A new method for the measurement of biogenic silica in suspended matter of coastal waters: using Si:Al ratios to correct for the mineral interference

Measurement of biogenic silica (BSi) in the suspended matter of estuarine and coastal waters is not trivial because of the interference of lithogenic silica (LSi) that can represent up to 90% or more of the total particulate silica within the water column. Till date, no method has provided a satisfying way to correct this mineral interference when the most common wet-alkaline digestions are used. The most recently published method, Ragueneau and Tréguer (Mar. Chem., 45 (1994) 43–51), presents the disadvantage to be site-specific. In this paper, we present a robust method to measure BSi in all estuarine and coastal waters and correction for LSi interference. It consists of a double wet-alkaline digestion where the filter sample is submitted to a first digestion (0.2 M NaOH, pH 13.3) at 100 °C for 40 min. At the end of this first leach, all the BSi and part of the LSi have been converted into Si(OH)4. Si and Al concentrations ([Si]1 and [Al]1) in the supernatant are analyzed. After rinsing and drying, the filter is submitted to a second digestion, exactly identical to the first one, leading to the determination of the (Si:Al)2 ratio that is characteristic of the silicate minerals present in the sample. The corrected biogenic silica concentration is thus given by [BSi]=[Si]1–[Al]1(Si:Al)2. There are three basic assumptions behind the proposed method: (1) all the Al measured during the first leach is derived from silicate minerals, (2) all the biogenic silica has been digested during the first leach, so the second leach addresses the Si:Al ratio of silicate minerals only, and (3) silicate minerals dissolve during the second digestion exactly as during the first digestion, i.e. with a similar Si:Al ratio. Theoretical considerations and field experiments demonstrate that the above assumptions are accurate. Blanks and repeatability tests indicate that the detection limit ranges from 0.6 to 1.5 nmol of BSi present on the filter and that the uncertainty is close to 10%. This method was successfully applied to more than 200 samples from five different coastal areas, demonstrating the importance of this correction, which can represent up to 92% of the apparent BSi concentration determined after the first digestion