Elementary Lattice Dynamics

Syllabus: Lattice Vibrations and Phonons: Linear Monoatomic and Diatomic Chains. Acoustical and
Optical Phonons. Qualitative Description of the Phonon Spectrum in Solids. Dulong and Petit's Law,
Einstein and Debye theories of specific heat of solids. T3 law.
1. Introduction:
Every atom in a crystal is surrounded by other atoms by the influence of attractive force. If an atom is
displaced from its equilibrium position, the force due to other atoms comes into act to restore its equilibrium
position. The lattice structure can be modeled as the atoms are connected with each other by spring in three
dimensions. A schematic structure in two and three dimensions are shown here.

If the position of an atom is displaced by some means, it will cause displacement of neighbourhood atoms
i.e., the vibration of an atom in the crystal is coupled vibration. The vibration in crystal may originate from
thermal motion – the atoms vibrate in normal modes that yields thermal properties of solids like the specific
heat, thermal conductivity. Another type of vibration in crystal is forced vibration which is caused due to
the propagation of acoustic wave or the electromagnetic wave through the crystal that yields acoustical and
some optical properties of solids. We will study such coupled vibration in one dimensional chain of lattice;
we will discuss a monoatomic chain of lattice and a diatomic chain of lattice.
Here are few assumptions necessary for the study:
1. Motion of one atom is experienced by the nearest neighbouring atom only
2. The motion follows Hooke’s law – that is the motion is simple harmonic 𝐹 = −𝛽𝑢
3. Only longitudinal motion is considered.

2. Linear Monoatomic Lattice Chain:

© DR. P. MANDAL 1
Equation of Motion:
𝐹𝑛 = 𝛽(𝑢𝑛+1 − 𝑢𝑛 ) − 𝛽(𝑢𝑛 − 𝑢𝑛−1 )
⇒ 𝑚𝑢̈ 𝑛 = 𝛽(𝑢𝑛+1 + 𝑢𝑛−1 − 2𝜇𝑛 ) … (1)
Here one motion is resistive and the other motion is supportive. So −𝑣𝑒 sign come between two relative
displacements. 𝐹𝑛 = −𝑣𝑒 will make it a restoring type force.
Trial solution:
𝑢𝑛 = 𝑢0 𝑒 𝑖(𝜔𝑡−𝑘𝑛𝑎)
̅̅̅̅̅̅𝑎) ̅̅̅̅̅̅𝑎)
𝑢𝑛−1 = 𝑢0 𝑒 𝑖(𝜔𝑡−𝑘𝑛−1 = 𝑢𝑛 𝑒 𝑖𝑘𝑎 , 𝑢𝑛+1 = 𝑢0 𝑒 𝑖(𝜔𝑡−𝑘𝑛+1 = 𝑢𝑛 𝑒 −𝑖𝑘𝑎
𝑢̈ 𝑛 = −𝜔2 𝑢0 𝑒 𝑖(𝜔𝑡−𝑘𝑛𝑎)
Substituting the trial solution in equation (1)
𝑖𝑘𝑎 𝑖𝑘𝑎 2
−𝑚𝜔2 = 𝛽(𝑒 𝑖𝑘𝑎 + 𝑒 −𝑖𝑘𝑎 − 2) = 𝛽 (𝑒 2 − 𝑒−2 )

⇒ −𝑚𝜔2 = −4𝛽𝑠𝑖𝑛2 𝑘𝑎/2

4𝛽 2 𝑘𝑎 4𝛽 𝑘𝑎 𝑘𝑎
or, 𝜔2 = sin ⇒𝜔 =±√ sin = ±𝜔0 sin … (2)
𝑚 2 𝑚 2 2

The longitudinal stiffness factor and linear mass density are respectively defined as 𝐶 = 𝛽𝑎 and 𝜌 = 𝑚/𝑎.
Thus equation (2) can be rewritten as (substituting 𝛽/𝑚 = 𝑐/𝜌𝑎2 )

2 𝐶 𝑘𝑎 2 𝑘𝑎
𝜔 = ± √ sin = ± 𝑣𝑠 sin
𝑎 𝜌 2 𝑎 2

where 𝑣𝑠 = √𝐶/𝜌 is constant and has the dimension of velocity.

Phase velocity:
𝜔 𝜔0 𝑘𝑎
𝑣𝑝 = = sin
𝑘 𝑘 2
Group velocity:
𝑑𝜔 𝑑 𝑘𝑎 𝜔0 𝑎 𝑘𝑎
𝑣𝑔 = = 𝜔0 sin = cos
𝑑𝑘 𝑑𝑘 2 2 2

© DR. P. MANDAL 2
Case 1: If the value of 𝑘 is small (i.e., wavelength 𝜆 is large, or frequency is small)
𝜔0 𝑘𝑎 𝜔0 𝑎
lim 𝑣𝑝 ≈ = = lim 𝑣𝑔
𝑘→0 𝑘 2 2 𝑘→0

Thus for 𝑘 → 0 the group velocity and the phase velocity are same which signifies that the medium will not
act as dispersive medium for the long wavelength limit (𝜆 large, frequency small). If acoustic wave passes
the medium then it will not be dispersive as 𝑣𝑝 = 𝑣𝑔 .
Case 2: For a larger value of 𝑘, the group velocity differs from the phase velocity (𝑣𝑝 ≠ 𝑣𝑔). The medium
thus behaves as dispersive medium for relatively shorter wavelength.
𝑘𝑎 𝑛𝜋
If 𝜔 = 𝜔0 , sin = 1 ⇒ 𝑘𝑛 = ± , 𝑛 = 1, 2, 3, …
2 𝑎
In this case we get standing wave energy will be confined in a region. Energy will not pass through medium.
Group velocity 𝑣𝑔 = 0 here as energy is not carried out through the medium.

3. Linear Diatomic Lattice Chain:

Figure: Solid State Physics by S. P. Kuila

Equations of motion:
𝐹2𝑛 = 𝑚𝑢̈ 2𝑛 = 𝛽(𝑢2𝑛+1 − 𝑢2𝑛 ) − 𝛽(𝑢2𝑛 − 𝑢2𝑛−1 )
= 𝛽(𝑢2𝑛+1 + 𝑢2𝑛−1 − 2𝑢2𝑛 ) … (1)
𝐹2𝑛+1 = 𝑀𝑢̈ 2𝑛+1 = 𝛽(𝑢2𝑛+2 + 𝑢2𝑛 − 2𝑢2𝑛+1 ) … (2)
Trial solution:
𝑢2𝑛 = 𝐴𝑒 𝑖(𝜔𝑡−𝑘.2𝑛.𝑎) = 𝐴𝜙 where 𝜙 = 𝑒 𝑖(𝜔𝑡−𝑘.2𝑛.𝑎)
̅̅̅̅̅̅̅̅.𝑎)
𝑢2𝑛+1 = 𝐵𝑒 𝑖(𝜔𝑡−𝑘.2𝑛+1 = 𝐵𝜙𝑒 −𝑖𝑘𝑎
̅̅̅̅̅̅̅̅.𝑎)
𝑢2𝑛+2 = 𝐴𝑒 𝑖(𝜔𝑡−𝑘.2𝑛+2 = 𝑢2𝑛 𝑒 −2𝑖𝑘𝑎 = 𝐴𝜙𝑒 −2𝑖𝑘𝑎
̅̅̅̅̅̅̅̅.𝑎)
𝑢2𝑛−1 = 𝐵𝑒 𝑖(𝜔𝑡−𝑘.2𝑛−1 = 𝐵𝜙𝑒 𝑖𝑘𝑎
Substituting the trial solutions in equations (1) and (2), we have

© DR. P. MANDAL 3
 𝑚(−𝜔2 𝐴) = 𝛽{𝐵(𝑒 𝑖𝑘𝑎 + 𝑒 −𝑖𝑘𝑎 ) − 2𝐴}
or, −𝑚𝜔2 𝐴 = 𝛽(2𝐵 cos 𝑘𝑎 − 2𝐴)
⇒ (2𝛽 − 𝑚𝜔2 )𝐴 − 2𝛽 cos 𝑘𝑎 𝐵 = 0 … (3)
and
𝑀(−𝜔2 𝐵) = 𝛽{𝐴(𝑒 𝑖𝑘𝑎 + 𝑒 −𝑖𝑘𝑎 ) − 2𝐵}
or, −𝑀𝜔2 𝐵 = 𝛽(2𝐴 cos 𝑘𝑎 − 2𝐵)
⇒ 2𝛽 cos 𝑘𝑎 𝐴 − (2𝛽 − 𝑀𝜔2 )𝐵 = 0 … (4)
Equations (3) and (4) yield nontrivial solutions for 𝐴 and 𝐵 if
2𝛽 − 𝑚𝜔2 −2𝛽 cos 𝑘𝑎
| |=0
2𝛽 cos 𝑘𝑎 −2𝛽 + 𝑀𝜔2
or, (2𝛽 − 𝑚𝜔2 )(𝑀𝜔2 − 2𝛽) + (2𝛽 cos 𝑘𝑎)2 = 0
or, −4𝛽 2 + 2𝛽𝑀𝜔2 + 2𝛽𝑚𝜔2 + 𝑀𝑚𝜔4 + 4𝛽 2 cos 2 𝑘𝑎 = 0
⇒ 𝑀𝑚𝜔4 − 2𝛽(𝑀 + 𝑚)𝜔2 + 4𝛽 2 sin2 𝑘𝑎 = 0
1 1 4𝛽 2
⇒ 𝜔4 − 2𝛽 ( + ) 𝜔2 + sin2 𝑘𝑎 = 0 … (5)
𝑀 𝑚 𝑀𝑚

1 1 1 1 1 2 16𝛽 2 2
∴ 𝜔2 = [2𝛽 ( + ) ± √4𝛽 2 ( + ) − sin 𝑘𝑎]
2 𝑀 𝑚 𝑀 𝑚 𝑀𝑚

1 1 1 1 2 4
√
= 𝛽 [( + ) ± ( + ) − sin2 𝑘𝑎] … (6)
𝑀 𝑚 𝑀 𝑚 𝑀𝑚

Equation (5) yields two roots (two frequency relations between 𝜔 and 𝑘 – the dispersion relations) as in
equation (6):

1 1
2 = 𝛽 [( + ) + √(
1 1 2 4
𝜔+ + ) − sin2 𝑘𝑎] … (7)
𝑀 𝑚 𝑀 𝑚 𝑀𝑚

1 1 1 1 2 4
𝜔−2 = 𝛽 [( + ) − √( + ) − sin2 𝑘𝑎] … (8)
𝑀 𝑚 𝑀 𝑚 𝑀𝑚

𝜔+ corresponds to optical mode of vibration while 𝜔− corresponds to the acoustic mode of vibration (as
the optical frequency is very large corresponds to acoustic frequency 𝜔+ ≫ 𝜔− ).
Case 1: Let us now consider a special case 𝑘𝑎 → 0, sin 𝑘𝑎 ≈ 0 – at the center of the Brillouin zone.
(a) Optical Branch: In this case (𝑘𝑎 → 0, sin 𝑘𝑎 ≈ 0) equation (7) yields

1 1
𝜔+ = √2𝛽 ( + )
𝑀 𝑚

© DR. P. MANDAL 4
(b) Acoustic Branch: In this case (𝑘𝑎 → 0, sin 𝑘𝑎 ≈ 𝑘𝑎) equation (8) yields

1 1 1 1 2 4𝑘 2 𝑎2
𝜔−2 = 𝛽 [( + ) − √( + ) − ]
𝑀 𝑚 𝑀 𝑚 𝑀𝑚

1 1 4𝑘 2 𝑎2 1 1 1
= 𝛽[ −√ 2 − ] where ( + = )
𝜇 𝜇 𝑀𝑚 𝑀 𝑚 𝜇

𝛽 4𝜇2 𝑘 2 𝑎2 𝛽 2𝜇2 𝑘 2 𝑎2
= [1 − √1 − ] ≈ [1 − (1 − )]
𝜇 𝑀𝑚 𝜇 𝑀𝑚

2𝛽𝜇𝑘 2 𝑎2 2𝛽𝑘 2 𝑎2
= =
𝑀𝑚 𝑀+𝑚

2𝛽
∴ 𝜔_ = √ 𝑘𝑎
𝑀+𝑚

If 𝑘 = 0 (at the center of Brillouin Zone) then the acoustic mode will disappear, optical mode will remain.
Case 2: Let 𝑘𝑎 → 𝜋/2, sin 𝑘𝑎 ≈ 1 – at the boundary of the Brillouin Zone
(a) Optical Branch: In this case (𝑘𝑎 → 𝜋/2, sin 𝑘𝑎 ≈ 1) equation (7) yields

1 1 1 1 2 4
2 = 𝛽 [(
𝜔+ + ) + √( + ) − ]
𝑀 𝑚 𝑀 𝑚 𝑚𝑀

1 1 1 1 2
= 𝛽 [( + ) + (− + )] = 𝛽 ( )
𝑀 𝑚 𝑀 𝑚 𝑚

2𝛽
⇒ 𝜔+ = √
𝑚

(b) Acoustic Branch: In this case (𝑘𝑎 → 𝜋/2, sin 𝑘𝑎 ≈ 1) equation (8) yields

1 1 1 1 2 4
𝜔−2 = 𝛽 [( + ) − √( + ) − ]
𝑀 𝑚 𝑀 𝑚 𝑚𝑀

1 1 1 1 2
= 𝛽 [( + ) − (− + )] = 𝛽 ( )
𝑀 𝑚 𝑀 𝑚 𝑀

2𝛽
⇒ 𝜔− = √
𝑀

Relative amplitudes and phases for optical and acoustic mode of vibration:
Let, 𝑘 → 0, cos 𝑘𝑎 → 1.

© DR. P. MANDAL 5
 1 1 2𝛽
𝜔+ = √2𝛽 ( + ) , 𝜔_ = √ 𝑘𝑎
𝑀 𝑚 𝑀+𝑚

For 𝜔 = 𝜔+ , we have from equation (3)

2 )𝐴 − 2𝛽 cos 𝑘𝑎 𝐵 = 0
(2𝛽 − 𝑚𝜔+
1 1
⇒ [2𝛽 − 𝑚2𝛽 ( + )] 𝐴 − 2𝛽. 𝐵 = 0
𝑀 𝑚
2𝑚 𝐴 𝑀
⇒− 𝛽𝐴 = 2𝛽. 𝐵 ⇒ = −
𝑀 𝐵 𝑚
Thus the atoms move out of phase in optical mode of vibration.
For 𝜔 = 𝜔− , we have from equation (3)
2 )𝐴 − 2𝛽 cos 𝑘𝑎 𝐵 = 0
(2𝛽 − 𝑚𝜔+
2𝛽
⇒ [2𝛽 − 𝑚 𝑘 2 𝑎2 ] 𝐴 − 2𝛽. 𝐵 = 0
𝑀+𝑚
⇒𝐴=𝐵
In acoustic mode of vibration, both atoms vibrate with same amplitude and in same phase.

Figure: Solid State Physics by S. P. Kuila

© DR. P. MANDAL 6
Specific heat of solids
The specific heat of solids is defined as 𝐶𝑣 = 𝑑𝑈/𝑑𝑇 where U is the energy of the system.
A. Dulong–Petit Law:
Specific heats of all solids are same and that is 6cal/mol/℃
This law is based on the classical explanation and the law of equipartition energy. It assumes the atoms in
a crystal as 3 dimensional harmonic oscillators. By the theorem of equipartition of energy, the average
kinetic energy of 3D oscillators is 3𝐾𝑇/2. Therefore, the total average energy of an oscillator is 3𝐾𝑇.
(average potential energy of a harmonic oscillator is same as the average kinetic energy).
Therefore, the molar energy
𝑈 = 3𝑁𝐴 𝐾𝑇 = 3𝑅𝑇
𝑑𝑈
⇒ 𝐶𝑉 = = 3𝑅 = 6 cal/mol/℃
𝑑𝑇
This is the same for all solids and independent of temperature. In practice, the observation is different.
Dulong–Petit law is consistent at higher temperatures only.
Limitations:
(i) Dulong–Petit law is consistent with experimental observation at room temperature or above. At lower
temperature 𝐶𝑣 varies with 𝑇.
(ii) 𝐶𝑣 below the room temperature depends on the material.

© DR. P. MANDAL 7
B. Einstein Model:
Phonons: Quantum of the vibrational energy. Energy of phonon 𝐸𝑛 = 𝑛ℎ𝜈 = 𝑛ℏ𝜔
It follows the similar property as photon does.
Momentum 𝑝 = ℎ𝜈/𝑐 = ℏ𝜔/𝑐 where 𝑐 is the speed of the acoustic wave.
No of states between 𝜔 and 𝜔 + 𝑑𝜔:
4𝜋𝑉 (ℏ𝜔)2 ℏ
𝑁(𝜔)𝑑𝜔 = 𝑑𝜔
ℎ3 𝑐 2 𝑐
4𝜋𝑉 𝑉
= 3 3 𝜔2 𝑑𝜔 = 2 3 𝜔2 𝑑𝜔
8𝜋 𝑐 2𝜋 𝑐
No of states between 𝜈 and 𝜈 + 𝑑𝜈:
4𝜋𝑉 (ℎ𝜈)2 ℎ
𝑁(𝜈)𝑑𝜈 = 𝑑𝜈
ℎ3 𝑐 2 𝑐
4𝜋𝑉
= 3 𝜈 2 𝑑𝜈
𝑐
In Einstein model, the energy is calculated in the semi-classical approach as follows. It uses the MB
statistics and phonon hypothesis.
𝑁𝑛 = 𝑁0 𝑒 −𝜀𝑛 /𝑘𝑇 = 𝑁0 𝑒 −𝑛ℎ𝜈/𝑘𝑇
∑∞
𝑛=𝑜 𝜖𝑛 𝑁𝑛 ∑∞
𝑛=𝑜 𝑛ℎ𝜈 𝑒
−𝑛ℎ𝜈/𝑘𝑇 ℎ𝜈
𝜖̅ = ∞ = ∞ −𝑛ℎ𝜈/𝑘𝑇 = ℎ𝜈/𝐾𝑇
∑𝑛=𝑜 𝑁𝑛 ∑𝑛=𝑜 𝑒 𝑒 −1
Molar energy
ℎ𝜈
𝑈 = 𝑁𝐴 𝜖̅ = 𝑁𝐴
𝑒 ℎ𝜈/𝐾𝑇 −1
𝑑𝑈 ℎ𝜈𝑁𝐴 (−1)𝑒 ℎ𝜈/𝐾𝑇
ℎ𝜈 1
∴ 𝐶𝑉 = = ℎ𝜈/𝐾𝑇 2
(− 2 )
𝑑𝑇 (𝑒 − 1) 𝐾 𝑇
This is for one dimensional oscillator.
For a 3D oscillator
𝑑𝑈 ℎ𝜈 2 𝑒 ℎ𝜈/𝐾𝑇
𝐶𝑣 = 3 = 3𝑁𝐴 𝐾 ( )
𝑑𝑇 𝐾𝑇 (𝑒 ℎ𝜈/𝐾𝑇 − 1)2
𝑘𝜈/𝐾 dimensionally represents a temperature, say 𝜃𝐸 . the Einstein temperature.
𝑒 𝜃𝐸 /𝑇
∴ 𝐶𝑣 = 3𝑅(𝜃𝐸 /𝑇)2
(𝑒 𝜃𝐸 /𝑇 − 1)2
Note that, if quantum mechanical expression of the energy were taken 𝜖𝑛 = (𝑛 + 1/2)ℎ𝜈, the average
energy would be
ℎ𝜈 ℎ𝜈
𝜖̅ = +
𝑒 ℎ𝜈/𝐾𝑇 −1 2
However, the zero point energy (ℎ𝜈/2) would not contribute to the specific heat as it is independent of
temperature.

© DR. P. MANDAL 8
High temperature limit:
Let the temperature is large and hence ℎ𝜈/𝐾𝑇 is small.
ℎ𝜈
∴ 𝑒 ℎ𝜈/𝐾𝑇 ≈ 1 +
𝐾𝑇
ℎ𝜈
⇒ 𝜖̅ = = 𝐾𝑇
𝑒 ℎ𝜈/𝐾𝑇 −1
⇒ 𝑈 = 3𝑁𝐴 𝜖̅ = 3𝑁𝐴 𝐾𝑇 = 3𝑅𝑇
𝜕𝑈
∴ 𝐶𝑣 = = 3𝑅
𝜕𝑇
Low temperature limit:
At the low temperature, ℎ𝜈/𝐾𝑇 is large and hence 𝑒 ℎ𝜈/𝐾𝑇 ≫ 1.
ℎ𝜈
∴ 𝜖̅ = ≈ ℎ𝜈 𝑒 −ℎ𝜈/𝐾𝑇
𝑒 ℎ𝜈/𝐾𝑇 −1
⇒ 𝑈 = 3𝑁𝐴 𝜖̅ = 3𝑁𝐴 ℎ𝜈 𝑒 −ℎ𝜈/𝐾𝑇
𝑑𝑈 ℎ𝜈 1
∴ 𝐶𝑉 = = 3𝑁𝐴 ℎ𝜈𝑒 −ℎ𝜈/𝐾𝑇 (− ) (− 2 )
𝑑𝑇 𝐾 𝑇
ℎ𝜈 2
= 3𝑁𝐴 𝐾 ( ) 𝑒 −ℎ𝜈/𝐾𝑇
𝐾𝑇
𝜃𝐸 2 −𝜃 /𝑇
= 3𝑅 ( ) 𝑒 𝐸
𝑇
Experimental Evidences
(i) The specific heat of solid at very low temperature follows a 𝑇 3 law that is 𝐶𝑉 ∝ 𝑇 3 at low temperature.
C. Debye’s Theory:
Debye assumed that the vibration of solid should be consider as a whole, rather than the vibration of a single
atom (as Einstein did) because the atoms in the solid form a coupled system.
Since the vibration of a solid as a whole is considered, corresponding to every longitudinal mode there will
be two transverse modes of vibration. Thus, the number of modes corresponding to the frequency between
𝜈 and 𝜈 + 𝑑𝜈 is
4𝜋𝑉 2
𝑍(𝜈)𝑑𝜈 = 𝜈 𝑑𝜈
𝑐3
Let longitudinal wave speed is 𝑐𝑙 while the transverse wave speed is 𝑐𝑡 . Thus, total number of modes
considering one longitudinal and two transverse modes of vibration between the frequency range 𝜈 and 𝜈 +
𝑑𝜈 is
1 2 2
𝑍(𝜈)𝑑𝜈 = 4𝜋𝑉 ( 3 + 3 ) 𝜈 𝑑𝜈 … (1)
𝑐𝑙 𝑐𝑡
Debye pointed out that the maximum frequency must be restricted due to the fact that total number of modes
of vibration must be 3𝑁𝐴 (𝑁𝐴 is the Avogadro number) i.e. there is a cut off frequency called the Debye
frequency 𝜈𝐷 .

© DR. P. MANDAL 9
 𝜈𝐷
∫ Z(𝜈)𝑑𝜈 = 3𝑁𝐴
0
𝜈𝐷
1 2
⇒ 4𝜋𝑉 ( + ) ∫ 𝜈 2 𝑑𝜈 = 3𝑁𝐴
𝑐𝑙 3 𝑐𝑡 3 0
4𝜋𝑉 1 2
⇒ ( 3 + 3 ) 𝜈𝐷3 = 3𝑁𝐴
3 𝑐𝑙 𝑐𝑡
9𝑁𝐴 1 2 −1
or, 𝜈𝐷3 = ( 3 + 3 ) … (2)
4𝜋𝑉 𝑐𝑙 𝑐𝑡
Equation (2) represents the expression for the Debye cut off frequency 𝜈𝐷 .
Debye took the expression for average energy as calculated by Einstein, i.e.
ℎ𝜈
𝜖̅ =
𝑒 ℎ𝜈/𝐾𝑇 −1
Thus, the total energy
𝜈𝐷
𝑈 = ∫ Z(𝜈)𝜖̅𝑑𝜈
0
𝜈𝐷
1 2 ℎ𝜈
= 4𝜋𝑉 ( 3 + 3 ) ∫ ℎ𝜈/𝐾𝑇
𝜈 2 𝑑𝜈
𝑐𝑙 𝑐𝑡 0 𝑒 − 1
9𝑁𝐴 ℎ 𝜈𝐷 𝜈 3 𝑑𝜈
= ∫ … (3)
𝜈𝐷3 0 𝑒 ℎ𝜈/𝐾𝑇 − 1
Specific heat:
𝜈𝐷
𝑑𝑈 9𝑁𝐴 ℎ 𝜈 3 𝑑𝜈(−1)𝑒 ℎ𝜈/𝐾𝑇 ℎ𝜈 1
𝐶𝑉 = = 3 ∫ ℎ𝜈/𝐾𝑇 2 (− 2 )
𝑑𝑇 𝜈𝐷 (𝑒 − 1) 𝐾 𝑇
0
𝜈𝐷
9𝑁𝐴 ℎ ℎ 𝑒 ℎ𝜈/𝐾𝑇 𝜈 4
= 3 ∫ ( 2 ) ℎ𝜈/𝐾𝑇 𝑑𝜈 … (4)
𝜈𝐷 𝐾𝑇 (𝑒 − 1)2
0

Let ℎ𝜈/𝐾𝑇 = 𝑥, Note that ℎ𝜈/𝐾 dimensionally represents a temperature. For 𝜈 = 𝜈𝐷 , let ℎ𝜈𝐷 /𝐾 = 𝜃𝐷 –
the Debye temperature. Thus 𝑥 = 𝑥𝑚 = ℎ𝜈𝐷 /𝐾𝑇 = 𝜃𝐷 /𝑇. Substituting in equation (4), we have
𝑥𝑚
9𝑁𝐴 ℎ ℎ 𝑒 𝑥𝑥4 𝐾𝑇 4 𝐾𝑇
𝐶𝑉 = 3 ∫ ( 2 ) 𝑥 ( ) ( ) 𝑑𝑥
𝜈𝐷 𝐾𝑇 (𝑒 − 1)2 ℎ ℎ
0
𝑥𝑚
9𝑁𝐴 ℎ 𝐾 4𝑇 3 𝑒 𝑥𝑥4
= 3 ∫ ( 4 ) 𝑥 𝑑𝑥
𝜈𝐷 ℎ (𝑒 − 1)2
0
𝑥𝑚
𝐾𝑇 3 𝑒𝑥
= 9𝑁𝐴 𝐾 ( ) ∫ 𝑥 𝑥 4 𝑑𝑥
ℎ𝜈𝐷 (𝑒 − 1)2
0

© DR. P. MANDAL 10
 𝑥𝑚
3 𝑒𝑥
= 3𝑅 ( 3 ) ∫ 𝑥 𝑥 4 𝑑𝑥 = 3𝑅𝐷(𝑥) … (5)
𝑥𝑚 (𝑒 − 1)2
0

where, 𝐷(𝑥) is called the Debye function and is defined as

𝑥𝑚
3 𝑒𝑥
𝐷(𝑥) = ( 3 ) ∫ 𝑥 𝑥 4 𝑑𝑥 … (6)
𝑥𝑚 (𝑒 − 1)2
0

For large temperature 𝐶𝑉 ≈ 3𝑅, and thus from equation (5) it is clear that the limiting value of the Debye
function in this case is 1 i.e. 𝐷(𝑥 → 0) ≈ 1.
High temperature limit: T→ large, 𝑥 = ℎ𝜈/𝐾𝑇 →small.
𝑒𝑥 ≈ 1 + 𝑥 ⇒ 𝑒𝑥 − 1 ≈ 𝑥
From equation (5)
𝑥𝑚
3 1
𝐷(𝑥) = ( 3 ) ∫ 2 𝑥 4 𝑑𝑥 = 1
𝑥𝑚 𝑥
0

∴ From equation (4), 𝐶𝑉 = 3𝑅.

Alternative approach: For higher temperature, ℎ𝜈/𝐾𝑇 is small and hence
ℎ𝜈
𝑒 ℎ𝜈/𝐾𝑇 ≈ 1 +
𝐾𝑇
Thus from equation (3)
9𝑁𝐴 ℎ 𝜈𝐷 𝜈 3 𝑑𝜈 9𝑁𝐴 ℎ 𝜈𝐷 𝜈 3 𝑑𝜈
𝑈= ∫ ≈ ∫
𝜈𝐷3 0 𝑒 ℎ𝜈/𝐾𝑇 − 1 𝜈𝐷3 0 (ℎ𝜈/𝐾𝑇)
𝜈𝐷
9𝑁𝐴
= 3 𝐾𝑇 ∫ 𝜈 2 𝑑𝜈 = 3𝑅𝑇
𝜈𝐷
0
𝑑𝑈
∴ 𝐶𝑉 = = 3𝑅
𝑑𝑇
Low temperature limit:
Recall 𝑥 = ℎ𝜈/𝐾𝑇, 𝑥𝑚 = ℎ𝜈𝐷 /𝐾𝑇. For low temperature, 𝑥𝑚 → ∞. From equation (3), we have
9𝑁𝐴 ℎ 𝜈𝐷 𝜈 3 𝑑𝜈 9𝑁𝐴 ℎ 𝑥𝑚 𝐾𝑇 4 𝑥 3 𝑑𝑥
𝑈= 3 ∫ ℎ𝜈
= 3 ∫ ( ) 𝑥
𝜈𝐷 0 𝜈𝐷 0 ℎ 𝑒 −1
𝑒 𝐾𝑇 − 1
9𝑁𝐴 ℎ 𝐾𝑇 4 ∞ 𝑥 3 𝑑𝑥 9𝑁𝐴 ℎ 𝐾𝑇 4 𝜋 4
= 3 ( ) ∫ 𝑥 = 3 ( ) ( )
𝜈𝐷 ℎ 0 𝑒 −1 𝜈𝐷 ℎ 15
𝑑𝑈 9𝑁𝐴 ℎ 𝐾 4 𝜋 4 3
12 4 𝑇 3
∴ 𝐶𝑉 = = 3 ( ) ( ) 4𝑇 = 𝜋 𝑅( )
𝑑𝑇 𝜈𝐷 ℎ 15 5 𝜃𝐷
∴ 𝐶𝑣 ∝ 𝑇 3 → at very low temperature as observed experimentally.

© DR. P. MANDAL 11
This is the famous Debye’s 𝑇 3 law and it agrees well with the experimental result at low temperature.
However, as T→0 K this law fails. In that region 𝐶𝑉 is found to vary linearly with 𝑇 i.e. 𝐶𝑣 ∝ 𝑇 as 𝑇 → 0.
This is not due to the lattice vibration. It is the electronic contribution of specific heat at very low
temperature. At low temperature 𝐶𝑉 = 𝐴𝑇 + 𝐵𝑇 3 .
Difference between the Einstein and the Debye Models:
Limitations of the Debye model
1. Debye assumed continuum model of solid which is approximately true only for long wavelength
vibrational modes. This assumption does not hold for short wavelength.
2. Debye assumed that the frequency of various modes of vibration is independent of temperature of solid
which is also not true.
3. It is assumed that 𝑐𝑙 and 𝑐𝑡 are independent of frequency of vibration that means we assumed a non-
dispersive medium which is not true in general.
4. Debye assumed that the total number of modes of vibration should be restricted to 3𝑁𝐴 . However, there
is no justification behind this assumption in a continuum model.
5. Debye did not consider the interaction between the atoms and the contribution of electrons to this specific
heat.

© DR. P. MANDAL 12