Week 3-Diffusion in

dilute solutions
Dr. Emad Alhseinat

CHEG324: Mass Transfer

Find the flux across a thin film in which diffusion varies sharply (i.e., the diffusion coefficient is
not a constant). Assume that below some critical concentration c1c, diffusion is fast, but above
this concentration it is suddenly much slower.

left d dc1
c10 l 0 j1 j1  D L
dz dz

DL
 j1dz  D L  dc1 c10  c1c 
zc c1c
c1c j1 
c1l 0 c 10 zc
zc
right 0   d j dc1
1 j1  D R
dz dz
large diffusion coefficient
small diffusion coefficient
DR
 j1dz  DR  dc1 c1c  c1l 
l c1l
j1 
zc c1c l  zc
DL c10  c1c  DR c1c  c1l 
The flux is the same across both films: j1 
l
In film 1: Large dc  smallD s
dz
zc c1c
dc 1 Ds
j1  D s   j1 dz   D s  dc 1 j1  (c 10  c 1c ) (1)
dz 0 c10 zc

In film 2: Small dc  large D 𝑙

dz
𝑙 c1𝑙
dc1 D𝑙
j1  D𝑙   j1 dz   D 𝑙  dc 1 j1  (c1c  c1𝑙) (2)
dz zc c1c 𝑙 z c

𝑙(c10 c1c )Ds 𝑙

(1) = (2)  zc  zc 
D s (c 10  c 1c )  D (c
𝑙 1c  c 1𝑙) 𝑙 (c 1c  c 1𝑙)
1 D
Ds(c10 c1c )

Ds(c10  c1c )  D (c
𝑙 1c  c 1𝑙)
The flux becomes then: j1 
𝑙
If
c c
c 1c  10 1𝑙 then D  Ds  D𝑙
2 2
Unsteady diffusion in a semiinfinite slab - free diffusion
• Any diffusion problem will behave as if the slab is
infinitely thick at short enough times.

At time zero, the concentration at z = 0 suddenly increases to c10

Mass balance on the thin layer Az

c10
rate of rate of diffusion
Solute
diffusion into out of the layer
time accumulation =
the layer at z at z + z

c1  
Azc1   A j1 z  j1 
t
zz

position z
 Mathematical Techniques
𝐶1∞
• Separation of variables
• Similarity method Z
• Laplace transforms 𝐶10

c1  c1 at t  0, for all z

c1  c10 at t  0, z  0
c1  c1 at t  0, z  

One-dimensional mass transfer in a semi-

The solution of this problem is easiest using the method of infinite solid
‘‘combination of variables.’’
 c1  2 c1

Azc1   A j1 z  j1  t
D
z 2
t
zz

Boundary conditions
Dividing A z
c1  c1 at t  0, for all z
c1  c10 at t  0, z  0
  j1 zz  j1 z 
c1     c1  c1 at t  0, z  
t  (z  z)  z  z

z  0 c1  c10 4Dt
 erf
c1  c1  c10 erf 
2


e s ds
2

 j1  0

t z j1  D
c1
z
c1
The solution j1  D  z2
z c1 D
of this j1  D  e 4 Dt c10  c1 
problem is z t
c1  2 c1
easiest using
D 2
the method of
t z
‘‘combination
j1 
D
c10  c1 
t
z0
of variables.’’ Fick’s second law
 The trick to solving this problem is to define a new variable

The differential equation can then be written as

or

In other words, the partial differential equation has been almost magically transformed
into an ordinary differential equation.

The magic also works for the boundary conditions:

and

The solution is now straightforward. One integration gives where a is an integration constant

A second integration and use of the boundary conditions give

where is the error function of . .

This is the desired concentration profile giving the variation of concentration with
position and time.
In many practical problems, the flux in the slab is of greater interest than the concentration profile itself. This
flux can again be found by combining Fick’s law with previous Eq.

One particularly useful limit is the flux across the interface at z = 0

This flux is the value at the particular time t and not that averaged over time
Example 2.3.1:Diffusion across an interface

The picture of the process in the Figure

implies that the concentration at z = 0 is continuous. This
would be true, for example,
if when z > 0 there was a highly swollen gel, and when z <
0 there was a stirred solution.
A much more common case occurs when there is a gas–
liquid interface at z = 0.
Ordinarily, the gas at z <0 will be well mixed, but the liquid
will not. How will this interface affect the results given
earlier?

Solution: Basically, it will have no effect. The only change will be a new boundary condition

where c1 is the concentration of solute in the liquid, x1 is its mole fraction, p10 is its partial
pressure in the gas phase, H is the Henry’s law constant, and c is the total molar concentration in the
liquid.
Example 2.3.2:Free diffusion with fast chemical reaction

In many problems, the diffusing solutes react rapidly and reversibly with surrounding material. The
surrounding material is stationary and cannot diffuse. For example, in the dyeing of wool, some dyes
can react quickly and reversibly with the wool as dye diffuses into the fiber.
 How does such a rapid chemical reaction change the concentration profile and the flux?

Solution: In this case, the chemical reaction can radically change the process by reducing the
apparent diffusion coefficient and increasing the interfacial flux of solute. These radical changes
stand in stark contrast to the steady-state result, where the chemical reaction produces no effect.

To solve this example, we first recognize that the solute is effectively present in two forms: (1) free solute that can
diffuse and (2) reacted solute fixed at the point of reaction. If this reaction is reversible and faster than diffusion,

where c2 is the concentration of the solute that has already reacted, c1 is the concentration of the unreacted solute
that can diffuse, and K is the equilibrium constant of the reaction. If the reaction is minor, K will be small; as the
reaction becomes irreversible, K will become very large.
With these definitions, we now write a mass balance for each solute form. These mass balances should have the
form

For the diffusing solute, this is

where r1 is the rate of disappearance per volume of species 1, the diffusing solute

Same mathematical treatment as before this becomes

 The term on the left-hand side is the accumulation; the first term on the right is the
diffusion in minus the diffusion out; the term r1 is the effect of chemical reaction.
When we write a similar mass balance on the second species, we find

We do not get a diffusion term because the reacted solute cannot

diffuse. We get a reaction term that has a different sign but the same
magnitude, because any solute that disappears as species 1
reappears as species 2.

To solve these questions, we first add them to eliminate the reaction term:

We now use the fact that the chemical reaction is at equilibrium:

This result is subject to the same initial and boundary conditions as before

As a result, the only difference between this example and the earlier
problem is that D/(1+K) replaces D.

 The chemical reaction has left the mathematical form of the answer unchanged, but it has
altered the apparent diffusion coefficient. The concentration profile now is

and the interfacial flux is

The flux has been increased by the chemical reaction.

 Thus a fast chemical reaction can tremendously influence the unsteady diffusion process.
The steady dissolution of a spherical particle
The sphere is of a sparingly soluble material, so that the sphere’s size
does not change much. However, the material quickly dissolves in the
surrounding solvent, so that solute’s concentration at the sphere’s
surface is saturated. The sphere is immersed in a very large fluid
volume, the concentration far from the sphere is zero. Find the
dissolution rate and the concentration profile around the sphere.

Mass balance on a spherical shell of thickness r

located at r from the sphere:
r
Solute rate of
rate of diffusion
accumulation = diffusion into
out of the shell
within the shell the shell


t
  
4r 2 rc1  4r 2 j1 r  4r 2 j1 r r 
s.s
Mass balance on a spherical shell:

 solute accumulation   diffusion    diffusion 

     
 within this shell   into the shell   out of the shell 


(4 r 2 r c1)  (4 r 2 j 1) r  (4 r 2 j 1) rr (1)
t
Divide both sides by the spherical shell’s volume, note that
LHS=0 at steady-state and take the limit as r  0

0
1 d 2
r 2 dr
 
r j1 (2)

Combine this with Fick’s law at spherical coordinates and D = const:

0  D d  r2 dc1 
r 2 dr  dr  (3)
Boundary Conditions: r  R0 c1  c1(sat) (4)
r c1  0 (5)

Integrating eqn. 3 gives: dc1  a (6)

dr r2
where a is an integration constant.

Integrating eqn. 6 again gives: c1  b a (7)

r
where b is another integration constant.

Using the B.C. gives b=0 from eqn. 5 and a =c1(sat)R0 from
eqn. 4 so eqn. 7 becomes
R
c1  c1(sat) 0 (8)
r
The dissolution flux can be found from Fick’s law:

 D  c1(sat) 0  
dc1 d R Dc1(sat)R 0
j1  D
dr dr  r  r2

which at the sphere’s surface is j1 D c1(sat)

R0
If you double the sphere (particle) size, the dissolution rate per unit
area is only half as large even though the total dissolution rate
over the entire surface is doubled.
Very important in pharmaceutics!!
Also in the growth of fog droplets and spraying, as well as in
growth of particles by condensation or by surface reaction limited
by transport.
 
0  4r 2 j1 r  4r 2 j1 r r  Boundary conditions
Dividing 4r2r r  R0 , c1  c1 (sat)
r  0
r  , c1  0
1 d
0 2
r dr
r2 j1  
c1
j1  D
r

D d r 2 dc1 R0
0 2 2 c1  c1 (sat)
r dr dr r
c1
j1  D
r

Example: R0
j1  D 2 c1 (sat)
The growth of fog droplets and the dissolution of drugs r
The diffusion of a solute into the cylinder
The cylinder initially contains no solute. At time zero, it is suddenly
immersed in a well-stirred solution that is of such enormous volume
that its solute concentration is constant. The solute diffuses into the
cylinder symmetrically. Find the solute’s concentration in this
cylinder as a function of time and location.

Mass balance on a cylindrical shell of thickness

r located at r from the central axis:

z Solute rate of
rate of diffusion
accumulation = diffusion into
out of the shell
r within the shell the shell


2rLrc1   2rLj1 r  2rLj1 r r 
t

2rLrc1   2rLj1 r  2rLj1 r r 
t
Dividing 2rLr Boundary conditions
r  0
t  0, all r, c1  0

c1    1  rj1 
t r r t  0, r  R0 , c1  c1 (surface)
c1
j1  D c1
r t  0, r  0, 0
r
c1 D  c1
 r
t r r r
c1 D  c1 t  0, all r, c1  0
 r
t r r r
t  0, r  R0 , c1  c1 (surface)

c1
t  0, r  0, 0
r
Dimensionless:

  1
c1

r   Dt2
c1 (surface) R0 R0

 1     0, all  ,   1
 
   
  0,   1,   0


  0,   0, 0

 1     0, all  ,   1
 
   
  0,   1,   0


  0,   0, 0


Assume:   g  f ( ) Using the method of “Separation of variables”

…

Please refer to my lecture note number 8 for the “applied mathematics”.


 
 
2  n
 0 n
J (  )e

n1  n 1
J ( n 
)
2 2
 D n t / R0
 J 0 ( n r )e
 1 2
c1 R0
c1 (surface) n1  n J 1 n r R 
 0
Diffusion across a thin, moving liquid film
The concentrations on both sides of this film are fixed by
electrochemical reactions, but the film itself is moving steadily.
Direction of diffusion Assumptions:
• the liquid is dilute
z • the liquid is the only resistance to mass transfer
• diffusion in the z direction
x • convection in the x direction
c10 c1l control volume

Mass balance on a control volume W x z :

Solute rate of rate of

accumulation = diffusion in diffusion in
moving liquid film in Wxz the z direction the x direction


c1Wxz   j1Wx z  j1Wx  c v Wz  c1v xWz 
t
zz 1 x x xx
 
c1 Wxz   j1 Wx z  j1Wx  c v Wz  c1v xWz 
t
zz 1 x x xx

s.s. Dividing Wx z Neither c1 nor vx change with x

x  0
z  0

dj1
0
dz
dc1
j1  D
dz

d 2 c1 z
0  D 2 c1  c10  (c1l  c10 )
dz B.C. l
z  0, c1  c10
j1 
D
c10  c1l 
l
z  l, c1  c1l
The flow has no effect!
Diffusion into a falling film
A thin liquid film flows slowly and without ripples down a flat
surface. One side of this film wets the surface; the other side is in
contact with a gas, which is sparingly soluble in the liquid. Find how
much gas dissolve in the liquid.

solute gas Assumptions:

z • the liquid is dilute
• the contact between gas and liquid is short
x • diffusion in the z direction
• convection in the x direction
l control volume
Mass balance on a control volume W x z :
Solute rate of rate of
Liquid with dissolved accumulation = diffusion in diffusion in
solute gas in Wxz the z direction the x direction


c1Wxz   j1Wx z  j1Wx  c v Wz  c1v xWz 
t
zz 1 x x xx
 
c1 Wxz   j1 Wx z  j1Wx  c v Wz  c1v xWz 
t
zz 1 x x xx

s.s. Dividing Wx z

x  0
z  0

j1 
0  c1v x
z x
dc1
vx ~ constant j1  D
dz

c1  2 c1 c1 z
 D 2  1 erf
 x / v x  z c1 (sat) 4Dx / v x
B.C.
x  0, all z, c1  0 vx
j1 z0  D c1 (sat)
x  0, z  0, c1  c1 (sat) x
x  0, z  l, c1  0
What we have done are:
1. We write a mass balance as a differential equation
2. Combine this with Fick’s law
3. Integrate this to find the desired result

D
j1  c1 For thin film
l
D
j1  c1 For thick slab
t

Fourier Number
2 Much larger than unity …………. Assume a semiinfinite slab
(length) Much less than unity ..…Assume a steady state or an equilibrium
(D)(time) Approximately unity ……..………. Used to estimate the process
detailed analysis
Example: Membrane for industrial separation:
Thickness = 0.01 cm
D = 10-11 m2/s
If the duration of the experiment is
4 2
1 10 cm
a) t=10 s
Fo 10 7 cm2 s 100
10s
This is a semi-infinite slab problem!

1 10 4 cm2 0.1
b) t=3 hrs 104 s 7 2 4
Fo 10 cm s 10 s
This is a thin film, steady-state problem.

The value of Fo = 1 indicates that mass transfer is significantly

advanced in a given process. As a result it can be used to estimate
the EXTENT (or DEGREE) of advancement (or progress) for
unsteady-state processes.
For example:
a) Guess how far gasoline has evaporated into the stagnant air in a
regular glass-fiber filter. Say that evaporation is going on for 10 min
and D = 10-5 m2/s.

1  length 2  1 length  8cm

Fo 10-5 m2 s 600s

b) Consider H2 diffusion in nickel making it rather brittle. If

D = 10-12 m2/s estimate how long it will take for H2 to diffuse 1
mm through the Ni specimen.

1 10 -6 m2
 -12 2  1  t  10 6s  11days
Fo 10 m s t
Another important difference of the two limiting cases stems
from the interfacial fluxes.

j1  D c 1 (thin film)
l

j1  D c (thick slab)
t 1

D D
Note that both l
and t
have velocity units (dimensions),

some people even call them “the velocity of diffusion”. In

fact these are equivalent to the mass transfer coefficients
we talked earlier on !!