Optimization of the optical characteristics of a fiber-optic-guided laser fluorescence technique for the in-situ evaluation of fuels in soils

We have developed a system for the in situ determination of petroleum hydrocarbons in soils. This system uses a pulsed N2 laser coupled with a photodiode array detectorto make fluorescent measurements via optical fibers. The measurement is made through a sapphire window on a probe that is pushed into the ground with a truck-mounted cone penetrometer. This is the first reported direct optical detector for contaminants in soils. Remote in situ fluorometric measurements over long lengths of optical fibers give rise to several complications not encountered with conventional laboratory fluorescence measurements. The effects of these issues on the calibration and response of the optical detector are discussed. While we specifically discuss calibration of measurements of diesel fuel marine (DFM) by UV fluorescence, we believe that the calibration techniques and optical issues we are addressing will be germane to most if not all in situ optical measurements of contaminants in soils. In orderto improve the in situfluorescent quantitation of petroleum hydrocarbons, we have calibrated the fluorescent response of fuels as a function of soil type and conditions. The fluorescent response of DFM vanes by an order of magnitude or more as a function of soil type. Experiments to determine the causes of this variability have shown that the controlling variable is surface area of the substrate, although there are secondary effects as a function of grain size, mineralogy, and degree of soil aggregation. We have found that normalizing contaminant concentration to available soil surface area allows for a much more predictable response factor. It should be noted, however, that the variation in fluorescent response of OEM in three of four EPA soils tested is relatively small, with only one showing a large divergence. Soils with mixtures of grain types and sizes fall in a relatively tight response range, while soils such as very pure sands or clays diverge significantly. Preliminary studies on the effects of moisture content on the fluorescent response of OEM in soils suggestthat the addition of waterto the matrix begins to exclude DFM from the grain surfaces, forcing the fluorophore into the grain interstices, and greatly decreasing the difference in fluorescent response between soil types. The moisture effect is the smallest in mixtures of sand and clay. We are currently developing specific fluorescence calibration algorithms as a function of soil type, and correlating actual soil types to soil classifications derived from cone penetrometer strain gauge data.