806
Langmuir 1988,4, 806-812
relating K and C,, to surfactant, water, and oil interactions on a molecular basis is needed.
Acknowledgment. We are grateful to E. Kaler, M. Kahlweit, and R. Strey for discussions. This work was
funded in part under NSF Grant No. PHY82-17853 supplemented by funds from NASA at the University of California at Santa Barbara and Grant No. CHE 8542620. D.R. is specially grateful to W. M. Gelbart for his kind hospitality at the University of California at Los Angeles.
Properties and Structures of Three- and Four-Component Microemulsionst J. E. Puig and I. Rodriguez-Siordia Facultad de Ciencias QuEmicas, Uniuersidad de Guadalajara, Guadalajara, Jal44430, Mexico
L. I. Rangel-Zamudio Departmento de Quimica, Cinuestav Mexico, D.F. 07000, Mexico
J. F. Billman and E. W. Kaler* Department of Chemical Engineering, BF-IO, University of Washington, Seattle, Washington 98195 Received January 26, 1988. I n Final Form: April 8, 1988 We report phase behavior, small-angle X-ray scattering (SAXS), fluorescencespectroscopy, conductimetry, and differential scanning calorimetry of didodecyldimethylammoniumbromide (DDAB)/cyclohexane/ water and DDAB/decane/water microemulsions with and without NaBr. We demonstrate that the hydrocarbon tails of the surfactant are fluid in microemulsions at room temperature. The observed phase behavior exhibits a complex dependence upon salt concentration, yet microemulsion structure as probed by SAXS and conductimetry is not altered by the addition of salt. Pyrene fluorescence is used to probe the internal interfacial layers, and evidence for the compression of the surfactant tails with decreasing curvature is obtained.
1. Introduction Surfactants have the remarkable ability to self-assemble in so1ution.l When that solution contains both oil and water there are diverse possibilities for the microstructure of the resulting microemulsions. The usual case is that the minor component-either oil or water-clusters into quasispherical droplets to disperse in the major component. Oil and water are separated by a surfactant-rich film. As the curvature of that film tends to zero, the droplets may merge and fuse into an equilibrium bicontinuous structure.2 The evidence for this progression of structure has been extracted from the results of experiments on microemulsions formed with five components: oil, water, surfactant, cosurfactant, and electr~lyte.~.~ Interpretation cf experiments on such systems is made difficult by the partitioning of cosurfactant and surfactant between the oil and water domains. In addition, the influence of critical points (i.e., critical opalescence) in the phase diagram can obscure some structural details. To remove some of the complexity attendant to the study of five-component systems, it is useful to detail the behavior of microemulsions found with the minimum number of components: oil, water, and surfactant. Once the behavior of the prototypical systems is understood then the roles of, for example, an electrolyte mitigating electrostatic interactions, a cosurfactant altering the rigidity of the surfactant film, or approach to a critical point can
be explored in a controlled fashion. Kahlweit and Stref and Shinoda and Friberg6 have focused on the phase behavior of simple nonionics with the question of structure in those mixtures just now being e ~ p l o r e d . Evans ~ and Ninham"12 and others13 have examined the structural details in microemulsions formed with the cationic surfactant didodecyldimethylammonium bromide (DDAB), water, and hydrocarbons. The structure and phase behavior of AOT-oil-water mixtures has also been explored.14J5 (1) Physics of Amphiphilic Layers, Meunier, J., Langevin, D., Boccara, N., Eds.; Springer-Verlag: Berlin, 1987. (2) Scriven, L. E. Nature (London) 1976,263,123. Scriven, L. E. In Micellization, Solubilization, and Microemulsions; Mittal, K., Ed.; Plenum, New York, 1977; p 877. (3) Auvray, L.; Cotton, J. P.; Ober, R.; Taupin, C. J.Phys. (Les Ulis, Fr.) 1984, 45, 913. (4) Kaler, E. W.; Davis, H. T.; Scriven, L. E. J. Chem. Phys. 1983, 79, 5685. (5) Kahlweit, M.; Strey, R. &new Chem. 1985, 24, 654. (6) Shinoda, K.; Friberg, S. Adu. Colloid Interface Sci. 1975,4, 281. (7) Lichterfeld, F.; Schmeling, T.; Strey, R. J.Phys. Chem. 1986,90, 5762. (8) Chen, S. J.; Evans, D. F.; Ninham, B. W. J.Phys. Chem. 1984,88, 1631. (9) Chen, S. J.; Evans, D. F.; Ninham, B. W. J.Phys. Chem. 1986,90, 842. (10) Evans, D. F.; Mitchell, D. J.; Ninham, B. W. J.Phys. Chem. 1986, 90, 2817. (11) Chen, V.; Evans, D. F.; Ninham, B. W. J.Phys. Chem. 1987,91, 1823. ~ .
* Author t o whom correspondence should be addressed. 'Presented at the symposium on "Fundamental and Applied Aspects of Microemulsions, 11",61st Colloid and Surface Science Symposium, Ann Arbor, MI, June 21-23, 1987; E. Kaler, Chairman.
0743-7463/88/2404-0806$01.50/0
~
_
(12) Blum, F. D.; Pickup, S.; Ninham, B. W.; Chen, S. J.; Evans, D. F . J. Phys. Chem. 1985,89, 711. (13) Samseth, J., in preparation. (14) Kotlarchyk, M.; Chen, S.-H.; Huang, J. S.; Kim, M. W. Phys. Reu. A 1984,29, 2054.
1988 American Chemical Society
Three- and Four- Component Microemulsions
Langmuir, Vol. 4, No. 4, 1988 807
DDAB was chosen for studies of three-component microemulsions because of its insolubility in both oil and water and because its molecular geometry satisfies steric constraints thought necessary for microemulsion formation. Yet even in this deceptively simple system there are surprises. Electrical conductivity and NMR measurements give a qualitative indication that a structure of connected, water-containing cylindrical regions is present at low water concentration.1° As the volume fraction of water increases, the structure undergoes a "reverse percolation" transition to a dispersion of disconnected water spheres dispersed in oil. The details of the cylinder/sphere transition depend on the curvature of the surfactant film, which is set by oil penetration into the tails of the surfactant and can be modified by variation of the ionic head group interactions with added electrolyte. The cylinder/sphere transition has been quantitatively documented very recently by using small-angle X-ray scattering (SAXS) interpreted with a model of "disordered open connected" structure.l@ls Here we report properties and structures of DDAB/ water/cyclohexane and DDAB/water/decane microemulsions made with and without NaBr. In light of previous work, cyclohexane is chosen to contrast with decane because it penetrates the DDAB tail region completely and so changes the curvature of the surfactant sheet.9 Because decane penetrates to a lesser extent,8 the opportunity to observe the effect of added electrolyte is enhanced. The probes of structure are small-angle X-ray scattering (SAXS), fluorescence spectroscopy, differential scanning calorimetry (DSC), conductimetry, and viscometry. We document that the DDAB tails are fluid at room temperature, report measurements of the specific oil/water interface, and show that addition of NaBr has a dramatic effect on phase diagrams but does not alter microstructure within the corresponding phases.
Vonk.wa The absolute scale was determined by measurement of a previously calibrated Lupolen platelet. The scattering patterns were corrected for sample cell scattering and sample transmission. The microemulsion samples were held in 1.0-mmdiameter thin-walled glass capillaries, and all SAXS experiments were performed at room temperature (22 f 1 "C). Conductance was measured at 1000 Hz with a Wheatstone bridge and a Jones-type cell (cell constant of 50 cm-') or an immersion cell (1 cm-'). Viscosities were measured with calibrated Cannon-Fenske viscometers (100,200,or 300). Thermograms were obtained with a Perkin-Elmer DSC-4 using a Perkin-Elmer Intracooler I refrigeration unit. All thermograms were determined with heating and cooling rates of 10 "C/min. Sample pans for volatile specimenta were used to minimize losses by evaporation. Uncorrected spectra were taken with an AMINCO SPF-500 ratio spectrofluorometer. The excitation wavelength for each sample was that of the maximum peak intensity of the excitation spectrum with an emission wavelength of 395 nm. Excimer formation intensities were measured at 394.6 and 470.6 nm, where molecular pyrene and excimer emit, respectively,with emission and excitation band-passes of 1 and 2 nm. The ratio of 11/13of pyrene was measured with emission and excitation band-passes of 0.5 and 2 nm. Fluorescence decay curves were obtained with an ORTEC time-correlated single-photon counting system with a Tracor Northern multichannel analyzer. A Photochemical Research Associates gated lamp, Model 510 B, running at 30 Khz and filed with nitrogen at 0.5 atm, served as a light source. The excitation light pulse was filtered by an Oriel interference filter 5259 with maximum transmittance at 338.1 nm and bandwidth of 8.9 nm. The fluorescence was monitored at right angles and the emission band selected by a Schoffel monochromator (Model 6M-100) at 400 nm with a bandwidth of 17 nm. To reduce the scattered light, an Oriel interference filter 5380 with a maximum transmittance at 339.5 nm was placed between the sample and the monochromator. Samples for fluorescence measurements were degassed by at least 10 freeze-pump-thaw cycles in order to eliminate oxygen, which is an efficient quencher of pyrene fluorescence.
2. Experimental Section Didodecyldimethylammoniumbromide (DDAB)from Eastman Kodak was recrystallized fmt from an acetone-ethyl ether mixture (5050 by volume) and then from ethyl acetate and dried in a
3. Results Partial phase diagrams in weight percent at 25 " C of mixtures of DDAB, NaBr brine, and cyclohexane or decane are shown in Figure 1. In the absence of NaBr, the one-phase region of the cyclohexane system is narrower and the minimum amount of water required for microemulsion formation is smaller than that in the decane system (cf. parts a and b of Figure 1,ref 9). Upon addition of NaBr, the one-phase region of the cyclohexane system shifts toward the DDAB-oil side of the diagram (Figure la), and the minimum amount of water (brine) becomes even smaller (ca. 1%). With decane, on the other hand, the one-phase region shrinks substantially with increasing NaBr concentration (Figure lb). NaBr brine is assumed to be a single component for the purpose of representing four-component mixtures on a ternary diagram. Figure 2 shows electrical conductivities of cyclohexane and decane microemulsions as a function of water content for various surfactant-to-oil ratios (s/o). For cyclohexane, the electrical conductivity is significant (10-35 ps/cm) when the microemulsion contains the minimum amount of water or brine. The conductivities increase and pass through a maximum upon brine addition. However, the conductivity drop after the maximum is smaller for microemulsions containing brine because of the smaller extent of the microemulsion phase (Figure 2a). Addition of salt to decane microemulsions does not alter their electrical conductivity. Without NaBr, there is a conducting-to-nonconducting transition at higher water
vacuum oven. Cyclohexene suitable for spectrophotometry was purchased from Burdick and Jackson. Gold Label decane was purchased from Aldrich. Sodium bromide was reagent grade, purchased from J. T. Baker. Pyrene, 99%+ from Aldrich, was recrystallized 3 times from ethanol. Water was doubly distilled with conductivity less than 1 hS/cm. Samples were made by weighing all the components in glass ampules that were sealed to prevent evaporation, gently handshaken, and equilibrated in a constant-temperature water bath (25 "C) before any measurements were made. In some cases, samples were prepared by diluting low water content, one-phase microemulsions with water or NaBr brine. The SAXS experiments were performed by using a slit-collimated Kratky camera with nickel- and cobalt-filtered Cu K, radiation.lg The entrance slit was varied from 25 to 200 fim to allow intensity measurements over a q range 0.01-0.3 A-1. q = (4*/X) sin (8/2), with 8 the scattering angle and X the X-ray wavelength (1.54 A). The scattered intensities were recorded on Du Pont NDT-75 X-ray film. The films were processed as recommended by the manufacturer, and the optical density of the scattering patterns was measu_red by using a slit-collimated densitometer. These intensities (Z)are smeared by the collimation system; the optical densities of the films were desmeared and converted to absolute intensity I (cm-') by using the methods of (15) Zuleuf, M.; Eicke, H. F. J. Phys. Chem. 1979, 83, 480. (16)Zemb, T.N.; Hyde, S. T.;Derian, P.-J.; Barnes, I. S.; Ninham, B. W. J. Phys. Chem. 1987,91, 3814. (17) Ninham, B. W.; Barnes, I. S.; Hyde, S. T.; Derian, P.-J.; Zemb, T.N.Europhys. Lett. 1987,4, 561. (18) Barnes, I. S.; Hyde, S. T.;Ninham, B. W.; Derian, P.-J.; Drifford, M.;Zemb, T.,in press. (19)Ross, P. A. J. Opt. SOC.Am. 1928, 16, 433.
(20) Vonk, C. G. J.Appl. Cryst. 1975,8, 340. (21) Vonk, C. G. J.Appl. Cryst. 1981, 14, 8. (22) Vonk, C. G. J. Appl. Cryst. 1971,4, 340.
Puig et al.
808 Langmuir, Vol. 4, No. 4, 1988
..
0 0 wt % No& I O u t % No& 2 0 vd% NaEr
WATER
CYCLOHEXANE o
1
~
I
l6
0
No& 2 . 0 wlXNaBr
o 0.0 w l %
3
,/
6
,,/
12
9
15
17
wt % WATER
0 0 wt % NaEr
o 01 w t % N a E r I 0 wt % NaEr
a
I9
16
14
12 I
x
-C IO
e WATER
DECANE
Figure 1. (a, Top) Triangular phase diagram of DDAB/cyclo-
hexane/brine mixtures showing the one-phase microemulsion region for brines containing 0.0, 1.0, and 2.0 wt % NaBr (cf. ref 24). The lines are the dilution paths with the S / O weight ratio of 0.25, 0.43, and 0.67. Phase boundaries were determined by visual observation. (b, Bottom) Triangular phase diagram of DDAB/decane/brine mixtures showing the one-phase microemulsion region for brines containing 0.0, 0.1, and 1.0 wt % NaBr (cf. ref 24). The line is the dilution path with S f 0 weight ratio of 0.67. The point is at 40% DDAB, 40% decane, and 20% water. Phase boundaries were determined by visual observation. content (see Figure 2b and Figure 4 of ref 8). With brine, however, the microemulsion conductivity is always high and the conducting-to-nonconducting transition is not observed. This is again due to the smaller extent of the microemulsion phase in the brine system. The viscosities of cyclohexane and decane microemulsions reflect the structural changes detected with cond ~ c t i v i t y . ~Viscosities ~ display a maximum near the conductance maxima and decreases as the water or brine content increases. As expected, microemulsion viscosities increase as the surfactant/oil ratio increases. Thermograms of pure DDAB, of 20 wt % DDAB in water (a lamellar liquid crystalline sample24), and of DDAB/cyclohexane/water microemulsions are depicted ~~
(23) Rodriguez-Siordia, I.; M.S. Thesis, University of Guadalajara (in progress). (24) Fontell, K.; Ceglie, A.; Lindman, B.; Ninham, B. W. Acta Chem. Scand. 1986, A40, 247.
6
20
30 Wt
40
so
% WATER
Figure 2. Inverse of specific conductance (IC1in k m ) for the cyclohexane (a, top) and decane (b, bottom) microemulsions at various S / O ratios and with 0.0 and 2.0 wt % NaBr.
in Figure 3. The pure surfactant shows no thermal activity in the temperature interval examined (from -60 to 30 "C). The 20% lamellar phase, in contrast, exhibits two thermal transitions between 0 and 30 "C (Figure 3b). The first one, at 0 "C, is due to the melting of bulklike water sandwiched between the surfactant bilayers. The transition at about 16 "C is due to the "melting" of surfactant hydrocarbon tails,23i.e., the so-called gel-to-liquid crystal transition.26 At low water content, cyclohexane microemulsions show three thermal transitions (Figure 3c). The one at 0 "C is again due to melting of bulklike water in the microemulsion phase. The transition at 6.5 "C is due to melting of cyclohexane. The transition starting at ca. 12 "C must be associated with the melting of the hydrocarbon chains of DDAB. Similar thermograms were obtained with decane and tetradecane microemulsions.23 Pyrene excimer formation (measured as the ratio of emission intensities of excimer to monomer, Ie/I,)26927 in (25) Chapman, D.; Williams, R. M.; Ladbrooke, D. H. Chem. Phys. Lipids 1967, I, 445.
Langmuir, Vol. 4, No. 4, 1988 809
Three- and Four-Component Microemulsions
Table I. Pyrene Monomer Lifetimes ( T ) , Ratios of Third to First Vibronic Intensities of Monomer Emission (ZJZl), and Electrical Conductivities (K)of Cyclohexane Microemulsions (S/O= l/*) as a Function of Water Content wt % HI0 T , ns" Zn/Zla K , &/cm 2.0 3.5 5.0 9.0 12.0 16.0
C
H
I
I
a
106.5 f 141.5 147.0 h 163.5 f 173.5 f 169.0 f
1.5
* 1.5 1.5 1.5 1.5 1.5
0.76 0.78 0.79 0.80
102 310 397 531 181
0.80 0.81
Measured in microemulsion made with 1.28 X lob M pyrene in
cyclohexane.
b a
-60 -40 -20 0 20 40 60 80
T,"C Figure 3. Thermograms of (a) pure DDAB, (b) 20% DDAB in water lamellar liquid crystal, and cyclohexane microemulsions with S / O = 0.25 containing 3% (c) and 9% (d) water.
IO
I
20'
(cm-I)
. 10'
0
0 10
0 20
0 30
q
Figure 5. Small-angle X-ray scattering spectra on an absolute scale for DDAB/cyclohexane/O.l w t % NaBr brine microemulsions ( S / O = 0.43) containing various amounts of brine (q has units of A-1).
Ol
0
2
6
4
a
1
PYRENE m y )
Figure 4. Excimer formation (Ze/Zm) as a function of pyrene concentration for varying amounts of water in cyclohexane microemulsions (S/O = 0.43). cyclohexane microemulsions increases as a function of pyrene concentration (Figure 4). For a given microemulsion, I e / I mis linear with pyrene concentration only when the excitation wavelength is the maximum of the excitation spectra (which varies from 337 to 360 nm). If a fixed excitation wavelength (e.g., 337 nm) is used,excimer formation increases nonlinearly with pyrene concentration, even though in pure solvents the ratio always grows linearly regardless of 'the excitation w a ~ e l e n g t h . ~ 'We ~ ~ ~discuss this nonlinear behavior further elsewhere. Lifetimes of pyrene monomer, T,, and the ratio of the intensities of the first and third vibronic bands of the monomer emission spectra (13/11) are reported in Table I as a function of water content in cyclohexane microemulsions. The ratio I I is sensitive to the polarity of hydrophobic is fairly constant for the
region^.^^.^^ k3/Il
(26)Birks, J. B. Photophysics of Aromatic Molecules; Wiley-Interscience: New York, 1970. (27)Thomas, J. K. The Chemistry of Excitation at Interfaces; ACS Monographs 181;American Chemical Society: Washington, D.C., 1984. (28)Rangel-Zamudio,L. I. Ph.D. Thesis, Cinvestav, Mexico (in progress). (29)Kalyanasundaran,K.; Thomas, J. K. J . Am. Chem. Soc. 1977,99, 2039.
various microemulsions (ca. 0.80) but is very different from the value measured in neat cyclohexane (1.68). Monomer lifetimes increase substantially with increasing water content up to about 9% water and then are constant within the experimental uncertainty. Plots of the X-ray-scattered intensity measured for cyclohexane microemulsions on an absolute scale as a function of q display a maximum (Figure 5). The intensity maximum decreases and moves to higher q as the water or brine concentration increases along the dilution lines shown in Figure 1. For large values of q, the absolute intensity from any sample with sharp interfaces between regions of different electron densities (e.g., oil and water) follows Porod's Law:31 lim (441) = 4 lim ($1) = ~ T ( s / v-)pw)2 (~, " ' 9
" ' 9
where S / V is the oil/water specific surface and po and pw are the electronic densities of oil and water. S / V can also be obtained from scattered intensities on a relative scale by normalizing the data by the invariant, given as lim 431 $,"q
1(q) dq;
-
- ' 9
so
X"q
Rs) dq
s/v
440 - +)
where 4 is the volume fraction of water. Porod plots of the data shown in Figure 5 display a broad region where Porod's law applies (Figure 6). This Porod law behavior was used to extrapolate the measured data to large q for (30)Glushko, V.;Thaler, M. S. R.; Karp, C. D. Arch. Biochem. Biophys. 1981,210, 33. (31)Kratky, 0.; Glatter, 0.Small Angle X-ray Scattering; Academic: London, 1982.
810 Langmuir, Vol. 4, No. 4, 1988
Puig et al. !
I
I
I
I
0 w t % NoBr I w t % NoBr
I 4
0.004
1
-
u
0 40
~
'
~ 0.2
0. I
0.3
q 0
0.008
0 004
0012
0.016
0 W I% NaBr
q1
Figure 6. Porod plots (1q3vs q 3 )of the smeared data in Figure 5. From top to bottom, the curves represent samples containing 3, 7, 10, and 13 wt % brine. Table 11. Compositions and SAXS Specific Surfaces for Cyclohexane Microemulsions" @rater S / V , A-1 E, A2 Sf0 wt % brine
a
0.25 0.25
0.1 0.1
0.025 0.061
0.011 0.010
56 50
0.43 0.43 0.43 0.43
0.1 0.1 0.1 0.1
0.025 0.065 0.082 0.113
0.014 0.016 0.013 0.015
47 51 46 54
0.67 0.67 0.67 0.67
0.1 0.1 0.1 0.1
0.027 0.060 0.085 0.112
0.020 0.022 0.022 0.020
44 52 56 53
0.25 0.25
0.0 2.0
0.040 0.040
0.010 0.010
49 49
Data are grouped according to water volume fraction, and is defined in the text.
&,tsmd
calculation of the invariant. The specific surface parameters for cyclohexane microemulsions are in Table 11. Scattering patterns were also measured for three microemulsions containing 40 wt % DDAB, 40 wt '70 decane, and 20 wt % of various brines (the point in Figure lb). The brines were 0, 0.1, and 1.0 wt '70 NaBr, and the three scattering patterns are nearly identical (Figure 7). 4. Discussion There is copious evidence that structure in DDAB microemulsions is set by the balance of surfactant head group repulsions and penetration of the tail region by hydrocarbon."'* A corollary is that addition of electrolyte mitigates head group interactions and influences structure.'l A model for the structure of microemulsions based upon simple geometrical arguments has been proposed.l6ls This model, the disordered open continuous (DOC) model, qualitatively reproduces measured SAXS patterns and is consistent with measured electrical conductivities.16-'8 Previous works have concentrated primarily on the penetration of the surfactant tail by various alkanes. Here we have quantified the effect of NaBr on the location of phase boundaries and on the structure of DDAB microemulsions with measurements of conductivity, fluorescence, and small-angle scattering. It is important to verify that DDAB tails are fluid at room temperature before proceeding with the analysis of structure in microemulsions at room temperature. One requirement for oil uptake and, hence, for microemulsion formation is that the surfactant hydrocarbon tails be fluidlike. Low water content microemulsions of cy-
I w t % NaBr 2 w t % NaBr 30
I (ern-')
IO
0 0
01
02
03
q
Figure 7. Small-angleX-ray scattering spectra on an absolute scale for (a, top) DDAB/cyclohexane/NaBr brine microemulsions ( S / O = 0.25) with 5 wt % of 0.0 and 1.0 wt % NaBr brine and (b, bottom) DDAB/decane/NaBr brine microemulsions ( S / O = 1.0) with 20 w t % of 0.0, 1.0, and 2.0 wt % NaBr brine (q has units of
A-1).
clohexane show three thermal transitions at 0,6.5, and 12 "C (Figure 3c). The 0 and 6.5 "C transitions are due to melting of water and cyclohexane, respectively. The peak near 12 "C was also detected in decane and tetradecane microemulsions and so must be due to a "melting" of the hydrocarbon chains of DDAB. The gel-to-liquid crystal transition in the DDAB/H20 lamellar phase occurs at 16 " C (Figure 3b), and it is likely that the difference in transition temperatures between the lamellar and microemulsion phases is due to oil penetration into the tails. Such penetration would alter the packing of the tails and the oil, and indeed for microemulsions of high water content the thermal transitions of cyclohexane and DDAB merge into one (Figure 3d). These results demonstrate for the first time that the hydrocarbon tails of DDAB in microemulsions a t these compositions are fluid at room temperature. Given that the tails of the surfactant are fluid at room temperature, simple arguments about the steric interactions of the surfactant molecules suggest that the curvature of the interface, and hence the microemulsion structure, is set by the ratio u/(aolc),where u is the volume of the hydrophobic portion of the surfactant, a, the area per head group, and 1, the tail length.l0 The swelling of the surfactant tails by a highly penetrating oil, such as cyclohexane, increases u/(aoZc)and the curvature of the interface. A similar effect should be observed with the addition of electrolyte due to screening of the head group interactions. Co-ion and counterion effects should be smaller in microemulsions made with highly penetrating oils as u / (a& is already large. Indeed, only small changes in the extent of the one-phase region of cyclohexane microemulsions are observed as the NaBr concentration of the brine is increased (Figure la). More dramatic effects are
Three- and Four-Component Microemulsions seen for the less penetrating oil decane (Figure lb). Here the one-phase region detaches from the oil corner and shrinks substantially with increasing NaBr concentration. Similar results are reported for octane microemulsions." In addition, variations of the counterion can also change the phase behavior dramatically and in a complex fashi0n.l' Such is also the case for other amphiphile-oil-water mixtures.32 To test whether the changes in phase behavior upon salt addition are related to structural changes, measurements of electrical conductivity and viscosity were made. In the absence of salt, both cyclohexane and decane microemulsions have high conductivities when little water is present and display a conducting-to-nonconducting transition a t higher water content. Samples containing salt have conductivities similar to those of the water samples. Moreover, the changes in conductivity values with brine content are the same as in the water case except that no percolation transition is detected because of the intrusion of the phase boundary. The fact that conductivities of microemulsions identical except for the presence of salt are the same strongly suggests that the microemulsion structure is not altered appreciably by the addition of salt. Further, since the addition of salt does not enhance conductivity, the fraction of counterions dissociated from the surfactant and so capable of carrying charge must be substantial. As an example, if 30% of the DDAB is assumed dissociated, then in a microemulsion with S / O = 0.25 and containing 10 wt % water, the water is 1.2 M in Br-. Evidently addition of 1 w t % NaBr (0.08 M) to the mixture has little effect on the conductivity. With the macroscopic evidence from conductivity in hand, we now turn to fluorescence and small-angle scattering experiments to probe the microscopic details of the microemulsion structure. Fluorescence probes are an important tool for physicochemical studies of multimolecular aggregate^.^^-^^ Studies with pyrene as a probe have received special c ~ n s i d e r a t i o n . ~Pyrene ~ ~ ~ ' has several interesting photophysical properties that make it adequate for use as an effective microstructural probe, mainly the long lifetime of the monomer (ca. 450 ns in cyclohexane), efficient formation of excimers, and the solvent (environmental polarity) dependence of the vibronic band intensities of monomer fluorescence. All values of 13/11 of pyrene solubilized in cyclohexane microemulsions are in the range of 0.80 regardless of water content (see Table I). These values are half that of pyrene in neat cyclohexane (1.68) but are remarkably close to those of pyrene solubilized in the hydrocarbonlike interior of dodecyldimethylammonium bromide (CTAB) micelles (-0.78).29 Since excitation conditions, quenching, and the presence of solubilizing hydrocarbons do not appear to affect the 13/11 values,30 the low values of 13/11 of the pyrene in CTAB micelles are likely due to the pyrene residing near the oil-water interface, nestled in the hydrocarbon chains of the surfactant tails.29136-39 In the (32) Knickerbocker, B. M.; Pesheck, C. V.; Davis, H. T.; Scriven, L. E. J. Phys. Chem. 1982,86, 393. (33) Gratzel, M.; Thomas, J. K. In Modern Fluorescence Spectroscopy Wehry, E. L., Ed.; Plenum: New York Vol. 2, 1976. (34) Thomas, J. K. Chem. Reu. 1980,80, 283. (35) Rushforth, D. S.;Sanchez-Rubio, M.; Santos-Vidals, L. M.; Wormuth, K. R.; Kaler, E. W.; Cuevas, R.; Puig, J. E. J. Phys. Chem. 1986. 90.6668.
Langmuir, Vol. 4, No. 4, 1988 811 micellar system pyrene partitions exclusively into the surfactant tail region closest to the interface where the hydrocarbon chains have hindered mobility.3639 This environment allows the planar pyrene to slip between the hindered chains and execute limited two-dimensional motion. In contrast, pyrene is excluded from the center of the micelle where the surfactant chains are fluid, and there are no preferred orientations or motions for the pyrene molecule. In cyclohexane microemulsions the 13/11 values suggest that the pyrene is likewise nestled in the hindered region of the surfactant tail chains adjacent to the interface and not solubilized in the bulk oil. The pyrene monomer lifetimes (Table I) and the ratios of excimer to monomer intensities, Ie/Im (Figure 4), in cyclohexane microemulsions are much lower than those in cyclohexane (480 ns26and Ie/Im = 7.3 for a 8 X lo9 M pyrene solution28). Hence, pyrene must be quenched by bromide ions. Since Br- is located in the aqueous domains and pyrene is in the "hindered" region of the surfactant tails, quenching must be by long-range coulombic interaction.40 Given the short lifetime of the pyrene monomer, the formation of pyrene excimers suggests that either the monomer can diffuse freely during its lifetime or the pyrene is confined to a limited region and is relatively immobile, although present a t high concentrations. The former explanation is unlikely as rapid diffusion is at odds with pyrene being amongst the surfactant tails, as indicated by the measured monomer lifetimes and 13/11 values. Finally, the slope of the IJImversus pyrene concentration curve is related to the rate of diffusion of pyrene or, equivalently, the rate of excimer formation. These slopes, the 13/11 values, and the monomer lifetimes all increase as the aqueous content of the microemulsion increases. The implications of these observations will be pursued after the DOC model and the SAXS results are reviewed. The DOC model of structure generates a variety of random closed and bicontinuous structures.16 The DOC structures relevant to this discussion are the "self-avoiding cylinder" models. These structures are generated by locating a set of random points (Poisson points) in space. Spheres are then centered about each of these points. In contrast to the Voronoi model in which the spheres are inflated until their edges form a three-dimensional network, in the DOC model a three-dimensional network is generated by connecting each of the spheres with a cylinder to Z of its nearest neighbors. Z is the coordination number of the sphere and cylinder network. This construct results in a bicontinuous network for values of 2 greater than 1.1 regardless of the internal volume fraction. The ratio u / (u,,lJ is fundamental to the application of this model to microemulsions as it prescribes the surface curvatures, thereby setting the coordination number and the ratio of the sphere radius to the cylinder radius. Interpretation of experiments in terms of the DOC model results in the following picture of the structural evolution of cyclohexane/DDAB/water microernulsions.l8 Microemulsions with tiny water volume fractions (ca. 0.01) are bicontinuous with coordination numbers near 1.1. As the water volume fraction increases to 0.03 the structure swells and Z increases to -3.0. Further increases of the water volume fraction produce a decrease in the coordination number and an increase in the sphere radii, cylinder radii, and the cylinder lengths. This continues until Z falls below 1.1,a t which point the microstructure looses bi-
(36) Dorrance, R. C.; Hunter, T. C. J. Chem. SOC.,Faraday Trans. 1 1972, 68, 1312.
(37) Dorrance, R. C.; Hunter, T. C. J . Chem. SOC.,Faraday Trans. 2 1974, 70, 1572. (38) Dorrance, R. C.; Hunter, T. C. J. Chem. SOC.,Faraday Trans. 2 1977, 73, 89.
~~
(39) Dorrance, R. C.; Hunter, T. C. J . Chem. SOC.,Faraday Trans. 1 1977, 73, 1891. (40) Rangel-Zamudio, L. I.; Rushforth, D. S.; Van Dvyne, R. P. J. Phys. Chem. 1986, 90,807 and references therein.
812 Langmuir, Vol. 4, No. 4, 1988 continuity. Subsequent increases in the water content result in rapid disappearance of the cylinders and swelling of the monodisperse spheres that remain. The SAXS measurements performed in this investigation corroborate the emerging picture of these microemulsions. A maximum in the scattered intensity is observed in each of the desmeared scattering patterns (Figure 5). The inverse of the q values a t which the maximum occurs (q-) is a measure of the characteristic size of the microstructure. This characteristic size increases with both an increase in the brine volume fraction and a decrease in the surfactant-to-oil ratio ( S / O ) . Hence the structure of the microemulsion grows to accommodate the increasing brine content a t constant S / O and becomes increasingly fine to allow an increase in S / O constant brine content. The position of the scattering peak predicted by the DOC model corresponds to the distance between the spheres, which is the cylinder length.16 In terms of the DOC model, this length increases with increasing water volume fraction for the cyclohexane/DDAB/water microemulsions, and so our conclusions are consistent with the predictions of the model. Swelling of the microemulsion structure with increasing water volume fraction as measured by SAXS is reflected in the pyrene fluorescence measurements. The inflation of the microstructure causes a decrease in the interfacial curvature and a compression of the oil-swelled surfactant tails. The compression of the DDAB tails decreases their fluidity and increases the hindered chain volume available to the pyrene molecules. Thus the average distance of the pyrene from the interface increases. An increase in the average separation of the pyrene and the interface with increasing brine content is evidenced by increases in the measured monomer lifetimes, rates of excimer formation, and 13/11values (Table I). The monomer lifetimes increase with aqueous content due to the increased separation of the monomer from the quenching Br-. A commensurate growth in the rate of excimer formation results from the increased monomer lifetime. The decrease in the 13/11 values with increasing brine volume fraction reflects the less frequent sampling of the polar interface by the pyrene. The specific oil-water interface for the cyclohexane microemulsions has been determined from the SAXS patterns (Table 11). This determination presumes the applicability of Porod’s law. The requisite q3 dependence of the smeared scattered intensity was observed over only a limited angular range in each of the SAXS spectra. The restricted range of Porod law behavior we observe diminishes the accuracy of the measured surface areas. The specific surface of the microemulsions decreases slightly as the aqueous volume fraction increases. The DOC model predictions and the observed increase in the rate of formation of pyrene excimer are consistent with a decrease in the specific surface with increasing brine content, but the specific surfaces determined from SAXS do not decrease as much as expected. The surfactant sheet of the microemulsion is densely packed with surfactant as indicated by the area per DDAB molecule values in Table 11. These values vary about a mean of 50 A2/molecule, which is comparable to the reported values for DDAB/cyclohexane/water microemul-
Puig et al. sions (55 A2 (ref 16) and 68 A2 (ref 18)). The density of DDAB a t the interface is not significantly altered by addition of NaBr. An alternative model for the structure of the microemulsions is a dispersion of spherical surfactant-coated water droplets interacting as hard spheres. The shape of the SAXS patterns for the microemulsion samples is consistent with the scattering from a dispersion of monodisperse hard spheres in which the volume fraction of the hard spheres is equated to the microemulsion water content. Given an arbitrary vertical scale this model yields good fits of the the SAXS patterns. However, the use of an absolute scale shows a hard-sphere model is incorrect, as the predicted scattered intensities are a fraction of those measured. Finally, SAXS measurements indicate that salinity has little effect on the structure of microemulsions formed with cyclohexane or decane (Figure 7). On the other hand, the effect of salinity upon the phase diagrams of these systems, especially the latter, is marked. This fact, and our observation that the area per surfactant is not significantly affected by salinity, challenges the proposition that the ratio u/(aOZc)solely determines the structure and phase behavior of DDAB microemulsions.
5. Conclusions This analysis of cyclohexane/DDAB/NaBr brine microemulsions with SAXS, pyrene fluorescence, conductivity, and DSC reveals that relations between microemulsion structure and phase behavior are more complex than those described by simple steric models. The steric model is based on the assumption that the “effective” geometry, u/(Zcao),of the surfactant sets the interfacial curvature, the structure, and ultimately the phase behavior of microemulsions. DSC measurements reveal that a premise of the steric model is correct in that the surfactant tails are fluid at room temperatures and that the effective geometry of the surfactant may be altered by penetrating oils. The structure of the microemulsion is insensitive to the strength of added brine, and structure evolves upon aqueous dilution in accord with the predictions of the DOC model. On the other hand, the phase behavior of these microemulsions is sensitive to the ionic strength of added brine. Our interpretation of measured pyrene flourescence shows it to be a probe of the effective geometry of the surfactant and to be sensitive to the compression of the surfactant tails caused by the aqueous dilution of the microemulsions. Thus fluorescence could be useful in comparing the effectiveness of oils in swelling surfactant tails.
Acknowledgment. This work was supported by the National Science Foundation through PYIA-8351179 and INT-8502390 and by CONACyT through PCCBNA 023157. 1.R.-S. and L.1.R.-Z. thank CONACyT for support. J.F.B. was supported in part by a SOH10 graduate fellowship, and we acknowledge useful conversations with K. R. Wormuth and the helpful comments of an anonymous reviewer. W s t r y No. DDAB, 3282-73-3; NaBr, 7647-15-6; cyclohexane, 110-82-7; decane, 124-18-5.