Subscriber access provided by NATIONAL TAIWAN UNIV
The Journal of Physical Chemistry B is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Article
Diffusiophoresis of Concentrated Suspensions of
Spherical Particles with Distinct Ionic Diffusion Velocities
Jyh-Ping Hsu, James Lou, Yan-Ying He, and Eric Lee
J. Phys. Chem. B, 2007, 111 (10), 2533-2539 • DOI: 10.1021/jp0659305Downloaded from http://pubs.acs.org on November 21, 2008
More About This Article
Additional resources and features associated with this article are available within the HTML version: • Supporting Information
• Links to the 1 articles that cite this article, as of the time of this article download • Access to high resolution figures
• Links to articles and content related to this article
Diffusiophoresis of Concentrated Suspensions of Spherical Particles with Distinct Ionic
Diffusion Velocities
Jyh-Ping Hsu, James Lou, Yan-Ying He, and Eric Lee*
Department of Chemical Engineering, National Taiwan UniVersity, Taipei, Taiwan 10617 ReceiVed: September 12, 2006; In Final Form: January 21, 2007
The diffusiophoresis of a concentrated spherical dispersion of colloidal particles subject to a small electrolyte gradient is analyzed theoretically for an arbitrary zeta potential and double layer thickness. In particular, the influence of the difference in the diffusivities of cations and anions is discussed. A unit cell model is used to simulate a spherical dispersion, and a pseudospectral method is adopted to solve the equations governing the phenomenon under consideration. We show that, as in the case of an infinitely dilute dispersion, when the diffusivities of cations and anions are different, the diffusiophoretic mobility is no longer an even function of the zeta potential or double layer thickness. In contrast to the case of identical diffusivity of cations and anions, a local electric field is induced in the present case due to an unbalanced charge distribution between higher and lower concentration regions. Depending upon the direction of this induced electric field, the diffusiophoretic mobility can be larger or smaller than that for the case of identical diffusivity. The diffusiophoretic mobility is influenced mainly by the induced electric field arising from the difference in the ionic diffusivities, the concentration gradient, and the effect of double layer polarization.
Introduction
Concentration gradients of chemical species in an isothermal gas or liquid mixture are known to cause movement of colloidal particles,1-2and the resulting motion is commonly referred to as diffusiophoresis.3The underlying concept was first proposed by Deryagin et al.,3-6where the importance of this effect was demonstrated in the kinetics of film formation from lattices by ion deposition. In particular, the original primary theory was further extended to cases of general electrolyte solutions.3The growth rate of the latex film on the solid shape with diffusio-phoresis would be comparable in magnitude to the conventional electrodeposition process, where hundreds of volts of electric power are required. Moreover, the Joule heating effect may be generated in the above-mentioned conventional process when an electric field is applied across conductive liquids. Such an effect is not welcomed in that the temperature will increase, which has a negative impact on electrophoretic motion.7-8The proposed new process with diffusiophoresis phenomena can avoid this undesirable heating effect.
A corresponding theoretical analysis was also carried out by Dukhin and Deryagin3,6 and experimentally verified in both electrolyte and nonelectrolyte solutions as well.9-11It was shown by Deryagin and Dukhin6that diffusiophoresis is caused by the polarization of the double layer under the influence of a bulk concentration gradient. The relative motion between electrolyte solutes and charged particles is very complicated, especially in electrolyte solutions. Under the conditions of a thin double layer, symmetric binary electrolyte, constant applied concentration gradient, and isolated sphere in an unbounded solution, Dukhin and Deryagin3derived
where U* is the diffusiophoretic velocity, ∇n∞is the applied
concentration gradient, ze is the charge carried by a single ion, kBis the Boltzmann constant, T is the temperature,ζ is the zeta
potential, and µ are respectively the permittivity and the
viscosity of the fluid, n0 is the bulk electrolyte concentration measured with the absence of colloidal particles and a concen-tration gradient, andζh ) zeζ/4kBT.β ) (D1- D2)/(D1+ D2), D1and D2are respectively the diffusion coefficients of cations and anions, andβ is a dimensionless parameter, which is an
experimentally measurable property. For example, in an aqueous solution,β values are respectively 0, -0.2, and 0.64 for KCl,
NaCl, and HCl.12A negative value ofβ implies that the diffusion velocity of anions is greater than that of cations. The appearance of β in eq 1 demonstrates clearly the involvement and
quantitative contribution of this important parameter in deter-mining the diffusiophoretic mobility, at least in that limiting case.
If the diffusion velocity of cations is different from that of anions,β * 0, the distribution of ionic species in the
neighbor-hood of a particle becomes asymmetric yielding an induced electric field. This induced field then exerts an extra electric force on the particle, in addition to that caused solely by the concentration gradient, as in the case whenβ ) 0. This can be
deduced directly from eq 1 also, where a nonzeroβ alters the
eventual diffusiophoretic velocity U*. Dukhin and co-workers13-15 were the first group to notice this phenomenon. They observed that in a suspension of colloids dispersed in a NaCl aqueous solution, as the sign of the colloidal surface potential varied from negative to positive, the reversion of diffusiophoresis direction might take place accordingly more than four times. In summary, there are two contributions to the diffusiophoresis in general: one due to the imbalance of electrolyte charge which induces an electric field and thus is of the nature of electro-phoresis (β * 0); the other due to the imbalance of the chemical
potential, or more specifically here, the osmosis pressure, associated with the imbalance of electrolyte concentration (β ) 0), sometimes referred to as the “chemiphoresis” by some
groups. Whenβ ) 0, only the chemiphoresis mechanism exists,
and whenβ * 0, both aspects are there.
* Corresponding author. Tel.: 886-2-23622530. E-mail: ericlee@ntu. edu.tw. U* )ζ µ kBT ze ∇n∞ n0 [β + ζh -1 ln(coshζh)] (1)
10.1021/jp0659305 CCC: $37.00 © 2007 American Chemical Society Published on Web 02/22/2007
Based on Gouy-Chapman model,1Prieve and co-workers16-20 further loosened the restriction of low zeta potential in their theoretical analyses and observed strong evidence supporting their proposed diffusiophoretic mechanism from experimental data. They found that colloidal particles might move toward the lower electrolyte concentration asβ * 0 even though the
zeta potential remained low. Dukhin21 provided a thorough physical analysis for this later on. It was shown that the external concentration gradient could induce both a concentration and an electrical dipole moment, causing a nonequilibrium double layer near the particle surface. Pawar et al.22also used the thin double layer model to demonstrate the mechanism for the reversion of the particle motion. This phenomenon arises from the concentration polarization of the double layer surrounding a particle, which induces a local electric field opposing to the normal diffusiophoretic motion.21 Aside from hard spherical particles, Baygents and Saville23also studied numerically the diffusiophoresis of a droplet and a small bubble suspended in an electrolyte solution. Misra et al.24 explored the diffusio-phoresis of a soft particle.
If the concentration of a colloidal dispersion is high, the existence of neighboring particles can no longer be ignored. In this case, the appropriate choice of conditions on the outer boundary is crucial. Therefore, theoretical studies on diffusio-phoresis have been focused on systems with boundaries, physical25or virtual,26,27in recent years. However, these results were restricted to the low surface potential, taking no account of the electroosmotic flow of ions in the double layer. In our recent work on the case ofβ ) 0,28we considered further the polarization effect at arbitrary zeta potential and double layer thickness on the diffusiophoretic mobility of colloidal particles. The osmotic flow of ions in the double layer is taken into account by the inclusion of convection term in the ion conservation equations. A particularly interesting result of our investigation was that the direction of particle motion could change at high zeta potentials and medium double layer thickness even whenβ ) 0. We found there that the
concentra-tion polarizaconcentra-tion of the double layer here produced a coupled mass and charge flow near the surface of a particle, which made its motion much more complicated at high zeta potentials and medium double layer thickness.
In this study, our previous analysis for the case whenβ )
028is extended to the general case whenβ * 0. The diffusio-phoretic behavior of a spherical dispersion under the conditions of arbitrary surface potential and double layer thickness is analyzed thoroughly, with special focus on the electroosmotic flow induced by the distinct ionic diffusion velocities.28The influences of the key factors such as the thickness of double layer, relative diffusion velocities between cations and anions, zeta potential, and volume fraction of a suspensions on the diffusiophoretic mobility are examined as well.
Theory
Let us consider the diffusiophoresis of concentrated rigid spherical particles of radius a in a aqueous solution of z1:z2 where electrolytes z1 and z2 are respectively the valences of cations and anions. The electroneutrality in the bulk liquid phase requires that n20 ) n10/R, n10and n20 being respectively the bulk concentrations of cations and anions, and R ) - z2/z1. The dispersion is simulated by the unit cell model of Kuwa-bara.29Referring to Figure 1, the dispersion is modeled by a representative particle of radius a surrounded by a concentric spherical liquid shell of radius b. Let φ ) (a/b)3, which a measure of the volume fraction of the present dispersion.
Suppose that a uniform concentration gradient∇n0is applied to the system in the z-direction, and as a response, the particle moves in the z-direction with a constant velocity U. The spherical coordinates (r,θ,φ) with the origin located at the center
of the representative particle are adopted.
For the present problem, we have to solve simultaneously the governing equations for the electric, the flow, and the concentration fields. These equations can be summarized as below:
In these expressions, φ is the electrical potential, v is the liquid velocity, F and are respectively the space charge density, and the permittivity of the liquid phase, and e is the elementary charge. The terms nj, fj, and Djare respectively the number
concentration, the concentration flux, and the diffusion coef-ficient of ionic species j. Also, p and µ are respectively the
pressure and the viscosity of the liquid phase.
It is known that in a static environment nj follows a
Boltzmann distribution at equilibrium, if the electrolyte con-centrations and particle’s potentials are not too high, which is the case under study here. The two factors in our system are sometimes several orders of magnitude smaller than where significant deviations from the Boltzmann distribution were reported experimentally. Furthermore, to account for possible concentration polarization arising from the movement of particles in the present problem, we assume a modified Boltzmann distribution of the form:30
Figure 1. Schematic representation of the system under consideration where a is the radius of a particle and b is that of a liquid cell. The applied concentration gradient∇n0is in the z-direction.
∇2 φ ) -F ) -
∑
j)1 2 z jenj (2) ∇‚v ) 0 (3) µ∇2v -∇p - F∇φ ) 0 (4) ∇‚fj) 0, j ) 1, 2 (5) fj) -Dj(
∇nj+njzje kT∇φ)
+ njv, j ) 1, 2 (6) nj) nj0exp(
- zje kBT(φe+ δφ + gj))
(7)That is, the electrical potential is decomposed into φe,δφ, and
gj, representing respectively the equilibrium electric potential
in the corresponding static problem, the induced electric potential arising from the movement of the liquid phase, and an equivalent perturbed potential arising from the convection of the electrolyte ions.
Note that diffusiophoresis can be characterized by an elec-trophoresis driven by the local gradient of electric potential coming from the bulk concentration gradient. This implies that the equations governing a diffusiophoresis problem can be deduced directly from those of the corresponding electrophoresis problem. It can be shown that the governing equations of the present problem, in dimensionless forms, are28,30
where a symbol with an asterisk represents a dimensionless quantity. The term κ ) [∑j)12 nj0(ezj)2/kBT]1/2is the reciprocal
Debye length κ, φr ) ζa/(kT/z1e) is the scaled zeta potential, withζaas the zeta potential of particles. Also, φe
/
) φe/ζa,δφ*
) δφ/ζa, gj/) gj/ζa, and nj /
) nj/n10. Pejis the corresponding
Peclect number of ion j, representing the effect of convection. We assume that the surface potential of a particle remains constant and it is nonconductive. Also, the stern layer
conduc-tivity is neglected.31The net ionic flux across the virtual surface of a cell vanishes, so does the ionic concentration gradient. The particle surface is no slip. For convenience, we assume that the fluid is flowing toward a stationary particle with a scaled velocity U*. On the basis of these assumptions, the boundary conditions associated with the present problem are as follows:
where r* ) r/a, (∇*n0/) )∇n0/(n10/a), and E*4) E*2E*2with
The approach of Prieve and Roman20is adopted where it was assumed that the concentration of solute is only slightly nonuniform over the length scale a, that is, a|∇n0| , n0. In this case, the present problem can be decomposed into two subproblems, where both are of linear nature.20 In the first subproblem, a particle moves with a constant velocity in the absence of the applied concentration gradient, and in the second subproblem, it is fixed in the space when the concentration gradient is applied. If we let F1and F2be respectively the forces acting on the surface of a particles in these two subproblems, then F1 ) f′1(∇*n0/) and F2 ) f′2U*, where f′1 and f′2 are proportional constants.20 Consequently, the scaled diffusio-phoretic mobility Um/ can be expressed as
Given the values of∇*n0/and U*, F1and F2are calculated first through solving the entire set of electrokinetic equations; f′1 Figure 2. Variation of the scaled diffusiophoretic mobility (U*/U0)
as a function of φrat various values of κa atφ ) 0.1, β ) -0.2, and
R ) 1. The dashed curves are results for low surface potential.
∇2 φe /) - (κa)2 (1 + R)φr[exp(- φrφe / ) - exp(Rφrφe / )] (8) ∇*2 δφ* ) - (κa) 2 (1 + R)φr{exp[-φr(φe / + δφ* + g1/ )] -exp[Rφr(φe/+ δφ* + g2/)]}+ (κa)2 (1 + R)φr[exp(-φrφe / ) - exp(Rφrφe/)] (9) ∇*2 gj/- φr∇*φe/‚∇*gj/- φr2Pejv*‚(∇*φe/+ ∇*δφ* + ∇*gj / ) - φr(∇*δφ* + ∇*gj/)‚∇*gj/) 0, j ) 1,2 (10) E*4ψ* ) - (κa) 2 1 + R
{
[
n1 /∂g1 / ∂r*+ n2 /∂g2 / ∂r*(Rn2 / )]
∂φ* ∂θ-[
n1/∂g1 / ∂θ*+ n2 /∂g2 / ∂θ*(Rn2 / )]
∂φ* ∂r*}
sinθ (11) φe/) 1, at r* ) 1 (12) ∂φe/ ∂r*) 0, at r* ) b a (13) ∂δφ* ∂r* ) 0, at r* ) 1 (14) δφ* ) - 1 φr(
1 Pe1 -1 R2 Pe2)(
1 Pe1+ 1 RPe2)
-1 (∇*n0/), at r* )b a (15) ∂g1/ ∂r*) ∂g2/ ∂r*) 0, at r* ) 1 (16){
(δφ* + g1 / ) ) - 1 φr(∇*n0 / ) (δφ* + g2/) )Rφ1 r (∇*n0/) , at r* )b a (17) ψ* ) 0, ∂ψ* ∂r* ) 0, at r* ) 1 (18) ψ* )1 2r* 2U* sin2 θ, at r* )b a (19) E2ψ* ) 0, at r* )b a (20) E*2) ∂ 2 ∂r*2 + sin θ r*2 ∂ ∂θ(
1 sinθ ∂ ∂θ)
(21) Um/ ) U* Ez/ ) -f′1 f′2 (22)and f′2are then determined directly by their definitions, and Um
/
is calculated by eq 22. Results and Discussion
Effect of Zeta Potential. The influence of the key parameters of the system under consideration on its diffusiophoretic mobility is examined through numerical simulation. For illustration, NaCl is chosen as the representative electrolyte, and we have R ) 1, Pe1) 0.39, Pe2) 0.26, and β ) -0.2. Figure 2 shows the variation of the scaled diffusiophoretic mobility (U*/U0) as a function of scaled surface potential φrat various values of κa, where U0) (/µa)(kT/z1e)2is a characteristic diffusiophoretic mobility, which is proportioned to the velocity of a correspond-ing isolated sphere in an unbounded electrolyte solution withβ
) 0, under a unit concentration gradient.3For comparison, the results for the corresponding limiting case of low surface potential at κa ) 1(ref 26) and 10 (ref 27) are also presented. Figure 2 reveals that the result of low surface potential can be recovered as the limiting case of the present analysis by assuming a sufficiently low level of φr. Also, a level of |φr| lower than unity can be considered as sufficiently low. Note that if φrexceeds this level, the deviation of the result based on the low surface potential assumption from the exact value becomes significant. For|φr| g 5, for instance, the deviation can be several fold.
Effect of Ionic Diffusion Coefficients. The influence ofβ
on the diffusiophoretic behavior of a dispersion can be seen by comparing Figures 2 and 3, whereβ ) 0 in the latter. Figure 3
indicates that if β ) 0, (U*/U0) is an even function of φr. However, as shown in Figure 2, the symmetric nature of (U*/ U0) with respect to φrno longer exists when β * 0. Similar behavior was also reported by Prieve20 in an analysis of the behavior of an infinitely dilute dispersion, and it was concluded that the behavior of diffusiophoretic mobility is complicated whenβ * 0. Dukhin21referred to this conclusion and gave a very theoretical explanation for the complicated flow field. He proposed that the diffusiophoresis of a particle arises not only from chemiphoresis but also electrophoresis since the diffusion velocity of cations is different from that of anions (Pe1* Pe2). Here, chemiphoresis refers to the fact that the motion of a charged particle is due to the nonuniform distribution of ions within the electric double layer, and electrophoresis refers to
the induced electric field arising from the difference in the motion speed of cations and anions. This induced electric field is directed toward lower NaCl concentrations since the diffu-sivity of Cl- is larger than that of Na+. Malkin et al.15first predicted the inversion of diffusiophoresis direction, such as that seen in Figure 2 for φr> 0, and later on, Dukhin21gave a detailed physical analysis of this mechanism. We find here that Dukhin’s analysis for a dilute dispersion is also applicable to a concentrated dispersion, as will be elaborated below.
The chemiphoresis and the electrophoresis are reflected by the governing equations and the corresponding boundary conditions in two separate ways: the contribution from the ionic concentration gradient in eq 17, and the two Peclet numbers in eq 15 are not identical (Pe1) 0.39 and Pe2) 0.26). Thus, the eventual speed and direction of particle motion is determined by the balance of these two forces. Besides, according to eq 15, the direction of the induced electric force is determined by the specific charged conditions of a particle and the relative magnitudes of Pe1and Pe2. For a negatively charged particle Figure 3. Variation of the scaled diffusiophoretic mobility (U*/U0)
as a function of φrat various values of κa atφ ) 0.1, β ) 0, and
R ) 1.
Figure 4. Variation of the scaled diffusiophoretic mobility (U*/U0)
as a function of φrat various values of κa atφ ) 0.1, β ) 0.2, and
R ) 1.
Figure 5. Variation of the scaled diffusiophoretic mobility (U*/U0)
as a function of κa at various values of φrfor φre 0, φ ) 0.1,
β ) -0.2, and R ) 1.
(φr< 0), the direction of the induced electric force is the same as that of the major driving force provided by the applied concentration gradient. This is why the direction of particle motion shown in Figure 3 is always the same as that of the concentration gradient in Figure 2 when φr is negative. As illustrated in Figure 3, although originally (U*/U0) increases monotonically with the increase in φr, it does not increase all the way with increasing φr. It actually reaches a local maximum first and then decreases ever after. This is due to the polarization effect generated at a high enough φr or concentrated enough electrolytes.28For the same values of κa and φr, the magnitude of (U*/U0) in Figure 2 is larger than that in Figure 3. This is because when the diffusion velocities of cations and anions are different a particle can feel an additional electric force. For the case of positively charged particles, the concentration gradient competes against the induced electric field. As mentioned previously, the direction of the induced electric force is the same as the sign of the charge carried by a particle. Therefore, if φr > 0 and sufficiently low, (U*/U0) is negative, as is seen in Figure 2. Moreover, as φrincreases, the effect of concentration gradient becomes significant gradually as well, and eventually when it surpasses the effect of ionic diffusion, (U*/U0) will change from negative to positive. This is why the value of (U*/ U0) on the left (right) half of Figure 2 is larger (smaller) than that of Figure 3. It is fully understandable from eq 1, which is based upon an unbounded solution, that (U*/U0) is not a simple even function of φr whenβ * 0. As illustrated in our recent
result,28the illustration in infinite solution is also applicable to a concentrated suspension. In other words, whether (U*/U0) is symmetric about φr) 0 is controlled by β.
We further examine the effect ofβ in more detail by letting
Pe1) 0.26 and Pe2) 0.39, which gives β ) 0.2 and a diffusion velocity of cations which is now greater than that of anions. In this case, instead of anions, cations are now cluster around the area of low concentration. Under this circumstance, the induced electric field is exactly opposite to that of the previous situation. Examining closely the overall behavior of (U*/U0) shown in Figure 4, one can see that the functional dependence of φrand
κa are exactly opposite to that shown in Figure 2. We thus conclude that the sign and the magnitude ofβ are the key factors
in the determination of the diffusiophoretic mobility.
Effect of Double Layer Thickness. The variations of (U*/U0) as a function of κa, a measure for the thickness of double layer, at various values of φrare presented in Figures 5 and 6. This type of figure is capable of providing more insights about the overall behavior of a colloidal dispersion. For example, the polarization effect can easily be observed by the presence of the local maximum and/or minimum. Briefly speaking, a higher κa results in larger (U*/U0) because a higher κa means higher ionic concentrations. In other words, as κa increases, the amount of electrolytes increases accordingly in the vicinity of the particle surface, and hence the electric force exerted upon it increases accordingly. As discussed previously for the case when β ) 0,28 the local maximum and minimum shown in Figure 5 over some range of κa and φrarise from the polarization effect of the double layer surrounding a particle, which induces a microscopic electric field opposing to the normal diffusio-phoretic motion. Moreover, for a positively charged particle, its behavior is much more complicated as it is determined by the competing concentration gradient and ionic diffusion under this circumstance,21especially when φr is high, as is seen in Figure 5. Besides, the polarization effect has to be considered as well when φr and κa are sufficiently large. Thus, a small Figure 6. Variation of the scaled diffusiophoretic mobility (U*/U0)
as a function of κa at various values of φrfor φr g 0, φ ) 0.1,
β ) -0.2, and R ) 1.
Figure 7. Contours of counterionic concentration n1/(a) and stream function (b) atφ ) 0.1, β ) -0.2, φr) 2.0, and R ) 1.
change in κa results in great variations in mass and electric potential flow. As a result, the direction of particle motion is altered more than once, as shown in Figure 6.
Streamline Plots. To shed more light on the shadow, we further examine both counterions’ distributions and the corre-sponding flow fields at various surface potential φr values, with κa fixed at 2.0; the simulated results are summarized in Figures 7 and 8. First, we observe that the distribution of counterions forms concentric spherical shells around a particle. With the increase of scaled surface potential φr, more counte-rions are attracted to the neighborhood of the particle surface. Checking further the concentration of counterions, we can determine whether a particle is negatively or positively charged. The concentration of anions decreases with the distance along the direction of concentration gradient, but that of cations increases. As explained earlier, this is because the diffusion velocity of anions is faster than that of cations, sinceβ < 0.
Due to the nonuniformity of the original overall bulk concentra-tion, ions have a tendency to diffuse from high to low concentration areas, and since anions move faster than cations do, more anions would cluster in the lower concentration area, whereas more cations would stay in the high concentration areas.
This induces a nonequivalent charge distribution and results in an additional electric force. If a particle is positively charged, this electric force is directed toward the area of lower ionic concentration and opposite to the direction of concentration gradient. The ultimate direction of diffusiophoresis is determined by net result of these two driving forces. Figure 7b shows that for a positively charged particle, a separate counterclockwise vortex flow arises around the particle. As mentioned before, the diffusion flows of cations and anions lead to an accumulation of charges of opposite sign in different areas. The constraint of absence of net electric current requires generation of an electric field which causes a charge flow in a direction to compensate the charge difference between two sides. The direction of this ionic charge flow is determined by the surface charge of a particle, exactly as in electroosmotic flow. Observing this electroosmosis flow closely, we find that this counterclockwise vortex flow appears in the vicinity of the particle surface, with an opposite flow direction to retard the motion of the particle otherwise. In Figure 8b, the particle moves downward to the lower concentration side, but the electroosmotic vortex flow is rotating along the opposite direction. The competition between particle motion otherwise and this electroosmotic flow becomes more and more obvious as φr increases. If the effect of the concentration gradient is greater than that of the electroosmotic flow resulting from the discrepancy of ionic diffusion velocities, the direction of particle movement will be the same as that of the concentration gradient; otherwise, they will be opposite to each other. As the surface potential or electrolyte concentration becomes higher, that is, larger φror κa values are exhibited, the competition of these two effects will continuously reconcile with each other and reach a new balance. The inside counter-clockwise vortex flow due to electroosmosis grows so strong that it eventually “swallows” the original outside flow due to the concentration gradient. That is why it is possible that the number of direction changes in diffusiophoresis can be more than one when β * 0. In contrast, if a particle is negatively
charged orβ ) 0, this phenomenon will not exist since the net
diffusion flow of cations and anions vanishes.
Effect of Volume Fraction. Finally, we examine the influ-ence of the particle concentration, measured by the volume fractionφ ) (a/b)3of the colloidal particle in the electrolyte solution. The results are shown in Figure 9. Double layer Figure 8. Contours of counterionic concentration n2/(a) and stream
function (b) atφ ) 0.1, β ) -0.2, φr) 2.5, and R ) 1.
Figure 9. Variation of the scaled diffusiophoretic mobility (U*/U0)
as a function of φrat various values ofφ for κa ) 1.0, β ) -0.2, and
R ) 1.
overlapping is allowed in the current analysis. It is found that the diffusiophoretic mobility decreases with increasing volume fraction. This is mainly due to the hindrance effect of neighbor-ing particles: the higher the concentration of colloidal particles, the more significant the overlapping of the neighboring double layers, leading to a greater hydrodynamic resistance for fluid flow in diffusiophoresis.
In summary, the diffusiophoresis of a concentrated colloidal dispersion is analyzed theoretically for the case when both the zeta potential and the thickness of a double layer can assume an arbitrary value, focused on the influence of distinct ionic diffusion velocities, measured byβ ) (D1- D2)/(D1+ D2), where D1and D2are respectively the diffusion coefficients of cations and anions. In contrast to the case whenβ ) 0, the
diffusiophoretic mobility exhibits specific oscillatory motion, and it is asymmetric with respect to the zeta potential whenβ
* 0. The movement of a particle is not always toward the higher bulk concentration of the electrolyte even when the sign of the zeta potential remains the same. The sign and the magnitude of
β are the key factors in determining the behavior of the
diffusiophoretic mobility. Furthermore, we find that the influ-ence of steric hindrance on the fluid flow must be taken into account as the colloidal dispersions become more and more concentrated.
Acknowledgment. This work is supported by the National Science Council of the Republic of China.
References and Notes
(1) Hunter, R. J. Foundations of Colloid Science; Clarendon Press: Oxford, 1989; Vols. I and II.
(2) Kosmulski, M.; Matijevic, E. J. Colloid Interface Sci. 1992, 150, 291.
(3) Deryagin, B. V.; Dukhin, S. S.; Korotkova, A. A. Kolloidn. Zh. 1961, 23, 409.
(4) Deryagin, B. V.; Dukhin, S. S.; Rulev, N. N. Colloid J. USSR 1978, 40, 531.
(5) Korotkova, A. A.; Deryagin, B. V. Kolloidn. Zh. 1991, 52, 861. (6) Dukhin, S. S.; Deryagin, B. V. Surface and Colloid Science; Wiley: New York, 1974; Vol. 7.
(7) Grushka, E.; Mccormick, R. M.; Kirkland, J. J. Anal. Chem. 1989, 61, 241.
(8) Knox, J. H.; Mccormick, R. M. Chromatographia 1994, 38, 207. (9) Ebel, J. P.; Anderson, J. L.; Prieve, D. C. Langmuir 1988, 4, 396. (10) Staffeld, P.; Quinn, J. J. Colloid Interface Sci. 1981, 130, 69. (11) Staffeld, P.; Quinn, J. J. Colloid Interface Sci. 1981, 130, 89. (12) Chun, B.; Ladd, A. J. C. J. Colloid Interface Sci. 2004, 274, 687. (13) Dukhin, S. S.; Malkin, E. S.; Dukhin, A. S. Colloid J. USSR 1978, 40, 536.
(14) Dukhin, S. S.; Malkin, E. S.; Dukhin, A. S. Colloid J. USSR 1979, 41, 734.
(15) Malkin, E. S.; Dukhin, A. S. Kolloidn. Zh. 1982, 44, 254. (16) Prieve, D. C.; Anderson, J. L.; Ebel, J. P.; Lowell, M. E. J. Fluid Mech. 1984, 148, 247.
(17) Prieve, D. C. AdV. Colloid Interface Sci. 1982, 16, 321. (18) Smith, R. E.; Prieve, D. C. Chem. Eng. Sci. 1982, 37, 1213. (19) Anderson, J. L.; Prieve, D. C. Langmuir 1991, 7, 403.
(20) Prieve, D. C.; Roman, R. J. Chem. Soc. Faraday Trans. II 1987, 83, 1287.
(21) Dukhin, S. S. AdV. Colloid Interface Sci. 1993, 44, 1.
(22) Pawar, Y.; Solomentsev, Y. E.; Anderson, J. L. J. Colloid Interface Sci. 1993, 155, 488.
(23) Baygents, J. C.; Saville, D. A. PCH Physicochem. Hydrodyn. 1988, 10, 543.
(24) Misra, S.; Varanasi, S.; Varanasi, P. P. Macromolecules 1990, 23, 4258.
(25) Keh, H. J.; Hsu, J. H. J. Colloid Interface Sci. 2000, 221, 210. (26) Wei, Y. K.; Keh, H. J. Langmuir 2001, 17, 1437.
(27) Wei, Y. K.; Keh, H. J. J. Colloid Interface Sci. 2002, 248, 76. (28) Lou, J.; Ho, Y. Y.; Lee, E. J. Colloid Interface Sci. 2006, 299, 443.
(29) Kuwabara, S. J. Phys. Soc. Jpn 1959, 14, 527.
(30) Lee, E.; Chu, J. W.; Hsu, J. P. J. Colloid Interface Sci. 1999, 209, 240.
(31) Delgado, A. V.; Gonzalez-Caballero, E.; Hunter, R. J.; Koopal, L. K.; Lyklema, J. Pure Appl. Chem. 2005, 77, 1753.
