Ultraviolet Spectrometric Characterization of Metal Decontamination

The viability of an ultraviolet fiber optic method for quantifying contamination levels on metal substrates was demonstrated. Contaminated metal samples obtained from industrial processing operations were cleaned with alternative solvents being used to replace chlorofluorocarbons, which are in the process of being phased out. Cleaned samples with varying amounts of oils remaining on them were analyzed for residual amounts using four testing methods: X-ray spectroscopy, infrared spectroscopy, ultraviolet spectroscopy, and gravimetric analysis. Multivariate data analysis techniques were used to prepare a model using ultraviolet and infrared data. Both principal component regression (PCR) and partial least squares (PLS) methods were used in the analysis of the data. A one-factor model was found to be adequate for the infrared data. Interpretation of the results using different types of cross validation showed that two factors were required for the ultraviolet model. Cleanliness levels were also determined using electron scanning for chemical analysis (ESCA) and gravimetric methods. The results from all four methods were expressed in terms of amount of contamination in milligrams per unit area. The values obtained were compared against actual contamination data, and the sum of squares of errors was calculated. The predicted residual sum of squares of error for an equal number of comparison points for IR, UV, ESCA, and gravimetry were 0.20, 0.84, 1.39, and 2.54 (mg. /ft.2)2 respectively. The predictive ability of each method was related to its cost and ease of use. The ultraviolet instrumentation tested in this study ranked second in predictive ability and second in cost, making it a viable option for cleanliness testing purposes.