ANALYTICAL METHOD

Traditionally, analyses of ¹³C of DIC in natural waters has been accomplished by either stripping the gases offline (Kroopnick, 1974) Ortiz et al., 2000) followed by injection to a mass spectrometer or by vacuum-extraction of the CO₂ by cryotrapping online with a mass spectrometer (Graber and Aharon, 1991). In addition, Salata et al. (2000) describe a method based on sample acidification and offline equilibration followed by headspace sampling and isotope measurement using a gas chromatography-combustion (GC-C) system online with a continuous-flow gas-source mass spectrometer. These methods range in analytical precision from ±0.04 to ±0.2 can be time consuming and require a minimum of 1 mL of pore water sample (see review of techniques in Torres et al., 2005).

We developed a new method for measuring ¹³C in DIC to allow for analysis of small volumes, which are typically available for pore water studies, while minimizing the manual labor similar to the GC-C injection method of Salata et al. (2000). The new technique, described by Torres et al. (2005), is based on direct injection of a pore water aliquot into a GasBench-II headspace autosampler, a continuous-flow interface that allows injections of several aliquots of a single gas sample into a mass spectrometer. The sample headspace is flushed automatically with helium to reduce residual air. Acid (100 µL of 43% H₃PO₄) is then injected with a gas-tight syringe and the samples are allowed to evolve DIC as CO₂ gas into the headspace. After an equilibration period of ~12 hr, the headspace gases are flushed with a helium stream, which passes through a sample loop of selected volume (50, 100, or 250 µL). The sample loop is charged with gas and a known volume of sample is then transferred to a second helium stream that flows through a gas chromatography column to separate the CO₂ from other gas compounds and a porous membrane trap to remove water. The dry sample stream is transferred to a Finnigan DELTAplusXL mass spectrometer, which integrates the relevant isotope masses (m/z 44, 45, and 46) as the CO₂ peak enters the source.

The method requires <0.5 mL of seawater sample (DIC > 2 mM), or ~12 µg C. For the Leg 204 pore water samples, with alkalinity values ranging from that of seawater to concentrations >100 meq/L, the sample volumes used ranged from 30 to 50 µL. This low volume requirement allowed duplicate measurements in selected samples to ascertain data precision. The technique involves little or no manual preparation of samples and allows throughput of ~80 samples per day in fully automated mode, including the delay time for equilibration. Based on multiple standard measurements, the overall precision of this technique is conservatively estimated to be better than ±0.15 (Torres et al., 2005). Standardization is provided by tank CO₂ referenced to an array of international standards, and analyses are monitored against a stock solution of reagent NaHCO₃.

The isotopic composition of the DIC in Leg 204 samples was measured in pore water subsamples preserved with HgCl₂ (5 µL/mL) and flame-sealed in 5-mL glass vials immediately after collection.