JOURNAL OF MATERIALS SCIENCE 29 (1994) 844-850 Modelling the behaviour of gas bubbles in an epoxy resin: evaluating the input parameters for a diffusion model using a solubility parameter approach J. R. WOOD*, M. G. BADER Department of Materials Science and Engineering, University of Surrey, Guildford, Surrey, GU2 5XH, UK Models based on mass diffusion theory successfully represent the growth and collapse of gas bubbles in an epoxy resin. Solution of the steady-state diffusion equations requires measurement of the diffusion coefficient and solubility of the mobile species within the resin pre-cursor. These parameters are affected by changes in temperature and/or pressure and are generally not measured as part of a processing schedule. Models have been evaluated that predict the prerequisite driving force'..in terms of a concentration gradient and the interaction with the processing variables from the chemistry of the resin molecule. A solubility parameter approach has been used to estimate the solubility of gas in the resin in conjunction with regular solution theory. The surface tension forces, which also play an active role in bubble stability and dynamics, have been estimated from molar attraction constants. 1. Introduction The objective behind the processing of thermosetting matrix composite materials is to produce void-free laminates of the specified dimensions with the opti- mum degree of cure of the resin. Porosity can occur due to inappropriate changes in the fabrication pro- cedure and will have a detrimental effect on the mech- anical properties of the final part [1]. In order to process high-performance structural laminates consis- tently, it is necessary to identify the material and processing inter-relationships which are critical to the production of void-free laminates and to formulate models that can be used to optimize the processing variables. Bubbles present within the resin will grow or col- lapse according to the temperature and hydrostatic pressure in the resin during the cure cycle. Bubble behaviour is influenced by such changes in accordance with the ideal gas law equations and due to the diffu- sion of mobile species across the bubble/resin inter- face, i.e. gaseous species in the bubble diffuse into the resin or dissolved gases in the resin diffuse into the bubble. An increase in temperature or decrease in pressure will cause the gas within the bubble to ex- pand resulting in bubble growth. Changes in pressure and temperature can also have a pronounced effect on the solubility of mobile species in the resin which may influence not only the magnitude of the driving force for diffusion but also the direction of the concentra- tion gradient. As an increase in temperature also causes an increase in the diffusion coefficient, the rate of mass transfer of molecularly mobile species will also increase, resulting in more rapid bubble growth or collapse depending on the direction of the diffusion gradient relative to the host bubble. [2]. Models based on mass diffusion theory have been investigated [-2, 3] and it is considered that entrapped bubbles can be collapsed or suppressed from growing by manipulating the process variables during the cur- ing operation, i.e. it is possible to influence bubble behaviour by changing the processing temperature and pressure. To implement these models successfully, a number of input parameters and their respective interactions with the process variables need to be evaluated before it is possible to predict the growth or collapse rate of gas bubbles in the resin pre-cursor. The material parameters include the diffusion coeffi- cient and the solubility of the mobile species in the liquid resin, both of which are functions of temper- ature and/or pressure. As a model is only as successful as the accuracy and availability of the input data, it is beneficial to have relationships that can predict these input parameters and/or their interaction with the changing process variables from fundamental mater- ials data or from more accessible physical quantities. The rationale behind the investigation is that the rela- tionships and the input parameters required for a void model can be translated to other systems with min- imal materials characterization and to different com- posite processing options. By minimizing empiricism it is also easier to confront the problem that there is no universal cure cycle but that it is necessary to deter- *Present address: Casali Institute of Applied Chemistry,The Hebrew University,Jerusalem, Israel. 844 0022-2461 1994 Chapman & Hall