The stem part of a dwarf type of napiergrass was dried and powdered by a blender until the powder passed through a sieve with 150 μm of mesh. The powdered napiergrass (5.0 g) was treated with a 1% aqueous solution of NaOH at 95 °C for 1 h in order to remove colored materials such as lignin and chlorophyll, which disturbed the light-absorption of the photocatalyst. The holocellulose (a mixture of cellulose and hemi-cellulose) was isolated as a pale yellow precipitate by centrifugation. The colored materials were dissolved in the aqueous NaOH solution. Lignin was collected as dark brown precipitate by centrifugation of the supernatant NaOH solution which was neutralized to pH 5.0 by a dilute HCl solution. As for the results, 2.70 g (54.0%) of holocellulose and 0.61 g (12.2%) of lignin were obtained from 5.0 g of powdered napiergrass. The residue (1.69 g, 33.8%) contained ash components and others.

The holocellulose was turned into the reducing saccharides (1) by three methods shown in Figure 1. In Method A, the holocellulose (2.5 g) was turned into pentose and hexose by hydrolytic enzyme (Equations 3 and 4). The SA was performed using a hydrolytic enzyme (250 mg, Acremozyme, cellulase from acremonium, Kyowa Kasei) [4] in an acetate buffer solution (60 mL, pH 5.0) for 48 h under vigorous shaking at 45 °C. The solution was subjected to centrifugation to produce the supernatant solutions (1a) containing both pentose (0.63 g) and hexose (1.27 g). HPLC analysis showed that the pentose and hexose were mainly xylose and glucose, respectively. Holocellulose was recovered in 0.49 g, which was 80.4% of the conversion of saccharification.

After the Method A, the 1a was subjected to the fermentation in an acetate buffer solution (180 mL, pH 5.0) containing the 1a (12.49 g) and the suspension solution (3.6 mL) of S. cerevisiae at 35 °C (Method B). The reaction was monitored by the CO₂-evolution (Equation 5). After the CO₂-evolution was stopped for 107 h, ethanol (3.2 g) was formed. Pentose (3.6 g) was remained without a reaction. After the ethanol was removed from the reaction mixture by evaporation under reduced pressure, an aqueous pentose solution (1b) was obtained and subjected to the following photocatalytic reaction.

In Method C, the holocellulose was converted into ethanol and pentose by the simultaneous saccharification and fermentation process (SSF) according to Equations 3, 4, and 5. SSF was performed in an acetate buffer solution (180 mL, pH 5.0) containing holocellulose (16.2 g), Acremozyme (3.0 g), and the suspension solution (3.6 mL) of S. cerevisiae. The CO₂-evolution was stopped for 89 h. The SSF process produced ethanol (3.69 g) and pentose (3.85 g). The ethanol was removed by evaporation of the solution to produce an aqueous pentose solution (1c). Table 1 summarizes the products yields based on 100 g of napiergrass.

The Pt/TiO₂ (100 mg) was suspended in an aqueous solution (150 mL) containing 1 (0–249 mg) and the oxygen was purged by bubbling with N₂. The suspended solution was irradiated by a high-pressure mercury lamp under vigorous stirring with a magnetic stirrer. The band gap of anatase-type of TiO₂ is known to be 3.20 eV which is corresponding to 385 nm. Therefore, the TiO₂ can be excited by 366 nm-emission from high-pressure mercury lamp [10]. The evolved gas was collected by mess-cylinder to measure the total volume of the evolved gas. The irradiation was performed until the gas evolution ceased. The quantitative analysis of hydrogen, oxygen, nitrogen, and carbon dioxide were performed by GLC. The results are shown in Table 2. Figure 3 shows the plots of the evolved H₂, CO₂, and O₂ volumes against the amounts of 1 used. The evolved H₂ and CO₂ volumes increased gradually with the increase of 1. When smaller amounts of 1 were used, the volume ratio of H₂ to CO₂ (H₂/CO₂) exceeded over 2.0 which was the stoichiometric value of Equations 6 and 7, because of the dissolution of CO₂ into aqueous reaction solution (Table 2). However, with the use of higher amounts of 1, the H₂/CO₂ became close to 2.0.

Moreover, it was confirmed that the H₂ evolution from water was small (2 mL) in the absence of the sacrificial reagent. At the same time, 18 mL of O₂ was evolved. Therefore, the volume ratio of H₂ became O₂ (H₂/O₂) in the photocatalytic reaction over stoichiometric value (2.0) [11]. Also, in the presence of sacrificial agent, a considerable amount of O₂ was evolved. At the present time, the evolution mechanism of O₂ is still under investigation. Other gases such as CH₄ and CO [12] were not observed in the evolved gas.

According to Equations 6 and 7, the 10 and 12 equivalents of H₂ were theoretically obtained from 1 mole of pentose and hexose, respectively. However, the molar ratio of the evolved H₂ to 1 (H₂/1) did not reach the theoretical values. Moreover, the H₂/1 depended on the amount of 1 used. Therefore, the H₂/1 values were plotted against the weight ratio of 1 to catalyst (1/catalyst), as shown in Figure 2. As the 1/catalyst values decreased, the H₂/1 values increased. The intercept of the plots represents the limiting H₂/1 values (N) at an infinite amount of a catalyst. The N values were nearly equaled to the theoretical values. Judging from these results, it is suggested that the deactivation of the catalyst occurred at the high turnover.

The amounts of the hydrogen obtained from the photocatalytic reaction of 1 were estimated by multiplying N by the moles of 1: The amount of H₂ in g = 2 N (W_P/150 + W_H/180) where W_P and W_H denoted the weight of pentose and hexose. The results are shown in Table 1. Thus, the amounts of H₂ were determined to be 4.23, 1.66, and 1.74 g from 1a, 1b, and 1c which were derived from 100 g of napiergrass, respectively.

The three processes were compared from the standpoint of the combustion energy (∆H/kJ mol⁻¹) of the bio-fuel produced by biological treatment and PC reaction, as shown in Table 1. The Method A process and PC produced 4.23 g of H₂. The total ∆H was calculated to be 603 kJ using the ∆H of H₂ (285 kJ mol⁻¹) and ethanol (1,367 kJ mol⁻¹) [9]. The Method B and PC processes produced ethanol (10.7 g) and H₂ (1.66 g) whose total ∆H was 555 kJ. Also the Method C and PC process gave ethanol (12.3 g) and H₂ (1.74 g) whose total ∆H was 615 kJ. Thus, the combination of the SSF process (Method C) with the PC process was most effective process. This ∆H was close to the 639 kJ which was ∆H of 27.4 g of hexose and 13.6 g of pentose which were formed from 100 g of napiergrass.

Anatase-type of TiO₂ (ST-01) was purchased from Ishihara Sangyo, Japan. According to the literature [13], an aqueous solution (50 mL) containing TiO₂ (1.0 g), K₂PtCl₆ (10–100 mg), and 2-propanol (0.38 mL) was irradiated by high-pressure mercury lamp for 24 h with stirring to give the Pt-loaded TiO₂ catalyst (Pt/TiO₂). The optimized Pt-content on TiO₂ was determined to be 2.0 wt% by the comparison of the amounts of hydrogen-evolution by the Pt-doped TiO₂ (100 mg) under irradiation by high-pressure mercury lamp for 6 h using glucose (100 mg) as sacrificial reagent. Thus the Pt/TiO₂ (2 wt% of Pt) was used throughout the present investigation.

The amount of the saccharides (1) formed by the enzymatic saccharification (SA) process was analyzed by the modified Somogyi-Nelson method assuming the composition of 1 as C₆H₁₂O₆ [14]. Also the amounts of hexose and pentose were analyzed by a Shimadzu LC-20AD high-performance liquid chromatography system using anion exchange column (Shodex Asahipak NH2P-50 4E). Ethanol concentrations were determined by a Shimadzu GC-2014 gas chromatograph using a glass column of 5% Thermon 1000 on Sunpak-A (Shimadzu) with 2-propanol as an internal standard. Hydrogen, carbon dioxide, and nitrogen were analyzed on a Shimadzu GC-8A equipped with TCD detector at temperature raised from 40 to 180 °C using a stainless column (3 mmΦ, 6 m) packed with a SHINCARBON ST (Shimadzu).

A dwarf type of napiergrass (Pennisetum purpureum Schumach; dwarf variety of late-heading type) [3] was cultivated in the Sumiyoshi Ranch, Faculty of Agriculture, University of Miyazaki. The powdered stem part of the napiergrass (30 g) was treated with a 1% aqueous solution of NaOH (400 mL) at 95 °C for 1 h. Holocellulose was isolated as pale yellow precipitate by centrifugation. The neutralization of the supernatant solution to pH 5.0 by a dilute HCl solution gave a dark brown precipitate of lignin which was collected by centrifugation.

The holocellulose (2.5 g) was dispersed in an acetate buffer solution (60 mL, pH 5.0) and a hydrolytic enzyme (Acremozyme, Kyowa Kasei, 250 mg) was added to the sterile solution. The saccharification was performed by under vigorous shaking at 45 °C for 48 h [4]. The solution was subjected to centrifugation at 12,000 rpm to produce the supernatant solutions of 1 which were used in the following process.

Saccharomyces cerevisiae NBRC 2044 was cultured at 30 °C for 24 h in a basal medium (initial pH 5.5) consisting of glucose (20 g L^–1), bactotryptone (1.0 g L^–1, Difco), yeast extract (1 g L^–1), NaHPO₄ (1 g L^–1), and MaSO₄ (3 g L^–1) [4]. After incubating for 24 h, the cell suspension solution of S. cerevisiae was obtained. The suspension solution (3.6 mL) of S. cerevisiae was added to a solution of 1 (180 mL). The fermentation (FE) was performed at 35 °C with stirring using a magnetic stirrer until the evolution of CO₂ was stopped.

The simultaneous saccharification and fermentation process (SSF) was performed as follows. The holocellulose (16.2 g) was dispersed in an acetate buffer solution (180 mL, pH 5.0) and was sterilized in autoclave at 120 °C. Acremozyme (3.0 g) and the cell suspension (3.6 mL) of S. cerevisiae were added to the suspension. The suspension was stirred vigorously with a magnetic stirrer at 35 °C until the evolution of CO₂ was stopped.

The catalyst (usually 100 mg) and the aqueous solution of 1 (usually 5.0 mL) were introduced to a reaction vessel. The volume of the reaction solution was adjusted to 150 mL with water (usually 145 mL). A high-pressure mercury lamp (100 W, UVL-100HA, Riko, Japan) was inserted into the reaction vessel, which was attached to a mess-cylinder to correct the evolved gas and set in a water bath to keep it at a constant temperature (usually 20 °C) (Figure 4). After the oxygen was purged by nitrogen gas, irradiation was performed with vigorous stirring using magnetic stirrer. The evolved gas was collected by a mess-cylinder to measure the total volume of the evolved gas (Figure 4(A)). The quantitative analysis of hydrogen, nitrogen, and carbon dioxide were performed by GLC using the peak area (A) measured by the GLC-analysis, the sample gas volume (V₃, 500 μL), and the slopes of the calibration curves (f) which were determined to be 3380, 334, 388 and 377 for H₂, N₂, O₂, and CO₂, respectively (Figure 4(B)). According to Equation (8) where V₁ was the total volume of evolved gas (mL) corrected by mess-cylinder, V₂ was the volume (mL) of the dead space of the vessel before reaction (usually 230 mL), the amounts of for H₂, N₂, O₂, and CO₂ were obtained

The amounts of gas (mL) = A (V₁ + V₂)/(f ×V₃) (8)

In order to efficiently irradiate the Pt/TiO₂ suspended in solution, the amount of the Pt/TiO₂ was optimized using 150 mg of 1a. When 100 mg (1.25 mmol) of the Pt/TiO₂ (2.0 wt% of Pt content) was used for an aqueous solution (150 mL), the largest amounts of H₂ evolved, as shown in Figure 5.