Previous   |    Next

doi:10.2204/iodp.proc.331.202.2015

Analytical methods

Total digestion methods

The details of our sediment digestion methods are outlined in Muratli et al. (2012). Briefly, dried and ground sediment samples were digested using a combination of inorganic acids (HCl, HNO₃, and HF) in a CEM MARS-5 microwave oven (CEM Corp., Matthews, NC). Postdigestion evaporation within the system was accomplished utilizing the MicroVap accessory. Following sample dilution with 5% HNO₃ into preweighed 10 mL HDPE bottles (this dilution is somewhat modified from Muratli et al., 2010a, 2010b, 2012), samples were heated in an analog block heater for ~24 h. Prior work identified the necessity of this step to redissolve any remnant fluoride-metal complexes that, although they might not be visually apparent in the digestion matrix, were found to impact analytical results (e.g., Muratli et al., 2010b, 2012).

ICP-OES

Samples for major elements (Tables T1, T2) were run on a Leeman Laboratories Prodigy inductively coupled plasma–optical emission spectrometer (ICP-OES) at the W.M. Keck Collaboratory at Oregon State University (USA). This instrument is capable of two viewing modes, axial and radial, and the various elements were run in one mode or the other. Table T1 also presents the specific emission wavelengths utilized for each element. Samples and standards were diluted twenty-fold with 1% quartz-distilled nitric acid. Repeated runs of the standard curve monitored instrumental drift, and for those elements influenced by drift we applied a correction factor. The reported data represent the mean of three replicate analyses for each element in each tube except in instances where we have results from multiple sample digestions. In this case, the mean is the average of the two results. Uncertainties for samples that were not digested in duplicate are derived from two sources: (1) the regression uncertainty, which is calculated using the standard error of the regression, and (2) the internal uncertainty calculated from the standard deviation of the three replicate analyses. These two uncertainties are combined as the square root of the sum of squares.

ICP-MS

Trace constituents were analyzed on a Thermo X-Series II inductively coupled plasma–mass spectrometer (ICP-MS), also at the W.M. Keck Collaboratory. IPC-MS results can drift over the course of the day’s run, and drift is corrected by spiking each sample with an internal standard solution. This solution consisted of 250 ng/mL Be and 50 ng/mL each of Rh and Bi or just 50 ng/mL of Rh and Bi. Measured counts throughout the run were corrected to the value of the first blank run of the day. Because of the varied matrix in these samples, Rh was used as the lone internal reference standard for the calculation of results. Although using only a single element for drift correction is less than optimal, samples had varying amounts of Be and Bi, thus making reliable concentration results difficult to obtain. Sample dilutions varied with concentration, but samples were diluted with 1% HNO₃. Sample uncertainties were calculated using the same approach employed for the ICP-OES results.

We used the following isotope masses to calculate concentrations: ⁶⁶Zn, ⁹⁸Mo, ¹¹¹Cd, ¹¹⁴Cd, ¹³³Cs, ¹⁴⁶Nd, ¹⁸⁵Re, ¹⁸⁷Re, and ²³⁸U. Multiple analyte masses have interferences; thus, some of the analytes above were used to estimate these interferences. For ¹¹⁴Cd we measured ¹¹⁷Sn to assess the level of interference of ¹¹⁴Sn:

An additional interference from ¹¹⁴MoO on ¹¹⁴Cd is estimated from the Mo results and the estimated MoO/Mo formation ratio. For this work, we only report results from the ¹¹⁴Cd analyses because we are more confident in the correction for this Cd mass. For ¹⁸⁷Re, a series of solutions containing Yb were analyzed and a YbO/Yb ratio was calculated. Counts of ¹⁷¹Yb were then used to estimate the interference on ¹⁸⁷Re. ¹⁸⁵Re has an interference by TmO; for this reason, we generally use ¹⁸⁷Re to calculate Re concentrations. However, at higher Re concentrations (and higher Re dilutions) this interference is small, and we use ¹⁸⁵Re for estimating concentrations.

Dithionite extractions of Fe and Mn

To determine “reactive” iron and manganese, we employed a single-step dithionite extraction at 60°C (Mehra and Jackson, 1960; Kostka and Luther, 1994; McManus et al., 2012; Roy et al., 2013). We used approximately 0.25 g of dried ground sediment to which we added ~10 mL of dithionite reagent. pH is buffered using a sodium acetate, sodium citrate solution (Roy et al., 2013). These extractions are generally thought to dissolve amorphous Fe oxides, some crystalline Fe oxides, and acid volatile sulfides (Kostka and Luther, 1994; Poulton and Canfield, 2005). However, for many of the samples digested here it is possible that the sediments are sufficiently reducing to the point where sulfides are the dominant reactive phase; thus, any future interpretations of the data presented here require caution regarding the specific phase being extracted using this procedure. Unlike Fe, this procedure is not well calibrated for Mn (McManus et al., 2012).

Quality control

To access accuracy and precision, we report values from replicate digestions of a PACS-2 standard and an in-house laboratory sediment standard (Tables T2, T3, T4). This latter standard is a continental margin sediment standard from the Chile margin (Muratli et al., 2010a, 2010b). We also report our own laboratory’s long-term average for these materials (Tables T2, T3). Finally, we redigested/reanalyzed a number of samples, which are indicated within the data tables. Because the samples that we analyzed as part of this study are quite unlike either of the reference material matrixes, the reproducibility of our reference materials may not be an accurate measure of the true reproducibility of the samples. Generally speaking, however, the samples did indeed reproduce quite well, with one exception. As noted in Table T3, the Zn and Cd results from one duplicate sample (331-C0014B-3H-5, 20–30 cm) did not agree, and the results for these two analytes (for one of the duplicates) were higher than all other samples measured during this study. We did not obtain an accurate concentration estimate for these two analytes because the concentrations were above our highest standard and we did not reanalyze this sample. We report here a lower concentration for the duplicate, for which we have a more analytically robust measure of these two analytes. We suspect that either the sample was somehow contaminated or it was not homogeneous. We do note, however, that other elements do show agreement between these two duplicates.

In the case of our Fe and Mn extractions, we also extracted these same sediment standards and report the data in Table T4. These values are not standardized and can thus only be used as a measure of technique precision. Furthermore, many of the reported results are the result of duplicate sediment extractions and analyses, and for these samples the average values are reported (± 1 standard deviation).

Top of page   |    Previous   |    Next