Download ECE692_3_1008

Review of Semiconductor Physics
Energy bands
• Bonding types – classroom discussion
• The bond picture vs. the band picture
Bonding and antibonding
Conduction band and valence band
• The band picture – Bloch’s Theorem
Notice it’s a theorem, not a law. Mathematically derived.
The theorem:
The eigenstates (r) of the one-electron Hamiltonian
2

Hˆ  
 2  V (r)
2m
where V(r + R) = V(r) for all R in a Bravais lattice, can be chosen to have
the form of a plane wave times a function with the periodicity of the
Bravais lattice:
 n,k (r)  eik run,k (r)
where un,k(r + R) = un,k(r) .
Equivalently,
 n,k (r  R)  eik R n,k (r)
Physical picture
- Wave function
- Band structure
1D case
3D case
Indirect gap
Direct gap
Limitations of the band theory
Static lattice: Will introduce phonons
Perfect lattice: Will introduce defects
One-electron Shrödinger Eq: We in this class will live with this
Justification: the effect of other electrons can be regarded as a kind of
background.
Semi-classic theory
Free electron
ħk is the momentum.
2k 2
E (k ) 
2m
1D
k 1 dE
v

m  dk
3D
k 1
v
  k E (k )
m 
Block electron
ħk is the crystal momentum, which is not
a momentum, but is treated as momentum
in the semiclassical theory.
n is the band index.
 2 | k  k 0 |2
E (k ) 
En(k) = En(k+K)
*
2m
1D
3D
 k (r)  e ikr
vn 
v n (k ) 
(k  k 0 ) 1 dEn

*
 dk
m
 | k k0 | 1
  k En (k )
*

m
 n,k (r)  eik run,k (r)
un,k(r + R) = un,k(r)
The Bloch (i.e. semiclassic) electron behaves as a particle following Newton’s laws.
(We are back in the familiar territory.)
• With a mass m*
• Emerging from the other side of the first Brillouin zone upon hitting a boundary
Newton’s 1st law: the Bloch electron moves forever – No resistance?
Newton’s 2nd law:
F = dp/dt = ħdk/dt
Oscillation in dc field. So far not observed yet.
Real crystals are not perfect. Defects scatter electrons.
On average, the electron is scattered once every time period . Upon scattering,
the electron forgets its previous velocity, and is “thermalized.”
vd 
F qE
 *   E
*
m
m
J  E  qnvd  qnE

q
m*
Mobility
q 2 n
  qn  *
m
Values of k
k = 2n/L, n = 1, 2, 3, …, N
Discrete but quasi-continuous
L = Na
Run the extra mile:
Show the above by using the “periodic boundary” condition.
Holes
A vacancy in a band, i.e. a k-state missing the electron, behaves like a particle with
charge +q.
Run the extra mile:
Show the above.
Review of Semiconductor Physics
Carrier Statistics
• Fermi-Dirac distribution
Nature prefers low energy.
Lower energy states (levels) are filled first.
Imaging filling a container w/ sands, or rice, or balls, or whatever
- Each particle is still
T=0K
- Each has some energy, keeping bouncing around
T>0K
• Density of States
How many states are there in the energy interval dE at E?
D(E)dE
1D case derived in class.
The take-home message: D(E)  E1/2
2D case
Run the extra mile
Derive D(E) in 2D.
Hint: count number of k’s in 2D.
The answer:
L2 m*
D( E ) 
2  2
The take-home message:
Or, for unit area
D(E) = constant
3D case
Run the extra mile
Derive D(E) in 3D.
Hint: count number of k’s in 2D.
For unit area,
2 (m* ) 3 / 2
D( E ) 
2 2  3
The take-home message: D(E)  E1/2
E
1 m*
D( E ) 
2  2
Things we have ignored so far: degeneracies
Spin degeneracy: 2
Valley degeneracy: Mc
Mc = 6 for Si
2 (m* ) 3 / 2
D( E )  2M c
2 2  3
E
Total number of carriers per volume (carrier density, carrier concentration)
Run the extra mile
Derive the electron density n.
Hint: Fermi-Dirac distribution approximated by Boltzmann distribution.
Results for n and p are given.
p is the total number of states NOT occupied.
Doping
One way to manipulate carrier density is doping.
Doping shifts the Fermi level.
np = ni2
One small thing to keep in mind:
Subtle difference in jargons used by EEs and physicists
We use the EE terminology, of course.
Fermi level
EF = EF(T)
Same concept
Physicists:
Chemical potential
Fermi energy
(T)
EF = (0)
We already used  for mobility.
Before we talk about device, what are semiconductors anyway?
Classroom discussion
Why can we modulate their properties by orders of magnitude?
Classroom discussion
We have mentioned defect scattering:
Real crystals are not perfect. Defects scatter electrons.
On average, the electron is scattered once every time period . Upon scattering,
the electron forgets its previous velocity, and is “thermalized.”
vd 
F qE
 *   E
*
m
m
J  E  qnvd  qnE

q
m*
Mobility
q 2 n
  qn  *
m
Any deviation from perfect periodicity is a defect. A perfect surface is a defect.
Phonons
Static lattice approximation
Atoms vibrate
Harmonic approximation
Vibration quantized
Each quantum is a phonon.
Similar to the photon:
E = ħ, p = ħk
Phonons scatter carriers, too.
The higher the temperature, the worse phonon scattering.
You can use the temperature dependence of conductivity or mobility to determine
the contributions of various scattering mechanisms.
Phonons
Sound wave in continuous media
 = vk
Microscopically, the solid is discrete.
Phonon dispersion
Wave vector folding, first Brillouin zone.
Watch video at http://en.wikipedia.org/wiki/File:Phonon_k_3k.gif
Recall that
Crystal structure = Bravais lattice + basis
If there are more than 1 atom in the basis, optical phonons
Phonons in the 3D world -- Si
In 3D, there are transverse
and longitudinal waves.
E = h = ħ
15 THz
62 meV
When electron energy is low, the electron
only interacts with acoustic phonons,
Optical phonons and transport
At low fields,
1 * 2 3
m vth  k BT = 38 meV
2
2
For Si,
vd
vd  E  vth
vsat
vth = 2.3 × 107 cm/s
E
At high fields, vd comparable to vth
Electrons get energy from the field, hotter than the lattice – hot electrons
When the energy of hot electrons becomes comparable to that of optical phonons,
energy is transferred to the lattice via optical phonons.
Velocity saturation
For Si, vsat ~ 107 cm/s
Alloys
Compounds, alloys, heterostructures
InP, GaAs, …, SiC
InxGa1-xAsyP1-y, …, SixGe1-x
Epitaxy
Band structure of alloys
Topics
• Review of Semiconductor physics
- Crystal structure, band structures, band structure modification by alloys,
heterostructurs, and strain
- Carrier statistics
- Scattering, defects, phonons, mobility, transport in heterostructures
• Device concepts
- MOSFETs, MESFETs, MODFETs, TFTs
- Heterojunction bipolar transistors (HBT)
- Semiconductor processing
- Photodiodes, LEDs, semiconductor lasers
- (optional) resonant tunneling devices, quantum interference devices,
single electron transistors, quantum dot computing, ...
- Introduction to nanoelectronics
We will discuss heterostructures in the context of devices.
More discussions on semiconductor physics will be embedded in the device
context.