Characterization of diesel fuel by chemical separation combined with capillary gas chromatography (GC) isotope ratio mass spectrometry (IRMS) Scott D. Harvey n , Kristin H. Jarman, James J. Moran, Christina M. Sorensen, Bob W. Wright National Security Directorate, Pacific Northwest National Laboratory, Richland, WA 99352, United States article info Article history: Received 10 February 2012 Received in revised form 18 May 2012 Accepted 22 May 2012 Available online 29 May 2012 Keywords: Differentiation of diesel fuel samples Compound specific isotope analysis (CSIA) of diesel fuel n-alkanes Hydrogen isotope ratio (dD) analysis Carbon isotope ratio (d 13 C) analysis Multi-element isotope ratio analysis Principal components analysis (PCA) Nearest neighbor classification of PCA scores abstract The purpose of this study was to perform a preliminary investigation of compound-specific isotope analysis (CSIA) of diesel fuels to evaluate whether the technique could distinguish diesel samples from different sources/locations. The ability to differentiate or correlate diesel samples could be valuable for discovering fuel tax evasion schemes or for environmental forensic studies. Two urea adduction-based techniques were used to isolate the n-alkanes from the fuel. Both carbon isotope ratio (d 13 C) and hydrogen isotope ratio (dD) values for the n-alkanes were then determined by CSIA in each sample. The samples investigated had d 13 C values that ranged from 30.1% to 26.8%, whereas dD values ranged from 83% to 156%. Plots of dD versus d 13 C with sample n-alkane points connected in order of increasing carbon number gave well-separated clusters with characteristic shapes for each sample. Principal components analysis (PCA) with d 13 C, dD, or combined d 13 C and dD data was applied to extract the maximum information content. PCA scores plots could clearly differentiate the samples, thereby demonstrating the potential of this approach for distinguishing (e.g., fingerprinting) fuel samples using d 13 C and dD values. & 2012 Elsevier B.V. All rights reserved. 1. Introduction The ability to determine unique chemical characteristics of a complex sample is critical for many applications. For petroleum- based materials, unique features aid in sample correlation and origin tracking and represent important applications in the geochemical and oil exploration areas. Likewise, the ability to track and identify sources for refined petroleum products, such as diesel fuel, also has important applications. One application of uniquely identifying and tracking diesel fuel is to provide a scientific tool to discover fuel tax evasion. A number of possible schemes have been devised to evade payment of fuel tax. One dated source places the United States (U.S.) loss of tax revenue from fuel tax evasion at billions of dollars annually [1]. A number of potential U.S. fuel tax evasion schemes have been described [2]. For example, tax-free fuel reported as exported to a foreign destination may actually be sold domestically as taxed fuel with the evader keeping the tax revenue. Similarly, schemes that involve out-of-country sale with subsequent redelivery of the same fuel back to the U.S. as taxed fuel have been documented. Another scheme involves addition of non-taxed and/or lower value products (e.g., non-fuel petroleum products and waste streams such as used oils and solvents), products taxed at a lower rate (e.g., jet fuel), or unwanted petroleum products (e.g., high-sulfur diesel and refining inter- mediates) to the diesel fuel and selling the higher-volume blended or ‘‘cocktailed’’ mixture as taxed fuel, again with the evader keeping the tax revenue. Additional schemes include complex daisy chain paper trails that make it appear tax has been paid when in fact it has not. Numerous other fuel tax evasion schemes also exist, but are not described here [2]. Various chemical analysis techniques provide valuable tools for discovering fuel tax evasion. One important tool is high- resolution capillary gas chromatography which is used to create fuel fingerprints based on the compositional profile. These pro- files, along with multivariate statistical analysis, can successfully allow sample correlation and determination of similar or common upstream distribution sources [3–6]. While this method allows reliable conclusions to be made, it can be difficult to differentiate fuels with very similar chromatographic profiles. It can also be problematic to determine if minor compositional changes arise from variation due to fuel source or from blending with other components or mixtures. Consequently, additional tools to complement and extend high-resolution, gas-chromatographic profiling techniques would be highly desirable. For instance, being able to fingerprint specific compounds in the chromato- graphic profile would add significant capability to discriminate unique sample attributes. In other words, using an orthogonal technique to ‘‘fingerprint the fingerprint’’ would provide enhanced capability. Contents lists available at SciVerse ScienceDirect journal homepage: www.elsevier.com/locate/talanta Talanta 0039-9140/$ - see front matter & 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.talanta.2012.05.049 n Corresponding author. E-mail address: scott.harvey@pnnl.gov (S.D. Harvey). Talanta 99 (2012) 262–269