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
Electrophoresis of a Rigid Sphere in a
Carreau Fluid Normal to a Large Charged Disk
Jyh-Ping Hsu, Li-Hsien Yeh, and Shu-Jen Yeh
J. Phys. Chem. B, 2007, 111 (43), 12351-12361 • DOI: 10.1021/jp0739317
Downloaded from http://pubs.acs.org on November 18, 2008
More About This Article
Additional resources and features associated with this article are available within the HTML version: • Supporting Information
• Access to high resolution figures
• Links to articles and content related to this article
Electrophoresis of a Rigid Sphere in a Carreau Fluid Normal to a Large Charged Disk
Jyh-Ping Hsu,* Li-Hsien Yeh, and Shu-Jen Yeh
Department of Chemical Engineering, National Taiwan UniVersity, Taipei, Taiwan 10617 ReceiVed: May 21, 2007; In Final Form: July 31, 2007
The electrophoresis of a rigid sphere in a Carreau fluid normal to a large disk is analyzed theoretically under the conditions of low surface potential and weak applied electric field. Previous analyses are extended to the case where a disk can be charged, and a more realistic electrostatic force formula is applied. We show that the qualitative behavior of a sphere depends largely on its distance from a disk, the thickness of double layer, and the nature of a fluid. In general, the presence of a disk has the effect of increasing the conventional hydrodynamic drag on a sphere, and a decrease in the thickness of the double layer surrounding a sphere has the effect of enhancing the shear-thinning effect. However, this might not be the case if a sphere is uncharged and a disk is charged, where the osmotic pressure field and the induced charge on the sphere surface can be significant. The shear-thinning effect is important only if the thickness of double layer is sufficiently thick. This result can play a significant role in practice such as in electrophoretic deposition, where the deposition electrode is charged and the fluid medium is usually of shearing-thinning nature.
1. Introduction
Since the report of the classic analytical result of electro-phoresis by Smoluchowski,1this electrokinetic phenomenon has been studied extensively in the past. Up to now, modifications of the original model so that it is capable of describing, more realistically, the conditions in practice are still one of the important issues in relevant fields. Electrophoresis has been applied widely in various areas serving both as an analytical tool to characterize the surface properties of an entity and as an operation to separate/differentiate entities of different natures. One of the key effects in conducting electrophoresis in practice is the boundary effect. The influence of a boundary can be significant, for instance, when electrophoresis is conducted in a relatively narrow space and/or when the behavior of a particle near a surface is of primary concern. In these cases, the governing electrokinetic equations need to be solved under appropriately defined boundary conditions. A considerable amount of theoretical efforts has been made in the past to simulate various types of boundary effect on electrophoresis. One of these is the electrophoresis of a particle normal to a boundary. Electrophoretic deposition (EPD), for example, involves this type of problem. Depending upon the geometry considered, previous results can roughly be divided into three categories: the electrophoresis of a particle normal to a plane,2-10to a disk,11,12and to a cavity.13-15In these cases, if a boundary is uncharged, its influence on the behavior of a particle is twofold: the increase in the viscous drag acting on a particle and the distortion of the local electric field near the particle. The situation becomes more complicated if a boundary is charged. In this case, several other factors need to be considered, for example, an electroosmotic flow field6,7,15-24 and an osmotic pressure field6,7,12,15-24will be established near the boundary, and charge will be induced on the particle surface as it approaches the boundary. These factors can have a profound influence on the electrophoretic behavior of a particle.6,7,12,15-24
Another factor which can be of practical significance is the nature of fluid medium. The introduction of surfactant and polymer into a colloidal dispersion to improve its stability,25 for example, can yield a non-Newtonian dispersion. These materials are also utilized often as additives to the electrolyte solutions employed in protein or DNA capillary zone electro-phoresis for minimizing protein or DNA-capillary wall interac-tions, for improving selectivity and resolution, and for control-ling the electroosmotic flow.26,27A dispersion can also be of non-Newtonian nature if the concentration of the dispersed phase becomes appreciable.28Several attempts were made recently to model the electrophoresis of a particle in a shear-thinning Carreau fluid. The geometry considered includes a sphere at the center of a spherical cavity,29a sphere at an arbitrary position in a spherical cavity,30a concentrated spherical dispersion,31,32 a sphere normal to a plane,33and a sphere along the axis of a cylindrical pore.34In general, the shear-thinning nature of a fluid is found to be advantageous to the movement of a particle, and the thinner the double layer surrounding a particle the more significant the effect of shear thinning.29-34
In this study, the influence of a charged boundary on the electrophoretic behavior of a particle in a Carreau fluid is investigated by considering the electrophoresis of a spherical particle normal to a large disk under the conditions of low surface potential and weak applied electric field. This extends * Corresponding author. Tel: 886-2-23637448. Fax: 886-2-23623040.
E-mail: jphsu@ntu.edu.tw.
Figure 1. Schematic representation of the problem considered where
a rigid sphere of radius a moves in a Carreau fluid normal to a large disk as a response to an applied uniform electric field E∞parallel to the z direction. The cylindrical coordinates (r,θ, z) with the origin at
the center of the disk are adopted; the disk lies on the plane z ) 0; and
h is the distance between the center of the sphere and the disk.
10.1021/jp0739317 CCC: $37.00 © 2007 American Chemical Society Published on Web 10/09/2007
previous analyses in that the fluid phase is non-Newtonian and that a boundary is charged. The latter is of practical importance in applications such as EPD.35,36 Also, a recently derived formula,37,38 which is more accurate than those used in the literature, for the evaluation of the electric force acting on a particle is adopted. The influences of the charged conditions of the sphere and the disk surfaces, the thickness of the double layer, and the nature of a fluid on the electrophoretic behavior of a sphere are discussed.
2. Theory
Referring to Figure 1, we consider the electrophoresis of a rigid, nonconductive, sphere of radius a in an incompressible Carreau fluid39,40normal to a large, conducting disk. Suppose that the disk is sufficiently large so that it can be treated as an infinite plane.4 The cylindrical coordinates (r,θ, z) with the
origin at the center of the disk are adopted, and the disk lies on the plane z ) 0. Since the present problem isθ-symmetric, only
the (r, z) domain needs to be considered; that is, it is of a two-dimensional nature. Let h be the distance between the center of the sphere and the disk. A uniform electric field E∞in the z direction is applied, and its strength is E∞. Both the surface of the sphere and that of the disk are maintained at constant potential, denoted asζpandζw, respectively.
The constitutive equation of a purely viscous non-Newtonian fluid can be expressed by
whereτ, η(γ˘), and γ3 are the stress tensor, the fluid viscosity, and the rate of strain tensor, respectively;γ3 )∇u + (∇u)T, and γ˘ is the magnitude of γ3; ∇, u, and T are the gradient operator, the fluid velocity, and the matrix transpose, respec-tively.40 For the Carreau fluid considered in this study, we choose39,40
whereη0andη∞are the viscosities corresponding to the zero-shear rate and the infinite-zero-shear rate, respectively; R is the
Figure 2. Typical mesh used in the resolution of the electric field (a) and that of the flow field (b) for a Carreau fluid with N ) 0.8 and RU/a ) 0.8. Other parameters used are ζp*( )ζp/(kBT/e)) ) 1,ζw*( )
ζw/(kBT/e)) ) 1, h/a ) 2, and κa ) 1.
Figure 3. Variation of the scaled mobilityµ*as a function of (h/a)
for various κa’s (a) and as a function of κa for various (h/a)’s (b) for the case of a positively charged sphere and an uncharged disk. Solid curves present results for a Newtonian fluid; dashed curves present results for a Carreau fluid with N ) 0.8 and RU/a ) 0.8; discrete symbols, numerical result based on Shugai and Carnie;7dashed-dotted
curves, analytical result predicted by Henry’s formula. Other parameters used areζp*) 1 and ζ
w*) 0.
τ ) -η(γ˘)γ3 (1)
η(γ˘) ) η∞+ (η0- η∞)[1 + (Rγ˘) β
](N-1)/β (2)
relaxation time constant, N is the power-law index, andβ is a
dimensionless parameter describing the transition region between the zero-shear-rate region and the power-law region. In practice,
β is close to 2 and η∞is small,39,40and therefore, eq 2 reduces to
Note that a Newtonian fluid can be recovered as a special case of the preset Carreau fluid by letting N ) 1 and/or R ) 0, and a power-law fluid can be obtained by assuming a large R.
We assume that the surface potential of a sphere and/or a disk is sufficiently low and E∞is relatively weak compared with that established by the sphere and/or the disk. The former is appropriate if the surface potential is lower than about 25 mV, and the latter is adequate if E∞is lower than about 25 kV/m, which is usually satisfied for conditions of practical signifi-cance.37Also, the system under consideration is at a pseudo-steady state;6,16-18that is, all of the dependent variables are time-independent. Under these conditions, it can be shown that the electric potential of the system,Ψ, can be described by41
where∇2is the Laplace operator,Ψ ) Ψ
1+ Ψ2, andΨ1and
Ψ2are the equilibrium potential in the absence of E∞and the
perturbed potential outside the sphere arising from E∞, respec-tively. κ ) [∑jnj
0 (ezj)2/k
BT]1/2is the reciprocal Debye length,
is the permittivity of the liquid phase, nj0and zjare the bulk number concentration and the valence of ionic species j, respectively, and e, kB, and T are the elementary charge, the Boltzmann constant, and the absolute temperature, respectively. We assume that both the sphere surface and the disk surface are maintained at constant potential, the sphere is nonconductive, and the disk is a perfect-conductor. Also, the equilibrium potential vanishes as z f∞and/or at a point far away from the sphere, and the perturbed electric field arising from E∞is normal to the disk. On the basis of these assumptions and the symmetric nature of the problem, the boundary conditions associated with eqs 4 and 5 can be expressed as
Here, n is the unit normal vector directed into the liquid phase, and ezis the unit vector in the z direction.
Figure 4. Contours of the scaled viscosityη*, viscN, for a Carreau fluid with N ) 0.8 and RU/a ) 0.8 for the case of Figure 3. (a) h/a ) 1.25 and κa ) 1, (b) h/a ) 2 and κa ) 0.5, (c) h/a ) 2 and κa ) 1, and (d) h/a ) 2 and κa ) 5.
η(γ˘) ) η0[1 + (Rγ˘) 2 ](N-1)/2 (3) ∇2 Ψ1) κ 2 Ψ1 (4) ∇2 Ψ2) 0 (5) Ψ1) ζp and n‚∇Ψ2) 0
on the sphere surface (6) Ψ1) ζw and Ψ2) constant at z ) 0 (7)
Ψ1) 0 and ∇Ψ2) -E∞ezas z f∞ (8)
n‚∇Ψ1) 0 and n‚∇Ψ2) 0
At steady state, the flow field of the present problem can be described by40
where p is the pressure. Fe ) ∑jzjenj 0
exp(-zjeΨ/kBT) is the space charge density, and Fe∇Ψ2is the electric body force acting on the fluid.6,7,15,24,37,38,42We assume that both the surface of a sphere and that of a disk are no-slip. Let U be the speed of the sphere in the z direction. Then the boundary conditions associated with the flow field under consideration can be expressed as
The mobility of a sphere can be determined from the fact that the total force acting on a sphere in the z direction vanishes at steady state;42that is,
where FEand FDare the z component of the electric force acting on a sphere at steady state and that of the hydrodynamic force
acting on a sphere at steady state, respectively. By letting S be the sphere surface, it can be shown that37,38
whereσp ) -n‚∇Ψ1 is the charge density on S, and Ez)
-∂Ψ2/∂z is the strength of the local external electric field in the z direction.
By letting t be the unit tangential vector on S, n be the magnitude of n, and tzand nzbe respectively the z component of t and that of n, FDcan be evaluated by37,43
For convenience, we let µ* ) U/U
refbe the scaled mobility where Uref) (kBT/e)E∞/η is the electrophoretic velocity of a sphere with a constant surface potential (kBT/e) when an electric field of strength E∞is applied.
The governing equations of the present problem and the associated boundary conditions are solved numerically by FlexPDE,44 a differential equation solver based on a finite element method. Throughout the computation, double precision is used and grid independence is checked to ensure that the mesh used is fine enough. In our case, a total of 19 890 and 1615 nodes are sufficient for the resolution of the electric field and the flow field, respectively, and a convergent result can be
Figure 5. Variation of the scaled mobilityµ*as a function of (h/a)
for various (RU/a)’s at N ) 0.8 (a) and that for various N’s at RU/a ) 0.2 (b). Other parameters used areζp/) 1, ζw*) 0, and κa ) 1.
Figure 6. Variation of the scaled mobilityµ*as a function of κa for
various (RU/a)’s at N ) 0.8 (a) and that for various N’s at RU/a ) 0.2 (b). Other parameters used areζp
/ ) 1, ζw*) 0, and h/a ) 2. FE)
∫ ∫
SσpEzdS (16) FD)∫ ∫
Sη ∂(u‚t) ∂n tzdS +∫ ∫
S-pnzdS (17) ∇‚u ) 0 (10) -∇‚τ - ∇p ) Fe∇Ψ2 (11)u ) Uez on the sphere surface (12)
u ) 0 at z ) 0 (13)
u ) 0 as z f∞or r f∞ (14)
FE+ FD) 0 (15)
obtained by setting an error limit of 10-6for the electric field and 10-3for the flow field. Figure 2 shows the typical mesh used. Note that an unstructured triangular mesh is used with a finer mesh near the particle and a coarser mesh away from it. The applicability and the accuracy of the software adopted were justified previously30,34and will also be discussed later. 3. Results and Discussion
The electrophoretic behavior of a sphere is examined through numerical simulation. Three representative cases are considered, namely, a positively charged sphere and an uncharged disk, an uncharged sphere and a positively charged disk, and a positively charged sphere and positively charged disk. Note that, because of the nonlinear nature of the present problem, the results of the last case cannot be obtained directly from a linear combina-tion of those of the first two cases.
3.1. Positively Charged Sphere and Uncharged Disk. Let us consider first the electrophoresis of a positively charged sphere normal to an uncharged disk. For a fixed sphere radius a, the influences of the scaled sphere-disk distance (h/a) and the thickness of double layer measured by κa on the scaled mobility of a sphereµ*is illustrated in Figure 3. For comparison, the corresponding results for the case of a Newtonian fluid and the result of Shugai and Carnie7for the case a sphere moving in a Newtonian fluid normal to an infinite plane are also
presented. Figure 3a reveals that the result based on the present numerical method for the case of a Newtonian fluid is consistent with that calculated by Shugai and Carnie,7which implies that the performance of the present numerical method is satisfactory. If κa is fixed, for both Newtonian and Carreau fluids, µ* decreases monotonically with the decrease in (h/a) andµ*
f 0 as h/a f 1, that is, as the sphere touches the disk. These are expected because the closer a sphere is to a disk, the more significant the hydrodynamic hindrance is on the movement of the sphere because of the presence of the disk. The same trend was observed by Shugai and Carnie.7Tang et al.,8Chih et al.,9 and Lee et al.,10 however, found that the closer the sphere-disk distance, the smaller the mobility of the sphere; but if a sphere is sufficiently close to a plane, its mobility may change the sign; that is, it moves in the opposite direction as that of the applied electric field. The latter arises if a problem is not of totally symmetric nature, such as the present one, an extraneous electric body force, Fe∇Ψ1, will be included in eq 16 and an extraneous electrostatic force arising from the equilibrium potential,∫∫S(∂Ψ1/∂n)(∂Ψ1/∂z) dS, is included in eq 17. These terms become significant when a sphere is close to a plane.15,37,38 Note that, because a nonslip boundary condition is assumed on the planar surface, the mobility of a sphere should vanish as it touches the plane. In general, the mobility of a sphere in a Carreau fluid is larger than that in the corresponding Newtonian fluid because of the shear-thinning nature of the former. We define33,34
whereµ*(Carreau) andµ*(Newtonian) are the scaled mobilities in the present Carreau fluid and in the corresponding Newtonian fluid, respectively. The larger the PD, the more important the effect of shear thinning of a Carreau fluid is. In Figure 3a, if
κa ) 1, PD are 6.44, 4.63, and 2.92% for h/a ) 1.25, 2, and 4, respectively; that is, the closer a sphere is to a disk, the more significant is the effect of shear thinning, as is also justified by Figure 4. Similar behavior was also observed by Lee et al.33 This can be explained by the presence of a disk having the effect of distorting the flow field below a sphere, implying that the spatial variation in the shear rate in the present case is greater than that in the case of an isolated sphere, and so is the spatial variation in the scaled viscosityη*) η/η
0, as can be seen in Figure 4. Note that if η*) 1, the fluid is Newtonian. For a fixed (h/a), Figure 3b indicates thatµ*increases monotonically with the increase in κa for both Newtonian and Carreau fluids, butµ*is more sensitive to the variation in κa in the latter. Also, for a fixed κa,µ*(Carreau) >µ*(Newtonian), and [µ*(Carreau)
- µ*(Newtonian)] increases with the increase in κa. Similar behavior was also observed in the electrophoresis of a rigid sphere at the center of a spherical cavity29and along the axis of a long cylindrical pore.34Note that the larger the κa, the thinner the double layer surrounding a sphere, the greater the absolute value of the gradient of the potential on its surface, the higher the surface charge density, and, therefore, the greater the electrical driving force acting on it. Also, because both the sphere surface and the disk are nonslip, a characteristic shear rate (U - 0)/κ-1can be defined,29which suggests that the larger the κa, the greater the shear rate and the more significant the shear-thinning effect of a fluid, as is justified in Figure 3. This explains why [µ*(Carreau) -µ*(Newtonian)] increases with the increase in κa. For h/a ) 2, PD are 2.05, 4.63, 17.39, and 44.21% for κa ) 0.1, 1, 5, and 10, respectively, which suggests that the thinner the double layer surrounding a sphere is, the
Figure 7. Variation of the scaled mobilityµ*as a function of (h/a)
for various κa’s (a) and as a function of κa for various (h/a)’s (b). Solid curves, present results for a Newtonian fluid; dashed curves, present results for a Carreau fluid with N ) 0.8 and RU/a ) 0.8; discrete symbols, numerical result based on Shugai and Carnie.7 Other
parameters used areζp
/ ) 0 and ζw*) 1. PD )|µ * (Carreau)| - |µ*(Newtonian)| |µ* (Newtonian)| × 100% (18)
more significant the effect of shear thinning is. This behavior is inconsistent with that reported by Lee et al.,33 where the behavior of the shear-thinning effect as κa varies is unclear. This is because the thicker the double layer surrounding a sphere, the more significant the presence of the disk, and the more serious the deviation arising from the consideration of the extraneous terms, Fe∇Ψ1and∫∫S(∂Ψ1/∂n)(∂Ψ1/∂z)dS, in the calculation of the mobility. According to Figure 4, the viscosity of the fluid varies most significantly in the gap between the sphere and the disk, and the minimum viscosity occurs on the sphere surface. This minimum declines with the decrease in (h/ a) and with the increase in κa. The latter was also observed in a study of the electrophoresis of a concentrated dispersion of spheres31and in the electrophoresis of a sphere along the axis of a cylindrical pore.34
Figures 5 and 6 illustrate the influences of the nature of a Carreau fluid on the scaled mobility of a sphere µ*. The qualitative behaviors ofµ*as the scaled sphere-disk distance (h/a) or the double layer thickness κa varies are similar to those for the case of a Newtonian fluid, also, the larger the deviation of N from unity and/or the deviation of RU/a from 0, the more significant the effect of shear thinning, and therefore, the larger theµ*, which is consistent with the relevant theoretical results in the literature.29-34Figure 6 shows that, if κa is small,µ*is more sensitive to the variation in RU/a than to the variation in
N. The same trend was also observed in the electrophoresis of a sphere at the center of a spherical cavity.29
3.2. Uncharged Sphere and Positively Charged Disk. The influences of the scaled sphere-disk distance (h/a) and the thickness of double layer κa on the scaled mobility of a sphere
µ*for the case in which the sphere is uncharged and the disk positively charged are presented in Figure 7. For comparison, the corresponding results for the case of a Newtonian fluid and the numerical results of Shugai and Carnie7 are also shown. The present results agree well with those of Shugai and Carnie.7 In the present case, negative charge is induced on the surface of a sphere, and it experiences an electric force in the -z direction, and an osmotic pressure field is established.12 As pointed out by Shugai and Carnie7 and Hsu et al.,12 for an isolated, charged disk, the spatial variation of mobile ions is Fe,iso(z) ) -κ2ζ
w exp(-κz), and an equilibrium osmotic or hydrostatic pressure field, piso(z) ) κζwE∞ exp(-κz), is established to balance the electric body force. If bothζw and
E∞ are positive, -dpiso/dz ) -Fe,iso(z)E∞ > 0, and both the sphere and the fluid feel a pressure in the z direction when the sphere disturbs the equilibrium pressure field. If we neglect the influence on the equilibrium system due to the presence of the sphere, the z component of the hydrostatic pressure acting on the sphere is Fp) 4πaζwE∞exp(-κh)[cos h(κa) - sin h(κa)/
κa], which is the last term of eq 12 of Shugai and Carnie.7 Figure 8. Contours of the scaled viscosityη*, viscN, for a Carreau fluid with N ) 0.8 and RU/a ) 0.8 for the case of Figure 7. (a) h/a ) 1.25 and κa ) 1, (b) h/a ) 2 and κa ) 0.5, (c) h/a ) 2 and κa ) 1, and (d) h/a ) 2 and κa ) 5.
Therefore, although the body force acting on the fluid is in the -z direction, its movement is influenced appreciably by a positive pressure drop, which is in the z direction. In the case of an uncharged sphere and a charged disk, the electrostatic force changes appreciably as the former approaches the latter because of a fast increase in the negative charge induced on the sphere surface. On the other hand, the hydrostatic pressure acting on a sphere increases with the decrease in h/a because both the sphere and the fluid are influenced by an appreciable increase in the positive pressure drop as stated previously. The competition of these two forces affect appreciably the behavior of a sphere and makes its electrophoretic behavior more
complicated than that in the case of a charged sphere and an uncharged disk. As can be seen in Figure 7a, for both Newtonian and Carreau fluids, |µ*| may have a local maximum as h/a varies, which was observed but not discussed in Shugai and Carnie,7where the fluid is Newtonian. The presence of the local maximum is the net results of the competition of the forces including the induced electric force in the -z direction, the osmotic pressure in the z direction, and the drag as h/a varies. If κa ) 1, a negative electric force dominates, and an increase in its magnitude leads to a more negativeµ*as h/a decreases from infinite. A further decrease in h/a causes a large increase in the drag, yielding a negative maximum ofµ*at h/a = 1.5. Figure 9. Variation of the scaled mobilityµ*as a function of (h/a) for various κa’s (a) and as a function of κa for various (h/a)’s (b) for the case whenζp/) 1 and ζw*) 1. Red curves, results of Figure 3b; solid curves, present results for a Newtonian fluid; dashed curves, present results for a Carreau fluid with N ) 0.8 and RU/a ) 0.8.
Similar reasoning is applicable to the curve corresponds to κa ) 3 in Figure 7a, except that a positive hydrostatic pressure due to the presence of a positive charged disk now dominates, leading to a positive maximum inµ*at h/a = 1.7. Therefore, the existence of a local maximum in|µ*| seen in Figure 7a is expected becauseµ*
f 0 as a sphere touches a disk (h/a f 1), as is required by the non-slip condition assumed on the disk
surface, or it is far away from the disk surface (h/a f∞). Figure 7b indicates that, for both Newtonian and Carreau fluids,µ* may change its sign from negative to positive as κa increases, and it may have a local maximum; the later was not observed in Shugai and Carnie.7These phenomena arise from a sudden raise and fall in Fpfor κ(h - a) ) κa(h/a - 1) in the range [2, 4] and are the net result of the competition between the induced electric force in the -z direction and the osmotic pressure in the z direction mentioned previously. First, because the amount of negative charge induced on the sphere surface increases with the decrease in κa, the corresponding electric force acting on the sphere increases accordingly. Second, for a fixed h/a, as κa increases, the pressure acting on a sphere increases first and then decreases. The pressure vanishes as κa f∞. In short, the influence of the presence of a charged disk is twofold: it induces charge on the sphere surface, and it establishes a pressure field. The later is responsible for the presence of a local maximum in Fpas κ(h - a) varies in the range [2, 4]. The former yields an electric force, which declines monotonically with the increase of κ(h - a) in the range [2, 4]. Also, the electric force is much smaller than Fp. As is seen in Figure 7b, the net result is that
µ*has a local maximum and/or varies drastically with κa. The
|µ*| for a Carreau fluid is larger than that for the corresponding Newtonian fluid, which is consistent with the electroosmotic mobility of polymer solutions in a charged fused silica capil-lary45and can be explained by the shear-thinning effect of a Carreau fluid. In Figure 7, if κa ) 1, PD are 2.77, 1.25, and
Figure 10. Variation of∆ ){[µ*(Carreau)-µ*(Newtonian)] sphere and disk positively charged- [µ*(Carreau)-µ*(Newtonian)]sphere positively charged, disk uncharged}for the cases of Figures 3b and 9b.
Figure 11. Contours of the scaled viscosityη*, viscN, for a Carreau fluid with N ) 0.8 and RU/a ) 0.8 for the case of Figure 9. (a) h/a ) 1.25 and κa ) 1, (b) h/a ) 2 and κa ) 0.5, (c) h/a ) 2 and κa ) 1, and (d) h/a ) 2 and κa ) 5.
0.29% for h/a ) 1.1, 1.25, and 2, respectively, and if h/a ) 1.25, PD are 4.78, 2.54, 1.25, and 0.81% for κa ) 0.1, 0.5, 1, and 3, respectively. That is, the effect of shear thinning is significant if the boundary effect is important and/or the double layer is sufficient thick. The former is consistent with the results shown in Figures 3 and 4, where a sphere is positively charged and a disk is uncharged, and can be explained by the same reasoning. The later is inconsistent with the results shown in Figures 3 and 4 and has not been reported previously. The effect of shear thinning can also be justified by examining the contours ofη*shown in Figure 8, where the fluid viscosity near a sphere is small when κa is small. This is because if the double layer is thick, the amount of negative charge induced on the sphere surface is large, so is the electric force acting on the sphere. Also, the pressure is small at the same time, yielding a large |µ*| and a large variation in the velocity gradient. Figure 8b,c indicates that if κa is not too large, the minimal viscosity occurs on the sphere surface, which is consistent with the results reported in the literature.31,33,34However, if κa is sufficiently large, the minimal viscosity becomes large and occurs on the disk surface, as illustrated in Figure 8d. This is because if the double layer of a disk is sufficiently thin, the negative electric force and the positive pressure acting on a sphere is unimportant, leading to a small µ* and a small variation in the velocity gradient near the sphere. Since only a limited electroosmotic flow is generated in the double layer of the disk, the minimal viscosity occurs on the disk surface.
3.3. Positively Charged Sphere and Positively Charged Disk. Let us consider next the case in which both a sphere and a disk are positively charged. The influences of the scaled separation distance between a sphere and a disk (h/a) and the thickness of double layer κa on the scaled mobility of a sphere
µ*are illustrated in Figure 9a,b, respectively. For comparison, the results for the corresponding Newtonian fluid are also presented in this figure. For the present case, µ* is positive because the net electric force acting on a sphere is always positive. Figure 9a reveals thatµ*declines monotonically with the decrease in h/a, implying that, if the surface potential of a sphere is the same as that of a disk, the mobility of the former is dominated by the viscous effect arising from the nonslip condition on the latter, as is in the case in which a sphere is positively charged and a disk is uncharged. However, the influence of the disk onµ*is significant only for κ(h - a) < 3, where double layer distortion is important. As can be seen in Figure 9b, for both Newtonian and Carreau fluids,µ*increases monotonically with the increase in κa. However, µ* is more sensitive to the variation of κa for a Carreau fluid than that for the corresponding Newtonian fluid. For fixed κa and h/a,µ* -(Carreau) >µ*(Newtonian), and [µ*(Carreau) -µ*(Newtonian)] increases with the increase in κa, which is consistent with the result of Lee et al.29for the case of a charged sphere at the center of an uncharged spherical cavity. Let∆ ){[µ*(Carreau)
- µ*(Newtonian)]
sphere and disk positively charged- [µ*(Carreau) -µ* -(Newtonian)]sphere positively charged, disk uncharged}. Then, according to
Figure 12. Variation of the scaled mobilityµ*as a function of κa for
various (RU/a)’s at n ) 0.8 (a) and that for various N’s at RU/a ) 0.2 (b). Other parameters used areζp
/
) 0, ζw*) 1, and h/a ) 1.25.
Figure 13. Variation of the scaled mobilityµ*as a function of (h/a)
for various (RU/a)’s at N ) 0.8 (a) and that for various N’s at RU/a ) 0.2 (b). Other parameters used areζp
/
Figure 10, if κa is small,∆ < 0. Also, ∆ has a local maximum as κa varies for κ(h - a) in the range [2, 4] and vanishes when
κa is sufficiently large. These phenomena have not been reported in the literature and can be explained by comparing the behavior of µ* as κa varies in a Carreau fluid with that in the corresponding Newtonian fluid or by examining the contours ofη*shown in Figure 11. Figure 9b also suggests thatµ* is more sensitive to the variation in κa when both a sphere and a disk are positively charged than when a sphere is positively charged and a disk is uncharged. We conclude that the presence of a charged disk can be advantageous to the movement of an identically charged sphere in that its mobility can have a local maximum and is larger than that for the case when a disk is uncharged. However, it becomes disadvantageous if κa is sufficiently small, that is,µ*(charged disk) <µ*(uncharged disk). Figure 9b suggestsµ*
f 0 as κa f 0. A similar result was also obtained in previous studies16,19,20and was proven by Zydney16 for the case of a charged sphere in a charged spherical cavity. The phenomenon are that (i)µ*is more sensitive to the variation in κa when both a sphere and a disk are positively charged than when a sphere is positively charged and a disk is uncharged, (ii) the presence of a charged disk can be advantageous to the movement of an identically charged sphere in that its mobility can have a local maximum and is larger than that for the case when a disk is uncharged, and (iii)µ*
f 0 as κa f 0 arise from the behaviors of the driving forces acting on a sphere as
κ(h - a) varies. For a fixed h/a, the electric force acting on a sphere is positive (so isµ*) and increases with the increase in
κa. Also, the hydrostatic pressure acting on a sphere arising from the osmotic pressure,12which is in the same direction as that of the electric force, increases first and then decreases as
κa increases for κ(h - a) in the range [2, 4]. In the limit κa f
∞, a sphere feels no pressure. If κa is small, the amount of negative charge induced on the sphere surface due to the presence of the positively charged disk, which has the effect of decreasing the net electric force acting on the sphere, is large. Also, the pressure hydrostatic force is small at the same time. Therefore,µ*(both sphere and disk positively charged) < µ* -(sphere positively charged, disk uncharged). However, the situation becomes reversed if κa is large, that is,µ*(both sphere and disk positively charged) > µ*(sphere positively charged, disk uncharged). Since the larger the κa, the greater the net driving force (electric+pressure) acting on a sphere, the larger theµ*, the greater the shear rate, and the more significant the effect of shear thinning, we conclude that, if κa is large, the effect of shear thinning for the case when both a sphere and a disk are positively charged is more important than that for the case when a sphere is positively charged and a disk is uncharged, and the reverse is true if κa is small. Note that a further increase in κa for κ(h - a) in the range [2, 4] leads to a decline in the hydrostatic pressure, which vanishes as κa f∞. That is, for both Newtonian and Carreau fluids, µ*(both sphere and disk positively charged) is close toµ*(sphere positively charged, disk uncharged), and they become the same when κa is sufficiently large. This explains that, why ∆ has a local maximum as κa increases for κ(h - a) in the range [2, 4], it vanishes when κa is sufficiently large. It is interesting to note that the value of κa at which∆ becomes positive increases with the decrease in h/a, as shown in Figure 10. This is because, if κa is small, the smaller the h/a, the more significant the influence of the negative charge induced on the sphere surface. Figure 11 reveals that the minimum viscosity of the system under consideration occurs on the sphere surface, and this minimum decreases with the decrease in h/a and/or the increase in κa. Similar behavior is also observed in Figure 4. A comparison between Figure 4b-d and Figure 11b-d reveals that, as mentioned previously, if κa is large, the effect of shear thinning in which both a sphere and a disk are positively charged is more important than that in which a sphere is positively charged and a disk uncharged, but the reverse is true if κa is small.
Figures 12 through 14 illustrate the influence of the nature of a Carreau fluid on the scaled mobility of a sphereµ*for the case in which a disk is positively charged. In general, the larger the RU/a and/or the smaller the N, that is, the more significant the effect of shear thinning, the larger the |µ*|. These are consistent with those of the previous studies where a boundary is uncharged.29-34It is interesting to note that, if κa is small,
µ*is more sensitive to the variation in RU/a than that in N. This phenomenon is similar to the case in which a sphere is at the center of a spherical cavity,29even if the cavity is uncharged. 4. Conclusions
The boundary effect on the electrophoretic behavior of a particle in a non-Newtonian fluid is studied by considering the electrophoresis of a sphere in a Carreau fluid normal to a large disk under the conditions of low surface potential and weak applied electric field. We show that the mobility of a sphere is influenced by three key factors: the separation distance between the sphere and the disk, the thickness of double layer, and the nature of a Carreau fluid. Also, the shear-thinning effect of a Carreau fluid correlates positively with the boundary effect. If a sphere is positively charged and a disk is uncharged, the
Figure 14. Variation of the scaled mobilityµ*as a function of κa for
various (RU/a)’s at N ) 0.8 (a) and that for various N’s at RU/a ) 0.2 (b). Other parameters used areζp
/
) 1, ζw*) 1, and h/a ) 2.
presence of the latter has the effect of retarding the movement of the former. Also, the shear-thinning effect becomes significant if a double layer is thin. On the other hand, although the presence of a charged disk does not cause an electroosmotic flow parallel to the applied field, it influences appreciably the movement of a sphere through inducing charge on its surface and establishing an osmotic pressure field. If a sphere is uncharged and a disk is positively charged, the charge induced on the sphere surface and the osmotic pressure field established by the disk can yield a local maximum in the mobility of the sphere or even change its sign. Also, the shear-thinning effect is pronounced if the double layer is thick, which has not been reported in the literature. If both a sphere and a disk are positively charged, the mobility of the sphere is always positive. For a fixed separation distance between a sphere and a disk, if the double layer is thick, the mobility in the present case is larger than that in the case in which a sphere is positively charged and a disk is uncharged, but the reverse is true if the double layer is thin. The mobility of a sphere is more sensitive to the variation in the thickness of the double layer in the case of a Carreau fluid than in the case of the corresponding Newtonian fluid. If the double layer is thick, the shear-thinning effect in the present case is less important than that in the case in which a sphere is positively charged and a disk is uncharged.
Acknowledgment. This work was supported by the National Science Council of the Republic of China.
References and Notes
(1) von Smoluchowski, M. Z. Phys. Chem. 1918, 92, 129.
(2) Morrison, F. A.; Stukel, J. J. J. Colloid Interface Sci. 1970, 33, 88. (3) Keh, H. J.; Anderson, J. L. J. Fluid Mech. 1985, 153, 417. (4) Keh, H. J.; Lien, L. C. J. Chin. Inst. Chem. Eng. 1989, 20, 283. (5) Feng, J. J.; Wu, W. Y. J. Fluid Mech. 1994, 264, 41.
(6) Ennis, J.; Anderson, J. L. J. Colloid Interface Sci. 1997, 185, 497. (7) Shugai, A. A.; Carnie, S. L. J. Colloid Interface Sci. 1999, 213, 298.
(8) Tang, Y. P.; Chih, M. H.; Lee, E.; Hsu, J. P. J. Colloid Interface
Sci. 2001, 242, 121.
(9) Chih, M. H.; Lee, E.; Hsu, J. P. J. Colloid Interface Sci. 2002,
248, 383.
(10) Lee, E.; Lou, S. H.; Hsu, J. P. Chem. Eng. Sci. 2006, 61, 4550. (11) Keh, H. J.; Lien, L. C. J. Fluid Mech. 1991, 224, 305. (12) Hsu, J. P.; Ku, M. H.; Kuo, C. C. Langmuir 2005, 21, 7588.
(13) Hsu, J. P.; Hung, S. H.; Kao, C. Y. Langmuir 2002, 18, 8897. (14) Yu, H. Y.; Hung, S. H.; Hsu, J. P. Colloid Polym. Sci. 2004, 283,
10.
(15) Hsu, J. P.; Yeh, L. H.; Chen, Z. S. J. Colloid Interface Sci. 2007,
310, 281.
(16) Zydney, A. L. J. Colloid Interface Sci. 1995, 169, 476. (17) Lee, E.; Chu, J. W.; Hsu, J. P. J. Colloid Interface Sci. 1997, 196, 316.
(18) Lee, E.; Chu, J. W.; Hsu, J. P. J. Colloid Interface Sci. 1998, 205, 65.
(19) Liu, H.; Bau, H. H.; Hu, H. H. Langmuir 2002, 20, 2628. (20) Hsu, J. P.; Ku, M. H.; Kao, C. Y. J. Colloid Interface Sci. 2004,
276, 248.
(21) Hsu, J. P.; Ku, M. H. J. Colloid Interface Sci. 2005, 283, 592. (22) Hsu, J. P.; Kuo, C. C.; Ku, M. H. Electrophoresis 2006, 27, 3155. (23) Hsu, J. P.; Kuo, C. C. J. Phys. Chem. B 2006, 110, 17607. (24) Hsu, J. P.; Yeh, L. H. J. Phys. Chem. B 2007, 111, 2579. (25) Hunter, R. J. Foundations of Colloid Science; Oxford University Press: Oxford, 1989; Vol. 1.
(26) Corradini, D. J. Chromatogr. B 1997, 699, 221. (27) Heller, C. Electrophoresis 2001, 22, 629.
(28) Hunter, R. J. Foundations of Colloid Science; Oxford University Press: Oxford, 1989; Vol. 2.
(29) Lee, E.; Huang, Y. F.; Hsu, J. P. J. Colloid Interface Sci. 2003,
258, 283.
(30) Hsu, J. P.; Hung, S. H.; Yu, H. Y. J. Colloid Interface Sci. 2004,
280, 256.
(31) Hsu, J. P.; Lee, E.; Huang, Y. F. Langmuir 2004, 20, 2149. (32) Lee, E.; Tai, C. S.; Hsu, J. P.; Chen, C. J. Langmuir 2004, 20, 7952.
(33) Lee, E.; Chen, C. T.; Hsu, J. P. J. Colloid Interface Sci. 2005, 285, 857.
(34) Hsu, J. P.; Yeh, L. H.; Ku, M. H. Colloid Polym. Sci. 2006, 299, 886.
(35) Popa, A. M.; Vleugels, J.; Vermant, J.; van der Biest, O. J. Eur.
Ceram. Soc. 2006, 26, 933.
(36) Besra, L.; Liu, M. Prog. Mater. Sci. 2007, 52, 1.
(37) Hsu, J. P.; Yeh, L. H.; Ku, M. H. J. Colloid Interface Sci. 2007,
305, 324.
(38) Hsu, J. P.; Yeh, L. H. J. Chin. Inst. Chem. Eng. 2006, 37, 601. (39) Yasuda, K.; Armstrong, R. C.; Cohen, R. E. Rheol. Acta 1981, 20, 163.
(40) Bird, R. B.; Armstrong, R. C.; Hassager, O. Dynamics of Polymer
Liquids; Wiley: New York, 1987; Vol. 1.
(41) Henry, D. C. Proc. R. Soc. London Ser. A 1931, 133, 106. (42) Hsu, J. P.; Kao, C. Y. J. Phys. Chem. B 2002, 106, 10605. (43) Backstrom, G. Fluid Dynamics by Finite Element Analysis; Studentlitteratur: Sweden, 1999.
(44) FlexPDE version 2.22, PDE Solutions Inc., U.S.A.
(45) Bello, M. S.; de Besi, P.; Rezzonico, R.; Righetti, P. G.; Casiraghi, E. Electrophoresis 1994, 15, 623.
