(519f) Design of Biofuel Additives Using Chemometric Modeling and Molecular Design Techniques

Any type of feedstock that contains free fatty acids and/or triglycerides such as vegetable oil, waste oil, animal fat, and waste grease can be converted into biodiesel. Because of its biological origin, biodiesel has many advantages compared to its petroleum derived counterparts. However, it does exhibit several limitations like oxidative stability and low temperature operability that have impaired its wide use and commercialization. Simultaneous solution of these problems has proven difficult as improvements in one area tend to impair another. As a result, design of an optimal fuel for all environments can be a quite difficult task. In addition, feedstock origin will impact the final quality of the biodiesel product. Therefore, to produce fuel grade biodiesel, the characteristics of the feedstock are very important during the initial research and production stage.

Several approaches are possible for improving the fuel properties of biodiesel. Modification of the fatty acid composition through physical processes or uses of additives are the most prevalent. Fuel additives, such as antioxidants, cetane enhancers, or cold-flow improvers have become indispensable tools not only to alleviate the drawbacks described above, but also to assure that any fuel blend will meet international and regional standards regardless of origin. However, an additive solution to one problem often aggravates another problem. Furthermore, the questions of additive compatibility, required additive levels, the effect on other properties, and whether these additives function as designed for biodiesel fuels with differing fatty acid profiles still remain challenges and that need further investigation.

Therefore, it is desired to molecularly design biodiesel additives that simultaneously account for the unintended effect on other fuel properties in the neat and the blend fuel in order to achieve the performance properties of the petroleum based fuel. In this way, biofuels can be formulated that are adaptable to a range or blend of feedstocks and the desirable fuel characteristics like oxidative stability and wide operating temperature range.

In this work, we aim to identify all possible compounds which, when added to off spec biodiesel, results in a fuel that satisfies performance standards such as ASTM D6751 in the US and EN14214 in Europe. To meet this end, multivariate characterization data obtained from IR spectroscopy of common additives were combined with decomposition and property clustering techniques in a reverse problem formulation. In this approach, the fuel additive property targets are identified in the first reverse problem followed by molecular design to match the targets. The characterization data consists of multitude of properties of interest (such as cetane number, melting point, and kinematic viscosity) to ensure adequate performance. To facilitate an efficient design we consolidated these various properties into a latent property domain using principle component analysis (PCA) techniques. Finally, characterization based molecular design using group contribution parameters are then used to build novel additives that match the fuel specifications in the latent property space.