Biohydrogen production and bioprocess enhancement: a review.

This paper provides an overview of the recent advances and trends in research in the biological production of hydrogen (biohydrogen). Hydrogen from both fossil and renewable biomass resources is a sustainable source of energy that is not limited and of different applications. The most commonly used techniques of biohydrogen production, including direct biophotolysis, indirect biophotolysis, photo-fermentation and dark-fermentation, conventional or "modern" techniques are examined in this review. The main limitations inherent to biochemical reactions for hydrogen production and design are the constraints in reactor configuration which influence biohydrogen production, and these have been identified. Thereafter, physical pretreatments, modifications in the design of reactors, and biochemical and genetic manipulation techniques that are being developed to enhance the overall rates and yields of biohydrogen generation are revisited.


Introduction
The world population and consequently energy demands seem to grow with an exponential rate (Antonopoulou et al., 2007). The impending shortage of energy resources together with the environmental fall off due to the unreasonable use of fossil fuels, has led many scientists to search for alternative energy sources (Antonopoulou et al., 2007). Among others, research has focused on the hydrogen production field, either by physicochemical or biological methods. Hydrogen is a clean (Kovács et al., 2006) and environmentally friendly fuel (Shin et al., 2010), which produces water instead of greenhouse gases when combusted. It can be produced by renewable raw materials, such as organic wastes, and possesses a high-energy yield (122 kJ/g) due to its light weight, which is 2.75 times greater than the hydrocarbon , and it could be directly used to produce electricity through fuel cells (Lay et al., 1999;Benemann, 1996).

Biohydrogen as a green fuel
Hydrogen has been an unrealized "fuel of the future" for over 30 years, but there are signs that hydrogen may finally become an important component of the energy balance of a global economy (Logan et al., 2002) arising out of the projection of a fossil fuel shortfall towards the middle of the twenty first century (Kotay and Das, 2008). The demand for hydrogen is not limited to its utilization as a source of energy, as hydrogen gas is also a widely used for the production of chemicals, for hydrogenation of fats and oils in the food industry for margarine production, processing steel, and also for the desulfurization and reformulation of gasoline in refineries (Kapdan and Kargi, 2006). Low-cost hydrogen based fuel cells, which have

REVIEW ARTICLE
Biohydrogen production and bioprocess enhancement: A review been expensive or not readily available, are now entering commercial production and are finding applications in residential housing and buses. Despite the "green" nature of hydrogen as a fuel, it is still primarily produced from nonrenewable sources such as natural gas and petroleum hydrocarbons via steam reforming. In order for hydrogen to become a more sustainable and green source of energy, hydrogen must be produced by biological or biochemical reaction pathways (Logan et al., 2002;Han and Shin, 2004). With sustainable development and waste minimization issues, biohydrogen gas production from renewable sources has received considerable attention in recent years (Kapdan and Kargi, 2006). The production of biohydrogen fits very well with the emerging "green chemistry" concept. In point of fact, green chemistry relates to the utilization of a set of principles that reduces or eliminates the use or generation of hazardous substances in the design, manufacture, and application of chemical products (Kidwai and Mohan, 2005). In practice, green chemistry is taken to cover a much broader range of issues than the definition suggests. As well as using and producing better chemicals with less waste, Green chemistry also involves reducing other associated environmental impacts, including a reduction in the amount of energy used in chemical processes (Kidwai and Mohan, 2005). The reader is directed to the following excellent publications which present and discuss the salient aspects of green chemistry and green engineering: Anastas and Warner (1998), Kirchhoff (2003), Anastas and Kirchhoff (2002), Anastas and Zimmerman (2003), Anastas and Lankey (2000), Clark (2006), Höfer and Bigorra (2007), Kirchhoff (2003), Lankey and Anastas (2002), Ren et al. (2008), Tang et al. (2008), and Tundo et al. (2000).

A glimpse of biohydrogen chemistry
Biohydrogen may be produced by cyanobacteria and algae through the biophotolysis of water (Asada and Miyake, 1999) or by photosynthetic and chemosyntheticfermentative bacteria. Some species of cyanobacteria naturally produce hydrogen gas as a byproduct of anaerobic fermentation at night using fixed-carbon compounds (Carrieri et al., 2008). Also, anaerobic fermentative bacteria produce hydrogen without photoenergy (Shin et al., 2010), and so the cost of hydrogen production is 340 times lower than the photosynthetic process. The main source of hydrogen during a biological, fermentative process is carbohydrates, which are very common in plant tissues, either in the form of oligosaccharides or as their polymers, cellulose, hemicellulose and starch. Thus, the biomass of certain plants with a high carbohydrate content has been earnestly considered as a very promising substrate for biohydrogen production. In addition, using properly selected microorganisms, many rural residues, and waste organic materials can be processed and degraded for biohydrogen production, in addition (Venkata . Mohanakrishna et al. (2010) verified that domestic sewage supplementation as co-substrate with composite vegetable based market waste could increase hydrogen production in a fermentative process and maintain a good buffering microenvironment that supports the fermentation process and in addition provides micronutrients, organic matter, and microbial biomass. The maximum theoretical hydrogen yield is 4 moles per mole of utilized carbohydrates, expressed as glucose equivalents when carbohydrates are used as substrate (Nandi and Sengupta, 1998;Logan et al., 2002).
Fermentative hydrogen production from biomass can be achieved either by using mixed acidogenic microbial cultures or a pure culture of a saccharolytic strain. Ruminococcus albus is a non spore-forming, obligatory anaerobic bacterium, the natural habitat of which is the first stomach (rumen) of ruminants. It produces extracellular hydrolytic enzymes (exoglucanases and endoglucanases), which break down cellulose and hemicellulose, whereas it cannot break down pectin and starch (Antonopoulou et al., 2007). The oligosaccharides produced from cellulose and hemicellulose degradation -cellobiose, glucose and the respective pentoses, xylose and arabinose, are further metabolized (Lou et al., 1997). Logan et al. (2002) analyzed the biological production of hydrogen from the fermentation of different substrates in batch tests using heat-shocked mixed cultures with two techniques: an intermittent pressure release method (Owen method) and a continuous gas release method using a bubble measurement device (respirometric method). Lay et al. (2004) demonstrated the optimal substrate concentration and pH for generating biohydrogen gas in composting enriched from heat-shocked anaerobic microbes of cow compost. Under otherwise identical conditions, the respirometric method resulted in the production of 43% more hydrogen gas from glucose than the Owen method. Logan et al. (2002) noted that the biohydrogen conversion efficiencies were similar for sucrose (23%) and lower for molasses (15%) but were much lower for lactate (0.5%) and cellulose (0.075%).
This paper reviews the essential biohydrogen generation processes, identifies the key limitations to a more efficient biohydrogen production, and thereafter probes into some selected recent research findings which report the enhancement achieved in the overall rates and yields of biohydrogen production.

Biohydrogen production methods
Biohydrogen can be produced biologically by biophotolysis (direct and indirect), photo-fermentation and darkfermentation or by a combination of these processes (such as integration of dark-and photo-fermentation, or biocatalyzed electrolysis). At the laboratory scale, biological hydrogen has been produced continuously (Manish and Banerjee, 2008); however biohydrogen production at the commercial scale has not been reported in the literature and challenges regarding process scale up remain (Hawkes et al., 2002;Vatsala et al., 2008;Li et al., 2009;Krupp and Widmann, 2009;Tian et al., 2010).
Biohydrogen production processes are fundamentally dependent upon the presence of a hydrogen (H 2 ) producing enzymes. These enzymes catalyze the chemical reaction 2H+ + 2e − ↔ H 2 . A survey of all presently known enzymes capable of hydrogen evolution shows that they contain complex metallo-clusters as active sites (Manish and Banerjee, 2008). At present three enzymes carrying out this reaction are known; nitrogenase, Fe-hydrogenase, and NiFe-hydrogenase (Hallenbeck and Benemann, 2002). The Fe-hydrogenase enzyme is used in the biophotolysis processes whereas photo-fermentation processes utilize nitrogenase. A brief description, condensed from Manish and Banerjee (2008) of these processes is provided below (photo-fermentation technology). In summary, the major bioprocesses utilized for hydrogen gas production can be classified according to the following categories: Biophotolysis of water by algae. 1.
Dark-fermentative hydrogen production during the 2.
acidogenic phase of anaerobic digestion of organic matter. Two stage dark/photo-fermentative production of 3. hydrogen.

Biophotolysis
The direct biophotolysis method is similar to the processes found in plants and algal photosynthesis. In this process, solar energy is directly converted to hydrogen via photosynthetic reactions 2H 2 O + hv → 2H 2 + O 2 where hv represent the energy from a photon in light (h is the Planck constant and v is the frequency of the light). The indirect biophotolysis method circumvents any problems of sensitivity of the hydrogen evolving process by separating temporally and/or spatially oxygen evolution and hydrogen evolution. Thus, indirect biophotolysis processes involve separation of the H 2 and oxygen evolution reactions into separate stages, coupled through CO 2 fixation/evolution. Our survey of the literature shows that cyanobacteria are the only bacteria capable of performing oxygenic photosynthesis in which they harness solar energy and convert it into chemical energy stored in carbohydrates, and under specific conditions, cyanobacteria can also use solar energy to produce molecular hydrogen (Allahverdiyeva et al., 2010). The overall mechanism of hydrogen production in cyanobacteria can be represented by the following reactions: 12H 2 O + 6CO 2 + hv → C 6 H 12 O 6 + 6O 2 and C 6 H 12 O 6 + 12H 2 O + hv → 12H 2 + 6CO 2 .

Photo-fermentation
Photosynthetic bacteria evolve molecular hydrogen catalyzed by nitrogenase under nitrogen-deficient conditions using light energy and reduced compounds (organic acids) (Levin et al., 2004). These bacteria themselves are not powerful enough to split water. However, under anaerobic conditions, these bacteria are able to use simple organic acids, like acetic acid as electron donors (Manish and Banerjee, 2008). These electrons are transported to the nitrogenase by ferredoxin using energy in the form of adenosine triphosphate (ATP). When nitrogen is not present, this nitrogenase enzyme can reduce the proton into hydrogen gas again using extra energy in the form of ATP (Akkerman et al., 2002). The overall reaction of hydrogen production can be given as: The fermentation process for hydrogen production has been widely reported but there is observably a lack of information related to detailed kinetic studies. Our review of the literature has shown that the kinetic analysis of biohydrogen production has been mostly performed using the modified Gompertz equation for fitting the experimental data of accumulative hydrogen production (Lay, 2001;Wu and Lin, 2004;Fang et al., 2005;Van Ginkel et al., 2005;Mu et al., 2006;Gadhamshetty et al., 2010). The modified Gompertz equation is: where H(t) is the accumulative hydrogen production (l) during the fermentation time t(h), P the (maximum) hydrogen production potential (l), R m the maximum production rate (l/h), λ the lag-phase time (h), and e is 2.7182818. The values of P, R m and λ are normally determined by best fitting the experimental hydrogen producing data using suitable software (Fang et al., 2002).

Dark-fermentation
Dark-fermentation is one of the most powerful processes because of a relatively higher rate of hydrogen production, and many researchers have studied biohydrogen production by fermentative bacteria, such as Escherichia coli (Yoshida et al., 2005), Enterobacter species (Palazzi et al., 2000;Kurokawa and Tanisho, 2005;Zhang et al., 2005;Shin et al., 2007) and Clostridium species (Jo et al., 2008). Many studies on hydrogen production have been performed using facultative anaerobes because of a difficulty in maintaining the strict anaerobic condition, which is necessary for obligate bacteria. Strict anaerobes, such as Clostridium species, are very sensitive to trace amounts of dissolved oxygen, resulting in the necessity of expensive reducing agents to be added in the culture medium (Shin et al., 2010). However, facultative anaerobes, such as Enterobacter species, are less sensitive to dissolved oxygen, and the activity of the enzyme involved in hydrogen production can be rapidly recovered from the oxygen damage when depleted in the culture medium (Shin et al., 2010). The majority of microbial hydrogen production is driven by the anaerobic metabolism of pyruvate breakdown, formed during the catabolism of various substrates, and the breakdown of pyruvate is catalyzed by one of two enzyme systems (Hallenbeck and Benemann, 2002;Manish and Banerjee, 2008;Hallenbeck, 2009;Vijayaraghavan and Soom, 2006) given below.

Two-stage process with integration of dark-and photo-fermentation
In fermentation, complete oxidation of 1 mole of glucose yields 12 moles of hydrogen. However, complete oxidation of glucose into hydrogen and carbon dioxide is not possible as the corresponding reaction is not feasible thermodynamically. With an external energy supply (photon-energy in photo-fermentation) theoretically 12 moles of hydrogen per mole of glucose can be produced. However, this process cannot be operated in the absence of light. On the other hand, in the absence of external energy (in the case of dark-fermentation), oxidation of glucose by fermentative bacteria results in other byproducts also and only a maximum of 4 moles of hydrogen are produced per mole of glucose consumption (C 6 H 12 O 6 + 2H 2 O → 4H 2 + 2CO 2 + 2CH 3 COOH, (▵G o = −206 kJ). Acetate produced in the dark-fermentation stage can be oxidized by photosynthetic bacteria to produce hydrogen (CH 3 COOH + 2H 2 O + hv → 4H 2 + 2CO 2 , (▵G o = +104 kJ). Hence, continuous production of hydrogen at maximum yield can be achieved by integrating the dark-and photo-fermentation methods. Yang et al. (2010) recently reported enhanced biohydrogen production rates by integrating dark-fermentation with the photo-fermentation process for pretreated corn cob. In the first step, the maximum biohydrogen yield and rate from corn cob by dark-fermentation was 120.3 mL H 2 /g corn cob and 150 mL H 2 /L/h, respectively. In the second step, a hydrogen yield of 713.6 mL H 2 /g COD was obtained from digesting the effluent of dark-fermentation by photosynthetic bacteria.

Substrates for biohydrogen production
In order to reduce carbon dioxide release, hydrogen gas will need to be produced from renewable sources (Van Ginkel et al., 2005;Refaat and El Sheltawy, 2008). Most hydrogen gas produced is obtained from thermocatalytic and gasification processes using natural gas (50%), petroleum-derived napthenes and distillates (30%), and coal (18%), with the remainder from electricity (2%) (Van Ginkel et al., 2005). Substrates are present in very large quantities as products or wastes from agriculture, crop residues, the food industry and market waste, animal waste, and organic matter of municipal solid waste (Forster-Carneiro et al., 2008). Biohydrogen production from the fermentation of renewable carbohydrate-rich and non-toxic raw materials (Kapdan and Kargi, 2006) is one promising alternative although the use of commercially produced food products, such as corn and sugar, is not yet economical (Benemann, 1996). Substrates used for biohydrogen production have ranged from simple sugars such as glucose , sucrose (Antonopoulou et al., 2007), starch containing waste such as cassava wastewater (Sangyoka et al., 2007), dairy wastewater (Venkata Mohan et al., 2007a), sweet potato starch residue (Yokoi et al., 2001), cheese whey (Davila-Vazquez et al., 2009) and food waste (Ruknongsaeng et al., 2005). Other substrates for biohydrogen production are listed in Table 1.
Wastewaters show great potential for economical production of hydrogen because producing a product from a waste could reduce waste treatment and disposal costs (Van Ginkel et al., 2005). Hydrogen has so far been produced from the organic fraction of municipal solid wastes (Okamoto et al., 2000) and cellulose (Lay, 2001). Batch tests using various wastes and wastewaters suggest that hydrogen production is more efficient from carbohydrates than other materials (Logan et al., 2002). Simple sugars, such as sucrose and glucose, are converted at elevated temperatures to hydrogen with high conversion efficiencies. Yields of 28% were obtained with glucose, and 26% with sucrose, at 30°C, while hydrogen produced from molasses, lactate, and cellulose were 15%, 0.5% and 0.075%, respectively (Logan et al., 2002). These results indicate that high-carbohydrate wastewaters will be seemingly the most useful for industrial production of hydrogen. Wu and Lin (2004) have conducted batch experiments to convert molasses wastewater (10-160 g chemical oxygen demand (COD)/L) into hydrogen at 35 °C at various pHs (4-8). The maximum hydrogen productivity (HP) and hydrogen production rate (HPR) reached 47.1 mmol H 2 /g COD and 97.5 mmol H 2 /L/d, respectively, at a substrate concentration of 40 g COD/L and pH 6, and the methane-free biogas contained up to 50% (v/v) of hydrogen. O- Thong et al. (2007) seeded thermophilic microflora into an anaerobic sequencing batch reactor for hydrogen production from palm oil mill effluent (POME) and supplemented the reaction mixture with nitrogen, phosphorus, and iron sources for biostimulants. O-Thong et al. (2007) noted that the nutrient supplementation strategy had increased the bacterial diversity in the reactor and promoted in particular the growth of the hydrogen-producing bacteria Thermosaccharolyticum which ultimately increased the hydrogen production yield from 1.6 to 2.24 mol H 2 /mol hexose and hydrogen production rate from 4.4 to 6.1 l H 2 /L POME/d.
Cellulose is a predominant constituent of agricultural waste and waste generated by the pulp and paper industry. To generate hydrogen directly from cellulose materials using dark fermentation requires expensive pretreatment processes such as delignification and hydrolysis to dissolve organic matter from a lignocellulose complex (Taguchi et al., 1996). Lay (2001) investigated the potential of producing hydrogen from microcrystalline cellulose using mesophilic digestion with heat-shocked sludge. With a 4 day lag, a maximum hydrogen yield of 4.36 mg/g cellulose was produced from suspensions containing 12:5 g cellulose/L. The metabolites were predominantly alcohols, followed by volatile fatty acids. Liu et al. (2003) determined that their mixed culture comprising microbes closely affiliated with the genus Thermoanaerobacterium produced hydrogen that peaked at 7:56 mg H 2 /g cellulose and a maximum rate of 21:2 mg H 2 /g VSS/d from a 5 g cellulose/L suspension maintained at pH 6.5 and 55°C. The metabolites observed were primarily acetate, butyrate, and ethanol.

Chemical reaction related limitations
For hydrogen generation, the current biomass technologies include: gasification, pyrolysis, liquefaction, hydrolysis, and conversion to liquid fuels by supercritical extraction followed in some cases by reformation, and biological hydrogen production (Holladay et al., 2009). The gasification technology of biomass or wastewater is commonly used in many processes but for biological hydrogen it has substantially increased over the last several years. Sewage sludge of wastewater treatment plants is composed largely of organic matter such as carbohydrates and proteins (Weemaes and Verstraete, 1998;Xiao and Liu, 2009) and the anaerobic digestion technique has been employed to treat sludge and obtain methane (Reith et al., 2003). Hydrogen is an intermediate product of anaerobic sludge digestion and is quickly consumed by hydrogen-consuming bacteria such as methanogens and sulfate-reducing bacteria. In order to harvest hydrogen from anaerobic sludge digestion, the activity of consuming-hydrogen bacteria must be inhibited at the hydrogen and acetic acid forming stage and the consumption of hydrogen must be blocked (Hawkes et al., 2002). In cornstalk wastes conversion into hydrogen, the acetate, propionate, butyrate, and the ethanol were the main by-products in the metabolism of hydrogen fermentation. Also, the test results showed that the acidification pretreatment of the substrate plays a crucial role in conversion of the cornstalk wastes into biohydrogen gas by cow dung composts generating hydrogen (Zhang et al., 2007a). Additionally, the engineering challenges of scale-up and a shift in the type of biomass substrates from starch-based food crops to lignocellulosic feedstock and wastes that are economically and environmentally less costly to produce, yet more difficult to biochemically process, present technical challenges that are inherent to the biohydrogen production process (Jones, 2008).

Reactor design related limitation
Biohydrogen production by anaerobic fermentation has attracted worldwide attention owing to the fact that hydrogen can be produced substantially at a high rate from renewable organic matters (Benemann, 1996). Biohydrogen production systems are to become commercially competitive they must be able to synthesize H 2 at rates that are sufficient to power fuel cells of sufficient size to do practical work (Logan, 2010).
The studies on continuous fermentative hydrogen production in the laboratory-scale had been conducted using suspended-cell systems and immobilized-cell systems since the 1980s (Chen and Lin, 2003;Fan et al., 2006). The HPR has been considered as an important index to evaluate the performance of continuous hydrogen-producing processes . However, the continuous stirred tank reactor (CSTR) process, a typical representative of suspended-cell systems, usually exhibits poor performance in HPR since it is unable to maintain high levels of hydrogen-producing biomass at a short hydraulic retention time (HRT) due to its intrinsic structure (Zhang et al., 2007a). To achieve satisfactory HPR, immobilized-cell systems have become a popular alternative to suspended-cell systems for continuous biohydrogen production since they are more capable of maintaining higher biomass concentration even at lower HRTs (Wu et al., 2002). More recent studies by other authors conclude that, at low hydraulic retention time, acidogenic anaerobic digestion of organic waste reaches the top speeds of hydrogen production, while contributing to the elimination of contaminating waste (Kapdan and Kargi, 2006;Kotsopoulos et al., 2006;Kyazze et al., 2005).

Enhancement of biohydrogen production
Under anaerobic conditions, hydrogen is produced as a by-product during the conversion of organic wastes into organic acids which are then used for methane generation. Anaerobic digestion allows the stabilization of the waste disposal or in conjunction with hydrogen production at rates higher than that of other biological processes (Lee et al., 2004;Valdez-Vazquez et al., 2006) and conform a steady-state model for biological hydrogen production in a fermentation process (Whang et al., 2006). In the acidogenic phase of anaerobic digestion of wastes can be manipulated to improve hydrogen production. Photosynthetic processes include algae which use CO 2 and H 2 O for hydrogen gas production. However, the rate and yield of H 2 production has been found to be relatively low (Kapdan and Kargi, 2006;Das, 2009) and hence the biohydrogen technology has been thoroughly researched (Rachman et al., 1998;Levin et al., 2004). Currently, laboratory-scale studies on anaerobic hydrogen fermentation technology are being conducted by a large number of research groups in different countries over the world (Fang and Liu, 2002;Lin and Jo, 2003). This technology exhibits positive features in hydrogen production such as a high production rate, low energy demand, easy operation and high sustainability. However, it is yet to compete with those thermochemical processes converting hydrogen from fossil fuels in cost, performance or reliability (Das and Veziroglu, 2001). As a result, current research on anaerobic hydrogen fermentation has been focused on improving the microbial hydrogen conversion rate and unit volumetric production rate. The former could be achieved by screening efficient hydrogen-producing bacteria and optimizing the operational conditions, while the latter is substantially influenced by the reactor biomass retention. To achieve such purposes, immobilization processes of hydrogen-producing culture have become most popular and have been developed extensively, due to the elevated biomass retention as compared to suspended-cell systems . Low yields and the rates of hydrogen formation may additionally be overcome by selecting and using more effective organisms or mixed cultures, developing more efficient processing schemes, optimizing the environmental conditions, improving the light utilization efficiency, and developing more efficient photo-bioreactors. Due to inhibition of biohydrogen production by oxygen and ammonium-nitrogen, microbial growth and hydrogen formation steps may need to be separated in order to improve the hydrogen productivity (Kapdan and Kargi, 2006). A possible alternative is to increase the production of hydrogen using chemical inhibitors of methanogenesis. Sparling and collaborators (1996) have shown how to apply low concentrations of acetylene (0.5−1% v/v) to the reactor atmosphere, an effective method of preventing methanogenesis in reactors designed for hydrogen production. Another strategy that has been applied is the thermal shock treatment of the inoculum used (Lay et al., 2003).
Many bacteria contain enzymes (hydrogenases) that can produce hydrogen during the fermentation of a variety of substrates. ATP is produced by substratelevel or electron transport phosphorylation, but the ATP yields of fermentation are quite low as compared to those of aerobic oxidation reactions. Fermentation reactions can produce many different end products such as hydrogen, acetate, and ethanol. The hydrogen-acetate couple produces more ATP per mole of substrate than alcohols such as ethanol and butanol and is the energetically "preferred" bacterial fermentation product for a sugar (Logan et al., 2002). In mixed anaerobic cultures, the accumulation of hydrogen is normally balanced by rapid hydrogen consumption by methanogens resulting in little net hydrogen accumulation in the system, and the individual and interactive effects of pH, temperature, and glucose concentration on H 2 production could be evaluated (Mu et al., 2009). If high concentrations of hydrogen are desired, a system must be designed to remove hydrogen before it can lead to repression of its production and to prevent interspecies hydrogen transfer leading to methanogenesis. The culture conditions that can adversely affect hydrogen production are only beginning to be studied and are therefore not so well-understood. Batch tests using mixed cultures have demonstrated that very low pH and high substrate concentrations can reduce biohydrogen production, while increasing the substrate loading increases the relative production of volatile acids and decreases the pH. Heat shocking has been used to reduce the concentration of non-spore forming bacteria such as methanogens, but the effect of this procedure on the storage of the material and the differences between different batches of mixed cultures has not been tested. Venkata  have observed that heat-shock pretreatment (100°C; 1 h) evaluated for selectively enriching the hydrogen producing mixed culture using dairy wastewater as substrate resulted in a relatively low H 2 yield. Furthermore, the optimization of the nutritional and environmental conditions has also been demonstrated to play an important role in developing hydrogen producing bioprocesses and improving their performance (Kumar and Satyanarayana, 2007). Moreover, given a selected substrate, its concentration appears to be critical in terms of hydrogen production, being in most cases a factor to be explored (Akutsu et al., 2009;Han and Shin, 2004).

Physical pretreatments and operating conditions
With regards to H 2 generation during anaerobic wastewater treatment whereby hydrolysis is the rate limiting step (Li and Noike, 1992), several pretreatments (listed in Table 2) of the parent anaerobic inoculum have been analyzed to accelerate the hydrolysis step reducing the impact of the rate limiting step and augmenting anaerobic digestion to enhance H 2 generation (Zhu and Béland, 2006). Xiao and Liu (2009) have assessed acid pretreatment, alkaline pretreatment, thermal pretreatment, and ultrasonic pretreatment to enhance biohydrogen production from sewage sludge. Their experimental results showed that the four pretreatments could all increase the soluble chemical oxygen demand (SCOD) of sludge and decrease the dry solid (DS) and volatile solid (VS) because the pretreatments could disrupt the floc structure and even the microbial cells of sludge. Additionally, the results of batch anaerobic fermentation experiments demonstrated that all of the four pretreatments could enhance hydrogen production with a hydrogen yield of the alkaline pretreated sludge reaching 11.68 mL H 2 /g VS and that of the thermal pretreated sludge 8.62 mL H 2 /g VS. Earlier, Cai et al. (2004) had performed batch tests to analyze influences of alkaline pretreatment and initial pH value on biohydrogen production from sewage sludge. It was noted that the biohydrogen yield had increased from 9.1 mL of H 2 /g of dry solids (DS) of the raw sludge to 16.6 mL of H 2 /g of DS of the alkaline pretreated sludge.
Higher yields of hydrogen gas can be recovered from the microbial fermentation of organic substrates at high concentrations when interspecies hydrogen transfer to methanogens is prevented. Bearing this metabolic requirement in focus, Oh et al. (2003a) have used two techniques to limit methanogenesis in mixed cultures: heat treatment, to remove non-spore forming methanogens from an inoculum, and low pH during culture growth. It was found that high hydrogen gas concentrations (57−72%) were produced in all tests and that heat treatment of the inoculum (pH 6.2 or 7.5) produced greater hydrogen yields than low pH (6.2) conditions with a non-heat treated inoculum. With regards to operational control of pH, Mohanakrishna et al. (2010) observed a significant improvement in H 2 production and substrate degradation upon supplementing the waste with domestic sewage, and much less variation in the outlet pH in supplementation experiments compared to normal operation. Supplementation of waste with co-substrate seemingly helps to maintain a good buffering microenvironment, supports the fermentation process, and in addition provides micro-nutrients, organic matter, and microbial biomass.

Modified reactor configurations
The reactor design and process configuration also have a bearing on the overall chemistry of the hydrogen producing reactions. In this respect, researchers have also studied several new configurations of experimental set ups to optimize the hydrogen production rates and yields (Maag et al., 2009). Studies of the feasibility of the anaerobic digestion process in separate phases show that the hydrogen production could continuously maintain and effluent with low concentrations of volatile fatty acids (VFA). Similar results have been obtained by other authors for the treatment of fruit and vegetable waste (Bouallagui et al., 2004); waste mills (Raposo et al., 2004), and municipal solid waste (Ueno et al., 2006). Remaining studies have developed a new carrierinduced granular sludge bed (CIGSB) bioreactor and it was shown to be very effective in hydrogen production (Lee et al., 2006). However, since mechanical agitation was not employed to enable sludge granulation, the CIGSB system might still encounter problems with poor mass transfer efficiency during prolonged operations. Lee et al. (2006) designed the CIGSB to improve the mixing efficiency of CIGSB for better biomass-substrate contact by adjusting the height to diameter (H/D) ratios of the reactor and by implementing appropriate agitation device. Reactor designs with an H/D ratio of 8 gave better H 2 production performance with a H 2 production rate of 6.87 l/h/L and a H 2 yield of 3.88 mol H 2 /mol sucrose, suggesting that the effectiveness of H 2 production in the CIGSB system can be enhanced by using a proper upflow velocity and physical configuration of the reactor. Lee et al. (2006) deepened their analysis and following the supply of additional mechanical agitation for the CIGSB reactor at a H/D = 12, the sludge piston flotation was dampened and this led to further increases in the H 2 production rate and H 2 yield to 9.31 l/h/L and 4.02 mol H 2 /mol sucrose, respectively. Ran et al. (2007) investigated the optimal fermentation type and the operating conditions of anaerobic process in continuous-flow acidogenic reactors for the maximization of biohydrogen production using mixed cultures. They reported a maximum hydrogen production of up  Penaud et al. (1999), Xiao and Liu (2009); Venkata  to 14.99 l/d for organic loading rate (OLR) of 86.1 kg COD/m 3 /d. Zhang et al. (2007b) have examined a new approach to immobilize mixed cultures of hydrogen-producing bacteria by growing these on granular activated carbon in an anaerobic fluidized bed reactor, with the production of hydrogen assessed by the immobilized culture at a consistent pH of 4 and at a temperature of 37°C. It was observed that the hydrogen production rate and specific hydrogen production rate were linearly correlated to the effective OLR, which was calculated on the basis of the organic loading and glucose conversion rate, giving the respective maximum rates of 2.36 Ll/L/h and 4.34 mmol H 2 /gVSS/h. Zhang et al. (2007b) concluded that a substantial quantity of the retained biomass would enable the reactor to run at these high organic loading rates and thus enhance the production rates of hydrogen gas. Later, Zhang et al. (2008a) used biofilm sludge and granular sludge to convert glucose into hydrogen in two anaerobic fluidized bed reactors (AFBRs) operated at a pH of 5.5 and 37°C. The influence of the HRT and glucose concentration on hydrogen production in the reactors was examined at a constant organic loading rate of 40 g glucose/L/h by varying the HRT from 0.125 to 3 h and the glucose concentration from 5 to 120 g/L. The hydrogen yield obtained in both reactors ranged between 0.4 and 1.7 mol H 2 /mol glucose, with a maximum yield occurring at an HRT of 0.25 h and a glucose concentration of 10 g/L. It was noted that the biofilm had been washed out substantially in the biofilm reactor and the reactor biomass was replaced by granules during the 50 day operation, and consequently no apparent variation in hydrogen production was observed as the biofilm was replaced by granules. Zhang et al. (2008a) deduced that as compared with the carrier-based biofilm reactor, the granule-based reactor indicated an advantage of better biomass retention without subject to washout of the support carriers. Later, Zhang et al. (2008b) equally concluded that a granule-base column-shaped reactor system appears to be the preferred process for continuous hydrogen fermentation on glucose substrate.
Another community reactor optimized for biohydrogen production is the upflow anaerobic sludge blanket (UASB) and the operating conditions of the acidogenic reactor (concentration of solids in the feed, retention time, organic loading density, pH, and flow recirculation) were extensively studied to maximize hydrogen production Zhao et al., 2008). Mu and Yu (2007) studied the performance of a granulebased H 2 -producing upflow anaerobic sludge blanket (UASB) reactor simulated using a neural network and genetic algorithm and a model was designed, trained, and validated to predict the steady-state performance of the reactor. The H 2 concentration, H 2 production rate, H 2 yield, and effluent total organic carbon were the inputs of the model, and the simulation results demonstrated that the model was able to effectively describe the daily variations of the UASB reactor performance, and to predict the steady-state reactor performance at various substrate concentrations and HRTs. The response surface methodology (RSM) was used by Zhao and collaborates (2008) to evaluate the biohydrogen production from sucrose in a granule-based upflow anaerobic sludge in the blanket (UASB) reactor.
Recently, an anaerobic sequencing batch reactor (ASBR), adopted from the classical reactor for wastewater treatment, has shown promising results in hydrogen production by changing the time of each cycle, and the authors concluded that the pH and the cyclic duration of the operations profoundly impacted fermentative hydrogen production (Chen et al., 2009).
Yet another factor influencing the bacterial productivity and total yield of hydrogen is the partial pressure of produced gas. A novel solution to enhance the bacterial productivity was through reduction of the gas pressure has been proposed by Alshiyab et al. (2009). An increase in the reactor size showed an enhancement in the bacterial production of hydrogen. This technique of increasing reactor size resulted in enhancement of the 1 glucose utilized to maximum yield-hydrogen yield from 269 mL/g glucose utilized by using 125 mL and 2 l reactor size of 448 mL/g, respectively. The hydrogen productivity was also enhanced from 71 mL/h to a maximum of 91 mL/h. Alshiyab et al. (2009) concluded that by using a bigger reactor size, the effect of gaseous products in the fermentation medium was reduced and thereafter enhanced both bacterial productivity and biomass concentration.

New bacterial strains
Microbial H 2 production is an attractive process accounting for a significant share of the H 2 required for the near future. The biochemical hydrogen potential (BHP) tests were conducted to investigate the metabolism of different inoculums fermentation and evaluate the hydrogen potential of bacterial strain species growing on different substrates. Lin et al. (2007) investigated the metabolism of glucose fermentation of four Clostridial species, including C. acetobutylicum M121, C. butyricum ATCC19398, C. tyrobutyricum FYa102, and C. beijerinckii L9 and the results were able to accurately describe the profile of glucose degradation as well as production of biomass, butyrate, acetate, ethanol, and a significant amount of hydrogen gas in the batch tests. Other microbial species, belonging to the genera Enterobacter, Citrobacter, Bacillus, and Clostridium are reported to produce hydrogen through dark fermentation (Nandi and Sengupta, 1998). Apart from pure cultures, various mixed microflora and co-cultures have also been explored for hydrogen production from carbohydrates (Das and Verziroglu, 2001). Nevertheless, the search for "ideal" and more selective microbe(s) for microbial H 2 production have thrust the researchers to screen various sources.
Isolating strains that can effectively utilize cellulose materials to produce hydrogen at room temperature is also of great practical interest. Oh et al. (2003b) had isolated a newly isolated Citrobacter sp. Y19 for CO-dependent H 2 production for its capability of fermentative H 2 production in batch cultivation. When glucose was used as the carbon source, the pH of the culture medium significantly decreased as fermentation proceeded and H 2 production was seriously inhibited but fortified by phosphate at 60-180 mmol/l as it alleviated this inhibition. The maximal H 2 yield and H 2 production rate were estimated to be 2:49 mol H 2 /mol glucose and 32:3 mmol H 2 /g cell/h, respectively. According to Oh et al. (2003b), the overall performance of Y19 in fermentative H 2 production was quite similar to that of most H 2 -producing bacteria previously studied, especially to that of Rhodopseudomonas palustris P4, and that indicated that the attempt to find an outstanding bacterial strain for fermentative H 2 production might be very difficult. In this case the glucose present in a medium of Citrobacter Y19 being used for biohydrogen generation, the glucose is believed to serve a double role in enhancing the sustained production rate of hydrogen (Pandey and Pandey, 2008).
Nevertheless, several other studies followed in this direction and the results seemed promising. Wang et al. (2008) isolated a strain (X 9 ), a member of clostridia genera (Clostridium acetobutylicum, ATCC 824), from a hydrogen-producing reactor, and determined hydrogen production potential by dark fermentation of this strain from microcrystalline cellulose suspensions at 37°C. In their work, Wang et al. (2008) also tested whether this stain could work with another strain, Ethanoigenens harbinense B49, which could produce hydrogen efficiently from monosaccharides, for bioaugmented biohydrogen production from microcrystalline cellulose. At 37°C and pH 5, the mono-culture of X 9 yielded hydrogen with a 5 h time lag and the end liquid products contained primarily of acetate and butyrate. The co-culture of X 9 with E. harbinense B49 produced more efficiently the biohydrogen via an ethanol-type fermentation metabolism compared with the mono-culture X 9 test. This meant that the bioaugmentation with X 9 +B49 improved cellulose hydrolysis and the subsequent hydrogen production rates as compared with that of monoculture bioaugmentation with X 9 . Earlier, Venkata Mohan et al. (2007b) studied the feasibility of a bioaugmentation strategy in the process of enhancing biohydrogen production from chemical wastewater treatment for an OLR of 6.3 kgCOD/m 3 /d in anaerobic sequencing batch biofilm reactor (AnSBBR) operated at 28°C under acidophilic microenvironment (pH 6) with a total cycle period of 24 h. A parent augmented inoculum, kanamycin resistant, was acquired from an operating upflow anaerobic sludge blanket (UASB) reactor treating chemical wastewater and subjected to selective enrichment by applying repetitive/cyclic pretreatment methods altering between heat-shock treatment at 100°C, 2 h and acid treatment at pH 3, 24 h to eliminate non-spore forming bacteria and to inhibit the growth of bacteria involved in the synthesis of methane. In the case of food waste at upflow anaerobic sludge blanket (UASB) reactor treating wastewater the specific methanogenic activity (SMA) of the granule was the highest for butyrate, and the lowest for propionate, also Methanosaeta-like bamboo-shaped rods were present in abundance (Han et al., 2005). From the data obtained, Venkata Mohan et al. (2007b) showed a positive influence of the bioaugmentation strategy on overall H 2 production with a specific H 2 production almost doubling after augmentation from 0.297 to 0.483 mol H 2 /kg COD/d. The survival and retention of the augmented kanamycin resistant inoculum and its positive effect on process enhancement was most seemingly attributable to the adopted reactor configuration and operating conditions.
Additionally, analyses of evolution of the microbial community was studied during reactor operation using molecular biology tools (T−RFLP, 16S rRNA cloning library and FISH) and conventional microbiological techniques to examine the feasibility of producing hydrogen by dark fermentation (Castelló et al., 2009). The results showed that hydrogen can be produced but in low amounts and microbiological studies showed the prevalence of fermentative organisms from the genera Megasphaera, Anaerotruncus, Pectinatus, and Lactobacillus, which may be responsible for hydrogen production. According Venkata Mohan and collaborators (2007b), the image analyses made on the scanning electron microscope (SEM) confirmed the selective enrichment of a morphologically similar group of bacteria capable of producing H 2 under acidophilic conditions in an anaerobic microenvironment at a specific H 2 production which almost doubled after augmentation from 0.297 to 0.483 mol H 2 /kg COD reduced /d. The survival and maintenance of the augmented consortia suggested that the growth rate of the organism might have been higher than the washout and the activity of the grazers was negligible. The SEM images (×5 K) of augmented mixed consortia had shown slightly bent, rod-shaped, and thick fluorescent capsid bacteria. It was presumed from the image visibility that the adopted selective enrichment procedure might have resulted in an enrichment of a morphologically similar group of rod-shaped bacteria capable of producing H 2 .

Genetic manipulation
Improved biohydrogen production rates will clearly benefit from both the selection of a suitable phototroph and the engineering of its biochemical pathways (Kruse et al., 2005). The majority of microorganisms currently studied for hydrogen photoproduction have been selected because of their ease of cultivation, which is often consistent with slow growth rates (Kruse et al., 2005). However, further efforts to overcome existing issues of low rates and yields of biohydrogen production in optimized reactor configurations will need to rely on the ability to analyze, predict, and engineer microbial metabolism in native H 2producing strains as well as genetically engineer strains with constructed H 2 -metabolism (Vignais et al., 2006;Jones, 2008).
References and further reading may be available for this article. To view references and further reading you must purchase this article.
Recently, there has been substantial progress in identifying the relevant bioenergy genes and pathways, and powerful genetic manipulations have been developed to engineer some strains via the targeted disruption of endogenous genes and/or transgene expression (Beer et al., 2009). Collectively, the progress that has been realized in these areas is rapidly advancing the ability of researchers and engineers to genetically optimize the production of targeted biofuels including biohydrogen. Akhtar and Jones (2008) have recently engineered a synthetic hydF-hydE-hydG-hydA operon for biohydrogen production while Xing et al. (2008) reported genomic evidence for the presence of novel H 2 -producing bacteria in acidophilic ethanol-H 2 -coproducing communities that were enriched using molasses wastewater. Xing et al. (2008) reported H 2 production rates that reached 0.48 L/g VSS/d. Earlier, Melis et al. (2007) have examined physiological and genetic engineering approaches to improve the hydrogen metabolism characteristics of microalgae. Melis et al. (2007) discussed the application of sulfurnutrient deprivation to attenuate oxygen-evolution and to promote H 2 -production, as well as the genetic engineering of sulfate uptake through manipulation of a newly reported sulfate permease in the chloroplast of the model green alga Chlamydomonas reinhardtii. Franchi et al. (2004) have constructed three differently metabolically engineered strains, 2 single PHA − and Hup − mutants and one double PHA − /Hup − mutant, of the purple nonsulfur photosynthetic bacterium Rhodobacter sphaeroides RV, were constructed to improve a lightdriven biohydrogen production process combined with the disposal of solid food wastes. These phenotypes were designed to abolish, singly or in combination, the competition of H 2 photoproduction with polyhydroxyalkanoate (PHA) accumulation by inactivating PHA synthase activity, and with H 2 recycling by abolishing the uptake hydrogenase enzyme. With lactic acid-based synthetic medium, the single Hup − and the double PHA − / Hup − mutants, but not the single PHA − mutant, exhibited increased rates of H 2 photoproduction, about one third higher than that of the wild-type strain. All three mutants sustained a longer-term H 2 photoproduction phase than the wild-type strain, with the double mutant exhibiting overall the largest amount of H 2 evolved. The work of Franchi et al. (2004) hence demonstrated the feasibility of single and multiple gene engineering of microorganisms to redirect their metabolism for improving H 2 photoproduction using actual waste-derived substrates. Yet another interesting advance was made by Yoshino et al. (2007) where a strategy to establish cyanobacterial strains with high levels of H 2 production that involved the identification of promising wild-type strains followed by optimization of the selected strains using genetic engineering was developed. Yoshino et al. (2007) selected the Nostoc sp. PCC by virtue of it having the highest nitrogenase activity. After sequencing the uptake hydrogenase (Hup) gene cluster as well as the bidirectional hydrogenase gene cluster from the strain, and constructing a mutant (ΔhupL) by insertional disruption of the hupL gene, H 2 was produced a rate three times that of the wild-type. Lately, Kars et al. (2008) improved the hydrogen producing capacity of cells by introducing a suicide vector containing a gentamicin cassette in the hupSL genes into R. sphaeroiodes O.U.001. The wild-type and the mutant cells showed similar growth patterns but the total volume of hydrogen gas evolved by the mutant was 20% higher than that of the wild type strain. NH 4 + is typically an inhibitor to hydrogen production from organic wastewater by photo-bacteria. Recently, Zheng et al. (2009) found the biohydrogen generation with wild-type anoxygenic phototrophic bacterium R. sphaeroides was to be sensitive to NH 4 + due to the significant inhibition of NH 4 + to its nitrogenase. In order to avoid the inhibition of NH 4 + to biohydrogen generation by R. sphaeroides, a glutamine auxotrophic mutant R. sphaeroides AR-3 was obtained by mutagenizing with ethyl methane sulfonate. Zheng et al. (2009) noticed that the AR-3 mutant could generate biohydrogen efficiently for hydrogen production medium with a higher NH 4 + concentration, because the inhibition of NH 4 + to nitrogenase of AR-3 was released. Under suitable conditions, Zheng et al. (2009) successfully demonstrated that AR-3 could effectively produce biohydrogen from tofu wastewater, which normally contained 50-60 mg/L NH 4 + , with an average generation rate of 14.2 mL/L/h. The salient improvement was that the biohydrogen generation rate had more than doubled compared with that from the wild-type R. sphaeroides. Several other related studies have indicated that genetic, and hence metabolic, engineering (Vignais et al., 2006;Mathews and Wang, 2009) is a promising approach to the improvement of biological hydrogen production by existing microorganisms, particularly concerning the redirection and optimization of the flow of reducing equivalents to the H 2 -producing enzymes, nitrogenase or hydrogenase.

Conclusion
Hydrogen from both fossil and renewable biomass resources is a sustainable source of energy. Several of the methods and experimental techniques/technologies so far developed and tested point to the potential for practical and/or industrial application. Additionally, the use of modern bioreactors and specific substrates (food crops and lignocellulosic wastes) will be economically and environmentally less costly. The most commonly used technologies for biohydrogen production include direct and indirect biophotolysis, photo-fermentations, and dark-fermentation. These biohydrogen production technologies are still in the very early stage of research and development (R&D). Optimization of bioreactor designs and operational conditions for pH and microbial flora, testing and validation of biological, chemical and physical pretreatments, rapid removal and purification of gases, and genetic modifications of enzymatic metabolic pathways that compete with hydrogen producing enzyme systems offer exciting prospects for improving biohydrogen production systems. The specific areas of research for enhanced biohydrogen production would be (i) the reengineering of photosynthetic microorganisms for achieving high hydrogen production capacity; (ii) development of techniques for effectively separating and refining the hydrogen formed; and (iii) design of integrated systems for biohydrogen production followed by technical evaluations and cost-benefit analyses.

Declaration of interest
The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.