]\gt ; Recombination of Ions at Different ) By P. PHILLIPS , D.Sc .
, B.A. , Royal Veterinary , Camden Town .
( Communicated by Prof. F. T. outon , F.R.S. Received November 10 , 1909 , \mdash ; Read January 13 , 1910 .
) Doubts were felt about the results of McClung 's experiments on this subject very soon after he published them , and Langevin indicated the weak point in them by showing that , in the apparatus used , diffusion would have a large effect and would be responsible at the higher temperatures for most of the disappearance of the ions .
The present experiment was therefore carried out under such conditions that the diffusion could be safely ected .
At highest temperature attained it will be seen that the diffusion is probably not quite negligible , but the error even here is certainly not a very large fraction of the whole .
Theory of the Method .
Alayer of air of uniform thickness situated between two parallel electrodes is ionised by a single flash from a Bontgen bulb .
The quantity of electricity received by either electrode will depend upon the field established between them .
It is easy to show that where is number of ions received per unit area of the electrode ; and are the velocities in unit field of the positive and negative ions respectively ; X is the field ; is the coefficient of recombination ; is number of ions per cubic centimetre originally produced by the flash ; is thickness of the layer of air ionised .
If , therefore , the number of ions received per square centimetre is when the field is and when the field is , then .
This experiment measures , a deflection which is proportional to , and the differences of potential and between the electrodes which are proportional to the fields and Let and and respectively , and being constants .
of Ions at Different Temperatures .
Throughout this experiment volts an volts and therefore the equation becomes .
will be constant if the apparatus is unmoved and therefore put Then .
First and are determined at the temperature of the laboratory where both and are well known .
The constant can , therefore , be calculated from the equation .
The same determination is repeated at the yher temperature and , using the calculated value of is calculated from the equation .
Since this experiment only determines the values of relative to its value at the temperature of the laboratory , this last value is taken as unity .
Apparatus.\mdash ; The general arrangement is shown in the figure .
The Rontgen bulb , X , sends a single flash of rays through windows cut in the lead screen , , and ionises a layer of air between each of the two pairs of electrodes , A and B. Slits placed in front of A and regulate the thickness of the ionised layer .
The pair of electrodes , , is simply an adjustable standard .
It serves to correct for irregularities in the through the bulb , and to balance or nearly balance the effect in , so that small differences in charge in A may be determined more accurately .
One electrode is maintained at a constant positive potential by means of cells , and the other may be earthed , insulated , or connected to the electrometer .
By means of a movable shield some of the rays may be cut off from , so Dr. P. Phillips .
[ Nov. 10 , that it may be adjusted to exactly balance the effect of A. In order to keep the temperature constant the electrodes are contained in a double-walled vessel through which tap water may circulate .
The pair of electrodes A is the real working pair .
One electrode is maintained at any desired negative potential by means of cells , and the other may be earthed , insulated , or connected to the electrometer .
The charge received is evidently of opposite sign to that received in B. The electrodes are contained in a double-walled vessel through which the vapour from a boiling liquid may circulate .
In order that the vessel may be used at high temperatures , all the joints are brazed and all the insulation is quartz .
In order to vary the sensitiveness of the electrometer at will , A and have each a variable capacity , , attached to them .
These are air condensers , of which each electrode consists of 10 parallel plates .
One electrode may slide up to the other so that the plates are interleaved and the capacity is large , or it may slide quite clear so that the capacity is small .
The electrometer is of the Dolezalek pattern , with a fine platinum suspension .
As used , the sensitiveness is about 100 whole divisions for 1 volt .
Readings talcen .
T. Ratio of the of the stem.\mdash ; We must first know the ratio etween the capacities of the system in the various conditions used .
These conditions Variable capacities , , both open , and A and connected ether to the electrometer .
( b ) A alone connected to the electrometer with its variable capacity closed .
( c ) alone connected to the electrometer with its variable capacity closed .
By putting a stop on the variable capacity connected to , this third capacity was adjusted to equality with the second , and by sharing a charge on the system with the same Leyden jar , the ratio of the capacity in the case ( a ) to that in the case ( b ) or ( c ) was found to be II .
Deterrnination of \mdash ; One electrode in is at volts throughout the series of readings .
( a ) One electrode in A is at volts , and with the variable capacities open both it and are connected to the elecbrometer .
The shield in front of is then adjusted until there is no deflection , when a flash passes through the bulb .
This means that the charge received by is equal and opposite to that received by A when its potential is volts .
( b ) A is raised to the highest available negative potential to volts ) and the variable capacities are both closed .
A single flash passes in the bulb .
and A and are separately co1mected to the electrometer .
Call the deflection when is connected the first deflection , and when 1909 .
Recombination of Ions Different Ternperatures .
A is connected the second deflection .
second deflection cvives if we may assume that there is no recombination with this large field .
( c ) A is charged to volts , the capacities are both closed , and the first and second deflections are obtained as in ( b ) .
These two deflections should be equal , and each should be proportional to ( d ) A is still to 100 volts , and both capacities are closed .
After the flash through the bulb , is connected to the electrometer giving the first deflection .
Then without disconnecting , A is also connected to the electrometer and the capacities are opened .
This gives the third deflection , which should be zero if the first adjustment was accurate .
( e ) A is charged to volts , and the process in ( d ) is repeated .
The third deflection here is proportional to third deflection will evidently be this , for the capacity when the third deflection is measured is time the capacity when the first deflection is measured .
We thus have and , and we know and at the temperature of the laboratory .
We can therefore calculate The series of readings is then completely repeated with A at an arranged higher temperature .
At the higher temperature we can calculate , using the value of which we have determined .
At some of the higher temperatures has been determined , at these temperatures , therefore , we can calculate One source of trouble in the experiment is the irregularity of the bulb .
In spite of trying several forms of contact breaker and putting a large capacity across the break , the variability was never quite suppressed .
Usually the bulb would give a whole series of uniform flashes and then some ular ones .
When the bulb was variable , and the first deflection therefore inconstant , a number of readings was taken under each of the headings , ( b ) , , ( d ) , .
The first and second and the first and third deflections were then plotted ainst one another , and from these curves the values of the second and third deflections for a particular value of the first deflection were read off .
The following is a series of taken on Apri129 , 1908:\mdash ; P. Phillips , " " Ionic Velocity at Different Temperatures ' Roy .
Soc. Proc , 1906 .
Dr. P. Phillips .
[ Nov. 10 , Temperature C. is taken as unity at cm .
per sec. per volt per cm .
; therefore .
The values of the function were calculated for different values of A and a curve was drawn from these values .
The value for the function is given by a value for ; hence so that Temperature with vatel in the boiler .
Thus ; whence , from the curve , so that 1909 .
] Recombination of Ions at Different After a series of with water in the boiler , anilin was next used .
At first the circulation of the anilin vapour was not fast enough , and th6 temperature only rose to C. As this remained fairly constant , however , several determinations were made at this temperature .
On a larger gas burner the temperature rose to its proper value and readings were taken at this temperature .
Finally , the boiler was filled with a naphthalene monobromide , and a temperature of C. was attained .
The boiling point of this substance gradually fell on prolonged , due evidently to decomposing a little .
In the first determination the temperature was C. , and in the last 27 C. , and therefore the mean temperature was taken as the temperature throughout .
The following are the values obtained at the different temperatures for From the curve given in the author 's paper mentioned above , at 10 is highest temperature attained in the experiment described in that paper was 13 C. , and therefore one must extrapolate in order to get the values of at higher temperatures .
These values are at at , and at The first two extrapolations are small , but the last is too large to be safe .
Using these values for , we get the following values for Beside the extrapolation being very large for the last temperature , it will be seen later that at this temperature diffusion probably begins to make itself felt , and in consequence is probably considerably too low a value .
from the curves given below , the author would make a rough estimate that it is about 30 per cent. below its proper value , i.e. that is about Dr. P. Phillips .
[ Nov. 10 , The curve showing the relation between and the temperature shows that the recombination decreases somewhat rapidly with rise of temperature , but less rapidly as the temperature rises .
L. L. Hendren*showed experimentally that in the conditions employed by him diffusion was negligible , and from his conditions it should also be negligible here .
In order to test this the width of the slit in front of A was reduced , and thus the layer of air ionised was made thinner .
Under these conditions diffusion would be more effective and would apparently decrease the value of was therefore determined with different widths of slit , and the value of was plotted against the width of the slit .
The following values were obtained at 10 C. :\mdash ; At C. the deflections were so small when the width of the slit was reduced that it was not possible to measure , and so the test could not be carried out at this temperature .
* Phys. Review , ' November , 1905 .
1909 .
] Recombination of Ions Temperatures .
The curves make it evident that at 10 C. we are safely within the limits where diffusion is negligible .
At we are evidently just within the limits , and therefore we are probably outside the limits at 27 C. , and the value of will be too low .
just published a paper on the same subject as this paper , but he uses an entirely different method .
The results , as far as they may be compared , are nearly the same ; but the comparison cannot be made accurately , as these experiments were carried out at constant pressure , while his were at constant density .
The variation of with density only been determined at room temperatures , and therefore we cannot correct for the difference in density at the temperatures .
If Hendren 's results ( in the paper before mentioned ) could be extended to higher temperatures , i.e. if varies with the density in the same wa .
at the higher tennperatures , then Erikson 's value for at C. would be 10 per cent. below the value found by this experiment .
It is difficult to estimate what would be the effect of the lack of 'uniformity of the ionisation in his experiment , but it might account for the difference .
In conclusion , I wish to thank Prof. Trouton and Prof. Porter for their kind encouragement and valuable suggestions during the ress of the experiment which was carried out at University College , and at the same time to acknowledge the Government grant which defrayed the expenses .
'Phil .
Mag August , 1909 .

