which were always performed in triplicate. As can be seen, good recoveries higher than 90% were reached for the samples with low salinity.
On the other hand, in the analysis of a sea water (saline sample) the recovery was 79.5%, evidencing the effect of ionic
strength already verified in the study of the influence of thisparameter.
So on, this sample was analyzed by standard addition , which provides similar result to that obtained by ETAAS (limit of detection = 0.60 g l−1), proving that is possible to achieve good accuracy, when standard addition method is employed. Copper was also determined (in triplicate) in all other samples by ETAAS not presenting statistical difference between results obtained, when a paired t-test was applied (95% confidence level).
All these results are summarized in Table 3.