Langmuir 2004, 20, 2579-2583
2579
Nuclear Magnetic Resonance Investigation of the Micellar Properties of Two-Headed Surfactant Systems: The Disodium 4-Alkyl-3-sulfonatosuccinates. 2. The Dynamics of the Chains Comprising the Interior of Two-Headed Surfactant Micelles Michael J. Doyle and D. Gerrard Marangoni* Department of Chemistry, St. Francis Xavier University, P.O. Box 5000, Antigonish, Nova Scotia, B2G 2W5, Canada Received May 28, 2003. In Final Form: January 22, 2004 The dynamics of the carbons comprising the micelles of two members of the family of two-headed surfactants, the disodium 4-alkyl-3-sulfonatosuccinates, has been determined via the application of the two-step model to the 13C relaxation rates and the nuclear Overhauser enhancements (nOe’s) at 200 MHz. The NMR relaxation times, determined from the inversion recovery method, increase steadily as we descend the chain from the headgroup region. The relaxation rate profiles and the order parameters have been calculated from the two-step model for the micellar sulfosuccinate aggregates. We note that the order parameter profile and the fast motion correlation time profile for these two-headed surfactants are distinctly different from those of a typical single-headed, single-tailed surfactant such as dodecyltrimethylammonium bromide, particularly in the headgroup region of the micelle. All these results are interpreted in terms of the effect of adding a second headgroup to a single-headed, single-tailed surfactant.
Introduction In recent years, there has been increasing interest in the dynamics of the self-assembly process. NMR spectroscopy has been emerging as a significant tool in these investigations. NMR self-diffusion experiments are a powerful method for examining the translational properties of micelles and surfactant monomers.1-3 In addition, the molecular dynamics and structural organization of micelles,4-7 polymer/surfactants,8 and biomembrane9-11 systems have been determined from an analysis of the NMR spin-lattice (T1) and spin-spin (T2) relaxation times and nuclear Overhauser enhancement (nOe) factors as a function of the magnetic field strength. For surfactant micelles, the 13C spin-lattice relaxation times are usually frequency dependent, leading to nuclear Overhauser factors (nOe values) that are less than their maximum value. This means that the motions of the chains inside the surfactant micelles are not simple in nature. The “twostep” model has been proposed to account for the nonsimple * To whom correspondence should be addressed. Phone: (902) 867-2324. Fax: (902) 867-2414. E-mail:
[email protected];
[email protected]. (1) Stilbs, P. J. Colloid Interface Sci. 1982, 87, 385-391. (2) Carlfors, J.; Stilbs, P. J. Colloid Interface Sci. 1985, 104, 489499. (3) Stilbs, P. Prog. NMR Spectrosc. 1987, 19, 1-45. (4) Nyden, M.; So¨derman, O.; Wiedmer, S. K.; Riekkola, M. L. J. Dispersion Sci. Technol. 2000, 21, 209-227. (5) Wennerstrom, H.; Lindman, B.; So¨derman, O.; Drakenberg, T.; Rosenholm, B. J. Am. Chem. Soc. 1979, 101, 6860-6864. (6) Schonhoff, M.; So¨derman, O.; Li, Z-.X.; Thomas, R. K. Langmuir 2000, 16, 3971-3976. (7) Jansson, M.; Li, P.; Stilbs, P. J. Phys. Chem. 1987, 91, 52795286. (8) Gao, Z.; Wasylishen, R. E.; Kwak, J. C. T. J. Phys. Chem. 1990, 94, 113-116. (9) Brown, M. F.; Thurmond, R. L.; Dodd, S. W.; Otten, D.; Beyer, K. J. Am. Chem. Soc. 2002, 124, 8471-8484. (10) Douliez, J. P.; Ferrarini, A.; Dufourc, E. J. J. Chem. Phys. 1998, 109, 2513-2518. (11) Saiz, L.; Klein, M. L. J. Am. Chem. Soc. 2001, 123, 7381-7387.
alkyl chain motions in the micelles.12-14 This model consists of a combination of fast internal motions with a correlation time (τfc, usually ascribed to the reorientations of the chains in the micellar interior) and a “slow” motion (τsc, due to the tumbling of the whole micelle and/ or diffusion of surfactant monomers over the micelle surface). The relaxation of the carbon-13 nuclei of surfactant alkyl chains is dominated almost exclusively by the 1H-13C dipolar interactions with directly bonded protons.15,20,21 Hence, the spin-lattice relaxation time is given by
R1 )
1 N ) (χD)2 [J(ωH - ωC) + 3J(ωC) + T1 4 6J(ωH + ωC)] (1)
where N is the number of directly bonded protons, ωH and ωC are the Larmor frequencies (in rad/s) for the 1H and 13 C, respectively, and J(ω) represents the spectral density function. χD is the dipolar coupling constant which is given by
χD )
µo pγCγH 3 4π r
(2)
C-H
where γH and γC are the proton and carbon magnetogyric ratios and rC-H is the effective C-H bond distance and is taken as 1.11 Å.20 The nuclear Overhauser enhancement (12) So¨derman, O.; Walderhaug, H.; Henriksson, U.; Stilbs, P. J . Phys. Chem. 1985, 89, 3693-3701. (13) Ellena, J. F.; Dominey, R. N.; Cafiso, D. S. J. Phys. Chem. 1987, 91, 131-137. (14) So¨derman, O.; Henriksson, U.; Olsson, U. J. Phys. Chem. 1987, 91, 116-120. (15) So¨derman, O.; Stilbs, P. Prog. NMR Spectrosc. 1994, 26, 445482.
10.1021/la034931a CCC: $27.50 © 2004 American Chemical Society Published on Web 03/03/2004
2580
Langmuir, Vol. 20, No. 7, 2004
Doyle and Marangoni
factors (η values) are given by the following equation.
η)
γH[6J(ωH + ωC) - J(ωH - ωC)] γC[J(ωH - ωC) + 3J(ωC) + 6J(ωH + ωC)]
(3)
According to the two-step model proposed first by Wennerstrom et al.,5 the spectral densities in eqs 1 and 2 above are given as a sum involving the spectral densities for the fast (Jf(ω)) and slow motions (Js(ω)) as follows.
J(ω) ) (1 - S2)Jf(ω) + S2Js(ω)
(4)
Here, S is the order parameter of the motions and is given by
1 S ) 〈3 cos2 θ - 1〉f 2
(5)
where θ is the angle between the C-H bond vector and the local director which is taken to be perpendicular to the surface of the two-headed micelle.16 A simplistic interpretation of the order parameter involves the number of possible reorientations of the carbon atom in the alkyl chain.17 A high value of the order parameter indicates that the number of possible reorientations of the C-H bond vector is limited. The lower the value of S, the more isotropic in nature are the motions of the C-H bond vector (an S value of 0 indicates completely isotropic motion). Finally, the fast and slow motions are described by singleexponential functions of the respective correlation times.
Jf,s(ω) )
{
}
τf,s 2 c 5 1 + (ωτf,s)2 c
(6)
Although there are a number of investigations of the dynamics of the micelles on typical surfactants such as sodium dodecyl sulfate (SDS) and dodecyltrimethylammonium bromide (DTAB), only a few investigations have dealt with nontypical surfactants.7,18,19 In this paper, the investigation of the molecular dynamics in a two-headed anionic surfactant system will be presented. The 13C NMR T1 values and the nOe’s have been measured at 200 mHz for two sulfosuccinate surfactants (disodium 4-heptyl-3sulfonatosuccinate or C7SS and disodium 4-octyl-3-sulfonatosuccinate or C8SS). The structure of the sevencarbon sulfosuccinate surfactant (C7SS) is given below.
The molecular dynamics of the surfactant molecules comprising the sulfosuccinate micelles has been deter(16) Walderhaug, H.; So¨derman, O.; Stilbs, P. J. Phys. Chem. 1984, 88, 1655-1662. (17) Brown, M. F.; Williams, G. D. J. Biochem. Biophys. Methods 1985, 11, 71-75. (18) Wong, T. C.; Ikeda, K.; Meguro, K.; So¨derman, O.; Lindman, B. J. Phys. Chem. 1989, 93, 4861-4867. (19) Furo´, I.; Sitnikov, R. Langmuir 1999, 15, 2669-2673.
mined from the application of the two-step model using the single-field method of Lipari and Szabo.20,21 In the case of this single-field method, the slow motion correlation time is input into the calculations; hence, it is of interest to determine if the molecular dynamics is particularly sensitive to our choice of the τsc value. Our results indicate that the values of S and τfs in the headgroup region of the surfactant are quite different from those of a comparable single-headed surfactant and that the numerical values of S and the fast motion correlation time are relatively insensitive to the choice of τsc used in their determination. Experimental Section The synthesis of the sulfosuccinate surfactants has been described in detail previously.22 The purified surfactants were characterized by NMR and FT-IR spectroscopy and by a two phase surfactant titration. Surface tension versus log(csurf) plots were free of any minima, indicating that the final product had no surface active impurities. The NMR relaxation times (1H and proton decoupled 13C) for the sulfosuccinate surfactants were obtained on a Bruker AC200 at St. Francis Xavier (St. F.X.) University; additional proton and carbon-13 spectra were obtained on a Bruker AMX-400 NMR at the Atlantic Region Magnetic Resonance Centre (ARMRC) at Dalhousie University. A typical 13C NMR spectrum for the C7 sulfosuccinate is presented in Figure 1. Nuclear overhauser enhancements (nOe values) were obtained on the Bruker AC200 at St. F.X. from the ratio of the integrated intensity of 13C peaks in two different experiments. In the first, the 13C spectrum was obtained with the proton decoupler on during the pulse delay and data acquisition; the second spectrum was collected with the proton decoupler off during the pulse delay but on during acquisition. The pulse delays used in nOe measurements were >10 T1. 13C T1 values were obtained on both the AC-200 spectrometer at St. F.X. and the AMX-400 NMR spectrometer at Dalhousie University, using the standard pulse programs from the Bruker software library. D2O was used as the solvent for all the NMR experiments. The 1H spectra were referenced to the HOD peak (δ ) 4.81 ppm), while the 13C chemical shifts were referenced to the deuterium lock signal using the method of So¨derman.23 Correlation times and order parameters were extracted from the relaxation time data and the nOe values using the Solver add-in in Microsoft Excel.
Results and Discussion A typical carbon-13 spectrum for C7SS is presented in Figure 1; the assignments of the carbon peaks were discussed in a previous publication.24 The relaxation rate data (i.e., R1 ) 1/T1) and the nOe values at 200 MHz are given in Table 1 for a 0.65 M C8SS solution (critical micelle concentration (cmc) value ) 0.154 M) and in Table 2 for a 0.90 M C7SS surfactant solution (cmc value ) 0.254 M). The R1 values and the nOe values were corrected for the monomer contributions using the two-state model. We can clearly see from Tables 1 and 2 that the R1 values decrease steadily as we descend the chain from the headgroup region to the tail region, in agreement with the R1 profiles observed for single-headed, single-tailed surfactants in the literature.5,6,8,12 For both surfactants, the relaxation times for the carbon nuclei in the headgroup (20) Lipari, G.; Szabo, A. J. Am. Chem. Soc. 1982, 104, 4546-4558. (21) Lipari, G.; Szabo, A. J. Am. Chem. Soc. 1982, 104, 4559-4565. (22) Boucher, G. D.; MacDonald, A. C.; Hawrylak, B. E.; Marangoni, D. G. Can. J. Chem. 1998, 76, 1266-1273. (23) So¨derman, O.; Geuring, P. Colloid Polym. Sci. 1987, 265, 7681. (24) MacDonald, A. C.; Boucher, G. D.; Hawrylak, B. E.; Marangoni, D. G. Can. J. Chem. 1998, 76, 1266-1273.
NMR Investigation of Two-Headed Surfactants
Figure 1. Typical
13C
Langmuir, Vol. 20, No. 7, 2004 2581
NMR spectrum for C7SS.
Table 1. Spin Lattice Relaxation Rates and nOe Values for 0.65 M C8SS carbon no. R1/s-1 nOe (1 + η) carbon no. R1/s-1 nOe (1 + η) 2 3 5 6 7
5.13 2.90 3.18 2.92 2.41
2.39 2.46 2.55 2.55 2.56
8,9 10 11 12
2.00 1.02 0.68 0.42
2.63 2.63 2.67 2.81
Table 2. Spin Lattice Relaxation Rates and nOe Values for 0.90 M C7SS carbon R1/s-1 carbon R1/s-1 no. (200 MHz) nOe (1 + η) no. (200 MHz) nOe (1 + η) 2 3 5 6 7
4.27 2.41 2.82 2.67 1.99
2.32 2.46 2.55 2.57 2.56
8 9 10 11
1.67 1.22 0.82 0.39
2.59 2.60 2.67 2.78
region are short, indicating efficient relaxation of the 13C nuclei, which may be due to the fact that the headgroup carbons of the C7SS and the C8SS are highly constrained. In addition, if we examine the nOe values (also given in Tables 1 and 2 for the individual surfactants) we clearly see that the nOe values are substantially less than the maximum value of 2.988. This observation is consistent with relaxation of the 13C nuclei being dominated by motions that are outside of extreme narrowing;5,12,20,21 that is, the correlation function J(ω) is a combination of both fast and slow motions. In aggregated systems, it is best to estimate the molecular dynamics from the variable frequency relaxation data, provided the relaxation data are available at very low magnetic field strengths (as they are dominated by the slow motions). The high-field relaxation rate data are not particularly sensitive to the rates of the slow motions; hence a reasonable estimate of the rate of the slow motion should be sufficient to determine the order parameter profile and the τfc values for the carbon nuclei comprising the chains of the surfactant. In this study, the relaxation time data and the nOe data at 200 mHz were used to compute the order parameter profiles and the τfc values from the single-field method of Lipari and Szabo.20,21 Hence, we are left with estimating the τsc values from
some other method. In the literature, for surfactants having a micellar radius comparable to that of the sulfosuccinate micelles, the correlation times for the slow motions have been estimated to be around 2.0 ns.5 We can also estimate a value of the slow motion correlation time as described below. The slow motion correlation time is mainly due to two contributions, each with a characteristic correlation time. The first is a contribution due to the rotation tumbling of the whole micelle,
τR )
4πrmic3η 3kBT
(7)
where rmic is the micellar radius, η is the viscosity of the solvent medium, and kB is the Boltzmann constant. The second contribution is from the monomer translational diffusion in the micelle over the micelle surface, with an effective correlation time τD given by
τD )
rmic2 6D
(8)
Here, the monomer diffusion coefficient is that of the monomer over the micellar surface; the rate-limiting step in this process is expected to be due to the diffusion of the doubly charged polar headgroup. Both rotational tumbling and monomer diffusion contribute to the effective slow motion correlation time, τs, as shown below.
1 1 1 ) + τs τR τD
(9)
Using the micellar radius from the literature25 and a diffusion coefficient estimated from the molar conductivities of the alkyl sulfosuccinate ions from previous conductivity data,26 we calculate the slow motion correlation time for both C7SS and C8SS to be approximately 2.4 and 2.8 ns, respectively. (25) Jobe, D.; Reinsborough, V. C. Aust. J. Chem. 1984, 37, 303309. (26) Wiseman, P.; Kennedy, C. A.; Palepu, R.; Marangoni, D. G. J. Solution Chem. 1998, 27, 217-233.
2582
Langmuir, Vol. 20, No. 7, 2004
Doyle and Marangoni
see from Figures 4 and 5 that although there are some differences in the magnitudes of the calculated correlation times, the order parameter profiles and the τfc values are relatively insensitive to the choice of the slow motion correlation time.
Figure 2. Order parameter profile for 0.90 C7SS ([) and 0.65 M C8SS (9) as a function of the carbon chain position.
Figure 3. Fast motion correlation time (τfc) profile for 0.90 C7SS ([) and 0.65 M C8SS (9) as a function of the carbon chain position.
Figure 4. Order parameter profiles (S) for 0.90 M C7SS as a function of the carbon chain position using different values of the slow motion correlation time: (, 1.5 ns; 9, 2.0 ns; 2, 2.5 ns; ×, 3.0 ns; *, 5.0 ns.
Figure 5. Fast motion correlation times (τfc) for 0.65 M C8SS as a function of the carbon chain position using different values of the slow motion correlation time: (, 1.5 ns; 9, 2.0 ns; 2, 2.5 ns; ×, 3.0 ns; *, 5.0 ns.
In Figures 2 and 3, the order parameter profiles and the fast motion correlation times for C7SS and C8SS using the method of Lipari and Szabo, incorporating the τcs times calculated above, are presented. We note that these profiles will be sensitive to the chosen value of τsc. Hence, in Figures 4 and 5, we have plotted the order parameters for C7SS and the τfc values for the C8SS obtained from our experimental nOe’s and R1 values using almost a 1/2 order of magnitude variation in the τsc values. We can clearly
The order parameter profiles and the τfc values for C8SS are substantially different for the headgroup carbons of the sulfosuccinates than for a typical single-headed, singletailed surfactant. These high S values and the τfc values are consistent with the motions of the headgroup carbons in these micelles being more restricted than in the case of a typical surfactant like DTAB or SDS.8,27 These values are even larger than the S values and the τfc values found for a zwitterionic surfactant ((decyldimethylammonio)propanesulfonate) by Jansson et al.7 This is in excellent agreement with what we observed for the 13C chemical shift changes upon micelle formation for the sulfosuccinates previously.22 For the C7SS surfactant, we note that the τfc values are smaller than those of the eight-carbon C8SS surfactant. This may mean that the motions of the carbon nuclei undergoing relaxation are not as spatially restricted and that the conformation of the carbons bearing the headgroups becomes significantly more constrained as the chain length of the surfactant increases. This is also consistent with our 13C NMR chemical shift data. As the chain length of the sulfosuccinate surfactants increased, the 13C chemical shift difference between the monomeric and the micellized surfactants increased. Hence, in the case of the sulfosuccinate surfactants, the increase in the surfactant chain length decreases the area/headgroup, and the packing of the two negatively charged headgroups on the micellar surface becomes more difficult. From the 13C chemical shift measurements, we proposed that, on average, the 1-4 carbons in the headgroup region adopted a predominately gauche configuration and that the carbons bearing the headgroups become significantly more constrained as the surfactant chain length increases.24 This is in excellent agreement with the order parameter profile and the fast motion correlation time profiles obtained from the relaxation time measurements. In addition, for the alkyl chains, the chemical shift differences are observed to be consistent with the micellar interior becoming more fluidlike, especially near the chain ends (this was interpreted in terms of an increase in the number of trans conformers as the sulfosuccinate surfactants form micelles). Note that the interpretation of the order parameter profiles and the fast motion correlation times is again in accord with the 13C δsurf,mic - δsurf,aq values, in that there is more motional freedom in the micelle toward the chain ends, which is in agreement with the order parameter profiles for other surfactant micelles.4,5,7,8,12,14,16,20,21,27 Conclusions The order parameters and the fast motion correlation times have been obtained for sulfosuccinate surfactants from the analysis of the 13C NMR relaxation data. The profiles for the C8SS and the C7SS imply that the motions of headgroup carbon atoms are more spatially restricted than those of typical ionic surfactants. Both systems show a decrease in the order parameters and the τfc values as the carbon atoms are further away from the headgroup (27) Monduzzi, M.; Ceglie, A.; Lindman, B.; So¨derman, O. J. Colloid Interface Sci. 1990, 136, 113-123.
NMR Investigation of Two-Headed Surfactants
region, similar to the profiles of typical surfactants such as SDS and DTAB. However, the large values of S and τfc for the carbons in the headgroup region of the surfactants suggest that the structure of the sulfosuccinate micelles consists of a very constrained region near the micellar palisade layer, whereas the values near the chain ends are similar to those of SDS and DTAB, suggesting a fluid hydrocarbon-like region in the core.
Langmuir, Vol. 20, No. 7, 2004 2583
Acknowledgment. The financial support of NSERC (research grant, D.G.M) and the St. F.X. University Council for Research is greatly appreciated. M.D. is grateful to St. F.X. University for financial support. Discussions with Drs. Michael Lumsden and Donald Hooper of the ARMRC are gratefully acknowledged. LA034931A
