A water-in-oil (W/O) cultivation technology has the potential of overcoming the problems related with high broth viscosity in xanthan fermentations. The aqueous broth is dispersed in a continuous oil phase. Consequently, the broth thickening mechanisms are con®ned within the aqueous droplets without signi®cantly increasing the overall viscosity. To better characterize the mixing and oxygen transfer in the complex multiple-phase (G-O-W) systems involved, the W/O dispersions of xanthan solutions in either n-hexadecane or vegetable oil were examined in this study. The experiments with n-hexadecane indicated that the coef®cient for oxygen transfer from gas bubbles to the oil, i.e., k L a gao , was much smaller than that for transfer from the bulk oil phase to the droplet surface, i.e., k L a oaw . The oxygen partial pressure at the surface of aqueous droplets, p R , was therefore close to that in the bulk oil phase.The experiments with vegetable oil were conducted under various combinations of operating conditions: agitation speed (N) ± 400, 600, and 775 rpm; aeration rate (GaV) ± 0.25, 0.5, and 0.875 vvm; aqueous-phase volume fraction (/ w ) ± 0.2, 0.3, 0.4 and 0.5; and aqueous-phase xanthan concentration (Xn) ± 10, 20, and 40 kg/m 3 . The correlations developed for the power input of agitation (P g , in W), droplet diameter (d p , in lm), and k L a gao H o (in kg mol/m 3 s atm) are:where H o is the Henry's law solubility for oxygen in the oil phase (kg mol/m 3 atm) and v s is the super®cial gas velocity. The dependencies associated with P g , d p , and N are consistent with those reported in the literature for simpler systems although no previous correlations exist for complex G-L-L systems. The dependencies associated with Xn are intuitively plausible while the responsible mechanisms for the observed dependencies on / w are less clear. List of symbolsA integration constant in Eq. (5) a speci®c interfacial area, m À1 C i empirical constant: C 1 , C 2 and C 3 in Eqs. (11)±(13), C 5 in Eq. (24), and C 6 in Eq. (26) are dimensional while C 4 in Eq. (17) is dimensionless D i impeller diameter, m D w oxygen diffusion coef®cient in the aqueous phase, m 2 /s d p droplet diameter, lm g gravitational constant, m 2 /s G volumetric aeration rate, m 3 /s H Henry's law solubility constant for oxygen, de®ned as the ratio of dissolved concentration to partial pressure, kg mol/(m 3 atm) k L a volumetric oxygen transfer coef®cient, s À1 k s rate constant for the second-order reaction of oxygen consumption in Eq. (3), kg mol/(m 3 s atm 2 ) N impeller agitation speed, rpm N CD minimum agitation speed required for complete dispersion of gas bubbles (N CDYG ) or liquid dispersed phase (N CDYL ), rpm N JS minimum agitation speed at which solid particles or liquid droplets are just fully suspended, rpm N JSG minimum agitation speed at which particles are fully suspended under gassed conditions, rpm P power input for agitation, W p oxygen partial pressure, atm q O speci®c oxygen uptake rate of microorganisms, kg mol/[(kg dry cells) s] q p...

No abstract

Water-in-oil (W/O) fermentation technology has the potential for overcoming the problems related with high broth viscosity in xanthan fermentations. By dispersing the aqueous broth in a continuous organic phase, the broth-thickening mechanisms are confined within the aqueous droplets without significantly increasing the overall viscosity. In this study, xanthan fermentations were made with perfluorocarbon (PFC) or vegetable oil as the organic phase. The results were compared with those obtained previously using n-hexadecane as the organic phase, to evaluate the effects of various properties. PFC provided easy phase separation at the end of fermentation but required higher power input for agitation, a direct result of its high density. The aqueous droplets formed were large (400-450 microm), limiting the cell concentration employable due to the occurrence of oxygen starvation in the inner core. One main advantage of using vegetable oil was its low cost. In addition, vegetable oil provided much finer droplets (<120 microm) and produced high xanthan concentrations (>100 g l(-1)). However, complete phase separation for product recovery was difficult to achieve. Fermentations in both organic phases were terminated by the occurrence of phase inversion to highly viscous O/W dispersions at aqueous-phase volume fractions of 0.53-0.56. The initial fraction was 0.3 but changed due to base addition for pH adjustment and nutrient addition for prolonged production.