T. C. Hebb.: (Second (.SERIES
T. C. Hebb.: (Second (.SERIES
r
I ^HE results of Davis and Goucher1 and of Bishop2 seem to prove the
-*- correctness of the suggestion of Van der Bijl3 that the apparent
ionization of mercury vapor when bombarded with electrons possessing a
velocity of 4.9 volts, as observed by Franck and Hertz 4 and by Newman,5
was due to the photo-electric action of the radiation X = 2536.7 acting
on the receiving plate of the ionization chamber. Their results are also
in harmony with those obtained by McLennan and Henderson6 and
also by Tate, 7 viz., that the ionization potential of mercury vapor is
10.3 volts.
Neither the above suggestion nor the experimental results quoted,
however, explain the results obtained by the writer: 8 viz., that the
complete spectrum of mercury vapor appeared at 4.9 volts, and that an
arc struck at that voltage. Millikan,9 however, has suggested that the
same radiation X = 2536.7 acts photo-electrically on the mercury vapor
and produces the necessary ionization. If this should prove to be true
then the fact will cast some light on the photo-electric action.
If, however, it should be found that photo-electric action is not suffi-
cient to explain the arc at 4.9 volts, it would seem to be necessary to
assume that under certain conditions mercury vapor can be ionized by
collision with electrons moving with a velocity of 4.9 volts.
The experiments reported in this paper were undertaken in the hope
that the above question might be decided. The results obtained will
be dealt with under the following five heads: (1) Arcing voltages; (2)
Current-potential Relations; (3) Striations; (4) Ionization Potential
and (5) Photo-electric Action.
1
Davis and Goucher, P H Y S . R E V . , Vol. 10, p . 101, Aug., 1917.
2
Bishop, P H Y S . R E V . , Vol. 10, p. 244, Sept., 1917.
3
Van der Bijl, P H Y S . R E V . , Vol. g, p . 173, Feb., 1917.
4
F r a n c k and Hertz, Deutsch. P h y s . Gessell. Verh., Vol. 11, p . 512, 1914.
5
N e w m a n , Phil. Mag., Vol. 28, p . 753, Nov., 1914.
6
M c L e n n a n and Henderson, Proc. Roy. S o c , A, Vol. 9 1 , 1915.
7
T a t e , P H Y S . R E V . , Vol. 7, p . 686, J u n e , 1916.
8
H e b b , P H Y S . R E V . , Vol. 9, p. 371, M a y , 1917.
9
Millikan, P H Y S . R E V . , Vol. 9, p. 378, M a y , 1917.
;^ 3 X L ] IONIZATION POTENTIAL OF MERCURY VAPOR. IJl
i. ARCING VOLTAGES.
The apparatus employed was essentially the same as that used in the
previous work on the mercury arc. A horizontal section is shown in
Fig. i. A was a glass tube about 20 cm. long and about 2.5 cm. in diam-
teter. B and C were iron caps which were fastened to the tube with
Khotinsky cement. Through B and C passed the iron electrodes which
carried the anode D and the cathode E. Pump connections were made
at both F and G, so that in case there was a small leak the resultant air
did not have to pass across the arc space DE. All the joints were sealed
with Khotinsky cement. The anode D was of platinum foil usually about
one centimeter square. The Wehnelt cathode was also of platinum
foil .003 cm. thick and about 1.0 cm. in length. Its width was usually
about .4 cm. The current used to heat the cathode varied between 10
and 20 amperes. Directly under D and E the glass tube was expanded
into a depression in order to hold the mercury and the expansion was
graded from the two ends of the tube, so that as fast as the mercury
condensed at the ends it ran back. This kept a constant supply of
mercury under the arc DE. The two iron caps, B and C, were surrounded
by cooling vessels through which water circulated. The central part
of the tube was surrounded by a gas-heated asbestos furnace with a
sheet-iron bottom. The mercury evaporated at the center of the tube
and passed both ways to the ends where, as stated above, it was con-
densed and returned to the center. As a result there should be produced
in the region DE an atmosphere of the purest mercury vapor. This
should be true even though the vacuum produced by the pump was
not very high. As a matter of fact the pump used gave a minimum
pressure of .25 mm.
With the apparatus as outlined above it was possible for me to sub-
stantiate my previous result, viz., that the arc could be caused to strike
at a potential difference as low as 4.9 volts. But in doing so I found that
it could be caused to strike at any potential difference above 4.9 volts
by varying some or all of the following factors: (1) The temperature of
the cathode, (2) the temperature of the furnace, (3) the distance between
the anode and cathode and (4) the pressure as recorded by a McLeod
gauge.
[SECOND
172 T. C. HEBB. LSERIES.
Curve A, Fig. 2, gives the relation between the striking voltage and
the amperes through the cathode for a particular case. T h e McLeod
gauge reading was 1.9 m m . In considering this curve it should be noted
t h a t the distance between the mercury and the cathode was a b o u t one
centimeter and hence a rise in temperature of the cathode caused a more
rapid evaporation of the mercury. T h a t this m a d e a difference was
proved by the observation t h a t the slope of the curve decreased when the
mercury was not directly under t h e anode and cathode.
In regard to the effect of the second factor mentioned above, it m a y
be stated t h a t the striking voltage decreases with a rise in the tempera-
ture of the furnace, although I have no exact d a t a to offer.
And in regard to the third factor I found t h a t the striking voltage
increased with the distance between the anode and cathode.
Curve B, Fig. 2, shows t h e relation between t h e striking voltage and
t h e reading of the McLeod gauge for a particular case. Everything
else was kept as constant as possible. I t will be noticed t h a t the striking
voltage decreases with t h e pressure, reaches a minimum and then rises
again. N o significance should be attached to t h e fact t h a t the curve
starts a t a b o u t 10 volts and rises again to t h a t value. I t could have
been extended and was in some cases. T h e minimum point of t h e curve
only reaches the value of 6 volts, b u t curves could be obtained in which
t h e minimum potential difference had a n y value above 5 volts.
As a result of m y experiments on arcing voltages I have come to the
conclusion t h a t the striking of the arc a t 4.9 volts depends on (1) the
density of the electron stream, (2) the density of the mercury vapor,
and (3) t h e purity of t h e mercury vapor. If t h e electron discharge is
weak or if t h e density of t h e mercury vapor is low, then there will be no
arc formed a t these low voltages. F u r t h e r if there is the slightest trace
of a foreign gas present, then, even though other conditions are favorable,
the arc will n o t strike as low as 4.9 volts. This last condition is as would
be expected if one considers t h e p a t h of an electron which leaves t h e
cathode and moves towards the anode through a dense atmosphere of
mercury vapor. Owing to t h e elasticity of t h e collisions between elec-
trons moving with speeds of less than 4.9 volts and molecules of mercury
vapor, the electron probably makes m a n y excursions back and forth
past a certain point before passing on to the anode. If there were a
molecule of an inelastic gas a t t h a t point the probability of collision with
this molecule and t h e consequent loss of t h e electron's energy would
be great.
2, CURRENT-POTENTIAL RELATIONS.
Using t h e same a p p a r a t u s with a low resistance galvanometer in
series with t h e experimental tube I made a study of the current-potential
VOL. XI. 1 IONIZATION POTENTIAL OF MERCURY VAPOR.
No. 3. J 173
3. STRIATIONS.
The conditions in a vacuum tube with a Wehnelt cathode are favorable
to the study of striations, and as the striations radiate the complete
spectrum of the element they appear to afford a method of studying
the minimum voltage necessary to produce this radiation. In order to
prevent .the arc from striking, however, it is necessary to work with a
comparatively cool cathode. I found that a red-hot cathode separated
about 5 mm. from the anode gave very satisfactory results when the
pressure was from 1 to 3 mm. If, with these conditions and with the
temperature of the furnace low, the potential difference between the
anode and cathode was raised, light appeared on the surface of the anode.
The potential difference at which this occurred was never low, but usually
in the neighborhood of 10 to 12 volts. If then the voltage was still
further raised, the light on the anode grew towards the cathode and a
portion of it separated from the main body of light on the anode, In
all cases I found that the increase in potential difference necessary to
produce this separation was 5 volts. If after the first striation was formed
the potential difference was further increased the phenomenon repeated
itself, the first striation in the meantime having moved nearer the
cathode. The formation of this second striation also required the addi-
tion of 5 volts. The production of each new striation required an extra
5 volts. I have had as high as four distinct striations and the light on
the anode with a potential difference of 32 volts. In this case the initial
light was produced at 12 volts. On the other hand, with the furnace
at a high temperature I have had two striatiohs and the light on the
anode for a potential difference of 15 volts.
The difference between the case where the temperature of the furnace
is low and the case where it is high probably lies in the purity of the vapor
between the anode and cathode. If the temperature is low, then the
evaporation of the mercury will not be sufficient to drive away all foreign
gases and as a consequence the electrons will lose energy in passing from
the cathode to the anode. As a consequence a potential difference
greater than 5 volts is required to produce light on the anode. In spite
of this loss, however, the addition of 5 volts will make a second ionization
possible. In the case where the furnace was at a high temperature,
however, the evaporation of the mercury was sufficiently rapid to drive
away all foreign gases and as a consequence the electrons lost no energy,
other than that due to ionization, in passing from the cathode to the
anode.
In connection with striations it may be of interest to state that I have
had them so close together and so close to the cathode that I could only
see the faintest dark line separating them.
^• 3 X L ] IONIZATION POTENTIAL OF MERCURY VAPOR. 175
4. IONIZING POTENTIAL.
It was considered both of interest and of value to determine the
ionizing potential of mercury vapor under conditions somewhat similar
to those in the preceding experiments, that is, to determine the ionizing
potential in an absolutely pure atmosphere of mercury vapor having a
pressure of one or more millimeters. In order to make this determina-
tion the anode of Fig. 1 was replaced by an ionizing chamber. The
chamber consisted of a platinum cylinder about 4 cm. in length and 1.5
cm. in diameter. The end of the cylinder near the cathode was covered
with platinum foil containing three slits. The central slit was about
1 cm. in length and about .3 cm. in width. The others were somewhat
smaller. A small receiving disk was placed about 2 cm. from the cathode
end of the cylinder. The cathode was separated 1-3 mm. from the end
of the cylinder. Mercury stood under both cathode and ionization
chamber and the apparatus was heated as usual. The ionization chamber
was kept charged to a constant positive potential of 24 volts. The
cathode was charged to a positive potential of less than 24 volts and
hence the electrons were accelerated as long as they were between
the cathode and ionization chamber. As soon, however, as they got
inside of the chamber they were retarded.
The gold-leaf electroscope was set to a sensitiveness of about .05 volt
per division. There was a condenser in parallel with it and the two
together-—condenser and instrument—had a capacity of about 230 e.s.u.
With it, therefore, it was possible for me to measure currents as large as
io~ 9 amperes, when charging it to 5 volts. I was not able, however, to
measure very small currents accurately, for I found that the passage of
the mercury vapor over the receiving disk charged it positively. This
was reduced to a minimum by arranging the apparatus so that the receiv-
ing disk was near the center of the furnace. But even under these
conditions many observations were vitiated, apparently, by a sudden
rush of vapor.
The results plotted in curve A', Fig. 4, were taken without the use
of the capacity mentioned above. The McLeod gauge registered 1.2
mm. The distance between the anode and cathode was 1.0 mm. The
cathode was new and uniformly coated with BaO and was heated by a
current of 19.5 amperes. The current flowing between the anode and
cathode varied from 14 X io~ 5 amperes at 4.5 volts to 23 X io~ 5 amperes
at 5.3 volts. The minimum potential difference between anode and
cathode has been used as abscissa. It is quite evident that ionization
must have started at about 4 volts, and as the drop in potential along
the cathode was .8 volt the ionization potential must have been in the
vicinity of 4.8 volts.
[SECOND
i;6 T. C. HEBB. L SERIES.
The results obtained with an old cathode are sometimes very marked
due to the fact that the BaO wears down to narrow patches at either end
of the cathode. Curve B\ Fig. 4, was plotted from data taken with
such a cathode. The McLeod gauge reading was .85 mm. It will be
noticed that the curve is very steep. As a matter of fact the ionization
current increased over one hundred times when the potential difference
between the anode and cathode was changed from 3.9 to 4.0 volts. It
was possible to estimate quite closely the potential drop in the cathode
at this point and this value—.9 volt—added to 4.0 volts gives 4.9 volts.
Fig. 4.
The curves shown—A' and Bf, Fig. 4—are similar to those obtained
by other experimenters but I do not believe that the results can be
explained on the assumption that the radiation X = 2536.7 has acted
photo-electrically on the receiving plate of the ionization chamber.
Although such an action must have existed, the current produced by such
action in these experiments must have been very small compared with
the currents measured. This was especially true as the receiving plate
had an area of only .25 square centimeter. And even when the receiving
plate consisted of a small platinum wire sealed in glass it was found that
the ionization current was still large.
Although ionization of mercury vapor occurred at 4.9 volts under
favorable conditions, as shown above, it was also possible to get it to
occur at any potential difference above 4.9 volts. This was accomplished
^• 3 XL ] IONIZATION POTENTIAL OF MERCURY VAPOR. LjJ
5. PHOTO-ELECTRIC EFFECT.
_ * •
Fig. 5.
set at 7.9 volts and the mercury arc in air started. If it sets up ionization
in the tube of sufficient amount, then one would expect the arc in the
vacuum to strike lower than 8 volts. No such effect was observed.
The ionization chamber was then arranged as in Fig. 5.
Two platinum cylinders A A were separated by a quartz test tube B and
together with C as a receiving plate constituted the ionization chamber.
A and A were joined electrically and a potential difference of 24 volts
was applied. No electrons from the cathode D could get into the ioniza-
tion chamber K. Mercury was kept in the chamber K as well as under
IONIZATION POTENTIAL OF MERCURY VAPOR. IJg
the arc DA. The current flowing between the anode and cathode was
measured and hence I could tell when ionization began. Some of the
radiation produced should pass into the chamber K and produce ioniza-
tion. Of course some of the radiation was absorbed by the quartz test
tube which was about I mm. in thickness and some was absorbed by the
mercury vapor which formed an unavoidable layer between two cylinders
A A. This layer was about .5 mm, in thickness. As mentioned before
the passage of mercury vapor over the receiving plate C causes it to
be charged with positive electricity. I could not obviate it in this case
as in the previously mentioned one and hence it was not possible for me
to detect extremely small currents but in no case did I detect any current
due to the radiation X = 2536.7. That the apparatus would have
responded as expected if the radiation X = 2536.7 had produced ioniza-
tion in sufficient quantity was proved by the fact that the slightest arc
between D and A produced a rapid charging of the electroscope. My
experiments along this line, therefore, have not shown, so far, any evi-
dence of an ionization of mercury vapor by the ratiation X = 2536.7.
SUMMARY.
1. These experiments prove conclusively that mercury vapor may be
ionized when bombarded with electrons moving with a velocity acquired
in falling through 4.9 volts and that the complete spectrum of mercury
is produced as a result.
2. Experimental evidence is given to show that this ionization is not
produced by the radiation X = 2536.7 acting photo-electrically on the
mercury vapor.
UNIVERSITY OF BRITISH COLUMBIA,
VANCOUVER, B. C.