We measured silica concentration in sediment samples using a base dissolution technique described by Mortlock and Froelich (1989). We modified the technique to use smaller centrifuge tubes so that we could have a larger sample throughput. All reagents were decreased by a factor of three. Sample weights were decreased accordingly to assure that the saturation of silica hydroxide was not exceeded (the equivalent of 8 mg of Si or a solution concentration of 60 mM) (Mortlock and Froelich, 1989). We found that the replication of the analyses was improved if we followed the peroxide addition with 60 min of sonication (removal of organic carbon).
Splits of 10-cm3 samples were freeze-dried, lightly ground, and sieved to assure a uniform size distribution. Weighed samples (ranging from 0.03 to 0.05 g) were processed in replicate. A bulk consistency standard (composed of a mixture of sediments from Site 1098) was run as a sample at the beginning and the end of each group of 40 samples. Si concentration in known volumes of extractants (and thus in the solid samples) were determined using an automated spectrophotometric flow injection analysis system (Lachat QuickChem 8000). Results are reported as the mean ±1 s (sample standard deviation, N = 4).
The detection limit, assessed as three time the standard deviation of replicate measurements of a blank, was equivalent to 0.9 wt% for a typical sample size. Detection limits may be decreased by increasing the mass of sediment extracted. As these sediments were generally silica rich, this was not an issue. The bulk consistency standard had an average relative standard deviation of 4% within each run and a long-term mean (24 separate runs) of 23 ± 2 wt% biogenic opal.