CH002.

qxd 2/24/11 12:22 PM Page 79

2.2 䊏 The Thermal Properties of Matter 79

in equipment and process design or failure to meet performance specifications that

were attributable to misinformation associated with the selection of key property values
used in the initial system analysis. Selection of reliable property data is an integral part of
any careful engineering analysis. The casual use of data that have not been well character-
ized or evaluated, as may be found in some literature or handbooks, is to be avoided.
Recommended data values for many thermophysical properties can be obtained from
Reference 19. This reference, available in most institutional libraries, was prepared by the
Thermophysical Properties Research Center (TPRC) at Purdue University.

EXAMPLE 2.1

The thermal diffusivity ␣ is the controlling transport property for transient conduction.
Using appropriate values of k, ␳, and cp from Appendix A, calculate ␣ for the following
materials at the prescribed temperatures: pure aluminum, 300 and 700 K; silicon carbide,
1000 K; paraffin, 300 K.

SOLUTION

Known: Definition of the thermal diffusivity ␣.

Find: Numerical values of ␣ for selected materials and temperatures.

Properties: Table A.1, pure aluminum (300 K):

冧
␳  2702 kg/m3
k 
cp  903 J/kg 䡠 K ␣  ␳c 237 W/m 䡠 K
k  237 W/m 䡠 K
p 2702 kg/m3 903 J/kg 䡠 K

 97.1 106 m2/s 䉰

Table A.1, pure aluminum (700 K):

␳  2702 kg/m3 at 300 K

cp  1090 J/kg 䡠 K at 700 K (by linear interpolation)
k  225 W/m 䡠 K at 700 K (by linear interpolation)
Hence

␣  ␳kc  225 W/m 䡠 K  76 106 m2/s 䉰

p 2702 kg/m3 1090 J/kg 䡠 K

Table A.2, silicon carbide (1000 K):

冧
␳  3160 kg/m3 at 300 K
cp  1195 J/kg 䡠 K at 1000 K ␣  87 W/m 䡠 K
3
k  87 W/m 䡠 K at 1000 K 3160 kg/m 1195 J/kg 䡠 K

 23 106 m2/s 䉰
CH002.qxd 2/24/11 12:22 PM Page 80

80 Chapter 2 䊏 Introduction to Conduction

Table A.3, paraffin (300 K):

冧
␳  900 kg/m3
cp  2890 J/kg 䡠 K ␣  ␳kc  0.24 W/m 䡠 K
3
k  0.24 W/m 䡠 K
p 900 kg/m 2890 J/kg 䡠 K

 9.2 108 m2/s 䉰

Comments:
1. Note the temperature dependence of the thermophysical properties of aluminum and
silicon carbide. For example, for silicon carbide, ␣(1000 K) ⬇ 0.1 ␣(300 K); hence
properties of this material have a strong temperature dependence.
2. The physical interpretation of ␣ is that it provides a measure of heat transport (k) relative
to energy storage (␳cp). In general, metallic solids have higher ␣, whereas nonmetallics
(e.g., paraffin) have lower values of ␣.
3. Linear interpolation of property values is generally acceptable for engineering
calculations.
4. Use of the low-temperature (300 K) density at higher temperatures ignores thermal
expansion effects but is also acceptable for engineering calculations.
5. The IHT software provides a library of thermophysical properties for selected solids,
liquids, and gases that can be accessed from the toolbar button, Properties. See Exam-
ple 2.1 in IHT.

EXAMPLE 2.2

The bulk thermal conductivity of a nanofluid containing uniformly dispersed, noncontacting

spherical nanoparticles may be approximated by
kp  2kbf  2␸(kp  kbf)
knf  冤 k  2k
p bf  ␸(kp  kbf) 冥kbf

where ␸ is the volume fraction of the nanoparticles, and kbf, kp, and knf are the thermal con-
ductivities of the base fluid, particle, and nanofluid, respectively. Likewise, the dynamic
viscosity may be approximated as [20]
␮nf  ␮bf (1  2.5␸)
Determine the values of knf, ␳nf, cp,nf, ␮nf, and ␣nf for a mixture of water and Al2O3 nanoparticles
at a temperature of T  300 K and a particle volume fraction of ␸  0.05. The thermophysical
properties of the particle are kp  36.0 W/m 䡠 K, ␳p  3970 kg/m3, and cp,p  0.765 kJ/kg 䡠 K.

SOLUTION

Known: Expressions for the bulk thermal conductivity and viscosity of a nanofluid with
spherical nanoparticles. Nanoparticle properties.
CH002.qxd 2/24/11 12:22 PM Page 81

2.2 䊏 The Thermal Properties of Matter 81

Find: Values of the nanofluid thermal conductivity, density, specific heat, dynamic viscos-
ity, and thermal diffusivity.

Schematic:
Water
Nanoparticle

kp ⫽ 36.0 W/m·K
ρp ⫽ 3970 kg/m3
cp,p ⫽ 0.765 kJ/kg·K

Assumptions:
1. Constant properties.
2. Density and specific heat are not affected by nanoscale phenomena.
3. Isothermal conditions.

Properties: Table A.6 (T  300 K): Water; kbf  0.613 W/m K, ␳bf  997 kg/m3,
cp,bf  4.179 kJ/kg K, ␮bf  855 106 N s/m2.

Analysis: From the problem statement,

kp  2kbf  2␸(kp  kbf)
knf  冤 k  2k
p bf  ␸(kp  kbf) 冥kbf

 冤36.036.0W/m
W/m 䡠 K  2 0.613 W/m 䡠 K  0.05(36.0  0.613) W/m 䡠 K 冥
䡠 K  2 0.613 W/m 䡠 K  2 0.05(36.0  0.613) W/m 䡠 K

0.613 W/m 䡠 K
 0.705 W/m K 䉰

Consider the control volume shown in the schematic to be of total volume V. Then the con-
servation of mass principle yields

␳nfV  ␳bfV(1  ␸)  ␳pV␸

or, after dividing by the volume V,

␳nf  997 kg/m3 (1  0.05)  3970 kg/m3 0.05  1146 kg/m3 䉰

Similarly, the conservation of energy principle yields,

␳nfVcp,nf T  ␳bfV(1  ␸)cp,bf T  ␳pV␸cp,p T

Dividing by the volume V, temperature T, and nanofluid density ␳nf yields
␳bf cp,bf (1  ␸)  ␳pcp,p␸
cp,nf  ␳nf
997 kg/m3 4.179 kJ/kg 䡠 K (1  0.05)  3970 kg/m3 0.765 kJ/kg 䡠 K (0.05)

1146 kg/m3
 3.587 kJ/kg 䡠 K 䉰
CH002.qxd 2/24/11 12:22 PM Page 82

82 Chapter 2 䊏 Introduction to Conduction

From the problem statement, the dynamic viscosity of the nanofluid is

␮nf  855 106 N 䡠 s/m2 (1  2.5 0.05)  962 106 N 䡠 s/m2 䉰

The nanofluid’s thermal diffusivity is

k 0.705 W/m 䡠 K
␣nf  ␳ cnf   171 109 m2/s 䉰
nf p,nf 1146 kg/m3 3587 J/kg 䡠 K
Comments:
1. Ratios of the properties of the nanofluid to the properties of water are as follows.
knf 0.705 W/m 䡠 K ␳nf 1146 kg/m3
  1.150 ␳ bf  997 kg/m3  1.149
kbf 0.613 W/m 䡠 K
cp,nf 3587 J/kg 䡠 K ␮nf 962 106 N 䡠 s/m2
cp,bf  4179 J/kg 䡠 K  0.858 ␮bf  855 106 N 䡠 s/m2
 1.130

␣nf 171 109 m2 /s

␣bf  147 109 m2/s
 1.166

The relatively large thermal conductivity and thermal diffusivity of the nanofluid
enhance heat transfer rates in some applications. However, all of the thermophysical
properties are affected by the addition of the nanoparticles, and, as will become evi-
dent in Chapters 6 through 9, properties such as the viscosity and specific heat are
adversely affected. This condition can degrade thermal performance when the use of
nanofluids involves convection heat transfer.
2. The expression for the nanofluid’s thermal conductivity (and viscosity) is limited to
dilute mixtures of noncontacting, spherical particles. In some cases, the particles do not
remain separated but can agglomerate into long chains, providing effective paths for
heat conduction through the fluid and larger bulk thermal conductivities. Hence, the
expression for the thermal conductivity represents the minimum possible enhancement of
the thermal conductivity by spherical nanoparticles. An expression for the maximum
possible isotropic thermal conductivity of a nanofluid, corresponding to agglomeration
of the spherical particles, is available [21], as are expressions for dilute suspensions of
nonspherical particles [22]. Note that these expressions can also be applied to nanostruc-
tured composite materials consisting of a particulate phase interspersed within a host
binding medium, as will be discussed in more detail in Chapter 3.
3. The nanofluid’s density and specific heat are determined by applying the principles of
mass and energy conservation, respectively. As such, these properties do not depend
on the manner in which the nanoparticles are dispersed within the base liquid.

2.3 The Heat Diffusion Equation

A major objective in a conduction analysis is to determine the temperature efi ld in a

medium resulting from conditions imposed on its boundaries. That is, we wish to know
the temperature distribution, which represents how temperature varies with position in the
medium. Once this distribution is known, the conduction heat flux at any point in
the medium or on its surface may be computed from Fourier’s law. Other important
CH002.qxd 2/24/11 12:22 PM Page 87

2.3 䊏 The Heat Diffusion Equation 87

are heat flux components in the radial, circumferential, and axial directions, respectively.
Applying an energy balance to the differential control volume of Figure 2.12, the following
general form of the heat equation is obtained:

冢 冣⭸r r 2 ⭸␾ ⭸␾ ⭸z 冢 冣
1 ⭸ kr ⭸T  1 ⭸ k ⭸T  ⭸ k ⭸T  q˙ ␳c ⭸T
r ⭸r ⭸z p 冢 冣
⭸t
(2.26)

Spherical Coordinates In spherical coordinates, the general form of the heat flux vector
and Fourier’s law is

q  kT  k i 冢 ⭸T⭸r  j 1r ⭸T⭸␪  k r sin␪

1 ⭸T
⭸␾冣
(2.27)

where
⭸T ⭸T k ⭸T
qr  k q␪   kr q␾   (2.28)
⭸r ⭸␪ r sin␪ ⭸␾

are heat flux components in the radial, polar, and azimuthal directions, respectively. Apply-
ing an energy balance to the differential control volume of Figure 2.13, the following
general form of the heat equation is obtained:

r 2 ⭸r 冢
1 ⭸ kr 2 ⭸T 
⭸r 2 冣
1 ⭸
2 ⭸␾
r sin ␪
k
⭸T
⭸␾ 冢 冣
 1 ⭸
r sin␪ ⭸␪
2 冢
k sin␪
⭸T
⭸␪ 冣
 q˙ ␳ cp
⭸T
⭸t
(2.29)

You should attempt to derive Equation 2.26 or 2.29 to gain experience in applying conser-
vation principles to differential control volumes (see Problems 2.35 and 2.36). Note that the
temperature gradient in Fourier’s law must have units of K/m. Hence, when evaluating
the gradient for an angular coordinate, it must be expressed in terms of the differential
change in arc length. For example, the heat flux component in the circumferential direction
of a cylindrical coordinate system is q␾  (k/r)(⭸T/⭸␾), not q␾  k(⭸T/⭸␾).

EXAMPLE 2.3

The temperature distribution across a wall 1 m thick at a certain instant of time is

given as

T(x)  a  bx  cx2

where T is in degrees Celsius and x is in meters, while a  900 C, b  300 C/m, and
.
c  50 C/m2. A uniform heat generation, q  1000 W/m3, is present in the wall of area
10 m2 having the properties ␳  1600 kg/m3, k  40 W/m 䡠 K, and cp  4 kJ/kg 䡠 K.
CH002.qxd 2/24/11 12:22 PM Page 88

88 Chapter 2 䊏 Introduction to Conduction

1. Determine the rate of heat transfer entering the wall (x  0) and leaving the wall (x  1 m).
2. Determine the rate of change of energy storage in the wall.
3. Determine the time rate of temperature change at x  0, 0.25, and 0.5 m.

SOLUTION

Known: Temperature distribution T(x) at an instant of time t in a one-dimensional wall

with uniform heat generation.

Find:
1. Heat rates entering, qin (x  0), and leaving, qout (x  1 m), the wall.
2. Rate of change of energy storage in the wall, E˙st.
3. Time rate of temperature change at x  0, 0.25, and 0.5 m.

Schematic:

A = 10 m2 q• = 1000 W/m3
k = 40 W/m•K
ρ = 1600 kg/m3
cp = 4 kJ/kg•K

T(x) =
a + bx + cx2
•
Eg
•
E st

qin qout

L=1m
x

Assumptions:
1. One-dimensional conduction in the x-direction.
2. Isotropic medium with constant properties.
.
3. Uniform internal heat generation, q (W/m3).

Analysis:
1. Recall that once the temperature distribution is known for a medium, it is a simple
matter to determine the conduction heat transfer rate at any point in the medium or at
its surfaces by using Fourier’s law. Hence the desired heat rates may be determined by
using the prescribed temperature distribution with Equation 2.1. Accordingly,
⭸T
qin  qx(0)  kA 兩  kA(b  2cx)x0
⭸x x0
qin  bkA  300 C/m 40 W/m 䡠 K 10 m2  120 kW 䉰
CH002.qxd 2/24/11 12:22 PM Page 89

2.3 䊏 The Heat Diffusion Equation 89

Similarly,
⭸T
qout  qx(L)  kA 兩  kA(b  2cx)xL
⭸x xL
qout  (b  2cL)kA  [300 C/m

 2(50 C/m2) 1 m] 40 W/m 䡠 K 10 m2  160 kW 䉰

2. The rate of change of energy storage in the wall E˙st may be determined by applying an
overall energy balance to the wall. Using Equation 1.12c for a control volume about
the wall,
E˙in  E˙g  E˙out  E˙st

where E˙g  q˙AL, it follows that

E˙st  E˙in  E˙g  E˙out  qin  q˙AL  qout

E˙st  120 kW  1000 W/m3 10 m2 1 m  160 kW

E˙st  30 kW 䉰

3. The time rate of change of the temperature at any point in the medium may be deter-
mined from the heat equation, Equation 2.21, rewritten as

⭸T k ⭸2T  q˙
 ␳c
⭸t p ⭸x2 ␳cp

From the prescribed temperature distribution, it follows that

⭸2T

⭸ ⭸T
⭸x2 ⭸x ⭸x 冢 冣
⭸
 (b  2cx)  2c  2(50 C/m2) 100 C/m2
⭸x

Note that this derivative is independent of position in the medium. Hence the time rate
of temperature change is also independent of position and is given by

⭸T 40 W/m 䡠 K
 (100 C/m2)
⭸t 1600 kg/m3 4 kJ/kg 䡠 K

 1000 W/m3
1600 kg/m3 4 kJ/kg 䡠 K

⭸T
6.25 104 C/s  1.56 104 C/s
⭸t

 4.69 104 C/s 䉰

CH002.qxd 2/24/11 12:22 PM Page 90

90 Chapter 2 䊏 Introduction to Conduction

Comments:
1. From this result, it is evident that the temperature at every point within the wall is
decreasing with time.
2. Fourier’s law can always be used to compute the conduction heat rate from knowledge of
the temperature distribution, even for unsteady conditions with internal heat generation.

Microscale Effects For most practical situations, the heat diffusion equations generated
in this text may be used with confidence. However, these equations are based on Fourier’s
law, which does not account for the finite speed at which thermal information is propagated
within the medium by the various energy carriers. The consequences of the finite propaga-
tion speed may be neglected if the heat transfer events of interest occur over a sufficiently
long time scale, t, such that
␭mfp
1 (2.30)
ct
The heat diffusion equations of this text are likewise invalid for problems where boundary
scattering must be explicitly considered. For example, the temperature distribution within
the thin film of Figure 2.6b cannot be determined by applying the foregoing heat diffusion
equations. Additional discussion of micro- and nanoscale heat transfer applications and
analysis methods is available in the literature [1, 5, 10, 23].

2.4 Boundary and Initial Conditions

To determine the temperature distribution in a medium, it is necessary to solve the appro-

priate form of the heat equation. However, such a solution depends on the physical condi-
tions existing at the boundaries of the medium and, if the situation is time dependent, on
conditions existing in the medium at some initial time. With regard to the boundary condi-
tions, there are several common possibilities that are simply expressed in mathematical
form. Because the heat equation is second order in the spatial coordinates, two boundary
conditions must be expressed for each coordinate needed to describe the system. Because
the equation is first order in time, however, only one condition, termed the initial condition,
must be specified.
Three kinds of boundary conditions commonly encountered in heat transfer are summa-
rized in Table 2.2. The conditions are specified at the surface x  0 for a one-dimensional
system. Heat transfer is in the positive x-direction with the temperature distribution, which
may be time dependent, designated as T(x, t). The first condition corresponds to a situation
for which the surface is maintained at a fixed temperature Ts. It is commonly termed a
Dirichlet condition, or a boundary condition of the rfi st kind. It is closely approximated, for
example, when the surface is in contact with a melting solid or a boiling liquid. In both
cases, there is heat transfer at the surface, while the surface remains at the temperature of the
phase change process. The second condition corresponds to the existence of a fixed or con-
stant heat flux qs at the surface. This heat flux is related to the temperature gradient at the
surface by Fourier’s law, Equation 2.6, which may be expressed as
⭸T
qx (0) k 兩  qs
⭸x x0
CH002.qxd 2/24/11 12:22 PM Page 91

2.4 䊏 Boundary and Initial Conditions 91

TABLE 2.2 Boundary conditions for the heat

diffusion equation at the surface (x  0)
1. Constant surface temperature Ts

T(0, t)  Ts (2.31)
T(x, t)

2. Constant surface heat flux

(a) Finite heat flux
qs''
⭸T
k 兩  qs
⭸x x0
(2.32) T(x, t)

(b) Adiabatic or insulated surface

⭸T
兩 0
⭸x x0
(2.33) T(x, t)

3. Convection surface condition T(0, t)

⭸T
k 兩  h[T앝  T(0, t)]
⭸x x0
(2.34)
T∞, h

T(x, t)
x

It is termed a Neumann condition, or a boundary condition of the second kind, and may be
realized by bonding a thin film electric heater to the surface. A special case of this condi-
tion corresponds to the perfectly insulated, or adiabatic, surface for which ⭸T/⭸x冷 x0  0.
The boundary condition of the third kind corresponds to the existence of convection heat-
ing (or cooling) at the surface and is obtained from the surface energy balance discussed in
Section 1.3.1.

EXAMPLE 2.4

A long copper bar of rectangular cross section, whose width w is much greater than its
thickness L, is maintained in contact with a heat sink at its lower surface, and the tempera-
ture throughout the bar is approximately equal to that of the sink, To. Suddenly, an electric
current is passed through the bar and an airstream of temperature T앝 is passed over the top
surface, while the bottom surface continues to be maintained at To. Obtain the differential
equation and the boundary and initial conditions that could be solved to determine the tem-
perature as a function of position and time in the bar.

SOLUTION

Known: Copper bar initially in thermal equilibrium with a heat sink is suddenly heated
by passage of an electric current.
CH002.qxd 2/24/11 12:22 PM Page 92

92 Chapter 2 䊏 Introduction to Conduction

Find: Differential equation and boundary and initial conditions needed to determine tem-
perature as a function of position and time within the bar.

Schematic:

Copper bar (k, α)

T(x, y, z, t)
T(x, t)
z
Air y
T∞, h x

Air w
T∞, h
L
T(L, t) Heat sink
I To
q
•
L x

To = T(0, t)

Assumptions:
1. Since the bar is long and w  L, end and side effects are negligible and heat transfer
within the bar is primarily one dimensional in the x-direction.
2. Uniform volumetric heat generation, q˙.
3. Constant properties.

Analysis: The temperature distribution is governed by the heat equation (Equation 2.19),
which, for the one-dimensional and constant property conditions of the present problem,
reduces to
⭸2T q˙ 1 ⭸T
  (1) 䉰
⭸x2 k ␣ ⭸t

where the temperature is a function of position and time, T(x, t). Since this differential
equation is second order in the spatial coordinate x and first order in time t, there must be
two boundary conditions for the x-direction and one condition, termed the initial condition,
for time. The boundary condition at the bottom surface corresponds to case 1 of Table 2.2.
In particular, since the temperature of this surface is maintained at a value, To, which is
fixed with time, it follows that

T(0, t)  To (2) 䉰

The convection surface condition, case 3 of Table 2.2, is appropriate for the top surface.
Hence
⭸T
k 兩  h[T(L, t)  T앝]
⭸x xL
(3) 䉰

The initial condition is inferred from recognition that, before the change in conditions, the
bar is at a uniform temperature To. Hence

T(x, 0)  To (4) 䉰
CH002.qxd 2/24/11 12:22 PM Page 93

2.4 䊏 Boundary and Initial Conditions 93

.
If To, T앝, q, and h are known, Equations 1 through 4 may be solved to obtain the time-varying
temperature distribution T(x, t) following imposition of the electric current.

Comments:
1. The heat sink at x  0 could be maintained by exposing the surface to an ice bath or by
attaching it to a cold plate. A cold plate contains coolant channels machined in a solid
of large thermal conductivity (usually copper). By circulating a liquid (usually water)
through the channels, the plate and hence the surface to which it is attached may be
maintained at a nearly uniform temperature.
2. The temperature of the top surface T(L, t) will change with time. This temperature is
an unknown and may be obtained after finding T(x, t).
3. We may use our physical intuition to sketch temperature distributions in the bar at
selected times from the beginning to the end of the transient process. If we assume that
T앝
To and that the electric current is sufficiently large to heat the bar to temperatures
in excess of T⬁, the following distributions would correspond to the initial condition
(t  0), the final (steady-state) condition (t l 앝), and two intermediate times.

T(x, ∞), Steady-state condition

T∞
T(x, t)

T∞
b a
T(x, 0), Initial condition
To

0 L
Distance, x

Note how the distributions comply with the initial and boundary conditions. What is a
special feature of the distribution labeled (b)?
4. Our intuition may also be used to infer the manner in which the heat flux varies with
time at the surfaces (x  0, L) of the bar. On qx  t coordinates, the transient variations
are as follows.

+
q"x (L, t)
q"x (x, t)

0
q"x (0, t)

–
0 Time, t

Convince yourself that the foregoing variations are consistent with the temperature
distributions of Comment 3. For t l 앝, how are qx (0) and qx (L) related to the volu-
metric rate of energy generation?