Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Download as pdf or txt
Download as pdf or txt
You are on page 1of 11

Motion of the atoms: Lattice vibrations

Properties of solids can be divided (roughly) into phenomena that are related to motion of the atoms (around their equilibrium position) and those determined by electrons. This division is justifiable since the motion of the nuclei is much Slower (larger mass) than electron motions (e- remain in ground state). From Heisenbergs uncertainty principle it is clear that even at absolute zero, the atoms must vibrate around their equilibrium position (zero point energy). For many phenomena electron motion leads to transport phenomena (bands, scattering, thermoelectric effects). These are not considered here. Lattice vibrations can explain: Sound velocity Thermal properties such as Heat capacity Specific heat Thermal expansion Thermal conductivity (for insulators and semicoductors) Elastic properties such as Temperature dependence of elastic constants Hardness Optical properties such as Infrared absorption

Linear chain of coupled harmonic oscillators


Longitudinal waves

r k
un-1 un un+1 un+2 un+4

a Transversal waves
un-1 un un+1 un+2 un+4

r k

a Key assumptions: Periodic boundary conditions Only nearest neighbor interactions Small displacements (harmonic oscillations: F~x, V~x2)

Dispersion (k) for linear atomic chain

r r r 2 m k and k are physically equivalent: k ' = k + a

(m integer)

Only wavelengths longer than 2a are needed.

All physical information (all frequencies, group velocities) are r contained in the interval of
a <k a

This corresponds to the first Brillouin zone!

Group velocity
From the solution it follows that the Group velocity is
d Ca 2 ka vg = = cos dk M 2

Note that:

1. There is a frequency maximum:


=2 C M
0 wavevector k
/a

2. On the edge of the Brillouin zone


vg = d =0 dk

The wave is a standing wave hence no energy is transmitted!

3. Long wavelength limit: For ka<<1 the group velocity does not depend on k. No dispersion! The velocity of sound does not depend on frequency.

group velocity vg=d/dk

Summary last time


We looked at lattice vibrations for a linear chain of identical masses M (classically) coupled harmonically by the force constant C. The displacement can be described i ( kna t ) as a wave

un = Ae

1 frequency

Long wavelength limit ka << 1

= ka

It is sufficient to only consider kvectors in the 1st Brillouin zone. There are N normal modes, the frequency is = a C cos ka
M 2

C M

There is a maximum frequency:


max C =2 M

0 wave vector k

at k=/a

The group velocity on the edge of the Brillouin zone disappears standing waves (no transport of energy). For long wavelengths (ka<<1 : typically sound waves) the group velocity does not depend on k (no dispersion).

d Ca 2 C ka ka <<1 vg = cos v g = a = const. = dk M M k 2

Diatomic of coupled harmonic oscillators


Longitudinal waves

vn-1 un-1 un

vn un+1

vn+1

r k

a Same assumptions: Periodic boundary conditions Only nearest neighbor interactions Small displacements (harmonic oscillations: F~x, V~x2) Solving the two equations of motion yields
m+M 2 m+M (1 cos(ka )) 2 = C C mM mM mM
2

Lattice vibrations for diatomic chain


m+M 2 m+M (1 cos(ka )) =C C mM mM mM
2 2

For each k value there are two values of . The branches of (k) Nomenclature: The branches are referred to as acoustic and optical branches.

1.5

frequency

Optical

1.0

Only one branch behaves like 0.5 sound waves (/kconst. For k0). For the optical branch the atoms are oscillating in antiphase and in an ionic 0.0 crystal these charge oscillations -/a (magnetic dipole moment) couple to electromagnetic radiation (optical waves).

Acoustic

0 /a wave vector k

2/a

Definition: All branches that have 0 at k=0 are optical. This does not necessarily mean optical activity! What do the different branches mean? Our detailed calculation shows: Optical branch acoustic branch

Phonons
The energy of lattice vibrations is quantized! The quants of lattic vibrations are called phonons (similar to photons the quants of the electromagnetic field). Experimental proof: Inelastic neutron scattering E and k of neutrons are multiples of phonon energies. Energy modes of quantum harmonic oscillator: 1 En = n + h n = 0,1, 2,... 2 Consider the state of energy En as state with n phonons, each with energy h Phonons are bosons with spin 0. Many quants can have same energy! (gi is degree of degeneracy)
E e k B T 1 N e = gi gi e
E k T B 1

+ Fermi -Bose Boltzmann

Phonon number n may take any value and change with time. Remember: Why do we need phonons to explain indirect band transitions?

Experiments to determine phonon properties (k)


I. Scattering with Neutrons II. Scattering with electromagnetic radiation (visible light):

If you would scatter with E&M radiation, which wavelength do you suggest? What do you think of inelastic X-ray scattering for determining (k)? Brillouin scattering (off acoustic phonons) Raman scattering (off optical phonons) Which part of (k) can one determine? III. Direct absorption of E&M radiation Which wavelength would you use?

Experiments in real materials


Theory our (simple) model for diatomic base General case in 3D: p atoms/unit cell We expect 3 acoustic and 3(p-1) optical branches. What do we find experimentally? Experiments: 1. Neutron scattering Neutrons excite lattice vibrations: measure E and k of neutrons (transfer to crystal) Ge [111] direction, T=80K Expect 3 acoust. & 3 opt. Branches KBr in [111] direction, 90K

Some branches are degenerated!

More experiments
2. Raman scattering

You might also like