Flow injection on-line dilution for zinc determination in human saliva with electrothermal atomic absorption spectrometry detection

An automated method is described for the determination of zinc in human saliva by electrothermal atomic absorption spectrometry (ET AAS) after on-line dilution of samples with a significant reduction of sample consumption per analysis (<0.4 mL including the dead volume of the system). In order to fulfill this aim without changing the sample transport conduits during the experiments, a flow injection (FI) dilution system was constructed. Its principal parts are: one propulsion device (peristaltic pump, PP) for either samples, standards or washing solution all located in an autosampler tray and for the surfactant solution (Triton X-100) used as diluent, and a two-position time based solenoid injector (TBSI1) which allowed the introduction of 10 μL of either solution in the diluent stream. To avoid unnecessary waste of samples, the TBSI1 also permitted the recirculation of the solutions to their respective autosampler cups. The downstream diluted solution fills a home made sampling arm assembly. The sequential deposition of 20 μL aliquots of samples or standards on the graphite tube platform was carried out by air displacement with a similar time based solenoid injector (TBSI2). The dilution procedure and the injection of solutions into the atomizer are computer controlled and synchronized with the operation of the temperature program. Samples or standards solutions were submitted to two drying steps (at 90 and 130 °C), followed by pyrolysis and atomization at 700 and 1700 °C, respectively. The aqueous calibration was linear up to 120.0 μg L−1 for diluted standard solutions/samples and its slope was similar (p > 0.05) to the standard addition curve, indicating lack of matrix effect. The precision tested by repeated analysis of real saliva samples was less than 3% and the detection limit (3σ) was of 0.35 μg L−1. To test the accuracy of the proposed procedure, recovery tests were performed, obtaining mean recovery of added zinc of 97.8 ± 1.3%. Furthermore, Zn values estimated by the procedure developed in this work were compared with those obtained by a standard addition flame-AAS method applied to 20 randomly selected saliva samples. No significant differences (p > 0.05) were obtained between the two methods. Zinc levels in saliva samples from 44 healthy volunteers, 15 male and 29 female, with ages between 20 and 51 years (mean 30.50 ± 9.14 years) were in the range 22–98 μg L−1 (mean of 55 ± 17 μg L−1), similar to some and different from others reported in the literature. It was found that zinc values for male were statistically higher (p = 0.006) than for female.