The application of experimental design to investigate the solvent matrix effects observed during the Determination of Rhodium (Rh) in organic media by Graphite Furnace Atomic Absorption Spectrometry (GFAAS)

Abstract

In an industrial application a GFAAS method for monitoring the Rh concentration in process streams is being used. Matrix effects are known to exist with the application of this technique; in fact, it was observed that different solvents lead to different results. Therefore, standard additions have to be employed for quantitative determinations, resulting in high costs and long analysis times. In an attempt to understand these interfering effects, fractional factorial designs were proposed to determine whether any GFAAS parameter was responsible for, or related to, the matrix effects. Seven GFAAS parameters were investigated: final temperature, ramp time and hold time of the transitions step (from the dry step); final temperature, ramp time and hold time of the ashing/pyrolysis step; ramp time of the atomisation step. The results showed that the matrix effects were not related to any specific parameter. A complete factorial design was implemented to demonstrate the fundamental role of the atomisation temperature. SEM analysis showed that the surface of the graphite tubes might be affected in different ways by different solvents. A Principal Component Analysis demonstrated that the matrix effects may be related to the viscosity and melting point of the solvents and may be independent of their molar mass. To identify the origins of these effects, an investigation on the link between the tube surface-sample matrix interactions and the physical properties of the matrices is recommended. Since GFAAS parameters cannot compensate for the matrix effects, standard additions remain the preferred mode of operation as it accounts for the effects in-situ.