J. Chem. Soc., Faraday Trans. 1, 1982, 78, 1369-1376 Absorption Spectra and Micellisation of a Surface-active Dye in Aqueous Methanol Solutions TOYOKO IMAE, CHISAKO MORI AND SHOICHI IKEDA* Department of Chemistry, Faculty of Science, Nagoya University, Chikusa, Nagoya 464, Japan Received 19th January, 198 1 The absorption spectra of aqueous methanol solutions of a surface-active dye, p-t-octylphenol yellow amine poly(ethy1ene oxide), have been measured at different dye concentrations, and the absorption bands are assigned to tautomeric forms of the 2’-hydroxy-4-dimethylaminoazobenzene derivative. The K band of the dye appears at 394 nm in methanol+ water solutions when the dye concentration is lower than mol dmP3. Hypochromism occurs and the shoulder band (B band) at 460-480 nm becomes manifest when the primary micelle is formed: the primary micelle is stabilised by the stacking interaction of dye molecules and the increased amount of their o-quinoid hydrazone tautomer.At higher dye concentrations the K band appears at 413 or 406 nm and its intensity increases. This step can be attributed to the second micellisation of dye molecules due to hydrophobic interaction. The methanol+O. 1 mol dm-3 HC1 solutions of the dye exhibit two absorption bands (K’ and Q bands) at 320 and 528 nm, which can be assigned to the ammonium form and the resonance-stabilised azonium tautomer, respectively, of the protonated dye, when the dye concentration is dilute. Above the critical micelle concentration the absorption spectra have the K band at 405 nm, indicating that the dye is deprotonated in the micelle.It has been observed that a surfactant with a chromophoric group undergoes a sudden change in the absorption spectra of its aqueous solutions when it associates into micelles as the concentration increases. Harkins et a1.l were the first to observe this phenomenon for dodecylpyridinium iodide, and Ray and Mukerjee2v measured its absorption spectra in different solvents. Ikeda and Fasman4 found a sudden increase in the molar extinction coefficient of the absorption bands of polyoxyethylene p-t-octylphenyl ether above its critical micelle concentration, and Rehfeld5 observed an abrupt decrease in the absorption intensity of sodium phenylundecanoate upon micellisation. The surface-active dye, p-t-octylphenol yellow amine poly(ethy1ene oxide), is expected to show some spectral changes when micelles are formed in aqueous solutions, because the molecule has an azo group sited between two phenyl groups.From measurements of the surface tension of aqueous solutions of the dye it was found that two-step micellisation occurs for the dye in methanol +water solutions, while the common micelle is formed in methanol + 0.1 mol dm-3 HCl solutions.6 The dye is a derivative of 2’-hydroxy-4-dimethylamino-azobenzene, and the complicated behaviour of the absorption spectra of its aqueous methanol solutions must be understood by referring to its tautomeric forms and resonance. In order to obtain detailed information on the structure of the micelles from the electronic structure of molecules in micelles, we measure the absorption spectra of the dye in methanol + water and methanol + 0.1 mol dmP3 HCl solutions, together with the spectra in various organic solvents, and examine the micellisation of the dye as welQ as its structure.45 1369 FAR 11370 ABSORPTION SPECTRA OF A SURFACE-ACTIVE DYE EXPERIMENTAL p-t-Octylphenol yellow amine poly(ethy1ene oxide) samples were the same as previously used,6 and were kindly donated by Dr F. Tokiwa of the Kao Soap Co. Ltd. They had an average degree of polymerisation of the polyoxyethylene parts of x + y = 10 and 20, respectively. Water, dioxan, 1,2-dichloroethane and 1,2-dibromoethane were redistilled before use, after the necessary purifications. Benzene, chloroform, ethyl acetate and concentrated HCl solution were special-grade reagents.The other organic solvents were spectrograde reagents of the Nakarai Chemical Co. Ltd. The dye was sparingly soluble in water, ethyl ether and cyclohexane. The method of preparation of the aqueous methanol solutions of the dye was the same as previously described.6 The absorption spectra were measured on a Shimadzu UV-200s spectrophotometer and recorded on a U-l25MU recorder. Quartz cells with path lengths of 10, 5, 2 and 1 mm were used. The temperature of the cell chamber was adjusted to 25 f 0.05 O C by circulating water of constant temperature from a Haake thermobath FS. RESULTS The absorption spectra of the surface-active dye having x+y = 10 in 2 and 20% methanol+water solutions are illustrated in fig. 1 and 2, respectively.At concentrations lower than mol dm-3 the absorption spectra have a band at 0' I I I I I 350 400 4 5 0 500 550 w avelengt h/n m FIG. 1.-Absorption spectra of the dye having x+y = 10 in 2% methanol+water solution. Concentration/mol dm-3: -, 4.66 x lop6; ---, 1.86 x lop5; ---, 7.76 x . . . , 6.21 x 394 nm with a shoulder at ca. 450 nm. The spectra of the dye show a feature of the 4-aminoazobenzene derivative in aqueous ethanol solutions, and the main band termed the K band7-10 can be assigned to the lowest n-n* transition of azo d ~ e . ~ ? l l The shoulder band appearing at 460-480 nm cannot be attributed to the n-n* transition of azo dyes since its molar extinction coefficient is too high. In fact, the spectral feature is more similar to that of 2-hydroxyazobenzene12~ l3 and 1 -phenylazo- 2-naphtholl49 l5 in aqueous ethanol solutions, which have a shoulder at wavelengths 50-70 nm longer than the K band.The subsidiary band, originally termed the B band,13 can be assigned to the hydrazone tautomer, as will be described later. With increasing concentration of the dye, the K band shifts to the red and the B band is stronger, but both bands are hypochromic. The concentration dependence of the wavelength of the main bands is shown in fig. 3. The spectra do not change withT. IMAE, C. MORI A N D S. I K E D A 1371 dye concentration in 50% methanol + water solutions. The red shift and hypochromism occur at the range of concentrations around the first c.m.c. and may be attributable to the stacking interaction of the dye molecules.We have previously6 found that the primary micelle has an aggregation number of, at most, 15 and consequently we can imagine that it is composed of a stack of dye molecules. 2 - I I 1 I . . . . . FIG. 2.-Absorption spectra of the dye having x+y = 10 in 20% methanol+water solution. Concentration/mol dm-3: -, 3.82 x ---, 3.05 x lop5; - . - . -, 6.36 x 10-5; ---, 1.27 x 10-4; . . . . -, 5.09 x 10-4. 5 420 x" 400 . 2 - 6 - 5 - 4 - 3 log (C/mol dm-3) FIG. 3.-Wavelength of the main band of the dye having x+y = 10 plotted against concentration in methanol+water solutions. Methanol content (volume %): 0, 1 ; @, 2; 0, 5; 0, 10; (0, 20; $, 50. In the region of the second c.m.c. or exceeding it the K band remains at 413 or 406 nm. The observed spectral shift will be caused by the facts that the molecules in the secondary micelle are free from hydrogen bonding with solvent, as shown below, and they are in a hydrophobic environment.Fig. 4 shows the absorption spectra of the surface-active dye having x+y = 10 in 2% methanol+O.l mol dm-3 HCl solutions. The dye is protonated in these solvents 45-21372 ABSORPTION SPECTRA OF A SURFACE-ACTIVE DYE in which the apparent pH is 1.1, since the pK of 4-dimethylaminoazobenzene is ca. 3.5 in water16 and 2.2 in 50% ethan01.~ The spectra have bands at 320 and 528 nm at concentrations < mol dm-3, and they are similar to those observed for amino- substituted azobenzenes in acid media.99 lo, 1 7 9 l8 The two bands can be assigned as the K' and Q bands, which are attributed to the ammonium form and the resonance- stabilised azonium form, respectively, as will be discussed later.When the dye concentration exceeds l 0-4 mol dm-3 the absorption spectra change drastically and the two bands are replaced by a single band at 405 nm having a shoulder at ca. 480 nm. The main band at 405 nm can be identified with the K band, and in this concentration range the micellisation is manifest from the surface tension I ' I I 1 - 0 E € 1 2 m P 0 1 n W I I 300 400 500 600 wavelength/nm FIG. 4.-Absorption spectra of the dye having x + y = 10 in 2% methanol+O.l mol dm-3 HCl solution. ---, 9.35 x Concentration/mol dm-3: -, 1.87 x ---, 4.67 x . . . . ., 9.35 x C W Vl ? n 2.5 2.0 1.5 ~~ - 5 - 4 - 3 log (C/mol dm-3) FIG. 5.-Molar extinction coefficient of the main bands of the dye having x+y = 20 plotted against concentration in methanol +O.1 mol dm-3 HCI solutions. Top: the Q band; bottom: the K band. Methanol content (volume %): (>, 2; a, 5 ; a, 20.T. IMAE, C. MORI A N D S. I K E D A 1373 measurements.G Thus the surface-active dye must be deprotonated and non-ionic in the micelle, even if the outside media are strongly acidic and the free dye molecules are totally protonated. The dye having x+y = 20 behaves similarly to that having x+y = 10, in both methanol + water and methanol + 0.1 mol dm-3 HC1 solutions. Fig. 5 illustrates the molar extinction coefficients of the two bands at ca. 400 and 530 nm plotted against the dye concentration in methanol + 0.1 mol dmP3 HCl solutions. The plots clearly show the presence of a concentration where the molar extinction coefficient at 530nm suddenly decreases at each methanol content.The break points, (7.1, 9.8 and 25) x mol dm-3 at 2, 5 and 20% methanol contents, can be well compared with the c.m.c. found by the surface tension measurements, (8.02, 10.6 and 34.5) x mol dm-3. TABLE WAVELENGTH OF THE K BAND OF THE DYE IN VARIOUS ORGANIC SOLVENTS carbon tetrachloride benzene toluene c yclohexane ethyl ether acetone ethanol methanol water ethyl acetate dioxan chloroform 1,2-dichloroethane 172-dibromoethane 382.5 387 388 377 388 395 396 395 (394) 39 1 390 385 388 388 - 415 417 41 5 420 398 379 40 1 417 412 - - 20 -21 - 20 ( - 26) -7 1 1 - 16 -31 - 24 Table 1 shows the wavelength of the K band of the dye having x+y = 10 dissolved in various organic solvents.It was found that the wavelength, A, of the K band in five solvents, i.e. ethyl ether, toluene, benzene, carbon tetrachloride and cyclohexane, follows the McRae equation19 - = - + A ( + - ) 1 1 z2 -1 (-L) D-1 C 2 - 1 a aG 2fiD+1 + B D+2 iib4-2 where aG is the wavelength of the K band of the dye in the hypothetical gaseous state, 6, is the refractive index of the solvent at the D line and D is the dielectric constant (relative permittivity) of the solvent. The parameters, aG, A and B, have been evaluated (2) as Table 1 gives the wavelength of the K band calculated for the other solvents using eqn (1) and (2). Although some polar halogenated hydrocarbons must be excluded, the deviation, A1 = Lobs-acal, can be taken as representing the effect of hydrogen bonding between the dye and solvent.In some strongly hydrogen-bonding solvents such as alcohols and acetone, the observed wavelength is 20 nm blue shifted from the calculated one. The calculated wavelength for these solvents, ca. 415 nm, is AG = 309 nm, A = - 28 576 cm-l, B = 4830 cm-l.1374 ABSORPTION SPECTRA OF A SURFACE-ACTIVE DYE approximately equal to the observed wavelength for the secondary micelle. Thus we can imagine that the dye molecules are free from hydrogen bonding in the secondary micelle and are effected only by the polarity of the environment within the micelle. DISCUSSION STRUCTURE OF THE DYE I N NEUTRAL SOLUTIONS Since the stability or presence of the tautomeric forms and resonance species of 2-hydro~yazobenzene~~9 1 3 9 20* 21 and 4-amino- or 4-dirnethylamino-azoben~ene~~~ 18* 22 has been subject to controversy, and since neither 2’-hydroxy-4-aminoazobenzene nor 2’-hydroxy-4-dimethylaminoazobenzene has ever been described, we examine the tautomeric equilibrium and resonance of the surface-active dye in aqueous methanol solutions on the basis of the absorption spectra observed.APPEARANCE OF THE SHOULDER BAND AT 460-480 nm 0spensonl2 and Burawoy and Chamberlain13 showed that 2-hydroxyazobenzene exhibited two absorption bands in aqueous ethanol solutions, one being the strong K band’. l2 and the other the B band at a higher wavelength.l29 l3 For l-phenylazo- 2-naphthol the K band is weaker relative to the B band.l2, 1 3 9 l5 The K band has been assigned to the azo form and the B band to its o-quinoid hydrazone form.l2? 1 4 9 1 5 9 23 In a similar way we may postulate the keto-enol tautomerism of 2’-hydroxy- 4-dimethylaminoazobenzene represented by Q--N;.p-Q--NRR~ / \ == Q N - y o N R R / 0-H O b b e H (I) in which both the azo (I) and o-quinoid hydrazone (11) forms are associated with the intramolecular hydrogen Evidence for the formation of o-quinoid forms of some derivatives of 4- aminoazobenzene can be found in their absorption spectra.Ross and Warwick17 reported that 2’-hydroxy-5’-ni tro-4-di(2-chloroethyl)aminoazobenzene exhibited only the B band in 95% ethanol, which we interpret as the complete shift of the keto-enol equilibrium towards the hydrazone form. They also found that the introduction of a 2’-carboxyl group in 4-dimethylaminoazobenzene or 4-N-methyl-N-(2-chloro- ethy1)aminoazobenzene enhanced the intensity of the B band and they ascribed this effect to the stabilisation of the p-quinoid form.The resonance Raman spectra of aqueous solutions revealed that Tropaeolin 0, i.e. 2,4-dihydroxyazobenzene-4’- sulphonic acid, can assume either or both the quinoid forms (0- and p-) at neutral pH.25 The present surface-active dye has a characteristic shoulder band at 460-480 nm, in addition to the K band, in methanol+water solutions, while the shoulder is not observed in methanol nor in any other organic solvents. As we have referred above to the tautomeric equilibrium of the derivatives of 2’-hydroxy-4-aminoazobenzene, the B band of the dye can be interpreted as being due to the formation of the o-quinoid hydrazone form.Thus we can conclude that the tautomeric equilibrium is displaced towards the azo form in methanol or other organic solvents, while it is considerably shifted towards the hydrazone form in methanol + water solutions. Note also that the hydrazone form is more stabilised in the primary micelles than in the monomer form or in the secondary micelles.T. IMAE, C. MORI AND S. IKEDA 1375 SUPPRESSION OF THE RED SHIFT OF THE K BAND IN WATER Forbes and Milligan18 reported that 4-dimethylaminoazobenzene had the K band at 406 nrn in ethanol and at 446 nm in water, and that Methyl Orange, i.e. 4-dimethylaminoazobenzene-4’-sulphonic acid, had the K band at 463 nm in water, which was shifted to the red by 45 nm from the K band in ethanol. They proposed various modes of solvation of the 4-aminoazobenzene derivatives and attributed the spectral difference to the formation of a hydrated form that could stabilise a polarised resonance species through an equilibrium.Brode et aZ.22 also explained the spectral shift by postulating the formation of a hydrated form of the dye which was hydrogen- bonded with water at the azo nitrogen atom. On the other hand, the surface-active dye has the K band at 394 nm in both methanol and methanol+water solutions. This behaviour of the dye can again be explained by the effect of substitution of a 2’-hydroxyl group. From the scheme by Forbes and Milligan, it is likely that the substitution of a 2’-hydroxyl group prevents the azo-hydrated form of the dye from being formed, because the azo group is protected from water molecules by its intramolecular hydrogen bond.Then some of resonance species, such as the polarised one, which should be more stabilised in more polar solvents, are not formed even in aqueous solutions, and the number of resonance species remains the same in aqueous solutions as in methanol. Thus the location of the K band of the dye is not influenced by the nature of the hydrogen-bonding solvent. MICELLISATION OF THE DYE IN NEUTRAL SOLUTIONS The surface-active dye in methanol + water solutions shows spectral changes in two steps, approximately following the two-step micellisation.6 The first step includes the red shift and hypochromism of the main bands, which can be attributed to the stacking interaction of dye molecules, leading to the formation of the primary micelle.The second step is characterised by a further spectral shift of the K band and a steady increase in intensity of the main bands. These changes can be ascribed to the release of an azo chromophore from hydrogen bonding and its transfer into the more hydrophobic environment. The former can be seen from the results of eqn (1) and (2) in table 1. The latter is supported by the observation that the K band of 4-dimethylaminoazobenzene shifts from 420 nm to 401 or 398 nm on going from 50% ethanolg* lo to cyclohexane18 or iso-octane,26 if the spectral shift is regarded as a blue shift on going from the primary micelle to the secondary micelle. The B band becomes stronger relative to the K band upon the first micellisation, indicating formation of the o-quinoid form, but it becomes weaker at the second micellisation.STRUCTURE AND MICELLISATION OF THE DYE IN ACIDIC SOLUTIONS The two absorption bands exhibited by the dye in acidic solutions can be attributed to the free protonated dye, i.e. the monomeric cation, which can assume tautomeric isomers and resonance species as shown by + + Q . N = J O ! R ~ - i_ Q ~ + ~ N R R / t-) Q - ~ - ; ~ ~ ~ R I 0-H 0-H 0--H (rm (IYa 1 ( IYb 1 While the K’ band at 320 nm is assigned to the ammonium form (111), the Q band at 528 nm is associated with the azonium form (IVa), which is stabilised through resonance with the p-quinoid immonium species (IVb). The formation of p-quinoid1376 ABSORPTION SPECTRA OF A SURFACE-ACTIVE DYE species of 4-dime thylaminoazobenzene derivatives in acidic solutions was demonstrated by means of absorption spectra9? lo, 1 7 9 27 and resonance Raman 28 In acidic solutions it is uncertain whether or not the o-quinoid hydrazone form would arise from the ketwmol tautomerism.mol dm-3 has been interpreted as being due to the micellisation accompanying deprotonation of the dye molecules. Thus the K band appearing at ca. 405 nm means that the dye is non-ionic in the micelles. The pK value of the dye is much lower in the micelles than in the monomeric cations. From the spectral and surface tension behaviour, the micelle of the dye in acidic solutions would have a structure similar to the secondary micelle in neutral solutions. A similar but opposite effect of micellisation on the protonation reaction was also found in the case of dimethyldodecylamine oxide:29 it has a pK value of 4.78 in the free cationic form in water, while it becomes less acidic in the micelle, giving an intrinsic pK value ca.0.85 pH units higher.30 More indirect micellar effects on acid-base equilibria of pH indicators have been the subject of recent investigations to elucidate the polarity and electrostatic potential of m i ~ e l l e s . ~ ~ ~ ~ ~ When Methyl Orange in acidic solutions forms a complex with octadecyltrimethylammonium chloride, in which the surfactant concentration is generally higher than the c. m. c., the deprotonation of Methyl Orange cations occurs, as revealed in the absorption It is likely that the pK value of a N-dimethylamino group is lowered by the micellar The drastic change in spectra of the dye at 33 W.D. Harkins, H. Krizek and M. L. Corrin, J. Colloid Sci., 1951, 6, 576. A. Ray and P. Mukerjee, J. Phys. Chem., 1966, 70, 2138. P. Mukerjee and A. Ray, J. Phys. Chem., 1966, 70, 2144. S. Ikeda and G. D. Fasman, J. Polym. Sci., Part A-I, 1970, 8, 991. S. J. Rehfeld, J, Colloid Interface Sci., 1970, 34, 518. ti T. Imae, C. Mori and S. Ikeda, J. Chem. Soc., Faraday Trans. I , 1982, 78, 1359. A. Burawoy, J. Chem. Sac., 1937, 1865. P. Birnbaum, J. H. Linford and D. W. Style, Trans. Faraday SOC., 1953, 49, 735. G. M. Badger, R. G. Buttery and G. E. Lewis, J. Chem. Soc., 1954, 1888. lo G. E. Lewis, Tetrahedron, 1960, 10, 129. l1 M. B. Robin and W. T. Simpson, J. Chem. Phys., 1962, 36, 580. l2 J. N. Ospenson, Acta Chem. Scand., 1951, 5, 491. l3 A. Burawoy and J. T. Chamberlain, J. Chem. Soc., 1952, 3734. I4 J. N. Ospenson, Acta Chem. Scand., 1950, 4, 1351. l5 A. Burawoy, A. G. Salem and A. R. Thompson, J. Chem. SOC., 1952, 4793. l6 I. M. Klotz, H. A. Fiess, J. Y. Chen Ho and M. Mellody, 1. Am. Chem. SOC., 1954, 76, 5136. l7 W. C. J. Ross and G. P. Warwick, J. Chem. SOC., 1956, 1719. la W. F. Forbes and B. Milligan, Aust. J. Chem., 1962, 15, 841. 2o A. Burawoy and I. Markowitsch, Liebigs Ann., 1933, 503, 180. 21 D. Hadii, J. Chem. SOC., 1956, 2143. 22 W. R. Brode, I. L. Seldin, P. E. Spoerri and G. M. Wyman, J. Am. Chem. SOC., 1955, 77, 2762. 23 R. Kuhn and F. Bar, Liebigs Ann., 1935, 516, 413. 24 S. B. Hendricks, 0. R. Wulf, G. E. Hilbert and U. Liddel, J. Am. Chem. SOC., 1936, 58, 1991. 25 Y. Saito, B-K. Kim, K. Machida and T. Uno, Bull. Chem. SOC. Jpn, 1974, 47, 21 1 1 . 26 W. R. Brode, J. H. Gould and G. M. Wyman, J. Am. Chem. Soc., 1953,753, 1856. 27 A. Hantzsch and A. Burawoy, Chem. Ber., 1930, 63, 1360. 28 K. Machida, B-K. Kim, Y. Saito, K. Igarashi and T. Uno, Bull. Chem. SOC. Jpn, 1974, 47, 78. 28 F. Tokiwa and K. Ohki, J. Phys. Chem., 1966, 70, 3437. 30 H. Maeda, M. Tsunoda and S. Ikeda, J. Phys. Chem., 1974, 78, 1086. 31 M. S. Fernandez and P. Fromberg, J. Phys. Chem., 1977,81, 1755. 32 C. A. Bunton, L. S. Romsted and L. Sepulveda, J. Phys. Chem., 1980, 84, 261 1 . 33 C. F. Hiskey and A. T. Downey, J. Am. Chem. Soc., 1954, 58, 835. E. G. McRae, J. Phys. Chem., 1957, 61, 562. (PAPER 1 /085)