Attempts to Produce the Rare Gases by Electric Discharge .
549 upon electrolysis accompanying the conduction of a current through the oxide .
Various suggestions as to the origin of the electron emission from glowing solids are discussed in the paper , and it is concluded that the experimental evidence points to the correctness of the original theory that an electron emission can arise as a result of thermal agitation alone , but that this emission may be aided or hindered in several ways .
Attempts to Produce the Rare Gases by Electric Discharge .
By Thomas B. Merton , B.Sc. ( Oxon .
) .
( Communicated by Prof. J. N. Collie , F.R.S. Received June 10 , \#151 ; Read .
June 25 , 1914 .
) The object of the present investigation was to obtain some further evidence on the apparent production of neon and helium by electric discharges in vacuum tubes , which has been described by Collie and Patterson.* Struttf has recently attempted to repeat these experiments , using methods somewhat different to those of Collie and Patterson , with negative results .
It is evident that in experiments of this kind the most rigorous precautions must be taken to exclude air from the apparatus , for although the rare gases are only present in the atmosphere in relatively small quantities , an exceedingly small quantity of air is required to give evidence of their presence .
Strutt ( loc. cit. ) has been able to detect with his apparatus the neon in 0*01 c.c. of air and has expressed the opinion that under favourable conditions the neon might be detected in 0*001 c.c. The method of the present investigation differs essentially from that used either by Collie and Patterson or by Strutt .
The apparatus , shown in the figure , consists of a bulb A connected to a glass tube L , which passes very nearly to the bottom of the bulb B. From the tube L are two branch tubes C and D , of which C leads to the sparking tube S and palladium tube P , and D leads to the stopcock H and thence to the mercury pump .
The tube D is connected by a bye-pabs E , with a stopcock , to the bulb B , which is provided with another branch tube with a stopcock F. At the top of the bulb A is a fine capillary tube K , into which is sealed a fine * 'Trans .
Chem. Soc.,5 vol. 103 , p. 419 ( 1913 ) .
t ' Proc. Roy .
Soc. , ' A , vol. 89 , p. 499 ( 1914 ) .
Mr. T. R Merton .
Attempts to platinum wire .
There is also a platinum wire sealed into the tube L. P is a short length of palladium tube soldered with gold to a platinum tube , which To Mercury Pump f is in turn sealed into the glass tube C. The sparking tube S was provided with stout electrodes fastened to platinum wires and sealed into the tube in the usual way .
The experiments were conducted as follows : The bulb B was nearly filled with mercury .
Taps G and H were opened and F was closed , and the apparatus was exhausted by the mercury pump until no discharge could be passed through the sparking tube S. During this process , a heavy discharge from a ten-inch induction coil was passed through the tube , which was also heated at intervals with a bunsen burner .
Hydrogen was admitted by heating the palladium tube P in a bunsen flame , and the hydrogen so admitted was pumped off .
This process of admitting and pumping off ' hydrogen was frequently repeated in order to wash any traces of other gases completely from the apparatus .
Stopcocks G and H were then closed and air wTas cautiously admitted into the bulb B until the mercury stood at the level shown by the dotted line in the figure .
A and S Produce the Rare Gases by Electric Discharge .
551 were thus completely protected from any possible leakage from the stopcocks by the mercury seal .
A moderately heavy discharge from the induction coil was then passed through the sparking tube , hydrogen being admitted through the palladium tube at intervals when the vacuum became too high .
After the conclusion of the sparking the hydrogen remaining in the apparatus was almost completely removed by heating the palladium tube by means of a small spiral of platinum wire maintained at a red heat by an electric current .
This operation presumably reduced the pressure of hydrogen in the apparatus to the partial pressure of hydrogen in the atmosphere .
During this process the sparking tube , which was usually coated writh a dense metallic mirror from the electrodes , was frequently heated .
The gas remaining in the bulb A was then driven into the capillary by slowly admitting air into the bulb B through the stopcock F , and examined spectroscopically by passing a weak discharge from the induction coil through the platinum wire in the capillary , connection to the mercury being made through the wire at L. Before the commencement of the sparking , the absence of atmospheric contamination could be shown by driving the gas in the bulb A into the capillary and examining it spectroscopically .
In the first four experiments moderately heavy discharges were passed through the sparking tubes , which were provided with aluminium electrodes , for periods of from six to fourteen hours .
The residual gases in these experiments showed , after the sparking , only lines due to mercury with a trace of hydrogen and sometimes a trace of carbon .
The latter was rapidly absorbed by the mercury when the discharge was passed .
An experiment in which the electrodes were of platinum and another in which the sparking tube contained a globule of mercury , also gave negative results .
In the next two experiments aluminium electrodes were used and the discharge from the coil was rectified by inserting a spark gap in the circuit , consisting of a fine point and a flat metal plate , the latter being connected to the negative pole of the coil , the experiments being in other respects conducted in exactly the same manner .
In these two experiments the residual gases showed the spectrum of argon .
No trace of nitrogen was found .
Three more experiments were conducted without the rectifier and in no case was any argon found .
In the next experiment a valve tube rectifier was substituted for the spark gap which had been previously used , and a trace of argon was again found .
The experiments thus far seemed to indicate that , whatever might be the source of the argon , a rectified discharge was essential to its appearance , but Mr. T. R. Merton .
Attempts to numerous subsequent experiments have shown that there is no systematic connection between the character of the discharge and the appearance of the argon in the residual gases .
Two experiments were carried out to discover whether the argon could be traced to the gases liberated from the electrodes .
In one experiment the electrodes were only sparked for a short time during the process of pumping and washing out the tube , and at the conclusion of the experiment a trace of argon was found .
In the second experiment no discharge was passed until the pumping and washing out was concluded .
A large quantity of gas was liberated from the electrodes on sparking and was rapidly absorbed by the aluminium mirror , but no argon was found .
In both of these experiments the discharge was rectified with the valve tube , and it is concluded that the argon cannot be traced to the gases liberated from the electrodes .
An experiment was conducted in which the washing out of the apparatus with hydrogen was omitted .
A trace of air therefore remained in the apparatus .
At the conclusion of the experiment no nitrogen , oxygen or argon were found .
Both the diatomic gases and the argon had evidently been absorbed by the electrodes or the aluminium mirror and the amount of argon driven off on heating the tube was too small to detect .
A similar experiment in which the sparking was only continued for about twenty minutes , so that no appreciable aluminium mirror had formed , gave a distinct argon spectrum , with no nitrogen .
It follows from these two experiments that the presence of argon in the residual gases cannot be explained by the presence of a minute quantity of air remaining in the apparatus if the sparking is continued for a considerable time .
The presence of the argon , however , can be accounted for by an exceedingly small , but continuous leak , since the diatomic gases appear to be far more rapidly absorbed by the aluminium mirror than the argon .
With the apparatus used a minute continuous leak can occur under certain circumstances .
It was found that a small amount of dirt had collected on the tube L which constitutes the barometric seal .
A new apparatus was , therefore , taken and carefully cleaned with fuming nitric acid , and the mercury , which had been purified by passage in a fine stream through dilute nitric acid , was poured into the apparatus without coming into contact with any tap grease .
Since these precautions were adopted , no trace of argon has been found after an electric discharge had been passed in the manner described above ; many experiments have been made with electrodes of aluminium , copper , nickel , silver , platinum and palladium , and all have given negative results .
In some experiments an air trap was inserted in the tube L as an additional Produce the Rare Gases by Electric Discharge .
precaution , but it seems as if this is unnecessary if the mercury and glass are scrupulously clean .
In order to test whether the appearance of argon could occur through a slow leak of air into the apparatus through stopcocks , three experiments have been carried out in which the mercury was not driven up to the level of the dotted line in the figure , the vacuum being maintained by the stopcocks .
These stopcocks gave every appearance of being reliable , but in two cases out of the three a trace of argon was found .
The spectroscopic test of argon appears to be more delicate than that of neon , and it follows therefore , from Strutt 's ( loc. ait .
) estimate for neon , and from the fact that there is about a thousand times as much argon as neon in the atmosphere , that a millionth of a cubic centimetre of air might be sufficient for the detection of argon .
From the above experiments it is considered doubtful whether stopcocks can be relied on in dealing with quantities of gas of that order of magnitude .
These experiments seem to suggest that the appearance of argon should be treated as an exceedingly delicate test for a leak in the apparatus , and the entire results of the investigation have , therefore been negative , since in the numerous experiments under different conditions in which no argon has been found , no trace of neon or helium has been observed .
It must be pointed out that both Collie and Patterson ( loc. cit. ) and Strutt { loc. cit. ) removed the diatomic impurities in the residual gases by means of charcoal cooled with liquid air , a process which would also have removed any argon present .
In the present investigation , it is highly probable that in those experiments in which argon was found in sufficient quantity , the neon and helium would have been detected if the argon had been removed in this way .
It must , however , be remembered that the leak of air which can be revealed by argon is exceedingly small , and about one thousandth of the quantity which could be detected if the test used were that of the recognition of neon , and that therefore the presence of recognisable quantities of neon , and more particularly of helium , would indicate a leak greater than it appears possible to attribute to the inefficiency of stopcocks .
I wish to express my indebtedness to Prof. Collie and to Mr. Masson for their great kindness in giving me the benefit of their experience and for much invaluable advice .
It is a disappointment to me to find that I have been unable to reproduce the conditions necessary for the production of neon or helium , and have obtained only negative results .
I also wish to thank Prof. H. Jackson for the valuable advice which he has given me .

