Scientists do stupid looking things sometimes

(though not too unsafe if they made the material carefully enough)

A white – hot cube

Only seconds after removed from hot furnace
Material Silica fiber insulation
Heat transfer from the surface rapid
Thermal conductivity of the material too low
Interior temperature 1250oC
Outside temperature: Room temperature

Tiles in Space shuttle orbiters

 Heat Capacity
Heat capacity is the measure of the energy required to raise the temperature of a material
From another point of view it is the increase in energy content per degree of temperature rise

It is normally measured as the heat capacity at constant pressure C p,

but theoretical calculations it is

reported in terms of the heat capacity at Constant Volume C v

𝐶 𝑃= ( ) ( )
𝑑𝑄
=
𝑑𝐻
𝑑𝑇 𝑃 𝑑𝑇 𝑃
𝐶𝑉 =
𝑑𝑄
( ) ( )
=
𝑑𝐸
𝑑𝑇 𝑉 𝑑𝑇 𝑉
𝐶 𝑃 − 𝐶𝑉 =
𝛼2𝑉 𝑜 𝑇
𝛽

Q is the heat exchange, the volume thermal expansion coefficient

E the internal energy, the compressibility
H the enthalpy, Vo the molar volume

the difference between Cp and Cv is

negligibly small for condensed phase for most applications but significant
at elevated temperatures
 PROPERTIES FROM BONDING: Energy versus bond length

• Bond length, r F
F

• Bond energy, Eo r

Energy (r)

unstretched length
ro
r
Eo =
“bond energy”
 Thermal Expansion
o The specific volume of any given crystal increases with temperature, and the crystal
tends to become more symmetrical

o The general increase in volume with temperature is mainly determined by the

increased amplitude of atomic vibrations about a mean position
 The repulsion term between atoms changes
more rapidly with atomic separation than
does the attraction term

 Consequently, the minimum-energy (trough

is non-symmetrical) as the lattice energy
increases,

 The increased amplitude of vibration

between equivalent energy positions leads to
a higher value for the atomic separation
corresponding to a lattice expansion

 Thermodynamically, the structure energy

but the entropy decreases.
Thermal Expansion : CTE or 𝛼

• Coefficient of thermal expansion, a

length, Lo
Thermal Expansion coefficient (TEC)
unheated, T 1
L coeff. thermal expansion

L
= (T2 -T 1 )
heated, T 2 Lo

• 𝛼 ~ symmetry at ro
Energy

ro
r a is larger if Eo is smaller
larger  and very asymmetric.

small er 
 Atomic positions and vibrations
• The minimum in an atomic energy vs.
interatomic distance curve yields the near
neighbor distance (bond length).

• The width of the curve is proportional to

the amplitude of thermal vibrations for
an atom.
T3

• If the curve is symmetric, there is no shift

T2
in the average position of the atom (the
center of the thermal vibrations at any
T0
given T).

• The coefficient of thermal expansion is

negligible for symmetric energy wells.
Thermal Expansion

• If the curve is not symmetric, the average position in which the

atom sits shifts with temperature.

• Bond lengths therefore change (usually get bigger for increased T).

• Thermal expansion coefficient is nonzero.

Bond energy
r(T 1 )
r(T 5 )

Bond length (r)

increasing T

T5
bond energy vs bond length
curve is “asymmetric”
T1
 Thermal Expansion
the length and volume change is associated with temperature
At any particular temperature, the coefficient is define as
1 𝑑𝑙 1 𝑑𝑉 𝑙=𝑙𝑜 ( 1+𝛼 𝑇 )
linear 𝛼=
𝑙 𝑑𝑇 Volume 𝛽=
𝑉 𝑑𝑇 𝛽=3 𝛼
(1+T)

Heat Conduction The rate of heat transfer under temperature gradient

The basic equation for thermal conductivity which serves to define this term is
dQ is the amount of heat flowing normal to the area A in time dt

The heat flow is proportional to the temperature gradient, (-dT/dx), the proportionality factor
being a material constant, the thermal conductivity (k) 𝑑𝑄 𝑑𝑇
=− 𝐾𝐴
𝑑𝑡 𝑑𝑥
𝑇 2− 𝑇1
𝑞=− 𝐾𝐴
Under steady state condition Heat flux 𝑋 2 − 𝑋1

For radial heat flow through cylinder of length l inner diameter D1 and Outer diameter D2
𝑇 2− 𝑇1
𝑞=− 𝐾 ( 2 𝜋 𝑙 )
ln ( 𝐷2 / 𝐷1 )
 Thermal Conductivity
• General: The ability of a material to transfer heat.
• Quantitative: temperature
dT gradient Fick’s First Law
q  k
heat flux dx
(J/m2-s)
k=thermal conductivity (J/m-K-s):
Defines material’s ability to transfer heat.

T1 T2 > T1
x1 heat flux x2

Atomic view:
Electronic and/or Atomic vibrations in hotter region carry
energy (vibrations) to cooler regions. In a metal, electrons are
free and thus dominate thermal conductivity. In a ceramic,
phonons are more important.
Thermal Conductivity

• Non-Steady State: dT/dt is not constant.

T   T  T   2T  Fick’s ndSecond Law

 k  if K  f T    k  2  Fick ' s 2 Law
t x  x  t  x 

T1 T2 > T 1
x1 heat flux x2
Thermal Conductivity
K=kl+ke: Again think about band gaps: metals have lots of free electrons
(ke is large), while ceramics have few (only k l is active).

Material k (W/m-K) Energy Transfer

• Metals
Aluminum 247 By vibration of
Steel 52 atoms and
Tungsten 178 motion of
Gold 315 electrons
• Ceramics
Magnesia (MgO) 38
increasing k

Alumina (Al 2 O 3 ) 39 By vibration of

Soda-lime glass 1.7 atoms
Silica (cryst. SiO 2 ) 1.4
• Polymers
Polypropylene 0.12 By vibration/
Polyethylene 0.46-0.50 rotation of chain
Polystyrene 0.13 molecules
Teflon 0.25
Transient state heat flow
If the temperature is not constant

its rate of change with time depend up on

the ratio of the thermal conductivity to
the heat capacity per unit volume pCp

This ratio is called the thermal diffusivity k /pCp

and has the same units as material diffusivity, square centimeters per second.

dx

[ ( )
]
𝑑𝑇
𝑑 𝐾𝑥

( )
𝑑𝑇 𝑑𝑇 𝑑𝑥
𝑄𝑖𝑛 =− 𝐾 𝑥 ( 𝑑𝑦𝑑𝑧 ) 𝑑𝑡 𝑄𝑜𝑢𝑡 =− ( 𝑑𝑦𝑑𝑧 ) 𝐾 𝑥
𝑑𝑥
+
𝑑𝑥
𝑑𝑥 𝑑𝑡
𝑑𝑥 dy

(
𝑑 𝐾𝑥
𝑑𝑇
)
𝑄𝑖𝑛 −𝑄 𝑜𝑢𝑡 =( 𝑑𝑥𝑑𝑦𝑑𝑧 )
𝑑𝑥
𝑑𝑥
(
𝑑𝑡=( 𝑑𝑥𝑑𝑦𝑑𝑧 ) 𝜌 𝐶 𝑃
𝑑𝑇
𝑑𝑥 )
𝑑𝑡
𝑑
( 𝐾 𝑑𝑇
𝜌 𝐶 𝑝 𝑑𝑥 )
𝑑𝑇
=
𝑑𝑡 𝑑𝑥
 Heat Capacity analysis
The energy required for raising the temperature of a material goes into

• Vibrational energy
• Rotational energy for molecules in gases, liquids and crystal
• Raising the energy level of electrons in the structure
• Creating the defects in the crystal lattice eg Schottky & Frenkel

These changes correspond to increase in internal energy and accompanied by an increase in

configurational entropy
According to classical kinetic theory each atoms have an average
1 1
𝐾𝐸 = 𝐾𝑇
2
𝑃𝐸=
2
𝐾𝑇 Total energy 𝑇𝐸= 𝐾𝑇
For three degree of freedom Total energy E = 3KT

Energy per gram atom E = 3NKT

𝐶𝑉 = ( )
𝑑𝐸
𝑑𝑇 𝑉
=3 𝑁𝐾 =3 𝑅=5.96 𝑐𝑎𝑙/𝑔𝑚𝑎𝑡𝑜𝑚𝑜𝐶

Valid for high temperature

 Heat Capacity Temperature
At low temperature = C = Debye Temperature
𝐶𝑉
3 𝑁𝐾 (
=𝑓
𝑇 ) ( )
h 𝜈 𝑚𝑎𝑥 / 𝐾
=𝑓
𝐷
𝑇
h = Planck Constant
𝜈 = Vibrational frequency
At low temperatures the heat capacity The temperature () at which the heat
is proportional to capacity becomes constant or only
whereas at higher temperatures slightly varying with temperature
approaches unity depends on
so that the heat capacity becomes independent of temperature the bond strength, elastic
constants, and melting point of
the material and varies widely
for different materials.
At higher temperatures the heat capacity
at constant pressure also increases more
rapidly and deviates to a greater extent
from the constant-volume values.
Development of Frenkel and
Schottky defects, magnetic disorder,
electronic energy contributions, and
so on, contribute to the increased
value of heat capacity at higher
temperature.
Heat Capacity – The Dulong-Petit Law

3N0kb

D = Temperature at which CV  0.96% of 3 N o k B

 Heat Capacity

Heat capacity at order-disorder transformations of (a) Glass Transition 0.15Na2O-0.85B2O3

hydrogen bonds in KH2PO. and (b) Fe3+ ions in Fe2O3.
Heat capacity of various forms of
CaO +Si02 in 1: 1 molar ratio Values of the heat capacity increase particularly rapidly
the heat content changes discontinuously at a during the course of a cooperative process such as the
polymorphic transformation the change in heat order-disorder transformation
capacity, although also discontinuous, is usually Changes in heat capacity at magnetic and ferroelectric
not large transformations
most of the transformation from the ordered to the disordered structure occurs rapidly over a limited
temperature range
The molar heat capacity for crystalline materials is not structure-sensitive,
the value reported as volume heat capacity does depend on the porosity,
since the mass of material in a unit volume is decreased in proportion to the pore spaces present

The heat energy required to raise the temperature of an insulating firebrick is much lower than that required to raise the
temperature of a dense firebrick
 der
sor
-di
d er
Or e
civ
pla
Dis
Phase Transformation

Reconstructive
 Density and Crystal Structure
The volume of crystalline materials and their volume changes with temperature are closely related to the
crystal structures
The density is directly determined by the crystal structure, that is, the efficiency of atomic packing

Oxides have a basic structure of oxygen ions with interstices filled or partly filled by cations
Structures with close-packed oxygen ions have a high density of atomic packing and high values of
atoms per cubic centimeter
The density depends on Packing Density and on the atomic weight of the constituents
Structures with simple cubic packing of the oxygen atoms, such as UO2, have relatively low density of
atoms per cubic centimeter; with high-atomic-weight cations, such as uranium, the measured specific
gravity is high
Aluminum and silicon have similar atomic weights, but the density of Al 203 (hexagonal cIose packing of
oxygen ions) is 3.96 g/cm3 compared with a value of 2.65 for quartz, the common form of silica
This low density for silicates results from the network structures required by the low coordination number
and high valence of the silicon ion.
Higher-temperature forms of silica have an even lower density (cristobalite 2.32, tridymite
2.26), corresponding to more open structures
high-temperature polymorphic forms have a higher specific volume (lower density) than the low-
temperature forms. This corresponds to a discontinuous increase in volume at the transformation
temperature
Linear and non-linear Thermal Expansion 𝑙=𝑙𝑜 ( 1+𝛼 𝑇 ) (1+T)

oxides

Glass transition

polymorphic transformations
 Specific heat and thermal expansion
The change in volume due to lattice vibrations is closely related to the increase in energy
content
Consequently, changes in the thermal expansion coefficient, with temperature are parallel to
changes in heat capacity
The thermal expansion coefficient increases rapidly at low temperature and reaches a nearly
constant value above the Debye characteristic temperature (D)
Normally there is a continued increase observed above this temperature, resulting from the
formation of Frenkel or Schottky defects
 Anisotropy in thermal expansion

For nonisometric crystals the thermal expansion varies along different crystallographic axes.
In tetragonal crystals the c/a ratio decreases as the temperature is increased. The ratio of the
expansion coefficients ac /aa tends to decrease as the temperature is raised.
Anisometric expansion of layered crystalline structures (low TEC materials)

Graphite - the bonding is strongly directional and expansion is much lower in the plane of the
layer (1 x 10-6/oC ) than normal to it (27x 10-6/oC)
For strongly anisometric crystals the expansion coefficient in one direction may be negative, and
the resulting volume expansion may be very low.
Materials such as this are useful for thermal-shock applications.
Other materials aluminum titanate, cordierite, and various lithium aluminium silicates

𝛽-eucryptite, the overall volume expansion coefficient is negative.

the small or negative volume expansion is related to highly anisotropic structures

The absolute value of the expansion coefficient is closely related to the crystal structure and bond strength

tungsten, diamond, and silicon carbide, have low coefficients of thermal expansion due to high bond strength
 Small/Negative thermal expansion
• Invar (Ni-Fe alloy) is the most common low thermal exp material: α = 1.6*10-6 /
degree
• Some materials have α<0 in one dimension and >0 in others.
• It is possible, though not intuitive, for materials to have a negative thermal
expansion in all dimensions.
– An increase in temperature causes the crystal to shrink.
• ZrW2O8: contracts continuously and linearly from 2 to 1050K
• Composites could allow zero thermal expansion components (superb for
optics, engine parts, etc).
 Firing of poly crystalline body vis-a vis development of micro stress at the grain Boundary

Firing/Densification of a poly crystalline solid

A dense coherent structure results from viscous flow, diffusion, or solution and
precipitation on firing a polycrystalline body consisting of a mixture of crystalline
phases, or a mixture of crystals and glasses

Development of Micro stress

Stress free contraction

If the expansion coefficients in the different crystalline directions are not the
same or if the various phases present have different coefficients of thermal
expansion, different grains present have different amounts of contraction on
cooling and cracks open up between grains if occurs.

actual contraction

In practice, however, each grain is restrained by the surrounding grains so that,

instead of grain separation, micro stresses are developed which are proportional
to the difference between the stress free contraction and the actual contraction
 Thermal Expansion of composites
Assumptions:
no cracks develop, that the contraction of each grain is the same as the overall
contraction, and that all micro stresses are pure hydrostatic tension and compression
Then the stresses on each particle are given by 𝜎 𝑖 =𝐾 ( 𝛼𝑐 −𝛼 𝑖 ) ∆ 𝑇

where and are the average volume expansion coefficient and the volume expansion coefficient
for particle i, is the temperature change from the stress free state, and K is the bulk modulus,
𝑃 𝐸
𝐾 =−=
( ∆ 𝑉 /𝑉 ) 3 ( 1− 2 𝜇 )
where P is the isotropic pressure, V the volume, E the elastic modulus, and  Poisson's ratio.)

If the stresses are nowhere large enough to disrupt the structure, the summation of stresses
over an area or volume is zero. Consequently, if VI and V2 are the fractional volumes for a
mixture of materials,

Then
𝐹 𝑖𝑉 𝑐 𝜌𝑐
𝐾 1 ( 𝛼 𝑐 −𝛼 1 ) 𝑉 1 ∆ 𝑇 + 𝐾 2 ( 𝛼𝑐 −𝛼 2 ) 𝑉 2 ∆ 𝑇 +…=0 𝑉 1+𝑉 2 +…=𝑉 𝑐 𝑉 𝑖=
𝜌𝑖
where Fi is the weight fraction of phase i, and ps and Pi are the mean and individual phase
densities.
 Thermal Expansion of composites
Substituting in and eliminating ΔT, s and Vs gives an expression for the coefficient of volume
expansion of the aggregate (originally obtained by Turner)
𝛼1 𝐾 1 𝐹 1 / 𝜌 1 +𝛼2 𝐾 2 𝐹 2 /𝜌 2+…
𝛼𝑐 =
𝐾 1 𝐹 1 / 𝜌 1 + 𝐾 2 𝐹 2 /𝜌 2+ …
An alternative model for the expansion behavior of composite materials takes into account
shear effects at the boundaries between the grains or phases. It is here assumed that the overall
dilation of a composite body is as ΔT

where as is the overall expansion coefficient and ΔT

is the temperature difference between the initial
unstressed state and the final stressed state of
the body.

Analyzing the displacement of the individual

components and applying continuity relations at
the interfaces, the overall expansion coefficient can
be expressed by the relation first obtained by
Kerner

𝐾 1 ( 3 𝐾 2 +4 𝐺1 ) + ( 𝐾 2 − 𝐾 1 ) ( 16 𝐺1+12 𝐺1 𝐾 2)
2 2

𝛼 𝑠=𝛼1 +𝑉 2 ( 𝛼 2 − 𝛼1 ) G shear modulus

( 4 𝐺1 +3 𝐾 1 ) [ 4 𝑉 2 𝐺1 ( 𝐾 2 − 𝐾 1 ) +3 𝐾 1 𝐾 2+ 4 𝐺1 𝐾 1 ]
Effect of Polymorphic Transformations

Thermal expansion of two porcelain compositions.

Body A contains cristobalite as the silica phase; body
B contains both cristobalite and quartz.
 Hysteresis (microstresses)

The stresses developed as a result of

large differences between the expansion
coefficients of two materials or in
crystallographic directions in a
composite are sufficient to cause
microcracks to occur in the body.

One important result of these

microfissures is hysteresis in the thermal
expansion coefficient measured for a
polycrystalline aggregate or a composite
body in which large stresses are
developed.

Polycrystalline titania, (TiO2) when cooled from the firing temperature, develops microcracks,
and as these form, the overall observed expansion coefficient is lower than the expansion
coefficient of the individual crystals. On heating, these cracks tend to close, and at low
temperatures abnormally low expansion coefficients are observed.
This kind of expansion hysteresis occurs particularly in polycrystalline compositions when the
expansion coefficient is markedly different in different crystalline directions. It also occurs for
mixtures of materials having different expansion coefficients.
 Grain Boundary Fracture
Microstress fractures occur both within grains and along grain boundaries,
However most commonly observed at grain boundaries

The boundary stresses developed are independent of grain size

grain boundary cracking and thermal expansion hysteresis
occur predominately in large-grain samples

Microstresses can lead to microcracks and failure,

microstresses are developed in bodies that are restrained from
expansion by being held in a fixed support or by being attached to a material of
different expansion coefficient.
If a bar of material is completely restrained from
expanding, by application of restraining forces due to
the design of the part, the stress is given by
𝐸𝛼∆𝑇
𝜎=
1 −𝜇

where E is Young's modulus, ΔT is the temperature change, and

 is Poisson's ratio.
 Glaze stresses
Stresses are caused by the difference between the expansion coefficient of a glaze or enamel
and that of the underlying ceramic or metal.

On a stress free surface at To, the stresses developed depend on

the new temperature T', the elastic properties of the material, and on
the coefficients of expansion.

For a thin glaze on an infinite slab the stresses are given by (elastic properties of glaze and
body are the same)
𝜎 𝑔𝑙= 𝐸 ( 𝑇 𝑜 − 𝑇 ) ( 𝛼𝑔𝑙 − 𝛼𝑏 ) ( 1 −3 𝑗+6 𝑗 )
′ 2
A positive stress denotes tension
𝜎 𝑏=𝐸 ( 𝑇 𝑜 −𝑇 ) ( 𝛼𝑏 −𝛼 𝑔𝑙 ) ( 𝑗 ) ( 1− 3 𝑗+6 𝑗 )
′ 2
j is the ratio of glaze to body thickness.

To is taken as the setting point of the glaze, generally in its annealing range

For a thin glaze on a cylindrical body, the corresponding expressions are

𝐸 𝐴𝑏
𝜎 𝑔𝑙= ( 𝑇 − 𝑇 ) ( 𝛼 𝑔𝑙 − 𝛼𝑏 ) 𝐴
1 −𝜇 𝑜
′

where A, Ab , and Agl are, respectively, the cross-sectional

𝐸 𝐴𝑔𝑙 areas of the overall cylinder, body and glaze
𝜎 𝑏=
1− 𝜇
( 𝑇 𝑜 − 𝑇 ) ( 𝛼𝑏 −𝛼 𝑔𝑙 )
′
𝐴
 Crazing of Glaze
In order to obtain satisfactory fit between a glaze and body or porcelain-enamel and metal, it is
desirable that after cooling to room temperature the glaze be in a condition of compression

If tensile stresses develop, the glaze tends to craze

When the tensile stress developed become more
than the tensile strength of the glaze
If the glaze is put under compressive stress, this
type of failure does not occur, and it is only with
substantial stresses that the reverse, shivering, or
failure under compressive stresses takes place.
Crazing of Glaze
Typical glaze and enamel stresses in cooled ware are about 10,000 -psi
compression
Delayed crazing failures may occur in service in the bodies with compressive stress
Silicate bodies tend to increase in volume because of moisture absorption
The expansion of the body decreases the compression of the glaze and
transforms the stress into tension, and after sufficient time the ware tends to craze.
Prevention of (delayed) Crazing
it is desirable to have substantial initial compressive stress
Compositions adjustment to have minimum moisture expansion
This can be done with more vitreous compositions or by using alkali free compositions
Steatite tile, for example, is much superior to feldspar-clay-flint tile in this regard
The stresses developed on cooling from the firing temperature, where
the glaze or enamel is sufficiently fluid to relieve stresses, begin to
increase when the transformation range is passed
The actual temperature at which stresses begin to increase depends on
the rate of cooling
The change in stresses with temperature during cooling depends on
details of the cooling curves
The overall stress developed at room temperature depends only on the
difference between the total expansion of the glaze and body between
the stress point and room temperature and is independent of the
changes in stress during the cooling process

This is illustrated for two bodies in which the change in stress

with temperature is quite different for the two but the
resulting stress is similar
glazes have been used for decorative effects and to render ceramic
bodies impermeable to liquids
they are used to provide large surface compressions and hence to
strengthen ceramic bodies
moderate-expansion glazes are used to strengthen high-expansion
glass-ceramic bodies in the Na20-BaO-Al2O3-Si02 system
 Thermal Conduction
The heat-energy transfer under the influence of a temperature gradient depends on
the energy concentration present per unit volume,
Its velocity of movement, and
its rate of dissipation with the surroundings.
In gases,
the individual atoms or molecules exchange kinetic energy by collision
the heat energy present is simply equal to the heat capacity per unit volume
the velocity of molecular motion can be calculated from kinetic theory, and
the rate of energy dissipation depends on the rate of collision between atoms or molecules .
Let us consider a temperature gradient in which
the concentration of molecules is N and their average velocity is v,
the average rate at which molecules pass a unit area in the x direction is equal to 1/3Nv.
If energy equilibrium is obtained by collisions between molecules and the average distance between
collisions, the mean free path, is l,
molecules moving parallel to the x axis have an energy E0+ l(dE /dx),
where E0 is the mean energy at x = 0, l is the mean free path, and dE/dx is the energy gradient
in the x direction.
Combining these relations, the net energy flux in the x direction is given by

( ) ( )( ) ( )
𝑞 𝑑𝑇 1 𝑑𝐸 𝑑𝐸 𝑑𝐸 𝑑𝑇 𝑑𝑇 1
=𝑘 = 𝑁𝑣𝑙 𝑁 =𝑁 =𝑐 𝑘= 𝑐𝑣𝑙
𝐴 𝑑𝑥 3 𝑑𝑥 𝑑𝑥 𝑑𝑇 𝑑𝑥 𝑑𝑥 3
where c is the heat capacity per unit volume
Phonon Conductivity

The conduction of heat in dielectric solids may be considered

either the propagation of an harmonic elastic waves through a continuum or
the interaction between quanta of thermal energy called phonons
The frequency of these lattice waves covers a range of values, and scattering mechanisms or wave
1
interactions may depend on the frequency
The thermal conductivity can be represented in a general form
𝑘=
3
∫ 𝑐 ( 𝜔 ) 𝑣𝑙 ( 𝜔 ) 𝑑𝜔

where c() is the contribution to the specific heat per frequency interval for lattice waves of that frequency and l() is the
attenuation length for the lattice waves.
The major process giving rise to a finite thermal conductivity and energy dissipation from thermal elastic
waves is phonon-phonon interactions corresponding to phonon scattering called Umklapp processes.
(important over a wide temperature range)
Other Processes:
Anharmonicities due to lattice imperfections-----phonon scattering------- Decrease in mean free path
(above room temperature the imperfection scattering is independent of temperature and vibrational frequency of imperfection)
At low temperatures a variety of different scattering mechanisms resulting from the lattice imperfections gives rise to
a number of specific results.

If more than one process is operative, the effective value of 1/l is found by adding 1/l for each process;
this corresponds approximately to additivity of the corresponding thermal resistances for operative
processes.
 Photon Conductivity
In addition to the vibrational energy in solids, a much smaller fraction of the energy
content results from higher frequency electromagnetic radiation energy.

This fraction of the total energy is usually neglected in discussing heat capacity and thermal
conductivity, but it becomes important at high temperatures

because it is proportional to the fourth power of temperature.

The energy per unit volume of blackbody radiation at temperature T is given by

𝐸 𝑅=
4 𝜎 𝑛3 𝑇 4
𝑐
Heat capacity 𝐶 𝑅=( )
𝑑𝐸 16 𝜎 𝑛3 𝑇 3
𝑑𝑇
=
𝑐

where is the Stefan-Boltzmann constant (1.37 x 10-12 cal/cm2 sec °K4), c is the velocity of light (3 x 1010 cm/sec), and n
is the refractive index. Since the velocity of this radiation is v = c/n,

1 16 2 3
𝑘= 𝑐𝑣𝑙 Radiant energy conductivity 𝑘𝑟 = 𝜎 𝑛 𝑇 𝑙𝑟
3 3

where Ir is the mean free path of the radiant energy

 Temperature Dependence
The temperature dependence of phonon conduction in single crystal aluminum oxide

At very low temperatures the phonon mean free

path becomes of the same magnitude as the sample
size,
Boundary effects predominate, and the conductivity
decreases to zero at 0oK.

At some low temperature the thermal conductivity

reaches a maximum, and phonon-phonon
interactions lead to .

This exponential temperature dependence changes

to as the temperature level is raised above the
Debye temperature.

If the temperature is raised to a sufficiently high

level, the mean free path decreases to a value near
the lattice spacing, and the conductivity is expected
to be independent of temperature.
 Thermal Conductivity of oxides

Thermal conductivity for several oxides shows are

proportional to the inverse temperature above the
Debye temperature

The phonon mean free path in these materials ( I)

from Wave-velocity values

which is determined from the modulus of elasticity

by the relation

Elastic modulus of these polycrystaIline oxides,

decrease rapidly above 700 to 1000°C, result from
grain-boundary relaxation and creep

The low temperature linear relationship between E and T was extrapolated to obtain
high-temperature wave-velocity values. Values of the mean free path at room
temperature vary from a few to more than a hundred angstroms.
 Thermal conductivity
• Thermal conductivity is temperature
dependent.

– Analagous to electron scattering.

– Usually first decreases with

increasing temperature

• Higher Temp=more scattering

of electrons AND phonons,
thus less transfer of heat.

– Then increases at still higher

temperatures due to other
processes we haven‘t considered
in this class (radiative heat
transfer—eg. IR lamps).
 Mean Free path _Temperature

Three general kinds of behavior are observed.

At temperatures below the Debye
temperature,
as in Al203 , BeO, and MgO, the
inverse mean free path increases
more rapidly than linearity with
temperature.

Over a wide temperature range

near and above the Debye
temperature,
a linear increase in inverse mean
free path with temperature is
observed.
At high temperatures,
where the mean free path has decreased to a few angstroms, its value becomes
fixed independent of temperature, as observed for Ti02 and mullite.
Influence of Structure and Composition of Pure Materials
The extent of phonon scattering and the value of the mean free path should depend on the
anharmonicity of the lattice vibrations
Materials with complex structures
have a greater tendency toward thermal scattering of lattice waves and
consequently have a lower thermal conductivity

Magnesium aluminate spinel has a lower conductivity than either AI 203 or MgO,
although each of these has similar structure, expansion coefficient, heat capacity, and elasticity.

Mullite a complex structure, has a much lower thermal conductivity than

magnesium aluminate spinel.

The thermal conductivity for anisotropic crystal structures is also found to vary with the
direction in the crystal. This variation tends to decrease as the temperature level is raised

since anisotropic crystals always become more symmetrical as the temperature l evel
is raised

the thermal conductivity is greatest in the direction having the lowest thermal expansion
coefficient for Si02 (quartz), Ti02 (rutile), and graphite.
 Boundary Effects

The scattering of phonons at grain

boundaries results in a maximum
for thermal conductivity at very
low temperatures

The thermal conductivity of

polycrystalline and single-crystal
samples was identical at
temperatures below about 200°C.

At higher temperatures the single

crystal values depart from the
polycrystalline sample values as a
result of photon conductivity
 The three photon energy-transfer processes
between two boundaries separated by any material

 energy transferred directly from one boundary to the other without interaction
with the intervening material

the only effect of the material is to alter the photon velocity, and
consequently the rate of heat transfer, but its own temperature is not
affected by and has no influence on the rate of heat transfer

 energy transferred by photon energy exchange between the boundary and the
material

This transfer is limited to a region having a thickness of the order of

magnitude of the mean free path of the photons, and the rate of energy
exchange is determined by the temperature of the boundary and the
temperature gradients in the material

 Energy transferred by photon processes within the material, independent of the

boundaries

the only process which can be described as a material property

 Boundary effect of photon conductivity
Boundary effects are important as a practical matter because the values of photon mean
free paths commonly found for ceramic materials, 0.1 to 10 cm, are the same order of
magnitude as common sample sizes

if the distance between boundaries is d, a value of effective conductivity for the heat
transfer due to pure boundary processes can be defined in terms of the effective emissivity
e/(2-e) as

𝑞= ( 𝑒
2 −𝑒 )
2 3
4𝑛 𝜎𝑇 ∆𝑇 𝑘𝑏=𝑞 ( )
𝑑
∆𝑇
𝑘𝑏= ( 𝑒
2− 𝑒 )
4 𝜎 𝑛2 𝑇 3 𝑑

The ratio of the observed radiant-energy transfer when the mean free path is large
compared with the sample size and the maximum photon conductivity for an infinitely
large sample is, to a first approximation, given by
4 𝜎 𝑛2 𝑇 3 [ 𝑒 / ( 2− 𝑒 ) ] 𝑑 3
( )
𝑘𝑏 𝑒 𝑑
=𝑝= = 𝑓𝑜𝑟 𝑝<1
𝑘𝑟 2
(16 / 3 ) 𝜎 𝑛 𝑇 𝑙𝑟3
4 2− 𝑒 𝑙𝑟
Conductivity of Multiphase Ceramics
conductivity of a multiphase body depends on the amounts and arrangement of each phase
present as well as their individual conductivity
understanding of the microstructure is essential in interpreting
thermal-conductivity data

Idealized phase arrangements. (a) Parallel slabs((b) continuous major phase; (c) continuous minor phase
the parallel-slab arrangement (a) is not commonly observed, but a continuous major phase (b) with a minor amount of a
discontinuous second phase is typical of many microstructures. It is characteristic of the form of porosity normally
present. A discontinuous major phase (C) with a minor phase appearing as a continuous boundary material, for
example, is typical of the glass bond distribution the metal distribution In metal-bonded carbides, and the porosity In
insulating powders.
Thermal conductivity of composite (model)
For the simplest geometry, a series of slabs,
Heat flow
parallel to the plane of the slabs equivalent to a perpendicular to the plane of the slabs equivalent
parallel electrical circuit each slab has the same to an electrical series circuit the heat flow through
thermal gradient most of the heat flow is through each slab is equal the temperature gradients are
the better conductor The thermal conductivity is different The total conductivity is given by
given by 1 𝑣1 𝑣2 𝑘1 𝑘2
= + 𝑘𝑚=
𝑘𝑚=𝑣1 𝑘1+ 𝑣 2 𝑘2 𝑘𝑚 𝑘1 𝑘2 𝑣 1 𝑘2 +𝑣 2 𝑘1
𝑘2
𝑘1 ≫ 𝑘 2 𝑘𝑚 ≅ 𝑣 1 𝑘 1 𝑘 𝑚=
𝑣2

v1 and v2 are the volume fraction of each component, equal to the fraction of cross-sectional area

realistic approaches toward the structure of actual ceramics are illustrated Fig. b and c. (pre-page)
For these structures, relationships for the resultant conductivity derived by Maxwell have been-discussed in
terms of thermal conductivity by Eucken. If the continuous phase has a conductivity kc and the dispersed
phase has a conductivity kd , the resultant conductivity of the mixture is given by

𝑣 𝑑 ( 1 − 𝑘𝑐 / 𝑘 𝑑 )
1 +2 𝑘𝑐 >𝑘𝑑 ; 𝑘𝑚= 𝑘𝑐 [ ( 1− 𝑣 𝑑 ) / ( 1+ 𝑣 𝑑 ) ]
( 2 𝑘𝑐 / 𝑘𝑑 + 1 )
𝑘 𝑚= 𝑘𝑐
1−
𝑣 𝑑 ( 1 − 𝑘𝑐 / 𝑘 𝑑 ) 𝑘𝑑 >𝑘𝑐 ; 𝑘𝑚= 𝑘𝑐 [ ( 1+2 𝑣 𝑑 ) / ( 1 −𝑣 𝑑 ) ]
( 𝑘 𝑐 / 𝑘 𝑑 +1 )
Thermal conductivity of composite

a minor phase dispersed in a continuous major phase, the

conductivity changes along an S-shaped curve

Above 40% forsterite, the forsterite phase is continuous with

dispersed MgO grains within it; in samples containing smaller
amounts of forsterite, it is present as dispersed grains in a
continuous MgO matrix

glass phase is continuous in vitreous ceramics, so that the

conductivity of porcelains and fire clay is closer to that of the
glass contained than the conductivity of the crystalline phase

Thermal conductivity in the two-phase system MgO-Mg2SiO4

Effect of Porosity as secondary phase
 Pores acts as the scattering centers for photons
 A small fractions of porosity strongly reduce the photon mean free path and severely limit this conduction
mechanism.
 The effect of porosity on the phonon conductivity of ceramic systems is in approximate agreement with the
disperse phase equation
 At low temperatures the porosity has a low conductivity as compared with any solid phases present, and a
nearly linear decrease in conductivity with increasing porosity is found for dispersed pores in a solid.
 However; in addition to true conduction, radiation across pores contributes to heat transfer at high
temperature.
 Porosity conductivity High temperature
If the material surrounding the pores is opaque, an effective radiation conductivity of a pore can be
defined when the temperature gradient is small

since the radiant-energy transfer across a pore is proportional to the temperature difference
between pore surfaces
𝑞=𝑛2 𝜎 𝑒 𝑒𝑓𝑓 𝐴 ( 𝑇 41 −𝑇 42 )

In this equation can be factored, so that 𝑞=4 𝑛2 𝜎 𝑒𝑒𝑓𝑓 𝐴 𝑇 3𝑚 ∆ 𝑇

Considering heat transfer across a flat cavity, with parallel sides and thickness dp , an effective
conductivity is defined by
∆𝑇
𝑞=𝑘𝑒𝑓𝑓 𝐴
𝑑𝑝

An effective conductivity is thus defined which gives the proper heat flux when inserted in the normal
thermal conductivity relations

𝑘𝑒𝑓𝑓 =4 𝑑 𝑝 𝑛2 𝜎 𝑒𝑒𝑓𝑓 𝑇 3𝑚
Consequently the effect of radiation on pore conductivity is proportional to the pore size and to the third
power of temperature. Pores of larger size contribute to increasingly high conductivity at high
temperatures; small size porosity remains a good barrier to heat flow