Lecture Notes # 3 Lecture Notes # 3 • Understanding Density of States – Solve 1-D Schrödinger equation for particle-in-a-box – Extend to 3-D – Invoke periodicity requirement – Solve for density of states
Dec 19, 2015
Lecture Notes # 3Lecture Notes # 3• Understanding Density of States
– Solve 1-D Schrödinger equation for particle-in-a-box
– Extend to 3-D
– Invoke periodicity requirement
– Solve for density of states
Review of Quantum MechanicsReview of Quantum Mechanics
H
• Often times you do not know or , but you have boundary conditions and want to solve for possible values of and a functional form of
electron that of
energies quantized allowableor Energy
space/timein electron of nature
depictingion wavefunctMathematic
operatorn Hamiltonia
H
Review of Quantum MechanicsReview of Quantum Mechanics
zyxUm
H ,,2
22
• Most general case: Time independent
Kinetic E. Potential E.
• How do we know first part is K.E?
22
2
22
2
22
1 So,
operator momentum quantum is ˆ22
1
yclassicall 2
1
mmv
ipm
pmv
mvKE
Review of Quantum MechanicsReview of Quantum Mechanics
ap
rikr
ˆ
expFor
rkrikirr
irip
ˆ
particle ofvelocity
m
kv
kmvp
Momentum operator Eigenvalue
a = Momentum Eigenvalue
• Free electron floating around in vacuum• Let’s impose some boundary – confine it to a region of space, a box or a unit cell in 1-D
Particle in a 1-D BoxParticle in a 1-D Box
UUU
U = 00 L x
Confined e-
Confine it by settingoutside the box andinside the box
U = 0
• Since U(x) = 0 for 0<x<L, we can drop U(x) out of the Hamiltonian, which becomes
22
2
mH
• Because outside the box, we know the e- CANNOT be there, so we get the boundary condition:
Particle in a 1-D BoxParticle in a 1-D BoxU
00 and
where
pickmust we,eigenvaluean back
givemust n Hamiltonia theBecause
00
Lff
xAfxfxf
Lxx
OK 0)sin( ,for
OK 0)0sin( ,0for
...4,3,2,1 ,sin
nLLx
x
nxL
nA
• We do not care what happens between 0 and L, so the simplest solution is just:
• Plug into the Schrödinger equation to make sure H=
Particle in a 1-D BoxParticle in a 1-D Box
xL
nA
L
n
m
xL
nA
L
n
m
xL
nA
L
n
dx
d
m
xL
nA
dx
d
m
sin2
1
sin2
cos2
sin2
2
2
222
2
2
22
Energy, E
• Energy values are quantized since n is an integer
• n=1 is lowest energy state, n=2 has higher energy, etc.
Particle in a 1-D BoxParticle in a 1-D Box
0
2
4
6
8
10
12
14
16
18
0 1 2 3 4
2
222
2 x
mL
nE
0 L
E
9x
4x
n=4
n=3
n=2
n=1
n
• We can map out (x,n) vs. E
• Allowed energy states
• Now let’s fill up the states with electrons. Suppose we have N e - we want to pour into our 1-D box.
Particle in a 1-D BoxParticle in a 1-D Box
• For N e- you can calculate the energies since we know we can have 2e-/n states (two spins).
• So N electrons fills nF= N/2 states.
• The highest energy state, nF, gives F, the Fermi energy.
L
n
mF
F
2
2
• Fermi energy is well defined at T = 0 K because there is no thermal promotion
• At high T, there is thermalization, so F is not as clear
Fermi-Dirac DistributionFermi-Dirac Distribution
0
2
4
6
8
10
12
14
16
18
0 1 2 3 4
2
222
2 x
mL
nE
n
1/exp
1
Tkf
B
• Officially defined as the energy where the probability of finding an electron is ½• This definition comes from the Fermi-Dirac Distribution:
• This is the probability that an orbital (at a given energy) will be filled with an e - at a given temperature
• At T=0, =F and = F, so f(F)=1/2
F
• Let’s confine e- now to a 3-D box• Similar to a unit cell, but e- is confined
inside by outside the box• Schrödinger’s equation is now
Particle in a 3-D BoxParticle in a 3-D Box
zyxzyxzyxm
,,,,2 2
2
2
2
2
22
U
• You can show that the answer is:
z
L
ny
L
nx
L
nAzyx zyx
n
sinsinsin,,
• We now have 3 quantum numbers nx, ny, and nz that are totally independent
• (1,2,1) is energetically degenerate with (2,1,1) and (1,1,2)
• What’s different about this situation?– U(x,y,z)=0– No region where U = infinity
• So, there’s really no reason that
• We don’t need those boundary conditions anymore
• Now let’s repeat this box infinitely in each direction to get a repeated “unit cell”
Particle in a 3-D BoxParticle in a 3-D Box
LxUxU
Lxx
0 since
00
• For now, we don’t need such a strict boundary condition• Make sure is periodic with L, which would make each 3-D box identical• Because of this, we’ll have a periodic boundary condition such that
Periodic Boundary ConditionPeriodic Boundary Condition
zyxzyLx ,,,, • Wave functions that satisfy this periodic B.C. and are solutions to the
Schrödinger equation are TRAVELING WAVES (not a standing wave anymore)
• Bloch function
Periodic Boundary ConditionPeriodic Boundary Condition
rkirk
exp
;...4
;2
;0LL
kx
xik
Lnxi
niLnxi
LLxniLxik
x
x
exp
/2exp
2exp/2exp
/2expexp
• Wave vector k satisfies
• Etc. for ky and kz
• Quantum numbers are components of k of the form 2n/L where n=+ or - integer
• Periodicity satisfied
• Substitute
Back to SchrBack to Schrödinger Equationödinger Equation
• Important that kx can equal ky can equal kz or NOT
• The linear momentum operator
2222
22
2
2
2
2
2
22
22
gives
2
into
exp
zyxk
kkk
k
kkkm
km
rrzyxm
rkir
ip̂
m
kv
k
r
rkrirp
rkir
k
kkk
k
isk orbitalin velocity particle theand ,
of eigenvaluean with momentumlinear of
ioneigenfunctan is waveplane theso
ˆ
expfor
• Similarly, can calculate a Fermi level
Fermi Level in 3-DFermi Level in 3-D
ky
kFkz
kx
Fermi level
• Inside sphere k<kF, so orbitals are filled. k>kF, orbitals are empty
• Quantization of k in each direction leads to discrete states within the sphere• Satisfy the periodic boundary conditions at ± 2/L along one direction
• There is 1 allowed wave vector k, with distinct kx, ky, kz quantum #s for the volume element (2/L)3 in k-space
• So, sphere has a k-space volume of
22
2 FF km
Vector in 3-D space
NOTE: This is a sphere only if kx=ky=kz. Otherwise, we have an ellipsoid and have to recalculate everything. That can be a mess.
Sphere: GaAs (CB&VB), Si (VB)
Ellipsoid: Si (CB)
3
3
4FkV
• Number of quantum states is
• Since there are 2 e- per quantum state
Number of Quantum StatesNumber of Quantum States
33
23
4
state quantized allowed 1 of volume
volumetotal
L
kF
31
2
32
32
3
3
3
3
,for Solve3
323
42
V
Nk
k
kV
N
kL
L
kN
F
F
F
F
F
• Depends on e- concentration
• Plug kF into
• Relates Fermi energy to electron concentration
• Total number of electrons, N:
Density of StatesDensity of States
32
22
22
3
2
2
V
N
m
km
F
F
• Density of states is the number of orbitals per unit energy
23
22
2
3
mV
N
212
3
22
2
2
mV
d
dND
Relate to the surface of the sphere. For the next incremental growth in the sphere, how many states are in that additional space?
• Divide by V to get N/V which is electron density (#/cm3)• Volume density of orbitals/unit energy for free electron gas in periodic potential
Density of StatesDensity of States
212
3
22
2
2
1
m
D
21
23
2
*
2
2
2
1CB
e EEm
D
21
23
2
*
2
2
2
1EE
mD VB
h
Effective mass of e-
Effective mass of h+
Starting point energy
CB
VB
Start from VB and go down
Concentration of ElectronsConcentration of Electrons
kTEkTm
n Ce /exp
22
23
2
*
CE
ee dEEfEDn fun
23
2
*
22
kTm
N eC
kTEENn FCC /exp
“EF”
Effective density of states in CB
Writing it with a minus sign indicates that as E difference between EC and EF gets bigger, probability gets lower
Concentration of HolesConcentration of Holes
kTEkTm
p Vh /exp
22
23
2
*
VE
hh dEEfEDp
23
2
*
22
kTm
N hV
kTEENp VFV /exp
Effective density of states in VB
eh ff 1
Intrinsic Carrier ConcentrationIntrinsic Carrier Concentration
kT
EEEENNpn VFFC
VC exp
kTENNpn GVC /exp
gap band the,GVC EEE
kT
EENNpn VC
VC exp
Entropy term Enthalpy term2inpn
ionconcentratcarrier intrinsic in
Constant for a given temperature.Intrinsic = undoped
Intrinsic Carrier ConcentrationIntrinsic Carrier Concentration
kT
ENNn G
VCi 2exp
ni
Ge: 2.4 x 1013 cm-3
Si: 1.05 x 1010 cm-3
GaAs: 2 X 106 cm-3
At 300 K
it fromaway and mequilibriu
under holds which too,constant, is
thatmeans Tgiven afor constant
pn
ni
• At temperature T, n = p by conservation• Add a field and np = constant, but n does not equal p• As n increases, p decreases, and vice versa
• Useful to define Ei, which is Ei = EF when it is an intrinsic semiconductor (undoped), so n = p = n i
kT
EEN
kT
EENn Vi
ViC
Ci expexp
Intrinsic Fermi LevelIntrinsic Fermi Level
C
VViiC
N
N
kT
EE
kT
EEln
kT
EEN
kT
EENn Vi
ViC
Ci expexp
C
ViVCi N
NkTEEEE ln
C
VCVi N
NkTEEE ln2
C
VVCVi N
NkTEEEE ln2
C
VGVi N
NkTEEE ln
22
Intrinsic Fermi LevelIntrinsic Fermi Level
C
VGVi N
NkTEEE ln
22
• This says that the intrinsic Fermi level (relative to the valence band) is about mid-gap ± the (kT/2)ln(NV/NC) scaling factor
Gefor meV 7-
GaASfor meV 35
Sifor meV -13ln2
C
V
N
NkT EG (eV)1.12
1.420.67
• So Ei for Si and Ge is slightly below mid-gap. Ei for GaAs is slightly above. It is minor compared to EG, but just so you know
• All of this has been intrinsic with no dopants
DopantsDopants
DCBD EE
Di Nnn carriers
• Let’s consider adding dopants
ECB
Ei
EVB
ED
n-type
Depends on EG, T
Depends on D, T, and dopant density
At T=0, all donor states are filled. Hence, n = 0.But at room temperature in P doped Si, 99.96% of donor states are ionized.
At mid temperature,
At high temperature, such that
DiD NnnN then , if
ii nnn then ,ND