|
The principle of a continuous gas expansion is widely used in high vacuum standards. The number density of the gas, n, in the calibration chamber, is calculated with the following formula: dN/dt=Sefn
where the gas flux dN/dt is measured.
There are two ways of determining the effective pumping speed Sef. Traditionally Sef =Cβ is determined with the conductance C-calculated, and the β-factor (that includes the backstreaming effect) measured. In this model the conductance is calculated assuming a uniform and cosine gas flux distribution over the orifice surface. It is also assumed that there are uniform distributions of the gas flux along the calibration and pumping chamber walls.
The new, so-called ‘global’ model permits us to determine the molecule's paths from the point where the molecules appear in the calibration chamber to their outflow from the pumping chamber and their disappearance in the pump. A description of the boundary conditions includes the shapes and dimensions of all the segments of the vacuum standard and also the process of scattering of gas molecules on the chamber walls. On this base the effective pumping speed is evaluated as Sefg=vaVg/Lg, where va is the mean arithmetic velocity, Vg is the gauge volume, Lg is the mean molecule path length inside this volume.
The paper presents results of analyses of the high vacuum standard obtained with the use of Direct Sampling Monte Carlo method (DSMC) where the both models were applied. It was found that the gas flux density distribution over the orifice plane is not uniform and its angular distribution also is not strictly cosine.
Also the distributions of the gas flux along the calibration and pumping chamber walls are not uniform. These effects cause that the effective pumping speed calculated with both models differs and the relative difference equals 4×10-4. Additionally it was found out that the uncertainty of the vacuum standard (due to the orifice dimensions measurements imprecision) computed using the global model is twice smaller. |