2.2.3. Carbon Dioxide Standards and Reference Gas Calibrations

A manometric absolute calibration system was developed for performing primary calibrations of the mole fraction of CO₂ in reference gas mixtures. Figure 2.4 shows a schematic diagram of the system. The brief specifications are as follows:

Accuracy: 0.07 µmol mol^-1
Precision: 0.06 µmol mol^-1
Gas consumption: 15 L in 100 Kpa
Calibration time: 4 hours

Fig. 2.4. Schematic diagram of manometric system to determine CO₂ mole fraction in samples of dry air.

There are 15 primary CO₂-in-air WMO standards in aluminum high-pressure cylinders ranging in CO₂ mole fraction from approximately 250 to 520 µmol mol^-1. These primary standard cylinders were analyzed by the manometric calibration system from September 1996 through February 1997. Each standard cylinder was analyzed at least three times, often with several days of separation or more between calibrations. A total of 65 individual manometric determinations were made. The results of analyses are summarized in Table 2.6. For comparison, the CO₂ mole fractions measured by the Scripps Institution of Oceanography (SIO) in these same cylinders relative to the WMO X93 scale (maintained at SIO until 1997) are also shown in Table 2.6. The mean precision of the manometric measurements indicated in Table 2.6 as the standard deviations is about 0.07 mol mol^-1 for a total of 65 measurements. Figure 2.5 gives CO₂ mole fraction comparisons between CMDL manometric determinations and SIO total means with error bars. The concentrations analyzed by two laboratories for the primary standards in the atmospheric CO₂ concentration range of 300 to 400 µmol mol^-1 are in good agreement, so that the switch from SIO to CMDL for the maintenance of the WMO mole fraction scale will not result in a noticeable shift of the scale. The function of the primary standards is to provide continuity to the WMO scale as well as a quality control check on the performance of the manometric system. The assigned values of the primaries are ultimately determined through the manometric analyses repeated annually.

TABLE 2.6. Measurements of NOAA Primary WMO CO₂ Standards

*Note: SIO measurements on these cylinders were performed during 1991-1993 and during 1996. Between parentheses are the means of SIO measurements during 1996 alone, perhaps a better comparison with the CMDL measurements during late 1996 and early 1997.

Fig. 2.5. Differences between CMDL determinations and Scripps total mean CO₂ measurements for 15 WMO primary standards. CMDL manometric determinations are shown as triangles, SIO total mean for each cylinder is shown as a square with the error bar.

Because we want to maximize the useful life span of the primaries, the calibration scale is transferred via NDIR measurements, approximately twice a year, to a set of secondary standards. These are in turn used to calibrate, again via NDIR comparisons, every other cylinder. During 1996 about 400 cylinders were assigned a WMO mole fraction value for CO₂ in this way. The reproducibility, in the range between 250 and 520 µmol mol^-1, has generally been better than 0.1 µmol mol^-1, and we expect the accuracy to be similar, pending a comparison with gravimetric mixtures. On request, calibrations can be performed with the manometric system well outside of the range of atmospheric CO₂ values. Cylinders prepared by CMDL with a specified CO₂ concentration have undergone the following procedures at our clean air pumping station at a high elevation on Niwot Ridge, Colorado, in the mountains east of Boulder:

New or recently hydrotested cylinder. (1) Cylinder is vented and then pressurized twice with dry natural air to about 20 atm (300 psi) and vented again. (2) Filled to about 34 atm (510 psi) with dry natural air and stored for several weeks. (3) Before the final fill the cylinder is first vented and then spiked with either 10% or zero CO₂-in-air, depending on the desired final mole fraction. (4) Final fill with ambient and dried natural air to 135 atm (2000 psi), during which the ambient CO₂ mole fraction is monitored. (5) Water vapor content of the filled cylinder is measured and must be less than 5 ppm; it is usually less than 1 ppm. Drying is accomplished through magnesium perchlorate Mg(ClO₄)₂. The pump is a RIX oil-less diving compressor.

[BACK] [CONTENTS] [NEXT]