International Journal of Engineering Research & Technology (IJERT) ISSN: 2278-0181 Vol.

1 Issue 4, June - 2012

Transport of Aerosols in the Presence of Electric Field with Interphase Mass Transfer
Meena Priya. P* and Nirmala P.Ratchagar#

Mathematics Section, Faculty of Engineering and Technology, Annamalai University, Annamalai Nagar- 608 002, India
Abstract We have developed a mathematical model for evaluating the unsteady convective dispersion of aerosols with inter phase mass transfer by poorly conducting couple stress fluid in a channel in the presence of a transverse electric field. The validity of time dependent dispersion coefficient is widened by using a generalized dispersion coefficient. The exact expression is obtained for the dispersion coefficient (K2) which shows it decreases with increase in couple stress parameter a and reaction rate . The anlaysis leads to the novel result for convection coefficient (K1) and dispersion coefficient (K2) (which is a measure of the longitudinal dispersion coefficient of the aerosol). It is found that the value of K2 depends on the value of reaction rate ( ) and electric field (we) whereas the values of K1 is constant in both the cases. Finally, the effect of a, and the electric field (we) on the axial distribution of the mean concentration m is investigated in detail. keywords: electrodes, generalized dispersion, aerosols, chemical reaction. 1. Introduction The problem caused by the air pollution are complex and they affect natural processes, strongly influencing the ecological balance [2]. For this reason, it is important to study and understand the dispersion process of pollutants in the atmosphere in order to predict the possible impact of the pollution on the diverse ecosystems involved. Pollutants released from various sources into the atmosphere in the form of suspended ultrafine particles are aerosols [1] affect the environment directly or indirectly. The couple stress fluid theory developed by stokes [9] represents the simplest generalization of the classical viscous fluid theory that sustains couple stresses and the body couples. The presence of earth's surface influences the concentration of ions, aerosols and radioactive particles, through its control over the wind, temperature and water vapour distributions [4], [8]. Atmospheric electricity abounds in the environment [7]. Recently it has become evident that lightening is a form of electricity associated with thunderstorms [10]. The resulting ground level concentration patterns have to be estimated for a wide variety of air quality analysis for social planning and industrial growth [3]. It is not always feasible to measure/monitor concentration of species at various vulnerable points of a city. However man insight in this regard could be achieved with the help of suitable mathematical models. In this paper, we developed a mathematical model to study the unsteady convective diffusion of atmospheric aerosols with interphase mass transfer in a couple stress fluid flow through a channel in the presence of electric field. The atmospheric flow and concentration change of aerosols are commonly described by a set of partial differential equations, which are mathematical formulation of one or more of the conservation law of physics. These include the equations of mass, momentum species along with Maxwell's equation which involve advection and diffusion terms as a main constituent. They have been solved using generalised dispersion model [5] with appropriate boundary conditions and the results have been depicted graphically. 2. Mathematical Formulation The physical configuration shown in the figure 1 consists of an infinite horizontal couple stress fluid layer bounded on both sides by electroconducting impermeable right plates embedded with electrodes located at y = h and electric

www.ijert.org

International Journal of Engineering Research & Technology (IJERT) ISSN: 2278-0181 Vol. 1 Issue 4, June - 2012

potentials
V V x at y -h and x x0 at y h h h are maintained on theseboundarieswhere V is potential.
electrodes y

The equation (7) is zero because in a poorly conducting fluid, induced magnetic field is negligible and there is no applied magnetic field.
J E (Ohm' s law) (8)

V (x h

x0 )
Couple stress fluid -Xs/2

The above equations are solved using the following boundary conditions on velocity and potential are,
y=+h

No slip condition,
slug x Xs/2 y=-h

at

h, h

(9)

V h

The couple stress condition,

d 2u dy 2 0 at y h (10)

Fig. 1. Physical Configuration

In this paper, we make the assumption that the electrical conductivity ( ) is negligibly small and hence the magnetic field is negligible. This assumption makes the electric field conservative. i.e. =The basic equations are
0

, to be

V x at y h h V ( x x0 ) at y h

(11) h

(1)

In Cartesian form, using the approximation equation (3) becomes

p x
2 2

above

Conservation of mass for an incompressible flow . =0 (2)

Ex ,

y2

Conservation of momentum q 2 4 ( q. ) q p q q t Where is a couple stress parameter. Conservation of species

C t (q. )C D
2

In a poorly conducting fluid, the electrical conductivity is assumed to vary linearly with temperature in the form
e

E (3)

[1 0

(T T )] h b 0

(12)

where h is the coefficient of volumetric expansion. We assume the flow is fully developed and unidirectional in the x direction. This means the velocity is independent of time and all physical quantities except pressure and concentration are independent of x, so that the velocity and temperature will be functions of y only. Using the following dimensionless quantities,
y* y ; u* h P
2

(4) Conserva

tion of charges
e

(q. )

.J

(5)

Maxwells equation

u h

.E
E

; Ex * x h

e 0

(Gauss law)
law)

(6)
P*
(7)

Ex ; V h

e 0V h2

; x*

0 (Faraday ' s

www.ijert.org

International Journal of Engineering Research & Technology (IJERT) ISSN: 2278-0181 Vol. 1 Issue 4, June - 2012

Where V is electric potential, we get electric potential through electrodes. Equations (3) to (11) becomes,

Tb 1 at y 1 Tb 1 at y 1

(19)

is Tb=y

(20)

d u dy 4

h d u l 2 dy 2

h l2

2 e

Ex

V2

vh l
2

p x

Therefore equation (12) becomes =

0[1+ h

Ty] = =

0 (1+ h

y)

0e

We assume that the fluid with pollutants is isotropic and homogenous so that molecular diffusivity D, viscosity are all constants [6].

(Where

T) (21)

From equations (14) and (21) we get,

u 2 d 2u 2 a a weP E e x y4 dy2
4

a2 P

(13)

d2 dy2

d dy

(22 )

where
we
0

V2

;P

P ,l x

,a

h l

Its solution satisfying the boundary condition (17) is

x x0 2 sin h [e e
y

( 23)

is the couple stress parameter Equation (5) becomes, we get, (

2

=0 using equation (1) (14)

using the dimensionless quantities and equation (23), equation (6), (7) and (8) reduce to
e

)+

=0

= . =-

=y

2 2 x0 e 2 sinh

The boundary conditions on velocity, couple stress and electric potentials after dimensionless are u = 0 at y= 1 (15)

eEx=

2 2 x0 e 2 sin h

( 24 )

d 2u dy2

The solution of equation (13) satisfying the condition (15), (16) is

at

(16)

x at y 1 x x0 at y 1

(17)

P 2 y b2 sinh ay 2 b3 cosh ay b4e y u b0 y b1

Where b0
a0 sinh( ) = 2 a2 a0 cosh( ) 2 a2 1
2

(25)
1 2 1 a2

The solution for , according to (14)depends on which in turn depends on the temperature Tb as in (12). In a poorly conducting fluid, <<1 and hence any perturbation on it is negligible and hence it depends on the conduction temperature Tb namely,

1 ; b1 a2 1 a2 1
2

b2 b3 b4

d 2Tb dy2

a0 sin h ; 2 sinh a a 2 a 2 1 P a2 cosha a0

2 2

(18)

a0 cosh ; 2 a2

Hence the solution satisfying the boundary conditions

a2

www.ijert.org

International Journal of Engineering Research & Technology (IJERT) ISSN: 2278-0181 Vol. 1 Issue 4, June - 2012

and a0

We x0 2 sin h

Where U= u , u is the average velocity of the

u

The average velocity is given by,

u 1 udy 20
1

flow, Peis the Peclet number and is the dimensionless reaction rate constant, and the initial conditions reduce to (0,X, ) = (X)Y( ),
at at 1, (30b) 1,

b1

P 6

b3 sinh a a (26)

(30a)

b4 sinh

Into this flow, introduced a slug input of concentration C(0,x,y) and the local concentration C(t,x,y) of the solute which satisfies the convective diffusion equation
C t u ( y, t ) C x
2

, ,

, ,

(30c)

C x2

C y2

(27)

In the generalized dispersion model [5], we let

k

along with the initial condition

,X,
k

fk ,

m k'

(31)

C (0, x, y )

C 0 1 ( x)Y1 ( y )

(28a )
Where
1 , X , ) d is m 2 1 average concentration.
1

Where C0 is a reference concentration. The corresponding boundary condition at the channel walls are
D C K s C at y C D K s C at y y y h and (28b) h

the

dimensionless (31a)
2

Equation (29) is multiplied throughout by 1 and integrated with respect to y from -1 to 1. Wh Introducing the definition of
m,

we get

ere Ks is the reaction rate constant catalysed by the walls.

C t, , y C t, , y x 0 (28c)

m
1

1 Pe2

X d

m 2

1 2

On introducing the following non-dimensional quantities,

1 U , 2 x 1

(32)

Using equation (31) in (32), the dispersion model

Dt ; h2

C ;X C0

Dx ; h 2u

y ; h

for

is obtained as

u U( , ) ; Pe u

Ks H hu ; D D

m
i 0

Ki ( )

m i

(33)

Equation (27) and (28) become

U 1 2 Pe2 X 2
2 2

Where Kis are given by

(29 )
Ki( ) 1 2
1

P fi
1

i2 2 e

1 fi 2 , U ,

,1 d (34)

www.ijert.org

International Journal of Engineering Research & Technology (IJERT) ISSN: 2278-0181 Vol. 1 Issue 4, June - 2012

Where f-1=0 and defined by

1,
ij

i2

is the Kroneckar delta

fk

,1

f k ,1

(37b)

i i

j j

fk

,0

(37 c)

0,

The exchange coefficient K0 ( ) accounts for the nonzero solute flux at the channel wall and negative sign indicates the depletion of solute in the system with time caused by the irreversible reaction, which occurs at the channel wall. The presence of non-zero solute flux at the walls of the channel also affects the higher order Ki due to the explicit appearance of fi/ ( ,1) in equation (34). Equation (33) can be truncated after the term involving K2 without causing serious error because K3, K4, etc. become negligibly small compared to K2.The resulting model for the mean concentration is
2

1 2

fk , d
1

k0 , k

0,1,2 (37 d )

The function f0 and the exchange coefficient K0 are independent of the velocity field and can be solved easily. It should be pointed out here that a simulatenous solution has to be obtained from these two quantities since K0, which can be obtained from (34) as
K0 1 2 f0 y
1

(38)
1

Substituting k =0 in (36) we get the differential equation or f0 as

f0
2

f0
2

f0 K0

(39 )

(35) X X To solve this equation, we need the coefficients Ki( ) in addition to the appropriate initial and boundary conditions. For this, the corresponding function fk must be determined. So, substituting (31) into (29) and using equation (32), the following set of differential equations for fk are generated.
m

K0 ( )

K1

K2

m 2

We now derive an initial condition for f0 using (31a) by taking = 0 in that equation to get

0, X

1 2

0, X , d
1

(40)

Substituting =0 in (31) and setting fk( ) = 0 (k=1,2,3) gives us the initial condition for f0 as

fk
k

2 fk 2

Uf k K

1 fk Pe2

f 0 0,
2

0, X , m 0, X

(41)

Ki f k
i 0

0,1,2,...

(36)

where f-1 = f-2 = 0. We note that to evaluate Kis we need to know the fks which are obtained by solving (36) for fks subject to the boundary conditions.

We note that the left hand side of (41) is function of only and the right hand side is a function of both X and . Thus clearly the initial concentration distribution must be a separable function of X and . This is the justification for the chosen form of (0,X, ) in (41). Substituting (30) into (41), we get

f 0 0,

f k ,0

finite

(37 a)

1 2

(42) d

The solution of the reaction diffusion equation (39) with these conditions may be formulated as
www.ijert.org

International Journal of Engineering Research & Technology (IJERT) ISSN: 2278-0181 Vol. 1 Issue 4, June - 2012
t

f0 ,

g0 ,

exp
0

K 0 ( )d

(43)

From which it follows that g0 ( , ) has to satisfy

g0
2

ensured convergence of the series seen in the expansions for f0 and K0. Having obtained f0, we get K0 from (38) in the form
9

g0
2

(44 )

An K0
n 0 9 n 0

exp exp

2 n 2 n

sin sin

along with the conditions

An
n n

(50)
n

f 0 0,

g 0 0,

1 2

(45a) K 0 d

is independent of velocity distribution.

As , we get the asympotic solution for K0 from (50) as (45b) K0 ( ) =

2 0

g0 ( ,0)= finite

(51)

g0

,1

g 0 ,1

(45c)

The solution of (44) subject to conditions (45) is

g0
n 0

An cos

exp

2 n

(46 )

Where 0 is the first root of the equation (47). Physically this represents first order chemical reaction coefficient having obtained K0( ), we can now get K1( ), from (34) with (i=1) knowing f0 ( , ) and f1( , ). Likewise K2( ), K3( ), require the knowledge of K0, K1, f0, f1 and f2. Equation (49) in limit reduces to

f0
Where
ns

sin

cos
0

(52)

are the roots of (47)

ntan n=

, n=0,1,2,3..

and Ans are given by

1

We then find f1, K1,f2, and K2. For asymptotically long times, i.e., , (34) and (36) give us Kis and fks as
1

2 An
1

( ) cos

d (48)

Ki

i2 2 e

fi

,1

Uf i
1

, d , (53)

sin 2 n 1 1 ( )d 2 n 1 Now from (43) is follows that

2g0 ,
1

(i 1,2,3) d 2 fk 2 0 fk d 2
1 Pe2 K2 fk 2 ,

K1 f k (K

1,2)

(54)

f0 ,

g0 ,
1 9

d cos

The fks must satisfy the conditions (31a) and this permits the eigen function expansion in the form
n

An exp
n 0 9 n 0

2 n 2 n

An
n

(49 ) exp sin

n

fk

,
j 0

B j ,k cos(

),

1,2,3,....

(55)

The first ten roots of the transcendental equation (47) are obtained using mathematica and are given in Table 1. We find that these ten roots

Substituting (55) in equation (54) multiplying the resulting equation by Cos and integrating with respect to we get after simplification

and
j

from -1 to 1,

www.ijert.org

International Journal of Engineering Research & Technology (IJERT) ISSN: 2278-0181 Vol. 1 Issue 4, June - 2012
K 2 i 1

B j ,k

1
2 j 2 0

1 B j ,k Pe2
1 9

ki B j , k

sin 2 2
j

B j,k 1 I ( j ,1)
j 0

(56)

1,2

Where
1

I j, l

U cos
1

. cos

I (l , j )

(57)

and the parameters of the problem a and . This distribution is valid only for long time and is a gross approximation at short and moderate times. The initial conditions for solving (33) can be obtained from (30a) by taking the cross-sectional average. Since we are making long time evaluations of the coefficients, we note that the side effect is independent of m on the initial concentration distribution. In view of this, the solution of (33) with asymptotic coefficients can be written as
m

,X

1 2 Pe X K2 K1 4K 2
2

B j,

0, B j ,0

for

j 1 to

(58) exp K 0
Where
m

(63)

The first expansion coefficient B0,k in (55) can be expressed in terms of Bj,k (j=1 to 9) by using the conditions (37) as
9

0,

B0,k

sin

B j ,k
0 j 1

sin
j

(64)

, k 1,2,3,...

(59)

3. Results and Discussions We have modeled the solvent as a couple-stress fluid and studied dispersion of solute in a couplestress fluid bounded by electrodes considering heterogenous chemical reaction, on the interphase. The walls of the channel act as catalysts to the first order reaction and the effect of wall reaction on dispersion is discussed. The problem brings into focus three important dispersion coefficients namely the exchange coefficient (-K0), the convective coefficient (-K1) and diffusive coefficient (K2). The exchange coefficient arises only due to the interphase mass transfer and hence it is unaffected by the flow and thus is independent of the couple-stress parameter a. The asymptotic values of these three coefficients are plotted in figure 3 to 6 for various value of a and reaction rate parameter . From these figures we predict the following. Figure 2 is a plot of the velocity against the nondimensional transverse coordinate for different values of electric number We. It is seen that the velocity profile is a parabolic curve. We note that the velocity increases with the increase in electric number. From figure 3, it is evident that K0( ) increases with increase in the value of the wall reaction parameter and is unaffected by the couple-stress parameter a.

Further, from (55) and (52) we find that

B0, 0

sin

(60 )
0

Substituting i=1 in (53) and using (57), (58) and (60) in the resulting equation, we get
I 0,0 (61) sin 2 0 1 2 0 Substituting i=2 in (53) and using (56), (57) and (60) in the resulting equation, we get K1
K2 1 Pe2
0

Sin

9 0 0 j 0

sin 2 1 2 0

B j ,1 .I j ,0

(62)

WhereBj - 1 =

2 j

2 0

sin 2 0 2 0

1 0

sin

.I ( j,0)
0

Using the asymptotic coefficients K0( ),K1( ) and K2( ), in (33). One can determine the mean concentration distribution as a function of X,

www.ijert.org

International Journal of Engineering Research & Technology (IJERT) ISSN: 2278-0181 Vol. 1 Issue 4, June - 2012

The convective coefficient -K1 is plotted in figure 4 versus wall reaction parameter for different couple stress parameter a with a fixed value of slip parameter = 0.1. From these figures we conclude that increase in as well as decrease in a is to increase K1. This is advantageous in maintaining the laminar of flow. Figure 5 is a plot of dispersion coefficient K2 against electric number We for different values of . From this we conclude that increase in and We is to decrease the effective dispersion coefficient K2. The scaled dispersion coefficient K2- Pe 2 is plotted versus in figure 6 for different values of a. From this figure, it is clear that the increase in a and , the effective dispersion coefficient decreases. These are useful in the control of dispersion of a solute. The cross sectional average concentration m is plotted versus x in figures 7 and 8 respectively for different values of a, and for fixed values of the other parameters given in these figures. It is clear that the increase in and a increases m as expected on the physical grounds. m is also plotted in figures 9 and 10 against the dimensionless time for different values of and a for fixed values of the other parameters given in these figures. We note that the peak of occuring at m decreases with an increase in the lower interval of time . We also note that the peak increases with an increase in a but occurs at almost at the same interval of time . These informations are useful to understand the transport of solute at different times. The proposed model and analysis presented here also suggests that to remove the pollutants from the atmosphere, external species may be introduced in the atmosphere which can interact with the pollutant and remove it by some suitable removal processes based upon the physical and chemical properties of the pollutants as well as that of externally introduced species.

Acknowledgment: The author's interest in this work was originally stimulated by correspondence with Dr. N. Rudraiah, Honorary Professor, UGC-Centre for Advanced Studies in Fluid Mechanics, Department of Mathematics, Bangalore University. The authors are thankful for his helpful suggestion and constructive criticism. 4. References
[1] Albrecht, B.A. Aerosols, Cloud microphysics and fractinoal cloudiness. Science, 245, (1989), 1227. [2] Brooke, J.W., Kontomaris and Hanratty, T.J. Turbulent deposition and trapping of aerosols at a wall. Phy. Fluids A4, (1992), 825. [3] Despiau, S., Cougnenc, S. and Resch, F. Concentrations and size distributions of aerosol particles in coastal zone. Journal of Aerosol Science 27. (1996), 403-415. [4] Finlayson-Pitts, B.J. and Pitts, J.N. Tropospheric air pollution: Ozone, airborne toxics polycyclic aromatic hydrocarbons and particles. Science, 276, (1997), 1045. [5] Gill, W.N. and Sankarasubramanian, R. Exact analysis of unsteady convective diffusion. Proc. Roy. Soc. London. Series, A 316, (1970), 341. [6] Gupta, P.S., and Gupta, A.S. Effect of homogenous and heterogenous reactions on the dispersion of a solute in the laminar flow between two plates. Proc. roy. Soc. London., A 330, (1972), 89. [7] Jayaratne, E.R., and Verma, T.S. Environmental aerosols and thier effect on the earths local fair weather electric field, meteorol. Atmos. Phys. 86, (2004), 275. [8] Nirmala P. Ratchagar and Meena Priya P. Effect of couple Stress on the unsteady convective diffusion of atmospheric aerosols in the presence of electric field. International journal of mathematical Sciences and Engineering Applications (2011), 5(4): 293-310. [9] Stokes, V.K. Couple stress in fluids. Phys. fluids 9, (1966), 1709-1715. [10] Vinayak Vidyadhar Barve. Simulation of gravitational setting under the electric Reids, Master of Science in Engineering Thesis, (2002), University of Texas,

www.ijert.org

International Journal of Engineering Research & Technology (IJERT) ISSN: 2278-0181 Vol. 1 Issue 4, June - 2012

Table 1: Roots of the equation

0 1 2 3 4 5

n tan
6

=
7 8 9

10-2 0.05 10-1 0.5 1.0 5.0 10.0 100.0

0.099834 0.22176 0.311053 0.653271 0.860334 1.31384 1.42887 1.55525

3.14477 3.15743 3.1731 3.29231 3.42562 4.03357 4.3058 4.66577

6.28478 3.15743 6.29906 6.36162 6.4373 6.9096 7.22811 7.77637

9.42584 6.29113 9.43538 9.47749 9.52933 9.89275 10.2003 10.8871

12.5672 9.43008 12.5743 12.606 12.6453 12.9352 13.2142 13.9981

15.7086 12.5703 15.7143 15.7397 15.7713 16.0107 16.2594 17.1093

18.8501 15.7111 18.8549 18.876 18.9024 19.1055 19.327 20.2208

21.9916 21.9934 21.9957 22.0139 22.2126 22.2126 22.4108 23.3327

25.1331 25.1347 25.1367 25.1526 25.1724 25.3276 25.5064 26.445

28.2747 28.2761 28.2779 28.292 28.3096 28.4483 28.6106 29.5577

Figure 2. Velocity profiles of aerosols or different We

www.ijert.org

International Journal of Engineering Research & Technology (IJERT) ISSN: 2278-0181 Vol. 1 Issue 4, June - 2012

Figure 3. Plots of exchange coefficient versus reaction rate parameter

Figure 4. Plot of convective coefficient K, with wall reaction parameter for different values of a.

Figure 5. Effect of electric number We on depersion coefficient K2 for different values of

Figures 6. Plots of scaled dispersion coefficient K2( )- Pe-2 against for different values of a.

Figures 7. Mean concentration ( m) varying along axial distance x for different values of a and for fixed = 0.06, =0.01

Figures 8. Mean concentration ( m) varying along axial distance x for different values of and for fixed = 0.06, a=1
10

www.ijert.org

International Journal of Engineering Research & Technology (IJERT) ISSN: 2278-0181 Vol. 1 Issue 4, June - 2012

Figures 9. Mean concentration ( m) varying Figures 10. Mean concentration ( m) varying along dimensionless time for different values of along dimensionless time for different values of a at x=0.6, =0.01 at x=0.6, a=1

www.ijert.org

11