Abstract
This thesis has identified a suite of hydrocarbon diagnostic indices to be applied in the forensic fingerprinting of diesel oil characterisation and source identification using compound-specific isotope analysis (CSIA). The diagnostic criteria for discrimination were applied in the differentiation of commercial diesel fuels from service stations to explore the discriminative power of the CSIA technique. Following that, the diagnostic indices were also used to generate isotopic data of the δ13C and δD compositions of the compounds from samples involved in a diesel theft case. The acquired data sets, when processed using multivariate statistical techniques, were presented as chemical evidence in a legal proceeding in a New Zealand court. In a parallel study, several key experiments were carried out to monitor and assess the physical and microbial degradation effects on the isotopic values of the compounds to provide an exemplary validation process. These method validations were considered necessary as a part of a strategy in a methodological framework when approaching a particular forensic investigation using CSIA technique.
This research commenced by refining and validating sample processing methods for extracting hydrocarbon compounds from a solid matrix prior to GC analysis. Rotary evaporation and Kuderna-Danish concentration system were found to introduce negligible isotopic fractionation during sample concentration. Soxhlet extraction and silica gel separation were used to extract target analytes from a solid matrix. Non-polar compounds, specifically n-alkanes were extracted with high efficiency using n-pentane and n-hexane with insignificant isotopic fractionation. Polycyclic aromatic hydrocarbons (PAHs) were extracted with good efficiency using mixtures of solvents, n-pentane:DCM (1:1) and n-hexane:DCM (1:1), also with negligible C and H isotope fractionation. However, when PAHs were extracted using n-hexane:acetonitrile (1:1) and n-hexane:ethyl acetate (1:1), low extraction efficiency was observed (less than 70 %). The same solvent mixtures introduced isotopic fractionation of up to –1.8‰ and –10‰ (for C and H, respectively). The validated sample processing methods were subsequently applied in the latter part of this thesis.
Forensic discrimination of refined petroleum oil using hydrocarbon diagnostic indices, n-alkanes, was carried out for commercial diesel fuels obtained from 45 service stations in various locations of the South Island of New Zealand. The isotopic signatures of the diesel oils varied significantly and samples from different locations were distinguishable indicating CSIA is potentially useful in source apportionment of petroleum-type hydrocarbons. Further, the discrimination of diesel fuel samples involved in a diesel theft case was also carried out and combined with the data sets from the commercial diesel fuels analysis which provided substantial sample population covering the widest range of sources, the results were reliably put into context by recognising the known or potential error rate of the technique.
Investigation of the effects of physical and microbial degradation on the δ13C and δD compositions of the hydrocarbon indices was performed to identify link(s) between isotopic shifts and compound degradation which could explain the discrimination observed among commercial diesel fuels and in the diesel theft case. This information also provided insight on the mechanism(s) that impart the changes in the isotopic compositions hence validating and fulfilling the criteria of evidence to be admissible in a court of law.
Carbon isotopic fractionation was absent during physical degradation of diesel n-alkanes. On the other hand, D depletion occurred in the shorter chain n-alkane residues during progressive evaporation, indicating an inverse trend in isotopic fractionation, where heavier isotopes of the element were preferentially removed. As for the longer chain n-alkanes (C18+), including pristane and phytane, the D values were stable over the course of the experiment. These findings suggest that kinetic modelling can be applied in monitoring natural attenuation for remediation work utilising shorter chain n-alkanes as organic tracers.
As for microbial degradation experiments, small shifts were observed in the δ13C values of the shorter chain n-alkanes after slight to moderate degradation (up to 0.6‰ for nC12 and nC13), but for the rest of the n-alkanes including the isoprenoids, the δ13C values remained conservative. On the contrary, the microbial degradation caused substantial D enrichment in the shorter chain n-alkanes in the residual oil with high values of enrichment (>20‰) after only 8 days of incubation, especially when the soil was biostimulated. The hydrogen isotopic values of the longer chain n-alkanes (C20+) and the isoprenoids remained unchanged during the course of the biodegradation experiments. The opposing trends observed in the δD isotopic shifts in the n-alkanes caused by physical degradation, where isotopic values were depleted, and microbial degradation, where isotopic values were enriched, generally followed the Rayleigh model where nC12 and nC13 usually showed strong correlation but the rest were weakly correlated. The weak correlations observed in the longer chain n-alkanes produced a curvature which deviated from the theoretical model. This deviation indicated that the microbial degradation occurring in the environment was a complex process where unquantifiable organic chemical transformations and isotopic exchanges occur.
The stability of the δ13C and δD compositions of the longer chain n-alkanes indicates huge potential for CSIA as a tool to correlate petroleum-type contaminants to their origin. Further, monitoring isotopic fractionation of two or more elements would be more effective and can yield stronger evidence in an environmental forensic investigation as well as providing fundamental data in monitoring in situ petroleum oil contamination. This study highlighted the power of the CSIA technique and potential application in various environmental research areas and could provide a strong platform for future research.