J. CHEM. SOC. FARADAY TRANS., 1994, 90(4), 629-640 Influence of the Cetyltrimethylammonium Chloride Micellar Pseudophase on the Protolytic Equilibria of Oxyxanthene Dyes at High Bulk Phase Ionic Strength Nikolay 0. Mchedlov-Petrossyan' and Valentina N. Kleshchevnikova Department of Chemistry, Kharkov State University, Kharkov, 310077,Ukraine The acid-base equilibria of fluorescein, sulfonefluorescein, 2,7-dichlorofluorescein, eosin and ethyl eosin have been studied spectrophotometrically in the micellar pseudophase of cetyltrimethylammonium chloride, the charge of micelles being strongly shielded by high counter-ion concentrations in the bulk phase (4.00 mol dm-3 KCI). The 'apparent' ionization constants (K:) and molar absorptivities of all the species have been determined.The completeness of solubilization is confirmed by several approaches, including examination of emission and excitation spectra. Conclusions concerning tautomerism of dye molecules and ions were deduced from absorp- tion spectra; the fractions (a) of tautomers as well as apparent microscopic ionization constants (k")have been evaluated. The medium effects, ApKZ (=pK: -pK:), vary from -1.5 to +2.0and are partly controlled by shifts in the tautomeric equilibria. For the Apk" values (= pk" -pk") the relationship Apk,,,, > ApcC&, is valid. Con- sidering the pKz values at 4.00 mol dm-3 KCI as the first approximation to the 'intrinsic' values (pK!J, and comparing them with the pK,s in aqueous acetone and other solvents, although the influence of micellar micro- environment is qualitatively similar to that of the organic solvents, there are some doubts about the possibility of modelling the values of the whole set of parameters in the pseudophase by that in a 'unique' water-organic compound mixture.Micellar solutions of cationic surfactants are widely used as solvents for various chemical reactions. 'v2 Acid-base proper- ties of substances solubilized by micelles are markedly differ- ent from those in aqueous solution^.^-^ Acid-base indicators are useful probes for the investigation of mi~elles,~*~9~-~since they are generally believed to be located in the Stern layer of ionic miceile~.'*~*~.~.~ The principal parameter measured in such studies is the so-called 'apparent' ionization constant, Ki,,435*8 which is obtained via standard spectrophotometric methods ; the pH values utilized in calculations characterize only the bulk phase and are as a rule measured with a glass electrode.Such systems are actually microheterogeneous ; the dispersed micellized surfactant is considered as a micellar pseudo- phase., The electrostatic approa~h~-~~~*'*-'~ permits the pKP value for the given completely solubilized acid-base indicator couple, HB/B, to be explained in terms of transfer activity coefficients (7) and the potential of Stern layer (Y): YB f"pKf = pK," + log -+ log 2--YF (1)YHB fEB 2.303RT where K," is the thermodynamic ionization constant in water, F is the Faraday constant, R is the gas constant, T is abso- lute temperature and f "are activity coefficients of solubilized species.f" values are not readily available.Since the Stern layer is assumed to be a concentrated salt solution we, as is widely ass~med,~.~.~~*'take fEfEBx 1. The so-called 'intrinsic' ionization constant, Ki , describes the two-phase acid-base equilibria:" PKi = PK: + log(YB/Y,B) (2) The ionization constant in the micellar pseudophase, Kr = (HL)(B,)/(HB), ,is related to Ki by PKF = PK: + log YH (3) Eqn. (3) is often used in attempts to compare the solvent properties of micelles with those of water-organic compound mixture^.^*^* In the case of the micelles of cetyltrimethyl- ammonium chloride (CTAC) and bromide (CTAB), dodecyl- trimethylammonium chloride (DTAC) and bromide (DTAB) as well as cetylpyridinium bromide (CPB) at low bulk phase ionic strength (d0.05mol dm-3) Y > 90 mV.8.'0*13-'6 Th e ApK: values (= pKi -pK,") are usually negative [sometimes even -(3-4) pK, units"].With increased ionic strength of the bulk phase, shielding of the micellar charge occurs, which results in Y decreasing, and the pK: values increasing. The same effect for pK: is explained in the pseudophase ion-exchange mode12.6*10 in terms of the corresponding equilibria shift, e.g.: OH; + Cl; eOH; + Cl, (1) The suffixes m and w denote the micellar and water phases, respectively. The number of counter-ions per micellized sur- factant ion is regarded as constant,2*6*10 but there is evidence that this is not always the case.' At high salt concentrations in the aqueous phase, e.g.4 mol dm-3 Na(K)Cl(Br), the exact residual Y values are, strictly speaking, unknown, but according to some data, they are rather small (e.g.+47 mV for DTAC and + 18 mV for DTAB16). Numerous reports show that under such condi- tions the apparent acidity of solubilized reagents, HB, decreases by ca. 2 orders of magnitude; the pKi values are strongly shifted toward the pKi values.' 'vl' Information con- cerning ionic equilibria in the micellar pseudophase with a (partly) reduced potential is sparse. The present investigation was undertaken to study the pro- tolytic equilibria of fluorescein and its derivatives in the cetyl- trimethylammonium micellar pseudophase at a bulk phase ionic strength of 4.00 mol dm-3 KCl.These dyes, possessing different functional groups (OH, CO,H), ionize in a stepwise manner : H3R+ =H,R + H+ ; K,, (11) H,R=HR- + H'; K,, (111) HR-$RZ-+ H'; K,, (IV) and also undergo tautomeric interconversion (in the case of H,R and, sometimes, HR -). Tautomeric interconversion is quite sensitive to the nature of the solvent and to substituent effects. Previously, we examined in detail the protolytic equi- 630 J. CHEM. SOC. FARADAY TRANS., 1993, VOL. 89 1 100- 100- iI 1 I I I ! ! r c I I 1 I 5 I E II E c I-7 50 E -0 m2--. 15'->1: \ lo\ l/n m 440 460 480 500 520 A/nm Fig. 1 Absorption spectra of fluorescein in CTAC micellar solutions (4 rnol dmP3 KCl); pH = 9.55 (1, spectrum of R2-,7a), 7.94 (2), 7.32 (3), 7.18 (4), 7.02 (9,6.78 (6),6.28 (7), 6.02 (8), 5.86 (9), 5.10 (10); pH, = 0.70 (ll), 0.40 (12), 0.22 (13), -0.40 (14,near to spectrum of H,R+, (la); dotted line: spectrum of HR-(5a), obtained using eqn.(5) libria of fluorescein and its derivatives in in mix- tures of water with acet~ne,~l-~~ and1,4-dio~ane~~~~' DMS0,26*27as well as in micellar solutions of non-ionic surf act ant^^^,^^ and of cetylpyridinium chloride (CPC) at low ionic strength (0.05 mol dmP3 KCl).30-32 Information about the acid-base behaviour of oxyanthene dyes in water in the Fig. 2 Absorption spectra of 2,7-dichlorofluorescein in CTAC micellar solutions (4 mol dm-3 KCI); pH = 9.10 (1, spectrum of R2-, 7b), 7.05 (2), 6.16 (3), 5.95 (4), 5.72 (5), 5.55 (6), 5.25 (7), 5.09 (8)4.89 (9), 4.67 (lo),4.07 (ll), 3.47 (12); dotted line: spectrum of HR-(5b=6b), obtained using eqn.(5) ments were performed at 25.0 & 0.1 "C on a P 363-3 poten- tiometer and pH-121 pH-meter equipped with a ESL-43-07 glass electrode and an Ag/AgCl reference electrode in a cell with liquid junction (3.5 mol dm-3 KCl). Standard buffers (pH 4.01, 6.86 and 9.18) and dilute HCl solutions were used for cell ~alibration.~~ Experimental error did not exceed 0.02 pH units. The dyes were purified as described Cetyltrimethylammonium bromide (Serva, Berlin) was used as received; its purity was determined by two-phase titration withpresence of surfactants is also available in the literat~re.~~-~~ The ionization of lipoidal fluorescein and eosin in a micellar microenvironment l4 as well as the spectra of 2,7-dichloro-fluorescein in reversed anionic micelle~~~ was also described.The aim of the present paper was to obtain the whole set of equilibrium parameters of dyes including the microscopic ionization constants in micellar solutions at high counter-ion concentration and to compare them with the data referring to aqueous and organic media, as well as with those obtained in CPC micellar solutions at low Cl, c~ncentration.~'-~~Such a system (CTAC-4 mol dm-3 KCI) was chosen because the size of C 6H33NMe,f micelles with increasing counter-ion concentration is known to be much smaller in the case of C1-than for Br-.2*37 Experimental Apparatus and Reagents Absorption spectra of dye solutions were measured using SP-46 apparatus (of USSR origin).Emission and excitation spectra were obtained with a Hitachi F-4010. pH measure- a standard solution of sodium lauryl sulfate (H,O-CHCl, , using erythrosin as indicator). Suitable pH values were provided with analytical-grade reagents : hydro-chloric and acetic acid, sodium hydrophosphate and borax ; the standard sodium hydroxide solution was prepared using C0,-free water. Potassium chloride of high quality was further purified by recrystallization. Procedure Measurements were carried out at surfactant concentrations c, = 3 x loP3rnol dm-3. The completeness of CTAB trans- formation into CTAC in 4 mol dmP3 KCl solution results from the value of the ion-exchange constant, Kt;: Cl; + Br; eC1; + Br, ; Kt; (V) According to the various data including those obtained at high ionic strength the K$ values are within the interval 2.7- 5.0.39*40The c.m.c. of CTAC, (1-1.8) x mol dm-3 in water, decreases by more than an order of magnitude at the ionic strength ~hosen.~'.~~ While calculating the quantity of KCl needed for a total ionic strength of 4.00 mol dm-3, the J. CHEM.SOC. FARADAY TRANS., 1994, VOL. 90 63 1 100 r I E c I-0 E m 50 m0 F.. PH PH Fig. 3 Absorption of the dyes in CTAC micellar solutions (4 rnol dm-j KCI) us. pH. (a)fluorescein, 450 nm (1); fluorescein, 500 nm (2); sulfonefluorescein,450 nm (3); sulfonefluorescein, 5 10 nm (4);(b)2,7-dichlorofluorescein,5 15 nm (1); 2,7-dichlorofluorescein,535 nm (2); 2,7-dichlorofluorescein,E (530 nm) + E (535 nm) instead of E (3); eosin, 525 nm (4); eosin, 545 nm (5).ionic strength resulting from the buffer mixtures (acetate, phosphate, borate), not exceeding 0.04 mol dm-3, was taken into account. Hydrochloric acid was used to prepare solu- tions having pH < 3.5, the sum of KCl and HCl concentra- tions being maintained constant (4.00 mol dm-3). Based upon values of ion-exchange constants = 0.50, K;ro4 = 0.78 and K307= 0.7939 the replacement of Cl, by buffer anions is unlikely. In the present study, it was assumed that the pH measurements gave the hydrogen ion activity in the bulk: (H'), = h = In acidic media (pH d 1) the scale pH, = -log cHClwas used in calculations.All experi- ments were carried out at 25 "C. The working solutions were nearly transparent and all spectra were referenced against solvent blanks containing all components except the dyes. The path length was 1-5 cm. During pK: determination concentrations of the dyes were held constant at ca. 8 x mol dmP3 (further lowering leads to the same E values at given pH), but while determin- ing &HrR of fluorescein and 2,7-dichlorofluorescein they were increased to (4-8) x lo-' mol dm-3. There was no indica- tion that the measured results depended upon the increased CTAC concentration. Emission and excitation spectra were measured at dye concentrations of 2 x mol dm-3 (pH 9-1 1).Representative pH-dependences of absorption spectra are depicted in Fig. 1 and 2. The number of different solutions used for pK: determinations was 50, 42, 28, 24 and 13 for fluorescein, 2,7-dichlorofluorescein, eosin, sulfonefluorescein and ethyl eosin, respectively. In the last two cases analytical positions near Amax of intensely coloured species were used for the pK: calculations, while for more complicated equilibria a greater number of wavelengths (15-1 8) were utilized. Typical &-pH dependences are given in Fig. 3. Results Calculation of Ionization Constants and Molar Absorptivities The calculation of pK,, and pK,, of fluorescein from the &-pa: curves in H,O-DMSO mixtures was recently described in The approaches, based on utilizing the whole data at various (with the help of a computer were used for final pK:calculations in the present study.For 2,7- dichlorofluorescein the calculations were carried out analo- gously, except those approaches which are based upon the individual peculiarites of fluorescein The Thamer- Voigt method43 at appropriate A was used to obtain the first approximations of K:, and K:, for further iterations. Differ- ences or sums of E values at various 1 were also used (instead of E) to obtain more marked maxima on the 'titration curves' (Fig. 3). For sulfonefluorescein only equilibria (111) and (IV) were taken into acc~unt,~~,~'-~~ while for ethyl eosin equi- librium (VI) is valid: HReR-+ H+; K,, (VI) Here the calculations were made by the standard procedure. The pKf, value of fluorescein was obtained by treating equi- librium (11) as isolated.The sums of molar absorptivities at 440,445 and 450 nm for solutions from pH, = 0.00 to pH, = 0.70 were used in the standard formula (4) pK:O = pHc + log[(& -&HzR)/(&H~R+ -(4) while the sum of E of the solution with cHCl= 2.5 mol dmP3 and cKC, = 1.5 mol dmW3 at 435, 440 and 445 nm, multiplied by 1.086, was used as &H3R+. Thus corrections were made for the solvatochromic shift of the H3R+ band (see below) and incomplete transformation of the dye into the cationic form at pH, = -0.4 (E,,, = 50 x lo3, while in water at pH, d 0 E,,, = 54.3 x lo3 dm3 mol- cm-').The results are presented in Table 1 together with the pKr values. The pK: values in CPC micelles (0.05 mol dm-3 KCl), obtained previo~sly,~~-~~ are also tabulated. Direct compari- son of our results with those obtained for octadecanoyl aminofluorescein (OAF) and hexadecanoyl aminoeosin (HAE)14 is impeded. In DTAC micelles (c, = 0.05 mol dm-3, without salt additives): pKt, = 5.78, pKf, = 3.71 (+0.05), pK;, < 0.5 (OAF) and pKE, = 2.87 k0.05, pK:, = 0.8 L-0.2 (HAE).l4 In DTAB micelles (4 mol dm-3 NaBr) the pK: values are reported as 7.47, 5.66 and 0.82 (k0.07) for OAF and 4.53, 2.32 (k0.05) for HAE.14 The pKr values in bromide systems at ionic strength in the bulk 4 mol dm-3 solution are 632 J. CHEM.SOC. FARADAY TRANS., 1994, VOL. 90 Table 1 The apparent ionization constants of oxyxanthenes in micellar (3 x rnol dm-3) solutions of cationic surfactants and the ther- modynamic pK, values in aqueous solutions ~~~ ~ CPC micelles, CTAC micelles, water 0.05 rnol dmP3 KC1" 4.00 rnol dm-3 KCl dye PKZO Pq PKZ2 PK:o PK:, PK:, PK:o PK:, PK:, fluorescein 2.14' 4.45' 6.80' 0.98 3.60 5.54 0.60 k0.10' 6.41 & 0.10 7.17 f0.06 sulfonefluorescein -3.22d 6.76d -0.9 5.46 -2.33 k0.05 7.00 f0.012,7-dichlorofluorescein 0.35' 4.00" 5.19' -3.58 3.70 < -03 5.50 f0.05 5.79 f0.06 eosin -2.@ 2.818 3.7Y -0.5' 2.82 -1.83 f0.07h 5.76 f0.06'' ethyl eosin -1.91' --0.V --1.11 & 0.03' -" From ref. 30-32; 20 "C.'From ref. 19; 25 "C. 'In the scale pH, = -log cHC,.From ref. 30-32; 20 "C. From ref. 21 ; 25 "C. f In the H, scale; -2.04 f0.03 (H,O-H,SO,), -1.98 k0.12 (H,O-HCI), -1.83 f0.09 (H,O-HCIO,); from ref. 42. From ref. 20; 20°C. In the pre- liminary comm~nication~~ the values 1.81 k0.11 and 5.89 k0.11 were reported as obtained from a smaller number of experimental data. Obtained by extrapolating the dependence of pK,, on 1/D in H,O-Me,CO mixtures; from ref. 23. j From ref. 31. known to be CQ. 0.5 units higher than those in chloride should be carried out: systems.l3.1 5.16.18.44 Spectral characteristics of R2-ions are measured directly at pH 9.5-11.5 and for R- of ethyl eosin at pH 6-7 (because of the C02Et group hydrolysis in aqueous alkali). The molar The H,R spectra are given in Fig. 5. Comparing them against absorptivities of HR- ions were obtained together with Kf, experimental curves (dotted lines) shows the necessity of the and Kf2 and then recalculated within the whole visible region use of eqn. (6).by using the pH area of maximal yield of the sought species Spectral characteristics of all the species differ from those in water (Tables 2 and 3), but are similar to those obtained in in solutions (pKf, < pH < pKf,): the presence of CPC micelles (0.05 mol dm-3 KCl).30332 The cases with the spectra of the H3R+ ion of fluorescein and H2R of sulfonefluorescein are more complicated. While for eosin there exists a narrow pH interval near pH ca. 4.0 where the directly measured E values are close to eHR-(Fig. 4), and for sulfonefluorescein this interval is wider, in the case of fluorescein and its dichloro derivative the HR--spectra, obtained by eqn.(5) (Fig. 1 and 2: dotted lines), do 4.5 not coincide with any spectra, measured directly. The stan- dard deviations of EHR-values are &(5-7)%. Because of weak intensities of H2R species of fluorescein and 2,7-dichlorofluorescein exact numerical &HzR values are 4.0 -not needed for pKf calculations. Traces of intensely coloured species, however, may strongly influence the spectra mea- sured in the pH region of H2R predominance. In order to 'remove' the ion contribution the following &HzR verification 3.5. , 2 .-I E c I-2.5.E m E U m z---. 40-450 500 550 ..-Alnm I ..,, I,--, 500 550 Fig.5 Absorption spectra of neutral forms of the dyes in CTAC A/n m micellar solutions (4 rnol dm-3 KCI). (1)-(3): H,R (3e4); (1) fluo- rescein, (2) 2,7-dichlorofluorescein, (3) eosin; (4) H,R of sul-Fig. 4 Absorption spectra of anions in CTAC micellar solutions (4 fonefluorescein (8); (1') and (2') spectra of fluorescein and 2,7-rnol dm-3 KCI): (1) R2- of sulfonefluorescein (lo), (2) RZ-of eosin dichlorofluorescein at pH 3.47; (3') spectrum of eosin (cHC,= 1.1 rnol (7c), (3) HR-of eosin (k),obtained using eqn. (5), (4) R-of ethyl dm-3); (5) HR of ethyl eosin in 2.5 rnol dm-3 HCI (11).(6) HR- of eosin (12) sulfonefluorescein(9). J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Table 2 Spectral characteristics of various species in micellar solutions and in water AmJnm(~JlO3 dm3 mol-' cm-') CTAC micelles, dye, species 4.00 mol dm -KCl H20a fluorescein H3R+ (la) 444-445 437 (54.3) H2R (2a e3a F? 4a) 465 (0.393), 490 (0.361) 437 (13.9), 470-485 (4-3.1) HR-(5a) 455 (24.9), 480 (24.0) 454-474 (32.7-33.8) R2-(7a) 500 (85.9) 491 (88.0) sulfonefluorescein H2R (8) 452 (56.6) 440 (52.2)HR-(9) 460 (29.9), 490 (25.5) 455, 480 (29) R2-(10) 512 (83.7) 495 (84.0) 2,7-dichlorofluorescein H2R (3b +4b) 470 (0.822), 500 (0.766) 460 (8.96), 485 (8.70) HR -(5be6b) 525 (57.4) 465-470 (24.1), 490 (28.3) R2-(7b) 515 (80.1) 502 (75.02) eosin H,R (3c F? 4c) 480 (7.92) 480-485 (8.5)* HR-(6c) 538-540 (97.6) 517-519 (81.9) R2-(7c) 526.5 (103.4) 5 15 (96.7) ethyl eosin CHR (11) 485.5 (21.6) -R-(12) 542.5 (97.1) 520 (97.3)' a As reported in ref. 19-21,42.'pH 0-0.5 (H2S0,). Low solubility. 5% EtOH or Me2C0. Location of Dye Species in the Microheterogeneous System The HR -spectrum of 2,7-dichlorofluorescein (Fig. 2, Table The bathochromic shifts (9-17 nm) of R2-absorption bands 2) differs decisively from that in water (where its configu- as compared to the spectra in water (Table 2) are usual for ration is similar to that of the HR-spectrum of these dyes in the case of replacing H20 by an organic fluorescein2'). Similar changes in the spectra of monoanionic The same is true for the HR- ion of eosin 2,7-dichlorofluorescein in water-acetone mixtures have been in CTAC micelles the shifts shown to be connected with the shift of equilibria from car- and R- ion of ethyl eosin;23,25*26 are 21-22 nm. For the HR- ions of fluorescein and sul- boxylate tautomer towards the phenolate one.21 fonefluorescein the solvatochromic shifts are less dramatic, The bathochromic displacement of the H2R band of sul- but they are still registered upon transferring the species from fonefluorescein (Table 2) is appreciable (12 nm); for this dye water to micelles of cationic surfactants.The closeness of A,,, the formation of colourless tautomers even in pure organic and in micellar solutions the values of R2- bands of fluorescein and eosin and HR- of solvents is not specific,21*26,27*42 eosin in micellar media, obtained in the present study (Table E,,, value is even higher than in water (Fig.5, Table 2). 2), to those of the lipoidal (solubilized) analogues under All the mentioned spectral effects, occurring in micellar similar conditions'" also confirms the completeness of solu- media (as compared to aqueous solutions), as well as pK: bilization. The deviation of the A,,(HR-) value of OAF values, are almost constant with a tenfold or more decrease (494-496 nm'") from our value for fluorescein (Table 2) may in the ratio cdye/c,. These observations are evidence of soh-be caused by the difference in assigning bands because the bilization of the species studied. If the concentrations of solu- HR- spectrum was not measured directly, see eqn. (5). bilized and non-solubilized species are commensurable, then The decrease in intensity of H2R spectra for fluorescein a shift of the distribution equilibria should be observed and 2,7-dichlorofluorescein (Fig.5), as compared to those in during the variations in the number of micellized surfactant aqueous media (Table 2), is attributed to shifts in the tauto- molecules (ions) per dye molecule. meric equilibria. The solubility of neutral forms of eosin In Table 3 the A,, values of fluorescence and excitation (H2R) and ethyl eosin (HR) increases essentially in the pres- spectra of the most hydrophilic species (R2-) are presented. ence of micelles as compared to water, where solubility is The data prove that the dianions R2- are completely solu- extremely low (ca. 2 x mol dm-3 and <loF6 mol bilized. The A,, of excitation are independent of the analyti- dm-3, re~pectively~'*"~). cal wavelengths used, i.e.no traces of unsolubilized (hydrated Table 3 The characteristic fluorescence and absorption (A,.Jnm) of R2- ions in water, aqueous KCl and in micellar (3 x mol dm-3) solutions of cationic surfactants emission excitation absorption CTAC, CTAC, CTAC, 4 rnol d m-3 4 mol dm-3 4 rnol d mP3 4 mol dmP3 4 rnol dm-3 dye H20 KCl CPC KCl H,O KCl CPC KCl H20 CPC KCl fluorescein 515 518 530 526 492 492 505 500 491 504.5 500 sulfonefluorescein 518 52 1 535 533 496 500 514 512 495 512 512 2,7-dichlorofluorescein 525 526 538 537 504 504 517 516 502 514.5 515 eosin 537 540 548 548 514 515 528 527 515 527 526.5 in the bulk phase) dye ions are revealed.The legitimacy of emission and excitation spectra used for such purposes (without trying to obtain information on peculiarities of solu- bilization of bound dyes in the pseudophase) is based on the much shorter fluorescence lifetime of the xanthene dyes (R2-)45-47 than the lifetime of micelle~.~' At pH, 2 0 I,,,(H3R+) of fluorescein ($44-445 nm) differs markedly from that in water [437 nm in HCl solutions (pH 0 ~~-~~to l), as well as at c~ = 4.0 mol dm-3], which allows one to consider the cationic species to be solubilized. For the OAF cation in DTAB micelles (4 mol dm-3 NaBr) I,, = 448nm.I4 At pH, < 0 the superseding of H3R+ of fluorescein from CTAC micellar surface or some other effects of HCl are possible [A,,,(H3R+) = 439-440 nm].In the system CPC-0.05 mol dm-3 KC1-HC1, 4,,,,(H3R+) is at 437 nm, as in water. Apparently the cation H3R+ is unsolubilized at high Y values; under this condition the pKf, lowering (=0.98; pKro = 2.14) occurs only because of H2R solubilization for the completely solubilized OAF in DTAC micelles at low [Cl-1,: pKi,(0.514. The other species of fluo-rescein and its derivatives, judging by the I,,, values and molar absorptivities as well as solubilities of neutral forms, are solubilized in CPC rni~elles.~'-~~.~~ The bathochromic shift of the H2R band of sulfonefluorescein in CPC micelles (5 nm) is less than in the micellar pseudophase with reduced Y value (12 nm), but this may be connected with the pecu- J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 liarities in the position of cationic chromophores within the micelles of both types. Discussion Detailed Scheme of Protolytic Equilibria in the Micellar Pseudo phase ~ Scheme 1, describing the interconversions of protolytic species (1 -, . . -,7), was discussed in previous pub-lication~.'~-~~'~~-~~The main assumptions, used in the cal- culation of the fractions (a) of various tautomers of H2R and HR-, are: (i) The lactone 4 is colourless and (ii) the ioniza- tion of the carboxylic group exercises a relatively slight influ- ence on the absorption band of the xanthene chromophore in the visible region. Hence the spectra of species 2, 3 and 6 for the dye with the given substituents in the positions 2, 4, 5, 7 or/and in the phthalic residue, are close to those of 1,5 and 7 respectively; in the last case the chromophore is most sensi- tive to 2'-substituents, and the band of 6 is red-shifted (ca.3-15 nm) compared to that for 7.20-26 The spectra (Fig. 1 and 4) show that the HR- ion of fluo-rescein exists as Sa, while the HR- ion of eosin exists as 6c. This is proved by comparing the spectra of these two mono- anionic dyes with the spectra of the HR- form of sul-fonefluorescein (9, Scheme 2) and of R- of ethyl eosin (12, IbC02-I -b""; 7 Scheme 1 Protolytic equilibria of oxyxanthenes. la-7a: unsubstituted fluorescein, lb7b: 2,7-dichlorofluorescein, lc-7c: 2,4,5,7-tetra-bromofluorescein (eosin) J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 635 8 9 10 Scheme 2 Ionization of sulfonefluorescein Br Br t3r Br C02Et C02Etty 0 11 12 Scheme 3 Ionization of ethyl eosin Scheme 3) respectively (Fig. 4 and 5). The replacement of the CO, group by SO, results in an additional red shift of the R2-ion band2' (Fig. 1 and 4, Tables 2 and 3). The HR-absorption intensity of eosin is somewhat lower than that of R2- [&,,(HR-) = 0.94 cmaX(R2-)], but this ratio remains almost constant (ca. 0.9) in various organic sol-yentS,20-26.29-32.42 which permits it to be regarded as a manifestation of 6c chromophore properties rather than a sign of the presence of 5c (see below). In the case of the HR-ion of 2,7-dichlorofluorescein in water the 5b tautomer pre- dominates,2' which may explain its incomplete transform- ation into 6b in micellar media.For calculating asb and a6b 2 = 525 nm was used: Taking &,,(6b) equal to &,,(7b) = &,,(R2-) [actually reali- zing the correction for the shift in bands, caused by 2'-substituent replacement (CO, 4 C02H)] and neglecting the 5b absorption in this region one can obtain ash = 0.72. Pro- ceding from the ~,,,,,(6)/~~~~(7) of 2,4,5,7-tetra-ratio halogenoderivatives in cationic micelles (average value : 0.9030-32.42 ) and assuming that the &Sb/&6bratio at Amax(6b)is equal to ratio of ethyl eosin at 542.5 nm (=0.082, Fig. 4 and 5), we have calculated the values a5b= 0.222 and a6b = 0.778, which we consider most reliable. Then KTx= [6b]/[5b] = 3.5.Similarly in CPC micelles (0.05 mol dm-3 KCl), in these for &,,(R2 -) of 2,7-dichloro- fluorescein in H20 the mean value of 101 x lo3 (503 nm)48 and 97 x lo3 dm3 mol-' cm-' (502.5 nm)49 was postulated and the apparent &,,(R2-) value in micellar solutions, reported previo~sly,~~,~~ is also higher than in Table 2. For the spectra of neutral forms (Fig. 5) the following equation is derived : 'H2R = '2 '2 + '3 '3 (9) The H2R spectra of fluorescein does not contain obvious fea- tures of 2a (which must absorb as la and 8, Fig. 1 and 5 and Table 2); the coexistence of 3a [absorbing as 5a (Fig. 1) and 9 (Fig. 5)] with the colourless lactone 4a is evident. In the case of eosin, the zwitterionic tautomer 2c is highly improbable because of the strong acidity of the oxy groups.The coloured fraction of H2R (3c) has a spectrum analogous to that of HR (11) of ethyl eosin (Fig. 5). The a3 values were calculated as : '3 = &H2dE3 (10) from the data near AmaX(H2R).The value e3, = 27.4 x lo3 dm3 mol-' cm-' [average of E,,, of 5a and 91 was used to offind afa = 0.014, a4' = 0.986. ~,,,~,(ll) ethyl eosin, 21.6 x lo3 dm3 mol-' cm-', was used in the case of eosin: a3, = 0.37, a4, = 0.63. For the dihalogeno-derivative the uti- lization of the average value 24.5 x lo3 dm3 mol-' cm-' leads to a3b = 0.034, a4b = 0.966. Thus the KT (tautomerisation constant) values, KT = [4]/[3], are equal to 70.4, 28.4 and 1.70 for fluorescein, 2,7-dichlorofluorescein and eosin, respectively. Apparent Microscopic Ionization Constants, k' It is easy to show that: pKfO = pkt,OH + log a3 = pk", C02H + log a2 (l pKfl = pk?,COzH -log '3 + log 'S = pk",oH -log a3 + log 66 = pkt, -log a2 + log a5 (12) pKf2 = pk?,OH -log '5 = pk", CO2H -log a6 (13) where k" are microscopic ionization constants (actually 'apparent' microscopic ionization constants): k", C02H = hc21/[11, k", OH = hC31/C11, kt,Z = h[51/c21, k",COzH = hc51/~31, k", OH = h[61/[31, k?, OH = hcwc5i, k;.CO2H = h[7]/[6], see Scheme 1. The pk" values are presented in Table 4. The fraction of the zwitterion aZn was undetectable by means of the absorption spectra (see above). However, it can be shown that: 10g(a3a/a2a)= pk", C02H -pk\, OH = pk?,C02H -pkt,Z (14) Using the pk;,, value of sulfonefluorescein as the corre-sponding value for fluorescein and assuming that pk", C02H is not more than 1 unit lower than pk",,,,, (see below), it is easy to evaluate: a3JaZp x lo2-lo3, which agrees with the spectral data.The pk;, C02H value of eosin (= pKi2) is higher than pk", COzH of fluorescein. The reason may be the influence of the negative charge delocalized in the xanthene moiety of 6c. Such an influence (and, in turn, the influence of the charge J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Table 4 Apparent microscopic ionization constants (P)of the oxyxanthene series in CTAC micellar pseudophase (4.00 mol dmF3 KC1 in the bulk) and the medium effects (ApP, AApP) fluorescein, k,,, OH sulfonefluorescein, k eosin, k 1, OH ethyl eosin, k 1, OH 2,7-dichlorofluorescein,k OH fluorescein, k2, OH sulfonefluorescein, k2,OH 2,7-dichlorofluorescein, k2,OH 2,7-dichlorofluorescein,k 1, mzH fluorescein, k 1.C02H 2,7-dichlorofluorescein,k2, COzH eosin, k2, COzH 2.45 3.10 -0.65 0.1" 2.33 3.22 -0.9 1.4 1.40 2.40 -1.0 0.9' 1.11 4.14 (1.9) - ( -0.8) - 1.1' 1.3 7.17 6.80 0.37 1.6 7.00 6.76 0.24 1.5 (5.14) 5.19 (0) (1.8) (4.68) 4.56 3.5 3.49 (1.2)1.1 (1.7)2.4 5.68 - - 2.2 5.76 3.75 2.0 2.9 ~~~~ The cationic species are probably desolubilized in CPC micelles (0.05 mol dm-3 KCl). The pk.,, OH of eosin and ethyl eosin in CPC micelles are obtained in the pH, scale. in the phthalic residue on the acidity of OH groups) may be qualitatively expressed through the Bjerrum-Kirkwood-Westheimer equation :50 6 = 241/Deffp where 6 is the additional rise in pk of the ionizing group, p/A is the distance between the latter and the negatively charged group, Deff is the 'effective' relative permittivity.If the xanth- ene moiety is positively charged (la species) the pk of the carboxylic group (pk,, C02H) can be expected to be lower than pk~,,,,,. In the media under study the pk:,, of sulfone- fluorescein is even less than pkt, OH of fluorescein, against eqn. (15) and contrary to H,O-DMSO, where pk,,, is 0.12-0.90 units higher than pk,,oH.27 The reasons may be the follow- ing: (i) pKf, ,used for pk", OH determination, is calculated on the pH, scale (perhaps at high KCl concentrations f; > 1); (ii) the peculiarities of the arrangement of solubilized cations la may be significant; in CTAC micelles these species may be shifted toward the water phase.The pk;,,, of eosin satisfactorily coincides with pKr, (=pk", OH) of ethyl eosin (Scheme 3), the latter value being 0.2-0.5 lower than the former in various organic solvents [e.g., see Fig. qb)]. While the pkt, C02H and pki, COzH values of 2,7-dichloro- fluorescein are close to the corresponding values of fluores- cein and eosin, respectively, the pk:,,, and pki,,, of the dihalogeno derivative are essentially higher than pk:, OH of eosin and ethyl eosin, but essentially lower than pk;,,, of fluoresecein and sulfonefluorescein (Table 4).This is undoubtedly the consequence of substitution; the pki, OH values of phenolsulfonephthalein and its tetrabromo-derivative in CTAC micellar pseudophase (4.00 mol dm- KC1 in the bulk) are 8.71 and 3.86, respectively.'* From Scheme 1 the following equation can be derived: Using the pk;,,,, pki,C02H values for eosin and pk;,,,,,, pk;,,, values for fluorescein and taking into account eqn. (15) KTxcan be calculated: KTx > lo3 (eosin) and KTx x 0.01 (fluorescein). The results agree with the absorption spectra of the HR -forms. Taking into account the degree of precision in confirming the assumptions used in spectra modelling of single tauto- mers as well as the errors of E values, used in a calculations, the a errors can reach ca.20%. Hence the pk" standard devi- ations are on the whole greater than those of pKr. The least exact is the aSb value (to emphasize this fact the values pk;, C02H and pki, OH of 2,7-dichlorofluorescein are given in Table 4 in parenthesis). If + a6b, then pk?,,, = 4.03, pk", C02H = 5.79. The Stern layer is in reality a region of high electrolyte concentration (ca. 3-7 mol dm-3).6,10 It should be noted that the salt effects on the pK,,, pK,, and pK,, of fluorescein in the H,O-KCl system (to 2.3 mol dm-3 KC1) are of standard nature (the values vary within the ranges 2.14-2.40,4.45-4.18 and 6.80-6.26 respectively), and aza, as, and a,. values are not influenced by cKcI.42That is why the effects observed in CTAC micelles cannot by caused by the electrolytic environ- ment itself. Medium Effects for Macroscopic and Microscopic Apparent Ionization Constants (ApK ,AN') The ApKr [k(0.07-0.191 in CTAC micellar solutions vary within the range of 3.4 units: from -1.4 (fluorescein, ApKt,), -0.98 (eosin, ApKf,) and -0.89 (sulfonefluorescein, ApKf,) to 2.01 (eosin, ApK;,) and 1.96 (fluorescein, ApKf,).The ApKr values can be devided into Apk" (=pk" -pk") and A log a [eqn. (11-13)]. The Apk" values are given in Table 4; though they are relatively small, the differences in medium effects are subjected to the same regularities as in organic solvents2'-' and aqueous solutions of non-ionic polymer^,^' e.g. for fluorescein and sulfonefluorescein : OH (and 'Pk",, Z) < 'Pk;, OH < 'Pk?, COzH (17) Here, as usua1,38*50*52 the medium effects for carboxylic acids are higher than for phenols.The charge types of cationic and zwitterionic acids are known to cause even the decrease in their pK, values (here pkt,,, and pk",,,) on addition of organic co-solven ts.' 9 38*5' The value for eosin is somewhat lower again (Table 4), but this is perhaps the result of the differentiating impact of the micellar environment. Apki, C02H (eosin) is higher than Apk", C02H (fluorescein), as is usual for anionic and neutral acids with the same ionizing gro~p.~' Medium Effects in Aqueous Acetone Here the ApKf and Apk" values are compared with the medium effects in the H,O-Me,CO system; it does not mean that we regard the solvent properties of these mixtures as being the closest to those of micellar pseudophase.The pK, values compiled from previous studies' 1-23 are plotted against acetone concentrations (wt.%) (Fig. 6). The thermody- J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Jf~'""'''20 40 60 80 J\ -' 20 ' 40 ' 60 ' 80 ' 20 40 60 80 wt.% acetone Fig. 6 Dependences of pK, on wt.% of acetone. (a)fluorescein: (1) pK,, ,(2) pK,, ,(3) pK,, ( =pk2, OH)r (4) pk,, OH, (5) pk,, COIH. (6)eosin: (1) pK,, ,(2) pK,, (= pk,, C02H), (3) pk,. OH;(4)pK,, (= pk,, OH) of ethyl eosin. (c) 2,7-dichlorofluorescein:(1) pK,, ,(2) pK,, . namic pK, values (usually kO.05) relate to pa; scales. For and pK,, values of fluorescein and 2,7-dichlorofluorescein fluorescein the same trends as in H,O-DMSO mixtures27 are became closer, while pK,, and pK,, values of eosin and pK,, evident.HR- exist as 5a, the 2a tautomer disappears: and pK,, values of fluoresecein grew further apart. In all the cases a3a< a3b < age.APKao = AP~,,OH + A log ~3n (18) ApKal = APkl,C02H-A log u3a (19) The Properties of the CTAC Micellar Pseudophase at High ApKa2 = Apk2, OH (20) Ionic Strength of the Bulk Phase as compared with Water-rganic Compound Mixtures The sharp decrease in aSa(from 0.111 in H20 to 5 x in 90% Me,CO), U-shaped pk,, OH dependence on Me2C0 The pKr values [see eqn. (3)] correspond to the pK, in content, typical for cationic acid^,^^.^^ and more marked H20-Me2C0 at certain unknown acetone content. Even if medium effects for carboxylic groups than for oxy pKt at 4 mol dmP3 KCl in the bulk phase can be considered groups 2 77387 50.5 2 as an approximation of pKa (neglecting the 'residual' quan- tity YF/2.303RT), the log yH value [see eqn.(3)] is unknown ApkO,OH < Apk2,OH < APkl,CO2H (21) (essentially ex trathermod ynamic), and no attempts were 7, made here to evaluate it. However, if the and Y values are lead to different changes of Kao/K,, and K,,/K,, ratios: the same for all the solubilized dyes, it could be expected that the whole set of parameters (at least pk" values) would corre- spond to one fixed Me,CO content. COzH -AP~,A(PKa1 -PKaJ = AP~,, OH -2A log ~3a (22) This is, however, not the case. From the curves in Fig. 6 (plotted on a large scale) it is found that the 'apparent' pK:,, A(pKa2 -PKa1) = A~k2,OH -AP~,,C02H + A log a3a pKf2 (= pk", OH) a;nd pkt, C02H of fluorescein in CTAC micel- (23) les (4 mol dm- KCl) correspond to 27, 19.4 and 21% Me2C0, respectively, pKf , and pK:, of 2,7-dichloro-As a result K,,/K,, increases from 204 in H,O to 3.8 x 10" fluorescein and pK:2 ( =pk", C02H) of eosin correspond to 16, 22 and 28.5% Me,CO, respectively.In the cases of pK:, andin 90% Me,CO, while K,,/Ka2 decreases from 224 in H,O to 0.05 (inversion of ionization constants) in 90% Me2C0. In contrast, for eosin HR- exists as 6c and aJCdoes not decrease as sharply (a3c= 0.057 in 90% Me2CO): pk:,,, of fluorescein, pK:, and pk;.,, of eosin and pK:, (=pk;,,,) of ethyl eosin all the values are lower than corre- sponding pK, values in H20-Me,CO mixtures.In H20-EtOH42 the K,, , K,, and kl,C02H of fluorescein, equal to the apparent (micellar) values, are measured at, respectively, 44-45, 25 and 25 wt.% of EtOH, pK,, = 5.76 of A(pKa2 -PKa,) = Apk2, cO~H-AP~,,OH + A log a3c (24) As pk2,co2H rises more sharply than pkl,OH, K,,/Ka2 increases from 8.7 in H,O to 4 x lo3in 90% Me,CO. In the case of 2,7-dichlorofluorescein the interpretation is more complicated because, the shift not only from 3b to 4b (a3b = 0.010 in 90% Me,CO), but also from 5b to 6b occurs. As a result K,,/K,, changes from 15.5 in H,O to 1.3 in 90% Me,CO. All the trends observed for the dyes under discussion in aqueous acetone as well as in other solvents (Table 5) qualit- atively followed those stated in the present paper for the CTAC micellar pseudophase (4 mol dmP3 KCl); the pK,, eosin corresponds to 45% EtOH.However, the values, equal to pKto and pkt, OH of fluorescein, pKr, and pk", OH of eosin and pK:, (= pk;, OH) of ethyl eosin, are below the correspond- ing pK, values within the whole range of EtOH content.42 While the k~,,,,, and k:,,, values of fluorescein corre- spond to the k values in ca. 22-42% DMS0,27 the values, equal to K:,, K:, and Kr2 (=k;,,,), were recorded in CQ. 50% DMS0.27 In the H20-1,4-dioxane system4, the values K,, and K,, of fluorescein, equal to K:, and K:,, are obtained at 27-30% of the organic co-solvent; k1,C02H, equal J. CHEM. SOC.FARADAY TRANS., 1994, VOL. 90 Table 5 The values of log(Ka,/Ka2) and of the quinoid tautomer fractions (a3)for fluorescein, 2,7-dichlorofluorescein and eosin in various media lodKailKa2) = ~Ka2-PKai a3 2,7-dichloro- 2,7-dichloro- media fluorescein fluorescein eosin fluorescein fluorescein eosin H,O" 21 wt.% ficolld 2.35 1.04 1.19 - 0.94 1.64 O.lllb 0.0823' 0.32 - 0.39' 0.21 16 wt.% Me2COf 1.40 0.5 2.06 0.072 0.10 0.40 CPC micellese 1.94 0.12 2.32 0.038 0.108 -h CTAC micelles' 0.76 0.29 3.93 0.014 0.034 0.37 64wt.% l,4dioxaneJ 0.1 0.19 3.33 5.7 x 10-3 0.022 0.09 91 wt.% DMSO~ -1.35 0.07 3.91 1.7 x 10-3 0.014 0.1 From ref. 19-21 and 42. * aZn= 0.218. 'The cJe value is taken to be 21.6 x lo3 dm3 mol-' cm-'.From ref. 51; ionic strength: 0.05 rnol dm-3 NaCI. aZ. = 0.0287. f From ref. 21-23, 42. From ref. 30, 32; ionic strength: 0.05 rnol dm-3 KCI. The value cmaX(H,R) = 21 x lo3 dm3 mol-' cm-' is calculated from equilibrium data in the acidic pH, region. '4.00 mol dm-3 KCl. j From ref. 25,42. From ref. 26,27, 42. to kt,,2H, corresponds to 20% 1,4-dioxane, while all pK,, and pk,,oH within the range 0-64% 1,4-dioxane are higher than the values of corresponding apparent (micellar) con- stants. The KT values in micellar medium can be directly com- pared with those in water-rganic compound mixtures without having Y and yH values at our disposal. While KT values, equal to 70 and 28, are determined for fluorescein and 2,7-dichlorofluorescein, respectively, in 36.5% Me,CO, the value K, = 1.7 of eosin corresponds to ca.18% Me2C0. The KT value for the latter dye is less sensitive to variations in en~ironment~~.~~.~~(Table 5). The 'micellar' value for fluo- rescein is given for 40% 1,4-dioxane and 52% DMSO. The D values of these three mixtures lie within the wide range 44.4- 75.5. Much closer are the E: (Dimroth-Reichardt parameter) values (0.78, 0.75 and 0.75 for 36.5% Me,CO, 40% 1,4-dioxane and 52% DMS053), which emphasises the role of the H-bonding ability of the solvents in the determination of the KT value. The E: value for CTAC micelles is not available in the literature; for DTAC micelles Ef =0.74 and 0.70 in aqueous solutions and at cNaCl= 4 mol dm-3.15 The signifi- cance of H-bonds in 3a stabilization stresses the fact that even in pure EtOH (D = 25, EF =0.66) the K, value (= 58)42 is lower than in CTAC micelles or in mixtures of water with 'aprotic' solvents with ET =0.75-0.78. The differences of Apk" values for stepwise ionization of fluorescein [eqn.(25)] and of eosin [eqn. (26)] may be expressed as 'Pk;, OH -'Pk;, CO2H = log Y7n +log Y3n -2 log ysn = -0.7 (25) The decrease in values on introducing four bromine atoms occurs for all the eosin species and to some extent this effect is mutually compensated in eqn. (26). Therefore, the essential difference in the above log y combinations for fluorescein and eosin reflects more strongly the stabilization of the anion 6c with delocalized charge as compared to the carboxylate anion (5a) in micellar pseudophase. This effect, being of the same nature as those described by expressions (17) and (21), is typical for the transfer from water to media with lower H- bonding ability.' 7s2s4 The modern structural models of micelles of the type discussed' ,2*46*5 presume the entry of hydrocarbon chains into the Stern layer (or rather 'Stern region'').Such proxim- ity of -CH2- groups to the dyes examined hinders H- bonding with water molecules, causing the effects described above. Such a mixture of water, electrolyte (3-7 mol dm-3) and hydrocarbon is unattainable in homogeneous systems, and this is the reason for difficulties in modelling the Stern region effect on the protolytic equilibria through comparing with medium effects in aqueous acetone, 1,4-dioxane, ethanol, DMSO etc.The location of different dyespecies within the Stern region, in particular the depth of penetration, can also contribute to the resulting medium effects (ApKf:, Apk", A log a), but the simple model used in the present study presup- poses an identical microenvironment [eqn. (1)-(311 with a fixed set of parameters of the pseudophase for all the particles solubilized. Transfer from CPC Micelles, [C1-] =0.05 mol dm-3, to ~CTAC Micelles, [C1-] = 4.00 rnol dm-Following Moller and Kragh-Han~en~~the differences between pK," (or pk") in CTAC micelles (4.00 mol dm -KCl) and in CPC micelles (0.05 rnol dm-3 KCl) are denoted as AApKf: (or AApk"). The AApKf: C-t(O.08-O.l8)] vary from 1.11 (AApKr, of ethyl eosin) to 2.94 (AApKf:, of eosin).As follows from eqn. (11)-(13) both pk" and log a cause the resulting effect. [Note that pKf:, values of eosin and ethyl eosin in CPC micelles (Table 1) are expressed in the pH, scale.] A special case for fluorescein (AApKf:, = -0.38) can be elucidated by the fact that in CPC micelles (0.05 mol dm-3 KCl) only the H2R (not H3R+) species are solubilized. As a whole the AApk" values vary within a narrower range than Apk" (Table 4); lowering the potential of the Stern layer leads to ca. 10-1000-fold decrease in the k" values. The effect differs for phenolic and carboxylic groups: (i) The average value for eight AApk$H is 1.4 & 0.3 (the value AAPG,~~is excluded, see above).For eight sulfonephthaleins' the AApk;,.. values are in the range 1.63-1.94; (ii) The average AApk:o2H value is 2.3 f0.5 (four pk" values). Thus if the decrease in Stern layer potential is evaluated as A'P = 59AApk" (27) rather close values (96 to 114 mV, average: 107 f7 mV) may be obtained for eight sulfonephthaleins, but if using AApk;, C02H (fluorescein) and AApk;, COzH (2,7-dichloro-fluorescein, eosin) the values AY = 142 and 130-171 mV differ more significantly (the calculations are conventional as the pk" values in CPC micelles are obtained at 20°C). The observed effects can be explained through the peculiarities of the solubilization of carboxylates in cetylpyridinium micelles (the similarity of pKr in CTAC and CPC micelles at equal [Cl-1, values is still proved only for ~ulfonephthaleins~~), J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 or/and it must be assumed that the rise in [Cl-J, results not only in Y lowering. The salt additives are known sometimes to increase the thickness of the Stern layer.' Besides, with increase in the counter-ion bulk concentration, the micelles growth and their changes in shape take place to some extent in chloride systems too.2 Note that the growth of micelles occurring both during c, increase (DTAB) and addition of salts to DTAC and DTAB (4 mol dm-3 NaCl and NaBr) leads to the decrease in Eq values;" for CPB and CTAB micelles (c, = 0.05 mol dmP3) without salt additives the E: values practi- cally coincide." Thus it can be supposed that the properties of the Stern region at the increase in ionic strength of the bulk phase became still more 'non-aqueous'.This may be the reason for the additional increase of pk;,,, as compared to p@, (see the AApk" in Table 4). The K, values of all the three dyes increase by ca. three times their original values in transferring from one micellar system to the other (Table 5). Moreover, the decrease of pK,, -pK,, values in the case of fluorescein and their increase in the case of eosin (in both cases with reference to aqueous medium) are more marked in micellar solutions with high [Cl-1, values [Table 5; see also eqn. (25) and (26)]. Following this logic the pKi (and pk') values at high ionic strength of the bulk phase may, generally speaking, differ from those at low electrolyte concentration.Further research is necessary in order to clarify this problem. Organic anions inside the micelles are supposed to be associated with cationic head gro~ps.'~'~' In this case the competition between CO; groups (with localized charge) and C1- may occur in their interactions with cationic surfactants. If the increase in [Cl -1, results in 'dye-surfactant associates' dissociation as well (as a consequence of further saturation of the Stern layer with the ele~trolyte'.~), then the contribution of the formation of such associates to the medium effects (ApKf, Apk") at low ionic strength of the bulk phase should be recognized. Such additional stabilization of carboxylates can be an additional source of hindrance in attempts to model pKi and pKY using pK, in water-organic compound mixtures.The errors (sometimes k0.2) and number of AApk" obtained (Table 4) do not allow more definite conclusions to be made. On the other hand if the CO, and SO, groups of HR-species (5a, 9) can be regarded as 'neutralized' with surfactant cations, then the pK;, values of fluorescein and sul-fonefluorescein must be equal to the pKr, value of 6-hydroxy-9-phenylfluoron (Scheme 4). But in CPC micellar solution (0.05 mol dm-3 KCl) this pK;, is found to be 4.67 & 0.07,30-32 i.e. 0.8-0.9 pK, units lower than pK;, of fluorescein and sulfonefluorescein. In water this difference, 6 [see eqn. (15)], is 0.5.19It can be considered as evidence of oversimplicity of the model of complete neutralization of anionic groups in the Stern layer of cationic surfactants.Conclusions According to the electrostatic approach, reflected in eqn. (l), increase of the counter-ion concentration in the bulk phase to 13 14 Scheme 4 Ionization of 6-hydroxy-9-phenylfluoron 4 mol dm-3 leads principally to micellar charge shielding and to the reduction of the contribution of the quantity YF/ 2.3RT. It allows the consideration of the set of pK: and k" values as a rough approximation to that of pKi (pk'). In several cases the values are essentially higher than pKz (pk"). The influence of salt additives on the specific interactions cannot be entirely excluded either.The Stern region is known to be highly aqueous, contain- ing high concentrations of ionic head groups and counter ions.1~2*6~10*3'*5'However, the equilibrium parameters, deter- mined for the solubilized oxyxanthene series, differ signifi- cantly from those in water or in water-salt systems. Three main features, typical for organic microenvironments, can be pointed out: (i) red shift of absorption and emission bands for species of type 7;(ii) differentiating effect of the pseudophase; different medium effects, Apk, reflect the charge type of the acid-base couple and the nature of ionizing group (ApPoH < Apkco2"); (iii)shifts of tautomeric equilibria. All these factors, controlling the intrinsic pKi values, cannot be caused by anything except the proximity of the hydrocar- bon core.It is difficult to compare directly the pk" (~pk') values with those in water-organic compound mixtures because of the extrathermodynamic nature of yH values. However, the data give rise to some doubts about the correspondence of the pseudophase solvent properties to those of a certain water- organic compound mixture. The pk;, and pk;,,, values of solubilized dyes show some- what differing increases at increasing ionic strengths of the bulk phase. A preliminary hypothesis for explanation of this and some other effects is that the micellar pseudophase (at least the Stern region) becomes more 'non-aqueous' at high salt concentrations in the water phase. Note Added in Proof In the paper of N.0. Mchedlov-Petrossyan and R. Salinas Moyorga, J. Chem. Soc., Faraday Trans., 1990, 90, 3025, the following corrections should be made: Within Scheme 1, "97 \ Aorokshould appear as References 1 E. J. R. Sudholter, G. B. Van de Langkruis and J. B. F. N. Eng-berts, Red. Trav. Chim. Pays-Bas, 1980, 99, 73. 2 C. A. Bunton and G. Savelli, Adv. Phys. Org. Chem., 1986, 22, 213. 3 P. Mukerjee and K. Banejee, J. Phys. Chern., 1964,68,3567. 4 N. Funasaki, Nippon Kagaku Kaishi, 1976, 722. 5 F. Dorion, G. Charbit and R. Gaboriaud, J. CoZloid Interface Sci., 1984, 101, 27. 6 L. S. Romsted, J. Phys. Chem., 1985,89, 5107. 7 G.S. Hartley and J. W. Roe, Trans. Faraday SOC.,1940 36,101. 8 M. S. Fernandez and P. Fromherz, J. Phys. Chem., 1977, 81, 1755.9 M. E. Diaz Garcia and A. Sanz-Medel, Talanta, 1986,33, 255. 10 N. Funasaki, J. Phys. Chem., 1979,83, 1998. 11 G. Charbit, F. Dorion and R. Gaboriaud, J. Chim. Phys., 1984, 81, 187. 12 C. J. Drummond, F. Grieser and T. W. Healy, J. Chem. SOC., Faraday Trans. 1, 1989,85,521. 13 C. J. Drummond, F. Grieser and T. W. Healy, J. Chem. SOC., Faraday Trans. I, 1989,85,537. 640 J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 14 J. Kibblewhite, C. J. Drummond, F. Grieser and P. J. Thistle- 34 Z. Bi, Huaxue Shiji, 1986,8, 324. twaite, J. Phys. Chem., 1989,93, 7464. 35 S. B. Savvin, R. K. Chernova and S. N. Stykov, Surfactants, Izd. 15 C. J. Drummond, F. Grieser and T. W. Healy, Faraday Discuss. Chem. SOC.,1986,81,95. 36 Nauka, Moscow, 1991.E. Bardez, B. Larrey, X. X. Zhu and B. Valeur, Chem. Phys. 16 C. J. Drummond, F. Grieser and T. W. Healy, Chem. Phys. Lett., 1987,140,493. 37 Lett., 1990, 171, 362. J. N. Ness and D. K. Moth, J. Colloid Interface Sci., 1988, 123, 17 M. J. Minch, M. Giaccio and R. Wolf, J. Am. Chem. SOC., 1975, 546. 97, 3766. 38 R. G. Bates, Determination ofpH (Russ. Transl.), Izd. Khimiya, 18 N. 0.Mchedlov-Petrossyan, V. N. Kleshchevnikova, M. I. Rubtsov, L. L. Lukatskaya, R. Salinas Mayorga and V. I. 39 Leningrad, 1972. D. Bartet, C. Gamboa and L. Sepulveda, J. Phys. Chem., 1980, Kuchtik, Dokl. Akad. Nuuk SSSR, 1989,308,122. 84,272. 19 N. 0.Mchedlov-Petrossyan, Zh. Anal. Khim., 1979,34, 1055. 40 J. A. Loughlin and L. S. Romsted, Colloids Surf., 1990,48, 123.20 N. 0. Mchedlov-Petrossyan, L. P. Adamovich and L. E. Niki- 41 J. KuliE and T. Vitakk, Sb. VZd. Praci, Vys. fkola Chem. shina, Zh. Anal. Khim., 1980,35, 1495. Technol., Pardubice, 1986,49, 369. 21 N. 0. Mchedlov-Petrossyan, M. I. Rubtsov and L. L. Lukat- 42 N. 0. Mchedlov-Petrossyan, Doctoral Dissertation, Kharkov 22 skaya, Dyes. Pigm, 1992,18, 179. N. 0. Mchedlov-Petrossyan, V. F. Mindrina and L. E. Niki- 43 University, 1992. B. J. Thamer and A. F. Voigt, J. Phys. Chem., 1952,56,225. shina, Dokl. Akad. Nauk. SSSR, 1985,283,670. 44 J. V. Moller and U. Kragh-Hansen, Biochemistry, 1975, 14, 2317. 23 24 N. 0.Mchedlov-Petrossyan, Ukr. Khim. Zh., 1987,53, 1304. N. 0. Mchedlov-Petrossyan, V. F. Mindrina and L.E. Niki- 45 46 M. A. J. Rodgers, Chem. Phys. Lett., 1981,78,509. W. Read, M. J. Politi and J. H. Fendler, J. Am. Chem. SOC., 1981, shina, Zzv. Vuzou SSSR, Khim. Khim. Technol., 1985,28,29. 103,4591. 25 N. 0. Mchedlov-Petrossyan, T. A. Chernaya and A. Yu. Per- 47 P. Bilski, R. Dabestani and C. F. Chignell, J. Phys. Chem., 1991, everzev, Zh. Anal. Khim., 1992,47,598. 95,5784. 26 N. 0. Mchedlov-Petrossyan, R. Salinas Mayorga and Yu. N. Surov, Zh. Obshch. Khim., 1991,61,225. 48 H. Leonhardt, L. Gordon and R. Livingston, J. Phys. Chem., 1971,75, 245. 27 N. 0.Mchedlov-Petrossyan and R. Salinas Mayorga, J. Chem. SOC.,Faraday Trans., 1992,88,3025. 49 50 E. Gandin, J. Piette and Y. Lion, J. Chromatogr., 1982,249, 393. R. P. Bell, The Proton in Chemistry (Russ. Transl.), Izd. Mir, 28 N. 0.Mchedlov-Petrossyan, M. 1. Rubtsov, L. L. Lukatskaya, Moscow, 1977. T. A. Chernaya and A. Yu. Pereverzev, Dokl. Akad. Nauk. SSSR, 51 B. Yu. Zaslavsky, L. M. Miheeva N. D. Gulaeva, A. A. Boro- 29 1988,299,921. N. 0. Mchedlov-Petrossyan, L. L. Lukatskaya and M. I. vskaya, M. I. Rubtsov, L. L. Lukatskaya and N. 0. Mchedlov-Petrossyan, J. Chem. SOC.,Faraday Trans., 1991,87,931. 30 Rubtsov, Ukr. Khim. Zh., 1990,56,69. N. 0.Mchedlov-Petrossyan and V. N. Kleshchevnikova, Depos- ited Doc., 1989, SPSTL 424 Khp-D89, Ref. Zh. Khim., 1989, 52 53 N. A. Izmailov, Electrochemistry of Solutions, Izd. Khark. Univ., Kharkov, 1959. T. M. Krygowski, P. K. Wrona, U. Zielkowska and C. Rei- 31 32 17B3 198. N. 0. Mchedlov-Petrossyan and V. N. Kleshchevnikova, Dokl. Akad. Nauk SSSR, 1990,312,397. N. 0. Mchedlov-Petrossyan and V. N. Kleshchevnikova, Zh. 54 55 chardt, Tetrahedron, 1985,41,4519. L. M. Epstein, Usp. Khim., 1979,48, 1600. P. Fromherz, Chem. Phys. Lett., 1981,77,460. 33 Obshch. Khim., 1990,60,900. T. Nash, J. Phys. Chem., 1958,62, 1574. Paper 3/02592F; Received 6th May, 1993