J. Phys. Chem. 1992,96,9555-9557
9555
Does the DLVO Account for Interactlons between Charged Spheric Vesicles? A. M.Carmona-Ribeiro Departamento de Bioquimica, Instituto de Quimica, Universidadede Sa0 Paulo, CP 20780. sa0 Paulo, SP, Brazil (Received:June 5, 1992; In Final Form: July 30, 1992)
{-Potentials for dioctadecyldimethylammonium chloride (DODAC) and sodium dihexadecylphosphate (DHP) vesicles as a function of NaCl concentration (C,) are obtained over an extensive range of salt concentration (10-'-10-' M)using particle microelectrophoresis. These estimates of the surface potential at a fmed vesicle size allow the calculation of theoretical stabilities for the vesicles (W)using a DLVO model adapted for interacting spheric vesicles without free parameters. The Wvalues calculated are at least a factor 10'" higher than the experimental stability ratios (We)from vesicle flocculation under identical experimental conditions. The DLVO does not account for interactions between charged spherical vesicles.
Introduction Charged synthetic amphiphile vesicles are becoming increasingly important as biomimetic systems,ld convenient media for charge separation and storage of light energy,68 model colloids for testing theory?JOand interfacial agents able to cover and modify surfaces at very low concentration." Nevertheless, the interaction forces which are involved in their colloidal stability remain still poorly understood over a range of NaCl concentration (C,). The dependences between initial flocculation rates and NaCl concentration and between the critical coagulation concentration and the sixth power of the counterions valency were those expected from the DLVOS3No correlation was found between d log vo/ d log C, and vesicle size.' The controversial DLVO relationship between size and surface potential stems to be followed by these charged spheric vesicles.1o At low ionic strength, hydration forces were not detected for interacting bilayers of dihexadecyldimethylammoniumacetate (DHDAA) or bromide (DHDAB) by using the surface force a~paratus'z~~ but were detected by the osmotic stress technique.14 At moderate and high ionic strength, direct measurements of the interaction forces between synthetic amphiphile bilayers have not been reported. At pH 5.5-6 and low ionic strength, an additional attraction between adsorbed dihexadecylphosphate (DHP) layers on mica was measured using the surface force apparatus.15 This additional attraction amounts to no more than a doubling of the expected Hamaker constant and is probably due to ion-ion correlation effects. For sonicated DHP vesicles, at pH 5.5-6 and moderate to high ionic strength, a much higher additional attraction was inferred from surface potential datal6 using the kinetical DLVO a p p r ~ a c h .However, ~ sonicated dispersions are not convenient systems to check DLVO because bilayer fragments present in the preparation are nonspheric and possibly have hydrophobic edges.5 In this respect, the large and spheric surfactant vesicles prepared by chloroform vaporization'J would have been suitable. Unfortunately, the absence in the literature of surface potential data for these large vesicles precluded further checking of the kinetical DLVO a p p r ~ a c h . ~ Recently the f-potential for the large vesicles prepared at very low ionic strength was determined.'O There is a concomitant variation of surface potential and vesicle size as a function of the NaCl concentration in the vesicles preparation medium.l0 For checking the kinetical DLVO approach the vesicle size has to be constant over a range of salt concentration. In this work the surface potential for large synthetic amphiphile vesicles at a fixed vesicle size is estimated using vesicle microelectrophoresis. From this, we compare experimentally derived stabilities with those from the kinetical DLVO approach. Our strategy is as follows. Theoretical stability ratios (w) are calculated from f-potential and vesicle size using the kinetical DLVO approach without free parameters? Wvalues are com0022-3654/92/2096-9555$03.00/0
pared with experimental stability ratios (We)from NaC1-induced vesicle flocculation.'
Material and Metbods Dioctadecyldimethylammoniumchloride (DODAC) was obtained from Herga Industrias Quimicas do Brasil. Dihexadecylphosphoric acid (DHPH) was obtained from Sigma (St. Louis,MO) and was used to prepare sodium dihexadecylphosphate (DHP) as described in ref 2. All other reagents were analytical grade and were used without further purification. Water was MILLI-Q quality. Large DODAC and DHP vesicles (LV) were prepared by injecting a chloroform solution of DODAC or DHP into an aqueous D-glucose solution, 75 OC, at pH 5-5.5. DODAC and DHP LV were prepared at a Pglucose concentration (C,) equal to 0.43 and 0.55 M, respectively. The reason for using these C, values was the need to reproduce precisely the experimental conditions for NaC1-induced vesicle flocculation.' The same vesicle size was measured before and after addition of composite isoosmolar solutions of NaCl and D-glucose. The size was found to be constant as a function of time or NaCl concentration up to 0.045 and 0.069 M NaCl, for DODAC and DHP LV, respectively. Sizes were measured using a Malvern 47OOc PCS apparatus employing a Coehrent Innova 90 laser and correcting the diffusion coefficient for the viscosity and refractive index of the vesicles medium. The size quoted is the mean harmonic r-average diameter (D,) of at least 15 independent measurements at 25 OC. Electrophoretic mobilities (EM) were determined at 25 OC using a Rank Brothers Mark I1 particle microelectrophoresis apparatus in the configuration for the flat cell. DHP and DODAC concentration (C) were determined by inorganic phosphorus analysis'' and by amphiphile-dye solubilization in micelles,18respectively. Because of the large effect of amphiphile concentration (C) on EM,'O mobilities had to be determined as a function of amphiphile concentration at a given NaCl concentration. The extrapolation of the EM against C curve to zero amphiphile concentration yielded the EMo value. EMo was corrected to take into account the viscosity of the vesicles medium (EM,). EMoc was then transformed into the reduced mobility, E.I9 From E and D,, the {-potential was calculated using the OBrien and White theory.19 From f-potential and D,, the charge density was calculated using an approximated solution of the nonlinear Poisson-Boltzmann equation for the case of spherical double layers at high potentials as previously described.1°
Computations The total potential energy (V)of interaction as a function of the reduced separation distance ( u ) was calculated from vesicle 0 1992 American Chemical Society
9556 The Journal of Physical Chemistry, Vol. 96, No. 23, 1992
Carmona-Ribeiro
TABLE I: The E M of N8Cl C-th (C,) OE f-potcatirls (f) d C h g e t k d t i e @( 0 ) Of amphiphile C, (M) 7/qo C, (mM) KD2/Zb EMob EMkb DODAC 0.43 1.23 0.09 4 3.0 3.7 0.425 0.4 13 0.394 0.379 0.346 0.550 0.548 0.534 0.501 0.464 0.436
DHP
1.23 1.22 1.21 1.20 1.18 1.33 1.31 1.28 1.20 1.11 1.04
2.70 9.00 18.00 27.00 45.00 0.09 0.90 9.00 27.00 45.90 62.10
23 42 54 73 94 6 17 55 96 125 145
DODAC E* 2.8 2.8 3.0 3.0 2.8 2.6 2.8 3.1 3.2 3.2 3.2 3.1
3.7 4.0 4.0 3.7 3.5 3.7 4.2 4.2 4.3 4.2 4.1
3.0 3.3 3.3 3.1 3.0 2.8 3.2 3.3 3.6 3.8 3.9
WIP V d c V {(mV) 85 62 58 55 49 46 -7 8 -69 -59 -57 -54 -5 1
u (10'
3.2 10.0 16.0 21.0 22.0 26.0 2.7 6.6 16.0 26.0 32.0 35.0
Due to the experimental approach, there is no variation of vesicle size as a function of C,. DODAC and DHP vesicles are prepared in 0.43 and 0.55 M ~+txsc solutions, respectively. In order to vary NaCl concentration, isoosmolar composite solutions of NaCl and D-glucosc were added to the vesicle dispersions. Large DHP vesicles are at pH 5-5.5. D, is 235 and 344 nm for large DODAC and DHP vesicles, respectively. the polydispersity index is 0.3 16 and 0.402, respectively. EMo is the electrophoretic mobility extrapolated to zero amphiphile concentration. Mobilities are corrected considering the viscosity of the composite solution relative to the viscosity of water (7/7,J, Le., EMk = (v/qo)EM,. K is the inverse of the Debye length, in cm-I. u is calculated from {-potentials using nonlinear Poisson-Boltzmann'O 104 cm2 V-I s-l.
4
A
2
PI, \
A
0
20
40
60
80
0
20
40
60
80
(A) Figure 2. The total energy of interaction (V)as a function of the separation distance (in A) as calculated for interacting DODAC (A) and DHP (B) vesicles with 123 and 172-nmmean radius, respectively. The hydrocarbon thickness was taken as 50 (DODAC) and 46 A (DHP), the Hamaker constant, as 5.4 X lo-" ergs in all cases and the NaCl concentration, as 0.04 (a), 0.0631 (b), 0.1 (c, e), 0.126 ( f ) , 0.1585 (d, g), and 0.200 M (h). Stem potential values were assumed to be equal to the {-potentials in Figure 1. One should notice that the magnitude of the resultant repulsive energy is about two orders of magnitude above kT. Thus very high theoretical stabilities are predicted. separation distance
Figure 1. The {-potential of large DODAC (A) and DHP (B) vesicles as a function of the logarithm of the NaCl concentration (C,). The large DHP Vesicles are at pH 5-5.5. The insert show the electrophoretic mobility (EM) of large DODAC vesicles as a function of DODAC concentration (C).The large DODAC vesicles prepared in D-ghCQSe 0.43 M are added to a composite isoosmolal D-ghcose/NaCl solution. For the curve in the insert, EM was measured at a final D-ghCOSe (C,) and NaCl concentration (C,) of 0.346 and 0.045 M, respectively. EM at C = 0, is., EMo is obtained by extrapolation (insert). From EMo corrected to the viscosity of the solution, the {-potential is calculated using the thcory of O'Brien and White.15 The arrows indicate the points taken from the experimental curve for calculating the total energy of interaction as a function of the separation distance between interacting vesicles using the DLVO model adapted for ve~icles.~
radius (0,/2), Stern potential ({-potential), NaCl concentration (Ce),and 5 X lO-" ergs as the Hamaker constant (A) using the DLVO model adapted for vesicles without free parameters? The
computer was programmed to tabulate and plot the coordinates of (V/kT, u) curves. The unit of the inverse of the Debye length ( K ) is cm-', and if the {-potential is expressed in millivolts and the vesicle radius in cm, the unit of Vis erg. The dimensionless dielectric constant of water at 25 OC was taken as 78.54. The stability ratio ( W)was obtained by numerical integration of eq 8 in ref 9. In all computations of Vand Wonly aqueous solutions of 1:l electrolyte at 25 OC were considered. The experimental from vesicle flocculation data' were taken stability values (We) from ref 9. Results .ad Discussion The electrophoretic mobility of the large vesicles (EM)increases with the amphiphile concentration ( C ) . ' O For the vesicles prepared in wglucose solution and with addition of an isoosmolal NaCl/ D-glucose solution, an analogous amphiphile concentration effect was obtained (Figure 1A, insert). Thus,the extrapolation to zero amphiphile concentration in the EM against C curve was necessary for each NaCl solution added to the vesicles. The {-potential decays exponentially with NaCl concentration for both vesicle types (Figure 1, Table I). From the nonlinear Poisson-Boltzmann this is the expected pattern for a single planar surface in a monovalent salt solution.*O As the vesicles are large,
The Journal of Physical Chemistry, Vol. 96, No. 23, 1992 9557
Interactions between Charged Spheric Vesicles
TABLE II: Theoretical ( W ) rad E x p e h W ( W e )Sbbilitics of Large DODAC md DHP Veaich Y I Fuoctioll of Salt Coaccatmtioll (C.)"
amphiphile DODAC
DHP
6 (A) 50
46
D,/2 (X107 cm) 123
172
C, (M)
0.040 0.063 0.100 0.159 0.10 0.126 0.159 0.200
1% c, -1.4 -1.2 -1.0
t (mv)
-0.8 -1 .o -0.9 -0.8 -0.7
40
48 45 42 -50.4 -49.3 48.5 -47.6
VnlU/kT 179 145 116 90 242 22 1 202 183
W,
W
102.8 102.2 101.5 10'.0
TH TH
103.1
TH
102.0 100.7
TH TH
100
TH
1047 1036
Wevalues9 were calculated from initial flocculation rates for NaCI-induced vesicle flocc~lation.~W values were calculated from the {-potentials (Figure l), the vesicle size (Dz/2), the Hamaker constant for two hydrocarbon layers interacting across water (5 X lo-'' ergs), the bilayer thickncs (a), and the NaCl concentration (C,) using the DLVO model adapted for vesicles? The maximum energy in the interaction energy profile as a function of the separation distance between vesicles is VmX. T H means high" to be calculated, Le., the computer could not carry out the numerical integration. the assumption of a planar bilayer is not very far from real. The surface potential in sonicated DHP vesicles also decays exponentially with C?I6 possibly because sonication is a disruptive process which generates planar bilayer fragments.21 Under the present experimental conditions which keep the vesicle size fmed, surface potential and surface charge vary as a function of the added electrolyte concentration (Table I). Nevertheless, by increasing the electrolyte concentration from lo4 to 0.045 M the surface potential decreasesfrom 85 to 46 mV and from -78 to -5 1 mV, for DODAC and DHP vesicles, respectively (Figure 1 and Table I). The range of surface potential variation is narrower for DHP than for DODAC (Table I). Tbe reliability of our determinations of surface potential can be checked by comparing some values for the surface charge density (Table I) with those obtained previously by using the surface force apparatus at low ionic strength. At pH 5.5 and 0.2 mM NaC1,62 nm2per charge or 2.6 mC/m2 were obtained from nonlinear Poisson-Boltzmann and direct surface force measurements.15 At pH 5.5 and 0.1 mM NaCl we estimated 2.7 mC/m2 (Table I). At pH 5.5 and 1 mM NaCl, the surface force technique yielded 31 nm2 per charge (5.2 mC/m2),Is and our mobility measurements for DHP vesicles under similar experimental conditions gave 6.6 mC/m2 (Table I). At 2 mM KBr, for DHDAA bilayers adsorbed on mica, a value of 18 mC/m2 was reported.12 We calculated 10 mC/m2 for DODAC vesicles at 2.7 mM NaCl. The dependence of the {-potential on NaCl concentration (C,) over a C, range where vesicle flocculation is absent allows extrapolation of the surface potential at higher C, (arrows in Figure 1). From the estimated surface potential, the potential energy of interaction as a function of the separation distance between vesicles was calculated using the DLVO model (Figure 2). The resultant repulsive energy has maxima which are very high and vary between 80 and 250 kT depending on NaCl concentration (Figure 2). As a consequence, theoretical stabilities (W)are very high or wen too high to be calculated (Table 11). Wvalues are (Table much higher than the experimental stability values (We) 11). This large discrqmcy indicates the existence of an attractive interaction between vesicles which is clearly not included in the DLVO model. The origin of this additional attractive interaction could be the ionion correlations which increase with electrolyte concentrationu and/or the appearance of defective (hydrophobic) regions in the bilayer due to salt addition. The very large discrepancy between
theoretical and experimental stability ratios (Table 11) and the occurrence of fusion between large DODAC vesicles upon monovalent salt addition4 favor the second hypothesis. The additional attraction observed for DHP layers adsorbed on mica is small and probably due to ion-ion correlation effects.15 For the large DHP vesicles, however, such small effects cannot account for the very large additional attraction presented h m which is possibly caused by the assymmetric ion concentration in the inner and outer vesicle
compartment^.^^ Acknowledgmenr. CNPq, FAPESP, and BID are gratefully acknowledged.
Refeream md Notes (1) Carmona-Ribeiro, A. M.; Chaimovich, H. Biochim. Biophys. Acta 1983, 733, 172. (2) Carmona-Ribciro, A. M.; Yoshida, L. S.;Sesso, A.; Chaimovich, H. J . Colloid Interface Sci. 1984, 100, 433. (3) Carmona-Ribeiro, A. M; Yoshida, L. S.;Chaimovich, H. J. Phys. Chem. 1985,89,2328. (4) Carmona-Ribeiro, A. M.; Chaimovich, H. Biophys. J. 1986,50,621. (5) Carmona-Ribeiro, A. M. Chem. Soc. Reu. 1992, 21 (3). 209. (6) Fendler, J. H. Acc. Chem. Res. 1980, 13, 7. (7) Bratt, P.; Kang, Y. S.;Kcvan, L.; Nakamura, H.; Matsuo, T. J . Phys. Chem. 1991,95,6399. (8) Kotchcvar, A. T.; Krtcvoy, M. M. J. Phys. Chem. 1991, 95, 10345. (9) Carmona-Ribeiro, A. M. J. Phys. Chem. 1989, 93, 2630. (10) Carmona-Ribeiro, A. M.; Midmore, B. R. J. Phys. Chem. 1992,%, 3542. (11) Carmona-Ribeiro. A. M.: Midmore. B. R. Lunmtuir 1992.3. 801. (12j pashley, R. M.; M&ui&n, P. M.; Ninham, B. Brady, J ; Evans, D. F. J. Phys. Chem. 1986, 90, 1637. (13) Marra, J. J. Phys. Chem. 1986, 90, 2145. (14) Parscgian, V. A,; Rand, R. P.; Fuller, N. L. J . Phys. Chem. 1991,95, 4777. (15) Clatsson, P. M.; Carmona-Ribciro, A. M.; Kurihara, K. J . Phys.
c.;
----.--.--
Chem. ..-.... 1989. 93. 917.-
(16) Lukac, S . J. Phys. Chem. 1983,87, 5045. (17) Rouscr, G.; Fleischer, S.;Yamamoto, A. Lipids 1970, 5, 494. (18) Stelmo, M.; Chaimovich, H.; Cuccovia, I. M. J. Colloid Interface Sci. 1987, 117, 200. (19) OBrien, R.; White, L. R. J. Chem. Soc. Faraday Tram. 2 1978.74, 1674. (20) Ccvc, G.; Marsh, D. Phospholipid Bilayers-Physical Principles and Models; Wiley-Interscience: New York, 1987. (21) Carmona-Ribeiro, A. M.; Castuma, C. E.; Sesso, A.; Schreier, S.J. Phys. Chem. 1991, 95, 5361. (22) Attard, P.; Mitchell, J.; Ninham, B. W. J . Chem. Phys. 1980, 89, 4358. (23) Carmona-Ribeiro, A. M.; Hix. S.;Sesso, A. Submitted for publication.
