182 Prof. W. Ramsay and Miss E. Aston .
[ June 14 , it is quite unnecessary to take measurements at temperatures very close to the critical temperature , because , if that temperature is known , there can he only one curve joining the points experimentally found at somewhat lower temperatures and the critical temperature .
The form of the curve is such that no doubt can exist as to its course .
Indeed , with chlorobenzene , measurements were not carried out at all in the immediate neighbourhood of the critical point , hut only at much lower temperatures , and yet there could be no question as to the course of the curve , when it was mapped .
But these are minor points ; and it is very gratifying to find that the material provided by Dr. Shields and myself affords such a remarkable confirmation of the justice of Professor van der Waals ' views.\#151 ; 1 1894 .
III .
" The Molecular Surface-Energy of Mixtures of non-associating Liquids .
" By Professor William Ramsay , Ph. D. , F.R.S. , and Miss Emily Aston , B.Sc. Received April 26 , 1894 .
It has been shown in the previous paper that it is possible to calculate the degree of association of an associating liquid such as alcohol , on the assumption that molecules of less complexity remain uniformly distributed along with molecules of greater complexity throughout the liquid , and that no one kind of molecule tends to congregate on the surface to the exclusion of the other .
It is necessary , however , to justify this assumption ; and for this reason experiments have been made on mixtures of liquids the molecules of which do not unite to form complex groups ; such are most of the liquids investigated by Ramsay and Shields ( 'Trans .
Chem. Soc. , ' vol. 63 , p. 1099 , et seq. ) .
The experiments of which an account is here given , show that while the height to which a mixture of two liquids ascends in a capillary tube is not the mean of the heights to which each singly would ascend at the same temperature , while the surface-tensions and the surface-energies are not necessarily the mean of those possessed by the liquids unmixed with each other , regard being paid to their relative proportion in the mixture , yet the coefficient of decrease of molecular surface-energy , and consequently the calculated molecular weights , are true means of those of the two liquids .
The substances used in these experiments were chosen in pairs , and as it was necessary in closing the tubes to evaporate some of the contained liquid in order to ensure the expulsion of air , mixtures of such liquids were taken as possess approximately equal boiling points , so 1894 .
] Molecular Surface-energy of Mixtures of Liquids .
183 that each should evaporate to nearly the same extent .
For this reason the following liquids were chosen:\#151 ; J Toluene L10o-6 at 761-2 mm. pressure .
1 Piperidine 105\#151 ; 106'*2 at 769 mm. pressure .
* J Benzene Constant , about 80 ' .
\ Carbon tetrachloride Constant " 77 ' .
f Chlorobenzene Constant , , 132 ' .
1 Ethylene dibromide .
Within 0'5 ' , about 131 ' .
f Carbon disulphide . . .
Constant , about 46 \#166 ; 2 ' .
L Chloroform Constant , about 62 ' .
The amount of toluene distilled was 750 c.c. The thermometer did not vary during the distillation by the fiftieth part of a degree .
The amount of piperidine was much less , about 75 c.c. The alteration of boiling point appears to be due , in part at least , to its eager absorption of carbon dioxide .
The benzene was part of a large stock which had been repeatedly frozen and thawed .
It was free from thiophene , and had an absolutely constant boiling point .
The carbon tetrachloride boiled constantly while 400 c.c. passed over .
The chlorobenzene was part of a stock used for securing constant temperatures , and had been repeatedly fractionated ; it boiled with absolute constancy while 750 c.c. passed over .
The ethylene dibromide was not quite so pure ; the rise of 0 5 ' , however , was spread over 200 c.c. ; while the purity of the carbon disulphide and the chloroform was guaranteed by the constancy of boiling point while large quantities distilled .
The molecular surface-energies of the pure substances were first determined .
They are given in the tables which follow:\#151 ; t \#151 ; temperature .
h = rises in centimetres in capillary tube .
p = density of liquid .
7 = surface-tension calculated by the equation 7 = \rqhp .
r = radius of tube .
7 ( Mw)i = molecular surface-energy , where M = molecular weight , and v = volume of one gram .
* The constancy of the hoiling-point of a liquid is the guarantee of its purity , provided a considerable quantity boils at a constant temperature .
The determination of the actual temperature involves the accuracy of the thermometer .
0-01843 184 Prof. W. Ramsay and Miss E. Aston .
[ June 14 , * .
2-068 2-168 2-055 rf* 2-125 g 9 04 dn* / " \#151 ; N | \lt ; p N 00 ^ VO O r\#151 ; 1 O CO O Cl CD lO vO CO e*9 ) a ip VO CD VO 492-6 426-3 .g \#163 ; 4\gt ; r-H CD 04 CD ( M 'H ip 0\gt ; CD 04 CD 04 04 04 i\#151 ; i \lt ; \#163 ; r2 *o 1 \#163 ; o 04 22 -90 19-26 \lt ; 3 p-i S cL 0 -8670 0 -8363 0-8048 0-7256 -2 a O -e c3 Q cL 1-6063 O \#187 ; H 04 i\gt ; rri x\gt ; JO rH iH \#171 ; S 3-785 3 -466 3-087 2-457 KS* 1 -861 1-643 1-442 !
04 CD H* VO 'VD CD GO 04 rH X\gt ; CO r\#151 ; 1 a o cc 00 .to 00 \#187 ; H 9 9 CD 00 H* X\gt ; 2-137 2-299 2-007 rH o o II 2-104 2-162 etn 1 631-0 563 -9 490-8 382 -2 eta* 'S ' | 575-6 502-3 433-1 aS S QO O CO CO rH CD C* 9 oo h* o vo 04 04 04 t\#151 ; l a5 S 29 -21 rH O ip ip O Oi 04 r2 ro EH d. 0-8682 0 -8380 0 -8080 0-7535 s pq du j 0 -8871 1 0-8500 j 0-8147 * 3-590 3-247 2 -867 2-280 -s I 3 -642 3-213 2-810 04 CD ip 'lO CD GO OJ rH ^ 4\gt ; CO rH 11'-2 ?
9 S s r = 001046 cm .
Chlorobenzene .
Ethylene dibromide .
t. Ji .
1 7* 7 ( MV ) ! .
Jc .
t. Ji P. 7* 7 ( Mu ) ! .
Jc .
9-5 45-6 77-8 131-3 5 -875 5-290 4-790 3-950 1-1182 1 -0795 1-0444 0-9836 33-71 29-30 25-66 19 -93 729 -1 648-8 581-0 469-8 2-225 2-104 2-079 12'*2 44-9 77-2 131-3 3-467 3-180 2-887 2-395 2 -1873 2 -1189 2 -0502 1 -9315 38-91 34-57 30-37 23 -74 757-7 687-7 617-6 502 -2 2-140 2-170 2-133 r \#151 ; 0 01046 cm .
r = 0-01708 .
Chloroform .
Carbon disulphide .
t. P- 7- 7 ( Mv)S .
Jc .
t. h P- 7- 7(Mw)j- Jc .
10'-2 45-5 77 -6 3-570 3-120 2-700 1-5077 1 -4385 1-3773 27-62 23-03 19-98 509 -7 438-5 374-0 2-016 2-010 o 9-7 46 -0 61-0 3-057 2-700 2-525 1 -2773 1 -2224 1 -1980 32 -73 27-68 25 -35 499-2 434-8 403-6 1-802 2-078 1894 .
] Molecular Surface-energy of Mixtures of Liquids .
185 186 Prof. W. Ramsay and Miss E. Aston .
[ June 14 , These resnlts call for no special remark , except that , contrary to the experiments of Ramsay and Shields , carbon disnlphide appears to associate somewhat at low temperatures .
Further experiments will be made on this matter at still lower temperatures .
The result given here may he taken as reliable , for it was carefully repeated several times , special precautions being taken to ensure the absolute purity of the bisulphide , and using a different capillary tube .
The densities were taken from the following sources :\#151 ; Toluene , Nasini and Pagliani , ' Jahresb./ 1862 , p. 63 .
Piperidine , Beilstein , ' Organische Chemie , ' vol. 3 , p. 616 .
Benzene , Kopp , 'Annalen , ' vol. 64 , p. 215 .
Carbon tetrachloride , Thorpe , ' Trans. Chem. Soc. , ' vol. 37 , p. 200 .
Chlorobenzene , determined by ourselves jit the temperatures chosen .
Ethylene dibromide , Thorpe , ibid. , p. 197 .
Chloroform , ibid. , p. 197 .
Carbon disulphide , ibid. , p. 364 .
Determination of the Capillary Rise of Mixtures .
I. Toluene and Piperidine .
( a. ) 5C6H5.CH3 to ICsHjo^H .
In filling the tube with this mixture O'113 gram was lost out of a total of 4 grams , or a little over 2 per cent. It may be assumed that , owing to these liquids having so nearly the same boiling point , no material alteration of their ratio is due to this cause .
The density of the mixture was assumed to be the mean of those of the constituents , taken in the proportion in which they were present .
As will be afterwards shown , no appreciable error is involved in this assumption .
The values of h are the mean of four observations in each case .
In the columns with the heading " calculated " the mean height , surface-tension , and surface-energy have been inserted , together with the mean values of Jc .
5C6h5-ch3 to ic5h10:nh .
Found .
Calculated , t. A. p(calctd .
) .
7- 7 ( Mr)* .
Jc .
A. 7- 7 ( Mr)* .
Jc .
14-5 3-647 0 '8684 28 *63 635*5 2-191 3-622 28-63 631-8 2-079 46 6 3 *285 0 '8377 24-88 565 -6 2-032 3-283 24-531 565 " 0 2 -277 78-4 2 -945 I 0 -8077 21-54 501-0 2-123 2 -904 21-19 492-6 2 013 132-5 2 *323 0-7535 15 -82 386*1 2*309 15-73 383-7 1894 .
] Molecular Surface-energy of Mixtures of Liquids .
187 The mean value of K , calculated over the whole range of temperature , is 2-115 ; the mean value found is 2-123 ; hence it may be concluded that these liquids are without influence on one another .
It will also be noticed that the found and calculated values of 7 and of .y(Mt\gt ; )s are in very close correspondence .
To check these results the proportions were reversed , and a mixture of ( b ) 5C5H10'.NH to 1C6H5CH3 was investigated .
5c5h10:nh to 106H3CH3 .
Found .
Calculated .
t. h. p(calctd .
) .
7- k. h. 7- 7(Mt/ )3 .
k. 14-9 46-6 78-4 132 5 3-757 3-404 3-046 2-422 0-8674 0-8366 0-8052 0 -7527 29-46 25-74 22-17 16-48 632-0 565-6 499-8 388-6 2 -094 2-069 2-056 3-752 3-429 3 -050 2-427 29-42 25-94 22 -21 16-52 634-9 569 6 499-9 389-2 2 -060 2 -190 2 -047 The mean value of Jc , calculated over the whole range of temperature , is 2-087 ; the mean value found is 2*067 .
Here again the found values of surface-tension and of surface-energy agree very closely with those calculated .
It is possible , without assuming a molecular weight , to calculate the mean molecular weights of such mixtures by means of the equation l ML : L \#151 ; -0 \lt ; yV\#187 ; r- In doing this , the mean value of h for each mixture has been taken , and the value of M has been calculated between extreme limits of temperature .
For the first mixture the mean molecular weight found is 90*61 ; that calculated for a mixture of five molecules of toluene to one of piperidine is 90*83 ; for the second mixture the mean molecular weight found is 86*12 , and the calculated value for a mixture of five molecules of piperidine to one of toluene is 86*17 .
This , of course , constitutes only an arithmetical check on the other results , but if a mean value for h had been chosen , e.g. , 2*121 , the results would have been practically the same .
It is obvious that in this case the liquids are without influence on each other .
188 Prof. W. Ramsay and Miss E. Aston .
[ June 14y II .
Benzene and Carbon Tetrachloride .
( a. ) 1C6H6 to ICCI4 .
Similar experiments were made with the above mixture ; the densities of the mixture were determined experimentally , and a comparison is given at each temperature of the numbers found , with the mean density of the mixture , calculated from the found densities of the components .
The loss of weight on sealing showed that 4*5 per cent , of the mixture had evaporated ; hut here , too , the boiling points of benzene and of carbon tetrachloride are so near that it is probable that no important change in composition occurred .
1C6H6 to ICCI4 .
h. P* 7- 7(Mw)\#187 ; .
1c Hd nd n3 pc 'T ?
T5 -e nd T3 t. d r-\#171 ; O d " o d O d 0 d -S d d d d *3 d f=i O O 1=5 O O O 16*0 2*421 2-709 1 -2597 1 -2558 27-70 27-58 561*9 562-4 2-331 2*148 46*2 2-150 2 -428 1 -2095 1-2098 23-50 23-80 492-0 497-5 2-110 2*118 78 *2 1-880 2-126 1 -1596 1-1590 19-71 19-98 424 -5 429 *7 The calculated values of h are the means of the heights of benzene and carbon tetrachloride at the respective temperatures .
It is to be noticed that the observed heights are widely different .
But , owing to the different densities of the two liquids , the calculated values of 7 , the surface-tensions , are nearly the means of those of each taken singly ; and the agreement of the found and calculated molecular surface-energy , \lt ; y(MnP , is also a close one .
The value of exaggerates the error of experiment , yet , on the whole , the agreement is satisfactory .
It would also appear that the operation of mixing does not affect the density of either liquid appreciably .
( 5 ) and ( c ) .
The tables which follow show the effect of varying the relative proportions :\#151 ; ( 5 .
) 10C6H6 to 17CC14 .
h. P* 7- 7(Mi/ )3 .
Jc .
nd nd nd nd nd nd nd nd ns nd t. d 0 d d 0 0 p 0 d " o 0 d d d ro3 O \#163 ; O O 0 \#163 ; O 18*2 2*265 2 '508 1-3509 1 *3505 27*66 27 -64 568-1 566 -9 2-127 2-119 , 46*6 2*010 2*224 1-2942 1*2963 23*51 23-57 497-0 496-1 2-290 .119 78*4 1 *740 1*948 1 -2411 1*2422 19 *52 19-79 424-2 428-7 1894 .
] Molecular Surface-energy of Mixtures of Liquids .
189 ( C\ 2C6H6 to ICCI4 .
t. h. P- 7- 7(Mu ) .
k. r* 5 pr 0 rH 6 d g 0 if o nd d d 0 Pr T2 0 s 0 T5 d d 0 Pr " 3 O T3 d O pR .4\#151 ; \gt ; O 10'8 46-2 78-2 2-769 2-440 2 121 3 -052 2 -689 2-354 1-1384 1 -0877 1 -0431 1-1395 1 -0899 1-0445 28-55 23-99 20-00 28 -37 24-11 20-22 576-2 499-5 428-1 574-1 499-1 430 -8 2-198 2-231 2-083 2-133 The same remarks apply to these results .
The mean values of k , calculated over the whole range of temperature are given in the next table , together with the values of M , also deduced between the extremes of temperatures by the equation already given .
k. M. Found .
Calculated .
Found .
Calculated .
a. 2-125 2*131 113 *1 112 -9 b. 2 -207 2*124 125 *8 125 -72 c. 2-077 2*106 103 *5 103 -3 Here , again , within limits of experimental error , it is seen that the values of 7 of 7(Mr)5 , and of k are uninfluenced by the operation of mixing , and that the mean molecular weight of the mixture is calculable from the data found .
III .
Mixtures of Chlorobenzene and Ethylene Dibromide .
A mixture of equal molecular proportions of these liquids gave the following results:\#151 ; Prof. W. Ramsay and Miss E. Aston .
[ June 14 , h. ? .
7- y(Mv)l .
k. Hd rd nd T ?
TS H3 nd Tj *-6 t. P Q -g p O P g P " o d *3 d 0 *1 d d *3 d \#163 ; O O \amp ; O \#163 ; O Pr O 10'-2 4-292 4-666 1 -6064 1 -6107 35 -38 36 -40 729 -0 748*5 2-147 2-190 45*6 3 -890 4-232 1 -5519 1 -5564 30 -97 31-89 653-0 671 -0 2-152 2-142 77 *8 3-515 3 -836 1 -5014 1 -5056 27-08 27 -98 583 -7 602-0 1-989 2-113 131 -3 2-925 3 -173 1 -4154 1-4184 21 -23 21 -83 477 -3 488-9 The radius of the tube was 0'01046 cm .
The calculated height given is the mean of the heights of chlorobenzene and ethylene dibromide , corrected for temperature-difference , on the assumption ( which is practically without error for such small differences ) that the variation of height with temperature is a linear one .
The calculated density is a similar mean .
But these heights and densities are not made use of in calculating the values of the " calculated " 7 .
It , too , is the mean of the found values ( see p. 184 ) , and similarly the values of 7(Mw)s are calculated from the calculated values of 7 , and from the calculated densities .
The molecular weight , computed by means of the equation given on p. 187 , for the whole range of temperature employed , is 148'6 instead of the theoretical mean 150*25 .
IV .
Mixture of Chloroform and Carbon Disulphide .
Here also equal molecular proportions were used .
r = 0*01046 cm .
i. h. P ' 7 .
k. Z3 P O PR Tj * 0 P O PR -g O P d \amp ; nd " 0 O nd p 0 PR T5 -g O T5 P O Pr g O 9-0 44-9 61*0 4*050 3*560 3*300 4-723 3-788 3-520 1*4026 1 -3406 1*3128 1 -4132 1 *3496 1*3213 29 -16 24 -49 22*23 30 *30 25 *47 23 -23 493 -7 427 -4 392 -5 510*6 442 *5 409*3 1- 847 2- 168 1 -897 2-062 I As the value of k for this mixture points to association at low temperatures , the mean molecular weight has not been calculated .
1894 .
] Molecular Surface-energy of Mixtures of Liquids .
191 We see , from these experiments , that a mixture of two liquids may either behave as a mean , or the liquids may influence each other .
With the first two pairs , toluene and piperidine , and benzene and carbon tetrachloride , the liquids belong to very different chemical types .
With the second pair , the heights and densities differ very greatly from each other , and although the values of 7 and of y(Mr)* approach more nearly , there is still a marked difference .
Yet the values of p , the density , 7 , the surface-tension , and 7(Mr)* , the molecular surface-energy of the mixtures , are identical with those calculated , within limits of experimental error .
Perhaps this statement should , in strictness , not apply to the densities , yet there is no great divergence from the mean .
It is therefore legitimate to state that in certain cases the molecular surface energy of a mixture is the mean of those of its constituents determined at the same temperature .
The third pair of liquids , chlorobenzene and ethylene dibromide , give results belonging to a different category .
Here the calculated density is greater than that found .
This implies expansion on mixing .
The values of 7 are also greater , and together with these the values of 7(Mr)i5 .
But the rate of alteration of 7(Mr)5 with temperature is practically normal , and the mean molecular weight can therefore be calculated with fair approach to accuracy .
The fourth pair of liquids give still more abnormal results , owing probably to the fact that one of them has some power of association .
It would be premature to discuss these results without much more extended experimental evidence .
Experiments have already been made with mixtures of alcohol and other liquids , and an investigation of mixtures of acetic acid is still in progress .
The problem is a complex one ; we have to deal with the extent to which the association of an associated liquid is altered by dilution ; and it will form the subject of a further communication .
We have , however , thought it advisable to bring forward some results in this paper to avoid the possible generalisation from the behaviour of the first two pairs of liquids that the molecular surface-energy of all liquids is the mean of those which they possess when unmixed .
One question remains to be considered ; it is this : Is it justifiable to assume that the molecular surface-energy of a mixture of the associated and dissociated molecules of a substance whose molecular complexity alters with temperature is the mean of each taken singly ?
For , on that assumption , the mean molecular weights of associating liquids have been calculated .
In our opinion it is ; but , as direct experimental evidence is not as yet attainable , it may be well to bear in mind that , although a fair working hypothesis , it cannot be taken as a proved fact .

