Ultraviolet–Visible-Near InfraRed spectroscopy for assessing metal powder cross-contamination: A multivariate approach for a quantitative analysis

Abstract

The last few years have seen an increasing use of spherical metals powders to produce bulk parts through metal forming technologies like Additive Manufacturing and Metal Injection Molding. This, coupled with the wide availability of metal powders, leads to a critical issue: contamination across different systems in different process steps. Consequently, it is necessary to find a new, fast, and reliable analysis sensible to tiny traces of contamination. This work evaluates the applicability of Ultraviolet–Visible-Near InfraRed (UV–Vis-NIR) spectroscopy, a technique providing information on powders’ reflectance, for studying contaminated powders. This work focuses on assessing 3 binary systems obtained from the cross-contamination of 3 components (A92618, C10200 and S31603) in a low contamination range (from 0.5 vol% to vol. 6%) and in a high contamination range (25 vol% and vol.50%). After the UV–Vis-NIR analysis, multivariate analysis has been used to obtain quantitative results. Results show that, as the contamination level increases in the binary system, the shape of spectra changes and becomes progressively more similar to the contaminant one. The chemometric analysis allows the detection of the contaminant type and its concentration percentage in the contaminated powder.