A study of structural effects on oxidative stability of soybean oils

Environmentally friendly lubricants have recently gained interest in many lubrication applications. Vegetable oils, which are the main base stocks of these lubricants, have an inherit problem due to reduced oxidative stability. This thesis studies the structural effects on the oxidative stability and additive response of different soybean oils and TMP ester. The experimental techniques used in this study include the pressurized differential scanning carolimeter (PDSC), the Penn State Micro-oxidation test (PSMO), gel permeation chromatography (GPC), Fourier Transform Infrared Spectroscopy (FTIR), and a carbon analyzer. Epoxidized soybean oils (ESB) show higher oxidative stability than other soybean oils due to the absence of allylic hydrogens in the acid chain. Aminic antioxidant and copper naphthenate perform much better than other antioxidants in ESB in both PDSC and PSMO. ZDDP promotes the oxidation in ESB because of acid-catalyzed ring opening. Phenolic antioxidant shows higher induction time/temperature than other antioxidants in the PDSC while ZDDP shows lower deposit formation in PSMO for soybean oils with various oleic acid contents. The difference in results between PSMO and PDSC is due to the different stages of oxidation. The PDSC measures the inhibited stage while the PSMO measures the final stage of oxidation. Soybean oils with higher oleic acid content show higher oxidative stability and better antioxidant response. However, the high oleic content in those oils leads to more volatile oxidative products due to α-scission of larger fragments. A statistical model of volatility loss shows that the factors affecting the volatility loss at 175°C on LCS coupon include oxidation time, % oleic, % linolenic. % ZDDP, and % Cu while the factors affecting the amount of deposit formation at the same condition include oxidation time, % oleic, % linolenic, % aminic AO, % phenolic AO, % ZDDP, %Cu, and % volatility loss. An artificial neural network was able to precisely predict the results from PSMO.