First results on Fe solid-phase extraction from coastal seawater using anatase TiO2 nano-particles

This paper describes the application of TiO(2) nano-particles (anatase form) for the solid-phase extraction of iron from coastal seawater samples. We investigated the adsorption processes by infra-red spectroscopy. We compared in batch and on-(mini)column extraction approaches (0.1 and 0.05 g TiO(2) per sample, respectively), combined to external calibration and detection by inductively coupled plasma mass spectrometry at medium mass resolution. Globally, this titania phase was slightly more efficient with seawater than with ultra-pure water, although between pH 2 and pH 7, the Fe retention efficiency progressed more in ultra-pure water than in seawater (6.9 versus 4.8 times improvement). Different reaction schemes are proposed between Fe(III) species and the two main categories of titania sites at pH 2 (adsorption of [FeL(x)]((3-x)+) via possibly the mediation of chlorides) and at pH 7 (adsorption of [Fe(OH)(2)](+) and precipitation of [Fe(OH)(3)](0)). Under optimised conditions, the inlet system was pre-cleaned by pumping 6% HCl for approximately 2 h, and the column was conditioned by aspirating ultra-pure water (1.7 g min(-1)) and 0.05% ammonia (0.6 g min(-1)) for 1 min. Then 3 g seawater sample was loaded at the same flow rate while being mixed on-line with 0.05% ammonia at 0.6 g min(-1) to adjust the pH to 7. The iron retained on the oxide powder was then eluted with 3 g 6% HCl (<0.002% residual salinity in the separated samples). The overall procedural blank was 220 +/- 46 (2 s, n = 16) ng Fe kg(-1) (the titania was renewed in the column every 20 samples, with 2-min rinsing in between samples with 6% HCl at 1.5 g min(-1)). The recovery estimated from the Canadian certified reference material CASS-2 was 69.5 +/- 7.6% (2 s, n = 4). Typically, the relative combined uncertainty (k = 2) estimated for the measurement of approximately 1 microg Fe kg(-1) (0.45 microm filtered and acidified to pH 1.5) of seawater was approximately 12%. We applied our method to a similar sample, from the coastal region of the North Sea. The agreement well within stated uncertainties of our result with the value obtained independently by isotope dilution mass spectrometry further validated our method.