DETERMINATION OF TOTAL MERCURY IN SCALP HAIR OF HUMANS BY GOLD AMALGAMATION COLD VAPOR ATOMIC-ABSORPTION SPECTROMETRY

Cold vapour atomic absorption spectrometry (CVAAS) with a laboratory-built system using a Au-Pt grid for mercury amalgamation, was applied in the determination of total mercury (Hg-T) in human scalp hair. Two sample dissolution procedures by HNO3 digestion were tested and compared: in a poly(tetrafluoroethylene) (PTFE) bomb for 1.5 h at 110-degrees-C (procedure 1), and in sealed Pyrex ampoules for 24 h at 50 +/- 10-degrees-C (procedure 2). After optimization of the aeration gas flow rate (1 00 ml min-1), de-amalgamation temperature (700-degrees-C) and releasing time (19 s), the analytical methodology was characterized and validated. The linear working range extended from 0.5 to 12.5 ng, and the characteristic mass was 0.29 ng of Hg (mass of analyte giving an absorbance of 0.0044). The repeatibility of measurements (within-day variation), expressed as percent relative standard deviation (RSD%) was 5.5% at 1.25 ng of Hg (n = 8) and 3.7% at 12.5 ng (n = 14), and the mean reproducibility (between-day variation), estimated from calibration curves on different days, was 6.4 +/- 1.1% (n = 5). The absolute detection limit (3 x sigma(b) was 0.1 3 ng of Hg and the limit of quantitation (1 0 x sigma(b)) was 0.43 ng of Hg (almost-equal-to 0.11 mg kg-1 in hair). Analytical precision (8.4 +/- 4.0% RSD) and accuracy (4.7 +/- 2.5% mean relative error) were satisfactory for ppm and sub-ppm levels of Hg-T in several biological and environmental certified and standard reference materials (CRMs and SRMs) including hair. Procedure 2 was selected as it is much simpler, requires inexpensive reagents and is more amenable for routine application. Mean recovery of Hg-T in hair spiked with Hg standard solutions was 105% in the range 0.83-1.27 mg kg-1. In addition, two human head hair samples were assayed as prospective laboratory control materials. Thus, sub-sample homogeneity, analytical intra-laboratory variability and external quality control were assessed to confirm the methodology in use. Between-day variation in hair analysis for Hg-T conducted on pooled scalp hair over a period of 5 months was 4.8% RSD (1.10 +/- 0.053 mg kg-1 for the 95% confidence interval; n = 16). Instrumental neutron activation analysis used as an independent method, showed significant correlation in results with CVAAS, both in several biological and environmental SRMs and CRMs and in human hair samples of pregnant and nursing women (Hg concentration range = 0.1 - 6.9 mg kg-1; n = 21; r2 = 0.880, p < 0.0001).