]\gt ; Photo-electric Emission from Sodium-Potassium Alloy .
359 If the maximum velocity at every point is , and if at any latitude the area of the land is a third that of the sea , the mean rate of of energy would be , from equation ( 26 ) , where is the radius of the eartl ] .
Hence from equation ( 33 ) or feet per second knots .
My thanks are due to Sir J. Larmor for the interest he has taken in the work , and to Dr. Proudman for help and advice on many points which have arisen .
Complete Photo-electric Emission the Alloy of Sodium and By WILLIAM WILSON , Ph. D. ( Communicated by Prof. O. W. Richardson , F.R.S. Reccived February 17 , 1917 .
) The experimental work described in the present paper suggested itself to the writer in connection with an earlier on the law the temperature variation of the complete photo-electric emission from a hot body , i.e. the photo-electric emission fro1n a body in equilibrium with the full ( black body ) radiation corresponding to its temperature .
By making uae of hypotheses contained in the quantum theory , the writer obtained the following expression for the current per unit area where is the work done in removing an electron from the hot body , and is equal to being the lowest frequency of the radiation capable of producing a photo-electric emission , and being Planck 's constant .
The quantity is the gas constant\ldquo ; reckoned for one molecule , and A is a quantity independent of , and characteristic of the substance .
As the expression inside the brackets in the above formula does not appreciably from unity , the latter is substantially of the same type as Richardson 's equation , ( 1 ) * The expenses of this research were partly defrayed by the aid of a grant from the Government Grant Committee of the Royal Society , to whom my thanks are due . .
Wilson , ' Annalen der Physik , ' vol. 42 , p. 1154(1913 ) .
Dr. W. Wilson .
The Complete Photo-etectric for the thermionic emission .
Richardson*has also shown that it follows , from thermodynamic considerations , that this law overns the complete photo-electric emission .
There is reason to believe that the thermionic emission is not wholly photo-electric in but it is clear that some portion of it is the complete auto-photo-electric emission of the substance concerned , and that the law goyerning its temperature variation should be the same as that for the whole thelmionic emission .
We are thus led to expect that , when a body is exposed to an external source of full radiation , the same law will govern the variation of its complete photo-electric emission with the temperature of the source of the radiation .
This expectation has been confirmed by experiments on the alloy of sodium and potassium , the description of which constitutes the subject of the present paper .
Description of the rimented Angements .
The part of the apparatus containing the alloy is illustrated by figs. 1 and 2 .
It consisted of a glass vessel completely sealed up and without taps , many attempts to prepare the alloy , and introduce it in satisfactory condition , failed on of the great difficulty of sufficiently good stop-cocks .
The sodium and potassium , in atomic proportions , were introduced into the portion .
This was afterwards sealed on to the rest of the apparatus , which was kept inverted during the preparation of the alloy , and not as actually shown in the figure .
The naphtha , in which the metals were immersed , was run off as far as possible before the apparatus was finally closed .
The vessel was then evacuated by a Toepler pump and a charcoal tube immersed in liquid air .
The metals were then melted , and the alloy formed , by Bunsen flame to the exterior of the tube .
After the apparatus had been evacuated as completely as possible , it was sealed off the pump , and the alloy run through a plug of glass wool and brought into the position ( figs. 1 and 2 ) by inverting the whole vessel , so that the liquid filled the -shaped portion .
About 3 cm .
above the surface of the alloy was a horizontal platinum wire ( figs. 1 and 2 ) , about 9 cm .
in length * ' Phil. Mag vol. 23 , pp. 619 and 620 ( 1912 ) .
See a paper by O. W. Richardson , ' Phil. Mag vol. 31 , p. 149 ( 1916 ) .
Emission from the Alloy of Sodium Potassiurn .
and mm. in diameter .
The ends of this wire were fused on , by means of gold , to thicker platinum leads sealed into mercury tubes FF ( fig. 2 ) .
Surrounding the wire , but not actually touching it , was a shield of silver foil and This was slightly open at the lower side , to allow the radiation from the platinum wire to reach alloy .
arrangement of silver shield and platinum wire constituted an approximately full radiator .
A battery of 8 or 10 accumulators was used to heat the wire of the radiator , , at the same , the connections were so arranged that the wire and its leads sCituted one arm of a Wheatstone bridge ( see fig. 3 ) , the remaining arms consisting of a resistance of ohms of thick eureka or platinoid wire mersed in paraffin oil , and the two large resistances ( of the order of 4000 ohms ) contained in a post-office box .
In addition to these four resistances , a fifth one was connected with the bridge .
This consisted of a very short piece of the same wire as that of the radiator , and was furnished with mercury tubes and leads exactly similar to those of the diator wire .
The whole ement is shown in fig. 3 .
By means of the mercury cups , , and , one of the terminals of the bridge vanometer could be connected in such a way that the short wire and its leads could be included Dr. W. Wilsorl .
The Complete Photo-electric in the radiator arm of the bridge , or , when desired , in the arm containing By this means the resistance of a uniformly hot portion of the wire of the radiatol could be found .
If this resistance be denoted by , and that of the ' leads and colder ends of the wire by , we can determine from the equations , ( 2 ) where and are the resistances in the arms A and , when the mercury cups and are connected ; and , the corresponding resistances when and are oined .
The bridge galyanometer was of the moving coil type , and had a resistance of , approximately , 100 ohms .
The whole arrangement was very sensitive ; in fact , such errors as may have entered into the measurements cannot be ascribed to the individual measurements , since a change of 1 in 4000 or 5000 of the arms A and produced an easily discernible galvanometer deflection .
Measnrement of th Photo-electric Current .
The methods employed to measure the photo-electric current from the alloy , when subjected to the radiation from the hot wire radiator , are shown rammatically in .
The larger currents were measured by a moving FIG. 4 .
Emission from the Altoy of Sodium and coil galvanometer ( fig. 4 , M ) , which gave a scale deflection of 1 cm .
for amperes , and was provided with a shunt , so that its sensitiveness could be uced , when desired , to of this .
One terminal of the galvanometer was " " earthed and other ( fig. 4 ) could be connected with the alloy by means of a platinum wire , sealed in the lass .
The radiator wire ( or one end of it ) and th shield could be maintained at a fixed posiCive potential by means of a battery of 240 small dry ceils ( fig. 4 , H ) , the negative pole of which was " " earthed The smaller photoelectric currents were measured by means of an electrometer and a condenser K. The capacity of the latter could be varied from microfarad to microfarad to suit the requirements of individual measurements .
The sensitiveness of the electrometer amounted to nearly 16 cm .
scale deflection for 1 volt .
The currents were measured by the time required by the electrometer and condenser to charge up to the potential represented by a deflection of 10 cm .
on the scale .
The sensitiveness of the electrometer was determined immediately before and after use .
The condenser , electrometer , the part of the apparatus containing the alloy and radiator , and all the connections were enclosed in metal and tubes , which were all connected to the same " " earth\ldquo ; as the pole of the -tension battery .
Sat of Currents .
number of measurements were carried out for the purpose of ascertaining what potential would saturate the currents .
The form of the apparatus and the comparatively large magnitude of the ) rendered it necessary to use a relatively potential for this purpose .
Iu fact , the upper limit of the photo-electric currents , which could be satisfactorily measured , was largely determined by the voltage available for them .
The dependence of the -electric current on the temperature of the radiator , when an insufficiently high voltage was maintained between the alloy and radiator , is illustrated by fig. 5 , which is plotted from actual measurements .
The ordinates represent the currents in arbitrary units , and the abscissae approximate temperatures .
The applied potential was 100 volts .
It will be seen that the current increases rapidly with the temperature up to a certain point , beyond which it rises less and less rapidly , and ultimately becomes stationary .
This phenomenon is already well known , and is explained by the fact that at the tures the emission is very and the applied potential is not sufficiently .
for saturation , owing to the mutual repulsion of the emitted electrons .
This kind of effect was first observed by J. E. Lilienfel accounted for it along lines similar to ' Phys. Zeitschrift , ' vol. 9 , p. 193 ( 1908 ) .
Dr. W. Wilson .
The Complete Photo-electric the above .
It has also been considered by C. D. Child , at greater lengt more recently , by I. Langmuir .
In fig. 6 is shown the dependence of the current on the applied potential when the radiator temperature was 57 C. The abscissae will represent volts when multiplied by .
The current , as will be seen from the , is practically saturated with 200 volts .
In the actual measurements described below , the applied potential was 350 volts , so that the currents obtained with the temperature stated above , and with all lower temperatures , were undoubtedly saturated .
fher radiator temperatures were employed , but the saturation of the currents was less satisfactory .
The saturation curve corresponding to a radiator temperature of 62 C. is shown in .
In this case uration appears to be just reached at the yhest potential at my disposal , namely , volts .
( The abscissae in fig. 7 to be multiplied by to give volts .
) It is that the current 'Phys .
Rev vol. 32 , p. 492 ( 1911 ) .
'Pbys .
Rev vol , 2 , p. 453 ( 1913 ) .
Emission from the Alloy of Sodium and Potassium .
\mdash ; FIG. 6 .
FIG. 7 .
VOL. XCi II .
Dr. W. Wilson .
The Complete Photo-electric measured , when the radiator temperature was used , slightly below the saturation value , but is not likely that it is much since the point corresponding to it in fig. 9 is not much below the straight line on which the other points may be approximately said to lie .
The Positive bission from the Radiator Wire .
The experimental method described above precludes any possibility ol errors due to a negative thermionic emission from the radiator wire , since the direction and magnitude of the applied potential were such as to prevent any negative ions , which might be emitted from the radiator , from ] reaching the alloy .
It is , however , obvious that any positive emission will be included in the currents measured by the galvanometer or electrometer .
The silver shield , although it could be insulated from the wire , was useless as a means of testing or for a possible remanent positive emission from the latter , since the heating process , necessary to get rid of this emission , associated with some condensation of metallic vapour on the shield .
It therefore necessary to subject the wire for a period to a con .
siderably higher temperature than those employed for the measurements ol the complete photo-electric emission .
During this heating a positive potential was applied to the wire , as the iter and other investigators*have observed that the rate of decay of the positive emission is dependent on the applied potential .
When the final measurements of the complete emission were carried out , the wire had been heated one or two hours a week over a period of about three months , and during these measurements the phenomenon of decay was quite absent .
This fact indicates that She positive leak , included in the measured currents , was of negligible magnitude .
Before the prolonged heating , the decay characteristic of the positive emission was very marked .
An accidental circumstance enabled an upper limit to be assigned to any positive emission included in the currents which are regarded as photoelectric ones .
During the earliest heating of the radiator wire , the latter emitted a small quantity of gas .
This showed itself by a small , but easily observable , movement of the alloy in the -shaped bend , and also by attacking the alloy .
A thin layer , showing the colours of thin films , formed over the alloy surface .
This had the effect of greatly reducing the ordinary photo-electric emission from the alloy .
When it was in this condition the currents observed were much smaller than those ultimately measured under the same conditions , but with a renewed alloy surface , and after the * W. Wilson , 'Phil .
Mag vol. 21 , p. 634 ( 1911 ) .
See also O. W. Richardson , Emission of Electricity from Hot Bodies , ' p. 227 et seq. Emission the Alloy of Sodium Potassium .
positive emission and decay had been got rid of .
In some of the earliest series of measurements , currents as low as 2 amperes were observed with a radiator temperature of 66 C. ; whereas with the same temperature and a clean alloy surface , and after long heating of the radiator wire , but otherwise similar conditions , the current rose to 48 amperes .
These numbers show that the positive leak was less than 1/ 24 of the total current observed during the linal series of measurements , and as the radiator wire was heated till the decay ceased to be perceptible , it may be safely assumed that it was practically absent .
It may be mentioned here that the direct thermionic emission from the alloy , which was undoubtedly heated to some extent by the radiation falling on it , did not contribute sensibly to the measured currents .
This was established by maintaining the radiator at the highest temperature employed in the experiments for a sufficiently long period to give the alloy at least as high a temperature as any it could have during the actual photo-electric measurements .
No emission from the alloy could be detected with the galvanometer after the current in the radiator wire had been switched off .
Had the thermionic contribution been only 1/ 150 of the total emission from the alloy it could easily have been detected .
and Results .
The measurements of the complete photo-electric emission from the Na-K alloy are given in Table I. Four fiducial temperatures were employed for determining the temperatures of the radiator .
These were : Melting point of potassium sulphate , 1070o C. Melting point of sodium chloride , so C. Melting point of lead chloride , 44 C. Melting point of naphthalene , C. The corresponding ' platinum\ldquo ; temperatures were calculated by means of Callendar 's formula , where is the temperature ( Centigrade ) and the ' platinum\ldquo ; temperature .
A small crystal of each substance was placed on a radiator wire , and the resistance of the wire ( less leads and cold ends ) determined when it was just hot enough to melt the crystal .
These resistances and the ' platinum\ldquo ; temperatures of the corresponding melting points are given in Table II .
Dr. W. Wilson .
The Complete Photo-electric Table I. cm .
ampere .
1 cm .
ampere .
NOTE.\mdash ; An electrometer deflection of cm .
represents volts .
A and are the tances in ohms of the corresponding bridge arms ( fig. 3 ) .
, indicate that the corresponding mercury cups were connected .
Table II .
Substance .
Melbing point .
Potassium sulphate 914 Sodium chloride 717 Lead chloride Naphthalene 80 These measurements are very consistent with one another , as will be seen from fig. 8 .
The ' platinum\ldquo ; temperatures , corresponding to the measurements given in Table I , were calculated by means of the formula , ( 3 ) where is the ' platinum\ldquo ; temperature , and the resistance of the wire .
For the constants , and , the following values , deduced from the fiducial data in Table II , were employed:\mdash ; The calculated results are given in Table III .
Column 2 contains the resistance of the radiator wire ( corresponding to different temperatures ) as calculated by equations ( 2 ) from the ements in Table I. In column 3 Emission from the Alloy of Sodium Potassium .
are the corresponding\ldquo ; platinum\ldquo ; tenlperatures obtained by equation ( 3 ) and the values of and given above .
The Centigrade temperatures computed by Callendar 's formula are in column 4 .
The remaining four colurnns contain , respectively , the absolute temperatures , their reciprocals , the common Table III .
C. Dr. W. Wilson .
The Complete Photo-electric logarithms of the photo-electric currents and the common logarithms of the photo-electric currents divided by the squares of the absolute temperatures .
In order to test the validity of equation ( 1 ) , which represents the expected relation between the complete photo-electric emission and the temperature of the radiator , it is convenient to put in the following form : log , ( 4 ) from which it will be seen that there is a linear relationship between and .
These two quantities are plotted in fig. 9 from the results given in FIG. 9 .
Table III ( being taken to be 2 ) .
The points , as will be seen , lie very close to a line over the whole of the very wide range of the photo-electric currents given in the Table .
The logarithm of the largest current exceeds that of the smallest by approximately , so that the extreme values of the currents differ by a factor of 46600 .
The tangent of the angle , under which the straight line in meets the horizontal ( 1/ ] : ) axis , is to a close nation 4 , and therefore , by equation ( 4 ) Emission from the Alloy of Sodium and Potassium .
where 0.4343 is .
The entropy constant is known to be ergs per degree , so that we get ergs , ergs , or joules .
For the purpose of comparison with thermionic or other measurements of , it will be convenient to give its value in equivalent volts .
The charge on an electron is very near coulombs , and therefore we find for the drop of potential equivalent to joules per coulomb , i.e. volts .
This result is in satisfactory agreement with the corresponding quantity as determined by Richardson from thermionic measurements on sodium , value being volts .
It should be said , however , that Prof. Richardson considers that his measurements may be subject to considerable errors through lack of saturation and the possible occurrence of chemical effects .
We can deduce the threshold frequency from the value of given above , that is to say , the lowest frequency of the radiation which is capable of causing a photo-electric emission .
This is simply effected by the equation which is based on the quantum theory .
The quantity is Planck 's constant and may be taken to be sec. The remaining quantity is the threshold frequency .
We have , therefore , sec. or sec. This corresponds to a wave-length ( in air ) of approximately .
Richardson and find for sodium or sec. * O. W. Richardson , 'Emission of Electrons from Hot Bodies , ' p. 69 . .
W. Richardson and K. T. Compton , ' Phil. Mag vol. 24 , p. 586 ( 1912 ) .
372 Photo-electric Emission from Sodium-Potassium Alloy .
Millikan*finds for the same metal or There is no reason to suppose that is smaller for the Na-K than for sodium .
The explanation of the discrepancies in the numbers just given lies in the fact , illustrated by the curves relating to sodium , in a paper by Richardson and Compton , that ( and therefore the threshold frequency ) is a function of the state of oxidation of the surface of the metal .
The writers just mentioned ( loc. cit they worked with a good vacuum , noticed a well marked " " fatigue They found , for example , that the photo-electric currents measured hours after the formation of the fresh sodium surface were considerably smaller than those measured only one hour afterwards .
The present writer also noticed the " " fatigue\ldquo ; phenomenon .
which is , of course , very well known .
After the formation of a new alloy surface , the photo-electric current for a given radiator temperature diminished for some time , but soon reached a stationary value , so that the same current could be measured days , provided the same radiator temperature was employed .
Though the alloy surface appeared perfectly clean and right ( and remained so ) , there is no doubt that , during the time which had to elapse before a series of measurements could be carried out , a very thin film of oxide formed over the surface , so that acquired a somewhat greater yalue than that corresponding to a perfectly clean surface .
Conclusion .
The law governing the variation of the complete photo-electric emission from the alloy of sodium and potassium , with the temperature of the source of radiation , has been experimentally investigated , and found to be well expressed ( as , indeed , theoretical considerations suggest it should be ) by Richardson 's formula for the temperature variation of the thermionic emission from metals .
I wish to take this opportunity of expressing my thanks to Prof. .
W. Richardson for his valuable advice , and for the interest he has taken in this investigation .
* Millikan , ' Phys. Rev vol. 7 , p. 380 ( 1916 ) .
O. W. Richardson and K. T. Compton , 'Phil .
Mag vol. 26 , p. 563 ( 1913 ) .

