale, R., Hawkes, D.L., 2003. Continuous fermentativehydrogen production from a wheat starch co-product by mixed microflora.Biotechnology and Bioengineering 84 (6), 619–626.

Lay, J.J., Lee, Y.J., Noike, T., 1999. Feasibility of biological hydrogen productionfrom organic fraction of municipal solid waste. Water Research 33 (11),2579–2586.

Midilli, A., Dincer, I., 2007. Key strategies of hydrogen energy systems forsustainability. International Journal of Hydrogen Energy 32 (5), 511–524.

Nandi, R., Sengupta, S., 1998. Microbial production of hydrogen: an overview.Critical Reviews in Microbiology 24 (1), 61–84.

Ueno, Y., Otsuka, S., Morimoto, M., 1996. Hydrogen production from industrialwastewater by anaerobic microflora in chemostat culture. Journal ofFermentation and Bioengineering 82 (2), 194–197.

Valdez-Vazquez, I., Rios-Leal, E., Esparza-Garcia, F., Cecchi, F., Poggi-Varaldo, H.A.,2005a. Semi-continuous solid substrate anaerobic reactors for h-2 productionfrom organic waste: mesophilic versus thermophilic regime. InternationalJournal of Hydrogen Energy 30 (13–14), 1383–1391.

Valdez-Vazquez, I., Sparling, R., Risbey, D., Rinderknecht-Seijas, N., Poggi-Varaldo,H.M., 2005b. Hydrogen generation via anaerobic fermentation of paper millwastes. Bioresource Technology 96 (17), 1907–1913.

Van Ginkel, S.W., Oh, S.E., Logan, B.E., 2005. Biohydrogen gas production from foodprocessing and domestic wastewaters. International Journal of HydrogenEnergy 30 (15), 1535–1542.

J. Zhu et al. / Bioresource Technology 100 (2009) 5472–5477 5477

Yokoi, H., Saitsu, A., Uchida, H., Hirose, J., Hayashi, S., Takasaki, Y., 2001. Microbialhydrogen production from sweet potato starch residue. Journal of Bioscienceand Bioengineering 91 (1), 58–63.

Yu, H.Q., Zhu, Z.H., Hu, W.R., Zhang, H.S., 2002. Hydrogen production from ricewinery wastewater in an upflow anaerobic reactor by using mixed anaerobiccultures. International Journal of Hydrogen Energy 27 (11–12), 1359–1365.

Zhu, H.G., Ueda, S., Asada, Y., Miyake, J., 2002. Hydrogen production as a novelprocess of wastewater treatment – studies on tofu wastewater with entrappedr-sphaeroides and mutagenesis. International Journal of Hydrogen Energy 27(11–12), 1349–1357.