A Chemically Active Modification of Nitrogen .
219 determination must be uncertain , since a wide variation in the assumed degree of dissociation corresponds in most instances to a very small change in the solubility , any decrease in undissociated substance dissolved being almost entirely balanced by an increase in dissociated , and vice versa .
This fact explains also why , although the degrees of dissociation of the strong acids used in the above research are not yet accurately determined , being liable to an uncertainty of several per cent , in dilute solutions , the solubility results of other acids with them , as calculated from these values , are subject to a much smaller degree of error .
In conclusion , I have much pleasure in expressing my thanks to Prof. Walker , at whose suggestion and under whose direction the above research was carried out , for his advice and assistance during the whole period of its execution .
Bakerian Lecture.\#151 ; A Chemically Active Modification of Nitrogen , Produced by the Electric Discharge .
By the Hon. B. J. Strutt , F.B.S. , Professor of Physics , Imperial College of Science , South Kensington .
( Received March 16 , \#151 ; Lecture delivered April 6 , 1911 .
) [ Plate 8 .
] .
) , \#166 ; UP S 1 .
Afterglow It is known that vacuum tubes frequently show a luminosity of the\gt ; contained gas after discharge is over .
In a previous paper* I was able to show that this effect , as it occurs in air , is of the nature of a phosphorescent combustion , and is due to the mutual reaction of nitric oxide and ozone , each formed in the discharge .
In a second paperf it was shown that other phosphorescent combustions can be observed in ozone , notably of sulphur , sulphuretted hydrogen , acetylene , and iodine .
Some of these give continuous spectra , but the majority band spectra .
In the first paper it was stated that pure nitrogen gives no afterglow whatever , and , with the simple induction coil discharge with which I was then working , this has been frequently verified since .
Mr. Percival Lewis has however , described an afterglow obtained in nitrogen when a jar * ' Phys. Soc. Proc. , ' Dec. 15 , 1910 , vol. 23 , p. 66 .
t Ibid. , Feb. 15 , 1911 , vol. 23 , p. 147 .
Hon. R J. Strutt .
A Chemicallij [ Mar. 16 , discharge with spark-gap is used.* I had no difficulty in obtaining this glow as soon as the jar discharge was used , and have applied to its examination the method used in the former papers .
This is due to Sir James Dewar , f and consists in drawing a current of gas through the vacuum tube into an observing vessel , where the glow is developed , and thence into an air pump , which must be a mechanical one of good construction , driven by power .
It is thus possible to examine the properties of the glowing nitrogen much more satisfactorily than can be done by intermittent examination after successive discharges .
As Lewis observed , the glow has a characteristic band spectrum not known in any other connection .
Its most conspicuous features in the visual region are a green , a yellow , and a red band , of not very unequal intensity .
The yellow one is usually the brightest , and imparts a yellow colour to the glow , not very dissimilar to that of the afterglow in air , due to the union of nitric oxide and ozone .
The two glows are , however , instantly distinguished by the spectroscope , the air glow giving a continuous spectrum .
I have obtained the nitrogen afterglow intense enough to be conspicuous to an observer 30 feet off , when it was 18 inches below a 32-candle electric lamp .
With regard to the conditions for its production , my observations do not altogether agree with those of Lewis .
He obtained it with nitrogen from sodium nitrite and ammonium chloride , but was unable to do so from atmospheric nitrogen .
He was disposed to regard the presence of a trace of nitric oxide as necessary .
I obtained it first from air which had passed over red-hot copper , and afterwards from air which had been freed from oxygen in the wet way by Hempel 's copper and ammonia method .
J It has been shown by Yon Mose'ngeilS that nitrogen completely freed from oxygen by sodium can give the afterglow .
I have confirmed this by bubbling atmospheric nitrogen through melted phosphorus , and thence into the vacuum tube .
The glow seemed rather improved than otherwise by this treatment .
No luminosity was observed in the phosphorus .
Nitrogen prepared by burning phosphorus in air under a large bell-jar gives the glow , but not well .
It is not free enough from oxygen .
In short , all the evidence obtained points to the conclusion that the glow is connected with nitrogen , and nothing but nitrogen .
For experiments on the glow7 , particularly if they are to be prolonged , it is * 'Ann .
d. Phys. , ' 1900 , vol. 2 , p. 466 ; 'Phys .
Zeit .
, ' 1904 , vol. 5 , p. 546 .
t 'Roy .
Inst. Proc. , ' 1888 , and 'Engineering , ' June 18 , 1909 .
f It is necessary to remove carbon dioxide and ammonia by appropriate absorbents when using this method , or the glow fails .
Special drying of the nitrogen , on the other hand , seems to be of no advantage .
S ' Ann. d. Phys. , ' 1906 , vol. 20 , p. 833 .
Active Modification of Nitrogen .
1911 .
] convenient to make use of commercial cylinders of compressed nitrogen , provided with the usual automatic regulators as used in connection with the limelight .
Commercial nitrogen is not free enough from oxygen for immediate use , but it can readily be made so by passage through a tube filled with freshly cut or cast fragments of phosphorus .
A glow is seen where the nitrogen enters , indicating that an absorption of residual oxygen is in progress .
This only extends a short way down the tube , and the darkness of the remainder is a guarantee that absorption is complete , and that the issuing nitrogen is fit for use .
The phosphorus tube should not be unnecessarily exposed to daylight , as this causes deterioration of the absorbent surface .
S 2 .
Effect of Temperature .
If a long tube , through which a stream of glowing nitrogen passes , is moderately heated , the glow is locally extinguished .
As the gas passes on to a cooler part of the tube its luminosity is recovered .
If , on the other hand , the gas is led through a tube immersed in liquid air , it glows out with increased brilliancy where it approaches the liquid air .
The luminosity is completely and finally extinguished when , or before , the fully cooled portion of the tube is reached .
This increased brilliancy is obviously the counterpart of the temporary obscurity observed with heat .
Whatever the atomic or molecular change may be , of which the afterglow is a sign , these experiments show that it is accelerated by cooling , and retarded by heating .
The kind of change which may be expected to behave in this way is an association , e.g. of dissociated nitrogen atoms into molecular nitrogen .
It may be doubted , however , whether the visual glow is a perfect measure of the changes in progress .
I find that if a tube carrying the glow is strongly heated the glow is permanently extinguished .
It is clear that the view suggested above is incomplete .
S 3 .
Effect of an Electric Field .
The question will naturally present itself whether the gaseous ions of the discharge , which no doubt remain in the glowing nitrogen , have any direct connection with its peculiarities .
If this were the case , we should expect that by passing the glowing gas through an electric field so as to remove the ions , the glow , and its power of exciting metallic spectra , to to be presently described , would be destroyed .
Glowing nitrogen was passed down a tube about 40 cm .
long and 2 cm .
in diameter , provided internally with tin-foil strips along the whole length on the opposite sides .
The glow was in no way affected at any part of the tube , if an E.M.F. Hon. II .
J. Strutt .
A Chemically [ Mar. 16 , of 200 volts was applied to these strips .
Nor was its power of exciting the spectrum of some sodium vapour at the further end at all diminished , ( see below , S 5 ) .
S 4 .
Action of the After-glowing Nitrogen on the Non-metcds .
The glowing nitrogen has remarkable chemical properties .
Drawing it by the pump over a small pellet of phosphorus a violent reaction occurs , red phosphorus is formed , * and the yellow glow is quenched .
At the same time the gas is absorbed .
This latter experiment can readily be shown by means of a tube illustrated in fig. 1 .
The tube is in this case closed , and has no current of nitrogen through it .
The discharge passes between the electrodes A and B as indicated by the dotted line .
The afterglow diffuses into the annex C , where its characteristic colour and spectrum may be seen .
On the other hand , phosphorus vapour diffuses out from the pellet of that substance at D , and they meet at a point E , where the reaction occurs , and red phosphorus is deposited as a film along the walls of the tube .
The changed appearance of the discharge shows that absorption is in progress .
Beginning with enough nitrogen to give a threadlike discharge , in a few minutes the pressure is sufficiently reduced to broaden out the discharge so as to fill the tube .
Shortly after that the vacuum becomes quite high , and conduction ceases .
Pteturning to the experiment where a continuous stream of nitrogen is maintained by the pump , it is found that if the Leyden jar is suppressed , and the ordinary coil discharge employed , a piece of phosphorus in the stream of nitrogen is very little acted on .
There is just perceptible action in a few minutes , and a faint greenish flame may be seen round the phosphorus .
The same flame may be seen more conspicuously with the jar discharge , but only for a moment ; for it is almost immediately obscured in this case by the opaque deposit of red phosphorus on the walls of the tube .
No doubt the ordinary coil discharge has the same kind of effect as the jar discharge , but in a degree too slight to make the afterglow visible , and only apparent by prolonged action on phosphorus .
* Other cases are known where the reaction of phosphorus with another body convert the excess into red phosphorus .
Iodine behaves thus .
FIG. 1 .
Active Modification of Nitrogen .
1911 .
] The novelty of these observations lies in the fact that the phosphorus vapour employed has no access to the region of discharge .
It has long been known that phosphorus vapour in the region of discharge would combine with nitrogen , and the fact has been made use of since 1893 in the exhaustion of incandescent lamps.* Discharge takes place through the rarefied air between the ends of the filament .
The E.M.F. of , say , 100 volts between these is amply sufficient to produce it with a white-hot cathode .
Sir Oliver Lodge , too , has used phosphorus in exhausting his vacuum valves.f No one , however , seems to have suspected that nitrogen could continue to react with phosphorus after it had left the discharge , and even after it had been deprived of ions by passing through an electric field .
Nor has the reaction been connected with the afterglow .
I have made use of the reaction with phosphorus to determine what percentage of active nitrogen is present in the gas shortly after leaving the discharge tube .
The principle of the method is to determine the gain in weight of the phosphorus after a measured volume of nitrogen has been passed over it .
Fig. 2 shows the apparatus .
Active nitrogen enters from a vacuum tube attached by a short indiarubber connector to the stopcock A. It passes over fragments of phosphorus at B , where it is absorbed with deposition of red phosphorus , and the glow disappears .
The gas then passes through a U-tube C , cooled in liquid air .
This is to prevent any phosphorus vapour being carried off .
Finally , the gas leaves through a stopcock D. The entire arrangement can be detached and hung on the balance by the wire loop E. The phosphorus is originally introduced through F , which is then sealed .
The weighings were always carried out with the tube vacuous , and every precaution was taken against the entrance of moisture , which might lead to erroneous results .
In a typical experiment 2540 c.c. of nitrogen were passed , and the gain in weight was 15'5 mgrms .
Thus the nitrogen absorbed was 12'2 c.c. , , * See , for instance , J. Rodet , 'Lampes a Incandescence , ' p. 107 , Paris , Gauthier-Villars , 1907 . . ?
+ Patent No. 25047 , 1905 .
FIG. 2 .
Hon. R. J. Strutt .
A Chemically [ Mar. 16 , 1/ 210 part of the whole .
It appears that the percentage of nitrogen converted into the active form is comparable with the percentage of oxygen converted in an ozoniser .
The glowing nitrogen also exhibits remarkable phenomena when led over iodine .
Its normal yellow glow is replaced by a magnificent light blue flame , at the place where it mingles with the iodine vapour .
The appearance is represented in the coloured plate .
A slight rise of temperature is observed where the blue glow originates .
The iodine is quite volatile enough without the application of heat .
I have not succeeded in demonstrating the absorption of nitrogen in this case , nor have I been able to see any signs of a compound deposited further on in the tube .
It may be masked by the excess of iodine .
The flame gives a magnificent spectrum of broad bands .
Details will follow in a later paper .
Brief notes will now be given on the behaviour of a few other non-metallic elements .
Sulphur , when warmed in the current of glowing nitrogen , extinguished it ; and , with somewhat stronger heating , a blue flame or glow was apparent , though not comparable in volume or intensity with that of iodine .
A transparent green deposit is formed on the glass .
Selenium had no effect .
Arsenic , heated in the glow , gave a not very conspicuous green flame , with a spectrum which seemed continuous , or , at all events , lacking in conspicuous features , and extended over the whole visual region .
Antimony and carbon gave no effect .
When hydrogen is admitted to mix with the glowing nitrogen , after the latter has been through the vacuum tube , no effect is produced , beyond a mere dilution of the glow .
Oxygen , on the other hand , destroys the glow without any other luminosity being developed in the process .
| 5 .
Action on the Metals\#151 ; Production of Metallic Line Spectra .
When the glowing nitrogen is led over a fragment of sodium , which is heated a little above its melting point , the sodium spectrum is developed with great brilliancy .
Indeed , the power of developing the D line is a more sensitive test for the presence of the active nitrogen than any other .
When the sodium is more strongly heated , say , to 250 ' C. , a curious effect is observed .
The denser vapour in the immediate neighbourhood of the metal becomes visually green , showing the green line E very strongly , while the D line is scarcely visible .
On either side of the central green light is an outer luminous region in which the D line predominates .
Strutt .
Koy .
ooc .
Froc .
, vot .
8d , 8 .
19H .
] Active Modification of Nitrogen .
I have been able to observe the absorption of the glowing nitrogen by sodium , by the same method as was employed with phosphorus .
The sodium vapour was not allowed to penetrate into the discharge , and its spectrum was not observed there.* There seems to be no reason for doubting that the sodium spectrum seen in these experiments is simply the flame spectrum of sodium burning spontaneously in the active nitrogen , to form the nitride .
It opens up a new field of experiment to be able to produce metallic spectra in a vessel at so low an average temperature , and in the absence of an electric field .
The line spectra of many other metals have been obtained similarly .
The following have been observed:\#151 ; Cadmium , magnesium , mercury , potassium , zinc , lead .
Thallium was tried in the form of chloride , and gave a magnificent green light , in striking contrast to the yellow afterglow which it replaces .
The sodium spectrum , too , can be obtained when the glow passes through a soda glass tube , heated to near the softening point , but in this case the I ) line alone appears .
In the case of metallic mercury , absorption of nitrogen was proved , as with sodium and phosphorus , though , working in a closed vessel without a current of gas , it was not found feasible in practice to prevent mercury vapour finding its way into the discharge .
There is no doubt that this is the action described by Threlfall.f He observed the absorption of nitrogen when a jar discharge was passed through it at a low pressure in presence of mercury , though there is no reference to the afterglow in his paper .
He observed , too , the formation of a compound which decomposed with a slight explosion when moderately heated .
Although in the absorption experiments mercury could not be prevented from getting into the discharge , it was easy to avoid this when the air pump was used , and a continuous current of nitrogen drawn through .
Working thus , I have observed that Threlfall 's explosive nitride is formed when the mercury spectrum is developed by the afterglow .
This seems conclusive evidence that the production of these spectra is a direct result of the chemical union of the metal with the active nitrogen .
It may be remarked that , though the hot metals ( .
, zinc ) are rapidly skinned over with nitride , cold metals remain bright in the nitrogen glow .
This was tried particularly with sodium , and with mercury\#151 ; the latter in the form of a bright film on a copper plate .
The clean cold metals have no unfavourable influence on the luminosity of the glow itself .
* This distinguishes the present action from the well-known absorbing power of the alkali metals when used as cathode in a discharge .
+ 'Phil .
Mag. , ' Jan. , 1893 , p. 1 .
Hon. R. J. Strutt .
A Chemically [ Mar. 16 , S 6 .
Action on Compound Bodies .
Some gases and vapours mixed with the glow merely dilute it .
Water and carbon dioxide are examples .
In one case an obvious chemical attack has been observed .
This was with naphthalene , which turns brown , a brown deposit forming also in the walls of the tube where the nitrogen is passed over it .
The nitrogen glow was destroyed , but no luminous phenomena accompanied the action .
Heat was not applied .
Ammonia was another case where the original glow was destroyed without attendant luminous effects .
It seems likely that a chemical action occurs here , but proof has not been obtained .
It is certain , however , that in some cases destruction of the glow by contact with another substance may occur without chemical action .
Manganese dioxide and copper oxide are at once fatal to it .
The analogy to the known destruction of ozone by these substances cannot fail to attract attention .
A small roll of superficially oxidised copper gauze was placed in a glass tube and carefully weighed .
It was then inserted between the vacuum tube and the pump by means of indiarubber connectors , and a stream of the glowing nitrogen passed over it for about half an hour .
The glow abruptly stopped at the surface of the copper oxide .
The tube was detached and reweighed .
No change of weight amounting to 1/ 10 mgrm .
could be detected , though the quantity of active nitrogen passed , judged by the phosphorus experiments described above , must have amounted to several milligrammes .
I conclude that the glow is destroyed in these cases by a catalytic action .
Another class of compound substances become luminous with characteristic band spectra when vaporised in the glow .
Stannous and stannic chloride both give the same brilliant and voluminous blue glow .
Most of the light comes from one broad symmetrical band in the blue and violet regions .
There are , in addition , a number of lines in the ultra-violet , chiefly due to tin .
Mercuric iodide gives a violet glow with a strong unsymmetrical violet band distinct from any in the iodine afterglow spectrum mentioned in S 4 .
Distinctive glows have also been obtained from other mercury salts .
Mercurous chloride gives a green one .
Cuprous chloride gives a blue-green glow , with a spectrum agreeing with that shown by the same compound in the Bunsen flame , but with additional features not present in the latter .
Cyanogen gas fed into the glow gives the lilac flame of cyanogen with its well-known spectrum .
Active Modification 1911 .
] The band spectrum of iodine in the afterglow ( see S 4 ) should probably be regarded as of similar origin with these compound spectra .
They are I think , essentially flame spectra , as are also the line spectra of metals produced in the afterglow ; but the afterglow must be regarded as a much more searching kind of flame than the ordinary high-temperature flames .
In much of the literature of flame spectra it is implied that temperature is a measure of the spectrum-developing quality of a flame , and no doubt the fact that the oxy-acetylene and oxy-hydrogen flames are , each in its own degree , capable of developing more lines than the Bunsen flame , is favourable to this view .
Yet the balance of evidence is greatly against the idea that heat alone is capable of exciting line and band spectra at all ; and if we reject this idea it is evidently unreasonable to assume that temperature alone governs the number of lines or bands emitted .
Regarding , then , the afterglow as a flame capable of developing spectra , we have the means of examining the spectra of a number of compounds which will not endure the temperature of an ordinary flame without decomposition .
Moreover , the spectrum is better developed.- Cuprous chloride , which gives a characteristic spectrum in the Bunsen flame , gives the same spectrum , with additional details , in the afterglow .
Compounds like mercuric iodide , instantly decomposed in the Bunsen flame , give in the glow a spectrum not before recognised .
It may be possible to obtain valuable generalisations from the examination of a long series of such spectra .
Nitric oxide , allowed to mix with the active nitrogen , shows a very strange behaviour .
A greenish flame , possessing a continuous spectrum , is produced , and heat is developed at the point of confluence .
To test whether any gas condensable at \#151 ; 180 ' was produced , the gases from the flame were led through a U-tube cooled in liquid air , and a dark blue substance was condensed out .
This melted to an indigo-blue liquid , and finally revealed itself as nitrogen peroxide by evaporating off into an orange gas , soluble in caustic alkali .
It is very surprising that a reaction between nitrogen and nitric oxide should lead to the formation of a substance , not less , but oxidised than the latter .
A critic may naturally object that the formation of nitrogen peroxide may have been due to oxygen which accidentally gained admission to the nitric oxide , quite independently of the active nitrogen .
I wish to emphasise a test experiment , which disproves this hypothesis .
The influence of the active nitrogen may be removed by simply turning off the electric discharge .
The apparatus being in adjustment , with a suitably regulated flow of the gases , the U-tube was kept cool with liquid air for Hon. R J. Strutt .
A Chemically [ Mar. 16 , 10 minutes , after which the liquid air was removed , and the tube allowed to warm up somewhat , for the removal of hoar frost from the outside .
No trace of the indigo liquid could be seen .
Now the discharge was started , and the tube kept in liquid air for only two minutes .
At the end of this time the indigo liquid was extremely conspicuous .
The tube could be warmed up to allow the stream of gases to wash out the product , and the experiment could be repeated , on and off , as often as desired .
The supply of gas , adjustment of taps , etc. , were left absolutely untouched throughout .
It would seem that the reaction which occurs may be of this kind :\#151 ; 2NO + N = N02 + N2 .
However , I lay no stress on this .
I wish only to insist on the definite facts above stated .
Brilliant glows are obtained from some carbon compounds containing the halogens , for instance , a lilac glow from ethyl iodide .
This , examined spectroscopically , shows a magnificent cyanogen spectrum .
The fact is evidence that the active nitrogen attacks the carbon compound , combining with the carbon ; iodine is set free , and may be collected as a conspicuous sublimate by cooling the resulting vapours in liquid air .
What becomes of the hydrogen has not yet been determined .
Chloroform and carbon tetrachloride also give a fine cyanogen spectrum .
In this case the glow is visually orange , owing to greater relative intensity in the red portion of the cyanogen spectrum .
Chlorine is set free , and can be collected in liquid air .
Again , the lilac cyanogen glow is obtained when acetylene is fed in .
The experiment mentioned above , where cyanogen itself is shown to give this glow under the same conditions , suggests very strongly that in all these cases the active nitrogen attacks the carbon compound , forming cyanogen , which is merely stimulated , without further change , by the peculiar conditions existing in the glow .
The formation of cyanogen was proved more directly as follows:\#151 ; The gases from the acetylene experiment were condensed by liquid air , and afterwards collected through a Toepler pump .
They were agitated with caustic potash solution to absorb any cyanogen present from the excess of acetylene .
The presence of a cyanide in this solution was readily proved by abundant formation of Prussian blue , and by the ferric sulphocyanate reaction .
It should be mentioned that , in addition to the condensed gases , some black tarry substance was present in the tube .
Finally , the cyanogen spectrum has been observed when the active nitrogen is allowed to react with methane , pentane , ethylene , Icohol , ether , and 1911 .
] A dive Modification of Nitrogen .
benzol .
It is not , however , so conspicuous in these cases , and probably does not indicate the main course of the reaction .
S 7 .
Summary and Conclusion .
The leading facts established are:\#151 ; ( 1 ) That pure nitrogen , from whatever source , subjected to the jar discharge , undergoes some modification which causes it to glow for a short time after it has left the discharge .
( 2 ) The glow which is emitted while the gas returns to its normal condition is not affected by the removal of ions .
It is weakened by heating , intensified by cooling .
This seems to favour the view that it is due to the recombination of dissociated atoms .
( 3 ) The modified nitrogen acts on ordinary phosphorus , combining with it , and at the same time forming much red phosphorus .
( 4 ) It combines with sodium and also with mercury at a gentle heat , forming in the latter case an explosive compound , and in each case developing the line spectrum of the metal concerned .
It also develops the line spectra of other metals , probably combining with them too .
( 5 ) It develops the band spectra of compounds when they are vaporised in it , giving , in many cases , spectra of substances tot ) unstable to be examined at the temperature of the Bunsen flame .
( 6 ) It attacks nitric oxide , with formation , strangely enough , of nitrogen peroxide , a more oxidised substance .
( 7 ) It attacks acetylene and the halogen derivatives of organic radicles , setting free the halogen , where one is present , and combining with the carbon to form cyanogen .
This is proved by the brilliant cyanogen spectrum produced , and by direct chemical tests .
It may perhaps be felt that more detailed study of the compounds produced by the active nitrogen should have been made .
The importance of this is not underrated , but the difficulties of working with such small .
quantities of material are considerable , and have led to a postponement of this branch of the work .
It may be pointed out how little has been done in studying the compounds produced by ozone .
The spectroscopic data in this paper are merely general and preliminary .
An accurate investigation of the various spectra by photographic methods , in collaboration with Prof. A. Fowler , F.R.S. , is in progress .
VOL. lxxxv.\#151 ; A. s

