Food Hydrocolloids Vol. 10no. 1pp.75-82, 1996 Isolation and partial characterization of cellulose produced by Acetobacter xylinum* Milda E.Embuscado, James N.BeMiller and Jay S.Marks Department of Food Science, Purdue University, West LaFayette, IN 47907-1160, USA Abstract A mechanical separation method and an alkali treatment for the isolation of bacterial cellulose were developed. Response surface methodology was employed to determine the optimum conditions for the isolation of bacterial cellulose. A homogenizing mill was used to shear and separate the cellulose fiber from the bacterial cells. Cells were then removed by centrifugation and washing with distilled water. The optimum conditions obtained for the mechanical separation method were 1 :1.9 gel to water ratio, 15 min of homogenization and 5 centrifugation-washing cycles. The isolated cellulose had 0.46 % nitrogen and 1.46% ash (dry weight basis). The whiteness index (WI) of the product was 30.3 as measured by a Hunterlab colorimeter. For the alkali extraction method, the optimum conditions for the isolation of bacterial cellulose were 0.30 molldm' NaOH (-1.2%) at 900C for 25 min. The product had 0.08 % nitrogen, 0.66% ash and a WI of 44.2. The isolated bacterial cellulose had 3-5 times higher water holding capacity than commercial cellulose powders as determined by the centrifugation method. Any form of drying reduced its water holding capacity, but it was higher than that of commercial celluloses. Introduction One of the attractive features of bacterial cellulose is the absence of polysaccharide impurities and waxes within its network. Wood pulp and cotton linters, the traditional sources of cellulose, are difficult to purify because the polysaccharide impurities are intimately associated with the cellulose. Chemical extraction at high temperatures must be employed to remove these impurities. This is not the case with bacterial cellulose which contains bacterial cells, organic acids, salts, residual sugars and metabolites as impurities. Because these impurities are not closely associ- ated with the cellulose, they can be removed easily by washing, filtration, mild chemical extraction methods or a combination of these treatments. Various isolation methods have been used to separate bacterial cellulose from other components (1-4). The usual method of isolating bacterial cellulose is by hot caustic treatment followed by a series of washing and dewatering steps (4-6), which gives a product containing 0.2-0.4 % nitrogen and 0.3% ash (4,6) . In the studies reviewed the presence of other components and other important prop- erties of bacterial cellulose were not determined. • Journ al Paper no. 14.541 of Purdue Agricultural Experiment Station. © Oxford University Press The isolation and purification methods cited above employed different concentrations of alkali (1-8%), tem- peratures and times of treatment. It appeared that simple isolation methods for bacterial cellulose could be formu- lated based on the solubility of the bacterial cell proteins, its main contaminant, and that optimum purification conditions could be established using optimization tech- niques such as response surface methodology (RSM). An optimized method would not only standardize the purifi- cation process, but also yield a cost-effective method. RSM determines the optimum levels of process variables by combining special experimental designs with modeling using first and second order polynomial equations (7). The objective of this study was to determine the optimum conditions for isolating bacterial cellulose through response surface methodology and to establish simple and effective method(s) of isolating bacterial cellulose. In the process the isolated product was also partially characterized. Material and methods Materials Bacterial cellulose 'gels' were produced as described in Embuscado et al. (7) using a medium containing 24.8 g