On Contact Electricity of Metals .
On Contact Electricity of Metals .
" By J. Erskine-Murray , D.Sc .
, E.R.S.E. , Heriot-Watt College .
Communicated by Lord Kelvin , G.C.V.O. , F.R.S. Received August 4 , 1897 , \#151 ; Read November 18,1898 .
I. Introductory .
II .
Method of Experiment .
III .
Effects of Different Methods of Cleansing the Metallic Surfaces .
IY .
Thin Solid Films of Oxides , Iodides , \amp ; c. Y. Atmospheric or Time Effects .
VI .
Very Thin Liquid Surface Films on Metals .
VII .
Thick Liquid Films on Metals , fill .
Films formed from Gases .
IX .
Temperature Variations .
X. Elimination of Metal-Air Potentials by Solid Non-conducting Films on the Metallic Surfaces .
I. Introductory .
S 1 .
The experimental investigation described in this communica-ion had as its primary object the elucidation and measurement of he variations of .Volta contact electricity of a pair of conductors , due 0 changes in the state of that portion of the surface of each ondudor which was separated from the other conductor hy an insulating nedium .
S 2 .
The discovery of contact electricity of dry metals in air by 7olta at the beginning of the present century , extended a quarter of 1 century later by PfafE to dry varnished metals in other gases , has ) even confirmed by many subsequent experimenters .
The reality of he electrostatic force in air near an interface between copper and fine , inferred as an obvious consequence from it by Lord Kelvin , was jxperimentally demonstrated by him in 1861* .
In the next twenty fears many investigations were made , the more important being hose of Hankel , Gerland , Clifton , Ayrton and Perry , and von Zahn .
in 1881 a paper of great importance was published by M. Pellat , f md as the present communication is , in some respects , only an impiification and extension of his work , it may be of advantage to 'ire a short resume of it before going further .
M. Pellat 's most important results appear to be ( 1 ) his demonstration of the influence of the physical condition of the metallic surfaces m their Volta-potential ; thus he found that a sharply scratched plate is positive to a more smoothly polished one of the same metal , he metals being washed with alcohol after polishing in both cases md allowed to dry before the measurement of their potential ; ( 2 ) * ' Electrostatics and Magnetism , * S 400 et f ' Ann. Chim .
Phys. , ' 1881 .
VOL. LXIII , K 1X4 Dr. J. Erskine-Murrayv \gt ; his measurements of the temporary variations of potential due change of temperature of a copper , iron ; or zinc plate ; and ( 3 ) experiments which prove the smallness of the changes produced the potential-.difference of : copper and zinc by varying either t pressure or the nature of the gas surrounding them .
S 2a .
I shall now give a short summary of the results detailed the present paper :\#151 ; ( ?
) Metals coyered with non-conducting solid films of waxS glass , except at their point of contact , give nearly the same potenti as the bare metals in air .
The substitution of wax for air next tl metal only causes a small change which may be in the same directic and of approximately equal amounts for metals whose potentials in a are very different ; e.g. , the Volta-potential-differencebetween zinc as copper when both are coated with solid paraffin wax is very nearl the same as that between bare zinc and copper .
In this connectio I may mention that I have measured the potential of sodium coat with wax and glass , and find it to be about 3'56 volts positive to standard gold plate .
( ?
) A metal cleaned , by careful polishing and scratching with emery-paper or glass is less positive when its surface is in a sharp !
scratched condition than when smoothed or burnished , the differeno frequently amounting to 02 or 03 of a volt .
This result is not in opposition to that of M. Pellat mentionei above ( S 2 ( 1 ) ) , for the conditions were different , as his plates wert washed with alcohol after polishing , while mine were not ( set Chap. Ill , and also SS 39 et seq. ) .
( c ) The temperature variations , between 15 ' C. and 60 ' C. , of th( Volta-potential of many metals have been determined , both for clear dry metals in air and for metals coated with liquid or non-conducting solid1 films ; and it has been found that they are of considerable magnitude in both cases .
The curves representing the variation5 ol potential with temperature appear as if they should meet at a point below 200 C. , at a potential about 0'4 volt positive to a standard gold plate at 16 ' C. ; this suggests that Volta-potential-differences may possibly vanish at a very low temperature ( see Chap. IX ) , ( d ) A liquid film , even if of extreme thinness , may cause a considerable change in the potential of a dry polished plate , which continues permanent for many hours and even days after the disappearance of the film .
Two films of the same liquid opposed to one another on the surfaces of two plates of different metals do not usually give zero potential-difference , as solid conducting films of one material would do , but give nearly the same potential-difference as the dry metals on which they lie ( see Chaps .
VI and VII ) .
( e ) A very thin film of oxide on a metal produces only a very On Contact Electricity of Metals .
lall change in the potential , and every increase in thickness of the m is attended by a farther change in potential until a limiting lue is reached , which is that of a mass of the oxide ( see lap .
IV ) .
( f ) Exposure to the atmosphere at ordinary temperatures does \gt ; t , as a rule , produce any rapid change in Volta-potential , especially the air be comparatively dry and free from dust .
The timate change is usually in the negative direction ( see Chap. V ) .
( g ) I have extended Lord Kelvin 's experiments , on the effect of mporary immersion of a metal in a gas to the cases of copper , zinc , n , and silver in oxygen , and find that copper , zinc , and silver icome temporarily positive , while tin becomes negative in consequence ' this treatment ( see Chap. VIII ) .
The research , suggested by Lord Kelvin , was carried out in the atural Philosophy Laboratory of Glasgow University during the essions 1893-94-95 , and during 1895-96 in the Cavendish Laboratory Cambridge University .
My thanks are due to Lord Kelvin for any suggestions and much valuable advice , both in regard to cperiments and to the discussion of results , and to Professor J. Thomson for similar kindnesses during my work in the avendish Laboratory .
A small portion of this investigation , on the feet of Rontgen arrays on the contact electricity of metals , was ablished in the 'Proceedings of the Royal Society ' for March , 1896 .
I have also to thank Professor James Holm , M.A.* and Mr. eorge E. Allan , B.Sc. , for the part they took in the earlier portion f the work .
II .
Method of Experiment .
S 3 .
The measurement of the natural potential-difference between ny pair of conductors was usually made , in air , by the null method escribed very briefly by Lord Kelvin in the Report of the British tssociation for 1880 , p. 494 .
To make clear the exact circumstances i which the potentials were measured it will be advantageous to escribe the apparatus and general method of experiment in detail .
S 4 .
A circular disc of one of the metals , usually about 9 cm .
a diameter and 02 cm .
thickness , is insulated and in permanent onnection with the insulated pair of quadrants of an electrometer .
l similar plate of the other metal placed parallel to the first , at a istance of a few millimetres from it , is uninsulated and in connection dth the uninsulated pair of quadrants .
While in this position they re temporarily connected with one another through a simple form f potential-divider in which a slope of potential is maintained by a * When this was written last August , my friend and former fellow-worker was 'rofessor of Applied Mathematics in the South African College , Cape Town .
He ied in October . .
" c. r ... .
\#166 ; e- - * 116 Dr. J. Erskine-Murray .
Daniell cell in the direction opposite to that given by the expei mental plates .
The temporary connection is now broken and i plates are separated .
In doing so the capacity of the condenser reduced , hence if there be any electric charge on them it will 1 indicated by a further deflection of the electrometer .
This operatic is repeated with different values of the counterpotential until separ ; tion of , the plates produces no change in the deflection of tli electrometer .
Since it has annulled the charge , the countei potential must nqw be equal and opposite to the natural potentia difference .
S 5 .
During most of the experiments a piece of apparatus mad by Lord Kelvin in or about 1861 Was used .
In it the plates wei surrounded by a cylindrical zinc case .
This was made in two pari for convenience in manipulation , the lower fixed to a heavy cast-iroi base plate , while the upper stood upon small brackets attached ti the top edge of the lower part , its position being regulated by i hole-slot-and-plane arrangement .
Openings in the sides of tht upper part facilitated the adjustment of the experimental plates , bui these were closed by a sliding cover during a measurement o !
potential .
The lower plate was supported by an insulating glass stem which was kept dry by means of pumice and sulphuric acid .
On the top ol this stem was a brass cap screwed to fit the sockets which were soldered to the backs of the plates , and from the brass cap a still wire passed out through a hole in the case and formed the connection between the insulated plate and the electrometer .
A piece of platinum foil was soldered round a part of this wire to give a clean contact for the temporary connection between the plates through the potential-divider , and the wire from the divider had likewise a platinum end piece .
The upper plate hung , by a ring at the centre of its back , on a hook at the lower end of a vertical metal rod , which could be drawn : up so as to increase the distance between the plates .
A small disc I fixed on the rod at right angles to it just above the hook had three : screws passed through it and pressed against the plate , holding it i firmly in the hook ; by this means the upper plate can be set parallel : to the lower one in a very short time .
This apparatus , which is the same as that used by Lord Kelvin in his experiments many years ago , was found to be very convenient , the arrangements enabling us | to take out and replace a plate very rapidly , a matter of great i importance when observing temporary variations , while at the sam\#174 ; time both plates were firmly attached to their supports when in position .
S 6 .
The divider , whose total resistance was about 2400 ohius , ] was made to divide the potential-difference between its outer , jj On Contact Electricity of Metals .
rminals into 100 equal parts .
The potential applied was valley that given by a Daniell cell of Lord Kelvin 's gravity type , id was tested frequently by comparison with a standard Clark cell , by means of a standard resistance and a current balance .
S 7 .
The permanent connections were as follows :\#151 ; ( 1 ) The lower ate to the insulated pair of quadrants of the electrometer ; ( 2 ) le terminals of the Daniell cell to the ends of the divider ( there as a reversing plug in this circuit ) ; ( 3 ) one end of the divider to le upper plate and uninsulated quadrants of the electrometer .
The temporary connection ( S 5 ) was from the sliding contact piece F the divider to the lower plate .
S 8 .
Each experiment was as a rule begun by polishing a metal late on clean glass-paper or emery cloth .
Its contact potential ith a standard plate , generally of electrolytically deposited gold .ashed some hours previously with alcohol , as used by M. Pellat , .as then measured .
The plate was next subjected to some particular reatment ; for instance it was filed , or burnished , or polished on 3ather or paper , or washed with water , alcohol , or turpentine , or eated in air and oxidized , or not oxidized , or exposed to steam , or xygen , or fumes of iodine , or hydrogen sulphide , or simply left to Iter under the influences of the atmosphere and its own molecular orces .
Its potential with the same standard plate was again measured , and the change due to the treatment its surface had mdergone noted .
S 9 .
The metal which requires to be joined to the zinc end of the \gt ; attery , in order to effect a balance , will be called 'positive , dhus iluminium is positive to zinc and zinc positive to copper .
Also , when a plate becomes more positive its potential will be said to rise ; when more negative , to fall .
When the " potential " of a plate is nentioned , without other qualification , its contact potential-difference with a standard plate ( S 8 ) is meant .
S 10 .
The results given are not in the order in which they were \gt ; btained , but are classified in such a way as to show more clearly jheir true import , and those of many experiments , which are not mentioned in the text , will be found in the tables .
$ III .
Effects of different Methods of Cleansing the Metallic Surfaces .
S 11 .
It seems probable that the very conflicting results obtained s by different experimenters for the potential-difference of any given pair of metals , in air , must be due , to a large extent , to differences in the methods by which the surfaces have been prepared .
In order to obtain a metallic surface as free as possible from all contamination it is clear that a hard polishing agent , such as a clean steel file , or emery-cloth , or glass-paper , must be used , as a softer material leaves Dr. J. Erskine-Murray .
more of itself On the plate .
A liquid is quite unsuitable , since fever particle of the metal probably retains particles of the liquid adheri , to it .
Thus the smell of a liquid remains long after the plate appeal to be quite dry ; and it has been found that such a film as must exl to cause the odour is quite sufficient to alter the potential ye considerably .
, \#166 ; S 12 .
In order to obtain uniformity of action the clean glass-pan or emery-cloth used to polish a plate was fixed on a wooden roll , made to revolve with circumferential velocity of about 100 ci per second .
Care was taken to hold the plate so that the scratch* caused by the polisher , should all be parallel .
Thus little or no could lodge in the surface , which would have occurred had tl scratches crossed one another .
A record was kept of the nature \lt ; the polishing agent in every case.- Apiece of glass-paper , or otto polisher , was seldom used more than once or twice , and was never used for any different metal .
S 13 .
Mi Pellat found that every change in the smoothness of ; surface is accompanied by a change in its contact potential ; but a his experiments were limited to metals washed with alcohol , it wa of interest to extend them to more general cases .
Thus a plate o zmc which had been polished on clean glass-paper , and had therefore , a surface sharply scratched in parallel lines , was found to be 0 ' 70 volt positive , to the standard gold plate .
It was next tarnished with a tool of hardened steel , and with the same standard plate it now gave 094 volt .
Two hours latOr the same plate gave 0 9^ volt , showing that , the effect is nearly permanent .
By burnishing the p ate again a finer polish was obtained .
With the same standard the potential is now TOO volt , still further burnishing giving T02 volt .
If the.zinc be now polished on glass-paper it remms to its original po en ia .
pw steel , of which , the burnisher was made , is negative to zinc ; hence this rise cannot have been due to particles of steel on the zinc , but must have been caused by some change involved in the smoothing process to which the zinc had been subjected , possibly by a hardening of the surface layer .
On Contact Electricity of Metals .
112 S 14 .
Intermediate states of polish give intermediate ( values of otential .
For instance , if a zinc plate be filed it will not be so tiarply scratched as it would be by glass-paper , and of course not so nootb as if burnished .
Its potential is also found to lie between lose of scratched and burnished zinc .
Each polishing , !
with any laterial , was usually sufficiently thorough to efface the effects of all revious polishings : the potential observed thus depended only on he state of surface produced by the polisher used just before the bservation .
The results of experiments show that the'smoother he surface the more positive it becomes .
S 15 .
The generalization given in S 14 is supported by the results , f over 100 similar experiments , with a number of metals .
The actual results are given in Table I ; it will be well , , however , , o discuss some of these ih greater detail than is possible in tabular orm .
S 16 .
In order to eliminate the use of different polishing agents , tried the effect of producing different states of surface by rubbing he plate against another of the same metal .
Two copper plates which had been polished op glass-paper gfive ( ?
) + O'Ol volt , ( ?
) -fO'04 " | .
, | -v- * .
... - ... ... . .
... ... .
' ... , ... ... .
U with the standard plate .
They were next , gently rubbejl together until parts of each were shiny and now gave , with j the same standard , j , s ( a ) +0*07 volt , j j ( b ) +0-07 " ; 1 V y i i . !
In another experiment two coppers which had been pplished on medium emery-cloth ( Davies 's No. l|r ) gave , with th^ standard plate , -i : \gt ; .
.'\gt ; i ' \#166 ; j ' .
: ' i ( a ) \#151 ; O'11 volt .
( b)-0-06 " !
' When slightly burnished by rubbing them together , they gave La ) \#151 ; 0-02 volt .
( b)-0-Q2 " : Thus , smoothing by mutuftl friction made both more positive .
The amount of the change is iniot great , either in potential or in smoothness , as it is very difficult to polish copper on copper , but the direction of change is the same as previously found , namely , a smooth surface is more positive than a sharply scratched pne .
S 17 .
On account of the difficulty of obtaining a burnished surface by simply rubbing two pieces of the same metal together , I tried another form of experiment in which the possible effect of the Note .
The potentials were taken with a standard gold plate ( see S 8 ) , which had been wasl Dr. J. Erskin e-Murray .
.3 fc* O \lt ; D \#163 ; JO *3 T5 O *3 Aluminium .
Bemarks .
Average Poten- tial . ?
~1 Potential in volts .
+i*u + 1 " 04 +1 *13 + 1*15 + 1-21 Zinc .
Bemarks .
Bather smoother than last .
i Potential in volts .
+0*67 j +0-70 } Average Potential .
SPSS SSiSgSKSS . ? ? ? ? ? ? ? ? ? ? ? ?
Tin .
Average Poten- tial .
+0*60 Potential in volts .
S8SS .
o o o o : + + + + Copper .
High conductivity . .
Average Poten- Bemarks .
tial .
r i 8 S ? ? !
Potential in volts . ? ? ?
?\#165 ; =53 ? ? ?
5 n o\#187 ; I " 3 .
) Bemarks .
Consecutive polishings on same paper .
5 \#169 ; Average Poten- tial .
S f Z " 9 i i ?
t ?
\#151 ; 1 v Potential in I volts .
-0-19 -0*21 [ -0-19 \#151 ; 0*19 -0*24 : -0*20 -0-24 -0*19 \#151 ; 0'14 -0*21 -0-12 -0*11 ' \#151 ; 0*C8 -0-07 .
-0*11 -0*06 -0*09 \#151 ; 0 'll PolisliiDg agents in order of sharpness of surface produced .
1 g s. | .** J 1 \lt ; S if ' i * i " J J1 1 S ' .* 1 S\#166 ; i f .
t a 1 | On Contact Electricity of Metals .
Burnished with glass knob .
S S3 Si r-4 rH \#187 ; -H + 4* 4 r~i r- , r-4 HH + 444 + + Bather rougher than last .
Bather rougher than last .
Very bright surface .
Finer polish .
These four exps .
were consecutive .
S 9 9 % i i i + s ? ? ? ? ? ?
: ?
+++ ?
+ g .
S : o : o 4 + ?
* o : o 4 + These were consecutive experiments .
The plate was burnished before each observation .
$ ?
S o o o 4 +4S sSS| SS ?
?= ? .
o OOOO OOOOOO : 4- 4-+ 4-4- 4* 4 " 4* 4* 4* 4 : : : : s S s S o o o o l 4 4 4 ' ? ? ? ?
S\#163 ; S OOOO OOO OOOOOO OOOO | | | I 14- 4- 4- 4- 4- 4- 4* 4 4 4-4File Glass-paper which had been worn smooth Steel burnishing tool Washleather ( after glass-paper ) * 122 .
Dr.* J/ IliskinerMurray .
material of the polishing agent is eliminated by using the same to\lt ; to produce different states of surface .
' Thus a copper plate wa scratched very roughly with a steel tool , and gave \#151 ; 007 volt with the standard plate , then burnished slightly with the same toe it gave \#151 ; 0,02 volt .
It had thus risen 0*05 volt on account of the alteration producec in smoothing its surface , j Thus these results also confirm the con elusion given in S 14 .
S 18 .
It should he noticed that , although individual result occasionally appear to conflict , the average value of the potential of any fifiven metal when polished with a certain agent is almost always in proportion to the smoothness of the surface produced .
However much one may try , it is impossible to get exactly the same state of surface over again , but ; by taking the average of a considerable number of experiments with one polishing agent , a general value is got which represents the potential given by the plate in a certain state of polish .
It is so arranged that the sharply scratched surfaces come at the head of the table , and each succeeding polishing agent m the list produces a smoother surface .
S ^ ^t should be observed that with soft metals such as tin , i erent polishers produce but little change in the contact potential , is , though at first sight apparently contrary to the results for harder metals , is really in accordance with them , for different po is ers do not produce [ appreciably different states of polish on a so ^ metal .
This remark applies equally to the fact that a smaller vana ion was observed with ''high conductivity " copper than with eo copper plate , as the old plate was found to be considerably harder than the purer " high conductivity " copper .
* * . ? .
' ^ *8 P^^able t^at the variation described in this chapter is no c irectly due to roughness or smoothness , but rather to altera* tion , either by change in the outer layer of the- metal , or-in the film o\#177 ; condensed air which no doubt exists on it , of the interface between metal and air .
TV .
Thin Solid Films of , , fyc .
, ^ -^-mong the earlier experiments were many in which the changes due to films of oxide of different thicknesses were studied ; j \#153 ; 1 \S was primary object of investigation suggested by vin ' \#174 ; Qt constant difficulties and ambiguities presented emse ves , hindering the interpretation of even the simplest On Contact Electricity of Metals .
123 xperiments , and it was found necessary to enlarge the scope of the esearch iu order to remove or explain them .
S 22 .
The experiments detailed below are among the most definite f those carried out .
They show large changes in potential on ccount of very thin films of oxide , and prove that the amount of liange is dependent on the thickness of the film when below a ertain limit .
S 23 .
Two plates of cast zinc , which had been carefully polished m medium glass-paper the day before , gave with one another -h 0*02 volt , the upper plate being positive .
The upper plate was now taken out and heated on the back by means of a very small blowpipe flame .
Its face became slightly blistered , but not much discoloured , and when sold it gave with the other , which had not been altered in any way , \#151 ; 0,44 volt , the " minus " indicating that the tipper is now negative .
This shows a permanent fall of 0*46 volt due to heating in air .
S 24 .
A plate of rolled zinc , which gave \#151 ; 0-04 volt with another plate of the same material , was heated as in the last experiment , but to a higher temperature , very nearly to its melting point .
Its face became a streaky yellowish-brown , slightly purple towards the centre ( the brown parts turn purple if moistened ) .
When compared , cold , with the clean plate it now gave -0-79 volt , a fall of 0'75 volt .
Thus this oxidized zinc plate has practically the same potential as copper .
This great fall of potential does not , take place by exposure to the atmosphere at ordinary temperatures , unless possibly after the lapse of many yeai 's , for the potential of zinc plates which have been left unpolished for many months is only two or three tenths of a volt below that of clean zinc ( see S 34 ) .
S 25 .
Similar experiments were made with copper .
The film of oxide was gradually increased in thickness by repeatedly heating the plate , and the potential measured in each stage of oxidation , the plate being cold .
There was a change of about 0'03 volt in the negative direction before the oxide-film became visible , and further increase in the thickness of the film was judged by the gradual change of colour , until the well-known dull purplish-black of massive copper oxide was attained .
The potential , measured each time when the plate was cold , gradually fell until it reached a limit of about 0'30 volt 124 Dr. J. Erskine-Murray .
negative to a standard gold plate , which value may therefore t I taken as the potential of a mass of copper oxide at 15 ' C. It ma he remarked that there are good reasons , which will be given ii Chapter IX , for supposing that oxidation does not commence in ai unless the temperature of the plate be raised above 80 ' C. S 26 .
A plate of copper which had been polished on glass pape 1 and then on wash-leather gave + 0'20 volt , with a standard gold plate , and was next held in iodine vapour for i 1 short time .
It looked moist at first , but quickly became dry .
In t this state it gave \#151 ; 0-34 volt , with the same standard , a change of 0 54 volt in the negativt f direction .
The surface had a dull colour with a whitish " bloom " j on it , the tint resembling that of clean copper , though rather darker .
| Exposure to iodine vapour thus makes the surface of copper I nearly as negative as continued exposure to the atmosphere at .
a h high temperature .
S 27 .
The potential of a clean zinc plate fell about 0*12 volt when j it had been exposed to iodine vapour .
S 28 .
The iodine compounds formed on the metallic surfaces were I not stable , as the appearance and potential of the plates altered con- j siderably during twenty hours ' exposure to the atmosphere .
S 29 .
The effects of sulphur fumes , hydrogen sulphide , \amp ; c. , were I tried , but though , in most cases considerable changes were observed , i the results obtained were not satisfactory , Y. Atmospheric or Time Effects .
S 30 .
A small variation of the potential is usually found to take j place during a short time after the plate has been polished , and before f it has settled down to a more or less permanent value .
This is partly j U.6 .
that the plate has been slightly warmed during j ?
' 18p1^ " ' ^a^es some time to come to the atmospheric temperature ; j ( see ap .
) .
As regards changes which are not due to variation j ot temperature , I shall give some typical experiments in the following sections of this chapter , S 31 .
Two plates of " high conductivity " copper were polished on j \#174 ; glass-paper , and several experiments by burnishing them by I mutual friction were then made ( see S 16 ) .
Forty-seven minutes alter the original polishing , their potentials with the standard plate I were * ( a ) +0'035 volt .
( b ) +0*045 " On Contact Electricity of Metals .
he plate ( a ) was left in the apparatus all night , that is to say , in a osition sheltered from dust , and in air partially dried by the ilphuric acid in the lower part of the case ; the potential next mrning , 20 hours 3 mins , after the measurements given above , was + 0"045 volt rith same standard plate .
It was thus nearly constant .
The plate b ) was left in a rack in the laboratory without special protection , nth the result that next morning the potential , owing to exposure to he air of the room for about 20 hours , had fallen to -\#151 ; O'056 volt , i change of \#151 ; O'OOl volt .
The plate ( 6 ) was now placed in the ipparatus , and remained nearly constant for 2 hours at least .
Plate a ) was left in the rack , and its potential fell 0'065 volt in about 2 lours .
It is thus evident that at ordinary temperatures in a rather Iry atmosphere , and in a place protected from dust and light , the rate }f variation of the contact-potential of copper is very small ; in an exposed place , however , it is no means negligible .
S 32 .
A plate of block tin was polished on glass-paper .
Its potential varied with time as follows , the time being counted from die completion of the polishing:\#151 ; Hours .
Mins .
Yolts .
0 22 ... +0-515 0 27 ... +0-520 0 46 ... +0-535 1 4 ... +0-535 1 23 ... +0-515 3 18 ... +0-495 4 6 ... +0-495 4 42 ... +0-495 71 12 ... +0-465 The plate was kept in the apparatus all the time and its potential taken with a standard gold plate .
S 33 .
In another experiment a tin plate polished on the finest glass-paper gave , after 0 hours 15 mins .
, + 0'525 volt , After 23 hours 30 mins , it gave + 0-535 volt .
This experiment was made in June , while the previous onejwas made in December .
Hence probably the greater rate of variation in the former was due to a moister and less pure atmosphere .
1M I)r/ J. Erskine-Murray .
S 34 .
An old zinc plate which had probably not been polished f( many years gave + 0*37 volt with the standard gold plate .
It was next polished on glass-papei and gave , v + 0'75 volt with the same standard .
This shows that the ultimate effect o j exposure is to make the potential of zinc fall .
I have found that ii i some cases this fall is preceded by a slight rise , as in the case of th i ( SS 32 , 33 ) ; but the ultimate effect is in the negative direction .
S 35 .
The potential of a silver plate which had been polished on | glass-paper remained constant for an hour .
The experiment was not * continued further .
, , S 36 .
Aluminium becomes gradually negative m air .
In one case * the potential fell about O'18 volt in a week .
The time-change of this j metal , mainly on accountof its large temperature-variation ( se'e \ Chap. IX ) , is rather difficult to determine .
S 37 .
An iron plate which had not been cleansed for two months i rose 0-05 volt when polished .
Hence the effect of exposure had been j to make its potential fall slightly .
S '38 .
I have included these results , for which , as for many others j like them , I can as yet give no definite explanation , in the hope that i they may serve as a basis for future experiment and generalization , i The great difficulty in all su\lt ; ?h " experiments on contact electricity is to define the circumstances and , , with them , the causes of any given variation .
t VI .
Very Thin Liquid Films on Metals .
v ear^er experiments it was found that , as a rule , copper polished on glass-paper or emery-cloth was about 020 volt negative to the standard alcohol-washed gold plate .
On comparing this value with M. Pellat 's results* I found that his value for copper was about 020 positive to the same standard .
I at .r+if .f ffrmine\lt ; ^/ r\gt ; ' meii10d ofeleansing the plate so as to find if the difference ( 0*40 volt ) were due to that alone .
Tor this purpose the^ta^d^ d^6 WaS Care^U^^ P'iished on glass-paper .
It gave , with \#151 ; 0'20 volt .
It was then washed with alcohol and allowed to .dry in air . .
It now gave .
+ 0:22 volt with the standard .
The change due to the treatment is there-^ fore +0-42 , volt , and continues permanent many hours .
* * ' Aup .
Chim .
PHys .
, ' 1881 .
; v iciu-sol On Contact Electricity of Metals .
117 S 40 .
A plate of thickly oxidized copper giving , with the standard ate , \#151 ; 024 volt iv , after it had been washed with alcohol , and allowed to dry , O'OO volt , rise of O'24 volt .
S 41 .
A zinc plate , when cleaned on glass paper , gave + 0'61 volt .
t was then washed with alcohol , and when apparently quite dry :ave + 094 volt , emaining constant at this value for many hours .
The rise in this as is 0'33 volt , i S 42 .
The results of many-similar experiments made with alcohol nd other liquids on gold , silver , zinc , copper , iron , tin , lead , and luminium will be found in Table II .
It is notable that in almost very case the contact-potential of a metal which has been polished m a hard dry material rises at least O'15 volt in consequence of deohol washing , and that this change is nearly permanent for many xours after the plate appears to be quite dry .
S 43 .
It may be remarked that , as M. Pellat appears to have nvariably washed the metal with alcohol before making a measure-nent of its potential , his results do not apply directly to metals which have been cleaned by means of a hard dry agent only .
Even in the experiment which he gives on pp. 79\#151 ; 80 of his paper as a proof that alcohol-washing does not permanently alter the potential of a clean gold plate , he appears to have used a plate which had already been washed with alcohol ; his result , therefore , does not prove that alcohol-washing makes no change in the contact-potential of a clean plate , but only that subsequent washings do not alter the state of surface , i.e. , that the effect is nearly permanent .
This agrees with the results given above and in Table II .
S 44 .
The alcohol film , whether in combination with the metal or not , must in any case displace the air from the surface of the plate .
Other liquids seem to leave films in much the same way , though the change of potential is different for each liquid ; in some cases , e.g. turpentine on zinc and copper , it is in the positive direction for the positive metal and in the negative for the other .
VII .
Thick Liquid Films on Metals .
S 45 .
In experimenting with films of sensible thickness the liquid was usually placed on a flat metal plate , its surface-tension being Dr. J. Ers kin e-Murray .
Table II.\#151 ; Permanent Changes produced in Contact-potential c Metals by washing with Alcohol and drying in Air .
Metal .
Previous treatment .
Copper # *. .
Polished on glass-paper \gt ; ... ... . .
jy . . . .
yy . . . ?
Polished on emery-cloth ... ... . .
Piled yy * .
** Oxidised t# Gold Polished with dry " plate powder " Silver ... .
* .
Polished on glass-paper Iron * .
Dry polished , t # , Polished on emery-cloth T t T T ## Lead Scraped thoroughly with a knife Tin Polished on glass-paper. .
Zinc ... ... .
Aluminium \#187 ; ) ) ... ... . .
i# \#187 ; )\#187 ; ... ... . .
\#187 ; \#187 ; " i ' ' \#187 ; ... ... . .
Potential with standard gold plate .
Before .
volt .
-0-20 + 0-09 + 0-02 -0*13 \#151 ; 0 " 24 + 0*10 + 0*08 + 0*17 + 0-61 + 0-56 + 0*61 + 1-05 + 1*04 + 1-13 + 1-17 After .
volt .
+ 0-22 + 0*29 + 0-14 + 0*14 0-00 + 0*20 + 0-16 + 0-29 + 0-80 + 0*66 + 0-94 + 1*10 + 107 + 1-28 + 1-26 Change .
volt , s t-0 *42 + 0*20 + 0*12 + 0*27 + 0*24 + 0*05 + 0*10 + 0*081 + 0 *12 I + 0*171 + 0*14 + 0*19 i + 0*10 1 + 0*33 I + 0 + 0 + 0 + 0 05 03 sufficient to retain it if its depth were not more than 0*3 or 0*4 cm. .
In other cases a shallow metal dish was used .
The behaviour of water puzzled me for some time , and made me realize the extreme i cu tie of obtaining a clean liquid surface .
After some preliminary experiments I found that distilled water which has not been exposed n cf a^mosphere gives fairly constant results when lying to a depth ot 0*2 cm .
on one of the standard gold plates .
When water on gold formed the one plate and a dry standard gold plate ( see S 8 ) the 0 er ' ^ e va^ue f'nnd was 0*10 volt , the water being negative .
Thus clean water in contact with gold is about 0*10 volt negative to it .
1 he value found is quite different if the water has been exposed to the air for some time , and is usually positive instead of negative to the standard plate . .
exPe^e\lt ; ^ / rom lungs produces a very marked change in the contact-potential of water .
By blowing through a fine glass On Contact Electricity of Metals , ibe which dipped under the surface of a layer of distilled water on a \gt ; ld plate , the potential was changed from \#151 ; 0*10 volt ith the standard plate to 4-016 volt , nd remained at this latter value for more than half an hour , showing rat the change was not due to a rise in temperature .
The curious oint is that the change seems to he only in the surface of the liquid ; )r when most of it was shaken off the potential fell to \#151 ; 005 volt , lthough the plate was still wet all over .
This appears as if the Iteration were due to a surface film of oily or dusty matter on the rater .
It was found that distilled water which had been exposed to he air of the room for some time gave the same value as water which iad been breathed through .
S 47 .
A lead plate and a flat circular leaden dish were scraped lean with a knife , and next morning , when compared with each there , they gave + 0T0 volt , lie plate being positive to the dish .
Water was now poured into the lish until it was nearly full , 04 cm .
deep , the potential remaining he same ; i.e. , the lead plate gave with the water in the leaden dish , + 0T0 volt , water in leaden dish being negative .
S 48 .
The plate and dish were again scraped , and this time they ^ave + 005 volt , +e plate being positive .
Water was next poured into the dish , which gave , with the dry lead plate , + 0'09 volt , the water in the dish being negative as before .
The dish was now removed , and a standard gold plate put in its place ; with this the lead plate gave + 0'52 volt , the lead being positive .
Thus water in contact with lead is about 0'9 volt negative to dry lead .
S 49 .
Turpentine seems to produce opposite effects on zinc and copper , i.e. , copper becomes more negative , and zinc more positive , VOL. LX1II .
L Dr. J. Erskin e-Murray .
when wet .
Thus copper and zinc which had been dry-polished gav , with one another O'77 volt , zinc positive .
With the zinc wet with tnrpentine , 0*96 volt .
And with both zinc and copper wet about 1*20 volts .
Their potential difference increased somewhat as the turpentine dried up ; and when apparently quite dry its value was still at least 1*20 volts , the plates , however , smelt strongly of turpentine , showing that an invisible layer remained on their surfaces .
S 50 .
A crystal of copper sulphate gave , with a standard gold plate + 0'02 volt , copper sulphate positive .
This result was obtained in connection with some experiments on the contact-potential of an aqueous solution of copper sulphate .
It was found that such a solution in a copper dish gave about + 0-07 volt with a dry standard gold plate , e so ution being positive .
Blotting-paper saturated with copper sulphate gay about +0*10 volt with the same standard plate .
As no special precautions were taken in these experiments to obtain a per ec y pure liquid surface , one cannot , especially in view of the changes described in S 46 , be quite certain that they represent the rue potential of copper coated with a solution of its sulphate .
They ' agree , however , with Professors Ayrton and Perry 's values for satu-rated and non-saturated solutions .
n-n ?
't !
he standard g'ld plates which had been polished with Hollis s plate powder used dry gave , with a similar plate which nad a layer of alcohol on it , \#151 ; O'13 volt .
When the alcohol had dried up the value was \#151 ; 0'05 volt , the polished plate being negative as before . .
wki'h I draw from the experiments describee m this and the preceding chapter and in the tables is , that a layer oi .
, '^1 a metallic surface does not give , with a metal separated om it by air , a definite potential-difference of its own , as in similar circumstances a solid conducting film would do , but merely adds a On Contact Electricity of Metals .
13 f ertain amount to that of the plate on which it lies .
Thus two differ-nt metals coated with layers of the same liquid do not , as a lmle in ir , give zero potential-difference , but usually give nearly the same otential-difference as the dry metals .
For instance , the potential of 3ad with an alcohol layer OT cm .
deep on it is about 0T3 volt higher han that of dry lead , while that of wet copper is about the same mount higher than the average value for dry copper .
There is no endency shown for a liquid film to take up a definite potential inde-xendent of that of the metal on which it lies , with any metal separated from it by a dielectric , as a solid conducting film would do .
Phis seems to be the most impoi'tant distinction between solid and iquid conductors , and it is in accord with what is known of voltaic jells ; for if the potential-differences in the chain copper-water-zinc were equal and opposite to that of zinc-copper , as copper-iron-zinc is jo zinc-copper , we should have no electromotive force in the circuit , when the materials are all at one temperature .
In most of my experiments the type is copper-water-nonconductor ( air)-water-zinc , xnd not copper-water-zinc ; the members of the first and last pairs ire in contact , while a non-conductor intervenes between the two free water surfaces , these being able to take up their natural contact-potential-differences with the metals they touch ; but the sum of the potentials , copper-water and water-zinc , is not equal to copper-zinc , thus proving that the two free surfaces of water are not at the same potential .
This is directly demonstrated by the experiments of Professors Ayrton and Perry , and by the results given in Chaps .
VI , VII , and X of this communication .
If the intervening layers of non-conductor be removed by joining the liquid surfaces so that there is but one mass of liquid between the plates , conduction at once tends to reduce the whole liquid to the same potential , leaving the contact-potential differences , now unbalanced by the removal of the non-conducting medium which was capable of sustaining the stress , to act as external electromotive force .
This then shows the connection between contact-potentials , measured electrostatically by the method described in this paper , and the electromotive force of a voltaic cell .
S 53 .
The contact-potential of a liquid with a metal is clearly , if the air-potentials be neglected , the difference between the potential of the dry metal and that of the metal when wet with the liquid ; the same standard plate being used as zero of potential in both cases .
The results given in the tables must no doubt be in some cases complicated by the formation of solid compounds in the interface between liquid and metal , so that the liquid is no longer in contact with clean metal , and the liquid must also displace any film of condensed air which may exist on the metal .
The latter influence will be discussed in Chaps .
VIII and X. Dr. J. Erskin e-Murray .
Table III.\#151 ; Potential of Metals covered with visible layer of Alcohol .
Metal .
State of surface .
Potential with standard gold plate .
Dry .
Wet .
Change .
Copper ... .
volt .
volt .
volt .
In Filed ; wet with alcohol , -0-13 + 0*19 + 0-32 Gold Dry-polished ; wet with alcohol .
4 . .
+ 013 " Silver Glass-paper polished .
Wet , 0*05 cm .
deep + 0-10 + 0*18 .
+ 0 -08 : " thin film . .
+ 0-19 + 0-09 ?
Lead Scraped thoroughly Wet , 0*1 cm .
deep + 0-61 + 0-74 + 0-13 % 1^ * ' / f " thin film only .
.* + 0*74 + 0-13 \ Table IY.\#151 ; Change in Contact-potential of Zinc and Copper due to Layers of Turpentine , Rosin Oil , and Indiarubber Solution .
Liquid .
Metal .
Potential , volts .
Kemarks .
Change .
Wet with turpentine Wet with rosin oil Copper , polished on glass-paper Zinc , polished on glass-paper Copper , polished on glass-paper Zinc , polished on glass-paper . . .
-0*11 + 0-20 -0-12 \#151 ; 0 *02 Note.\#151 ; The plates 11 were compared 1 with one another , !
1 and not with the IS standard plate .
II Hence only the || change of poten-J| tial is given .
Wet with india-rubber solution Copper , polished on glass-paper Zinc , polished on glass-paper : : + 0'02 + 0-07 VIII .
Films formed on Metals by Gases .
S 54 .
The potential of a metal is usually altered by soaking the p ate m a gas other than air .
If the gas be oxygen , this alteration is as a rule only temporary , and apparently depends on the formation of a surface film , or rather on a change in the film which doubtless a rea y exists .
It will be shown that the change of potential of zinc On Contact Electricity of Metals .
ue to soaking in oxygen is nearly equal to that of copper ; hence ie film of oxygen acts like a liquid film ( Chap. VII ) , but its effect less permanent .
Previous experimenters , except Lord Kelvin , ppear to have neglected the existence of these films , and to have .oked only to the nature of the body of gas between the plates .
Of jurse I do not here allude to the " doable-layer " which has been ffered as an explanation of the phenomenon , but to a layer in lechanical and electrical contact with the metal .
If the contact-otential of two metals immersed in a gas were the sum of the lotential-differences between each metal and a skin of gas close to if , ve should have no slope of potential in the body of the gas between he plates .
That a slope does exist , however , is proved by Lord Kelvin 's earliest experiments with the divided ring of copper and \#166 ; inc .
In this connection Dr. Bottomley 's research on contact-jlectricity in high vacua * is of great importance as showing that he volta-potential of metals is not sensibly different in different rases so long as the metals are not chemically affected , and is not sensibly altered by a great reduction of pressure .
The body of the ^as may possibly have some influence , but the variations which have 'requently been attributed to it may usually be more satisfactorily explained as being due to change in the film in contact with the metallic surface .
In the case of a gas which acts vigorously on the metal at ordinary temperatures , the film is permanent , and is probably a solid compound ; in other cases it is not permanent , and hence probably not solid .
S 55 .
In 'Nature ' for 1881 , Lord Kelvin describes some very important experiments on this subject .
As these appear to be but little known I shall give some extracts from his paper before describing my own results .
Under the date November 23 , 1880 , Lord Kelvin says :\#151 ; " I have found that a dry platinum disc , kept for some time in dry hydrogen gas , and then put into its position in dry atmospheric air in the Volta-condenser , becomes positive to another platinum disc which had not been so treated , but had simply been left undisturbed in the apparatus .
The positive quality thus produced by the hydrogen diminishes gradually , and becomes insensible after two or three days .
P.S.\#151 ; On December 24 , 1880 , one of the platinum plates in the Volta-condenser was taken out ; placed in dined oxygen gas for forty-five minutes ; taken out , carried by hand , and replaced in the Volta-condenser at 12.30 on that day .
It was then found to be negative to the platinum plate , which had been left undisturbed .
The amount of the difference was about O'33 of a volt .
The plates were left undisturbed for seventeen minutes in the condenser , and were tested again , and the difference was found to have fallen to 0'29 of a volt .
At noon on the 25tli they were again * ' Brit. Assoc. Report , ' 1885 .
134 Dr. J. Erskine-Murray .
tested , and the difference found to be 018 .
The difference had ha4 tested from time to time since that day , the plates having been left in the condenser undisturbed in the intervals .
The following taS shows the whole series of these results :\#151 ; Electric difference between surfaces of a platinum plate in natural condition m- and a platinum plate after 45 mins ' Dec. 24 , 12.30 , M ... ... ... . .
" 17 7TZt .
I \#187 ; 24 , i2*47 " ... ... ... ... ... ... 0-29 " 25 , noon ... ... ... ... ... ... ... .
018 ' *\gt ; 27\gt ; \#187 ; ... ... ... ... ... ... 0-116 " " 28 , 11.20 a.m ... ... ... ... ... .
0'097 " 31 , noon ... ... ... ... ... ... ... .
0047 Jan,.4 , 11 a.m ... ... .
... ... ... ... 0 042 " 11 , 11.40 a.m ... ... ... ... ... . .
0-020 After detailing some experiments in which the plates were coated wi h the gases by electrolysis , Lord Kelvin concludes : " Thus in the case of polarization by oxygen , as well as in the case of polarization f r ?
gen ' .
the \#174 ; ffect of exPosure to the dry gas was considerably g eater than the effect of electroplating the platinum with the gas byl tne electromotive force of one volt .
" ' The large effects on contact-potential produced by films formed !
ItTirr\#153 ; T^I SiWn ^ thGSe -P---ts of Lord Kelvin 's !
of oLT d Platlniim other metals have the property !
the condor ?
UantitieS 'f gaS in thGir Swface andPthai for the !
DSei V^ the H^uid State ' which would account !
for ^e similarity between the effects of liquids and of gases on^ were which my first experiments on this subject polished on fi^- ' ^ lgh COnductivity " copper .
They had been Potential had ** .glasS'paPer 6 hours before , and their mutual potential had remained constant at 0"02 volt whicK oxvo* ^'Wer P^ate was then put into a glass vessel , into 1 \#187 ; h chw3nMgaS a cylinder , and the o^gen , !
mi ll manufacture , guaranteed 93-95 per cent. 1 thinirhthe of M n8n.a Iy the on , J '"Purity ) , was allowed to stream ] thTSt anf ?
w 'ontzmmg \#171 ; \#187 ; Plate for some minutes , and I laterfhe nf , t .
T68Scl Were th " closed .
Forty-flve minutes 1 the othe/ oW\#153 ; ^11 !
0^ and its P'tential again measured with 1 CoundnJtfm 1 ' ^ remained in air ^ring the interval .
1 Counting time from the moment at which the plate was taken out of 1 On Contact Electricity of Metals .
he oxygen , the potential varied as shown below , the oxygenized date being positive to the other in all cases .
Time .
Contact-potential .
Hrs .
Mins .
Yolt .
0 3 ... .
0-12 0 19 ... .
0-08 0 30 ... .
0-06 17 20 ... .
0-03 Thus the effect of increasing the proportion of oxygen in the surface film was to make the copper more positive .
S 57 .
This variation is in the opposite direction to that found by Lord Kelvin for platinum which has been soaked in oxygen .
In order to make sure that this difference was not the effect of some impurity in the oxygen , I repeated his experiment , obtaining the same result as he had obtained .
This shows that the result given above was not likely to be due to an impurity which had influenced the action of the oxygen .
S 58 .
Without any further treatment the same copper plate was placed in the bell-jar and the oxygen admitted .
After 45 minutes it was taken out and its potential again measured by comparison with the other plate .
The results are given in the following table , time being counted from the moment the plate was taken out of the oxygen:\#151 ; Time .
Contact-potential .
Hrs .
Mins .
Y olt .
0 2 ... ... ... ... .
0T3 0 10 ... ... ... . .
0-08 0 21 ... ... ... ... 0-07 0 29 ... ... ... ... 0-06 The experiment was discontinued before the plate had returned to its original value ; but it would no doubt have done so in a few hours , for the amount of change , and its rate , are almost exactly the same as in the former experiment ( see S 56 ) .
S 59 .
In the following experiment the conditions were somewhat varied .
A copper plate was polished on glass-paper .
It gave , with a standard copper plate , \#151 ; 0'05 volt .
A jet of oxygen was now sent against its surface for 2 or o minutes , and with the same standard plate it now gave \#151 ; 0'06 volt .
It was then left in oxygen for 25 minutes , and on being taken out gave 130 Dr. J. Erskine-Murray .
Time .
Contact-potential .
Hrs , Mins .
Yolt .
0 4 ... ... ... . .
+0-060 0 13 ... ... ... . .
+0035 which shows that its immersion had raised its potential 0*12 voltl how warmed it with a soldering-boit applied to its back .
Wh( ahont 4/ 0 .
( see Chap. IX ) it gave Time .
Contact-potential , Hrs .
Mins , Yolt .
0 16 ... ... ... ... .
+0020 It wras again warmed slightly , and when at about 30 ' C. gave Time .
Contacfc-potentiai , Hrs .
Mins .
Volfc .
When 16 ' C. , ' 42 ... ... " 0-010 4 48 ... ... . .
\#151 ; 0040 lt8TPh''"*ial ?
hns " " " 8 baok *0 very nearly the original value , time .
fn~PAerimen.f S. ^ : ^ cliang\#174 ; requires considerable 2 or 8 mi , T * fa'J 7 Str'DS Jet of Paying on the plate for in still T P " eS n ' aPPredabI\#174 ; \#174 ; \#171 ; oct , while 25 minutes does n'tn 7fn 04,1863 * T '* ''20 toI* ' ( 2\gt ; tba ' gentle heating ( see Chan lx ?
W'M d ' With \#187 ; '"\#187 ; " evygenized copper when'heit was ' " + b\#174 ; remembered that the copper was , alone would h aPP 16 , a ready above the potential to which heat ccTec^on h r ra'S !
lt ; We thiS experiment does not show any Bln .
, as mighuT fat siXb^ppTif\#153 ; d\lt ; 5"Si'7 'f 'X7ge " with ?
standard ^pe^plate \#151 ; b66n poIi8bed on g^s-paper gave 0 81 volt , zinc positive .
takerouteit8Pn'tint,0 iM7gen and * of * for 16 " " " tea .
After being plate , and'was m folto"e ?
!lagam " easnred with the same standard Hr^tLs .
C\#153 ; ta^P\#171 ; \#171 ; ntiJ .
0 2* .
_______ 0-89 0 1'- ... ... ... 0-87 4 35* ... ... ... . .
0-85 s 61 z^nc-a^so_^s More positive after immersion in oxygen , gave ' 6 ZmC Pkte Was aSain polished on glass-paper and +0*73 volt On Contact Electricity of Metals .
137 itli the standard copper .
A short time later it gave + 0-70 volt .
was now put into oxygen for 10 minutes , and after being taken at gave Time .
Contact-potential .
Hrs .
Mins .
Yolt .
0 3 ... ... ... .
+0-80 0 13 ... ... ... . .
+078 17 1 ... ... ... . .
+0-74 S 62 .
A tin plate was polished on clean glass-paper , and gave with copper plate + 0'40 volt , in being positive .
Fifteen minutes later the potential of the plates lad not altered .
I now put the tin into the bell-jar and turned on he oxygen .
After it had soaked for 47 minutes it was taken out and ompared with the same copper plate .
It gave Time .
Contact-potential .
Hrs .
Mins , Yolt .
0 5 ... .
+032 0 13 ... .
+0-32 0 35 ... , +0-35 Thus the variation of tin appears to be in the negative direction , like that of platinum .
S 63 .
A silver plate , polished on glass-paper , gave with a standard copper plate \#171 ; \#151 ; 0'04 volt , and remained constant during an hour .
It was then put into oxygen for 15 minutes , and when taken out its potential was found to be Time .
Contact-potential .
Ilrs .
Mins .
Yolt .
0 3 .
, .
, 0 19 ... .
-o-oi 0 40 ... .
-0-02 0 45 ... .
-0-02 0 59 ... .
-0-03 Thus by immersion in oxygen for 15 minutes it had risen 0'06 and had fallen to nearly its original value in an hour in air .
S 64 .
Silver polished on clean " fine " glass-paper gave with a standard copper \#151 ; 0'095 volt .
138 Dr. J. Erskine-Murray .
After it had been 24 minutes in oxygen it gave in air + 0*015 volt , and 12 minutes later + 0*010 volt , showing that the potential had in 24 minutes in oxygen become 0*11C more positive .
S 65 .
It is noticeable that the amount of change in these experiments on silver is to some extent proportionate to the time of exposure to oxygen .
Thus , in S 63 , 15 minutes in oxygen caused a rise of 0*06 volt ; while , in S 64 , 24 minutes in oxygen caused a rise of 0*11 volt , but there is no doubt a limit to the change .
S 66 .
In searching for an explanation of the temperature-variations described in Chapter IX , I compared them with those given above .
In the case of copper , the oxygen-film variation is in the same direction as^the temperature-variation of copper in air , which suggests the possibility of the latter being caused by an increase of the proportion of oxygen in the film at higher temperatures on account of a greater attraction between the elements .
The same reasoning holds as regards zinc and tin , but the results for silver are in direct opposition ; while the further experiments described in Chapter X , which show that the temperature-variations exist in cases where air ts entirely excluded from the metallic surface , render such an ex-planation ^ very doubtful .
Probably , therefore , the temperature-variation is the more general of the two , it being a change in contact-potential of the metals , which , if they are exposed to a gas , may be complicated by alteration of the surface-film .
IX .
Temperature Variations .
S 67 .
A large number of determinations were made of the variation of contact-electricity with the temperature of the conductor .
This was one by heating one plate while the other was kept cool , and t err potential-difference was measured from time to time as the warm plate was cooling , their temperatures being observed at the same time .
^ In the diagrams , the abscissae represent temperature an e ordinates potential ; each Curve , therefore , shows the tempera ure-variation of the contact-potential of a particular metal .
*or instance , the potential represented by the point which corre-I spends to 16 ' C. on the gold line is Sailed zero in this and the other c ap srsof the present communication .
Thus a standard gold plate ^see S 8 ) at 40 ' 0 .
is 0*04 volt negative to one at 16 ' C. , and an aiumimum plate at 40 ' 0 .
is 0*10 volt , positive to aluminium at 16 ' C. , or 1*20 volts positive to a gold plate at 16 ' C. It must be remembered that unless stated otherwise these varia* On Contact Electricity of Metals .
1 39 \gt ; ns are for metals in air .
In the experiments described in Chapter however , the metallic surfaces were protected by solid non-nducting films , and were , not in contact with the atmosphere ; jvertheless , temperature-variations were found , which in the case silver were actually larger than those which took place in air .
S 68 .
One of the copper plates used was hollow and could be led with wrater and a thermometer inserted , but with the other ates other methods of measuring temperature had to be adopted .
1 some cases the temperature was measured thermo-electrically , hile in others a simpler and more rapid mode of measurement was sed , which , though not very accurate , is quite reliable within 3rtain limits .
S 69 .
By touching the back of the plate I found that its tempera-ire could be judged as " tepid , " " slightly warm , " " warm , " " very rarm , " and so on .
It was found experimentally that these terms orrespond to constant temperatures ; or rather that each term enotes a small range of temperature , the middle point of which xay be taken as corresponding to the term .
In determining the alues of these terms a plate was used in which a thermometer was lserted .
One observer touched the plate with the tips of the rst and second fingers and judged its state , naming it by one f the terms , " warm , " " tepid , " \amp ; c. ; the other observed the tber-lometer , and the temperatures found to correspond to each term are s follows :\#151 ; Cold ... ... ... ... ... ... ... ... ... ... ... ... ... .
16 ' C. Quite cool ... ... ... ... ... ... ... ... ... ... ... 24 Cool ... ... ... ... ... ... ... ... ... ... ... ... ... .
28 Rather cool ... ... ... ... ... ... ... ?
... ... . .
30 Tepid ... ... ... ... ... ... ... ... ... ... ... ... ... .
35 Slightly warm ... ... ... ... ... ... ... ... ... ... . .
40 W arm ... ... ... ... ... ... ... ... ... ... ... ... ... .
47 Yery warm ... ... ... ... ... ... ... ... ... ... ... ... 50 Hot ... ... ... ... ... ... ... ... ... ... ... ... ... ... 53 Yery hot ... ... ... ... ... ... ... ... ... ... ... .
57 Too hot to touch continuously ... ... ... ... .
63 Too hot to touch for more than one second . .
73 This method of measuring temperature is rough and ready , but since the possible errors are within limits of a very few degrees , one only requires to take the average of a considerable number of results in order to arrive at a very fair approximation to the true values .
In experiments on contact electricity in air there are so many possible causes of disturbance that extremely accurate measurement of the temperature is of little use , especially if it require that much time be spent over each reading .
Dr. J. Erskine-Murray .
S 70 .
As a rule the upper plate was heated , in order that tl lower plate might not he affected by draughts of hot air , as won have been the case if the lower had been hot and the upper col Sometimes , during the time of cooling , the npper part of tt apparatus , including of course the upper plate , was removed aftc each observation and was replaced only the moment before the ner In other experiments the upper plate was merely drawn up as far t possible ( about 10 cm .
) to prevent its warming the lower one .
Th temperature of the lower plate was also observed , but as a rule i varied only a very few degrees .
S 71 .
At first I used to apply a hot soldering bolt to the back o the plate in order to heat it , but latterly I heated two or three small blocks of tinned copper and placed them on the back of th upper plate .
By this second method it was possible to observe tht variation of the potential during the rise as well as the fall ol temperature .
S 72 .
In attempting to determine the temperature-coefficient oi copper we were long baffled by curious anomalies .
Sometimes the plats was positive when hot , other times negative , and occasionally it did not vary at all .
The clue to this was found in observing that during one experiment while the copper was cooling it was at first positive , then negative , and then it gradually became positive again , though never quite reaching its original value , hiow it had been found that copper oxide is negative to copper , and that it became temporarily more negative when hot ; hence it was guessed that the successively positive and negative variation must be due to hot clean copper being positive to cold copper , bat that it had finally become oxidized and therefore negative whether hot or cold , the small permanent change being due to the thin coating of oxide formed .
S 73 .
The copper was heated much more gently next time , and gave , the expected result " that clean copper becomes rapidly more positive as its temperature rises , and that , on cooling , its potential returns to its original value unless the temperature has exceeded a certain limit .
If this limit has been exceeded its potential rapidly becomes negative and does not return to its original value .
S 74 .
At ordinary atmospheric temperatures the surface of clean copper remains for a long time almost unaltered either visibly or electrically ( see S 31 ) , and the film which ultimately forms on the surface cannot be pure copper oxide , because the potential of tarnished copper is higher than that of copper oxide obtained by heating in however , the temperature of the copper is raised to about 80 0 .
it immediately begins to oxidize , though heating to a temperature below this limit does not rapidly produce any permanent change , hus there is , as it were , an ignition point for copper and .
oxygen in On Contact Electricity of Metals .
r ; below it , little action takes place ; above it , combination .oceeds vigorously .
S 75 .
I shall now give a specimen experiment .
The two standard \gt ; ld plates gave \#151 ; 0'02 volt , ie minus sign indicating that the upper plate is negative .
I now eated the upper : when " tepid , " iabout 35 ' C. , it gave with the old plate \#151 ; 0'045 volt .
t was next heated further until " veryhot " ( 57 ' C. ) and gave \#151 ; 006 volt .
Hien it bad cooled down to " tepid " ( 35 ' C. ) , it gave \#151 ; 004 volt .
Vhen " cool " ( 28 ' C. ) , \#151 ; 0'03 volt . !
ome hours later , when both plates were cold , they gave as at first \#151 ; 0'02 volt .
Hence the potential of gold which has been washed with alcohol " nd allowed to dry falls temporarily about 0-0016 volt per degree ; entigrade rise of temperature .
S 76 .
Most of the temperature experiments on copper were made with the hollow plate previously mentioned , which was filled with lot water in which the bulb of a thermometer was placed .
In many if the experiments on zinc , and also on aluminium , a thermo-electric irrangement was used , and the results obtained with it do not differ materially from those obtained by the above method ( S 69 ) .
These md other details are noted on the diagram .
S 77 .
The diagram gives the temperature-variations of all the substances studied .
The curves in it are plotted by taking the results for each metal of those experiments which are most free from all complication or cause of doubt .
If the curves be prolonged m the direction of lower temperature they appear to meet somewhere below \#151 ; 200 ' C. , and probably asymptotically to the line representing 04 volt positive to standard plate at 16 ' C. Within their range they show contact-potential-differences diminishing with lowered temperature .
The only apparent exceptions are clean copper , and silver coated wTith glass ; but both their curves are distinctly bent between 16 ' C. and 50 ' C. , so that probably they are directed towards the same point as the others at lower temperatures .
Thus it appears that at about -200 ' C. the contact-potential-differences of metals FbtentiaJ with s L\gt ; r J. Erskine-Murray .
concfUco/ V/ jury copl On Contact Electricity of .
ay vanish , and that a plate of any metal at that temperature would about 0'4 volt positive to a standard gold plate at 16 ' C. S 78 .
The approximate numerical values of the temperature-iriations are given in the following table :\#151 ; Table V. Metal .
Approximate range of temperature .
Potential of metal with standard gold plate .
Both at 16 ' C. Variation of potential per 1 ' C. Standard plate kept always at 16 ' C. Aluminium , polisLed on glass-paper ' p 16\#151 ; 50 volts .
+ 1-10 volts .
+ 0-0043 " waxed 16\#151 ; 40 + 0-98 + 0-0032 " alcohol-washed , dry . .
16\#151 ; 47 + 1-30 + 0-0045 Zinc , polished on glass-paper 15\#151 ; 62 + 0-73 + 0-0013 Lead , scraped 16\#151 ; 75 + 0-58 + 0-0016 Tin , polished on glass-paper 16\#151 ; 33 + 0-52 About\#151 ; O'OOIO Iron , polished on emery-cloth 16\#151 ; 55 + 0-20 -0-0022 , r alcohol-washed , dry 16-65 + 0*28 -0-0007 t_ 0-0007 Silver , polished on glass-paper ... .
16\#151 ; 60 + 0-12 " alcohol-washed , dry 16\#151 ; 70 + 0-16 -00004 " coated with glass 16\#151 ; 30 -0-05 About \#151 ; 0'0035 30\#151 ; 50 , , About \#151 ; 0"0110 Copper , polished on emery-cloth . .
16\#151 ; 30 + 0-04 very small .
7 } )\gt ; 5 ) . .
30\#151 ; 60 , # About + 00015 Gold , alcohol-washed , dry 16-60 o-oo -0-0016 Oxidised copper .
16\#151 ; 65 -o-ll -0-0016 It must he clearly understood that these are true .ariations and not permanent changes in the plate caused by exposure o a high temperature .
K. Elimination of Metal-Air Potentials by Solid Non-conducting Films on the Metallic Surfaces .
S 79 .
As very great differences of opinion seemed to exist as to lie part played by the layer of air which is close to the metallic mrface , I devised a method in which it should be removed and a film \gt ; f solid non-conducting material of a very different chemical nature put in its place .
A copper plate which had been polished on glass-paper uvas filed with a clean dry file which had not been used for any other metal .
Its potential with the standard gold plate being + 0-045 volt .
A zinc plate was prepared in an exactly similar way , and with the copper plate gave + 0-655 volt .
144 Dr* J. Ersldn e-Mu rray .
The copper plate was now gently heated with a bolt until it was It enough to melt paraffin-wax ; the temperature , about 50 ' C. , requitj for this is not sufficient to cause sadden permanent change of copper surface ( see S 74 ) .
Paraffin-wax was then poured on , a[ the plate was filed with its own file while covered with molten wj Thus the fresh surface exposed by the filing came directly in , contact with the wax .
More wax was poured on and the filiu drained off , the plalfe remaining well covered with wax all the whi It was then allowed to cool , and gave with the bare zinc plate +0*555 volt , zinc being positive as before .
Thus the change due to substituti : paraffin-wax for air next the copper is not more than + 0T00 voi I now waxed the zinc in exactly the same way .
When it was qui cool it gave with the Waxed copper + 0*602 volt , showing that waxing the zinc had raised its potential j +0*047 volt .
So , on the whole , the substitution cf wax for air on both copper aD zinc had only decreased their mutual potential by +0*053 volt and the potential of the waxed plates remained nearly constant fc several hours .
The changes due to waxing the plates as given abov were confirmed by the independent comparison of each plate with tb standard gold plate .
It does not follow that even the small change which did occur were due solely to the substitution of wax for air , they may have been caused by slight changes in the surface on accoun of the filing .
\#166 ; S 80 .
An aluminium plate was coated with wax in the waj escribed in S 79 , a knife being used to scrape the surface under the molten wax .
When cold , this waxed plate gave with a bare zinc ob( + 0*36 volt , which is about the usual value for bare aluminium and zinc .
The removal of the air had therefore not appreciably altered the potential .
S }* .
_now warmed the waxed aluminium slightly .
Its potential with , the zinc varied as follows :\#151 ; ^ ^ ... ... ... ... ... +0*41 volt .
35 ' 0 ... .
; ... ... . .
+0-37 " 28 ' O. ... ... ... . .
+0*35 " This gives a variation of about 0*0032 volt per degree centigrade , On Contact Electricity of Metals .
145 ich is nearly the same as the temperature-variation of hare a minium in air .
j 82 .
A plate of silvered glass used with the glass side facing a ndard plate gave almost the same potential as clean silver in air .
; is plate , which was practically silver coated with glass , gave a \#166 ; aperature-variation larger than that of silver in air .
In this case v have glass in contact with the silver surface instead of air , but 13 change does not alter the potential .
It may be mentioned that \gt ; 3 back of the silver film was painted black , and not coated with iss ; but this is of small consequence , since it has been proved by periment that the condition of the back of a plate does not sensibly ect the volta contact-potential ; or , more generally , that if parts of ; olate be in different conditions , the potential observed will be the Ban of the potentials of the different parts , the importance of each rt being proportional to its capacity .
S 88 .
On account of the great attraction of sodium for oxygen , it :emed of interest to measure its potential in circumstances which eluded that gas from the surface of the metal .
In order to effect is two pieces of thin sheet-glass , each about 6 cm .
square , were put to a dish of melted paraffin-wax together with some clean sodium , , da large drop of the sodium was put between the plates of glass-d squeezed out into a small plate of 2 or 3 square cm .
area .
The ass plates , with sodium between them , were taken out of the melted ix and allowed to cool .
Since the glass plates were of much rger diameter than the sodium , the edges of the latter were proofed by the wax which filled up the space between the plates not icupied by sodium .
The flat faces of Ihe sodium were apparently contact with the glass .
The sodium was connected to the elec-ometer by a fine copper wire .
The greatest part of the sodium irface was bright or only slightly tarnished , and it remained in most the same condition for many days , being protected by the ass plates and by the wax which filled the space betwreen them loccupied by sodium .
The first measurements gave sodium 2-86 volts jsitive to a tarnished zinc plate , i.e. , about 3'56 volts asitive to the standard gold plate .
This potential gradually deceased .
S 84 .
Experiments were made to make sure that the result was at due to temporary electrification of the glass .
For instance I reathed on the glass , causing a conducting layer of impure water to \gt ; rm on its surface .
Repeated measurements , made by the usual lethod , showed that the potential at once fell to a small fraction of VOL. LX1II .
M 146 Prof. J. C. Bose .
On Rotation of Plane of a volt , but slowly rose again to nearly its original value as the f1 evaporated .
This shows that the electrification was not a tempor^ one of the glass surface , for that would not have returned to* definite value .
Heating the plate by radiation or washing the g]s with benzol caused the potential to rise further , but in no case the potential quite so high as when the plate was first formed .
4L even more convincing proof that the potential measured was rea-that of the sodium , was found in the fact that the sensibility of t\gt ; apparatus was such as would be given by a plate the size of i sodium .
If the electrification had been on the whole surface of t glass , the sensibility , on account of the larger surface , would ha been at least ten times as great as that observed .
S 85 .
The experiments described in this chapter show that | when two metals are coated with the same non-conductor , such wax or glass , their potential is not sensibly different from that oft bare metals in air ; ( ii ) that temperature-variation still takes pla\lt ; though air be excluded .
These results seem to prove that gaseo films play no essential part in the phenomenon .
On the Rotation of Plane of Polarisation of Electric Wav by a Twisted Structure .
" By Jagadis Chunder Bos M.A. , D.Sc .
, Professor of Physical Science , President College , Calcutta .
Communicated by Lord Rayleig : F.R.S. Received February 14 , \#151 ; Read March 10 , 1898 .
j In my previous papers* I have given accounts of the doub refraction and polarisation of electric waves produced by variot \#166 ; crystals and other substances , and also by strained dielectrics .
A account was there given of the polarisation apparatus with which the effects were studied .
In the present investigation effects ha to be studied which were exceedingly feeble .
The apparatus ha therefore , to be made of extreme sensitiveness ; but the secondar isturbances became at the same time more prominent , and the grea 1_ Cult^ \#174 ; xperieneed was in getting rid of these disturbances .
In one of my communications I alluded to the fact that these secondary disturbances are to a great extent reduced when the radia tois are made small .
The advantage of a large radiator is the com parative ease with which the receiver can be adjusted to respond 1 t e waves , but this advantage is more than counterbalanced by th increased difficulty with the stray radiation and other disturbance ; * " On the Polarisation of the Electric Ray by Double-refracting Crystals , Journal of the Asiatic Society of Bengal , ' May , 1895 , and " On a New Electro polanscope , " ' The Electrician , ' December 27 , 1895 .

