Optimization of the production of aroma compounds by Kluyveromyces marxianus in solid-state fermentation using factorial design and response surface methodology

Studies were carried out for the production of aroma compounds in solid-sta te fermentation using factorial design and response surface methodology (RS M) experiments. Five agro-industrial residues were evaluated as substrate f or cultivating a strain of Kluyveromyces marxianus. The results proved the feasibility of using cassava bagasse and giant palm bran (Opunitia ficus in dica) as substrates to produce fruity aroma compounds by the yeast culture. In order to test the influence of the process parameters on the culture to produce volatile compounds, two statistical experimental designs were perf ormed. The parameters studied were initial substrate pH, addition of glucos e, cultivation temperature, initial substrate moisture and inoculum size. U sing a 2(5) factorial design, addition of glucose and initial pH of the sub strate was found statistically significant for aroma compounds production o n palm bran. Although this experimental design showed that addition of gluc ose did not have a significant role with cassava bagasse, 2(2) factorial de sign revealed that glucose addition was significant at higher concentration s. Head-space analysis of the culture by gas chromatography showed the prod uction of nine and eleven compounds from palm bran and cassava bagasse, res pectively, which included alcohols, esters and aldehyde. In both the cases, two compounds remained unidentified and ethyl acetate, ethanol and acetald ehyde were the major compounds produced. Esters produced were responsible f or the fruity aroma in both the cases. With palm bran, ethanol was the comp ound produced in highest concentration, and with cassava bagasse (both supp lemented with 10% glucose), ethyl acetate was produced at highest concentra tion, accumulating 418 and 1395 mu mol l(-1) head-space g(-1) substrate in 72 h, respectively. (C) 2000 Elsevier Science S.A. All rights reserved.