The Direct Synthesis of Ammonia .
fluorescent light .
It is true that it would not depend upon the amplitudes of the vibrations ; but if new free periods are excited as I have described , the intensity and the absorption would both depend upon the number , and the duration of the periods thus produced , and it is this which , I think , the change of absorption in fluorescence most distinctly proves .
The Direct Synthesis of Ammonia .
By Edgar Philip Perman , D.Sc .
, Assistant Professor of Chemistry at University College , Cardiff .
( Communicated by Principal E. H. Griffiths , F.R.S. Received January 3 , and in revised form , March 4 , \#151 ; Read March 30 , 1905 .
) it was shown in a recent paper on " The Decomposition of Ammonia by Heat , " * that ammonia is decomposed almost ( if not quite ) completely when heated in a porcelain vessel at about 800 ' to 1100 ' , and that there is no sign of equilibrium between the ammonia and its decomposition products at any of the temperatures employed , 677 ' to 1111 ' .
In order to discover if there is any such state of equilibrium , it was thought better to attempt to reach that state by synthesising ammonia instead of decomposing it , as the testing for and estimation of the ammonia could then be carried out with much greater accuracy .
Preparation of the Mixed Gases.\#151 ; In the first series of experiments , the mixture of nitrogen and hydrogen was made by passing ammonia gas , from a strong aqueous solution , through a red-hot iron tube heated in a gas furnace , and the resulting gases were collected in a large gasholder and stored over dilute sulphuric acid .
It may be objected to this method of preparation that carbon monoxide , hydrogen , and other gases would percolate through the iron , and contaminate the product obtained .
In order to test this point , a careful analysis of the gases was made , with the result that no carbon monoxide could be detected by the blood test ; no carbon dioxide was found after exploding the gases with oxygen ; and the ratio of nitrogen to hydrogen was found to be correct .
The mixture of gases made by this method will be referred to as " Mixed Gases I. " It was thought , nevertheless , that traces of foreign gases might have escaped detection , and might possibly have influenced the results .
Consequently a second series of experiments was carried out with a mixture of * 'Roy .
Soc. Proc. , ' vol. 74 , p. 110 , 1904 .
VOL. LXXVI.\#151 ; A. N Prof. E. P. Perman .
[ Mar. 4 , nitrogen and hydrogen made in an entirely different way .
Nitrogen gas was made by heating gently a solution of equivalent quantities , of ammonium chloride and sodium nitrite , and hydrogen was prepared from a concentrated potash solution heated with metallic aluminium , the gas being passed through two Drechsel flasks containing potassium permanganate solution .
Each gas was stored in a gasholder , and a mixture of the two was then made , in the proportion of 1 volume nitrogen to 3 volumes hydrogen , in a third gasholder .
The volume of gases was measured by the volume of water drawn off , due regard being paid to the " head " of water .
This mixture will be called " Mixed Gases II .
" Attempted Synthesis by Heat .
( Mixed Gases I , and II , separately.)\#151 ; The mixture was proved to be free from ammonia by testing with Nessler 's solution .
It was then passed through a hard glass tube heated in a combustion furnace .
Glass was chosen as likely to have no chemical action on ammonia or its constituents .
The temperature was varied from about 600 ' to 1000 ' C. The resulting gases were bubbled through dilute hydrochloric acid solution ; this was afterwards made alkaline with potash , and Nessler 's solution added .
The result was that no ammonia could be , whatever the temperature or state of gases as to moisture .
In some experiments the mixture was freed from traces of oxygen by bubbling through alkaline pyrogallate solution , and then dried by strong sulphuric acid ; in other experiments these precautions were omitted , but the result was always the same , and was obtained many times .
The experiment was varied by filling a porcelain globe with nitrogen and hydrogen ( Mixed Gases I ) , and heating it in a furnace to a bright red heat for about 1J hours .
The gases were then tested for ammonia in the usual way , but not a trace was found .
We may conclude , therefore , from these experiments , that ammonia cannot be synthesised from nitrogen and hydrogen by heating in vessels of glass or porcelain , or that , if it is formed , it is not in sufficient quantity to be detected by Nessler 's solution .
Synthesis by Heat in the Presence of Iron .
( Mixed Gases I.)\#151 ; Some of the mixed gases were then passed through an iron tube heated to redness , or in some cases a glass tube containing iron nails , and it was found that , when moisture was present , traces of ammonia were formed ; if , however , care was taken to exclude moisture by passing the gases through alkaline pyrogallate solution and sulphuric acid , and reducing any iron oxide , then no ammonia could be detected .
This result was obtained also by 11 am say and Young.* In order to form an idea of the amount of ammonia produced , known * ' Chem. Soc. Jovirn.f vol. 44 , p. 88 , 1884 .
1905 .
] The Direct Synthesis of Ammonia .
volumes of the mixed gases were passed through a hard glass tube packed with " French nails , " and the ammonia was estimated by Kessler 's solution , as in the method used in water analysis .
The following results were obtained :\#151 ; Vol. of mixed gases , c.c. Approximate rate , litres per hour .
10 10 Ammonia .
milligrammes .
003 006 0-08 -0T0 Ko ammonia could be detected in any case unless the iron was at a bright red heat , about 800 ' to 900 ' .
With Mixed Gases II:\#151 ; Time .
Vol. of mixed gases .
Ammonia .
/ n c.c. milligrammes .
8 20 250 0-20 2 45 250 0-20 0 42 250 017 It will be noticed that the maximum amount of ammonia was formed when the gases were passed at the middle rate , indicating that the mixture had come into equilibrium .
More than twice as much ammonia per litre of mixed gases was formed in this series as in the first , and it appeared to be formed at a lower temperature .
Moreover , when the gases were carefully dried by sulphuric acid , traces of ammonia were still found .
Whether this difference was due to greater purity of the gases or to any other variation in the conditions ( e.g. , the new French nails ) I am unable to say .
My attention has been called to a paper by Haber and van Oordt* in which some very similar experiments are described .
The proportion of ammonia formed in the experiments of these investigators was about 02 to 1000 possible ( if completely converted ) , which is considerably less than obtained by me , but the temperature and other conditions were different in the two cases .
Haber and van Oordt have attempted to find the dissociation constant at different temperatures , but it appears to me that the available data are entirely insufficient for the purpose .
Moreover , the part played by the iron is not yet completely understood .
My experiments show that the quantity of ammonia formed depends on the amount of moisture present , but Haber and van Oordt appear to have overlooked this point , and say simply that their gases were dry .
' Zeits .
fur anorg .
Chemie , ' vol. 43 , p. Ill , 1905 .
Prof. E. P. Perman .
( Mar. 4 , Synthesis by Heat in the Presence of other .\#151 ; Similar experiments were made with a number of other metals .
Copper , zinc , nickel , cobalt , palladium , aluminium , and magnesium , all gave traces of ammonia , but usually less than iron .
Platinum sponge yielded traces , whilst platinised asbestos and platinum foil produced very minute and scarcely perceptible quantities .
Zinc and copper in contact gave no more than when present separately .
In all these cases , which were made with " Mixed Gases I , " no attempt , was made to exclude moisture .
Effect of Large Surface.\#151 ; ( Mixed Gases I. ) The following substances were selected in order to test the influence ( if any ) of large surfaces , on the synthetical formation of ammonia : pipe stems , pumice , broken porcelain , asbestos .
A hard glass tube was packed with the substance , and the experiment conducted in the usual way .
As the result , traces of ammonia were found in each case , except with the porcelain .
All these substances , except the porcelain , contained a notable quantity of iron , and I believe that it is owing to its presence that the ammonia was formed .
The porcelain was from a broken globe , see former paper.* The pipe-stems altered in colour , under the influence of the gases , from a yellowish tint to a dull grey , which I ascribe to the reduction of the iron present .
Although my conclusion may be questioned , I believe that ( in this case ) the extent of surface has no effect , unless the substance with which the gases are in contact reacts chemically with them .
Synthesis by Explosion.-\#151 ; It was noticed that if an explosion of the mixed gases and air took place in the hard glass tube , traces of ammonia were formed , and the effect was further investigated by exploding the gases with oxygen in a eudiometer and testing the resulting gases for ammonia .
The following are the results:\#151 ; Mixed gases .
Oxygen .
Result .
c.e. c.c. 10-5 2 Trace of ammonia 15 5-2 ditto ( but less ) 16 3-5 ditto 16-3 3-4 ditto 16 9 ditto ( but less ) It will be seen that the quantity of ammonia formed diminishes if the oxygen is in excess .
The quantities were very small , but nevertheless considerably greater , in proportion to the volume of gases taken , than those * Loc .
cit. 1905 .
] The Direct Synthesis of Ammonia .
171 produced by the action of iron .
Similar effects have been noticed by other investigators.* Synthesis by Sparking.\#151 ; It is well known that ammonia can be synthesised in small quantities with the aid of high potential electric discharges , and I have now attempted to bring the gases into a state of equilibrium during sparking , i.e. , into such a state that the rate of decomposition is equal to the rate of formation of the ammonia , and to reach that condition from each direction .
Mixed Gases I. The experiments were carried out in a glass bulb of about 250 c.c. capacity , provided with two tubes and stop-cocks , and with platinum wires for sparking , the sparking distance being about 25 mm The bulb was placed in a thermostat and maintained at a temperature of 40 ' C. , it was tilled with the mixed gases , and the platinum wires connected with the terminals of an induction coil capable of giving an 8-inch spark .
After the sparking the gases were aspirated into dilute hydrochloric acid solution and nesslerised .
The various results are put together in the following tables:\#151 ; Pressure .
Time of sparking .
Ammonia formed .
Remarks .
mins .
milligrammes .
Atmospheric 5 0-02 moist gases 15 0-06 yy \#187 ; 15 0-02 very thin spark yy 5 0*021 gases dried yy 10 0-03J by H2S04 yy 15 006 yy yy 20 007 yy yy 80 010 yy yy 45 010 yy 2 atmospheres 60 019 yy yy 30 019 yy From these results it will be seen that:\#151 ; ( 1 ) Under atmospheric pressure equilibrium ( as before defined ) was reached when 0T milligramme ammonia had been formed .
( 2 ) Under a pressure of two atmospheres equilibrium was reached when 0T9 milligramme was present .
( 3 ) The amount of ammonia formed depends on the quantity of electricity passing , thus a very thin spark produced only one-third as much ammonia as a " fat " spark in the same time .
* See Watts ' Dictionary , 1st ed. , Art .
" Ammonia .
" N 2 Prof. E. P. Perman .
[ Mar. 4 , Decomposition of Ammonia by Sparking.\#151 ; Attempts were made to reach the same equilibrium points by starting from the opposite end of the reaction .
It was found to require long sparking before equilibrium was nearly reached .
The following are the results :\#151 ; Pressure .
Time of sparking .
Ammonia remaining .
hrs .
milligrammes .
Atmospheric , allowed to blow off every few minutes 2-5 0-13 Atmospheric , rising to two atmos .
at end 5 42 Two atmos .
, commencing with a mixture of N and II ( 1 : 3 ) , and 2 per cent. NH3 2 0-56 Two atmos .
, commencing with a mixture of N and II ( 1 : 3 ) , and 1 per cent. NIL 2-5 0-32 At atmospheric pressure decomposition was rapid , and the equilibrium point was very nearly reached , synthesis giving 0T0 milligramme and analysis 0T3 milligramme .
When the volume was kept constant , decomposition was very slow and the point of equilibrium was approached only by starting a long distance from the beginning of the decomposition .
Starting with a mixture containing 1 per cent , of ammonia , after 2S hours ' sparking , 0-32 milligramme ammonia remained instead of 019 milligramme by the synthetical method .
Mixed Gases II.\#151 ; In this series the length of the spark was 11 mm. , and the capacity of the globe 262 c.c. At the conclusion of the experiment the gases were pumped out through dilute acid .
The temperature was 39a8 C. In primary coil .
-------X_______ Time .
Current .
Voltage .
Ammonia .
mins .
15 amperes .
2 4 milligrammes .
0-08 30 2 4 0-30 15 1-5 4 0-25 30 5*5 4 0-44 45 5-5 .
4 0-37 22-5 5*5 4 0-41 15 2-75 2 0-07 68 Decomposition of ammonia by sparking .
5-5 4 041 1905 .
] The Direct Synthesis of Ammonia .
L7 3 From these results it is seen that\#151 ; ( 1 ) Equilibrium was reached in about 22 minutes , the gases then containing ( Ml milligramme ammonia .
On sparking ammonia at atmospheric pressure exactly the same point of equilibrium was reached .
( 2 ) More ammonia is formed than in the first series , owing to the shorter spark ; also the rate of decomposition and the rate of formation are quicker .
( 3 ) Other conditions remaining the same , the amount of ammonia formed is roughly proportional to the current strength .
( 4 ) If the current strength remains the same , the quantity of ammonia formed is much influenced by the voltage .
( It must be remembered that the electrical measurements here mentioned are those of the primary coil .
) Since writing the above , I have discovered that Berthelot carried out experiments of a somewhat similar nature .
He found that , starting either with ammonia or with nitrogen and hydrogen , sparking left finally a minute quantity of ammonia which was " of the same order " in the two cases.* Using the silent discharge , he obtained the same point of equilibrium , starting from either direction , viz. , a mixture containing 3 per cent , ammonia .
Reference should be made also to the work of Hemptinne , who has investigated the synthesis of ammonia under various conditions . !
Summary .
( 1 ) So far as can be shown by one of the most delicate tests known to chemists , ammonia cannot be synthesised by heat ( except under special conditions specified below ) .
The decomposition of ammonia by heat may , therefore , be regarded as an irreversible reaction .
( 2 ) Ammonia may be synthesised in small quantities from its constituent elements ( a ) by heating with many of the metals ; ( b ) by exploding with oxygen ; ( c ) by sparking .
These are reversible reactions .
( 3 ) It would appear that the synthesis of ammonia is effected only when the gases are ionised ; the ionisation would be brought about by sparking , or by the high temperature of an explosion of hydrogen and oxygen .
The immediate decomposition of the ammonia formed would be prevented by its sudden cooling .
The metals in the presence of moisture also produce " nascent " or ionised hydrogen .
( 4 ) It does not appear that nitrides of the metals form an intermediate stage in the formation of ammonia , for it was found that metals readily forming nitrides , e.g. , magnesium , did not produce more ammonia than the others .
* 4 Mecanique chimique , ' vol. 2 , p. 358 , 1879 .
t 4 Bull .
Acad. Roy .
Belgique , ' 1902 , p. 28 .
N 3 Prof. E. P. Perman and Mr. J. H. Davies .
[ Mar. 8 , ( 5 ) There is a close analogy between ozone and ammonia with regard to their synthesis and decomposition ; both are formed by sparking , and both are completely decomposed by heat .
In conclusion , I wish to express my thanks to Mr. G. A. S. Atkinson , B.Sc. , and to Mr. J. H. Davies , B.Sc. , for valuable assistance rendered during the earlier and later portions of the work respectively .
Determination of Vapour-pressure Air-bubbling .
By Edgar Philip Perman and John Hughes Davies .
( Communicated by Principal E. H. Griffiths , F.R.S. Received March 8 , \#151 ; Read March 30 , 1905 .
) It was shown recently by one of us* that the vapour-pressure of water can be determined with a considerable degree of accuracy by bubbling a current of air through water in a thermostat , and estimating the amount of water evaporated by absorbing it in strong sulphuric acid .
The accuracy of the method has since been questioned , f supersaturation being specially suggested as likely to cause error .
We have therefore made experiments in order to discover what error ( if any ) is introduced by supersaturating the air with moisture before it enters the water in the thermostat .
The effect of dust in the air and of electrification have also been investigated .
In each case the arrangement of the apparatus was as described in the previous paper .
Supersaturation.\#151 ; Before passing into the flasks in the thermostat , which was maintained at 70 ' , the air was bubbled through a large wash-bottle containing water at about 85 ' .
The wash-bottle was connected by a short rubber tube with the flasks at 70 ' .
Otherwise the experiment was conducted as already described .
The following results were obtained:\#151 ; w. P. T. y. P- V apour-pressure .
gramme .
mm. ' C. litres .
mm. mm. 0-6757 753-2 286-1 2-005 736-4 234-7 0 6706 749-3 288-1 2-005 730-2 234-8 The numbers obtained in the previous experiments were 234-2 , 233-2 , 234-5 , 235-0 , 233*5 , and 233-5 , while Regnault 's number ( corrected as described * ' Roy .
So .
Proc. , ' vol. 72 , p. 72 , 1903 .
t ' Journ. Phys. Cheni .
, ' vol. 8 , pp. 299 and 313 , 1904 .

