Professional Documents
Culture Documents
Sse PDF
Sse PDF
Textbook:
“Semiconductor Physics and Devices”
By Donald A. Neamen, 1997
Reference:
“Advanced Semiconductor Fundamentals”
By Robert F. Pierret 1987
“Fundamentals of Solid-State Electronics”
By C.-T. Sah, World Scientific, 1994
Homework: 0%
Midterm Exam: 60%
Final Exam: 40%
Introduction
Classification of materials
Crystalline and impure semiconductors
Crystal lattices and periodic structure
Reciprocal lattice
Crystallinity
– Single crystalline, polycrystalline, and amorphous
Imperfection
– A solid is imperfect when it is not crystalline (e.g., impure) or its atom are
displaced from the positions on a periodic array of points (e.g., physical defect).
– Defect: (Vacancy or Interstitial)
– Impurity:
Based on the atomic forces (binding force) that bind the atom together,
the crystals could be divided into:
– Crystal of Inert Gases (Low-T solid):
Van der Wall Force: dipole-dipole interaction
– Ionic Crystals (8 ~ 10 eV bond energy):
Electrostatic force: Coulomb force, NaCl, etc.
– Metal Crystals
Delocalized electrons of high concentration, (1 e/atom)
– Hydrogen-bonded Crystals ( 0.1 eV bond energy)
H2O, Protein molecules, DNA, etc.
Unit cell: is a small volume of the crystal that can be used to represent
the entire crystal. (not unique)
Primitive unit cell: the smallest unit cell that can be repeated to form
the lattice. (not unique) Example: FCC lattice
To denote the crystal directions and planes for the 3-d crystals.
Plane (h k l)
Equivalent planes {h k l}
Direction [h k l]
Equivalent directions <h k l>
In 3-d solids, there are 7 crystal systems (1) triclinic, (2) monoclinic, (3)
orthorhombic, (4) hexagonal, (5) rhombohedral, (6) tetragonal, and (7)
cubic systems.
N-fold symmetry:
⇒6-fold symmetry
With 2π/n rotation, the
crystal looks the same!
Solid-State Electronics 21 Instructor: Pei-Wen Li
Chap. 1 Dept. of E. E. NCU
Basic Cubic Lattice
Consider a BCC structure and the (110) plane, the surface density is
found by dividing the number of lattice atoms by the surface area;
Surface density = 2 atoms
(a1 )(a1 2 )
v 1 1 1 v 1 1 1
a = (− , ,− ), b = ( , ,− )
4 4 4v 4 4 4
v
a •b 1 θ=109.4o
θ = cos −1 v v ,θ = cos −1( ) = 109.4o
a || b | 3
1 1 1
1 1 1 ( , ,− )
(0,0,0)
( − , ,− ) 4 4 4
4 4 4
The space lattice of diamond is fcc. It is composed of two fcc lattices
displaced from each other by ¼ of a body diagonal, (¼, ¼, ¼ )a
Compound Semiconductors:
(SiC, SiGe, GaAs, GaP, InP, InAs, InSb, etc)
– Has the same geometry as the diamond structure except that zinc blende crystals
are binary or contains two different kinds of host atoms.
Compound Semiconductors
(ZnO, GaN, ALN, ZnS, ZnTe)
– The adjacent tetrahedrons in zinc blende structure are rotated 60o to give the
wurzite structure.
– The distortion changes the symmetry: cubic →hexagonal
– Distortion also increase the energy gap, which offers the potential for optical device
applications.
Every crystal structure has two lattices associated with it, the crystal
lattice (real space) and the reciprocal lattice (momentum space).
The relationship between the crystal lattice vector ( aˆ , bˆ, cˆ ) and
reciprocal lattice vector ( Aˆ , Bˆ , Cˆ ) is
ˆ ˆ
ˆA = 2π bxcˆ ; Bˆ = 2π cˆxaˆ ; Cˆ = 2π aˆxb
aˆ ⋅ bˆxcˆ aˆ ⋅ bˆxcˆ aˆ ⋅ bˆxcˆ
The crystal lattice vectors have the dimensions of [length] and the
vectors in the reciprocal lattice have the dimensions of [1/length],
which means in the momentum space. (k = 2π/λ)
A diffraction pattern of a crystal is a map of the reciprocal lattice of the
crystal.
– The force acting on the j-th particle comes from all the other 1023-1
particles.
– The rate of collision between particles is very high, 1013
collisions/sec
⇒average electron motion instead of the motion of each electron at a
given instance of time are interested. (Statistical Mechanics)
equilibrium statistical mechanics:
Observation:
– the existence of a peak in the density of scattered electrons can be
explained as a constructive interference of waves scattered by the periodic
atoms.
– the angular distribution of the deflected electrons is very similar to an
interference pattern produced by light diffracted from a grating.
− h 2 1 ∂ 2ψ ( x) 1 ∂φ (t )
+ V ( x ) = j h =E
2m ψ ( x) ∂x 2
φ (t ) ∂t
where E is the total energy, and the solution of the eq. is
and the time-indep. Schrodinger equation can be written as φ (t ) = e − j ( E / h )t
∂ 2ψ ( x) 2m
+ 2 ( E − V ( x))ψ ( x) = 0
∂x 2 h
The physical meaning of wave function:
– Ψ(x,t) is a complex function, so it can not by itself represent a real
physical quantity.
– |Ψ2(x,t)| is the probability of finding the particle between x and
x+dx at a given time, or is a probability density function.
– |Ψ2(x,t)|= Ψ(x,t) Ψ*(x,t) =ψ(x)* ψ(x) = |ψ(x)|2 -- indep. of time
∞ 2
1. ∫−∞
ψ ( x) dx = 1
since |ψ(x)|2 represents the probability density function, then for a
single particle, the probability of finding the particle somewhere is
certain.
If the total energy E and the potential V(x) are finite everywhere,
2. ψ(x) must be finite, single-valued, and continuous.
3. ∂ψ(x)/∂x must be finite, single-valued, and continuous.
Boundary conditions:
1. ψ(x) must be continuous, so that ψ(x = 0) = ψ(x = a) = 0
⇒A1 = A2sinKa ≡ 0 ⇒ K = nπ/a, where n is a positive integer.
∞ 2
a 2
2. ∫−∞ ψ ( x ) dx = 1 ⇒ ∫ A2
2
sin
0
2
Kxdx = 1 ⇒ A2 =
a
So the time-indep. Wave equation is given by
2 nπx
ψ ( x) = sin( ) where n = 1,2,3...
a a
The solution represents the electron in the infinite potential well is in a
standing waveform. The parameter K is related to the total energy E,
therefore, h 2 n 2π 2
E = En = where n is a positive integer
2ma 2
That means that the energy of the particle in the infinite potential well
is “quantized”. That is, the energy of the particle can only have
particular discrete values.
In region I, V = 0,
And the general solution of this equation is
∂ 2ψ 1 ( x) 2mE
+ 2 ψ 1 ( x) = 0
∂x 2
h
2mE
ψ 1 ( x) = A1e jK x + B1e − jK x ( x ≤ 0) where K1 =
1 1
h2
In region II, V = Vo, if we assume E < Vo, then
∂ 2ψ 2 ( x) 2m
− (Vo − E )ψ 2 ( x) = 0
2 2
∂x h
Solid-State Electronics 17 Instructor: Pei-Wen Li
Chap. 2 Dept. of E. E. NCU
The Step Potential Function
The results implies that there is a finite probability that a particle will
penetrate the barrier, that is so called “tunneling”.
The transmission coefficient is defined by T = A3 ⋅ A *
3
A1 ⋅ A1*
If E<<Vo, E E
T ≅ 16 1 − exp(− 2 K 2 a )
Vo Vo
This phenomenon is called “tunneling” and it violates classical
mechanics.
4πε o r
2.1
2.15
2.23
Recall from the previous analysis that the energy of a bound electron is
quantized. And for the one-electron atom, the probability of finding the
electron at a particular distance from the nucleus is not localized at a
given radius.
Consider two atoms that are in close proximity to each other. The
wave functions of the two atom electrons overlap, which means that
the two electrons will interact. This interaction results in the discrete
quantized energy level splitting into two discrete energy levels.
Kronig-Penny Model
u1 (a) = u2 (−b) ⇒ Ae j (α − k ) a + Be − j (α + k ) a − Ce − j ( β − k ) b − De j ( β + k ) b = 0
du1 du2
= ⇒ (α − k )Ae j (α − k ) a − (α + k )Be − j (α + k ) a − (β − k )Ce − j ( β − k )b + (β + k )De j ( β + k )b = 0
dx x=a dx x=−b
Solid-State Electronics 7 Instructor: Pei-Wen Li
Chap. 3 Dept. of E. E. NCU
Kronig-Penny Model
There is a nontrivial solution if, and only if, the determinant of the
coefficients is zero. This result is
− 1 ≤ f (ξ ≡ E / Vo ) =
(
− α2 + β2 )
(sin αa )(sin βb) + (cos αa )(cos βb) = cos k (a + b) ≤ 1
2αβ
The above equation relates k to the total energy E (through α) and the
potential function Vo (through β). The allowed values of E can be
determined by graphical or numerical methods.
Energy 4
3
2
1
sin αa
Consider the function of f (αa) = P' + cos αa graphically,
αa
This shows the concept of the allowed energy bands for the particle
propagating in the crystal.
Solid-State Electronics 14 Instructor: Pei-Wen Li
Chap. 3 Dept. of E. E. NCU
Reduced k-space
Energy Band
n
The drift current due to the motion of electrons is J = −e∑ vi
i =1
where n is the number of electrons per volume and vi is the electron
velocity in the crystal.
m* is called the effective mass which takes into account the particle
mass and the effect of the internal forces.
Recall for a free electron, the energy and momentum are related by
p 2 h 2k 2 dE h 2 k hp 1 dE p
E= = ⇒ = = ⇒ = =v
2m 2m dk m m h dk m
– So the first derivative of E w.r.t. k is related to the velocity of the particle.
In addition, d 2E h2 1 d 2E 1
= ⇒ 2 =
dk 2 m h dk 2 m
– So the second derivative of E w.r.t. k is inversely proportional to the mass
of the particle.
In general, the effective
mass could be related to
1 1 d 2E
=
m* h 2 dk 2
m* >0 near the bottoms of all band; m* <0 near the tops of all bands
m* <0 means that, in response to an applied force, the electron will accelerate
in a direction opposite to that expected from purely classical consideration.
In general, carriers are populated near the top or bottom band edge in a
semiconductor—the E-k relationship is typically parabolic and, therefore,
d 2E
= constant ...E near E edge
dk 2
thus carriers with energies near the top or bottom of an energy band typically
exhibit a CONSTANT effective mass
Solid-State Electronics 20 Instructor: Pei-Wen Li
Chap. 3 Dept. of E. E. NCU
Concept of Hole
Valence Band:
– In all cases the valence-band maximum occurs at the zone center, at k = 0
– is actually composed of three subbands. Two are degenerate at k = 0,
while the third band maximizes at a slightly reduced energy.
The k = 0 degenerate band with the smaller curvature about k = 0 is called
“heavy-hole” band, and the k = 0 degenerate band with the larger
curvature is called “light-hole” band. The subband maximizing at a
slightly reduced energy is the “split-off” band.
– Near k = 0 the shape and the curvature of the subbands is essentially
orientation independent.
Conduction band:
– is composed of a number of subbands. The various subbands exhibit
localized and abssolute minima at the zone center or along one of the
high-symmetry diirections.
– In Ge the conduction-band minimum occurs right at the zone boundary
along <111> direction. ( there are 8 equivalent conduction-band minima.)
– The Si conduction-band minimum occurs at k~0.9(2π/a) from the zone
center along <100> direction. (6 equivalent conduction-band minima)
– GaAs has the conduction-band minimum at the zone center directly over
the valence-band maximum. Morever, the L-valley at the zone boundary
<111> direction lies only 0.29 eV above the conduction-band minimum.
Even under equilibrium, the L-valley contains a non-negligible electron
population at elevated temp. The intervalley transition should be taken
into account.
Direct bandgap: the valence band maximum and the conduction band
minimum both occur at k = 0. Therefore, the transition between the
two allowed bands can take place without change in crystal momentum.
Solid-State Electronics 27 Instructor: Pei-Wen Li
Chap. 3 Dept. of E. E. NCU
Constant-Energy Surfaces
A 3-D k-space plot of all the allowed k-values associated with a given
energy E. The geometrical shapes, being associated with a given
energy, are called constant-energy surfaces (CES).
Consider the CES’s characterizing the conduction-band structures near
Ec in Ge, Si, and GaAs.
2m 2
e
The relative sizes of ml* and mt* can be deduced by inspection of the Si
and Ge constant-energy plots.
length of the elliosoid
m
*
along the axis of revolution
l
=
m*
max. width of the ellipsoid
t
perpendicu lar to the axis of revolution
For both Ge and Si, ml* > mt*. Further, ml*/mt* of Ge > ml*/mt* of Si.
The valence-band structure of Si, Ge, and GaAs are approximately
spherical and composed of three subbands. Thus, the holes in a given
subband can be characterized by a single effective mass parameter, but
three effective mass (mhh*, mlh*, and mso*) are required to characterize
the entire hole population. The split-off band, being depressed in
energy, is only sparsely populated and is often ignored.
h2 a
Solid-State Electronics 35 Instructor: Pei-Wen Li
Chap. 3 Dept. of E. E. NCU
Density of State Function
2mE
Recall that k2 =
h2
We can determine the density of states as a function of energy E by
4π 3
gT ( E )dE = ⋅ ( 2 m ) 2
⋅ E ⋅ dE ⋅ a 3
h 3
mn* and mp* are the electron and hole density of states effective masses.
In general, the effective mass used in the density of states expression
must be an average of the band-structure effective masses.
Solid-State Electronics 38 Instructor: Pei-Wen Li
Chap. 3 Dept. of E. E. NCU
Density of States Effective Mass
For T > 0K, electrons gain a certain amount of thermal energy so that
some electrons can jump to higher energy levels, which means that the
distribution of electrons among the available energy states will change.
3.5
3.8
3.16
Carriers in Semiconductors
Dopant Atoms and Energy Levels
Extrinsic Semiconductor
Statistics of Donors and Acceptors
Charge Neutrality
Position of Fermi Energy
po = ∫
Ev (
4π 2m p)
* 32
− ( EF − E )
Ev − E exp
2πm*p kT
dE = 2
32
− ( E F − Ev )
exp
h3
−∞ 2
kT h kT
32
2πm*p kT
We may define N v = 2 h 2 , (at T =300K, Nv ~1019 cm-3), which
is called the effective density of states function in the valence band
When the donor impurity atoms are added, the density of electrons is
greater than the density of holes, (no > po) ⇒ n-type; EF > EFi
When the acceptor impurity atoms are added, the density of electrons
is less than the density of holes, (no < po) ⇒ p-type; EF < EFi
nn
⇒ no
(N − N a ) +
= d
Nd − Na
2
+ ni
2
2 2
If Na > Nd, ⇒ po (N − N d ) +
= a
Na − Nd
2
+ ni is used to
2
2 2
calculate the conc. of holes in valence band
4.18
4.20
4.24
Carrier Drift
Carrier Diffusion
Graded Impurity Distribution
Hall Effect
Homework
E=0
In semiconductors, holes/electrons are involved in collisions with
ionized impurity atoms and with thermally vibration lattice atoms. As
the hole accelerates in a crystal due to the E-field, the velocity/kinetic
energy increases. When it collides with an atom in the crystal, it lose s
most of its energy. The hole will again accelerate/gain energy until is
again involved in a scattering process.
If the mean time between collisions is denoted by τcp, then the average
drift velocity between collisions is
eτ cp vdp eτ cp
vd = * E ≡ µ p E µp = = *
m E mp
p
where µp (cm2/V-sec) is called the hole mobility which is an important
parameter of the semiconductor since it describes how well a particle
will move due to an E-field.
Two collision mechanisms dominate in a semiconductor:
– Phonon or lattice scattering: related to the thermal motion of atoms; µL ∝T-3/2
– Ionized impurity scattering: coulomb interaction between the electron/hole and the
ionized impurities; µI ∝T3/2/NI., N I = N d + N a : total ionized impurity conc. ↑, µI ↓
+ −
If T↑, the thermal velocity of hole/electron ↑⇒carrier spends less time in the
vicinity of the impurity. ⇒ less scattering effect ⇒ µI ↑
V
ρ = 2πs Fc ; Fc : correction factor
I
For GaAs, the electron drift velocity reaches a peak and then decreases
as the E-field increases. ⇒negative differential mobility/resistivity,
which could be used in the design of oscillators.
This could be understood by considering the E-k diagram of GaAs.
In the lower valley, the density of state effective mass of the electron
mn* = 0.067mo. The small effective mass leads to a large mobility. As
the E-field increases, the energy of the electron increases and can be
scattered into the upper valley, where the density of states effective
mass is 0.55mo. The large effective mass yields a smaller mobility.
The intervalley transfer mechanism results in a decreasing average
drift velocity of electrons with E-field, or the negative differential
mobility characteristic.
dφ d ( E /( −e)) 1 dE
The induced E-field is defined as E x = − = − Fi
= Fi
dx dx e dx
that is, if the intrinsic Fermi level changes as a function of distance
through a semiconductor in thermal equilibrium, an E-field exists.
If we assume a quasi-neutrality condition in which the electron
concentration is almost equal to the donor impurity concentration, then
E − Ei N d ( x)
no ≈ ni exp F ≈ N d ( x ) ⇒ E F − Ei = kT ln
kT ni
d ( EF − Ei ) d (− Ei ) kT dN d ( x)
⇒ = =
dx dx N d ( x) dx
kT 1 dN d ( x)
⇒ E x = −
e N d ( x) dx
So an E-field is induced due to the nonuniform doping.
Einstein relation says that the diffusion coefficient and mobility are not
independent parameters.
I x Bz
for a n-type, n=−
edVH
Once the majority carrier concentration has been determined, we can
calculate the low-field majority carrier mobility.
For a p-semiconductor, Jx = epµpEx. ⇒ µ p = IxL
epVxWd
For a n-semiconductor,
IxL
⇒ µn =
enVxWd
5.14
5.20
For the direct band-to-band generation, the generation rates (in the unit
of #/cm3-sec) of electrons and holes are equal; gn’ = gp’ (may be
functions of the space coordinates and time)
An electron in conduction band may “fall down” into the valence band
and leads to the excess electron-hole recombination process.
Band-to-Band generation:
Impact-Ionization:
An e-h pair is produced as a result of the
energy released when a highly energetic
carrier collides with the crystal lattice. The
generation of carriers through impact ionization
routinely occurs in the high e-filed regions of
devices and is responsible for the avalanche
breakdown in pn junctions.
where τn0 is referred to as the excess minority carrier lifetime (τn0 ∝1/p0)
δn(t )
and the recombination rate of excess carriers Rn’ = Rp’=
τ n0
For a n-type material (no >> po) under low-level injection,
− t / τ pn 0 Rn’ = Rp’= δp (t )
δp (t ) = δp (t = 0)e
τ p0
Under low-level injection, excess carriers are created and the quasi-
Fermi level for holes (minority), EFp, is significantly different from EF.
[ ]
Rn = Rcn − Ren = Cn N t n (1 − f F ( Et ) − n ' f F ( Et ) )
In Process 3 and 4, the net rate at which holes are captured from the
valence band is given by R p = C p N t [pf F ( Et ) − p ' (1 − f F ( Et )) ]
− ( Et − E v )
p ' = N v exp
kT
In semiconductor, if the trap density is not too large, the excess
electron and hole concentrations are equal and the recombination rates
of electrons and holes are equal.
Cn n + C p p '
⇒ f F ( Et ) =
Cn ( n + n ' ) + C p ( p + p ' )
CnC p N t (np − ni2 )
and Rn = R p = ≡R
Cn ( n + n ' ) + C p ( p + p ' )
In thermal equilibrium, np = ni2 ⇒ Rn = Rp = 0
6.14
6.17
6.19
P-Semiconductor N-Semiconductor
EC EC
EFi EF
EFi
EF
EV EV
P-Semiconductor
EC N-Semiconductor
EC
EFi
EF EF
EFi
EV
EV
Solid-State Electronics 2 Instructor: Pei-Wen Li
Chap. 7 Dept. of E. E. NCU
P-N Junction
P-Semiconductor
EC N-Semiconductor
EC
EFi
EF EF
EFi
EV
EV
Depletion region
Semiconductor B
Semiconductor A
EC
EC
EFi EF
EF EFi
EV
EV
Electron
Confinement
Hole
Confinement
Solid-State Electronics 5 Instructor: Pei-Wen Li
Chap. 7 Dept. of E. E. NCU
Strained Si1-xGex Strained Si
.. .. .. .. .. .. .. .. .. .. .. .. .. .. .. ..
Relaxed Si1-xGex Relaxed Si1-xGex
misfit dislocation misfit dislocation
bulk Si bulk Si