Capillary microelectrodes for dissolved oxygen can discriminate between fluid elements with different oxygen concentration.' Near the interface between air and a solution to which oxygen is being transferred, the microelectrode signal is erratic as either enriched elements from the surface or oxygen-poor elements from the main bulk of the solution are encountered. When the interface is contaminated by surface active substances. circulation of fluid elements is retarded, and exchange between elements results in a much less erratic microelectrode signal.
Tsao and Lee' made traverses near the interface between air and an aqueous surfactant solution. The microelectrode signal had less amplitude with surfactant than with water alone close to the interface, but the amplitude increased to about that for water after a certain depth was reached. The zone where the signal was attenuated was interpreted as a fairly rigid "skin" formed by a high concentration of surfactant molecules attracted to the interface. We have used Tsao and Lee's procedure with suspensions of bacterial cells to ascertain the properties of the skin at the interface.
The bacterium Pseudornonus owlis was chosen because it oxidizes glucose to gluconic acid, and titration of the acid in the bulk solution provides a convenient index of the amount of oxygen transferred. Strain ATCC 8209 was cultured on a medium of glucose, urea, yeast extract, and inorganic compounds. After 24 hr. cells were collected by centrifuging, washed, and added to sterile nitrogen-free medium in a vessel with a magnetic stirrer. The vessel was filled until the liquid rose into a chimney with an area of I .53 cm2. With a silver-silver chloride reference electrode contacting the medium, an oxygen microprobe with a I-pm tip was lowered through the interface in increments while recording the electrode current. An unbuffered medium was used in runs where the overall oxygen transfer rate was determined from the rate at which base was added to hold the pH at 7.0. Buffered medium was used in runs in which attention was paid only to signal fluctuations during traverses with the microprobe. The apparatus was much the same as reported previously.'.' Most of the electrodes were obtained from Dr. W. H. Whalen ofSt. Vincent Charity Hospital, Cleveland, OH, but Dr. Young Lee of Drexel University donated one of his electrodes for comparison.
Results as tracings of strip chart recordings are shown in Figures and. The scale for the abscissa is irregular because the microelectrode was lowered quickly to the indicated depth, held for a brief period to get a typical oscillating signal, and advanced quickly to the next indicated depth. Dissolved oxygen concentration is approximately linear between the two calibration points shown on the ordinate. Bacterial populations less than about 10h/ml gave recordings that were similar to Figure for which no organisms were present. The dissolved oxygen concentration shows a sigmoidal drop from the interface to a fairly constant concentration representative of the main bulk of solution. The main difference for Figure where 3.49 x 10' cellsiml were present is a shift of the sigmoidal curve to the right. The shift corresponded to a depth of about 40 p n which was taken as the "skin" thickness. In other runs, the amplitude of oscillations of the electrode signal was attenuated in the skin areas lending confirmation to the estimates of skin thickness. The initial oxygen concentration was relatively constant indicating the stagnant nature of the film resulting from cells congregating at the interface. Diffusion of glucose is slow into the stagnant region, thus oxygen demand is low. Thickness of the skin was less for lower concentrations.
The enhancement factor. defined as the ratio of absorption rate with cells to that without cells under identical conditions, was less than one (Table ) for several different cell con-