Abstract:Frequency domain (FD) fluorescence lifetime data was collected for a series of 20 crude petroleum oils using a 405 nm excitation source and over a spectral range of ~426 to ~650 nm. Average fluorescence lifetimes were calculated using three different models: discrete multi-exponential, Gaussian distribution, and Lorentzian distribution. Fitting the data to extract accurate average lifetimes using the various models proved easier and less time consuming for the FD data than with Time Correlated Single Photon Counting (TCSPC) methods however the analysis of confidence intervals to the computed average lifetimes proved cumbersome for both methods. The uncertainty in the average lifetime was generally larger for the discrete lifetime multi-exponential model when compared to the distributionbased models. For the lifetime distributions, the data from the light crude oils with long lifetimes generally fit to a single decay term. Heavier oils with shorter lifetimes required This is the author version of the paper: Frequency Domain Fluorescence Lifetime Study of Crude Petroleum Oils. P. Owens, A.G. Ryder, and N.J.F. Blamey. Journal of Fluorescence, 18(5), 997-1006Fluorescence, 18(5), 997- , (2008. DOI: 10.1007/s10895-008-0330-5Page 2 of 21 multiple decay terms. The actual value for the average lifetime is more dependant on the exact fitting model employed than the data acquisition method used. Correlations between average fluorescence lifetimes and physical and chemical parameters of the crude oils were made with a view to developing a quantitative model for predicting the gross chemical composition of crude oils. It was found that there was no significant benefit gained by using FD over TCSPC other than more rapid data analysis in the FD case. For the FD data the Gaussian distribution model for fluorescence lifetime gave the best correlations with chemical composition allowing a qualitative correlation to some bulk oil parameters.