Variation in the determination of platinum group metals using ICP OES induced by the effect of complex matrices and the correction method based on multivariate calibration

Abstract

The presence of interelements due to sample complexity can produce significant effects on the quantitative determination of platinum group metals (PGMs) by using inductively coupled plasma optical emission spectrometry (ICP OES). However, research on such an important issue is still limited and poses a considerable challenge towards increasing the quality of PGM determination. In the present work, the deviation of PGM determination induced by a single and a mixture of matrix elements was investigated based on the selection of PGM emission lines, the variation of the analytes (Pt, Pd and Rh) and the concentration of interferents (Al, Ca, K, Na). The addition of selected interferents could result in an underestimation of the measured concentration (negative effect on Pt and Pd at all emission lines, and Rh at 233.477 nm) or an overestimation (positive effect on Rh 343.489 nm). Among the examined emission lines of PGMs, Pt 214.423 nm and Rh 233.477 were the most sensitive wavelengths responding to the effect of the interelements, which produced intensive bias between the measured and actual concentrations of PGMs up to ∼30%. This level of effect was significantly extended when the concentration of matrix elements increased, except for Rh 343.489 nm. A 2² factorial design was applied to describe the combination effect between two interelements, and the model could express the behavior by the agreement between the statistical significance of the interaction coefficients and the depression in the combined effects compared to the sum of individual effects. A correction method based on the multivariate calibration technique was employed to reduce the effect of matrix complexity, namely multiple linear regression (MLR). The MLR method performed excellently to minimize the interferences and enhance the accuracy of PGM determination.