574 



NATURE 



[January 23, 1913 



MODERN PUMPS FOR HIGH VACUA. 



THE widespread researches on the phenomena 

 in electrical discharge tubes, which form so 

 important a feature of modern physics, directed 

 much attention to the question of obtaining' high 

 vacua. In 1888, as Lenard tells us,i an efficient 

 vacuum pump was by no means an essential part 

 of the equipment of a physical laboratory : at the 

 present time it emphatically is so. In the follow- 

 ing a brief account will be given of the modern 

 forms of the different types of pumps, especial 

 reference being made, however, to a pump recently 

 invented by Dr. Gaede, as it depends on a prin- 

 ciple never before applied, and seems from present 

 information more efficient than any of its prede- 

 cessors. 



All vacuum pumps except this latest one of 

 Gaede 's make use of the principle employed 

 by Otto von Guericke in the first air-pump — 

 that is, the intermittent separation and discharge 

 of a fraction of the gas from the reservoir to be 

 exhausted by means of a piston, which in the 

 mercury pumps takes a liquid form. We can, 

 in reviewing the modern forms, divide these pumps 

 into three classes : the solid piston pump, the hand 

 mercury pump, and the automatic mercury pump. 



The solid piston pump has preserved much of 

 its original arrangement of valves, but has been 

 modified in the Geryk pump, which may be taken 

 as a modern example, by the use of layers of 

 a particular oil in the place of packing. The 

 valves are always covered by the oil, which takes 

 up all clearance, and hence leakage is largely 

 avoided, but the vapour pressure of the oil, 

 though very small, prevents the highest vacuum 

 being produced ; however, o'ooo2 mm. of mercury 

 can be attained. In a still more recent pattern, 

 the " Rose " pump manufactured by Messrs. 

 Cosser, there is no piston rod, the piston being 

 of iron and moved by electro-magnets oscillating 

 outside the pump cylinder. 



The forms of hand mercury pump now used 

 are all modifications of the well-known Toepler 

 pump. One of the simplest and most successful 

 is that devised by Antropoff, in which the usual 

 bulb is replaced by one of cylindrical form 

 arranged obliquely instead of vertically. 



The desire to reduce the time and labour attach- 

 ing to the hand pump has led to the construction 

 of a large number of mercury pumps which can 

 be operated mechanically ; in experiments such as 

 those of Prof. Wien on canal rays such a con- 

 tinuously running pump is a necessity. The most 

 convenient of these are the various rotary pumps, 

 of which the first was devised by Schulze-Berge, 

 and of which Kaufmann in 1905 brought out a 

 pattern which has been considerably used. The 

 essential of this is an inclined spiral tube which 

 rotates continuously ; a thread of mercury running 

 in it cuts off and forces out a fraction of the air 

 at every rol.-ition. There are two such tubes; the 

 pump, though efficient, is somewhat fragile and 

 complicated. 



' Nnlwl d^coiirve. 1006, p. 3. 



NO. 2256, VOL. 90] 



The rotary mercury pump most in use at the 

 present time is that of Dr. Gaede. It consists of 

 an outer closed drum half filled with mercury, 

 in which a second drum rotates. This drum is 

 divided into chambers, which in turn become con- 

 nected to the vessel to be exhausted ; by the rota- 

 tion they are filled alternately with gas and 

 mercury, the gas being displaced into the outer 

 space between the two drums and cut off from 

 return by the mercury. The system is similar 

 to the gas meter, only in this the moving gas 

 effects the rotation, while in the Gaede pump the 

 rotation sets the gas in motion. With this pump 

 the pressure must first be reduced to a few milli- 

 metres of mercury by any rough preliminary pump, 

 as otherwise the difference of pressure between 

 the outside and inside of the rotating drum will 

 become sufficient to drive the gas back into the 

 drum again. 



In the past year, however, Dr. Gaede described 

 an air-pump depending on a new principle, which 

 he calls the molecular air-pump. Maxwell 

 assumed, and Knudsen has recently verified ex- 

 perimentally, that if a gas be in contact with a 

 solid surface, the gas molecules are reflected from 

 it in all directions independently of the angle of 

 incidence, or "diffusely reflected." This is due to 

 molecular irregularities of the surface. Gaede 

 has shown that 

 for pressures 

 above o'ooi mm. 

 of mercury the 

 above assump- 

 tion is not ex- 

 perimentally veri- 

 fied, and he 

 attributes this to 

 the formation of 

 a film of ad- 

 sorbed gas on 

 the solid surface, 

 which covers and 

 conceals the 

 molecular irregu- y^^ _„ 

 larities. The 

 surface then presents only 

 larities, and the result is 

 travelling over a surface 

 preferentially thrown back in the direction from 

 which they came, as they fall in general on 

 small slopes of the irregularities facing their 

 direction of drift. In both this case and that of 

 diffuse reflection the new pump is effective, 

 but the point is of interest in considering 

 the theory of the pump, and it was considera- 

 tions of this kind which led Gaede to its 

 construction. 



The new pump depends for its action on the 

 dragging of the gas by a rapidly moving surface.^ 

 Consider a cylinder A rotating in a clockwise direc- 

 tion in a case B ; \n B there are two openings n 

 and m connected by a slot (Fig. i). The gas will 

 be dragged by the cylinder from n to m, and in 



- For the illustrations which accompany this article, we are indebted to 

 the makers of the new pump— E. Leybold's Nachfolger, COln. 



:iple of molecular air pump. 



mechanical irregu- 

 that if a gas be 

 the molecules are 



