Fundamental Thermo-Mechanical Property Modeling of Triglyceride-Based Thermosetting Resins
John J. La Scala, Army Research Laboratory, Building 4600, Aberdeen Proving Ground, MD 21005 and Richard P. Wool, University of Delaware, Composites Manufacturing Science Laboratory, University of Delaware, Newark, DE 19716.
Triglyceride-acrylates were prepared from various plant oils and model compounds using the unsaturation sites on the triglycerides. Unlike petroleum-derived resins, acrylated plant oils have a distribution of functionality ranging from 0 through 9 acrylates per triglyceride. This distribution was calculated using the distribution of unsaturation sites found on plant oils and then using successive binomial relationships to determine the probability of functionalizing a given functionality for a measured extent of functionalization. The cross-link densities of the resulting polymers were calculated using Flory-Stockmayer theory accounting for the extent of intramolecular cyclization. The glass transition temperature was clearly a function of the level of acrylation of triglyceride-based polymers and was modeled using simple empirical relationships. The glass transition temperatures of n-acrylated triglycerides were used as the relaxation modes for plant oil-based polymers. Using these relaxation modes, the Twinkling Fractal Theory of Tg was then used in conjunction with the calculated triglyceride distribution in order to predict the storage modulus, tan(d), and loss modulus as a function of temperature. Essentially, the percentage of n-acrylated triglycerides with Tg less than that of the ambient temperature, the rubbery fraction, is the amount the modulus drops with temperature. The rubbery moduli for these polymers were calculated using the Theory of Rubber Elasticity and the calculated cross-link density. The tan(d) was predicted based on the percentage of rubbery segments that were formed as the temperature was incremented from one relaxation mode to the next. The product of the storage modulus and tan(d) is the loss modulus, and thus the loss modulus was also be predicted based on the polymer composition. There were some deviations between the experimental and predicted dynamic mechanical properties, but overall the model predictions matched the experimental results well. Nonetheless, this method offers an excellent method to approximate the dynamic mechanical properties as a function of temperature for a resin containing a distribution of n-functional cross-linkers.