The mercury vapour pumping effect for 3 He, 4 He, H 2 , D 2 , Ne, Ar, Kr, Xe, N 2 , O 2 , CO 2 , CH 4 , C 2 H 4 , and C 2 H 6 : its precise measurement and convenient application in McLeod gauge metrology

Abstract

The McLeod gauge, an amplifying device for the mercury column, has been used since its inception in 1874 as the basic instrument for measurement of pressure or number density of gases at pressures below 1 Torr (1 Torr ≈ 133.32 Pa). In practically all cases for direct pressure measurement or for use in calibration of other gauges it must be used with a trap to remove mercury. In the 1960s it became generally known that the movement of mercury vapour from the gauge toward the trap increases the permanent gas pressure beyond the trap above that at the McLeod gauge, and that if this is overlooked the pressures assigned to the working region under common laboratory conditions may be in gross error, of a magnitude of up to hundreds of times the probable error of reading the gauge. We have found it quite impossible to gather from the literature the means for accurate correction of the McLeod pressure readings adequate for use in our clean surface thermal accommodation work, so have undertaken much more extensive sets of measurements than have previously been available. These are described in this paper. A new apparatus and method has been applied under carefully controlled conditions to supply 250 mercury pumping pressure measure­ments in 27 sets involving 14 gases over relevant ranges of pressure, temperature, and pumping tube radius and length. A semi-empirical equation has been developed for calculating the mercury pumping pressures for any set of conditions and has been thoroughly tested over the ranges and the gases used against our measurements, and comparison of results is made with other selected papers. The result of the work for us has been renewed confidence in the ability of the McLeod-gauge—cold-trap system to render precisely known pressures and, when applied in our measurements of thermal accommodation coefficients, it constitutes a ‘break-through’ in the accuracy and consis­tency of the measurements. Because of this, the paper is presented here as a necessary preliminary to two experimental papers on clean surface accommodation coefficients, which we intend to offer shortly in honour of and as an extension of the work of J. K. Roberts (1930), who first emphasized the importance of controlling surface conditions and first undertook bare surface measurements in this area. Our expression of appreciation for the work of Roberts belongs with these papers to follow.