Mass Transfer Operations

CHMT 3005

Prof. Geoffrey S. Simate

School of Chemical and Metallurgical Engineering

University of the Witwatersrand, Johannesburg

RW 517
 LECTURE 1/2

Introduction to Basic Concepts of

Mass Transfer

Prof. Geoffrey S. Simate

Room RW 119

2
Introduction
What happens if a few crystals of a coloured material like CuSO4 are placed at the
bottom of a tall bottle filled with water?
 The colour will spread slowly through the bottle until the solution appears
homogeneous
 Process responsible for the movement of the CuSO4 is termed molecular diffusion
 Molecular diffusion is the transfer or movement of individual molecule from a
higher concentration to a lower concentration in a stagnant medium.
To increase the rate of mixing of CuSO4, the liquid can be mechanically agitated, and
thus convective mass transfer will occur
 Mass transfer occurring under the influence of motion in a fluid medium is called
‘convective mass transfer’, e.g., delivery of oxygen in body fluid
Therefore, mass transfer which is the movement of material from one homogeneous
phase to another; with or without phase change comprises of:
 Molecular diffusion
 Convection
3
Introduction
The transport or migration of one constituent from a region of higher concentration to
that of a lower concentration is known as mass transfer
Mass transfer operations depend on molecules diffusing from one distinct phase to
another and are based upon differences in the physico-chemical properties of the
molecules, such as vapour pressure or solubility
For interphase mass transfer, there is a concentration gradient between bulk and
interface, however under steady state, at interface equilibrium is assumed
Mass transfer operation plays an important role in many industrial processes. A group
of operations for separating the components of mixtures is based on the transfer of
material from one homogeneous phase to another. These methods are covered by the
term mass transfer operations which include techniques like gas absorption and
stripping, liquid-liquid extraction, leaching, distillation, humidification, drying,
crystallization and number of other separation techniques

4
Two basic models of diffusion
Imagine two large bulbs of equal volume at constant temp and press are connected by
a long thin capillary tube (see Figure below). One bulb contains CO2, and the other is
filled with N2.

If you measure the concentration of CO2 in the bulb that initially contains N2, you find
that the concentration of CO2 varies linearly with time (see figure above). From this,
the amount transferred per unit time (i.e., gradient) is known.

Now develop a generalised expression for the transfer of diffusing species, that is
applicable not only to this experiment but also in other experiments (or physical
properties ), by introducing the term flux defined as:

(Flow quantity)
Flux = (1) 5
(time) (area)
Two basic models of diffusion
The term ‘flux’ means the net rate at which a species in a medium passes through a
unit area, which is normal to the direction of diffusion, in unit time. It is expressed in
kg/m2s or kmol/m2s

The flow quantity in Eqn (1) may be taken as mass, momentum, energy, electrons,
neutron, etc

In the case of CO2 in the given illustration,

(Amount of gas removed)

CO2 flux = (2)
(time) (cross sectional area of capillary)

Next assume that the flux is proportional to the gas concentration:

CO2 flux = k (CO2 concentration difference) (3)

The proportionality constant k is called a mass transfer coefficient. Its introduction

signals one of the two basic models of diffusion (to be discussed in detail in the
6
subsequent lectures)
Two basic models of diffusion
Alternatively, we can recognize that increasing the capillary’s length will decrease the
flux, and we can then assume that

(CO2 concentration difference)

CO2 flux = D (4)
length of capillary

The new proportionality constant D is the diffusion coefficient. Its introduction implies
the other model for diffusion, the model often called Fick’s law

7
Choice between the two models
The choice between the two models outlined represents a compromise between
ambition and experimental resources

For fundamental studies where you want to know concentration versus position and
time, use diffusion coefficients

For practical problems where you want to use one experiment to tell how a similar one
will behave, use mass transfer coefficients

The former approach is the distributed-parameter model used in chemistry, and the
latter is the lumped-parameter model used in engineering

Both approaches are used in medicine and biology, but not always explicitly.

8
Concentrations
The concentration of particular species is expressed in variety of ways

In mass transfer operation, the concentration gradient is the driving force when
other driving forces (temperature, pressure gradients, etc) are kept constant

 The actual driving force for mass transfer (gradient of chemical potential
between two points) is a function of all external forces

The concentration gradients are generally expressed in terms of mass

concentration of component, molar concentration of component and mass or mole
fraction of species

9
Mass concentration
The mass concentration of species i is expressed as ρi

It is defined as the mass of i per unit volume of a multicomponent mixture i.e.,

mi
i = (5)
V
It has the same unit as density

Total mass concentration within a mixture is equal to overall density which can be
expressed as
n
=  i (6)
i=1

where n is the number of species in the mixture

10
Mass fraction
The mass fraction of species i is defined by the ratio of mass concentration of
species i to the total mass density

It can be expressed as

i i
wi = = (7)
n 
 i
i =1

From Eqns (6) and (7), it is shown that

n n 
 wi =  
i = 1 (8)
i 1 i 1

11
Molar concentration
The molar concentration of a component is denoted by Ci.

It is defined as moles of ith component per unit volume of mixture

The total concentration in the system can be obtained by summing up all

molar concentrations for all species which can be represented as

n
CT =  Ci (9)
i=1

For an ideal gas mixture the molar concentration of species i can be obtained
from the ideal gas law (PV = nRT) as

pi
Ci = (10)
RT 12
Molar concentration
The total concentration in the gaseous system can be represented by

1 n PT
CT =  p
RT i=1 i
=
RT
(11)

where PT is the total pressure of the system which is sum of partial pressures of
all components

To convert from mass to molar concentration, divide the mass concentration of

species i by the molar weight of that species

13
Mole fraction
The mole fraction of species i is found by dividing the molar concentration of
species i by total concentration in the system, and is expressed as

For liquids and solids:

Ci
xi = (12)
CT

For gasses:
pi
yi = (13)
PT

If it is summed over all species, then one can get

 xi  1 and  yi  1 (14) 14
Average velocities
In a system, several molecular species move with different average velocities,
therefore, a frame of moving reference must be chosen

The important moving references are mass average, molar average and
volume average velocities

15
Mass average velocity
The mass average velocity can be defined in terms of the mass concentration and
the velocity of species i based on fixed axis

It is expressed as

n
  i ui n
1
u = i =1
n
=   i ui
 i =1
(15)
 i
i =1

Where: ui = the linear velocity of the ith species in the concerned direction

16
Molar average velocity
The molar average velocity can be expressed by the expression analogous to the
mass average velocity

It can be represented by replacing mass concentration of species i, ρi, with

molar concentration of species i, Ci

n
 Ci ui
1 n n
U = i =1
n
= 
CT i =1
Ci ui =  yi ui (16)
 Ci i =1
i =1

17
Diffusion velocity
Diffusion velocity, uiD, defined in terms of Ji, is relative to molar average velocity,
U, and is defined as the difference between species velocity, ui, and the molar-
average mixture velocity
Ji
uiD = = ui  U (17)
Ci

When solving mass transfer problems involving net mixture movement (bulk flow),
fluxes and flow rates based on U as the frame of reference are inconvenient to
use

It is thus preferred to use mass transfer fluxes referred to stationary to

coordinates.

Thus:

 From Eqn (17), the species velocity relative to stationary co-ordinates is:

ui  U + uiD (18)
18
Volume average velocity
For experimental analysis the volume average velocity is important due to a fixed
system of constant volume

The volume average velocity can be expressed by

n
v=  v Ci u i (19)
i =1

where v is the partial molar volume of species i

19
Fluxes
The flux is defined as the rate of transport of species i per unit area in a direction
normal to the transport

It is expressed in kg/m2s, kmol/m2s, lbmol/ft2h

Three frames of reference (or coordinate systems) are commonly defined for
expressing the flux of diffusing species

Assume that in any frame of reference, there is an ‘observer’ who ‘observes’

or ‘measures’ the velocity or flux of species in a mixture

ui The observer is stationary

u ui - u The observer is moving with

the mass average velocity

U
The observer is moving with
ui - U the molar average velocity
20
Frames of reference for mass flux
The mass fluxes in the three frames of reference is expressed as

(i) Relative to a stationary observer

ni = iu i (20)
(ii) Relative to an observe moving with the mass average velocity

ii = i (u i  u ) (21)
(iii) Relative to an observe moving with the molar average velocity

ji = i (u i  U ) (22)

Where: i  mass concentration of species i in a solution, in kg/m 3

ui = the linear velocity of the ith species in the concerned direction
U = molar average velocity of a mixture; u = mass average velocity
of a mixture

21
Frames of reference for molar flux
The molar fluxes in the three frames of reference is expressed as

(i) Relative to a stationary observer

Ni = Ciu i (23)
(ii) Relative to an observe moving with the mass average velocity

Ii = Ci (u i  u ) (24)
(iii) Relative to an observe moving with the molar average velocity

J i = Ci (u i  U ) (25)

Where: Ci  molar concentration of species i in a solution, in kmol/m

3

ui = the linear velocity of the ith species in the concerned direction

U = molar average velocity of a mixture; u = mass average velocity
of a mixture
22
Example
One kmole of gas mixture at a total pressure of 250 kPa and 303 K contains 10%
CH4, 30% C2H6, and 60% H2 by volume. The absolute velocities of each species
are -10 m/s, -5 m/s, and 15 m/s, respectively, all in the direction of the z-axis

(a) Determine the molar average velocity, U for the mixture

(b) Evaluate JCH4
Solution
Mole fractions: CH4 = 10% = 0.1, C2H6, = 30% = 0.3, H2 = 60% = 0.6
n
 Ci ui n
U = i =1
n
=  yi ui
 Ci i =1
i =1
= (0.1)(-10) + (0.3)(-5) + (0.6)(15) = 6.5 m/s

Concentrations:
P 250
CT = = = 0.09924 kmol/m3 = 99.24 mol/m3
RT (8.314x303)

CCH 4  (0.1)(99.24)  9.924 mol/m3

 J CH 4 = CCH 4 (u CH  U )  (9.924)(-10  6.5)

4
=  163.746 mol/m 2s
Fick’s 1st law of diffusion (steady state)
Adolf Fick (1955) first described the molecular diffusion in an isothermal, isobaric
binary system of components A and B

Fick’s first law applies to steady state systems, where concentration gradient is
constant
Mass transfer of A occurs from the zone of high concentration to that of low
concentration, the driving force being the concentration difference between the
two zones

C1 C
slope is
z
concentration C Flux J
concentration
C2
C1
direction of C2
diffusion of z1 z2
z
z substance A When concentration C1 > C2,
flux is positive from z1 to z2
25
Fick’s 1st law of diffusion (steady state)
According to Fick’s idea of molecular diffusion, the molar flux of a species relative
to an observer moving with molar average velocity is proportional to the
concentration gradient in a certain direction

dCA
JA  (26)
dZ

dCA dx A
 J A =  D AB =  CT D AB (27)
dz dz
dC dx
J B =  D BA B =  CT D BA B (28)
dz dz
When diffusion occurs in one direction in a binary mixture of A & B , the molar rate
of transfer of A (or B), per unit area due to molecular motion is given by Eqn (27)
and (28):

Equations 27 and 28 are mathematical representations of Fick’s first law

26
Fick’s 1st law of diffusion (steady state)
Where:
 JA is the molar rate of diffusion per unit area (molm-2s-1)
 D (m2s-1) is the diffusivity of A (or B), a physical property which describes how
fast or slow an object diffuses
 CA & CB (moldm-3) are the molar concentrations; CT is the total molar
concentration
 d represents the words ‘change in’. It refers to subtracting the beginning value
from the final value of the symbol following the d
 z is the distance in the direction of diffusion;
 XA & XB are the mole fractions

The diffusional flux JA is a positive quantity by convention

Since diffusion occurs in the direction of decreasing concentration [(dCA/dz) < 0],
the negative sign is incorporated in Eqns (27) & (28) to make it consistent with
respect to sign 27
Fick’s 1st law of diffusion (steady state)
When diffusion occurs in more than one direction, Fick’s law has to be written in
vector form:

J =  DC (29)
dC dC dC
C means 'grad' C = i+ j+ k (in a cartesian co-ordinates)
dx dy dz

Essentially this is just 3

equations rolled into one

28
Molar flux NA in a stationary frame of reference
Fick’s law given by Eqn 27 (and/or 28) expresses the molar flux JA with respect to
an observer moving with the molar average velocity

In practice, however, the molar flux NA in a stationary frame of reference is more

useful

An expression for NA can be developed by using Eqns (27) & (25)

dCA
J A =  DAB (27)
dz

J i = Ci (u i  U ) (25)

 J A = C A (u A  U ) (30)

Fluxes of A & B with respect to a stationary frame [From Eqn 23] are:

N A = C Au A and N B = C BuB (31) 29

Molar flux NA in a stationary frame of reference
And molar average velocity [From Eqn 16] is:
1 n
U = 
CT i =1
Ci ui =
1
CT

C Au A  C B u B  (16)

Therefore, Equating Eqns (27) & (30) gives:

dCA
J A =  DAB = C A (u A  U ) (32)
dz
dCA
  DAB = C A (u A  U )
dz
= C Au A  C AU
 1 
= N A  CA   C Au A  C B u B  
 CT 
C
= N A  A  N A  NB 
CT

CA dCA
 NA =  N A  NB   DAB (33) 30
CT dz
Molar flux NA in a stationary frame of reference
The flux NA can be viewed as consisting of two terms as follows:

The term representing bulk flow, i.e.

C
 N A  NB  A (34)
CT

The term representing molecular diffusion, i.e.

dCA
 DAB (35)
dz

If the concentration of A in a mixture is small (dilute solution), the contribution of

the bulk flow term in Eqn (34) becomes small too

In such a case:

dCA
N A  J A  D AB (36)
dz

31
Molar flux NA in a stationary frame of reference
For gas phase diffusion, Eqn (33) may be written in terms of partial pressure too

If the gas phase is assumed to behave ideally, i.e.

p P
C A  A and C  ,
RT RT
where T is the uniform temperture of the gas (in K)

p A DAB dpA
 NA =  N A  NB   (37)
PT RT dz

32
Fick’s 2nd law of diffusion (unsteady state)
Fick’s first law applies to steady state systems, where concentration gradient is constant

In many cases of diffusion, the concentration changes with time, therefore, Fick’s second
law must be used which is:

dC d 2C
D (one dimensional diffusion) (38)
dt dz 2

or

dC
= D 2 C (multi-dimensional diffusiion) (39)
dt

d 2C d 2C d 2C
where  C =2
2
i+ 2
j+ 2
k (in a cartesian co-ordinates)
dx dy dz

33
Derivation of Fick’s 2nd law of diffusion
One-dimensional diffusion
Consider an elementary volume (or box) with dimensions dx, dy, dx with a
concentration gradient in the z direction

Jz+dz Flux due to diffusion into the box =J Z per unit area
Flux due to diffusion out of the box =J z+dz per unit area

dy Amount of substance in the box = CdV

where: dV = volume of box = dxdydz
C = concentration of substance in the box
Jz
d (CdV)
Rate of change of amount of substance in the box =
dt
This is equal to difference in diffusional fluxes across boundaries

d (CdV)
  J z dxdy  J z +dz dxdy (40) 34
dt
 Derivation of Fick’s 2nd law of diffusion
Rate of change = difference in diffusional fluxes

d (CdV)
  J Z dxdy  J Z+dZ dxdy (40)
dt

Dividing through by volume (which is constant) we obtain:

dC J z  J z+dz  dJ z
   (41)
dt dz dz
(taking limits as dz  0)

dC
From the Fick's first law: J z =  D ;
dz
dJ z d 2C
  D 2 (first derivative of Fick's first law)
dz dz
dJ z
Combining Eqn (41) with the first derivative of Fick's law to eliminate yields:
dz
dC d 2C
D 2 (38)
dt dz 35
Mass, heat and momentum transfer analogies
The various unit operations can be classified into three fundamental transfer (or transport)
processes: momentum transfer, heat transfer, and mass transfer
The fundamental process of momentum transfer occurs in such units operations as fluid
flow, mixing sedimentation, and filtration. Heat transfer occurs in conductive and convective
transfer of heat, evaporation, distillation, and drying. The third fundamental transfer process,
mass transfer, occurs in distillation, absorption, adsorption, drying, liquid-liquid extraction,
and membrane processes

When mass is being transferred from one distinct phase to another or through a single
phase, the basic mechanisms are the same whether the phase is a gas, liquid or solid. This
is the case in heat transfer, where the transfer of heat by conduction follows Fourier’s law in
a gas liquid or solid. In mass transfer, the transfer process is governed by Fick’s law
(Equation 25)

All the three molecular transport processes are characterised by the same general type of
equation
driving force
rate of transfer = (42)
resistance
Refer to question 2 of tutorial 1 36
Diffusion in the gas phase
(1) Equimolecular counter diffusion

(2) Diffusion through a stationary gas

37
Equimolecular counter diffusion
Assume two gases A & B at constant pressure are diffusing at equal and opposite rates

A
B

The total flux of A (= NA) is made up two components, namely that resulting from the bulk
flow of A (NXA) and that resulting from molecular diffusion JA:
Where:
N = NA + NB
XA is the fraction of A in the binary mixture

Therefore, the total molar flux NA, for a binary system at constant temperature and pressure
is described by:
 N A = NX A + J A
C dCA  CA 
 N A = (N A +N B ) A  D AB ;  where = XA  (43)
CT dz  C T 
Equimolar counter diffusion is given by N A =  N B , which reduces Eqn (43) to:

dCA
 N A =  D AB (44) 38
dz
Equimolecular counter diffusion
For steady state diffusion, Eqn (44) may be rearranged and integrated using the
boundary conditions as:
Z2 CA2

 N A  dz =  D AB  dC A (45)
Z1 CA1

D AB
 NA =  CA1  CA2  (46)
Z2  Z1
For ideal gases
nA p P y
 CA =  A = T A (47)
V RT RT

Therefore, Eqn (46) becomes

D AB D AB
 NA =  A1 A2 
p  p  PT  yA1  yA2  (48)
RT  Z2 -Z1  RT  Z2 -Z1 

Eqn (46) & (48) are equations describing the molar flux for steady-state equimolar
counter diffusion Similarly, Eqn (46) & (48) apply to NB which is equal to -NA 39
Example
A binary gaseous mixture of components A & B at a pressure of 1 bar & temperature of
300 K undergoes steady-state equimolar counter diffusion along a 1-mm thick diffusion
path. At one end of the path the mole fraction of component A is 70%, while at the
other end it is 20%. Under these conditions DAB = 0.1cm2/s. Calculate the molar flux of
component A.

40
Solution
Equation (48) applies

DAB = 0.1cm2/s = 1.0x10-5 m2/s ;

PT = 1bar = 1.0 x 105 Pa;
z2 - z1 = 1mm = 0.001m;
R = 8.314 Pa.m3/mol.K;
T = 300K;
yA1= 0.7, yA2 = 0.2

D AB PT (y A1  y A2 )
 NA =
RT (z 2  z1 )

(1.0x10-5 ) (1.0x105 )(0.7  0.2)


(8.314)(300) (1x10-3 )

 0.20 mol/m 2 .s
41
Diffusion through a stationary gas
The case of diffusion of A through stagnant or non-diffusing B at steady state often occurs.
In this case one boundary at the end of the diffusion is impermeable to component B, so it
can not pass through.

An example shown in the figure below occurs in the absorption of NH3 (A) vapour which is in
air (B) by water. The water surface is impermeable to air, since air is only very slightly
soluble in water. Thus, since B cannot diffuse, NB = 0.

Fig. 1. Diffusion of A through stagnant, non-diffusing B

42
Diffusion through a stationary gas
The derivation of the case for A diffusing in stagnant, non diffusing B is as follows

Fig. 3. Mass transfer through stationary gas B

The figure shows a system where the bulk of A & B flow towards a surface that is absorbing
component A only. The absorption of component A on the surface will create a partial
pressure gradient causing component A to diffuse towards the surface and component B
away from it. As the process continues, A being absorbed & B diffusing away, a total
pressure gradient will result causing a bulk fluid motion towards the surface. However, as B
is diffusing away on one hand and moving with the bulk towards the surface (in the opposite
direction), there is no net motion for B; the bulk motion must exactly balance the diffusive
43
motion.
Diffusion through a stationary gas
The rates of diffusion of A & B are given previously by Fick’s law as

dCA D dPA
JA =  D  . (49)
dz RT dz
dCB D dPB
JB =  D  . (50)
dz RT dz
dCA D dPA
J B = +D + . (51)
dz RT dz
The total mass transfer rate of B, NB = 0, & there must be a bulk flow of the system away
from the absorbing surface exactly to counterbalance the diffusional flux towards the surface
(Fig. 2)

dCA
Bulk flow of B =  D (52)
dz

Bulk flow of A & B are related to their concentrations as follows:

Bulk flow of A C
= A (53)
Bulk flow of B CB
44
Diffusion through a stationary gas
Therefore, the corresponding bulk flow of A must be CA/CB times bulk flow of B, since bulk
flow implies that the gas moves en masse (moves together)
dCA C A
Bulk flow of A =  D (54)
dz C B
The total flux of A, NA, is therefore given by:
N A = diffusional flux + bulk flow
dCA dC A C A
NA =  D  D
dz dz C B
dCA CT
= D (55)
dz C B
Equation (55) is known as Stefan’s law of diffusion

Writing Equation (55) as

CT dCB
NA = D (56)
C B dz
On integration for the whole system between point 1 & 2:
DCT C 
NA = ln  B2  (57) 45
z2  z1  CB1 
Diffusion through a stationary gas
By definition, CBM, the logarithmic mean of CB1 & CB2, is given by
C B2  CB1
C BM = (58)
ln(C B2 / C B1 )
Thus:
D CT
NA =  CB2  CB1 
z2  z1 CBM
D CT
= (C A1  C A2 ) (59)
(z 2  z1 ) C BM

In terms of mole fraction:

DCT 1
NA = (x A1  x A2 ) (60)
(z 2  z1 ) x BM

In terms of partial pressure

D PT D PT
NA = (p A1  p A2 ) = PT (y A1  y A2 ) (61)
RT(z 2  z1 ) p BM RT(z 2  z1 ) p BM

D PT2
= (y A1  y A2 )
RT(z 2  z1 ) p BM 46
Example
O2 (A) is diffusing through CO (B) under steady state conditions, with CO nondiffusing. The
total pressure is 1 X 105 N/m2, and temperature is 0°C. The partial pressure of O2 at two
planes 2.0 mm apart is 13000 & 6500 N/m2, respectively. The diffusivity for the mixture is
1.87x10-5 m2/s. Calculate the rate of diffusion of O2 in kmol/s through each square metre of
the two planes. Assume that the system is isothermal.

47
 Solution
Equation (61) applies

DAB = 1.87x10-5 m2/s O2;

PT = 105 N/m2 ;
z2-z1 = 0.002m;
R = 8314 N.m/kmol.K,
T = 0°C = 273K;
pA1= 13000, pA2 = 6500, pB1= 105-13000=87x103, pB2= 105-6500=93.5x103, all in N/m2

pB2  pB1 (93.5  87)(103 )

pBM = = = 90 200 N/m2
ln(p B2 /p B1 ) ln(93.5/87)

D PT
 NA = (p A1  pA2 )
RT(z 2 -z1 ) pBM

(1.87x10-5 ) (105 )
= 3
(13  6.5)(10 3
)
8314(273)(0.002) (90.2x10 )

48
= 2.97x10-5 kmol/m 2 .s
Diffusion in liquids
The rate of diffusion of solution in a liquid is governed by the same equations as for the gas
phase, employing molar concentration terms instead of partial pressures.

Fick’s law
dCA
J A =  DL (62)
dz
Where DL is the liquid diffusivity

Integrating Equation (62) for the equimolar counter diffusion system gives
C A2  C A1 C  CA2
N A =  DL = D L A1 (63)
z2  z1 z2  z1
Where CA1 & CA2 are molar concentrations of A at two positions z1 & z2,
respectively

Alternatively,
x A2  x A1 x  x A2
N A =  D L ( /M) av = D L ( /M)av A1 (64)
z2  z1 z2  z1

Where (ρ/M)av is the average molar concentration of the liquid 49

Diffusion in liquids
For diffusion through a stagnant liquid B:

DL CT
NA =   CA2  CA1  (65)
(z 2  z1 ) CBM

Where CA1, CA2 are molar concentrations of A at two points; CT is the total molar
concentration; z2-z1 is the equivalent thickness of the liquid film through which diffusion is
taking place; and CBM is the logarithmic mean of the molar concentrations of B on each side
of the liquid film

Alternatively,

DL ( /M) av
NA =   x A2  x A1  (66)
(z 2  z1 ) x BM

Where (ρ/M)av is the average molar concentration of the liquid

50
 Diffusion in solids
Steady state
substance A at
substance A at concentration CA2
concentration CA1
direction of diffusion
of substance A

z
 Fick’s law is applied in the form used for fluids

dC A
J A =  D AB (67)
dz
 If DA is constant, integration of Fick’s law equation for diffusion through a flat slab results in

D A (CA1 - CA2 )
JA= (68)
z
which is similar for expressions obtained for fluids. CA1 and CA2 are concentrations at
opposite sides of the slab

 For other solid shapes, an appropriate average cross sectional area SAV is applied.
Diffusion in solids
 For other solid shapes, an appropriate average cross sectional area SAV is applied.

D ASAV (CA1 - CA2 )

JA= (69)
z
 cross sectional area SAV for radial diffusion through a solid cylinder of inner & outer radii
a1 & a2, respectively, & of length L,
2πL(a 2 -a1 )
SAV = and z = a 2 - a1 (70)
ln( a 2 a1 )

 For radial diffusion through a spherical shell of inner & outer radii a1 & a2,

SAV = 4πa1a 2 and z = a 2 - a1

 Refer to example 4.1 in Treybal

52
Unsteady-state diffusion
 Since solid are not easily transported through equipment as fluids, unsteady-state
diffusional conditions arise much more frequently

 Where there is no bulk flow, and in the absence of a chemical reaction, Fick’s second law
can be used to solve problems of unsteady-state diffusion by integration with appropriate
boundary conditions.

 Fick’s second law is derived from the continuity equation (Navier-Stokes equation) in a
special case where the velocity equals zero, and there is no chemical reaction

C A   2C A  2C A  2C A 
 DAB    2 
(71)
   x 2
 y 2
 z 
 Several examples are available where Fick’s second law is applied

 Diffusion from a slab with sealed edges

 Diffusion from a rectangular bar with sealed ends

 Diffusion from a sphere

 etc, etc
 Diffusion from solids (e.g., slab) with sealed edges
 Consider a slab of thickness 2a, with sealed edges on four sides, so that diffusion can only
take place only toward and from the flat parallel surface
average Conc of average conc of
substance A inside substance A at
initially (t=0) is CA0 surface (t=∞) is CA ∞
direction of diffusion of
average conc of substance A substance A
inside over time (t=θ) is CAθ
2a
 If diffusion were allowed to continue indefinitely, the concentration would fall to the uniform
value CA ∞ , therefore, CA 0 – CA ∞ is a measure of the amount of solute removed. If diffusion
were stopped at time θ, CA θ - CA ∞ is a measure of the amount of solute unremoved

 The fraction unremoved, Ea, is given by integration of Fick’s second law as

CA   CA   D   a 
Ea   f  2   1  erf   (72)
CA 0  CA   a   2 D  
 The above relationship is given graphically in Figure 4.2 of Treybal

 Similarly, Ea, can be derived for several other shapes and circumstances (e.g., sphere,
cylinder)
Types of solid diffusion
 The structure of the solid and its interaction with the diffusing substance have a profound
influence on how diffusion occurs

Diffusion through polymers

 Diffusion resembles that through liquid solutions

gas at partial
gas at partial
pressure PA2
pressure PA1
direction of diffusion of
gaseous species

z
 The gas dissolves in the solid at the faces exposed to the gas in accordance with Henry’s
law, i.e., concentration directly proportional to pressure

 Since at the two faces of the membrane the equilibrium solubility of the gas in the polymer
is proportional to the pressure, Fick’s law becomes

D A s A (PA1  PA2 )
JA = (73)
z
 The permeability, P, of the membrane for the gas is defined as

P = DAsA (74)
Example
Calculate the rate of diffusion of CO2 through a membrane of valcanised rubber 1 mm thick at
25°C if the partial pressure of CO2 is 1 cm Hg on one side and zero on the other. Calculate
also the permeability of the membrane for CO2. At 25°C the solubility coefficient is 0.90 cm3
gas (stp)/cm3 atm. The diffusivity is 1.1x10-10 m2/s.
Solution
(i) Obtain the solubility coefficient in terms of cm Hg as ratio of 76 cm Hg standard

Since 1 std atm = 76 cm Hg press, the solubility coefficient in terms of cm Hg is

0.90/76 = 0.01184 cm3 gas (stp)/cm3.cm Hg = SA, z = 0.1 cm

(i) Given the thickness, and the two pressures apply the equations to calculate the rate of
D A s A (p A1 -p A2 )
diffusion JA= and permeability P = DAs A
z

PA1 = 1 cm Hg, PA2 = 0

DA = (1.1x10-10)(104) = 1.1x10-6 cm2/s

(1.1x10- 6 )(0.01184)(1.0-0)
JA = =0.13x10- 6 cm3 (STP)/cm 2 .s
0.1

P = (1.1x10- 6 )(0.01184) = 0.13x10- 7 cm3 (STP)/cm 2 .s.(cmHg/cm)

57
Types of solid diffusion
Diffusion in porous solids

 The prediction and correlations of fluxes for diffusion of fluids through porous solids, such
as catalysts and adsorbents, are of considerable importance.

 transport of reactant molecules through the catalyst pores to the active catalyst
surfaces

 transport of molecules through adsorbent pores

 Most transport in a porous solid is complex

 Several mechanisms of transport (molecular diffusion, Knudsen diffusion, viscous and slip
flow, surface migration) may contribute to the flux of fluid through the pores of the solid

 Most models for prediction of gaseous diffusion and flow in porous solids are based on well
developed theories for diffusion and flow in capillaries
Diffusion and flow in capillaries
 Consider a single cylindrical capillary of radius r and length L through which steady-state
transport of components A and B of a binary gas mixture occurs. The gas mixture at ends
of the capillary are maintained at constant pressures P1 and P2, and constant compositions
y1 and y2, respectively
L
P1 P2
yA1 r yA2
yB1 yB2

 If the radius of the capillary and the gas pressure is such that the mean free path λ
(distance between molecular collisions) is large compared to the diameter d of the pores
(d/ λ < 0.2), the rate of transport of the molecules A and B is governed only by the
collisions with the capillary wall. This type of transport is referred to as Knudsen diffusion
 Diffusion and flow in capillaries
 Knudsen showed that
dC A  2-f 
J A =  D KA , where D KA = 2/3rvA   = Knudsen diffusion coefficient (75)
dz  f 

 The mean molecular velocity is given by the kinetic theory of gases as

 8g RT 
vA =  c  (76)
  MA 
 The fraction f of the molecules, which undergo diffuse reflections at the walls is usually taken
to be zero, thus the Knudsen coefficient is
1/2
d  8g RT 
D KA =  c  (77)
3   MA 
 At steady-state, Equation 75 is integrated to give Equation (78) and (79) for A and B,
respectively DK , A ( PA1  PA 2 )
NA  (78)
RTL
DK , B ( PB1  PB 2 )
NA  (79)
RTL
 At constant pressure, Equation (78) and (79) add to give

JA J
+ B =0 (80)
D KA D KB
 Diffusion and flow in capillaries
1/2
d  8g c RT 
D
 By use of , KA =   the ratio of fluxes is given by
3   MA 

1/2
NA M 
= - B  (81)
NB  MA 

 If d/ λ is greater than approximately 20, ordinary molecular diffusion predominates

NA DAB , eff pT N /( N A  N B )  y A 2
NA  ln A (82)
N A  NB RTL N A /( N A  N B )  y A1

 In the range d/ λ from 0.2 to 20 (transition range), both molecular and Knudsen diffusion
have influence (use Treybal Eqn.4.23)

NA DAB , eff PT N A /( N A  N B )(1  DAB , eff f / DKA )  y A 2

NA  ln (83)
N A  NB RTL N A /( N A  N B )(1  DAB , eff / DKA )  y A1
Diffusion and flow in capillaries

 Since length of various pores and their cross sectional area are not constant, flux is based
on gross external surface of the membrane or pellets, and the length on a readily
measured distance y such as membrane thickness, pellet radius, and the like

 Since the fluxes will be smaller than the true values based on the true fluid cross section
and length , an effective diffusivity, DAB, eff, smaller than true DAB must be used

 For the same types of processes DAB / DAB, eff in a given solid, the ratio will be constant,
and once measured can be applied to all solutes

 For a given solid DAB,eff / DK,A,eff = DAB / DK,A

 The mean free path can be estimated as

0.5
3.2  RT 
   (84)
PT  2 gC M 
Example (Refer to Treybal: Illustration 4.5

63
Diffusion with a difference in absolute pressure
Hydrodynamic flow of gases
 If there is a difference in absolute pressure across a porous solid, a hydrodynamic flow of
gas through the solid will occur

 The flow of gas through the solid pores (or capillaries) may be laminar or turbulent
depending on whether Reynold’s number is below or above 2100

 For a single gas at low velocities (laminar flow), the flow is described by Poiseuille’s law
for a compressible fluid obeying the ideal gas law

d 2 gc
NA  pT ,av ( pT 1  pt 2 ) (85)
32 RTL

where :
pT 1  pT 2
pT ,av  (86)
2
 Assumes entire pressure difference is the result of friction in the pores and ignores
entrance and exit losses and kinetic energy effects
64
Diffusion with a difference in absolute pressure
 Since the pores are neither straight nor of constant diameter, just as diffusive flow, NA is
best based on external cross section of the solid (use Treybal, Eqn.4.26)

 If the conditions for pore diameter and pressure occur for which Knudsen flow prevails (d/λ
> 0.2) then the flow will be described by Knudsen’s law

65
Example (Refer to Treybal: Illustration 4.6)

66