MSN 501 - Thermal

Properties

Cüneyt Şahin | Spring 2024

Discover the world at Leiden University 1

 Thermal Properties
Thermal property refers to the response of a material to the application of heat. As a solid absorbs energy
in the form of heat, its temperature rises and its dimensions increase.

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 2

 Thermal Properties
Thermal property refers to the response of a material to the application of heat. As a solid absorbs energy
in the form of heat, its temperature rises and its dimensions increase.

The energy may be transported to cooler regions of the specimen if temperature gradients exist, and
ultimately, the specimen may melt.

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 3

 Thermal Properties
Thermal property refers to the response of a material to the application of heat. As a solid absorbs energy
in the form of heat, its temperature rises and its dimensions increase.

The energy may be transported to cooler regions of the specimen if temperature gradients exist, and
ultimately, the specimen may melt.

Heat capacity, thermal expansion, and thermal conductivity are properties that are often critical
in the practical utilization of solids.

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 4

 along their lengths. A change in temperature will cause this strip to bend; upon

1
Thermal Properties
heating, the metal having the greater expansion coefficient will elongate more pro-
ducing the directionrefers
of bending shown in Figure of
(a).a In these thermostats, the
Thermal property to the response material to the application C h absorbs
of heat. As a solid a p t eenergy
r
bimetallic strip is as a coil or spiral [Figure (b)]; this configuration
in the form of heat, its temperature rises and its dimensions increase. provides for a
relatively long bimetallic strip, more deflection for a given temperature change, (c)

Theand
energy
greatermay be transported
accuracy. to the
The metal having cooler
higherregions ofcoefficient
expansion the specimen if on
is located temperature
the underside gradients exist,
of the strip such and
that,
ultimately, thethe
upon heating, specimen
coil tends may melt.Attached to the end of the coil is a mercury switch—a small glass bulb that contains
to unwind.
several drops of mercury [Figure (b)]. This switch is mounted such that deflections of the coil end (when temperature is
Heat capacity, thermal expansion, and thermal conductivity are properties that are often critical
changed) tips the bulb one way or the other; accordingly the blob of mercury will roll from one end of the bulb to the other.
in the practical utilization of solids.
When temperature reaches the set point of the thermostat, electrical contact is made as the mercury rolls to one end; this
switches on the heating or cooling unit (e.g., furnace or air condi-
tioner). The unit shuts off when a limit temperature is achieved and
as the bulb tilts in the other direction, the blob of mercury rolls to
the other end, and electrical contact is broken.
The photograph of Figure (d) shows the consequences of a
scorching heatwave that occurred in Melbourne, Australia: rail lines
buckled as a result of stresses from unanticipated thermal expansion.
[Figures (a) and (b) courtesy of Electrical-Forensics, Ray Franca,
PhD, PE. Photograph of Figure (c) from iStockphoto. Photograph of
Figure (d) from Calum Robertson/Newspix.]
(a)
MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 5
 Phonon Heat Capacity
Consider phonons as a classical gas of particles transporting energy (hw) at a velocity (vg=group velocity).
Hot regions have a higher density of phonons than the cool regions.

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 6

 Phonon Heat Capacity
PHONON HEAT CAPACITY
Consider phonons as a classical gas of particles transporting energy (hw) at a velocity (vg=group velocity).
Hot regions have a higher density of phonons
By heat capacitythanwetheshall
cool regions.
usually mean the heat capacity at constan
ume, which is more fundamental than the heat capacity at constant pre
Specific heat capacity
which is what the experiments determine.1 The heat capacity at constan
ume is defined as CV " (%U/%T)V where U(constantis the energy
volume)and T the temper
The contribution
where of the
U is the energy and phonons to the heat capacity of a crystal is
T the temperature
the lattice heat capacity and is denoted by Clat. The total energy
phonons at a temperature "(" kBT) in a crystal may be written as the s
the energies over all phonon modes, here indexed by the wavevector
polarization index p:

Ulat " ##U

K p
K,p " # # !n
K p
K,p "!!K,p ,

where !nK,p " is the thermal equilibrium occupancy of phonons of wave

K and and
MSN 501 - Atomic Structure, Mechanical polarization p. The
Thermal Properties form of !nK,p " is given by the Planck distri
of Materials 7
 goes to a non-zero finite value in the long wave limit.
e is defined as CV " (%U/%T) Phonon V where Heat U isCapacity
the energy and T the temperatur
is further divided into two main categories: transver
PHONON HEAT
phonons. TheCAPACITY
terms “transversal” and “longitudinal” re
The contribution
Consider of the
phonons as a classical phonons
gas of to the heat
particles transporting atomiccapacity
energy (hw) ofu (t )a(vwith
at a velocity
displacements gcrystal
=group is direction
calle
velocity).
respect to
Hot regions have a higher density of phonons
By heat capacitythanwetheshall
cool regions.
usually mean for thetransversal
heat capacity at constan
lattice heat capacity and is denoted bygenerally Clat. twoThe
perpendicular
total energy
and parallel
of th
for long
ume, which is more fundamental than the heat capacity at constant pre
transverse and one longitudinal
Specific heat capacity acoustic phonons. Furthermore, the dispersion relati
branch
onons at a temperature what kthe
which is "(" BT) in a crystal
experiments may be
determine. 1
The
isotropic, meaning
written
heat
that the capacity
as the sum
at constan
phonon dispersion
o
relatio
energies over allume phonon
is defined modes,
as CV " here
(%U/%T)indexed
V where U by
different symmetry
is the
(constanttheenergywavevector
directions
volume) within
and T thethe K an
crystal.
temper
arization index p: The contribution
where of the
U is the energy and phonons to the heat capacity of a crystal is
T the temperature E
the lattice heat capacity and is denoted by Clat. The total energy
LO
phonons at a temperature "(" kBT) in a crystal may be written asTOthe s
Ulat
the energies
"
polarization
## UK,p
over
K index
p
all "
p:
##
!nK,p "!!
phonon modes,
K p
here ,
K,pindexed by the wavevector
TO (1
LA

ere !nK,p " is the thermal equilibrium #K #p K,p #K #p K,p K,p

Ulat "occupancy
U " of phonons
!n "!! of
, wavevecto
TA
TA
and polarization p. The form of !nK,p " is given by the Planck distributio
k/kma
nction: where !nK,p " is the thermal equilibrium occupancy of phonons of wave
0 1
K and and
MSN 501 - Atomic Structure, Mechanical polarization p. The
Thermal Properties form
of Materials of !nK,p " is given by the Planck distri
8
Fig. 5.9. Typical phonon dispersion spectrum for a three-dimens
 goes to a non-zero finite value in the long wave limit.
e is defined as CV " (%U/%T) Phonon V where Heat U isCapacity
the energy and T the temperatur
is further divided into two main categories: transver
PHONON HEAT
phonons. TheCAPACITY
terms “transversal” and “longitudinal” re
The contribution
Consider of the
phonons as a classical phonons
gas of to the heat
particles transporting atomiccapacity
energy (hw) ofu (t )a(vwith
at a velocity
displacements gcrystal
=group is direction
calle
velocity).
respect to
Hot regions have a higher density of phonons
By heat capacitythanwetheshall
cool regions.
usually mean for thetransversal
heat capacity at constan
lattice heat capacity and is denoted bygenerally Clat. twoThe
perpendicular
total energy
and parallel
of th
for long
ume, which is more fundamental than the heat capacity at constant pre
transverse and one longitudinal
Specific heat capacity acoustic phonons. Furthermore, the dispersion relati
branch
onons at a temperature what kthe
which is "(" BT) in a crystal
experiments may be
determine. 1
The
isotropic, meaning
written
heat
that the capacity
as the sum
at constan
phonon dispersion
o
relatio
energies over allume phonon
is defined modes,
as CV " here
(%U/%T)indexed
V where U by
different symmetry
is the
(constanttheenergywavevector
directions
volume) within
and T thethe K an
crystal.
temper
arization index p: The contribution
where of the
U is the energy and phonons to the heat capacity of a crystal is
T the temperature E
the lattice heat capacity and is denoted by Clat. The total energy
LO
phonons at a temperature "(" kBT) in a crystal may be written asTOthe s
Ulat
the energies
"
polarization
## UK,p
over
K index
p
all "
p:
##
!nK,p "!!
phonon modes,
K p
here ,
K,pindexed by the wavevector
TO (1
LA

ere !nK,p " is the thermal equilibrium

3 U
Debye model: T dependence
lat " #Kat#low
occupancy
p
U " #K #p K,p K,p
of
temperature
K,p phonons
!n "!! of
, wavevecto
TA
TA
(acoustic phonons)
and polarization p. Einstein
The form of !naccurate
model: mostly K,p
" is forgiven by the Planck distributio
optical phonons k/kma
nction: where !nK,p " is the thermal equilibrium occupancy of phonons of wave
0 1
K and and
MSN 501 - Atomic Structure, Mechanical polarization p. The
Thermal Properties form
of Materials of !nK,p " is given by the Planck distri
9
Fig. 5.9. Typical phonon dispersion spectrum for a three-dimens
 ration of a pair of atoms at a temperature
Anharmonic Effects - T. We take the
Thermal potential energy of the
Expansion
Weatoms atpotential
take the a displacement from at
x atoms
energy of the their equilibrium
a displacement separation
x from at absolute
their equilibrium zero
separation as
at absolute
zero as
U(x) ! cx2 " gx3 " fx4 , c, g, f positive (38)
harmonic
with c, g, and f all positive. The term in x3 represents the asymmetry of the
mutual repulsion of the atoms and the term in x4 represents the softening of the
vibration at large amplitudes. The minimum at x ! 0 is not an absolute mini-
mum, but for small oscillations the form is an adequate representation of an in-
teratomic potential.
We calculate the average displacement by using the Boltzmann distribu-
tion function, which weights the possible values of x according to their
thermodynamic probability:

% "!
!
dx x exp["!U(x)]
!x" !
MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials
! , 10
 ration of a pair of atoms at a temperature
Anharmonic Effects - T. We take the
Thermal potential energy of the
Expansion
Weatoms atpotential
take the a displacement from at
x atoms
energy of the their equilibrium
a displacement separation
x from at absolute
their equilibrium zero
separation as
at absolute
zero as
U(x) ! cx2 " gx3 " fx4 , c, g, f positive (38)
harmonic
with c, g, and f all positive. The term in x3 represents the asymmetry of the
the softening of the vibration
mutual repulsion ofasymmetry
the atomsof the
and the term in x4 represents
mutual
at large the softening of the
amplitudes
repulsion of the atoms
vibration at large amplitudes. The minimum at x ! 0 is not an absolute mini-
mum, but for small oscillations the form is an adequate representation of an in-
teratomic potential.
We calculate the average displacement by using the Boltzmann distribu-
tion function, which weights the possible values of x according to their
thermodynamic probability:

% "!
!
dx x exp["!U(x)]
!x" !
MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials
! , 11
 with and all positive. The term in 3
represents
teratomic potential.
thepotential
asymmetry of theof the
ration
separation of c,
afrom
pair of fatoms
g,Anharmonic
their at a temperature
equilibrium
Effects x We
position.
- T.
Thermal take theExpansion energy
mutual repulsion of the atoms and the term in x 4
represents We
the calculate
softening ofthe
the average
Weatoms atpotential
take the a displacement from at
x atoms
energy of the their equilibrium
a displacement x fromseparation at absolute
their equilibrium separationzero as
at absolute
𝑈(at𝑥 )large
zero as vibration = 𝑐𝑥 2
− 𝑔𝑥 3The
amplitudes. 𝑓𝑥 4
− minimum at x ! 0tionis notfunction,
an absolutewhich mini- weights
mum, but for small oscillations U(x)the
! form
cx2 "isgx
an3 " fx4 , thermodynamic
adequate representation of probability:
c, g, f positive an in- (38)
teratomic potential.
%
tive. harmonic
with c, g, and f all the
We calculate averageThe
positive. displacement
term in xby3 using the Boltzmann distribu-
represents the asymmetry of the
tionrepulsion
mutual function, of
which
asymmetry
the weights
of the
atoms thethe
possible
mutual
and term values
in x 4 theof x according
softening
represents the to theirof the
of the vibration
softening
n valuethermodynamic
of the displacement
repulsion using
of the
probability: atoms the Boltzmann distribution
at large amplitudes function!x" !
vibration at large amplitudes. The minimum at x ! 0 is not an absolute mini-
mum, but for small∫oscillations
〈𝑥 〉 = value
teratomicexpectation
potential.
∞
−∞
𝑑𝑥 𝑥𝑒 %
−𝛽𝑈(𝑥)
the!
form
dx x
is an adequate representation of an in-
exp["!U(x)]
!x"
"!
,with ! ! 1/kBT . For displacem
%
∞ !
−𝛽𝑈(𝑥)
We calculate the ∫−∞ 𝑑𝑥 𝑒 displacement
average
!
dx exp["!U(x)] by using energy the Boltzmann
are small in distribu-
comparison
tion function, which weights the possible values of x according to their
"!

−𝛽𝑈!(𝑥!) 1/k represents the probability ofsuch beingthat

at a particular position xterms
at a given temperature
term 𝑒with
thermodynamic T .
represents
B For displacements
probability: the probability of being# dx the anharmonic particular$the
at xaexp("!U) in position x
# dx [exp("
energy are small in comparison with kBT , we may expand the integrands as

%
f an Arrhenius equation in which ! high-energy processes are exponentially # dx exp("!U)
2dx x exp["!U(x)]
4 5 1/2 5/2 "3/2
# dx x exp("!U)
ure). The denominator is thus "! $ # dx [exp("!cx )](x $
a normalization!gx $ !fx ) !
factor (3" #4)(g
which #c )!
sums ; up all of
MSN 501 - Atomic Structure, Mechanical !x"
# dx exp("!U)
and ! $
Thermal # dx
Properties of Materials 2) ! ("
exp("!cx whence
# ,
!c) 1/2
, the thermal (39)expansion12
!
 tion
withfunction,
and which weights
all positive. the
The term possible
in 3
represents
teratomic
values of x potential.
according
thepotential
asymmetry of to their
theof
ration
separationof c,
afrom
pair of fatoms
g,Anharmonic
their at a temperature
equilibrium
Effects x We
position.
- T.
Thermal take the
Expansion energy the
thermodynamic
mutual repulsion probability:
of the atoms and the term in x 4
representsWe
the calculate
softening ofthe
the average
atoms at a displacement x from their equilibrium separation at absolute zero as
We take the potential energy of the atoms at a displacement x from their equilibrium separation at absolute
2
𝑈(𝑥 ) = 𝑐𝑥 − 𝑔𝑥 − 𝑓𝑥 3 4 tion function, which weights
zero as vibration at large amplitudes. The! minimum at x ! 0 is not an absolute mini-
mum, but for small oscillations
teratomic potential.
U(x)the
! form
cxdx" %
gx " fx4 , thermodynamic
2 is an3 adequate
x exp["!U(x)] representation of probability:
c, g, f positive an in- (38)

%
"!
!x" ,
%
tive. harmonic ! !
Weand
with c, g, calculate
f all the averageThe
positive. displacement
term in xby
3 using the Boltzmann distribu-
represents the asymmetry of the
dx exp["!U(x)]
tionrepulsion
mutual function, of
which
asymmetry
the weights
of the
atoms the
mutual
and the
"! possible
term values
in x 4 theof x according
softening
represents the to theirof the
of the vibration
softening
n valuethermodynamic
of the displacement
repulsion using
of the
probability: atoms the Boltzmann distribution
at large amplitudes function!x" !
vibration
with ! ! at 1/k
large
BT . amplitudes.
For displacementsThe minimum such that x ! 0anharmonic
at the is not an absolute terms inmini- the
mum,
energy butareforsmall
small∫
〈𝑥 〉 = value
teratomicexpectation
potential.
∞
inoscillations
comparison
−∞
𝑑𝑥 𝑥𝑒 %
−𝛽𝑈(𝑥)
the
with
!
form kBTis, we
an
dx x exp["!U(x)]
adequate
may expand representation
the integrandsofasan in-
!x"
"!
,with ! !1/2 1/kBT . 5/2 For "3/2displacem
%
∞ !
−𝛽𝑈(𝑥)
#We exp("!U)the
dx xcalculate ∫ 𝑑𝑥
$ #average
dx 𝑒
[exp("!cx ! 2
)](x $ !gx
displacement by
4
$using
dx exp["!U(x)]2 energy are
5
!fx )the ! (3" #
Boltzmann4)(g#c distribu-
)! ;
−∞ 1/2 small in comparison
tion function, which # dx exp("!U)
weights the $ # possible
"! dx exp("!cx ) ! ("
values of#!c) ,
x according (39)
to their
−𝛽𝑈!(𝑥!) 1/k represents the probability ofsuch beingthat
at a particular position xterms at a given temperature
term 𝑒with
thermodynamic B T .
represents For displacements
probability:
whence the thermal expansion the probability
is of the anharmonic
being # at
dx a
x particular
exp("!U) in
$the
position
# dx x
[exp("
energy are small in comparison thermal
withexpansion
kBT , we maybecomesexpand the integrands as

%
f an Arrhenius equation in which 3g
! high-energy processes are exponentially # dx exp("!U)
!x"
2dx! x kB4T
exp["!U(x)] 5 1/2 5/2 "3/2 (40)
ure). The # dxdenominator
x exp("!U) $ # dx is [exp("!cx
thus "! )](x $ 2
4c!gx $ !fx factor
a normalization ) ! (3" which #4)(g#c sums)! ; up all of
MSN 501 - Atomic Structure, Mechanical !x"
# dx exp("!U)
and ! $
Thermal # dx
Properties of Materials 2) ! ("
exp("!cx whence
# ,
!c) 1/2
, the thermal (39)expansion 13
!
 5 Phonons II. Thermal Properties
Anharmonic Effects - Thermal Expansion
Lattice constant of solid argon as a function of temperature

5.46 Triple point 1.62

5.42
Lattice constant, in A°

1.66

Density, g cm#3
5.38
1.70

5.34
1.74

5.30
1.78
0 20 40 60 80 Figure 15 Lattice constant of solid argon
Temperature, in K function of temperature.
MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 14
 5 Phonons II. Thermal Properties
Anharmonic Effects - Thermal Expansion
At an arbitrary temperature, we are seeking the thermal expansion coef
Lattice constant of solid argon as a function of temperature
We are assuming an isotropic material such that the volume thermal exp
5.46 Triple point 1.62
linear thermal expansion coefficient. thermal expansion coefficient

5.42 1 𝜕𝑙 1 𝜕𝑉
𝛼= ( ) = ( )
Lattice constant, in A°

1.66
𝑙 𝜕𝑇 𝑃 3𝑉 𝜕𝑇 𝑃

Density, g cm#3
5.38 The bulk modulus of a material is the volume contraction in response to
1.70
compression: slope of the Lattice constant vs. Temperature

5.34
1.74 𝜕𝑃
𝐵 = −𝑉 ( )
𝜕𝑉 𝑇
5.30
1.78
0 Plugging
20 in
40 the bulk
60 modulus
80 above and noting chain rule
Figure 15 Lattice constant of solid argon
Temperature, in K function of temperature.
MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 1 𝜕𝑃 15
 5 Phonons II. Thermal Properties
Anharmonic Effects - Thermal Expansion
At an arbitrary temperature, we are seeking the thermal expansion coef
Lattice constant of solid argon as a function of temperature
We are assuming an isotropic material such that the volume thermal exp
5.46 Triple point 1.62
linear thermal expansion coefficient. thermal expansion coefficient

5.42 1 𝜕𝑙 1 𝜕𝑉
𝛼= ( ) = ( )
Lattice constant, in A°

1.66
𝑙 𝜕𝑇 𝑃 3𝑉 𝜕𝑇 𝑃

Density, g cm#3
5.38 The bulk modulus of a material is the volume contraction in response to
1.70
compression: slope of the Lattice constant vs. Temperature

5.34
1.74
at𝐵low
𝜕𝑃
= −𝑉 ( )
T: α~T 3

𝜕𝑉 𝑇
at high T: α~constant
5.30
1.78
0 Plugging
20 in
40 the bulk
60 modulus
80 above and noting chain rule
Figure 15 Lattice constant of solid argon
Temperature, in K function of temperature.
MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 1 𝜕𝑃 16
 The thermal conductivity coefficient K of a solid is defined with respect to
Thermal
the steady-state flow of heat down Conductivity
a long rod with a temperature gradient
dT/dx
Thermal :
conductivity is defined as heat flow in response to a temperature gradient
dT
,
jU " #K (41)
dx
is the energy transmitted
where jU isacross a unit
the flux area per
of thermal unit or
energy, time
theand K is the
energy thermal conductivity
transmitted across unit coefficient.
area per unit time.
This form implies that the process of thermal energy transfer is a random
process. The energy does not simply enter one end of the specimen and pro-
ceed directly (ballistically) in a straight path to the other end, but diffuses
through the specimen, suffering frequent collisions. If the energy were propa-
gated directly through the specimen without deflection, then the expression
for the thermal flux would not depend on the temperature gradient, but only
on the difference in temperature !T between the ends of the specimen, re-
gardless of the length of the specimen. The random nature of the conductivity
process brings the temperature gradient and, as we shall see, a mean free path
into the expression for the thermal flux.

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 17

𝑗
 The thermal conductivity coefficient K of a solid is defined with respect to
Thermal
the steady-state flow of heat down Conductivity
a long rod with a temperature gradient
dT/dx
Thermal :
conductivity is defined as heat flow in response to a temperature gradient
dT
,
jU " #K (41)
dx
is the energy transmitted
where jU isacross a unit
the flux area per
of thermal unit or
energy, time
theand K is the
energy thermal conductivity
transmitted across unit coefficient.
area per unit time.
why is there a negative sign?
This form implies that the process of thermal energy transfer is a random
process. The energy does not simply enter one end of the specimen and pro-
ceed directly (ballistically) in a straight path to the other end, but diffuses
through the specimen, suffering frequent collisions. If the energy were propa-
gated directly through the specimen without deflection, then the expression
for the thermal flux would not depend on the temperature gradient, but only
on the difference in temperature !T between the ends of the specimen, re-
gardless of the length of the specimen. The random nature of the conductivity
process brings the temperature gradient and, as we shall see, a mean free path
into the expression for the thermal flux.

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 18

𝑗
 The thermal conductivity coefficient K of a solid is defined with respect to
Thermal
the steady-state flow of heat down Conductivity
a long rod with a temperature gradient
dT/dx
Thermal :
conductivity is defined as heat flow in response to a temperature gradient
dT
,
jU " #K (41)
dx
is the energy transmitted
where jU isacross a unit
the flux area per
of thermal unit or
energy, time
theand K is the
energy thermal conductivity
transmitted across unit coefficient.
area per unit time.
why is there a negative sign?
This form implies that the process of thermal energy transfer is a random
the negative sign The
process. reflects the intuitive
energy does notfact that enter
simply ‘heat’ flows from
one end ofhot
theregions intoand
specimen coldpro-
regions.
ceed directly (ballistically) in a straight path to the other end, but diffuses
through the specimen, suffering frequent collisions. If the energy were propa-
gated directly through the specimen without deflection, then the expression
for the thermal flux would not depend on the temperature gradient, but only
on the difference in temperature !T between the ends of the specimen, re-
gardless of the length of the specimen. The random nature of the conductivity
process brings the temperature gradient and, as we shall see, a mean free path
into the expression for the thermal flux.

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 19

𝑗
 The thermal conductivity coefficient K of a solid is defined with respect to
Thermal
the steady-state flow of heat down Conductivity
a long rod with a temperature gradient
dT/dx
Thermal :
conductivity is defined as heat flow in response to a temperature gradient
dT
,
jU " #K (41)
dx
is the energy transmitted
where jU isacross a unit
the flux area per
of thermal unit or
energy, time
theand K is the
energy thermal conductivity
transmitted across unit coefficient.
area per unit time.
why is there a negative sign?
This form implies that the process of thermal energy transfer is a random
the negative sign The
process. reflects the intuitive
energy does notfact that enter
simply ‘heat’ flows from
one end ofhot
theregions
specimenintoand
coldpro-
regions.
ceed directly (ballistically) in a straight path to the other end, but diffuses
throughconductivity
We encounter thermal the specimen, suffering
in many waysfrequent collisions.
in the real world If the energy were propa-
• When we heat
gatedupdirectly
a large pan on a tiny
through the electric
specimen burner and deflection,
without we have to wait
thenfor
thethe edges to reach the
expression
same temperature as the center
for the thermal flux would not depend on the temperature gradient, but only
on the difference in temperature !T between the ends of the specimen, re-
gardless of the length of the specimen. The random nature of the conductivity
process brings the temperature gradient and, as we shall see, a mean free path
into the expression for the thermal flux.

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 20

𝑗
 The thermal conductivity coefficient K of a solid is defined with respect to
Thermal
the steady-state flow of heat down Conductivity
a long rod with a temperature gradient
dT/dx
Thermal :
conductivity is defined as heat flow in response to a temperature gradient
dT
,
jU " #K (41)
dx
is the energy transmitted
where jU isacross a unit
the flux area per
of thermal unit or
energy, time
theand K is the
energy thermal conductivity
transmitted across unit coefficient.
area per unit time.
why is there a negative sign?
This form implies that the process of thermal energy transfer is a random
the negative sign The
process. reflects the intuitive
energy does notfact that enter
simply ‘heat’ flows
one endfromofhot
theregions
specimen intoand
coldpro-
regions.
ceed directly (ballistically) in a straight path to the other end, but diffuses
throughconductivity
We encounter thermal the specimen, suffering
in many waysfrequent collisions.
in the real world If the energy were propa-
• When we heat
gatedupdirectly
a large pan on a tiny
through the electric
specimen burner and deflection,
without we have to wait
thenfor
thethe edges to reach the
expression
same temperature as the center
for the thermal flux would not depend on the temperature gradient, but only
• When we touch
on thetwo different in
difference materials in thermal
temperature equilibrium
!T between the with
endsthe
of same room, such
the specimen, as wood and
re-
marble, and the latter
gardless feelslength
of the cooleroftothe
thespecimen.
touch (theThehigher thermal
random conductivity
nature of marble draws heat
of the conductivity
away from your fingers to the ‘cooler’ side which you are not touching)
process brings the temperature gradient and, as we shall see, a mean free path
into the expression for the thermal flux.

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 21

𝑗
 The thermal conductivity coefficient K of a solid is defined with respect to
Thermal
the steady-state flow of heat down Conductivity
a long rod with a temperature gradient
dT/dx
Thermal :
conductivity is defined as heat flow in response to a temperature gradient
dT
,
jU " #K (41)
dx
is the energy transmitted
where jU isacross a unit
the flux area per
of thermal unit or
energy, time
theand K is the
energy thermal conductivity
transmitted across unit coefficient.
area per unit time.
why is there a negative sign?
This form implies that the process of thermal energy transfer is a random
the negative sign The
process. reflects the intuitive
energy does notfact that enter
simply ‘heat’ flows
one endfromofhot
theregions
specimen intoand
coldpro-
regions.
ceed directly (ballistically) in a straight path to the other end, but diffuses
throughconductivity
We encounter thermal the specimen, suffering
in many waysfrequent collisions.
in the real world If the energy were propa-
• When we heat
gatedupdirectly
a large pan on a tiny
through the electric
specimen burner and deflection,
without we have to wait
thenfor
thethe edges to reach the
expression
same temperature as the center
for the thermal flux would not depend on the temperature gradient, but only
• When we touch
on thetwo different in
difference materials in thermal
temperature equilibrium
!T between the with
endsthe
of same room, such
the specimen, as wood and
re-
marble, and the latter
gardless feelslength
of the cooleroftothe
thespecimen.
touch (theThehigher thermal
random conductivity
nature of marble draws heat
of the conductivity
away from your fingers to the ‘cooler’ side which you are not touching)
process brings the temperature gradient and, as we shall see, a mean free path
• A challenge in modern
into electronics
the expression is tothermal
for the draw heat
flux.away from the processor, and appropriate materials
with high thermal conductivity can serve this purpose

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 22

𝑗
 the steady-state flow of heat down a long rod with a temperature gradient
dT/dx: Thermal Conductivity
dT
jU " #K , thermal conductivity (41)
=K
dx
where jU is the fluxheat
of thermal energy,
flow is not or the process
a ballistic energy transmitted
? across unit
area per unit time.
This form implies that the process of thermal energy transfer is a random
process. The energy does not simply enter one end of the specimen and pro-
ceed directly (ballistically) in a straight path to the other end, but diffuses
through the specimen, suffering frequent collisions. If the energy were propa-
gated directly through the specimen without deflection, then the expression
for the thermal flux would not depend on the temperature gradient, but only
on the difference in temperature !T between the ends of the specimen, re-
gardless of the length of the specimen. The random nature of the conductivity
process brings the temperature gradient and, as we shall see, a mean free path
into the expression for the thermal flux.

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 23

 the steady-state flow of heat down a long rod with a temperature gradient
dT/dx: Thermal Conductivity
dT
jU " #K , thermal conductivity (41)
=K
dx
where jU is the fluxheat
of thermal energy,
flow is not or the process
a ballistic energy transmitted
? across unit
area per unit time.
enter one end of the specimen and proceed directly (ballistically) in a straight path to the other end
This form implies that the process of thermal energy transfer is a random
but rather, involves a lot of collisions
process. The energy does not simply enter one end of the specimen and pro-
ceed directly (ballistically) in a straight path to the other end, but diffuses
through the specimen, suffering frequent collisions. If the energy were propa-
gated directly through the specimen without deflection, then the expression
for the thermal flux would not depend on the temperature gradient, but only
on the difference in temperature !T between the ends of the specimen, re-
gardless of the length of the specimen. The random nature of the conductivity
process brings the temperature gradient and, as we shall see, a mean free path
into the expression for the thermal flux.

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 24

 the steady-state flow of heat down a long rod with a temperature gradient
dT/dx: Thermal Conductivity
dT
jU " #K , thermal conductivity (41)
=K
dx
where jU is the fluxheat
of thermal energy,
flow is not or the process
a ballistic energy transmitted
? across unit
area per unit time.
enter one end of the specimen and proceed directly (ballistically) in a straight path to the other end
This form implies that the process of thermal energy transfer is a random
but rather, involves a lot of collisions
process. The energy does not simply enter one end of the specimen and pro-
ceed directly (ballistically) in a straight path to the other end, but diffuses
If the energy were propagated directly through the specimen without deflection, then the expression for
through the specimen, suffering frequent collisions. If the energy were propa-
the thermal flux would not depend on the temperature gradient, but only on the difference in temperature
ΔT between thegated
ends directly through regardless
of the specimen, the specimen without
of the length deflection, then the expression
of the specimen.
for the thermal flux would not depend on the temperature gradient, but only
on the difference in temperature !T between the ends of the specimen, re-
gardless of the length of the specimen. The random nature of the conductivity
process brings the temperature gradient and, as we shall see, a mean free path
into the expression for the thermal flux.

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 25

 marble draws heat away from
the steady-state your
flow of heatfingers
down to the ‘cooler’
a long rod withside which you are
a temperature not touching)
gradient
 dT/dx
A challenge :
in modern Thermal
electronics Conductivity
is to draw heat away from the processor, and appropriate
materials with high thermal conductivityj can dT this purpose
serve
" #KU , thermal conductivity (41)
=K
dx
mperature is anwhere incoherent phenomenon—changes
jU is the fluxheat
of thermal energy,
flow is not
in the
or temperature
a ballistic
of a system
energy transmitted
process ?
areunit
across associated with
nges in entropy. areaThus, heattime.
per unit flow is not a ballistic process, but rather, involves a lot of collisions. A
enter one end of the specimen and proceed directly (ballistically) in a straight path to the other end
ting point, we continueThis form
withimplies that the process
a particle-like of thermal
treatment energy transfer
of phonons. As shown is aearlier,
randomat higher
but rather, involves a lot of collisions
process. The energy
mperature, there is higher occupation doesornot simply enter
population one end ofmodes
of phonons the specimen and pro-
with energy less than or
ceed directly (ballistically) in a straight path to the other end, but diffuses
mparable to the temperature.
If the energy were As these
propagated directly phonons
through diffuse without
the specimen to the cooler sidethen
deflection, of athesolid, they will
expression forscatt
through the specimen, suffering frequent collisions. If the energy were propa-
the thermal
ny times and theflux would
value not
of depend
the on theconductivity
thermal temperature gradient,
coefficientbutwill
onlydepend
on the difference
on how in temperature
often they scatt
gated directly through the specimen without deflection, then the expression
ΔT between the ends of the specimen, regardless of the length of the specimen.
for the thermal flux would not depend 1 on the temperature gradient, but only
on the difference in temperature 𝐾 !T= between
𝐶𝑣ℓ the ends from of
thethe
kinetic theory re-
specimen, of gases
3
gardless of the length of the specimen. The random nature of the conductivity
C is heat
the the heat process
capacity
capacity brings
(at(atconstantthevolume,
constant temperature
volume, per gradient
perunit and,v as
unitvolume),
volume), we
isvthe shall see,particle
average
is the averagea mean free path
velocity,
particle the ℓ i
and l is and
velocity,
mean free path into
between collisions for the thermal flux.
the expression
mean free path between collisions. The factors of v and ℓ are fairly intuitive (why?), but C needs
mean free path: the average length a phonon (particle) propagates between two collisions!
me explaining.

en a particle
MSN moves
501 - Atomic from
Structure, a region
Mechanical with temperature
and Thermal 𝑇 + Δ𝑇 to a region with temperature T¸it will
Properties of Materials 26
 Thermal Conductivity
Examples
Table 2 Phonon mean free paths
[Calculated from (44), taking v ! 5 & 105 cm/sec as a representative sound velocity.
The !’s obtained in this way refer to umklapp processes.]

Crystal T, %C C, in J cm"3K"1 K, in W cm"1K"1 !, in Å

Quartza 0 2.00 0.13 40

"190 0.55 0.50 540
NaCl 0 1.88 0.07 23
"190 1.00 0.27 100

a
Parallel to optic axis.

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 27

 Thermal Conductivity
The factors of v and l are fairly intuitive, but C needs some explaining.

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 28

 Thermal Conductivity
PHONON HEAT CAPACITY
The factors of v and l are fairly intuitive, but C needs some explaining.
By heat capacity we shall usually mean the heat capacity at constan
When a particle moves from a region with temperature + Δ to a region with temperature T ̧it will give
ume, which is more fundamental than the heat capacity at constant pre
up Δ of energy
which is what the experiments
Why? determine.1 The heat capacity at constan
ume is defined as CV " (%U/%T)V where U is the energy and T the temper
The contribution of the phonons to the heat capacity of a crystal is
the lattice heat capacity and is denoted by Clat. The total energy
phonons at a temperature "(" kBT) in a crystal may be written as the s
the energies over all phonon modes, here indexed by the wavevector
polarization index p:

Ulat " ##U

K p
K,p " # # !n
K p
K,p "!!K,p ,

where !nK,p " is the thermal equilibrium occupancy of phonons of wave

K and polarization p. The form of !nK,p " is given by the Planck distri
MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 29
𝑐
𝑇
𝑇
𝑇
 3
Thermal Conductivity
PHONON HEAT CAPACITY
heat capacity (at constant volume, per unit volume), v is the average particle velocity, and ℓ is
The factors of v and l are fairly intuitive, but C needs some explaining.
n free path between collisions. The factors of v and ℓ are fairly intuitive (why?), but C needs
By heat capacity we shall usually mean the heat capacity at constan
When a particle moves from a region with temperature + Δ to a region with temperature T ̧it will give
plaining.
ume, which is more fundamental than the heat capacity at constant pre
up Δ of energy 1
particle moves from a region which
withistemperature
what the experiments
𝑇 +Why? determine.
Δ𝑇 to a region The heatT¸it
with temperature capacity
will at constan
Δ𝑇 of energy. ume is defined as CV " (%U/%T)V where U is the energy and T the temper
If this particle moves of
Theofcontribution
a distance one
of theinphonons to the heat capacity of a crystal is
rticle moves a distance one mean freemean free
path in path
this this process, between the ends of a free path of
process,
the particle is given by the lattice heat capacity and is denoted by Clat. The total energy
phonons at a𝑑𝑇temperature
𝑑𝑇 "(" k T) in a dT may be written as the s
crystal
Δ𝑇 = ℓ𝑥 = 𝑣𝑥 𝜏 B U = c vxτ
the energies𝑑𝑥 over all𝑑𝑥phonon modes, here dx indexed by the wavevector
isWhere is thetime
the average average time between
polarization
between collisions.
collisions.index p:the temperature
The subscript x has beengradient is only inthe
added because the x-direction for now
ture gradient is only in the x-direction for now

lux of energy is:

Ulat " ##U
K p
K,p " # # !n
K p
K,p "!!K,p ,

2
𝑑𝑇 1 2
𝑑𝑇
𝑗where
𝑢 = 𝑛𝑐 <
!nK,p𝑥" >
𝑣 is 𝜏the = 𝑛𝑐 〈𝑣 equilibrium
thermal 〉𝜏 occupancy of phonons of wave
𝑑𝑥 3 𝑑𝑥
MSN 501 - Atomic
K and polarization p. The form of !nK,p " is given by the Planck distri
30
2 Structure, Mechanical and Thermal Properties of Materials
a factor of 𝑣 because the first factor of 𝑣 comes from the amount of energy the particle gives
𝑐
𝑇
𝜏
𝑇
𝑇
 3
Thermal Conductivity
PHONON HEAT CAPACITY
heat capacity (at constant volume, per unit volume), v is the average particle velocity, and ℓ is
The factors of v and l are fairly intuitive, but C needs some explaining.
n free path between collisions. The factors of v and ℓ are fairly intuitive (why?), but C needs
By heat capacity we shall usually mean the heat capacity at constan
When a particle moves from a region with temperature + Δ to a region with temperature T ̧it will give
plaining.
ume, which is more fundamental than the heat capacity at constant pre
up Δ of energy 1
particle moves from a region which
withistemperature
what the experiments
𝑇 +Why? determine.
Δ𝑇 to a region The heatT¸it
with temperature capacity
will at constan
Δ𝑇 of energy. ume is defined as CV " (%U/%T)V where U is the energy and T the temper
If this particle moves of
Theofcontribution
a distance one
of theinphonons to the heat capacity of a crystal is
rticle moves a distance one mean freemean free
path in path
this this process, between the ends of a free path of
process,
the particle is given by the lattice heat capacity and is denoted by Clat. The total energy
phonons at a𝑑𝑇temperature
𝑑𝑇 "(" k T) in a dT may be written as the s
crystal
Δ𝑇 = ℓ𝑥 = 𝑣𝑥 𝜏 B U = c vxτ
the energies𝑑𝑥 over all𝑑𝑥phonon modes, here dx indexed by the wavevector
isWhere is thetime
the average average time between
polarization
between collisions.
collisions.index p:the temperature
The subscript x has beengradient is only inthe
added because the x-direction for now
ture gradient is only in the x-direction for now dT 2 dT 1 2 dT
ju = (−nvx)(cU "vxτ) =Uncτv = "!!ncτ <, vx >
The net flux of energy is:
lux of energy is: dx
lat #K #p K,p #Kdx#p K,p3 K,p
" x !n dx
rate of flow of n 2
𝑑𝑇amount
1 of2energy 𝑑𝑇
𝑗where
𝑢 = 𝑛𝑐 !n< K,p
𝑣𝑥" >is 𝜏theone
= 𝑛𝑐 〈𝑣 equilibrium
thermal
particle 〉𝜏
gives up occupancy of phonons of wave
particles 𝑑𝑥 3 𝑑𝑥
K and polarization p. The form of !nK,p " is given by the Planck distri
MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 31
a factor of 𝑣 2 because the first factor of 𝑣 comes from the amount of energy the particle gives
𝑐
𝑇
𝜏
𝑇
𝑇
 3
Thermal Conductivity
PHONON HEAT CAPACITY
heat capacity (at constant volume, per unit volume), v is the average particle velocity, and ℓ is
The factors of v and l are fairly intuitive, but C needs some explaining.
n free path between collisions. The factors of v and ℓ are fairly intuitive (why?), but C needs
By heat capacity we shall usually mean the heat capacity at constan
When a particle moves from a region with temperature + Δ to a region with temperature T ̧it will give
plaining.
ume, which is more fundamental than the heat capacity at constant pre
up Δ of energy 1
particle moves from a region which
withistemperature
what the experiments
𝑇 +Why? determine.
Δ𝑇 to a region The heatT¸it
with temperature capacity
will at constan
Δ𝑇 of energy. ume is defined as CV " (%U/%T)V where U is the energy and T the temper
If this particle moves of
Theofcontribution
a distance one
of theinphonons to the heat capacity of a crystal is
rticle moves a distance one mean freemean free
path in path
this this process, between the ends of a free path of
process,
the particle is given by the lattice heat capacity and is denoted by Clat. The total energy
phonons at a𝑑𝑇temperature
𝑑𝑇 "(" k T) in a dT may be written as the s
crystal
Δ𝑇 = ℓ𝑥 = 𝑣𝑥 𝜏 B U = c vxτ
the energies𝑑𝑥 over all𝑑𝑥phonon modes, here dx indexed by the wavevector
isWhere is thetime
the average average time between
polarization
between collisions.
collisions.index p:the temperature
The subscript x has beengradient is only inthe
added because the x-direction for now
ture gradient is only in the x-direction for now dT 2 dT 1 2 dT
ju = (−nvx)(cU "vxτ) =U− ncτv ="!!
− ncτ < vx >
The net flux of energy is:
lux of energy is: dx
lat #K #p K,p #K #pdx K,p K,p3
" x !n , dx
rate of flow of n 2
𝑑𝑇amount
1 of2energy 𝑑𝑇
𝑗where
𝑢 = 𝑛𝑐 !n< K,p
𝑣𝑥" >is 𝜏theone
= 𝑛𝑐 〈𝑣 equilibrium
thermal
particle 〉𝜏
gives up occupancy of phonons of wave
particles 𝑑𝑥 3 𝑑𝑥
K and polarization p. The form of !nK,p " is given by the Planck distri
MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 32
a factor of 𝑣 2 because the first factor of 𝑣 comes from the amount of energy the particle gives
𝑐
𝑇
𝜏
𝑇
𝑇
 3
Thermal Conductivity
PHONON HEAT CAPACITY
heat capacity (at constant volume, per unit volume), v is the average particle velocity, and ℓ is
The factors of v and l are fairly intuitive, but C needs some explaining.
n free path between collisions. The factors of v and ℓ are fairly intuitive (why?), but C needs
By heat capacity we shall usually mean the heat capacity at constan
When a particle moves from a region with temperature + Δ to a region with temperature T ̧it will give
plaining.
ume, which is more fundamental than the heat capacity at constant pre
up Δ of energy 1
particle moves from a region which
withistemperature
what the experiments
𝑇 +Why? determine.
Δ𝑇 to a region The heatT¸it
with temperature capacity
will at constan
Δ𝑇 of energy. ume is defined as CV " (%U/%T)V where U is the energy and T the temper
If this particle moves of
Theofcontribution
a distance one
of theinphonons to the heat capacity of a crystal is
rticle moves a distance one mean freemean free
path in path
this this process, between the ends of a free path of
process,
the particle is given by the lattice heat capacity and is denoted by Clat. The total energy
phonons at a𝑑𝑇temperature
𝑑𝑇 "(" k T) in a dT may be written as the s
crystal
Δ𝑇 = ℓ𝑥 = 𝑣𝑥 𝜏 B U = c vxτ
the energies𝑑𝑥 over all𝑑𝑥phonon modes, here dx indexed by the wavevector
isWhere is thetime
the average average time between
polarization
between collisions.
collisions.index p:the temperature
The subscript x has beengradient is only inthe
added because the x-direction for now
ture gradient is only in the x-direction for now dT 2 dT 1 2 dT
ju = (−nvx)(cU "vxτ) =U− ncτv ="!!
− ncτ <v>
The net flux of energy is:
lux of energy is: dx
lat #K #p K,p #K #pdx K,p K,p3
" x !n , dx
rate of flow of n 2
𝑑𝑇amount
1 of2energy 𝑑𝑇
𝑗where
𝑢 = 𝑛𝑐 !n< K,p
𝑣𝑥" >is 𝜏theone
= 𝑛𝑐 〈𝑣 equilibrium
thermal
particle 〉𝜏
gives up occupancy of phonons of wave
particles 𝑑𝑥 3 𝑑𝑥
K and polarization p. The form of !nK,p " is given by the Planck distri
MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 33
a factor of 𝑣 2 because the first factor of 𝑣 comes from the amount of energy the particle gives
𝑐
𝑇
𝜏
𝑇
𝑇
 3
Thermal Conductivity
PHONON HEAT CAPACITY
heat capacity (at constant volume, per unit volume), v is the average particle velocity, and ℓ is
The factors of v and l are fairly intuitive, but C needs some explaining.
n free path between collisions. The factors of v and ℓ are fairly intuitive (why?), but C needs
By heat capacity we shall usually mean the heat capacity at constan
When a particle moves from a region with temperature + Δ to a region with temperature T ̧it will give
plaining.
ume, which is more fundamental than the heat capacity at constant pre
up Δ of energy 1
particle moves from a region which
withistemperature
what the experiments
𝑇 +Why? determine.
Δ𝑇 to a region The heatT¸it
with temperature capacity
will at constan
Δ𝑇 of energy. ume is defined as CV " (%U/%T)V where U is the energy and T the temper
If this particle moves of
Theofcontribution
a distance one
of theinphonons to the heat capacity of a crystal is
rticle moves a distance one mean freemean free
path in path
this this process, between the ends of a free path of
process,
the particle is given by the lattice heat capacity and is denoted by Clat. The total energy
phonons at a𝑑𝑇temperature
𝑑𝑇 "(" k T) in a dT may be written as the s
crystal
Δ𝑇 = ℓ𝑥 = 𝑣𝑥 𝜏 B U = c vxτ
the energies𝑑𝑥 over all𝑑𝑥phonon modes, here dx indexed by the wavevector
isWhere is thetime
the average average time between
polarization
between collisions.
collisions.index p:the temperature
The subscript x has beengradient is only inthe
added because the x-direction for now
ture gradient is only in the x-direction for now dT 2 dT 1 2 dT
ju = (−nvx)(cU "vxτ) =U− ncτv ="!!
− ncτ <v>
The net flux of energy is:
lux of energy is: dx
lat #K #p K,p #K #pdx K,p K,p3
" x !n , dx
rate of flow of n 2
𝑑𝑇amount
1 of2energy 𝑑𝑇
𝑗where
𝑢 = 𝑛𝑐 !n< K,p
𝑣𝑥" >is 𝜏theone
= 𝑛𝑐 〈𝑣 equilibrium
thermal
particle 〉𝜏
gives up
C=ncof wave
occupancy of phonons
particles 𝑑𝑥 3 𝑑𝑥
K and polarization p. The form of !nK,p " is given by the Planck distri
MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 34
a factor of 𝑣 2 because the first factor of 𝑣 comes from the amount of energy the particle gives
𝑐
𝑇
𝜏
𝑇
𝑇
 Thermal Conductivity?
If, as for phonons, v is constant, we may write this as

1 2 dT
jU = − ncτ < v > with ! ! v! and C ! nc
3 dx

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 35

 𝑗𝑢 = 𝑛𝑐 < 𝑣𝑥 >
Thermal Conductivity?
2
There is a factor of 𝑣 write this as the first factor of 𝑣
𝑥 because
If, as for phonons, v is constant, we may
up and the1second factor dT comes from the rate of f
jU = − ncτ < v >2
denoted by3 n and nc=Cdxiswith the heat v! and Cper
! ! capacity ! nc
un
1 𝑑𝑇
Thus, 𝑗𝑈 = − 𝐶𝑣ℓ and the previous ‘assertion
3 𝑑𝑥

Causes of thermal resistivity

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 36

 𝑗𝑢 = 𝑛𝑐 < 𝑣𝑥 >
where ! is the average time between
Thermal collisions.
Conductivity?
The net fluxIf,ofas There
energy is (from
a factor both
of 𝑣𝑥
senses
2
because ofasthe
the particle
first flux)
factor of 𝑣
THERMAL CONDUCTIVITY
for phonons, v is constant, we may write this
up and the1second factor dT comes from the rate of f
j = − ncτ2 < v dT
> 2
with
1
! 2! v! dTand C ! nc
The thermal conductivity jU !
denotedU "n!v
coefficient
by3 n xand" c!Knc=C
of !a
dx "
is 3 n!v
solid
the is"defined
heat c!capacity. with
per res
un
dx dx
teady-state flow of heat down a1 long𝑑𝑇rod with a temperature g
x: If, as for phonons, v is
Thus, 𝑗𝑈 =
constant, − we𝐶𝑣ℓ may write
and the (43) as
previous ‘assertion
3 𝑑𝑥

Causes of thermaldT dT
resistivity
1
jU ! "3Cv!
jU " #K , ,
dx dx
e with
jU is!the flux of thermal energy,
! v! and C ! nc. Thus K ! 3Cv!. or
1 the energy transmitted acro

per unit time. the previous ‘assertion’ for the expression for K is confirmed
This form implies that the process of thermal energy transfer is a r
MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 37
 Thermal Resistivity
Phonons are not conducted ballistically from one end of the sample to another (if this were the case,
thermal conductivity would solely be determined by sound velocity). Instead, a number of scattering
processes introduce a ‘thermal resistivity’ (resistivity is inverse of conductivity). These include:

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 38

 Thermal Resistivity
Phonons are not conducted ballistically from one end of the sample to another (if this were the case,
thermal conductivity would solely be determined by sound velocity). Instead, a number of scattering
processes introduce a ‘thermal resistivity’ (resistivity is inverse of conductivity). These include:
• Imperfections in the crystal such as impurities, inhomogeneities, and crystal grain boundaries
• The edge or surface of the crystal
• Scattering off another phonon (this requires an anharmonic potential, which typically holds for
real materials because the atomic potential is not perfectly quadratic)

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 39

 Thermal Resistivity
Phonons are not conducted ballistically from one end of the sample to another (if this were the case,
thermal conductivity would solely be determined by sound velocity). Instead, a number of scattering
processes introduce a ‘thermal resistivity’ (resistivity is inverse of conductivity). These include:
• Imperfections in the crystal such as impurities, inhomogeneities, and crystal grain boundaries
• The edge or surface of the crystal
• Scattering off another phonon (this requires an anharmonic potential, which typically holds for
real materials because the atomic potential is not perfectly quadratic)

The theory of the effect of anharmonic coupling on thermal resistivity predicts that l is proportional to
1/T at high temperatures, in agreement with many experiments.

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 40

 Thermal Resistivity
Phonons are not conducted ballistically from one end of the sample to another (if this were the case,
thermal conductivity would solely be determined by sound velocity). Instead, a number of scattering
processes introduce a ‘thermal resistivity’ (resistivity is inverse of conductivity). These include:
• Imperfections in the crystal such as impurities, inhomogeneities, and crystal grain boundaries
• The edge or surface of the crystal
• Scattering off another phonon (this requires an anharmonic potential, which typically holds for
real materials because the atomic potential is not perfectly quadratic)

The theory of the effect of anharmonic coupling on thermal resistivity predicts that l is proportional to
1/T at high temperatures, in agreement with many experiments.

We can understand this dependence in terms of the number of phonons with which a given phonon can
interact: at high temperature the total number of excited phonons is proportional to T. The collision
frequency of a given phonon should be proportional to the number of phonons with which it can
collide, whence l ~ 1/T.

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 41

 Thermal Resistivity
Let’s start with this argument:
• Scattering off another phonon (this requires an anharmonic potential, which typically holds for
real materials because the atomic potential is not perfectly quadratic)

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 42

ources and sinks. Net energy transport under a temperature gradient occurs as in (b).
Thermal Resistivity
Let’s start with this argument:
• Scattering off another phonon (this requires an anharmonic potential, which typically holds for
real materials because the atomic potential is not perfectly quadratic)
Ky Ky

in 2D

Imagine a process where two phonons of momentum

K2
and collide and produce a third phonon with
K1 K1
momentum . Another way to say this is that two
K x phonons annihilate and produce a new one with Kx
K2 momentum = + . K1 ! K2
K3

K3 G

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 43

𝑲
𝑲
𝑲
𝑲
𝑲
𝑲
 Thermal Resistivity
ng this process, the occupation numbers of these three modes is as follows:
Following this process, the occupation numbers of these three modes is as follows:
𝑛𝑲𝟏 → 𝑛𝑲𝟏 − 1 = 𝑛𝑲𝟏 ′

𝑛𝑲𝟐 → 𝑛𝑲𝟐 − 1 = 𝑛𝑲𝟐 ′

𝑛𝑲𝟑 → 𝑛𝑲𝟑 + 1 = 𝑛𝑲𝟑 ′

conservation requires that:

𝑛𝑲𝟏 ℏ𝜔(𝑲𝟏 ) + 𝑛𝑲𝟐 ℏ𝜔(𝑲𝟐 ) + 𝑛𝑲𝟑 ℏ𝜔(𝑲𝟑 ) = 𝑛𝑲𝟏 ′ℏ𝜔(𝑲𝟏 ) + 𝑛𝑲𝟐 ′ℏ𝜔(𝑲𝟐 ) + 𝑛𝑲𝟑 ′ℏ𝜔(𝑲

= 𝑛𝑲𝟏 ℏ𝜔(𝑲𝟏 ) − ℏ𝜔 (𝑲𝟏 ) + 𝑛𝑲𝟐 ℏ𝜔(𝑲𝟐 ) − ℏ𝜔 (𝑲𝟐 ) + 𝑛𝑲𝟑 ℏ𝜔(𝑲𝟑 ) + ℏ𝜔(𝑲𝟑 )

0 = −ℏ𝜔(𝑲𝟏 ) − ℏ𝜔(𝑲𝟐 ) + ℏ𝜔 (𝑲𝟑 )

ntum conservation requires that:
MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 44
 Thermal Resistivity
ng this process,
ollowing thethe
this process, occupation numbers
occupation numbersof
ofthese threemodes
these three modesisisasas follows:
follows:
Following this process, the occupation numbers of these three modes is as follows:
𝑛𝑲
𝑛𝟏𝑲𝟏→→𝑛𝑛𝑲𝑲𝟏𝟏 −
−1 =
= 𝑛𝑛𝑲𝑲𝟏𝟏′ ′
𝑛𝑲𝟐→→𝑛𝑛𝑲𝟐 −
𝑛𝑲 −11=
= 𝑛𝑛𝑲𝟐 ′ ′
𝟐 𝑲 𝟐 𝑲𝟐
𝑛𝑲𝟑 → 𝑛𝑲𝟑 + 1 = 𝑛𝑲𝟑 ′
𝑛𝑲𝟑
→ 𝑛𝑲𝟑 + 1 = 𝑛𝑲𝟑 ′
nergy conservation requires that: Energy conservation requires that:
conservation requires that:
𝑛𝑲𝟏 ℏ𝜔(𝑲𝟏 ) + 𝑛𝑲𝟐 ℏ𝜔(𝑲𝟐 ) + 𝑛𝑲𝟑 ℏ𝜔(𝑲𝟑 ) = 𝑛𝑲𝟏 ′ℏ𝜔(𝑲𝟏 ) + 𝑛𝑲𝟐 ′ℏ𝜔(𝑲𝟐 ) + 𝑛𝑲𝟑 ′ℏ𝜔(𝑲𝟑 )
𝑛𝑲𝟏 ℏ𝜔(𝑲𝟏 ) + 𝑛𝑲𝟐 ℏ𝜔(𝑲𝟐 ) + 𝑛𝑲𝟑 ℏ𝜔(𝑲𝟑 ) = 𝑛𝑲𝟏 ′ℏ𝜔(𝑲𝟏 ) + 𝑛𝑲𝟐 ′ℏ𝜔(𝑲𝟐 ) + 𝑛𝑲𝟑 ′ℏ𝜔(𝑲
= 𝑛𝑲𝟏 ℏ𝜔(𝑲𝟏 ) − ℏ𝜔 (𝑲𝟏 ) + 𝑛𝑲𝟐 ℏ𝜔(𝑲𝟐 ) − ℏ𝜔 (𝑲𝟐 ) + 𝑛𝑲𝟑 ℏ𝜔(𝑲𝟑 ) + ℏ𝜔(𝑲𝟑 )
= 𝑛𝑲𝟏 ℏ𝜔(𝑲𝟏 ) − ℏ𝜔 (𝑲𝟏 ) + 𝑛𝑲𝟐 ℏ𝜔(𝑲𝟐 ) − ℏ𝜔 (𝑲𝟐 ) + 𝑛𝑲𝟑 ℏ𝜔(𝑲𝟑 ) + ℏ𝜔(𝑲𝟑 )
0 = −ℏ𝜔(𝑲𝟏 ) − ℏ𝜔(𝑲𝟐 ) + ℏ𝜔 (𝑲𝟑 )
0 = −ℏ𝜔
Momentum conservation requires that:
(𝑲𝟏 ) − ℏ𝜔 (𝑲𝟐 ) + ℏ𝜔 (𝑲𝟑 )

ntum conservation requires that: 𝑲 +𝑲 = 𝑲 +𝑮

𝟏
MSN 501 - Atomic Structure, Mechanical and Thermal Properties of𝟐Materials
𝟑 45
 Thermal Resistivity
ng thisthis
lowing
ollowingprocess, thethe
thisprocess,
process, occupation
theoccupationnumbers
occupation numbersof
numbers ofthese
of these threemodes
these three
three modesisisisas
modes asas follows:
follows:
follows:
Following this process, the occupation numbers of these three modes is as follows:
𝑛𝑛𝑲 →
𝑛𝑲𝟏𝑲𝟏𝟏→→𝑛𝑛𝑲𝑲𝑲𝟏𝟏𝟏 −
−11=
= 𝑛𝑛𝑲𝑲𝑲𝟏𝟏𝟏′′

𝑛𝑛𝑲 →
𝑛𝑲𝑲𝟐𝟐→
𝟐
→𝑛𝑛𝑲𝑲𝟐𝟐 −
𝑲𝟐
−11=
= 𝑛𝑛𝑲𝟐𝟐 ′′
𝑲𝟐
𝑛𝑛𝑲𝑲𝟑𝟑 →
→ 𝑛𝑲𝑲𝟑𝟑 + 1 = 𝑛𝑲𝟑 ′
𝑛𝑲𝟑 → 𝑛𝑲𝟑 + 1 = 𝑛𝑲𝟑 ′
ergy
nergyconservation
conservationrequires
requiresthat:
that: Energy conservation requires that:
conservation requires that:
𝑛𝑛𝑲𝑲𝟏 ℏ𝜔
𝟏
(𝑲𝟏𝟏))++𝑛𝑛𝑲𝑲𝟐𝟐ℏ𝜔
ℏ𝜔(𝑲 ℏ𝜔((𝑲
𝑲𝟐𝟐))+ ℏ𝜔((𝑲
+𝑛𝑛𝑲𝑲𝟑𝟑ℏ𝜔 𝑲𝟑𝟑 ) = 𝑛𝑲𝟏 ′ℏ𝜔(𝑲𝟏𝟏) + ′ℏ𝜔((𝑲
+ 𝑛𝑛𝑲𝑲𝟐𝟐′ℏ𝜔 𝑲𝟐𝟐))++𝑛𝑛𝑲𝑲𝟑𝟑′ℏ𝜔
′ℏ𝜔((𝑲𝑲𝟑𝟑))
𝑛𝑲𝟏 ℏ𝜔(𝑲𝟏 ) + 𝑛𝑲𝟐 ℏ𝜔(𝑲𝟐 ) + 𝑛𝑲𝟑 ℏ𝜔(𝑲𝟑 ) = 𝑛𝑲𝟏 ′ℏ𝜔(𝑲𝟏 ) + 𝑛𝑲𝟐 ′ℏ𝜔(𝑲𝟐 ) + 𝑛𝑲𝟑 ′ℏ𝜔(𝑲
ℏ𝜔((𝑲𝑲𝟏𝟏))−−ℏ𝜔
==𝑛𝑛𝑲𝑲𝟏 𝟏ℏ𝜔 ℏ𝜔((𝑲
𝑲𝟏𝟏))+ ℏ𝜔(𝑲𝟐 ) − ℏ𝜔(𝑲𝟐 ) +
+ 𝑛𝑛𝑲𝑲𝟐𝟐ℏ𝜔 ℏ𝜔((𝑲
+ 𝑛𝑛𝑲𝑲𝟑𝟑ℏ𝜔 𝑲𝟑𝟑))++ℏ𝜔
ℏ𝜔((𝑲𝑲𝟑𝟑))
= 𝑛𝑲𝟏 ℏ𝜔(𝑲𝟏 ) − ℏ𝜔 (𝑲𝟏 ) + 𝑛𝑲𝟐 ℏ𝜔(𝑲𝟐 ) − ℏ𝜔 (𝑲𝟐 ) + 𝑛𝑲𝟑 ℏ𝜔(𝑲𝟑 ) + ℏ𝜔(𝑲𝟑 )
00 = −ℏ𝜔((𝑲
= −ℏ𝜔 𝑲𝟏𝟏)) − ℏ𝜔(𝑲𝟐𝟐 ) + ℏ𝜔 (𝑲
𝑲𝟑𝟑))
0 = −ℏ𝜔
Momentum conservation requires that:
(𝑲𝟏 ) − ℏ𝜔 (𝑲𝟐 ) + ℏ𝜔 (𝑲𝟑 )
omentum conservation requires that:
ntum conservation requires that: 𝑲 +𝑲 = 𝑲 +𝑮
𝑲𝟏𝟏 + 𝑲of𝟐𝟐Materials
𝑮
MSN 501 - Atomic Structure, Mechanical and Thermal Properties = 𝑲𝟑𝟑 + 46
 Thermal
𝑛𝑲𝟏 → 𝑛𝑲Resistivity
− 1 = 𝑛𝑲𝟏 ′
ng thisthis
process, thetheoccupation numbers 𝟏
lowing
ollowing thisprocess,
process, theoccupation
occupation numbersof
numbers ofthese
of these threemodes
these three
three modesisisisas
modes asas follows:
follows:
follows:
Following this process, the occupation numbers of these three modes is as follows:
𝑛𝑛𝑲𝑛𝑲𝑛𝟐𝑲𝑲𝟏→
→→𝑛
→ 𝑛𝑛𝑲
𝑲
𝑲
𝑲𝟐
−11==𝑛𝑲𝑛𝑲𝟏𝑲′′𝟐 ′
−
𝟏 − 1 = 𝑛𝑲
𝟏𝟏 𝟏𝟏 𝟏𝟏

𝑛𝑛𝑲𝑛𝑲𝑛𝟑𝑲𝟐𝑲𝟐𝟐→
→→
→𝑛𝑛𝑛𝑲 −
+
𝑲𝟐𝟐 −
𝑲
𝑲𝟐𝟑
11=
1 ==𝑛𝑛𝑲𝑛𝟐𝟐 ′′ ′
𝑲𝑲 𝟐 𝟑

𝑛𝑛𝑲𝑲𝟑𝟑 →
→ 𝑛𝑲𝑲𝟑𝟑 + 1 = 𝑛𝑲𝟑 ′
servation requires that: 𝑛𝑲𝟑 → 𝑛𝑲𝟑 + 1 = 𝑛𝑲𝟑 ′
ergy
nergyconservation
conservationrequires
requiresthat:
that:
Energy conservation requires that:
ℏ𝜔(𝑲𝟏 ) + 𝑛𝑲𝟐 ℏ𝜔(𝑲𝟐 ) + 𝑛𝑲𝟑 ℏ𝜔(𝑲𝟑 ) = 𝑛𝑲𝟏 ′ℏ𝜔(𝑲𝟏 ) + 𝑛𝑲𝟐 ′ℏ𝜔(𝑲𝟐 ) + 𝑛𝑲𝟑 ′ℏ𝜔
conservation requires that:
𝑛𝑛𝑲𝑲𝟏 ℏ𝜔
𝟏
(𝑲𝟏𝟏))++𝑛𝑛𝑲𝑲𝟐𝟐ℏ𝜔
ℏ𝜔(𝑲 ℏ𝜔((𝑲
𝑲𝟐𝟐))+ ℏ𝜔((𝑲
+𝑛𝑛𝑲𝑲𝟑𝟑ℏ𝜔 𝑲𝟑𝟑 ) = 𝑛𝑲𝟏 ′ℏ𝜔(𝑲𝟏𝟏) + ′ℏ𝜔((𝑲
+ 𝑛𝑛𝑲𝑲𝟐𝟐′ℏ𝜔 𝑲𝟐𝟐))++𝑛𝑛𝑲𝑲𝟑𝟑′ℏ𝜔
′ℏ𝜔((𝑲𝑲𝟑𝟑))
𝑛𝑲𝟏 ℏ𝜔(𝑲𝟏 ) (+ 𝑛𝑲)𝟐 ℏ𝜔(𝑲(𝟐 ) +)𝑛𝑲𝟑 ℏ𝜔(𝑲𝟑 )( = 𝑛)𝑲𝟏 ′ℏ𝜔((𝑲𝟏 ) )+ 𝑛𝑲𝟐 ′ℏ𝜔(𝑲(𝟐 ) +) 𝑛𝑲𝟑 ′ℏ𝜔
( (𝑲
= 𝑛𝑲𝟏 ℏ𝜔 𝑲𝟏ℏ𝜔(−
==𝑛𝑛 ℏ𝜔 (𝑲𝑲ℏ𝜔 𝑲𝟏((𝑲𝑲+))+𝑛
))−−ℏ𝜔
ℏ𝜔 +𝑲 ℏ𝜔
ℏ𝜔(𝑲
𝑛𝑛𝟐 ℏ𝜔
− ℏ𝜔 𝑲 + 𝑛𝑛+ 𝑛
𝑲𝟐) − ℏ𝜔(𝑲 ) 𝟐+ ℏ𝜔
ℏ𝜔
ℏ𝜔
𝑲(𝟑(𝑲 𝑲ℏ𝜔 +
𝟑 ((𝑲
𝑲 ))++ℏ𝜔
ℏ𝜔 𝑲𝟑
𝑲 ))
𝑲𝑲
𝟏𝟏 𝟏𝟏 𝟏𝟏 𝑲𝟐𝟐
𝑲 𝟐 𝟐 𝑲
𝑲𝟑𝟑 𝟑𝟑 𝟑𝟑
= 𝑛𝑲𝟏 ℏ𝜔(𝑲𝟏 ) − ℏ𝜔 (𝑲𝟏 ) + 𝑛𝑲𝟐 ℏ𝜔(𝑲𝟐 ) − ℏ𝜔 (𝑲𝟐 ) + 𝑛𝑲𝟑 ℏ𝜔(𝑲𝟑 ) + ℏ𝜔(𝑲𝟑 )
0 =00−ℏ𝜔 (𝑲
= −ℏ𝜔
= −ℏ𝜔 𝟏𝟏𝟏)))−
((𝑲
𝑲 ℏ𝜔((𝑲𝑲
− ℏ𝜔 𝟐)
𝟐 𝟐)
++ 𝑲𝟑𝟑())𝑲𝟑 )
ℏ𝜔ℏ𝜔
(𝑲
0 = −ℏ𝜔
Momentum conservation requires that:
(𝑲𝟏 ) − ℏ𝜔 (𝑲𝟐 ) + ℏ𝜔 (𝑲𝟑 )
omentum
m conservation
conservation requires
requires that:
that:
ntum conservation requires that: 𝑲 +𝑲 = 𝑲 +𝑮
𝑲𝟏𝟏 + 𝑲of𝟐𝟐Materials
MSN 501 - Atomic Structure, Mechanical and Thermal Properties 𝑮 = 𝑲𝟑𝟑 + 47
 0 =Thermal
−ℏ𝜔(𝑲𝟏 ) − Resistivity
ℏ𝜔(𝑲𝟐 ) + ℏ𝜔 (𝑲𝟑 )
mentum conservation requires that:
Momentum conservation requires that:

𝑲𝟏 + 𝑲𝟐 = 𝑲𝟑 + 𝑮
Where G is a reciprocal lattice vector.
ere G is a reciprocal lattice vector.

make both of these expressions completely generic, we can write:

 Energy conservation:∑𝑝 ℏ𝜔𝑝 (𝑲)𝑛𝑲,𝑝,𝑖𝑛𝑖𝑡𝑖𝑎𝑙 − ∑𝑝 ℏ𝜔𝑝 𝑛𝑲,𝑝,𝑓𝑖𝑛𝑎𝑙 = 0

 Momentum conservation: ∑𝑝 𝑲𝑛𝑲,𝑝,𝑖𝑛𝑖𝑡𝑖𝑎𝑙 = ∑𝑝 𝑲𝑛𝑲,𝑝,𝑓𝑖𝑛𝑎𝑙 + 𝑮

ere the sum over p denotes a sum over all phonon branches (“polarization branches”).

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 48

 0 =Thermal
−ℏ𝜔(𝑲𝟏 ) − Resistivity
ℏ𝜔(𝑲𝟐 ) + ℏ𝜔 (𝑲𝟑 )
mentum conservation requires that:
Momentum conservation requires that:

𝑲𝟏 + 𝑲𝟐 = 𝑲𝟑 + 𝑮
Where G is a reciprocal lattice vector.
ere G is a reciprocal lattice vector.

make both of these expressions completely generic, we can write:

To make both of these expressions completely generic, we can write:
 Energy conservation:∑𝑝 ℏ𝜔𝑝 (𝑲)𝑛𝑲,𝑝,𝑖𝑛𝑖𝑡𝑖𝑎𝑙 − ∑𝑝 ℏ𝜔𝑝 𝑛𝑲,𝑝,𝑓𝑖𝑛𝑎𝑙 = 0
Energy conservation: ∑ ħ ( ) ,, −∑ ħ , , =0
 Momentum conservation: ∑𝑝 𝑲𝑛𝑲,𝑝,𝑖𝑛𝑖𝑡𝑖𝑎𝑙 = ∑𝑝 𝑲𝑛𝑲,𝑝,𝑓𝑖𝑛𝑎𝑙 + 𝑮
Momentum conservation: ∑ , , =∑ , , +
ere the sum over p denotes a sum over all phonon branches (“polarization branches”).

Where the sum over p denotes a sum over all phonon branches (“polarization branches”).

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 49

𝑲
𝜔
𝑛
𝑲
𝑛
𝑲
𝑛
𝜔
𝑛
𝑮
 Thermal Resistivity
The factor of G is needed because phonon momentum is only defined modulo a reciprocal lattice vector, so
if the sum of two phonon momenta is located outside the first Brillouin zone, it is mapped onto an
equivalent momentum in the first Brillouin zone by translating by a reciprocal lattice vector.

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 50

 Thermal Resistivity
nTheU processes
factor ofthere is a large
G is needed net change
because phononin phonon momentum
momentum is onlyindefined
each collision
modulo a reciprocal lattice vector, so
al ifnet
thephonon
sum offlux
twowill rapidlymomenta
phonon decay as we move tooutside
is located the right.
theThe
firstends may act
Brillouin as it is mapped onto an
zone,
ks. Net energy
equivalent transport under
momentum in thea temperature gradient
first Brillouin zone byoccurs as in (b).by a reciprocal lattice vector.
translating

If 0 is needed to make the equality above hold, the scattering is called an Umklapp process.

Ky Ky

K2
K1 K1
max K=?
Kx Kx
K2 K1 ! K2
K3

K3 G

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 51

𝑮
≠
 Thermal Resistivity
nTheU processes
factor ofthere is a large
G is needed net change
because phononin phonon momentum
momentum is onlyindefined
each collision
modulo a reciprocal lattice vector, so
al ifnet
thephonon
sum offlux
twowill rapidlymomenta
phonon decay as we move tooutside
is located the right.
theThe
firstends may act
Brillouin as it is mapped onto an
zone,
ks. Net energy
equivalent transport under
momentum in thea temperature gradient
first Brillouin zone byoccurs as in (b).by a reciprocal lattice vector.
translating

If 0 is needed to make the equality above hold, the scattering is called an Umklapp process.

If G=0, it is called a normal scattering process. It should be noted that only Umklapp processes
Ky
contribute to a material’s thermal resistivity. Ky

K2
K1 K1
max K=?
Kx Kx
K2 K1 ! K2
K3

K3 G

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 52

𝑮
≠
 Thermal Resistivity
It should be noted that only Umklapp processes contribute to a material’s thermal resistivity.

To show that only umklapp processes contribute to thermal resistivity, we consider the sum over all
phonon scattering processes

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 53

 mensions, and for a square lattice, the intersection points between the first Brillouin zone an
𝐾𝑦 axes are at ±
𝜋 Thermal Resistivity
It should be 𝑎
noted that only Umklapp processes contribute to a material’s thermal resistivity.
thatToonly
showumklapp processes
that only umklapp contribute
processes to thermal
contribute resistivity,
to thermal weconsider
resistivity, we considerthethe
sumsum
over over
all all
phonon processes
scattering scattering processes

𝑲𝑡𝑜𝑡 = ∑ ∑ 𝑲𝑛𝑠 (𝑲)

𝑝 𝐾∈1𝑠𝑡 𝐵𝑟𝑖𝑙𝑙𝑜𝑢𝑖𝑛 𝑧𝑜𝑛𝑒

r sum sums over all phonon branches (e.g. all acoustic and optical
counts all
dispersions) and the inne
over all phonon branches
nts all momenta in the first Brillouin zone. The terms being momenta
(e.g. all acoustic and
added together
in the include the
first Brillouin
optical
um (vector) dispersions)
of each phonon mode being considered (K) multiplied by how many of them th
zone
𝑲)).

ormal scattering processes are included 𝑲𝑡𝑜𝑡 = 0 because for every K, there is a –K in the fir
1
zone and 𝑛𝑝 (𝑲) = 𝑛𝑠 (−𝑲) = ℏ𝜔𝑝 /𝑘𝐵 𝑇
𝑒 −1
MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 54
 mensions, and for a square lattice, the intersection points between the first Brillouin zone an
𝐾𝑦 axes are at ±
𝜋 Thermal Resistivity
It should be 𝑎
noted that only Umklapp processes contribute to a material’s thermal resistivity.
thatToonly
showumklapp processes
that only umklapp contribute
processes to thermal
contribute resistivity,
to thermal weconsider
resistivity, we considerthethe
sumsum
over over
all all
phonon processes
scattering scattering processes

𝑲𝑡𝑜𝑡 = ∑ ∑ 𝑲𝑛𝑠 (𝑲)

𝑝 𝐾∈1𝑠𝑡 𝐵𝑟𝑖𝑙𝑙𝑜𝑢𝑖𝑛 𝑧𝑜𝑛𝑒

r sum sums over all phonon branches (e.g. all acoustic and optical
counts all
dispersions) and the inne
over all phonon branches
nts all momenta in the first Brillouin zone. The terms being momenta
(e.g. all acoustic and
added together
in the include the
first Brillouin
optical
um (vector) dispersions)
of each phonon mode being considered (K) multiplied by how many of them th
zone
𝑲)).
The terms
ormal being added
scattering together
processes include
are included 𝑲𝑡𝑜𝑡 = 0(vector)
the momentum becauseof each
for phonon mode
every K, being
there is aconsidered
–K in the fir
(K) multiplied by how many of them there1are ( ( )).
zone and 𝑛𝑝 (𝑲) = 𝑛𝑠 (−𝑲) = ℏ𝜔𝑝 /𝑘𝐵 𝑇
𝑒 −1
MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 55
𝑛
𝑲
 Thermal Resistivity
If only normal scattering processes are included = 0 because for every K, there is a –K in the first
Brillouin zone with the same occupation ns(K)

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 56

𝑲
 Thermal Resistivity
If only normal scattering processes are included = 0 because for every K, there is a –K in the first
Brillouin zone with the same occupation ns(K)

In the context of thermal conductivity, what this means is that if you heat up one side of a rod, you will
change the occupation numbers of phonon modes, but if these phonons can only scatter via normal
processes, they will propagate down the length of the rod with no thermal resistance.

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 57

𝑲
 𝑝

Thermal
f only normal scattering processes are Resistivity
included 𝑲𝑡𝑜𝑡 = 0 because for every K, there is a –K in the first
1
If onlyzone
Brillouin normal
and 𝑛𝑝 (𝑲) =processes
scattering 𝑛𝑠 (−𝑲)are
= included
ℏ𝜔𝑝 /𝑘𝐵 𝑇
= 0 because for every K, there is a –K in the first
Brillouin zone with the same occupation𝑒 ns(K) −1
n the context of thermal conductivity, what this means is that if you heat up one side of a rod, you will
In thethe
hange context of thermal
occupation conductivity,
numbers what modes,
of phonon this meansbutisifthat if you
these heat upcan
phonons oneonly
sidescatter
of a rod,
viayou will
normal
change the occupation numbers of phonon modes, but if these phonons can only scatter via normal
processes,
processes,they
theywill
willpropagate
propagate down thelength
down the lengthofofthe
therod
rod with
with nono thermal
thermal resistance.
resistance.

Nowwhere
Now consider a situation consider a situation
Umklapp where Umklapp
processes processes are allowed.
are allowed.

𝑲𝑡𝑜𝑡 = 𝑲𝒊𝒏𝒊𝒕𝒊𝒂𝒍 − 𝑲𝒇𝒊𝒏𝒂𝒍 = ∑ 𝑲𝑛𝑲,𝑝,𝑖𝑛𝑖𝑡𝑖𝑎𝑙 − ( ∑ 𝑲𝑛𝑲,𝑝,𝑓𝑖𝑛𝑎𝑙 + 𝑮)

𝑝 𝑝

𝑮 = 𝜈1=𝒃𝟏1 ++𝜈22𝒃𝟐++3 𝜈3is𝒃any reciprocal

𝟑 is any lattice
reciprocal vector,
lattice so when
vector, so we
whenhave
wemany
haveterms
manycontributing to G, we to
terms contributing
accommodate
G, we accommodate thisthis
by making the integers
by making 1, 2, 𝜈1
the integers , 𝜈2large
3 as , 𝜈3 as
as large
needed
as needed

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 58

𝑮
𝜈
𝒃
𝜈
𝒃
𝜈
𝒃
𝜈
𝜈
𝜈
𝑲
 𝑝

Thermal
f only normal scattering processes are Resistivity
included 𝑲𝑡𝑜𝑡 = 0 because for every K, there is a –K in the first
1
If onlyzone
Brillouin normal
and 𝑛𝑝 (𝑲) =processes
scattering 𝑛𝑠 (−𝑲)are
= included
ℏ𝜔𝑝 /𝑘𝐵 𝑇
= 0 because for every K, there is a –K in the first
Brillouin zone with the same occupation𝑒 ns(K) −1
n the context of thermal conductivity, what this means is that if you heat up one side of a rod, you will
In thethe
hange context of thermal
occupation conductivity,
numbers what modes,
of phonon this meansbutisifthat if you
these heat upcan
phonons oneonly
sidescatter
of a rod,
viayou will
normal
change the occupation numbers of phonon modes, but if these phonons can only scatter via normal
processes,
processes,they
theywill
willpropagate
propagate down thelength
down the lengthofofthe
therod
rod with
with nono thermal
thermal resistance.
resistance.

Nowwhere
Now consider a situation consider a situation
Umklapp where Umklapp
processes processes are allowed.
are allowed.

𝑲𝑡𝑜𝑡 = 𝑲𝒊𝒏𝒊𝒕𝒊𝒂𝒍 − 𝑲𝒇𝒊𝒏𝒂𝒍 = ∑ 𝑲𝑛𝑲,𝑝,𝑖𝑛𝑖𝑡𝑖𝑎𝑙 − ( ∑ 𝑲𝑛𝑲,𝑝,𝑓𝑖𝑛𝑎𝑙 + 𝑮)

𝑝 𝑝

𝑮 = 𝜈1=𝒃𝟏1 ++𝜈22𝒃𝟐++3 𝜈3is𝒃any reciprocal

𝟑 is any lattice
reciprocal vector,
lattice so when
vector, so we
whenhave
wemany
haveterms
manycontributing to G, we to
terms contributing
accommodate
G, we accommodate thisthis
by making the integers
by making 1, 2, 𝜈1
the integers , 𝜈2large
3 as , 𝜈3 as
as large
needed
as needed

When we consider Umklapp processes, we do not have the same cancellation as we did without and
= contributes a net drift velocity between collisions which is not zero.
MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 59
𝑮
𝑲
𝜈
𝒃
𝑮
𝜈
𝒃
𝜈
𝒃
𝜈
𝜈
𝜈
𝑲
sider two temperature regimes:
Thermal Resistivity
1) 1) Temperature much
At high temperatures (much larger than
higher than the
Debye Debye temperature (𝑻 ≫ 𝜽) :
temperature)
1 𝑘𝐵 𝑇
at high temperature 𝑛𝑝 (𝑲) = ℏ𝜔/𝑘𝐵 𝑇 ≈ (from taylor
(Taylor expanding
expansion)
𝑒 −1 ℏ𝜔𝑝 (𝑲)
order) At high temperatures C is constant (at low temperatures T3)
1 2〉
Using 𝐾 = 𝐶𝑣ℓ = 〈 𝑣 𝜏,
High temperature not only ensures that CCisis independent
constant as a functionof
of temperature
temperature, but itin
alsothis high
ensures thatte
3
Umklapp processes have a reasonable probability of happening because it allows for sufficient thermal
𝜏 is not.
population of higher momentum (which are also higher frequency) phonons.
1
The collision frequency (𝜏 ) should be proportional to the number of phon
1 1
phonon can collide with and thus 𝐾 ∝ 𝜏 ∝ 𝑛 ∝ 𝑇 . In reality, 𝐾 ∝ 1/𝑇 𝑥 w
and 2.
High temperature not only ensures that C is constant as a function of tem
ensures that Umklapp processes have a reasonable probability of happen
MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 60
sider two temperature regimes:
Thermal Resistivity
1) 1) Temperature much
At high temperatures (much larger than
higher than the
Debye Debye temperature (𝑻 ≫ 𝜽) :
temperature)
1 𝑘𝐵 𝑇
at high temperature 𝑛𝑝 (𝑲) = ℏ𝜔/𝑘𝐵 𝑇 ≈ (from taylor
(Taylor expanding
expansion)
𝑒 −1 ℏ𝜔𝑝 (𝑲)
order) At high temperatures C is constant (at low temperatures T3)
1 2〉
Using 𝐾 = 𝐶𝑣ℓ = 〈 𝑣 𝜏,
High temperature not only ensures that CCisis independent
constant as a functionof
of temperature
temperature, but itin
alsothis high
ensures thatte
3
Umklapp processes have a reasonable probability of happening because it allows for sufficient thermal
𝜏 is not.
population of higher momentum (which are also higher frequency) phonons.
1
The collision frequency (𝜏 ) should be proportional to the number of phon
1
2) At low temperatures: only low energy (and low momentum) acoustic1 modes will have substantial 𝑥
phonon
thermal can collide
population, so Umklapp with andwill
processes thus 𝐾 ∝ 𝜏and
be unlikely ∝ phonon-phonon
𝑛
∝ 𝑇
. In reality,
scattering 𝐾 ∝
will 1/𝑇
not w
contribute to thermal resistivity
and 2.
High temperature not only ensures that C is constant as a function of tem
ensures that Umklapp processes have a reasonable probability of happen
MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 61
 Thermal Resistivity - Imperfections
Geometric effects can decrease the thermal conductivity if they increase a length scale shorter than l. For
instance, if the material is in a reduced geometry (e.g. a thin film or a nanowire) with minimum dimension
length D
≈
This also holds if the material consists of small crystallites of dimension D—phonons can scatter at the
boundary between crystallites of different orientations and reduce the thermal conductivity

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 62

𝐾
𝐶
𝑣
𝐷
 dimensionThermal
length D Resistivity - Imperfections
Geometric effects can decrease the thermal conductivity if they increase𝐾 ≈a 𝐶𝑣𝐷
length scale shorter than l. For
instance, if the material is in a reduced geometry (e.g. a thin film or a nanowire) with minimum dimension
length D This also holds if the material consists of small crystallites of dimension D—phonon
≈
boundary between crystallites of different orientations and reduce the thermal con
This also holds if the material consists of small crystallites of dimension D—phonons can scatter at the
boundary between
Example: crystallites of different orientations and reduce the thermal conductivity

Example:
Thermal conductivity
Thermal conductivity of insulating
of insulating solids solids at room
at room temperature temperature
(insulating so we don’t(insulating
have to deal so we do
with
electrons’ contribution to thermal conductivity)
electrons’ contribution to thermal conductivity)

Material K (W/m K) v (m/s) Debye T

Diamond 2200 12,000 1320K
Silicon 130 5000 645K
KCl 6.5 2273 240K
SiO2 130 3962 470K
MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 63
𝐾
𝐶
𝑣
𝐷
 dimension length D
Thermal Resistivity - Imperfections
𝐾 ≈ 𝐶𝑣𝐷
• Thisillustrates
This table also holdsthe if the material
general consists
scaling between of small
thermal crystallites
conductivity, ofvelocity,
sound dimension D—phonon
and Debye
temperature,
boundarywhichbetween
makes sense since bothof
crystallites K and depend
different on v
orientations and reduce the thermal con
• Diamond is characterized by a high thermal conductivity, and this is how one tests to see if a material
Example:
is a diamond. Physically, the large sound velocity, which produces high thermal conductivity and
Debye temperature, originates from very light atoms (carbon) connected by very stiff covalent bonds
Thermal conductivity of insulating solids at room temperature (insulating so we do
electrons’ contribution to thermal conductivity)

Material K (W/m K) v (m/s) Debye T

Diamond 2200 12,000 1320K
Silicon 130 5000 645K
KCl 6.5 2273 240K
SiO2 130 3962 470K

MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 64

𝜃
 important when we discuss thermal conductivity. In discussing thermal conductivity, we often think of a
Comparison
‘gas’ of phonons, and the table belowwith the
makes this ideal
more gas
concrete viaof molecules
a comparison to a classical gas.

Classical gas (molecules) Phonon gas

Number fixed Population depends on
frequency of mode and
temperature:
1
〈𝑛〉 = ℏ𝜔/𝑘 𝑇
𝑒 𝐵 −1
(What are the limits of this
expression for very small T and
for very large T)?
Small T: 𝑒 ℏ𝜔/𝑘𝐵 𝑇 ≫ 1 →
〈𝑛〉~𝑒−ℏ𝜔/𝑘𝐵 𝑇
ℏ𝜔
ℏ𝜔
Large T: 𝑒 𝑘𝐵 𝑇 ~1 + 𝑘
𝐵𝑇
〈𝑛〉~𝑘𝐵 𝑇/ℏ𝜔
Relationship between energy 𝐾𝐸 = 𝑝2 /2𝑚 Depends on which branch you
and momentum For each particle are considering (see chapter 4),
but for acoustic phonons near
MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials 65
 important when we discuss thermal conductivity. In discussing thermal〈𝑛 〉~𝑒
conductivity, we often think of a
ℏ𝜔
Comparison
‘gas’ of phonons, and the table belowwith the
makes this ideal
more concretegas of
Largemolecules
via a comparison
T: 𝑒 𝑘𝐵to𝑇 ~1
ℏ𝜔 gas.
a classical
+ 𝑘𝐵 𝑇
Classical gas (molecules) Phonon gas 〈𝑛〉~𝑘𝐵 𝑇/ℏ𝜔
Relationship between energy
Number fixed 𝐾𝐸 = 𝑝2 /2𝑚 Depends ondepends
Population which branch
on you
and momentum For each particle are considering
frequency (seeand
of mode chapter 4),
but for acoustic phonons near
temperature:
k=0, 𝜔~𝑣𝑠 𝑘 where1k is crystal
〈𝑛〉 = ℏ𝜔/𝑘 𝑇
momentum and 𝑒 𝑣𝑠𝐵is speed
− 1 of
(What
sound are the limits of this
Relationship between For an N-particle gas 𝑈 = ∑𝐾 ∑𝑝for
expression < 𝑛very
𝐾,𝑝 > ℏ𝜔𝐾,𝑝
small T and
=
temperature and total internal 𝑈 = 3𝑁𝑘𝐵 𝑇 for very large T)? ℏ𝜔𝐾,𝑝
energy Small ∑𝑒 ℏ𝜔/𝑘
= T: ∑ 𝐵ℏ𝜔 𝑇
≫ 1 𝐵→𝑇 − 1
𝑒 𝐾,𝑝/𝑘
𝐾 𝑝𝐵 𝑇
〈𝑛〉~𝑒−ℏ𝜔/𝑘
Collisions Molecules collide with each Phonons collide
ℏ𝜔 with each
ℏ𝜔
other and with walls of vessel Large 𝑒 𝑘𝐵surface
other,T:with 𝑇 ~1 +
of crystal,
𝑘𝐵 𝑇
and with impurities
〈𝑛〉~𝑘𝐵 𝑇/ℏ𝜔
Energy conserved
Relationship in collisions?
between energy Yes 𝐾𝐸 = 𝑝2 /2𝑚 Yes
Depends on which branch you
Momentum
and momentum conserved in Yes,each
For except for at walls
particle Yes,considering
are modulo a reciprocal
(see chapter lattice
4),
collisions? vector
but forGacoustic phonons near
Number of particles conserved Yes No 𝜔~𝑣𝑠 𝑘 where k is crystal
k=0,
in collision? momentum and 𝑣𝑠 is speed of
MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials sound 66
 important when we discuss thermal conductivity. In discussing thermal〈𝑛 〉~𝑒
conductivity, we often think of a
ℏ𝜔
Comparison
‘gas’ of phonons, and the table belowwith the
makes this ideal
more concretegas of
Largemolecules
via a comparison
T: 𝑒 𝑘𝐵to𝑇 ~1
ℏ𝜔 gas.
a classical
+ 𝑘𝐵 𝑇
Classical gas (molecules) Phonon gas 〈𝑛〉~𝑘𝐵 𝑇/ℏ𝜔
Relationship between energy
Number fixed 𝐾𝐸 = 𝑝2 /2𝑚 Depends ondepends
Population which branch
on you
and momentum For each particle are considering
frequency (seeand
of mode chapter 4),
but for acoustic phonons near
temperature:
k=0, 𝜔~𝑣𝑠 𝑘 where1k is crystal
〈𝑛〉 = ℏ𝜔/𝑘 𝑇
momentum and 𝑒 𝑣𝑠𝐵is speed
− 1 of
(What
sound are the limits of this
Relationship between For an N-particle gas 𝑈 = ∑𝐾 ∑𝑝for
expression < 𝑛very
𝐾,𝑝 > ℏ𝜔𝐾,𝑝
small T and
=
temperature and total internal 𝑈 = 3𝑁𝑘𝐵 𝑇 for very large T)? ℏ𝜔𝐾,𝑝
energy Small ∑𝑒 ℏ𝜔/𝑘
= T: ∑ 𝐵ℏ𝜔 𝑇
≫ 1 𝐵→𝑇 − 1
𝑒 𝐾,𝑝/𝑘
𝐾 𝑝𝐵 𝑇
〈𝑛〉~𝑒−ℏ𝜔/𝑘
Collisions Molecules collide with each Phonons collide
ℏ𝜔 with each
ℏ𝜔
other and with walls of vessel Large 𝑒 𝑘𝐵surface
other,T:with 𝑇 ~1 +
of crystal,
𝑘𝐵 𝑇
and with impurities
〈𝑛〉~𝑘𝐵 𝑇/ℏ𝜔
Energy conserved
Relationship in collisions?
between energy Yes 𝐾𝐸 = 𝑝2 /2𝑚 Yes
Depends on which branch you
Momentum
and momentum conserved in Yes,each
For except for at walls
particle Yes,considering
are modulo a reciprocal
(see chapter lattice
4),
collisions? vector
but forGacoustic phonons near
Number of particles conserved Yes No 𝜔~𝑣𝑠 𝑘 where k is crystal
k=0,
in collision? momentum and 𝑣𝑠 is speed of
MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials sound 67
 important when we discuss thermal conductivity. In discussing thermal〈𝑛 〉~𝑒
conductivity, we often think of a
ℏ𝜔
Comparison
‘gas’ of phonons, and the table belowwith the
makes this ideal
more concretegas of
Largemolecules
via a comparison
T: 𝑒 𝑘𝐵to𝑇 ~1
ℏ𝜔 gas.
a classical
+ 𝑘𝐵 𝑇
Classical gas (molecules) Phonon gas 〈𝑛〉~𝑘𝐵 𝑇/ℏ𝜔
Relationship between energy
Number fixed 𝐾𝐸 = 𝑝2 /2𝑚 Depends ondepends
Population which branch
on you
and momentum For each particle are considering
frequency (seeand
of mode chapter 4),
but for acoustic phonons near
temperature:
k=0, 𝜔~𝑣𝑠 𝑘 where1k is crystal
〈𝑛〉 = ℏ𝜔/𝑘 𝑇
momentum and 𝑒 𝑣𝑠𝐵is speed
− 1 of
(What
sound are the limits of this
Relationship between For an N-particle gas 𝑈 = ∑𝐾 ∑𝑝for
expression < 𝑛very
𝐾,𝑝 > ℏ𝜔𝐾,𝑝
small T and
=
temperature and total internal 𝑈 = 3𝑁𝑘𝐵 𝑇 for very large T)? ℏ𝜔𝐾,𝑝
energy Small ∑𝑒 ℏ𝜔/𝑘
= T: ∑ 𝐵ℏ𝜔 𝑇
≫ 1 𝐵→𝑇 − 1
𝑒 𝐾,𝑝/𝑘
𝐾 𝑝𝐵 𝑇
〈𝑛〉~𝑒−ℏ𝜔/𝑘
Collisions Molecules collide with each Phonons collide
ℏ𝜔 with each
ℏ𝜔
other and with walls of vessel Large 𝑒 𝑘𝐵surface
other,T:with 𝑇 ~1 +
of crystal,
𝑘𝐵 𝑇
and with impurities
〈𝑛〉~𝑘𝐵 𝑇/ℏ𝜔
Energy conserved
Relationship in collisions?
between energy Yes 𝐾𝐸 = 𝑝2 /2𝑚 Yes
Depends on which branch you
Momentum
and momentum conserved in Yes,each
For except for at walls
particle Yes,considering
are modulo a reciprocal
(see chapter lattice
4),
collisions? vector
but forGacoustic phonons near
Number of particles conserved Yes No 𝜔~𝑣𝑠 𝑘 where k is crystal
k=0,
in collision? momentum and 𝑣𝑠 is speed of
MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials sound 68
 important when we discuss thermal conductivity. In discussing thermal〈𝑛 〉~𝑒
conductivity, we often think of a
ℏ𝜔
Comparison
‘gas’ of phonons, and the table belowwith the
makes this ideal
more concretegas of
Largemolecules
via a comparison
T: 𝑒 𝑘𝐵to𝑇 ~1
ℏ𝜔 gas.
a classical
+ 𝑘𝐵 𝑇
Classical gas (molecules) Phonon gas 〈𝑛〉~𝑘𝐵 𝑇/ℏ𝜔
Relationship between energy
Number fixed 𝐾𝐸 = 𝑝2 /2𝑚 Depends ondepends
Population which branch
on you
and momentum For each particle are considering
frequency (seeand
of mode chapter 4),
but for acoustic phonons near
temperature:
k=0, 𝜔~𝑣𝑠 𝑘 where1k is crystal
〈𝑛〉 = ℏ𝜔/𝑘 𝑇
momentum and 𝑒 𝑣𝑠𝐵is speed
− 1 of
(What
sound are the limits of this
Relationship between For an N-particle gas 𝑈 = ∑𝐾 ∑𝑝for
expression < 𝑛very
𝐾,𝑝 > ℏ𝜔𝐾,𝑝
small T and
=
temperature and total internal 𝑈 = 3𝑁𝑘𝐵 𝑇 for very large T)? ℏ𝜔𝐾,𝑝
energy Small ∑𝑒 ℏ𝜔/𝑘
= T: ∑ 𝐵ℏ𝜔 𝑇
≫ 1 𝐵→𝑇 − 1
𝑒 𝐾,𝑝/𝑘
𝐾 𝑝𝐵 𝑇
〈𝑛〉~𝑒−ℏ𝜔/𝑘
Collisions Molecules collide with each Phonons collide
ℏ𝜔 with each
ℏ𝜔
other and with walls of vessel Large 𝑒 𝑘𝐵surface
other,T:with 𝑇 ~1 +
of crystal,
𝑘𝐵 𝑇
and with impurities
〈𝑛〉~𝑘𝐵 𝑇/ℏ𝜔
Energy conserved
Relationship in collisions?
between energy Yes 𝐾𝐸 = 𝑝2 /2𝑚 Yes
Depends on which branch you
Momentum
and momentum conserved in Yes,each
For except for at walls
particle Yes,considering
are modulo a reciprocal
(see chapter lattice
4),
collisions? vector
but forGacoustic phonons near
Number of particles conserved Yes No 𝜔~𝑣𝑠 𝑘 where k is crystal
k=0,
in collision? momentum and 𝑣𝑠 is speed of
MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials sound 69
 important when we discuss thermal conductivity. In discussing thermal〈𝑛 〉~𝑒
conductivity, we often think of a
ℏ𝜔
Comparison
‘gas’ of phonons, and the table belowwith the
makes this ideal
more concretegas of
Largemolecules
via a comparison
T: 𝑒 𝑘𝐵to𝑇 ~1
ℏ𝜔 gas.
a classical
+ 𝑘𝐵 𝑇
Classical gas (molecules) Phonon gas 〈𝑛〉~𝑘𝐵 𝑇/ℏ𝜔
Relationship between energy
Number fixed 𝐾𝐸 = 𝑝2 /2𝑚 Depends ondepends
Population which branch
on you
and momentum For each particle are considering
frequency (seeand
of mode chapter 4),
but for acoustic phonons near
temperature:
k=0, 𝜔~𝑣𝑠 𝑘 where1k is crystal
〈𝑛〉 = ℏ𝜔/𝑘 𝑇
momentum and 𝑒 𝑣𝑠𝐵is speed
− 1 of
(What
sound are the limits of this
Relationship between For an N-particle gas 𝑈 = ∑𝐾 ∑𝑝for
expression < 𝑛very
𝐾,𝑝 > ℏ𝜔𝐾,𝑝
small T and
=
temperature and total internal 𝑈 = 3𝑁𝑘𝐵 𝑇 for very large T)? ℏ𝜔𝐾,𝑝
energy Small ∑𝑒 ℏ𝜔/𝑘
= T: ∑ 𝐵ℏ𝜔 𝑇
≫ 1 𝐵→𝑇 − 1
𝑒 𝐾,𝑝/𝑘
𝐾 𝑝𝐵 𝑇
〈𝑛〉~𝑒−ℏ𝜔/𝑘
Collisions Molecules collide with each Phonons collide
ℏ𝜔 with each
ℏ𝜔
other and with walls of vessel Large 𝑒 𝑘𝐵surface
other,T:with 𝑇 ~1 +
of crystal,
𝑘𝐵 𝑇
and with impurities
〈𝑛〉~𝑘𝐵 𝑇/ℏ𝜔
Energy conserved
Relationship in collisions?
between energy Yes 𝐾𝐸 = 𝑝2 /2𝑚 Yes
Depends on which branch you
Momentum
and momentum conserved in Yes,each
For except for at walls
particle Yes,considering
are modulo a reciprocal
(see chapter lattice
4),
collisions? vector
but forGacoustic phonons near
Number of particles conserved Yes No 𝜔~𝑣𝑠 𝑘 where k is crystal
k=0,
in collision? momentum and 𝑣𝑠 is speed of
MSN 501 - Atomic Structure, Mechanical and Thermal Properties of Materials sound 70