Fundamentals of Mass Transfer

The previous chapters dealing with the transport phenomena of momentum

and heat transfer have dealt with one-component phases that possessed a
natural tendency to reach equilibrium conditions. When a system contains two
or more components whose concentrations vary from point to point, there is a
natural tendency for mass to be transferred, minimizing the concentration
differences within the system. The transport of one constituent from a region
of higher concentration to that of a lower concentration is called mass transfer.
Many of our day-to-day experiences involve mass transfer. A lump of
sugar added to a cup of black coffee eventually dissolves and then diffuses
uniformly throughout the coffee. Water evaporates from ponds to increase the
humidity of the passing air stream. Perfume presents a pleasant fragrance that
is imparted throughout the surrounding atmosphere. Mass transfer is the basis
for many biological and chemical processes. Biological processes include the
oxygenation of blood and the transport of ions across membranes within the
kidney. Chemical processes include the chemical vapor deposition (CVD) of
silane (SiH4) onto a silicon wafer, the doping of a silicon wafer to form a
semiconducting thin film, the aeration of wastewater, and the purification of
ores and isotopes.
Mass transfer underlies the various chemical separation processes
where one or more components migrate from one phase to the interface
between the two phases in contact. For example, in adsorption or
crystallization processes, the components remain at the interface, whereas in
gas absorption and liquid–liquid extraction processes, the components
penetrate the interface and then transfer into the bulk of the second phase.
If we consider the lump of sugar added to the cup of black coffee,
experience teaches us that the length of time required to distribute the sugar
will depend upon whether the liquid is quiescent or whether it is mechanically
agitated by a spoon. The mechanism of mass transfer, as we have also
observed in heat transfer, depends upon the dynamics of the system in which
it occurs. Mass can be transferred by random molecular motion in quiescent
fluids, or it can be transferred from a surface into a moving fluid, aided by the
dynamic characteristics of the flow. These two distinct modes of transport,
molecular mass transfer and convective mass transfer, are analogous to
conduction heat transfer and convective heat transfer. Each of these modes of
mass transfer will be described and analyzed. As in the case of heat transfer,
we should immediately realize that the two mechanisms often act
simultaneously. However, in the confluence of the two modes of mass
transfer, one mechanism can dominate quantitatively so that approximate
solutions involving only the dominant mode need be used.

MOLECULAR MASS TRANSFER

As early as 1815, Parrot observed qualitatively that whenever a gas mixture
contains two or more molecular species, whose relative concentrations vary
from point to point, an apparently natural process results, which tends to
diminish any inequalities of composition. This macroscopic transport of
mass, independent of any convection within the system, is defined as
molecular diffusion.
In the specific case of gaseous mixtures, a logical explanation of this
transport phenomenon can be deduced from the kinetic theory of gases. At
temperatures above absolute zero, individual molecules are in a state of
continual yet random motion. Within dilute gas mixtures, each solute
molecule behaves independently of the other solute molecules, as it seldom
encounters them. Collisions between the solute and the solvent molecules are
continually occurring. As a result of the collisions, the solute molecules move
along a zigzag path, sometimes toward a region of higher concentration,
sometimes toward a lower concentration.
Let us consider a hypothetical section passing normal to the
concentration gradient within an isothermal, isobaric gaseous mixture
containing solute and solvent molecules.
The two thin, equal elements of volume above and below the section
will contain the same number of molecules, as stipulated by Avogadro’s law.
Although it is not possible to state which way any particular molecule will
travel in a given interval of time, a definite number of the molecules in the
lower element of the volume will cross the hypothetical section from below,
and the same number of molecules will leave the upper element and cross the
section from above. With the existence of a concentration gradient, there are
more solute molecules in one of the elements of volume than in the other;
accordingly, an overall net transfer from a region of higher concentration to
one of lower concentration will result. The net flow of each molecular species
occurs in the direction of a negative concentration gradient.
As pointed out in Chapters 7 and 15, the molecular transport of
momentum and the transport of energy by conduction are also due to random
molecular motion. Accordingly, one should expect that the three transport
phenomena will depend upon many of the same characteristic properties, such
as mean free path, and that the theoretical analyses of all three phenomena
will have much in common.
The Fick Rate Equation
The laws of mass transfer show the relation between the flux of the diffusing
substance and the concentration gradient responsible for this mass transfer.
Unfortunately, the quantitative description of molecular diffusion is
considerably more complex than the analogous descriptions for the molecular
transfer of momentum and energy that occur in a one component phase. As
mass transfer, or diffusion, as it is also called, occurs only in mixtures, its
evaluation must involve an examination of the effect of each component. For
example, we will often desire to know the diffusion rate of a specific
component relative to the velocity of the mixture in which it is moving. As
each component may possess a different mobility, the mixture velocity must
be evaluated by averaging the velocities of all the components present.
In order to establish a common basis for future discussions, let us first
consider definitions and relations that are often used to explain the role of
components within a mixture.
Concentrations. In a multicomponent mixture, the concentration of a
molecular species can be expressed in many ways. Figure 24.1 shows an
elemental volume dV that contains a mixture of components, including species
A. As each molecule of each species has a mass, a mass concentration for each
species, as well as for the mixture, can be defined. For species A, mass
concentration, A, is defined as the mass of A per unit volume of the mixture.
The total mass concentration or density, , is the total mass of the mixture
contained in the unit volume; that is,

where n is the number of species in the mixture. The mass fraction, A, is the
mass concentration of species A divided by the total mass density
The molecular concentration of species A, cA, is defined as the number of
moles of A present per unit volume of the mixture. By definition, one mole of
any species contains a mass equivalent to its molecular weight; the mass
concentration and molar concentration terms are related by the following
relation:

where MA is the molecular weight of species A. When dealing with a gas

phase, concentrations are often expressed in terms of partial pressures. Under
conditions in which the ideal gas law, pAV = nART, applies, the molar
concentration is

where pA is the partial pressure of the species A in the mixture, nA is the

number of moles of species A, V is the gas volume, T is the absolute
temperature, and R is the gas constant. The total molar concentration, c, is the
total moles of the mixture contained in the unit volume; that is,

or for a gaseous mixture that obeys the ideal gas law, c = ntotal/V = P/RT, where
P is the total pressure. The mole fraction for liquid or solid mixtures, xA, and
for gaseous mixtures, yA, are the molar concentrations of species A divided by
the total molar density

For a gaseous mixture that obeys the ideal gas law, the mole fraction, yA, can
be written in terms of pressures
For a gaseous mixture that obeys the ideal gas law, the mole fraction, yA, can
be written in terms of pressures

A summary of the various concentration terms and of the interrelations for a

binary system containing species A and B is given in Table 24.1.

See EXAMPLE 1
Velocities. In a multicomponent system the various species will normally
move at different velocities; accordingly, an evaluation of a velocity for the
gas mixture requires the averaging of the velocities of each species present.
The mass-average velocity for a multicomponent mixture is defined in
terms of the mass densities and velocities of all components by

where vi denotes the absolute velocity of species i relative to stationary

coordinate axes. This is the velocity that would be measured by a pitot tube
and is the velocity that was previously encountered in the equations of
momentum transfer. The molar-average velocity for a multicomponent
mixture is defined in terms of the molar concentrations of all components by

The velocity of a species relative to the mass-average or molar-average

velocity is termed a diffusion velocity. We can define two different diffusion
velocities
vi - v; the diffusion velocity of species i relative to the mass-average velocity
and
vi - V, the diffusion velocity of species i relative to the molar-velocity average
According to Fick’s law, a species can have a velocity relative to the mass- or
molar-average velocity only if gradients in the concentration exist.

Fluxes. The mass (or molar) flux of a given species is a vector quantity
denoting the amount of the species, in either mass or molar units, that passes
per given increment of time through a unit area normal to the vector. The flux
may be defined with reference to coordinates that are fixed in space,
coordinates that are moving with the mass average velocity, or coordinates
that are moving with the molar-average velocity.
The basic relation for molecular diffusion defines the molar flux
relative to the molar average velocity, J A. An empirical relation for this molar
flux, first postulated by Fick and, accordingly, often referred to as Fick’s first
law, defines the diffusion of component A in an isothermal, isobaric system:

For diffusion in only the z direction, the Fick rate equation is

where J A,z is the molar flux in the z direction relative to the molar-average
velocity, dcA/dz is the concentration gradient in the z direction, and DAB, the
proportionality factor, is the mass diffusivity or diffusion coefficient for
component A diffusing through component B.
A more general flux relation that is not restricted to isothermal, isobaric
systems was proposed by de Groot who chose to write

or

As the total concentration c is constant under isothermal, isobaric conditions,

equation (24-15) is a special form of the more general relation (24-16). An
equivalent expression for jA,z, the mass flux in the z direction relative to the
mass-average velocity, is

where dA/dz is the concentration gradient in terms of the mass fraction.

When the density is constant, this relation simplifies to

Initial experimental investigations of molecular diffusion were unable

to verify Fick’s law of diffusion. This was apparently due to the fact that mass
is often transferred simultaneously by two possible means: (1) as a result of
the concentration differences as postulated by Fick and (2) by convection
differences induced by the density differences that resulted from the
concentration variation. Steffan (1872) and Maxwell (1877), using the kinetic
theory of gases, proved that the mass flux relative to a fixed coordinate was a
result of two contributions: the concentration gradient contribution and the
bulk motion contribution.
For a binary system with a constant average velocity in the z direction,
the molar flux in the z direction relative to the molar-average velocity may
also be expressed by

Equating expressions (24-16) and (24-18), we obtain

which, upon rearrangement, yields

For this binary system, Vz can be evaluated by equation (24-14) as

Substituting this expression into our relation, we obtain

As the component velocities, vA,z and vB,z, are velocities relative to the fixed z
axis, the quantities cAvA,z and cBvB,z are fluxes of components A and B relative
to a fixed z coordinate; accordingly, we symbolize this new type of flux that
is relative to a set of stationary axes by

NA = cAvA and NB = cBvB

Substituting these symbols into equation (24-19), we obtain a relation for the
flux of component A relative to the z axis

This relation may be generalized and written in vector form as

It is important to note that the molar flux, NA, is a resultant of the two vector
quantities:

Either or both quantities can be a significant part of the total molar flux, NA.
Whenever equation (24-21) is applied to describe molar diffusion, the vector
nature of the individual fluxes, NA and NB, must be considered and then, in
turn, the direction of each of two vector quantities must be evaluated.
If species A were diffusing in a multicomponent mixture, the expression
equivalent to equation (24-21) would be

where DAM is the diffusion coefficient of A in the mixture.

The mass flux, nA, relative to a fixed spatial coordinate system, is
defined for a binary system in terms of mass density and mass fraction by
If a balloon, filled with a color dye, is dropped into a large lake, the dye will
diffuse radially as a concentration gradient contribution. When a stick is
dropped into a moving stream, it will float downstream by the bulk motion
contribution. If the dye-filled balloon were dropped into the moving stream,
the dye would diffuse radially while being carried downstream; thus both
contributions participate simultaneously in the mass transfer.
The four equations defining the fluxes, JA, jA, NA, and nA are equivalent
statements of the Fick rate equation. The diffusion coefficient, DAB, is identical
in all four equations. Any one of these equations is adequate to describe
molecular diffusion; however, certain fluxes are easier to use for specific
cases. The mass fluxes, nA and jA, are used when the Navier–Stokes equations
are also required to describe the process. Since chemical reactions are
described in terms of moles of the participating reactants, the molar fluxes, J A
and NA, are used to describe mass-transfer operations in which chemical
reactions are involved. The fluxes relative to coordinates fixes in space, nA
and NA, are often used to describe engineering operations within process
equipment. The fluxes JA and jA are used to describe the mass transfer in
diffusion cells used for measuring the diffusion coefficient. Table 24.2
summarizes the equivalent forms of the Fick rate equation.

24.2 THE DIFFUSION COEFFICIENT

Fick’s law of proportionality, DAB, is known as the diffusion coefficient. Its
fundamental dimensions, which may be obtained from equation (24-15)
are identical to the fundamental dimensions of the other transport properties:
kinematic viscosity, , and thermal diffusivity, , or its equivalent ratio, k/cp.
The mass diffusivity has been reported in cm2/s; the SI units are m2/s, which
is a factor 10-4 smaller. In the English system ft 2/h is commonly used.
Conversion between these systems involves the simple relations

The diffusion coefficient depends upon the pressure, temperature, and

composition of the system. Experimental values for the diffusivities of gases,
liquids, and solids are tabulated in Appendix Tables J.1, J.2, and J.3,
respectively. As one might expect from the consideration of the mobility of
the molecules, the diffusion coefficients are generally higher for gases (in the
range of 5×10-6 to 1×10-5 m2/s), than for liquids (in the range of 10-10 to 10- 9
m2/s), which are higher than the values reported for solids (in the range of 10
-14 to 10 -10 m2/s).

In the absence of experimental data, semitheoretical expressions have

been developed which give approximations, sometimes as valid as
experimental values due to the difficulties encountered in their measurement.

Gas Mass Diffusivity

Using a similar kinetic theory of gases approach for a binary mixture of
species A and B composed of rigid spheres of unequal diameters, the gas-
phase diffusion coefficient is shown to be

Unlike the other two molecular transport coefficients for gases, the viscosity
and thermal conductivity, the gas-phase diffusion coefficient is dependent on
the pressure and the temperature. Specifically, the gas-phase diffusion
coefficient is
As equation (24-31) reveals, and as one of the problems at the end of this
chapter points out, the diffusion coefficients for gases DAB = DBA. This is not
the case for liquid diffusion coefficients.
Hirschfelder et al. (1949) using the Lennard–Jones potential to evaluate
the influence of the molecular forces, presented an equation for the diffusion
coefficient for gas pairs of nonpolar, nonreacting molecules:

where DAB is the mass diffusivity of A through B, in cm2/s; T is the absolute

temperature, in K; MA, MB are the molecular weights of A and B, respectively;
P is the absolute pressure, in atmospheres; AB is the ‘‘collision diameter,’’ a
Lennard–Jones parameter, in A; and D is the ‘‘collision integral’’ for
molecular diffusion, a dimensionless function of the temperature and of the
intermolecular potential field for one molecule of A and one molecule of B.
Appendix Table K.1 lists D as a function of T/AB,  is the Boltzmann
constant, which is 1.38 × 10-16 ergs/K, and AB is the energy of molecular
interaction for the binary system A and B, a Lennard–Jones parameter, in
ergs, see equation (24-31). Unlike the other two molecular transport
coefficients, viscosity and thermal conductivity, the diffusion coefficient is
dependent on pressure as well as on a higher order of the absolute temperature.
When the transport process in a single component phase was examined, we
did not find any composition dependency in equation (24-30) or in the similar
equations for viscosity and thermal conductivity. Figure 24.2 presents the
graphical dependency of the ‘‘collision integral,’’ D, on the dimensionless
temperature, T/AB.
For a binary system composed of nonpolar molecular pairs, the Lennard–
Jones parameters of the pure component may be combined empirically by the
following relations:
Simplifying equation (24-33), we can predict the diffusion coefficient at any
temperature and at any pressure below 25 atm from a known experimental
value by

See Example 2

The empirical correlation recommended by Fuller, Schettler, and Giddings

permits the evaluation of the diffusivity when reliable Lennard –Jones
parameters, i and i, are unavailable. The Fuller correlation is

where DAB is in cm2/s, T is in K, and P is in atmospheres. To determine the v

terms, the authors recommend the addition of the atomic and structural
diffusion-volume increments v reported in Table 24.3.
See Example 3

Liquid-Mass Diffusivity
Wilke and Chang have proposed the following correlation for nonelectrolytes
in an infinitely dilute solution:

where DAB is the mass diffusivity of A diffusing through liquid solvent B, in

cm2/s; B is the viscosity of the solution, in centipoises; T is absolute
temperature, in K; MB is the molecular weight of the solvent; VA is the molal
volume of solute at normal boiling point, in cm3/g mol; and B is the
‘‘association’’ parameter for solvent B.
See Example 5

Pore Diffusivity
Knudsen diffusion. Consider the diffusion of gas molecules through very
small capillary pores. If the pore diameter is smaller than the mean free path
of the diffusing gas molecules and the density of the gas is low, the gas
molecules will collide with the pore walls more frequently than with each
other. This process is known as Knudsen flow or Knudsen diffusion. The gas
flux is reduced by the wall collisions.
The Knudsen number, Kn, given by

If the Kn number is much greater than one, then Knudsen diffusion can be
important. At a given pore diameter, the Kn number goes up as the total system
pressure P decreases and absolute temperature T increases.
For Knudsen diffusion, we replace path length  with pore diameter dpore, as
species A is now more likely to collide with the pore wall as opposed to
another molecule. In this instance, the Knudsen diffusivity for diffusing
species A, DKA, is

This simplified equation requires that dpore has units of cm, MA has units of
g/mol, and temperature T has units of K. The Knudsen diffusivity, DKA, is
dependent on the pore diameter, species A molecular weight, and temperature.
We can make two comparisons of DKA to the binary gas phase diffusivity,
DAB. First, it is not a function of absolute pressure P, or the presence of species
B in the binary gas mixture. Second, the temperature dependence for the
Knudsen diffusivity is DKAαT1/2, vs. DAB αT3/2 for the binary gas phase
diffusivity.

The above relationships for the effective diffusion coefficient are based on
diffusion within straight, cylindrical pores aligned in a parallel array.
However, in most porous materials, pores of various diameters are twisted and
interconnected with one another, and the path for diffusion of the gas molecule
within the pores is ‘‘tortuous.’’ For these materials, if an average pore
diameter is assumed, a reasonable approximation for the effective diffusion
coefficient in random pores is

 is the volume void fraction of the porous volume within the porous material.
This ‘‘void fraction’’ is usually experimentally determined for a specific
material. The four possible types of pore diffusion are illustrated in Figure
24.3, each with their respective diffusivity correlation. The first three, pure
molecular diffusion, pure Knudsen diffusion, and Knudsen and molecular
combined diffusion, are based on diffusion within straight, cylindrical pores
that are aligned in parallel array. The fourth involves diffusion via ‘‘tortuous
paths’’ that exist within the compacted solid.

See Example 6
24.3 CONVECTIVE MASS TRANSFER
Mass transfer between a moving fluid and a surface or between immiscible
moving fluids separated by a mobile interface (as in a gas/liquid or
liquid/liquid contactor) is often aided by the dynamic characteristics of the
moving fluid. This mode of transfer is called convective mass transfer, with
the transfer always going from a higher to a lower concentration of the species
being transferred. Convective transfer depends on both the transport
properties and the dynamic characteristics of the flowing fluid. As in the case
of convective heat transfer, a distinction must be made between two types of
flow. When an external pump or similar device causes the fluid motion, the
process is called forced convection. If the fluid motion is due to a density
difference, the process is called free or natural convection.
The rate equation for convective mass transfer, generalized in a manner
analogous to Newton’s ‘‘law’’ of cooling, equation 15.11 is

where NA is the molar mass transfer of species A measured relative to fixed

spatial coordinates, CA is the concentration difference between the boundary
surface concentration and the average concentration of the fluid stream of the
diffusing species A, and kc is the convective mass-transfer coefficient.
As in the case of molecular mass transfer, convective mass transfer occurs in
the direction of a decreasing concentration. Equation (24-68) defines the
coefficient kc in terms of the mass flux and the concentration difference from
the beginning to the end of the mass transfer path. The reciprocal of the
coefficient, 1/kc, represents the resistance to the transfer through the moving
fluid. Chapters 28 and 30 consider the methods of determining this coefficient.
It is, in general, a function of system geometry, fluid and flow properties, and
the concentration difference CA.
See Example 7