General Description: The Sami pCO2 analyzer utilizes a gas permeable membrane, which permits CO2 from seawater to diffuse to an indicator solution. The CO2 combines with water in the solution to form bicarbonate and carbonate, thus decreasing the pH of the indicator solution. Once the indicator solution has equilibrated with the seawater (ca. 5 min.), the pH of the solution is read using a spectrophotometer using three frequencies (2 frequencies for the pH (620 nm and 434 nm), 1 frequency as a blank to test for solution degradation and drift (740 nm?)). (Degrandpre et al 2000, see Wanninkhof et al 2013).

Accuracy: ±3 µatm (Sunburst website)

Precision: <1 µatm (Sunburst website)

Drift: <1 µatm over 6 months (Sunburst website)

Equilibration Method: Gas permeable membrane allows CO2 to equilibrate with an indicator solution, changing the pH of the solution.

Sensor: Spectrophotometric using 3 wavelengths. Can support up to 3 external instruments.

Notes: Max 600 m depth. Fresh or seawater. Approximate year-long lifespan before it must be recovered for maintenance and calibration (ca. 10000 measurements). Subject to biofouling if not equipped with biofouling package. (Sunburst website)

Calibrations: Calibration before and after deployment at Sunburst. Suggested additional bottle calibrations when SAMI is deployed and recovered.

Portability: High. Must be deployed and recovered at minimum annually.

Installation and Spatial requirements: Length 55 cm, Diameter 15.2 cm. Weight 7.6 kg in air, 1.1 kg in seawater. (Sunburst Website)

Potential Sources of Systematic Error:

1.     Temperature offset between seawater temperature and calibration temperatures. The pKa and carbonate equilibria are temperature dependent, so a correction is needed for the discrepancy. This is normally done during calibration (?). (Degrandpre et al 2000).

2.     Biofouling of the instrument. Biofouling can markedly change the pCO2 of the intake seawater. This is especially important for in-situ instruments. Sunburst has developed several mitigation tactics to help alleviate this concern.

3.     Fouling of the indicator solution. If the indicator solution fouls, all subsequent data will likely be unusable. Unfortunately, due to its in-situ design, any fouling of the solution will not be detected until after the SAMI has been retrieved from its mooring/buoy.

4.     Degradation of the gas permeable membrane. Similar to fouling of the indicator solution, degradation of the gas permeable membrane will not be detected until after the SAMI has been retrieved.

5.     Temperature and pressure effects on the gas permeable membrane. Some membrane based sensors have not been fully tested for the impacts of temperature and pressure on the permeability of the membrane. This may not be an issue with the SAMI, since the membrane is not externally located on the instrument, rather internal with seawater pumped using a solenoid valve.

6.     Indicator solution is not fully equilibrated with seawater. The systematic offset will depend on whether the atmospheric pCO2 is greater or less than the pCO2 of the water being measured. Normally this offset can be mitigated by maintaining a consistent and diffuse spray from the shower head and minimizing the amount of unequilibrated air introduced to the equilibrator. The latter is normally accomplished by adding a secondary equilibrator from which the primary equilibrator draws replacement air. This two stage approach prevents unequilibrated air from directly entering the primary equilibrator and supposedly allows some residence time in the secondary equilibrator for the new air to equilibrate. QCs of the in/out-flow to the secondary equilibrator are suggested to check that new air is not entering the system too quickly. Furthermore, in areas with large pCO2 gradients, the equilibration time of head space gas may not be sufficient to handle the gradient. This could result in a smearing of the gradient (i.e. the gradient appearing spatially larger than actual) or a complete mischaracterization of the gradient.