zotero-db/storage/66ILTZMT/.zotero-ft-cache

The
Through Gases

1
k

THE DISCHARGE OF ELECTEICITY THROUGH GASES

THE
DISCHARGE OF ELECTRICITY THROUGH GASES
BY
J. J. THOMSON, D.Sc, F.R.S.
Cavendish Professor op Experimental Physics, Cambridge
WITH DIAGRAMS
WESTMINSTER ARCHIBALD CONSTABLE & CO.
1898

Copyright, 1898, by Charles Scribner's Sons, for the United States of America.
Printed by the University Press, Cambridge, Mass., U.S.A.

THIS BOOK IS DEDICATED
TO
THE MEMBERS OF THE CLASS WHO ATTENDED
THE AUTHOR'S LECTURES AT
PRINCETON,
WHOSE SYMPATHY AND KINDNESS HE CAN NEVER FORGET.

PREFACE
The following pages contain an expansion of
four lectures on Discharge of Electricity through
Gases, given at the University of Princeton, New
Jersey, in October, 1896. In the hope of making the work more useful, I have added some results which have been published between the delivery and the printing of the lectures.
J. J. THOMSON.
Cambridge, August, 1897.

CONTENTS

PAGE
The Discharge of Electricity Through Gases . . 3

Commuuication of a Charge of Electricity to a Gas .

5

.... Electrification of Gas by Chemical Means

7

Electrification of Gases liberated by Electrolysis . . 10

... Formation of a Cloud round Electrified Gas

11

.... Electrification by the Splashing of Liquids

17

Electrification of a Gas by the Aid of Rontgen Rays 29

Uranium Radiation

56

Photo-Electric Effects

61

.... Electrification of Gases by Glowing Metals

81

Electrification in the Neighbourhood of an Arc Dis-

charge

86

The Arc in Hydrogen

89

Conduction through Hot Gases

97

Conduction by Flames

101

Effect of a Discharge in making a Gas a Conductor . 105

Electrolysis in Gases

121

Short Sparks

129

Medium Sparks

130

Long Sparks

130

Transport of one Gas through Another

134

X

CONTENTS

FAOE

Cathode Rays

137

Properties of the Cathode Rays

139

Thermal Effects produced by the Rays

145

Mechanical Effects produced by the Cathode Rays . 146

Action of a Magnet upon the Cathode Rays . . . 150

Paths of the Rays in Different Gases

155

Shape of the Path of the Rays

161

Electric Charge carried by the Cathode Rays . . . 161

Repulsion of Cathodic Streams

169

Diffuse Reflection of Cathode Rays

179

Transmission of Cathode Rays

181

Lenard's Experiments

182

Magnetic Deflection of the Rays

188

Theories of the Nature of the Cathode Rays

.

.

.

189

THE
DISCHARGE OP ELECTRICITY
THROUGH GASES

THE
DISCHAEGE OF ELECTRICITY
THROUGH GASES
The variety and complexity of the electrical phenomena which occur when matter is present in the electric field are in marked contrast to the simplicity of the phenomena when the ether alone is involved. The latter, as far as our knowledge of them extends, are fully explained by laws which
can be expressed mathematically by six very simple differential equations. Indeed, the phenomena in the ether appear to be even simpler in character than we could expect a priori, for the equations we have alluded to as covering all the phenomena which have been observed are true only when the ether is at rest. The agreement of theory and experiment justifies the tacit assumption involved in their use, that the ether remains at rest even when exposed to considerable mechanical forces, as it is when an electric wave is passing through it. The negative result of experiments made to detect the

4 DISCHARGE OP ELECTRICITY THROUGH GASES
motion of the ether in the electro-magnetic field also tends to justify this assumption.
When, however, we consider electric phenomena in which matter plays a part, we find qualities coming into prominence which hardly appear at all as long as we confine our attention to the ether ; thus the idea of a charge of electricity, which is perhaps in many classes of phenomena the most prominent idea of all, need not arise, and in fact does not arise, as long as we deal with the ether alone.
The questions which occur when we consider the relation between matter and the electric charge
— carried by it such as, the state of the matter
when carrying this charge, the effect produced on this state when the sign of the charge is changed
— are some of the most important in the whole
range of Physics. The close connection which exists between electrical and chemical phenomena
— — as shown, for example, in electrolysis indi-
cates that a knowledge of the relation between matter and electricity would lead not merely to an increase of our knowledge of electricity, but also of that of chemical action, and might indeed lead to an extension of the domain of electricity over that of chemistry.
If we wish to study the relation between matter and electricity, the most promising course is to begin with the relation between electricity and mat-

DISCHARGE OF ELECTRICITY THROUGH GASES 5

ter in the gaseous state; for the properties of a gas and the laws it obeys are simpler than for either a solid or a liquid, it is the state of matter which has been most studied, while the Kinetic Theory
of Gases supplies us with the means of forming a mental picture of the processes going on in a gas which is lacking for matter in its other states.

Communication of a Charge of Electricity to a Gas.

One of the most striking phenomena connected

with the electrical properties of gases is the diffi-

culty of directly communicating a charge of elec-

A tricity to a gas in its normal condition.

very

simple instance will suffice to show this: let us

take the case of a charged metal plate which is

insulated so well that there is no leakage of electri-

city across its supports ; let this plate be in contact with air or any other gas at a moderate temperature,

and let it be screened off from ultra-violet light
— and Rontgen rays, then the evidence of the best

experiments we have, proves that under these con-

ditions the plate will suffer absolutely no loss of

charge, provided the surface density of its electrification is less than a certain value. Thus, though

myriads of molecules of the gas strike against the charged surface, they rebound from it without any

electrification. To fully appreciate the significance

of this result, we must remember the very large

6 DISCHARGE OF ELECTRICITY THROUGH GASES

charges that can be carried by the gas under other conditions. The phenomena of electrolysis show

that the charge on each unit of surface of the plate

could be carried thousands of millions times over by

a cubic centimetre of hydrogen at normal pressure

We and temperature.

must, I think, conclude that

the inability of a gas (which when in a certain state

has such an enormous capacity for carrying electri-

city) to take up when in its normal condition any

of the charge of electricity from a body against

which it strikes is very significant and sug-

gestive.

Another fact which exhibits in perhaps even a more striking way the inability of the molecules of

a gas to take up an electric charge is that the

vapour arising from an electrified liquid is quite

We free from any charge of electricity.

owe this

discovery to two American men of science ; for in

1761, Kinnersley of Philadelphia, in a letter to

Franklin, stated that he found the steam arising

from electrified water was not itself electrified.

This result seems to have been overlooked for a

long time ; and in 1883 another American physicist, Blake, made a very complete investigation of the

subject,^ and found that the vapour arising from boil-

ing mercury was not electrified, however strongly

the mercury itself might be. Blake's results have

1 Wiedemann's Annalen, 19, p. 518, 1883.

DISCHARGE OP ELECTRICITY THROUGH GASES 7
been confirmed by Sohncke,^ and quite recently by Scliwalbe.^ In these experiments the molecules of the unelectrified vapour came from or through an electrified surface, and the conditions here seem such that if a molecule could ever get electrified by contact, it would do so in these experiments. Thus we see that when an electrified liquid evaporates,
the electrified particles are left behind, just as the salt in a salt solution is left behind on evaporation.
Electrification of Gas by Chemical Means.
Chemical action is so frequently attended by electrical separation that we might expect that the most likely way to electrify a gas would be to make it one of the parties in a chemical reaction. Examples of electrification, apparently in a gas, produced by this method have been known for a long time, though it is only in recent years that experiments have been made to show that the elec-
trification in these cases is not, in all cases, carried
by dust which may either be present in the gas to
begin with, or produced by the chemical reaction
itself.
One of the earliest-known cases of electrification produced in a gas by chemical action is that of the combustion of carbon. Pouillet ^ found that when
1 Wiedemann's Annalen, 34, p. 925, 1888.
2 Ibid., 58, p. 500, 1896.
3 Pogg. Ann., ii. 422.

8 DISCHARGE OF ELECTRICITY THROUGH GASES
a carbon cylinder was being burnt, the cylinder was negatively electrified, while there was positive elec-
trification in the gas over the cylinder. Lavoisier
and Laplace* showed that the same effect takes place with glowing coal. Reiss "^ proved that there was positive electricity in the air near a glowing platinum spiral. Pouillet * found that when a jet of hydrogen was burnt in air there was negative electrification in the unburnt hydrogen in the jet.
Another case of electrification produced in a gas by chemical action was discovered by Lavoisier and Laplace,* who found that when liydrogen is rapidly liberated by the action of sulphuric acid on iron, a strong positive electrification comes off. This, with other cases of electrification by chemical action, was investigated by Mr. Enright. ^ As a good deal of spray is produced by the bubbling of the hydrogen through the sulphuric acid, it has
been suggested that the spray and not the gas may
in this case be the carrier of the electrification.
To test this point, Mr. Townsend recently made in the Cavendish Laboratory a series of experiments on the electrification produced when a gas
is liberated by chemical action. He found that
1 Phil. Trans., 1782. 2 Reiss, Reibungselektricitat, vol. i. p. 267. 8 Pogg. Ann., ii. 426.
* Memoires de I'Acad. des Sciences, 1782. * Phil. Mag. [5], 29, p. 56, 1890.

DISCHARGE OP ELECTRICITY THROUGH GASES 9
when the hydrogen produced by the action of strong sulphuric acid on iron or zinc was passed
through tubes fitted with plugs of tightly packed glass wool it retained after its passage through these plugs a strong positive electrification, thus showing that no ordinary spray could be the carrier of the electrification.
Using an open beaker, Mr. Townsend^ found that when the mixture of sulphuric acid and iron was heated to about 94° C, there was at first strong positive electrification at the mouth of the beaker when the chemical action was very vigorous and the gas was coming off with great rapidity. But as the temperature fell and the rate of evolution of hydrogen diminished, the positive electrification diminished and finally changed to negative. It was found, however, that the negative electrification, unlike the positive, was completely stopped not only by a plug of glass wool, but even by a layer of wire gauze. This seems to indicate that the negative electrification is carried by coarse spray, while the positive is on the hydrogen, or at any rate on much smaller carriers. If there is positive electrification in the hydrogen, there must be an equal quantity of negative in the sulphuric acid
;
if this is dashed into spray, the spray will be negatively charged. The experiments indicate that
1 Proceedings of Cambridge Philosophical Society, 1897.

10 DISCHARGE OF ELECTRICITY THROUGH GASES

when the gas is not coming off very vigorously,

the negative electrification carried from the vessel

exceeds the

positive carried

off

by the

gas ;

while

when the gas comes off with great rapidity, the

positive electrification carried by it far exceeds the

negative carried by the spray.

Mr. Townsend investigated several other cases

of electrification produced when a gas is liberated
by chemical action. He found that when chlorine

was liberated by the action of hydrochloric acid

on manganese dioxide, the chlorine had a strong

positive electrification, and that when potassium

permanganate was heated, there was positive elec-

trification in the oxygen evolved.

Electrification of Gases liberated by Electrolysis.
Mr. Townsend has also found that when a strong current is sent through a solution of sulphuric acid, so that there is a copious liberation of hydrogen at one terminal and of oxygen at the other, there is
positive electrification in the hydrogen, while the oxygen is either apparently unelectrified or has a very small positive charge. It is perhaps to the point to mention that in this case the oxygen is liberated by a secondary process ; the negative ion is SO4, and the oxygen is liberated by the chemical
action of this ion on the water. The positive electrification in the hydrogen is very much influenced

;
DISCHARGE OF ELECTRICITY THROUGH GASES 11
by temperature. At the ordinary temperature of
the Laboratory there is very little electrification when, however, the temperature is raised to 4050° C, the electrification is very strong.
When a solution of caustic potash is electrolyzed,
there is, on the other hand, very little electrification in the hydrogen, while the oxygen is negatively electrified, though the amount of this electrification is not nearly so large as that on the hydrogen in the preceding experiment. In this case the hydrogen is liberated by secondary chemical action, and the great diminution in its electrification as compared with the previous case seems to show that a gas is more likely to be electrified from electrolysis
when it forms one of the ions than when it is liberated by secondary chemical action. The amount of electrification in the oxygen rapidly increases with the temperature. If oil is added to the caustic
potash solution, the sign of the electrification in the oxygen changes, and the oxygen is positively
instead of negatively electrified. The nature of the electrodes has a considerable influence on the amount of electrification which comes off with the
gas.
Formation of a Cloud round Electrified Gas. Mr. Townsend has discovered that electrified gas possesses the remarkable property of producing

12 DISCHARGE OF ELECTRICITY THROUGH GASES
a fog when admitted into a vessel containing aqueous vapour. This fog is produced even though the vessel is far from saturated with moisture and does not require any lowering of temperature such as would be produced by the sudden expansion of the gas in the vessel in which the fog is produced.
The method used by Mr. Townsend was to send the electrified gas, produced most conveniently by
the rapid electrolysis of a solution of sulphuric acid or caustic potash, according as the gas was to be positively or negatively electrified, through a
series of tubes, some of which were tightly packed with glass wool, whilst others contained strong sulphuric acid through which the gas bubbled. The electrified gas finally passed into the atmosphere, where it formed a dense cloud which slowly settled down.
We can, from observations made on this cloud,
calculate the charge carried by each of the electrified particles of the gas. For by placing the beaker in which the cloud is formed inside an insulated metal vessel of known capacity connected with an electrometer, we can calculate, from the deflection of the electrometer when the electrified gas goes into the beaker, what is the total charge of electricity in the cloud. Thus, if we know the number of water particles in the cloud, we can
We calculate the charge of electricity on each.

DISCHARGE OP ELECTRICITY THROUGH GASES 13

can get the number of particles in the cloud in the following way. The rate at which the cloud falls gives us the radius and therefore the weight of each drop, since

= a^ 4.5 ^-—i

a

7

where v is the velocity with which a drop falls, a
the radius of the drop, /x the coefficient of viscosity
of the gas through which the gas falls. The weight of the whole collection of drops
forming the cloud can be determined by weighing the beaker with the cloud in, then blowing out the cloud and weighing again. Dividing the weight of the cloud by the weight of a drop, we get the number of drops in the cloud, and then dividing the charge of electricity on the whole cloud by the number of drops, we get the charge carried by

each drop.
The results of a series of measurements made by Mr. Townsend were that the radius of the drop in the cloud formed by negatively electrified oxygen was 8. 1 X 10~^ cm., the charge carried by it 3.1 X lO"^*' electrostatic units, the size of the drop formed
by positively electrified oxygen was 6.8 x 10"^ cm., and the charge carried by it, 2.8 X 10"^" electrostatic
units. The difficulty of the measurement prevents us from attaching importance to the slight difference between the positive and negative charges.

14 DISCHARGE OF ELECTRICITY THROUGH GASES
The charge on the drop formed by positively electrified hydrogen was about half that of oxygen.
The charges carried by the electrified particles of oxygen and hydrogen agree within the limits of errors of experiment with those deduced from the
electro-chemical equivalents of these substances.
The calculation of the atomic charge from the electro-chemical equivalent involves a knowledge of the number of molecules in a cubic centimetre of a gas at standard temperature and pressure, and all that we at present know about this number is that it lies between the limits 10^^ and 10^^ (Boltzmann,
Vorlesungen liber Theorie der Gase) ; thus all that
we can infer about the atomic charge from the
values of the electro-chemical equivalent is that it
is between 10"^ and x lO"".
It is probable that in the present state of our knowledge as to the mass of a particle of oxygen or hydrogen, the atomic charge can be determined more accurately from observations on the cloud formed by the electrified particles than from the electro-chemical equivalent. In calculating the charge carried by each particle from the weight
and rate of fall of the cloud, we have assumed
that each drop in the cloud is associated with one and only one charged particle.
In a book on the Applications of Dynamics to Physics and Chemistry (p. 164) I have shown that

DISCHARGE OF ELECTRICITY THROUGH GASES 15
the presence of an electric charge on a drop of water tends to prevent evaporation, and will when the drop is very small neutralize the effect of surface tension, which tends to promote evaporation. It is only, however, for exceedingly small drops that the effects of electrification balance those of surface tension; the size of the drop, when these effects Just neutralize each other, is given by the equation.

IGttT

where a is the radius of the drop, e the charge of
T When electricity on it, and the surface tension.

the radius is greater than the value given by this

equation the drop evaporates more readily than a

plane surface. With the charge deduced from the

preceding experiments the limiting size of the drop

would be

only

10~^ ,

whereas

the

rate

of

fall

shows

that the radius of the drop is in reality about 8 x

10"^ .

This seems to indicate that the drops con-

tain something besides pure water.

The appearance of the cloud and the size of its

particles depends upon the sign of the electrifica-

tion; thus the particles in the cloud formed by

negatively electrified oxygen are larger than those

formed by positively electrified oxygen. This

would seem to indicate that a positively electrified

drop of water evaporated more rapidly than a nega-

16 DISCHARGE OF ELECTRICITY THROUGH GASES
tively charged one of the same size. Few direct
experiments on the evaporation of electrified water surfaces seem to have been made. Mr. Crookes,*
from some experiments he made on this point, came to the conclusion that a negatively electrified surface of water evaporated more rapidly than an unelectrified one. Mascart ^ came to the conclusion that an electrified surface, whether the electrification was positive or negative, evaporated more rapidly than an unelectrified one; while Wirtz^ found that electrification diminished the rate of evaporation of dust-free water, and that positive electrification had more effect than negative.
Even when there is no cloud to be seen, there is evidence that an electrified particle of gas is the centre of an aggregate of some kind which is very large compared with the dimensions of a molecule. This is shown by an experiment made by Mr. Townsend, where dust-free electrified hydrogen was put into a porous pot. The rate at which hydrogen in its normal condition escaped from the pot was determined, and then by connecting the porous pot with an inductor connected with an
electrometer the rate at which the charge escaped from the pot was determined : if the electrified
1 Crookes, Proc. Roy. Soc, 50, p. 88, 1891. 2 Mascart, Comptes Rendus, 86, p. 575, 1878. 3 Wirtz, Wied. Auu., 37, p. 516, 1889.

DISCHARGE OF ELECTRICITY THROUGH GASES 17

particles had been simple molecules, the rate at which the charge escaped would have been equal to the rate of escape of the hydrogen, whereas the experiment showed that in reality the rate at which the charge escaped was only a small fraction of the rate at which the hydrogen escaped. This was not due to the electrified gas sticking in the pores of the pot, for if the gas were afterwards blown out of the pot, all but a fraction of the charge went out with it.

Electrification by the Splashing of Liquids.

One of the most effectual ways of charging a

gas with electricity is by means of the splashing of liquids. It had been known for a long time

that there was something anomalous in the condi-

tion of atmospheric electricity at the feet of water-

falls, where the water fell upon rocks and broke

into spray. Lenard ^ investigated the subject with great thoroughness, and found that when a drop of

water splashed against a metal plate, a positive

charge went to the water, while there was negative

electrification in the surrounding air. This elec-

trical separation is even more marked in the case

A of mercury than in that of water.

very simple

way of showing the effect is to shake some mer-

cury up vigorously in a bottle and then draw off

i Leuard, Wied. Ann., 46, p. 584, 1892.
2

18 DISCHARGE OF ELECTRICITY THROUGH GASES

the air. This will be found to have a negative charge ; this charge is carried neither by dust nor by spray, for it will remain after the air has been sucked through glass wool, or even through a coarse porotis plate. Lord Kelvin^ has shown that the reciprocal process of bubbling air through

water also gives rise to electrification, the air which has bubbled through the water being nega-

tively electrified.
To investigate the laws of electrification produced by splashing, it is often more convenient to

measure the positive charge on the drop rather

than the negative charge in the air. This can

conveniently be done by the arrangement repre-

A sented in Fig. 1.

known quantity of the liquid

to be investigated is allowed to fall through the
A funnel and falls on the metal plate and sau-

cer C, which is carefully insulated and connected with one pair of quadrants of an electrometer ; a

blower sends a strong current of air over the
plate and blows away the air from the neighbour-

hood of the plate, so as to prevent the negative

electrification in the air interfering with the indications of the electrometer. The deflection of the
electrometer when a given quantity of liquid falls on the plate may be taken as an indication of the

electrification produced by the splashing of the

1 Kelvin, Proc. Roy. Soc, 47, p. 335, 1894.

DISCHARGE OF ELECTRICITY THROUGH GASES 19
liquid. By the aid of an instrument of this kind we can investigate the circumstances which affect
the electrification. This electrification is influ-
SlectromeAsr
2'oEart^

Fan.

Ta -Xlectrometer

Fig. 1.
enced to a remarkable extent by minute changes in the composition of the liquid. Lenard found that the electrification in the air in the neighbourhood of the splashing place, which was very marked with

20 DISCHARGE OF ELECTRICITY THROUGH GASES
the exceptionally pure water of Heidelberg, was almost insensible with the less pure water of Bonn ; while the splashing of a weak salt solution electrified the air in the neighbourhood positively instead of negatively as when the water was pure. Thus, while the splashing of rain electrifies the air negatively, the breaking of waves on the sea-
shore electrifies it positively.
In some experiments which I made on this subject, I found that the effects produced by exceedingly minute traces of some substances were most
surprisingly large. Kosaniline, for example, is a substance which has very great colouring power,
so much so, indeed, that the colour imparted to a large volume of water by a small quantity of
rosaniline is sometimes given as an instance of the extent to which matter can be subdivided yet I
;
found that the change produced in the electrification, due to the splashing of drops, was appreciable in a solution so weak as to show no trace of colour. The effect of fluorescent solutions on the electrification of drops is especially great, but different kinds of solution act in different ways. Thus rosaniline and methyl violet reverse the
— effect, that is, they make the electrification on the
drop negative in the air; while eosine and fluorescein, on the other hand, increase the normal effect,
— that is, they make the drop more strongly elec-

DISCHARGE OP ELECTRICITY THROUGH GASES 21

trifled positively than a drop of pure water, and pro-

duce a greater negative electrification in the air.
The electrical effects of weak solutions are much

greater than those

of strong ones ; in

fact, strong solu-
ftions of all sub-
stances tested gave

little or no elec-

trical effects by

splashing.

The

effect produced by

the addition of for-

eign substances to
water may be rep-

resented by curves

in which the ab-

scissae are propor-

tional to the

amount of the sub-

stance added to

the water, and the

ordinates to the electrification pro-
by splashing.

Fig. 2
We find that these curves

Lduced are of three types, a, y8, y. Fig. 2. Curves of the type a represent the behaviour of

solutions of phenol, eosine, fluorescein, where the

;
22 DISCHARGE OP ELECTRICITY THROUGH GASES
addition of a small quantity of the substance increases the electrical effect.
Curves of the type ^ represent the behaviour of solutions of potassium permanganate, chromium
binoxide, hydrogen peroxide, rosaniline, and methyl violet; here the addition of the substance begins by diminishing the electrification and finally re-
verses it. Curves of the type y represent the behaviour of
solutions of zinc chloride, hydrochloric and hydriodic acids, and, in fact, of most inorganic salts and acids ; here the addition of the substance produces a diminution but not a reversal of the electrification.
The addition of strong oxidizing agents to the water seems, on the whole, to tend to reverse
— the normal effect, that is, it tends to make the — electrification in the air positive, while the addi-
tion of reducing agents seems to increase the normal effect.
The amount and even the sign of the electrification depends on the gas through which the drops fall ; thus I found that when the drops fell through steam no electrification was produced ; while when they fell through hydrogen which had been very carefully freed from any trace of air, the electrification in the hydrogen was very small but positive
that is, of the opposite sign to the electrification in air.

;
DISCHARGE OF ELECTRICITY THROUGH GASES 23
There does not appear to be any appreciable separation of the positive and negative electrification until the drop splashes against the plate thus a drop of rain falling through the air does not appear to leave a trail of negative electricity behind it.
We can explain the electrification produced by
the splashing of drops if we suppose that at the surface of each drop we have a double electrical
layer ; that is, a spherical layer of electrification of one kind surrounded by a closely fitting concentric layer of electricity of the opposite sign, the charge of positive electricity on one layer being equal to the charge of negative electricity on the other. In the case of pure water the positively electrified layer is next the water, the negatively electrified one next the air.
When the water-drop strikes against the plate,
this double layer gets very roughly treated whilst the great increase which takes place in the area of
the drop is in progress. The double layer gets
torn asunder, the negative coating staying in the air, the positive one on the drop.
The suggestion that the electrification due to the splashing of drops is produced by the rupture of an electrical double layer is due to Lenard ; and though it is perhaps somewhat surprising that the disturbance produced by splashing should produce

24 DISCHARGE OF ELECTRICITY THROUGH GASES

SO great a separation of the coats of the layer, the

existence of the layer is proved in many cases by

the definite contact difference of potential which

occurs at the junction of two dissimilar substances.

Though this contact difference of potential has

only been measured when one of the substances is

a conductor, we must, if we accept the electrifica-

tion due to the splashing of drops as evidence of

the existence of this double layer, suppose that the

double layer is almost invariably present at the

We junction of different substances.

find, for

example, that electrification is produced by the

splashing of drops of substances as unlike as

paraffin-oil, water, and mercury.

The electrification produced by splashing shows

that the coatings of the layer can be separated by

mechanical means ; but if this is the case, and all

substances are covered with an electrical double

layer, the production of electrification by rubbing

is very simply explained.

The existence of an electric double layer at the

surface of separation of two bodies implies the

existence at the surface of the body of a layer of

matter which is neither quite of the nature of one

substance or of the other ; it thus implies a certain

amount of chemical combination, or rather the first

stages of an uncompleted chemical combination,

since complete chemical combination is electrically

DISCHARGE OF ELECTRICITY THROUGH GASES 25
neutral. The phenomena of electrification by splashing shows that this must occur in cases when the two substances are not known to exert any chemical action when tested under ordinary con-
ditions.
We must remember, however, that the layers of
substance next the surface are affected by circumstances which are not allowed to exert any influence in ordinary chemical operations. One of the most important differences is the effect which surface tension can exert on chemical actions taking place in the surface layers. The production of a drop of liquid in air involves an increase in potential energy equal to the area of the drop multiplied by the surface tension between the drop and the air; this in the case of water and air would be about 78 ergs per square centimetre of surface of the drop. The production of this energy will be a tax laid upon a thin layer of water whose thickness is comparable with the range of molecular forces : let us take this as 10"^ then 78 ergs per centimetre would if converted into heat be able to raise the temperature of this layer by about 200° C. •, a saving then of a fraction in the surface tension might be sufficient to convert the chemical action which effected this saving from one attended by an absorption into one attended by an evolution
of heat.

26 DISCHARGE OP ELECTRICITY THROUGH GASES
Now Lord Kayleigh ^ has shown that any dimi-
nution in the abruptness of transition between two surfaces diminishes the surface tension, and therefore the energy required for the production of the surface. Now, in the case of the drop of water it seems reasonable to suppose that the abruptness of the transition would be mitigated if there was between the water and the air a layer formed of a quasi-compound of the two substances. This layer must not be too thin, for the effect of a transitional layer in diminishing the surface tension diminishes very rapidly with the thickness of the layer when this sinks below a certain value.
The double electric layer must, since it can be separated by mechanical means, have the positive and negative charges separated by a distance much greater than that which separates the positively from the negatively electrified atoms in an ordinary molecule. On the view we have been discussing, the reason the coatings of the electric layer remain at a considerable distance apart is that any further approach of these coatings would, by diminishing
the thickness of the transitional layer, increase the
potential energy due to surface tension more than it would diminish the potential energy due to the
electrostatic attraction of the constituents of the layer.
1 Kayleigh, Phil. Mag., 33, p. 468, 1892.

;
DISCHARGE OF ELECTRICITY THROUGH GASES 27
With regard to the great effect produced by the addition of a small quantity of some substances to the water, we must remember that the layer whose
disruption produces, on this theory, the electrification in the splashing of drops, is due to a partial chemical combination between the liquid and the air. Anything which alters the conditions under
which this combination takes place may be expected to alter the electrification produced by the
drops. Now tlie conditions are altered when a
foreign substance is added to the water, for the water which is required for any compound of water and air has to be torn from, if not a chemical compound, yet that connection which exists be-
tween a salt and a solvent ; it may require more work to tear the water away from this connection
than could be compensated for by the diminution in the surface tension, and thus the layer might not be formed. In those cases where the addition of the foreign substance reverses the electrification, a layer of a different kind must be formed
this may consist of a combination with the air
of the substance added to the water ; or possibly a compound into which all three enter.
The magnitude of the effects produced by small
quantities of foreign substances is very surprising, but it must be remembered that the chemical actions which produce the layer, such as that be-

28 DISCHARGE OF ELECTRICITY THROUGH GASES
tween water and the air, are not of a very energetic character, and might therefore be easily affected by apparently trivial circumstances.
The amount of electrification produced by the fall of a drop from a given height depends not merely on the charge of electricity per unit area of the double layer, but also on the ease with which the layers can be separated by the shock produced by the impact of the drop with the plate. This makes the interpretation of the results of experiments on the electrification produced by drops, ambiguous, as we cannot tell whether, when we get a big effect, it is due to a highly charged double layer, or to one with a small charge but of which
the coatings are exceptionally easily separated. In connection with this subject it is interesting
to find that Kenrick,^ who measured the potential differences between various liquids and air, found that the substances which when added to water produced a large effect on the potential difference were those which produced a large effect upon the
surface tension.
Holmgren^ has shown that when two wet cloths are rapidly brought together and then pulled
1 Keurick, Zeitschrift fiir physikalische Chemie, 19, p. 625, 1896.
2 Holmgren, Sur le Developpement de I'electricite au contact tie I'air et de I'eau. Societe pliysiograpliique de Lund,
1894.

DISCHARGE OF ELECTRICITY THROUGH GASES 29
quickly apart, electrification is produced, the positive electrification being on the cloth, the nega-
tive in the air. He also found that when the area
— of a surface of water is changing rapidly, as, — for example, when ripples are travelling over it,
there is again electrification, the water being positively, the air negatively, electrified.
The fact that electrification is produced by a sudden diminution in the area of a water surface maybe of considerable importance in meteorology
;
for the coalescence of small drops of water to form a large one would be accompanied by a diminution in the water surface and would give rise to electrification. Thus, as Sir George Stokes has suggested, the large drops of rain which fall in a
thunder-storm may be the cause rather than the
consequence of the storm.
The production of electrification by the flow of steam through pipes, as in Armstrong's hydro-electric engine, is an example of electrification by the splashing of drops, as Faraday showed that for this electrification to occur it was essential
that water-drops should be present in the steam.
Electrification of a Gas hy the Aid of Rontgen Rays.
Of all the methods by which we can put a gas into a state in which it can receive a charge of

30 DISCHARGE OF ELECTRICITY THROUGH GA^ES
electricity, none is more remarkable than that of the Rontgen rays. These rays, when they pass through a gas, turn it into a conductor and enable it to receive a charge of electricity, and the gas retains its conductivity and its power of being charged for some time after the rays have ceased
to pass through it. It may at the outset be worth
while to distinguish between the effects of Rontgen rays and those of rays of ultra-violet light. The latter only make the gas a conductor when the light is reflected from a fluorescent substance or from the surface of a metal immersed in the gas ; and the gas is only able to discharge a charged body in its neighbourhood which is not illuminated by the ultra-violet rays when the charge on the body is positive. The Rontgen rays, on the other hand, make the gas through which they pass a conductor independently of any reflection of the rays, and the gas when in a conducting state is able to discharge negatively as well as positively charged bodies when it comes into contact with them.
To show this effect, the apparatus used to pro-
— duce the rays the Ruhmkorff coil and the — exhausted tube is placed inside a closed iron
tank; the iron of the tank stops the rays, so to allow them to get out a hole is cut in the tank, and this hole is covered with a window made

DISCHARGE OF ELECTRICITY THROUGH GASES 31
of a thin sheet of aluminium which allows the rays to get through. The exhausted tube is put into such a position that the rays emanating from it strike against the window and emerge into the air outside the tank. If now an insulated charged body is placed outside the tank in such a position that when the coil is in action the rays play upon the body, it will be found that even though the insulation is perfect when the coil is not working, yet as soon as the coil is started and the Rontgen rays fall upon the charged body, the charge, whether it be positive or negative, will
rapidly disappear.^
Next take the insulated body and place it near the tank, but in such a position that the rays do not fall upon it ; the insulation will be unimpaired even when the coil is in action ; now, with a pair of bellows blow on to the charged body the air from above the aluminium window. This is air through which the Eontgen rays have passed ; the charged body will now rapidly lose its charge. The air traversed by the rays has thus retained its conductivity during the time taken by it to pass from the neighbourhood of the aluminium window to the charged body.
To examine in greater detail the properties of
1 For a qualification of this statement when the initial
charges are very small, see p. 54.

32 DISCHARGE OF ELECTRICITY THROUGH GASES

gas through which Kontgen rays have passed, the

author and Mr. E. Eutherford ^ used the following

A arrangement.

closed aluminium vessel was

placed in front of the window through which the

A rays passed.

tube through which air could be

blown by a pair of bellows led into this vessel; a

plug of glass wool was placed in this tube to keep

out dust, and a gas meter was placed in series

with the tube to measure the rate at which the air

passed through. The air left the aluminium vessel

through another tube, at the end of which was

placed the arrangement for measuring the rate

at which electricity leaked through the gas. This

was usually a wire charged to a high potential

placed in the axis of an earth-connected metal tube

through which the stream of gas passed ; the wire

was connected with one pair of quadrants of an

electrometer. The wire was carefully shielded

from the direct effect of the rays, and there was no

leak unless a current of air was passing through

the apparatus ; when, however, the current of air

was flowing, there was a considerable leak, showing

that the air after exposure to the rays retained

its conducting properties for the time (about .5

second) it took to pass from the aluminium vessel

to the charged electrode.

1 J. J. Thomson and E. Rutherford, Phil. Mag. [5] 42, p.
392, 1896.

DISCHARGE OF ELECTRICITY THROUGH GASES 33

Heating the gas when in this state does not

seem to impair its conductivity to any considerable extent ; for when we inserted a piece of porcelain

tubing between the aluminium vessel and the testing apparatus, we found we could raise the tube to

a white heat without affecting the conductivity of

the gas, though the gas after coming through the

tube was so hot that it could hardly be borne by

the hand.

If the gas in its passage from the aluminium

vessel to the tester was made to bubble through

water, every trace of conductivity seemed to

disappear.

The gas also lost its conductivity when forced

through a plug of glass wool, though the rate of

flow was kept the same as in an experiment with-

out the plug when there was a rapid leak. If the

plug was inserted in the system of tubes before

the gas reached the vessel where it was exposed to

the Eontgen rays, the conductivity was not dimin-

ished if the rate of flow was kept constant. These

experiments seem to show that the structure in the

gas in virtue of which it conducts is such that

it is not able to pass through the fine pores in a

A plug of glass wool.

diaphragm of fine wire

gauze or muslin does not seem to diminish the

conductivity of the gas passing through it.

The effect of passing a current of electricity

3

34 DISCHARGE OF ELECTRICITY THROUGH GASES
through the gas on its way from the aluininiura vessel where it is exposed to the Rontgen rays to the place where its conductivity is tested is very suggestive. This was done by inserting a metal tube in the circuit and fixing along the axis of this
tube an insulated wire connected with one terminal of a battery of small storage cells, the other terminal of this battery was connected with the outer
tube ; in this way a current of electricity was sent
through the gas as it passed through the tube. The passage of a current from a few cells was sufficient to diminish greatly the conductivity of the gas passing through the tube and by increasing the number of cells the conductivity of the gas could be entirely destroyed. Thus the peculiar state into which the gas is thrown by the Rontgen
rays is destroyed when a current of electricity
passes through it. It is the current which effects this destruction,
not the electric field; for when the central wire was enclosed in a glass tube so as to stop the
current but maintain the electric field, the gas passed through with its conductivity unimpaired.
The current produces the same effect on the gas as it would produce on a very weak solution of an electrolyte. For imagine such a solution to pass through the tubes instead of the gas, then if enough electricity passed through the solution to decom-

DISCHARGE OP ELECTRICITY THROUGH GASES 35 pose all the electrolyte, the solution when it emerged would be a non-conductor, and this is precisely what happens in the case of the gas. The analogy between a dilute solution of an electrolyte and gas exposed to Kontgen rays is complete over a wide range of phenomena.
The fact that the passage of a current of electricity through a gas destroys its conductivity ex-
plains a very characteristic property of the conduction of electricity through gases exposed to Eontgen rays ; that is, for a given intensity of radiation the current through the gas
does not exceed a certain maximum value, whatever the electro-motive force may be,^
£ M.F,
Fig. 3.
— the current, as it were, gets saturated. The
relation between the electro-motive force and the current is shown in the above curve, Fig. 3,
1 J. J. Thomson, Nature, Apr. 23, 1896.

36 DISCHARGE OF ELECTRICITY THROUGH GASES
where the ordinates represent the current, and the abscissae, the electro-motive forces. For small values of the electro-motive force the curve is straight, showing that the conduction follows Ohm's law; as the electro-motive force increases, the current increases, but at a slower rate than the electro-motive force. The rate of increase of the current gets slower and slower as the electromotive force continually increases, until finally the current reaches a constant value, when an increase in the electro-motive force produces no effect upon the current. It is evident that there must be a limiting value of the current if the passage of the current destroys the conductivity of the gas, and
that the maximum current will be the current
which destroys the conductivity at the same rate as the Rontgen rays produce it.
If we suppose, as seems proved by some of the
preceding results, that the conduction through a gas exposed to the Rontgen rays is electrolytic in character, the rays dissociating the gas and producing a supply of oppositely charged ions, which
carry the current, we can easily find an expres-
sion connecting the electro-motive force with the current. Let n be the number of ions with charges of one sign in unit volume of the gas, q the rate at which these are produced by the rays ; if e is the charge carried by one of the ions, then the passage

DISCHARGE OF ELECTRICITY THROUGH GASES 37

E through the gas of a quantity oi electricity will E destroy je of these ions. The oppositely charged
ions will come into collision and will tend to recombine. The number of recombinations we shall assume to bear a constant ratio to the number of collisions. The number of collisions is proportional to n"^. Let a nP' be the number of ions in unit volume which recombine in unit time ; let i be the current through unit area of the gas, I the distance between the electrodes. Then we have

d-n—

=

7

—

a w/

—

—i •

dt ^

le

(^ 1)^

So that when the gas is in a steady state,

o = q-an^-^^'

(2)

Let us assume that the velocity of the ions is
U proportional to the potential gradient ; let be the
sum of the velocities of the positive and negative

velocities when the potential gradient is unity.

E Then if

is the difference of potential between

the plates, the sum of the velocity of the positive

and negative ions is £J U/l, and hence

. 7lUU

li

Substituting this value of n in equation (2), we

get

_

al^v^

i

38 DISCHARGE OF ELECTRICITY THROUGH GASES
or

We U see from this equation that when is infinite, i

approaches the value leqi if we call / the limiting

value of i when HJ is infinite, we have

and from (3)

= 1 qle,

(4)

(5)

or writing G for al^ejW we get

^-'=

(C)

E^

as the relation between the current and the electro-

motive force. The quantity a is not one that we

can measure directly ;

we

shall try to

express

it in

terms of quantities which can be determined ex-

T perimentally. Let be the time which elapses

after the rays have ceased to pass through the gas

before the number of ions falls to one half its

initial value, no current passing through the gas

in the interval, then we have by equation (1)

dn

„

integrating this equation, we get

DISCHARGE OF ELECTRICITY THROUGH GASES 39

where t is the time which has elapsed since the
N rays were turned off, and is the number of ions

at the instant when the rays ceased to pass through

the gas ; we see by equation (1), since no current is

passing through the gas in this case, that

q = aN\

(8)

Now when t = T^n = \ N.

Hence, by (7)

N = aT)

or Thus,

i;i'=-
le
T^q

Making this substitution, equation (5) becomes

1 V^

l"

— *)

J^i rJ^'i JJ-2

(9)

This equation, when i is small compared with /, its limiting value, becomes

I~ I'

(10)

thus the current is proportional to the electro-mo-
tive force.
The formula (9) thus coincides with experiment
in indicating a limiting value to the intensity of the
current, and in making the conductivity obey Ohm's

40 DISCHARGE OF ELECTRICITY THROUGH GASES

law for small values of the current. The truth of

the formula was much more severely tested by the

determination by Mr. Kutherford and myself of a

long series of determinations of the current corre-

sponding to electro-motive forces varying from 1 or

2 volts to the electro-motive force required to pro-

duce saturation. The experiments were made with

air, hydrogen, coal gas, chlorine, sulphuretted hy-

drogen, and mercury vapour ; and the intensity of the Eontgen rays as well as the nature of the gas

was varied. The results of these experiments are

given in the paper by Mr. Rutherford and myself

already quoted ; the agreement between theory

and experiment was very close, never exceeding an

amount which might fairly be attributed to errors

of experiment. There are some points in the form-

We ula which call for special attention.

see from

equation (4) that, with a constant intensity of radiation, 7, the limiting current, is proportional to Z, the
distance between the electrodes. Thus, when we

approach saturation, the current will increase as

the distance between the electrodes increases, and

we get what is at first sight the paradoxical result

that a thin layer of gas offers a greater resistance

to the passage of a current than a thicker one.

This is, however, easily explained if we remember

that the current destroys the conductivity of the

gas, and that as in a thicker layer there are more

DISCHARGE OF ELECTRICITY THROUGH GASES 41

conducting particles than in a thinner one, the cur-
rent required to destroy them all will be greater. The experiments show that the effect of the dis-
tance between the electrodes (two parallel plates) on the current is very marked. The following tables show the result of some experiments on this
point.

Potential Difference between Electrodes, 60 Volts,

Distance between electrodes in millimetre.
.1
.12 .25
.5
1 1.5 3 8

Current (arbitrary
scale).
9 15 21 37 50 62 91 110

The next table contains measurements with a small

potential difference of 1.3 volts.

.25

10

J5

32

2

48

3

53

8

53

18

40

We see, as we should expect from the theory,

that the effect of distance is not nearly so well

42 DISCHARGE OF ELECTRICITY THROUGH GASES

marked when the potential difference is small as

when it is large.

The measurement of the rate of leak when the

current is saturated enables us to form an estimate

of the number of ions produced by the liontgen

rays in unit time, as in this case the number of

ions produced by the rays is equal to the number

destroj^ed by the current. Let us take the case of

an experiment with hydrogen : when the current

was saturated, the rate of leak between two plates,

each about 10 sq. cm. in area and 1 cm. apart, was

in one second about equal to the quantity required

to raise a capacity of 30 cm. to the potential of

1 volt. Thus, the quantity of electricity passing between the plates in one second was about 10~i

x electrostatic units, or 1/3 10^^ electro-magnetic

units ; and this quantity is sufficient to neiftralize

all the ions produced in one second by the Ront-

Now gen rays.

1 electro-magnetic unit of electricity

can set free about 10~* grammes of hydrogen, or

about 1 cc. of the gas at standard temperature

X and pressure; hence, 1/3

10^^ units of electri-

city would be carried by 1/3 10" cc. of hydrogen,

so that if the ions in the gas carry the same

charge as they do in an electrolyte, the volume

occupied by the ions produced in one second by

X the Rontgen rays would only be 1/3 10" cc. at

standard temperature and pressure. The volume

DISCHARGE OF ELECTRICITY THROUGH GASES 43

of the gas exposed to the rays was, however, about 10 cc, so that in this experiment the amount of
X gas ionized was only 1/3 10^^ of the amount of
gas exposed to tHe rays. This result shows that it is not surprising that experiments to see if any alteration in the volume of a gas under constant pressure was produced by the Itontgen rays should have led to negative results. The conductivity of
iodine and mercury vapour is much greater than that of hydrogen, so that the number of ions produced by the rays would be greater ; but even in the case of the best-conducting vapours the num-

ber of ions produced is an exceedingly small fraction of the number of molecules in the gas.
When the current through the gas is small com-

pared with the saturation current, we have by

equation (10),

i _ EUT I~ '
I''

Now EXJjl is the sum of the velocities of the
positively and negatively charged ions in the elec-

tric field. Thus, this equation leads to the result
that when the current is small, the ratio it has to the maximum current is equal to the ratio of the

space passed over by the charged ions in the time
T to the space between the electrodes. In an ex-
periment when I was about 1 cm., the rate of leak through air when the potential difference was 1

44 DISCHARGE OF ELECTRICITY THROUGH GASES

volt was about 1/30 of the maximum rate of leak;

T hence, the charged ions must in the time have

moved through about .03 cm. The time T depends

A upon the intensity of the radiation.

rough esti-

mate gave T as about .1 second in the experiment

under discussion ; this would make the sum of the

velocities of the ions in air, under a potential gra-

dient of 1 volt per cm., equal to .33 cm. / sec.^ This velocity is very large compared with the velocities

of the ions in the electrolysis of liquids ; it is,

however, small compared with the velocity with

which an atom carrying an atomic charge of elec-

tricity would move under the potential gradient

through a gas at atmospheric pressure. If we cal-

culate by the Kinetic Theory of Gases this velocity'',

we find that for air it is of the order 50 cm. / sec. The magnitude of this velocity compared with that

of the ions seems to show that the ions in a gas

exposed to Eontgen rays are the centres of an

aggregation of a considerable number of molecules.

The result is thus in accordance with the view to which we are led by many other independent con-

siderations, that there is a structure in a gas-con-

veying electricity whose structure is exceeding

coarse compared with a structure consisting of

single molecules uniformly distributed.

1 More accurate experiments have shown that this is about 3.2 cm. per second.

DISCHARGE OF ELECTRICITY THROUGH GASES 45

When the intensity of the Kontgen rays is

>ns^°^

allowed, the alteration in the intensity of the cur-
E rent is not the same at all points on the i and
curve. When the intensity of the rays is changed,

the saturation current is increased or diminished

in a larger proportion than the current for small

rXx^

/ /^ §

/ yV*^

*i.

jf

* **

/

E.M.F.
Fig. 4.

electro-motive forces. This is shown by the above
E diagram. Fig. 4, which represents the i and
curves for conduction through chlorine gas for different intensities of the Eontgen rays ; the weak radiation was got by interposing between the bulb and the gas a thick plate of aluminium. In the figure the ordinates for the weak radiation
have been increased so as to make the ordinate for

46 DISCHAKGE OF ELECTRICITY THKOUGH GASES
the saturation current of the weak radiation the
same as that of the strong. When this is done
the rest of the " weak " curve is above that of the strong, showing that the diminution in the intensity of radiation has affected the saturation current to a greater extent than the weak current.
These results follow at once from equation (4). /, the saturation current, is given by the equation
1=. qle,
so that it is proportional to the number of ions produced by the rays in unit time. On the other hand, when the current i is small compared with /, we have
Wb E
.
thus the small current is only proportional to the square root of q, and therefore does not vary so quickly with q as the saturation current.
If the number of ions produced by the Kontgen rays in a gas were proportional to the number of
— molecules of the gas in unit volume that is, to the — pressure of the gas, then q would be proportional
to the density of the gas thus the saturation cur;
rent would be directly proportional to the pressure, while if the current were a long way from saturation it would be inversely proportional to the square root of the pressure, as C/" is inversely pro-

DISCHARGE OF ELECTRICITY THROUGH GASES 47
portional to the pressure. The results obtained bydifferent observers as to the effect of pressure on the conductivity of a gas exposed to Rontgen rays are not in accordance. M. Perrin, using an elec-
tro-motive force sufficiently large to saturate the
gas, found that the conductivity was proportional to the pressure; in some experiments made by Mr. McClelland and myself, and in others by M. Humuescu, the conductivity varied as the square root of the pressure. This discrepancy arises in M. Perrin's opinion from an effect produced by the rays at the surface of a metal on which they fall.
Different gases conduct when under the influence of Rontgen rays with very different degrees of facility. Using a constant intensity of radiation, the saturation current for mercury vapour was about twenty times as large as that for air, while that for air was half as large again as the saturation current through hydrogen. Hydrogen has a smaller saturation current than any gas I have tried. Experiments made on the following gases showed that their saturation currents were in the order, Hydrogen, Nitrogen, Air, Oxygen, Carbonic Acid, Sulphuretted Hydrogen, Hy-
drochloric Acid, Chlorine, Mercury. The magnitude of the saturation current does
not seem to depend altogether upon the density of the gas ; thus it is very much greater for HgS

48 DISCHARGE OF ELECTRICITY THROUGH GASES
than for air, although the densities of these gases are nearly equal. Chlorine, Bromine, Iodine, Sulphur compounds and Mercury vapour have large saturation currents, and these elements and their
compounds have when in their gaseous state abnor-
mally high specific inductive capacities in compari-
son with their valencies. The high conductivity of mercury vapour is very remarkable, as this gas is often regarded as monatomic.
Mr. Rutherford^ has found that the absorption of the Rontgen rays by different gases is proportional to the saturation current through these
— gases, that is, if the intensity of the radiation
after passing through a stratum of thickness, d, be represented by e~^'^, then X is proportional to the saturation current through the gas. If each of the ions carry the same charge, then the saturation current is proportional to the number of molecules of the gas dissociated in unit time, so that
another way of expressing this result would be
that the dissociation of a molecule into ions causes the same absorption of Rontgen rays, what-
ever may be the nature of the gas. The proportion between the conductivities of
different gases depends upon the electro-motive force used thus for small electro-motive forces the
;
conductivity of hydrogen is greater than that of
1 Rutherford, Phil. Mag.
n

DISCHARGE OF ELECTRICITY THROUGH GASES 49
air, though the saturation current through air is
much greater than that through hydrogen. This is shown in the following diagram, which repre-
E sents the / and curves for hydrogen and air.
The curves intersect, the hydrogen curves being
above the air curve for small values of the electro-
motive force and below it for large values. The saturation current depends merely on the number

£. M. F.

Fig. 5.

of conducting particles produced by the rays while the current in the early part of the curve depends upon the space described by the conducting parti-
T cles in the time (see equation 10) ; and we infer
from these curves that more ions are produced by the rays in air than in hydrogen, but that the product of TJ, the velocity of the ions, and T, a
4

50 DISCHARGE OF ELECTKICITY THROUGH GASES

time which is proportional to the time the ions linger after the rays are cut off, is greater for hydrogen than it is for air.
In deducing equation (6), we assumed that the only way in which when a current is not passing through the gas the ions disappeared was by recombining ; if the gas is contained in a vessel it may be that some of the ions would give up their charges when they struck against the sides of the vessel. The number so giving up their charges would be proportional to n and not to ii^^ and would equal kn where A; is a coefficient depending on the shape and size of the vessel. Thus equation (1) would, if this term were included, become

dn

i

so that the corresponding relation between the current and the electro-motive force would be

a i^

ki

i

The effect of the term we are considering would be to diminish the conductivity when the electromotive force is small ; it does not affect the limiting
value of the current. Mr. Eutherford^ has investigated the electrifi-
cation of a gas exposed to Eontgen rays. The
1 Rutherford, Phil. Mag., 43. p. 241, 1897.

DISCHARGE OF ELECTRICITY THROUGH GASES 51

arrangement used is represented in Fig. 6. The gas was blown rapidly through the tube ; after passing out of this tube the gas rushed along the wire placed along the axis of the tube through which the gas was blown into the vessel. The gas as it passed along the wire was exposed to Rontgen rays. After leaving the wire the gas went into

Jrutuctcr
ym

A

D

B

EC
.

I
loBatttru

Fig. 6.

the metallic vessel connected with the electrome-

ter. The gas entering this vessel was found to be

charged with electricit}'-, and the electrification was

of the opposite sign to that on the wire with which

the gas had been in contact. This effect admits of

a simple explanation on the view that when a gas is

exposed to Rontgen rays, positively and negatively

charged ions are produced in the gas. Let us sup-

pose that the wire is charged with positive electri-

city ;

then

it

will

attract

the

negatively

charged

ions

in its neighbourhood and repel the positive ones.

In the layer of gas next the wire there thus would

52 DISCHARGE OF ELECTRICITY THROUGH GASES
be an excess of negative over positive ions ; and if this layer gets swept past the wire before all the negative ions can strike against the wire and give up their charges, the gas after passing the positively charged wire will have a charge of negative electricity. The greater the velocity of the ions under a given electro-motive intensity, the more likely will the negative ions be to move up to the wire and give up their charges before they are swept past it, and thus contribute nothing to the negative electrification of the gas in the metal
vessel.
The fact that tne electrification in the gas is of opposite sign to that on the wire seems very strong
evidence in favour of the view that there are charged ions in the gas exposed to the Eontgen rays, as if the ions carried no charge to begin with, the gas after passing the wire would, if it were charged at all, be charged with electricity of the same sign as that on the wire.
If the charged wire is coated with a layer of paraffin, or placed inside a glass tube so as not to be in actual contact with the stream of gas, the gas is electrified and the sign of the electrification is opposite to that on the wire.
The properties of gases electrified when under the influence of the Eontgen rays differ somewhat from those of gases electrified by other methods.

DISCHARGE OF ELECTRICITY THROUGH GASES 53
Thus the electrification in a gas electrified under Rontgen rays cannot pass through glass wool, or bubble through water; it thus resembles the conducting property conferred on the gas by these rays. On the other hand, as we have seen when the electrification in the gas is produced by chemical action, the electrification is able to pass through glass wool, or bubble through conducting liquids.
Another property of the electrification produced in a gas when exposed to Eontgen rays is that it is rapidly taken out of the gas not merely by conductors, but also by insulators in contact with the gas;^ the electrification cannot survive the passage of a gas through any considerable length of fairly wide bore tubing, whether this be made of metal or glass ; while the electrification due to chemical action would not be appreciably affected by the
passage through similar tubes. This may be the reason why the electrification produced under
the rays cannot pass through plugs of glass wool, or bubble through water or sulphuric acid.
Mr. Rutherford also found that the negative ions gave up their charges more readily to metals than the positive ions. The difference was more marked with some metals than with others ; thus the difference was greater for zinc than for copper. In fact, the more electro-positive the metal, the more
1 Rutherford, Phil. Mag., 43, p. 241, 1897.

54 DISCHARGE OP ELECTRICITY THROUGH GASES
did the rate at which the negative charge was given up exceed that for the positive. ISTo difference of this kind was observed in the case of insulators. Mr. Erskine-Murray ^ found that a gas exposed to the Rontgen rays produced the same effect on the potential difference between two metals immersed in the gas as is produced when the two metals are connected by an electrolyte ; that is, the two metals are brought to the same potential, in the sense that if, being connected by the conducting gas, the rays are cut off so that the metallic plates are insulated from each other, no change of potential is produced by separating the plates. If the plates had been connected by a metallic conductor and then disconnected, the difference of potential between the plates would increase as the distance between them was increased. Lord Kelvin ^ has shown that the flame gases arising from the combustion of a spirit lamp produced a similar effect on the potential difference between two metals to that produced by the Eontgen rays.
Professor Minchin has shown that metal plates previously uncharged get, when exposed to Eontgen rays, charges of electricity, positive in some cases, negative in others. If a metal which when originally uncharged gets negatively electrified when
1 Erskine-Murray, Proc. Roy. Soc, 49, p. 333, 1896. 2 Kelvin, Reprint of Papers on Electrostatics and Magnetism.

DISCHARGE OP ELECTRICITY THROUGH GASES 55

exposed to the rays is charged with negative elec-

tricity, it will not lose the whole of its charge ; while,

on the other hand, if it is initiall}^ charged posi-

tively, it will not only lose its positive charge, but

will acquire a small negative one. The potentials

to which these residual charges raise the metals

are not large, being generally much less than a

volt, except in the case of sodium or (what is much

more convenient to work with) sodium amalgam,

when it may amount to several volts. The charge

assumed by a metal plate will probably depend on

the nature of the metals with which it is con-

nected by gas traversed by Kontgen rays. If these

are electro-negative to the metal under considera-

tion, it will get a negative charge ; if, on the other

hand, they are electro-positive to it, it will get a

We positive charge.

may look upon the system of

metals and Rontgenized air as analogous to metals

immersed in an electrolyte and forming a galvanic

battery ; in this case, if the electro-positive metal

is connected to one quadrant of an electrometer, it

will show a negative charge if the other metal and

the other quadrant of the electrometer are copnected

with the earth.

Perrin has lately discovered that the Rontgen

rays, in addition to the effect they produce on the

gas through which they pass, exert a special effect

on the gas in the neighbourhood of a metallic sur-

66 DISCHARGE OF ELECTRICITY THROUGH GASES
face on which they impinge. The layer of gas adjacent to the metal acquires an abnormally great conductivity. The magnitude of the effect varies very much with the nature of the metal and of the
gas ; it is very small for aluminium, but considerable for gold, zinc, lead, and tin.
Uranium Radiation,
The salts of uranium as well as the metal itself were found by Becquerel to emit something which produced similar effects to those produced by the Rontgen rays. The radiation from uranium, like the Rontgen radiation, can affect a photographic plate
after it has passed through films of thin metal which are opaque to ordinary light ; it also, like the Rontgen radiation, renders a gas through which it passes a conductor of electricity. The gas through which the uranium radiation has passed has properties quite analogous to those of gas through which the Rontgen rays have passed. Thus the gas retains its conducting properties for some time after the rays have ceased to pass through it ; the
— laws of conduction that is, the connection be— tween the electro-motive force and the current are
the same in the two cases. The current gets " saturated" and does not increase beyond a certain point, however much the electric intensity is in-
creased; the saturation current is greater for a

DISCHARGE OF ELECTRICITY THROUGH GASES 57
thick layer of the gas than for a thin one.
The conducting property is destroyed by passing the gas through glass wool, or by sending a current of electricity through it. Again, the rate of leak through different gases varies; it is greatest for those gases which give a large rate of leak under Rontgen rays. Mr. Rutherford has lately found that the velocity of the ions when a gas is traversed by uranium rays is the same, if the electric field is the same, as when it is traversed by Rontgen rays. The resemblance between the conduction through gases under Becquerel radiation and under Rontgen radiation is so complete as to make it almost certain that the conduction is produced by
the same mechanism. We have seen that the
hypothesis that the gas is ionized when exposed to Rontgen rays explains the laws of conduction in that case; we therefore conclude that a gas is ionized by the uranium radiation.
Becquerel has shown that the uranium compounds emit this radiation apparently with undiminished energy, even though they have been kept in the dark for months before they are
tested. He found, moreover, that solutions of
uranium salts emit this radiation. Becquerel found that this radiation could be
reflected, refracted, and polarized, so that it is
evidently a form of light. He found that the

58 DISCHARGE OF ELECTRICITY THROUGH GASES refractive index of glass for this light was not very different from the value for ordinary light. If this result should be confirmed, it would seem to show that the wave-length of the uranium light was not an excessively small fraction of that of ordinary light, as on all theories of dispersion the refractive index for light of an infinitely small wave-length is unity.
n

PHOTO-ELECTRIC EFFECTS

;

PHOTO-ELECTRIC EFFECTS

The discovery by Hertz/ in 1887, that the in-
cidence of ultra-violet light on a spark gap facilitates the passage of a spark, led to a series of investigations by Hallwachs,^ Hoor,^ Kighi,* and Stoletow^ on the effect of ultra-violet light on electrified bodies. It was found that a freshly cleaned surface of zinc, if charged with negative electricity, rapidly loses its charge, however small
this may be, when ultra-violet light falls upon it
whereas a similar surface charged with positive electricity suffers no loss of charge, and further that the surface if uncharged acquires a positive
charge when exposed to the light. The ultra-violet light may be obtained from an arc-lamp, the effect of which is increased if one of the terminals is of

1 Hertz, Wied. Ann., 31, p. 983, 1887.

2 Hallwachs, Wied. Ann., 33, p. 308, 1888.

8 Hoor, Repertorium der Physik, 25, p. 105, 1889.

* Righi, C. R., 107, p. 560, 1888.

5

Stoletow, C. R.,

106,

pp.

1149,

1593 ;

107, p. 91 ;

108, p. 1241.

See also Physikalische Revue, bd. 1, 1892.

62

PHOTO-ELECTRIC EFFECTS

zinc or aluminium, the light from these substances
being very rich in ultra-violet rays; it may also
be got from burning magnesium, or by sparking with an induction coil between zinc or cadmium terminals. Ordinary sunlight is not rich in ultraviolet light, and does not produce anything like so great an effect as the arc-light. Elster and Geitel,*
who have investigated the effects of light on elec-
trified bodies with great success, found that the more electro-positive metals lose negative charges
when exposed to ordinary light and do not require the presence of the ultra-violet rays. They found that amalgams of sodium and potassium enclosed in a glass vessel lose a negative charge when ex-
posed to daylight, though the glass stops the small amount of ultra-violet light left in sunlight after its passage through the atmosphere. If sodium or potassium, by themselves, instead of their amalgams, be used, or (what is often more convenient) the curious liquid obtained by mixing sodium and potassium in the proportion of the combining weights of these metals, they found that the photo-electric effects are produced by the light from an ordinary petroleum lamp. While if the still more electro-positive metal, rubidium, is used, the photo-

1 Elster audGeitel, Wied. Ann., 38, pp. 40, 497, 1889; 41, p. 161, 1890; 42, p. 564, 1891 ; 43, p. 225, 1892 ; 52, p. 433, 1894; 55, p. 684, 1 895.

PHOTO-ELECTRIC EFFECTS

63

electric effects due to the light from a glass rod just heated to redness can be distinctly perceived. They found, however, that the eye was more sensitive to radiation than the rubidium cell, and no photo-electric effects could be detected before the radiation from the glass rod was visible.
Elster and Geitel arrange the metals in the following order for their photo-electric effects:

Rubidium. Potassium. Alloy of Sodium and Potassium. Sodium. Lithium. Magnesium. Thallium.
Zinc.

With copper, platinum, lead, iron, cadmium, carbon, and mercury, the effects with ordinary light
are too small to be appreciable. This order is the same as that in Volta's electro-chemical series.
Elster and Geitel found that the ratio of the photo-electric effects of two metals exposed to approximately monochromatic light depended upon the wave-length of the light, different metals
exhibiting a maximum sensitiveness in different
parts of the spectrum. Thus, in a solar spectrum obtained by a glass prism, the blue is the part

64

PHOTO-ELECTRIC EFFECTS

wliich produces the greatest effect on potassium. The following table for the alkaline metals, given by Elster and Geitel,^ shows how the photo-electric effect for a particular metal depends upon the character of the incident light. The numbers in the table are the rates of emission of negative electricity under similar circumstances. The rate of emission for each of the cells under the white light from a petroleum lamp is taken as unity.

Blue.

Yellow. Orange.

Eed.

Bb

.16

.64

.33

.039

K

.57

.07

.04

.002

Na

.37

.36

.14

.009

This table indicates that the absorption of light by the metal has a great influence on the photo-electric effect, for while potassium is more sensitive to blue light than sodium, the strong absorption of yellow light by sodium makes it more than five times more sensitive to this light than potassium. Stoletow very early called attention to the necessity of
strong absorption for photo-electric effects. He
showed that water, which does not absorb the visible or ultra-violet rays much, does not'lose a charge of negative electricity when illuminated, while strongly coloured solutions and especially solutions of fluorescent substances, such as methyl green and methyl

1 Elster and Geitel, Wied. Ann., 52, p. 438, 1894.

PHOTO-ELECTRIC EFFECTS

65

violet, do so to a very considerable extent ; he found indeed that a solution of methyl green was very
much more sensitive than zinc. Phosphorescent
substances, such as Balmain's luminous paint (sulphide of calcium) show this photo-electric effect, so also, as Elster and Geitel ^ have shown, do various specimens of fluor-spar, the magnitude of the effect depending to a very great extent on the
colour of the spar. As phosphorescence and fluor-
escence is probably accompanied by a very intense absorption by the surface layers of the substance, the evidence seems very strong that in order to get the photo-electric effect there must be strong absorption of some kind of light, whether this be ultra-violet light or light of longer wave-length. Hallwachs ^ has shown that, in the case of liquids, there is always strong absorption whenever the
liquid shows photo-electric effects; we may have,
however, strong absorption without photo-electric
effects. When there is strong absorption of light
of a particular frequency, say p, we should be led by
all theories of dispersion to believe that abnormally great effects on the surface would be produced by
all light vibrations having a frequency between p
m and ^ -f- w, where is a positive finite quantity de-
pending on the nature of the absorbing substance.
1 Elster and Geitel, Wied. Ann., 44, p. 722, 1891. 2 Hallwachs, Wied. Ann., 37, p. 666, 1889.
5

66

PHOTO-ELECTRIC EFFECTS

Let us take Helmholtz' ^ theory of dispersion as an example. The intense absorption of light is due to the frequency of the absorbed light being equal to one of the free periods of the substance. Helmholtz gives the relation between the frequency of vibration and the refractive index of a substance which has one free period of vibration. This relation is represented by the curve, Fig. 7,

V

f

A' 1
,

\

h

1

Cb

^ vj

TV^.O

7c

Fig 7.

in which the ordinates represent the refractive in-

dex, and the abscissae the frequency of the vibra-

tion ;

the

refracting substance is supposed to

have

a free period whose frequency is represented by

al. It will be noticed that there is a gap in the

curve just after passing this frequency ; thus light whose frequency is between al and ak cannot pass

through the medium. This light will be totally re-

flected from the surface, and this total reflection

Ilelmlioltz, Collected Works, vol. iii., p. 505.

PHOTO-ELEGTRIC EFFECTS

67

will be accompanied by an intense agitation of the molecules in a very thin layer close to the surface. Though the preceding curve only applies to the extreme case when the refracting substance has only
one free period, yet when there are many free
periods the effects, though more complicated, will probably be of the same general character. It is

Fig. 8.
— this intense agitation of the surface layer that is,
the possession of the molecules in this layer of an
— abnormal amount of energy which seems neces-
sary for the production of photo-electric effects.
The laws which regulate the flow of negative electricity from the illuminated surface are very interesting. They have been investigated by Stole-

68

PHOTO-ELECTRIC EFFECTS

tow/ Righi,2 and Elster and Geitel.^ The apparatus used by Stoletow is shown in Fig. 8. The light from an arc-lamp, A, passed through a hole in a metal screen; it then fell on the metal j)lates, C, connected with a battery and galvanometer in the way shown in the figure. The plate nearest the light was perforated, and the light passed through the perforations on to the sensitive plate connected with the negative pole of the battery. The current passing between the plates was measured by the galvanometer. By means of this apparatus Stoletow investigated the relation between the current and the potential difference between the plates, varying in his experiments the distance between the plates ; his results are represented by the curves in Fig. 9. The curves show that except for small electro-motive forces the conduction does not obey Ohm's law; the current does not increase as rapidly as the electro-motive force. For large values of
the electro-motive force the curves give indications of becoming straight lines slightly inclined to the axis along which the electro-motive force is measured. In conduction through a gas illuminated by
Rontgen rays, these lines become parallel to the

1 Stoletow, Journal de Physique, (2) 9, p. 468-473, 1890.
2 Righi, Mem. della R. Ace. de Bologna, (4) 10, pp. 85-114,
1890.
« Elster and Geitel, Wied. Ann., 52, p. 438, 1894.

^

^y

y ^ yi^ 'J

^

y y y^^
/^

^^
^ L -«\

y /
/
// y

/

,

/
/

/

y

Y/
//
//

y^
/y
/

^ ^^ 45
^r' ^ Kt )0
r-

'^
y^

—

//
//
/

y y -^
/.
/

^

/

/

//

/

//

1

/

r^

,

^ y

5

y\

,^

/

/

/

J//

/- r

y

/

.

^

//
I

11

^

j

/

/r
/ /

^ ^^ 5c- -5<

f/—

/

II
/

/
f—

y

y y TJ /

/
/

/

/ ^ y y ^ y 1/
^

^ -^

y

^ ^ ^ u-

JQ. ^^l

^

E»5 10 »S 20 €5 30

40

50

60

70

90 100

70

PHOTO-ELECTRIC EFFECTS

axis. Stoletow and Righi have investigated the

effect of the pressure of the gas on the rate of

leak. They found that the current increased as

the pressure diminished until a certain pressure

was reached, when any further diminution in the

pressure caused a diminution in the current. The

change in the current was small compared with

the change in the pressure. The following are

some measurements by Stoletow with an electro-

motive force equal to that of 65 Clark cells. The

distance

between

the plates

was

3.71 mm. ;

p

stands for the pressure, and c for the correspond-

ing current.

p = 754 = c 8.46
p = 0.64 = c 108.2

152 13.6 0.52 102.4

21 26.4 0.275 82.6

8.8 32.2 0.105 65.8

3.3 48.9 .0147 53.8

2.48 74.7 .0047 50.7

1.01 106.8 .0031
49.5

If p is the pressure at which the current is a maximum, then Stoletow states that p l/E is
E constant, where is the electro-motive force and
I the distance between the plates ; according to
Righi p is the pressure at which sparks pass
most easily between the plates. The nature of the gas between the plates has a
considerable effect upon the rate of leak. Elster and GeiteP investigated the rate of leak from an illuminated surface through air, carbonic acid

1 Elster and Geitel, Wied. Ann., 41, p. 161, 1890.

PHOTO-ELECTRIC EFFECTS

71

gas, oxygen, and hydrogen ; they found that the rate of leak through carbonic acid was much faster than for any of the other gases.
Elster and Geitel ^ found that the plane of polarization of the incident light had considerable influence upon the amount of the photo-electric effect. The effect produced by light polarized in a plane at right angles to the plane of incidence is for oblique incidence much greater than that produced by light polarized in the plane of inci-
dence. This effect is much more marked in the
case of the liquid alloy of potassium and sodium surrounded by gas at a low pressure and exposed to visible light than it is for a zinc plate surrounded by air at atmospheric pressure and exposed to ultra-violet light.
In light polarized in a plane at right angles to that of incidence, the periodic electric intensity, which, according to the electro-magnetic theory of light, exists in the incident light wave, has a component at right angles to the reflecting surface, and produces a periodic effect on the electrifica-
tion of the surface. When the light is polarized
in the plane of incidence, the electric intensity is parallel to the reflecting surface and the electrification over this surface has no tendency to a periodic variation.
1 Elster and Geitel, Wied. Ann., 55, p. 684, 1895.

72

PHOTO-ELECTRIC EFFECTS

The existence of a periodic electric intensity at right angles to the reflecting surface seems to facilitate the escape of negative electricity from the surface. Another circumstance which may be connected with the effect produced by the posi-
tion of the plane of polarization is the fact discovered by Quincke/ that light polarized at right angles to the plane of incidence penetrates more deeply into metals than light polarized in that plane.
Elster and Geitel^ also found that the rate of escape of negative electrification from a charged surface is diminished by a magnetic field when the lines of magnetic force are parallel to the reflecting surface. This effect is small at ordinary pressures, but increases rapidly as the pressure diminishes. In one of the cases investigated by Elster and Geitel, when the gas was oxygen at a
pressure of .0002 mm. of mercury, the rate of leak
in a magnetic field (whose strength is not given)
was only half the rate when there was no magnetic
force.
Stoletow,' Eighi,^ and Arrhenius ^ have shown that when two different metals, Jf and Jf \ are used
1 Quincke, Pogg. Ann., 129, p. 177, 1866. 2 Elster and Geitel, Wied. Ann., 41, p. 166, 1890. 8 Stoletow, Physikalische Revue, 1, p. 76.'}, 1892. * Righi, Journal de Physique, (2) 7, p. 153, 1888. * Arrhenius, Wied. Ann., 33, p. 638, 1888.

PHOTO-ELECTRIC EFFECTS

73

for the perforated and continuous plate in an arrangement like that in Fig. 8, the total electromotive force acting round the circuit is not E, the electro-motive force of the battery, but
E ± M/M\ where MjM'^ is, approximately at any
rate, the contact difference of potential between the metals when these are not exposed to ultraviolet light. Thus the air traversed by ultraviolet light behaves like an electrolyte. Stoletow has verified by direct experiment that two differ-ent metals when immersed in the illuminated air are at the same potential, just as Lord Kelvin has
shown they are when immersed in an electrolyte, and as they are when immersed in air near a flame or in air traversed by Rontgen rays.
When the plates are of the same material, there
is no polarization produced by the passage of a current through the illuminated gas.
Stoletow constructed a battery of two parallel plates made of different metals ; one of these was perforated, and ultra-violet light passed through the openings and fell upon the other plate. For this battery to work, it is necessary that the perforated plate should be made of the more electropositive metal, as in this case the current goes round in such a direction that the negative electricity passes from the plate on which the ultraviolet light shines to the perforated plate. If the

74

PHOTO-ELECTRIC EFFECTS

plates were reversed then for the current to pass, the illuminated plate would have to discharge positive electricity, and this, as we have seen, it is unable to do.
Stoletow* made a series of experiments to see if the discharge from a negativel}^ electrified plate lasted for an appreciable time after the ultra-violet
light was cut off. He was not able to obtain any
evidence that the cessation of the discharge was not absolutely contemporaneous with the withdrawal of the light, and was able to prove that there was no leakage after the rays had been cut off for 1/1000 of a second.
The investigations by Lenard and Wolf ^ on the behaviour of a steam jet placed near a surface illuminated by ultra-violet light are of great
interest in connection with these photo-electric
effects. These observers found that when ultra-
violet light falls on a negatively electrified plati-
num surface, a steam jet in the neighbourhood of
the surface shows by its change of colour that the steam in it has been condensed. The metals tried by Lenard and Wolf were zinc, mercury, platinum, brass, copper, tin, lead, gold, and silver, and in all these cases condensation of the steam occurred when the metallic surface was negatively electri-
1 Stoletow, Physikalische Revue, 1, p. 747, 1892. 2 Lenard and Wolf, Wied. Ann., 37, p. 443, 1889.

PHOTO-ELECTRIC EFFECTS

75

fied. They found too that though no condensation is produced when the light falls on a water surface, yet it occurs when the water surface is replaced by one of the fluorescent solutions, such as rosaniline or methyl violet, which give photoelectric effects. They found also that condensation occurs, though to a much smaller extent, if the surfaces are not electrified ; but no condensation can be detected when the surfaces are charged with positive electricity. It will be noticed that some of the metals which are effective in producing con-
densation are not very sensitive to the photo-
electric effect, and surfaces of quartz or gypsum also produce condensation glass and mica, on the
;
other hand, give no appreciable effect. Lenard and Wolf attributed this condensation
of the steam in the jet to dust emitted from the illuminated surface, the dust, in accordance with Aitken's experiments,^ producing condensation by forming nuclei around which the water-drops
collect.
The indications of a steam jet are, however, very ambiguous, as condensation is promoted not only by dust, but also by chemical action or electrification in its neighbourhood. Thus Lenard and Wolf's experiments do not determine whether metallic
dust or the gas is the carrier of the electrification.
1 Aitken, Trans. Roy, Soc, Edinburgh, 30, p. 337, 1887.

76

PHOTO-ELECTRIC EFFECTS

Lenard and Wolf observed a slight roughening of the metal at the places exposed to the ultra-violet
light.
As the direct evidence as to the nature of the
carriers of the electrification in these photo-electric
effects is inconclusive, we must have recourse to
the indirect evidence afforded by the laws obeyed by the convection currents which start from the illuminated surface ; those seem to point to the gas playing a very considerable part in the discharge. The fact that the rate of discharge depends upon the nature of the gas is not conclusive, as, if the
discharge were carried by the metallic dust, we
might expect that the rate of diffusion of the dust through the surrounding gas would vary with the nature of the gas. The way in which the rate of leak varies with the pressure seems, however, inconsistent with the idea that the metallic dust carries the greater part of the charge. As we diminish the pressure the rate of leak does not continually increase ; it increases indeed until the pressure sinks to a certain critical value, but when the pressure falls below this value, any further decrease in the pressure produces a decrease in the rate of leak. This seems to point to the gas as the carrier of the greater part of the convective current. It is, however, to be remembered that the gas is not put into the

PHOTO-ELECTRIC EFFECTS

77

state in which it can act as a carrier of electricity merely by the passage through it of ultra-violet light ; this state of the gas is manufactured only when the light strikes against the surface of the electro-positive metal or the solution of a phosphorescent substance. Its production depends on the material of which the reflecting surface is made, and even, when the light is polarized, on the orientation of this surface. It therefore differs essentially from the conductivity produced in a gas by the passage through it of Rontgen rays, which is not dependent upon the existence of suitable surfaces for these rays to strike against.
The sign of these photo-electric effects is opposite to that produced when the surface is made so hot as to be luminous ; for in air, as Elster and Geitel have shown, the hot metal becomes nega-
tively, the surrounding air positively, electrified.
The sign of the electrification in the air produced by the action of light is the same as that produced in air by the splashing of drops of mercury or water; and as the electrification, in that case, was explained by the existence of a double layer of electrification, the positive side of which was on the metal, the negative on the gas, the negative layer getting partially torn away by the shock produced by the splash, so it would seem that the photo-electric effects might be produced by the

78

PHOTO-ELECTRIC EFFECTS

tearing away of the outer layer of the double sheet by the action of the incident light. There seem reasons for thinking that the incident light might produce this result ; the photo-electric effects are accompanied by an intense absorption of light
at the surface layer of the reflecting body, so that
a considerable amount of energy will flow into the molecules of this layer, which forms one of the sides of the electrical double layer at the surface. These particles, having a large amount of kinetic energy, may by communication so increase the kinetic energy of the gaseous particles which form the opposite side of the double layer, that it gets sufficiently large to enable some of them to escape from the neighbourhood of the reflecting surface and diffuse into the surrounding gas ; as the
particles of gas in the outer layer are negatively charged, a negative charge will diffuse into the gas and a positive one be left on the metal. If the metal itself is charged negatively, the repulsion
exerted by this charge on the negatively electrified particles of gas will make these move more quickly and so accelerate the rate of leak ; a positive charge, on the other hand, would retard it. The effect of pressure on the rate of leak is consistent with this view, for reduction of pressure would, on the one hand, facilitate the leak by increasing the rate of diffusion of electrified particles through it; on the

;

PHOTO-ELECTRIC EFFECTS

79

other hand, it would retard the rate of leak if it affected the number of particles available for forming the outer layer ; it would seem likely that the
pressure would have to be very low before much effect was produced on the number of charged particles in the outer layer, so that it would not be
until the pressure is very low that a decrease in the pressure would cause a decrease in the rate of leako The action of a magnetic field on the rate of leak is also consistent with this view, as a magnetic force acting parallel to the reflecting surface would act upon the moving charged particles with a force acting at right angles both to the direction of motion of the particles and to the magnetic force. This force will leave the velocity unaltered, but will constrain the particles to follow a curved path; this will diminish the component of the velocity parallel to the electric intensity and so diminish the
rate of leak.
This view also affords an explanation of the curious connection between the rate of leak and the electro-motive force illustrated by Fig. 9.
This view seems also to explain the existence of a pressure at which the current is a maximum.
For we may suppose that the effect of a diminu-
tion in pressure is twofold : the first an effect on the velocity of the electrified particles through the gas, the velocity varying inversely as the pressure

80

PHOTO-ELECTRIC EFFECTS

the second an effect upon the number of active particles produced by the rays, the number diminishing when the pressure is reduced, but varying more slowly than the pressure. Starting with a
higher pressure, the effect of diminution of the pressure will be to increase the velocity with which the particles move, in a greater ratio than
it diminishes the number of particles thus we ;
shall get an increase in the current. This will go on until the current is so great that to carry it requires all the active particles produced by the ultra-violet light j as soon as this is the case, the current depends only upon the number of particles
produced, and not upon their velocity. Any further
diminution of the pressure will therefore produce a diminution in the current, since it diminishes the number of active particles. There will thus be a
certain pressure when the current is a maximum.
Stoletow found that at this pressure an increase in the potential difference produced an increase in the current. This shows, if the preceding explanation is the correct one, that the greater the density of the negative electricity on the plate on which the ultra-violet light fell, the greater the number of active particles produced ; this would happen if the effect of the ultra-violet light was to produce a definite difference of potential between the gas and the metal.

ELECTRIFICATION OF GASES

81

A very interesting result obtained by Stoletow
was that, for very low pressures, the maximum
current was independent of the pressure ; this would seem to indicate that the current in these cases was carried either by the mercury vapour
from the pump or by particles from the metal
surface.

Electrification of Gases hy Glowing Metals.
Elster and Geitel ^ found that when a platinum
wire is heated to luminosity in air, the air near the wire acquires a charge of positive electricity, the wire itself a charge of negative electricity ; if, however, the wire is heated in hydrogen, the electrification in the hydrogen is negative, that on the wire positive.
Hydrogen was the only gas amongst those they tried (oxygen, carbonic acid gas, water vapour, and the vapours of sulphur and phosphorus and mercury) which acquired a negative charge the others
;
all got charged positively, with the exception of mercury vapour, which did not get charged at all.
The same effects were observed when palladium and iron wires were used instead of platinum. When carbon filaments were heated, the electrifica-

1 Elster and Geitel, Wied. Ann., 16, p. 193, 1882; 19, p. 588,

1883 ;

22, p. 123, 1884 ;

26, p.

1,

1885 ;

31, p.

109, 1887 ;

37, p.

315, 1889. (This paper contains a summary of their results.)

82

ELECTRIFICATION OF GASES

tion in the gas was always negative ; these filaments, however, give off so much gas that the conditions of the experiments become indefinite.
In these experiments precautions were taken to get rid of all dust from the gas ; but though the gas
may be freed from dust at the beginning of the
experiment, yet inasmuch as glowing metals give off either metallic dust or vapour, the gas is liable to get charged with metallic dust as the experiment goes on. According to Nahrwold,^ little if any metallic dust is given off when platinum is raised to incandescence in hydrogen.
If air is blown past an incandescent platinum wire, it comes off positively electrified ; thus heating the wire to redness, and blowing air in its normal electric condition past the wire, produces electric effects of the same sign as when the wire is cold, and air through which Kontgen rays have recently passed is blown past it. If another (cold) platinum wire is placed in the neighbourhood of the one raised to luminosity, its potential will be raised in consequence of the positive electrification in the air around it. The potential attained by this wire seemed almost independent of the pressure of the gas. It depends, however, to a great extent on the temperature of the incandescent
wire, the potential attaining a maximum at a
1 Nahrwold, Wied. Ami., 35, p. 107, 1888.

ELECTRIFICATION OF GASES

83

bright yellow heat, after which any increase in temperature causes a diminution in the potential. In hydrogen, where the electrification is negative, an increase in the temperature seems always to be accompanied by an increase in the negative electrification of the gas. At very high temperatures the positive electrification in air gets exceedingly small ; indeed, some observers have thought that at very high temperature they detected a tendency for the electrification in the air to become negative instead of positive ; and thus to have the same sign as air in the neighbourhood of a clean metallic sur-
face reflecting ultra-violet light. When very thin
wires are used and the pressure is very low, then a conductor in the neighbourhood gradually acquires, after the glow has lasted for a long time, a negative charge. The wires become brittle, and their resistance is changed.
Branly ^ investigated this effect of glowing bodies by a slightly different method : he suspended near the glowing body an insulated, charged conductor, and observed whether this was dis-
charged or not. Where the glowing body was a platinum spiral, he found that when this was at
a dull red heat, a neighbouring conductor lost a charge of negative electricity, but not of positive. This is what we should expect from Elster and
1 Branly, C. R., 114, p. 1531, 1892.

84

ELECTRIFICATION OF GASES

GeitePs result, that there is positive electrification
in the air around the glowing body. When the
platinum spiral was made to glow very brightly, Branly found that the conductor lost its charge,
whatever the sign of it might be. He found that
the effects at a dull red heat depended upon the nature of the glowing body. Where this was a lamp-shade covered with bismuth oxide or lead oxide, he found that this discharged a positively electrified body in its neighbourhood, but not a negatively electrified one. This is a reversal of the effect observed with clean metallic surfaces.
Mr. Stanton ^ found that a hot surface of clean copper discharges a negatively electrified conduc-
tor in its neighbourhood, but ceases to do so when
the copper surface gets coated with a layer of oxide ; thus the discharge goes on as long as oxidation is taking place, but ceases as soon as this stops. When, however, the hot oxidized copper and the positively charged conductor are placed in a vessel containing hydrogen, so that the oxide gets reduced, the conductor is discharged as long as the reduction is going on, but ceases as soon as the reduction is completed, so that hot, clean copper in hydrogen is unable to discharge a conductor placed in its neighbourhood if this is positively charged ; on the other hand, the con-
1 Stanton, Proc. Roy. Soc, vol. xlvii., p. 559, 1889.

ELECTRIFICATION OF GASES

85

ductor is discharged if it originally has a negative charge. Thus, in hydrogen, hot copper can retain a charge of negative, but loses one of positive,
electricity.
Another aspect of the electrification produced by glowing bodies is what is known as "unipolar
— conduction," that is, when a hot body loses elec-
tricity of one sign more easily than it does that of the opposite sign. Thus a hot platinum wire in air, since it, if unelectrified to begin with, acquires a negative charge, will evidently lose a positive charge more quickly than a negative one. This difference between the rate of escape of the two electricities from hot bodies was known before the production of electrical separation by glowing metals was directly demonstrated. Guthrie,^ who was the first to call attention to phenomena of this kind, observed that an iron sphere in air cannot, when white hot, retain a charge either of positive or of negative electricity, and that as it cools it acquires the power of retaining a negative charge before it can retain a positive one. If the sphere is connected with the earth and held near a charged body, then, when the sphere is white hot, the body soon loses its charge, whether this be positive or negative ; when the sphere is somewhat colder, the body is discharged when negatively, but not when positively, electrified.
1 Guthrie, Phil. Mag., [4] 46, p. 257, 1873.

86

ELECTRIFICATION OF GASES

Elster and GeiteP made the very interesting observation that at low pressures the positive electrification in the case of air is increased by
a magnetic field, while the negative electrification in hydrogen is diminished ; the latter result is the moftt marked of the two.

JElectrification in the Neighbourhood of an Arc
Discharge.

The author has found that electrical effects, very

similar to those produced by incandescent solids,

occur near the arc. The following arrangement

was used for these experiments. An arc dis-
B charge between the platinum terminals A^ (Fig.

10), was produced by a large transformer, which

A transformed up in the ratio of 400 to 1.

current

of about 40 amperes, making 80 alternations per

second, was sent through the primary. When the

gases were at atmospheric pressure the method

A used was as follows.

current of the gas under

examination entered the vessel through a glass

tube, C, and blew the gas in the neighbourhood of

the arc against the platinum electrode, E, which

was connected with one quadrant of an electrome-

ter, the other quadrant of which was connected
E with the earth. To screen from external elec-

tric influences, it was enclosed in a platinum tube,

1 Elster and Geitel, Wied. Ann., 38, p. 27, 1889.