(358i) Experimental Study of Gas Liquid Mass Transfer Enhanced By Bottom Shear Turbulence

The study of mass transfer at the interface between a gas and a liquid is of great importance in many environmental phenomena and industrial processes. In the industry, the physics of mass transfer defines the efficiency of diphasic mixing and degassing in stirred tanks, bubble columns, or even more complex mixing tank geometries. From an environmental point of view, the same mechanisms also control the dissolution of atmospheric gases into the oceans, rivers and lakes, and have therefore a huge impact on climate and biodiversity.

Gas dissolution at flat interface is controlled by gas and liquid turbulence characteristics [1]. Three possible turbulence sources may be considered [2]: gaseous phase turbulence, similar to that of wind above oceans or lakes; liquid turbulence caused by bottom shear found in rivers and stirred cells, and convective turbulence created by temperature gradients between the free surface and the bottom of the fluid.

To implement reliable models of mass transfer into CFD codes, a better understanding of the mechanisms ruling the transfer, like the influence of interface’s shape or flow regime, is still needed.

The aim of the present work is to study experimentally the mass transfer of a low solubility gas at a flat gas-liquid interface, and how it is enhanced by bottom shear turbulence. The dissolution of carbon dioxide (CO₂) into water is first studied. The same experimental procedure will then be applied to the observation of CO₂ dissolution into dilute polymer solutions with viscoelastic, shear-thinning behavior.

Experimental setup

The experimental bench is a prismatic tank with transparent Plexiglas® walls and lid allowing flow visualization. Liquid phase turbulence is generated by an oscillating grid appartus similar to the one used by Hopfinger and Toly [3].

Carbon dioxide (99,99% purity) is introduced above the liquid at atmospheric pressure. The overall carbon dioxide mass transfer is monitored by a pH sensor placed at the bottom of the tank.

Two non-intrusive optical techniques are used simultaneously to measure both velocity and dissolved gas concentration fields at a good spatial resolution (typically 0.4 mm for the velocity field and up to 0.008 mm for the concentration field out of the Laser sheet thickness). The Region of Interest (ROI) is a vertical plane underneath the interface at the center of the tank.

Stereoscopic Particle Image Velocimetry (SPIV) allows for the measurement of the three components of the velocity field by stereoscopic reconstruction of particle images of two identical cameras looking at the same region of the flow. As the measurements are made inside the liquid phase, a Scheimpflung arrangement for the lenses and the use of a liquid prism are necessary [4].

An original Quenched Planar Laser Induced Fluorescence (Q-PLIF) technique is used for the determination of the dissolved gas concentration field [5]. A fluorescent dye (here Fluorescein disodium salt) is previously mixed into the liquid at low concentration (typically of 10^-7M). The fluorescence of this dye strongly depends on the pH, which is itself modified by the acid-base reactions of dissolved carbon dioxide in aqueous media. A rigorous calibration of the relations between carbon dioxide concentration and pH, and between pH and fluoresced intensity enables the transformation of fluorescence intensity fields into dissolved gas concentration fields, and consequently the measurement of local gas concentration.

External timing of the two techniques based on the grid’s oscillation, and a common spatial calibration make it possible to correlate spatially and temporally the two measured fields and to compute the mass fluxes in the ROI.

Results and discussions

From PIV measurements, it has been shown that for a given grid oscillation frequency, velocity magnitude and fluctuations decrease toward the free surface. Vertical distribution of turbulence can be related to the grid oscillation frequency f, the stroke S, the mesh parameter M (distance between two consecutive grid bars), and the distance to the interface, as shown by Hopfinger and Toly [3]. Turbulence was also found to become almost homogeneous with no periodic oscillation of the velocity field in an horizontal plane at a given distance from the grid’s average position [6].

When approaching the interface, the root mean square of velocity fluctuations profile is modified. The vertical component, which is the strongest fluctuation in the bulk fluid, becomes lower than the two horizontal components. All fluctuations decrease rapidly from a given depth to the interface, verifying at least from a statistical point of view, the concept of a viscous sublayer expressed in Hunt [7].

The dissolved-gas fields obtained by Q-PLIF show the existence of a thin subsurface layer in which the concentration rapidly decreases from the surface saturation concentration. Different types of scalar structures modifying the shape of this sublayer are observable.

- Renewal eventscorrespond to “fresh” low concentration fluid being brought to the interface by an upwelling eddy to replace saturated fluid consequently enhancing the diffusion and the mass transfer

- Injection events or “peeling” [8] which consist in high concentration filaments being brought down by swirling velocity structures down to the bulk fluid where it is mixed.

Using coupled measurements, the correspondence between scalar and velocity structures can easily be done. But since the velocity field under the interface tends to be highly three dimensional, 3D velocity structures appear and are also found to be correlated with some complex injection and renewal events [6].

Current work and perspectives

From the previously cited measurements, it is possible to compute the correlation between two velocity components fluctuations and between velocity and concentration fluctuations, i.e. the Reynolds shear stresses and the turbulent contribution to the overall mass flux. The conditioning of this fluxes using quadrant analysis [9] based on the vertical velocity fluctuation and the concentration, and a further extension to an octant analysis [10] including the horizontal velocity fluctuation should give a better insight about the events that dominate turbulent mass transfer.

The next step of the study consists in the addition of a small concentration of polymer to the liquid phase. It has been known for a long time that such additives tend to lower the flow friction at rigid boundaries and modify turbulence dissipation [11], [12]. It has also been observed that mixing of passive scalar [13] and gas liquid mass transfer [14] in such media differ from the Newtonian case, but to the author’s knowledge, no fundamental study has yet focused on the turbulent mass transfer events at a gas liquid interface for these complex fluids.

[1] D. E. Turney and S. Banerjee, “Air–water gas transfer and near-surface motions,” J. Fluid Mech., vol. 733, pp. 588–624, Oct. 2013.

[2] H. Herlina, “Gas transfer at the air-water interface in a turbulent flow environment,” Universitätsverlag Karlsruhe, Karlsruhe, 2005, PhD Thesis.

[3] E. J. Hopfinger and J.-A. Toly, “Spatially decaying turbulence and its relation to mixing across density interfaces,” J. Fluid Mech., vol. 78, no. 1, pp. 155–175, Nov. 1976.

[4] A. K. Prasad and K. Jensen, “Scheimpflug stereocamera for particle image velocimetry in liquid flows,” Appl. Opt., vol. 34, no. 30, p. 7092, Oct. 1995.

[5] P. Valiorgue, N. Souzy, M. EL Hajem, H. B. Hadid, and S. Simoëns, “Concentration measurement in the wake of a free rising bubble using planar Laser-induced fluorescence (PLIF) with a calibration taking into account fluorescence extinction variations,” Exp. Fluids, vol. 54, no. 4, pp. 1–10, Apr. 2013.

[6] T. Lacassagne, M. EL Hajem, F. Morge, S. Simoens, and J.-Y. Champagne, “Study of Gas Liquid Mass Transfer in a Grid Stirred Tank,” Oil Gas Sci. Technol. – Rev. D’IFP Energ. Nouv., vol. 72, no. 1, p. 7, Jan. 2017.

[7] J. C. R. Hunt, “Turbulence Structure and Turbulent Diffusion Near Gas-Liquid Interfaces,” in Gas Transfer at Water Surfaces, W. Brutsaert and G. H. Jirka, Eds. Springer Netherlands, 1984, pp. 67–82.

[8] null Herlina and G. H. Jirka, “Experiments on gas transfer at the air–water interface induced by oscillating grid turbulence,” J. Fluid Mech., vol. 594, pp. 183–208, Jan. 2008.

[9] E. A. Variano and E. A. Cowen, “Turbulent transport of a high-Schmidt-number scalar near an air–water interface,” J. Fluid Mech., vol. 731, pp. 259–287, Sep. 2013.

[10] J.-Y. Vinçont, S. Simoëns, M. Ayrault, and J. M. Wallace, “Passive scalar dispersion in a turbulent boundary layer from a line source at the wall and downstream of an obstacle,” J. Fluid Mech., vol. 424, pp. 127–167, Dec. 2000.

[11] K. R. Sreenivasan and C. M. White, “The onset of drag reduction by dilute polymer additives, and the maximum drag reduction asymptote,” J. Fluid Mech., vol. 409, pp. 149–164, Apr. 2000.

[12] M. Q. Nguyen, A. Delache, S. Simoëns, W. J. T. Bos, and M. EL Hajem, “Small scale dynamics of isotropic viscoelastic turbulence,” Phys. Rev. Fluids, vol. 1, no. 8, p. 83301, Dec. 2016.

[13] T. D. Nguyen, “Influences des propriétés non-Newtoniennes sur un mélange de scalaire passif,” Lyon, INSA, 2013, PhD Thesis.

[14] V. R. Ranade and J. J. Ulbrecht, “Influence of polymer additives on the gas-liquid mass transfer in stirred tanks,” AIChE J., vol. 24, no. 5, pp. 796–803, Sep. 1978.