During the expedition, 1099 TALK samples were collected according to methods outlined in the DOE Handbook (DOE 1994) and stored in 250-mL borosilicate glass bottles until the samples could be processed. Samples were poisoned immediately after collection with 200 L of a saturated solution of HgCl2 in deionized water to minimize biological activity prior to analysis. A total of 879 samples were analyzed at sea, and the remaining 220 were returned to the laboratory for analysis. The samples returned for shore-based analysis were from six stations: 30, 41, 54, 71, 77, and 86.
Samples analyzed at sea were brought to 25°C in a water bath, then pumped into a water- jacketed, 100-cm3, closed titration cell with a peristaltic pump. The cell design is described in the DOE Handbook (DOE 1994). The acid titrant was made by adding concentrated HCl to a 0.5 molar NaCl solution to give a normality of approximately 0.2. The calibrated acid normality (0.195091 ± 0.000041 µmol/kg solution) was determined immediately after the cruise by A. Dickson of SIO using the coulometric technique described by Millero et al. (1993). The titration system consisted of a personal computer, a Metrohm 665 Dosimat digital burette, an Orion 720A pH meter, and a Corning semi micro pH electrode. Using a program similar to those used by previous investigators (Bradshaw and Brewer 1988; Millero et al. 1993), the system automatically titrated the sample past the carbonic acid endpoint with electromotive force (emf) intervals of approximately 13 mV. The TALK was calculated from the full titration curve using the nonlinear least-squares approach described in the DOE Handbook (DOE 1994), with the exception that the effect of temperature on the dissociation constants was determined for every titration point based on the measured sample temperature. Sample weight was calculated from the cell volume, the measured sample temperature, and the bottle salinity value measured by the WHOI CTD group on all but three samples. The bottle salinity values from station 18, Niskins 25, 26, and 27, were significantly different (>0.1) from the CTD salinity values and from other bottle salinity values from equivalent depths for the surrounding casts. This difference was large enough to result in an error in the calculated density that was greater than the sample precision, therefore the sample density was recalculated for those three samples using the CTD salinity.
To evaluate the performance of the alkalinity system, CRMs were titrated between each set of samples from a station. Two titration cells were used to run the at-sea samples. The first cell was used for stations 1 to 33. This cell was replaced with a new cell before running the station 36 samples, because the CRM values titrated on first cell had a steady upward drift with time for the 8 days it was in use. The replacement cell gave a much more satisfactory performance and was used for the remainder of the cruise. Sodium carbonate standards were dried and stored in airtight vials following the procedures described in the DOE Handbook (DOE 1994). Sets of four solutions were made in precalibrated volumetric flasks using a 0.7 molar NaCl solution with nominal concentrations of 600, 1200, 2400, and 2800 µmol/kg. Four replicate titrations of each solution were made to generate a calibration curve for the cell. Fresh standards were prepared and titrated at the beginning of the cruise, before replacing the first alkalinity cell, after the replacement cell was in place, and at the end of the cruise.
The time-dependent increase in the first-cell TALK values was removed using a linear fit of the CRM TALK values as a function of time (Fig. 4). The mean TALK of the corrected CRM values was adjusted to match the mean of the second cell CRM values. The final calibration of all samples run at sea was then determined by fitting the sodium carbonate titration data as outlined in the DOE Handbook (DOE 1994). The fit of the nominally measured alkalinity vs the alkalinity calculated from the weight of sodium carbonate gave a slope of 0.987 ± 0.006 and an intercept of 12 ± 11 (Fig. 5). The final cell volume was determined by dividing the nominal cell volume by the slope. Applying this correction resulted in a mean batch 15 CRM alkalinity value of 2207.2 ± 3.9 µmol/kg (N = 98) for the cruise. The mean difference for sample replicates run at sea was 0.5 ± 3.8 µmol/kg (N = 71).
The P10 alkalinity samples returned to the laboratory for analysis were titrated using the same methods described for the at-sea samples. However, since accurate analytical balances were available, the cell volume was determined by careful weighing of the titration cell both empty and full of deionized water at a known temperature. The sample weight was converted to a cell volume based on density. A CRM sample (batch 17) was titrated every working day that the P10 samples were titrated to confirm the stability of the titration cell. The mean value for CRMs titrated over this period (~35 days) was 2207.6 ± 3.5 µmol/kg (N = 40). The mean difference between duplicate P10 samples run in the lab was 3.8 ± 5.2 µmol/kg (N = 12), not quite as good as the samples run at sea but not significantly different from zero. The mean difference between the samples run in the laboratory and the replicates run at sea immediately after collection was 3.8 ± 4.0 µmol/kg indicating that the storage of the samples did not affect the TALK values.