USING RESPONSE-SURFACE METHODOLOGY TO EVALUATE XYLITOL PRODUCTION BY CANDIDA-GUILLIERMONDII BY FED-BATCH PROCESS WITH EXPONENTIAL FEEDING RATE

Xylitol production by Candida guilliermondii FTI 20037 from sugar cane bagasse hydrolyzate by fed-batch process using exponential feeding ra te was evaluated. A factorial design was employed to study the effects of the xylose concentration in the fermenter (S-0) and the xylose con centration in the feeding solution (S-i) upon the xylitol/xylose yield coefficient and the xylitol concentration. It was found that none of the variables had a significant effect on the xylose-to-xylitol conver sion factor. However, their effects on the xylitol concentration were significant. A response-surface methodology permitted us to obtain a p olynomial model for xylitol production (C-P) at 95% confidence level: C-P= 15.896 - 1410 S(0-)0.487 S-i - 0.032 S-0(2). The best experimenta l results achieved for xylitol production was 44 g l(-1), with a xylit ol yield coefficient of 0.78 g g(-1) and a volumetric productivity of 0.62 g l(-1) h(-1). (C) 1998 Elsevier Science B.V. All rights reserved .