1506 LAING AND McBAIN' THE INVESTIGATION O F CLXVI1.- The Investigation of Sodium Oleate Solutions in the Three PhysicaZ States of Curd Gel and Sol. By MARY EVELYN LAING and JAMES WILLIAM MCBAIN. ONE of us (M.E.L.) in continuing previous work (l'. 1919 115, 1280) has now found that aqueous sodium oleate a t temperatures between Oo and 25O can be bronght at' will into any one of three typical states namely clear oily liquid sol clear transparent elastic gel or white opaque solid curd all at one and the same concentration and temper a tur e. Hitherto the last two types have not been differentiated, althciugh as will be shown they are wholly independent and exhibit very different properties. Probably nearly all of the observations on solidified soap systems recorded in the literature refer to what we have here defined as soap curd and some con-fusion would have been avoided if the existence of these two separate types had been recognised (see for example the con-troversy bet'weeln Zsigmondy Bachmann and von Weimarn with regard to whether or not soaps are "gels").The real soap gels are unmistakable since they resemble gels of gelatin. It has become evident that under suitable conditions all soaps may exist in these three forms and each form is important both for the theory and manufacture of soaps. The chief experimental discovery described in this paper is that soap sol and soap gel are identical in all respects with the excep-tion of elasticity and rigidity which are characteristic of ihe gel form alone. The following properties are all identical in sol an SODIUM OLEATE SOLUTIONS ETC.1507 gel (i) electrical conductivity; * (ii) lowering of vapour pressure, which measures the amount of substance in true solution; (iii) refractive index; and (iv) as Salmon has recently shown in this laboratory (this vol. p. 536) the concentration of the sodium ion. A soap curd on the contrary is a sol or gel from which a part or nearly all of the soap has been abstracted through the formation of white curd fibres as Darke McBain and Salmon have shown in another communication. The distinctive structural feature of soap curd is the separation of hydrated soap as a felt of long, white fibres these fibres being of barely microscopic diameter. Zsigmondy maintained that the formation of the soap curds which he studied was essentially a process of crystallisation; the analogy is certainly close.It would seem an unavoidable conclusion from the observed identity of so many properties of the sol and gel of soap that the chemical equilibria are the same and that therefore the colloidal particles present in the two are identical both in nature and amount. It is evident that this is of direct significance in the general theory of the structure of gels, .It has long been known that salt solutions retain the greater part of their conductivity and diffusibility when gelatinised by the addition of gelatin. This was however less surprising since the conductivity could be largely ascribed to the electrolyte and to this extent it was a mattea. of indifference what happened to the gelatin on gelatinisation.To those who regarded the gelatin solu-tion as heterogeneous whether in the form of sol or gel any alteration in the degree of dispersion even amounting to removal of the gelatin from solution need not necessitate any great change in the conductivity of the dissolved salt. No such explanation is possible in the case of soap solutions where no extraneous chemical is present and where the colloid arises from and is in true equilibrium with the crystalloidal con-stituents. Any essential alteration in amount or degree of dispersion of the soap colloid would have been instantly reflected in changed conductivity well illustrated in the behaviour of curds. A further point to be mentioned here is that the gel in its genesis is much more closely related to the curd than to the sol.With the less concentrated solutions i t was found almost impossible * Since this paper was submitted for publication i t has been pointed out t o us that Arrhenius so long ago as 1887 (Oefvers. Stockholm Akad. 6 121) showed that the conductivity of gelatin-water-salt systems is the same whether in the sol or gel condition. This was quite unknown to LIB and, so far as we know this important observation has been overlooked. 3 9 1508 LAMG AND McBAIN THE INVESTIGATION O F to prepare a gel except through formation of a curd which was then gently warmed until it became clear. On careful cooling it could be retained in the form of gel. This is a t complete variance with the emulsion conception of gelatinisation as involving merely a large increase in viscosity on lojwering through a small zone of temperature.It would seem necessary throughout the discussion of colloids to distinguish clearly between gelatinisation and coagulation that is, between the formation of a gel and a curd or coagulum whether the latter is irreversible as in the case of a boiled egg or reversible, as in the case of a soap curd. The experimental data here presented include first systematic measurements of the conductivity of sol gel and curd of sodium oleate over the whole range of temperature 0-28O a t which it is posible to obtain gel or curd. These are supplemented by measurements of refractive index and of vapour tension and by direct analyses. Finally some observations on the behaviour of curds of mixtures of soaps are recorded.EX P E R I Y E N TAL. The earlier experiments had to be carried out with a Eohlrausch (Kohler) apparatus whereas with the final work a set of Washburn (Leeds and Northrup) apparatus was available. Even when not using the pure sine wave current from the Siemens high-frequency machine or the Vreeland oscillator but with an ordinary induction coil the position of the minimum was defined with an eight-fold accuracy by the Washburn set as com-pared with the Kohler apparatus. This is important. for it means that even small differences in conductivity between gel and sol had they existed would have been measurable on such a bridge even if they would not have been detected on the ordinary metre wire. All resistances bridge wires thermometers etc.were carefully calibrated. The thermostat temperatures were constant within O - O l O . The special conductivity cells of borosilicate glass have already been described (see McBain Laing and Titley T. 1919, 115 1279). The cell constants a t Oo were 13.06 and 0.1664 and a t ZOO' 13.20 and 0-1690 (co'mpare McBain and Coleman Trans. Faraday SOC. 1917 15 36 who obtained a similar variation with temperature when using Kohlrausch's standard values). Preparation of the Soaps. Apart from more than a year's delay in securing the necessary oleic acid there was unexpected difficulty awing to the lack o SODIUM OLEATE SOLUTIONS ETC. 1509 purity of the materials ultimately obtained. Pending the routine tests soaps were prepared immediately on the arrival of various specimens which had proved satisfactory in pre-war work.Table I contains some of the iodine values and molecular weights as obtained by titration of specimens supposed to be pure. Speci-mens marked ‘ ( K ” were palmitic and oleic acid “Kahlbaum.” The iodine value was determined by the Wijs method for which pure palmitic and oleic acids should give the values 0 and 90 respectively. TABLE I. Tests of Fnttp -4cids. Specimen. Palmitic ‘‘ K ” pre-war ............ Palmitic “ K ” post-war ......... British ost-war ..................... Oleic “ K ” pre-war sealed.. ....... British post-war sealed ............ Oleic f; ’’ pre-war bottled ...... Oleic “ K ” post-war sealed ...... Iodine value. 1.0 3.0 1Q.O 66.6 90.14 90-14 88.9 Mol.wt. & Titration. Theory. 256.1 256.1 256.9 256.1 262-8 256.1 299.5 282.2 - 282.2 385.5 282.2 286.5 282.2 The soap solutions were prepared from oleic acid and sodium drippings free from carbon dioxide in stoppered bottles of Jena glass in the manner previously described. The hydroxide was standardised against hydrochloric acid (method of Hulett and Bonner J . Amer. Chem. SOC. 1907 31 390) as well as against pure oxalic acid. Concentrations are invariably expressed in weight-normality gram-molecules per 1000 grams of water. In all four series of measurements were carried out with sodium oleate. In series A solutions of good sodium oleate were employed, although they had been kept since the previous communication (T.1919 115 1279). Although the actual data are not here recorded they in themselves sufficed to establish the main conclusions now confirmed and extended. Series B was carried out using the oleic acid of the fourth hori-zontal row of table I. Evidently it had become partly oxi+sed by exposure to air as shown by the iodine and titration values. Here again sol and gel are identical and the values for curd are of interest as showing the high conductivity of a curd which was not more than a few hours old. Series C made with the oleic acid of the sixth horizontal row in table I contained 1.5 per cent. excess of alkali. All the data obtained in this the principal series have been corrected by the 3.8 per cent. predicted from the measurements of McBain and Taylor (Zeitsclt.phtysikal. Chem. 1911 76 178) and it will b 1510 LAINCr AND McBAIN THE INVESTIGATION OF noted that Lh0y agree within a fraction of 1 per cent. with the three measurements already published. Series L) was carried out wit'h the same speciinen of oleic acid, using neutral solutions prepared by taking the experimentally determined quantities of acid required as recorded in table I. Density Meusure me n t s . The density was found for each concentration a t these low temperatures with a 10 C.C. pyknometer. The densities found give straight-line graphs when plotted against concentration the values, except for 0*2N above 18O being only slightly greater than unity. TABLE 11. Densities of Sodium Oleate Solutions. Con centration. . Temperature. 0*2N,,.0.4N,. 0.4N,,,i 0" 1.0012 1-0040 1.00GS 10 1.0005 1.0029 1.0062 18 0.9997 1.0020 1.00 5 0 (Values above 18" extrapolated.) Method of obtaining t h e Curd or Gel. There are definite temperature limits to the existence of each of the states gel and curd. Thus a 0.4N-gel could not be obtained a t Oo or the 0.4N-curd above 2 3 O ; 0.6N-gel and sol were not obtained below Go or the O.6N-curd above 2 5 O . Both Oa4N- and O*GN-gel rapidly liquefied at' 2 5 O . The gel was obtained in almost every case by very slow warming of the curd which gradually lost its opacity first becoming cloudy and then turning to a clear gel of definite shape. I n the case of 0.2N- and O*4N-soap the gel was always surrounded by a film of sol a t least 2 mm. thick in the centre of the cell and for 4-6 mm.above each electrode. The gel column is firmly attached to each electrode and the sol surrounds it. With rising temperature the gel shrinks more and more remaining attached to the electrodes a t each end until there is only a fine diaphanous thread connect-ing the two small masses of gel. The O*GN-solution was always obtained as either sol or gel not having a boundary line as in the other cases mentioned. With this concentration alone a gel was obtained by cooling the sol. With slow cooling the whole viscous sol set to a gel which curded very gradually (that is after two days). The curd formed in each iimb spread towards the centre in long silky fibres which graduall SODIUM OIiEATE SOLUTIONS ETC. 1511 became more and more opaque until the whole mass was white curd with individual fibres no longer distinguishable.The 0-2N- and Om4N-sols when cooled developed large white nuclei in isolated spots and these nuclei grew towards each other by thickening filaments. Thus a clear gel was not obtained in these cases from the cooling side. Whenever the nuclei appeared, the conductivity always dropped appreciably but when a8 in the case of 0-6N-gel the fine fibres appeared it was not until the whole was cloudy that the lowering of conductivity became appreci-able. When the nuclei formed the soap was evidently being removed from the sol in considerable quantities whereas for a few T h e Conductivity Data. TABLE 111. Molecirlar Conductivity of Sodium Oleate. (Series B alkaline see p. 1509.) 0.2.N.0.4N. Tempera- < wk-ture. Sol. Gel. Ckrd. Sol. Gel. Curd.* o*oo 13.42 13-38 5.88 14.16 - 7.19 5.5 14.63 14.82 6.97 16.77 16.78 8.97 11.0 17-54 17.33 9.55 19-70 20.04 10-43 16.0 2 1.25 21-25 - 24.48 24-43 21.32 - - - 26.02 26.04 - 20.0 22.0 - - - 27.00 27.00 -* Fresh curd not more than a few hours old in each case; often a day elapsed between measurement of gel and sol at any one temperature but the order in which these were measured did not affect the result. TABLE IV. Molecular Conductivity of Sodifurn Oleate Solutions. (Series C corrected for alkalinity.) Tempera-ture. 0.0" 5.0 10.0 16.0 18.0 22.0 25.0 7 Sol. 13.94 16.22 19.13 20.95 22.62 25.87 -0*4N,,,. Gel. 13.94 16.22 19-13 20.94 22.62 25.81 1 -0*6N,.- A t- I Curd.* Sol. (3.01. Curd. * 3253 - - 2.547 4-248 18-10 - 3.046 6.015 16.95 16.94 4.341 - 20.37 20.33 6.393 - 21-65 21-65 12.96 - 22.64 22-64 - 25.97 25.97 -* The curds were kept four days at o" and during two subsequent days were gradually warmed t o the ordinary temperature. The temperature was kept constant at each intermediate temperature before the conductivity was noted. Sometimes a week elapsed between the measurements of gel and sol without affecting the constant value obtained before and after such treatment 1612 LAMG AND McBAIN THE INVESTIGATION O F TABLE V. Comparison of Specific Conductivities of Curds and Gels. (Series C.) Curds. Gel (or sol.). Tempera- A- - ture. 04V. 0.6N. 0 4 N . 0.W. 0" 0.01168 0*001304 - -6 0.001 119 0.001555 0.004983 0.007729 10 0.002150 0*002213 0.005805 0.008632 15 0.003169 0.003255 0.00682 1 0.01038 18 0.004529 0.006594 0-007494 0.01101 25 - - 0.009196 0*01320 20 0.005597 - o.on8065 0.01 161 TABLE VI.Specific and Molecular Conductivity of Cwrds of Sodium Oleate at Oo. (Series D neutral.) Days. 1 2 3 7 14 21 28 40 56 05 0.2N. -, k. P* 0'0009661 4.860 0-0008908 4.478 0.0008672 4.362 0-0008339 4.195 0.0008106 4.078 0.0008066 4.057 0.0008028 4.038 0 0008138 (4.094) I - - -0.m. - k. 0.001 171 0.001 137 0.001 114 0.001072 0*001014 0.0009858 0.0009734 0.0039580 0.000941 7 0*0009328 - P. 3.268 3.176 3.111 2.994 2.832 2,754 2.719 2.676 2.630 2.609 0.6N & 0.001162 2.273 0.001 140 2-229 Om0O1135 2.219 0.001126 2.203 0*001114 2.178 0*001112 2.174 k.P* - -- -- -- -TABLE VII. Specific and Molecular Conductivity of ,4ged Curds * of Sodium Oleate from Oo' t o 20° (Series D neutral.) Tempera- - * - ture. P* k. P- k. P-0*2N. 0-41y. 0-6N. 0' 0.0008028 4.038 0.009328 2.609 0.001112 2.174 6 0.0009666 4-867 0*001160 3.243 0.001277 2.479 10 0.001190 6.994 0.00170 4.752 0.00179 3.506 16 0.001930 9.735 0.00282 7-893 0.003146 6.167 18 0.002720 13-11 0.00414 11.32 0.004700 9.217 20 0.003572 18-01 0-005501 15.40 0.006577 12.90 * The ciirds in the thermostat after completion of table VI at O" were allowed to attain the ordinary temperature while these measurements were being taken ; this occupied four more days.Figures 1 and 2 present the conductivity data and show a t a glance the effect of time of temperature and of concentration besides showing the relative magnitudes of the specific conductivities oL sols gels and of aged curds SODIUM OLEATE SOLUTIONS ETC. 1513 faint fibres only very small amounts were being removed. In every case where the gel exhibited a conductivity slightly less than that of the sol this could he directly traced to the presence of a few curd fibres. The sol was always obtained by warming the gel or curd to 20° or 30°. Analysis Refractive Index and Vapour Pressure of Sols and Gels. It was first established that the refractive index of 0-5N-sodium oleate as gel is identical to eight significant figures with that of the same solution as sol.A Zeiss interferometer with 3-em. cell was employed. Into one side was placed some of the sol and, after it had been caused to gelatinise more of the same sol was poured into the other side for direct comparison without causing alteration of more than one division from the position of the zero reading corresponding with a difference in concentration of 10-6A7. This confirms the result of the conductivity measurements and also of Salmon's measurements of sodion concentration using the sodium electrode (Zoc. cit.). A similar result has been obtained for gel and sol of gelatin by Walpole. The identity in properties of sol and gel is shown also by the following measurements of vapour tension," using the dew-point method of McBain and Salmon (Proc.Roy. Soc. 1920 [ A ] 97, 44; J . Amer. Chem. Soc. 1920 42 426). Dew-point Lowering of Sodium Oleate at 18O. Concentration. Gel. 0.6N 0.06" 0.LV 0.04 Sol. 0-06" 0.04 Experiments were next carried out to test whether appreciable segregation of soap occurs during gelatinisation. A 0.4N-curd was introduced into a clean dry test-tube and very gently warmed. The curd very gradually became a column of gel with sol surrounding it. After some time the contents were poured through silver gauze into a dry weighed flask the sol passing through the gel remaining on the gauze. Some of this gel was removed into a second tared vessel and both weighed speci-mens were carefully tested both for sodium and for oleic acid (or oleate) by titration with about iV/4-solutions of aqueous and * Vapour pressure is used as a measure of the concentration of crystalloidal constituents present; the vapour pressure of gels and sols 'containing no crystalloidal material does not possess any such significance since it ie necessarily the 8am0 as that of the solvent itself.3 K 1514 LAING AND McBAIN THE INVESTIGATION OF alcoholic sodium hydroxide standardised against the oleic acid used (molecular weight 286.5 instead of the theoretical 282.2). Two such experiments are recorded below the sodium and oleic acid content being expressed in C.C. of O.2N-acid. Soap Difference Weight Oleic calculated sol/gel taken. Type. Sodium. acid. per 100 grams. per cent. 6.4344 sol 9.43 4.081 4 sol 7-11 7-00 34-56 0-9 0.08 10.55 10.48 33.8 9.42 34.7 1.6.2272 gel 2. 2.2807 gel 3.94 3.93 34.53 It is evident from the above table that the sol and gel have nearly the same composition. Even after allowing considerable time for shrinking and possible syneresis the result expected would be similar to the above since the conductivity of an old gel is the, same as that pf freshly formed gel and as that of the sol into which it turns. A similar experiment was made with 0.5N-sodium oleate using the Zeiss interferometer with 5 cm. cell. The difference in con-centration bettween lumps of gel and the sol in which they were floating was not more than 0.003N. Perhaps even this small difference may have been due to segregation caused by previous curding which could not be equalised again without melting the gel for the purpose.From all the above it is evident that in complete contrast to formation of curd formation of gel takes place without appreciable segregation of the soap. The refractive index of a soap solution is an accurate measure of the total concentration of soap whether as sol or gel. Discztssio?L Theory of Gels. It would appear from the present work that a gel is identical with a sol except for its mechanical properties.* Osmotic activiey, electromotive force and conductivity alike prove that the chemical equilibria are identical in gel and sol. Furthermore since the conductivity of concentrated gel and sol is thus identical the hypothesis of a closeld cellular spongy or honeycomb structure or other similar structure is disproved and even a similar structure with partly open pores is rendered! extremely unlikely [see (1) and (2) below].* Previous communications have indicated that soaps are to be classed with proteins gelatin salts dyes etc, m colloidal electrolytes. The similarity in the behaviour of their sols is so definite that m y general theory must be consistent with the facts of all these cases. Hence the significance of the identity of soap sol and gel must extend to these other systems SODIUM OLEATE SOLUTIONS ETC. 1515 A difficulty in reviewing the various opinions held with regard to the structure of gels is that a definition of the word phase is primarily involved. The old diversity still subsists as to whether typical colloidal sols and gels are to be called one-phase or two-phase systems.Thus whereas most authors are agreed that sols and gels are to be regarded as heterogeneous systems a few such as Proctor Hayes and Brailsford Robertson prefer to call them one-phase systems even where they postulate a molecular network throughout the gel. It was arguable to regard sols and gels of protein salts as one-phase systems containing only molecularly dis-persed matter until the facts with regard to soap sols and gels had been established. I n the case of soap solutions the formula and molecular weight of the chemical substance concerned are known, and it is impossible to ascribe the colloidal properties to the presence of unknown molecules of enormous molecular weight. So long as the soap is in molecular dispersion it behaves as an ordinaq crystalloid.Apart from the above conception of a moleoular network there are four suggestions which have been advanced to account for the properties of gels namely : (1) Closed honeycomb solid-liquid ; Freundlich. (2) The porous but continuous solid cellular framework ; (3) The emulsion liquid-liquid ; Wo. Ostwald. (4) The micellar theory of Nageli Pauli Zsigmondy and the present authors. Various authors whilst adhering to one or other of the above conceptions have argued that the phenomenon of gelatinisation is a process of crystallisation. The evidence here presented is how-ever incompatible with this suggestion if crystallisation is taken to be removal of substance from the colloidal solution (see for instance Levites 1908 ; Zsigmondy and Bachmann 1912 etc.; Bradford Biochem. J. 1913 12 351; 1920 14 91; Bogue 1920; Lloyd Biochem. J . 1920 14 147). Thus coagulation and form-ation of curd are differentiated from gelatinisation. The present data give but little information with regard to the question as to whether the individual colloidal particles of soap gels and sols are to be regarded as (‘ crystalline” or amorphous but the fact remains that in the process of gelatinisation they are not sufficiently removed from solution to disturb the equilibria into which they enter. Since they are identical in sol and gel and since they exhibit properties of aggregation or orientation it is more difficult to conceive of them as liquids. These terms however lose much of their meaning when applied to particles of ultramicroscopic and Hardy and Lloyd.3 K’ 1516 LAINU AND MCBAIN THE~INVESTIGATION OF amicroscopic dimensions such as exist in these typical “ emulsoid ” systems. As a result of the optical observations of Biitschli (1892-1900) and of Hardy (1899) and Quincke an emulsion theory of gels, whether (1) or (2) or (3) above was regarded as finally established (see for instance Freundlich “ Kapillarchemie,” 1909 p. 475). Most authors considered that the emulsion had solidified to a cellular or honeycomb structure a conception which is irrecon-cilable with the definite experimental evidence here presented. Wo. Ostwald (1905) regarded the emulsion as still consisting of two liquids and several authors have suggested that gelatinisation consisted of the inversion of these two liquids the dispersed phase becoming in its turn the dispersion medium.This again is clearly contradicted by the experimental data for soap gels since gelatinisation leaves conductivity quite unaffected. Further, Hatschek (Trans. Faraday SOC. 1916 12 17) showed that the viscosity of a gel does not exhibit the behaviour characteristic of an emulsion of two liquids. The very earliest explanation of gels was the micellar theory of Frankenheim (1835) and Nageli (1858). They considered that gels owe their characteristic propertiw to a loose network or aggrega-tion of ultramicroscopic or amicroscopic solid particles. Zsigmondy and Bachmann resuscitated this theory producing strong experi-msntal evidence in its favour as was pointed out by Pauli a t the time.They showed that the heterogeneity of gels is of a different order of magnitude from that assumed by Biitschli Hardy etc. whose observations were made before the invention of the ultramicro-scope. The heterogeneity is amicroscopic instead of microscopic, involving distances of less than one hundredth of a micron. They emphasised the very important point that in all the positive ex-periments of Butschli and of Hardy some procedure was adopted, or chemical added to produce the structure observed in the micro-scope as had indeed been pointed out by Pauli in 1902. The micellar theory is in harmony with the phenomena of syneresis double refraction swelling peptisation the definite form and elasticity of gels coagulation dehydration and vapour-pressure curves pleocroism optical and ultramicroscopic pheno-mena the behaviour of protected colloids such as gold particles, which retain their identity even after repeated transformations, such as gelatinisation in other words all the characteristic proper-ties of gels.All that is necessary is to assume that the particles become stuck together or oriented into loose aggregates which may be chance granules or more probably threads. This hypothesis i SODIUM OLEATE SOLUTIONS ETC. 1517 supported by many other phenomena such as those observed by Garrett (Phil. Mag. 1903 [vi] 6 374) Shreve (Science 1918, 48 324) Holmes Kaufmann and Nicholas ( J . Amer. Chem. SOC., 1919 41 1329) Shoji (Riochem. J. 1919 13 227) Lloyd (ibid., 1920 14 147) Bogue (Chem. Met. Eng. 1920 23 62) and Lifschitz and Brandt (Rolloid Zeitsch.1918 22 133). The conception of micellar orientation to which the present evidence so directly leads is supported by many other facts such as the following. Thus on heating a gel molecular movement becomes so intensified that the forces holding the particles are overcome and we have the familiar phenomenon of the ‘‘ melting ” of the gel. Again nitrecotton which is completely gelatinised in a minute or so by a urethane is converted into a sol in a period ranging from a week down to a few minutes depending entirely on the degree of mechanical stirring employed which is in accord-ance with our theory that a sol is formed by mechanically breaking the orienting bonds between the particles. The third significant fact for the special case of soap is that we have never observed Brownian movement under the ultramicroscope in a soap which we had reason to believe was in the form of gel.Bachmann (Zeitsch. anorg. Chem. 1912 73 125) has made similar observa-tions with gels of gelatin. This conception explains moreover the fact that the apparent viscosity of a sol frequently depends on its previous treatment and! history. Many sols must be in process of forming such orienta-tions between their particles and hence the viscosity must depend on how far this has gone. Thus for instance a sol of boiled starch, the delicate incipient structure of which had been destroyed by previous shearing exhibited thereafter much lower viscosity (Hatschek Kolloid Zeitsch. 1913 13 881). Again it is quite char why supersaturation and hysteresis with regard to gelatin-isation so frequently occur.Once more this conception links up with the observed facts of liquid crystals. It would seem that there must be some connexion between the orienting forces observed in crystalline liquids and in gels. Indeed one of the typical soaps ammonium oleate is a well-known liquid crystal. Reference may also be made to Sand-qvist ’s b ram op h en an t hr en esul phonic acid which is obviously a1 so a colloidal electrolyte the solutions of which however can be obtained under certain conditions as crystalline liquids. Vorlander has observed that long molecules are required for forming liauid crystals which agrees with Bose’s theory of swarms of long molecules. The same appears to be true of colloidal electro-lytes such as sodium palmitate or hexadecylsulphonio acid.1618 LAING AND McBAIN THE INVESTIGATION OF decoate in concentrated solutions is a colloidal electrolyte; on the other hand sodium and naphthalene-a- and -P-sulphonates and sodium naphthionate with an equal number of carbon atoms, contain only a small amount of colloid if any. The tendency of soaps to form long strings of inolecules* or of colloidal particles is demonstrated by the lung ultramicroscopic fibres which are the characteristic feature of curds (and possibly gels) of sodium soaps (observations of Darke McBain and Salmon). An exceedingly fine filamentous structure may account for the elasticity of gels f and also for the fact that they exhibit more or less clearly oriented properties such for instance as the lenti-cular (that is not isotropic) form of bubbles generated within gels, as described by Hatschek.Freundlich has published an interest-ing study of vanadium pentoxide sols which had been aged many years and in which he found that a t the boundaries or through-out the sol when the sol was set in movement all the anisotropic and other characteristics of a crystalline liquid were exhibited. The theory of gels here deduced also leads to a prediction of the characteristic phenomenon of syneresis. Thus if there is an orienting force between the particles there must necessarily be in that force a component of attraction and hence the gel structure of oriented particles must exhibit a distinct tendency to shrink.Even i f tbis attractive force is only feeble i t must in course of time produce syneresis since in dilute gels itl is opposed only by the viscosity of a fluid. The swelling of gelatin salts is not in conflict with this view because the ionic inicelle of gelatin and proteins unlike that of soap does not become crystalloidal in dilute solution and so continues to be retained within the gel. The question now arises as to what are the colloidal particles which are linked together in the case of soap to form the gel structure. There are but two possibilities neutral colloid and ionic micelle. It is uncertain as to how much of the neutral colloid is included in the ionic micelle and it is just possible, * It might explain the fact that true reversible equilibria obtain in soap solutions if each of the true colloidal particles of soap were essentially composed of strings or even sheets of molecules thus harmonising colloidal behaviour with molecular reactivity and also resolving the problem of the number of phases present in a gel.The mechanical properties of the sols would however require further consideration. Indeed the stable existence of any colloidal aggregate has not yet been explained. t Each of the innumerable threads consisting of colloidal particles stuck together could exhibit mechanical elasticity in itself. Owing to the arnicro-scopic degree of dispersion there would be such great frictional resistance t o their displacement in the liquid that this property of elasticity would be transmitted t o the mass of gel itself for a temporary period; the time of relaxation would be dependent upon the viscosity of the intramioellar liquid SODIUM OLEATE SOLUTIONS ETC.1519 although unlikely that all of i t is so included and that therefore these conducting particles are those the orientation of which gives to the gel its structure. Were this possibility to be entertained it would have to be assumed that the gel structure as a whole would carry half the current the corresponding sodium ions being merely interspersed in the intramicellary liquid of the gel. Much more probably the neutral colloidal particles are thus linked up. An important fact t o remember in the case of soap gels and one that distinguishes them from salt solutions which have gelatinised with gelatin or agar-agar is that here more than half the current is carried by the ionic micelle instead of by ordinary ions.Thus these colloidal particles under the influence of electrical as distinguished from mechanical forces must pass as freely through the open network of the gel as they do through the sol. This is quite consonant with the fact that the neutral colloidal particles in soap sol and gel are identical in nature and amount. It is also supported by the observation of Salmon (Zoc. c i t . ) that t,he diffusion potential of soap against a solution of chloride is the same for a gel as for the sol which indicates that t,he diffusibility of both sodium ion and ionic niicelle are unaffected by gelatinisation. The Theory and Structure of Soap Curds. Whereas sol and gel are identical save for some coherence or orientation of the colloidal particles curds as will be shown are the result of an actual removal of soap from the solution in the form of white fibres the product of a process clcsely related t o crystallisation.Figs. 1 and 2 show that the specific conductivity of gel or sol is nearly proportional to its concentration and that that of the corresponding curd is much smaller and that in the curd it is comparatively independent of the concentration a t any one temperature. The curd fibres therefore enmesh a solution much more dilute than the original sol or gel of concentration which is roughly fixed for any onei temperature. Were curd fibres a true crystalline phase of constant composi-tion it would follow that the solution they enmssh would exhibit a definite saturation concentration for each temperature and this concentration would be the definite solubility of such curd fibres a t that temperature.The specific conductivity of a curd a t any temperature would then be nearly independent of the total original concentration of that curd although only approximately so on account of the mechanical diminution of the cross-section o 1520 LAING AND MCBAIN THE INVESTIGATION OF 0.0015 0*0010 0.0006 o*ooo the electrolyte through the presence of masses of curd fibres in concentrated curds. One would expect for this reason that the conductivity of a concentrated curd would be low as compared with that of a dilute curd. The contrary is however the case. The conductivity of a 0-6N-curd is always distinctly greater than that of a 0-2N-curd and it is sometimes nearly twice as great.Again for a true crystalline phase of constant composition the * - 6.6N. - 0,m a L Q c a * * b C U R 0 5 0.1 rJ -. . . . -3 0.0040 Q 8 0.0030 B 0~0020 " w2 solubility would be independent of time whereas for curd fibres this is not so. The curd falls off rapidly in conductivity during the first twelve hours and the conductivity continues to diminish slowly for some months. As shown in table VI and Fig. 2 the conductivity of 0.2N-curd fell steadily for fourteen days then remained nearly constant up to forty days a t Oo. The 0-6N-curd was altering so slowly that the experiment was discontinued a SODIUM OLEATE SOLUTIONS -" ETC. 1521 twenty-one days but the later experiment with 0.4N-curd where the conductivity was still falling a t sixty-three days shows that changes must extend over a period of months a t least.Hence although formation of curd means the separation of neutral soap from solution in the form of curd fibres and is analogous to crystallisation these fibres constitute a variable phase. For the present case the phase rule assumes the form f'+ F = C + 3 where the extra degree of freedom is the diameter of the component particles of the curd fibre or the diameter of the fibres themselves i f each of these is one solid whole. There are two possible factors of variability namely the FIG. 2 1.6 '"1 I t 0.2 6" lo0 160 2b" 2iY Temperature. Specifl conductivity o j soh gels and curds of sodium oleate at variow c temperatures.diameter of the fibres and their degree of hydration. The first factor ceases to cause appreciable change as soon as the fibres have become microscopic. In any case it can scarcely be the chief factor here concerned for the specific conductivities of fresh 0.2 and O-4N-curds differ too greatly whilst those of 0-4 and 0.6N-curds which a t first nearly coincide diverge on keeping although by hypothesis the mother liquor would be identical in both cases. The general nature of the facts in tables I11 to VII is the same, whether fresh or aged curds are considered whether equilibrium is approached from one side or the other and whether much or but little time has been taken in so1 approaching equilibrium 1522 LAING AND McBAIN THE INVESTIGATION OF It would appear that the important factor involved is degree of hydration and that the more heavily hydrated fibres are the more insoluble and1 the more stable.Taking the last point first it follows from the phase rule that for any definite diameter of fibre (or of component particles) the most stable form is that which is least soluble and in fact the conductivity does diminish with time. As regards degree of hydration a series of investigations in this laboratory using several methods one of which has been published (McBain and Taylor T. 1919 115 1300) has shown that hydra-tion of curd fibres is greatly influenced by the vapour pressure of the mother liquor. Thus in the presence of 3'ON-sodium hydr-oxide the composition of the palmitate fibres was NaP,3*2H20, whilst in the presence of 1-5~~-sodium hydroxide it was NaP,6*5H20.According to this therefore the curd fibres which first separate out from O'GN-oleate sol or gel are less hydrated and also more soluble than fibres which separate from 0.2N-oleate. Such fibres are moreover unstable in that greater hydration is required to correspond with the residual diluted mother liquor with which they are finally in contact. It should be mentioned that concentration, not temperature as such is the chief factor in determining the degree of hydration. Equilibrium in a soap curd can therefore only be attained by a process of recrystallisation with formation of more heavily hydrated fibres a process which requires much time since convec-tion or stirring is impossible and the diffusibility of soap is low (Salmon Zoc.c i t . ) . It seems certain however that the concen-tration of the mother liquor enmeshed in a soap curd kept a t constant temperature would ultimately reach a value independent of its original concentration or previous history. This view receives further support from a comparison of tables I11 or V with VII which show that an aged 0-6N-curd exhibits a lower con-ductivity than a moderately fresh 0-4N-curd ; similarly with 0.4N-and 0.2N-curds. Further a curd which has been well aged at Oo possesses on subsequent keeping at any higher temperature a conductivity which shows no tendency to rise. Discussing conductivity as distinguished from concentration it is a matter for some surprise that the heavy mass of fibrm in a 0-6N-curd does not cause such mechanical obstruction to the passage of the current through reduction of the effective cross-section of the liquid as to reduce the specific conductivity to a much lower value than that of a 0*2N-solution in which only about one-quarter as much fibre is present.The apparent absenc SODIUM OLEATE SOLUTIONS ETC. 1623 of any such effect might possibly be due to electric endosmosis; the study of the problem is being continued with the object of elucidating the phenomena which accompany the passage of current in gels and sols as well as curds. The conductivity data here given suffice to show that if the two effects suggested are appreciable they nearly counterbalance each other. The conductivity curves in Fig.2 show that with rising tempera-ture the solubility of the curd fibres increases and the conductivity rapidly rises towards that of the sol or gel. The conductivity of curd sol and gel will become identical a t the point where the solubility of the soap fibres becomes equal to the gross concentra-tion of soap present where the last fibre just dissolves or where the first fibre just appears. Such a solubility curve is indicated by the line of dashes in Fig. 2 although its exact position varies somewhat with the previous history of the curd fibres present. The rapid increase of the solubility of the curd fibres with rise of temperature is in accordance with that required by thermo-dynamics as deduced from the heat of solidification.The heat evolved on solidification in the case of a soap solution which has been cooled well belojv the temperature of initial solidification results in a noticeable rise of temperature. This conception of the solidification of soap solution as being essentially of the nature of a crystallisation process with separ-ation of curd fibres also explains the observation of McBain and Martin with regard to the increased alkalinity of curds a t low temperaturw as compared with sols a t somewhat higher tempera-tures. They found that the hydrolytic alkalinity of the sol increases within certain limits with decrease of concentration or with rise of temperature. That it is greater when curd has formed is now evidently due to the fact that the residual solution is more dilute and hence more hydrolysed.It is thus not necessary to conclude that the fibres in soap curd are anything but neutral (hydrated) soap whereas previously it might have been suspected that they were very slightly acid soap (see also the analyses below). The behaviour of the sodium oleate systems here investigated is all in agreement with the incidental observations of McBain, Cornish and Bowden on sodium rnyristate (T. 1912 181 2053), only they like all authors hereto confused gels and curds and they did not allow sufficient time in the measurements of curds. They concluded that the formation of a soap curd is not a process continuous with the adjustment of the degree of dispersion of a colloid in the sol since in a sol supersaturated with respect to formation of curd true reversible equilibrium subsists 1524 LAING AND McBAIN TRE INVESTIGATION OF Composition and Solubility of Curd Fibres.In order to obtain a direct experimental control of the deduc-tions made on the preceding section curds and qother liquor were analysed. From curds of O'6N-oleate as much as possible of the enmeshed mother liquor was removed by pressure and suction in an atmosphere free from carbon dioxide and the amounts of sodium and fatty acid radicle were determined. Three specimens thus examined all showed a slight alkalinity the proportion of sodium to fatty acid radicle being 100.4 100.9 100.4 equivalents of sodium to 100 equivalents of oleic radicle. This slight alkalinity may be attributed to oxidation of the oleate. It is thus established that the curd fibres consist essentially of (hydrated) neutral soap; the small amounts of acid soap which correspond with the hydrolytic alkalinity of all soap solutions are evidently completely submerged in these high concentrations of these neutral soap fibres.N o data previously existed with regard to this point, Tor although many authors had analysed the sediments and suspensions from various soap solutions since these always referred to low concentrations of solid they. found acid soaps of varying composition. Analysesi made of the mother liquor of 0-6N-curd a t various temperatures also showed equivalence of the sodium and oleic radicle within a small fraction of 1 per cent. Each of these determinations was carried out in duplicate involvinq four in-dependent analyses for each case.The resulting solubilities thus determined are for 0*590N-curd of sodium oleate about sixteen hours old 0-391N a t 1 8 O and 0-261N a t loo and for 0.1905N-sodium oleate and of about the same age a t 18O 0.0998N. The solubility obtained from a 0-590N-curd after three days a t Oo was 0*114N. These analyses thus substantiate the conclusions of the preceding section and the lack of quantitative agreement with the con-ductivity measurements must partly a t least be attributed to the difference in the age and previous history of the rmpective speci-mens. Further measurements will be undertaken with well-aged curds the conductivity of which will be taken just before analysis. Preliminary Experiments on the Behauiour of Mixed Cwrds of Oleate and Palmitate.In view of the fact that all commercial soaps are made from mixtures of fatty acids it is of importance to. inquire as to whether these different sodium salts separate out independently or whether the individual fibrw are each composed of mixture of soaps. I SODIUM OLEATE SOLUTIONS ETC. 1525 will be seen that although to a large extent the oleate and palmitate remain independent the separation is nevertheless in-complete. Sodium palmitate was chosen because its molecular weight and especially its iodine value differ from those of the oleate and because its solubility is less than N / 4 0 a t 25-30', temperatures a t which sodium oleate does not exist in the form of curd. Equivalent quantities of solutions of the oleate and palmitate were melted together to form a clear sol and then allowed to cool until curd fibres separated out.In the first experiment the snmeshed liquor was withdrawn at 3 5 O by means of a pipette to a tiny piece of filter paper which withheld the soap and allowed the soap solution to pass. The liquor was then weighed for analysis. In the second experiment the mixed soap curd was transferred a t 30° to a vacuum filter and the enmeshed liquor drawn through. Air passing through the soap caused frothing of the withdrawn liquor. This was condensed by warming the flask and weighed f or analysis. In the third experiment the treatment was similar to that in the second except that air deprived of ,carbon dioxide was drawn through the soap the temperature this time being 2 5 O (but see below).In the second and third cases comparative experiments were made with aqueous sodium palmitate of the same concentration as in the mixture to ascertain what concentration of the more insoluble sodium palmitate existed a t that temperature in the enmeshed liquor. The total concentration of soap in the filtrate was determined by decomposing and titrating with aqueous acid as well as with alcoholic hydroxide; the molecular weight of the fatty acid was deduced from its weight and the latter titration and finally the iodine value of the fatty acid was determined. TABLE IX. Analysis of Mother L i p o r in Mixed Curd of Sodium Oleate and Patmitate. Experi- Tempera- Mol. Iodine ment. ture. Concentration. Filtrate. wt.* va1ue.f I...3 6" 0.1680N each 0.1732N 286.9 82.2 2... 30-35 0.1903N each 0.2303N 27993 71.8 2. .. 30 0.191 4N NaP 0.01 778N 256.3 -3. .. 25-17 0'2486N each (0'2152N) 282.7 81.2 3 . . . 25 0.2500N NaP 0.00819N - -* H P=25&1 by theory; H 01=282.3 by theory t H P= 0 0 by theory; H 01= 90.1 by thoory; see table I 1526 LAING AND MCBAIN THE INVESTIGATION 0%’ Table IX shows that most of the oleate remains in the mother liquor and that the palmitate mostly separates out as curd fibres. This is shown by the high iodine values of the mixed fatty acids in the filtrate namely 82 72 and 81 comparable with fairly pure oleic acid as was shown in table I. The separation is however not quite complete for there is some oleate in the curd fibres. This was shoiwn by a continuation of experiment (3) in which the residual curd was washed with pure water and then a sample tested for the molecular weight of the fatty acid in the curd fibres.The result was 263.1 instead of the theoretical value 256.1 for palmitic acid. After about half of the curd had been washed away a sample was taken for determin-ation of the iodine value of the fatty acids present. The result was 11.2 instead of the 3.0 for the original palmitic acid used. Hence the palmitate fibres contained 10-20 per cent. of oleate alsot although unfortunately at the end of this experiment the temperature was low enough for oleate itself to) curd. It is fortunate that the behaviour of a mixture is even so far additive, since in the corresponding sol the behaviour is much more com-plex.Further experiments a t exact temperatures are contem-plated. Characterisation of a Commercial Soap. It is now possible to attempt a definition of a commercial soap. A transparent soap is a gel. All other hard commercial soaps are gels containing a felt of curd fib&. The gelatinisation may be due to the dissolved soaps or it may be partly due to other gelatinking filling agents. Curd fibres enmeshing a sol instead of a gel would probably not have the physical properties required. Thus some of the soap is separated out as curd fibres whilst some is in solution. The latter is absolutely necessary if the soap is to exhibit detergent action at ordinary temperatures. A soap containing only palmitate and stearate would not be a detergent a t ordinary temperatures.The value of the presence of oleate lies not only in its solubility but in the persistence of its colloidal constituents in unusually dilute solutions. A good illustration of the dual qualities required of a com-mercial soap is to be found in the behaviour of remnants of used soaps which have largely lost their detergent properties through extraction of the more soluble constituents. Holde (Chem. Urnschaw F e t t . Ind. 1920 27 56; SeifeiLfabi-., 1920 40 113) in a recent summary of the facts with regard to residues of soap cakes considers the further factor which helps in the accumulation of insoluble matter in these residues namely SODIUM OLXATE SOLUTIONS ETC. 1527 the precipitation of calcium salts through reaction with the impuritie~ in the water.Further interesting corroboration of the views here put forward is observed in the formation of the feathery clusters which separate in a soft soap on keeping known as figging. Soft soaps made froin linseed oil (and therefore containing chiefly very soluble soaps if unsaturated fatty acids show no “fig”). A small proportion of tallow rich in the lass soluble stearate ensures figging. Another example is that mentioned by “ H.A.” (Seifensied. Zed. 1920 47 646). A sodium soap made from coconut oil alone is very hard and very crumbly. Substitution of potassium for some of the sodium together with the addition of some potassium chloride to develop formation of a colloid makes this soap more like ordinary soap sol that it can be stamped out into cakes.For completeness it should be mentioned that many commercial soaps are formed by the curding of a mixture of two liquids. When salt is added tol a soap solution the viscosity rises enormously until suddenly the solution breaks into two layers and the mixture becomes of manageable effective mobility in the soap pan. Thus normally in the process of soap boiling the soap is probably never allowed to be in one homogeneous solution. This statement is the subject of a further communication froin this laboratory to be made by Mr. Buriiett. Summary. (1) We have discovered t’hat a soap solution of one and the same concentration a t any definite temperature may be prepared in three characteristic states ; namely clear fluid sol transparent elastic gel and white1 opaque cnrd.The latter has oiten but erroneously been called a gel. (2) The sol and gel forms of a solution of sodium oleate are identical in osmotic activity concentration of sodium ions con-ductivity and refractive index ; this proves that identical chemical equilibria and constituents are present in the two cases. The sol and gel differ only through the mechanical rigidity and elasticity of the gel form. (3) The quantitative identity of conductivity in sol and gel is irreconcilable with all theories of gel structure hitherto advanced, with the exception of the micellar theory of Nageli which was resuscitated by Zsigmondy and Bachmann in 1912 and is strongly supported by many lines of evidence referred to in the presen 1528 THE INVESTIGATION OF SODIUM OLEATE SOLUTIONS ETC.paper. The colloidal particles in sol and gel are the same but whereas in the former they are independent in a fully formed gel they stick together probably t o form a filamentous structure. It is probably the particles of neutral soap and not of ionic micelle that exhibit this behaviour. (4) The formation of soap curd in clear contradistinction from gelatinisation is analogous to a process of crystallisation neutral soap separating from the solution in the form of curd fibres of microscopic or ultramicroscopic diameter. This is shown by the drop in conductivity and osmotic activity and confirmed by direct and indirect analysis in addition to observations with the ultra-microscope. Coagulation and crystallisation are thus sharply distinguished from gelatinisation. This we consider to be the chief theoretical result of the present paper. (5) The curd fibres consist of hydrated neutral soap the hydra-tion of which depends on their origin and previous history. Within corresponding limits their solubility is definite for each temperature. The so-called melting points of soap curds are the temperatures a t which the solubility curve rises to a value equal to that of the total concentration of the soap and a t which the last curd fibre just dissolves. (6) All the above results are of general applicability both on account of the detailed similarity of soaps to protein and gelatin salts etc. as well as on account of the precision of the methods available in the investigation of these simple systems. (7) I n a curd formed from a mixture of palmitic and oleic acids, the two1 soaps are largely independent. In conclusion our thanks are d m to the Colston Society of the University of Bristol and to the Research Fund of the Chemical Society for their generous grants which enabled this work to be carried out. THE UNIVERSITY, BRISTOL. [Received October 18th 1920.