5888
J . Phys. Chem. 1989, 93, 5888-5894
EXAFS study of Mo/AI2O3 catalysts (prepared by equilibrium adsorption) that the basic hydroxyl groups of the A1203 lead to the adsorption of tetrahedral molybdenum species and other sites on the A1203,presumably coordinatively unsaturated AI3+ sites, are responsible for the adsorption of octahedrally coordinated Mo species. Thus, it is reasonable to tentatively assign surface species A to a tetrahedral surface molybdate species and surface species B to surface polymolybdate species with octahedral coordination. Finally, it is necessary to see whether the Raman spectral features are consistent with the tentative assignments for surface species A and B. It must first be realized, however, that Raman bands characteristic of tetrahedral and octahedral Mo species overlap extensively in the 900-1000-~m-~region.2g Thus, the change observed in the 950-960-cm-l peak shape as molybdenum loading is increased from 2.5 to 8 wt % MOO, (Figure 8) may be due to formation of an additional Raman band characteristic of octahedral polymolybdate species. Features in the Raman spectra of supported Mo catalysts have also been observed in the 800-900-cm-' region (e.g., ref 29) and have been attributed to either M e O - M o or Mo-0-X modes or to a combination of both (X refers to the support, e.g., X = Ti for Ti02) (ref 29 and references herein). The broad background on the low-frequency side of the 950-960-cm-' peak that we observe for Mo/TiO, catalysts with Mo loadings of >2.5 wt % MOO, suggests a species with a high degree of disorder. Thus, it is possible that this broad "peak" on the low-frequency side of the 950-960-cm-' peak is due to a Mo-0-Mo mode, characteristic of polymolybdate, and that the high degree of disorder is a result of the binding of a welldefined polymolybdate species to heterogeneous sites on TiO, (presumably acidic hydroxyl groups) or to the interaction of an (28) Mensch, C. T.J.; van Veen, J. A. R.; van Wingerden, B.; van Dijk, M. P. J . Phys. Chem. 1988, 92, 4961. (29) Payen, E.: Grimblot, J.; Kasztelan, S. J . Phys. Chem. 1987, 91, 6642.
ill-defined polymeric phase with one type of TiO, site.
Conclusions Raman data and thiophene HDS activity measurements suggest that three molybdenum species are present on oxidic M o / T i 0 2 catalysts. A surface Mo interaction species, tentatively assigned to tetrahedral surface molybdate species, is formed up to Mo loadings of 2.5 wt % MOO, and is the oxidic precursor to the most active species for thiophene HDS (intrinsic thiophene HDS activity of -816 cm3 of C4 product/(h g of MOO,)). A second Mo interaction species, tentatively assigned to surface polymolybdate species, is formed for higher M o loadings. The polymolybdate species contributes to the background on the low-frequency side of the broad 950-960-cm-l Raman peak and is less active for thiophene H D S (intrinsic activity of 250 cm3 of C4 product/(h g of MOO,)). ESCA data showed that both of these Mo surface species have similar dispersions. Catalysts with Mo loadings of >8-9 wt % MOO, also contain bulk MOO,. The distribution of these three Mo species as a function of molybdenum loading can be derived from Raman and H D S measurements. Acknowledgment. We gratefully acknowledge Dennis Finseth and Leo Makovsky from the Pittsburgh Energy Technology Center for stimulating discussion and for use of the Raman spectrometer. W e also thank Douglas P. Hoffmann for assistance with data analysis and Thomas Gasmire for machine shop work. This work was supported by the Department of Energy under Grant DEAC02-79ER10485. R.B. Quincy acknowledges the Gulf Oil Corp. and the A. W. Mellon Educational and Charitable Trust for Predoctoral Fellowships. Registry No. Moo3, 1313-27-5; MoS,, 1317-33-5; TiO,, 13463-67-7; thiophene, 110-02-1; n-butane, 106-97-8; I-butene, 106-98-9;butadiene, 106-99-0; cis-2-butene,590-18-1; trans-2-butene, 624-64-6; molybdenum, 11098-99-0.
Threadlike Micelles from Cetyltrimethylammonium Bromide in Aqueous Sodium Naphthalenesulfonate Solutions Studied by Static and Dynamic Light Scattering Wyn Brown,* Karin Johansson, and Mats Almgren Institute of Physical Chemistry, Box 532, University of Uppsala, 751 21 Uppsala, Sweden (Received: November 4 , 1988; In Final Form: January 31, 1989)
Dynamic and static light scattering measurements have been made on both dilute and semidilute solutions of the threadlike micelles formed in equimolar mixtures of CTAB and sodium naphthalenesulfonate in aqueous solutions. In dilute solution (1 mM), the first normal mode of the chains (7,)was isolated from the CONTIN decay time spectra as a function of measurement angle. The value of T~ agreed with that estimated by the free-draining Zimm model. At higher, semidilute concentrations (C > 0.02 M) above which the solutions become viscoelastic, the decay time distribution is bimodal with well-separated components on the time scale. The faster (q2-dependent) component reflects the cooperative motions of the transient network formed through interchain entanglements. The slow (q-independent) component of large amplitude apparently reflects the disruption/coalescence kinetics of the micellar aggregates, which are characterized by a strong positive concentration dependence of the measured relaxation rate (i.e., a relaxation time which decreased with increasing concentration of the CTAB/ naphthalenesulfonate complex) and an activation energy of about 84 kJmol-'.
Introduction ~n recent years, considerable interest has been taken in the solution properties of the extended micellar structures formed when salts are added to suspensions of cationic surfactants. When the salt is a simple one (e.g., NaBr) and is present a t high concentration, long, flexible, threadlike micelles are formed. This was shown, for example, by ~ ~et ~ using~ magnetic l birefrinl ( I ) Porte, G.; Appell, J.; Poggi, Y . J . Phys. Chem. 1980, 84, 3105. (2) Porte, G.; Appell, J . J . Phys. Chem. 1981, 85, 2511.
0022-3654/89/2093-5888$01.50/0
gence and light scattering on dilute solutions of cetylpyridinium bromide in solutions of NaBr of high concentration. It was shown that there are similarities between the properties of such solutions and those of conventional Polymer solutions. Subsequently, there have been numerous studies highlighting the properties of both the dilute surfactant solutions and also more concentrated ones, exploring the dynamical behavior above the overlap concentration where network behavior similar to that of semidilute solutions
(e)
( 3 ) Appell, J.; Porte, G.; Poggi, Y. J . Colloid Interface Sci. 1982, 87, 492.
0 1989 American Chemical Society
Threadlike Micelles from Cetyltrimethylammonium Bromide of flexible polymers has been found."" In the presence of strongly binding counterions, such as salicylate and thiocyanate, more extreme behavior is observed in that highly extended micellar structures are formed a t substantially lower concentrations of surfactant or added salt and these solutions are typified by the striking onset of viscoelastic behavior. Thus, for example, by adding sodium salicylate to an equimolar solution of the surfactant (eg., CTAB), pronounced viscoelastic characteristics are displayed already at very low concentrations.llbsc Various systems of this type have been the subject of extensive investigations by Hoffmann and c o - w o r k e r ~ , with ' ~ ~ ~emphasis ~ on the measurement of the rheological properties, and by Olsson et al.IIa by N M R methods. A more detailed investigation of the dynamical behavior as revealed by light-scattering techniques seemed warranted. Recently, Rao et aI.l4 have proposed a micellar chain model, on the basis of N M R data, in which the micellar beads are linked through molecules of sodium salicylate to form polymer-like chains. This model, however, is at variance with other N M R studies1*J6 where, for instance, a broadening of the IH absorption band has been observed. It is also unlikely as an explanation for such a general phenomenon. The transition to rheopectic and/or viscoelastic solutions was actually first observed in ammonium oleate solutions1' and has also been found to occur at exceedingly low concentrations on addition of certain polar compounds to CTAB, e.g., 1 mM each of @-naphthol, CTAB, and KBr.I8 We have recently observed that when equimolar solutions of CTAB and sodium naphthalenesulfonate are mixed one obtains viscoelastic solutions analogous to those observed with salicylate ions. At high concentrations (C > 0.1 M), more or less rigid, optically clear gels are formed. Preliminary dynamic lightscattering measurements (DLS) showed that, in contrast to the single-exponential correlation functions observed in the CTAB/KBr s y ~ t e m the , ~ ~decay ~ time spectra are multimodal. It was decided to examine the structure of CTAB/naphthalenesulfonate suspensions over a wide range of concentrations ( 10-3-0.2 M) by static and dynamic light scattering to elucidate the origins of this complexity. This concentration range corresponds to 0.15C*-30C* where the overlap concentration for flexible coils is given by C* = 3 M/4?rR,3NA. Here, Rg is the radius of gyration and NA is Avogadro's number.
Experimental Section Materials. Cetyltrimethylammonium bromide (CTAB) and sodium naphthalenesulfonate were purchased from Merck, Darmstadt, FRG, and used without any further purification. The solutions were prepared by weight prior to mixing in appropriate quantities the equimolar solutions of CTAB and sodium naphthalenesulfonate in distilled water. All solutions were filtered through 0.22-pm Millipore filters into Hellma cylindrical lightscattering cells (0.d. = 25 mm) for static light-scattering or IO-" (4) Candau, S. J.; Hirsch, E.; Zana, R. J . Colloid Interface Sci. 1985, 105, 52 I (5) Candau, S. J.; Hirsch, E.; Zana, R.; Adam, M. J . Colloid Interface Sci. 1988, 122, 430. (6) Imae, T.; Kamiya, R.; Ikeda, S. J . Colloid Interface Sci. 1985, 108, 21 5
(7) Imae, T.; Ikeda, S. J . Phys. Chem. 1986, 90, 5216. (8) Imae, T.; Ikeda, S. Colloid Polym. Sci. 1987, 265, 1090. (9) Imae, T.; Abe, A,; Ikeda, S. J . Phys. Chem. 1988, 92, 1548. (IO) Flamberg, A,; Pecora, R. J . Phys. Chem. 1984, 88, 3026. ( 1 1) (a) Olsson, U.; Soderman, 0.;GuCring, P. J . Phys. Chem. 1986, 90, 5223. (b) Gravsholt, S. J . Colloid Interfqce Sci. 1976, 57, 575. (c) Ulmius, J.; Wennerstrom, H.; Johansson, L. B.-A,; Lindblom, G.; Gravsholt, S. J . Phys. Chem. 1979, 83, 2232. (12) Angel, M.; Hoffmann, H.; Lobl, M.; Reizlein, K.; Thurn, H.; Wunderlich, I. f r o g . Colloid Polym. Sci. 1984, 69, 12. (13) Rehage, H.; Hoffmann, H . J . Phys. Chem. 1988, 92, 4712. (14) Rao, U. R. K.; Manohar, C.; Valaulikar, B. S.; Iyer, R. M. J . Phys. Chem. 1987, 91, 3286. (15) Anet, F. A. L. J . Am. Chem. Soc. 1986, 108, 7102. (16) Ulmius, J.; Lindman, B.; Lindblom, G.; Drakenberg, T. J . Colloid Interface Sci. 1978, 65, 88. (17) Hatschek, E. Kolloid-Z. 1925, 37, 25. Hatschek, E.; Jane, R. S. Kolloid-Z. 1926, 38, 33. (18) Nash, T. J . Colloid Sci. 1958, 13, 134; Ibid. 1959, 14, 59. (19) d e Gennes, P . 4 . Scaling Concepts in Polymer Physics; Cornell University Press: London, 1979.
The Journal of Physical Chemistry, Vol. 93, No. 15, 1989 5889 55O
/
0 02 OOL 006 008 Figure 1. Static light-scattering data on the CTAB/sodium naphthalenesulfonate system: concentration dependence of the reduced scattering intensity a t three temperatures.
precision-bore N M R tubes for dynamic light-scattering measurements. Solutions of molar concentration greater than 0.05 M were allowed to stand in the oven at 50 OC for 6 weeks prior to the light-scattering measurements to achieve homogeneity. Static Light Scattering. Intensity light-scattering measurements were made with a photon-counting apparatus supplied by Hamamatsu. The light source was a 3-mW He-Ne laser. The optical constant for vertically polarized light is
K = 4~n,-,~(dn/dc)~/N~X~
(1)
where no is the solvent refractive index, dnldc the refractive index increment (=0.153 mL@ at 25 OC6), and X the wavelength (633 nm). The reduced scattered intensity, K C / R e , was measured on the same solutions as were used for the dynamic light-scattering measurements. Here, Re is the Rayleigh ratio obtained by calibration measurements with benzene: R9,, = 8.51 X 10" at 25 OC.*O As an additional internal standard, intensity measurements were also made on dilute solutions of an essentially monodisperse fraction of poly(ethy1ene oxide), M , = 280000, from Toya Soda Ltd., Tokyo. From the angular dependence of KCIRe in the range 45-135O, the radius of gyration, R,, was estimated. All measurements were made in a concentration range in which multiple scattering effects could be neglected. Dynamic Light Scattering. The experimental arrangement has been described earlier.21 The light source was a He-Ne 50" Spectrophysics laser. An ALV 3000 multibit, multi-r correlator was operated with 23 simultaneous sampling times (covering about 8 decades in delay time) using 191 exponentially spaced channels. All measurements were made in the homodyne mode. The data were routinely analyzed by the cumulants method (two and three terms). The distributions of decay times were obtained by two methods: ( I ) M A X E N T (Maximum Entropy Analysis). A review of its applications to problem areas involving image reconstruction has been given by Gull and Skilling22and its application to DLS data by Livesey et al.23324 The method provides a unique, smoothed, solution that is robust to noise and only shows peaks if demanded by the data. (20) Pike, E. R.; Pomeroy, W. R. M.; Vaughan, J. M. J . Chem. Phys. 1975, 62, 3188.
(21) Brown, W.; Johnsen, R. M.; Stilbs, P. J . Phys. Chem. 1983,87,4548. (22). Gull, S. F.; Skilling, J. In Indirect Imaging; Roberts, J. A,, Ed.; Cambridge University: Cambridge, U.K., 1987. (23) Livesey, A. K.; Delaye, M.; Licinio, P.; Brochon, J . E. Faraday Discuss. Chem. Soc. 1987, No. 83, 247. (24) Livesey, A. K.; Licinio, P.; Delaye, M. J . Chem. Phys. 1986, 84, 5102.
The Journal of Physical Chemistry, Vol. 93, No. 15, 19‘89
5890
Brown et al.
Ox)
002
Figure 2. Angular dependence of the reduced scattering intensity at various concentrations; data at 25 OC.
TABLE I
A. Data on the CTAB/Naphthalenesulfonate System from Static Light-Scattering Measurements A2 X lo5/
temp, ‘C
&fwX lo-’
mL.mol.g-*
R,/A
C*
25
4300 2900 3100
5.4 6.4 1.2
890 660 690
6.6 X IO-’ M (2.4 X g.mL-l)
40 55
8. Molecular Length Parameters
Darameter L (contour length)
I (persistence length) l / r ( K u h n segment length) H (Gaussian coil end-to-end distance) R, (radius of gyration)
vaIueI.4 6830” 360b 9.1‘ 2160d 882‘
I? L = M / M L , where M , = 630 A-’ is the molecular weight per unit length for CTAB.6 bFrom 6Rg2/2I2 = L / l - [ I + e~p[-(L/l)]].~ ‘Kuhn’s length, I / ? = 21. d R 2 = L 2 / y L . eR,2 = ( R 2 / 6 ) for Gaussian coils.
(2) CONTIN. This is a constrained Laplace inversion program provided by P r o v e n ~ h e r . ~It~ provides various solutions with different degrees of smoothing and also a “chosen” solution. In both ( I ) and (2), the moments are given in the output, yielding the relative amplitude and the relaxation rate (r)for each resolved component. In general, we find that the solution chosen from CONTIN is identical with the MAXENT solution. One advantage of CONTIN is that with minimum smoothing it allows the resolution of components that are smoothed out in MAXENT,and this means that, in the current project, the first normal mode for the Gaussian coil could be separated with good precision. The data a t low concentration were also fitted to the doubleexponential model (see Results and Discussion).
0 25 05 0 75 Figure 3. Dynamic light scattering on the CTAB/naphthalenesulfonate system: Values of the average diffusion coefficient obtained by the cumulants method as a function of angle for a concentration of lo-’ M. The increase is due to the progressively increasing contribution of the internal modes (see eq 2).
Increasing temperature always results in a decrease in the size of ionic surfactant micelles in contrast to the situation prevailing in suspensions of nonionic surfactants where micellar growth and/or micellar aggregation is anticipated with increasing temperature and also concentration.21,26No conclusions can be. drawn regarding the shape of the micellar aggregates from the data in Table IA. However, the parameters given in Table IB indicate that the micellar aggregates are bulky, highly extended structures. The measured value of R, (890 A) agrees reasonably with the calculated Gaussian coil quantity (882 A). These qualitative conclusions are in accord with those drawn by other groups on analogous system^.^^^ The often applied term “rodlike” to describe the micellar aggregates, including those formed in the presence of simple salts, is a misnomer since it implies a stiffness that is clearly absent. Dynamic Light Scattering. As mentioned in the Introduction, the present study encompasses a wide concentration range. This was necessary since the decay time spectra found in different concentration regions differ in essentials and reflect different dynamic processes. We will consider these sequentially. ( I ) Low Concentrations [C = l P 3M (0.15C)I. With a flexible long-chain macromolecule, internal relaxation processes, which include random motions of sections of the chain as well as rotational diffusion, also contribute to the line width of the scattered light. In the case of the Gaussian coil, Pecoraz7 has derived an expression for the first-order correlation function, having the following form
g‘”(t) = S,(m exp(-q2Dt)
+ S 2 ( X )exp[-(q2D + 2 / r , ) t ] + ... (2)
where the S ’ s are the dynamic form factors, D is the translational diffusion coefficient, and r 1is the longest intramolecular relaxation time. X = (qRJz with the scattering vector q = (4an/X) sin (e/2). Pecoraz7and Perico et aL2*have calculated amplitude factors as a function of the variable X for the first five terms in eq 2. The relative contributions of the translational diffusion and the various internal modes depend on qR, as follows: At qR,