No abstract

Articles you may be interested inNew spiral molecular drag stage design for high compression ratio, compact turbomolecular-drag pumps J. Vac. Sci. Technol. A 28, 931 (2010); 10.1116/1.3386591 General geometry calculations of onestage molecular flow transmission probabilities for turbomolecular pumps J. Vac. Sci. Technol. A 13, 2954 (1995); 10.1116/1.579621Comparison between several new molecular drag pumps and turbomolecular pumps Conventional turbomolecu1ar pumps consist of a succession of rotating and stationary bladed disks. In order to increase compression ratios, the open angle of the disks is decreased from the upper to the lower stages of the pump: a compression ratio of > 10 7 can be obtained for nitrogen in nine stages. However, the maximum compression ratios are obtained only when the pump is working in free molecular flow conditions, corresponding to a foreline pressure in the range 10 -3_10 -2 mbar; at higher pressures the compression ratio decreases dramatically thus limiting the total efficiency of the pump. Molecular drag pumps, on the contrary, can work with high compression ratios also in the transition flow region, in the range from 10 -2 to 10 mbar. Combination of turbomolecular pumping stages and drag stages have been tested. Molecular drag stages of a new design have been incorporated in a conventional turbomolecular pump without adding to its dimensions. From these tests it appears possible to have high compression ratios with the foreline pressures up to 10 mbar: the forevacuum pump could be a conventional, double-stage rotary pump or a less conventional diaphragm oil-free pump. The behavior for different gases has been analyzed and advantages of this arrangement are examined.

Modern turbomolecular pumps include a drag stage in the exhaust, operating roughly in the pressure range of 10 mTorr–10 Torr. Flow conditions range from molecular flow at the drag inlet, to viscous flow at the outlet, known as ‘‘transition’’ flow. In general, models of transition flow in drag pumps have not been developed. Moreover, the model of a Gaede pump given in journals and textbooks up to the present, gives values of compression ratio that are orders of magnitude too high. In 1913, Gaede proposed a differential equation for transition flow in a drag pump. He did not solve the general equation, and the model was incomplete. We have developed a new model that takes transition flow in a differential element and integrates it over the length of the pump. This model is modified by a ‘‘pumping leak’’ expression for the gas stripper, which separates the inlet from the outlet. The result is compared with experimental measurements, and good agreement is obtained over the entire pressure range from molecular, through transition, and into viscous flow. Up to a critical pressure in viscous flow, compression ratio is constant as a function of exhaust pressure, within a factor of 2. Within this factor, increasing compression arises from the reduced pressure drop across the inlet aperture as its conductance increases in the transition flow regime. Above the critical pressure, compression drops rapidly as laminar backflow increases. This critical pressure is controlled by the dimensions of the channel. Below the critical pressure, compression is determined by the pumping leak, and is somewhat independent of molecular weight. If the surface velocity is zero, the model reduces to a capillary leak. Predictions of our model agree with Knudsen’s data for capillary leaks in transition flow, in addition to giving a better account of the ‘‘conductance minimum.’’ ‘‘Slip flow’’ is not an obvious factor, and it cannot be distinguished from the right combination of viscous and molecular flow.

No abstract

The history of the method of differential probability in molecular flow is reviewed, beginning with the little known derivation by D. Santeler (5th Annual Symposium on Space Environmental Simulation, Arnold Air Force Station, TN, May, 1964), based on the equation of C. W. Oatley [Br. J. Appl. Phys. 8, 15 (1957)]. This method contains the aperture correction within the theory, without phenomenological assumptions. A new equation of this type, for molecular pumping, is derived by differential reduction of the Kruger–Shapiro equations. A simple solution of the differential equations yields results of good accuracy for engineering use. The physical characteristics of molecular pumping are clarified by describing the pressure distribution within the pumping tube as if it were a conductance. By this method the calculated performance of a model pump is shown to be in satisfactory agreement with a Clausing-type solution from a previous publication.