COLLABORATIVE SAMPLING ERROR ASSESSMENT OF TRACE-METAL DETERMINATION IN SEAWATER

An international intercomparison of sampling devices was conducted to assess different collection procedures in gathering seawater samples from the Mediterranean for determination of Pb, Cd and Cu. Four kinds of samplers (Go-Flo, Niskin, van Dorn and Ruttner) with capacities of between 2 and 30 1 were used. The 20 1, Teflon-coated Go-Flo sampler was considered as the reference. Samples were collected at 20 and 50 m depth at four sites. Measurements were carried out by Differential Pulse Anodic Stripping Voltammetry (DPASV) both on board, immediately after collection, and on stored samples in the reference laboratory. Analytical quality control was guaranteed by the reference laboratory. The mean concentration values obtained using the reference sampler were (in nM unit): 0.34 (SD 0.04) for Pb, 0.058 (SD 0.012) for Cd and 3.7 (SD 0.6) for Cu, in agreement with recent, reliable literature data. Results obtained with other samplers showed that they became progressively cleaner during use (and the continuous washing programme) with concentration values tending to those obtained with the reference sampler. Better results were obtained for high-capacity samplers, indicating that the volume of sampler (or rather the ratio between the volume and the contact surface) can have a considerable effect on sample contamination.