In the present work, quantitative analysis of major and minor elements in aluminum alloys is investigated using chemometrics and laser-induced plasma spectroscopy with a commercially available laser-induced breakdown (LIBS) spectrometer. Multivariate calibrations use the entire signal matrix for all elements in a single multivariate regression model. This enables accounting for the correlation between variables often referred to as matrix effects in conventional univariate modeling. Modeling the entire signal matrix improves robustness over traditional univariate calibration since it can compensate for matrix effects. Several nonlinear data pretreatment methods have been used to correct for nonlinear behaviors of the analytical signals prior to performing the multivariate calibration. The use of multivariate calibration in combination with cubic implicit nonlinear data pretreatment showed the most accurate results. The accuracy reported with the developed multivariate calibration is better than 5% for the major alloying elements. Based on the results obtained, the use of chemometrics and laser-induced plasma spectroscopy have been successfully applied to the quantitative analysis of major and minor alloying elements in aluminum.

Electrical spark-optical emission spectroscopy (OES) has been used for over 50 years for the direct analysis of alloys in the aluminium industry. This method has undergone continual improvement over time, but conventional OES can barely reach the new specification limits required for some critical aluminium alloys. As part of a continuous improvement process, instrument manufacturers, academic and industrial researchers are now showing a particular interest for laser induced plasma spectroscopy (LIPS) as an alternate and/or complementary approach for the analysis of aluminium alloys. In this work, by using the same optical mounting (i.e. same spectrometer model with the same spectral lines), we have directly compared the analytical performances of these two sources. By comparing the calculated concentrations obtained with these two sources using the same samples and the same standards, we show that in general the two sources give comparable quantitative results for major, minor and trace elements. The precision obtained with LIPS is improved by a factor of two when compared to traditional spark-OES. The results obtained in this work establish LIPS as a potential method for the quantitative analysis of aluminium alloys in industry.

No abstract