CHE 306: TRANSPORT PHENOMENA NOTES

RYAN GODIN

1. Mass Balance
Transport Phenomena relies on a number of material/energy balances to cal-
culate useful things for a system. For example, the mass balance can be used to
calculate volumetric flow rates and the change of mass in a system. Similarly,
the momentum balance can be used to calculate the forces acting on a pipe bend
from fluid flow. The majority of the conservation equations use the control volume
approach and are calculated from the Reynold’s Transport Theorem:
Theorem 1.1 (Reynold’s Transport Theorem). For a given ”fluid” property, N
with specific property η, the change of that property per unit time in the system is
given by: ! " "
dN ∂ #‰ #‰
= ηρdVV– + ηρV · d A
dt system ∂t CV CS

! Main Idea: Here, the system refers to the group of objects (fluids or particles)
under study. While these objects can move through the control volume (CV), they
cannot leave the system!
○ Danger: Since objects have a defined mass, we cannot add or subtract mass
to the system without changing the system under study (conservation of mass)!.
Thus, when N = m, we have:
!
dm
= 0. (Conservation of Mass)
dt system

We will now use this theorem to derive the conservation of mass formula. This is
otherwise known as the equation of continuity! Plugging in the appropriate values
for N and η, we have:

! 0 (Conservation of"Mass)
dm ✟✟✟ ✯ ∂ " #‰ #‰
✟ = ρdV
V– + ρV · d A (1.1)
✟ ✟
dt system ∂t CV CS
" "
∂ #‰ #‰
0= ρdV
V– + ρV · d A (1.2)
∂t CV CS
Now, noticing that the first term on the right hand site equals the change of
mass in the control volume and that the second term represents the mass flow in
and out through the control surfaces, we have the following:

Date: January 29, 2022.

1
2 RYAN GODIN

!
dm
0= + φm,out − φm,in (1.3)
dt CV
!
dm
= φm,in − φm,out (Mass Balance)
dt CV

# Careful: Notice that in Equation (1.3), we have flow out minus flow in! This is
because flow into a control surface is defined to be negative and that out is defined
to be positive!
Therefore, we have arived at the mass continuity balance as desired. As expected,
mass flow into the system equals mass flow out of the system.

2. Momentum Balance
From the above equation, we can similarly derive the momentum balance. First,
#‰ #‰
we plug in the appropriate variables into Theorem 1.1 knowing that P = mV to
arrive at the following:
#‰ # " "
dP ∂ #‰ #‰ #‰ #‰
= V ρdV
V– + V ρV · d A (2.1)
dt ∂t CV CS
system

$ Problem-Solving Strategy: To make calculations easier, the above equation

is usually broken down into the velocity components (e.x. x, y, and z).
#‰
dP #‰
Now, since = F , we know that the change of momentum of a system is
dt #‰
equivalent to the net force acting on that system, F system . Furthermore, since in
deriving the Reynold’s Transport Theorem the system and control volume coincided
#‰ #‰
at t = 0, it follows that F system = F CV . Assuming that the only forces acting on
our control volume are surface and body forces, we have:

" "
#‰ #‰ ∂ #‰ #‰ #‰ #‰
F surf ace + F body = V ρdV
V– + V ρV · d A (Momentum Balance)
∂t CV CS

2.1. Surface Forces. In most fluid mechanics problems, surfaces forces result only
from the pressure of the surroundings acting onto the control volume. Therefore,
in most applications, we have the following where p is the pressure of the system:
"
#‰ #‰
F surf ace = − pd A (2.2)
CS

$ Problem-Solving Strategy: The negative sign is needed since pressure acts

onto the control volume. In most applications, the pressure distribution along a
control surface is uniform so we can simplify this to:
"
#‰ #‰
F surf ace = − pd A (2.3)
CS
$
= pi Ai . (2.4)
CS
 CHE 306: TRANSPORT PHENOMENA NOTES 3

This is usually evaluated component wise, and the pressures are taken to be
positive when they act in the direction of their component (pressure pointing in
positive x-direction is positive).
○ Danger: Here, the pressure on ALL sides of the CV must be taken into account,
not just those control surfaces that have an inflow/outflow. You also need to make
sure that the directions are correct!
However, in some situations, a force needs to be applied to the fluid to keep the
#‰
system stationary, such as in a pipe. This force is dubbed the reaction force, R,
and, in these situations, we have:
"
#‰ #‰ #‰
F surf ace = − pd A + R (2.5)
CS

#‰
2.2. Body Forces. In most problems, the body force is just gravity, F g . However,
one must be careful to only consider it in dimensions in which it acts.
○ Danger: Gravity and other forces can’t act (are equal to 0) in directions perpen-
dicular to the force direction! For example, if gravity is acting in the z-direction,
then:
#‰ #‰ #‰
F g,x = 0; F g,y = 0; F g,z = msystem #‰
g

3. Energy Balance
Like the mass balance, the energy balance is probably one of the most use-
ful/fundamental equations in transport phenomena. We derive it as follows by
plugging the appropriate values into the Reynold’s Transport Theorem:
! " "
dE ∂ #‰ #‰
= eρdV
V– + eρV · d A (3.1)
dt system ∂t CV CS

Now, from the first law of thermodynamics, we know that the following relation-
ship holds where Q̇ is the rate of heat transfered to the system and Ẇ is the rate
of work done by the system:
!
dE
= Q̇ − Ẇ (3.2)
dt system
Making this substitution and simplifying the notation, we arrive at the following
version of the energy balance:
!
dE
= ein φm,in − eout φm,out + φq − φw + Ėproduced (3.3)
dt CV
This is probably the most basic form of the energy balance. However, in practice,
it is often difficult to calculate e experimentally. Thus, the substitution e = u +
1 2
2 v +gz is usually made to obtain the following form of the energy balance equation:

! % & % &
dE 1 1
= φm,in u + v 2 + gz −φm,out u + v 2 + gz +φq −φw +Ėproduced
dt CV 2 in 2 out
(3.4)
4 RYAN GODIN

Further simplification of the above equation is also useful. However, before this
simplification can be done, we first need to discuss they type of work that can occur
in our system and act upon the control volume.
# Careful: From now on in the discussion, we will assume that the Ėproduced
term is equivalent to 0 and will ignore it. I’m not sure yet in what situations it
becomes relevant.
○ Danger: Energy provided by a heating coil is usually written as φq and not
electrical work, φw !
3.1. Work. In transport phenomena, we mainly deal with three different kinds
of work: work done by shear stress at the control surface (Wshear ), work done
by the normal stress at the control surface (Wσ ), shaft work (Wshaf t ), and other,
situational work depending on the type of problem such as electrical work (Wother ).
Therefore, we have the following equation:
Ẇ = Ẇshaf t + Ẇσ + Ẇshear + Ẇother (3.5)
We will now go on to discuss Wσ and Wshear in more detail so that we can learn
how to calculate them!

3.1.1. Work Done by Normal Stress. The normal stress, σii , is essentially a type of
stress that acts perpendicular onto the control surface. Since it is a stress, it has
units of force per areas. Thus, if we have a differential element, the force applied
by the normal stress on that area is given by:
#‰ #‰
dF normal = σii d A CS (3.6)
Since work is equivalent to the Force multiplied over a distance, if we take a
differential distance element, d #‰
s , the work applied to the differential area over that
distance is given by:
#‰
δW normal =F normal d #‰
s (3.7)
Now, if we divide by time and take the limit as ∆t goes to 0, we get the rate of
work done by normal stress as:
#‰ #‰ #‰
δ Ẇnormal = F normal d A · V (3.8)
#‰ #‰
= σii d A · V (3.9)
Integrating this we then get the following equation for the rate of work done by
the normal stress to the control volume, Ẇnormal :
"
#‰ #‰
Ẇnormal = σii V · d A (3.10)

○ Danger: Since we took work to act by the control volume to the outside, we
actually have to take the negative of this work. Thus, Ẇnormal is actually given
by:
"
#‰ #‰
Ẇnormal = − σii V · d A (3.11)

In a good proportion of transport phenomena problems, we can assume that

viscous effects are negligible. In these cases, the normal stress, σii is equivalent to
 CHE 306: TRANSPORT PHENOMENA NOTES 5

the pressure acting onto the control surface: σii ≃ −p. Therefore, we can effectively
calculate Ẇnormal as:
"
#‰ #‰
Ẇnormal = pV · d A (3.12)

Now, as a consequence of this relation, we can actually write down our energy
balance in an alternate form that includes both the mechanical and thermal energy
balances as follows:
! % & % &
dE 1 2 1 2
= φm,in u + v + gz − φm,out u + v + gz + φq − φw,normal
dt CV 2 in 2 out
− φw,shear − φw,shaf t − φw,other (3.13)
= φm,in ein − φm,out eout + φq − φw,normal − φw,shear − φw,shaf t − φw,other
(3.14)
= φm,in ein − φm,out eout + φq − φw − φw,normal (3.15)
"
#‰ #‰
= φm,in ein − φm,out eout + φq − φw − pV · d A (3.16)
"
p #‰ #‰
= φm,in ein − φm,out eout + φq − φw − ρV · d A (3.17)
ρ
' !
p p
= φm,in ein − φm,out eout + φq − φw − φm,out − φm,in (3.18)
ρ out ρ in
! % & % &
dE 1 p 1 p
= φm,in u + v 2 + gz + − φm,out u + v 2 + gz + + φq − φw
dt CV 2 ρ in 2 ρ out
(3.19)
In making these simplifications, we essentially leave work to include terms that
we can straightforwardly compute. However, some further simplification and group-
ing is necessary to get the above equation in a form reminiscent of the thermal and
mechanical energy balances! This is done as follows:
! Main Idea: Since h = u + pv = u + ρp , the above equation essentially reduces
down to:
! % & % &
dE 1 2 1 2
= φm,in h + v + gz − φm,out h + v + gz + φq − φw (3.20)
dt CV 2 in 2 out

This is equivalent to the original energy balance except now we take work done
by normal stress and put it into the convective transport terms. Thus, it is not
included in φw !
# Careful: In the above equations, work is negative if done to the control volume
which will make the energy provided positive! Furthermore, φw is equivalent to the
sum of all forms of work besides the normal work!
! % & % &
dE 1 p 1 p
= φm,in u + v 2 + gz + − φm,out u + v 2 + gz + + φq − φw
dt CV 2 ρ in 2 ρ out
(3.21)
% ' ! & % ' ! &
1 2 p 1 2 p
= φm,in v + gz + − φm,out v + gz +
2 ρ in 2 ρ out
+ φm,in uin − φm,out uout + φq − φw (3.22)
6 RYAN GODIN

Now, if we take ef r φm to represent all the work and heat flow terms in the above
equation related to heat loss by friction (such as shear work), then we can write
the following mechanical and thermal energy balances where we have:
! ! !
dE dE dE
= + (3.23)
dt CV dt CV,mechanical dt CV,thermal
! ' ! ' !
dE 1 2 p 1 2 p
= φm,in v + gz + − φm,out v + gz +
dt CV,mechanical 2 ρ in 2 ρ out
+ φ w − ef r φ m (3.24)
# Careful: In the above equation, the sign of φw was switched, so now we are
assuming that it is positive if it is acting onto the control volume.
!
dE
= φm,in uin − φm,out uout + φq + ef r φm (3.25)
dt CV,thermal
Since from thermodynamics we have du = Cv dT ≃ Cp dT = dh for liquids and
solids, this equation is often rewritten as:
!
dE
= φm,in Cp,in Tin − φm,out Cp,out Tout + φq + ef r φm (3.26)
dt CV,thermal
! Main Idea: Since ef r φm is always positive, the above two equations essentially
state that heat loss by friction is always from mechanical energy to thermal energy,
never the other way around!
3.1.2. Bernoulli’s Equation. Notice that the mechanical energy balance looks very
similar to the Bernoulli Equation. We can actually derive the Bernoulli Equation
from the mechanical energy balance if we make the following assumptions:
!
dE
(1) Steady flow: This assumption allows us to cancel out
dt CV,mechanical
from the equation since it is equal to 0.
(2) Incompressible flow: This implies that the densities are the same at each
point!
(3) Frictionless flow: This allows us to cancel out the friction term ef r φm .
Furthermore, combined with the incompressible flow assumption, the con-
vective (bulk) heat transfer and heat transfer by convection (φq ) cancel
out.
(4) Flow along a streamline: This implies that the mass transfer in and
out remains constant since it cannot pass through a streamline (veloc-
ity/direction is tangent).
(5) No work happens along the streamline between the two points
3.1.3. Work Done by Shear Stress. Similar to the work done by the normal stress,
the work done by the shear stress only applies at the control surface boundaries. It
can be calculated by:
#‰
dF shear = #‰
τ dA (3.27)
#‰
Multiplying it by a differential distance, d s , dividing by ∆t and taking the limit
as ∆t → 0, we have:
#‰
δ Ẇshear = #‰
τ · V dA (3.28)
 CHE 306: TRANSPORT PHENOMENA NOTES 7

Integrating both sides and taking the negative of the integration, we can calculate
the work done by shear stress as:
"
#‰
Ẇshear = − #‰ τ · V dA. (3.29)

# Careful: This can only be calculated at control surfaces! Thus, even though
the fluid molecules enact a shear stress on each other, there is no work done since
energy is transfered between molecules of the system, not to or from the system!
#‰
$ Problem-Solving Strategy: At solid surfaces, V = 0 by the no-slip condition
so shear stresses do not occur at (non-moving) solid surface boundaries. If the solid
#‰
is moving relative to the control volume than V ∕= 0.
3.2. Alternative Energy Balances. Using the energy balance and some useful
relations, you can convert the energy balance to a number of useful balances such as
heat or enthalpy balances. We will now discuss these “alternative energy balances.”
3.3. Heat Balance. A heat balance is essentially just a temperature balance.
Thus, to write one, we need to figure out a way to re-write the energy balance
so that it describes a change of temperature. This can be done relatively easily if
we ignore any change in kinetic or potential energy of the system (CV) and make
the following two substitutions:
! !
dE dU
= and u ≃ Cp T
dt CV dt CV
Now, to get the left hand side be a derivative of temperature, we can make the
substitution UCV = mCV Cp T . If the mass of the system does not change, then we
can just take the mCV and Cp out of the derivative to arrive at our heat balance:
!
dT
mCV Cp,CV = φm,in Cp,in Tin −φm,out Cp,out Tout +φq +ef r φm (3.30)
dt CV,thermal
3.3.1. Enthalpy Balance. We know that enthalpy is given by the relation: dh =
Cp dT . Thus, if we make the substitution h = Cp T into the (thermal) energy
balance, we’ve effectively created an enthalpy balance:
!
dE
= φm,in hin − φm,out hout + φq + ef r φm (3.31)
dt CV,thermal
However, unless we are at steady state, we must also find a way to replace the
E term representing the energy of the control volume. We can do this by using the
relation: !
dE dU dEkinetic dEpotential
= + + (3.32)
dt system dt dt dt
Specifically, if we assume that the kinetic and potential energy of the system
(CV) does not change, then we can effectively assume that:
! !
dE dE
= (3.33)
dt system dt system
Furthermore, since du = Cv dT ≃ Cp dT = h, we can effectively rewrite the
previous equation as: ! !
dE dH
= (3.34)
dt system dt system
8 RYAN GODIN

Making the appropriate substitution into we finally arrive at our enthalpy bal-
ance as desired:
!
dH
= φm,in hin − φm,out hout + φq + ef r φm (3.35)
dt CV,thermal

4. Drag Force
The drag force is the force an object experiences when traveling through a
medium such as air or water. It is a result of pressure and frictional forces on
the object and it impedes its movement. However, it acts in the same direction
as the relative fluid flow. The amount of drag force experienced by an object is
dependent upon three factors:
(1) Velocity of the object;
(2) Area of the object;
(3) Fluid properties (µ and ρ).
The equations for calculating drag force depend on if the flow is laminar or
turbulent. As a reminder, whether a flow is turbulent depends on the Reynold’s
number and, for pipe flow, we have laminar flow if:
ρDv
Re = < 2300. (4.1)
µ
If there is only frictional drag with a negligible pressure difference between the
front and back of our object, then this is given by the following equation where vr
is the relative velocity of the object and A is the area of the object in contact with
the fluid:
ρv 2
FD = C f r A (4.2)
2
% Notation: If the body is not moving, then vr is usually written as v∞ and is
called the free-stream fluid velocity.
However, in the case that the geometry of the body causes a significant separation
in the streamlines which results in a pressure gradient between the front and back,
this creates what is called form drag. In this case, friction and form drag contribute
to the drag force and it is instead written as follows where CD is the drag coefficient
and AP is the projected area (normally the maximum projected area):
ρvr2
FD = C D AP (4.3)
2
vr2
% Notation: The term 2 is usually referred to as the dynamic pressure.
$ Problem-Solving Strategy: The maximal projected area is calculated by
imagining what 2-D shape the object would look like if viewed straight on. For a
sphere, this is a circul with a radius equal to that of a sphere. Thus,
AP = πr2 (4.4)

The relationship between the Reynold’s number and the drag coefficient, CD ,
can be divided into four different flow regimes:
 CHE 306: TRANSPORT PHENOMENA NOTES 9

(1) Regime 1: Re < 1. In this regime, the flow is laminar and the the
streamlines stick very closely to the body of the object. Thus, in this case,
frictional drag predominates and there is very little form drag. This flow
is called ”creeping” or Stokes flow. In this case, we have the following
relationship:
24
CD = (4.5)
Re
(2) Regime 2: 1 ≤ Re < 103 . In this case, the streamlines behind the object
get disturbed and we have a bit of form drag beginning to develop.
(3) Regime 3: 103 ≤ Re < 5 · 105 . In the third regime the point of flow
separation stabilizes at a point about 80 degrees from the forward stagna-
tion point. The wake is no longer characterized by large eddies, although
it remains unsteady. The flow on the surface of the body from the stag-
nation point to the point of separation is laminar, and the shear stress in
this interval is appreciable only in a thin layer near the body. The drag
coefficient levels out at a near-constant value of approximately 1.
(4) Regime 4: 5 · 105 ≤ Re. Specifically, one may observe the same sharp
decrease in CD to a minimum value near a Reynolds number value of 5 ·
105 . This is again due to the change from laminar to turbulent flow in the
boundary layer.
For Stoke’s flow around a sphere then, we can derive formula for the drag force,
FD , acting on the sphere as follows:
1
FD = CD ρvr2 AP (4.6)
2
24 1 2
= ρv AP (4.7)
Re 2 r
24µ 1 2 D2
= ρv π (4.8)
ρvr D 2 r 4
FD = 3πµvr D (4.9)
For the above equation, 2/3 of the drag force is a result of viscous (friction) drag
and the rest is a result of pressure (form) drag. In other words, this can be written
out as:
FD = 2πµvr D + πµvr D (4.10)
( )* + ( )* +
viscous drag form drag

5. Boundary Layer Concept

The boundary layer represents the small area of fluid flow around an object where
we have no pressure gradient (just a constant pressure) and in which a velocity
profile exists. Specifically, inside of the boundary layer, we have the following:
(1) No pressure gradient (constant pressure)
(2) Velocity profile
(3) Viscous flow (boundary layer is viscous layer)
However, outside of the boundary layer, we have the fluid engaged in a constant
free-stream flow. Specifically, the following holds:
(1) Inviscid flow (neglect friction)
(2) Incompressible flow
(3) Steady flow (constant velocity)
10 RYAN GODIN

(4) Bernoulli’s equation applies!

Often, it is useful to be able to determine the boundary layer thickness, δ. As we
will show, this value is often used to calculate a number of useful properties such
as the velocity profile and the drag force that an object experiences. The boundary
layer thickness, δ, around a small plate is a function of x, the distance along the
plate. Specifically, it is defined to be the height at which the velocity obtains 99%
of the free-stream velocity. To figure out δ for a flate-plate, it is useful to define the
local Reynold’s number, Rex , which only applies at the specified distance, x, along
the plate:
ρv∞ x
Rex = (5.1)
µ
For the local Reynold’s number, Rex , we have laminar flow when Rex < 2 × 105
and turbulent flow when Rex > 3 × 106 . In between these two values we have a
flow transition.

5.1. Laminar Flow. The solutions for δ and other values depend on whether we
have laminar or turbulent flow. If we have laminar flow, Rex < 2 × 105 , then the
boundary layer thickness at a point x is defined by:
5x
δ=√ (5.2)
Rex
Now, the drag experienced by an object with a boundary layer is only viscous
(friction) drag since there is no pressure gradient! The friction drag coefficient
(coefficient of skin friction), Cf x , at a point, x, is given by the following formula:
0.664
Cf x = √ (5.3)
Rex
Since we have a flat plate, we may integrate Cfx over the length of the plate to
arrive at the drag coefficient, CfL (CD ), where w is the width of the plate:
"
1
C fL = Cf dA (5.4)
A A x
"
1 0.664
= √ dA (5.5)
wL A Rx
" L
1
= 0.664w , dx (5.6)
0 µ
ρv∞ x

1.328
=, (5.7)
ρv∞ L
µ
1.328
C fL = √ (5.8)
ReL
# Careful: The new Reynold’s number, ReL , is defined by:
ρv∞ L
ReL = (5.9)
µ
 CHE 306: TRANSPORT PHENOMENA NOTES 11

The drag force for the parallel plate setup can then be calculated as follows,
where A is the area of the plate in contact with fluid, not the projected area, AP !
vr2
FD = CfL Aρ (5.10)
2
○ Danger: In the above equation, A is the area of the plate in contact with fluid,
not the projected area, AP !
# Careful: Make sure to know if the flow is on both sides of the plate or not! If
so, you will need to multiply the area by 2!
The velocity distribution is given by the following equations. Note that we can
only really evaluate it at specific x values!
vx y -
= Rex (5.11)
v∞ 2x
.
vx y νx
= (5.12)
v∞ 2 v∞
5.2. Turbulent Flow. Similar to above, we also have equations for the boundary
length, δ, and coefficient of friction drag when we have turbulent flow:
0.376x
δ= 1/5
(5.13)
Rex
0.0576
C fx = 1/5
(5.14)
Rex
"
1
C fL = Cf dA
A A x
0.072
C fL = 1/5
(5.15)
ReL
6. Useful Assumptions
Below, we will list a number of assumptions that can be made to help simplify
a problem and what they allow you to do!
6.1. Steady-State. If you assume that a system or control volume are operating
under steady state, then none of its properties change with time. In other words,
if the steady-state assumption is valid then for any property, N :
∂N
= 0.
∂t
6.2. Incompressible Liquid. If you assume that your liquid is incompressible,
then you can assume that its density, ρ, is constant over any variable. Thus, you
can either pull it out of an equation or assume that:
∂ρ
= 0.
∂t
6.3. Well-Mixed. If you assume that a fluid or substance inside a control volume
is well-mixed, then whatever is leaving the system has to equal what is inside the
system!
! Main Idea: Concentration out = concentration inside!
12 RYAN GODIN

6.4. Incompressible Flow. In incompressible flow, the density of the fluid does
not change with the flow. In other words, density is constant with respect to spatial
derivatives. However, this does not mean it is constant with respect to time!
∂ρ
# Careful: Incompressible flow implies that ∂x is constant where x is any spatial
dimension. However, it does not imply that ∂∂tρ is constant!

7. Fundamentals of Heat Transfer

Now that we’ve talked about momentum transfer, we will begin to talk more
about heat transfer so that we can calculate φq in the energy balance for various
situations. Heat can be transferred in three different ways: conduction, convection,
and radiation. In transport phenomena, we will focus on conductive and convective
types of transport.

7.1. Conduction. Conductive heat transport occurs between materials that are in
the same phase. For example, between liquids and liquids or solids and solids. One-
dimensional conductive heat transport can be described in a simplified way using
Fourier’s Law. This law s related to momentum transport, and this makes sense
since heat is related to the velocity of molecules in a substance and they transfer
momentum in the same way they transfer heat, by colliding with each other.
Definition 7.1 (Fourier’s Law). One-dimensional, conductive heat transport in
the x-direction can be described as follows, where A is the conductive area and k
is the thermal conductivity:
dT
φq = −kA (7.1)
dx
○ Danger: The following equation assumes a NEGATIVE temperature gradient.
If it is positive, then we must switch the sign of φq in Fourier’s Law!
! Main Idea: To see how Fourier’s Law is related to Newton’s law of viscosity, we
need to re-write the above equation to solve for φq/A. This is essentially a ”stress”
just like shear stress, τ :
φq dT
= −k (7.2)
A dx

# Careful: Like viscosity, µ, thermal conductivity, k, is unique to the medium.

○ Danger: One needs to be careful about denoting heat flux through an object
with varying area like through a cylinder. The inside and outside areas can result
in drastically different heat fluxes!
7.2. Convection. Heat transfer due to convection involves the energy exchange
between a surface and an adjacent fluid (or gas). A distinction must be made
between forced convection, wherein a fluid is made to flow past a solid surface by
an external agent such as a fan or pump, and free or natural convection wherein
warmer (or cooler) fluid next to the solid boundary causes circulation because of
the density difference resulting from the temperature variation throughout a region
of the fluid. The rate equation for convective heat transfer was first expressed by
Newton in 1701, and is referred to as the Newton rate equation, or Newton’s law
of cooling:
 CHE 306: TRANSPORT PHENOMENA NOTES 13

Definition 7.2 (Newton’s Law of Cooling). One-dimensional, convective heat

transport can be described as follows, where A is the convective transport area
and h is the convective heattransfer (film) coefficient:

φq = hA∆T (7.3)

7.3. Combined Mechanisms of Heat Transfer. In some situations, we will

have heat transfer that utilizes conduction through multiple materials or multiple
mechanisms such as convective and conductive heat transport. In these situations,
we can figure out the heat transport, φq , relatively simply if we assume that φq is
constant and the area normal to heat transfer is constant. First, notice in this case
that if we have a substance of thickness, L, Fourier’s Law simplifies to:

dT
φq = −kA (7.4)
dx
φq dx = −kAdT (7.5)
" x=L " Tcold
φq dx = −kAdT (7.6)
x=0 Thot
" x=L " Thot
φq dx = kA dT (7.7)
x=0 Tcold
φq L = kA∆T (7.8)
kA
φq =
∆T (7.9)
L
○ Danger: Be careful of the sign! The negative sign is removed because heat goes
from hot (higher) to cold (lower).
Note that Equation (7.9) is very similar to Newton’s Law of cooling. If we
re-write in terms of the driving force, the temperature difference ∆T , then both
equations are in the form ∆T = φq Rthermal where Rthermal represents the resistance
to heat transfer. In an analogous way to Ohm’s Law, if we have a series of heat
transfer mechanisms, then through the whole system, we have:
$
∆Ttotal = φq Rq,total = φq Rq,i (7.10)
i∈I

We can then re-write the above equation to solve for the actual heat transport,
φq , through the system in a straightforward manner.
! Main Idea: Note that the above equation relies on the assumption that the
heat transfer, φq through one part of the system has to equal the heat transfer into
the other part. Thus, no φq is lost in the transition!
# Careful: The resistance terms have to be re-written for different geometries,
but they can be applied in the same way to find the overall heat transfer!
It is also common in these situations to express the heat transfer in terms of an
overall heat transfer coefficient, U , which functions almost exactly as h where:

φq = U A∆T (7.11)
14 RYAN GODIN

We can calculate U by re-writing and substituting for φq in Equation (7.10) to

arrive at the following equation:
1
U= / (7.12)
A R
& Question: Is U also referred to as the equivalent conductance? This appears
to be the value for k where that corresponds with ∆x = L = 1, no matter the unit.
$ Problem-Solving Strategy: If the material inside a window or equivalent is
stagnant, then we can assume there is no convection, just conduction!

8. Steady-State Conduction
We will now spend a little more time studying different types of conduction,
usually steady-state, and usually one dimensional. First, we will describe how to
calculate the temperature profiles in materials at steady-state.
8.1. One-Dimensional Conduction. For steady-state conduction in the absence
of any internal energy generation, the general differential equation representing heat
transfer is given by the Laplace Equation:
∇T 2 = 0 (8.1)
However, in the case that we have one-dimensional heat flow simplifies further
to the following: ' !
d dT
xi =0 (8.2)
dx dx
# Careful: The above equation assumes that we have one-dimensional and steady-
state heat transfer with no heat/energy generation!
In the above simplification of the Laplace equation, the exponent on the x term,
i, is dependent upon the geometry of the problem.
(1) For rectangular coordinates: i = 0 −→ xi = 1
(2) For cylindrical coordinates: i = 1 −→ xi = x
(3) For spherical coordinates: i = 2 −→ xi = x2
Now that we have these equations, we will apply them to some common heat-
transfer problems to see how they can be utilized. We will begin with considering
steady-state conduction through a flat plate.
8.2. Flate Plate. If we have one-dimensional, steady-state heat transfer through
a flat wall, then we have rectangular coordinates so Equation (8.2) is written as:
' !
d dT
=0 (8.3)
dx dx
Integrating the above equation, we find that the temperature profile is given by
the following equation where c1 and c2 are constants:
T = c1 x + c2 (8.4)
Clearly, this means that our temperature profile in a flat plate is linear and to
find the values for c1 and c2 , we need to apply the boundary conditions:
(1) At x = 0, T = T1 ;
(2) At x = L, T = T2 .
 CHE 306: TRANSPORT PHENOMENA NOTES 15

Applying the boundary conditions and solving for c1 and c2 , we find that the
temperature profile is given by:
T2 − T1
T = x + T1 (8.5)
L
Now, since we know that heat transfer occurs from the higher to lower temper-
ature and we know that T1 > T2 , the above equation is often rewritten as:
T1 − T2
T (x) = T1 − x (8.6)
L
We can actually use this temperature profile as another way to derive an equation
for conductive heat transfer through a flate plate which is given by the Fourier
equation. This is possible because the temperature gradient can be obtained by
differentiation the temperature profile with respect to x.
dT
φq = −kA (8.7)
dx
' !
T1 − T2
= −kA −
L
T1 − T2
= kA
L
kA
φq = ∆T (8.8)
L
% Notation: Notice how the above equation can be written in the form φq Rq =
∆T where Rq = L/kA. For that reason, L/kA is referred to as the thermal resis-
tance against the heat transfer for a flat plate. The thermal resistances for other
geometries can be found in a similar fashion.
8.3. Composite Walls. We will now consider that we have steady-state, one-
dimensional conduction through composite walls. We have already considered if
the walls are arranged in series. However, now we will also consider what happens
if they are arranged in parallel. In fact, it is very common for walls to have combine
both series and parallel heat transfer. One example is where we have steel-beam
reinforced concrete.
From previous experience in physics II where electrical transfer through circuits
is often discussed, we know that the formula for parallel resistances is given by:
0 #−1
$ 1
RT,parallel = (8.9)
Ri
i∈I

Furthermore, when resistances are connected in parallel, we are able to calculate

the heat flow, φq , through each resistance utilizing the fact that the temperature
difference must be the same and that the flow through each parallel resistance must
sum to the total flow. Thus, we have:

φq,1 R1 = φq,2 R2 (8.10)

$
φq = φq,i (8.11)
i∈I
16 RYAN GODIN

8.4. Hollow Cylinder. For a hollow cylinder, we have cylindrical coordinates, so

we know that i = 1. Therefore, it follows that:
' !
d dT
r =0 (8.12)
dr dr
Integrating the above equation and applying the boundary conditions, we then
find the temperature profile in a hollow cylinder:
' !
Ti − To r
T (r) = Ti − ln (8.13)
ln (Ro/Ri ) Ri
Similar to the flat plate, we can use the temperature profile to solve for the
temperature gradient and determine the heat transfer, φq :
dT
φq = −kA (8.14)
dr ' ' !!
d Ti − To r
= −k(2πrL) Ti − ln
dr ln (Ro/Ri ) Ri
' !
Ti − To 1
= −k(2πrL) − R
ln ( /Ri ) r
o

Ti − To
= 2πkL
ln (Ro/Ri )
∆T
φq = 2πkL (8.15)
ln (Ro/Ri )
8.5. Hollow Sphere. Following the same procedure for the hollow sphere as was
used for the flat plate and hollow cylinder, we find that the temperature profile is
given by: ' !' !
Ti − To 1 1
T (r) = Ti − 1 − (8.16)
/Ri − 1/Ro Ri r
Taking the derivative to find the temperature gradient, we find that the heat
transfer through the hollow sphere is given by:
4πk∆T
φq = (8.17)
− 1/Ro
1/Ri

9. Convective Heat Transfer

In our previous work, we were usually given the convective heat transfer value
for a problem. We will now learn how to calculate the convective heat transfer
coefficient, h, for various geometries and situations. The calculation of this value
relies heavily on the boundary layer concept discussed previously.

9.1. Significant Parameters in Convective Heat Transfer. Before we go on

to discuss how to calculate h, we will first list some of the important dimensionless
parameters that we will utilize.
µ
ν≡ (Momentum Diffusivity)
ρ
k
α≡ (Thermal Diffusivity)
ρCP
 CHE 306: TRANSPORT PHENOMENA NOTES 17

ν µCP
Pr ≡ ≡ (Prandtl Number)
α k
hL
Nu ≡ (Nusselt Number)
k
9.1.1. Forced Convection Parameters. If we have forced convection through a pipe,
then we also have the following important parameters:

h
St ≡ (9.1)
ρv̄CP
9.1.2. Natural Convection Parameters. If we have natural convection from a plane
wall to an adjacent fluid, then we also have the following important parameters:

βgρ2 3
Gr ≡ ·L ∆T (9.2)
µ2
( )* +
A. I

$ Problem-Solving Strategy: The characteristic length is defined differently

for different geometries:
(1) Pipe: L = δ (thickness);
(2) Plate: L = D (diameter);

9.2. Exact Solution for Thermal Boundary Layer (Laminar). When we

have a flat plate geometry and laminar flow, we are able to obtain exact and
approximate equations for the thermal boundary layer. Specifically, we are able to
solve for Nux , NuL , and, by extension, h. Furthermore, we are also able to solve
for the thermal boundary layer thickness, δt . These equations are given by the
following:

hx x 1/3
Nux = = 0.332Re1/2
x Pr (9.3)
k
hL L 1/2
NuL = 2Nux=L = = 0.664ReL Pr1/3 (9.4)
k
δ
δt = (9.5)
Pr1/3
$ Problem-Solving Strategy: For this chapter, all fluid properties will be eval-
uated at the film temperature, Tf , where:
Tsurface + T∞
Tf = (9.6)
2

9.3. Approximate Solution for Thermal Boundary Layer. For other geome-
tries, or if our flow is not laminar, then we can use the following approximation for
the local Nusselt number, Nux . In this case, we have:

1/3
Nux = 0.36Re1/2
x Pr (9.7)
18 RYAN GODIN

9.4. Energy and Momentum Transfer Analogies. If we don’t have laminar

flow over a flat plate and don’t want to rely on the integral approximation for the
Nusselt number, we can use the following momentum transfer analogies to describe
heat transfer with moving fluids.

9.4.1. Reynold’s Analogy. The Reynold’s analogy is valid when the following con-
ditions hold:
(1) Laminar or turbulent flow;
(2) Pr = 1;
(3) No form drag, only viscous drag.
For the Reynold’s analogy, we are able to calculate the convective heat transfer
using the following equation:

h Cf
St ≡ = (9.8)
ρv∞ CP 2
$ Problem-Solving Strategy: When we have flow through a closed conduit, Cf
is equivalent to the Fanning friction factor, ff . Thus, Cf = ff .
9.4.2. Colburn’s Analogy. The Colburn analogy is valid when the following condi-
tions hold:
(1) Laminar or turbulent flow;
(2) 0.5 < Pr < 50;
(3) No form drag, only viscous drag.
For the Colburn analogy, we are able to calculate the convective heat transfer
using the following equation:

Cf
StPr2/3 = (9.9)
2
9.4.3. Prandtl’s Analogy. The Prandtl analogy is valid when the following condi-
tions hold:
(1) Turbulent flow only;
(2) 1 < Pr;
(3) No form drag, only viscous drag.
For the Prandtl analogy, we are able to calculate the convective heat transfer
using the following equation:

Cf /2
St = - (9.10)
1+5 Cf /2(Pr − 1)

9.4.4. Karman’s Analogy. The Karman analogy is valid when the following condi-
tions hold:
(1) Turbulent flow only;
(2) 1 < Pr;
(3) No form drag, only viscous drag.
For the Karman analogy, we are able to calculate the convective heat transfer
using the following equation:
 CHE 306: TRANSPORT PHENOMENA NOTES 19

Cf /2
St = - 1 2 (9.11)
1 + 5 Cf /2(Pr − 1 + ln 1 + 56 (Pr − 1) )

10. Convective Heat Transfer Correlations

Above, we presented some results for calculating convective heat transfer that
were derived using analytical means. However, there are a number of other situa-
tions in which convective heat transfer is important. For those situations, we need
to rely on experimentally determined correlations to determine h.

11. Heat Exchangers

One important application of heat transfer in chemical engineering specific appli-
cations is for heat transfer in heat exchangers. A heat exchanger typically involves
two flowing fluids separated by a solid wall. Heat is first transferred from the hot
fluid to the wall by convection, through the wall by conduction, and from the wall
to the cold fluid again by convection. Thus, the overall thermal resistance is given
as follows:
1 1
Rq,total = + Rwall + (11.1)
h i Ai h o Ao
In the case of a tube, which is the usual geometry that we will encounter, the
equation can be written as follows:
1 ln (Ro/Ri ) 1
Rq,total = + + (11.2)
h i Ai 2πkL h o Ao
Now, for heat exhanger applications, it is usually more useful to talk about the
overall heat transfer coefficient, U , instead of the overall thermal resistance, Rq,total .
Thus, when using heat exchangers, the basic equation that we will be relying on
most often is:
φq = U Ao ∆T (11.3)
Now, we know from our previous study of convection and conduction that the
overall heat transfer coefficient, U , is given by the following equation:
1 1 ln (Ro/Ri ) 1
= Rq,total = + + (11.4)
U Ao h i Ai 2πkL h o Ao
If we assume that we either have (1) a thin wall (Ri ≈ Ro ) or if (2) the thermal
conductivity of the wall is high (k → ∞), then the thermal resistance from the wall
goes to 0. In the case that we make the assumption that the wall is very thin, then
it follows that we are also assuming that Ao = Ai . In this case, the equation for
the overall heat transfer coefficient, U , simplifies to:
' !−1
1 1
U= + (11.5)
hi ho

! Main Idea: Note that in the following equation, the term U is dominated
by the smaller heat transfer coefficient, U ≈ hsmall . In this case, the smaller heat
transfer coefficient creates a bottleneck impeding heat transfer. This situation arises
frequently when one of the fluids is a gas and the other is a liquid.
20 RYAN GODIN

11.1. Heat Transfer Analysis. Heat exchangers are commonly used in practice,
and an engineer often finds himself or herself in a position to select a heat exchanger
that will achieve a specified temperature change in a fluid stream of known mass
flow rate, or to predict the outlet temperatures of the hot and cold fluid streams in
a specified heat exchanger. We will now lay the groundwork for two approaches to
accomplishing these tasks: the log mean temperature difference (or LMTD) method
and the effectiveness-NTU (the number of transfer units) method.
$ Problem-Solving Strategy: The above methods are suited two accomplishing
one of the aforementioned tasks. Specifically:
(1) Log mean temperature difference (LMTD): best suited for selecting a heat
exchanger that will achieve a specified temperature change in a fluid stream
of known mass flow rate.
(2) NTU method: best suited for predicting the outlet temperatures of the hot
and cold fluid streams in a specified heat exchanger.

Since heat exchangers usually operate for extended periods of time with no
change in their operating conditions, we are able to model them as steady-flow
processes. Thus, if we take control volumes for the hot and cold streams, we are
able to derive the following equations assuming constant fluid properties:
φq = φm,c CP,c (Tc,out − Tc,in ) (11.6)
φq = φm,h CP,h (Th,in − Th,out ) (11.7)
# Careful: Note that the temperatues are selected so that ∆T is positive. This
makes φq positive and the same for both streams even though the heat transfer is
from hot to cold streams.
In heat exchanger analysis, it is often convenient to combine the product of the
mass flow rate and the specific heat of a fluid into a single quantity. This quantity
is called the heat capacity rate and is defined as follows for the hot and cold fluid
streams:
Ch = φm,h CP,h and Cc = φm,c CP,c (11.8)
As a result of the above definitions, the energy balances for the cold and hot
fluid streams can be rewritten as:
φq = Cc (Tc,out − Tc,in ) (11.9)
φq = Ch (Th,in − Th,out ) (11.10)
Note that as a result of the above equation, it follows that the only time the
temperature change for the hot fluid is equal to the temperature change of the cold
fluid is when their heat capacity rates are equal :
Cc (Tc,out − Tc,in ) = Ch (Th,in − Th,out )
Cc ∆Tc = Ch ∆Th
Cc = Ch
Now, returning back to the general analysis of a heat exchanger, we know that
out fundamental relationship is given by:
φq = U Ao ∆T (11.11)
 CHE 306: TRANSPORT PHENOMENA NOTES 21

However, the overall heat coefficient and the overall temperature difference may
change over the course of the heat exchanger. Thus, to take this into account, we
use the average of the convective heat transfer coefficients and the temperature
change. For the convective heat transfer coefficients, this is done by evaluating at
T∞ = (Tin + Tout )/2. However, for the temperature change, we shall see that the
average is actually logarithmic in nature.

11.1.1. Phase Change. A special situation for a heat exchanger is if we have a phase
change. In this case, the temperature of the stream undergoing the phase change
remains constant and the energy balance is instead derived as:
0 (SS)
!
dE #
✒
# CV = φm,in uin − φm,out uout + φq
dt
#
0 = φm,in Cp,in Tin − φm,ou Cp,out Tout + φq
φq = φm (Cp,out Tout − Cp,in Tin )
φq = φm (CP,out − CP,in )T
φ q = φ m hf g (11.12)

12. The Log Mean Temperature Difference Method

It can be shown that the heat-transfer in a double-pipe heat exchanger can be
written as follows:
φq = U Ao ∆Tlm (12.1)
In the above equation, the term ∆Tlm represents the log mean temperature
difference for the heat exchanger. As mentioned previously, it represents the actual
average temperature difference for the heat exchanger. ∆Tlm is calculated as follows:
∆T1 − ∆T2
∆Tlm = (12.2)
ln(∆T1/∆T2 )

In the above equation, the two temperature differences, ∆T1 and ∆T2 , are defined
according to which type of flow we have in our double-pipe flow heat exchanger:
(1) Parallel-flow: If we have parallel flow, then ∆T1 = Th,in − Tc,in and
∆T2 = Th,out − Tc,out .
(2) Counter-flow: If we have counter flow, then ∆T1 = Th,in − Tc,out and
∆T2 = Th,out − Tc,in .
! Main Idea: Here ∆T1 and ∆T2 represent the temperature difference between
the two fluids at the two ends (inlet and outlet) of the heat exchanger. It makes no
difference which end of the heat exchanger is designated as the inlet or the outlet.
The log mean temperature difference equation are valid for any heat exchanger
provided that the end point temperature differences are defined properly. However,
for multi-pass or cross-flow heat exchangers, a correction factor needs to be intro-
duced into the equation. Before we move on to discuss these different arrangements,
a couple points need to be touched upon:
(1) We always have ∆Tlm < ∆Tam . Therefore using ∆Tam will overestimate
the heat transfer.
22 RYAN GODIN

(2) When ∆T1 and ∆T2 differ from each other by no more than 40%, then
∆Tlm ≈ ∆Tam .
Now, for the counter-flow arrangement, some important points also need to be
discussed. Specifically,
(1) Note that the hot and cold fluids enter the heat exchanger from opposite
ends, and the outlet temperature of the cold fluid may exceed the outlet
temperature of the hot fluid. However, the outlet temperature of the cold
fluid cannot exceed the inlet temperature of the hot fluid!
(2) Heat transfer is always greater for counter-flow since: ∆Tlm,CF > ∆Tlm,PF .
(3) If Ch = Cc , then ∆Tlm is indeterminate. However, it can be shown that
∆Tlm = ∆T1 = ∆T2 as expected.

12.1. Multipass and Crossflow Heat Exchangers: Use of a Correction

Factor. We are now in a position to discuss what happens if our geometry varies
from the double-pipe single pass case. In this situation, we need to calculate a
correction factor, F , where:
∆Tlm = F ∆Tlm,CF (12.3)

φq = U Ao F ∆Tlm,CF (12.4)

! Main Idea: The correction factor F for a heat exchanger is a measure of

deviation of the ∆Tlm from the corresponding values for the counterflow case.
Now, to determine the correction factor, we need to use a graphical approach.
However, this relies on us first needing to calculate the following two parameters:
Ttube,out − Ttube,in
Y /P = (12.5)
Tshell,in − Ttube,in
(φm CP )tube side Ctube Tshell,in − Tshell,out
Z/R = = = (12.6)
(φm CP )shell side Cshell Ttube,out − Ttube,in
$ Problem-Solving Strategy: It makes no difference whether the hot or the
cold fluid flows through the shell or the tube.
$ Problem-Solving Strategy: In the case of a boiler or condenser with no
temperature change F = 1, regardless of the configuration of the heat exchanger.
12.2. Selecting a Heat-Exchanger (LMTD). With the LMTD method, the
task is to select a heat exchanger that will meet the prescribed heat transfer re-
quirements. The procedure to be followed by the selection process is:
(1) Select the type of heat exchanger suitable for the application.
(2) Determine any unknown inlet or outlet temperature and the heat transfer
rate using an energy balance.
(3) Calculate the log mean temperature difference ∆Tlm and the correction
factor F , if necessary.
(4) Obtain (select or calculate) the value of the overall heat transfer coefficient
U.
(5) Calculate the heat transfer surface area As /Ao .
! Main Idea: The task is completed by selecting a heat exchanger that has a
heat transfer surface area equal to or larger than As /Ao .
 CHE 306: TRANSPORT PHENOMENA NOTES 23

13. The Effectiveness-NTU Method

As previously mentioned, the log mean temperature difference (LMTD) method
is good for when the temperatures at the inlet and outlet are defined. However,
there are some situations in which you want to determine the heat transfer rate
and the outlet temperatures of the hot and cold fluids for prescribed fluid mass
flow rates and inlet temperatures when the type and size of the heat exchanger
are specified. In these situations, an alternative method, the effectiveness-NTU
method, is more appropriate. This method is based on a dimensionless parameter
called the heat transfer effectiveness ε, defined as
φq Actual heat transfer rate
ε= = (13.1)
φq,max Maximum possible heat transfer rate
In these situations, the inlet temperatures and flow rates for the two liquids are
usually given. From this information, we can calculate φq,max . The effectiveness
term, ε, is then usually looked up in a chart which allows us to calculate the actual
heat transfer rate, φq , and subsequently the outlet temperatures.
To calculate the maximum heat transfer rate, we first need to recognize that the
maximum heat transfer rate occurs alongside the maximum temperature change,
∆Tmax . Specifically, The heat transfer in a heat exchanger will reach its maximum
value when (1) the cold fluid is heated to the inlet temperature of the hot fluid or
(2) the hot fluid is cooled to the inlet temperature of the cold fluid. Thus, this
maximum is equal to the temperature difference between the inlet streams:
∆Tmax = Th,in − Tc,in (13.2)
However, we need to recognize that both fluids may not experience this tem-
perature change. These two limiting conditions will not be reached simultaneously
unless the heat capacity rates of the hot and cold fluids are identical (i.e., Cc = Ch ).
When Cc ∕= Ch , which is usually the case, the fluid with the smaller heat capacity
rate, Cmin , will experience a larger temperature change, and thus it will be the first
to experience the maximum temperature change, at which point the heat transfer
will come to a halt. Thus, the max heat transfer rate is given by:
φq,max = Cmin ∆Tmax = Cmin (Th,in − Tc,in ) (13.3)

14. Radiation
Radiation differs from the other two heat transfer mechanisms in that it does not
require the presence of a material medium to take place. In fact, energy transfer
by radiation is fastest (at the speed of light), and it suffers no attenuation in a
vacuum. Also, radiation transfer occurs in solids as well as liquids and gases. In
most practical applications, all three modes of heat transfer occur concurrently at
varying degrees. But heat transfer through an evacuated space can occur only by
radiation. For example, the energy of the sun reaches the earth by radiation.

14.1. Blackbody Radiation. The amount of radiation energy emitted from a

surface at a given wavelength depends on the material of the body and the condition
of its surface as well as the surface temperature. Therefore, different bodies may
emit different amounts of radiation per unit surface area, even when they are at the
same temperature. Thus, it is natural to be curious about the maximum amount
of radiation that can be emitted by a surface at a given temperature.
24 RYAN GODIN

! Main Idea: A blackbody is defined as a perfect emitter and absorber of

radiation. At a specified temperature and wavelength, no surface can emit more
energy than a blackbody.
For a blackbody, we can derive the Stefan-Boltzmann Law that relates the ther-
mal radiation of a blackbody to its surface temperature:
φq,rad,blackbody
= σT 4 (14.1)
A
○ Danger: For radiation, T always has to be given in absolute scale!
From the idea of a blackbody, we can move on to describe how other materials
interact with thermal radiation. We do this now.

14.2. Emissivity. The emissivity, ε, of a surface represents the ratio of the ra-
diation emitted by the surface at a given temperature to the radiation emitted by
a blackbody at the same temperature. In other words:
φq,rad,emitted
ε= (14.2)
φq,rad,emitted,blackbody
As a result of this equation, if we plug in for φq,rad,emitted,blackbody , we will find
that:
φq,rad,emitted
= εσT 4 (14.3)
A
$ Problem-Solving Strategy: The area of the blackbody and the non-blackbody
here are assumed to be equivalent. That is how we arrive at this equation!
14.3. Radiation Properties. Everything around us constantly emits radiation,
and the emissivity repre- sents the emission characteristics of those bodies. This
means that every body, including our own, is constantly bombarded by radiation
coming from all directions over a range of wavelengths. If we consider the irradiation
flux on a surface, φ′′q,irrad , then some of that irradiation, is absorbed, reflected by,
and transmitted through the surface. From this we arrive at the following balance
where α represents the absorptivity, ρ represents the reflectivity, and τ represents
the transmissivity of the material:
αφq,irrad + ρφq,irrad + τ φq,irrad = φq,irrad (14.4)
From this, we arrive at the following useful relationship:
α+ρ+τ =1 (14.5)

$ Problem-Solving Strategy: We have the following useful simplifications for

different types of substances:
(1) For blackbodies: α = 1 =⇒ ρ, τ = 0.
(2) For opaque surfaces such as most solids and liquids, τ = 0 =⇒ α + ρ = 1.
(3) For most gases, the reflectance is absent, ρ = 0 =⇒ α + τ = 1.

Objects that we encounter in everyday life are referred to as gray bodies. For
these objects, the absorptivity is a constant that needs to be less than or equal to
one: 0 < α < 1.
 CHE 306: TRANSPORT PHENOMENA NOTES 25

14.4. Kirchoff ’s Law. Consider a small body of surface area As , emissivity ε, and
absorptivity α at temperature T contained in a large isothermal enclosure at the
same temperature. We know that at steady-state, the radiation absorbed by the
small body must be equivalent to the radiation emitted by it:
As εσT 4 = As ασT 4 (14.6)
From this situation, we can therefore derive Kirchoff’s Law as:

α(T ) = ε(T ) (14.7)

! Main Idea: The above equation essentially states that the absorptivity of a
body is equivalent to its emissivity if in a large enclosure (which functions like a
blackbody)!
$ Problem-Solving Strategy: The above relationship holds whenever we have
an isolated object since we can assume it is in a large room!
14.5. Net Radiation. We will now discuss how to calculate φq,rad for a certain
special situations. Let’s imagine that we have an object contained in an enclosure
(or the atmosphere/large room). We can treat the atmosphere or enclosure as a
blackbody and combined with Kirchoff’s Law, we are able to derive an equation
for the net radiation transfer. To see this, note that the net radiation emitted by a
body is given by the following balance:
φq,rad $ $
= Eabsorbed − Eemitted (14.8)
As
For our specific case, this simplifies to:
φ′′q,rad = φ′′q,rad (absorbed from enclosure) − φ′′q,rad (emitted from body) (14.9)
4 4
= ασTenclosure − εσTbody (14.10)
4 4
= εσTenclosure − εσTbody (by Kirchoff’s Law)

φ′′q,rad = εσ(Tenclosure
4 4
− Tbody ) (14.11)

Now, if we try and come up with a heat transfer analogy for radiation in the
form of convective heat transfer, φ′′q,rad = hrad (Tenclosure − Tenclosure ), then it is easy
to see that:
3 4
4 4
εσ Tenclosure − Tbody
hrad = (14.12)
(Tenclosure − Tbody )

14.6. Special Cases and Geometries. We will now develop equations for net
heat transfer in other geometries/cases in a similar way to a body in a large enclo-
sure. These are:
1 2
As σ T14 − T24
φq,rad (1→2) = 1 1 (Infinitely large parallel plates)
ε1 + ε2 − 1

1 2
φq,rad (1→2) = A1 σε1 T14 − T24 (Small object in large cavity)
26 RYAN GODIN

14.6.1. Radiation Shields. If we have a series of radiation shields between to flat

plates giving of thermal radiation, then the equation for the overall heat transfer
between the plates can be given by:
1 2
Aσ T14 − T24 1
φq,rad (1→2), N shields = 1 2= φq,rad (1→2), no shields
(N + 1) 1ε + 1ε − 1 N +1
(14.13)
# Careful: This is assuming that the emissivity, ε, for all the surfaces are equiv-
alent!
! Main Idea: At steady-state, the net transfer from surface 1 → 2 and from 2 to3
are equivalent to the net transfer. In other words,

φq,rad (1→2) = φq,rad (2→3) = φq,rad, net (1→3) (14.14)

15. Fundamentals of Mass Transfer

We will now begin to discuss how to derive equations concerning mass transfer.
It shares a number of similarities to heat transfer, so the development of the topic
will follow a similar path. However, mass transfer is a little more complicated
since we have two components, A (solute) and B (solvent) which interact with each
other and can both diffuse. Thus, we need to take into account the behavior of
both species when studying mass transfer.
! Main Idea: Mass transfer has three requirements: (1) that transfer occur only
in a mixture, (2) that at least one substance within the mixture move from its
source to its sink, and (3) that the rate of mass transfer is proportional to the
concentration gradient.
15.1. Fick’s Rate Equations. Being a little more complicated than heat and
momentum transfer because of the dual components, we will work to develop a
practical mass conservation equation known Fick’s Rate Equation.
! Main Idea: Diffusion depends on the concentration and relative mobility (ve-
locity) of each species.
15.1.1. Concentrations. Since mass transfer only occurs in mixtures, we need to
have a discussion about concentrations. In a binary mixture, only two components
are present. In a multi-component mixture, more than two components are present,
but no population of any given component dominates. In a pseudo-binary mix-
ture, one component is dominant, and the population of the other species is small,
relative to this dominant component. For gases, this dominant component is the
often called the carrier gas, and for liquids, it is called the solvent. First, we
have the mass concentration, otherwise known as the density of our mixture:

mass of component A
ρA = (15.1)
unit volume of mixture in a given phase
 CHE 306: TRANSPORT PHENOMENA NOTES 27

We also have the density of our mixture which is defined as one would expect,
except that the partial densities sum up to the total.
$ n
total mass in mixture
ρ= = ρi (15.2)
unit volume of mixture in a given phase i=1

The mass fraction of component A in a mixture is then defined, again, as one

would expect. By definition,
/n these components are not all independent since they
must sum up to one, i=1 wi = 1:
ρA ρA
w A = /n = (15.3)
ρ
i i ρ
What is usually encountered more often in solutions, however, is the molar con-
centration, cA . It is related to the mass concentration (density) as follows where
MA is the molecular weight of species A:
ρA
cA = (15.4)
MA
When dealing with a gas phase, concentrations are often expressed in terms of
partial pressures. Under conditions in which the ideal gas law applies, the molar
concentration is:
nA pA
cA = = (15.5)
V RT
As one would expect, the total concentration for a mixture is defined using
summability:
$ n
total moles in mixture P
c= = ci = (15.6)
unit volume of mixture in a given phase i=1
RT

Depending on whether we have a liquid or a gas phase, the mole fractions are
written with different symbols. This is necessary since these two phases often
coexist.
cA pA
x A = yA = = (15.7)
c P

15.1.2. Velocities. Since we have a mixture, the velocity for our mixture needs to
be averaged from the velocities of our individual components. The mass average
velocity, v is given by:
/n /n
ρi v i i=1 ρi vi
v = /i=1n = (15.8)
i=1 ρ i ρ
Similarly, the molar average velocity, V , is defined by:
/n /n
ci vi i=1 ci vi
V = /i=1
n = (15.9)
i=1 ci c
Now, the velocity of a component relative to the average velocity of the mixture
is an important value. It is known as the diffusion velocity.
! Main Idea: The diffusion velocity only exists if we have a concentration gradi-
ent.
28 RYAN GODIN

15.1.3. Fick’s Rate Equation. In the most simplified form, Fick’s rate equation
(first law) is given by:
dyA
φ′′n,A = −cDAB (15.10)
dz
Here, c represents the total concentration of the mixture at constant pressure
and temperature. However, it is not necessarily constant! When it is constant, we
can write:
dcA
φ′′n,A = −DAB (15.11)
dz
Similarly, we also have the equivalent form for the mass flux:
dwA
φ′′m,A = −ρDAB (15.12)
dz
When the total density of the mixture is constant, than this equation can also
be simplified to the following equation:
dρA
φ′′m,A = −DAB (15.13)
dz
However, where this equation goes wrong is that if we have a moving fluid that
diffusion is taking place in, then mass transfer is happening because of both simple
diffusion and convective bulk transport. Thus, the more general form of the above
equation is:
dyA 1 2
φ′′n,A = −cDAB + yA φ′′n,A + φ′′n,B (15.14)
dz
This equation can be written out in extended form as:
dyB
cA vA,z = −cDAB + yA (cA vA,z + cB vB,z ) (15.15)
dz
c φ
$ Problem-Solving Strategy: The fluxes are given by φ′′n,i = i AV,i V–
= ci vi,z .
Thus, if diffusion is slow (relative diffusion speed is approximately 0, then fluxes of
A and B in the convective term are negligible.
From this equation, we see that if B is a non-moving carrier gas, then its flux is
zero. Therefore, we can neglect the φ′′n,B term and arrive at:
−cDAB dyA
φ′′n,A = (15.16)
1 − yA dz
Thus, we see that the presence of our carrier gas modulates the diffusion of A
from Fick’s first law by 1 − yA . Thus, if A is very dilute in B (yA → 0), then we can
further simplify the equation to Fick’s first law which is analogous to the Fourier
Law in heat transfer:
dyA
φ′′n,A = −cDAB (15.17)
dz
15.2. Convective Mass Transfer. For mass transfer between a moving fluid and
a surface or between immiscible moving fluids separated by a mobile interface, we
have convective mass transfer. Analogous to Newton’s Law, the convective mass
transfer equation is given by:
φn,A = kc ∆c = kc A (CA,S − CA,∞ ) (15.18)
 CHE 306: TRANSPORT PHENOMENA NOTES 29

For stagnant film theory, we are able to relate the convective transport coefficient,
kc , to the mass (concentration) transport boundary layer thickness, δc , by the
following equation:
DAB
kc = (15.19)
δc
However, δc is often difficult to calculate/determine. Therefore, we usually rely
on boundary layer theory which gives us the following exact solution for laminar
flow over a flat plate where Sc is the Schmidt number and L is the (characteristic)
length of the plate:
DAB 1/2 1/3
kc = 0.664 ReL Sc (15.20)
L
15.2.1. Important Dimensionless Numbers for Mass Transfer. The following are
the important dimensionless numbers for mass transfer:
ν µ
Sc = = (Schmidt Number)
DAB ρDAB
k
Le = (Lewis Number)
ρCp DAB
kc L
Sh = (Sherwood Number)
DAB
When we have laminar flow over a flat plate, we are able to come up with an
exact equation for its local solution:
kc,x x
= Shx = 0.332 Re1/2
x Sc
1/3
(15.21)
DAB
$ Problem-Solving Strategy: We can simplify this equation when Shx = 1 and
the concentration of A is very dilute.
Integrating the above equation, we see that the mean sherwood number is given
by the following equation:
kc,L L 1/2
= ShL = 0.664 ReL Sc1/3 (15.22)
DAB
For the approximate solution, the value compares quite nicely. Specifically, the
local Sherwood number is given by:
kc,x x
= Shx = 0.323 Re1/2
x Sc
1/3
(15.23)
DAB
A benefit of the approximate solution is that we can use it find an estimate for
the Sherwood number during turbulent flow (Rex > 3 × 106 ).
Shx = 0.0289Re4/5
x Sc
1/3
(15.24)
We are also able to derive the following relationship between the concentration
boundary layer and the hydrodynamic boundary layer:
1
δc = δSc− 3 (15.25)
Since we already have equations or calculating the hydrodynamic boundary layer
over a flat plate, we can use the above equation to derive usable equations for the
30 RYAN GODIN

concentration boundary layer. For laminar flow over a flat plate:

.
5x νx
δc = 1/3 √ = 5Sc−1/3 (15.26)
Sc Rex v ∞