Fermentative hydrogen production from agroindustrial lignocellulosic substrates

To achieve economically competitive biological hydrogen production, it is crucial to consider inexpensive materials such as lignocellulosic substrate residues derived from agroindustrial activities. It is possible to use (1) lignocellulosic materials without any type of pretreatment, (2) lignocellulosic materials after a pretreatment step, and (3) lignocellulosic materials hydrolysates originating from a pretreatment step followed by enzymatic hydrolysis. According to the current literature data on fermentative H2 production presented in this review, thermophilic conditions produce H2 in yields approximately 75% higher than those obtained in mesophilic conditions using untreated lignocellulosic substrates. The average H2 production from pretreated material is 3.17 ± 1.79 mmol of H2/g of substrate, which is approximately 50% higher compared with the average yield achieved using untreated materials (2.17 ± 1.84 mmol of H2/g of substrate). Biological pretreatment affords the highest average yield 4.54 ± 1.78 mmol of H2/g of substrate compared with the acid and basic pretreatment - average yields of 2.94 ± 1.85 and 2.41 ± 1.52 mmol of H2/g of substrate, respectively. The average H2 yield from hydrolysates, obtained from a pretreatment step and enzymatic hydrolysis (3.78 ± 1.92 mmol of H2/g), was lower compared with the yield of substrates pretreated by biological methods only, demonstrating that it is important to avoid the formation of inhibitors generated by chemical pretreatments. Based on this review, exploring other microorganisms and optimizing the pretreatment and hydrolysis conditions can make the use of lignocellulosic substrates a sustainable way to produce H2.


Introduction
H 2 is a promising fuel: it is carbon-free and its combustion produces only water (Wang and Wan, 2009). Although H 2 constitutes a clean fuel, currently available methods leading to its production, such as methane reforming and partial oil and coal oxidation, demand fossil fuels and a high amount of energy (Chaubey et al., 2013). Biological approaches that produce H 2 offer several advantages over current physicochemical methods: they occur at ambient temperature and pressure, and they use renewable raw materials as substrates (Li and Fang, 2007;. A number of microbes belonging to a wide variety of bacterial groups can perform fermentative H 2 production, also called dark fermentation because it does not require light. The strict anaerobe Clostridium spp. and facultative anaerobes from the family Enterobacteriaceae are the most often cited H 2 -producing bacteria (Seol et al., 2008, Elsharnouby et al., 2013. (Valdez-Vazquez and Poggi-Varaldo, 2009, Kothari et al., 2012, Show et al., 2012Rafrafi et al., 2013).
However, it is the choice of substrate for fermentative H 2 production that determines the feasibility of the process. The substrate should (1) be carbohydrate-rich, (2) originate from renewable resources, (3) suffice for fermentation, and (4) promote energetically favorable energy recovery. In addition, any necessary pretreatment should be inexpensive (Wang and Wan, 2009;Chaubey et al., 2013). In this context, several investigators have turned to lignocellulosic materials to produce H 2 (Kapdan and Kargi, 2006;Lin et al., 2012). According to Kotay and Das (2008), if the use of these resources is appropriately controlled, they will become a major source of energy in the future. Unfortunately, these residues have a complex chemical structure and often call for previous treatment and/or hydrolysis to serve as substrate for biological H 2 production. Such pretreatment and/or hydrolysis could not only alter the physicochemical features of the waste, making carbohydrates available for fermentation, but also afford byproducts that negatively interfere in fermentative H 2 production.
This review compares the yields of fermentative H 2 production from (1) different agroindustrial lignocellulosic substrates without any chemical or biological pretreatment (2) lignocellulosic materials after a pretreatment step and (3) hydrolysates of lignocellulosic materials originating from a pretreatment step followed by enzymatic hydrolysis. The comparison of these results will show how the pretreatment and hydrolysis of lignocellulosic substrates affect fermentative H 2 production. In addition, this review will present the microorganisms involved in H 2 production from those materials.

Lignocellulosic Materials as Substrate for Fermentative H 2 Production
Lignocellulosic materials are the most abundant residues derived from agroindustrial activities; therefore, they can potentially become a significant source of renewable H 2 (Saratale et al., 2008;Levin et al., 2009;Ren et al., 2009;Hay et al., 2013). Agricultural residues from harvested crops are the cheapest and the most abundant readily available lignocellulosic organic waste; they include straw, stover, peelings, cobs, stalks, and bagasse (Guo et al., 2010a;. All these residues can undergo biological transformations to varying degrees, as well as conversion to hydrogen (Guo et al., 2010a).
Lignocellulosic materials consist primarily of cellulose, hemicelluloses, and lignin. Thus, the main products of the enzymatic, chemical, or thermochemical hydrolysis of lignocellulosic materials are hexoses, mainly glucose, and pentose sugars, mainly xylose.
In addition to H 2 , the anaerobic digestion of glucose by strict anaerobes or facultative microorganisms yields different final products. Depending on the bacterial species, pH, and H 2 partial pressure, the fermentation of glucose can result in H 2 , CO 2 , acetate and/or butyrate (Eqs. 1 and 2). Theoretically, when the final product is acetate only, 4 mol of H 2 /mol of glucose can emerge (Eq. 1). However, if the final product is butyrate, only 2 mol of H 2 /mol of glucose arises (Eq. 2).
Xylose is the major pentose derived from the hydrolysis of hemicelluloses, which in turn constitutes approximately 20 to 30% of plant biomass. It can be used for the growth and energy production of numerous microorganisms. The use of xylose as a substrate for ethanol production has been extensively studied (Sun and Cheng, 2002;Lin and Tanaka 2006;Sarks et al., 2014). However, only recently has attention been given to H 2 production from xylose fermentation. Theoretically, similarly to glucose fermentation, xylose fermentation can produce 3.33 mol H 2 /mol xylose when acetate is the fermentation product (Eq. 3). When butyrate is the fermentation product, 1.66 mol of H 2 /mol of xylose will emerge (Eq. 4) (Martin del Campo et al., 2013

1
(4) Figure 1 shows the main steps of the metabolic pathways and enzymes leading to H 2 production throughout glucose and xylose fermentation performed by anaerobic microorganisms. The figure shows that the enzyme xylose isomerase (XI) catalyzes the isomerization of xylose to xylulose. The latter is then phosphorylated by xylulokinase (XK), to afford xylulose-5-phosphate, one of the intermediates of the pentose phosphate (PP) pathway. Through the activities of epimerase, isomerase, transketolases, and transaldolases, enzymes of the PP pathway, xylulose-5phosphate is converted to fructose-6-phosphate and glyceraldehyde-3-phosphate. Both of these compounds are intermediates of the EMP pathway, through which they undergo conversion to pyruvate. The supposed activities of pyruvate, ferredoxin oxyreductase (PFOR) and ferredo-xin-dependent hydrogenase (Hyd) will produce H 2 , CO 2 , and acetate.
According to the Figure 1, glucose is converted to pyruvate, from which H 2 , CO 2 , and acetate are produced, as outlined above. It is noteworthy that for both carbohydrates, the consumption of reducing power to generate butyrate instead of acetate reduces the H 2 yield.
To produce H 2 by fermentation, it is possible to use (1) lignocellulosic materials without any chemical or biological pretreatment, (2) lignocellulosic materials after a pretreatment step, or (3) hydrolysates of lignocellulosic materials that normally originate after a pretreatment step followed by enzymatic hydrolysis. Another approach is to conduct simultaneous saccharification and fermentation (SSF), which consists in adding a hydrolytic enzyme(s) or microorganisms to a fermentation vessel (Quemeneur et al., 2012a).

Pretreatment of Lignocellulosic Materials for Fermentative H 2 Production
The complex nature of lignocellulosic substrates may adversely affect their biodegradability. Therefore, prehydrolysis, often referred to as pretreatment, is required to alter the structure of lignocellulosic biomass to make the sugars available for fermentation , Levin et al., 2009. Carbohydrate polymers (cellulose and hemi-cellulose) and lignin are the main components of lignocellulosic materials (Rezende et al., 2011;Mood et al., 2013). Agricultural residues such as wheat straw, corn stalk, sugarcane bagasse, and rice straw contain approximately 32-47% cellulose, 19-27% hemicellulose, and 5-24% lignin (Sun and Cheng, 2002). Although hemicellulose and lignin are minor components, they protect cellulose. Hence, it is necessary to hydrolyze these components, to efficiently use the cellulose (Mosier et al., 2005;Rezende et al., 2011). Thus, appropriate pretreatment steps reduce the cellulose crystallinity and/or polymerization degree and selectively remove hemicellulose and lignin to make carbohydrates from lignocellulosic materials accessible for enzymatic hydrolysis Monlau et al., 2013a).
The main pretreatment methods rely on mechanical, physical, chemical, and biological techniques or a combination thereof (Alvira et al., 2010;Guo et al., 2010b;Ogeda and Petri, 2010). These methods serve to prepare lignocellulosic materials for bioethanol production mainly, but most of them also find application in fermentative H 2 production (Guo et al., 2010a;Mood et al., 2013;Monlau et al., 2013a).
Physicochemical pretreatment includes steam explosion, steam explosion with ammonium, use of organic solvents and supercritical fluids, and use of diluted acids Fermentative hydrogen production 325 Figure 1 -Schematic view of the major metabolic pathways that lead to the production of H 2 , CO 2 , and acetate from the carbohydrate components obtained from the hydrolysis of lignocellulosic materials. EMP, Embden-Meyerhoff-Parma; Fd, oxidized ferredoxin; FdH 2 , reduced ferredoxin; Hyd, hydrogenase; PFOR, pyruvate: ferredoxin oxyreductase; PP, pentose phosphate; XI, xylose isomerase; XK, xylulokinase. The dashed arrows indicate multisteps of a metabolic pathway. and/or bases (Mosier et al., 2005;Vargas Betancur and Pereira Jr, 2010;Monlau et al., 2013b). Biological pretreatment relies on the ability of fungi and bacteria to produce enzymes such as lignin peroxidase and laccase, and hemicellulase, which help to remove lignin and hemicellulose from the lignocellulosic matrix, respectively (Ogeda and Petri, 2010). Various methods for pretreating lignocellulosic material exist; however, it is essential to select a method that minimizes carbohydrate degradation and avoids the formation of inhibitory compounds that are toxic to fermentative microorganisms (Alriksson et al., 2011;Rezende et al., 2011;Jonsson et al., 2013). Pretreatment at high temperatures rapidly degrades hemicellulose pentoses and to a lesser extent hexoses, producing acetic acid and furfurals, which constitute potential fermentation inhibitors (Alriksson et al., 2011;Jonsson et al., 2013). Figure 2 shows the main carbohydrate degradation products from hemicelluloses and cellulose hydrolysis, i.e., xylose and glucose, as well as furfural, hydroxymethylfurfural (HMF), and organic acids, such as formic and acetic acid (Palmqvist and Hahn-Hagerdal, 2000;Jonsson et al., 2013).
Furfural originates from pentose dehydration; its concentration in the liquid phase increases with rising pretreatment temperature, acid concentration, or pretreatment time (Chen et al., 2013). Furfural may react further, to yield formic acid, or it may polymerize. Hydroxymethylfurfural (HMF) stems from the dehydration of hexoses such as glucose; it can further react to yield levulinic and formic acid (Palmqvist and Hahn-Hagerdal, 2000;Chen et al., 2013;Jonsson et al., 2013). These inhibitors may interfere with cell functions and osmotic pressure; they can even directly inhibit the acid fermentation pathway (Palmqvist and Hahn-Hagerdal, 2000).
Acetic acid is an inhibitory substance that also exists in hydrolysates. It is formed by the hydrolysis of acetyl groups in hemicellulose and, to some extent, lignin (Klinke et al., 2004). In the undissociated form, acetic acid can pen-etrate the cell membrane and inhibit product formation, disrupting the pH balance at high concentration, inhibiting cell growth or even killing cells (Klinke et al., 2002). However, some strains can use acetic acid as a substrate to produce H 2 (Matsumoto and Nishimura, 2007;Xu et al., 2010).
Aromatics may arise in hydrolysates depending on the type of pretreatment applied and on the ratio of p-coumaryl alcohol, coniferyl, and sinapyl alcohol, the main lignin monomers. Pretreatment can transform lignin into a complex mixture of low-molecular-weight or "monomeric" phenolic compounds, especially by acid impregnation (Klinke et al., 2004;Chen et al., 2013). Phenolic compounds are well known for being toxic to microbial cells. They bear carboxyl, formyl, and hydroxyl groups, which increase the fluidity of the membrane and affect its permeability .
In summary, the pretreatment of lignocellulosic material to use it as a substrate for producing H 2 may generate fermentation inhibitors as well as other unusual substrates, such as pentose (xylose) and/or oligosaccharides (Maintinguer et al., 2011;Quemeneur et al., 2012b), which is a major drawback.
The use of xylose as a substrate appears to be less problematic than the presence of inhibitory compounds because xylose can be metabolized as illustrated in Figure 1. Indeed a series of H 2 -producing microorganisms, such as Clostridium spp. (Maintinguer et al., 2011); Enterobacter spp. CN1 (Long et al., 2010); and the thermophiles Thermoanaerobacterium saccharolyticum Shaw et al., 2008), Thermotoga neapolitana DSM 4359 (Ngo et al., 2012), Caldicellulosiruptor saccharolyticus (de Vrije et al., 2009) and Thermoanaerobacterium thermosaccharolyticum (Khamtib and Reungsang, 2012), can consume and produce hydrogen from xylose. Ren et al. (2008) reported that T. saccharolyticum W16 can ferment a mixture of glucose and xylose with a H 2 yield of up to 2.37 mol of H 2 /mol of substrate.
However, inhibitors such as furan derivatives and phenolic compounds negatively affect H 2 production by 326 Reginatto and Antônio mixed cultures. According to Quéméneur et al. (2012), furans exert a more negative effect than that induced by phenolic compounds. These authors found that Clostridium beijerinckii strains resisted these inhibitors better than other clostridial and non-clostridial bacteria did; therefore, C. beijerinckii is a promising microorganism for H 2 production from lignocellulosic hydrolysates. Tai et al. (2010) observed that higher phenol concentrations (1 g/L) significantly inhibited C. butyricum metabolism. Nevertheless, no metabolic inhibition or co-degradation occurred at concentrations of approximately 0.6 g/L. Veeravalli et al. (2013) observed that furans affected fermentative H 2 production by a mixed anaerobic culture. Furan levels of up to 1 g/L favored propionate and ethanol generation, decreasing H 2 production.
In conclusion, the main limitation of using pretreated lignocellulosic materials in fermentative H 2 production is the presence of these inhibitors.

H 2 Production From Non-Pretreated Lignocellulosic Materials
Because pretreatment processes are expensive and can produce inhibitory compounds, it would be beneficial to avoid pretreatment and directly convert lignocellulosic materials to H 2 (Levin et al., 2009;Raj et al., 2012).
Only a few reports concerning the production of H 2 from untreated lignocellulosic feedstocks exist in the literature , and most of them involve thermophilic microorganisms. For example, Clostridium thermocellum ATCC 27405 and C. saccharolyticus DSM 8903 can hydrolyze cellulose and hemicellulose to produce H 2 (Raj et al., 2012).
C. saccharolyticus can produce H 2 directly from mechanically comminuted switchgrass without any chemical or biological pretreatment (Talluri et al., 2013).
Some authors have resorted to co-cultures that allow for the use of lignocellulosic materials as substrates. Wang et al. (2008) reported that a co-culture consisting of Clostridium acetobutylicum and Ethanoigenens harbinense effectively hydrolyzed cellulose and produced H 2 from microcrystalline cellulose.  developed a co-culture of C. thermocellum and C. thermosaccharolyticum, to improve hydrogen production via the thermophilic fermentation of cornstalk waste. The authors achieved a hydrogen yield of 68.2 mL of H 2 /g of cornstalk, 94.1% higher than the yield obtained using a monoculture of C. thermocellum. Table 1 lists results for fermentative H 2 production from lignocellulosic materials without any chemical pretreatment, the employed inocula, and the H 2 yield obtained from these substrates. The results are presented as maximum assessed production yield, as indicated by the authors; when possible, we converted the data and expressed them as maximum calculated production yield (mmol of H 2 /g of substrate) for comparison. All the wastes included in Table 1 were milled before being assayed.
The temperature clearly affected the fermentative H 2 production yield from lignocellulosic residues. Most of the studies that used untreated lignocellulosic materials employed thermophilic conditions (10, n = 14) to provide yields approximately 75% higher than those obtained under mesophilic conditions. Although most studies employed a mixed culture as an inoculum, C. thermocellum and T. thermosaccharolyticum, previously known as C. thermosaccharolyticum were the thermophilic microorganisms most frequently employed to produce H 2 from untreated feedstock.
The untreated raw materials presented in Table 1 afforded an average maximum calculated H 2 production yield of 2.17 (± 1.84) mmol of H 2 /g of substrate; yields ranged from 0.12 to 11.2 mmol of H 2 /g of substrate. The only study on switchgrass furnished the highest yield -11.2 mmol of H 2 /g of substrate (Talluri et al., 2013). When we excluded this study from the calculations, the average H 2 production yield from untreated lignocellulosic substrates decreased to 1.41 (± 1.02) mmol of H 2 /g, where the highest average yield observed was that obtained for cornstalk -2.16 (± 1.17) mmol of H 2 /g.

H 2 Production From Pretreated Lignocellulosic Materials
Although some studies on the direct conversion of lignocellulosic materials to H 2 exist, most microorganisms require pretreated lignocellulosic material as a substrate to produce biohydrogen. The degree of pretreatment depends on the nature of the raw material and on the inoculated organism(s) .
Most pretreatment steps generate undesirable inhibitors, but they significantly enhance H 2 production. Zhang et al. (2007) improved biohydrogen production from cornstalk after acidification and heat pretreatment. The authors achieved maximum cumulative H 2 production of 150 mL of H 2 /g of VS after treating the substrate with 0.2% HCl; this production was 50 times higher than the value obtained without pretreatment. Cornstalks treated with NaOH (0.5%) furnished 57 mL of H 2 /g of VS, i.e., 19-fold the initial value obtained for the raw material (3 mL of H 2 /g of VS) (Zhang et al., 2007). Table 2 summarizes literature results concerning the use of pretreated lignocellulosic wastes, the pretreatment type, the inoculum, and the H 2 yield obtained from these substrates. The results shown in Table 2 refer to the maximum assessed production yield, as indicated by the authors; when possible, we converted the data and expressed them as maximum calculated production yield (mmol H 2 /g of substrate) for comparison.
Acid and base pretreatment have been the pretreatments most frequently employed to prepare lignocellulosic materials for biohydrogen production -11 and 6 studies, respectively, from the 21 publications presented in Table 2 have been reported. Enzymatic and/or biological pretreatment represent 3 of the 21 studies shown in Table 2. Only one study involved the use of temperature alone.
As indicated by the maximum calculated production yield data presented in Table 2, the biological pretreatment afforded the highest average yield 4.54 (± 1.78) mmol of H 2 /g of substrate compared with the acid and basic pretreatment (2.94 ± 1.85 and 2.41 ± 1.52 mmol of H 2 /g of substrate, respectively). Therefore, pretreatment effectiveness depended on the feedstock and pretreatment conditions, such as acid or base concentration, exposure time, and temperature.
According to Table 2, the average H 2 production yield from pretreated material was 3.17 (± 1.79), ranging from 0.68 to 8.11 mmol of H 2 /g of substrate for corn stover and cornstalk, respectively. Pretreated cornstalk furnished the highest average yield 4.74 (± 1.80) mmol of H 2 /g of substrate, which was approximately 2.2 times higher that yielded by untreated cornstalk (2.17 ± 1.84 mmol of H 2 /g of substrate, Table 1). Therefore, the pretreatment step enhances H 2 production.
Most studies used a mixed culture of microorganisms previously enriched with H 2 -producing bacteria as an inoculum. The thermophilic T. thermosaccharolyticum was the pure culture most frequently employed in the studies using pretreated lignocellulosic wastes as substrates.

H 2 Production From Lignocellulosic Materials Hydrolysates
The structural changes that prehydrolysis (pretreatment) promotes in a lignocellulosic matrix positively affect the subsequent enzymatic hydrolysis of lignocellulosic materials, increasing the saccharification yield . Several authors have used this strategy to increase the concentration of sugars in hydrolysates for H 2 production (de Vrije et al., 2009;Cui et al., 2010;Luo et al., 2011;Pan et al., 2011;Monlau et al., 2013b). Pan et al. (2011) pretreated cornstalk containing 81.7% TVS with dilute acid, i.e., 1.5% H 2 SO 4 , at 121°C for 60 min, followed by enzymatic hydrolysis at 52°C, pH 4.8, with an enzyme loading of 9.4 IU/g, to obtain a total soluble sugar content of 562.1 ± 6.9 mg/g of TVS during the stages of hydrolysis. The maximum hydrogen yield from this hydrolysate using 328 Reginatto and Antônio  an anaerobic mixed culture was calculated in terms of grams of cornstalk (TVS) as 209.8 mL of H 2 /g of TVS. Pretreatment followed by enzymatic hydrolysis is a very efficient method for saccharifying lignocellulosic substrates. However, depending on the type of substrate and pretreatment conditions employed, the hydrolysates could inhibit fermentative H 2 production. Monlau et al. (2013b) verified that hydrolysates from sunflower stalks pretreated with dilute acid negatively affected H 2 -producing microflora. The dilute acid pretreatment condition that these authors employed (170°C, 1 h, 4 g of HCl/100 g of TS) was highly efficient in hydrolyzing hemicellulosic material because approximately 3.14 g/L of xylose and only 0.28 g/L of glucose emerged in the slurry. In addition to the amount of xylose, other byproducts arose -formate (0.6 g/L) and acetate (0.81 g/L), and furan derivatives such as furfural (1.15 g/L) and HMF (0.13 g/L). In a batch system inoculated with mixed microflora, 15% of this hydrolysate completely inhibited H 2 production.
In a long-term experiment, Arreola-Vargas et al. (2013) observed that partial replacement of a synthetic medium containing glucose and xylose with an acid and with an enzymatic hydrolysate of oat straw, in a continuous reactor, diminished H 2 production. The acid hydrolysate consisted mainly of glucose 1.5 g/L and xylose 3.7 g/L as well as phenolic compounds, such as HMF (133.2 mg/L), furfural (0.6 mg/L), and vanillin (3.59 mg/L). The enzymatic hydrolysate contained 3.8 g/L of glucose and 1.3 g/L of xylose, but no HMF, furfural, or vanillin. Both hydrolysates were used to feed an anaerobic sequencing batch reactor by gradually substituting the glucose/xylose medium with the hydrolysates. The substitution of glucose/xylose by the acid hydrolysate disaggregated the granules and interrupted the process. On the other hand, the replacement of the glucose/xylose medium with the enzymatic hydrolysate without fermentation inhibitors elicited H 2 production. However, the H 2 yield and production rate decreased from 2 mol of H 2 /mol of sugar and 278 mL of H 2 /L.h to 0.81 mol of H 2 /mol of sugar and 29.6 mL H 2 /L.h, respectively, in going from the synthetic medium to the enzymatic hydrolysate (Arreola-Vargas et al., 2013).
Simultaneous saccharification and fermentation (SSF) has been successfully conducted to produce H 2 from pretreated or even untreated lignocellulosic substrates by adding hydrolytic enzyme(s) or by seeding hydrolytic enzymes produced in the same fermentation vessel. Thus, in this approach, no pretreatments or only mild conditions for pretreating substrates are necessary, diminishing the formation of fermentation inhibitors (see Figure 2) because most saccharification occurs simultaneously with the fermentation (Lakshmidevi and Muthukumar, 2010;Quemeneur et al., 2012a;. For example, Quemeneur et al. (2012a) used a mixed culture of microorganisms and evaluated the efficiency of exogenous enzyme addition during fermentative H 2 production from wheat straw. The authors used two experimental designs: a onestage system (direct enzyme addition) and a two-stage system (enzymatic hydrolysis prior to fermentation). H 2 production from untreated wheat straw ranged from 5.18 to 10.52 mL of H 2 /g of vs. H 2 production yields increased two-fold and ranged from 11.06 to 19.63 mL of H 2 /g of VS after the enzymatic treatment of the wheat straw. Direct addition of exogenous enzymes during one-stage dark fermentation was the best way to improve H 2 production from lignocellulosic biomass. Table 3 summarizes the lignocellulosic material hydrolysates used as substrates for fermentative H 2 production, the pretreatment and enzymatic hydrolysis methods used, the source of inoculum or the microorganisms involved in the fermentation, and the process yields and/or rates. Results regarding H 2 yields from hydrolysates are expressed in terms of mmol of H 2 /mmol of sugar or mmol of H 2 /g of substrate because it was not always possible to convert these units. In the last case, it was possible to compare data with the results of untreated and pretreated substrates (Table 1 and 2).
According to Table 3 the H 2 production yields from hydrolysates ranged from 0.45 to 13.39 mmol of H 2 /g of substrate, for wheat straw and sugarcane bagasse, respectively.
Cornstalk is the most often studied lignocellulosic substrate for H 2 production. The average yield using a cornstalk hydrolysate for biohydrogen production is 5.93 mmol of H 2 /g of substrate, which is approximately 270% and 25% higher than that afforded by the untreated (2.17 mmol of H 2 /g of substrate) and pretreated cornstalk (4.74 mmol of H 2 /g of substrate), respectively. The results demonstrated that after pretreatment and/or hydrolysis, this substrate is potentially applicable in biohydrogen production.
Although sugarcane bagasse afforded the highest yield -13.39 mmol of H 2 /g of TVS; this figure represents the results obtained in only one study (Chairattanamanokorn et al., 2009). The average H 2 production yield per mol of sugar of pretreated bagasse was 1.23 mol of H 2 /mol of glucose (Table 2); for the hydrolysates, this yield dropped to 1.12 (Table 3), demonstrating that H 2 production from hydrolysates of this substrate was slightly lower.
Excluding the work of Chairattanamanokorn et al. (2009) with sugarcane bagasse, the average H 2 production yield with sugarcane bagasse hydrolysates (Table 3) was 3.78 ± 1.92 mmol of H 2 /g, 20% higher compared with the average yields of pretreated substrates. However, this average H 2 production yield was lower than that of biologically pretreated substrates, 4.54 ± 1.78 mmol of H 2 /g. These results demonstrate the importance of avoiding the presence of inhibitors originating from chemical pretreatment methods. 330 Reginatto and Antônio Maximum production yield in terms of mmol of H 2 /g of substrate. b* Maximum production yield in terms of mmol of H 2 /g of total volatile solids (TVS) or volatile solids (VS) contained in the substrate. c Maximum production rate in terms of mmol of H 2 /h.g TVS. nd: not determined.

Conclusions and Perspectives
Based on this review, converting agroindustrial lignocellulosic substrates to H 2 by fermentative microorganisms is a feasible solution for producing H 2 sustainably. However, additional research into the pretreatment of lignocellulosic wastes for biohydrogen production is desirable to improve the yield and make the process economically viable. Efforts to control the formation (or removal) of toxic compounds (such as furan derivatives, phenolics, and organic acids, formed during the chemical pretreatment) are necessary because these could clearly inhibit H 2 fermentation. Biological pretreatment methods afford higher H 2 yields from lignocellulosic materials because they do not produce inhibitors.
The development of microbial strains or consortia resistant to inhibitors remains an important research area. Moreover, the discovery of novel H 2 -producing microorganisms able to use lignocellulosic derivatives is associated with different environmental conditions, particularly high temperatures.
Currently, results have shown that corn stalk submitted to a pretreatment step and/or hydrolysis furnishes a higher average yield of biohydrogen production than that afforded by other agroindustrial lignocellulosic substrates. Exploring other microorganisms and optimizing the pretreatment and hydrolysis conditions can make the use of this substrate and other agroindustrial residues a sustainable way to produce clean H 2 .