SUMMARY: Seawater samples are collected at discrete depths using CTD-rosette sampling protocols. Subsamples for dissolved inorganic carbon are collected, immediately preserved with HgCl2 and stored for subsequent analysis in the laboratory using a commercial CO2 coulometer.

1. Principle

The accurate and precise determination of dissolved inorganic carbon (DIC) concentrations over annual and interannual time scales is central to the achievement of GOFS objectives. In the central ocean basins, DIC concentration in surface seawater is believed to be controlled by air-sea exchange reactions. However, physical processes and biological activity also influence the concentration of DIC in surface waters. Beneath the mixed layer the concentration of DIC increases as a result of the decomposition of organic material.

DIC concentrations are determined using a commercial CO2 coulometer. The coulometric determination of carbon dioxide has the unique distinction of performing with a high degree of both precision and accuracy while maintaining relatively high sample throughput. The coulometric determination of carbon dioxide involves the stripping of an acidified seawater sample with a carbon dioxide-free air stream and subsequent absorption of the carbon dioxide by a solution of ethanolamine. The weak acid generated by carbon dioxide absorbed in ethanolamine is titrated by a strong base produced electrolytically. The equivalence point is detected photometrically with thymolphthalein as the indicator. The number of coulombs required to reach the end-point is proportional to the quantity of carbon dioxide evolved from the sample.

2. Precautions

DIC samples should be the first samples taken from the water bottle unless dissolved oxygen (DO) or pH is also sampled from the same hydrocast. In this case, DIC samples are collected immediately after the DO and/or pH samples.

Careful subsampling is important for all dissolved gases. Samples should be taken as soon as possible and in the same manner as DO samples (i.e., no bubbles, low turbulence with adequate flushing; see Chapter 5).

3. Water Sampling

3.1. Drawing the sample
3.1.1. Samples are drawn into clean 300 ml glass reagent bottles as soon as the rosette arrives on deck.
3.1.2. The drawing tube is completely filled with sample by raising the end of the drawing tube. Bubbles are simultaneously dislodged by manipulation of the tube. The drawing tube is flushed and inserted to the bottom of the sample bottle.
3.1.3. The sample bottle is overflowed with at least two volumes of sample.
3.1.4. The tube is slowly withdrawn from the bottle while the sample is flowing so that the bottle remains brimful when the tube is completely withdrawn.

4. Preserving the Sample

4.1. Some of the sample is removed from the reagent bottle using a plastic pipette equipped with a rubber bulb. Enough water is removed so that approximately 1 ml of air is contained in the bottle when the glass stopper is inserted.
4.2. 100 µl of saturated HgCl2 is added to each sample. The tapered ground glass bottle neck is dried with a Kimwipe wrapped on an applicator stick. The bottle is sealed with a ground glass stopper coated with a light covering of Apiezon grease. The stopper is pressed firmly into the bottle to make a good seal. The stopper is secured with polyethylene tape or a large rubber band.
4.3. The samples are stored in a cool location, in the dark.

5. Coulometric Determination of DIC

5.1. Maintenance of extraction and analysis system and temperature control
5.1.1. The glassware used in the extraction system is combusted (450°C, 3 hours) on a regular basis in order to prevent the buildup of organic films within the extraction system.
5.1.2. The titration cell and rubber stopper are dried overnight at 55°C before use.
5.1.3. The air stream leaving the extractor is passed through a condensor and then through a Balston air filter.
5.1.4. The temperature of the titration cell is maintained at 25°C with circulating water.
5.2. Analysis
5.2.1. A 50 ml plastic syringe is rinsed with sample and then filled and weighed on a microbalance.
5.2.2. Five ml of 6% phosphoric acid is added to the extractor and the acid is purged for 2-5 minutes with carbon dioxide-free carrier gas.
5.2.3. The contents of the syringe are injected into the extractor through a septum.
5.2.4. The syringe is weighed again and the mass of the extracted sample determined.
5.2.5. The acidified sample is purged with carrier gas. Successive coulometer readings are integrated at 1 minute intervals until they differ by less than 0.01%

6. Determination of the Coulometer Blank

The coulometer blank is determined at intervals throughout each day by allowing the coulometer to titrate a CO2-free air stream. The blank is taken as the µg C per minute value detected by the coulometer when a steady-state reading is achieved.

7. Calibration

Although the digital coulometer output is fairly accurate, the coulometer response per unit carbon may vary with time. In order to achieve maximum accuracy, it is necessary to calibrate the coulometer with samples containing known quantities of inorganic carbon. We are presently using anhydrous sodium carbonate standards. The dried (270°C for 3 hours) reagent is carefully weighed on a microbalance to the nearest 0.1 µg and introduced into a degassed acidified sample solution in a combusted aluminum boat through a port in the side of the extractor. Recoveries are generally slightly less than 100%.

8. Data Reduction and Calculations

In order to compute the absolute concentration of DIC in a water sample, the integrated reading given at the titration endpoint must be corrected for both the coulometer blank and the recovery of NaCO3 standards. These corrections are made by multiplying the blank µg C min-1 by the time taken to reach the endpoint and subtracting this value from the integrated reading. This value is then corrected for the recovery of standards by dividing by the average percentage recovery of known standards run on the day of analysis.

9. Precision and Accuracy

Three replicate samples from a single Niskin bottle generally yield a coefficient of variation of less than 0.1% (+2 µmol/kg). The precision of our replicate determinations of liquid standards averages approximately 0.5 µmol/kg over periods of months. The accuracy of our determinations is within 1 µmol/kg as determined by the routine analysis of liquid standards provided by Dr. Andrew Dickson of Scripps Institution of Oceanography under the U.S. Department of Energy Global Change Research Program.

10. Quality Control

As a safeguard on the quality of our results, we maintain a set of secondary seawater standards. These are made from a large surface seawater sample, which is preserved with saturated HgCl2 and subdivided into 300 ml bottles. These are sealed as described above and stored in the dark. These secondary standards are analyzed relative to a primary seawater standard provided by Dr. Andrew Dickson.

11. Equipment/Supplies

12. Reagents

13. References