UV Causes Dramatic Changes in Aggregation with Mixtures of

Jun 16, 2004 - In Final Form: May 4, 2004. Aqueous mixtures of photosensitive and inert surfactants have been prepared; photoreactions and changes...
1 downloads 0 Views 210KB Size
Download PDF
6120

Langmuir 2004, 20, 6120-6126

UV Causes Dramatic Changes in Aggregation with Mixtures of Photoactive and Inert Surfactants Julian Eastoe,* Margarita Sanchez Dominguez, Paul Wyatt, and Andrew J. Orr-Ewing School of Chemistry, University of Bristol, Bristol BS8 1TS, U.K.

Richard K. Heenan ISIS-CLRC, Rutherford Appleton Laboratory, Chilton, Oxon OX11 0QX, U.K. Received February 10, 2004. In Final Form: May 4, 2004 Aqueous mixtures of photosensitive and inert surfactants have been prepared; photoreactions and changes in aggregation after irradiation have been characterized. The photosensitive component was a stilbenecontaining gemini photosurfactant (E-SGP), and the inert surfactants were either DTAB (dodecyltrimethylammonium bromide) or one of two different gemini surfactants, 12-4-12 or 16-4-16 (butanediyl-1,4bis(dodecyldimethylammonium bromide) or butanediyl-1,4-bis(hexadecyldimethylammonium bromide)). Small-angle neutron scattering (SANS) studies revealed that in general the initial nonirradiated mixed systems form vesicle-type aggregates (100-200 Å radius), in equilibrium with some smaller charged spheroidal or ellipsoidal micelles (∼20 Å radius). In all cases, UV irradiation resulted in disruption of these vesicles and the formation of charged micelles. 1H NMR showed that the main photoproduct is the cis, anti, cis dimer of E-SGP (ZEZ-DiSGP); hence, photochemically induced changes in the reactive SGP drive significant changes in the preferred aggregation structure. These results demonstrate the utility of photoactive surfactants in mixtures with inert analogues.

Introduction New technologies and applications are driving the development of more “chemically sophisticated” or functionalized surfactant molecules. Nanotechnology, for example, is a latent “customer” of surfactant science, since the varied morphologies of surfactant aggregates offer multiple possibilities for nanoengineering. Control over aggregation by external triggers could represent an advantage toward the development of intelligent nanomaterials. Targeted delivery is another example where a straightforward switch between different aggregation states is needed. Several methods have been employed to induce large changes in aggregation: pH,1-3 electrochemical potential,4-6 temperature,7,8 addition of electrolytes,9,10 and light.11-21 A good example is an anionic * Corresponding author. Phone: +44-117-9289180. Fax: +44117-9250612. E-mail: [email protected]. (1) Cheng, W. J.; Li, G. Z.; Zhou, G. W.; Zhai, L. M.; Li, Z. M. Chem. Phys. Lett. 2003, 374, 482. (2) Bergsma, M.; Fielden, M. L.; Engberts, J. B. F. N. J. Colloid Interface Sci. 2001, 243, 491. (3) Fukuda, H.; Goto, A.; Yoshioka, H.; Goto, R.; Morigaki, K.; Walde, P. Langmuir 2001, 17, 4223. (4) Aydogan, N.; Abbott, N. L. Langmuir 2001, 17, 5703. (5) Tsuchiya, K.; Sakai, H.; Saji, T.; Abe, M. Langmuir 2003, 19, 9343. (6) Sakai, H.; Imamura, H.; Kakizawa, Y.; Abe, M.; Kondo, Y.; Yoshino, N.; Harwell, J. H. Denki Kagaku oyobi Kogyo Butsuri Kagaku 1997, 65, 669. (7) Yin, H. Q.; Zhou, Z. K.; Huang, J. B.; Zheng, R.; Zhang, Y. Y. Angew. Chem., Int. Ed. 2003, 42, 2188. (8) Majhi, P. R.; Blume, A. Langmuir 2001, 17, 3844. (9) Sein, A.; Engberts, J. B. F. N.; Vanderlinden, E.; Vandepas, J. C. Langmuir 1993, 9, 1714. (10) Ranganathan, R.; Tran, L.; Bales, B. L. J. Phys. Chem. B 2000, 104, 2260. (11) Kunitake, T.; Nakashima, N.; Shimomura, M.; Okahata, Y. J. Am. Chem. Soc. 1980, 102, 6642. (12) Sakai, H.; Matsumura, A.; Yokoyama, S.; Saji, T.; Abe, M. J. Phys. Chem. B 1999, 103, 10737. (13) Haubs, M.; Ringsdorf, H. New J. Chem. 1987, 11, 151.

ferrocenyl surfactant; this undergoes a monomer-to-vesicle transition after oxidation.4 On the other hand, pH-induced vesicle-to-micelle transitions with sugar-amine gemini surfactants2 and temperature-driven micelle-to-vesicle switches with mixtures of sodium dodecyl sulfate and dodecyltriethylammonium bromide7 have also been observed. The use of UV light is also promising, since there is no need to change the system composition or thermodynamic conditions. There are various examples of light-induced morphology changes.11-21 With a single-chain azobenzene photosurfactant, a transition from large globular micelles to short rods was confirmed by transmission electron microscopy (TEM).11 The same technique was used to identify a reversible destruction of vesicles formed by mixtures of a cationic azobenzene photosurfactant and sodium dodecylbenzene sulfonate;12 disruption of the vesicles was followed by precipitation. A similar change was obtained with a photodestructible surfactant comprising a benzylammonium moiety.13 For an azodibenzoic acid (related in structure to a gemini), a transition from infinite hydrogen-bonded linear tapes into cyclic tetramers, which further aggregate to form long rods through (14) Veronese, A.; Berclaz, N.; Luisi, P. L. J. Phys. Chem. B 1998, 102, 7078. (15) Haubs, M.; Ringsdorf, H. Angew. Chem., Int. Ed. Engl. 1985, 24, 882. (16) Rakotondradany, F.; Whitehead, M. A.; Lebuis, A. M.; Sleiman, H. F. Chem.sEur. J. 2003, 9, 4771. (17) Szczubialka, K.; Nowakowska, M. Polymer 2003, 44, 5269. (18) Taguchi, M.; Li, G.; Gu, Z.; Sato, O.; Einaga, Y. Chem. Mater. 2003, 15, 4756. (19) Deng, Y.; Li, Y.; Tuo, X.; Wang, X. Abstracts of Papers, 226th National Meeting of the American Chemical Society, New York, NY, Sept 7-11, 2003; American Chemical Society: Washington, DC. (20) Higuchi, M.; Kinoshita, T. J. Photochem. Photobiol., B 1998, 42, 143. (21) Eastoe, J.; Sanchez Dominguez, M.; Wyatt, P.; Beeby, A.; Heenan, R. K. Langmuir 2002, 18, 7837.

10.1021/la0496486 CCC: $27.50 © 2004 American Chemical Society Published on Web 06/16/2004

UV Causes Dramatic Changes in Aggregation

Langmuir, Vol. 20, No. 15, 2004 6121

Scheme 1. (a) E-SGP Surfactant and the Following Possible Photoreactions: (b) Cis-Trans Isomerization and (c and d) Dimerization

Figure 1. Surfactants used in mixtures with E-SGP.

stacking interactions, was suggested both by TEM and by dynamic light scattering (DLS).16 Recently, the photoinduced destabilization of vesicles formed by a stilbene-containing gemini photosurfactant (E-SGP, Scheme 1) was reported.21 Unlike previous examples where the final irradiated state became insoluble,12,13 soluble spherical charged micelles are formed after irradiation of the SGP vesicles.21 This effect was shown to be caused by dimerization of E-SGP (confirmed by 1H NMR). The spacer group in E-SGP was shown to undergo a transition from a stiff planar geometry to a twisted conformation, and this change promotes higher interfacial curvature, consistent with the different aggregation structures. Such light-induced vesicle-to-micelle transitions, which can be achieved after only minutes of irradiation, could represent a viable controlled-release system. This new paper builds on the initial work with pure SGP.21 In the present study, E-SGP was mixed with three different inert cationic surfactants and the presence of vesicles was confirmed by small-angle neutron scattering (SANS). The vesicle size distribution was found to be dependent on the amount of E-SGP present, suggesting possibilities for controlling vesicle size. These results show that expensive photosurfactants can be “diluted” into inert surfactant micelles and still at low levels can be used to drive dramatic light-induced changes in aggregation. Experimental Section Materials. Photosurfactant E-SGP was synthesized and characterized as described previously.21 The critical micelle concentration (cmc) of the initial E-SGP was 0.19 mM.21 Dodecyltrimethylammonium bromide (DTAB, Figure 1) was used as received (Lancaster, 97% minimum). Gemini surfactants 124-12 and 16-4-16 (shown in Figure 1) were synthesized as follows. N,N,N′,N′-tetramethyl-1,4-butane diamine (1 equiv, Aldrich 97%) and n-bromoalkane (2.2 equiv, n-bromododecane for 12-4-12 or n-bromohexanecane for 16-4-16, Aldrich 98%) were refluxed in dry ethanol for 48 h. The solvent was removed under reduced pressure; the crude product was recrystallized four times from either ethyl acetate and ethanol (12-4-12) or ethyl acetate and chloroform (16-4-16). The products were dried to yield white crystals. Characterization for 12-4-12. Elemental analysis. Found: C, 58.8; H, 11.6; N, 4.7; Br, 25.4. Requires C32H70N2Br2: C, 59.8; H, 10.9; N, 4.4; Br, 24.9. 1H NMR as expected. Critical micelle concentration (cmc, conductivity), 1.1 mM at 25 °C; reported,22 1.2 mM.

Characterization for 16-4-16. Elemental analysis. Found: C, 62.4; H, 12.0; N, 3.7; Br, 23.6. Requires C40H86N2Br2: C, 63.7; H, 11.4; N, 3.7; Br, 21.2. 1H NMR as expected. Critical micelle concentration (cmc, conductivity), 0.018 mM at 40 °C; reported,22 0.027 mM. Techniques. Glassware cleaning and conductivity (for cmc determinations) were conducted as previously reported.23-25 Ultraviolet-visible (UV-vis) absorption spectra were measured on a Unicam UV2 instrument in 1 cm path length cells. All samples were diluted before measuring the spectrum with a few drops of ethanol (to destroy aggregates, hence sharpening the spectra) and the appropriate amount of water. 1H NMR experiments were conducted in a JEOL Delta GX/270 instrument; samples were initially made up in D2O, irradiated (or not), and diluted by 50% with methanol-d4 (Cambridge Isotope Laboratories, 99.8%) before measurement. Small-Angle Neutron Scattering. SANS experiments were carried out on the time-of-flight LOQ instrument at ISIS, U.K., as previously reported (e.g., ref 25) but by using a 12 mm diameter beam. Hydrocarbon surfactant micelles were contrasted against D2O (Fluorochem, 99.9%), so that for a given shape the scattering intensity reports on particle number density (concentration). Absolute intensities were obtained by calibration with a partially deuterated polymer standard (described in ref 25). Measurements were conducted in circular quartz cells with a 5 mm path length and a 1.8 cm diameter. The Q range was 0.0073-0.23 Å-1. The samples were thermostated at 50 °C, since the Krafft temperature (TK) for E-SGP was found to be ∼45 °C.21 To follow UV-induced changes in aggregation, nonirradiated and irradiated samples were studied. The concentrations were chosen so that samples would be at least 1% in volume fraction, to obtain a convenient scattering intensity. To achieve this, different total surfactant concentrations (){[E-SGP] + [mixer]}) were used, 32, 16, and 14 mM for the E-SGP/DTAB, E-SGP/12-4-12, and E-SGP/16-416 series, respectively. Different mole fractions of E-SGP, defined as XSGP ) [E-SGP]/{[E-SGP]+ [mixer]} were studied, 0, 0.25, 0.5, 0.75, and 1.0. Note that, although the samples differ in the absolute level of photosurfactant, the relative concentrations (mole fractions (XSGP values)) are the same for all three systems. Scattering data were fitted using the interactive FISH program,26 a flexible multimodel suite that uses a standard iterative leastsquares method for a variety of different form factors P(Q), structure factors S(Q), and polydispersity functions. Full accounts of these scattering laws are given elsewhere.26-33 As detailed in (22) Zana, R.; Benrraou, M.; Rueff, R. Langmuir 1991, 7, 1072. (23) Eastoe, J.; Paul, A.; Rankin, A.; Wat, R. Langmuir 2001, 17, 7873. (24) Downer, A.; Eastoe, J.; Pitt, A.; Penfold, J.; Heenan, R. K. Colloids Surf., A 1999, 156, 33. (25) Summers, M.; Eastoe, J.; Davis, S.; Du, Z.; Richardson, R.; Heenan, R. K.; Steytler, D. C.; Grillo, I. Langmuir 2001, 17, 5388. (26) Heenan, R. K. FISH Data Analysis Program; Rutherford Appleton Laboratory; Report RAL-89-129; CCLRC: Didcot, U.K., 1989. (27) Hayter, J. B.; Penfold, J. J. Chem. Soc., Faraday Trans. 1 1981, 77, 1851. (28) Kotlarchyk, M.; Chen, S.-H.; Huang, J. S.; Kim, M. W. Phys. Rev. A 1984, 29, 2054. (29) Hayter, B. J. Mol. Phys. 1981, 42, 109. (30) Hayter, B. J. Mol. Phys. 1982, 46, 651. (31) Hayter, J. B.; Penfold, J. Colloid Polym. Sci. 1983, 261, 1022. (32) Livesey, I. J. Chem. Soc., Faraday Trans. 2 1987, 83, 1445.

6122

Langmuir, Vol. 20, No. 15, 2004

Eastoe et al.

Table 1. Composition of Irradiated Samples at Respective Photostationary States, as estimated by 1H NMRa mixer DTAB

XSGP % E-SGP % Z-SGP % ZEZ-DiSGP % EEE-DiSGP

0.25 0.50 0.75 1.0 12-4-12 0.25 0.50 0.75 1.0 16-4-16 0.25 0.50 0.75 1.0

16.6 11.2 7.7 4.7 22.6 12.5 3.7 3.2 34.2 13.7 6.2 4.2

12.9 6.9 5.1 4.7 11.1 7.5 0.5 2.3 19.4 8.0 5.2 4.0

66.5 80.0 85.8 88.9 66.3 76.9 93.6 91.1 46.4 73.9 86.1 88.5

4.0 1.9 1.4 1.7 0.0 3.1 2.3 3.4 0.0 4.5 2.5 3.3

a Total surfactant concentrations: DTAB series, 32 mM; 12-412 series, 16 mM; 16-4-16 series, 14 mM.

Figure 2. Spectral changes of mixtures of E-SGP with DTAB: (O) initial spectrum, same for all mixtures; (- - -) irradiated XSGP ) 0.25; (‚‚‚‚) irradiated XSGP ) 0.5; (-‚‚-) irradiated XSGP ) 0.75; (s) irradiated XSGP ) 1. All samples were diluted to an effective concentration of 0.04 mM E-SGP for comparison of the spectra. the Supporting Information, the model employed here was a distribution of large unilamellar, hollow-shell vesicles, polydisperse in radius, in equilibrium with small charged micelles: the balance of populations (vesicle-to-micelle transition) depends on composition (XSGP) and irradiation. Photoirradiation. The mixtures were irradiated in quartz vessels (5 mL) with a 100 W high-pressure Hg white source (unfiltered); the samples were thermostated at 50 °C and stirred with a magnetic flea. The photoreactions were monitored by UVvis; a photostationary state was achieved within 2-12 min, depending on the initial E-SGP concentration. For the NMR analyses of the photoproduct composition (Table 1), the solutions contained the same concentrations as those for the SANS analyses. For the reversibility experiments (Figure 2), monochromatic UV light was produced either by the Hg source (254 nm) or by a laser rig (266 nm). A visible wavelength beam is produced from an optical parametric oscillator (Spectra Physics MOPO 73010), pumped by the 355 nm output of a frequency tripled Nd: YAG laser (Spectra Physics Pro-250-10); the frequency was doubled into the UV using a KDP (potassium dihydrogen phosphate) crystal. The samples were stirred with a magnetic flea, and maximum conversion from dimer to trans was achieved within 12 min of irradiation. The irradiation cycles were carried out on post-cmc samples of pure SGP at 3 mM, which were then diluted to 0.027 mM to record the UV-vis spectra. (33) Eastoe, J. New Physico-Chemical Techniques for the Characterisation of Complex Food Systems; Dickenson, E., Ed.; Blackie: Glasgow, U.K., 1995; p 268.

Results and Discussion Photochemistry. General Observations with Mixed Surfactants. For stilbenoid compounds, it is well-known that there are two dominant UV-induced reactions:34-37 cis-trans isomerization and dimerization (Scheme 1). Extensive work has been carried out with pure E-SGP:21 above the cmc, UV-vis spectroscopy, liquid chromatography-mass spectroscopy (LC/MS), computer simulations, and 1H NMR showed that dimerization predominates over isomerization. For the present study, the E-SGP/inert surfactant mixtures were investigated by UV-vis spectroscopy, confirming attainment of a stationary state after sufficient UV irradiation. Figure 2 shows absorption spectra obtained before and after irradiation for the SGP/DTAB mixture series, at a total surfactant concentration of 3 mM. To allow for comparisons between spectra, the samples were diluted to an effective SGP concentration of 0.04 mM. In general, the spectral changes noted for the 12-4-12 and 16-4-16 mixture series were similar. Since all these surfactants are cationic, it is most likely that the mixed micelles are locally homogeneous; hence, it can be assumed that the photoreactions occur within a mixed micellar matrix. From Figure 2, it is clear that the extent of photoreaction depends on surfactant composition via the mole fraction (XSGP). Comparing the UV-vis spectra for XSGP ) 0.25 (- - -) and XSGP ) 1 (s, pure SGP), for the mixture, the absorbance band for the final dimer state is lower, while the residual peak for the initial E-SGP state is higher than that for the pure compound. This suggests that dilution of E-SGP into mixed micelles reduces the extent of photodimerization. Table 1 shows 1H NMR peak integrations to determine final compositions, in terms of the relative amounts of E-SGP, Z-SGP, ZEZ-DiSGP, and EEE-DiSGP species (as before21). The values represent an average of three separate analyses for each condition, each displaying good sample-to-sample reproducibility. The trends given in Table 1 are consistent with the UV-vis spectra (Figure 2), indicating that the extent of dimerization is lower as the mixed micelles become richer in inert surfactant (i.e., the fraction of ZEZ-DiSGP + EEE-DiSGP generated decreases with decreasing XSGP). This is likely to be due to a dilution effect, since, in mixed micelles with lower SGP content, the molecules become progressively isolated from each other, hence reducing the ability to dimerize. Reversibility. In previous studies,21 reversibility of the photoreaction was investigated with pre-cmc samples of SGP (10-2 mM) using 254 nm monochromatic light from a white Hg source: a photostationary mixture of ∼1:1 of the dimer and trans forms was achieved. To prove if reversibility was also possible in post-cmc samples (3 mM), tests were done using a 266 nm laser, chosen since absorption of E-SGP is a minimum at this wavelength. Figure 3 shows the UV-vis monitored recovery of E-SGP observed after irradiation, with either 254 nm monochromatic light from an Hg lamp or the 266 nm laser irradiation. For these post-cmc concentrations, the maximum regeneration of E-SGP from the DiSGP was seen with the laser, resulting in a recovery of ∼50% after 12 min of irradiation. These results are significant for demonstrating that partial reversibility is indeed possible in SGP micellar solutions. However, since the maximum (34) Meier, H. Angew. Chem., Int. Ed. Engl. 1992, 31, 1399. (35) Green, B. S.; Heller L. J. Org. Chem. 1974, 39, 196. (36) Trecker, D. J. Org. Photochem. 1969, 2, 63. (37) Hicks, J. M.; Vandersall, M. T.; Sitzmann, E. V.; Eisenthal, K. B. Chem. Phys. Lett. 1987, 135, 413.

UV Causes Dramatic Changes in Aggregation

Langmuir, Vol. 20, No. 15, 2004 6123

Figure 3. Reversibility of the photoreaction: (s) initial spectrum, E-SGP; (O) irradiated with 254 nm from an Hg lamp, mainly ZEZ-SGP; (0) irradiated with 266 nm laser light, ∼1:1 E-SGP/ZEZ-DiSGP. The cycle of irradiations was carried out on post-cmc samples at 3 mM but diluted to 0.027 mM to record the spectra.

regeneration of E-SGP is only ∼50%, this also highlights limitations to the uses of photosurfactants. Small-Angle Neutron Scattering (SANS). General Observations and Limitations of the Technique. SANS was employed to study changes in aggregation as a function of composition (variation of XSGP) and the effect of UV irradiation. The data are displayed in Figures 4, 5, and 6 for mixtures of E-SGP with DTAB, 12-4-12, and 16-416, respectively; preirradiated samples are shown in part a and postirradiated conditions in part b. Prior to irradiation, the scattering profile and intensity changed progressively with increasing XSGP in the mixtures (Figures 4a, 5a, and 6a). Irradiation of the samples of pure inert surfactant only (XSGP ) 0) had no effect on the SANS data. However, for samples containing E-SGP (XSGP > 0), irradiation caused dramatic changes in the SANS data (compare the data in parts a and b of Figure 4, etc.). Taken together with the UV-vis and NMR evidence, these results imply that the UV-induced changes in I(Q) are caused by photochemical transformation of SGP.21 There are obvious limitations to the analyses of these SANS data, owing to the multicomponent nature of the samples (four or five different surfactant species present) and importantly the difficulty for contrasting individual components by selective isotopic labeling. Since the surfactants and any photoproducts comprise hydrocarbon chains, the dominant contrast step is for D2O (Fm ) 6.37 × 1010 cm-2) against the H-surfactant mixtures (between -0.3 and 0 × 1010 cm-2, depending on the assumed molecular volume, given in the Supporting Information). A reasonable approximation is that the scattering length density (Fp) of the micelles remains essentially constant as a function of composition and irradiation: in any case, Fp would not differ much from that for normal cationic micelles and still be greatly different to that for D2O. Below, it is shown that certain examples of these SANS data are consistent with a mixed population of globular (spherical or ellipsoidal) aggregates coexisting with hollowshell vesicles (planar). Owing to these limitations of neutron contrast, it is not possible to ascertain how the different surfactant species are distributed among the

Figure 4. SANS data (a) before and (b) after irradiation for DTAB/E-SGP mixtures: (b) XSGP ) 0; (O) XSGP ) 0.25; (2) XSGP ) 0.5; (4) XSGP ) 0.75; (9) XSGP ) 1.0. The solid lines represent the fits described in the text. Total surfactant concentration, [DTAB] + [E-SGP] ) 32 mM.

aggregates of different curvatures. For any such complex mixture, the mean curvature (planar, cylindrical, or spherical) would be set by an average packing parameter for the equilibrium mixture of surfactants, which in this case have widely different molecular structures (single chain, gemini, dimer, etc.). It is likely that as the photochemical changes proceed the compositions of the globular and planar aggregates could become imbalanced, and partitioning of certain species into aggregates of preferred curvature could occur. However, it is not possible to follow this process by SANS, unless the surfactants are specifically labeled. Furthermore, the SANS final states (Figures 4b, 5b, and 6b) are consistent with small globular micelles, suggesting that the systems settle into one kind of preferred curvature. Hence, implicit in the analyses is that the surfactant species mix homogeneously and that the curvature is affected by composition (XSGP) and the photochemical state of the SGP, pre- and post-UV irradiation. Given these caveats, a detailed interpretation of these SANS data in terms of changes in aggregation structure is discussed below. Initial States: Nonirradiated Samples of Pure Inert Surfactants. The SANS data from pure inert micelles (XSGP ) 0) were consistent with previous studies,38-41 displaying a correlation peak characteristic of spherical or ellipsoidal charged micelles. For pure DTAB, the ellipsoid of revolu(38) Bergstro¨m, M.; Pedersen, J. S. Phys. Chem. Chem. Phys. 1999, 1, 4437. (39) Berr, S. S. J. Phys. Chem. 1987, 91, 4760. (40) Danino, D.; Talmon, Y.; Zana, R. Langmuir 1995, 11, 1448. (41) De, S.; Aswal, V. K.; Goyal, P. S.; Bhattacharya, S. J. Phys. Chem. 1996, 100, 11664.

6124

Langmuir, Vol. 20, No. 15, 2004

Eastoe et al.

Figure 5. SANS data (a) before and (b) after irradiation for 12-4-12/E-SGP mixtures: (b) XSGP ) 0; (O) XSGP ) 0.25; (2) XSGP ) 0.5; (4) XSGP ) 0.75; (9) XSGP ) 1.0. The solid lines represent the fits described in the text. Total surfactant concentration, [DTAB] + [12-4-12] ) 16 mM.

Figure 6. SANS data (a) before and (b) after irradiation for 16-4-16/E-SGP mixtures: (b) XSGP ) 0; (O) XSGP ) 0.25; (2) XSGP ) 0.5; (4) XSGP ) 0.75; (9) XSGP ) 1.0. The solid lines represent the fits described in the text. Total surfactant concentration, [DTAB] + [16-4-16] ) 14 mM.

tion model fits the data well, with semiminor axis a ) c ) 11.3 Å, semimajor axis b ) 28.9 Å, axial ratio (x) ) 2.55, and fractional charge (R) ) 0.22.38,39 For 12-4-12, the SANS data could be fitted with prolate ellipsoids, with a ) c ) 15.2 Å, b ) 34.2 Å, x ) 2.25, and R ) 0.20. Previous cryoTEM studies with 12-4-12 showed densely packed spherical or spheroidal micelles.40 For pure DTAB and 12-4-12, a spherical form factor was also tested; however, the fits for ellipsoidal aggregates were better in terms of scale factor testing. Bhattacharya et al. studied 16-4-16 by SANS:41 no correlation peak was observed, owing to the limited Q range in that study (0.02-0.3 Å-1). In the present work, a well-defined peak was observed around 0.017 Å-1, and data analysis was consistent with prolate ellipsoidal charged micelles. Several x values were tested, and the best fit was obtained for x ) 12, with a ) c ) 23.3 Å, b ) 280 Å, and R ) 0.42. This is in good agreement with the earlier cryo-TEM work by Zana et al., showing entangled threadlike micelles.40 Initial States: Nonirradiated Samples of SGP and Inert Surfactants. Addition of E-SGP results in clear Q-2 scattering for most mixtures, and this contribution to I(Q) is enhanced with increasing XSGP. Although this power law can be interpreted in various ways (e.g., Gaussian coils, lamellae, or hollow-shell vesicles), none of which can be completely ruled out, a polydisperse hollow-shell vesicle form factor is consistent with the scattering (Supporting Information). As seen before for pure E-SGP,21 initially, the mixtures had a milky appearance, which may be attributed to the presence of large aggregates, vesicles, or a lamellar dispersion, and thus, the use of a vesicle

Table 2. Structural Parameters Fitted to SANS Data from Mixed Surfactant Systemsa mixer

XSGP

vesicle thickness (∆)/Å

sphere radius (Rav)/Å

DTAB 12-4-12

0.25-1.0 0.25 0.50-1.0 0.50 0.75 1.0

20.0 20.0 20.0 24.0 24.0 20.0

20.0 23.2 20.0 26.9 23.3 20.0

16-4-16

a Total surfactant concentration: DTAB series, 32 mM; 12-4-12 series, 16 mM; 16-4-16 series, 14 mM.

form factor does have some basis. The oscillations at low Q values were accounted for by introducing the “free form” distribution N(R) for vesicle radii (ref 21 and Supporting Information). However, a polydisperse vesicle model alone did not fit the entire Q range adequately, especially for samples with lower XSGP values. To alleviate this, the model was combined with a spherical charged micelle scattering law, which worked well. This has been shown before for samples of pure E-SGP at intermediate stages of irradiation, where E-SGP and ZEZ-DiSGP coexist.21 Given the large number of free parameters in this combined model, a spherical rather than ellipsoidal form factor was eventually preferred. Thus, in general, the fitted lines for XSGP g 0.25 (Figures 4a, 5a, and 6a) represent different vesicle size distribution functions (Supporting Information), in equilibrium with smaller charged spherical micelles. Table 2 shows the vesicle thickness (∆) and sphere radius (Rav) for the different systems. Note that ∆ and Rav for 16-4-16 mixtures

UV Causes Dramatic Changes in Aggregation

Langmuir, Vol. 20, No. 15, 2004 6125

Table 3. Apparent Change in Vesicles: Charged Micelle Population as a Function of Composition Based on Analysis of SANS Data for Nonirradiated Samplesa mixer DTAB

12-4-12

16-4-16b

XSGP

% total surfactant in vesicles

% total surfactant in spheres

0.25 0.50 0.75 1.0 0.25 0.50 0.75 1.0 0.25 0.50 0.75 1.0

70 94 99 99 45 60 87 100 0 59 82 100

30 6 1 1 55 40 13 0 100 41 18 0

a Total surfactant concentration: DTAB series, 32 mM; 12-4-12 series, 16 mM; 16-4-16 series, 14 mM. b For 16-4-16 at XSGP ) 0.25, the fits represent elongated prolate ellipsoids rather than spheres.

are slightly higher than those of pure E-SGP or DTAB and 12-4-12 systems, and this is consistent with the presence of a larger alkyl chain (C16 vs C12). However, these shell thicknesses with geminis are inconsistent with the “normal” end-to-end structure seen for classic lipid bilayers. The typical thickness 20 Å, being just slightly larger than the molecular lengths, suggests some interdigitation of chains between layers. For geminis, the headgroup spacer can be envisaged to promote more effective packing, allowing for interdigitation compared with classical lipids. In the absence of detailed contrast variation experiments, it is not possible to resolve any further detail about the internal bilayer structure. As shown in the Supporting Information, the vesicle size distribution was found to vary with composition, and for XSGP ) 0.25 and 0.5, only one population was detected. For XSGP g 0.75, bimodal distributions gave a better description of the SANS data, hence accounting for differences in the oscillations at low Q values seen for certain compositions. The reasons for these apparent changes in vesicle populations are unclear. From absolute scattering intensities and fitted scale factors, it is possible to estimate the balance of population vesicles to micelles. In general, the volume fraction of surfactant in vesicles increases as XSGP increases. The data in Table 3 show that the formation of vesicles is preferred in mixtures with DTAB; this is readily seen as a large difference in the scattering profile when comparing XSGP ) 0 with XSGP ) 0.25 in Figure 4a. The most remarkable result is that for a DTAB mixture at XSGP ) 0.5; almost all surfactant appears to be present as vesicles. For 12-4-12, this fraction is lower, whereas, in the case of 16-4-16, vesicles were detected only for XSGP g 0.5%. The system consisting of 16-4-16 at XSGP ) 0.25 could not be fitted as vesicles; instead, prolate ellipsoidal charged micelles with semiminor axis a ) c ) 20.5 Å, semimajor axis b ) 185 Å, axial ratio (x) ) b/a ) 9, and fractional molecular charge (R) ) 0.38 worked well. This is similar to pure 16-4-16 (see above), and it appears that such elongated ellipsoidal micelles have a low enough curvature to accommodate E-SGP molecules. Given its rigid spacer, this surfactant alone would otherwise aggregate as a flatter vesicular phase.21 Final States: Irradiated Samples. All systems with added photosurfactant (XSGP g 0.25) exhibited large scattering changes after irradiation. This effect is seen clearly when comparing the SANS data (a) before and (b) after irradiation in Figures 4, 5, and 6. In all cases, after irradiation, I(Q) shows a maximum around Q ∼ 0.05 Å-1,

Figure 7. Structural parameters for irradiated systems: (a) ellipsoidal and (b) spherical charged micelle fits for (b) DTAB mixtures, (O) 12-4-12 mixtures, and (1) 16-4-16 mixtures. The main figures are for (a) semiminor axis a for the ellipsoidal fits and (b) the sphere radius (Rav) for the spherical fits. The inset in part a is for the axial ratio (x) for ellipsoids; note the secondary axis x2 is for 16-4-16 mixtures. The inset in part b shows polydispersity σ/Rav; the secondary axis (σ/Rav)2 is for DTAB mixtures.

suggesting the formation of ellipsoidal or spherical charged micelles. For most samples, no residue of the initial strong Q-2 decay remained, which indicates breakdown of the initial vesicles. On the basis of these observations, test fits were carried out using either the ellipsoidal or polydisperse spherical charged micelle models to elucidate if there were any major differences between the qualities of the fits. On balance, for the entire set of data, taking into account the agreements in absolute scale factors and chi-squared parameters, it was not possible to distinguish between the two models. Hence, all that can be said is that the final states represent small charged globular micelles, which (obviously) exhibit a different scattering profile from that of the initial vesicle (or vesicle + micelle) systems. Figure 7 summarizes the structural parameters obtained from ellipsoidal (Figure 7a) and spherical (Figure 7b) fits, and several trends can be seen. For the DTAB series, there is a steady increase in ellipsoid semiminor

6126

Langmuir, Vol. 20, No. 15, 2004

axis a and the axial ratio (x) (x ) b/a), whereas, for the 12-4-12 series, there is little change for compositions XSGP g 0.25. For the 16-4-16 systems, both a and x notably decrease, consistent with fewer long chain molecules (C16 vs C12) in the ellipsoidal aggregates as XSGP increases. (Details on the fractional molecular charge (R) for ellipsoidal and spherical fits can be found in the Supporting Information.) It is likely that the effective surfactant packing parameters (P ) v/a0lc) for 12-4-12 and ZEZ-DiSGP are quite similar, since these mixtures have very similar scattering profiles (Figure 5b) and fitted model parameters (Figure 7). This finding also supports the previous explanation for the formation of small spherical/ellipsoidal micelles by ZEZ-DiSGP. Even though this dimer is twice the size of E-SGP, it is plausible that the central ZEZ-cyclobutane group provides sufficient conformational disorder for the alkyl chains to allow small, spheroidal or ellipsoidal aggregates, in a fashion similar to the simpler molecule 12-4-12. Summary and Conclusions A stilbene-containing gemini photosurfactant (E-SGP) has been mixed with the inert cationic surfactants DTAB, 12-4-12, and 16-4-16, and photoreactions as well as changes in aggregation in aqueous systems have been characterized. The rationale for this work was to find out if large changes in aggregation, as seen in pure E-SGP solutions,21 could still be induced after dilution with lightinsensitive analogues. This could have implications for efficiency and cost-effectiveness if these systems were to be commercialized. As shown in Table 1, 1H NMR indicates that ZEZ-DiSGP is the dominant photoproduct, irrespective of the nature of the second surfactant or mixture composition. However, the extent of reaction and formation of Z-SGP depends on the initial composition, and this effect seems to be due to a diluting effect imposed by the inert molecules, coexisting

Eastoe et al.

in the aggregates with photoactive E-SGP. Partial photoreversibility has been demonstrated in post-cmc aqueous solutions of E-SGP. The maximum 50% recovery of the initial E-SGP, with 266 nm laser irradiation, points to limitations of the use of photosurfactants. SANS studies have shown that significant changes in aggregation can be induced by both mixture composition and UV irradiation. Owing to the multicomponent nature of these systems, limitations are imposed on the SANS analyses, which have been outlined. Given these limitations and a plausible first approximation model for the distribution of components among the aggregates, it has been shown that the SANS data are consistent with vesicle-like aggregates in equilibrium with spherical or ellipsoidal charged micelles. The proportion of vesicles depends on mixture composition and the nature of the mixer surfactant. The vesicle radius distribution was seen to depend on mixer composition and type, opening the possibility for vesicle size tuning. Irradiation causes large changes in aggregation, to form small spherical or ellipsoidal charged micelles. These fascinating systems may have potential applications for photocontrol over solubilization, delivery systems, and, possibly, rheology. Acknowledgment. M.S.D. is grateful to the Mexican organization CONACyT (National Council of Science and Technology, Grant No. 151737) for a scholarship. We also acknowledge the CLRC for allocation of beam time at ISIS and grants toward consumables and travel. Financial support from the EPSRC Portfolio Grant LASER is gratefully acknowledged. Supporting Information Available: A discussion of SANS theory, a figure showing fitted vesicle size distributions, and a table showing the S(Q) of irradiated samples. This material is available free of charge via the Internet at http://pubs.acs.org. LA0496486