Discrimination of glass sources using elemental composition and refractive index: development of predictive models

Refractive index and metal ion concentrations (Al, Ba, Ca, Fe, Mg, Mn, Sr, Ti and Zr) were determined for four product-use categories: headlamp glass; container glass, non-vehicle window float glass, and vehicle float glass. Linear discriminant modelling using these data allowed differentiation of the four glass types but further discrimination was not possible within a product-use category. For this specific set of glasses, the concentrations of a number of metal ions were found to be correlated in some cases. This precluded the use of traditional probability calculations in using elemental composition data for interpretation of glass evidence. Alternative approaches to glass data interpretation are suggested.