OCt., 1959
DETERMINATION OF C R I T I C A L MICELLE CONCENTRSTION
OF
ASSOCIATION COLLOIDS
1671
A NEW METHOD FOR THE DETERMINATION OF CRITICAL MICELLE CONCENTRATIONS OF UN-IONIZED ASSOCIATION COLLOIDS I N AQUEOUS OR I N NON-AQUEOUS SOLUTION BY SYDNEY Ross A N D J. P. OLIVIER Department of Chemistry, Rensselaer Polytechnic Institute, Trou, New York Received March 6 , 1969
A new method for the determination of critical micelle concentration (CMC) of un-ionized surface-active agents is developed, and shown to be applicable in & number of different solvents. The method de ends on the formation of a colored r ! t different agents and for different iodine-micelle complex. The absorption maximum of the complex is always a t 360 mp solvents. Critical micelle concentrations in water, benzene, carbon tetrachloride and petroleum ether solutions have been determined. The results by the new method are found to a pee with those obta.ined by independent methods, namely, measurements of static surface tension and relative differenti3 refractive index. Colored impurities in the agent are not troublesome except a t high concentrations, and even there an empirical correction can be applied. An isosbestic point in the absorption spectra is evidence that dissolved iodine is in equilibrium with the iodine in the micelle and that only one type of iodine-micelle complex is present in the solution.
Introduction The spontaneous molecular association of ionic surface-active agents in aqueous solution can be detected readily by measurements of electrical conductivity. The two chief advantages of this method lie in the availability of suitable instruments and the absence of any need to add additional substances to the solutions being measured. No experimental technique of comparable ease has yet been discovered for the study of the association of non-ionic surface-active agents, and methods based on solubilization or color changes of dyes are all suspect when the associated micelle is formed at concentrations of agent that are close to the concentration of the added solubilizate or indicator. The latter objection is frequently valid with nonionic surface-active agents, whose tendency to associate may be expected to be stronger than that of ionic agents, because molecules of the former have little or no electrical force to resist aggregation. A new method that is even relatively free from these objections is therefore important for the further study of the non-ionic association colloids. In organic solvents the distinction between ionic and non-ionic agents disappears, and we have no convenient method of detecting the association of the solute. Oil-soluble detergents, such as the petroleum mahogany sulfonates and petroleum naphthenates, are extensively used as additives in lubricating oils and as rust inhibitors both in lubricants and in preservative coatings. These practical applications are handicapped because so little is known about the state of dispersion, and the factors that affect it, of soap dispersions in oil. A method that could be adapted for the investigation of these systems would find a wide range of application. The method that we report is based on the color change of iodine that takes place when non-ionic association micelles are added to an iodine solution.1 The absorption spectrum that results always shows a development of a new maximum absorption peak a t a shorter wave length; the position of this maximum does not vary from one solvent to another. The method is equally applicable, therefore, to the determination of critical micelle con( 1 ) N. A. Allawala and 8.Riegelman, J . Am. Pharm. Assoo., Sei. Ed., 42, 267, 396 (1953).
centration (CMC) in aqueous or in non-aqueous solutions. Iodine also has the advantage of small molecular size compared to the molecule of the surface-active agent; it is less likely, therefore, to affect micelle formation as much as do the large dye molecules customarily used for CMC determination by spectral shift. This supposition has been found to be true in the few cases where the results of an independent determination of CMC of non-ionic agents have been available for comparison. Unfortunately, the large body of information about colloidal electrolytes in aqueous solution cannot be used for this comparison; no interaction takes place between iodine and negatively charged micelles, and a precipitate is formed with positively charged micelles. The method is restricted, therefore, to un-ionized surface-active agents, although agents that are colloidal electrolytes in aqueous solution are brought within the scope of the method when dissolved in non-ionizing solvents. Apparatus and Materials Most of the measurements for this study were made with a Beckman Model B spectrophotometer; a few sets of solutions were measured with the Perkin-Elmer Spectracord. The surface-active agents were all commercial samples of non-ionics; no attempt was made to purify them, or to remove water. The non-aqueous solvents were dried before use; benzene was distilled over sodium; carbon tetrachloride was distilkd. Measurements of static surface tension were made with a Cenco-du Nouy tensiometer. Differential refractive index was measured with a Brice-Phoenix laboratory differentialrefractometer, Model 1890, a t 546 mp. The stock solution of iodine ( A ) was selected to transmit 80% of the light transmitted by the pure solvent; the concentrations required for this purpose are: for water, 30 mg./ 1.; for benzene 25 mg./l.; for carbon tetrachloride, 21 mg./ 1. The stock solution of the surface-active agent (B) contained a known concentration of agent well above its CMC and the same concentration of iodine as in solution A. By diluting solution B with solution A , a series of concentrations of the surface-active agent could be obtained, each with the same concentration of iodine. This series had to include concentrations above and below the CMC to be determined, which usually made necessary a crude preliminary titration of B with A in order to fix an appropriate range of concentrations with which to work. The spectrophotometer readings are best made with solution A as the standard for 100% transmittance, although sometimes the pure solvent can be used. The failure of the reported curves to show 100% transmission in the absence of surface-active agent is caused by imperfect maiching of the cells used; :his effect, however, can have no influence on the determination of the CMC.
1672
SYDNEY Ross AND J. P. OLIVIER
Vol. 63
TABLEI MATERIAL5 USED Trade name
(DESCRIPTION AND SOURCE) AND RESULTS
Description
Brij 30 Brij 35 G-3701 Pluronic L62 Renex 36 Renex 698 Siponic.BC Span 40 Span 40 Span 80 Triton X-100
Source
4 Oxyethylene lauryl ether
Atlas Powder Co. Atlas Powder Co. Atlas Powder Co. Wyandotte Chemicals Corp. Atlas Powder Co. Atla8 Powder Co. American Alcolac Corp. Atlas Powder Co. Atlas Powder Co. Atlas Powder Co. Rohm and Haas Co.
C M C in g./dl.
0.70
23 Oxyethylene lauryl ether 0.0058 1 Oxyethylene lauryl ether 0.036 2.40 HO(C2H4O),( CsH~o)a(CaH40)oH Polyoxyethylene tridecyl alcohol 0.078 Polyoxyethylene alkyl aryl ether 0.0047 Ethoxylated branched chain alcohol 0.011 Sorbitan monopalmitate 0.66 Sorbitan monopalmitate 0.045 Sorbitan monooleate 3.5 CSH&HIO( C~H40)loH,etc.' 0.016; The absorption maxima of iodine in different environ- ues thus determined are reported in the diagram. ments2s*are given: Color transmitted
mN 12
512 610 300,490 450 360 288,353
vapor
1 2 in carbon tetrachloride
ISin benzene rz in water J2-micelle complex IY-ICJcomplex in water8
Violet Violet . Red
Brown Yellow Yellow
3fO mu I
I
IO 0
I I I
310
330
350
370
410
430
P.
450
470
490
510
+
Fig. 1.-Absorption spectra of Renex 648 iodine in iodine in carbon tetrachloride; and benzene; Brij 30 Brij 35 iodine in water. All these spectra have a characteristic maximum a t 360 mp.
+
+
The value of 360 mp for the iodine-micelle complex was found to hold constant with the different agents here tested, and in different solvents. All the spectrophotometer readings, therefore, were taken a t a wave length of 360 mp. In Fig. 1 we report the absorption spectra of three different agents, each above the CMC, and each in a different solvent, to illustrate the constancy of the absorption maximum of the iodine-micelle complex under different conditions of formation. The materials used and their commercial sources are given in Table I, which also reports the results of CMC determinations in various solvents by the present method.
Results In Fig. 2 we report the spectrophotometer readings of the intensity of transmitted light at 360 mp for three non-ionic agents as a function of their concentration in aqueous solution. We interpret the break in the curve as caused by the presence of micelles in appreciable amounts, and the CMC val(2) H. A. Benesi and J. H. Hildebrand, J . Am. Chem. SOC.,70,2832 (1948); 71, 2703 (1949).
(3) A. D. Awtrey and R. E. Connick. ibid., 73, 1842 (1961). (4) L. M. Kuahner and W. D. Hubbard, THISJOURNAL, S8, 1163 (1964).
No particular difficulties were encountered with these agents in the application of the *present.
method. The readings should be taken Pnthin an hour of the preparation of the solutions because of a slow fading of the color of the iodine-micelle complex. At high concentrations of agent (3 to 4 times CMC) the equilibrium is reached more slowly, but errors due to non-equilibrium in this range of concentration do not affect the determination of the CMC. The CMC of Siponic BC, determined by the present method (Fig. 2), is 0,0110 g./dl. Measure ments of static surface tension provide an independent method of finding CMC. The ensuing description6 refers to the determination of CMC by this method: "The static surface tension data when plotted against the log of the concentration give two straight Iines whose point of intersection can be located accurately. The CMC from this method is 0.0120 f 0.0005 g./dl. This value is based on two independent series of determinations. Fresh solutions were prepared for each series. Each series of determinations gave the same CMC." The CMC of Triton X-100, determined by the present method, is reported in Fig. 2 as 0.0158 g./ deciliter. This value can be compared with the CMC determined by the surface tension method by Hsiao, Dunning and lor en^,^ who reported, for a condensate product of octylphenol with a mole ratio of 8.5 ethylene oxide, a CMC of 180-230 pM. If we take a molecular weight of 646 (see formula in Table I) for Triton X-100, the CMC by the present determination is 245 pM, which gives us a reasonably good agreement with the results of the surface tension method. The CMC of Brij 35 has been determined in this' laboratory by Mr. R. C. Little from measurements of static surface tension, and found to be 0.006 g./ dl., which agrees with the value 0.0058 g./dl. obtained by the present method (Fig. 2). The method failed to give a CMC value when applied t o aqueous solutions of Myrj 59 and Myrj 52. The characteristic absorption color was visible, but measurements did not disclose the expected discontinuity in the curve for transmitted intensity a t 360 mp 8s. concentration; instead the development of (6) W. G. Cutler, Whirlpool Corporation Research Labs., private oommunioation. (6) L. Hsiao, H. N. Dunning and P. B. Loreni, THISJOURNAL, 60, 667 (1966).
Oct., 1959
DETERMINATION OF CRITICAL MICELLECONCENTRATION OF ASSOCIATION COLLOIDS 1673
the color seemed to take place gradually and continuously. We have traced this effect in these agents to a shift of the maximum absorption wave length, which is 360 mp only for high concentrations of agent (Le., about 5 7 3 ; on dilution with stock solution A, the absorption peak shifts gradually to 380 mp. A probable explanation is the presence of free polyethylene glycol in the sample. We find that polyethylene glycol alone in aqueous solution forms a complex with iodine, shown by the presence of a new broad and weak absorption maximum in the vicinity of 350 mp. This alone would not interfere with the determination of CMC, but in the presence of the non-ionic agent the effect is greatly enhanced. On adding polyethylene glycol to one of the agents reported in Fig. 2, namely, Brij 35, the same perturbing effect that had been encountered with Myrj 52 and Myrj 59 is reproduced. Apparently an interaction occurs between all three solutes when appreciable amounts of uncombined polyethylene glycol are present with the agent. In Fig. 3 we report the application of the iodine method to two agents dissolved in benzene. For comparison of CMC determinations by an independent method we have used relative differential measurements of the refractive index. For Renex 36 the iodine method gives 0.075 g./dl. compared with 0.10 g./dl. by differential refractive index; for Span 40 the two methods are in closer agreement, viz., 0.GG g./dl. and 0.65 g./dl., respectively. As measurements of the relative differential refractive index are seldom used for this purpose, especially in non-aqueous solution, we include (Fig. 4) the application of the method to Span 40 in benzene solution. It is important with benzene solutions that the comparison solution used in the spectrophotometer also contain iodine, because of a maximum absorption peak of iodine in benzene a t 318 mp. The absorption peak a t 318 mp becomes confused with the 360 mp absorption of the iodine-micelle complex, unless it is eliminated by a proper comparison. For aqueous and carbon tetrachloride solutions this factor does not arise, and it would then be permissible to use the pure solvent without iodine as the comparison solution. Another complication arises if the original agent is itself colored. Our work with Span 80 (sorbitan monooleate) introduced us to this difficulty. The original data failed to give the expected discontinuity. It then was necessary t o measure the transmittance a t 360 mp of Span 80 solutions in benzene in the absence of iodine, as a function of concentration of the agent. These data provided a correction curve for the light absorption due to the colored constituents in the Span 80; by means of this curve the readings for the iodine-micelle complex were corrected Total '% transmission Corrected % transmission = 70Transmission of agent x 100 the discontinuity then became apparent. We have applied the method to agents dissolved in carbon tetrachloride. This solvent disdavs the greatest change in absorption spectrum,- a