STUDIES OF THE CONSTITUTION OF SOAP SOLUTIONS. 41’7XLV1.-Studies of the Constitution of Soap Solutions.The Electrical Conductivity o j Potassium Salts o jFatty Acids?By HUGH MILLS BUNBURY and HERBERT ERNEST MARTIN.THE preseni; communication is a study of the conductivity ofsolutions of potassium salts of the saturated fatty acids, fromstearic acid down to acetic acid. For most of these aubstances nodata are in existence, and yet they present special points ofinterest. A number of these will be discussed in future communi-cations dealing with the boiling points and the degree of hydrolysisof soap solutions.The electrical conductivity of sodium soaps has already beenshown to be unexpectedly great, and the conductivity curves areanomalous in that they pass through a minimum in N/5- toN/10-solutions and a maximum in N- to N/2-solutions.It might have been expected that the potassium salts (soft soaps)would have exhibited a different behaviour, since they are lessviscous a t room temperature, and more nearly resemble ordinaryelectrolytes.It will be seen in the sequel that the contrary is thecase..EXPERIMENTAL.Preparation of Solutions of Pure Potassium Hydroxide.Potassium in the form of the commercial balls is thrown intoordinary ether in a wide-mouthed, loosely-stoppered bottle. Thestormy effervescence serve6 to keep the potassium cool enough toavoid explosion, and a t the same time its surface becomes cleanand highly polished. These balls are now thrown into moltenparaffin in a nickel crucible, where they unite, since their surfacesare clean.The large piece obtained is allowed to solidify, andremoved just before the paraffin solidifies. After it has beenf,hrown into clean ether and shaken, it is rapidly wiped with filter-paper and placed on nickel gauze above a nickel crucible in adesiccator containing alkaline water. To avoid explosion, thenickel gauze must be cleaned and thoroughly dried each time. A tfimt the absorption of oxygen is more rapid than the evolutionof hydrogen, and the potassium ultimately gets deeply coveredwith a slowly deliquescing layer of potassium hydroxide, so that* Previous communications from this laboratory : Ber., 1910, 43, 321 ; ztihch.physika2. Chem., 1911, 76, 1 7 9 ; T., 1911, 99, 191; 1912, 101, 2042; Trans.Faraday SOL, 1913, 9, 99 ; KolZoid-Zeitsch., 1913, 12, 256418 BUNBURY AND MARTIN:the desiccator and contents must be left undisturbed in a fire-proofroom.To obtain pure hydroxide, the desiccator should be of the vacuumtype, with ground-in glass stopper, and tap sealed to a U-tubehalf full of a concentrated solution of potassium or sodium hydr-oxide.It should be emphasised here that it has been foundnecessary to discard many months of work done in this laboratorywhere in dealing with solutions of alkali, acid, or soap the precau-tion had not been rigorously taken of quite thoroughly smearingwith vaselin all ground-in connexions and glass stoppers (rubberis worthless as a protection against carbon dioxide).Preparation of Soap Solutions.The solutions were made up in silver tubes, with the precautionsdescribed in previous papers, from Kahlbaum’s best acids,* andthe calculated amounts of potassium hydroxide solution.Thedensity of the potassium hydroxide solution was taken to be thatgiven in Landolt-Bornstein’s (‘ Tabellen ’’ (1913) a t 15O.f Theconcentrations are expressed in weight normalities (gram-moleculesof salt per 1000 grams water), which is not affected by the largetemperature interval involved.The time allowed for equilibrium to be attained varied from afew hours, in the case of the dilute solutions, to a’ day or twofor the more concentrated solutions of the higher homologues. Thesilver tubes were then taken from the shaking machine (80°),opened, and the solution transferred whilst hot and homogeneousto a tube of Jena-glass into which was inserted the dippingelectrode, also made of Jena-glass after the pattern already* Of these acids NMcBain and Taylor (Zoc.c i t . ) had shown that the conductivitywas unaltered, in the case of the special palmitic acid, by further purification;decoic and hexoic acids were synthetic acids, and the acetic acid has been shown tobe very pure. Less weight can be laid on the values for the myristates, laurates,and octoates as these were only froni Kahlbaum’s ordinary acids.j- I n practice the following simple calculation gives the correct amount of waterrequired to make the solutions exact within 0.1 per cent. in all cases. The weight(in air) of acid taken is divided by its molecular weight (which is diminished by0-11 per cent.in order t o give the apparent weight in air) and is multiplied by1000 and divided by the weight-normality desired. This is the weight of watertheoretically required, but it has to be corrected for two errors, both of which areproportional to the concentration, namely, the water introduced with thehydroxide solution and the amount of water formed by the union of the hydroxideand acid, the latter being 1-80 per cent. of the total water i n a N-solution. Forthis purpose the number obtained is converted from grams to c.c., and from it issubtracted for potassium salts, 1-14 per cent. (for sodium salts, 2’04 per cent.miiltiplied by the weight-normality desired ; finally, this volume of water miniisthe volume of alkali used is added to the solutionSTUDIES OF THE CONSTITUTION OF SOAP SOLUTIONS.419described. The cell constant is therefore found to be independentof the rwistance of the solution, and there are other obviousadvantages. The previous work of McBain, Cornish, and Bowdenhad shown that this use of Jena-glass would be permissible forthese comparatively mobile and homogeneous solutions. Themeasurements were carried out a t 90*OOo (con.).The conductivity data appear to be more accurate than thoseof McBain, Cornish, and Bowden (Zoc. cit.). Electrode XV wasused throughout; it had a cell constant in glass of 4.070 (H. M. B.)and 4.080 (H. E. M. eight months later).Duplicate determinations of the densities of N / 2- and moreconcentrated solutions were made, agreeing within 0.06 per cent.,using the special form of pyknometer (capacity about 10 c.c.)previously described.Previous work in this laboratory and else-where has shown that no error exceeding 0.1 per cent. is madeif the densities of solutions of intermediate concentrations areobtained graphically by linear interpolation from the densities ofthe O.5N-solutions and that of water (D:! 0*9653), using the concen-trations (not the dilutions) as abscissz. The densities obtainedby interpolation are enclosed in brackets in the tables.Jn all the tables the first column gives the weight-normality;the second the number of grams of potassium salt to 100 grams ofwater ; the third the specific conductivities observed ; the fourththe average specific conductivity; the fifth the density; and thelast the equivalent conductivity a t 90°.The equivalent conduc-tivity is derived by multiplying the weight of solution containing1000 grams of water by the specific conductivity, and dividing bythe weight-normality and the density. In each cam all the correc-tions previously enumerated have been applied. The duplicatedeterminations were carried out on entirely independent solutions,as usual.Po t assium Pdmit ate.Particular attention wm given to these solutions, since on theprevious careful study of the possible sources of error in theinvestigation of the conductivity of soaps, using sodium palmitate(McBain and Taylor, 1911, Zoc. cit.), is based the validity of allthe subsequent conductivity results. The 1.5N-solution proved tooviscous for introduction into electrode XV, although the momdilute solutions, of which eight concentrations were studied, seemmuch more mobile than the corresponding sodium soaps.Thefollowing data were obtained at 90'OOo420 RUNBURY AND MARTIN :TABLE I.Conductivity of Potassium Palmitate (C,,) at 90*OOo.I.1.007-N0.764-N0.6016-N0.2003-N0- 1001-N0.0500-N0.02000-N0~01000-N11.29-6422.1914-776-8962.9461.4720-58880.2944111.0.09320.093330.076280.076350.053650.053640.020270.020280.0 10060.0 10070.0052680.0052720.0025560.0025560.0016510.001650IV.0.09330.07630.053650.020280.010070.0052700.0025560.00165 1V.0.96790.96720.9668(0.9659)(0.9656)(0.9654)(0.9654)(0.9653)VI.124.2127.9127.011 1.0107.0110.8133.2171-6The results given in table I are shown graphically in Figs. 1 and2, where the ordinates are equivalent conductivities and theabscissze the dilutions in litres.As in the case of the sodium palmitate, an anomalous conduc-tivity curve is produced.The greater apparent solubility ofpotassium soaps, and the fact that they do not, like sodium soaps,form gels with alcohol, might have pointed to a more normalbehaviour as compared with the sodium soaps; the contrary isseen to be the case.The conductivity curve exhibits a maximum at 0.75N- and aminimum at 0-lrV-potassium palmitate. Sodium palmitate exhibitsits maximum at 0.5N and its minimum at 0.2-0.1N; thus theposition of the depression in the conductivity curve is the samefor potassium palmitate as in the case of the four sodium soapsalready described.The anomaly is accentuated in the case ofthe potassium salt as compared with sodium palmitate, for thedifference between the maximum and minimum conductivitiesamounts to 20.9 mhos in the former case and to only 7.1 in thelatter.The solutions from N / 2 downwards are quite fluid and clear at90°, but the l*OiV-solution was rather viscous, and it coagulated a t60-70°. A t room temperature the 0-01N- and 0*02N-aolutionsexhibit a large amount of fine, shiny sediment a t the bottom of anearly clear, slightly opalescent aolution.The separation is muchless complete in the case of the 0~05N-solution, where silky fibres orstriae fill the liquid even on long keeping. With increaae ofconcentration the viscosity rapidly increases, and the pearly appear-ance diminishes. A 0.5N-solution has the appearance of boileSTUDIES OF THE CONSTITUTION OF' SOAP SOLUTIONS. 42190 -80 *70 -60 -40 -30 -20 -10 ..!!? -starch, but the 1*5N-solution is nearly colourlese, and it is veryviscous, scarcely flowing a t room temperature, bubbles caught init never rising to the surface.These descriptions should be compared with similar data inprevious papers and with those of Zsigmondy and Bachmann(Rolloid-Zeitsch., 1912, 11, 145).FIG. 1.c l l 1 1 I I IPo tassiurn A ce tu t e.I n order to have a standard of comparison, measurements weremade of potassium acetate, which is always considered to be atypical electrolyte.Since this salt is very hygroscopic, the solutionswere made up from solutions of pure acetic acid and potassiumhydroxide, as in the case of the soaps proper. The resdtingsolutions were neutral to litmus, and did not affect phenol-phthalein. The conductivity data are given in table 11422 BUNBURY AND MARTIN:FIG. 2.260240220200180160140I2010080604020I.1.000-N0.500-A'0.2000-N1 I I 1 1 I1 2 5 10 20 litresEquil:alcnt concluctizilics of potassium salts of fatty acids at PO".TABLE 11.Conductivity of Potassiwm Acetate at 90.00°.0~1000-N0.0500-N0.0 1000-N11.9.8124-9061.9620.9810.49060.0981111.0.16280.16210.092680-092510.042280.0423 10-022710.022740.012030.012010.0026100.0026 14IV.0.16250,092600.042300.022730.012020.0026 12V.1.00900.98680,97400.96970.96750.9657VI.176.9196.6221.2236.5249-5270.4In order to evaluate these data it is necessary to know theconductivity a t infinite dilution at 90° of each of the ions conSTUDIES OF THE CONSTITUTION OF SOAP SOLUTIONS. 423cerned.This was done by a consideration of the following dataof Kohlrausch and of Noyes (Report of the Curnegie Zmstitute ofTVashington, 1907, 63, 1):A t 18O, KC1 = 130.1 ; NaCl= 109.0 ; NaC2H302= 78.1 ; NaOH =216.5; K'=64-7; C1/=65.4; Na'=43.6; C2H30,/=34*5; OH'=172.9; that is, the mobilities of the five ions named are as1 : 1.011 : 0.674 : 0.533 : 2.673.At looo, KC1 = 414 ; NaC1= 362 ; NaC,H30, = 285 ; NaOH = 594.I f the fairly reasonable assumption be made that the potassiumand chlorine ions have attained equal mobilities at this hightemperature (the very pronounced tendency of all ions a t hightemperatures), then K' = C1' = 207 a t looo, and Na' = 155 ;C2H,0,'=130, and OH/=439; that is, the mobilities at looo areas 1 : 1 : 0.749 : 0.628 : 2-120.If t,he further assumption be now made that these ratios betweenthe mobilities of the ions at infinite dilution change linearly withthe temperature, the magnitude of the effect is small between loooand 90°, being 2 per cent.for the acetate and 1 per cent.for thesodium ion, and even for the hydroxyl ion only 3 per cent; thusthe mobilities of the ions KO, Cl', Na', CaH302/, OH' become as1 : 1.001 : 0.740 : 0-616 : 2.187 a t 90°.Finally, if the value for potassium chloride at infinite dilutionbe interpolated graphically on the very flat curve obtained byplotting the values already cited together with KC1=321*5 a t 75O,232.5 a t 50°, and 152.1 a t 25O, the conductivity of potassiumchloride a t infinite dilution is fouhd to be 376.5 at 90° (compareBohi, Diss., Zurich, 1908, p. 37). From this we obtain the follow-ing final values for 90°:K. = 188 KCl = 376C1' = 188 NaCl = 327NaC2H,0, = 255KC2H,0, = 304C.,H,O,' = 116 NaOH =550Na* =139=411 KOH =599 0-H' -Utilising these data, we are now in a position to calculate thedegree of dissociation of the sodium acetate and hydroxide deter-minations of McBain and Taylor, as well as of the above solutionsof potassium acetate.Data for a fewintermediate concentrations have been obtained by interpolationon the graph of the molar conductivity plotted against the cube-rootof the concentrations; this graph deviates from a straight line onlyin the most concentrated solutions.This is done in table 111424 BUNBURY AND MARTIN:TABLE 111.Electrolytic Dissociation a.t 90°.Potassium Sodium sodiumacetate. acetate.* hydroxide. * - - - a. a. aPer cent. P. Per cent. p. Per cent.( 6 3 ) 50.3 - - - -71.6 176.9 58.2 129.7 51.1 392.8183.9 60.5 138-6 54.6 -196.6 64.8 154.0 60.6 427.3 77.822 1-2 72.8 178.9 70.4 458.4 83.5236.5 77.7 195.0 76.6 476.9 86.9249.5 82.1 207.8 81.7 491.1 89-5262.6 86.4 221.0 86.9 - -270-4 89.0 228.0 89.6 614.3 93.7304.0 100.0 254-5 100.0 648.9 100.0-* Measured a t 89.75" ; volume-normality at 18".2.0-N1.0-N0.75-N0.6-N0-2-N0.1-N0.05-N0.02-N0.01-N0The values obtained for sodium acetate fit in well with thosedetermined by Noyes (Zoe.eit.) at other temperatures. It will benoted that the dissociation of this sodium salt is very distinctlyless than that usually found for binary univalent electrolytes,although according to Noyes the dissociation of ammonium andsodium acetate are the same. The results in table I11 will proveuseful in estimating the significance of the data obtained for thevarious soap solutions.Other Potassium Salts.I n view of the interesting character of the conductivity data forpotassium palmitate, it appeared desirable to extend the measure-ments to the potassium salts of the other saturated fatty acids ofeven number of carbon atoms in the molecule, these being theonly saturated fatty acids occurring in soaps.I n each case thenormal acid was the one investigated. The data are given intables IV to IX.TABLE IV.Conductivity of Potassium Stearate (CIJ at 90*OOo.I, 11. 111. IV. V. VI.1.007-N 32-46 0.08301 0.0830 (0.9637) 113.4(0.9641) 112-6 0.754-N 24.30 0.065990.06567 0'065830.9645 113.9(0.9648) 100.016.17 0.047436-468 0.018200.04735 0.047390.01807 0.018140.5016-N0.2003-NE:EEi:;f 0.008980 (0.9652) 96.00.004839 0.004829 0.004834 (0.9652) 101.70-1001-N 3.2260.05001 -N 1.612~ : ~ ~ ~ $ $ ! 0-002397 (0.9653) 124-9 0*02000-N 0-64600.0014 18 0.001426 0.001421 (0.9663) 147.7 0.0 1000-N 0.322STUDIES OF THE CONSTITUTION OF SOAP SOLUTIONS.425At 90° 1.0N-potassium stearate is a clear, colourless liquid, buthighly viscous-at room temperature the white paste resulting isscarcely more viscous; it will hardly pour under the influence ofgravity. A t 90° the O’lN-potassium stearate is clear, but the0*05N-solution is distinctly cloudy, and the 0-02N- and O.02N-solu-tions are faintly opalescent.A t room temperatures the solutions from N/10 upwards arewhite pastes; the more dilute solutions are white and translucent,exhibiting very little shiny sediment, with the exception of the0~05N-solution, where the precipitate occupies about one-half ofthe volume.The silky appearance is not nearly so marked as inthe case of the palmitates.I n view of the fact that Kahlenberg and Schreiner (Zeitsch.physikal. Chem., 1898, 27, 552) have measured one concentrationof potassium stearate solution a t temperatures from 40° up to80°, and further t h a t their data for sodium oleate are twice a53great ils those of Dennhardt (Diss., Erlangen, 1898), it is interest-ing to note that the data for their .iV/4-stearate solution whenextrapolated for 90° (105-106 mhos) agree with ours within afew per cent.TOLE V.COl LductivityI. IL1.007-N 26.810.60 16- N 13.360-2003-N 5.3400.1001 -N 2.6650.0600-N 1.332of Potassium Myristate (C14) at 90.009111. IV.V. VI.0.1051 0.9718 136.20.10520.10510.9693 135.4 0*0.58030.05800 0.05801 !::if:: 0.02405 (0.9669) 130.8 8::::;: 0*01147 (0.9659) 121.80.006512 0.006506 0.006509 (0.9656) 136.6: : ~ ~ ~ ~ ~ ~ 0.003488 (0.9654) 181.60.0034860.002 15 10~02000-N 0.53260.002168 0.002159 (0.9654) 224.3 0*01000-N 0-2663On comparing the values in table VI for potassium laurate withthose of F. Goldschmidt (Zeitsch. Elektrochem., 1912, 18, 386) forthe potassium salt of the acids from palm kernel oil (“ chief consti-tuent potassium laurate ”*), his values are seen to vary somewhat* Oudemanns (J. pr. Che?n., 18’15, 11, 393) gives the following data for a palmkernel oil : oleic acid, 26.6 ; stearic, palmitic, aud myristic, 33 ; lauric, decoic,octoic, and hexoic acids, 44‘4 per cent.Lewkowitsch, however, in view ofValenta’s later data considered that lauric acid must be the chief constituent, andthat the oleic acid could not exceed 10 to 20 per cent.VOL. cv. F 426 BUNBURY AND MARTIK' :TABLE VI.Conductivity of Potassium LCHLTU~C (C,,) at 90.00'.I. 11. 111. IV. V. VT.0.1650 0.9835 123.5 2.028 - N 48.34 0.16530.16470.1134 0.9761 143.2 1-007-N 23.99 0.11350-11330.754-N 17-97 0.0888 0.0888 (0.9741) 142.60.9720 146.0 0.50 16-N 11.95 0.06363(0.9680) 144.2 0.2003-N 4.772 0.02665(0.9666) 159.7 0.1001 -N 2.384 0.01509: : ~ ~ " g ~ p 0.009341 (0.9660) 195.6 0.0500-N 1.1910.06357 o'063600.02671 0.026680.01508 0.015098::::::; 0.002245 (0.9654) 233.0 0.0 1000- N 0.2383irregularly between those here obtained for potassium laurate andmyristate.Some of his data are 137.3, 143.6, 141, 135, and 173.5for the 1.0, 0.74, 0.5, 0.2, and 0*05N-solutions respectively. Hisresult3 for 0.74N- and 0'5.N-solutions are relatively high, as ourresults show no such decided maximum. We have not yetmeasured mixtures of pure fatty acids.TABLE VII.Conductivity of Potassiam Decoate (Cl,,) at 90*OOo.I.1.007-N0.6016-N11. 111. IV.0.1185 21-17 0.11880.118210.54 0.069010.06927 0.06914V. VI.0.9827 145.90-9749 156.3(0.9691) 180.9 4.21 1 0.033682-104 0-01902 (0 9672) 200.6 0.01902 0.019020.03372 0*033700.2003-N0*1001-N0.0500-N 1.051 0*01014 0.01014 (0.9663) 211.90.002231 0.002239 (0.9655) 232.4 0.0 1000-N 0,2103 0.002241TABLE VIII.Conductivity of Potassium Octoate (C,) at 90*OOo.3.063 -N 55.82 0.2165 0.2165 (1.027) 107.3'I.11. 111. IV. V. VI.0.1251 0-9893 148.7 1.007-N 18.34 0-12510.12510.5016-N 9.14 0*07584 0.07672 0.9784 168.50.075 600.2003-N 3.746 0'03586 0.03578 (0,9706) 191.0*0.0357 10.1001 -N 1.823 o'ol 953 0.01 952 (0.9679) 206.20.019500.0505-N 0.91 1 o'01062 0.01061 (0*9666) 219-20.010600.0 1000-N o*1822 0*002310 0.002309 (0.9666) 239-6 0*002309* Dekrniinatiou by Miss CornishSTUDIES OF THE CONSTITUTION OF SOAP SOLUTIONS. 427TABLE IX.Conductivity of Potassium Hexoate (Cs) at 90.00°.I. 11. 111. IV. V. VI.1.007-N 15.62 0- 1300 0.1300 0.9974 149.50.9822 177.7 0.6016-N 7.74 0.081460.08098 o*08122(0.9721) 201.2 0.2003-N 3.088 0.037 740.03767 0.03771(0.9687) 216.5(0.9670) 227.70.1001 -N 1.643 0.020690.0600-N 0.771 0~010900.02064 0.020670.01095 0.010930~01000-N 0.1542 0.002368 0.002375 0.002371 (0.9656) 245.9At 90° all the potassium soaps from the myristate downwardsare clear liquids, below and including the l*ON-solutions.A t roomtemperature, to which the following notes refer, theae soaps exhibitnone of the silky appearance characteristic of the palmitates (andto a less extent of the stearates).The 1.0N- to 0'05N-myristate solutions are clear liquids (the1 *ON-solution being oily) with steadily increasing amounts of flakysediment, the O*OlN-solution containing most, and being alsoslightly opalescent.On long keeping large groups (2 cm.) offeathery crystals separate out from the 0*05N-solution.l*ON-Potassium laurate is clear except for a slight sediment.The 0.5N-solution contains much more of this flaky sediment, butthere is also a large amount of amorphous turbidity at the topof the liquid, which is clear in the middle layer. The 0-2N-solutionis similar with less sediment and less turbidity. The O*lN-solutionhas as much sediment as the 0'2N-solution, but there is only avery slight turbidity a t the top. The 0~05N-mlution has a slightsediment, is faintly opalescent, and shows no upper turbidity.With the decoates (Clo) the sediment is less than with the laurates( C Q , and it is a t a maximum in the 0.W-solution.The1'0N-decoate solution is very slightly viscous, and has only a veryslight sediment. Only in the 0.2, 0.1, and 0'05N-solutions is therea trace of the " upper turbidity"; the 0.01N-solution is almostclear, wikh a alight sediment.In the case of the octoates (Cs) no surface turbidity is observed;the maximum sediment is found in the 0*2N-solution, but there isonly a trace of sediment in the l'Ofl-solution. The 0*05fl-solutionis somewhat opalescent, the O'OIN-solution only faintly so.With the hexoates (Cs) the sediment is still less, being at amaximum in the 0-1N-solution ; the O*O5N-soluti~n is faintlyopalescent.It was impossible to measure the conductivity of l.5N-potassiumF F 428 BUNBURY AND MARTIN:palmitate, 2*ON-myristate, and 3.0N-laurate at 90°, as they areviscous jellies even a t looo, and conductivity electrodes could notbe filled with solution free from air-gaps even through severalhours' work.A t room temperatures the 2.ON-laurate and the3-ON-octoate solutions are oily liquids, with a distinct sediment ;the S*ON-solution, on the other hand, is a transparent jelly,preserving its form unaltered by the influence of gravity.Kashing Power."It is interesting and important to note the different washingpowers of the various hard and soft soaps. The potassium myristateand laurate are ideal soaps (for washing the hands), and the palmi-tate is almost as good, but the stearate is much less so. Sodiummyristate is good, but sodium palmitate and stearah are quiteuseless, feeling remarkably like greasy sawdust.Potassium hexoate (C,) is distinctly a soap in concentratedsolution, although this at om0 disappears on dilution.Potassiumoctoate (C,) is still more distinctly a soap, but this again disappearson dilution. The decoate (C,,) is the first to raise a typical lather,although this has not much body.Discussion of the Results.In order to facilitate comparison, the conductivity results aresummarised in table X. These molar conductivities are also showngraphically in Figs. 1 and 2, which should be compared with thecorresponding graphs (T., 1912, 101, 2049) of previous data forthe higher sodium salts.TABLE X.Molar Conductivity of Potassium Soaps a t 90*OOo.Concentration . . .Stearate, C,, ......Palmitate, C,, .. .Myristate, C,, . ...Laurate, C,, ......Decoate, C,, ......Octoate, C, .......Hexoate, C, .......Acetate, C, .......1.0113.4124.2136.2143-2145.9148.7149.6176.90.75 0.6112-6 113.9127.9 127.0 - 135.4142.6 146.0- 168.6 - 177.7183.9 196.6- 156.30.2100.011 1.0130.8144-2180-9191.0201.2221-20.196.0107.0121.8169.7200.6205.2216.6236.50.06101.7110.8136.6195-6211.9219.2227.7249.60.02 0.01124.9 147.7133.2 171.6181.6 224.3 - 233.0 - 232.4 - 239.6 - 246.9262-6 270.4It will be noticed at once how similar the behaviour of eachof these salts is to that of the corresponding sodium salts as far asThese notes are certainly not intended to be taken as a general discussion ofwashing power, which is known to be a complicated property, some of the factorsof which have received separate elucidation. On the contrary, the present remarksrefer only to the conditious stated (warm tap water ; pure, not mixed acids, etc.).STUDIES OF THE CONSTITUTION OF SOAP SOLUTIONS.429these have been measured. All these soap solutions conduct excel-lently. The nature of the constituents which exhibit this highconductivity will be elucidated in two further experimental com-munications from this laboratory now completed and awaitingpublication ; almost all authors have ascribed this conductivity(erroneously as will be seen) to free alkali hydroxide.The close similarity between the corresponding sodium andpotassium salts does not, however, amount to an identity, either inthe position or form of the conductivity curves; as is seen, forexample, in the following calculation : If the difference in theconductivity of the sodium and potassium soaps is essentially dueto the difference in mobility of the sodium and potassium ion, thisdifference might be allowed for by adding to the conductivity valuesfor sodium palmitate the excess in conductivity of potassium acetateover that of sodium acetate, taking the corresponding concentrationsin each case.(It so happens that the difference between the con-ductivity of potassium and sodium acetates is constant, within thelimits 41.2 and 42.6 mhos., over the whole range of concentrationfrom 0.5N to 0*01N.) The values thus predicted for potassiumpalmitate from the conductivity of sodium palmitate, and thedivergence between these calculated and the observed values are :Concentration .. . . . . 1.0 0.5 0-2 0.1 0.05 0.01KP Ob60I'Ved ......... 124.2 127.0 111.0 107.0 110.8 171.6-observed . ........ 7.7 5-1 13.7 17.0 19.5 8.5Difference per cent. 6-2 4.0 12.4 16.9 17.6 4.9NaP+(KAc-"eAc) 131.9 132.1 124.7 124.0 130.3 180.1Difference, theoryPotassium palmitate has thus an appreciably lower conductivitythan that predicted in this manner from the values for sodiumpalmitate; the difference being greatest in the N/ZO-solution,where it amounts t o over one-sixth of the total conductivity.Another mode of comparison, not a t all necessarily morelegitimate, is that to be shown graphically below (in Figs.3 and 4),which is mathematically equivalent to adding to the values forsodium palmitate only that fraction of the difference between thetwo acetates which the conductivity of sodium palmitate itselfbears to that of sodium acetate a t the same concentration.A second proof that the potassium ion as such is not entirelyresponsible for the differences between the conductivity of the hardand soft soaps is to be found in the fact that the differences betweenthe maximum and minimum conductivity values are much moremarked in the case of the potassium soaps. This is seen withparticular clearness in the case of the laurates, where the potassiumsalt exhibitg a well-marked maximum and minimum conductivity430 BUNBURY AND MARTIK:whilst these are entirely absent from the curve for the sodium salt,which exhibits merely a " step-out."Considering the molar conductivity values for the various potass-ium salts as a whole (if the specific conductivities were considered,the relationships would be similar), it is seen from the graphs inFigs.1 and 2, that in dilute solution, below N/20, the soaps fallinto two groups, the stearate and the palmitate, on the one hand,and from the laurate t o the hexoate (C,) on the other, with themyristate occupying an intermediate position.If solutions ranging between N/5 and N/20 be considered, theintermediate member is the laurate, the myristate here resemblingthe higher maps. I n X/2-solutions, the grouping is entirely lost,for the conductivity values are almost equidistant from each other.I n l'ON-solutions, however, the conductivities of potassium hexoate,octoate, decoate, and laurate are nearly identical, whilst the valuesfor myristate, palmitate, and stearate fall off regularly.It should be emphasised here that, in a sense, even the hexoates(C,) are soaps, for they precipitate quite appreciable amounts ofsediment( a t room temperature (of course, not silica, for it is ob-served just as well in solutions freshly taken from the silver tubesin which all our solutions are prepared), and, as was seen, theamount of this sediment is a t a maximum in O'lN-solution.Donnan and Potts (Rolloid-Zeitsch., 1910, 7, 208), on the basis ofsurf ace-tension experiments, conchded that sodium octoate (Cs)was the lowest soap, although Donnan, using diluh solutions(Zeitsch.physikd. Chem., 1899, 31, 42), found that sodium lauratewas the lowest member to give a great lowering of surface-tension.Krafft (Ber., 1896, 29, 1328), on the other hand, had found thatthe nonoate (C,), in very concentrated solution, was the first togive too high a molecular weight, whilst Mayer, Schaeffer, andTerroine (Compt. rend., 1908, 146, 484) state that the hexoate (Cs)is the lowest member showing ultramicroscopic particles.The data available are still quite inadequate to decide whethera gradual transition takes place from crystalloid to colloid onchange in concentration or change in the homologue, or whether,on the other hand, the transition may be completely accountedfor on the hypothesis that true electrolytes co-exist with colloidaland electrolytic colloidal (in the narrow sense) constituents, andthat these merely vary in relative amount.Perhaps the clearest insight into the complicated behaviour ofthese conductivity data may be obtained through a direct compari-son of all the other salts with the corresponding acetate.In thisway the effect of change in degree of dissociation (and also thedifferent influence of the two cations on the degree of dissociation)STUDIES OF THE CONSTITUTION OF SOAP SOLUTIONS. 431which mwt be, to some extent, superimposed upon the otherfeatures, is approximately eliminated. The further communicationsalready referred to will bring more justification for this mode ofcomparison, although owing to the difference in mobility of thevarious ions concerned a relative method (involving a ratio) suchas this can never be quite exact.An attempt will be made tocomplete this comparison by measurement of the concentrations ofthe potassium and sodium ions by a direct method.If the conductivity of potassium acetate be taken for the sakeof comparison to be unity at all concentrations, the conductivitiesof the potassium soaps may be expressed as fractions of thisstandard. Similarly, the conductivity of the sodium soaps so far1'000-900.800.700 *600.500'40FIG. 3.0.200 -10 0.10frl2 5 10 20 25 27 9899100Abnormality of conductivity of potassium soaps relative to potassiumacetate at 90".measured in this laboratory (Zoc.cit.) may be compared with thatof sodium acetate of the same concentration taken a.s unity. Theresults of thie comparison are best seen graphically in Figs. 3 and4, where the conductivities relative to the corresponding acetateare plotted against the dilution in litres.The effect on the appearance of the conductivity curves is reallystriking, and the graphs in Figs. 1 and 2 permanently lose muchof their appearance of irregularity.*The similarity in magnitude and form between the values forthe corresponding potassium and sodium salts is appreciably* The 1.0N-sodium stearate and the O-lN-potassium decoate now appear as ifa redetermination ought to give slightly lower values, although there is no otherreason t o suspect the accuracy of these particular solutions432 BUNBURY AND MARTIN:enhanced by this mode of comparison, although there is definitelyno identity in any case.The results for the potassium soaps nowlie a little higher than the curves for the corresponding sodiumsalts. The general remarks concerning the grouping of the resultsstill hold, but the regularities are much more distinct.Thus the curves, except for the hexoate (C,) and octoate (CJ,all exhibit minima. These minima are regularly graduated inposition and depth. The minimum for sodium stearate is a t N / 2 0 ,for potassium stearate between N / 2 0 and N/10; for sodium palmi-tate between N / 2 0 and N / l O , for potassium palmitate a t N/20;for the myristates ,V/lO in both cases; for the laurates N / 5 in bothcases; for potassium decoate (C,,) at N / 2 .ThO most concentratedFra. 4./Ih a i - - - - - - i0.90o .a00 *700.600 5 00.400.300.200.10I I I I I I 1 I ........ u 1 DL&&L in fmed43 1 2 3 4 5 10 20 25 27 9899Abnormality of conductivity of sodium soaps relative to soditcmacetate at 90".0'900 3 00 -700.600 -60P0.400'300.200.10100solutions of the lower homologues were not measured, so that itis not certain whether, for example, the decided dip of the curvefor l*ON-potassium hexoate (C,) leads to a minimum or not.*It was expected that the value for the 2*ON-potassium laurate(C,,) would be much higher, but it is, if anything, just less thanthe value for the 1-0N-solution.I n the graphs for the sodium soapsthe 1*5N-solutions are included in each case, with a very markedeffect on the appearance of the curves (compare, for example,sodium and potassium palmitates) ; the 1*5N-solutions have notbeen measured in the case of any potassium salt.If the hexoate had been taken as a standard of comparison in* The insertion. i r i thp proof of this paper, of the low value of the ratio forThe valueThus the relative conductivity of all the3*01\7-p btashiiim octoate (Cs), 0.79, practically excludrs this possibility.for 2-01\~-11otassiiirn laurate (C12) is 0.81.lower members rapidly falls off in concentrated solutionSTUDIES OF THE CONSTITUTION OF SOAP SOLUTIONS. 433place of the acetate, owing to the hexoate curve being horizontalbetween N / 2 and N/100, the resulh would have been uniformlyincreased by 0.10, except in the meet concentrated solutions, wherethe values would have been still further increased.Reference must here be made to the able arguments advancedby F.Goldschmidt (Zeitsch. Elektrochem., 1912, 18, 394) in favourof the view that the conductivity data are not wholly distorted bythe effects of the very considerable viscosities here met with insome of the solutions.It is indeed increasingly evident that one of the ultimate resultsof the study of colloidal electrolytes will be to throw new light oncome of the problems of the dissociation theory, both in aqueousand in non-aqueous solutions. Physical chemists were formerlyquite clear on the point that they did not know how to allow forthe effects of viscosity on the quantitative interpretation of conduc-tivity data, for this question has not 80 far received adequateexperimental treatment.In the last few years, however, it hasbecome customary to correct all conductivity data for the ratio ofthe viscosity of the solution compared with that of the solvent, orwith some power of this ratio, this being based on consideration ofStokes’s law, the temperature-coefficient of conductivity a t infinitedilution, or the effect of change of solvent on the conductivity a tinfinite dilution, or a few meagre data on the effect of an addednon-electrolyte on conductivity. This arbitrary procedure is but amake-shift until the effect of viscosity on diffusion has been experi-mentally elucidated, and even its empirical basis appears inad+quate until phenomena such as are met with here have beenreconciled with it.Densities.The densities observed promise to afford some help towards theelucidation of the soap problem.The results are collected in tableXI, and for convenience of comparison they are referred towater of the same temperature its unity; each result is, of course,the mean of several determinations differing by a few units in thefourth decimal place.TABLE XI.Densities D$ of Potassium Salts.Concentration ...............Stearate, C1, ...............Palmitate, C16 ...............Myristate, C,, ...............Laurate, (& ................Decoate, C,, ...............Octoate, C, ...............HeXO&t.e, C6 ...............Butyrate, C, ...............Acetate, C, ..................2.0-N.1.0-N - - - 1-0029 - 1,00671.0189 1.0112 - 1.0180 - 1.0249 - 1.0333 - 1-0398 - 1.04630.6-N. AI.0.0.9993 -1-0016 0.00381,0041 0-00451-0069 0-00681.0102 0.00691.0136 0-00841.0175 0.00651.0212 0.00551.0223 -2A0.,.0.00460.00500.00560-00660.00680.0078000.74000- 18 - -0.0003040160-00260.00240.00230.00170-0026.0.000434 STUDIES OF THE CONSTITUTION OF SOAP SOLUTIONS.The density values are analysed in the last three columns oftable XI; Al.,, is the amount which has to be added to the densityof a 1-ON-solution in order to bring it up to the value for the nextlower homologue, 2A0.5 is the corresponding difference for the0~5N-solutions, but doubled in order to compare them directlywith Al.o; and finally, A, is the difference between the density pre-dicted for the 1'0N-solution from that of the 0'5N-solution if theusual rule for electrolytes mere obeyed, that the divergence of thedensity of the solution from that of water is directly proportionalto the concentration. It will be noted that all the salts betweenthe acetate and the palmitate (for which it holds) depart decidedlyfrom this rule, for the density rises very much less rapidly thanproportional to the concentration.Thus the excess of density of2*0N-potassium laurate over that of water is only two-thirds of thatpredicted from the O.5N-solution.The effects are very regular and gradual, as may be seen from thetable, or still better from a graph.When plotted against theposition in the homologous series, the densities give quite smoothcurves (not straight lines), provided that the acetate is put intothe place of the butyrate, the latter being omitted. I n other words,there is a point of inflexion or a decided break in the curve betweenhexoate (C,) and acetate (C,). Thus again the hexoate betrays itsrelatedness to the higher soaps.The densities of the solutions of the corresponding sodium salts,as far as our data extend, lie on similar curves from 0.0060 to0-0070 units (for 1*0N and half that for 0.5X) below those intable XI, and, in this respect, again illustrate the parallel be-haviour of the potassium and sodium soaps.*Potassium and sodium stearates and sodium palmitate areexamples of the rare case that the density of a solution may be lessthan that of either coiistituent (compare A.L. Hyde, J . Amer.Chem. Soc., 1912, 34,1507).Summary.(1) The conductivities of the potassium salts (soft soaps) of thesaturated fatty acids of even number of ca.rbon atoms from thestearate down to the acetate have been measured a t 90° by thesomewhat laborious method previously described by McBain andTaylor.* The density of 1 'ON-sodium myristate appears to be about 0.15 per cent. toohigh, and the molar conductivity previously published should be increased by thisamount ; similarly, the 1*5N-sodinm myristate should be corrected by about0'34 per cent., whilst the conductivities from the 0'5N-solution downwards remainexactly as they standNON-AROMATIC DIAZONIUM SALTS. PART 111. 435(2) The conductivities of the potassium soaps are higher thanthose of the corresponding sodium soaps, but there it3 a generalresemblance between the form and position of correspondingcurves. Closer comparison shows an even greater tendency towardsabnormality on the part of the potassium salts, this not being dueto the potassium ion as such, well-developed maxima and minimain the conductivity curves being exhibited from the stearate asfar down as the laurate (CI2).(3) The appearance, washing power, density, and conductivitycurve of potassium hexoate (C,) distinctly mark the beginning ofthat deviation from the behaviour of the acetate, which rapidlyand regularly increasw through the otker members of the homo-logous series until it attains the typical character of the highersoaps.(4) In all cases where it is directly visible the depression in theconductivity curve occurs in the same region of concentration,independent of the nature of the acid or alkali taken. Furtherinvestigation might, however, show that the real abnormality isshifted in the case of the lowest homologues to regiogs of higherconcentration.In conclusion, w0 desire to express our thanks to the ResearchFund Committee of the Chemical Society, and especially to theColston Society of the University of Bristol, for grants towardsthe purchase of materials. and apparatus.Our thanks aro due to Dr. James W. McBain, a t whose sugges-tion this work was undertaken, for his constant interest and advice.THE CHEMICAL DEPARTMENT,THE UNIVERSITY, BKISTOL