]\gt ; Tlve Charges By JOHN S. TOWNSEND , , Wykeham Professor of Physics , Oxford .
( Received January 20 , \mdash ; Read January 23 , 1908 .
) In a paper on the Diffusion of Ions in Gases , described a method of comparing the charges on the ions generated in gases with ) on an ion in a liquid electrolyte .
If be the number of molecules in a cubic centimetre of a gas at standard pressure and temperature , aJJd e the on an ion , then N. , where is the velocity of an ion in a field of unit electric force and the coefficient of diffusion .
Thus from separate determinations of the quantities and , the product N. can be obtained .
If is the charge on a monovalent ion in a liquid electrolyte , N. , so that a comparison of the various may be made , and the calculations have shown that the charge on a positive or negative ion in a gas is equal to the charge E. There were , however , considerable discrepancies , particularly with positive ions , which gave of as great as in some cases .
Some time ago , while examining these investigations , I found that , by a somewhat erent method , it would be possible to arrange an experiment for determining N. directly , and to arrive at much more accurate than have been obtained hitherto .
In order to apply the principle , it is necessary to have the ions in a uniform field , and to find experimentally how they are ibuted in some particular case .
For this purpose two horizontal plates , A and , were set at a distance apart , and circular holes of the same radius were cut the plates , the centres of the circles being in the same vertical line .
of very fine wire was laid across the hole in the upper plate , and a metal disc on insulating support was fixed in the aperture of the lower plate , leaving a small air space between the disc and the plate , so that the should be insulated from each other . .
( / * John .
Townsend , ' Phil. Trans , 1899 , vol. 193 .
Prof. J. S. Townsend .
The on [ Jan. 20 , The plate was maintained at a potential , and a third plate above was maintained at a potential greater than and of the same sign .
The lower plate and the disc were connected alternately to a sensitive electrometer , and during the course of an experiment their potentials did nob differ appreciably from the zero .
The ions were generated in the air space between and by rays , and positive or negative ions were driven to the grating according as the potential of was positive or negative .
The electric force being in the same direction above and below the grating , some of the ions pass through and travel under a constant electric force towards the central disc D. At the same time the ions diffuse laterally , so that some arrive on the disc and some on the plate .
The greater the force the less time there will be for the diffusion , so the proportion of the total number that arrive at the disc increases with the force between A and B. The ratio of the number arriving at the disc to the total number was determined accurately with the aid of the electrometer .
The mathematical investigation of the distribution of the ions between the plates A and can easily be made by considering the equations for the variation of the partial pressure of the ions .
The three equations corresponding to the three ular axes are of the form .
X. where is the number of ions ) cubic centimetre ( which is proportional to , the velocity and X the electric force along the axis of Differentiating the three equations and in the equation of continuity for the steady state , the following equation for is obtained\mdash ; where is the pressure due to 760 mm. of mercury , at which pressure is reckoned .
This latter equation that at any point is a function of The constants in the solution are functions of the position of the point and at the lower plate , where , the ratio may be obtained by integration .
Thus , where is the distance from the centre of the disc , and being the radii of disc and the plate respectively .
1908 .
] Positive and Ions in Gases .
Hence It is not necessary to know the form of the function in order to see how to colupare the charge on a positive ion with that on a negative ion , or to compare the charges on ions in difierent gases .
Experiments made with positive and negative ions for the same intensity of force gave a value of for positive ions which was much greater than he value for ative ions .
This shows that the charge on a posiGive ion is reater than the charge on a negative ion .
The same value of , within 1 per cent. , was obtained for the two kinds of ions when the force used with ative ions was double that used with positive ions .
Letting the charge on a positive ion and the charge on a ative ion , then for all values of Z. Hence the charge on the positive ion is exactly double the char , on the negative ion .
This shows that when a molecule of air is ionised by Bontgen rays , one positive ion and two negative ions are produced , the charge on the former being double that on either of the latter .
In order to find the absolute value of N. , it is necessary to solve the differential equation in .
An exact solution can easily be obtained which satisfies all the surface conditions , since can be expanded in a series of Bessel 's functions .
The distribution being symmetrical round the axis of joining the centre of the with the centre of the disc , the differential equation reduces to the form\mdash ; where the distance of a point from the axis .
Letting the equation for is so that , and The values of and are determiued from the surface conditions and is positive root of the equation Hence can be found at any point of the lower plate , where , and VOL. LXXX.\mdash ; A. Prof J. S. Townsend .
The on [ Jan. 20 , the integrations and can easily be obtained from the known properties of Bessel 's functions .
The connection between and N. was thus found , and a curve was drawn to show the values of for a series of values of N. .
When is found experimentally the corresponding value of N. ( : .
can be found immediately from the curve .
The expression for in terms of N. is very complicated and I have to acknowledge my indebtedness to Mr. C. E. Haselfoot in having assisted in making the necessary calculations from the tables of Bessel 's functions .
I have made a number of determinations of with different forces , pressures , and intensities of ionisation , and they all agree in giving results in accurate accordance with the theory .
The values found for Ne are for positive ions and for ative ions .
Each of these numbers represents the mean result of several different experiments in which none of the determinations differ from the mean value by more than 4 or 5 per cent. , which shows that considerable accuracy can be obtained by this method .
When the value of N. is deduced by the first method which I gave , numbers corresponding to the positive ions were about 10 or 20 per higher those for the negative ions .
Thus , when the rates of diffusion are compared with the velocities given by Zeleny , the values of N. for positive ions in air , oxygen , and hydrogen are , and , and the numbers for the negative ions were These numbers taken alone would not justify the conclusion that the molecules are ionised in the way that I have just found , but a little consideration shows that the relative values are of the kind that should be expected .
If the effect of recombination be considered , it will be seen that a positive ion would rapidly combine with a negative ion when the charge on the former is 2 , as the force between them is .
After combining with one negative ion , the force between the positive ion and another negative ion would be , so that the second step towards complete recombination would not proceed so rapidly as the first .
The result of this would be that after the ) ositive and negative ions have been in the gas together for a short time , a large proportion of the positiye ions would have one-half of their original charge .
The determinations of the coefficients of diffusion were made with small ionisations that had been reduced by the process of recombination , and from a consideration of Zeleny 's method of finding the velocities it is evident * J. S. Townsend , .
cit. 'Phil .
Trans , 1900 , vol. 196 .
1908 .
] Positive and Negative that recombination must have been on in his experiments also .
Under these circumstances the numbers obtained for N. by the method correspond to positive ions , of which some have a charge and others a charge .
The discrepancies from the value obtained for the negative ions are not greater than what might have arisen from experimental errors .
In the experiments I have just now made the ination must have been ] ible , so that no appreciable number of the positive could have been reduced from to A full description of the experimental methods used in these determinations will be given in a future paper , when the experiments which are in ress with different gases are completed .

