The gas transfer process across the air–water interface in a turbulent flow environment, with the turbulence generated in the water phase far away from the surface, was experimentally investigated for varying turbulent Reynolds numbers ReT ranging between 260 and 780. The experiments were performed in a grid-stirred tank using a combined particle image velocimetry – laser induced fluorescence (PIV-LIF) technique, which enables synoptic measurements of two-dimensional velocity and dissolved gas concentration fields. The visualization of the velocity and concentration fields provided direct insight into the gas transfer mechanisms. The high data resolution allowed detailed quantification of the gas concentration distribution (i.e. mean and turbulent fluctuation characteristics) within the thin aqueous boundary layer as well as revealing the near-surface hydrodynamics. The normalized concentration profiles show that as ReT increases, the rate of concentration decay into the bulk becomes slower. Independent benchmark data for the transfer velocity KL were obtained and their normalized values (KLSc0.5/uHT) depend on ReT with exponent −0.25. The spectra of the covariance term c′ w′ indicate that the contribution of c′ w′ is larger in the lower-frequency region for cases with small ReT, whereas for the other cases with higher ReT, the contribution of c′ w′ appears to be larger in the higher-frequency region (small eddies). These interrelated facts suggest that the gas transfer process is controlled by a spectrum of different eddy sizes and the gas transfer at different turbulence levels can be associated with certain dominant eddy sizes. The normalized mean turbulent flux $\overline{c^\prime w^\prime}$ profiles increase from around 0 at the interface to about 1 within a depth of approximately 2δe, where δe is the thickness of the gas boundary layer. The measured turbulent flux (c′ w′) is of the same order as the total flux (j), which shows that the contribution of c′ w′ to the total flux is significant.

To elucidate the physical mechanisms that play a role in the interfacial transfer of atmospheric gases into water, a series of direct numerical simulations of mass transfer across the air-water interface driven by isotropic turbulence diffusing from below has been carried out for various turbulent Reynolds numbers (R T = 84, 195, 507). To allow a direct (unbiased) comparison of the instantaneous effects of scalar diffusivity, in each of the DNS up to six scalar advection-diffusion equations with different Schmidt numbers were solved simultaneously. As far as the authors are aware this is the first simulation that is capable to accurately resolve the realistic Schmidt number, Sc = 500, that is typical for the transport of atmospheric gases such as oxygen in water. For the range of turbulent Reynolds numbers and Schmidt numbers considered, the normalized transfer velocity K L was found to scale with R − 1 /2 T and Sc − 1 /2 , which indicates that the largest eddies present in the isotropic turbulent flow introduced at the bottom of the computational domain tend to determine the mass transfer. The K L results were also found to be in good agreement with the surface divergence model of McCready, Vassiliadou & Hanratty (AIChE J., vol. 32, 1986, pp. 1108-1115 when using a constant of proportionality of 0.525. Although close to the surface large eddies are responsible for the bulk of the gas transfer, it was also observed that for higher R T the influence of smaller eddies becomes more important.

No abstract

in Wiley InterScience (www.interscience.wiley.com).The aim of this study is to quantify the mass transfer velocity using turbulence parameters from simultaneous measurements of oxygen concentration fields and velocity fields. The surface divergence model was considered in more detail, using data obtained for the lower range of b (surface divergence). It is shown that the existing models that use the divergence concept furnish good predictions for the transfer velocity also for low values of b, in the range of this study. Additionally, traditional conceptual models, such as the film model, the penetration-renewal model, and the large eddy model, were tested using the simultaneous information of concentration and velocity fields. It is shown that the film and the surface divergence models predicted the mass transfer velocity for all the range of the equipment Reynolds number used here. The velocity measurements showed viscosity effects close to the surface, which indicates that the surface was contaminated with some surfactant. Considering the results, this contamination can be considered slight for the mass transfer predictions. V