Graduation Year


Document Type




Degree Name

Doctor of Philosophy (Ph.D.)

Degree Granting Department

Marine Science

Major Professor

Paula C. Rhodes, Ph.D.

Co-Major Professor

Kendra L. Daly, Ph.D.

Committee Member

Albert C. Hine, Ph.D.

Committee Member

Robyn N. Conmy, Ph.D.

Committee Member

Ana R. Arellano, Ph.D.

Committee Member

Mark E. Luther, Ph.D.


baffle flask, Deepwater Horizon, EEMs, fluorescence, Gulf of Mexico, wave tank


The genesis of this research was the Deepwater Horizon oil spill, which discharged petroleum and gas into the Gulf of Mexico for 87 days in 2010. High-resolution fluorescence spectroscopy was employed for the detection of petroleum in seawater samples following the oil spill. Fluorescence arises from the chemical structure of π-bonding in C=C bonds, especially those in aromatic structures. Spectrofluorometry was also used to observe and track the formation of petroleum plumes in seawater undergoing controlled physical dispersion in a wave tank, both with and without the addition of chemical dispersant. Further, the changing fluorescence characteristics of a broad range of 25 types of petroleum, with the addition of chemical dispersant at differing application rates, were investigated in the laboratory.

Following the guidance provided in the U.S. SMART protocol, many researchers employed a variety of in situ fluorometers to inform their water sampling efforts in tracking the oil spill, as well as to gauge the effectiveness of chemical dispersant application to surface slicks. Excitation emission matrix spectroscopy (EEMS) was performed on discrete water samples collected and analyzed, both at sea and in our laboratory in the year following the DWH oil spill, in order to investigate the optimal excitation and emission wavelengths for the detection of petroleum.

In order to further explore the performance of in situ fluorometers used following the DWH oil spill, EEMS analysis was performed on discrete water samples collected in a series of wave tank experiments conducted at the Bedford Institute of Oceanography (BIO) in Nova Scotia, Canada. In situ fluorometers were mounted within the wave tank, which was then filled with filtered seawater from Halifax Harbor. A randomized series of experiments using oil collected from the DWH oil spill, both fresh and weathered, with and without the addition of chemical dispersant, was conducted over a two-week period. High-resolution EEMs of water samples collected at specific time points were compared with the fluorescence signals collected with in situ instruments, as well as with chemical analysis by GC/MS.

Finally, a series of experiments was conducted to investigate the variation in fluorescence signals exhibited by a broad variety of oil types. EEMS analyses of 25 types of oil, both without the addition of chemical dispersant, and at three different dispersant to oil ratios (DORs) was performed using artificial seawater in baffled trypsinizing flasks on a shaker table. Chemical analysis was also performed by GC/MS on oil-in-water samples, with no chemical dispersant added, and on samples at the highest DOR of 1:20. Parallel Factor Analysis (PARAFAC) was utilized in an attempt to identify components specific to petroleum, dispersant, and/or natural colored dissolved organic matter (CDOM) both within each experimental series and across all samples. Four characteristic oil-type fluorescence peaks were identified in the EEMS analyses. A clear linear relationship was seen between fluorescence intensity and concentration of 2-ring polycyclic aromatic hydrocarbons (PAHs) in oil-water without chemical dispersant; however, the relationship between fluorescence intensity and PAH concentration at highest chemical DOR was not straightforward. Comparison of fluorescence intensity in the four peak regions enabled a division into two overarching oil types related to oil viscosity. As evidenced by EEMS, higher viscosity Type II oils do not respond well to the addition of chemical dispersant. PARAFAC analysis showed changes in the contribution of intensity from different fluorescence regions with increasing levels of dispersion, likely related to the action of chemical dispersant reducing oil droplet size, which in turn reduces reabsorption of fluorescence.

Results of EEMS analysis of wave tank samples provided good agreement with the signal from all in situ fluorometers tested and showed that all instruments would have been able to detect oil-type fluorescence in the field. Differences were noted in the evolution of fluorescence peak location over the 90-minute course of the experimental series between oil with and without chemical dispersant. Highest intensity oil-type fluorescence was found at the excitation and emission wavelength pair known to be characteristic of naphthalene. Chemical analyses showed a relationship between 2-ring and 3-ring PAHs only with dispersed oil. Good correspondence was also seen between total benzene, toluene, ethylbenzene and xylene (BTEX) concentration and a ratio of fluorescence intensity at two emission wavelengths. PARAFAC analysis showed agreement with components found in the baffle-flask series.

EEMS analyses of field samples collected in the Gulf of Mexico in the year following the DWH oil spill show correspondence between fluorescence intensity in the oil-type regions seen in both the bench-scale and mesoscale experimental series. An interesting evolution of oil-type fluorescence intensity over the course of the three research cruises showed the continued presence of petroleum at or near the surface, as well as a continued deep-water petroleum signature through May 2011. The interplay of fluorescence intensity at oil-type and protein-type fluorescence regions also appeared to show the response of oil-degrading bacteria.

This research has shown the presence of fluorescence peak regions characteristic of petroleum, which can be distinguished from protein-like and CDOM-like fluorescence naturally present in the marine environment. Further, fluorescence measurements can be accomplished with very small quantities of sample (3 mL), are relatively fast to process, do not involve complex pre-processing, and are sensitive down to the ppb range. PAHs are known to be toxic at very low concentrations. Chronic petroleum spills are ubiquitous, and with petroleum exploration in ever more extreme environments, future large-scale spills are unfortunately likely to occur. The ability to track petroleum spills, especially in deep sub-surface plumes, will facilitate rapid response efforts to protect vulnerable marine ecosystems, which are still little-understood or perhaps even remain undiscovered.