Search facility:  
BIPM Home | Site map | Metrologia | KCDB | JCTLM-DB | Contact us  
New measurement of the argon content of air
Ozone reference standards
International comparisons of ozone standards (BIPM.QM-K1)
Calibration of national ozone standards
Ozone cross-section measurements in the Hartley Band
Ozone cross-section measurements facility
UV laser ozone photometer
Gas phase titration
International comparisons of nitrogen dioxide gas standards
International comparisons of nitrogen monoxide gas standards
International comparisons of formaldehyde in nitrogen standards
New measurement of the argon content of air
Related articles
Mass comparisons using air buoyancy artefacts
Equation for the determination of the density of moist air (1981/91)
Direct access

Park S.Y., Kim J.S., Lee J.B., Esler M.B., Davis R.S., Wielgosz R.I., Metrologia, 2004, 41(6), 387-395

Picard A., Fang H., Gläser M., Metrologia, 41(6), 396-400

If asked to name the major chemical components of air, most of us would list oxygen (about 21 %), carbon dioxide (about 0.04 %) and water vapour (typically about 1 %). The principal component of air is nitrogen and the only other major component is argon (about 0.9 %). Argon is chemically inert and its presence in the atmosphere poses no problem to human well being. Old measurements dating from as early as 1903 gave the mole fraction (moles of argon/mole of dry air) as 0.934 %. The most recent value available until now was lower (0.917 %) and was thought to supersede the previous result. It is therefore surprising that the work reported in Metrologia gives (0.9332 plus-minus 0.0006) %, very close to the measurement results of 100 years ago. The uncertainty in the new measurement is given at the 95 % confidence limit and is unprecedented.

Figure 1. The gas analysis laboratory in Korea where the argon analysis was performed, using mass spectrometry.

Figure 2. Calibration gas mixtures for the argon analysis being prepared by careful weighing (gravimetry).
The analysis was performed at KRISS using high precision mass spectrometry (Figure 1). A set of air-like calibration gas mixtures was prepared (Figure 2) by very careful weighing of pure gases into high pressure cylinders. Analysis of these synthetic air mixtures along with samples of natural air contained in other high pressure cylinders yielded the result reported in Metrologia.

The argon mole fraction is important to a small community of scientists working on precision mass measurements. To understand why, think of the old puzzle: which weighs more, a kilogram of feathers or a kilogram of lead? If it were possible to do the weighing on a very precise balance, we would see that the balance readings are identical for the feathers and the lead if the weighing is carried out in a vacuum. But the feathers would produce a considerably lower balance reading for measurements in air. This is because feathers are more buoyant in air than is lead (Achimedes' Principle). Mass metrologists use an equation to correct for the effect of air buoyancy. The equation includes the air density which, in turn, includes a parameter for the mole fraction of argon in the atmosphere. The different historical values for argon mole fraction lead to a difference in air density of just under 0.01 %, or about 15 micrograms in the apparent mass of one kilogram made of stainless steel (15 parts in 109). The higher the argon mole fraction, the denser the air.

Even though the density of air is roughly 800 times smaller than water density, the effects of air buoyancy are easily seen in precise weighing. Thus the air density calculated from the new value of argon mole fraction should agree with precise data obtained from the feathers and lead experiment.

Figure 3. Test objects loaded onto the automated BIPM flexure-strip balance. The balance can operate either in vacuum or in air.

Figure 3 shows the two objects being weighed. The stainless-steel cylinder on the left is actually hollow inside, thereby representing the low-density feathers. The thick-walled tube on the right is a solid piece of stainless steel, thereby representing the high-density lead. The cylinder and tube have the same surface area, which simplifies analysis of the experimental data. The results of measurements with several different sets of hollow and solid objects are reported in a companion article in Metrologia, written by scientists at the BIPM and the Physikalisch-Technische Bundesanstalt, Germany. The results confirm the new argon mole fraction and thus can explain discrepancies that had already been observed using the previously-accepted value dating from the mid-20th century. Indeed, the new determination of argon mole fraction was motivated by numerous mass measurements which stubbornly failed to agree with the accepted formula for air density. The new results should lead to improved coherence among high-precision mass measurements.

Related articles

Mass comparisons using air buoyancy artefacts
Equation for the determination of the density of moist air (1981/91)