Solubilization and the Colloidal Micelles in Soap Solution

be utilized to throw light upon the nature of the micelles and solutions of colloidal electrolytes. The residís here tobe presented show that solu- b...
3 downloads 0 Views 565KB Size
Jan., , I943

SOLUBILIZATION AND THE COLLOIDAL MICELLES IN SOAP SOLUTION [CONTRIBUTION FROM

THE

9

DEPARTMENT OF CHEMISTRY, STANFORD UNIVERSITY]

Solubilization and the Colloidal Micelles i e Soap Solution BY J. W. MCBAINAND K.

E.JOHNSON

Solubilization is the power which even dilute scribed in references 3 and 4,namely, l-o-tolylazo-2-naphF. D. and C. orange no. 2; Orange OT, in the form of aqueous solutions of soaps, detergents and other thol; well-developed crystals. The potassium soaps were precolloidal electrolytes possess of bringing into pared by neutralization of Kahlbaum purest fatty acids thermodynamically stable’ colloidal solution with Kahlbaum, washed, carbonate-free potassium hyhydrocarbons, dyes and other substances insoluble droxide. These were made up in boiled-out distilled water and contained 0.4 equivalent per cent. excess of potassium in water. The commercial importance of this hydroxide to suppress hydrolysis Dye was added in only phenomenon has long been known, but it is only slight excess and the mixture was gently agitated in closed recently that its mechanism has been established containers in a thermostat until equilibrium was reached. as consisting of sorption upon and incorporation This required a t least sixteen to twenty-four hours for more dilute solutions and from seven to ten days for more conwithin colloidal micelles.2t584 centrated solutions of soap. When approaching equilibConversely, a study of solubilization may now rium from supersaturation, much longer times were rebe utilized to throw light upon the nature of the quired, presumably on account of the effect of protective micelles and solutions of colloidal electrolytes. action upon the dye separating out. The liquid was then kept motionless in a thermostat for twenty-four hours, or The resdts here to be presented show that solu- until examination in a strong light showed that all dye bilization cannot be accounted for by solution in particles had settled. Analysis was made by diluting a the hydrocarbon portion of the micelles, and they portion of the liquid with a mixture of acetone and water therefore show that assumed structures such as in such proportion as to keep the dye in solution, and measuring in a Klett-Summerson photoelectric colorimeter spherical micelles, in which this is the only mech- with a B-420 blue filter, calibrated against known beniene anism so far proposed, are wholly inadequate. solutions of dye. The latter gave similar results to dye On the other hand, they lend further support to solubilized in aqueous Aerosol OT. the conception that the solubilized material is Some Results with Commercial Detergents placed in layers within the lamellar micelles. Data for many detergents have been published Materials and Method in reference 4. We may supplement these by the The dye and methods used were the same as those de- results in Table I. TABLE I THE SoLUEILIZATION I N MILLIGRAIUS OF DYE PER HUNDREDCC. OF SOLUTION,MADE BY TAKING ONE G Detergent

MERCIAL Mg. of dye

Aerosol 18

DETERGENT IN ONE HUNDRED Cc. Type of detergent

OB

U OF COM-

WATER Manufacturer

2.60 Disodium monosulfostearylsuccinamide, 35%, American Cyanamide and 65% water and slight turbid impurities Chemical Co. Aerosol OS 3.65 Sodium isopropyl naphthalene sulfonate, 95% Same active alkyl naphthalene sulfonate-fairly pure Ammonium salt of Aerosol OT 1.44 Ammonium salt of dioctyl(2 ethyl hexanol) ester Same of sodium sulfosuccinic acid Zinc salt of Aerosol OT 6.66 Zinc salt of above Same Alrose Chemical Co. Alronol(lOO%) 3.00 Non-electrolytic detergent with trivalent nitrogen Alronol(90%) 3.4 Same Nacconol NRSF 3.23 Sodium alkyl aryl sulfonate, not a sulfated fatty National Aniline and Chemialcohol cal Co. Nopco 1067 1.17 An alkyl aromatic sulfonate National Oil Products Co. Nope0 1086-C 0.44 Same Same Procter and Gamble Orvus W. A. Flakes 3.70 Sodium lauryl sulfate Orvus W. A. Paste 1.85 Same Same Santomerse 1 1.25 Belongs to a homologous series of substituted Monsanto Chemical Co. aromatic sodium sulfonates (alkylated aryl sulfonate) Santomerse S 0.95 Same Same Sodium tetradecyl sulfate 0.13 Specially purified E. I. du Pont de Nemours Co. Triton W-30 0.57 A sulfonated aromatic ether alcohol RBhm and Haas (I) J. W. McBain and M. E.

+

L. McBain, THISJOURNAL, 68, 2610

(1936). (2) J. W. McBain, “Solubiliaation and other Factors in Detergeqt Action,” Advances ia Colloid Science I, 99-142, Intencience Publishera Company, Inc., Now York, 1942. (3) R. c. Mdurill, Jr., and J. W. McBain. J . P h w . Chem., 46, 10 (1942). (4) J. W. McBain and R. C. Meml 1 Jr.,Ind.

(1942).

En#. Chsm.. 84,916

The Solubilization Of Dye h a Homologous Series of Pure Potassium Soaps Containing Eight to Fourteen Atoms*-The lowest members of the homologous series of potassium salts of fatty acids, such as potassium are Ordinary propionate and strong electrolytes, whereas colloid appears in con-

J . W. MCRAIN.4ND K. E. JOHNSON

10

centrated solutions of potassium hexoate, and in progressively less concentrated solutions of the higher soaps. Since solubilization of dye is amanifcstation of the presence of micelles, it is of great interest to compare the solubilization of dye by various concentrations of typical homologs both to show in how far these are colloidal electrolytes, and also to throw light upon the nature of the colloidal micelles present in different concentrations of different homologs. Measurements are therefore presented in Tables 11, 111, IV and V for the solubilization of Orange OT in potassium octoate, decylate, dodecylate and tetradecylate containing the added alkali sufficient to suppress hydrolysis. The results are also shown graphically in Fig. 1. T o ensure attain-

Vol. 66

TABLE IV THES~LUBILITY OF ORANGE OT I N VARIOUS MOLALITIES OF AQUEOUSSOLUTIONS OF POTASSIUM LAURATE (Kcla) AT 25’ in

soap

0 05 .10 15 .20 .25 .50 .75 1 00

2.00

From undersaturation M dye/

?*1 dye/ 180 cc

1 94 6.90 10.74 15.65 20.87 39.54 59.59 83.15 164.5

mo%

KC,^

388.0 690 716 782 835 79 1 795 832 822

From oversaturation M .dye/ M .dye/ 180 cc. mofe KC^?

1.95 7.43 12.16 16.79 21.03 42.77 65.10 85.44

390 743 811 840 841 855 868 854

TABLEV THESOLUBILITY OF ORANGE OT I N VARIOUS MOLALITIES OF AQUEOUS SoLUTIONS OF POTASSIUM MYRIS7ATE (Kcid) AT 25”

w

B

0

0.2 0.4 0.6 0.8 1.0 2.0 Normality of soap. Fig. 1.-The solubility of orange OT in various molalities of aqueous solutions of a homologous series of pure potassium soaps at 25’.

m soap

Mg. dye/ 100 cc

M dye/ mofl KC14

0.025 05 I10 15 20

3.26 7.23 14.87 22.03 29.27 36 20 73.09 108.4 147.4

1306 1445 1487 1468 1463 1448 1462 1445

a5

.io 75 1 00

1474

%O%/

M dye/ mof; KCW

...

..

7.24 15.17 22.73 30.39 37 76 76.04 114.1 152 3

1448 1517 1515 1520 1510 1521 1521 1523

It will be seen from the tables and Fig. 1that the ment of the reversible equilibrium from both sides, amount of dye solubilized by one mole of each of the measurements with laurate and with myristate these four successive homologs of even carbon are approached from the side of undersaturation number stand in the proportions of 1 to 2.34 to and also, by first heating the solution to cause more 6.77 to 12.04, when the concentration is suffidye to dissolve, from the side of supersaturation. ciently high to ensure full colloidality of the soap. I t is seen from the graphs how closely the correIn other words, the solubilizing power per soap sponding values came together in the timeallotted. with molecule in a micelle increases enormously We may therefore be sure that the true solubilizarelatively small increase in the number of carbon tion was actually measured atoms in the molecule. Nevertheless, the soap solutions are strictly TABLE I1 comparable because, as Brady has shown,6 the THESOLUBILITY OF ORANGB OT IN VARIOUSMOLALITIES osmotic coefficients of all straight chain colloidal OF AQUEOUS SOLUTIONS OF POTASSIUM CAPRYLATE (KC*) electrolytes actually coincide when they are plotAT 25’ ted on a diagram in which the unit of concentration From undersaturation From oversaturation is in each case taken as the molality a t which the Mg. dye/ Mg. dye/ m soap Mltod~’ %,“E< 100 cc. mole KCs osmotic coefficient, g, equals 0.5. The &&cance 0.700 2.73 39 0 3.18 45.4 of these data will be discussed in a later section. .. .. 0.924 4.90 53.0 Next of interest was to ascertain the solubiliz2 00 53 0 115.2 26.7 133.7 ing power of equimolar solutions of laurate and myristate. The data given in Table VI and Fig- 2 TABLEI11 show that the solubilizing power of the mixture T H E SOLUBILITY OF ORANGE OT I N VARIOUS MOLALITIESlies not quite midway between that for the two OF AQUEOUS SOLUTIONS OF POTASSIUM CAPRATE (KCd pure soaps separately, namely, 9.09 times the AT 25’ solubilizing power of the caprylate. In addition to these data for 2 5 O , measurements given in Table VI1 and plotted in Fig. 2, a t No, show the very high terriperature coefficient of 0 1 1.81 181 1 19 149 solubilization. 186 .2 3.72 .4 1.0

10.33 29.21

258 292

11.26 28.98

282 290

(5) J W. McBa~nand A. P. Brady, THISJOURNAL, 66, 2072 (1943).

Jan., 1944

SOLUBILIZATION AND THE

COLLOIDAL MICELLES IN

11

SOAP SOLUTION

TABLE VI

THESOLUBILITY OF ORANGE OT IN VARIOUS MOLALITIES OF AN

EQUIMOLAR MIXTUREOF POTASSIUM LAURATE AND POTASSIUM MYRISTATEAT 25'

m

soap

0.05 .10 .15 .20 .25 .50 .75 1.00

From undersaturation Mg. dye/ Mg. dye/ 100 cc. mole soap

1019 1058 1127 1129 1105 1102 1106 1123

5.10 10.58 16.91 22.58 27.61 55.07 82.92 112.3

From oversaturation Mg. dye/ Mg. dye/ 100 cc. mole soap

5.17 10.89 17.14 22.99 29.21 57.17 84.83 114.2

1370 1519 1657 1741 1764 1746 1759 1751

d 2000 fl-0

% 1600

g

a

m soap

Mg. dye/100 cc. s o h . ClZ Cld Cl4

CI?

+

Mg. dye/mole soap CI?

cl? ..

+

cI4

+

at 50°

800.

u

$

400'

01 0.4 0.6 0.8 1.0 Normality of soap. Fig. 2.-The solubility of orange OT in pure potassium soap solutions at 50' and in an equimolar mixture of potassium laurate and myristate at 25'. 0

TABLE VI1 SOLUBILITY OF ORANGEOT IN VARIOUS MOLALITIES OF SOAPS AT 50"

c,, G,'

0.2

el4

of the micelle. This accords well with the positive temperature coefficient of solubilization, and it has been adopted by Hartley in spite of some large numerical discrepanciesbetween the amounts ... of dyes solubilized and their solubilities in a .. . similar weight of hydrocarbon. 33.30 The main result of the present communication 66.76 is to show that the solubilization of the same dye .., by a homologous series of soaps changes with mo132.7 lecular weight in a manner which does not support The solubility has therefore increased between the hypothesis that solubilization is solution in the 25 and 50' to an extent corresponding to a nega- hydrocarbon part of the soap. The results are extive heat of solubilization amounting to -3830, pressed in milligrams of dye solubilized per mole of 3460 and 3830 calories per mole of soap for lau- soap, because this allows of direct comparison with rate, for mixture of laurate with myristate, and for the number of CH2 groups per molecule of soap. myristate, respectively. These are similar to the The results show that for soaps containing 8, 10, values calculated from previous worka3 They have 12 and 14 carbon atoms, respectively, the solubilthe same order of magnitude as heats of solution or izations are in the proportions 1 :2.34:6.77: 12.04. as heats of sorption of the van der Waals type. Thus, for increases of 25, 20 and 16.7% in The increase in temperature also has a marked length of carbon chain, the increases in dye solueffect on the concentration of soap at which the bilized are 143.5, 190 and 79.5%> respectively. concentration necessary for full colloidality is For an increase of 50% in carbon atoms from caphigher at 50 than a t 25'. rylate to laurate, the solubilization has increased Solubikation and the Nature of Mkel1es.6.77-fold. The increments of these numbers are, Solubilization has been shown to he a property in the same units, 1.34, 4.43 and 5.27 for each of the micelles in solutions of colloidal electrolytes. added pair of CH2 groups, instead of being conConversely, solubilization can throw light upon stant. It should be recalled that these numbers the nature of these micelles. were strictly comparable because they refer to The first micelles proposed were the spherical concentrations sufficiently high to ensure full micelles of Reychler, McBain, and later Adam colloidality in each case; likewise the numbers and Hartley. The radius of the spherical micelle were constant'for each soap as soon as full is limited by the length of the hydrocarbon chain colloidality was reached. They are solubilization which is expelled from contact with the water as per molecule in the fully formed micelles. far as possible. Nevertheless, as Ward6points out, The only form of micelle for which there is seven-tenths of the surface of a spherical micelle direct evidence is the lamellar micelle, first procontaining 50 soap ions of 12 carbon atoms is still posed by McBain. Several investigators7 have hydrocarbon, even if it is assumed possible so to (7) P. Krishnamurti, I n d i a n J . Phys., 3, 307 (1929); K. Hess, pack the short, thick rods, each length 15 k.,cross H. Kiessig and W. Philippoff, Nafurwisscnschaffen,26, 184 (1938); section 20-25 sq. k.,without interspaces. K. Hess, W. Philippoff and H. Kiessig, Kolloid-Z., 88, 40 (1939b); It is an obvious suggestion that the solubilized H.Kiessig, {bid., 88, 213 (1942); H. Kiessig and W. Philippoff. 2'7, 593 (1939); W. Philippoff and K . Hess, dye merely dissolves in the hydrocarbon interior Nafuruisscnschaftcn, Ecr., TO, 1808 (1937); J. Stauff, Natuvwissenschaffcr, 97, 2tR 0.025 .05 .10 .15 * 20 .25 .50 .75 1.00

...

...

5.27 11.82

6.85 15.19 24.86 34.81 44.10 87.30 131.9 175.1

4.94 10.93 23.37 34.47 46.34 58.59 117.3 173.0 232.9

..

1054 1370 1182 1519 . . 1657 . . 1741 1332 1764 1335 1746 . . 1759 1327 1751

1796 2185 2337 2298 2319 2344 2346 2307 2329

(6) A . F. H. Ward, Proc. Roy. Soc. (London), 8176, 420 (1940).

(1039a); J. Stauff. Kolloid-Z., 89, 224 (1939b).

J. W. MCBAINAND K. E. JOHNSON

12

A

1000

1.00

1.00

900

0.90

0.90

v

2

VOl. titi

-

0.80

0.80

’io0

oio

0.70

w

600

.8E: 0.60

O.fiO

2

900

0.50

0.50

800

3 ,

u 0

k

400 ;+

5a

2:

G

0.40

0.40

2 0.30

0.30

-2

200

0“ 0.20

0.20

7

100

0.10

0.10

0

0.00

a,

.$‘ 300 . 4 c

9 2M

g

c)

P

0.00

0.0

0.8 1.0 1.2 1.4 1.6 1.8 2.0 Normality of soap. Fig. 3.--Comparison of the solubilization curve (SI, the osmotic coefficient curve i g ) , and the conductivity curve potassium laurate. 0.2

04

0.0

(01)

for

found that the micelles in soaps and synthetic detergent solutions give an X-ray diffraction

higher value at 0.75 m. The changes in the micelles indicated by conductivity and osmotic activpattern corresponding to lamellar micelles in ity are not reflected by any further change in soluwhich molecules are packed side by side and layer bilization. Most authors now agree that there are to layer. According to Hess and his collaborators, many sizes of micelles in different concentrations when benzene is solubilized by the soap solution, of the same colloidal electrolyte, ranging from spaces between the lamellar planes increase corre- “small micelles” or “ionic micelles” that raise the spondingly. This mechanism, perhaps supple- conductivity to larger lamellar micelles that mented by sorption on the exterior of the micelles, greatly lower it. would best accord with the present data. X comparison of the relevant data for lauryl It is of interest to compare the curve of solu- sulfonic acid is given in reference 2. bilization by the typical soap, potassium laurate, Effect of Added Salts on Solubilization.with the corresponding curves for its osmotic co- McBain and Merrillaj4advanced the plausible efficient, g, and conductivity ratio, a. This is hypothesis that addition of a salt to a colloidal done in Fig. 3. It is seen that solubilization at- electrolyte having a common ion would favor tains its constant value at 0.25 m,while the OS- formation of micelles in the dilute region where motic coefkicient is still definitely falling; whereas they were not already formed and tend to salt the conductivity has already reached its minimum out dye in higher concentrations where the mivalue at 0.2 m and is on its way to the much celles were already completely formed. This they found to be true for solutions of sodium desoxycholate. Table VI11 gives the data for the solubilization of two potassium soaps as affected by the presence 4000 of molal ‘potassium chloride in each case, and the results are plotted in Fig. 4. Solutions of 1 m myristate and 1 m and 0.5 m laurate could not be measured, since the dye particles would not settle in such viscous or jelly-like systems.

li

TABLE VI11 OT IN VARIOUS MOLALITIES OF POTASSIUM LAURATEAND POTASSIUM MYRISTATE AT 25” IN THE PRESENCE OF 1 1?z POTASSIUM CHLORIDE

SOLUBILITY OF ORANGE K LAURATE 4- 1 id KCI

,n

0.25 0.50 i J 7.7 Normality of soap. Fig. 4.--The solubility of orange OT a t 25” ill soluriuns of potassium laurate and potassium myristate with and without the presence of 3 111 potassium chloride 0

0 010 025 050 2ou 3OU

Laurate Mg. dye/ Mg. dye/ 100 cc. soh. mole soap

M ristate Mg. d y e r Mg. dye/ 100 cc. soln. mole soap

1 53 2 TU

1530

8.895

1080

944

5.66

1132 1OYt 10%

13 39 54 01

20 7 2 . i ; i 00

3896 3776 2678 2iOO

I t is seen at once that the potassiuin chloride does greatly enhance the solubilization in dilute

Jan., 1944

IONIZATION, RAMAN SPECTRUM AND VIBRATIONS OF PERCHLORIC ACID

solutions, where but little micelle was previously present. Indeed, the micelle so formed appears to have a greatly increased solubilizing power, not only over the ordinary micelle of the same soap, but even over the same micelle in higher concentrations in the presence of the potassium chloride. Furthermore, even in the concentrated solutions, the potassium chloride has increased the solubilization. This is the opposite of the expected salting out effect. Neither can it be reconciled with the hypothesis of solution in the hydrocarbon, which must be unaffected since it is in equilibrium in all cases with solid crystals of dye. We have no adequate explanation either for the improved solubilizing power in concentrated solutions or for its still greater enhancement in dilute solutions. However, a similar effect in dilute solutions was indicated in Hartley's data for cetylpyridinium chloride with sufficient addition of sodium chloride (see reference 2, page 129). Likewise, Merrill found similar effects with pure %-(laurylcolamino formyl-methyl)-pyridinium chloride, commercially known as Catol607, to which potassium chloride was added. On the other hand, Dr. R. B. Dean in unpub-

[CONTRIBUTION FROM THE

13

lished demonstrations in this Laboratory has shown that potassium chloride reduces the solubilizing power of 0.2 m potassium laurate for capryl alcohol. Addition of this chloride to a clear saturated solution of capryl alcohol in 0.2 rn laurate produces a very evident turbidity, which is cleared by heating or by further addition of potassium laurate solution free from capryl alcohol. Summary 1. The solubilization of water insoluble dye by four potassium soaps has been measured for equilibrium conditions over a range of concentrations. 2. The solubilization increases so rapidly with the higher soaps as to cast doubt upon the suggestion that it is solution in the hydrocarbon fraction of the molecule, but rather to favor its incorporation between the layers of lamellar micelles. 3. Potassium chloride not only greatly increases the solubilizing power of fully formed micelles, but it produces in dilute solution micelles of still higher solubilizing power. STANFORD UNIV., CALIF. RECEIVED SEPTEMBER 2, 1943

DEPARTMENT OF CHEMISTRY OF THE STATE

COLLEGE OF WASHINGTON]

The Ionization of Strong Electrolytes. II. Ionization, Raman Spectrum and Vibrations of Perchloric Acid B Y 0. REDLICH, E. K. HOLTAND

As discussed in the fist paper of this series,2 the Raman spectrum furnishes a fairly general and unambiguous criterion of the existence of molecules. This criterion, in addition, is cogent if one retains the historical concept of the molecule which characterizes the molecule as a mechanical unit without referring to the nature of the intramolecular forces. The principle of determining the degree of ionization by means of Raman spectra, its history, and some experimental details have also been discussed previously. In the present paper, a refinement of the experimental method and measurements on perchloric acid are reported. In a check of the Raman frequencies of anhydrous perchloric acid, a previously missing line was discovered by resolution of a doublet. Since this line completes the spectrum of the simplified model, our results for the frequencies of perchbric acid and the perchlorate ion, and an assignment of the frequencies to the vibration forms are included. The Experimental Method The spectrograph and the general equipment have been described p r e v i o u ~ l y . ~Cells ~ ~ of the (1) Present address: Department of Chemistry, Columbia University, New Yotk, N. Y. (2) 0. Redlich and J. Bigeleism, ~ W I SJOUXNAL, 66, 1883 ( I Y 4 3 ) . (3) 0. Redlich and .I E. Nlelsen, ibid., $6, 654 (1943).

J. BIGELEISEN'

smaller type (18 ml.) were used. The front face of the cells was ground and polished. The perchlorate line 931 cm.-', excited by the blue mercury line was used. Because of the instability of the acid, the violet and ultraviolet light was filtered out by a sodium nitrite solution. Exposure times ranged from three minutes with Eastman Kodak Co. plates 103a-0 and a slit width of 0.10 mm. to twenty-four minutes with plates 11-0and 0.25 mm. Like the nitrate line, the perchlorate line is fairly sharp and narrow in solutions of sodium perchlorate and in the practically completely ionized acid solutions of moderate concentrations, but considerably broader in concentrated solutions of the acid. Several attempts were made to diminish the error involved in the photographic comparison of the intensity of lines which appreciably differ in width. For convenience, the usual way of comparing the areas under the microphotometric curves will be called the first method. This method can be considered to be based on two assumptions: (a) that under appropriate experimental conditions, the microphotometric curves of two lines are equal only if the total intensities striking the plate are equal, and (b) that a microDhotometric curve is sufficiently characterized by its area.