Determination of peroxide value of edible oils by FTIR spectroscopy with the use of the spectral reconstitution technique.

Spectral reconstitution (SR), a technique that has been developed to facilitate mid-FTIR transmission analysis of inherently viscous samples, was applied to simplify and automate a previously reported FTIR method for the determination of peroxide value (PV) of edible oils. The basis of the PV determination is the rapid reaction of triphenylphosphine (TPP) with the hydroperoxides present in an oil to produce triphenylphosphine oxide (TPPO), which exhibits a readily measurable absorption band at 542 cm(-1). In the SR procedure, the viscosity of oil samples is reduced by mixing them with a diluent, which allows them to be readily loaded into a flow-through transmission cell. The spectra of the neat oil samples are then reconstituted from those of the diluted samples by using the absorption of a spectral marker present in the diluent to determine the dilution ratio. For the SR-based PV method, the TPP reagent was added to the diluent, which consisted of odorless mineral spirits (OMS) containing methylcyclopentadienyl manganese tricarbonyl (MMT) as the spectral marker. Sample preparation for PV analysis involved mixing approximately 10 ml of oil with approximately 25 ml of the TPP-containing diluent; accurate weighing or delivery of precise volumes was not required because the dilution ratio was determined spectroscopically from the intensity of the nu(CO) absorption of MMT at 1942 cm(-1) in the spectrum of the diluted sample relative to that in the spectrum of the diluent. Calibration standards, prepared by gravimetric addition of TPPO to a peroxide-free oil, were handled in the same manner, and a linear calibration equation relating the concentration of TPPO (expressed as the equivalent PV) to the absorbance of TPPO at 542 cm(-1) relative to a baseline at 530 cm(-1) in the reconstituted spectra was obtained, with a regression S.D. of +/-0.15 meq/kg oil. PV determinations on two sets of validation samples, spanning PV ranges of 0-20 and 0-2 meq/kg oil, were carried out in parallel by the AOCS titrimetric and SR-based FTIR procedures, and comparison of the results of duplicate analyses by the two methods indicated that the latter was more reproducible and slightly more sensitive. The SR-based PV method, when implemented on an autosampler-equipped FTIR system, allowed for the automated analysis of approximately 90 samples per hour.