Analytical methods are reported in detail in Mix et al. (1999) and Andreasen et al. (Chap. 8, this volume). Isotope data in this paper are from benthic foraminifers, primarily Cibicidoides wuellerstorfii and Uvigerina species.
All carbon analyses except the CaCO3 analyses from Core W8709-13 were run at Boise State University. CaCO3 analyses on core W8709-13 were run at Oregon State University using a UIC coulometer with an acidification module.
At Boise State University calcium carbonate was determined by coulometry using a UIC Inc. Model CM-5012 CO2 coulometer attached to a modified version of the CM-5120 combustion furnace, schematically shown in Figure 2. Approximately 30-70 mg of dried, homogenized sediment sample was combusted in an ultra-high-purity oxygen (>99.994% O2) atmosphere at 1000°C. The reported CaCO3 data were calculated from two independent analyses of each sampled interval: first we measured the total carbon concentration and then the Corg fraction was measured in the second analysis after pretreatment with a 10% hydrochloric acid solution to dissolve solid CaCO3. The carbonate carbon was calculated as the difference between the total carbon and the organic carbon fractions, and the CaCO3 weight percent was calculated by multiplying carbonate carbon by 8.33.
We modified the CM-5120 furnace for marine sediment analysis because the residual sea salt caused premature breakage of the combustion tube. The CM-5120 furnace was designed with post-combustion scrubbers to remove impurities from the CO2 gas evolved from sample combustion. Chlorine is removed as silver chloride by reaction with pure silver wire positioned at the end of the combustion tube. Residual sea salt produced excessive AgCl and formed a solid block in the combustion tube, resulting in breakage. We solved this problem by removing the silver from the UIC combustion tube and attaching a second tube furnace to the CM-5120 with teflon tubing for the chlorine scrubber. High-purity fused quartz tubing (7 mm outer diameter × 5 mm inner diameter × 610 mm length) was filled with ~15 g of pure silver wire and heated during the analysis to 475°C. The optimum operating temperature was determined to best remove chlorine from the gas stream after several trial and error experiments. This is based on the absence of a gas cloud in the cathode compartment of the titration cell, a stable cell current of zero mA (0 ± 1 mA) after sample analysis, and reproducible carbon measurements of in-house standards of several marine sediments also containing residual sea salt (Table 2).
We used the following acidification method to measure Corg. Approximately 70 mg of dried sample was placed inside a precleaned, custom made, high-purity fused quartz combustion boat (55 mm × 10 mm × 5 mm). The sample was wetted with four drops of water, and then with 10 drops of 10% HCl. The slurry was stirred with a glass stir rod and rinsed with four more drops of HCl, to wash back into the sample boat any sample adhering to the glass rod. This step prevented cross contamination between the sample boats and greatly improved the reproducibility of results. The fused quartz boats were heated at 110°C until sufficient solution evaporated to accommodate a second treatment with 10 drops of 10% HCl. Samples were oven dried and allowed to cool before analysis. We emphasize the importance of using fused quartz combustion boats. Previously, we had used a glazed porcelain boat (Coors Model #60032) but discovered the glaze deteriorated after about a dozen sample runs. Partial absorption of the sample slurry into the exposed porcelain produced erratic results and unacceptably high background readings.
The accuracy and precision of the data were estimated by including two standards in each sample run and by repeat analyses of the unknown samples. For the standards we used a reagent-grade calcium carbonate and an in-house standard (SX27547, Gulf of California marine sediment) of a similar matrix as Leg 167 sediments. Two samples of the calcium carbonate reagent were analyzed at the beginning and at the end each sample run. Four samples of the in-house standard were analyzed in a typical sample run: two samples of SX27547 were analyzed for total carbon and two for organic carbon following the acid treatment described above. Summary statistics for these standards are found in Table 2.
Precision was estimated by repeating the analysis of every fourth "unknown" sample in the sample run. The repeated analyses pertain to both the total carbon and organic carbon measurements. The average difference between the repeated samples is <0.01% carbon as shown in Table 3.