Novelli, P. C., J. W. Elkins and L. P. Steele, (1991), The Development and Evaluation of a Gravimetric Reference Scale For Measurements of Atmospheric Carbon Monoxide, Journal of Geophysical Research-Atmospheres, 96, D7, 13109-13121, JD01108


We have prepared a set of 17 carbon monoxide (CO) reference mixtures for use in the calibration of measurements of atmospheric concentrations of this gas. The mixing ratios of these standards range from 25 to 1003 ppb (parts per billion by mole fraction) in zero natural air and are contained in 5.9-L, high-pressure aluminum cylinders. Carbon monoxide was measured using gas chromatography with a mercuric oxide detector. The concentration range of the standards is sufficient to cover that of the background troposphere and also that found in remote locations affected by anthropogenic activities. The low concentration standards were prepared by gravimetric methods using one of three high concentration standards as the parent. Two of the parents were prepared by gravimetric methods starting from high-purity (99.97%) CO to have concentrations of about 250 ppm (parts per million) CO. A total of 14 atmospheric level primary standards were prepared from these two parents. The third parent was a NIST SRM (National Institute of Standards and Technology, Standard Reference Material) having 9.7 ppm CO, from which three standards were prepared. Monitoring of CO levels in the primary standards, relative to natural air contained in 29.5-L high-pressure aluminum cylinders, suggests that the CO content of some primaries may be increasing at rates of between 1 and 2 ppb yr?1. The CO concentration scale defined by the gravimetric standards was used to calibrate a set of 10 secondary standards. The secondary standards are all contained in 29.5-L high-pressure aluminum cylinders and range in concentration from 35 to 200 ppb. Examination of the CO content in several of the oldest secondary standards indicates their CO concentrations have not changed relative to each other over the 2 years they have been studied. Comparison of the low concentration standards derived from the gravimetric parents to those prepared from the NIST SRM show no difference to within 1% between the two scales. We also compared our standards to commercially available, NIST-traceable, CO standards (approximately 0.5 and 1 ppm of CO in air). The concentrations assigned these standards by the manufacturer agreed to within 3% of concentrations we calculated referenced to our standard scale. In addition, we compared our concentration scale to a CO standard used at the Commonwealth Scientific and Industrial Research Organization (CSIRO), Australia. Intercomparison of a cylinder of natural air between our laboratory and CSIRO (which used a CO reference gas traceable to the standards of the Oregon Graduate Institute for Science and Technology, formerly the Oregon Graduate Center) indicated that the CO concentration determined for this air based upon our reference scale was approximately 25% greater than the concentration determined by CSIRO. Carbon monoxide concentrations determined in flask samples collected at Mauna Loa, Hawaii, referenced to this concentration scale, are compared to the earlier reports of CO levels at this location by Seiler et al. [1976] and Khalil and Rasmussen [1988].