Agricultural and Food Science in Finland, Vol.11 (2002):51 –58 51 A G R I C U L T U R A L A N D F O O D S C I E N C E I N F I N L A N D Vol. 11 (2002): 51–58. © Agricultural and Food Science in Finland Manuscript received May 2001 Production of a cellulosic substrate susceptible to enzymatic hydrolysis from prehydrolyzed barley husks Ana Belén Moldes, José Manuel Cruz, José Manuel Domínguez and Juan Carlos Parajó Departamento de Enxeñería Química, Universidade de Vigo (Campus Ourense), Edificio Politécnico, As Lagoas, ES - 32004 Ourense, Spain, e-mail: jcparajo@uvigo.es An effective process for the chemical-biotechnological utilization of barley husks is reported. A first treatment with sulfuric acid (prehydrolysis) allowed the solubilization of hemicelluloses to give xy- lose-containing liquors (suitable to make fermentation media for xylitol production) and a solid phase containing cellulose and lignin. The solid residues from prehydrolysis were treated with NaOH in order to increase their cellulase digestibility. In the alkaline treatments, the effects of temperature (in the range, 50–130ºC), reaction time (10–60 min) and NaOH concentration (3–10 weight percent of solution) on the composition of solid residues were assessed by means of an experimental plan with factorial structure. The cellulose content increased with temperature and NaOH concentration, whereas the duration of treatments was not influential within the range tested. The treated samples showed high susceptibility toward the enzymatic hydrolysis with cellulases, leading to almost quantitative glucose yields under selected operational conditions. Key words: barley husks, hydrolysis, alkali treatment, cellulose, xylitol Introduction Barley husk is an agricultural byproduct whose direct utilization as a carbohydrate source (for example, as feed supplement or for manufacture of glucose or ethanol) is hindered by its low di- gestibility. On the other hand, the combustion of barley husk is difficult owing to its compara- tively high ash content. In this scope, the sequen- tial processing of barley husk with sulfuric acid and NaOH allows the separation of two fractions (hemicellulose as soluble sugars in the first processing step and a cellulosic solid phase in the second processing step) which can be uti- lized by fermentation and enzymatic hydrolysis, respectively. The processing of lignocellulosic substrates in acidic media (prehydrolysis) is carried out to convert the hemicellulose polysaccharides (xy- lan, mannan and galactan) into the correspond- ent monosaccharides (xylose, mannose and glu- cose). As xylan is the main hemicellulosic poly- mer in barley husk, xylose is the major sugar present in the prehydrolysis liquors obtained from this raw material. After neutralization and mailto:jcparajo@uvigo.es 52 A G R I C U L T U R A L A N D F O O D S C I E N C E I N F I N L A N D Moldes, A.B. et al. A process for the chemical-biotechnological utilization of barley husks nutrient supplementation, the hydrolyzates can be used as fermentation media to produce xyli- tol, a low caloric pentitol with sweetening pow- er which has negative heat of solution, anticari- ogenic properties and is suitable as sugar sub- stitute for diabetics (Parajó et al. 1997, Cruz et al. 2000a). Using continuous fermentation with cell recycle, the production of xylitol from bar- ley husk hydrolyzates can be carried out with productivities as high as 2.53 g/L · h (Cruz et al. 2000b). Owing to the selective hemicellulose removal caused by prehydrolysis, the solid residue from this processing step is enriched in both cellulose and lignin. In order to convert it in a suitable substrate for enzymatic hydrolysis, a delignifi- cation stage (for example, with NaOH) must be performed. As lignin forms a physical barrier hindering the access of enzymes to cellulose, delignification should result in increased acces- sibility, and so in an improved glucose yield with faster reaction rate. Alkaline treatments achieve other collateral effects (for example, alteration in the physicochemical features of cellulose such as crystallinity and surface area) leading to im- proved susceptibility toward enzymatic hydrol- ysis (Gharpuray et al. 1983). This study deals with the chemical-biotech- nological processing of barley husk. In a first step, the hemicellulosic fraction was converted into xylose-containing liquors suitable for xyli- tol production. The solid phase from prehydrol- ysis was treated with NaOH under a variety of operational conditions, and the effects of treat- ments on both composition and hydrolysis sus- ceptibility of cellulosic substrates were meas- ured. Material and methods Raw material Barley husk was kindly provided by San Martín (Ourense, Spain), where husks were separated from the grain by pneumatic conveying and cy- clone separation. Husks were stored in a dry and dark place at room temperature until utilization. Analysis of the raw material Aliquots from the homogenized lots were sub- jected to moisture determination and to quanti- tative hydrolysis in two-stage, acid treatment (the first step with 72 weight percent sulfuric acid at 30ºC during 1 hour, and the second one after di- lution of the media to 4 weight percent sulfuric acid at 121ºC during 1 hour) (Vázquez et al. 1991). The solid residue after hydrolysis was considered to be Klason lignin. Hydrolyzates were assayed by HPLC using an Interaction ION- 300 column (mobile phase, H2SO4 0.01 M; flow rate, 0.4 mL/min; IR and UV detection). This method allowed the direct determination of glu- cose, xylose, arabinose, acetic acid, ethanol, xylitol, furfural and hydroxymethylfurfural (HMF). Representative material balances are presented in Figure 1. Chemical processing of samples Ground samples of barley husks were hydrolyzed under selected conditions (3% H2SO4, 15 min, 130°C, liquid:solid ratio of 8:1 g/g). The solids from treatments were separated by filtration, washed with water, air dried and treated in auto- clave with solutions containing 3–10% NaOH at 50–130ºC during 10–60 min. In this step, the liquor/solid ratio was fixed in 10 g/g. At the end of treatments, the solid residues were separated by filtration, washed with water, air dried and analyzed as described for the raw material. Enzymatic hydrolysis The commercial enzyme concentrates (“Cellu- clast” and “Novozym”, with cellulase and β-glu- cosidase acitivities, respectively) used in exper- iments were kindly provided by Novo, Denmark. 53 A G R I C U L T U R A L A N D F O O D S C I E N C E I N F I N L A N D Vol. 11 (2002): 51–58. The cellulase activity of concentrates was as- sayed by the Filter Paper Activity test (FPA) ac- cording to Mandels et al. (1976) and expressed as Filter Paper Units (FPU)/mL. The β-glucosi- dase activity was measured according to Paquot and Thonart (1982). The operational conditions used in enzymatic hydrolysis were: temperature, 48.5ºC; pH, 4.85; liquor/solid ratio, 30 g/g; cel- lulase/substrate ratio, 28 FPU/g and cellobiase/ cellulase ratio, 13 IU/FPU (Moldes et al. 2000). Experimental design and statistical analysis In the alkaline stage of barley husk processing, an incomplete 33 factorial design (Box et al. 1978) was used to study the influence of tem- perature, reaction time and NaOH concentration on both the composition of the solid substrates from treatments and their susceptibility to enzy- matic hydrolysis. The experimental data were analyzed by the Response Surface methodology using the Statistica 5.0 software. The interrela- tionship between dependent and operational var- iables was established by a model including lin- ear, interaction and quadratic terms: Y = b0 + b1 · x1 + b2 · x2 + b3 · x3 + b12 · x1·x2 + b13 · x1 · x3 + b23 · x2 · x3 + b11 · x1 2 + b22 · x2 2 + b33 · x3 2 where Y is the dependent variable, b denotes the regression coefficients (calculated from experi- mental data by multiple regression using the least-squares method), and x denotes the inde- pendent variables. Fig. 1. Scheme proposed for the chemical-biotechnological processing of barley husks. 54 A G R I C U L T U R A L A N D F O O D S C I E N C E I N F I N L A N D Moldes, A.B. et al. A process for the chemical-biotechnological utilization of barley husks Results and discussion The sequential treatment of barley husk with sulfuric acid and sodium hydroxide allows the benefit of hemicelluloses and cellulose. Figure 1 shows some details on the composition of liq- uors and solid residues from treatments. The hydrolyzates coming from the first step have been successfully employed for the production of xylitol, and the corresponding results have been reported in a previous study (Cruz et al. 2000b). Using cell recycle after membrane sep- aration, the fermentation of hydrolyzates with the yeast Debaryomyces hansenii led to a maxi- mum volumetric productivity of 2.53 g/L · h at a dilution rate of 0.284 h–1. In order to assess the possibility of reaching a simultaneous benefit of both liquid and solid phases coming from the prehydrolysis step, pre- liminary enzymatic hydrolysis assays were car- ried out using the solid residues from prehydrol- ysis as substrates. As expected, these substrates showed a poor susceptibility toward enzymatic hydrolysis (data not shown) owing to their high lignin content. In order to improve the results, a delignification stage with NaOH was carried out before the enzymatic hydrolysis. The independent variables used in this study and their variation ranges were: temperature T, 50–130ºC; duration of treatments t, 10–60 min and NaOH concentration [NaOH], 3–10 weight percent of solution. The standardized (coded) adimensional variables employed, having vari- ations limits (–1,1), were defined as x1 (coded temperature), x2 (coded time) and x3 (coded NaOH concentration). The correspondence be- tween coded and uncoded variables was estab- lished by linear equations deduced from their respective variation limits (see Table 1). Table 1 also lists the dependent variables considered: the composition of delignified sol- ids was measured by variables y1 (cellulose con- tent of samples, g/100 g oven-dry substrate), y2 (hemicellulose content of samples, g/100 g oven- dry substrate) and y3 (Klason lignin content of samples, g/100 g oven-dry substrate); whereas y4 (defined as conversion of cellulose into glu- cose) was selected to measure the susceptibility of prehydrolyzed, alkali-treated samples toward enzymatic hydrolysis. It can be noted that the processed samples also contained other fractions different from those measured by y1, y2 and y3 (such as acid-soluble lignin, acetyl groups, ash- es, etc.) which were of minor importance for this study. Table 1. Variables used in this study. Variable Nomenclature Units Variation range Independent variables Temperature T ºC 50–130 Time t Min 10–600 NaOH concentration [NaOH] wt % 3–10 Dimensionless, coded independent variables Dimensionless temperature x 1 (T – 90)/40 (–1,1) Dimensionless time x 2 (t – 35)/25 (–1,1) Dimensionless NaOH concentration x 3 ([NaOH] – 6.5)/3.5 (–1,1) Dependent variables Cellulose content, g /100 g o. d. sample y 1 Hemicellulose content, g /100 g o. d. sample y 2 Lignin content, g /100 g o. d. sample y 3 Cellulose conversion into glucose, g glucose/100 g potential glucose y 4 55 A G R I C U L T U R A L A N D F O O D S C I E N C E I N F I N L A N D Vol. 11 (2002): 51–58. Since a systematic study of the effects caused by the operational variables on composition and hydrolysis susceptibility would require a great amount of experimental work, an incomplete, factorial design of experiments was carried out. Several research groups used phenomenological models based on experimental designs to study the chemical processing and/or bioconversion of lignocellulosic materials (Parajó et al. 1995, Roberto et al. 1995, Alves et al. 1998, Mayer- hoff et al. 1998, Silva and Roberto 1999). In this study, we utilized an incomplete factorial design, in which 3 dependent variables were assayed at 3 levels. Based on experimental data, equations including linear, interaction and quadratic terms were employed to describe the interrelationship between operational and experimental variables. Table 2 shows the set of experimental condi- tions assayed (expressed in terms of coded vari- ables), as well as the experimental data obtained for variables y1 to y4. The sequence for the ex- perimental work was randomly established to limit the influence of systematic errors on the interpretation of results. It can be noted that ex- periments 13–15 are replications in the central point of the design measuring the experimental error. Table 3 lists the regression coefficients and their statistical significance (based on a t-test). The same Table includes statistical parameters (r2 and F) measuring the correlation and the sta- tistical significance of the models, respectively. It can be noted that all the models showed good statistical parameters for correlation and signif- icance and allowed a close reproduction of ex- perimental data. In the range tested, the reaction time caused only minor effects on the cellulose content of samples, as it can be seen from the absolute val- ue of the corresponding coefficients. Figure 2 shows the predicted dependence of the cellulose content of samples (y1) on the most influential operational variables (T and [NaOH]) in experi- ments lasting 35 min. Increased severity (defined by high values of temperature and/or NaOH con- centration) resulted in remarkable increases in the cellulose content of samples. The effects caused by increased alkali concentrations were Table 2. Operational conditions considered in this study (expressed in terms of the coded independent variables dimension- less temperature x 1 , dimensionless time x 2 and dimensionless NaOH concentration x 3 ) and experimental results achieved for the dependent variables y 1 (cellulose content of samples, %), y 2 (hemicellulose content of samples, %), y 3 (lignin content of samples, %) and y 4 (cellulose conversion into glucose, %). Operational conditions Experimental results Exper. x 1 x 2 x 3 y 1 y 2 y 3 y 4 1 0 –1 –1 68.3 5.2 24.0 89.6 2 0 1 –1 64.2 3.0 27.0 99.7 3 0 –1 1 71.0 2.8 22.5 91.9 4 0 1 1 79.8 2.5 16.2 91.0 5 –1 –1 0 68.2 4.6 24.0 79.6 6 –1 1 0 69.3 4.2 23.0 86.5 7 1 –1 0 80.1 3.0 15.6 96.0 8 1 1 0 86.1 4.2 06.8 91.9 9 –1 0 –1 64.5 6.3 26.0 77.6 10 –1 0 1 70.2 3.4 23.5 86.8 11 1 0 –1 75.9 3.0 18.8 93.6 12 1 0 1 84.7 3.0 11.4 94.3 13 0 0 0 75.0 3.1 19.0 99.7 14 0 0 0 75.0 3.1 17.5 98.8 15 0 0 0 74.8 2.9 18.6 100.4 56 A G R I C U L T U R A L A N D F O O D S C I E N C E I N F I N L A N D Moldes, A.B. et al. A process for the chemical-biotechnological utilization of barley husks more marked when the experiments were car- ried out at the highest temperature assayed. Un- der the severest conditions considered (130ºC, 10% NaOH), the model predicted cellulose con- tents as high as 86%, confirming the suitability of the alkaline processing for delignification. The experimental results achieved for the hemicellulose content of samples (measured by y2) varied within a narrow range (2.9–6.3%). The surface response shown in Figure 3 (calculated for assays lasting 35 min) shows a continuous decrease in y2 with temperature for experiments with 3% NaOH. However, the model predicted a slight increase in the hemicellulose content of samples with temperature, which can be justi- fied on the basis of both experimental and fit- ting errors. Operating at 130ºC, the hemicellu- lose content of samples remained almost con- stant with the NaOH concentration, confirming that the hemicellulose solubilization was com- pleted at high temperature even in media con- taining the lowest NaOH concentration tested. Figure 4 shows the predicted dependence of variable y3 (lignin content of samples) on the most influential operational variables (T and [NaOH]) for treatments lasting 35 min. As ex- Table 3. Regression coefficients, significance level and statistical parameters (r2 and F) measuring the correlation an signif- icance of the models. a) Regression coefficients and significance Coefficients y 1 y 2 y 3 y 4 b 0 74.93* 3.033* 18.37* 99.63* b 1 6.84* –0.663* –5.49* 5.66* b 11 2.00* 0.758* –1.76* –8.06* b 2 1.46* –0.213* –1.64* 1.50* b 22 –1.03* 0.208** 0.74 –3.08* b 3 4.10* –0.725* –2.78* 0.42 b 33 –3.11* 0.133 3.32* –3.53* b 12 1.22* 0.400* –1.95* –2.76* b 13 0.78* 0.725* –1.23** –2.14* b 23 3.23* 0.475* –2.33* –2.76* ** Significant coefficients at the 95 % confidence level ** Significant coefficients at the 90 % confidence level b) Statistical parameters measuring the correlation and significance of models Variable R2 Corrected R2 F exp Significance level (based on the F test) y 1 0.9942 0.9839 90.99 >99% y 2 0.8779 0.6582 47.81 >99% y 3 0.9753 0.9310 05.25 >95% y 4 0.9414 0.8359 20.29 >99% Fig. 2. Dependence of the cellulose content of samples (var- iable y 1 ) on NaOH concentration and temperature predict- ed for samples delignified for 35 min. 57 A G R I C U L T U R A L A N D F O O D S C I E N C E I N F I N L A N D Vol. 11 (2002): 51–58. Fig. 3. Dependence of the hemicellulose content of sam- ples (variable y 2 ) on NaOH concentration and temperature predicted for samples delignified for 35 min. Fig. 4. Dependence of the lignin content of samples (varia- ble y 3 ) on NaOH concentration and temperature predicted for samples delignified for 35 min. pected, the response surface predicted lower lignin contents for higher temperatures at a giv- en NaOH concentration. Comparatively, the ef- fects caused by the NaOH concentration on the lignin content were of minor importance. The most important variable for the objec- tives of this work was the cellulose conversion achieved in the enzymatic hydrolysis step (y4). The experimental data listed in Table 2 show that all the alkali-treated samples were highly sus- ceptible toward the enzymatic hydrolysis, with 77.6–100% cellulose conversion into glucose. Cellulose conversions below 86% were obtained only in experiments 5 and 9, which correspond- ed to assays performed at the lowest tempera- ture considered (50ºC) with either short treat- ments (x2 = –1 in experiment 5) or low NaOH concentrations (x3 = –1 in experiment 9). When low-temperature delignification was combined with either prolonged reaction times (such as in experiment 6) or high NaOH concentrations (such in experiment 10), the cellulose conver- sion increased significantly (y4 up to 86.5– 86.8%). Harsher delignification conditions re- sulted in increased susceptibility toward enzy- matic hydrolysis. The experimental results of Table 2 show that enzymatic hydrolysis yields higher than 93% can be obtained under a variety Fig. 5. Dependence of the cellulose conversion into glu- cose (variable y 4 ) on NaOH concentration and temperature predicted for samples delignified for 35 min. of operational conditions. Figure 5, which shows the predicted dependence of the cellulose con- version on the operational variables T and [NaOH], confirms the above findings: the major effects on y4 are caused by temperature, particu- larly in the range 50–100ºC. The decreased cel- lulose conversions predicted at high tempera- tures and alkali concentrations can be justified on the basis of both experimental and fitting er- rors. For treatments lasting 35 min, 94–100% 58 A G R I C U L T U R A L A N D F O O D S C I E N C E I N F I N L A N D Moldes, A.B. et al. A process for the chemical-biotechnological utilization of barley husks cellulose conversion is predicted for samples delignified at temperatures above 80ºC even in treatments carried out with the minimum NaOH charge considered. In conclusion, the sequential treatment of barley husk with sulfuric acid and sodium hy- droxide allows a simultaneous benefit of the hemicelluloses and cellulose fractions to produce xylitol and glucose solutions respectively. The prehydrolyzed barley husk, after being subject- ed to an alkaline treatment, shows high cellu- lose content (up to 86%) and shows an excellent susceptibility toward enzymatic hydrolysis, with near quantitative glucose yields. Acknowledgements. Authors are grateful to “Xunta de Gali- cia” (Project XUGA 38303A98) and to the University of Vigo (Project 64502 K904) for the financial support of this work, as well as to Ms. Aida Ramos Nespereira and Anto- nia Rodríguez Jardón for their excellent technical assist- ance. References Alves, L.A., Felipe, M.G.A., Silva, J.B.A.E., Silva, S.S. & Prata, A.M.R. 1998. Pretreatment of sugarcane ba- gasse hemicellulose hydrolysate for xylitol produc- tion by Candida guilliermondii. Applied Biochemistry and Biotechnology 70–72: 89–98. Box, G.E.P., Hunter, W.G. & Hunter, J.S. 1978. Statistic for experimenters: an introduction to design, data analysis and model building. John Wiley, New York. p. 125–175. Cruz, J.M., Domínguez, J.M., Domínguez, H. & Parajó, J.C. 2000a. Preparation of fermentation media from agricultural wastes and their bioconversion into xyli- tol. Food Biotechnology 14, 1&2: 79–97. Cruz, J.M., Domínguez, J.M., Domínguez, H. & Parajó, J.C. 2000b. Xylitol production from barley bran hy- drolysates by continuous fermentation with De- baryomyces hansenii. Biotechnology Letters 22: 1895–1898. Gharpuray, M.M., Fan, L.T. & Lee, Y.H. 1983. Caustic pretreatment study for enzymatic hydrolysis of wheat straw. Wood and Agricultural Residues 1: 369–389. Mandels, M., Andreotti, R. & Roche, C. 1976. Measure- ment of saccharifying cellulose. Biotechnology Bio- engineering Symposium 6: 21–33. Mayerhoff, Z.D.V.L., Roberto, I.C. & Silva, S.S. 1998. Production of xylitol by Candida mogii from rice straw hydrolysate. Study of environmental effects using statistical design. Applied Biochemistry and Biotech- nology 70–72: 149–159. Moldes, A.B., Alonso, J.L. & Parajó, J.C. 2000. Multi-step feeding systems for lactic acid production by simul- taneous saccharification and fermentation of proc- essed wood. Bioprocess Engineering 22: 175–180. Paquot, M. & Thonart, Ph. 1982. Hydrolyse enzymatique de la cellulose regenerée. Holzforschung 36: 177– 181. Parajó, J.C., Alonso, J.L. & Santos, V. 1995. Enzymatic hydrolysis of wood: an engineering assessment. Bi- oprocess Engineering 12: 253–261. Parajó, J.C., Domínguez, H. & Domínguez, J.M. 1997. Improved xylitol production with Debaryomyces hansenii Y-7426 from raw or detoxified wood hydro- lyzates. Enzyme Microbiology Technology 21: 18–24. Roberto, I.C., Sato, S., Mancilha, I.M. & Taqueda, M.E.S. 1995. Influence of media composition on xylitol fer- mentation by Candida guilliermondii using response surface methodology. Biotechnology Letters 17, 11: 1223–1228. Silva, C.J.S.M. & Roberto, I.C. 1999. Statistical screen- ing method for selection of important variables on xylitol biosynthesis from rice straw hydrolysate by Candida guilliermondii FTI 20037. Biotechnology Techniques 13: 743–747. Vázquez, D., Lage, M.A., Parajó, J.C. & Vázquez, G. 1991. Transformación de materiales lignocelulósicos: Composición, fraccionamiento y aprovechamiento. Revista Agroquimica Tecnología de Alimentos 31, 2: 143–164. Title Introduction Material and methods Results and discussion References