Lecture 5

ChE-402

1
 Intended Learning Outcome

To solve transient problems involving both convection and diffusion.

To further analyze origin of convection.

To evaluate diffusion coef cients using derived as well as empirical models.

2
fi

 Transient convection and diffusion

Component 2 does not mix with liquid phase of 1

• In the steady case, we could use v2 = 0 (stagnant air);

• In this unsteady case, the ux of 1 varies with time and position
(for example, 1 displaces 2 immediately after t =0, so v2 ≠ 0).

De ne your system -
A
De ne an element to do mass balance z=l c1l

Apply mass balance

o o o o o
l
Accumulation * dV = Flu x ∣in * A − Flu x ∣out * A + Generation * dV − Consumption * dV
z=0 c10

Fast evaporation by
diffusion and
convection

3
fi
fi
fl

Transient convection and diffusion

∂c1 ∂n
=− 1
∂t ∂z

∂c1 ∂2c1 ∂
= D 2 − (c1v v)
∂t ∂z ∂z

A
Apply mass balance for component 2 z=l c1l
o o o o o
Accumulation * dV = Flu x ∣in * A − Flu x ∣out * A + Generation * dV − Consumption * dV dz
z
l

z=0 c10

Fast evaporation by
diffusion and
convection

4
Transient convection and diffusion
∂c1 ∂n1 ∂c2 ∂n2 − − − −
=− =− v
v = c1V1v1 + c2V2v2 = V1n1 + V2n2
∂t ∂z ∂t ∂z

Multiply each term with molar volume and add

A
z=l c1l

dz
z
l
Implications

z=0 c10

Fast evaporation by
diffusion and
convection

5
Transient convection and diffusion
∂c1 −
n1 ∣z=0 = −D ∣ + c1V1n1 ∣z=0
∂z z=0

A
z=l c1l

z=0 c10

Fast evaporation by
diffusion and
convection

6
Transient convection and diffusion
− ∂c1
V1
(1 − c V ) z=0 ∂z
∂c1 ∂2c1 ∂z ∂c1
= D 2 + D − ∣
∂t ∂z 1 1

Initial condition:

Boundary conditions

7
Summary: transient convection and diffusion
− ∂c1
V1

( ) ( π(1 + erf ϕ)ϕeϕ2 )

−1
1 ∂ζ − 1
ϕ =− − ∣z=0 V1csat
1 = 1 +
2 1 − c1V1

z c1 1 − erf (ζ − ϕ)
ζ= =
4Dt c1
sat 1 + erf ϕ

∂c1

πt ( 1 − V1c sat ) 1 + erf ϕ

−D D 1 e −ϕ 2

(1 − c V )
∂z
n1 ∣z=0 = − ∣z=0 = − c1sat
1 1
1

8
Summary: transient convection and diffusion
∂c1

πt ( 1 − V1c sat ) 1 + erf ϕ

−
−ϕ 2
V1 D 1 e
( )
ϕ =−
1 ∂ζ
∣z=0
n1 ∣z=0 = − c1sat
−
2 1 − c1V1 1

9
In which of the following cases, convection can not be
neglected

a) Diffusive ux at z = 0 is high
b) y1 at z =0 is 0.05
c) y2 at z = L is 0.5

10
fl

 Exercise problem
A container lled with liquid aniline (molar mass: 93.13 g/mol) is stored in a large room. The container
has a cylindrical tube with length of 10 cm and a diameter of 1 cm. Initially, the cylindrical tube is lled
with air (pressure of 1 atm), and aniline is prevented from evaporation. At time t = 0, the evaporation of
aniline is allowed.

Saturation vapor pressure of aniline: 10 kP

The diffusion coef cient of the aniline vapor in air: 9 x 10-2 cm2/s
Aniline vapor can be treated as ideal gas.
Assume a concentration of 0 at z = 10 cm, and saturation conditions at z =0

1. Neglecting convection, calculate the partial pressure of aniline vapor at z =1.5 cm after 25 seconds

2. Estimate the error in neglecting convection.

11
fi

fi

fi
.

Exercise problem

12
Diffusion coef cients in gases, liquids and solids

13
fi
 Understanding the driving force for diffusion
Diffusion ux is essentially a ow driven by force (the gradient of chemical potential).
If there is a ow, there should be frictional force opposing the ow.

Frictional drag = chemical potential force High Low

concentration concentration

dμ 1 dμ u
fu = − ⇒ u =− f
dz f dz Friction,

[ f d ln c ] dz
c dμ ̂
kBT d ln( f /P) dc dc
ux = uc = − =− =−D
f dz dz

̂ ̂ kBT
D =
kBT d ln( f /P)
= Do
d ln( f /P) Do =
f d ln c d ln c f

Diffusivity is inversely proportional to frictional force

Stoke’s Einstein Equation

14
fl

fl
fl
fl
fl

Comparison of diffusion coef cients

RT
Do =
f

fgas < < fliquid < < fsolid

Dgas > > Dliquid > > Dsolid

15
fi
Diffusion coef cient of gas
l = mean free path
Velocity of particles are thermally distributed
v
v mv 2
v
( 2πkBT )
3/2 −
m
g(v) = e 2kBT
z0

l
θ
l cos θ
z−
v
z0 = z− + l cos θ

dn
↑ = c ∣z− vz vz = v cos θ
dt

df
Employ Taylor series for c ∣z− f (x) = f (x0) + (x − x0) ∣x=x0
dx

dc
c ∣z− = c ∣z0 + (z− − z0) ∣z=z0
dz
dc
c ∣z− = c ∣z0 − l cos θ ∣z=z0
dz

16
fi
Diffusion coef cient of gas
⟨ dt ⟩ ⟨ dt ⟩
dn dn
J ∣z=z0 = ↑ − ↓

∫−∞ ( )
∞ ∞ ∞

∫−∞ ∫−∞
J ∣z=z0 = dvx dvy c ∣z− vz − c ∣z+ vz g(v) dvz

∫−∞ x ∫−∞ y ∫−∞ ( )

∞ ∞ ∞
dc
⇒ J ∣z=z0 = − dv dv 2l cos θ ∣z=z0 vz g(v) dvz
dz
−
v
6

(∫−∞ x ∫−∞ y ∫−∞ )

∞ ∞ ∞
dc dc 8kBT
⇒ J ∣z=z0 = − 2l ∣z=z0 dv dv cos θ vz g(v) dvz = − D ∣ −
dz dz z=z0 v =
πm

( )
dc v dc
⇒ J ∣z=z0 = − 2l ∣z=z0 =−D ∣z=z0
dz 6 dz

17
fi
 Diffusion coef cient of gas
1−
D = vl − 8kBT
σ
3 v =
πm

kBT/P volume of space per molecule

l = =

(4 )
π 2 cross−sectional area
σ

Mean free path is much bigger than average

σ distance between gas molecules

[( r ) (r ) ]
12 6
σ σ
VLJ = 4ϵ −

18
fi
Calculate mean-free path, mean velocity and D of
helium at 1 atm and 25 ºC.
kBT/P volume occupied by single molecule − 8kBT
l = = v =

(4 )
π 2 cross−sectional area πm
σ

m( π )
3/2
1− 8 2 kBT 1
D = vl =
3 3 Pσ 2

19
In previous problem, calculate how fast a gas
molecule is colliding. Also, calculate average
distance between molecules.

20
Binary diffusion coef cient for gas
Chapman-Enskog theory: Empirically derived (accurate to 8%)
Ω = Collision Integral
1.86 * 10−3 * T 1.5 * (1/M1 + 1/M2)0.5 D in cm2/s
D12 = T in Kelvin
Pσ122Ω P in atm
M in g/mole
σ12 in Angstrom

[( r ) (r ) ]
12 6
σ1 + σ2 σ σ
σ12 = ϵ12 = ϵ1ϵ2 VLJ = 4ϵ −
2

21

fi
 Calculate D for helium in argon at 1 atm and 25 ºC
with the Chapman-Enskog theory
D in cm2/s
1.86 * 10−3 * T 1.5 * (1/M1 + 1/M2)0.5 T in Kelvin
D12 = P in atm
Pσ122 Ω
M in g/mole
σ12 in Angstrom

22

Diffusion at high pressure

m( π )
3/2
8 2 kBT 1
D=
3 Pσ 2

23
 Diffusion coef cient in liquid

One can primarily use the Stoke’s-Einstein equation

kBT kBT
Do = = Stoke’s law f = 6πηR
f 6πηR

Rsolute
Valid when ≥5
Rsolvent

24

fi
 Shape effect

kBT kBT
Do = =
Spherical f 6πηR

kBT kBT
Do = =

[ a + (a 2 − b 2)1/2 ]
f (a 2 − b 2)1/2
Prolate ellipsoid 6πη

( )
ln
b

kBT kBT
Do = =

[ ]
Oblate ellipsoid f (a 2 − b 2)1/2
6πη

( )
(a 2 − b 2)1/2
tan−1
b

25
 Diffusion coef cient in liquid

When Rsolute ≈ Rsolvent 7.8 10−8 T ϕM2 Wilke-Chang correlation

D= − 0.6
ηV1

D is diffusion coef cient of solute in cm2/s

ϕ is empirical (1 for most organic solvents, 1.5 for alcohols, 2.6 for water)
−
V1 is molar volume of solute in cm3/mol

M2 is molecular weight of solvent in daltons

η is viscosity in centipoise

26

fi
fi
 Diffusion of solids

Di usion is activated

ΔE
−
D = Do e RT

Extremely small di usion coe cient

1) Almost all transport takes place through defects in the solid, especially along grain-boundaries.
2) Transport approaches the limit of semi-in nite media rather than the thin lm.
3) Diffusion of hydrogen in solids is an exception because hydrogen dissociates in atomic
hydrogen. The electrons of hydrogen disperse into the metallic electron cloud. The proton being
much smaller than inter-atomic distances can reside at the inter-metallic interstitial positions and
diffuse relatively fast through the lattice.

27
ff
ff
ffi
fi
fi

Diffusion coef cient of H2 at 1 bar in air is D0. At 3

bar, it would be

A. Do

B. 3 Do

C. Do^(1/3)

D. Do/3

28
fi