Submitted to: Journal of American Chemists Society
Publication Type: Peer reviewed journal
Publication Acceptance Date: 9/20/1998
Publication Date: N/A
Citation: Interpretive Summary: Concerns over the landfill burdens of plastics, as well as legislation to eliminate the dumping of plastics at sea (MARPOL Treaty), have increased awareness of the need for biodegradable plastics. Starch- based plastics use an inexpensive and renewable resource, yet such materials can be extremely brittle unless other materials, called plasticizers, are included to increase flexibility. Amino acids were tested and shown to be effective plasticizers with the added benefit of providing nitrogen required for biodegradation. From 20 tested amino acids, a predictive computer program was developed which will provide the basis for evaluating expensive compounds and designing new compounds as starch plasticizers. These plasticizers will produce nontoxic, environmentally friendly blends for single-use, disposable or compostable starch-based plastics. The research benefits the fledgling bioplastics industry, as well as the general American populous, by potentially replacing imported petroleum-derived plastic materials with renewable, home-grown commodities.
Technical Abstract: Twenty natural and synthetic amino acids (5 cyclic and 15 acyclic) were blended with a standard starch-glycerol mixture and extruded as ribbons. Glycerol was present in all blends as a co-plasticizer, permitting observation of both increase and decrease in sample flexibility resulting from amino acids. Mechanical testing of the ribbons revealed that amino acids had a dramatic effect on the percent elongation at break (%E) which varied from 13 to 379%. Tensile strengths of the ribbons also varied considerably from 0.96 to 6.29 MPa. In general, samples displaying the greatest elongation had the lowest tensile strength. FT-Raman results indicated that the amino acids in these blends existed predominately as zwitterions. Computational models of all test compounds were therefore generated as zwitterions, and the global minimum-energy conformation of each test compound was used as the basis for calculating molecular descriptors. Unexpectedly, only 2 descriptors (sum of absolute values of atomic charges and maximum positive charge on the molecule) of the 17 descriptors evaluated were needed to generate predictive quantitative structure-property relationships (QSPR) for both percent elongation and tensile strength data sets. By calculating these two descriptors from computer models, percent elongation and tensile strength can be predicted for blends with unknown or expensive compounds.