Combining Dispersed Particle Extraction with Dried-Droplet Laser Ablation ICP-MS for Determining Platinum in Airborne Particulate Matter.

A combination of analyte pre-concentration using dispersed particle extraction (DPE) and dried-droplet laser ablation inductively coupled mass spectrometry (LA-ICP-MS) was developed with the aim to quantify Pt and Pd in urban particulate matter with an aerodynamic diameter ≤2.5 µm (PM2.5 ). The PM2.5 aerosol was collected on cellulose ester filters during a sampling period of three days, with sampling intervals of 4 h only. Each of the filters was chemically digested, and the resulting solution was pre-concentrated using DPE. Droplets taken from the pre-concentrated sample were deposited on polymeric disks and dried. These dry spots were then analyzed by means of LA-ICP-MS. This approach allowed ICP-MS analysis of solutions with high content of dissolved sorbent particles coming from the DPE procedure. Furthermore, spectral interferences arising from sample-inherent matrix elements as well as solvent-related interferences could be removed by the proposed approach. The method was validated by determining the Pt concentration in Bureau Communautaire de Référence certified reference material (BCR CRM) 723 road dust certified reference material and a good agreement with the certified value was obtained. The temporal variation of Pt during the three-day sampling period is discussed, with respect to automotive traffic. The daily average of Pt measured in the air corresponds to typical values observed in urban areas in Central Europe. Although the pre-concentration of palladium is feasible with dispersed particle extraction, the method detection limits achieved here did not allow to quantify this element in the CRM or in the PM2.5 samples. The source for these high method detection limits for palladium are blank values arising from the filter material as well as the digestion procedure of the PM2.5 samples. Instrumental sensitivity of the approach would, however, suggest that palladium quantification is possible, provided the abovementioned blank issues are controlled better.