]\gt ; Some Physical Constants of Ammonia : a Study of the Effect of Change of and Pressure on an Easily Condensible Gas .
By EDGAR PHILIP PERMAN , D.Sc .
, Assistant Professor of Chemistry , and JOHN HUGHES DAVIES , B.Sc. , University College , Cardiff .
( Communicated by Principal E. H. Griffiths , F.R.S. Received Apri124 , \mdash ; Read May 17 , 1906 .
) This work was originally undertaken in order to obtain an accurate value for the density of ammonia , the only determination that appeared reliable that of Le Duc .
* While the experiments were in ress , the result of a determination by Guye was published ; the number differed from that of Le Duc and made a fresh determination still more desirable .
We ]lave now determined the density at temperatures ranging frum \mdash ; 20o C. , also the density at half an atmosphere pressure at C. From these data the coefficient of expansion and the compressibility be calculated .
In addition to this , measurements of the pressure , of vapour-pressure , and of the density of ammonia in the presence of air , have been made .
Vapour-den.sity of Ammonia .
Two methods were employed:\mdash ; ( 1 ) The ammonia was pumped from the lobe in which it was contained into concentrated sulphuric acid contained in -tubes , and so weighed .
The globe remained in position during each iment .
( 2 ) In a few experiments the ammonia was weighed directly .
Apparatus , Method ( l).\mdash ; The apparat , us may be divided into three parts : ( a ) the ammonia supply , ( b ) the globe and bath , ( c ) the absorption apparatus : The ammonia was contained in an iron tube cm .
long and cm .
in diameter , closed by a valve , obtained from the Scotch and Irish Company .
The outlet tube was connected by means of lead pipe ( about 1 foot ) and junctions of fusible alloy to the drying tubes and containing limy and recently fused sodium hydroxide respectively .
The tube then divided into two branches , one leading to the escape tube , about 900 mm. long , dipping under mercury , and the other to the globe G. Any 'Comptes Rendus , ' 1897 , vol. 125 , p. 571 .
'Comptes Rendus , ' ) , vol. 141 , p. 61 .
Some Physical Constants of ammonia bubbling through the mercury was absorbed by dilute sulphuric acid contained ) the wash-bottles F. ( b ) The globe of about 18litres capacity was placed in ( 1 ) a thermostat containing an Ostwald toluol regulator of special pattern , in an ice bath , ( 3 ) in a bath of alcohol and solid carbon dioxide , according to the temperature required .
was ordinary stop-cock , and a three-way tap leading to two sets of bulbs , containing strong uric acid .
( Only one set is shown in the diagram .
) The temperatures above were measured accurately to by means of a thermometer made by Fuess , of Berlin , and standardised at the Reichsanstalt .
For temperatures below a mercury thermometer was used which had been previously standardised by comparison with an air thermometer .
to the difficulty of maintaining the oonstancy of these low tenlperatures , we cannot claim that they were accurate to less than ( c ) The absorption apparatus consisted of three -tubes , , connected ether by thick rubber and were stoppered .
was connected with a tube from the globe by the round joint the -tube being eld in place by a rubber band .
The tube was connected with the gauge and a water air-pump not shown in the diagram .
and were stop-cocks .
Method of Work .
Filling the Globe .
Every n position , the globe and connecting tubes were exhausted to mm. by means of a Topler pump .
The valve was then opened until ammonia bubbled through the mercury in , the taps and were opened , and left open until the ammonia began to escape , when was immediately closed .
The main part of the ammonia was then } ) umped out by the water airpump through , and the remainder by the mercury pump through R. It should be mentioned that the drying tube of the pump contained strong sulphuric acid , which immediately absorbed the ammonia thus admitted .
The globe was again filled as before .
The temperature of the bath was carefully regulated , and the excess of ammonia allowed to escape into the air , the stop-cock being turned on from time to time .
Owing to the cooling caused by the expansion , some little time was required for the gas to acquire the exact temperature and pressure .
From half an hour to an hour sufficed .
order to indicate when the pressure was equal to that of the 'Roy .
Soc. Proc 1903 , vol. 72 , p. 74 .
Dr. E. P. Perman and .
Mr. J. H. Davies .
[ Apr. 24 , atmosphere , a small mercury gauge was connected with ; it showed finally no on turning R. In some of the earlier experiments the water air-pump was used in the filling ; in that case the globe was exhausted to 1/ 10 atmosphere and filled with ammonia , the process being carried out six times .
Absorption of the \mdash ; The tubes , includin , were freed from ammonia and moisture by repeated exhaustion and admission of air L. The -tubes containing sulphuric acid and glass wool soaked with it ( see fig. I ) were then connected up , and having been weighed .
was a guard tube to prevent the access of moisture from the pump .
connections were tested by exhausting .
the -tubes and noting any change in the on standing .
Leaks having been shown to be absent , air was admitted through , and the -tubes were connected with the globe through and R. The apparatus was then exhausted very slowly to about 1/ 10 atmosphere , very nearly all the ammonia being absorbed by the sulphuric acid in ; air was admitted through , and the apparatus again exhausted .
This was done { ive times in all .
Next , air was admitted through , and in order to make the pressure inside the tubes exactly that of the atmos.phere , connection was made through a calcium chloride tube fixed to and were closed and then removed and carefully .
The barometer was read immediately after the ammonia 1906 .
] Some Constants of in the globe had ceased to blow off , the readings being made accurately to 0.1 mm. Volume of the Globe.\mdash ; The globe was full of dry air at known temperature and pressure , and again full of vater at known temperature .
From these weighings the capacity of was calculated .
Correction was made for the displacement of air by the weights , and the used in this part of the work were compared with those used in the ammonia ( both sets ( been compared with a ) .
The coefficient of expansion of was found by experiments with a weight thermometer made from a broken globe .
Between and the coefficient was Jfethod ( 2).\mdash ; The ammonia was directly in a of about litre capacity .
The hings were made on a short beam Oertling balance , and were carried to ramme .
The were standardised with the same accuracy .
In this case the globe was attached to the rest of the apparatus by means of a ground glass joint above the stop-cock ; it was exhausted by a Fleuss pump to mm. , the read on a mercury gauge by means a cathetometer .
The empty globe was bed , then filled with onia and detached .
The excess of mmonia was allowed to escape , and the globe was again weighed , the usual precautions taken throughout .
A globe was used as a counterpoisc .
The volume of the globe was found by it , completely exhausted ; ( b ) full of dry sir at known temperature and pressure ; ( c ) full of water at known tenlperature .
These measurements gave two closely concordant values for the capacity of the globe .
The contraction of the globe on exhaustion was measured as follows:\mdash ; The globe was nearly filled with water , about 5 .
of air in the upper part .
This air was then out and the stop-cock closed .
The was next suspended in water and to ramme .
It was opened to the air and , when it was found to lost ramme .
The weight of air admitted was igible .
Testing for Possible Errors .
Adsorption of Ammonia by has been usually supposed that a considerable quantity of ammonia is orbed ' by glass , the adsorption ressing for some time .
If so adsorbed , it be again evolved on exhausting the globe , or it might be held permanently by the lass .
An experiment was made in which the bolobe was filled with nmmonia ) the atmospheric pressure , and allowed to stand 12 hours ; the of Dr. E. P. Perman and Mr. J. H. Davies .
[ Apr. 24 , ammonia was then allowed to escape , and the usual number for the density was obtained .
It is thus proved that if ammonia is adsorbed it is not given off on exhausting the globe .
As a further test a globe was filled with ammonia , exhausted to and allowed to stand .
The pressure was then read at intervals for some days by means of a cathetometer , but it remained constant .
Further , in determining the pressure coefficient , the same conclusion was arrived at , the pressure of the ammonia in an ammonia thermorneter was not found to alter after the lapse of a month .
From these experiments we are forced to the conclusion that , with carefully dried ammonia and glass , adsorption ( if it takes place at all ) is inappreciable .
Condensation on the Surface of the Glass.\mdash ; It has been thought that ammonia will condense in the same way as water vapour on the surface of glass , and so would cause an error in density determinations in glass globes .
If this were the case , different densities should be obtained if }obes of different sizes were employed .
We have found , however , that globes of litre and litres capacity respectively gave the same result .
The surfaces of these are as 1 : , whilst the volumes are as 1 : The only conclusion to be is that when the ammonia and the glass are carefully dried , tihere is no appreciable condensation of ammonia .
Loss of Am nonia ; Incomplete Absorption.\mdash ; The ammonia pumped out of the globe passed into two -tubes concentrated sulphuric acid ( including four plugs of ( flass wool soaked with acid ) , but as in all the exhaustions , except the first , the ammonia was mixed with a large proportion of air , it was thought possible that traces of ammonia might nevertheless escape absorption .
This was tested by allowing the mixture of air and amtnonia at the second and the fourth exhaustions to pass through two -tubes into an exhausted globe .
The contents of the globe were then tesCed for ammonia by Nessler 's solution ; but not a trace was found .
Moisture from the Air.\mdash ; The air made to expel the ammonia was first passed through two ' potash bulbs\ldquo ; containing strong sulphuric acid , and the bulbs were recharged every day .
On 5litres of air the weight of the -tubes remained unaltered , thus showing that none of the moisture escaped absorption .
Evaporation of Sulphuric Acid.\mdash ; In order to test whether any sulphuric acid was lost by evaporation , about 8litres of dry air were passed the first -tube , the sulphuric acid contained in it being heated to about the same temperature as it acquired during the absorption process .
The weight constant to milligramme .
1906 .
] Some Physical Constants of Dispfacemont of Air owing to Change in Volume of Sulphuric Acid caused by of Ammonia .
After the completion of an experiment , the density of the contents of the -tube was determined , also that of the original sulphuric acid : grammes .
20- .
specific gravity bottle bottle after expt . . . .
filling bottle .
bottlefresh Fresh filling bottle If no change in volume of acid had taken place , weight of mixture filling bottle would be grammes grammesweight of absorbed ) , Change in volume produced by absorption of .
approximately .
The correction for air displaced is therefore gramme .
Preparcction of jbonia .
Ammonia from various sources was employed : ( 1 ) Commercial anhydrous ammonia from an iron der .
( 2 ) Commercial ammonia was passed through a tube containing red-hot lime into pure hydrochloric acid ; the ammonium chloride formed was heated with solution of sodium hydroxide , and the ammonia evolved was thoroughly dried by over quick lime and sodium vdroxide .
This method was used by Guye .
* In our early ents t was condensed in a glass tube placed in a mixture of ether and solid carbon dioxide ; later an iron tube was used .
In order to prove the efficacy of this method of preparation in destl'oying pyridine , a slow stream of hydl.ogen was bubbled liquid contained in a small wash-bottle , and then passed over -hot lime .
No pyridine could be detected in the gases the hard-glass tube .
A sample of ammonium chloride made by the above-mentioned method was very kindly subjected to spectroscopic examination by Dr. J. J. Dobbie , who reported that he found no indication of the presence of pyridine .
( 3 ) Dobbie has shown that ammonium oxalate freed from pyridine by repeated recrystallisation .
Some ammonium oxalate was made 'Chem .
Soc. Journ , Trams .
, vol. 77 , p. 318 .
VOL. LXXYUI .
Dr. E. P. Perman and Mr. J. H. Davies .
[ Apr. 24 , from commercial oxalic acid and ammonia and recrystallised 10 times ( another sample six times , which proved sufficient ) ; the ammonia was set free by heating with potash , and condensed in a glass tube .
( 4 ) Sodium nitrite was reduced with aluminium and sodium hydroxide solution ; the ammonia was absorbed by hydrochloric acid , and again liberated by sodium hydroxide .
The main part of the ammonia used was made by method ( 2 ) , and was condensed in the iron tube ( see diagram ) placed in ether and solid carbon dioxide .
The tube was fitted up like a wash-bottle , the ammonia passing in through the long tube .
To the short tube was connected a drying-tube to prevent the entrance of moisture from the air .
When a sufficient quantity of ammonia had condensed , the tube was removed from the freezing mixture , placed in a vice , and the valve immediately screwed on , a leather washer being placed between the tube and the valve .
The valve was fully opened , and left for about 5 minutes in order to expel the air from the top of the tube by the rapid stream of ammonia .
After some preliminary experiments the following results were obtained:\mdash ; Method ( 1 ) .
Commercial Ammonia.\mdash ; Temperature .
Capacity of globe 17738litre .
Temperature Capacity of globe 17738litre .
*Left for 12 hours to test adsorption .
1906 : ] Some Physical Constants of Method ( 2 ) .
Capacity of globe .
Temperature The results now available for the vapour-density of ammonia at GuyeLeDuc .
Reduced t Perman and Davies ( 1 ) We may venture to say that Le Duc 's number is erroneous , owing to the use of unpurified ammonia .
Our final result is ( giving the greater weight to series ( 1 ) ) , which differs only by about one part in 15,000 from that obtained by Guye , and may be taken , we believe , as the most accurate value yet obtained for this constant .
Density of Ammonia at \mdash ; Capacity of globe litres .
Mean finaMe } Demity of at \mdash ; Capacity of globe ( new ) ] litres .
Dr. E. P. Perman and Mr. J. H. DavieS .
; Density of Ammonia at \mdash ; The temperature was obtained by shaking solid carbon dioxide into a bath of alcohol surrounded by thick felt and thoroughly stirred .
temperature was read on a mercury thermometer and was maintained constant within by adding solid carbon dioxide in small quantities as required .
The thermometer was standardised by means of an air thermometer .
Capacity of globe Coefficient of Exparosion of Ammonia.\mdash ; From the results just given the specific volumes Temperature .
Volume of 1 gramme .
119575litres 1.2973 From these numbers the coefficients of expansion for different ranges of temperature have been calculated:\mdash ; Temperature .
Between and\mdash ; 20o Coefficient of expansion .
The coefficient is seen to be much greater than that of the less easily condensible ases , and decreases with rise of tenlperature .
Attempts were made to calculate values of and in van der Waals ' equation from these data ; it was found , however , to be impossible , the equation not representing the facts with sufficient accuracy .
Taking the simple equation is found to vary about 2 per cent. between and 190.6 .
] Some of monia .
Temperature .
is here expressed in atmospheres and in litres .
Compressibility of Ammonia.\mdash ; By comparing the density at half an atmosphere with that at one atmosphere , the compressibility of ammonia may be determined .
The globe was filled with ammonia in the usual way , then connected with a pressure-gauge .
and about half the ammonia pumped out .
Next the ammonia was allowed to blow off until equilibrium had been attained , when the pressure was immediately read .
The readings were made on a mirror scale , the barometer standing in the same trough being read at the same time ( see fig. II ) .
to the difficulty of reading the pressure and maintaining its constancy , no great accuracy was attained ; the probable error is about mm. Dr. E. P. Perman and Mr. J. H. Davies .
[ Apr. 24 , Density of Ammonia at Half an \mdash ; Capacity of globe litres .
Temperature From these results As before stated , we cannot claim any great accuracy for this result .
and we prefer to make use of Lord Bayleigh 's number in proceeding to calculate the molecular weight of ammonia , the method and apparatus used by him specially adapted to the purpose .
* Lord Rayleigh found at , which , corrected in the way indicated by him , becomes at This gives , and the correction factor Molecular weight of ammonia Taking , the atomic weight of nitrogen is deduced as thus closely .
confirming the number obtained recently by several investigators .
Density of Ammonia in the Presence of Air.\mdash ; Experiments were made to test the effect of admixture with air on the density of ammonia .
Determinations were first made of the density of air freed from moisture and carbon dioxide by means of phosphorus pentoxide and soda-lime respectively .
Temperature .
Volume of globe Zeits .
fur Phys. Chem 1905 , vol. 52 , p. 705 .
1906 .
] Some Constants of Ammonia .
The globe was exhausted and weighed , about half filled with ammonia , and again .
It was then surrounded with ice , and air carefully admitted un til the mixture had attained the atmospheric pressure .
Another weighing was then made .
The following are the data : Temperature .
Volume of globe mmonia a The density calculated from 's value for the compressibility is .
The partial pressure of the ammonia in the above table is tained by subtracting the partial pressure of the air ( calculated from its weight ) from the total pressure .
It will be seen that the density is nearly one part in 1000 higher than that calculated from ths compressibility , and this represents the deviation Dalton 's law of a mixture of approximately equal volumes of ammonia and air .
The normal value of the density of ammonia , calculated from its molecular weight and the density of oxygen , is ; consequently the density of ammonia is still about 1 per cent. above the normal value , even when diluted to the extent mentioned .
Pressure-coeffici of simple form of constant volume airthermometer , with a globe of about S litre capacity , was filled with pure ammonia at about C. , and atmospheric pressure .
In order to test if there were any inaccuracy caused by adsorption , successive readings were taken at an interval of 48 hours , of 24 hours at , and of a month at but the readings in each case agreed within the limits of expel'imental error .
Allowance was made ior the expansion of the globe with rise of temperature .
The correction for the tube connecting the globe and the gaugc was inappreciable .
The following are the results ( see p. 40 ) .
It will be seen at once that the numbers show a considerable deviation from van der Waals ' equation , which may be written in the form where and are constants : is proportional to the pressure-coefficient and varies , therefore , by nearly 6 per cent. over the range of temperature employed .
The values of the pressure-coefficient are , as ] be expected , very close to those obtained for sulphur dioxide , and than those for carbon dioxide or-nitrous .
oxide .
: E. P. Perman and .
J. H. Davies .
[ Apr. 24 , Summary .
( 1 ) The vapour density of ammonia at has been found to be mass of llitre in grammes at latitude ) , previous results being by Guye and by Le Duc .
( 2 ) When the ammonia and the ylass vessel were thoroughly dried no preciable adsorption of ammonia by glass , or condensation of ammonia on the surface of glass , was found to take place .
( 3 ) From density determinations at different temperatures , the coefficient of expansion of ammonia has been deduced as between and , and between .
and ( 4 ) From Rayleigh 's determination of the compressibility of ammonia and our own value for the density , the molecular weight of ammonia has been calculated as , and the atomic weight of nitrogen as ( 5 ) Incidentally the density of air free from water vapour and carbon { lioxide has been determined as ( lat. ) .
( 6 ) The deviation from Dalton 's law for a mixture of approximately equal volumes of air and ammonia has been found to be about 1 part in ( 7 ) The pressure-coefficient of ammonia has been determined , the pressure being atmospheric at .
Between and the coefficient was , and between and it was The expenses incurred in the above research have been defrayed by a grant from the Royal Society .
1906 .
] Constants Pressurc of Liquid Ammoni and the of its Boiling Point .
By JOHN HUGHES DAVIES .
It was suggested to me by Dr. Pcrman , to whom I am much indebted for kindly advice nd assistance , that I should repeat the determination of the vapour pressure of liquid ammonia at some of the lower temperatures , using pure ammonia , in order to obtain accurate value for its point .
Apparatus .
This consisted 1 .
The bath , which was a cylindrical zinc pot about 10 inches high and 6 inches diameter , covered completely on the outside with a of very thick felt .
It contained alcohol and solid carbon dioxide , and was provided with a stirrer worked by a hot-air motor .
2 .
The iron tube , which contained the liquid ammonia .
This differed from the tube previously described only in having a narrow stem .
3 .
The pressure gauge .
Two forms of gauge were used , one for the lower temperatures , i.e. , temperatures at which the pressure was less than atmospheric , and the other for the crher temperatures and pressures .
The former is indicated in the by and the latter by .
The pressure was read on a millimetre scale .
is a movable glass mirror of rectangular shape with a horizontal line drawn across it ; by sliding it along the edge of the scale , and behind the glass tubes of the gauge , the pressure could easily be read to 1/ 10 mm. 4 , The thermometer .
A pentane thermometer was employed ; it was gladuated in single degrees , so that the temperature could be read accurately to .
It was standardised by means of an air thermometer .
Method of Procedure.\mdash ; The apparatus being in position with the suitable pressure gauge attached and the steel valve shut off , the leading tubes were exhausted to 1/ 10 mm. by means of a Fleuss pump , exhaustion taking place through th stop-cock in the case of the and through the side tube in the case of F. The stop-cock was then turned off , or the side tube drawn out and fused , as the case might be .
The bath was brought down to the required temperature by the addition of solid carbon dioxide , the valve turned on , the pressure indicated by the gauge read , and the temperature of the mercury in the gauge taken .
Results.\mdash ; Determinations were carried out over the range of temperature to Some of From these results the values of the vapour pressures at equal intervals of from to have been obtained by graphical interpolation .
Those pressures marked have been obtained from the curve , whilst the others were actually observed .
From this curve the boiling point of liquid ammonia at 760 mm. pressure is given as C. The results obtained are found to be in very good agreement with those obtained by Regnault*and Pictet , whilst they differ considerably from those obtained by Faraday and Blumcke .
S Also the value obtained for the boiling point , viz. , , is in close agreement with the value obtained by H. D. Gibbs Vide ' Landolt und Bornstein Tabellen , ' 1897 edition , pp. 77 , 78 .
Vide xbid .
'Phil .
Trans 1845 , p. 166 .
S 'Wied .
Ann vol. 34 , p. 10 , 1888 .
'Amer .
Chem. Soc. Jour vol. 27 , p. , 1906 .

