260 WORLEY: THE SURFACE TENSION OF MIXrURES. PART I.XXIX.-The Surface Tension of Mixtures. Part 1.Mixtures of Pal-tly Miscible Liquids and theInjueizce of Solubility.By RALPH PALLISER WORLEY.THE most marked peculiarities of the surface tension of mixturesof liquids are, first, the lowness of the surface tension of aqueoussolutions of liquids which are only partly miscible with water,and, secondly, the divergence of the surf ace tension-compositioncurves, in the case of liquids, which are perfectly miscible, fromthe straight line that would express the relationship between surfacetension and composition if the former were merely an additiveproperty; that is to say, by what is generally called the admixturerule. The following investigations were carried out in order t othrow more light, if possible, on these phenomena, and in the* While this paper was being written, I have been in communication with Prof.Kehrmann, who has forwarded me a dissertation of one of hie students. Confirma-tion of v.Liebig's results as to the different varieties of resorcinol-benzeiu was notobtainedWORLEY: THE SURFACE TENSION OF MIXTURES. PART I. 261present paper will be found the results of the study of the surfacetension of mixtures of partly miscible liquids. The case of mixturesof perfectly miscible liquids is dealt with in the succeeding paper.Before proceeding to a description of the experiments carriedout, a short account of the results obcained by some other observersis necessary to show wherein the peculiarities lie.Duclaux (Ann.Chim. Phys.; 1878, [v], 13, 76) carried out aseries of experiments on the surface tension of mixtures of thehomologous series of alcohols and fatty acids with water in varyingproportions. His measurements were made by means of the droppipette. He pointed out that a considerable drop in surfacetension occurred on passing from pure water to a weak alcoholicsolution, and that this drop increased in magnitude as the molecularweight of the alcohol increased. The following numbers are takenfrom his tables, and represent the surface tensions of 1 per cent.solutions by volume, that of pure water being taken as unity:bIeth3 1 alcohtll and water .......................... 0.962Ethyl , , ,, , ..........................0 -933isoProlly1 ), ,) ) ) ........................... 0 900isoButy1 ) ) ,) ,, ......................... 0-742Amy1 ,, ,, ) ) .......................... 0.594He pointed out, moreover, that the surf ace tension-compositioncurves were of the form of hyperbolas, and the following formulawas proposed to express the curves:y = k(ez - l),where y =surface tension, and x =percentage composition of themixture by weight.J. Traube (Ber., 1884, 17, 2294) determined the capillaryheights of mixtures of the alcohols, fatty acids, and isomeric esterswith water. Unfortunately he did not find the densities of themixtures, so it is impossible to convert his results into surfacetension expressed as dynes per centimetre. However, the numbersobtained agree relatively with those found by Duclaux.He alsonoticed the decreasing value of the surface tension of a solution ofgiven strength with increasing molecular weight of the solvent, andhe formulated the following expression to represent the relationshipbetween the two:where ha and hfl represent the heights to which solutions of thesame concentration of two substances the molecular weights ofwhich are Ma and Mp respectively will rise in a given capillarytube, and h', and Id, the heights corresponding with another equalconcentration of the same two liquids262 WORLEY: THE SURFACE TENSION OF MIXTURES. PART I.What neither of these investigators * pointed out, however, andwhat has hitherto escaped attention is that there is a marked fallin the surface tension of a solution of given strength on passingfrom the members of a homologous series which are miscible withwater in all proportions to those members which are only partlymiscible a t the ordinary temperature.This fact is clearly shown inthe above table. It seems remarkable, too, that such a weaksolution of amyl alcohol as 1 per cent. should have a surface tensiononly one-half as great as that of water, whilst a solution of ethylalcohol of the same composition has a surface tension almost asgreat as that of water itself. It seemed probable from such con-siderations as the above that a relationship exists between thesolubility of the solute and the surface tension of its solutions.With a view, therefore, to learn more concerning the relationof surf ace tension to mutual miscibility, experiments were madewith aniline, phenol, and isobutyl alcohol, the approximate solu-bilities a t 1 5 O being f o r aniline 1 part in 30, for phenol 1 part in 15,and for isobutyl alcohol 1 part in 10.The method adopted of measuring the surface tension was themethod of capillary rise, the surface tension being calculated fromthe formula :S = $rghd,where S =surface tension in dynes per centimetre ;T = radius of capinary tube in centimetres ;g=981 dynes;h = capillary rise in centimetres;d = density of liquid.Two factors have been neglected in the above formula, d’ thedensity of the vapour above the liquid which ought to be sub-tracted from d, and JT, which ought to be added to the capillaryheight h.Both of these quantities are, however, so small that noappreciable errors have been introduced in omitting them.The apparatus consisted of a tube 25 cm. long and 2 cm. wide,which contained the liquid to be examined. It was sealed off a tthe lower end, and closed by means of a rubber stopper at theupper end. The stopper was perforated to hold a glass rod whichpassed down into the tube, and to this rod the capillary tube wasfastened by means of fine platinum wire. By the vertical motion* The’ following investigators may also be referred to : Rodenbeck ( D i m ,Bonn, 1879) ; Whatmough (Zeitsch. physiknl. Chem., 1902, 39, 158) ; Volkmanii(Ann. Phys. Clmn., 1882, [iii]. 16, 321 ; Torch (Ann. Physik, 1905, [iv], 17, 744) ;Lewi3 (Phil.i M q . , 1908, [vi], 15, 499) ; Lqlinstein (Ann. Physik, 1906, [iv], 20,614) ; Ramsden (Pmc. Any. Soc., 1903, 72, 156) ; Drucker (Zeitsch. physikal.Chem., 1905, 52, 678) ; Milner (Phil. Mag., 1907, [vi], 13, 96)WORLEY: THE SURFACE TENSION OF MIXTURES. PART I. 263of the rod, therefore, the capillary tube could be raised or loweredin respect to the surface of the liquid. In order t o obtain accurateresults, it is absolutely necessary that the liquid be made to flowover the end of the capillary tube before each reading to ensurethe walls of the tube above the meniscus being wet, and the surfaceof the liquid fresh. The whole was contained in a jacket made ofglass tubing of wide bore. Water a t various constant temperatureswas passed through this jacket, the temperature being adjusted byincreasing or decreasing the rate of flow of the water.The diameters of the capillaries were measured by means of amicroscope furnished with a micrometer eyepiece.Measurementswere taken in four directions, and unless these agreed the tubeswere rejected.The tubes -were cleaned and dried by drawing through themboiling chromic acid, distilled water, alcohol, and dry ether in theorder named.The capillary heights were measured by means of a cathetometerwhich read to a hundredth of a millimetre. Four readings weretaken in each case, the tube being lowered and raised again bymeans of the glass rod before each observation. The rod wasadjusted each time so that the meniscus fell to a point 1 cm.fromthe top of the capillary tube. The capillary heights contained inthe tables are the means of the four observations.*The densities were measured by means of a 25 C.C. Sprengel tube.Effect on the Surface Tension of Decreasing the Solubility ofAniline in Water by the Addition of Common Salt.The solubility of aniline in water is greatly diminished by theaddition of salt. The effect of this change of solubility on thesurface tension of the mixture was the first thing to be investigated.Twenty-five C.C. of a solution of aniline in water (containing 1 partin 60) were taken, and salt was added in quantities of 1 gram a ta time until the eolution became turbid. #The capillary height andthe density were measured after each addition of salt, and theresults obtained are contained in table I.The symbols a t the headof each column refer to the quantities above mentioned, whilstG=number of grams of salt added. The temperature was 20°.* A tube of different diameter was always used in addition so as to keep a checkThe two tubes gave results which never differed by more thanThe results obtained from the second tube are not containedon the results.one-tenth of a dyne.in the tables264 WORLEY: THE SURFACE TENSIOX OF MIXTURES. PART I.TABLE r.G. T (cm.). d. h (cm.). x.0 0 02154 1 -000 5.1 91 54.8341 0'021 54 1.026 4'780 51.8162 0.02154 1.051 4.470 49'5903 0'02'154 1.070 4'222 47'7524 0.02154 1.094 3.989 45'9405 0.021 54 1-119 3*P75 45'803T i 3 solution became turbid when the fourth gram had beenadded, and the fifth gram had very little further effect.Thesurface tension of pure aniline a t 20° is 42.The change in surface tension brought about by the additionof salt to water is very small, and consists, moreover, in an increaseof the tension. Thus, according to Whatmough (Zeitsch. ph ysiknl.Chem., 1902, 39, 149), the surface tension of a 2N-solution ofsodium chloride is 79.35, that of water being 75.57. The coil-siderable fall in surface tension shown in the above table cannottherefore be due to the mere addition of the salt, but most probablyi t may be due t o the diminution of the solubility of the aniline inwater. The evidence, however, is not conclusive, and the followingfurther investigations were therefore carried out.The EfJect of Temperature and the Accompanying Changes inSolubility on the Surface Tension.Since the mutual miscibility of such liquids as aniline, phenol,and isobutyl alcohol with water is greatly dependent on the tem-perature, the liquids becoming soluble in all proportions a t hightemperatures, it seemed necessary t o find out how the surfacetension of these mixtures varied when the temperature wasincreased.A series of experiments was therefore made withaqueous solutions of varying strength of the three liquids men-tioned, and the surface tension was found a t several temperaturesfrom the room temperature up t o looo. The surface tension ofthe distilled water used t o make up the solutions was measured a tfour different temperatures.1.-Aniline and Water.The aniline used was of conrkant bciling point. Solutions ofdifferent concentration were made up, and the surface tensionwas determined a t various temperatures.The results are set forthin the following tablesWORLEY: THE SURFACE TENSION OF MIXTURES. PART I. 265TABLE 11.3-33 C.C. of Aniline and 100 C.C. of Water.t. r (cm.). d. h (cm.).15" 0.0250 1.0028 3.88755 0'0250 0.9920 4.07795 0.0250 0.9665 4'274TABLE 111.2.5 C.C. of Aniline and 100 C.C. of Water.11" 0 *0250 1.0027 4.10621 0.0250 3'0016 4.13132 0.0250 0.9997 4.16946 0'0250 0-9927 4.23161 0 0250 0'9850 4,28380 0.0250 0.9761 4.364100 0'0250 0-9656 4-455TABLE IV.2 C.C. of Aniline and 100 C.C. of Water.16" 0,02154 1 -001 8 5.06438 0.021 54 0.9942 5.10560 0,02154 0.9915 5.15680 0.02154 0.9747 5.273TABLE V.1.5 C.C. of Aniline and 100 C.C.of Water.16" 0.02154 1.0013 5.36438 0.02154 0'9925 5 *38960 0'02154 0.9900 5-35590 0.02154 0.9746 5.389TABLE VI.1 C.C. of Aniline and 100 C.C. of Water.15" 0.02154 1.0012 5.72053 0.02154 0'9904 5-54585 0'02154 0.9741 5.563TABLE VII.0.5 C.C. of Aniline and 100 C.C. of Water.0'02154 1 0011 6.21730 15" 0.06154 0.9978 6.119(Cu,rve 6.)S.48.20649'22250.729(Curve 5.)50.49650'70151.11151.50451'73252.23052'754(Curve 4.)53.59653.62354.01 254.300(Curve 3.)56.74456.51456.01255.499(Curve 2.)60.50558 02357'254(Curve 1.)65.75364'50556 0'021 54 0,9880 5 915 61.73380 0.02154 0 9758 5'724 59'00266 WORLEY: THE SURFACE TENSION OF MIXTURES.PART J.TABLE VIII.Surface Tension of Distilled Water.17" 0.0297 0.9988 7.188 72.88641 0'0207 0'9922 6.878 69.28359 0'0207 0.9840 6 666 66 57377 0.0207 0.9737 6.441 63.256f. r (cm.). d. h (em.). S.F I G . 1.45 -35 '-20" 40" 60' 80" 100" 120" 140"Temperatzcre.The above results are shown graphically in Fig. 1. It will beseen that increase of temperature has a very different effect fromthat observed in the case of pure liquids. The curves representingdilute solutions fall with rising temperature but less rapidly thanthe water curve. At higher concentrations the rate of decrease ofsurface tension becomes markedly less, until in the case of solutionswhich are nearly saturated the interesting fact is exhibited thaWORLEY: THE SURFACE TENSION OF MIXTURES.PART I. 267with increase of temperature the surface tension even rises. Atall concentrations increasing the temperature reduces theabnormally large difference between the surface tension of thesolutions and that of the solvent, and the surface tension of thesolutions tend to approach the value they would be expected tohave if the liquids were miscible in all proportions. A t lowtemperatures the surf ace tension of nearly saturated solutions isnot much greater than that of pure aniline, whilst a t high tem-peratures it becomes nearly as great as that of water.2.-Phenol and Water.The phenol was '' pure crystallised phenol " and was recrystallisedThe solutions were made up as before, but by weightThe surface tension of phenol above itsThe results of the experimentsbefore use.instead of by volume.melting point was also determined.were as follows:TABLE IX.Supface Tension of Phenol above its Melting Point.to. r (cm.).d. h (cm.).49" 0.01528 1.0514 4'63766 0.01528 1.0387 4'46181 0.01528 1.0264 4.311100 0 '01 528 1.0121 4'101TABLE X.6-66 grams of Phenol artd 100 C.C. of Water.15" 0.02038 1'0069 4.06535 0 *02038 1'0011 4.02857 0-02038 0.9931 4.06676 0?)2038 0.9840 4'11090 0*02038 0.9767 4-249TABLE XI.3-33 grams of Phenol and 100 C.C. of Water.15" 0.02038 1.0054 4'74835 0 -02038 0.9997 470253 0.02038 0.9925 4-72675 0 *02038 0.9825 4.77090 0.02038 0.9730 4'827TABLE XII.2 grams of Phenol and 100 C.C.of Water.14" 0.02038 1 '0028 5.36236 0 -0 2038 0-9976 5-28755 0.02038 0'9919 5302s.36.54034'72833'15931 -098(Curve 5.)40.88940.30040.36540'42541 '485(Curve 4.)47.70346.94146.88846'85046.945(Curve 3.)53.74852 71252.56652.048 75 0.02038 0.9815 5.30588 0 '02038 0.9729 5.285 51 '40268 WORLEY: THE SURFACE TENSION OF MIXTURES. PART I.TABLE XIII.1 gram of Phenol aizd 100 C.C. of Water. (Curve 2.)to. T (cm.). d. h (cin.). S.15" 0.02038 1.0018 6-029 60.36837 0.02038 0.9961 5-916 58.91160 0.02038 0.9897 5.888 58'14680 0 *02038 0.9776 5.815 56,82490 0.02038 0 9711 5.717 55.500TABLE XIV.0.5 gram of Phenol and 100 C.C. of Water. (Curve 1.)16" 0-02038 1.0017 6.583 65.84540 0.02038 0-9955 6'41 1 63 *79861 0.02038 0.9827 6.277 61.66081 0.02038 0.9760 6'112 59.662The above results are set forth graphically in Fig.2, and hereagain the same peculiarities as those shown by aniline are exhibited,FIQ. 2.20" 40" 60" 80" 100" 120"TemperatureWORLEY: THE SURFACE TENSION OF MIXTURES. PART I. 269although the curves do not show such a marked tendency to rise.The surface tension of the saturated solution is approximately thesame as that which would be obtained for phenol alone byextrapolation a t room temperature.isoBzcty1 Alcohol and Water.Unfortunately no pure isobutyl alcohol could be procured, butsome was prepared by saponifying isobutyl acetate. The latter witsKahlbaum's, and before use was redistilled.The alcohol kept forthe following experiments distilled a t 105-107*. Solutions of fourdifferent concentrations were made up, and their surface tensionsdetermined a t different temperatures. The surface tension of thepure alcohol was found a t one temperature only, the result agreeingwell with that found by other observers. The following tablescon taiii the results of the experiments.TABLE XV.Surface Tension of isoButyl Alcohol. (Curve 5.)to. r (ctn.). &. h (cm.). S.15" 0 01528 0.8094 3.073 22.919TABLE XVI.10 C.C. of isoButyl Alcohol and 100 C . C . of Water. (Curve 4.)15" 0.01528 0.9870 3 651 27 -01 144 0.01528 0'9785 3 205 23.51076 0.01528 0 9656 2.768 20.029TABLE XVII.6 C.C. of isoButyl Alcohol arm? 100 C.C. of Water.(Curve 3.)1 6" 0-01 528 0'9906 4.378 32'47046 0.01528 0.9808 3.956 29.07865 0 *O 1528 0.9724 3.708 27.021TABLE XVIII.3 C.C. of isoButyl Alcohol and 100 C.C. of Water. (Curve 2.)14" 0 01528 0 9962 5.410 40.39350 0.01528 0.9804 4.875 35231880 0.01528 0-9684 4.458 32.353TABLE XIX.1.5 C.C. of isoButyZ Alcohol and 100 C . C . of Water. (Curve 1.)14" 0.01528 0.9990 6.912 51.74946 0 '01 528 0'9891 6.449 47.80478 0.01528 0.9716 6.012 43'74270 WORLEY: THE SURFACE TENSION OF MIXTURES. PART I.The graphic representation of these results will be found inFig. 3. It is noticeable that the curves are markedly differentfrom those obtained with solutions of aniline and phenol, therebeing little o r no tendency for the abnormally low value of thesurf ace tension to disappear with rise of temperature.When the changes in solubility which are brought about byincrease of temperature as regards aniline, phenol, and isobutylalcohol are examined in detail, it is found that the changes are byno means the same in each case.I n the following table will beFIG. 3.Tempernt w e .found the percentage solubilities up to the critical solution tem-perature, the numbers f o r aniline and isobutyl alcohol being dueto Alex6ev (Ann. Phys. Chem., 1886, [iii], 28, 305), and those forphenol to Rothmund (Zeitsch. physikal. Chem., 1898, 26, 433):Aniline. - Per cell t.2 0" 3.140 3.360 3.880 6.5100 7-2120 9'1140 13-5167 48.6Phenol.7-l'er cent.209 8-4030 8.9240 9 *7845 10 6255 13.8860 17.1065 22-2668.8 35-90isoButyl alcohol.e-Per ccnt.O0 13.020 9'040 7.560 7.080 7.0100 8.0220 16.0133 40.WORLEY: THE SURFACE TENSION OF MIXTURES. PART I. 211It is seen that whereas the solubility of aniline and phenolincreases up to the critical temperature, that of isobutyl alcoholdecreases up to 70°, and then begins ta increase again until thecritical temperature is reached.On comparing the solubilities of these substances with the surfacetension of their solutions, it is evident that some close connexionexists between the two phenomena, and the ‘abnormally low valuesof the surface tension would appear to be connected with the degreeof immiscibility of the liquids. When the solubility is increasedby raising the temperature, as in the case of aniline and of phenol,the surface tension becomes less and less abnormal, and tends t oapproach the value it might be expected t o have if the liquidswere miscible in all proportions, whereas, when increase of tem-perature is not accompanied by increased solubility, as in the caseof isobutfl alcohol, there is no such tendency (see Fig.3). More-over, it was shown (table I) that on decreasing the solubility ofaniline in water by the addition of salt, the surface tension wasconsiderably diminj shed.It is not surprising that an aqueous solution of a liquid of lowsurface tension, miscible with water only in small proportions,should have an abnormally low surface tension when nearlysaturated. A t the point of saturation there is a tendency for theliquid to separate out, and even below this point it probably existsto some extent as molecular aggregates of low surface tension whichare continually being formed and resolved.Such a condition,however, could not be expected to exist in dilute solutions in whichthe surface tension is also abnormally low. The explanation of thelow values in this case lies in all probability in the fact that thesolute is not uniformly distributed, but is concentrated a t thesurf ace.According to Willard Gibbs ( I ‘ Thermodynamic Studies,” pp.219-300), a solute which increases the surface tension of thesolvent is pushed out of the surface because the molecules tend t oarrange themselves in a system having the least potential energy,and vice versa, a solute which decreases the surface tension isconcentrated in the surface.Various investigators have shown that the surface layer of asolution often has a different composition from the bulk.Most ofthe experiments have been carried out with colloidal substances,but the same phenomenon has been shown to exist in the case oftrue solutions. Reference may be made to the work of Dupr6(Ann. Chim. Phys., 1866, [iv], 7, 409), Rayleigh (Proc. Roy. Znst.,1890-1892, 13, SS), Milner (Phil. Mag., 1908, [vi], 15, 499),Ramsden (PP’oc. Roy. SOC., 1903, 72, 156), Lewis (Phil. Mag., 1908272 WORLEY: THE SURFACE TENSION OF MIXTURES. PART I.[vi], 15, 49; 1909, [vi], 17, 466), and Donnan and Barker (Yroc.Roy. ~'oc., 1911, A , 85, 557).There is thus good reason to believethat solutes act in general in the manner predicted by Gibbs.I n the case of solutions of aniline, phenol, and isobutyric acid,then it is probable that the low values of the surface tension, evenin dilute solutions, are due to the fact that the s01ut0 is concentrateda t the surface, which'is thus much nearer the point of saturationthan is the bulk of the solution. Increase of temperature increasesthe solubility of aniline and phenol, and consequently the surfacelayer, even if its concentration does not fall, becomes less nearlysaturated as the temperature is raised. The increased solubilitywould also possibly cause the concentration of the bulk to becomegreater a t the expense of the surface layer.I n the case of isobutyl alcohol, since increase of temperature isnot accompanied by increased solubility, the surface lay& does notbecome less nearly saturated as the temperature is raised, and,moreover, there is no reason to expect that the concentration ofthe surface layer would diminish.It is thus to be expected thatthe abnormal lowness of the surface tension will not tend todisappear when the temperature is raised. I n this connexionmention may be made of some experiments carried out on theduration of froth upon the different solutions. An aqueous solutionof aniline agitated violently f o r thirty seconds a t 20° formed acopious froth, which had not entirely disappeared after fifteenminutes had elapsed. The same solution when agitated similarly a t70° formed a froth which disappeared entirely in ten seconds. I nthe case of isobutyl alcohol the froth was not so copious. It lastedfor thirty seconds a t 20°, and for thirty-five seconds a t 50°.To conclude, there seems to be little doubt of the correctnessof the view advanced in an earlier part of the paper that the lowsurface tension of the aqueous solutions studied is intimatelyconnected with the comparative insolubility of the solutes. Thereduction of the surface tension of solutions of aniline accompanyinga reduction of solubility brought about by the addition of salt, andthe comparison of the effects of increase of temperature on thesolubilities of the liquids studied and on the surface tensions oftheir solutions ail support this conclusion. There is good reasonto suppose that the lowness of the surface tensions of the lessconcentrakd solutions is due to the concentration of the surfacelayer being greater than that of the bulk of the solution, and thusconsiderably nearer the point af saturation.UKIVERSITY COLLEQE,AUCKLAND, N. Z