Interphase mass transfer

Diffusion of A between a vapor and liquid phase:

Individual mass transfer coefficient,k:

NA = ky (yA,G – yA,i ) = kx (xA, --xA,L)

(yA,G – yA,i ) kx where k is individual mass transfer coefficient

--------------- = -- ---- suffix x---liquid phase; y---vapor phase
(xA, --xA,L) ky yA and xA are mole fractions in vapor and
liquid phase respectively.
Subscript G, L, i represent bulk gas phase, bulk
Liquid phase and interface respectively.
Overall mass transfer coefficient, K:
NA = Ky (yA,G –y*A) y*A = molefraction of A in vapor phase
= Kx (x*A—xA,L) that would be in equilibrium with xA,L ,
mole fraction of A in bulk of the liquid.
x*A = mole fraction of A in liquid phase that
would be in equilibrium with yA,G,
mole fraction of A in bulk of the gas.
NA NA m’ NA 1 1 m’
---- = ---- + --------- OR ---- = ---- + ------
Ky ky kx Ky ky kx

NA NA NA 1 1 1
---- = --------- + --------- OR ---- = --------- + ---------
Kx m’’ ky kx Kx m’’ ky kx

yA, ---y*A yA,G ---yA,

where m’ = -------------- & m’’ = ----------------
xA, --- xA,L x*A ---xA,

Gas film controlled:

If m’ is too small, 1/Ky = 1/ky and the rate of mass transfer is gas-phase-controlled;
yA,i = yA* ; A is very soluble in liquid.
Eg. Vapors like NH3 and SO2 being absorbed in water.

Liquid film controlled::

If m’’ is too large, 1/Kx = 1/kx and the rate of mass transfer is liquid-phase-
controlled;
xA,i = xA* ; A is relatively insoluble in liquid.
Eg. Gases like CO2 and O2 being absorbed in water.

If gas phase concentration is given in partial pressure , p A and liquid phase concentration
is given in concentration,c:
p Ai = H cA,i where H is Henry’s law constant.
The overall coefficients, K can be related to individual phase coefficients,k as:

1 1 H
---- = ---- + ------
KOG kG kL

1 1 1
---- = ---- + ------
KOL HkG kL

If H is too small, 1/KG = 1/kG and the rate of mass transfer

is gas-phase-controlled; pA,i = pA* ; A is very soluble in liquid.
If H is too large, 1/Kx = 1/kx and the rate of mass transfer
is liquid-phase-controlled; cA,i = cA* ; A is relatively insoluble in liquid.

Local Mass Transfer Coefficient,F:

F is the local mass transfer coefficient at a point on the gas-liquid interface.
F= DAB c / z
so that
NA N ( N A + N B ) − c Ai c
NA = F ln A
NA + NB N A ( N A + N B ) − c AG c

Relationship between F and Individual mass transfer coefficients,k:

Gas phase:
A through stagnant B Equimolar counterdiffusion of A & B
( ) ( ) ( ) (
NA = kG ∆ p = kG p AG − p A,i = kG’ ∆ p = kG’ p AG − p A,i )
= ky (∆y) = ky (yA,G –yA,i) = ky ‘(∆y) = ky’ (yA,G –yA,i)
= kc (∆c) = kc (cA,G –cA,i) = kc ‘(∆c) = kc’ (cA,G –cA,i)

Liquid phase:
A through stagnant B Equimolar counterdiffusion of A & B
NA = kL (∆c) = kL (cA,i –cA,L) = kL ‘(∆c) = kL’ (cA,i –cA,L)
= kx (∆x) = kx (xA,i –xA,L) = kx ‘(∆x) = kx’ (xA,i –xA,L)

F = kG p BM = ky p BM /pt = kc p BM / RT = kY/MB
= kG’pt = ky’ = kc pt /RT = kc’ c for gas phase
F = kx xB,M = kL xB.M c = kL’ c = kL’ ρ/M = kx’ for liquid phase
Units for mass transfer coefficients:
KOG , kG , kG’ ----moles/ (area)(time0(pressure)
---kmoles/(m2 .s.atm) or kmoles/(m2 .s.Pa)
Ky, ky , ky’, ----moles/ (area)(time)(mole fraction)
KOL , kL , kL’ ---kmoles/(m2 .s.)
kc , kc ’ ---- moles/ (area)(time)(mole/ volume)
--- kmoles/(m2 .s.kmol/m3)
Resistance in gas phase 1 / ky
---------------------------------------- = ---------
Total resistances in both phases 1 / Ky

Resistance in liquid phase 1 / kx

---------------------------------------- = ---------
Total resistances in both phases 1 / Kx

Estimation of k:
Experimental method: Wetted wall column
Flow may be laminar or turbulent.
For laminar flow of liquid in a wetted wall column: Mass transfer in a falling film
Mass transfer at the interface may be diffusional mass transfer or convective mass
transfer or both.
Diffusional mass transfer: Molecular diffusion is the mass transfer mechanism.
Convective mass transfer: Mass transfer is influenced predominantly by fluid flow
characteristics.
For laminar flow of liquids:
Mass Transfer from a gas into a falling liquid film:
Consider the absorption of solute from a gas into a falling liquid film as in wetted wall
column. Here liquid flow is laminar; diffusion is very slow. Hence the velocity field is
virtually unaffected by diffusion.
Solute A is slightly soluble in liquid B (eg. Air in water). As xA is low, viscosity is not
varying appreciably. A does not penetrate more into B. Hence δ c << δ.
CAG is constant throughout the length of the column. At equilibrium, cAi is also constant.
Film thickness = δ.
Fluid flow is in z direction. Velocity profile = Vz (x)
∂c
Mass flow (diffusion) is in x direction. Conc. gradient = A
∂x
Momentum transfer:
  x 2 
V z ( x ) = Vmax 1 −   
  δ  
Mass transfer:
volume element = W(∆x)( ∆z)
At steady state, about the volume element:
Rate in – Rate out = 0
N AZ z W∆x − N AZ z + ∆z W∆x + N Ax x W∆z − N Ax x + ∆x W∆z = 0
Where W is the width of the film.
Dividing by W(∆x)( ∆z), and letting (∆x) ,( ∆z) -
0,
∂  ∂
  N Ax +   N AZ = 0
 ∂x   ∂z 
One directional molar fluxes are:

 ∂c 
N Ax = − D AB  A  + x A ( N Ax + N Bx ` )
 ∂x 
 ∂c 
N Az = − D AB  A  + x A ( N AZ + N BZ )
 ∂z 

A transport in x direction: By diffusion

A transport in z direction: By convective mass transfer; diffusion is negligible

 ∂c 
N Ax = − D AB  A 
 ∂x 
N AZ = x A ( N AZ + N BZ ) = C AVZ ( x )

Substituting we get
 ∂c  ∂  ∂ 2c 
VZ  A  = [VZ C A ] = D AB  2A 
 ∂z  ∂z  ∂x 

Solute has penetrated only a very small distance and it gives the impression that it is
carried along with the film with a velocity= Vmax.
Hence at x=0, VZ = Vmax.
 ∂c   ∂ 2c 
Vmax  A  = D AB  2A 
 ∂z   ∂x 
This is solved by Laplace transform using the boundary conditions:
B.C.1; at z =0, CA=0
B.C.2; at x =0, CA=CA0
B.C.1; at x =∞, CA=0
Finally,
4(D ABVmax )
N A = C A0
πL
Where L is the length of the film (in z direction).

Hence k ∝ DAB 0.5

Relationship between k and D:
Experimental: k ∝ Dn where n = 0.8 to 0.9.
Film theory:
Entire conc. difference is attributed to molecular diffusion
Within an effective film of thickness,z F .
Validity: the film should be extremely thin, so that the solute quantity
within it is very small compared to that passing through it.
D
NA = kc (cA1 –cA2) = (c A1 − c A2 ) AB
zF
k∝ D
Penetration theory:
Time of exposure of fluid to mass transfer = θ. θ is constant and is
very short so that ZF is too large for solute to cross.
∂ cA ∂ 2cA
------ = DAB ----------
∂ θ ∂ z2
Boundary conditions:
cA0 at θ = 0 for all z
CA = cA,i at z = 0 for θ >0
cA0 at z = for all θ

NA,av = (cA,i -- cA0 ) DAB /π θ

k ∝ (DAB )0.5
Surface renewal theory:
Time of exposure of eddies of the fluid at the surface, θ is not constant.
The liquid-gas surface is a mosaic of surface elements of different θ.
The rate of solute penetration depends on θ and hence NA,av θ.
Assumption: The chance of surface element being replaced by another
is independent of θ.
NA,av = (cA,i -- cA0 ) D AB s
kL,av = D AB s
k ∝ (DAB )0.5

Combination of film –surface renewal theory: Dobbins’ theory.

Assumption: z = z b which is finite.
This way, it is different from surface renewal theory
Which assumes z = z .
The 3rd boundary condition is now as follows:
cA = cA0 at z = zb for all θ
sz b2
kL,av = D AB s coth
D AB

k ∝ (DAB )0.5
Corresponding Dimensionless Groups in Heat and Mass Transfer
No Mass Transfer Heat Transfer Significance
1 c A − c A1 t − t1 Fraction of quantity
c A2 − c A1 t 2 − t1 transferred
2 Reynolds No: Reynolds No: Inertial stress
luρ luρ Viscous stress
Re = Re =
µ µ

3 Schmidt No Prandtl No Momentum diffusivity

Sc = µ /(ρ DAB) = ν /DAB Pr = Cp µ / k = ν /α Sc = ---------------------------
where α = k / ρ cp Molecular diffusivity
4 Sherwood No Nusselt No
Fl Convective flux
Sh = where ----------------------------
cD AB hl
Nu = ( Molecular) diffusive flux
F= DAB pt / RTz k
Across a layer of thickness l
p B,M
= k G p B , M = ky
pt
p B.M
=kc = kY /MB =kG’pt
RT
=ky’ = k c‘ pt /RT = kc’c

5 Grashof No Grashof No (Buoyancy forces x inertia

forces)
gl 3 ∆ρ  ρ 
2
ρ
2
(viscous forces)2
GrD =   GrD = gl β ∆t  
3

ρ  µ  µ
6 Peclet No Peclet No Flux due to bulk flow of the
lu lu medium
PeD = ReSc= PeH = RePr =
D AB α diffusive flux
Across a layer of thickness l
7 Stanton No Stanton No Convective flux
Sh Sh F StH= Flux due to bulk flow of the
StD = = =
Re Sc Pe D cu Nu Nu h medium
= =
=F/G Re Pr Pe H c p uρ
8 jD = StD Sc2/3 jH = StH Pr2/3
 Analogy between Momentum, Heat and Mass Transfers

Flux Tranfer Operation

Momentum Heat Mass
Molecular du dt dc
Diffusion τ = −µ x q = −k J A = − D AB A
dz dz dz

k d ( ρc P t ) d (ρc P t )
=− = −α
ρc P dz dz
Eddy du x d ( ρc P t ) dc A
Difusion τ = − (µ + E v ) q = −(α + E H ) J A = −(D AB + E D )
dz dz dz
= τ molec + τ turb = q molec + q turb = J Amolec + J Aturb
Eddy  du  b H u ' z ∆ ( ρc P t ) bD u ' z ∆c A
E v = ρl 2  − x  EH = ED =
diffusivity
 dz  ∆ ( ρc P t ) / l ∆c A / l
Molecular µ α D AB
diffusivity

Prandtl Mixing Length,l : The mixing length is the distance that a fluid eddy particle will
keep its original characteristics before blending into the surrounding fluid.
u’x = ∆ux = -l dux / dz.

Mass Transfer Analogies

Reynold’s analogy:

relates the Fanning friction factor for fluid flow to heat transport:

f /2 = h / (ρ cp ux ) = StH

where the right hand side is the "Stanton Number".

Stanton number, StH = overall heat transport to the wall / convective effects in the
mainstream.

f / 2 =overall momentum transported to the wall / inertial effects in the main stream.

The Reynolds analogy says that these ratios are equal for mass.. It can be defined for
heat transfer or for mass transfer.
The Reynolds analogy gives reasonable values for gases where the Prandtl number is
roughly one.

Prandtl’s analogy:

The Reynolds analogy postulates direct interaction between the turbulent core of the flow
and the walls. It neglects the effect of laminar sublayer. If a laminar sublayer is included
between these, the Prandtl-Taylor analogy applies:

When the j-factors are used, the fluid properties in the Stanton number are evaluated at
the mean bulk average temperature and those for the Prandtl number at the film
temperature (this means two heat capacities!). Assumption: Ev =EH =0 in lam.sublayer.

Colburn analogy:
hav ½ f
St H,av = --------- = -----------
C p ux ρ Pr2/ 3

The group St H,av Pr2/ 3 is called j H.

hav
St H,av Pr = ----------- Pr2/ 3 = j H = ½ f
2/ 3
= φ(Re)
C p ux ρ
This represents the data even for large Pr.

Chilton- colburn analogy:

For mass transfer:
Fav
St D,av Sc = --------- Sc2/ 3 = j D = ½ f
2/ 3
= φ(Re)
C ux

When Sc is not too large.

The Colburn-Chilton analogy is simply

valid for turbulent flow in conduits with NRe > 10000, 0.7 < NPr < 160, and tubes where
L/d > 60 (the same constraints as Seider-Tate).

A wider range of data is correlated by the Friend-Metzner analogy:

which is valid when NRe > 10000, 0.5 < NPr < 600, 0.5 < NSc < 3000.

Reynolds Prandtl & Chilton-Prandtl Von-Karman

Assumes entire fluid Assumes fluid region consists of Considers all 3
region to be under turbulent core and viscous sub- regions of fluid
turbulent zone layer; neglects buffer region flow.
ν=α=D Sc ≠ 1 and Pr ≠ 1 Sc ≠ 1 and Pr ≠ 1
All molecular diffusivities
are equal. Hence,
ν /D =ν / α = 1
ie Sc = Pr =1