Improving the yield from fermentative hydrogen production - CiteSeerX

Biotechnol Lett (2007) 29:685–695 DOI 10.1007/s10529-006-9299-9

REVIEW

Improving the yield from fermentative hydrogen production Jeremy T. Kraemer Æ David M. Bagley

Received: 4 October 2006 / Revised: 20 December 2006 / Accepted: 21 December 2006 / Published online: 6 February 2007 Springer Science+Business Media B.V. 2007

Abstract Efforts to increase H2 yields from fermentative H2 production include heat treatment of the inoculum, dissolved gas removal, and varying the organic loading rate. Although heat treatment kills methanogens and selects for spore-forming bacteria, the available evidence indicates H2 yields are not maximized compared to bromoethanesulfonate, iodopropane, or perchloric acid pre-treatments and spore-forming acetogens are not killed. Operational controls (low pH, short solids retention time) can replace heat treatment. Gas sparging increases H2 yields compared to un-sparged reactors, but no relationship exists between the sparging rate and H2 yield. Lower sparging rates may improve the H2 yield with less energy input and product dilution. The reasons why sparging improves H2 yields are unknown, but recent measurements of dissolved H2 concentrations during sparging suggest the assumption of decreased inhibition of the H2–producing enzymes is unlikely. Significant

J. T. Kraemer (&) Department of Civil Engineering, University of Toronto, 35 St. George Street, Toronto, ON, Canada M5S 1A4 e-mail: [email protected] D. M. Bagley Department of Civil and Architectural Engineering, University of Wyoming, 1000 E. University Ave. Dept. 3295, Laramie, WY, USA 82071

disagreement exists over the effect of organic loading rate (OLR); some studies show relatively higher OLRs improve H2 yield while others show the opposite. Discovering the reasons for higher H2 yields during dissolved gas removal and changes in OLR will help improve H2 yields. Keywords Carbon dioxide Dissolved gases Heat treatment Hydrogen Organic loading rate Sparging

Introduction Biological H2 production via dark fermentation of organic wastes is being investigated as a potential source of renewable energy (Hawkes et al. 2002). Fermentative H2 production is carried out by anaerobic bacteria that ferment organic compounds to volatile fatty acids (VFAs), alcohols, CO2, and H2. Many different substrates can be fermented to produce H2 (Kapdan and Kargi 2006; Nishio and Nakashimada 2004), although carbohydrates (e.g. glucose) have been most commonly used. The reactor environmental conditions capable of achieving H2 production are well understood, as demonstrated by several recent reviews (Hallenbeck 2005; Hawkes et al. 2002; Nath and Das 2004). For example, completely-mixed reactors with a pH of 5.5 and solids retention time of 6–12 h can achieve H2 production.

123

686

Biotechnol Lett (2007) 29:685–695

The H2 yield is the moles of H2 produced per mole of substrate. Care must be taken when comparing studies because some authors report the yield per mol of substrate converted whereas others report per mol of substrate applied to the reactor. Additionally, the H2 yield is not applicable to complex substrates where ‘‘moles of substrate’’ cannot be measured (e.g. sludge). Kraemer and Bagley (2005) addressed these issues by proposing the use of the H2 productivity (HP), which they defined as the percent of influent substrate electrons distributed to H2 gas (gaseous + dissolved phases). The maximum H2 yield from fermentative H2 production is 4 mol H2/mol glucose (HP = 33%), which occurs if acetic acid is the only VFA produced and no electrons are used for growth (Angenent et al. 2004). The H2 yield is lower if other metabolites are produced, such as butyric acid (2 mol H2/mol glucose, HP = 17%) or ethanol (0 mol H2/mol glucose, HP = 0%). Thus, production of VFAs is preferred over alcohols and acetate is preferred over butyrate. Substrate electrons contained in the residual fermentation products can be converted into methane in a second-stage reactor (Kraemer and Bagley 2005). Current H2 productivities reported in the literature are 10-20% (Angenent et al. 2004; Benemann 1996; Logan 2004), much less than the theoretical maximum of 33%. Higher H2 yields would be beneficial for practical application of the technology. This paper reviews the efficacy of three operational techniques for increasing the yield from

Fig. 1 Interactions between fermenter, acetogen, and methanogen metabolic groups in a fermentative H2-producing system. KLa is the overall masstransfer coefficient assuming the liquid-phase is mass-transfer limiting

fermentative H2 production: (1) inoculum heat treatment, (2) dissolved gas removal, and (3) varying the organic loading rate. Additionally, this review identifies several recommendations for practical application and research.

Inoculum heat treatment In a reactor operated for fermentative H2 production, three groups of H2 metabolising bacteria appear to be important (Fig. 1): H2 producers (fermenters), H2-consuming methanogens, and H2-consuming acetogens. Using molecular phylogenetic techniques, researchers have identified Clostridium spp. as being the most common in continuous-flow bioreactors engineered for fermentative H2 production (Table 1). In addition, Bacillus, Enterobacter, and Thermoanaerobacterium spp. have also been regularly observed, but less frequently than Clostridium spp. Methanogenesis (Kraemer and Bagley 2005; Shizas and Bagley 2005) and acetogenesis (Park et al. 2005) have also been observed. The key to high H2 yields is to optimize the conditions for the H2 fermenters while preventing the methanogens and acetogens from consuming the produced H2. Heat treatment has been a common practice for killing methanogens contained in inocula (Ahn et al. 2005; Kim et al. 2006a, b; Oh et al. 2003a). Heat treatment at 80–104C and exposure times of 15–120 min have been used (Zhu and Be´land 2006). Heat treatment selects for bacteria

H2,gas

CO2,gas

CH4,gas

headspace liquid

KLaH2 glucose

H2 HCO3–

H2,dissolved CO2,dissolved HCO3–

acetate, butyrate ethanol, butanol acetogen

KLaCH4

KLaCO2

fermenter

CH4,dissolved H2 HCO3– methanogen

– H2 HCO3

acetate

123


687

Table 1 Microbial community assessments using molecular techniques for continuous-flow mixed-culture biohydrogen systems Heat Treatment

Inoculum

Observed Organisms

Study

no

secondary sludge

Fang et al. (2002a)

no

secondary sludge

no no no

digester sludge thermophilic digester sludge compost

yes

digester sludge

yes

soil

yes

digester sludge

yes yes

digester sludge soil

yes

sludge

Clones: 69.1% 4 Clostridium spp., 13.5% Sporolactobacillus racemicus, 5.8% no close relative, 11.5% phylogeny not checked Clones: 64.6% 3 Clostridium spp., 18.8% Enterobacteriaceae, 3.1% Streptococcus bovis, 13.5% 8 unidentified OTUs Methanogens observed Glucose: mostly Bacillus; cellulose: mostly Clostridia 10 of 15 bands either Clostridium or Thermoanaerobacterium, 1 Symbiobacterium, 1 Bacillus, 1 Sulfobacillus, 1 Ruminococcus, 1 Selenomonas Clostridia, Bacilli, Bacteroides; primarily Thermoanaerobacterium thermosaccharolyticum HRT = 30 h: Bacillaceae, Clostridiaceae, Enterobacteriaceae; HRT = 10 h: Clostridiaceae only 15 of 17 bands were Clostridium spp., 1 Gluconacetobacter, 1 Lactobacillus 9 of 10 bands were Clostridia, 1 Bacillus Most related to Clostridiaceae and Flexibacteraceae Prominent band was similar to Clostridium pasteurianum

that can form endospores, such as Clostridium, Bacillus and Thermoanaerobacterium (Table 1). The microbial community after heat treatment is relatively homogeneous, whereas microbial diversity is higher in reactors using non-heat-treated inocula; reactors without heat treatment contain more enteric and other non-spore-forming bacteria (Table 1). However, depending on whether the inoculum is dry or wet and the time and temperature of exposure, not all vegetative cells will be killed by heat treatment. For example, non-spore-forming bacteria such as Bacteroides, Lactobacillus and Enterobacter have been observed even though heat treatment was employed (Table 1) (Ahn et al. 2005; Iyer et al. 2004; Kim et al. 2006a). Although the intent has primarily been prevention of methanogenesis, the use of heat treatment does not select exclusively for H2-producing bacteria. This is because the characteristic of H2 production is not directly associated with the ability to form endospores. For

Fang et al. (2002b)

Shizas and Bagley (2005) Ueno et al. (2001a) Ueno et al. (2001b)

Ahn et al. (2005)

Iyer et al. (2004)

Kim et al. (2006a) Kim et al. (2006b) Oh et al. (2004a) Wu et al. (2006)

instance, non-spore-forming H2-producers include enteric bacteria like Enterobacter spp. (Nakashimada et al. 2002) and Citrobacter spp. (Oh et al. 2003b). There are also many H2consuming groups of bacteria that can form spores and therefore survive heat treatment, including acetogens (Acetobacterium, some Clostridium spp., Sporomusa), certain propionate and lactate producers (Propionibacterium, Sporolactobacillus), and a sulphate-reducer (Desulfotomaculum) (Madigan et al. 2000). For example, Kim et al. (2006b) observed the acetogen Clostridium scatologenes in a heat-treated H2-producing sludge. Current evidence does not prove that heat treatment actually increases the H2 yield compared to non-heat-treated systems. Only a few reports have assessed heat-treated versus nonheat-treated inocula in the same experiment, and all of these were batch studies. Oh et al. (2003a) found the H2 yield was higher with heat treatment compared to non-heat-treated or pH 6.2.

123

688


However, Zhu and Be´land (2006) observed that heat-treatment decreased H2 production compared to the use of bromoethanesulfonate (BES) and iodopropane. Similarly, Cheong and Hansen (2006) observed less H2 produced from wet- and dry-heat-treated sludge compared to BES and perchloric acid treatments. Kawagoshi et al. (2005) observed the same H2 production from non-heat-treated and heat-treated digester sludge. Instead of increasing the H2 yield, heat treatment may actually be detrimental to achieving maximum H2 production. Controlling methanogens by heat treatment is unnecessary because they can be easily controlled through operational means: operating at a pH near 5.5 and employing a solids retention time of 6–12 h can achieve methane-free biogas production (Hussy et al. 2005; Kraemer and Bagley 2006; Oh et al. 2003a). These conditions wash out the methanogens because they grow slower than the fermenting, H2-producing bacteria. One concern with untreated sludge inocula is the possible establishment of a methanogenic biofilm on the reactor walls (Kraemer and Bagley 2005; Shizas and Bagley 2005), but this can be avoided by transferring the sludge to a new vessel after the initial start-up period (Zhu and Be´land 2006).

Fig. 2 Metabolic pathways in Clostridium spp. fermenting glucose. Solids lines indicate substrate transformations, dotted lines indicate ATP creation or utilization, and dashed lines indicate electron flow. ATP = adenosine triphosphate; G3P = glyceraldehyde-3-phosphate; G3PDH = G3P dehydrogenase; 1,3-BPG = 1,3-bisphosphoglycerate; NADH = nicotinamide-adenine dinucleotide; Fd = ferredoxin; PFOR = pyruvate: Fd oxidoreductase; NFOR = NADH:Fd oxidoreductase; H2ase = hydrogenase; CoA = coenzyme A. Adapted from Madigan et al. (2000)

123

Dissolved gas removal The metabolic pathways for the fermentation of glucose by Clostridium spp. are well understood (Fig. 2) (Madigan et al. 2000). Glucose is fermented via glycolysis to pyruvate with electrons being transferred to nicotinamide-adenine dinucleotide (NADH). Pyruvateis oxidized by pyruvate:ferredoxin oxidoreductase (PFOR) to acetylCoA and CO2, with electrons being transferred to ferredoxin (Fd). Several end-products are produced from acetyl-CoA, including acetate, butyrate, ethanol, and butanol. Production of acetate or butyrate allows for ATP generation while alcohols do not, although the more reduced products utilize electrons from NADH thereby maintaining redox balance. NADH can alternatively be re-oxidized by electron transfer to ferredoxin by NADH:Fd oxidoreductase (NFOR). H2 is produced by the hydrogenase enzyme, which catalyzes proton reduction using electrons from ferredoxin. The pool of reduced Fd is generated from two sources: (1) pyruvate oxidation by PFOR and (2) NADH oxidation by NFOR. These enzyme systems can be thermodynamically regulated by the H2 concentration (Angenent et al. 2004): PFOR can function at

glucose ATP G3P G3PDH

NADH

1,3-BPG NFOR

ATP pyruvate

PFOR

H2ase

Fd

acetyl -CoA

ATP CO 2 acetate

H2

ethanol

butyryl- CoA

butyrate

butanol


689

the H2 concentrations observed in fermentative H2 systems (Angenent et al. 2004) so there will always be some H2 produced, whereas NADH oxidation by NFOR is inhibited for H2 >60– 100 Pa (~0.5–0.8 lM) (Angenent et al. 2004; Hallenbeck 2005). Various techniques have been used to remove metabolic gases (H2, CO2) from the liquid phase. Measuring the dissolved concentrations is important when assessing dissolved gas removal techniques because H2 and CO2 are supersaturated in fermentative H2-producing systems and therefore headspace concentrations are inaccurate (Kraemer and Bagley 2006). Gas sparging Gas sparging has been the most common method used to decrease dissolved gas concentrations in fermentative H2-producing reactors. In pure cultures, sparging has altered the relative amounts of metabolic products. Argon and H2 sparging in a culture of Enterobacter aerogenes decreased the production of succinate while increasing acetate (Tanisho et al. 1998). In a culture of Clostridium

butyricum Crabbendam et al. (1985) observed a lower acetate:butyrate ratio when N2 was passed over the fermentation liquid instead of through the liquid. These studies imply more H2 was produced with sparging. Also, the H2 yield increased by 47% when Rhodopseudomonas palustris P4 was intermittently purged with argon (Oh et al. 2002). In mixed cultures, gas sparging has increased the H2 yield noticeably in comparison to unsparged conditions (Table 2). The H2 yield improvement has been 20–70% for N2 sparging, 80–120% for CO2 sparging (discussed later), 88% for methane sparging, and 0–12% for H2/CO2 (biogas) sparging. However, there does not appear to be any relationship between the amount of sparging and the increase in H2 yield (Fig. 3). Even for the study by Kim et al. (2006a), where multiple sparging rates were tested within the same experiment, there was no meaningful relationship between sparging rate and H2 yield, although all sparging rates did provide higher H2 yields than the un-sparged case. These results indicate higher H2 yields may occur at sparging rates lower than those usually used in the liter-

Table 2 Comparison of gas sparging during fermentative H2 production in continuous-flow mixed-culture systems Sparge Gas

Study

Sparge flow, Qs (ml min–1)

Liquid volume, QS/V Yield [mol H2 V (l) (ml min–1 l–1) (mol hexose)–1]

Yield Increase (%)

No With Sparging Sparging N2

CO2

Mizuno et al. (2000) Hussy et al. (2003) Hussy et al. (2005) Kyazze et al. (2006) Kim et al. (2006a)

Kraemer and Bagley (2006) Kim et al. (2006a)

methane Liu et al. (2006) biogas Kim et al. (2006a)

a

110 58 55 630 100 200 300 400 160

2.3 2.3 2.3 9.3 5 5 5 5 2

48 25 24 68 20 40 60 80 80

0.85 1.26 1.00 1.23 0.77 0.77 0.77 0.77 1.30

1.43 1.87 1.80 1.65 0.91 0.92 0.95 0.92 1.80

68 48 66 34 18 19 23 19 38

100 200 300 400 2 100 200 300 400

5 5 5 5 0.4 5 5 5 5

20 40 60 80 5 20 40 60 80

0.77 0.77 0.77 0.77 n/r a 0.77 0.77 0.77 0.77

1.40 1.65 1.68 1.54 n/r 0.86 0.83 0.77 0.82

82 114 118 100 88 12 8 0 6

n/r = not reported

123

690


Increase in H2 Yield (%)

75%

50%

25%

0% 0

15

30

45

60

75

90

Specific Sparging Rate (ml min-1 l-liquid-1)

Fig. 3 Increase in H2 yield versus sparging rate. Data are from Table 2 for N2 sparging

ature. For example, Liu et al. (2006) used methane sparging at only 2 ml/min and observed an 88% increase in H2 yield. Non-sparging techniques Other techniques to decrease dissolved gas concentrations include increased stirring, decreasing the reactor headspace pressure (i.e. applying a vacuum), and using an immersed membrane to directly remove dissolved gases. Vigorous stirring of Clostridium thermocellum cultures decreased the dissolved H2 content by a factor of 3 and the ethanol/acetate ratio decreased compared to unstirred cultures (Lamed et al. 1988). Lay (2000) stated that increased mixing caused a release of H2. Thus, sufficient mixing is important for aiding mass-transfer of metabolic gases from the liquid to the reactor headspace. Kataoka et al. (1997) did not observe any significant difference between continuous cultures of Clostridium butyricum strain SC-E1 at 0.28 atm headspace pressure (vacuum operation) compared to the control (non-vacuum operation). In contrast, Mandal et al. (2006) observed double the H2 yield during batch culture of Enterobacter cloacae DM 11 at 0.5 atm headspace pressure compared to non-vacuum operation. The H2 yield during vacuum operation was 3.9 mol H2/mol glucose, which was extremely high considering enteric bacteria usually produce 2 mol/mol glucose (refer to Fig. 2 and discussion above). However, this has always been an assumption because the dissolved H2 concentrations were never measured. Recently, Kraemer and Bagley (2006) measured dissolved H2 and CO2 concentrations with and without N2 sparging. They demonstrated that although dissolved H2 and CO2 concentrations were significantly lower during sparging, H2 was still 1000fold higher than the regulatory level for the NFOR enzyme. It is possible for N2 sparging to alter the relative amounts of acetate and butyrate in pure-cultures of clostridia, but such a change was reported for sparging rates 10 times higher than those of Table 2 (Crabbendam et al. 1985). Thus, altered H2 thermodynamic regulation is unlikely to have been the cause of higher H2 yield.


691

Decreasing the dissolved concentrations of H2 and CO2 has been proposed for limiting H2 consumption by acetogens thereby increasing the H2 yield (Hussy et al. 2003; Kim et al. 2006a; Kraemer and Bagley 2006; Park et al. 2005). Acetogenesis was observed in batch culture by Park et al. (2005) from a heat-treated inoculum. Kraemer and Bagley (2006) observed decreases of 65% and 90% for dissolved H2 and CO2, so the effective reduction in substrate concentrations (since both gases are needed) would have been 96.5%. However, the authors did not measure VFAs or microbial community so changes in acetogens cannot be substantiated. Additionally, the extent to which acetogenesis actually occurs in continuous cultures is not yet known. High concentrations of CO2 are known to be inhibitory to bacteria and this fact has led to the use of CO2 in food packaging to prevent spoilage (Dixon and Kell 1989). Partial CO2 pressures >0.5 atm have been observed to decrease microbial yield in Clostridium sporogenes (Dixon and Kell 1989) and pure CO2 sparging drastically decreased microbial diversity in a continuous mixed culture (Kim et al. 2006a). Without sparging, Kraemer and Bagley (2006) measured the dis-

solved CO2 concentration to be in the inhibitory region (~0.56 atm) and sparging at 160 ml/min significantly reduced it (~0.06 atm). Thus, it is possible that relatively low rates of sparging could remove CO2 sufficiently to remove its inhibitory effect.

Varying the organic loading rate Changing the organic loading rate (OLR = feed concentration/hydraulic retention time) can increase the H2 yield. Table 3 summarizes studies that used mixed cultures and assessed multiple substrate concentrations at a constant hydraulic retention time (HRT). There is disagreement in the literature as to whether higher H2 yields are achieved with lower or higher OLRs, and this division is shown in Table 3. In some cases higher OLRs decreased the H2 yield whereas in others higher OLRs increased the H2 yield. In the latter case, as OLRs increased the H2 yield usually became constant or eventually began to decrease thereby providing an optimal OLR (maximum H2 yield) (Kim et al. 2006b; Wu et al. 2006; Yang et al. 2006).

Table 3 Comparison of studies that varied the organic loading rate (OLR) by changing the substrate concentration Study

Substrate

Lower OLR improves H2 production Kyazze et al. (2006) sucrose Oh et al. (2004b) glucose Van Ginkel and Logan (2005a) glucose Van Ginkel and Logan (2005b) glucose glucose Wu et al. (2006) sucrose sucrose Yu et al. (2002) rice winery WW Higher OLR improves H2 production Kim et al. (2006b) sucrose Lin et al. (2006) sucrose sucrose Yang et al. (2006) citric acid WW Zhang et al. (2004) glucose

S0 (g COD l–1)

Low

high

11.2 7.3 10.7 2.5 2.5 10 10 14

56.1 29.2 32 10 10 40 40 36

10 5 5 varied 5

60 40 40 varied 15

HRT (h) OLR (g COD l–1 d–1) H2 Yield [mol H2 (mol hexose converted)–1] low

high

low OLR

high OLR

12 4 10 10 2.5 6 2 2

22.4 43.9 25.6 6 24 40 120 168

112.2 175.4 76.8 24 96 160 480 432

1.65 1.30 2.20 2.80 2.40 1.84 2.10 1.89

0.81 1.05 2.00 2.20 1.90 1.36 1.96 1.79

12 8.9 6 varied 4.5

20 13.5 20 10 26.7

120 107.9 160 40 80

0.25 1.69 1.34 0.50 0.72

1.00 2.49 2.17 0.85 1.04

Abbreviations: S0 = influent substrate concentration, HRT = hydraulic retention time, WW = wastewater

123

692

Effects of OLR The reason for such diverse H2 yield observations for lower or higher OLRs is unknown. For those studies where higher OLRs decreased H2 production some possible reasons are: (1) increased inhibition by VFAs at higher OLR, (2) lower dissolved H2 concentrations at lower OLRs decrease thermodynamic regulation, (3) OLR affects acetogenic activity, and (4) lower dissolved CO2 concentrations decrease CO2 inhibition. These hypotheses will each be discussed. VFA inhibition at higher OLRs appears to be a valid explanation. The addition of external VFAs has been demonstrated to decrease or inhibit H2 production in mixed-culture continuous-flow systems, although there is consensus that butyrate causes greater inhibition than acetate. Kyazze et al. (2006) observed inhibition of H2 production when 4 g butyrate/l was added to the reactor (18.9 g butyrate/l in total) whereas no inhibition was observed when only 2 g butyrate/l was added (12.2 g butyrate/l in total). Van Ginkel and Logan (2005a) attributed the inhibition to the undissociated form of the VFAs because changing the pH from 5.5 to 5.0 (increasing the un-ionized fraction of the VFAs) at the same total acid concentration decreased H2 production. However, this hypothesis does not support their results when the OLR was changed. The H2 yield was constant at 2 mol H2/mol glucose for 10–30 g glucose/l (producing 42–116 mM total VFAs) and decreased to 1.6 mol H2/mol glucose at 40 g glucose/l (producing 75 mM total VFAs). Thus, the total VFAs (and un-ionized VFAs) were lower at 40 g/l than at 30 g/l, so undissociated VFAs may not be the only contributing factor to the inhibition of H2 production at higher OLRs. Van Ginkel and Logan (2005b) suggested lower OLRs may decrease the dissolved H2 concentration thereby removing the thermodynamic inhibition on hydrogenase. This is the same hypothesis as discussed above for dissolved gas removal. Although no one has measured dissolved gas concentrations while varying the OLR, it is unlikely that changes in the OLR can decrease the dissolved H2 concentration sufficiently to alter NADH:Fd oxidoreductase thermodynamics. The OLR used by Kraemer and

123


Bagley (2006) was 33 g COD l–1 d–1 for which the dissolved H2 concentration (without sparging) was 1.385 lM. The lowest OLR used to date has been 6 g COD l–1 d–1 (Table 3) (Van Ginkel and Logan 2005b), approximately 6-fold lower than Kraemer and Bagley (2006), but a 1000-fold decrease in dissolved H2 to 0.5 atm). Thus, lower OLRs may decrease CO2 inhibition. This could explain why Kyazze et al. (2006) could only achieve stable H2 production at an OLR of 112 g COD l–1 d–1 when N2 sparging was employed.

Practical implications and future research Heat treatment dramatically changes the microbial community. However, such a microbial community will not necessarily maximize H2 production because spore-forming H2-consuming bacteria will remain (Kim et al. 2006b; Park et al. 2005). Non-spore-forming facultative aerobes, such as Enterobacter spp., may be beneficial for consuming oxygen that enters the reactor with the feedstock. In addition, the decrease in microbial diversity caused by heat treatment could be undesirable when complex substrates are to be degraded, such as primary and secondary wastewater sludges, because greater microbial diversity would provide a higher number of metabolic degradation pathways.


At present, there is not enough evidence to support the use of inoculum heat treatment. Heat treatment does not maximize H2 yield compared to other pre-treatments (BES, iodopropane, perchloric acid) and at full-scale would be energetically costly. Moreover, non-sterile feedstocks would negate heat treatment because incoming organisms could consume H2 or decrease its production. Methanogen control through heat treatment should be replaced by operational controls using low pH and short solids retention time. Further research is necessary to determine the specific mechanisms by which sparging and other dissolved gas removal techniques cause the H2 yield to increase. For example, to what extent does H2-consumption (e.g. acetogenesis) exist in continuous-flow H2-producing systems? Can this H2-consumption be decreased by changes in dissolved gas concentrations, organic loading rate or other means? From a practical perspective, sparging is energy intensive and therefore minimizing the use of sparging will improve the overall energy balance of a fermentative H2 system. Lower sparging rates than previously reported in the literature should be investigated because our discussion above indicated that H2 yields may still be improved to the same extent as for higher sparging rates. This would be an important finding because the energy required for sparging would be reduced and the product gas would be less dilute. Nevertheless, full-scale implementations would require appropriate assessments to determine whether the increased H2 yield from sparging would be worth the added energy and economic cost of using sparging at all. The disagreement in the literature about the effect of OLR on H2 yields does not allow any meaningful conclusions to be drawn at present. Further research is needed to determine the specific mechanisms by which changes in organic loading rate affect the H2 yield. Practically, the highest allowable OLR would be preferred so as to minimize the size of the reactor. That is, there may be tradeoffs between H2 yield and reactor size. Even if lower OLRs increase the H2 yield, a minimum OLR will exist for practical H2 recovery so that it is not all lost in the dissolved phase.

693 Acknowledgements The authors would like to thank the William H. Doherty Ontario Graduate Scholarship in Science & Technology and the Natural Sciences and Engineering Research Council (NSERC) of Canada for funding.

References Ahn Y, Park EJ, Oh YK, Park S, Webster G, Weightman AJ (2005) Biofilm microbial community of a thermophilic trickling biofilter used for continuous biohydrogen production. FEMS Microbiol Lett 249(1):31–38 Angenent LT, Karim K, Al-Dahhan MH, DomiguezEspinosa R (2004) Production of bioenergy and biochemicals from industrial and agricultural wastewater. Trends Biotechnol 22(9):477–485 Benemann JR (1996) Hydrogen biotechnology: Progress and prospects. Nat Biotechnol 14(9):1101–1103 Cheong DY, Hansen CL (2006) Bacterial stress enrichment enhances anaerobic hydrogen production in cattle manure sludge. Appl Microbiol Biotechnol 72(4):635–643 Crabbendam PM, Neijssel OM, Tempest DW (1985) Metabolic and energetic aspects of the growth of clostridium-butyricum on glucose in chemostat culture. Arch Microbiol 142(4):375–382 Dixon NM, Kell DB (1989) The inhibition by CO2 of the growth and metabolism of microorganisms. J Appl Bacteriol 67(2):109–136 Fang HHP, Liu H, Zhang T (2002a) Characterization of a hydrogen-producing granular sludge. Biotechnol Bioeng 78(1):44–52 Fang HHP, Zhang T, Liu H (2002b) Microbial diversity of a mesophilic hydrogen-producing sludge. Appl Microbiol Biotechnol 58(1):112–118 Hallenbeck PC (2005) Fundamentals of the fermentative production of hydrogen. Water Sci Technol 52(1–2):21–29 Hawkes FR, Dinsdale R, Hawkes DL, Hussy I (2002) Sustainable fermentative hydrogen production: Challenges for process optimisation. Int J Hydrogen Energy 27(11–12):1339–1347 Hussy I, Hawkes FR, Dinsdale R, Hawkes DL (2005) Continuous fermentative hydrogen production from sucrose and sugarbeet. Int J Hydrogen Energy 30(5):471–483 Hussy I, Hawkes FR, Dinsdale R, Hawkes DL (2003) Continuous fermentative hydrogen production from a wheat starch co-product by mixed microflora. Biotechnol Bioeng 84(6):619–626 Iyer P, Bruns MA, Zhang HS, Van Ginkel S, Logan BE (2004) H2-producing bacterial communities from a heat-treated soil inoculum. Appl Microbiol Biotechnol 66(2):166–173 Kapdan IK, Kargi F (2006) Bio-hydrogen production from waste materials. Enzyme Microbial Technol 38(5):569–582 Kataoka N, Miya A, Kiriyama K (1997) Studies on hydrogen production by continuous culture system

123

694 of hydrogen-producing anaerobic bacteria. Water Sci Technol 36(6–7):41–47 Kawagoshi Y, Hino N, Fujimoto A, Nakao M, Fujita Y, Sugimura S, Furukawa K (2005) Effect of inoculum conditioning on hydrogen fermentation and pH effect on bacterial community relevant to hydrogen production. J Biosci Bioeng 100(5):524–530 Kim DH, Han SK, Kim SH, Shin HS (2006a) Effect of gas sparging on continuous fermentative hydrogen production. Int J Hydrogen Energy 31(15): 2158–2169 Kim SH, Han SK, Shin HS (2006b) Effect of substrate concentration on hydrogen production and 16S rDNA-based analysis of the microbial community in a continuous fermenter. Process Biochem 41(1):199– 207 Kraemer JT, Bagley DM (2006) Supersaturation of dissolved H2 and CO2 during fermentative hydrogen production with N2 sparging. Biotechnol Lett 28(18):1485–1491 Kraemer JT, Bagley DM (2005) Continuous fermentative hydrogen production using a two-phase reactor system with recycle. Environ Sci Technol 39(10):3819– 3825 Kyazze G, Martinez-Perez N, Dinsdale R, Premier GC, Hawkes FR, Guwy AJ, Hawkes DL (2006) Influence of substrate concentration on the stability and yield of continuous biohydrogen production. Biotechnol Bioeng 93(5):971–979 Lamed RJ, Lobos JH, Su TM (1988) Effects of stirring and hydrogen on fermentation products of Clostridium thermocellum. Appl Environ Microbiol 54(5):1216– 1221 Lay JJ (2000) Modeling and optimization of anaerobic digested sludge converting starch to hydrogen. Biotechnol Bioeng 68(3):269–278 Liang TM, Cheng SS, Wu KL (2002) Behavioral study on hydrogen fermentation reactor installed with silicone rubber membrane. Int J Hydrogen Energy 27(11– 12):1157–1165 Lin C, Wu S, Chang J (2006) Fermentative hydrogen production with a draft tube fluidized bed reactor containing silicone-gel-immobilized anaerobic sludge. Int J Hydrogen Energy 31(15):2200–2210 Liu DW, Liu DP, Zeng RJ, Angelidaki I (2006) Hydrogen and methane production from household solid waste in the two-stage fermentation process. Water Res 40(11):2230–2236 Logan BE (2004) Extracting hydrogen and electricity from renewable resources. Environ Sci Technol 38(9):160A–167A Madigan MT, Martinko JM, Parker J (2000) Brock biology of microorganisms, 9th edn. Prentice Hall, Upper Saddle River, NJ Mandal B, Nath K, Das D (2006) Improvement of biohydrogen production under decreased partial pressure of H2 by Enterobacter cloacae. Biotechnol Lett 28(11):831–835 Mizuno O, Dinsdale R, Hawkes FR, Hawkes DL, Noike T (2000) Enhancement of hydrogen production from

123

Biotechnol Lett (2007) 29:685–695 glucose by nitrogen gas sparging. Bioresour Technol 73(1):59–65 Nakashimada Y, Rachman MA, Kakizono T, Nishio N (2002) Hydrogen production of Enterobacter aerogenes altered by extracellular and intracellular redox states. Int J Hydrogen Energy 27(11–12):1399–1405 Nath K, Das D (2004) Improvement of fermentative hydrogen production: Various approaches. Appl Microbiol Biotechnol 65(5):520–529 Nishio N, Nakashimada Y (2004) High rate production of hydrogen/methane from various substrates and wastes. Recent Prog Biochem Biomed Eng Japan I 90:63–87 Oh SE, Lyer P, Bruns MA, Logan BE (2004a) Biological hydrogen production using a membrane bioreactor. Biotechnol Bioeng 87(1):119–127 Oh SE, Van Ginkel S, Logan BE (2003a) The relative effectiveness of pH control and heat treatment for enhancing biohydrogen gas production. Environ Sci Technol 37(22):5186–5190 Oh YK, Kim SH, Kim MS, Park S (2004b) Thermophilic biohydrogen production from glucose with trickling biofilter. Biotechnol Bioeng 88(6):690–698 Oh YK, Seol EH, Kim JR, Park S (2003b) Fermentative biohydrogen production by a new chemoheterotrophic bacterium Citrobacter sp. Y19. Int J Hydrogen Energy 28(12):1353–1359 Oh YK, Seol EH, Lee EY, Park S (2002) Fermentative hydrogen production by a new chemoheterotrophic bacterium Rhodopseudomonas palustris P4. Int J Hydrogen Energy 27(11–12):1373–1379 Park W, Hyun SH, Oh SE, Logan BE, Kim IS (2005) Removal of headspace CO2 increases biological hydrogen production. Environ Sci Technol 39(12):4416–4420 Shizas I, Bagley DM (2005) Fermentative hydrogen production in a system using anaerobic digester sludge without heat treatment as a biomass source. Water Sci Technol 52(1–2):139–144 Tanisho S, Kuromoto M, Kadokura N (1998) Effect of CO2 removal on hydrogen production by fermentation. Int J Hydrogen Energy 23(7):559–563 Ueno Y, Haruta S, Ishii M, Igarashi Y (2001a) Changes in product formation and bacterial community by dilution rate on carbohydrate fermentation by methanogenic microflora in continuous flow stirred tank reactor. Appl Microbiol Biotechnol 57(1–2):65–73 Ueno Y, Haruta S, Ishii M, Igarashi Y (2001b) Microbial community in anaerobic hydrogen-producing microflora enriched from sludge compost. Appl Microbiol Biotechnol 57(4):555–562 Van Ginkel S, Logan BE (2005a) Inhibition of biohydrogen production by undissociated acetic and butyric acids. Environ Sci Technol 39(23):9351–9356 Van Ginkel SW, Logan B (2005b) Increased biological hydrogen production with reduced organic loading. Water Res 39(16):3819–3826 Wu SY, Hung CH, Lin CN, Chen HW, Lee AS, Chang JS (2006) Fermentative hydrogen production and bacterial community structure in high-rate anaerobic

Biotechnol Lett (2007) 29:685–695 bioreactors containing silicone-immobilized and selfflocculated sludge. Biotechnol Bioeng 93(5):934–946 Yang HJ, Shao P, Lu TM, Shen JQ, Wang DF, Xu ZN, Yuan X (2006) Continuous bio-hydrogen production from citric acid wastewater via facultative anaerobic bacteria. Int J Hydrogen Energy 31(10):1306–1313 Yu HQ, Zhu ZH, Hu WR, Zhang HS (2002) Hydrogen production from rice winery wastewater in an upflow anaerobic reactor by using mixed anaerobic cultures. Int J Hydrogen Energy 27(11–12):1359–1365

695 Zhang JJ, Li XY, Oh SE, Logan BE (2004) Physical and hydrodynamic properties of flocs produced during biological hydrogen production. Biotechnol Bioeng 88(7):854–860 Zhu H, Be´land M (2006) Evaluation of alternative methods of preparing hydrogen producing seeds from digested wastewater sludge. Int J Hydrogen Energy 31(14):1980–1988

123