]\gt ; Mr. E. P. Perman .
[ Feb. 4 , " " The Evaporation of Water in a Current of Air By EDGAR PHILIP PERMAN , Assistant Lecturer in the University College of South Wales and Monmouthshire .
Communicated by Professor E. H. GRIFFITHS , F.R.S. Received February 4 , \mdash ; Read February 19 , 1903 .
Introdnction and Hisforical .
It may be thought that so simple a matter as the evaporation of water in a current of air has already been thoroughly investigated , but , so far as I been able to discover , this is not the case , although much concerning it has been taken for granted .
This research was undertaken in order to ascertain with what accuracy the vapour pressure of water could be calculated from the amount of water vapour carried off by an air current passed through the water , the temperature being maintained constant .
The work was begun in the Physikalisch-chemisches Institut , Leipzig , during the summer of 1902 , and completed at University College , Cardiff , during the same year .
Experiments of a similar nature have been made by Regnaultand by W. N. Shaw .
Regnault found that the weight of the vapour drawn off agreed within about 1 per cent. with that calculated from the apour pressure .
The temperatures employed extended up to C. Shaw 's experiments were at ordinary temperatures of the air , and a very close agreement was found between the usually accepted vapour pressures those calculated from the tmount of water vapour drawn off .
of The apparatus employed consisted of the following chief parts:\mdash ; ( 1 ) Thermostat , with thermometer and stirring arrangement .
( 2 ) Four wash-bottles to contain the water .
( 3 ) Absorption apparatus to take up the water-vapour as the moist air passed through it .
Aspirator .
( 1 ) The thermostat ( T ) , consisted of an iron pot 23 cm .
in diameter and 21 cm .
high , surrounded by thick felt , and heated by an Ostwald burner ( O ) , fig. 1 .
The thermometer was made by Fuess , of Berlin , and was standardised at the Reichsanstalt , Charlottenburg , number 17,000 .
It was divided into tenths of 1o C. , and was read to ; the final corrections vere also given to At the temperatures the stem was allowed to proiect from tlJe bath iust enough to make the thermometer readable , but at * ' Mem. de ' vol. 26 , p. 679 .
' Phil. Trans , vol. 179 , 1888 , Appendix No. III , 1903 .
] Evaporation of in a of Air .
Mr. E. P. Perman .
[ Feb. 4 , 1rrounded by a wide tube , open ork at the upper end .
The tube of cotton wool .
The arrangequid could be drawn up from the a small tube passing through the cork ; the tube was then closed by means of a clip on a rubber tube connected with the glass tube .
The temperature of the liquid in the lvide tube was kept uniform by tion currents .
This was shown to be the case by emptying the liquid into the bath , and quickly drawing up a fresh portion , when the reading of the thermometer remained unaltered .
In some of the experiments the thermometer was placed in a sloping position ( in order to get the whole column into the bath ) ; in that case a correction for internal pressure was made , being calculated from the coefficient given in the certificate .
A toluol gas regulator was employed ( not shown in fig. 1 ) , which usually mainGaiIIed the temperature constant to ; it rarely varied more , than this during an experiment , and I do not believe that the temperature error exceeds in any single experiment .
The regulator is of a new pattern , although it does not involve any new principle .
The gas can be regulated in either direction by the stop-cocks cury enclosed can be adiusted to arlge by means of the bulb , and os of the screw , the temperature be easily brought to any desired , oluol regulator , seo 'Physiko-chemische 1903 .
] The Evaporation of Wctter in a of : The stirring arrangement , fig. 1 , consists of a vertical glass tube , terminating in a horizontal -piece open at both ends .
In the vertical tube is an opening a little below the surface of the liquid .
The -tube was rotated abont a vertical axis by means of a hot-air engine .
The circulation of the ater in the thermostat was found to be very effective when the -piece made three revoIutions a second .
( 2 ) The wash-bottles , , fig. 1 , were each of aboub 100 and were all sealed together ( except in some of the earlier periments ) in order to prevent leaks and the access of grease or other foreigll matter .
The last flask which the air passed was connected a pressure-gauge ( read by a mirror scale ) by a tube a stopcock jnst above the bath .
( 3 ) The absorption apparatns , , fig. hich was colmectedhvith the tube from by a gromld-glass joint , consisted of two small wash-flasks containing con centrated sulphuric acid ; about 10 vere placed in the first flask and 2 c.c. in the second .
In the earlier experiments the two flasks connected by rubber tube , and weighed , but it lvas found that the second never gained mors than 1 .
during an experiment , and they were therefore sealed and eighed together .
As a fnrther proof of the completeness of the sorption , in two experiments at a tube containing phosphoric anhydride connected up with the second bulb , but it did not iucrease in weight by so much as ( 4 ) The aspirator consisted of a large bottle , , fig. 1 , placed in a large pan of water .
For the experiments at the lower temperatnres the bottle had a capacity of 13 litres , but for some of ths higher temperature experiments a 4-litre bottle employed .
The quantity of water outside was so large that the temperature did not change more than during an experiment , and hence a regulator considered unnecessary .
Vater drawn off through the tube into graduated flasks .
The pressure of the air in the bottIe read by means of a gauge , the joint and the bath , the projecting tube was surrounded by sheet brass and heated by a row of small gas jets ( in the earlier experiments by a coiled lvire , through which an electric current was passed ) , in order to prevent condensation of vapour in this part of the tube .
The three-way stop-cock was the dividing line between the flasks containing the water and the absorption apparatus ; by means of it connection could be made veen two -tubes ( not in the figure ) , containing qtrong sulphuric acid , and the absorption so that the tube between and might be cleared of moisture by a current of air from the outside , passing first the -tubes .
Between the absorption apparatus and the aspirator were a tube , Ca , and two small wash-bulbs containing water , the former to prevent water vapour from getting back to the absorption .
E. P. Perman .
[ Feb. 4 , apparatus , and the latter to saturate the air passing into the aspirator .
The weights employed were compared a 1 .
platinum weight , which had been standardised at the National Physical Laboratory .
The balances employed were tested for inequality of length of arm , but in each case this was found to be negligible .
Method of Work .
An experiment was carried out as follows:\mdash ; The four wash-bottles were thoroughly cleaned by placing them in an inverled position and blowing steam through them for half an hour .
This , or some similar process , is important , for if there is any grease present , erratic resuIts are obtained .
The flasks were then about two-thirds filled with distilled water , and placed in the thermostat .
The sulphuric acid was placed in the absorption apparatus , which was placed in position , and all necessary connections made .
A slow stream of air was aspirated through the whole apparatus , and the two pressure gauges were read .
The absorption apparatus was weighed aftsr standing in the balance case 15 or 20 minutes .
Meanwhile the tube CJ was cleared of moisture .
The absorption apparatus was then connected up again , the rubber junctions coated with wax , the pressure in the aspirator adjusted to that previously found , and the apparatus was left for some minutes to discover any possible leak .
The absence of any leak being proved , the clip and the stop-cock were opened and the experiment begun .
To bring the experiment to an end , the clip was first closed , and then the stop-oock C. The pressure gauge was read again ( it usually remained quite unaltered ) , and the moisture was driven from the tube CJ into the absorption apparatus ( by the method already described ) , which was then removed to be weighed .
The temperature of the aspirator and the barometer were read during the experiment .
At the lower temperatures the air was never passed through at a greater rate than litre per minute , while at the higher temperatures the rate was not more than per minute .
The air acquires the temperature of the bath and becomes saturated with aqueous vapour with surprising rapidity .
Special experiments were made to test this point .
of the Air .
If air is passed through one wash-bottle only , containing water , and placed in a thermostat at a temperature higher than that of the air , rapid fall of temperature takes place owing to the low temperature of the air and to the cooling by evaporation of the water .
Experiments were made at C. and at C. ; a thermometer which could be 1903 .
] Evaporation of Water in a Cur.rent of read to was placed in the wash-bottle through a cork in the neck , and air was drawn through at varying rates .
In most of the experiments the air passed first of all through another wash-bottle containing water , and through the flask containing the thermo- meter .
The results are here tabulated:\mdash ; Temperatureof Fall of thermostat .
bottle .
Air rrent .
ature .
C. 1 .
per mins .
, , 76 ] 0.08 \ldquo ; \ldquo ; 76 2 , , Slow fall .
76 2 , , Rapid fall .
It will be noted at , with only two flasks , no perceptible of temperature is produced by such a rapid current of air as 3 litres per minute , while at , with the same nnmber of flasks , a current of litrs per minute has no perceptible effect .
Since the rates used in the actual experiments for the determination of vapour pressure were much less than litre per minute , and four flasks ( in some cases five ) tvere employed , there can be no doubt as to the saturation of the air in these experiments .
This was confirmed also by varying the rate during the experiments , when the variation was not found to have any effect on the amount of water driven off .
Method of Calcnlation .
An experiment carried out in the manner described furnishes the following data:\mdash ; Weight of water drawn off .
Pressure in the last tion flask .
Absolute temperature of air in aspirator .
Volume of air through the water , at temperature and pressure of ator .
Pressure of air in aspirator .
If we assume that the vapour pressure of water is the same in the presence of air as in the absence of air or any other gas , and the partial pressure of each constituent of the gaseous mixtnre is the same as if each occupied the whole space ( Dalton 's laws ) , and the of saturated water vapour is normal , then the vapour pressure can be calculated from the relation 78 Mr. E. P. Perman .
[ Feb. 4 , of aqueous vapour Volume of aqueous vapour Total pressure Total volume The " " normal\ldquo ; value of the density of aqueous vapour is taken to be that caloulated from the density of oxygen and the molecular weight of lvater , and the corresponding specific volume , which is more convenient for purposes of calculation , is .
The vapour pressure is calculated then from the expression\mdash ; W. P. Iiesults .
The lowest temperature which the method was applied was C. The experiments are naturally very tsdious at low temperatures owing to the large volums of air which must be passed through the apparatns .
The whole of the experimental results are here tabulated , the actual data being given in an Appendix .
It will be noted that the temperatures are at exactly every ( with two exceptions ) ; the corrections were calcuIated beforehand , the thermostat carefully adjusted .
One exception is , which was the temperature in the only series of experiments made with a portion of the thermometer stem outside the bath ; a correction was made for this exposed portion in the usual way .
On calculating out t , he results , the vapour pressure appeared to be abnormally high , and consequently a large number of experiments were made to ascertain this point with certainty .
However , on proceeding to , normal vapour pressures were again obtained , and suspicions were aroused as to the exactness of the temperature reading at .
The thermostat was brought again to , and the thermometer placed first with its stem surrounded by the wide tube and hot liquid , and then placed in the original position , when it was found that the correction had been overestimated by The reason for this large difference between the calculated and the real correction is probably that in this class of thermometer the column of mercury projecting from the thermostat is surrounded by a bath of hot air enclosed by the outer tube of the thermometer .
The other exception , at WttS one of the earliest experiments carried out .
The vapour pressures given in the last column are taken from Ostwald and Luther 's 'Physiko-chemische Messungen ' ( 1902 ) , .
The numbers were obtained by plotting the differences between the results of Wiebe and Thiesen and Scheel , and the RegnauIt-Broch numbers , interpolating the curve from to , reading off the corrections from the smoothed curve , and applying them to the RegnauIt-Broch numbers throughout .
weight of a litre of oxygen has been taken as 1903 .
] The Evaporation of Water Vapour- Temperature .
pressure .
32.05 31.79 40 .
57.23 92.33 92.38 92.50 148.5 149.2 149.1 235.0 233.5 358.4 355.0 356 .
1 525.6 525.7 92 .
92.35 149.3 ( RegnauIt uncorr .
) 356.9 526 .
Mr. E. P. Perman .
[ Feb. 4 , It must be here remarked that the experiments at the lower temperatures are by no means so reliable as those at the higher temperatures , for the following reasons:\mdash ; They were the first carried o1lt , before experience in working the experiment had been gained .
2 ) The amount of water carried over was always very small , so that it is impossible for these experiments to be so accurate as those in which the water is carried off mors rapidly .
( 3 ) A rubber connection used at ( fig. 1 ) , whereas in the latsr experiments this was replaced by a ground-glass joint .
Also , the absorption flasks were weighed with glass stoppers connected by rubber tubes , and it was found that moisture from the air its way slowly through the rubber , although the glass tubes were made The absorption apparatus.was in this way found to take up sometimes as much as .
in an hour , which is quite sufficient to account for the highness of the numbers obtained in these experiments .
This may be , perhaps , included under reason ( 1 ) .
At the higher temperatures there was no rubber connection th ough which moisture could pass to the absorption apparatus , and the apparatus was closed with small rubber stoppers inserted into the tubes .
lVith this arrangement the weight was found to be constant .
Deductions from esults .
The vapour pressures obtained are seen to agree very closely with obtained by direct measurement .
Now , in calculation , certain assumptions have been made , viz. , Dalton 's law of partial pressures , and that the density of aqueous vapour is normal .
It would thus seem that these assumptions are justified .
It is highly improbable that deviations from Dalton 's law , and from the normal density , should so balance one another at all these temperatures as to give correct vapour pressures .
Fortunately there is independent evidence of the validity of Dalton 's law under the conditions of the experiments .
Galitzin by some well-devised experiments , showed that there was no appreciable deviation from Dalton 's law for air and aqueous vapour at moderate pressures below C. It may be concluded , then , that the density of saturated water vapour at the pressures and temperatures employed is very near the normal value .
This conclusion seems to the uthor quite justified by the experiments described , although it is at variance with the work of a number of other investigators , who have found the density to be greater than normal , usually from 1 to 2 per cent. , sometimes much more .
* Wied .
Aun 1890 , vol. 41 , p. 588 1903 .
] The Evaporation of Watcr in a Current of Air .
81 A list of papers on the subject is here given for reference : \mdash ; RegnauIt ' Ann. Chim .
Phys. ' ( 3 ) , 15 , 129 , 1845 ) .
airbairn and Tate ( ' Phil. Trans vol. 150 , p. 185 , 1860 ) .
Wiillner and Grotrian ( ' Wied .
Ann vol. 11 , p. 544 , 1880 ) .
Perot Rendus , .
Batelli ( ' Mem. dell ' Accad .
di Torino , ' ( 2 ) , vol. 41 , p. 33 , 1891 , and ( 2 ) , vol. 43 , p. 1 , 1892 ) .
Ramsay and Young Phil. Trans .
Bauer ( ' Wied .
Ann vol. 55 , p. 184 , 1895 ) .
Griffiths has calculated the density of saturated water-vapour from latent heat of evaporation by means of the thermo-dynamical equation The values of were taken from Broch 's reduction of Regnault 's experimental resuIts ; the author has recalculated the densities , replacing these values of by numbers culated from the vapour pressures iven by Luther .
Both series are given in the following table:\mdash ; This method is a very indirect one , but is fres the Qrrors involved in most of the experimental methods .
Since and have determined with great accuracy , the densities culated from them should be accurate at the higher temperatures where is known with fair accuracy .
The mean density so ) tained is very close to the normal density , I believe the greater number of the experiments quoted are rendered inaccurate by tlJe condensation of vapour on the sides of the glass vessel employed , owing to the hygroscopic nature of lass ; this is impossible to avoid , but in my experiments it will have * .
Trans , vol. 186 , 1895 , p. 325 .
'Pbysiko-chemische Messungen , ' -Luther , p. 166 .
Mr. E. P. Perman .
[ Feb. 4 , no effect on the result , and I believe it can be safely concluded from these experiments that the density of saturated water vapour under the conditions of the experiment ( that is , when mixed with half its .
volume to forty times its volume of air , at temperatures from to is normal .
Also , it is highly improbable that the density of the saturated vapour alone is more than slightly above normal at these temperatures , as the deviations from Dalton 's law have been shown to be very small .
This is also confirmed by the resuIts of Griffiths ( vide supra ) .
Conclusion .
( 1 ) When air is aspirated through water , it becomes saturated with aqueous vapour with great rapidity .
( 2 ) In the saturated air so obtained , the pressure of the aqueous vapour is the same as the vapour pressure of water when no other gas is present .
( 3 ) The density of the aqueous vapour in the mixture is ( 4 ) The density of saturated aqueous vapour ( without admixtnre ) is probably only very slightly ( if at all ) above normal at temperatures up to 1903 .
] The ) tion of Water in of Air .
APPENDIX .
aperimental Data .
T. 291.0 291.0 291.3 292.0 290 .
5 290.8 294.3 294.3 95.0 294.2 291.0 285.5 287.8 284.8 284.9 286.7 284.9

