A WILEY-INTERSCIENCE PUBLICATION JOHN WILEY

Transcription

1 Cellulose Structure, Modification and Hydrolysis Edited by RAYMOND A. YOUNG and ROGER M. ROWELL University of Wisconsin Madison, Wisconsin and USDA Forest Products Laboratory Madison, Wisconsin A WILEY-INTERSCIENCE PUBLICATION JOHN WILEY & SONS New York Chichester Brisbane Toronto Singapore

2 16 Kinetic Modeling of the Saccharification of Prehydrolyzed Southern Red Oak ANTHONY H. CONNER, BARRY F. WOOD, CHARLES G. HILL, JR., and JOHN F. HARRIS USDA, Forest Service, Forest Products Laboratory and Department of Chemical Engineering, University of Wisconsin, Madison, Wisconsin A renewed interest in the use of biomass as a chemical raw material or as an energy source has developed over the past decade due, in large part, to fuel shortages in the early 1970s, to present political unrest in the Middle East, and to dwindling petroleum supplies predicted in the near future. Pyrolysis, gasification, liquefaction, and saccharification (hydrolysis) of cellulosics have all been proposed as methods for utilizing renewable biomass as an energy or chemical source. For the near term, saccharification of cellulosics to glucose, which can be fermented to fuel alcohol or a variety of other chemicals, is the method closest to implementation. Hydrolysis of cellulosics can be carried out either with cellulolytic enzymes or with mineral acids. Enzymatic hydrolysis occurs under mild conditions and gives good sugar yields, but the cost of the enzyme is high and the conversion rate of cellulose to glucose is slow (1). Acid hydrolysis of cellulosics can be carried out either in concentrated [e.g., the Bergius process (2)] or dilute [e.g., the American process (3), the Scholler process (4), or the Madison process (5)] acid solution. Hydrolysis of cellulosics in concentrated mineral acids at low temperature can give near quantitative yields of glucose (6). However, large quantities of concentrated acid are needed, a fact compounded by incomplete hydrolysis at low liquid-to-solid ratios (7). Therefore, for economic reasons acid recycling is imperative. In addition, plant 281

3 282 KINETIC MODELING OF THE SACCHARIFICATION costs will probably be excessive owing to the need for expensive corrosion resistant equipment. Dilute acid hydrolysis of cellulosics occurs under rather severe reaction conditions and gives poorer sugar yields. However, the conversion of cellulose to glucose is rapid, and the acid catalyst is cheaper than enzymes and does not have to be recycled. Dilute acid hydrolysis of cellulosics is the method closest to commercial implementation and is preferable from a process point of view (8). Current efforts to hydrolyze cellulosics with dilute acid are diverse and sometimes highly innovative. Of these processes, Rugg and Brenner (9) described equipment for a screw-fed, continuous, single-stage hydrolyzer, and Thompson and Grethlein (10) investigated a plug-flow reactor. Both processes take advantage of the improved kinetics of cellulose hydrolysis relative to glucose decomposition at higher temperatures and shorter retention times to improve sugar yields. Both also have the advantage of being simple one-stage processes. The U.S. Forest Products Laboratory (FPL) in cooperation with the Tennessee Valley Authority (TVA) has been studying a two-stage dilute acid hydrolysis process (11) based in part on studies of Cederquist (12) in Sweden during the 1950's. The first stage (prehydrolysis) selectively removes the hemicellulosic sugars with dilute sulfuric acid at about 170 C (13) prior to hydrolysis of the lignocellulosic residue to glucose in the higher temperature (230 C) second stage. Saeman (14) pointed out that a two-stage dilute acid process has a number of important advantages: (a) The carbohydrates are fractionated into hemicellulosic sugars and glucose, which facilitates the separate utilization of each fraction; (b) glucose is isolated in moderately good yield (~ 50%); (c) the glucose solution from the second stage is moderately concentrated (~ 10-12%); and (d) the consumption of acid and steam is relatively low. Despite recent efforts such as these to produce glucose from agricultural residues and wood, little attention has been given to improving the kinetic modeling of cellulose saccharification with dilute acid. Kinetic modeling plays an important role in the design, development, and operation of many chemical processes. Kinetic data are equally important in the design, evaluation, and operation of processes to hydrolyze cellulosics to fermentable sugar. In this chapter we describe a new model for the dilute acid hydrolysis of cellulose that was developed at FPL in connection with our studies on the twostage dilute sulfuric acid hydrolysis process. The model incorporates the effect of the neutralizing capacity of the substrate, the presence of readily hydrolyzable cellulose, and the reversion reactions of glucose in acid solution. Although general in nature, the model was developed specifically for application to the dilute sulfuric acid hydrolysis of prehydrolyzed wood. A computer program to simulate the new model under various hydrolysis conditions is presented. This program can be used to predict yields of free glucose, reducing sugars, reversion material, remaining cellulose, and glucose loss due to dehydration as a function of acid concentration, temperature, and reaction time.

4 CELLULOSE HYDROLYSIS MODEL 283 CELLULOSE HYDROLYSIS MODEL Luers (15) first recognized that the rate of glucose production from cellulose depends on the reaction rates for its formation and destruction and that these reactions proceed independently. However, he concluded from limited data that the two reactions were similarly affected by changes in acidity and temperature, and thus maximum glucose yields were unaffected by such changes. Saeman (16) showed that the rate of the hydrolysis reaction is increased more rapidly than that of the sugar dehydration reaction by both increased temperature and increased acid concentration. He modeled the system as two consecutive pseudofirst-order reactions in which the rate constants were functions of the applied acid concentration and temperature: cellulose glucose dehydration products Saeman s model, based on two consecutive reactions, is currently the only model used to simulate cellulose saccharification. This model is, however, an oversimplification of the reactions that actually occur in dilute acid solution. He recognized that glucose yields would be affected by the presence of neutralizing mated (i.e., ash-forming constituents) in the cellulose substrate and by the formation of reversion materials from glucose in dilute acid solution. Since these effects were less pronounced at the reaction conditions of the percolation process he was investigating, he did not incorporate them into his model. However, both factors are especially important in predicting glucose yields at the lower acidities and liquid-to-solid ratios encountered in some of the current processes under study. In addition, Saeman s model does not include the fact that all celluloses contain material that is rapidly hydrolyzed. This factor can, however, be readily introduced (10) into his model. Our new model, which includes elements to correct the shortcomings of the model limited to two consecutive pseudo-first-order reactions, is outlined in Figure This model incorporates the fact that cellulose contains both an easily hydrolyzable portion and a resistant portion, that reversion products (primarily disaccharides and levoglucosan-type materials) are formed from glucose in dilute acid solution, and that the disaccharides formed during reversion can degrade at the reducing end, giving a glycoside that on hydrolysis gives glucose and degradation products. The model depicted in Figure 16.1 can be described in terms of the following differential equations: (1) (2)

5 284 KINETIC MODELING OF THE SACCHARIFICATION Figure 16.1 Model for dilute acid hydrolysis of cellulosics. (3) (4) (5) (6) If the values of the various rate constants for a particular saccharification condition are known or can be derived, one can numerically integrate this set of equations to determine concentrations of intermediates at those conditions as a function of time. Cellulose Hydrolysis Rate Constant The rate constant for the hydrolysis of the resistant, portion of cellulose in prehydrolyzed Douglas fir was determined by Saeman (16, 17) and Kirby (18). A recorrelation (11) of their data, accounting for the acid-neutralizing capacity of the cellulosic substrate, showed that the pseudo-first-order rate constant for cthe resistant portion, k c, could be described as a function of acid concentration, [H + ], and temperature as follows: (7) Throughout this chapter all rate constants are expressed in units of min-1.

6 CELLULOSE HYDROLYSIS MODEL 285 where [H + ] = molal hydrogen-ion concentration, R = universal gas constant (calories/gmole- K), and T = temperature ( K). This equation applies specifically to the sulfuric acid hydrolysis of cellulose contained in Douglas-fir lignocellulose. Saeman (16) measured the rates ofhydrolysis ofthe resistantcellulose in five different wood species using dilute sulfuric acid. His data showed that rates differing by as much as 20% accompany changes in the substrate. Thus while the functional form of an equation to describe the cellulose hydrolysis rate constant for other species will have the sameform as that of Eq. (7), the values of the parameters will change with substrate as well as catalyst. Assuming the hydrolysis rate for southern red oak (Quercus falcata Michx.) is affected by temperature and acidity in a manner similar to that for Douglas fir, the following equation to describe the rate constant for the hydrolysis of southern red oak lignocellulose is obtained from experimental data (Table 16.1) by a least-squares procedure: (8) This equation differs from Eq. (7) only in the pre-exponential factor. Comparison of Eqs. (7) and (8) shows that at the same temperature and acidity southern red oak will hydrolyze slightly faster than Douglas-fir. This difference was observed previously by Saeman (16). In addition to the resistant portion of cellulose, all celluloses contain an easily hydrolyzable portion that reacts at a much greater rate. The quantity of easily hydrolyzable material depends on both the origin and prior treatment of the cellulosic substrate (19). The data of Millett et al. (20) show that approximately 10% of the cellulose contained in wood pulps and native celluloses hydrolyzes at a rate times as fast as the remaining 90%. For purposes of the model it was assumed that 10% of wood lignocellulose hydrolyzes at 50 times the rate of the resistant portion-thatis, k 1 = 50 k c (Fig. 16.1). Glucose Dehydration Rate Constant Of the information available on glucose dehydration (16, 17, 21, 22), the data of McKibbins et al. (21, 22) include the complete range of temperature and acidity of interest in cellulose saccharification. They measured the rate constant for the disappearance of glucose as reducing sugars, k,, in dilute sulfuric acid solution. The formation of reversion products in acid solution (see following discussion) contributes to the reducing power of the sugar solutions. Hence, k g (Fig. 16.1), the true rate constant for glucose dehydration, is related to but cannot be equated with k,. Values for k g were calculated from McKibbins experimental values of k, using the submodel (see following discussion) developed for the reversion reactions (23). The calculated k g values were then correlated by a least-squares method (23). The final correlation between the glucose dehydration rate constant and acid

7 TABLE 16.1 Experimental Data for Saccharification of Prehydrolyzed Southern RedOak Lignocellulose a 286

8 CELLULOSE HYDROLYSIS MODEL 287 concentration and temperature was found to be: (9) where [G] = molal glucose concentration The rate constant k g relates to the acid-catalyzed degradation reaction of glucose, but glucose and other sugars are also destroyed by various base catalyzed and uncatalyzed reactions (24). Above 0.05 molal [H + ] the disappearance of glucose is correctly predicted by Eq. (9); however, at lower acid concentrations the non-acid-catalyzed reactions become significant. Therefore, Eq. (9) should not be used below an acid concentration of 0.05 molal. Since only limited data (21, 22) are available on the rates of glucose destruction at hydrogen-ion concentrations below 0.05 molal, the application of the saccharification model is also limited by this restriction. Reversion Reaction Rate Constants In aqueous acid solutions, glucose not only reacts to form small quantities of its isomers, mannose and fructose, but also undergoes reversible reactions that result in disaccharides, oligosaccharides, and anhydrosugars. These reactions are referred to as reversion reactions, and the reaction products as reversion products. The reversion reactions result in components bonded only by glucosidic bonds; no evidence of ether bond formation exists. Although most of the isomeric disaccharides of glucose have been isolated under reversion-type reaction conditions (25-27), it has been found that the preferential mode of condensation is that between the anomeric hydroxyl and the C6-hydroxyl (i.e., the most prominent constituents of the reversion mixture are [1,6]-linked glucosides). Minor (28) showed that at glucose concentrations below 20%, the reversion products formed from glucose in dilute sulfuric acid solution consist almost exclusively of the disaccharides gentiobiose and isomaltose and the internally [1,6]-linked glucosans levoglucosan and 1,6-anhydro- ß-D-glucofuranose. At 180 C the disaccharides accounted for most of the reversion products, the glucosans constituting less than 30%. However, at 230 C the glucosans were the major reversion products. Smith et al. (29) also observed large amounts of levoglucosan in reversion products obtained at 220 C. These facts led to the development of a submodel for the formation of reversion products from glucose. The reaction scheme for this submodel was incorporated into the cellulose saccharification model (Fig. 16.1). In this scheme, the reversion products are assumed to consist only of monomeric, nonreducing glucosans (exemplified by levoglucosan) and of reducing disaccharides such as gentiobiose and isomaltose. It is further assumed that cellulose hydrolysis produces only monomeric glucose and that the disaccharides result exclusively from the reversion reactions. Also incorporated into this part of the model was the fact

9 288 KINETIC MODELING OF THE SACCHARIFICATION that in acid solution the reducing end of the disaccharide can be dehydrated to give a glucoside. Indeed, 2-furfuryl-5-methylglucoside, the expected glucoside from a degradation reaction of this type, has been observed among reversion products (30). Because this reaction is identical to the reaction by which glucose is destroyed, dehydration of the disaccharides would proceed at approximately the same rate as glucose dehydration (i.e., k 6 = k g ). This is not a critical assumption. Large variations in the assumed value of k 6 may be tolerated, since the quantity of disaccharide seldom exceeds 5% of the total potential glucose. Equilibrium constants for the two reversion reactions shown in Figure 16.1 can be defined as: (10) (11) where KL and KD are the equilibrium constants for the formation of levoglucosan and disaccharides, respectively. Values for KL and KD at 180 and 230 C were determined by Minor (28). These can be expressed analytically as: (12) and (13) Minor s data were taken at a single acid concentration; however, the principle of thermodynamic consistency requires that the forward and reverse reactions be accelerated to the same extent by changes in acidity. Thus, the expressions obtained above for KL and KD are independent of acidity. If one assumes that the rate constants for the hydrolysis of 1,6-glucosidic bonds are similar (i.e., k 3 = k 5 = k 7 ), and if one knows this rate constant relative to the glucose dehydration rate constant (k g ) then Eqs. (10) and (11) can be used to obtain estimates of the rate constants k 2 and k 4, expressed in terms of k g. The rates for the hydrolysis of all glucosides (31) are much greater than that for dehydration of glucose. The rate of hydrolysis of isomaltose is approximately 35 times the glucose dehydration rate; the ratio is higher for

10 COMPUTER SIMULATION OF THE SACCHARIFICATION MODEL 289 gentiobiose. It is assumed that k 3 = k 5 = k 7 and that these rate constants are all equal to 35 k,. Hence as a conservative estimate, k 2 = KL 35 k g and k 4 = KD 35 k g. Effective Acid Concentration The acidity of the reacting system, [H + ], is a function of the amount and concentration of the applied acid solution, the neutralizing capacity of the substrate, and, for sulfuric acid, the extent of the secondary ionization. This latter factor may be ignored, since measurements of the bisulfate dissociation constant (32) indicate that under the usual conditions of cellulose saccharification less than 1% of the bisulfate ion dissociates. Thus, to evaluate the hydrogen-ion concentration the sulfuric acid is assumed to release only one hydrogen ion. It is also assumed that all the cations in the cellulosicsubstrateare readily accessible and are effective immediately. For prehydrolyzed lignocelluloses this assumption is quite valid, since most of the ash constituents of these materials result from ion exchange during the water wash that follows prehydrolysis. Thus, with the further assumption that the hydroxyl-ion concentration is negligible, the hydrogen-ion concentration is the difference between the molality of the added acid solution and the molality of the cations (expressed as eq/kg of water). At low liquid-to-solid ratios and low acid concentrations, the neutralizing capacity of the lignocellulose can significantly lower the effective acid concentration. COMPUTER SIMULATION OF THE SACCHARIFICATION MODEL From the preceding discussion, it is readily apparent that Eqs. (1) to (6), which describe the saccharification model, can be rewritten in terms of the known constants k c (the hydrolysis rate constant for resistant cellulose), k, (the glucose dehydration rate constant), KL (the equilibrium constant for the formation of levoglucosan from glucose), and KD (the equilibrium constant for the formation of disaccharides from glucose). A BASIC computer program (33) which numerically integrates these transformed differential equations by the Runge-Kutta fourth-order method (34) is listed in Appendix I. This program allows one to calculate yields of monomeric glucose, total anhydroglucose, reducing sugar, and the remaining cellulose as a function of acid concentration, temperature, and reaction time. In the program it is assumed that the temperature varies in a manner approximating that occurring in the heat-up of the experimental reaction vessel. The temperature of the contents is assumed to have no spatial variation but to change with time according to Newton s law: (14)

11 290 KINETIC MODELING OF THE SACCHARIFICATION where T = temperature of reactants, TB = temperature of bath, X = time (min), and a = characteristic heat transfer coefficient for reaction = 13.8 min -1 for 5-mm glass ampuls. Calculation is simplified by assuming the water content of the reactants to be constant throughout the reaction. Although water is involved in the reactions, being produced by glucose dehydration and consumed by cellulose hydrolysis, the net variation was less than 1% over the range of experimental data. This assumption is equivalent to assuming a constant acid molality throughout the reaction. The same assumption was used in correlating the rate constants. The variables used in the program are defined in the program listing. The program is also annotated at various points to document how the program was constructed from the parameters discussed above. Table 16.2 contains an example of the output generated by this program; the program can be easily modified to output yield data for any reaction product. DILUTE SULFURIC ACID HYDROLYSIS OF PREHYDROLYZED SOUTHERN RED OAK Chips of southern red oak, impregnated with dilute sulfuric acid, were prehydrolyzed by direct steam heating for 6 min at 170 C as previously described (13). One-half of the lignocellulosic residue was slurried, fiberized, and washed using distilled water; the other half was treated similarly using tap water. After being washed, the wet residues were pressed to remove as much water as possible and then air-dried. The ash contents were determined using ASTM method No. D 1102 (35). Ash samples were titrated with dilute acid to determine their neutralizing capacity (36) as meq/kg oven-dry lignocellulose. The distilled waterwashed residue contained 0.18% ash and had a neutralizing capacity of 27.2 meq/kg. Corresponding values for the tap water-washedmaterial were 0.47% and meq/kg. Samples of the two residues were reacted in glass ampuls (5 mm OD) at a liquid-to-solid ratio of 3, using 0.8% H 2 SO 4 at 210 and 230 C. Approximately 200 mg of residue was placed in a tared 20-cm-long glass ampul sealed at one end. The sample was then dried (60 C in vacuo). The glass ampul was fitted with a rubber septum and evacuated. To obtain the desired liquid-to-solid ratio, acid solution was injected into the ampul through the septum with a syringe. The contents were brought to atmospheric pressure with N 2 and sealed. The saccharification reaction was carried out by placing the ampul in a molten salt bath controlled to ±0.1 C. The reaction time was taken as the interval from immersion in the salt bath to quenching in an adjacent water bath. After reaction, the ampul was opened and its contents were washed onto a tared sintered glass funnel with hot water. The solid residue was washed with approximately 100 ml of boiling water, dried overnight (60 C in vacuo), and weighed. The collected filtrate was diluted to a known volume for analysis.

12 TABLE 16.2 Cellulose Hydrolysis Yields Calculated from Model a 291

13 292 KINETIC MODELING OF THE SACCHARIFICATION The filtrates were analyzed for both total reducing sugars and monomeric glucose. Reducing sugars were measured by Nelson's colorimetric modification of the Somogyi method (37). Glucose was determined by high-performance liquid chromotography (HPLC) (38) using a Bio-Rad HPX-85 carbohydrate column maintained at 85 C. The column was eluted with distilled water, and the effluent was monitored with a refractive index detector. Carbohydrate analyses on the solid residues were performed using ASTM method No. D 1915 (39), except that the sugars were analyzed by HPLC as described above rather than by paper chromatography. The original prehydrolyzed southern red oak contained 57% glucan, 3% xylan, and 33% Klason lignin (40), reported on an ash-free oven-dried basis. The unaccounted 7% is probably soluble lignin. COMPARISON OF MODEL WITH EXPERIMENTAL DATA The experimental data obtained on the saccharification of two prehydrolyzed southern red oak residues by the methods presented above are shown in Table The two substrates differed only in their neutralizing capacity. This resulted in a difference in the effective hydrogen-ion concentration during the saccharification of the two substrates. Thus four groups of data, constituting a 2 2 factorial experiment with two temperatures and two acidities, were obtained. At each set of reaction conditions, measurements of monomeric glucose yield, reducing sugar yield, and remaining cellulose were made at time intervals that covered a wide extent of reaction. Comparisons of the experimental data with values of monomeric glucose and reducing sugars calculated with the model are shown graphically in Figures 16.2 and These comparisons illustrate that the model fits the experimental data quite well. This model has several advantages for use in process calculations. The various forms of glucose-containing species in solution can be estimated and may be treated as entities in process schemes. The estimated values can be used to interrelate values obtained from the various analytical methods used to assay carbohydrate content-forexample, reducing sugars, HPLC, and enzymatic methods. The model also estimates the amount of material lost through dehydration. From this, one can then estimate the quantities of the various sugar degradation products (i.e., hydroxymethylfufural and levulinic acid) in solution. USING THE MODEL WITH OTHER SUBSTRATES AND CATALYSTS Although the model is general in nature, it was developed specifically for application to the dilute sulfuric acid saccharifiation of prehydrolyzed wood lignocellulose. In principle the model can be used if other substrates or catalysts are used, but the exact values of parameters in the various equations of the

14 Figure 16.2 Comparison of glucose yields calculated from the model (solid line) with experimentally determined glucose yields (open symbols) for hydrolysis of prehydrolyzed southern red oak wood. The temperature for the hydrolysis reaction and the neutralizing capacities of the substrates are indicated in the figure. The liquid-to-solid ratio was 3:1. The added acid was sulfuric. Figure 16.3 Comparison of reducing sugar yields calculated from the model (solid line) with experimentally determined reducing sugar yields (open symbols) for hydrolysis of prehydrolyzed southern red oak wood. The temperatures for the hydrolysis reaction and the neutralizing capacities of the substrates are indicated in the figure. The liquid-to-solid ratio was 3:1. The added acid was 0.8% sulfuric. 293

15 294 KINETIC MODELING OF THE SACCHARIFICATION model will probably be different. Millett et al. (20), for example, have shown large differences in the rates of cellulose hydrolysis depending on the cellulosic substrate used (Fig. 16.4). It has also been shown (41) that hydrochloric acid, for example, is a much more effective catalyst than sulfuric acid (compared at equal hydrogen-ion concentrations) for the degradation of glucose, but only limited data are available on its effect on the cellulose hydrolysis rate at similar conditions. Further experimental data are needed that will allow this model to be used with a wide variety of substrates and catalysts. Kinetic studies on the saccharification of other cellulosics are under way. Work is also planned that will permit a more accurate description of the reversion reactions. These data will extend the applicability of this model. SUMMARY AND CONCLUSIONS The model presented successfully simulates the dilute sulfuric acid saccharification of prehydrolyzed southern red oak lignocellulose as determined by comparison of the experimental data with the values predicted for a given set of reaction conditions. The model is a considerable improvement over the simple model based on two consecutive reactions. It can be used to advantage for process calculations, since the various forms of glucose-containing species in solution can be estimated separately and may thus be treated as entities in process schemes. It takes into consideration the facts that the neutralizing capacity of the cellulosic substrate lowers the effective concentration of the added acid, that glucose gives reversion products in dilute acid solution, and that all cellulosics contain material that is readily hydrolyzable. Although the model Figure 16.4 Weight loss occurring during the heterogeneous hydrolysis of cellulose (20). The semilogarithmic relation to time is illustrated for a range of natural and regenerated celluloses.

16 REFERENCES 295 was developed specifically for application to the dilute sulfuric acid saccharification of prehydrolyzed wood lignocellulose, it can in principle be used if other substrates or catalysts are used. However, the values of parameters in the various equations of the model will be different, and they must be evaluated for the specific system of interest. REFERENCES 1. M. R. Ladisch, Proc. Biochem. 14, 21 (1979). 2. F. Bergius, Saccharification of Ground Wood, German Pat (1941). 3. M. F. Ewen and G. H. Tomlinson, Process of Producing Fermentable Sugar from Lignocellulose. U.S. Pat 1,032,392 (1912). 4. H. Scholler. Saccharifying Cellulosics. French Pat (1930). 5. E. E. Harris and E. Beglinger, The Madison Wood-Sugar Process, Rep. R1617, USDA, Forest Serv., Forest Prod. Lab., Madison, WI (1946). 6. I. S. Goldstein, Tappi 63, 141 (1980). 7. I. S. Goldstein, H. Pereira, J. L. Pittman, B. A. Strouse, and F. P. Scaringelli, Fifth Symposium on Biotechnologyfor Fuel and Chemicals, Gattinburg, TN H. E. Grethlein, Biotechnol. Bioeng., 20, 503 (1978). 9. B. Rugg and W. Brenner, Proc. 7th Energy Technol. Conf. andexposition, Washington, D.C., D. R. Thompson and H. E. Grethlein, Ind. Eng. Chem. Prod. Res. Dev., 18, 166 (1979). 11. J. F. Harris, A. J. Baker, A. H. Conner, T. W. Jeffries. J. L. Minor, R. C. Pettersen, R. W. Scott. E. L. Springer, and T. H. Wegner, Two-Stage, Dilute Sulfuric Acid Hydrolysis of Wood, Gen. Tech. Rep. FPL-45. Madison, WI: U.S. Department of Agriculture, Forest Service, Forest Products Laboratory; p. 12. K. N. Cederquist. Some Remarks on Wood Hydrotyzation. Report of a Seminar on the Production and Use of Power Alcohol in Asia and the Far East, Lucknow, India, United Nations Rep. ST/TAA/Ser. C/10 (1954). 13. R. W. Scott, T. W. Wegner, and J. F. Harris, J.; Wood Chem. Technol., 3, 245 (1983). 14. J. F. Saeman, Bioenergy 80 World Congress and Exposition. Atlanta, GA, H. Luers, Z. Angew. Chem., 43, 455 (1930). 16. J. F. Saeman, Ind Eng. Chem., 37, 43 (1945). 17. J. F. Saeman, A. M. Kirby, E. P. Young, and M. A. Millett, Rapid High Temperature Hydrolysis of Cellulose, Res. Mark. Act Stud. Rep. 30, USDA, Forest Serv., Forest Prod. Lab., Madison, WI (1950). 18. A. M. Kirby. Jr., Kinetics of Consecutive Reactions Involved in Wood Saccharification. Masters Thesis (Chem. Eng.), University of Wisconsin-Madison, Madison, WI (1948). 19. E. L. Springer, Cellulose Chem. Technol., 17, 525 (1983). 20. M. A. Millett, W. E. Moore, and J. F. Saeman, Ind. Eng. Chem., 46, 1493 (1954). 21. S. W. McKibbins, Kinetics ofthe Acid Catalyzed Conversion ofglucose to 5-Hydroxymethyl 2-Furaldehyde and Levulinic Acid, Ph.D. Thesis (Chem. Eng.), University of Wisconsin- Madison, Madison, WI (1958). 22. S. W. McKibbins, J. F. Harris, J. F. Saeman, and W. K. Neill, Forest Prod. J., 12, 17 (1962). 23. J. F. Harris, to be published. 24. R. P. Bell, Acid-base Catalysis, Oxford University Press, New York

17 296 KINETIC MODELING OF THE SACCHARIFICATION 25. A. S. Spriggs, Studies on the Composition of Starch Hydrolyzates and Glucose Reversion Mixtures, Ph.D. Thesis, Washington University, St. Louis, MO (1954). 26. A. Thompson, K. Anno, M. L. Wolfrom, and M. Inatoma. J. Am. Chem. Soc., 76, 1309 (1954). 27. S. Peat, J. Chem. Soc., 586 (1958). 28. J. L. Minor, J. Appl. Polym. Sci. Symp., 37, 617 (1983). 29. P. C. Smith, H. E. Grethlein, and A. O. Converse, Solar Energy, 28, 41 (1982). 30. Von A. Sroczynski and M. Boruch, Die Starke, 16, 215 (1964). 31. J. Szejtli, Sauerhydrolyse Glykosidscher Bindungen, Akademiai Kiado, Budapest, M. H. Lietzke, R. W. Stoughton. and T. F. Young, J. Phys. Chem. 65, 2245 (1961). 33. A FORTRAN version is available from the authors. 34. A. C. Noms, Computational Chemistry: An Introduction to Numerical Methods, Wiley, New York, American Society for Testing and Materials, Standard Test Method for Ash in Wood, ASTM Designation D E. L. Springer and J. F. Harris, Ind Eng. Chem., Prod. Res & Dev. (accepted). 37. N. Nelson, J. Biol. Chem., 153, 375 (1944). 38. R. C. Pettersen, V. H. Schwandt, and M. J. Effland, J. Chrom. Sci. 22, 478 (1984). 39. American Society for Testing and Materials, Standard Method for Chromatographic Analysis of Chemically Refined Cellulose, ASTM Designation D M. J. Effland, Tappi, 60, 143 (1977). 41. M. Losin and D. Sumnicht, An Experimental Study of the Kinetics of Glucose Decomposition in Dilute Acid, unpublished student report (Chem. Eng.), University of Wisconsin-Madison, Madison, WI (1983).