Gravimetric air density determination gives lower uncertainties compared with conventional methods.

Figure 1: Stainless steel
working standard

The measurement of air density is critical for the dissemination of the mass scale. When comparing weights of different volumes a correction must be made for the buoyancy of the air in which the comparison is made. Conventionally, air density has been derived from measurements of temperature, pressure and humidity and calculated using the equation recommended by the Comité International des Poids et Mesures (CIPM) [1]. When comparing the national standard of mass, which is made of platinum-iridium, with working standards of stainless steel the calculation of air buoyancy is, by an order of magnitude, the largest source of uncertainty (Figure 1).

Figure 2: Air density
measurement artefact.

The National Physical Laboratory has developed artefacts for the direct measurement of air density (Figure 2). These artefacts can be used to measure air density with an uncertainty ten times lower than the parametric CIPM method.

Gravimetric air density measurement involves weighing a large volume artefact against a small volume artefact in the air you need the density of. The difference in weight between the two artefacts is directly related to the density of the air. Air density is calculated using the equation:

ρ_a = (ΔM_T - ΔW_A)/ΔV

where:

ΔM_T is the true mass of the large volume artefact less the mass of the small volume artefact, determined from weighing in vacuum.
ΔW_A is the weight of the large volume artefact less the weight of the small volume artefact in the air whose density is to be determined.
ΔV is the volume of the large artefact less the volume of the small artefact.

To minimise the effect of convection currents on the artefacts we determined the best shape of air density artefact NPL has modelled the airflow around different shaped artefacts (Figures 3 to 5). The results of the modelling have identified tube shapes as a better design for the small volume artefact than bobbin shapes.

Figure 3: Convection current modelling (cylinder).

Figure 4: Convection current modelling (tube).

Figure 5: Convection current modelling (bobbin).

For information on air density determination and buoyancy corrections go to the good practice guide ( PDF, 577 kB).

For more information on gravimetric air density measurement see NPL report ENG1.