350 Mr. C. E. Haselfoot .
[ June 26 , appears to rotate at its own proper speed , but in the reverse direction .
If both rotate in the same direction , but the object twice as fast as the prism , it does not appear to rotate at all .
I have examined the rotating soap film with this but have not obtained any striking results .
I have , however , observed the curious effect of angular error in position of the prism .
This causes a point in the axis of the motion to appear to describe a circle once for every turn and in the same direction .
If the prism were truly in position a point not in the axis would appear to turn twice for every turn and in the same direction .
If the prism is not truly placed a point not on the axis will appear to move with a combination of these movements , it will describe a trochoidal curve .
The Diffusion of Ions into Gases at Loiv Pressure .
By C. E. Haselfoot , Fellow of Hertford College , Oxford .
( Communicated by Prof. J. S. Townsend , F.R.S. Received June 26 , \#151 ; Read June 27 , 1912 .
) 1 .
In papers published in the ' Proceedings ' of the Royal Society , * ' the charges on ions produced by the action of Rontgen rays and radium on air * were determined by a method depending on the diffusion of the ions , and in the course of the investigations it was observed that the removal of water-vapour from the gas produced a large change in the motion of negative ions .
In this paper the results are given of some accurate experiments on the ions produced by radium , and experiments at low pressures on the motion of the ions produced by ultra-violet light are described .
In the latter case , effects similar to those observed by Prof. Townsend when the gas was ionised by Rontgen rays have been found , and some interesting results at pressures lower than those previously employed have been obtained .
2 .
The arrangement of the apparatus is here reproduced ( fig. 1 ) .
The ions are generated in the field A by radium placed in shallow horizontal grooves/ , covered with aluminium foil , in brass blocks F. They pass under the action of the electric force through the grating g and the aperture h into the field B , which was kept constant by means of the brass rings G maintained at definite potentials .
Here they diffuse and the ratio R of the charge received by the disc D to the charge received by the disc and the ring S together is measured .
This ratio is a known function of E7 ?
Z/ P , * Yol .
80 , p. 207 ; vol. 81 , p. 464 ; vol. 82 , p. 18 .
1912 .
] The Diffusion of Ions into Gases at Low Pressure .
351 where e is the charge on an ion , N the number of molecules in a cubic centimetre of air at pressure P at the temperature of the laboratory , and Z the electric force in the field B. -------G ' Field A Field B D 5 Fig. 1 .
The relation between R and c is shown graphically in fig. 2 , for the case where the aperture h is 7 cm .
from the disc D , the diameter of the aperture being 1*5 cm .
, and the diameters of the disc D and the ring S being 1*5 cm .
and 5 cm .
respectively .
From the value of R determined experimentally when an electric force Z is acting , the value of c is found by means of the curve , and No is deduced , being 3*042 x 108 x e/ Z , when N corresponds to a pressure of 760 mm. of mercury and Z is expressed in volts per centimetre .
Thus if No = 1*225 x 1010 the value of c corresponding to 1 volt per centimetre is approximately 40 .
The lower values of the ratio R in terms of c as calculated by the theory are given in the following table .
It will be noticed that as the effect of diffusion increases and c diminishes the ratio R falls but little below the value 0*2 , although the ratio of the area of the disc D to the area of the two electrodes D and S is 0*09 .
This is due to the effect of the rings G , which 2 b 2 352 Mr. C. E. Haselfoot . .
[ June 26 , are assumed , in the " theory , to discharge all ions that reach a distance of 2*5 cm .
from the central line of the field B. The density of ionisation near the edge of the ring S is therefore small and equal to zero at the ring G even when the lateral diffusion of the stream is very great .
c ... 2 4 8 12 16 20 24 28 32 R ... ... 0195 0-197 0-210 0-235 0-266 0-298 0-329 0-357 0-384 X ... ... o-i 0-323 0-661 0-831 0-915 0-953 0-975 0-986 0-992 Thus , as the stream spreads out , some of the ions reach the rings , and the 1912 .
] The Diffusion of Ions into Gases at Low Pressure .
353 proportion X of those that arrive at the electrodes to those that come through the aperture li may be calculated in terms of c , so that X is known when R has been determined experimentally .
3 .
The experiments on radium were made under the conditions stated in the previous paper as most favourable , viz. at pressures of 4 mm. and 7 mm. and with forces of 1 , 1*5 , and 2 volts per centimetre .
With these forces the ions do not diffuse beyond the ring , and the correction for self-repulsion , which depends on the force , the pressure , and the amount of ionisation , never exceeded 4 per cent. It is necessary to allow also for the effect due to ions generated by the emanation in the field B. In the case of negative ions it was found possible almost to eliminate this effect , but with positive ions it was seldom less than one-third of that due to the field A. When there was a very large amount of emanation present the values deduced for IsTe were generally larger than those obtained when the correction for the ionisation in B was small .
Of the experiments made , the seven which gave results closest to the mean were chosen , both with positive and negative ions .
The mean value of He deduced from these was in each case l-22 x 1010 , the value given by each experiment not differing from this result by more than 5 per cent. 4 .
The apparatus used for the ultra-violet light experiments was identical with that employed for radium , except that the blocks F were removed and a tube L closed by a quartz plate was inserted in the cover of the apparatus and so placed that the ultra-violet light after passing through it was incident on the zinc plate C ( fig. 1 ) , which forms the upper boundary of the field A. The rings G , G ' , were cut away in the path of the stream to avoid interference with the ultra-violet light and fine wires inserted to keep the potentials unchanged .
The ionisation was made uniform and of sufficient intensity by placing the spark-gap at the focus of a quartz lens which was mounted on a special frame , arranged so that the axis of the beam coincided with that of the tube .
The experiments have two advantages over those with Rontgen rays and radium .
In the first place , the ionisation due to the field B is always small ; and , secondly , the ionisation from the field A does not fall off as the pressure is reduced , so that observations may be made at low pressures ( and thus the correction for self-repulsion is reduced ) .
This method was therefore used especially to investigate the motion at very low pressures , when the abnormal increase of the rate of diffusion observed by Prof. Townsend in gases which have been thoroughly dried becomes very marked .
Large drying tubes were connected by wide tubing with the apparatus , and the value of R was determined at intervals .
It diminished gradually for about a month , after which Mr. C. E. Haselfoot .
[ June 26 , no further change was observed , and experiments were then made at pressures of 7 , 3*3 , 1*6 , and 08 mm. No correction was made for self-repulsion , as it is known that in dry gases the velocity of the ions is very large .
Preliminary experiments had shown that the ratio E given by the apparatus under ordinary conditions , when the gas had not been dried , was higher than that corresponding to the value l-23 x 1010 of Ne , the mean difference being about 5 per cent. This was probably due to an effect of contact potential , as it was observed when the apparatus was taken down that the ring and disc were unequally oxidised .
Allowance was made for this , and the results so corrected are shown in fig. 2 .
The dotted curves show the values of E for the different pressures .
The continuous curve represents the value of E when the gas is slightly moist and there is no abnormal diffusion .
It will be seen that there is a very large increase in the lateral diffusion ( E is diminished ) , becoming more marked as the pressure is diminished and the force increased .
The results may be compared with those found by Prof. Townsend , * who attributes the effect to an increase in the energy of the ions due to the electric force .
He points out that the large increase in the lateral diffusion shows that the mean kinetic energy of translation of the ions must exceed that of an equal number of molecules of the gas .
This effect occurs to a marked degree in gases which have been thoroughly dried , and increases as the electric force is increased and the pressure diminished .
The phenomenon is connected with the tendency of the ions to assume the electronic state , as it has been found that the prevalence of this state depends on exactly the same conditions as those under which the abnormal increase in the rate of diffusion occurs .
It is possible to deduce from the experiments an estimate of this increase of energy or , in other words , of the increase of the partial pressure of the ions .
It may be shown that if the partial pressure is increased in the ratio k : 1 , the effect on the ratio E is the same as if there had been no increase , but the electric force Z had been diminished in the ratio 1 : In the following table p is the pressure , Z the force in volts per centimetre , Zi the force , as deduced from the theoretical curve , which would give the same value of E when k \#151 ; 1 and Ne = P23 x 1010 , so that k is the ratio of Z to Zi .
In the apparatus used the radius of the ring E was 4'8 cm .
, it being necessary to have a space between it and the ring G , of inner radius 5 cm .
, which determined the field .
The values of found should probably therefore be slightly diminished , but this correction was not made as the other sources of error are considerable .
* ' Eoy .
Soc. Proc. , ' voh 81 , p. 468 .
1912 .
] The Diffusion of Ions into Gases at Low Pressure .
355 Dry Air .
p. Z. R. Zi .
Jc .
zip .
7 1 -03 0-296 0-488 2 -1 0-15 7 2-Q5 0-307 0-520 3 -9 0-29 7 4-13 0-342 0-646 6-4 0-59 7 6-07 0-357 0-695 8-7 0-87 3-3 1 -03 0-232 0-297 3-5 0-31 3-3 2-04 0-250 0-351 5-8 0-62 3-3 4-10 0-254 0-356 11 -5 1 -24 1 -6 1 *03 0-205 0-168 6-1 0-64 1-6 2-04 0-222 0-254 8-0 1 -3 1-6 4-10 0*206 0 171 24 2-6 0-8 1-03 0-205 0-168 6-1 \#187 ; 1 25 Partially Dried Air .
P- Z. R. Zi .
Jc .
Zip .
1 -03 1 -02 0-402 0-872 1 -2 1 1 -03 2 -04 0 -441 1 -08 1 -9 2 1 -03 4-04 0-266 0-406 10 39 0-48 1 -03 0*400 0-828 1 -2 2-15 0-48 2-04 0-357 0-693 2-9 4 -25 0-48 4-05 0-214 0-200 20 8-4 0-35 1 -03 0-374 0-752 1 -4 2-94 0-35 2-04 0-249 0-351 5-8 5 -83 0-26 1 -03 0-357 0-695 1 -5 4 0-26 2-04 0-219 0-235 8-7 7-8 0-26 4-09 0-211 0-196 21 15-7 Prof. Townsend points out that on his theory should be a function of Z/ p , and verified this in his experiments with Rontgen rays .
In the present experiments this is only approximately the case .
From the graphical representation ( fig. 3 ) , however , in which the values of k and are the co-ordinates , it appears that many of the points lie approximately on a straight line , meeting the axis of kat a distance 1 from the origin .
This would indicate that the factor k\#151 ; 1 by which the energy of translation of the ions is increased is proportional to the energy acquired by the ions as they move along their free paths under the electric force , since the latter quantity is proportional to Z]p .
Prof. Townsend 's experiments are indicated by the mark x. It should be noticed that the possibilities of error are very great when R is small .
In this case the charge received by the disc is small , and a considerable experimental error is possible , leading to a much greater error in k.For example , a 5-per-cent , error in R , when R is about 0-228 , might lead to a 20-per-cent , error in k. This cause of error is enormously increased when R approaches the value 0*2 , and little reliance 356 The Diffusion of Ions into Gases at Low Pressure .
can be placed on values of E smaller than 0:22 .
For this reason no experiments were made with perfectly dry air at a lower pressure than 08 mm. With air that has not been thoroughly dried , the pressure must be much reduced before these abnormally low values of E are reached .
Several experiments were made , and it was often found that E diminished as the force increased , that is , the lateral diffusion was actually greater with higher forces than with lower .
This is seen in fig. 3 , where a curve for k at a pressure of 1*03 mm. is shown with a dotted line .
It will be observed that the rate of increase of k with the force becomes very large as the force Effect of Magnetic Force on Motion of Negative Ions .
357 increases .
The values obtained , of course , depend on the degree of drying , so that observations for any one curve must be made on the same day .
A possible explanation of the diminution of R with X is that , in a gas which has been imperfectly dried , the number of ions which attain and continue in the electronic state increases more rapidly with the force than when the gas is perfectly dry .
Thus , with the larger values of X/ y\gt ; , the negative ions move as electrons in the gas even when some water vapour is present .
I wish , in conclusion , to express my thanks to Prof. Townsend , to whom I am greatly indebted for advice and assistance throughout .
Effect of a Magnetic Force on the Motion of Negative Ions in a Gas .
By J. S. Townsend , F.R.S. , M.A. , Wykeham Professor of Physics , Oxford , and H. T. Tizard , M.A. , Fellow of Oriel College , Oxford .
( Received June 26 , \#151 ; Read June 27 , 1912 .
) 1 .
When the velocity of a charged particle in a gas is proportional to the electric force and inversely proportional to the pressure , the size of the particle is unaltered either by changes in the pressure or in the force .
For a large range of pressures and forces the mass of an ion is thus shown to be constant , since the velocity is proportional to the ratio X/ P. At low pressures when the ratio X/ P exceeds a certain value the velocity of the negative ions undergoes large changes when small variations are made in the force or in the pressure .
The increase in the mobility may be explained on the hypothesis that the mass associated with the negative ion diminishes .
Thus in dry air at a pressure of 29 mm. the velocity of the negative ions is 926 cm .
per second , under a force of 2'3 volts per centimetre , * whereas if the ion travelled with the same mass that it has at atmospheric pressure the velocity would be about 114 cm .
per second .
Also under these conditions it may be shown that the velocity of a molecule of air with an atomic charge would be of the order 370 cm .
per second .
It thus appears that at low pressures the mass associated with the negative ions has an average value less than that of a molecule , so that along some of its free paths between collisions with molecules the ion moves as an electron .
It is of interest to make further investigations at lower pressures and to find to what extent the electronic state prevails as the pressure is reduced .
* R. T. Lattey , 'Roy .
Soc. Proc. , ' 1910 , A , vol. 84 , p. 173 .

