668 Electrophoresis 2013, 34, 668–673

Sourav Mondal
Sirshendu De
Research Article
Department of Chemical Mass transport in a porous microchannel
Engineering, Indian Institute of
Technology Kharagpur, for non-Newtonian fluid with electrokinetic
Kharagpur, India
effects
Received September 16, 2012 Quantification of mass transfer in porous microchannel is of paramount importance
Revised October 30, 2012
in several applications. Transport of neutral solute in presence of convective-diffusive
Accepted October 30, 2012
EOF having non-Newtonian rheology, in a porous microchannel was presented in this
article. The governing mass transfer equation coupled with velocity field was solved along
with associated boundary conditions using a similarity solution method. An analytical
solution of mass transfer coefficient and hence, Sherwood number were derived from
first principles. The corresponding effects of assisting and opposing pressure-driven flow
and EOF were also analyzed. The influence of wall permeation, double-layer thickness,
rheology, etc. on the mass transfer was also investigated. Permeation at the wall enhanced
the mass transfer coefficient more than five times compared to impervious conduit in
case of pressure-driven flow assisting the EOF at higher values of ␬h. Shear thinning fluid
exhibited more enhancement of Sherwood number in presence of permeation compared
to shear thickening one. The phenomenon of stagnation was observed at a particular
␬h (∼2.5) in case of EOF opposing the pressure-driven flow. This study provided a direct
quantification of transport of a neutral solute in case of transdermal drug delivery, transport
of drugs from blood to target region, etc.

Keywords:
Debye layer / EOF / Microchannel / Non-Newtonian fluid / Sherwood number
DOI 10.1002/elps.201200552

1 Introduction volving both fluid flow and mass transfer in a microchannel

with porous wall or membrane. Permeation of desired so-
Miniaturized devices in the dimensional scale of microm- lutes through the porous barrier (skin) is controlled by po-
eter have found wide spread applications recently, for ex- larization of larger neutral solutes over the membrane wall.
ample, microelectronics, diagnostics, and other bio-medical Relevant transport of species is entirely dictated by the mass
applications, different “Lab-on-Chip” systems, etc. [1–3]. The transfer boundary layer developed by the rejected neutral so-
striking features of such systems include easier flow control, lutes. Estimation of mass transfer coefficient is critical in such
improved heat, and mass transfer [4–6]. systems. Moreover, in actual physiological systems, the asso-
Flow occurs in microchannel or microtubes due to pres- ciated fluid has a predominantly non-Newtonian behavior.
sure gradient or electroosmotic effects or combination of Various studies involving power law and viscoelastic fluid
both. Plethora of information is available in literature to quan- models are available to characterize the flow behavior and
tify the velocity field and the coupled heat transfer problem in heat transfer to quantify the Nusselt number [17–19] in im-
such systems for Newtonian fluid rheology [7, 8]. These stud- pervious conduits.
ies have direct implication on various heat transfer devices The corresponding mass transfer problem in porous mi-
in microscale applications, for example, microchip cooling crochannel and microtubes was analyzed by Vennela et al.
[9], microscale heat exchanger [10, 11], etc. The correspond- [20, 21] for Newtonian fluids. They developed analytical solu-
ing mass transfer problems have applications in transdermal tion of Sherwood number in such geometry for Newtonian
drug delivery [12, 13], electrolyte transport in fuel cells [14], rheology. Mass transfer in a microchannel bioreactor partially
transport in hydrogel [15], electrokinetic separation [16], etc. filled with porous medium was studied numerically by Chen
These applications include study of transport processes in- et al. [22, 23]. However, they dealt with Newtonian fluid and
pressure-driven flow only. Quantification of mass transfer co-
efficient (or Sherwood number) in a porous microchannel for
non-Newtonian rheology, which is more relevant in physio-
Correspondence: Professor Sirshendu De, Department of Chem- logical system is not available. The transport of fluid in the
ical Engineering, Indian Institute of Technology, Kharagpur,
tissues and capillaries, transport of enzymes through blood
Kharagpur 721302, India
E-mail: sde@che.iitkgp.ernet.in
Fax: +91-3222-255303 Colour Online: See the article online to view Figs. 1, 4 and 5 in colour.


C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com
Electrophoresis 2013, 34, 668–673 Liquid Phase Separations 669

Figure 1. Schematic of a channel showing the (A) relative effects of electroosmotic phenomenon and pressure gradient and (B) various
concentration profiles and boundary layer.

brain barrier for Alzheimer’s patient involves knowledge of where, ␬ is the inverse of Debye layer thickness, ␬ =
the mass transfer of such physiological fluids [24–26]. This 2 2 1
( 2n␧k∞BzTe ) /2 , ␧ is the dielectric constant of the medium, E x
work presents an analytical expression for Sherwood number
is the electric field strength, and ␨ is the zeta potential. In-
of a neutral solute in a rectangular porous microchannel for
tegrating Eq. (3), using the boundary condition, du dy
x
= 0 at
a power law fluid under combined EOF and pressure-driven
y = h, the following expression in nondimensional form is
flow. The domain of validity of Sherwood number has also
obtained:
been identified.
n 
du∗x 1
= n Rn (n + 1)n (1 − y ∗ )
d y∗ n
2 Theory

(␬h) ∗
The 2D EOF, in the porous slit microchannel containing a ± sinh{␬h(1 − y )} , (4)
cosh(kh)
1:1 symmetric electrolyte solution with constant dielectric is
shown in Fig. 1. The inner surfaces of the channel are charged u
where, y ∗ = hy ; u∗x = uuHx ; R = pumax ; u pmax is the maxi-
at the wall potential, ␨. The fluid in the channel is driven by H
mum average cross-sectional velocity in the channel due to
the combined effects of a uniform external
 electric
 field of pressure-driven flow, which is u pmax = ( n+1 n n+1
)( pmx )1/n h n and
strength Ex , and a pressure gradient p x = − dd Px . It may be 1 − n/
noted that the coordinate system is fixed at the lower surface u H = n␬ n (− ␧Emx ␨ )1/n . The expression of uH reduces to
of the channel. Helmholtz–Smolchowski velocity for Newtonian fluid.
The x-component velocity (ux ) profile of a non-Newtonian It must be emphasized that the portion of the velocity
fluid, in a rectangular microchannel (0 < y < 2h) for pressure- profile within the mass transfer boundary layer is our region
driven flow of power-law fluid is given as [27]: of interest. Typical values of diffusivity of neutral solute to
 be considered are in the order of 10−11 to 10−15 m2 /s, which
 p 1/  n  n+1 n+1

n y n leads to Schmidt number in the order of 107 to 1010 , for

ux (y) =
x
h n
1 − 1 − (1)
m n+1 h non-Newtonian fluids. Since, the thickness of mass transfer
boundary layer is inversely proportional to Schmidt number,
The fluid rheology is defined as [27]: we can assume that in the microchannel, the mass transfer
boundary lies within 0.1% of the channel half height. This
 n
␶ = −m du x
for 0 ≤ y ≤ h assumption is more consolidated in the subsequent sections.

dy
n (2) Regarding the magnitude of R, it may be noted that R → 0
and ␶ = m − d y
du x
for h ≤ y ≤ 2h indicates purely EOF and 1/R → 0 implies a purely pressure-
driven flow.
where, ␶ is the shear stress, m is the consistency index, and n
is the power law exponent. Combing the effects of pressure-
driven and EOF, the governing equation of fully developed,
1D axial velocity profile (for 0 ≤ y ≤ h) under the Debye– 2.1 Approximated velocity profile for all values of ␬h
Huckel approximation can be expressed as [4]:
     As described above, for y*  1, Eq. (4) can be modified as:
d dux n px ␧E x ␨ ␬2
= ± −
dy dy m m cosh (␬h ) 1 n 1/
u∗x = R (n + 1)n ± (␬h )n tanh (␬h ) n y ∗ . (5)
cosh ␬ (h − y ) , (3) n


C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com
670 S. Mondal and S. De Electrophoresis 2013, 34, 668–673

double-layer thickness), the flow direction takes place in the

direction of pressure gradient. Similarly, as the double-layer
thickness decreases, by increasing solute concentration, the
electroosmotic effect becomes prominent and flow occurs in
the direction of the applied external electric field.

2.2 Solute transport in the slit microchannel

Mass transport of an electrically neutral solute in a slit mi-

crochannel having semipermeable wall is analyzed here. The
fluid rheology is considered to be non-Newtonian in nature
and is represented by power law model, as most of the phys-
iological fluid systems are generalized by the power law be-
havior [28]. Complete retention and insignificant electrical
or colloidal interaction of the solute with the channel or elec-
trolyte is assumed. The electric field gradient across the small
solutes is negligible to give rise to any electrokinetic forces
on the solutes. Under these conditions, the species balance
equation within the mass transfer boundary layer at steady
Figure 2. Comparison of the exact and approximate velocity state is:
profiles for different combinations of parameters and rheology
−
→ −
→2
(A) n = 0.5 and (B) n = 3.0. u ∇ c = D ∇ c. (6)

Since, the mass  transfer boundary layer is thin, y-

component velocity v y is approximately equal to permeation
velocity at the wall (vw ), i.e. v y ≈ −vw (x) [29]. Nondimen-
sional form of the above equation using the approximated
velocity profile from Eq. (5), is:
   
1 de 1 2n + 1 ∂c ∗ Pe w ∂c ∗
ReSc A (kh, R, n) y ∗ ∗ −
16 L R n+1 ∂x 4 ∂ y∗
∂ 2c ∗
= , (7)
∂ y2
where, the dimensionless quantities are defined as, di-
Figure 3. Phase space plot showing the direction of fluid velocity mensionless solute concentration, c ∗ = cc0 ; A (kh, n, R) =
in the channel. 1
n
|Rn (n + 1)n ± (␬h )n tanh (␬h )|1/n ; Reynolds number, Re =
␳u pmax de ␮e f f
␮e f f
; Schmidt number, Sc = ␳D ; and nondimensional
Comparison of the exact Eq. (4) and approximated veloc- permeation velocity, Pe w = vwDde . The term de is the equiv-
ity profile Eq. (5) for the range of parameters, within 0.1% of alent diameter of the channel (equals to 4h). The relevant
the channel half height is depicted in Fig. 2. It clearly shows boundary conditions for the above equation are:
that the velocity profile is linear and the approximate profile
does not deviate more than 2% of the exact profile. Eq. (5) at x = 0; c = c 0 (8a)
represents a linear form of the axial velocity with y* within
∂c
the mass transfer boundary layer. at y = 0; vw c + D =0 (8b)
In the case of opposing effects of EOF and pressure- ∂y
driven flow, a feasible range of the parameters R, n, and
at y → ∞; c = c 0 (8c)
␬h which suggests change in direction of flow is shown in
Fig. 3. The reversing of flow direction is accounted by the
negative sign in Eq. (5). The set of parameters, correspond- The above set of equations is solved for the lower sur-
ing to the region above or left of the curve simulates flow in face of the microchannel. Considering an order of magni-
the direction of the electroosmotic effect. Similarly, in case of tude analysis, at the edge of the mass transfer boundary layer,
the region toward the right or bottom of the curve signifies Eq. (7) leads to an expression of thickness of nondimensional
∗
flow in the direction of the pressure gradient. This can be mass transfer boundary layer, ␦∗ = ( xA )1/3 . Moreover, the lo-
explained from the definition of R, that the pressure-driven calization properties are also affected by the channel cor-
flow becomes more dominant as one increases R, and be- ner points, which are not considered in the present context.
yond a critical R (depending on the rheology and electric Eq. (7) can be solved with the similarity parameter, equals to


C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com
Electrophoresis 2013, 34, 668–673 Liquid Phase Separations 671

Figure 4. Enhancement of Sher-

wood number with scaled Dedye
length due to (A) the relative effect
of electroosmotic phenomenon
corresponding to n = 1.5, and (B)
rheology of fluid corresponding to
R = 1.

∗  1/3  
␩ = ␦y∗ = y ∗ xA∗ . In terms of the similarity parameter (␩), strated by considering no permeation case Pe w = 0 . The
the concentration profile can be expressed as [21]: enhancement factor (E) can be defined as ratio of Sherwood
␩  3  number to Sherwood number for purely pressure-driven
c 1 − B 0 exp − ␩9 − B␩ d␩ flow.
= ∞  3  , (9)
c0 1 − B 0 exp − ␩9 − B␩ d␩  1/
Sh L (␬h )n 3n
E= = 1± tanh (␬h ) (13)
 ∗ 1/3 Sh L R (n + 1)n
n
where, B = Pe4w xA . The constant B can be calculated R→∞

in
1 terms of length averaged permeate flux (Pe w ), as Pe w =
Pe w d x ∗
= 6B A 1/3
. The mass transfer coefficient (k) is de- The above expression is for electric field assisting (+) as
0
fined as: well as opposing (-) the flow. The variation of enhancement
  factor with ␬h for various combinations of R and n is pre-
  ∂c
k c| y=0 − c 0 = −D (10) sented in Fig. 4. The opposing flow effects clearly show a
∂ y y=0 stagnation point, indicated by zero E, corresponding to dif-
ferent magnitudes of R and n. Sherwood number can be
Computing the length averaged Sherwood number from enhanced by 2–5 times (depending on ␬h and shear thicken-
the mass transfer coefficient (k), we obtain: ing fluid, n = 1.5) when the electroosmotic effect is around
 1 tenfolds of the pressure gradient. Considering the rheology
Sh L = Sh(x∗ )d x∗ of the fluid, Sherwood number increases as the shear thin-
0
ning behavior is dominant. Sherwood number enhancement
  2n+1  1/ can be as large as 1.5–3 times for equally dominating effects
A(␬h,R,n) 3
6 1
ReSc dLe
=
16
⎛
n+1 R
⎞ . (11) (R = 1) corresponding to power law index n = 0.5.
∞ ␩3 Pe w
0
exp ⎝− − ␩⎠ d␩
9 1
6A (␬h, R, n) /3 3 Results and discussion
In case of no permeation, Pe w = 0 and purely pressure-
Effects of operating conditions and rheological properties of
driven flow 1/R = 0, the expression of average Sherwood
the fluid on Sherwood number are presented in Fig. 5. The
number becomes:
influence of Re.Sc. dLe on the mass transfer is described in
 1/  1/ Fig. 5A. As expected, average Sherwood number increases
1 3 de 3
Sh L = 1.284 2 + ReSc , (12) with Reynolds number due to enhanced forced convection
n L
that restricts the growth of mass transfer boundary layer.
which is identical with the expression of Hele–Shaw flow for In case of opposing flow, Sherwood number decreases up
power law fluid in a channel [30]. to ␬h = 2.5. This is directly interpreted from Fig. 3. At this
point, a stagnation zone is attained leading to no fluid flow
and associated mass transfer. Beyond ␬h = 2.5, the flow re-
2.3 Enhancement of Sherwood number verses its direction and finally, the Sherwood number in-
creases and matches with assisting flow case, because the
The relative improvement of mass transport with electroos- electric double layer becomes thinner at higher ␬h. A strik-
motic effect, in terms of Sherwood number can be demon- ing feature is that the stagnation region is always achieved


C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com
672 S. Mondal and S. De Electrophoresis 2013, 34, 668–673

Figure 5. Variation of
Sherwood under the
influence of different
operating conditions and
rheology of fluid. The
constant parameters for
figure (A) n = 1.5, R =
1, Pew = 100; (B) n =
1.5, R = 1, Re.Sc.de/L
= 107 ; (C) R = 1, Pew =
100, Re.Sc.de/L = 107 and
(D) n = 1.5, Pew = 100,
Re.Sc.de/L = 107 .

at ␬h = 2.5, irrespective of Re.Sc. dLe . This can be explained than 10. Since, physiological fluids exhibit shear thinning
from Eq. (11) as parameter A is independent of Reynolds behavior, the assisting flow should be preferred on account
number. Increase in permeation, which relates to increased of enhanced mass transfer. The relative dominance of the
permeability of porous wall, also affects the average Sher- pressure and electrokinetic effects (R) on Sherwood num-
wood number (Fig. 5B). The porous wall increases the mass ber is represented in Fig. 5D. To maintain a fixed Re.Sc. dLe ,
transport due to decrease in mass transfer boundary layer the electroosmotic velocity must decrease, leading to de-
[31, 32]. For highly porous membrane, the contribution of crease in overall fluid velocity, with R, which in turn de-
electrokinectic effects on the average Sherwood number is creases Sherwood number. However, at higher R, the elec-
masked and it approaches a constant value, independent of troosmotic effect becomes negligible leading to dominance
␬h. It can be noted that for transdermal drug delivery, a rel- of purely pressure-driven flow, and hence, Sherwood num-
atively less porous membrane (lower Pe W < 100) is suitable ber becomes independent of ␬h. At higher ␬h (lower elec-
for controlling the flow behavior and consequent transport of troosmotic effect), the stagnation point is attained in case of
solute by electrokinetic means. high R.
The effect of fluid rheology on mass transfer is illus- It may be mentioned here that the relative magnitude
trated in Fig. 5C. For a shear thickening fluid, the average of average Sherwood number is significantly influenced by
Sherwood number is relatively lower, compared to Newto- Re.Sc. dLe compared to other physical parameters.
nian and shear thinning fluid. At higher ␬h (beyond 10),
the Sherwood number for the opposing and assisting case 4 Concluding remarks
merges for the shear thickening fluid. For shear thinning
fluid, this effect is more gradual. It may be understood that Mass transfer coefficient, an important design parameter
the realistic range of ␬h in a microchannel is typically more was derived from first principles for flow through a porous


C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com
Electrophoresis 2013, 34, 668–673 Liquid Phase Separations 673

microchannel having non-Newtonian rheology. EOF had pro- [11] Jiang, L., Mikkelsen, J., Koo, J.-M., Huber, D., Yao, S.,
found impact on the Sherwood number. For EOF assisting Zhang, L., Zhou, P., Maveety, J. G., Prasher, R., Santiago,
the pressure-driven flow, Sherwood number could be en- J. G., Kenny, T. W., Goodson, K. E., IEEE Trans. Comp.
Packaging Technol. 2002, 25, 347–355.
hanced to more than five times for ␬h > 30, compared to
an impervious conduit. Enhancement of Sherwood number [12] Banga, A. K., Chien, Y. W., J. Control Release 1998, 7,
1–24.
was more in case of Shear thinning fluid compared to Shear
thickening one. Mass transfer increased with wall perme- [13] Brown, M. B., Traynor, M. J., Martin, G. P., Akomeah, F.
K., in: Jain, K. K. (Ed.), Drug Delivery Systems, Humana
ation, double-layer thickness, and flow velocity as the fluid
Press, New Jersey 2008, pp. 119–140.
behavior changed from pseudoplastic to dilatant rheology.
[14] Karimi, G., Li, X., J. Power Sources 2005, 140, 1–11.
However, the effect of flow velocity on mass transfer was
apparently more at the expense of more energy. The phe- [15] Matos, M. A., White, L. R., Tilton, R. D., J. Colloid Inter-
face Sci. 2006, 300, 429–436.
nomenon of stagnation was observed at a particular ␬h(∼ 2.5)
provided all other physical conditions were constant for a [16] Molla, S. H., Bhattacharjee, S., Langmuir 2008, 23,
10618–10627.
specific rheological fluid. Sherwood number expression de-
veloped in this work provided adequate quantification of the [17] Sadeghi, A., Saidi, M. H., Int. J. Heat Mass Transf. 2010,
53, 3782–3791.
transport of solutes (enzymes, biomolecules, microorganism,
etc.) in the tissues, capillaries, and target cells. More specific [18] Tang, G. H., Li, X. F., He, Y. L., Tao, W. Q., J. Nonnewton.
Fluid Mech. 2009, 157, 133–137.
non-Newtonian rheological fluid models can be investigated
to improve the predictive ability of the mass transport phe- [19] Vasu, N., De, S., Colloids Surf. A: Physiochem. Eng. Asp.
2010, 368, 44–52.
nomena. There is scope to validate the model developed in
this work against transport of real-life solutes in physiological [20] Vennela, N., Bhattacharjee, S., De, S., Chem. Eng. Sci.
2011, 66, 6515–6524.
systems of different microchannels.
[21] Vennela, N., Mondal, S., Bhattacharjee, S., De, S., AIChE
The authors have declared no conflicts of interest. J. 2012, 58, 1693–1703.
[22] Chen, X. B., Sui, Y., Lee, H. P., Bai, H. X., Yu, P., Winoto,
S. H., Low, H. T., J. Biomech. Eng. Trans. ASME 2010,
5 References
132, 061001.

[1] Stone, H. A., Stroock, A. D., Ajdari, A., Annu. Rev. Fluid [23] Chen, X. B., Sui, Y., Cheng, Y. P., Lee, H. P., Yu, P.,
Mech. 2004, 36, 381–411. Winoto, S. H., Low, H. T., Biochem. Eng. J. 2010, 52,
227–235.
[2] Kirby, B. J., Hasselbrink Jr., E. F., Electrophoresis 2004,
25, 187–202. [24] Zlokovic, B. V., Ghiso, J., Mackic, J. B., Mccomb, J.
G., Weiss, M. H., Frangione, B., Biochem. Biophys. Res.
[3] Erickson, D., Microfluid. Nanofluid. 2005, 1, 301–318. Comm. 1993, 197, 1034–1040.
[4] Masliyah, J. H., Bhattacharjee, S., Electrokinetic and Col- [25] Lundvall, J., Mellander, S., Westling, H., White, T., Acta
loid Transport Phenomena, Wiley, New Jersey 2006. Physiol. 1972, 85, 258–269.
[5] Chang, C.-C., Yang, R. J., Microfluid. Nanofluid. 2007, 3, [26] Merrikh, A. A., Lage, J. L., J. Biomech. Eng. 2005, 127,
501–525. 432–439.
[6] Simonnet, C., Groisman, A., Phys. Rev. Lett. 2005, 94, [27] Bird, R. B., Stewart, W. E., Lightfoot, E. N., Transport
134501. Phenomena, Wiley, New York 2007.
[7] Mala, G. M., Li, D., Dale, J. D., Int. J. Heat Mass Transf. [28] Truskey, G. A., Yuan, F., Katz, D. F., Transport Phenomena
1997, 40, 3079–3088. in Biological Systems, Prentice Hall, New Jersey, 2009.
[8] Patankar, N. A., Hu, H. H., Anal. Chem. 1988, 70, [29] De, S., Bhattacharjee, S., Sharma, A., Bhattacharya, P. K.,
1870–1881. J. Membr. Sci. 1997, 130, 99–121.
[9] Goodson, K. E., Chen, C. H., Huber, D. E., Jiang, L., Kenny, [30] Ranjan, R., Dasgupta, S., De, S., J. Food Eng. 2004, 64,
T. W., Koo, J. M., Laser, D. J., Mikkelsen, J. C., Santiago, 53–61.
J. G., Wang, E. N.-Y., Zeung, S., Zhang, L., US Patent
2005, 6, 942, 018. [31] De, S., Bhattacharya, P. K., J. Membr. Sci. 1997, 128,
119–131.
[10] Kandlikar, S., Garimella, S., Li, D., Colin, S., King,
M. R., Heat Transfer and Fluid Flow in Minichannels and [32] van Den Berg, G. B., Racz, I. G., Smolders, C. A., J.
Microchannels, Elsevier, Oxford 2006. Membr. Sci. 1989, 47, 25–51.


C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com