Total Petroleum Hydrocarbon Concentration Prediction in Soils Using Diffuse Reflectance Infrared Spectroscopy

Partial least squares (PLS) calibration models, from diffuse reflectance infrared Fourier-transform (DRIFT) spectra, were developed for the prediction of total petroleum hydrocarbon (TPH) concentrations in contaminated soils. Soils were collected from sites in southeastern Australia known to be contaminated with TPH. Regression models were derived for concentration ranges of 0 to 32,600, 0 to 5000, 0 to 15,000, and 15,000 to 32,600 mg kg(-1). For each range, PLS models using selected near-infrared (NIR) and mid-infrared (MIR) frequencies were tested. The aliphatic alkyl stretching vibration regions were the most sensitive to TPH: NIR frequencies at 4500 to 4100 cm(-1) and MIR at 3000 to 4600 cm(-1). The MIR range included two -CH3 peaks, one at 2950 cm(-1) and the other near 2730 cm(-1), having strong correlation with TPH at low and high TPH concentrations, respectively. These peaks were considered to be either weak or absent in natural soil organic matter relative to the usual -CH2 region at 2930 to 2850 cm(-1). The PLS regression analysis using the combined 2980 to 2950 and 2777 to 2650 cm(-1) MIR frequency regions for the 0 to 15,000 mg kg(-1) set resulted in a ratio of prediction deviation of 3.7 ("analytical quality"), coefficient of determination (R-2) = 0.93, and root mean square error of cross-validation of 564 mg kg(-1). Using these MIR frequencies, the DRIFT infrared technique showed the potential to be a rapid and accurate nondestructive method to determine TPH concentrations in contaminated soils, potentially adaptable to in-field use.