Evaluation of the determination of mercury at the trace and ultra-trace levels in the presence of high concentrations of NaCl by flow injection–cold vapour atomic absorption spectrometry using SnCl2 and NaBH4 as reductands

Abstract

The influence of different concentrations of NaCl (0–20%m/v) on the analytical figures of merit for the determination of mercury by flow injection–CVAAS was studied. The determination was performed with and without preconcentration by gold amalgamation and by employing different reducing agents, NaBH4 and SnCl2. When using the reductand SnCl2, stable operation both with and without preconcentration was obtained, resulting in RSDs below 5% (n 30), even in the presence of high concentrations of NaCl. Without preconcentration the sensitivity was a function of the concentration of NaCl (between 0.033 and 0.013 ml ng–1). With preconcentration the sensitivity could be improved by a factor of 2.5 (i.e., from 0.084 to 0.037 ml ng–1, depending on the concentration of NaCl). With the reducing agent NaBH4, sufficient stability for routine analysis was observed only without preconcentration (RSD ⩽ 2%, n = 45) and a sensitivity of 0.022 ml ng–1 was obtained for matrix-free solutions. In the presence of high concentrations of NaCl the sensitivity increased up to 0.030 ml–1 ng, but owing to the unsystematically unstable blank values it was not a monotonic function of the concentration of NaCl. The use of preconcentration in the case of NaBH4 was hampered by the gold trap being covered with gaseous reaction products or entrained droplets as a result of the generation of foam in the gas–liquid separator. This could be improved by the use of two gas-washing bottles (NaOH, H2O). With the different methods, for which the detection limits finally were between 15 and 230 pg ml–1 (without) and 2 and 11 pg ml–1 (with preconcentration), mercury was determined in five standard reference samples of soil and sedimented sludge at the ng g–1 to µg g–1 level subsequent to their dissolution with aqua regia.