14-Oct-21

General Case for Diffusion of Gases A & B Plus Bulk Movement

◼ Up to now we have considered Fick's law for diffusion in a stationary fluid;
that is, there has been no net movement or convective flow of the entire
phase of the binary mixture A and B.
◼ What happens if the whole fluid is moving in bulk or convective flow? To
a stationary observer, A is moving faster than the bulk of the phase since its
diffusion velocity is added to that of the bulk phase velocity .

Expressed mathematically: Expressed pictorially:

A
A = Ad + M
Ad M
A: velocity of A relative to the stationary point
Ad: diffusion velocity
M: molar average velocity of the mixture

18

18

The flux of species defined with reference to fixed spatial coordinates, NA is

𝑁𝐴 = 𝐶𝐴 𝑣𝐴
kg mol A m kg mol A
𝑐𝜐 ≡ flux
Substituting by A = Ad + M : m3 s m2 𝑠
𝑁𝐴 = 𝐶𝐴 Ad +𝐶𝐴 M

By definition, the molar average velocity of the mixture M

σ 𝐶𝑖 𝑣𝑖 ,C: total concentration of the mixture
𝑣𝑀 = 𝐶
𝐶 & 𝐶𝑖 is the mole fraction of i in the mixture

By Substitution: 𝐶𝐴
𝑁𝐴 = 𝐶𝐴 Ad + ෍ 𝐶𝑖 𝑣𝑖
𝐶
𝐶𝐴
𝑁𝐴 = 𝐶𝐴 Ad + 𝐶
𝐶𝐴 𝑣𝐴 + 𝐶𝐵 𝑣𝐵
𝐶𝐴
∴ 𝑁𝐴 = 𝐶𝐴 Ad + 𝑁𝐴 + 𝑁𝐵
𝐶
The first term on the right-hand side of this equation is the diffusional molar flux
of A (J*A: molecular diffusion), and the second term is the flux due to bulk
motion (convection term). 19

19

1
 14-Oct-21

Substituting:
cA 𝑑𝑐𝐴 𝑐𝐴
N A = J A* + ( N A + N B ) or: 𝑁𝐴 = −𝐷𝐴𝐵
𝑑𝑧
+ (𝑁𝐴 + 𝑁𝐵 )
𝑐
c
cB 𝑑𝑐𝐵 𝑐𝐵
N B = J B* + ( N A + N B ) or: 𝑁𝐵 = −𝐷𝐵𝐴
𝑑𝑧
+ (𝑁𝐴 + 𝑁𝐵 )
𝑐
c

Notes:

✓ The above equations are valid for diffusion in a gas, liquid, or solid

✓ To solve the above equation, the relation between the flux NA and NB must be
known.

✓ For equimolar counterdiffusion: NA = - NB, the convective term become zero

N A = J A* N B = J B*

20

20

Special case: A is Diffusing Through Stagnant or Nondiffusing B

The case of diffusion of A through stagnant or nondiffusing B at steady state
often occurs. One boundary at the end of the diffusing path is impermeable to
component B (NB = 0)
▪ Evaporation of pure liquid such as
2 pA2. air (B)

benzene (A) at the bottom of narrow tube, NA

and a large amount of air (B) passed over
the top.
- Benzene vapor (A) diffuses 1 pA1 .
liquid
through the air (B) in the tube
benzene (A)
- But at point 1, air (B) is insoluble
in liquid benzene (NB = 0)
. 1 NH3 (A)
Air (B)
▪ Another example: Removal of NH3 from air/NH3 NA
mixture by absorption
- Water surface is impermeable to air (B), since . 2
air is very slightly soluble in water (NB = 0) Liquid water
- NH3 can dissolve in water
21

21

2
 14-Oct-21

Derivation of expression for NA when NB = 0:

𝑑𝑐𝐴 𝑐𝐴 0
𝑁𝐴 = −𝐷𝐴𝐵 + (𝑁𝐴 + 𝑁𝐵 )
𝑑𝑧 𝑐
c = P/RT, cA = pA/RT, and cA/c = pA/P:
DAB dp A p A
NA = − + NA
RT dz P
Rearranging and integration:
 p 
 
z2 DAB pA2 dp A
D dp
N A 1 − A  = − AB A  N A dz = −
 P RT dz RT  pA 
1 − 
z1 p A1

DAB P  P − p A2  𝐷𝐴𝐵 𝑃 𝑝𝐵2  P

NA = ln  = ln
RT ( z 2 − z1 )  P − p A1  𝑅𝑇(𝑧2 − 𝑧1 ) 𝑝𝐵1

To make it easier, it is usually expressed in terms of log mean value of inert B:

pB 2 − pB1 P = pA1 + pB1 = pA2 + pB2  pB1 = P - pA1
pBM =
ln( pB 2 / pB1 )  pB2 = P - pA2
(𝑃 − 𝑝𝐴2 ) − (𝑃 − 𝑝𝐴1 ) 𝐷𝐴𝐵 𝑃
= 𝑁𝐴 = 𝑃 − 𝑃𝐴2
ln[ (𝑃 − 𝑝𝐴2 )/(𝑃 − 𝑝𝐴1 )] 𝑅𝑇 𝑧2 − 𝑧1 𝑃𝐵𝑀 𝐴1
22

22

Example: Diffusion of water Through Stagnant, Nondiffusing Air

Water in the bottom of a narrow tube is held at a constant temperature of 293 K.
The total pressure of air (assumed dry) is 1.01325 × 105 Pa and the temperature is
293 K. Water evaporates and diffuses through the air in the tube, and the diffusing
path z2 – z1 is 0.1524 m long. Calculate the rate of evaporation at steady state
kgmol/s.m2 for short times. The diffusivity of water vapor at 293 K and 1 atm
pressure is 0.250 × 10-4 m2/s. Assume that the system is isothermal. The saturation
vapor pressure of water at 293 K is 2.34 × 103 Pa atm
𝐷𝐴𝐵 𝑃
𝑁𝐴 =
𝑅𝑇 𝑧2 − 𝑧1 𝑃𝐵𝑀 𝐴1
𝑃 − 𝑃𝐴2
pA22 .
air (B)

z NA
z2-z1

1 . pA1
water (A)

23

23

3
 14-Oct-21

Example: Diffusion in a Tube with Change in Path Length (boundaries

move with time)
Diffusion of water vapor in a narrow tube is occurring as in the last example under
the same conditions. However, at a given time t, as diffusion proceeds, the level
drops (i.e the level z from the top increases) slowly. Derive the equation for the time
tF for the level to drop from a starting point of z0 at t = 0 to zF at t =tF.

𝑁𝐴 =
𝐷𝐴𝐵 𝑃
𝑃 − 𝑃𝐴2
𝑅𝑇𝑧 𝑃𝐵𝑀 𝐴1 z0
. air (B)

- The level is dropping very slowly

- Then we can assume pseudo-steady state z NA
zF
condition (i.e. the time scale over which the level
changes is large compared with the time scale
over which the diffusion process from the surface
. water (A)
reaches steady state.)
- Note that the above equation is derived
assuming steady state (used the steady state form
of Fick’s law). This means that at any time t, the
above equation holds, but z2 - z1 = z is now a
variable and hence NA is changing function of z
If the assumption of pseudo-steady state condition does not apply (i.e. The level
is dropping very quickly for example, we then have to use Fick’s second law
𝜕𝐶(𝑧,𝑡) 𝜕2𝐶(𝑧,𝑡) 24
( =𝐷 ), which is much harder to solve (partial differential equation)
𝜕𝑡 𝜕𝑧2

24

The molar flux NA is related to the amount of A leaving the liquid by the material
balance on the deferential element on the liquid surface:
(𝐴𝑑𝑧)𝜌𝐿𝐴 = 𝑁𝐴 × 𝐴 × 𝑀𝐴 × 𝑑𝑡 dz
𝜌𝐿𝐴 𝐴𝑑𝑧 A
𝑁𝐴 =
𝐴𝑀𝐴 𝑑𝑡
𝐷𝐴𝐵 𝑃 𝜌𝐿𝐴 𝑑𝑧
𝑅𝑇𝑧 𝑃𝐵𝑀 𝐴1
𝑃 − 𝑃𝐴2 =
𝑀𝐴 𝑑𝑡
z0
. air (B)

𝑡𝐹 𝑧𝐹 NA
𝐷𝐴𝐵 𝑃 𝜌𝐿𝐴 z
𝑃 − 𝑃𝐴2 න 𝑑𝑡 = න 𝑧𝑑𝑧 zF
𝑅𝑇 𝑃𝐵𝑀 𝐴1 0 𝑀𝐴 𝑧0
. water (A)
𝜌𝐿𝐴 𝑧𝐹 2 − 𝑧0 2 𝑅𝑇 𝑃𝐵𝑀
𝑡𝐹 =
2𝑀𝐴 𝐷𝐴𝐵 𝑃 𝑃𝐴1 − 𝑃𝐴2

The method can be used to determine DAB experimentally (Winkelmann's method).

In this experiment, the starting path length z0 is measured at t = 0 and z is measured
with time. From the slope, DAB can be calculated
2𝑀𝐴 𝐷𝐴𝐵 𝑃 𝑃𝐴1 − 𝑃𝐴2 25
𝑧 2 − 𝑧0 2 = 𝑡
𝜌𝐿𝐴 𝑅𝑇 𝑃𝐵𝑀
25

4
 14-Oct-21

Diffusion Through a Varying Cross-Sectional Area

So far, we only considered the cases of steady state where the flux of mass
transfer is constant in the integration. In these cases, the cross-sectional area A,
through which the diffusion occurs, has been constant with varying distance.
However, in some situations the area A may vary.

❑ Examples
▪ Evaporation of a drop of liquid
▪ Evaporation of a ball of naphthalene
▪ Diffusion through a pipe with non-uniform cross sectional area

𝑁𝐴
▪ The rate of mass transfer (𝑁𝐴) is related to the flux (𝑁𝐴) by 𝑁𝐴 =
𝐴

26

26

Diffusion from a Sphere (ignoring changes in the boundaries)

Derivation of expression for the concentration

r1 NA
profile and molar flux for diffusion of A through
pA2
a spherical shell of stagnant B (NB = 0). r dr
pA1
+ (N A + N B )
DAB dp A p A 0
NA = −
RT dr P

 p  D dp
N A 1 − A  = − AB A Take care: NA in this case changes with r
 P RT dr
NA D dp A
NA = = − AB
4 r 2
RT  p 
1 − A dr
 P 
r2 pA2
NA dr D dp A N A  1 1  DAB P P − p A 2
4  r 2
= − AB
RT 
p A1 1 −
pA   − =
4  r1 r2  RT
ln
P − p A1
r1
 
 P 

27

27

5
 14-Oct-21

pB 2
ln
4 DAB P pB1
NA =
RT  1 1 
 − 
 r1 r2 

pB2 − pB1 4𝜋𝐷𝐴𝐵 𝑃 (𝑝𝐴1 − 𝑝𝐴2 )

Letting pBM = 𝑁𝐴 =
𝑅𝑇 𝑃𝐵𝑀 1 1
ln( pB 2 / pB1 ) 𝑟1 − 𝑟2

1 4𝜋𝐷𝐴𝐵 𝑃 𝑟1
If 𝑟2 >> 𝑟1 , ≅0 𝑁𝐴 = (𝑝𝐴1 − 𝑝𝐴2 )
𝑟2 𝑅𝑇 𝑃𝐵𝑀

The flux of evaporation on the sphere surface can be determined

𝑁𝐴 𝐷𝐴𝐵 𝑃
𝑁𝐴1 = = 𝑝 − 𝑝𝐴2
4𝜋𝑟1 2 𝑅𝑇𝑃𝐵𝑀 𝑟1 𝐴1

28

28

The flux of evaporation on the sphere surface derived so far is

𝑁𝐴 𝐷𝐴𝐵 𝑃
= 𝑁𝐴1 = 𝑝 − 𝑝𝐴2
4𝜋𝑟1 2 𝑅𝑇𝑃𝐵𝑀 𝑟1 𝐴1

This equation can be simplified further. If pA1 is small compared to P (a dilute

gas phase) → PBM ≅ P. (take care that pA2 << pA1)

Also, setting 2r1 = D1 (diameter), and cA1 = pA1 /RT

This equation can also be used for any fluid,

NA
= N A1 =
2 DAB
(c A1 − c A2 ) where DAB is the diffusivity of A in the
4 r1
2
D1
liquid

29

29

6
 14-Oct-21

Example: Evaporation of a Naphthalene Sphere

A sphere of naphthalene having a radius of 2 mm is suspended in a large volume of
still air at 318 K and 1 atm. The surface temperature of the naphthalene can be
assumed to be at 318 K and its vapor pressure at 318 K is 0.555 mmHg. The DAB of
naphthalene in air at 318 K is 6.92 × 10-6 m2/s. Calculate the rate of evaporation of
naphthalene from the surface.

NA
= N A1 =
DAB P
( p A1 − p A2 )
4 r1
2
RT PBM r1

pB1=P-pA1=(1 – 0.555/760) ×1.013×105=1.01226×105 Pa

pB2=P-pA2=(1 – 0) ×1.013×105 =1.013×105 Pa
pBM= 1.01263×105 Pa (we can use the arthimatic average instead of the log
𝑝𝐵1 +𝑝𝐵2
average since pB1 and pB2 are very close, 𝑝𝐿𝑀 ≅ )
2
NA 6.92  10-6  1.013 105 0.555 1.013 105 − 0
= N A1 = = 9.6847 10-5 mol/m 2 .s  N A = 4.868 10-9 mol/s
4 r1
2
8.314  318  2 1000 1.01263 105
30

30

Example: Time needed for the sphere to completely evaporate

(boundaries change with time)
For the same previous example, calculate the time needed for the sphere to
evaporate completely (Problem Set 1 #8)

DO IT YOURSELF!
(Take care that the r is changing with time and thus the rate of MT
Take care also that that r is measured from the center of the sphere and with time
the change dr is negative (i.e. as m evaporates, r decrease))

31

31

7
 14-Oct-21

Multi-component Diffusion of Gases

Multicomponent component diffusion occurs when three or more components A, B, C, …
are present at constant total pressure.
Example: diffusion of A through a stagnant nondiffusing mixture of B, C, D, …

For NB = NC = … = 0
D Am P
NA = ( p A1 − p A 2 ) (Derived using Stefan-Maxwell method)
RT ( z 2 − z1 ) piM
pi 2 − pi1
where piM is the log mean of pi1 = P – pA1 and pi2 = P – pA2 piM =
ln( pi 2 / pi1 )
1
DAm =
Also, '
x / DAB
B + xC' / DAC +   

where x’i = mol i/mole inerts (non-diffusing):

xB xC
xB' = , xC' = 
1 − xA 1 − xA

32

32

Example: Diffusion of A Through Nondiffusing B & C

At 298 K and 1 atm total pressure, methane (A) is diffusing at steady state through
nondiffusing argon (B) and helium (C). At z1 = 0, the partial pressures in atm are pA1
= 0.4, pB1 = 0.4, and pC1 = 0.2, and at z2 = 0.005 m, pA2 = 0.1, pB2 = 0.6, and pC2 =
0.3. The binary diffusivities are DAB = 2.02 × 10−5 m2/s, DAC= 6.75 × 10−5 m2/s, and
DBC = 7.29 × 10−5 m2/s. Calculate NA.

1
DAm =
x / DAB + xC' / DAC
'
B

33

33

8
 14-Oct-21

34

34

Diffusion Coefficients for Gases

A. Experimental determination of diffusion coefficients: when no data is available
B. Experimental diffusivity data at certain T &P: See Table 6.2-1;
others are available in literature.
C. Empirical equations
• with deviation up to 8% up
Chapman and Enskog correlation for to 1000K
diffusion of non-polar binary system with
elastic sphere–type molecules: • AB and D,AB from
1/ 2 references (sometimes not
1.8583 10 −7 T 3 / 2  1 1 
DAB =  +  available or difficult to
P AB  D , AB  M A M B 
2
estimate and slightly f(T))

DAB is the diffusivity in m2/s, T temperature in K, MA molecular weight of A in

kg mass/kg mol, MB molecular weight of B , and P absolute pressure in atm
AB: average collision diameter
D,AB: collision integral based on Lennard-Jones potential; is a ratio giving the
deviation of a gas with interaction as compared to a gas of rigid, elastic spheres.
 D,AB = 1 with no interactions 35

35

9
 14-Oct-21

Fuller et al. correlation (semi-empirical) Does not include AB

DAB =
−7
110 T (1 / M A + 1 / M B )
1.75 1/ 2
and D,AB
P( A ) + ( B ) 
1/ 3 1/ 3 2
T 1.75
D AB 
DAB is the diffusivity in m2/s, T P
temperature in K, MA molecular weight
of A in kg mass/kg mol,
MB molecular weight of B , and P
absolute pressure in atm
A , B : sum of structural volume increments

Values of atomic and structural

diffusion elements are available in
Table 6.2-2
In gases, the effect of concentration of A in
B is very weak (<4%) and is not included
in above equations

36

36

Example: Estimation of Diffusivity of a Gas Mixture

Normal butanol (A) is diffusing through air (B) at 1 atm abs. Using the Fuller et
al. method, estimate the diffusivity DAB for the following temperatures and
compare with the experimental data:
110 −7 T 1.75 (1 / M A + 1 / M B )
1/ 2
(a) For 0oC. DAB =
(b) For 25.9oC. 
P ( A ) + ( B )
1/ 3
 1/ 3 2

(c) For 0oC and 2 atm abs.

For part (a), P = 1.00 atm, T = 273 + 0 = 273 K, MA (butanol) = 74.1, MB (air) = 29.

This value deviates by +10% from the experimental value of 7.03×10−6 m2/s
For part (b), T = 298.9K: DAB= 9.05×10−6 m2/s. This value deviates by +4% from
the experimental value of 8.70×10−6 m2/s.
For part (c), the total pressure P = 2 atm. Using the value predicted in part (a) and
correcting for pressure,
37

37

10