Download as pdf or txt
Download as pdf or txt
You are on page 1of 14

PCE 351

SEPARATION PROCESSES I

Lecture notes prepared by


Dr. Augustine Ntiamoah
Department of Chemical Engineering, KNUST.
Program: BSc (Hons) Petrochemical Engineering (3rd Year)

UNIT 2: MASS TRANSFER 2

2.4. PREDICTION OF DIFFUSIVITY


To be able to calculate the flux, we need to know the diffusivity, 𝐷AB. Due to the
difference in the mobility of molecules in different phases, the diffusivity is
dependent on the type of phase, i.e., gas, liquid, or solid.

Diffusivities are generally higher for gases than liquids, both of which are higher
than for solids. For example, the molecules in liquids are closer together than in
gases, and hence, diffusing molecules will collide more often and diffuse more
slowly than in gases. This explains why diffusion coefficients in liquids are
relatively smaller than in gases.

Diffusivity values
Gases: 5×10‒6 to 1×10‒5 m2 /s
Liquids: 1×10‒10 to 1×10‒9 m2 /s
Solids: 1×10‒14 to 1×10‒10 m2 /s

Measurement of diffusivity can be made in the laboratory, but there are also
semi-theoretical equations (correlated from experimental data) that can be used
to predict the values of diffusivities for different systems.

For Gases: Diffusivity is shown to be inversely proportional to pressure, directly


proportional to temperature, and almost independent of composition. The two
most commonly used equations for predicting the diffusivity of gases are:

1
(i) Fuller, Schettler, and Giddings Equation:
1⁄2
1.00 × 10−7 𝑇 1.75 (1⁄𝑀 + 1⁄𝑀 )
𝐴 𝐵
𝐷𝐴𝐵 = 1⁄ 3 1⁄ 3 2
𝑃𝑇 [(∑ 𝑣𝐴 ) + (∑ 𝑣𝐵 ) ]

𝑇 is absolute temperature (K); 𝑃 is total gas pressure (atm); 𝑀A and 𝑀B are molar
masses (g/mol); 𝑣𝐴, 𝑣𝐵 are atomic diffusion volumes (or structural volume
increments). Values of atomic diffusion volumes for common substances are
given in Table 18.2-2 (Geankoplis, 5th Ed.). Here, 𝐷AB is proportional to 1/𝑃 and
to 𝑇1.75.

(ii) Chapman-Enskog Equation (sometimes referred to as Hirschfelder equation)

1.8583 × 10−7 𝑇 3/2 1 1⁄ )


1⁄2
𝐷𝐴𝐵 = 2 ( ⁄𝑀𝐴 + 𝑀𝐵
𝑃𝑇 𝜎𝐴𝐵 Ω𝐷,𝐴𝐵

𝐷𝐴𝐵 = diffusivity in m2/s; 𝑃 = absolute pressure in atm; 𝜎𝐴𝐵 = average collision


diameter = (𝜎𝐴 + 𝜎𝐵 )⁄2 ; Ω𝐷,𝐴𝐵 =collision integral based on Lennard-Jones
potential (a ratio giving the deviation of a gas with interaction as compared to a
gas rigid, elastic spheres).

For Liquids: DAB = f(CA). DAB decreases for increasing CA.

𝑇
𝐷𝐴𝐵 𝛼 ( ), 𝜇 = viscosity
𝜇

There are two main theories for liquid diffusivities:

(i) Stokes-Einstein equation:

9.96 × 10−16 𝑇
𝐷𝐴𝐵 = 1/3
𝜇𝑉𝐴

where DAB is diffusivity in m2/s, T is temperature in K, μ is viscosity of solution in


Pa·s or kg/m·s, and VA is the solute molar volume at its normal boiling point in
m3/kgmol.

2
(ii) Wilke-Chang equation:
𝑇
𝐷𝐴𝐵 = 1.173 × 10−16 (∅𝑀𝐵 )1/2
𝜇𝐵 𝑉𝐴0.6

where MB is the molecular weight of solvent B, μB is the viscosity of B in Pa·s or


kg/m·s, VA is the solute molar volume at the boiling point. 𝛷 is the association
parameter for the solvent B. The recommended values of this parameter for
some common solvents are given below:

Solvent 𝜱𝑩
Water 2.6
Methanol 1.9
Ethanol 1.5
Others 1.0

Example 2.4
Estimate the diffusivity of methanol vapour in carbon monoxide gas at 1 atm and
100 °C, using the Fuller et al method.

Species A: Methanol (CH3OH): 1C, 4H, 1O

Species B: CO

Given: T = 373 K (100 °C) and P = 1 atm

Also, MA = 32.04 kg/kgmol; MB = 28.01 kg/kgmol

Fuller et al equation is given by:


1
1 1 2
1.00 × 10−7 T1.75 ( +
MA MB )
DAB =
1 1 2
P [(ΣVA )3 + (ΣVB )3 ]

3
From table 18.2-2:

ΣVA = ΣVcarbon + ΣVhydrogen + ΣVoxygen

ΣVA = 1(16.5) + 4(1.98) + 1(5.48) = 29.9

ΣVB = 18.9 (i.e., the value for CO, a simple molecule)

1
1.00 × 10−7 (373)1.75 (1/32.04 + 1/28.01)2
DAB =
1 1 2
P [(29.9)3 + (18.9)3 ]

DAB = 2.463 × 10−5 m2 /s

2.5 CONVECTIVE MASS TRANSFER


Convective mass transfer occurs due to macroscopic fluid motion. It can be due
to forced or free/natural convection. In the chemical industry, convective mass
transfer is more common because it results in much higher mass-transfer rates,
and thus, requires smaller equipment.

In many cases, to have a fluid in convective flow, the fluid velocity is increased
until turbulent diffusion or eddy current occurs. Convective mass transfer from
the surface to a fluid, or vice versa, can be represented as follows:

4
The simplest model is to assume a linear profile, then an equation like Fick’s law
can be written for the flux:

𝑑𝑐𝐴
𝐽𝐴 = −(𝐷𝐴𝐵 + 𝜀𝑀 )
𝑑𝑧

where DAB is the molecular diffusivity in m2/s and εM is eddy diffusivity in m2/s.
In practice 𝑑𝑧 (𝑖. 𝑒. 𝑧1 − 𝑧1 ) is very small ((𝜇𝑚). It cannot be predicted, hence
∆z is included in the proportionality constant referred to as convective mass-
transfer coefficient. The flux, denoted by NA is thus written as:

𝑁𝐴 = 𝑘𝑐 (𝑐𝐴1 − 𝑐𝐴2 )

𝑘𝑐 = convective mass transfer coefficient (MTC). MTC is similar to the heat


transfer coefficient h and is a function of the system geometry, fluid properties
and flow velocity.

Because there are many measures of composition (other than 𝑐𝐴 ) there are
equally many forms of the MTC.

Gas: 𝑁𝐴 = 𝑘𝑐 (𝑐𝐴1 − 𝑐𝐴2 ) = 𝑘𝐺 (𝑝𝐴1 − 𝑝𝐴2 ) = 𝑘𝑦 (𝑦𝐴1 − 𝑦𝐴2 )

𝑝𝐴 𝑅𝑇
For perfect gases: 𝑝𝐴 = 𝑦𝐴 𝑃; 𝑐𝐴 = ; 𝑘𝑐 = 𝑘𝐺 𝑅𝑇 = 𝑘𝑐 and 𝑘𝑦 = 𝑘𝐺 𝑃
𝑅𝑇 𝑃

Liquids: 𝑁𝐴 = 𝑘𝑐 (𝑐𝐴1 − 𝑐𝐴2 ) = 𝑘𝐿 (𝑐𝐴1 − 𝑐𝐴2 ) = 𝑘𝑥 (𝑥𝐴1 − 𝑥𝐴2 )

𝑐𝐴 𝑘𝑥
𝑥𝐴 = ; 𝑘𝑐 = 𝑘𝐿 =
𝑐 𝑐

Estimating the Mass Transfer Coefficient, 𝒌𝒄


It is often not feasible to experimentally measure the mass transfer coefficient.
Approximate approaches (correlations) can be employed to estimate the mass
transfer coefficient using dimensional analysis.

Variables needed to describe the system include those descriptive of the system
geometry, the flow and fluid properties and the quantity of importance, 𝑘𝑐 .

5
Analysis by the Buckingham π-method:

Variables: mass transfer coefficient (k c ), length scale (L or d), characteristic


velocity (υ), fluid properties (ρ, μ, D).

Variable Symbol Dimensions


Diameter d L
Density ρ ML-3
Viscosity μ ML-1T-1
Velocity υ LT-1
Diffusivity DAB L2T-1
Mass-transfer coefficient 𝑘𝑐 LT-1

There are six variables and three fundamental dimensions (M, L, T). Hence three
dimensionless groups (6 – 3 = 3) are needed to describe the system. Using the
core variables 𝐷,  and 𝑑 gives the following three dimensionless groups:

𝑘𝑐 𝐿
1. Sherwood number: 𝑆ℎ =
𝐷

𝐿υ𝜌
2. Reynolds number: 𝑅𝑒 =
𝜇

𝜇
3. Schmidt number: 𝑆𝑐 =
𝜌𝐷

The three relations show that 𝑆ℎ = 𝑓(𝑅𝑒, 𝑆𝑐). The general equation is of the
form:

𝑆ℎ = 𝑎𝑅𝑒m 𝑆𝑐n (𝑎, 𝑚, 𝑛 are constants)

Knowing the Sherwood number, the mass transfer coefficient, 𝑘𝑐 can be


determined. A single equation describes both gases and liquids if some accuracy
can be sacrificed:

𝑆ℎ = 0.023𝑅𝑒 0.83 𝑆𝑐 0.33

where Re > 2100; 0.6 < Sc < 3000

6
(Gases) (Liquids)

2.6 INTERPHASE MASS TRANSFER


Molecular diffusion and convective mass transfer are concerned with mass
transfer within a single phase. In many diffusional separation processes, mass
transfer takes place between two (insoluble) phases across their interface. When
the two phases are in contact, there is exchange of species across the interface
from one phase to another until equilibrium is reached. This is referred to as
interphase mass transfer.

The theoretical model commonly used to quantitatively describe the mass


transfer across an interface is the film model. The key assumption of the film
model is that the whole resistance to diffusional mass transfer entirely resides in
the fluids. (i.e. to diffuse into the bulk of a fluid, molecules must cross a thin,
stagnant region, called film, where the whole resistance to mass transfer is
present). The molecules do not encounter any resistance as they pass across the
interface.

The interface:
• Assumed to have zero thickness, hence offers zero resistance to mass transfer.
• The required driving force for any finite flux is zero. (A system with zero
driving force is an equilibrium system).

Interphase mass transfer involves three transfer stages:


Consider the transfer of component A from the gas to a liquid phase, where 𝑦Ai
and 𝑥 Ai are the interfacial mole fractions. Component A is transferred from:
- Bulk of gas phase G to the interface,
- Through the interface,
- From the interface to the bulk of liquid phase L.

7
We consider a two-film or two-resistance model, involving two mass transfer
films (gas phase and liquid phase films), each is characterized by a film mass
transfer coefficient, MTC (denoted by lower-case k).

(1) gas-film: 𝑁A = 𝑘y(𝑦A − 𝑦Ai)


(2) liquid-film: 𝑁A = 𝑘x(𝑥Ai – 𝑥 A)

where 𝑘y and 𝑘x are the locally applicable mass transfer coefficients for the gas
and liquid films.

At steady state, the flux of A is everywhere the same (through each film and
through the interface). Thus, the rate of transfer to the interface is equal to the
rate of transfer from the interface.

𝑁𝐴 = 𝑘𝑦 (𝑦𝐴 − 𝑦𝐴𝑖 ) = 𝑘𝑥 (𝑥𝐴𝑖 − 𝑥𝐴 )

(𝑦𝐴 − 𝑦𝐴𝑖 ) 𝑘𝑥
→ =−
(𝑥𝐴 − 𝑥𝐴𝑖 ) 𝑘𝑦

The ratio −𝑘x/𝑘y is the slope of the line connecting the equilibrium composition
with the bulk composition. The equation can be represented on a graph as
follows:

8
The equilibrium concentrations, 𝑦𝐴𝑖 and 𝑥𝐴𝑖 are related by equilibrium relations.
Note that these compositions do not exist physically. They are hypothetical
mixtures that would be in equilibrium with the bulk of the other phase. To
calculate the rate of mass transfer, a more appropriate expression for 𝑁A is used
which uses an overall mass transfer coefficient, normally indicated with an
upper-case letter, 𝐾.

Overall Mass Transfer Coefficient (MTC)


Experimentally, it is possible to determine bulk concentrations and not
interfacial concentrations. Overall mass transfer coefficients are therefore,
defined in terms of bulk concentrations.

In terms of an overall coefficient, the flux 𝑁A, is given by

𝑁𝐴 = 𝐾𝑦 (𝑦𝐴 − 𝑦𝐴∗ ). Flux based on overall gas-phase driving force.

𝑁𝐴 = 𝐾𝑥 (𝑥𝐴∗ − 𝑥𝐴 ). Flux based on overall liquid-phase driving force

Equilibrium Relations (Henry’s law)


The partial pressure of a component in a gas phase 𝑝A, is given by:

𝑝𝐴 = 𝑦𝐴 𝑃

where 𝑦𝐴 is the mole fraction of component A in the gas phase, and P the total
pressure.

9
Henry’s Law: For dilute solutions, the equilibrium relation between the partial
pressure of component A in the gas phase, 𝑝A, and mole fraction in the liquid
phase, 𝑥 A, is expressed by a straight-line Henry’s law equation.

𝑝𝐴 = 𝑥𝐴 𝐻

where 𝐻 is Henry’s law constant in atm/mole fraction for the given system.

Dividing both sides by total pressure, 𝑃

𝑦𝐴 = 𝑝𝐴 ⁄𝑃 = 𝐻′ 𝑥𝐴 , 𝐻 ′ = 𝐻 ⁄𝑃

where 𝐻′ is the Henry’s law constant in mole frac gas/mole frac liquid and is
equal to 𝐻/𝑃. (𝐻′ depends on total pressure, whereas 𝐻 does not).

Relations between overall mass transfer coefficients and individual mass


transfer coefficients
Expressions for the overall mass transfer coefficients for gas and liquid phases
can be represented graphically as follows:

10
As can be seen from the graph, 𝑥𝐴 is uniquely related to 𝑦𝐴∗ for a given condition
of 𝑇 and 𝑃. From the graph:

𝑦𝐴 − 𝑦𝐴∗ = (𝑦𝐴 − 𝑦𝐴𝑖 ) + (𝑦𝐴𝑖 − 𝑦𝐴∗ ) = (𝑦𝐴 − 𝑦𝐴𝑖 ) + 𝑚′ (𝑥𝐴𝑖 − 𝑥𝐴 )

Therefore
𝑁𝐴 𝑁𝐴 𝑁𝐴 1 1 𝑚′
= + 𝑚′ ⟹ = +
𝐾𝑦 𝑘𝑦 𝑘𝑥 𝐾𝑦 𝑘𝑦 𝑘𝑥

Similarly, since 𝑥𝐴∗ is a measure of 𝑦𝐴 , it can be used to define another overall


coefficient 𝐾𝑥 as:

1 1 1
= ′′ +
𝐾𝑥 𝑚 𝑘𝑦 𝑘𝑥

For dilute solutions, with 𝑦𝐴 = 𝐻′ 𝑥𝐴∗ (i.e., assuming the equilibrium curve is
linear), the slope of the equilibrium curve is the Henry’s law constant. Hence,

1 1 𝐻′ 1 1 1
= + and = ′ +
𝐾𝑦 𝑘𝑦 𝑘𝑥 𝐾𝑥 𝐻 𝑘𝑦 𝑘𝑥

• 1/ 𝐾y = overall resistance to mass transfer (i.e., inverse of mass transfer


coefficient)
• 𝐻’/𝑘x = the resistance in liquid film
11
• 1/𝑘y = the resistance in gas film

𝑅𝑒𝑠𝑖𝑠𝑡𝑎𝑛𝑐𝑒 𝑖𝑛 𝑔𝑎𝑠 𝑝ℎ𝑎𝑠𝑒 1⁄𝑘𝑦


=
𝑇𝑜𝑡𝑎𝑙 𝑟𝑒𝑠𝑖𝑠𝑡𝑎𝑛𝑐𝑒 1⁄𝐾𝑦

𝑅𝑒𝑠𝑖𝑠𝑡𝑎𝑛𝑐𝑒 𝑖𝑛 𝑙𝑖𝑞𝑢𝑖𝑑 𝑝ℎ𝑎𝑠𝑒 1⁄𝑘𝑥


=
𝑇𝑜𝑡𝑎𝑙 𝑟𝑒𝑠𝑖𝑠𝑡𝑎𝑛𝑐𝑒 1⁄𝐾𝑥

Limiting cases
• If 𝑘 y and 𝑘 x are nearly equal and the slope is so small so that solute A is
relatively insoluble in liquid, the term 𝑚’/𝑘x will be negligible compared to
1/𝑘y. Hence,
1 1

𝐾𝑦 𝑘𝑦

We say the rate of mass transfer is gas-phase controlled (i.e., the overall
resistance lies in the gas phase only).

• Conversely, if 𝑚′′ is very large, then the solute A is relatively insoluble in


liquid. The term 1/𝑚′′ 𝑘y will be negligible compared to 1/kx. Then

1 1

𝐾𝑥 𝑘𝑥

We say the rate of mass transfer is liquid-phase controlled.

• For cases where 𝑘y and 𝑘x are not nearly equal, then it will be the relative
size of the ratio 𝑘𝑥 ⁄𝑘𝑦 and of slope which will determine the location of the
controlling mass transfer resistance.

Example 2.4
For a system in which component (A) is transferring from the liquid to the gas
phase, the equilibrium is given by 𝑦A∗= 0.75 𝑥A. At one point in the apparatus
the liquid contains 90 mol% of (A) and the gas contains 45 mol% of (A). The
individual gas film mass transfer coefficient at this point in the apparatus is
12
0.02716 kmol/m2.s, and 70% of the overall resistance to mass transfer is known
to be encountered in the gas film. Determine:

(a) The molar flux of (A).


(b) The interfacial concentration of (A).
(c) The overall mass transfer coefficient for liquid and gas phases.

Solution:

Gas film Liquid film


𝛿𝐺 𝛿𝐿
𝑥𝐴 = 0.9

𝑦𝐴∗

𝑥𝐴∗ 𝑥𝐴𝑖

𝑦𝐴𝑖

𝑦𝐴 = 0.45

Distance, z

Part (a)

1 1 1 1 𝑘𝑚𝑜𝑙
= 0.7 ( ) → = 0.7 ( ) ⟹ 𝐾𝑦 = 0.019
𝑘𝑦 𝐾𝑦 0.02716 𝐾𝑦 𝑚2 . 𝑠

𝑁𝐴 = 𝐾𝑦 (𝑦𝐴∗ − 𝑦𝐴 )

𝑦𝐴∗ = 0.75𝑥𝐴 = (0.75)(0.9) = 0.675


𝑘𝑚𝑜𝑙
𝑁𝐴 = (0.019)(0.675 − 0.45) = 4.274 × 10−3
𝑚2 . 𝑠

Part (b)

𝑁𝐴 = 𝑘𝑦 (𝑦𝐴𝑖 − 𝑦𝐴 ) → 4.274 × 10−3 = (0.02716)(𝑦𝐴𝑖 − 0.45)


𝑦𝐴𝑖 = 0.607

13
Part (c)

1 1 𝐻′ 1 1 0.75
= + → = + ⟹ 𝑘𝑥 = 0.0476
𝐾𝑦 𝑘𝑦 𝑘𝑥 0.019 0.02716 𝑘𝑥

1 1 1 1 1 1
= ′ + = +
𝐾𝑥 𝐻 𝑘𝑦 𝑘𝑥 𝐾𝑥 (0.75)(0.02716) 0.0476

𝑘𝑚𝑜𝑙
⟹ 𝐾𝑥 = 0.0142
𝑚2 . 𝑠

14

You might also like