Saccharification of Cellulose Using a Hot-Compressed Water-Flow

Cellulose was hydrolyzed with hot-compressed water (HCW; ∼310 °C, 9.8 MPa) using an HCW-flow reactor. HCW was continuously delivered into a reactor...
0 downloads 3 Views 85KB Size
Ind. Eng. Chem. Res. 2002, 41, 661-665

661

KINETICS, CATALYSIS, AND REACTION ENGINEERING Saccharification of Cellulose Using a Hot-Compressed Water-Flow Reactor Tsuyoshi Sakaki,* Masao Shibata, Toshihisa Sumi,† and Seiji Yasuda National Institute of Advanced Industrial Science and Technology, Shuku-machi, Tosu, Saga 841-0052, Japan

Cellulose was hydrolyzed with hot-compressed water (HCW; ∼310 °C, 9.8 MPa) using an HCWflow reactor. HCW was continuously delivered into a reactor charged with cellulose. The effluent from the reactor was cooled and separated into a water-soluble fraction (WS) and a waterinsoluble fraction which deposited after cooling (DP). Cellulose started to decompose into WS and DP when HCW was delivered above 230 °C. The main components of the WS were hexose and oligosaccharides ranging from a dimer to a pentamer, and the DP consisted of polysaccharides ranging from a hexamer to an eicosamer or more polymerized saccharides. When HCW was delivered at 295 °C, nearly all of the cellulose decomposed after a 12-min flow of HCW, and 81 wt % of WS and 18 wt % of DP were obtained. The formation rates of WS and DP increased as the temperature of HCW increased. However, the composition of WS and DP barely changed with the HCW temperature under around 280 °C, which corresponds to the softening point of the cellulose. When the HCW temperature was further increased, the depolymerization of the products proceeded. The decomposition rate of cellulose itself was not affected by the flow rate of HCW, but the depolymerization of the products was suppressed when the flow rate of the HCW was increased. Introduction Cellulose is an abundant and renewable biomass resource that is receiving attention for its potential as an alternative to fossil fuels as an energy and chemical source.1 Cellulose is a linear polymer formed by β-1,4linked D-glucopyranose residues, and one of the major decomposition processes for cellulose is acid hydrolysis. However, reactor corrosion and wastewater treatment are problems associated with acid catalytic processes; therefore, noncatalytic hydrothermolysis is being studied.2-8 Bobleter et al.4 studied this process using cellobiose and indicated that hydrothermolysis has a reaction mechanism that is different from that of acidic hydrolysis. Cellulose was hydrolyzed to water solubles by water alone, and the decomposition rate was significantly increased when the water temperature reached the critical mark of water (374 °C).5 Furthermore, Sasaki et al.8 reported that the products of hydrolysis or saccharides could be produced with high yields from cellulose in supercritical water temperatures of around 400 °C; however, in subcritical water temperatures below 350 °C, the main components were the decomposition products of saccharides. On the other hand, we studied the saccharification of cellulose with hot-compressed water (HCW) at temperatures ranging from 300 to 400 °C in batch experiments6,7 and showed that hexose (glucose and fructose) * To whom correspondence should be addressed. Tel.: 81942-81-3631. Fax: 81-942-81-3692. E-mail: [email protected]. † Domestic Research Fellow of Japan Science Technology Corporation.

could be produced as main components, with yields as high as those reported by Sasaki et al.,8 under conditions as mild as those around 300 °C by adjusting the reaction time. The cause of the discrepancy between both results at subcritical temperatures is not clear, but a saccharification process at lower temperatures would be advantageous compared to a supercritical process from the viewpoints of energy and control. It is difficult to control the hydrolysis process at supercritical temperatures because of the extremely short reaction time,8 and, furthermore, the reactor material or nickel alloy has been reported to leach in supercritical water.9 However, the content of hexose in the water-soluble fraction did not exceed 50 wt % because of the secondary decomposition of the formed hexose in the closed reaction system.7 Therefore, we constructed a semicontinuous apparatus that permits the hot water including the solubilized products to be removed immediately from the reaction zone by delivering fresh hot water into the reactor continuously and investigated the saccharification of cellulose. As a result of using the HCW-flow reactor, the composition of the products was different from that obtained in the batch reactor, and oligo- and polysaccharides were the main components. These saccharides are available to the enzymatic monomerization10,11 for alcohol production. Furthermore, the production of these cellulosic oligo- and polysaccharides, which have been difficult to produce by the conventional methods such as acidic or enzymatic hydrolysis and synthesis, may lead to the creation of novel functional materials. In this paper, the formation behaviors of oligo- and polysaccharides from cellulose using an HCWflow reactor are reported.

10.1021/ie010614s CCC: $22.00 © 2002 American Chemical Society Published on Web 01/18/2002

662

Ind. Eng. Chem. Res., Vol. 41, No. 4, 2002

Figure 1. Schematic diagram of HCW-flow reactor: 1a,b, water tank; 2a,b, high pressure pump; 3a-c, valve; 4, heating coil; 5, salt bath; 6, line heater; 7, reactor; 8, cooler; 9, back-pressure regulator; 10, receiver; 11, nitrogen cylinder; 12a,b, pressure gauge; 13, thrmocouple; 14, recorder.

Experimental Section Materials. The cellulose used in this study is the same type as that used in the batch experiments.6,7 That is, microcrystalline cellulose powder (100-120 µm; Funakoshi Co.), which was prepared from refined pulp for column chromatography, was used. The reagents used for high-performance liquid chromatography (HPLC) analysis are glucose (Waco Pure Chemicals Industries, Ltd.), cellobiose (Funakoshi Co.), and acetonitrile (Wako Pure Chemicals Industries, Ltd.). Furthermore, the following reagents were purchased for the mass spectrometry analysis: 2,5-dihydroxybenzoic acid (2,5-DHB; Aldrich), amylose A (average molecular weight 1700; Nacalai Tesque Inc.), and cellotriose and cellohexaose (Seikagaku Corp.). HCW-Flow Reactor. The semicontinuous apparatus used is an HCW-flow reactor. A schematic diagram is shown in Figure 1. About 1 g of dried cellulose was charged in a reactor (SUS 316; 10.2 mm i.d × 46.7 mm length, 3.6 mL), and it was capped with gasket filters (average pore size 5 µm) so that powder samples would not flow out. The air in the system was replaced with nitrogen, and the back-pressure regulator was then set at 9.8 MPa. Distilled water was continuously delivered through a heating coil in the salt bath into the reactor by a high-pressure pump (2a). When the temperature of the reactor inlet reached 100 °C, the other pump (2b) was started to deliver distilled water to the reactor outlet at 2.5 mL/min to cool the products immediately and suppress secondary decomposition. The joined fluid was further cooled and was then collected at fixed intervals after passing the back-pressure regulator. The water-flow time was set at zero for convenience when the effluent first came out of the system. Also, the reaction temperature was represented by the HCW temperature measured at the inlet of the reactor. The temperature of the salt bath primarily controlled the HCW temperature. Moreover, the reactor and the pipe connecting the heating coil with the reactor were heated by a line heater to keep the temperature difference between the reactor inlet and the salt bath within 10 °C. The salt in the bath (3 L) was prepared by mixing 40 wt % NaNO2, 53 wt % KNO3, and 7 wt % NaNO3, which melted at around 160 °C. Two patterns of temperature control were adopted. The first called for raising the temperature to 285 °C for 5 °C/min after running water at 180 °C and 10 mL/ min for 20 min. After that, the heating was stopped,

Figure 2. Fractionation of decomposition products.

and the salt bath was lowered. Cold distilled water was then run to cool the reactor. The second pattern called for running 120 mL of HCW at fixed temperatures of 250-310 °C and at flow rates of 7.5-15 mL/min. After that, the reactor was cooled by the same procedure as that described above. The effluent from the system was not clear, but the suspended materials soon agglomerated and precipitated. The solid matters were not fine cellulose particles that came out through the gasket filter. Instead, they were found to be the polysaccharides formed through the hydrolysis of cellulose as described later. Therefore, as shown in Figure 2, the collected effluent was filtered using a No. 4 glass filter after allowing it to stand overnight. The water in the filtrate was distilled under vacuum, and the water-soluble fraction, WS, was obtained. The solid residue was dried, and the deposit fraction, DP, was obtained. WS and DP were weighed, and each product yield was calculated based on the dry cellulose weight. The decomposition yield of the cellulose was defined as the sum of the WS yield and the DP yield because the formation of gas was hardly recognized under the experimental conditions. Product Analyses. The components of WS were measured by HPLC equipped with two columns (SEC W12 + SEC W13; Yokogawa) and two detectors for ultraviolet (UV; 254 nm) and refractive index (RI) detection in series. HPLC was operated at 40 °C with a 0.8 mL/min flow of a mixture of water and acetonitrile (70/30 by volume) as an eluate. The DP was analyzed by MALDI-TOF MS. Details of this system can be obtained from our previous work.12 DP (2 µg) was suspended in 1 µL of water and mixed with 10 µL of a matrix solution (10 mg of 2,5-DHB in 1 mL of methanol). The mixture was vigorously shaken in a 1.5-mL Eppendorff tube on a vortex mixer. Then, 1 µL of the suspension was put on the sample plate and allowed to dry for about 5 min at room temperature. The plate was set in the Voyager RP (ion path length in the reflector mode, 3 m; matrix, 2,5-DHB; ionization, nitrogen laser; mode, positive ion mode) and analyzed using a Biospectrometry Workstation (PE Biosystems). A two-point external calibration was performed using cellotriose (m/z 504.4) and cellohexaose (m/z 990.9). Results and Discussion The group of Antal13,14 reported that cellulose was hardly solubilized when lignocellulosic biomass was treated with HCW at 190-230 °C. Therefore, we first

Ind. Eng. Chem. Res., Vol. 41, No. 4, 2002 663

Figure 3. Relation between decomposition of cellulose and temperature of HCW.

Figure 6. Mass spectrum of the DP fraction from 6-9-min HCW flow at 295 °C (see Figure 4). Values show the number of glucose units.

Figure 4. Decomposition of cellulose with HCW at a temperature of 295 °C and a flow rate of 10 mL/min.

Figure 7. Effect of HCW temperature on the formation of WS and DP at a flow rate of 10 mL/min. Figure 5. HPLC chromatogram of the WS fraction from 6-9min HCW flow at 295 °C (see Figure 4). Values show the number of glucose units.

examined the decomposition temperature of cellulose in HCW. HCW was delivered at 180 °C for 20 min, and then the HCW temperature was raised to 285 °C. The results are shown in Figure 3. Cellulose hardly decomposed at 180 °C and started to decompose to WS and DP when the temperature of the HCW reached around 230 °C. The integral decomposition yield of cellulose amounting to 79 and 21 wt % of the unreacted cellulose remained in the reactor when the HCW temperature reached 285 °C. When HCW was run at a fixed temperature of 295 °C, the cellulose was almost completely solubilized by the 12-min HCW flow including a 3-min heat-up period, and 81 wt % of WS and 18 wt % of DP were obtained,

as shown in Figure 4. Figure 5 shows the HPLC chromatogram for the main WS fraction, which was obtained during a 6-9-min HCW flow. Its main components were hexose and oligosaccharides ranging from dimer to pentamer, and the amount of secondary decomposition products such as furfurals was small compared to that in the WS obtained in the batch experiments.6,7 Concerning the DP, Sasaki et al.8 reported formation from the FTIR analysis of the precipitates that have cellulose-like molecular structure, but they could not specify the products. Figure 6 shows the mass spectrum for the main DP fraction from a 6-9-min HCW flow. The DP was found to consist of polysaccharides, from hexamer to eicosamer or more polymerized saccharides, indicating that even polymers formed from about 20 glucose units can be stripped off from cellulose by HCW

664

Ind. Eng. Chem. Res., Vol. 41, No. 4, 2002

Figure 8. Relation between the ratio of product yield and temperature of HCW. Product yields are obtained by delivering 120 mL of HCW at a flow rate of 10 mL/min.

at around 300 °C. These saccharides of high molecular weight would agglomerate and deposit by cooling. The effects of HCW temperature on the formation of WS and DP were examined by running HCW at temperatures from 250 to 310 °C and at a fixed flow rate of 10 mL/min. The formation rate of WS and DP increased as the temperature of HCW increased as shown in Figure 7. Although the decomposition of cellulose had not been completed by the 12-min HCW flow at 250 and 275 °C, the decomposition was nearly completed after a 6-min HCW flow at 310 °C. The final yield of DP obtained by the HCW flow at 310 °C was lower than that obtained at 295 °C, implying the conversion of DP to WS at a higher temperature. In fact, the distribution of polymers in the DP fractions was confirmed from the MS analysis to shift to a lower molecular weight by raising the HCW temperature from 295 to 310 °C. The saccharides, the degree of polymerization of which has been reduced to less than 5, are to be recovered as WS. Because quantitative analyses of oligomers that are larger than dimer are difficult, the ratios of the product yields of dimer to monomer and DP to WS are calculated to estimate the effect of HCW temperature on the composition of saccharides, where the product yields are those obtained by the 12-min HCW flow at 10 mL/min. As shown in Figure 8, both ratios barely changed at the lower HCW temperatures although the ratios decreased when the HCW temperature was further increased. The

HPLC analysis for the WS fractions and the MS analysis for the DP fractions also supported the little change in composition of the products with temperature at the lower HCW temperatures. The variation in decomposition behavior of cellulose around 280 °C may be related to the thermal softening of cellulose caused by the cleavage of hydrogen linkage in the cellulose crystal. We reported previously10 that the cellulose softened at around 280 °C. Besides, Sasaki et al.8 observed the shrinking rate of microcrystalline cellulose particles in HCW by using a diamond anvil cell and showed that the shrinking rate was extremely slow at around 250-280 °C compared to that at around 280320 °C. At temperatures lower than the softening point of cellulose, the hydrolysis would take place only on the surface of cellulose particles, and the formed species would be immediately separated from the surface and removed from the reactor, probably within 20 s in the case of 10 mL/min HCW flow. On the other hand, at temperatures above the softening point, HCW is assumed to penetrate into the cellulose particles and hydrolysis to take place inside the particles. The formed species would suffer from further decomposition to the smaller species during the dispersion in the particles. Figure 9 shows the effect of the flow rate of the HCW on the decomposition yield (WS + DP) of cellulose at 295 °C. The flow rate was found to barely affect the decomposition rate of cellulose itself under the experimental conditions. Consequently, the amount of water required for the decomposition of cellulose increased when the flow rate increased. To investigate the effect of the HCW-flow rate on the composition of the products, the ratios of each product yield of dimer to monomer and DP to WS are calculated and are indicated as a function of flow rate in Figure 10, where the product yields are obtained by delivering 120 mL of HCW at 295 °C. Both ratios increased with increasing flow rate, indicating that the depolymerization of the products was suppressed when the HCWflow rate was increased. These results reflect the change in the residence time of the products in the reactor. Consequently, the ratios of dimer to monomer and DP to WS could be controlled in the range of 0.32-0.55 and 0.10-0.33, respectively, by changing the HCW-flow rate from 7.5 to 15 mL/min.

Figure 9. Effect of the flow rate of HCW on the decomposition yield of cellulose at 295 °C.

Ind. Eng. Chem. Res., Vol. 41, No. 4, 2002 665

Figure 10. Relation between the ratio of product yield and flow rate of HCW. Product yields are obtained by delivering 120 mL of HCW at 295 °C.

Conclusions Oligo- and polysaccharides, which are formed from glucose molecules, could be produced in a hydrothermal process of cellulose at temperatures above 230 °C using an HCW-flow reactor. Hexose and oligosaccharides formed from 2-5 glucose units were recovered as WS, and polysaccharides formed from more than 5 glucose units were separated as DP. The temperature and the flow rate of HCW could control the degree of polymerization of the products. The depolymerization of the WS and DP proceeded with increasing HCW temperature above around 280 °C, which corresponds to the softening point of the cellulose, although the product composition barely changed with the HCW temperature at the lower temperatures. The decomposition rate of cellulose itself was not affected by the flow rate of HCW, but the depolymerization of the products was suppressed when the flow rate of the HCW was increased. Literature Cited (1) Antal, M. J., Jr.; Varhegyi, G. Cellulose Pyrolysis Kinetics: The Current State of Knowledge. Ind. Eng. Chem. Res. 1995, 34, 703-717.

(2) Bobleter, O.; Niesner, R.; Rohr, M. The Hydrothermal Degradation of Cellulosic Matter to Sugars and Their Fermentative Conversion to Protein. J. Appl. Polym. Sci. 1976, 20, 20832093. (3) Bobleter, O. Hydrothermal Degradation of Polymers Derived from Plants. Prog. Polym. Sci. 1994, 19, 797-841. (4) Bobleter, O.; Schwald, W.; Concin, R.; Binder, H. Hydrolysis of Cellulose in Dilute Sulfuric Acid and Under Hydrothermal Conditions. J. Carbohydr. Chem. 1986, 5, 387-399. (5) Adschiri, T.; Hirose, S.; Malaluan, R.; Arai, K. Uncatalytic Conversion of Cellulose in Subcritical and Supercritical Water. J. Chem. Eng. Jpn. 1993, 26, 676-680. (6) Sakaki, T.; Shibata, M.; Miki, T.; Hirosue, H.; Hayashi, N. Decomposition of Cellulose in Near-Critical Water and Fermentability of the Products. Energy Fuels 1996, 10, 684-688. (7) Sakaki, T.; Shibata, M.; Miki, T.; Hirosue, H.; Hayashi, N. Reaction Model of Cellulose Decomposition in Near-Critical Water and Fermentation of Products. Bioresour. Technol. 1996, 58, 197202. (8) Sasaki, M.; Fang, Z.; Fukushima, Y.; Adschiri, T.; Arai, K. Dissolution and Hydrolysis of Cellulose in Subcritical and Supercritical Water. Ind. Eng. Chem. Res. 2000, 39, 2883-2890. (9) Antal, M. J., Jr.; Allen, S. G.; Schulman, D.; Xu, X. Biomass Gasification in Supercritical Water. Ind. Eng. Chem. Res. 2000, 39, 4040-4053. (10) Sakaki, T.; Shibata, M.; Miki, T.; Hirosue, H. Oligosaccharification of Cellulose Using a Hot-Compressed-Water Flow Type Reactor and Enzymatic Hydrolysis of Formed Oligosaccharide. J. Jpn. Inst. Energy 1998, 77, 1111-1115. (11) Sakaki, T.; Shibata, M.; Miki, T.; Hirosue, H. Production method for water-soluble oligisaccharides and monosaccharide. Japan Patent 3,041,380, 2000. (12) Sumi, T.; Sakaki, T.; Ohba, H.; Shibata, M. Application of Matrix-assisted Laser Desorption/Ionization Time-of-flight Mass Spectrometry to Insoluble Glucose Oligomers in Decomposed Cellulose. Rapid Commun. Mass Spectrom. 2000, 14, 1823-1827. (13) Mok, W. S.-L.; Antal, M. J., Jr. Uncatalyzed Solvolysis of Whole Biomass Hemicellulose by Hot Compressed Liquid Water. Ind. Eng. Chem. Res. 1992, 31, 1157-1161. (14) Allen, S. G.; Kam, L. C.; Zemann, A. J.; Antal, M. J., Jr. Fractionation of Sugar Cane with Hot, Compressed, Liquid Water. Ind. Eng. Chem. Res. 1996, 35, 2709-2715.

Received for review July 19, 2001 Revised manuscript received November 13, 2001 Accepted November 21, 2001 IE010614S