Mechanism of bubble formation during the drying of polymer films

In creating thin films of polymers by solution processing, a common problem is the undesirable formation of bubbles during the drying process. Practical experience shows that bubbles can be created well below the boiling point of the solvent. Also, it has been observed that the degassing of the polymer solutions result in reduced bubble formation, indicating a relationship between the presence of air and bubble formation. This work is based on a hypothesis that if the solubility of air in the polymer solution increases with solvent concentration, then the solution can become super saturated with air as the concentration of the solvent is reduced during the drying process. To test this hypothesis we have chosen the system of polyvinyl acetate, toluene and nitrogen. Experimental methods were developed to measure the solubility of nitrogen in the polymer-solvent system as a function of solvent composition and temperature. The group-contribution lattice-fluid equation of state was used to correlate the thermodynamic behavior of the ternary system, utilizing experimental measurements to determine the interaction parameters in the equation of state. In addition, experiments were conducted to measure the diffusion coefficients of the nitrogen over a range of pressures and temperatures. Different diffusion models based on the friction coefficients and free volume model and were then used to correlate the diffusivity data so that the diffusional behavior of the ternary system can be predicted over a broad range of conditions. Finally, the thermodynamic and diffusivity correlations were incorporated into a multi-component drying model which included main and cross diffusion terms to predict saturation behavior in the polymer solution during the drying process. The model without the cross diffusion terms represents the ideal system in which the diffusion of one component does not affect the diffusion of others. The drying model did not predict supersaturation of nitrogen when cross diffusion terms were neglected. However, the model predicted supersaturation of nitrogen when the cross diffusion terms are included. Therefore, the cross diffusion terms in the mass transfer model are essential for the development of nitrogen supersaturation. Also different diffusion models based on the friction coefficients led to qualitatively similar predictions for the supersaturation of nitrogen. The simulation's results supported all our observations in experiments regarding bubble formation.