No abstract

SummaryDissolved oxygen measlirements were made in pilot (20 and 250 1,) arid production scale (15,000 1.) novobiociri fermentations. Bulk mixing was found to be incomplete in pilot tanks with turbine impellers of D / T = 0.40 (where D is impeller diameter, and T is tank diameter) but appeared homogeneous with impellers of D / T = 0.69. I n the former case, the respiration rate was presumably limited by insufficient oxygen supply in areas of poor bulk mixing, whereas, in the latter case, the major resistance was between the bulk of the liquid and the cell (intrachrmp resistance). Higher agitator speeds decreased the gas-liquid resistance proportionally more than they reduced the liquid-cell resistance. I n production fermentors, dissolved oxygen measurements indicated that bulk mixing was complete with each of three impeller sizes tested (D/T = 0.28, 0.33, and 0.43), but that the respiration rate was again limited, mainly by a resistance between the bulk of the liquid arid the cell.

SummaryStudies wert: made of oxygen transfer and power in 20-1. novobiocin fermentations with five sizes (23-60% of the tank diameter) of geometrically similar flatblade turbine impellers. The influence of changes in apparent viscosity on oxygen transfer and power was found to be related t o impeller diameter. At constant agitator speed the oxygen availability rate (OAR) decreased with increase in apparent viscosity, but the magnitude of the reduction was less with a small impeller than with a large one. At constant power input the small-diameter impeller (23% of tank diameter) gave a n OAR about 8 times that of the large one (60% of tank diameter) when tested in viscous beer, whereas with sulfite or yeast all impeller diameters gave the same OAR. Impeller tip velocity correlated with OAR independently of diameter for the five turbine sizes examined. The power required for optimal novobiocin yields increased with increase in turbine diameter, but the yields were independent of diameter at equal OAR.

SummaryThe performance of a multiple-rod mixing impeller was compared to that of conventional turbine impellers in viscous novobiocin beers. The advantages of the multiplerod impeller were found to be: ( I ) the power requirement was independent of changes in apparent viscosity of the fermentation beer; and ( 2 ) it gave the same novobiocin yield and oxygen-availability rate at about one-half of the power required by turbines.During the past few years, we have reported on the problem of oxygen transfer using the novobiocin fermentation as an example of a viscous non-Newtonian system.1-4 The general picture which emerged for turbine impellers under fully baffled conditions is this : In pilot-scale equipment, at least up to 250-1. batches, the oxygen requirement of the culture may be limited primarily by two main factors. The relative importance of these factors varies with the geometry of the system. First, with turbines of about 40% or less of the tank diameter, the oxygen requirement may be limited by poor bulk mixing. This situation is indicated by dissolved oxygen measurements that are low in one part of a fermentor and high in another part of a fermentor (see also ref. 5). Or, second, with largediameter turbines up to 69% of the tank diameter, where bulk mixing is apparently homogeneous as indicated by dissolved oxygen measurements, the respiration rate may be limited primarily by a resistance between the bulk of the liquid and the site of reaction in the cell. This type of behavior is indicated by uniform dissolved oxygen measurements throughout the fermentor, but with an abnormally high "apparent" critical dissolved-oxygen concentrat-ion, for example, values up to 50 or 60% of saturation. This is also the type of behavior we have observed in a 15,000-1. production fermentor.109

The. llpjolm C'ompniiy, Kalnmuzoo, .IIicliiyrcnSuninmy. Novobiocin fermentations in 20 1. baffled fermentors were studied as a function of the size and speed of the dual four-bladed flat-blade t.urbines used. Power input and sulphite oxidation rate measurements were made. The courses of pH, sugar utilization, mycelial dry weight,, a r b o n dioxide evolution and antibiotic titer were determined. Optimum antibiotic yields were achieved at a power input of 0.5 h.p./100 gal, equivalent t o a sulphite oxidation rate of 110 mmoles Oz/l.h, when the impeller diameters were 29 per cent or 39 per cent of the tank diameter. A power input of 0.75 h.p./100 gal, equivalent to a sulphite oxidation rate of 160 mmoles Oz/l.h, was required for equivalent results with the impeller diameter 49 per cent of the tank diameter. Some explanations of the lack of equivalent results with the large impeller are discussed.