Semiconductors Physics PDF

SEMICONDUCTORS
PHYSICS
Professor P. Lorenzini
Polytech’Nice Sophia
Electronics Department
Ph. Lorenzini
Semiconductors Physics 2
University of Nice (France)
Objectives
• To Understand the interest of semiconductors in the
production of electronic components
• To Understand transport mechanisms and physical
processes governing the operation of electronics devices.
• Mastering DC and AC operations of the basic PN junction
• To Have all the basics to understand the behavior of
electronic active devices
Ph. Lorenzini
Plan
1. Crystal structure and crystallography

2. Quantum / wave mechanics: the Schrödinger equation
3. Quasi free electrons: the Sommerfeld model
4. The electrons in a periodic structure: Bloch – Brillouin model
5. Current in solids and case of semiconductors
6. Statistical mechanics: the Fermi Dirac function and approximation
of Maxwell - Boltzmann
7. Semiconductor at equilibrium
8. Doping of semiconductor
9. Semiconductor at nonequilibrium: currents in the semiconductor
10. PN junction
Ph. Lorenzini
bibliography
• C. Kittel, « physique de l’état solide », dunod université, 5° ed., 1983
• H. Mathieu, « Physique des semiconducteurs et des composants

électroniques », dunod, 5° ed., 2004
• J. Singh, « physics of semiconductors and their

heterostructures »,Mc.Graw Hill, 1993
• D.A.Neamen, « semiconductor physics and devices: basic principles »,

Mc.Graw Hill, 2003
• Chenming C. Hu , « Modern Semiconductor Devices for Integrated

Circuits », 2009
• Cours de Physique des semiconducteurs, Pr. Rouzeyre, Université de

Montpellier II, 1985
• McMurry and Fay, « Chemistry », Prentice Hall; 4th edition (April 7,

2003) ( les figures du chapitre 1 proviennent majoritairement de cet
ouvrage)
CHAPTER 1
Crystal bonding and crystallography
Ph. Lorenzini
Crystal structure and crystallography
• Crystalline and amorphous states
• Crystalline bonds
• Geometry of the crystals
• Crystal diffraction - Reciprocal Lattice

Ph. Lorenzini
Amorphous and crystal states

• amorphous  glass
• crystal  quartz, SiO2
• distinction:
• At the macroscopic level:
• If the glass temperature is raised, there is a gradual transition from the solid to
the liquid state without bearing condition.
• For a crystal, there is a temperature plateau caused by coexistence (change)
phase.
• At the microscopic level:
• Amorphous: random distribution of the atoms
• Crystal: spatial periodic distribution of atoms.
Ph. Lorenzini
Amorphous and crystal states
(from Neamen)
Amorphous Polycrystalline Single Crystal

(no order) (order over many atomic (high degree of order)
dimensions)
Ph. Lorenzini
Crystal bond
• What are the strengths that allow atoms bond together and form such
structures?
The idea : U crystal  U free  0
• Several parameters / considerations to take onto account:
• Keep the positively charged ions away from each other

• Keep the negatively charged electrons away from each other
• Keep the electrons close to the ion
• Minimize the kinetic energy of electrons by spreading
Ph. Lorenzini
Electronegativity
• Electronegativity: it is the ability of an atom to attract
electrons shared in a covalent bond.
The electronegativity En
between two bonded atoms
could be null, low or high.
•ΔEN ~ 0: the electrons are equally
distributed
•ΔEN ~ 1: electrons are closest to
the most electronegative atom.
•ΔEN is high, the electrons are not
shared
Ph. Lorenzini
Crystal bond
• 4 differents types: • One commun point:
• metal • Atoms try to have their last

electronic layer empty or
• covalent full !
• ionic
• Van der Waals (noble gaz) or

molecular
Ph. Lorenzini
Metallic bond
• Most chemical elements have a more or less
pronounced metallic behavior.
• Built up from elements with few valence electrons
compared to their period or energy level
• Example:
• Sodium (Na)  1s22s22p63s1
• Cuivre (Cu) 1s22s22p63s23p63d104s1
Ph. Lorenzini
Metallic bond: copper
Configuration du Cu
(1s22s22p63s23p63d104s1)
Ph. Lorenzini
Metallic bond
• Outers electrons not strongly linked to atom
• Tis atom can « free » this (these) electron (s)
• nucleus+core electrons look like to a set of positives charges  positive ions
with saturated outer layer
• The cohesion is ensured by the negatively charged electron cloud
• Bonding’s force Coulomb attraction
• Weak bonds softer materials and low melting temperature
Ph. Lorenzini
Metallic bond
Electron-Sea Model: A metal crystal is viewed as a three-dimensional array
of metal cations immersed in a sea of delocalized electrons that are free to
move throughout the crystal.
Ph. Lorenzini
Covalent bond
• Crystals belong to column IV of periodic
table
• Hydrogen like bonds
• Hydrogen:
• 1 outer electron
• To complete the layer, try to « find, accept » an
extra electron
• A second electron from another Hydrogen will
allow to share outer electrons
• We get H2 molecule
H H
+ H
H
H2
Ph. Lorenzini
Covalent bond
• Example : Silicon
• 4 valence (outer) electrons Si
• Need four more electrons to complete the valence energy

shell
• It is sufficient to bring four other Silicon
A significant difference between the

Si covalent bonding of Hydrogen and
Silicon: Unlike H2, once saturated links,
Si Si Si ! the other four still have Silicon "dangling"
bonds  This process may continue 
then we form a crystal.
Si
Ph. Lorenzini
Covalent bond
• Lower energy if the electrons

wander near the 2 nucleus
• We can get the same result with II-
VI or III-V compounds (see after)
Ph. Lorenzini
Ionic bond
• Association of a strongly electronegative chemical element (7e)
and a strongly electropositive element (1e): ex NaCl
• The elements in group 1 of the periodic table tend to lose their
one electron (become positively charged) while the elements in
group VII gain one electron and become negatively charged 
the two ions experience coulomb attraction
3s²3p5
3s1
(D’après McMurry and Fay)

Ph. Lorenzini
Ionic bond
• The strengh of cohesion is due to the Coulomb
attraction of the two ions  ionic bond
• In fact, binding is "identical" to the covalent bond
except that atoms are very differents (not the same
column)
• Ionic/covalent border is not sharp: depends on the
electronic nature of the associated elements Col I – VII
 mainly ionic
• Col II-VI  80% ionic 20% covalent (CdTe)
• Col III-V  60% ionic 40% covalent (GaAs, GaP, InP)
• Col IV-IV  essentielly covalent (Si, Ge)
Ph. Lorenzini
Ionic bond
• the electron released from alcali meal (Na) is trapped by

the halogen (Cl)
• None electron released in the lattice
• Ionic crystals are generally insulating
• Bonds between atom very strong  very hard crystals
Ph. Lorenzini
Van der Waals bond

• Weakest of the chemical bonds
• Electronic distribution = free atom
• Most of this materials are in gazeous form at RT
• Electronics shells (layers) already saturated (col. VIII)
• Very weak binding energy (few% of the ionization
energy of the atom)  crystals melt at low
temperature
• Responsible for the cohesion of the molecules
Ph. Lorenzini
Crystal geometry
• The building block which is repeated infinitely, to
produce the crysal,may be a single atom or quite
complex (protein) : this is the basis.
• For Si,Ge basis : one atom
• For GaAs basis : 2 atoms
• For molecular atom basis : Protein 104-105 at.!!
• The lattice represents a set of points in space which

form a periodic structure
• 2 Dim: net (Réticule)
• 3 Dim: Lattice (Réseau )
Ph. Lorenzini
Crystal geometry
+ =
Lattice Basis (base): Crystal structure

(réseau) a builing block of atoms (structure cristalline)
Ph. Lorenzini
Crystal geometry
  directions
• Bravais Postulat : it exists in the
crystal a set of points Pn with
exactly the same microscopic  ’ ’’ 
environment that P0 (origin pt) ie
that the atomic landscape seen
from Pn is the same as P0 in P0
magnitude and direction. This P1’ Pn
set of points is a set of P1
congruent points or vertex (top)
of the lattice
• Bravais lattice: it is a sample of P2
the lattice that shows all the
translations of the lattice. There
are 14 different lattices
Ph. Lorenzini
Notation to denote planes and points in a

lattice: Miller indices
• Lattice planes: the points (nœuds) of the lattice can be grouped on sets of
parallel and equidistant planes : reticular planes (les plans réticulaires)
• Method :
• Take the intercepts of the plane along the axis
• x1 (distance / origin = x1.a ) in units of lattice constant a
• x2 (distance / origin = x2.b ) in units of lattice constant b
• x3 (distance / origin = x3.c in units of lattice constant c
• Take the reciprocal of the intercepts
• Reduce them to the smallest integers (in the same ratio !)
• We get the Miller indices
• Example:
• x1=4, x2=1, x3=2
• ¼ , 1, ½
• 1,4,2 (h,k,l) =(1,4,2)
• Nota : in case of negative number, we place a minus sign over h

• (h,k,l) defines a plane
Ph. Lorenzini
Crystalline planes (h,k,l)

• A1, A2, A3 : 3 nodes 
reticular plane  OAi  pi ai
z • consider M(OM  x1 a1  x2 a2  x3 a3 )
3
Plan  point of direct lattice
A3
2 •  plane can be described
y
A2 by : hx  ky  lz  N
a3
a2 • It can be rewritten:
a1
x y z
A1
1
  1
x p1 p2 p3
Using Miller index, we get : hx  ky  lz  N . If N=2,  plane is the second

plane (h,k,l) after origin.
Ph. Lorenzini
Crystallographic direction
• Any line through two points of the lattice defines a crystallographic
direction. It’s defined by 3 indices h,k,l smallest integers having the
same relationship to each other that the components of a vector
collinear to the line.
• Notation: [h,k,l]
• Paricular cases : one direction [h,k,l] is orthogonal to the plane of
same indexes (h,k,l) or [h,k,l] is the axial vector of the plane (h,k,l) .
Ph. Lorenzini
Miller Indices (continued)

• Allow us to define crystal directions and crystal planes:
(a,0,0) ou (1,0,0)
(From McMurry and Fay)

Ph. Lorenzini
Some examples.
z
(From McMurry and Fay)
y
Ph. Lorenzini
The 14 Bravais lattices in three dimensional system
Triclinic Monoclinic centered Orthorhombic centered

Monoclinic Orthorhombic
base centered faces centered

Hexagonal
Orthorhombic Orthorhombic Tetragonal
Rhomboédric centered
Tetragonal
Ph. Lorenzini
The 14 Bravais lattices in three dimensional system
Simple Cubic body centered face centered

(sc) (bcc) Cubic Cubic (fcc)
Ph. Lorenzini
Crystalline structures : multiplicity of the lattice
sc bcc
1 atom 2 atoms (From McMurry and Fay)

per unit cell per unit cell
Ph. Lorenzini
fcc
4 atoms (from McMurry and Fay)

per unit cell
Ph. Lorenzini
Hexagonal
Structure
fcc
Structure
(from McMurry and Fay)

Ph. Lorenzini
How to define the primitive unit cell

• Volume is the   directions
smallest possible
 ’ ’’ 
• Not unique way
• Method: P0 a2
• a1 to be the shortest
a1 P1’ Pn
period of the lattice P1
period of the lattice not
P2
parallel to a1
period of the lattice not
coplanar with a1 and a2
Primitive unit cell

Ph. Lorenzini
The body centered cubic (bcc)

z
a1
a2
a3
y
a1 a3
x
a2
a1  a x a
a1  ( x  y  z )
2
a2  a y or: a
a2  ( x  y  z )
2
a
a3  ( x  y  z ) a
a3  ( x  y  z )
2 2
Ph. Lorenzini
Primitive basis vectors for a fcc: the resulting lattice is a rhombohedric

lattice ( 60° for each angle)
From C. Kittel, Dunod, 5°ed.

Ph. Lorenzini
Atoms surface density

Plane (110)
a a 2
a
a
1  4 1 / 4 2 2
surface density    5.66  1014
at.cm
aa 2 (5  10 8 cm) 2  2
Ph. Lorenzini
Silicon and its « diamond » structure
• fcc
• Two atom basis
• Looks like two fcc
shifted/deplaced by ¼
diagonal
• Each atom has 4
nearest neighbors
(tetragonal bond)
Ph. Lorenzini
The reciprocal lattice

• a priori:
• All the physical phenomena are involved in real space (3D)  real
space is obviously the physical space of interest
• but:
• For many applications/studies  most astute to work in a virtual
space : reciprocal lattice.
• Reasons: fully understood when we discuss the problem of X-ray
diffraction and energy bandstructures of crystal.
Ph. Lorenzini
The reciprocal lattice: definition

• One of the reasons of reciprocal lattice:
• Periodic structures  physical properties likewise ( local
electron density, electrostatics potential, …)
• These properties can be defined with ad hoc functions which
are also periodic functions ( crystal periodicity)
f (r  l )  f (r )
Ph. Lorenzini

• One dimension:
• Structure with a period « a »
f ( x  l )  f ( x) avec l  l1a l1 entier

• Such function can be rewritten as a Fourier series
2
f ( x)   An e 2inx / a
 Ag e igx
avec g n  n g   m 1
n g a
Ph. Lorenzini
• verification
1

igx
f ( x)   Ag e igx Ag  f ( x )e dx
g
a cellule
f ( x  l )   Ag eig ( x l )   Ag eigx eigl  f(x)  Ag eigx
g
2
e igl
 1  gl  n l1a  nl1  2  integer  2
a
Ph. Lorenzini

• We start with three primitive vectors in the real space a1,
a2, a3.
• We take a « simple » transformation to get three new
vectors in the reciprocal lattice b1, b2, b3:
a2  a3 a3  a1 a1  a2
b1  2 b2  2 b3  2
(a1 , a2 , a3 ) (a1 , a2 , a3 ) (a1 , a2 , a3 )
bi .a j  2 ij with  ij  0 if i  j ,  ij  1 if i  j
Ph. Lorenzini
Reciprocal lattice building

• From the same origin O of • We have (scalar product):
the real space, we generate
ON .OP  2 hn  kp  lq 
a reciprocal lattice.
• if N, a point of the reciprocal
lattice
integer
2
g  ON  hb 1  k b 2  l b3 g  ON  hn  kp  lq 
OP
• if P, a point of the real lattice

Of course, unit of
l  OP  n a1  p a 2  q a 3
reciprocal lattice is :
• In general Reciprocal lattice
has not the same symmetries
than real space lattice

b 
1
length
 m 1
Ph. Lorenzini
Plane of the direct lattice

Point of the direct lattice
RR vector
l’
l’’
Ph. Lorenzini
Properties of the reciproqual lattice

• Each vector of the reciprocal lattice is normal to a set of lattice
planes
 of the direct lattice
g  g1 b1  g 2 b 2  g 3 b 3 
 g . l  2 ( g1l1  g 2l2  g 3l3 )  2N
l  l1 a1  l2 a 2  l3 a 3

• and g. l
represents the projection of the vector l on the direction of
g, which has the length « ? » .
• But there are infinitely many points of the lattice with this property (for
example l’1 = l1-mxg3, l’2 = l2-mxg3, l’1 = l1+mx(g2 +g1))

g . l '  2 ( g1l1  g 2l2  g 3l3 )  2N
• We have construct one of the lattice planes
Ph. Lorenzini

• If the components of g have no common factor, then |g| is inversely
proportionnal to the spacing of the lattice planes normal to g.
• if (g1,g2,g3) have no common factors, then we can always find a
lattice vector l’’ with components such that

g. l  2 ( N  1)
''
• So d’’ can be written:

2 ( N  1) 2N 2 2
d''    d  d hkl d hkl 
g g g g
• From these two geometrical results, we see that the simplest way of
characterizing the planes of a lattice is by their normals, expressed as
vectors of the reciprocal.
Ph. Lorenzini

• Choose a plane with its normal g (a vector of reciprocal lattice)

g  hb
1  k b 2  l b3

We get g. l  2 hn  kp  lq   2N ; verified for all points belong
to this plane. If we choose a point with p=q=0, we get n=N/h, so,
N
d1  a1
the choosen plane cut axis a1 at . Same relation with
d2 and d3. h
N N N N N N En unité des
d1  a1 , d 2  a2 , d 3  a3 d '1  , d '2  , d '3  vecteurs de
h k l h k l base
3 4 5
2 2
Example : d '1  , d '2  , d '3 
2
 , ,  3,4,5  h, k , l 
3 4 5 2 2 2
We obtain the same indexes for g (normal to the plane) and for the plane !
Ph. Lorenzini
[3,1,0]
d2
[1,0,0]
d1
Ph. Lorenzini
Physical meaning of planes – importance of

reciprocal lattice
• Consider a crystal traversed by a wave :
• EM (RX, R)
• Acoustic (phonons)
• Particle wave (electrons, neutrons)
Electrons (electrons cloud) move and radiate in all

directions. This is the diffusion effect (Rayleigh scattering)
Ph. Lorenzini

reciprocal lattice
• Interferences between diffusing waves:
• destrutively  no propagation (b)
• Constructively  we can detect (« see ») them(a)
Bragg’s diffraction condition
(a) (b)
Ph. Lorenzini

reciprocal lattice
• Bragg law determination : elementary method.
• Crystalline planes behave as mirrors
• We calculate de path difference (ddm) between the two beams
• Constructively interference  ddm = n
Reciprocal vector, for ex |ON|

k
k'
plane (h,k,l)
dhkl
ddm  2d hkl sin  2d hkl sin   n

Ph. Lorenzini
Physical meaning of planes – importance

of reciprocal lattice l l’
l’’
• Bragg’s law: condition on vector of reciprocal lattice
2d hkl sin   n
2
d hkl . ON  2 d hkl  with ON vector of RL
ON

2 sin   2
ON   2

2. n and wave vector k  ,1  , k .1  sin 
ON   ON 
     
k .1 k ON k ON ON
2. 2 2   n 2k  n ON , or k  n
ON ON ON ON ON 2
Ph. Lorenzini
Unity cell / Wigner-Seitz cell: first Brillouin zone

CHAPTER 2
Wave diffraction from a crystal
Reciprocal lattice
Brillouin zone
Ph. Lorenzini
Introduction:
• The diffraction of waves (k = 2 by a crystal is present
in two domains :
• Cristallography:
• X Rays difraction:
hc hc 12,4 
E  h    A
 E E (keV )
• Neutrons diffraction
Wave – particle duality (de Broglie)
• Electrons diffraction
h p2 h2  h 12
deBroglie  E  2 e (A)  
p 2mn / e 2 mn / e 2me E E (eV )
Ph. Lorenzini
Generality:
• The objective of crystallography: studying the directions in

which the diffracted energy is maximum and deduce the
distance between the lattice planes, the # of atoms of
each of lattice planes, all this by measuring the scattered
intensity
Ph. Lorenzini
Generality:
• The second field of interest :
• Solid state physics: on utilise la condition de Bragg (maximum
d’intensité diffractée) pour déterminer les  électroniques qui ne
peuvent se propager dans le cristal (elles sont diffractées!). On
associe à chacune des ces satisfaisant à la loi de Bragg une
bande d’énergie pour l’électron qu’on appelle bande interdite
(« gap »). We use the Bragg condition (maximum diffracted
intensity) to determine the electronic wave length  that can not
propagate in the crystal (they are diffracted!). Is associated with
each of those  satisfying the Bragg's law an energy band for the
electron, called band gap (or "gap").
Ph. Lorenzini
Bragg’s law
• Elementary method (see previous chapter)
ON
2d hkl sin   n 2k  n ON , or k  n
2
Example 1: 1D net
N
RD
a
2a
RR
O N
|ON|/2 • if –a < k < /a, wave is not diffracted. This
k is the first Brillouin zone
O
• if k = ± /a, wave is diffracted, k is in the
limit of the first Brillouin zone
Ph. Lorenzini
Bragg’s law / Brillouin zone

• Example 2: 2D lattice
A2
a2
1° ZB ┴ A1 (// a2) A1
a1
Reciprocal latticeR
Direct lattice
Waves with
wavevector inside the
1° ZB ┴ A2 (// a1)
Brillouin zone can
propagate in the
crystal.
Ph. Lorenzini
Brillouin Zone of FCC crystal
http://en.wikipedia.org/wiki/Brillouin_zone
Ph. Lorenzini
Bragg law: Laue equations

• Direct method: using scattering theory for optical waves
k k
D
r
scattering : incident waves send R, k’

off spherical wave

• k, k’ : incident/diffused wave vector

• : scattering atom position 0
• r : atom/detector distance
• R: position of a detector(far from crystal)
Ph. Lorenzini

• Consider a plane wave at the position  described by
the field : 
F (  )  F0 exp i ( k .  t )

• Wave with a wave vector k , angular frequency ω and
2
wave length   k
• We put the crystal into the beam. Origin O is
arbitrarily chosen
• Hypothesis: incident beam is not affected by the crystal
It means no energy loss
Ph. Lorenzini

• Field of incident beam at  is given by :
  
ik .   it
F (  )  F0e e (3D)

• at  , we have an intercation between atoms and incident wave:

we get a scattered spherical wave . The scatterd field at point r
( the detector) is given by:
 
ik  i ( k r t )
 e e k
Fsc (r )  fF0
Scattering factor: r
fraction of incident
wave scattered by scattering : creation
an atom of a spherical wave
• The detector position is far compared to the size of the crytal

R
r  R   cos( ρ, )  R
Ph. Lorenzini

• Scaterred wave:
R
 
ik  i ( k r t )
 e e r  R   cos( ρ, )  R
Fsc (r )  fF0
r
    
 it ik .   ik R ik cos(  , R )
 fF0 e e
Fsc (r )   Fsc ( R)
R
Ph. Lorenzini

   
• We writte ik .   ik cos(  , R ) differently:
       
ik cos(  , R )  ik cos(  , k )  ik '  cos(  , k )  ik .
 
• k  is the
 scattered wave vector along the direction , with the same norm
R
than k (elastic scattering)
• Scattering relation becomes:
    
fF0e  it ik .  ik R ik cos(  , R )
e fF0e  it ei k R i k . 
F ( R) 
R
F ( R) 
R
e k  k 'k
• We now sum the signal from all the atoms on the lattice sites:
F ( R)  
p R
e 
F0 ei k R e it i k .  p
fp 
F ( R) 
F0ei k R e it
R
f
p
p e 
i k .  p

Ph. Lorenzini

• Can the amplitude of the filed be zero?
F ( R) 
F0 ei k R e it
R
f
p
p e 
i k .  p
 0 ?
• For a large number of sites (identical atoms) and this is
the case, the sum is zero if :  p .k  2n
• However, this is the definition of a reciprocal lattice vector.

thus k ( ou scattering vector) has to be a reciprocal
lattice vector G, for non zero scattered intensity.
Ph. Lorenzini
• Conclusion: the condition of a high diffraction intensity is

satisfied only when the wave vector is scattered by a
reciprocal lattice vector
Ph. Lorenzini
Laue equations (Ewald sphere)

• Elastic scattering of photons (no enery loss) |k| =|k’| .
2
k
k: // of incident beam 
k’: // of scattered beam
k’ G

 k
RR
Ph. Lorenzini
• So far we have considered that scattering factor fp is

the same for all the sites and that there is only one
atom per basis. This is not true, specially for
semiconductors. We will now generalize the conditions
for diffraction peaks:
• n for # atoms in the basis
• p for the position of unity cell
• j for the position of atom j in the unit cell
j
p
Ph. Lorenzini

• Scattering factor
F ( R) 
F0 ei k R e it
R
fp
p e 
i k .  p
 F0 ei k R e  it
R
 p j

f j exp  i (  p   j ).k 
• We describe the scattering from each of the basis j atom by a factor fj (
depending of the nature of the atom)
• The summation is separated into two summations – one over the lattice
sites (4 in FCC !) and one over the atoms in the basis (2 in the diamond !)
F0ei k R e it
F ( R) 
R
 exp i
p
p .k  f j exp i j .k 
j
Structure description Basis description

Ph. Lorenzini
F0ei k R e  it
F ( R) 
R
 exp i
p
p .k  f j exp i j .k 
j
Structure description Basis description
 j .k   j .G
 j  x j a  y j b  z j c et G  h A  k B  l C
 j .G  ( x j a  y j b  z j c) .(h A  k B  l C )
 j .G  2 ( x j h  y j k  z j l )
 p .G  2 ( x p h  y p k  z p l )
Ph. Lorenzini
Structure factor and atomic form factor

• Atomic form factor: represents how the atom scatters the X-
Rays impinging upon it. It depends essentially of electron
cloud (Coulomb charge density), so the nature of the atom.
2 differents atoms in
the basis Each node have two diffrents atoms  we have to
modify the relation
p : position of the p node in the lattice

j : position of the atom j in the basis (here 2 atoms
in the base so two differents « j » atoms).
S hkl   f j exp[i 2 ( x j h  y j k  z j l )] exp[i 2 ( x p h  y p k  z p l )]

j p
We take into account the degenerescence of the basis

Ph. Lorenzini
Structure factor and atomic form factor

S hkl   f j exp[i 2 ( x j h  y j k  z j l )] exp[i 2 ( x p h  y p k  z p l )]
j p
Altough the Bragg (Laue)

condition is verified, the
periodicity of the crystal
conducts to supplementaries
« zero conditions »(Shkl = 0 ),
more or less numerous,
depending of the crystalline
structure.
Ph. Lorenzini
X-Rays and diffraction methods.

• X-Rays:
• Electrons are accelerated
• Electrons hit a target (Cu, Mo, Al,
…)
• 2 effects:
• braking deceleration radiation
(“brehmsstrahlung spectrum”)
• Recycling of X-Rays produce by
braking  core electrons can be
excited  electrons from higher
energy levels then fill up the
vacancy and X-ray photons are
emitted
Ph. Lorenzini
X-Rays and diffraction methods.

• X-Rays:
• Discrete Spectrum : signature of the target (Cu, Mo, Al …)
Ph. Lorenzini
Experimental methods for X-rays diffraction

• Laue Method:
• monocrystal
• Polychromatic X-rays
• Immobile crystal
•To detect crystal symmetry.
•To orient the crystal

Ph. Lorenzini

• Powder method:
• monochromatic X-Rays (hkl)2
(hkl)1
Powder sample:
Thousands of small „crystals“ (1 - 10 µm)
=> Statistical orientation of crystallites
1 monocrystal  1 plane (hkl) has
=> Bragg‘s condition most likely valid for each
the « right » angle of Bragg  1 lattice plane (in fact all the planes present in the
scattered beam crystal)
=> Constructive interference of diffracted beams
Ph. Lorenzini
• Powder method:
• R: radius of the Debye –Scherrer chamber

• W= R : half perimeter of the chamber
S1  2 R1
Measure S1 allow to determine the Bragg angle
Ph. Lorenzini
• Powder method:
X-ray detector
Intensity (c/s)

Ph. Lorenzini
• Powder method:
• Function of the structure, particular peaks are present or not

 we can determine the crystalline structure.
• Peak intensity is also an important parameter.
• Allow to determine lattice parameter
• It works also in mixed powder (mixed crystal)

CHAPITER 3
Wave mechanics :
Schrödinger equation
Ph. Lorenzini
The end of the classical mechanics?

• mater: • light:
• Position • Electric field
• speed (momentum) • Magnetic field
• Newton law • Maxwell relation
Six variables can perfectly we can not transform the radiation

define the state of each particle into particles maintained localized
in space  interference and
diffraction process
Ph. Lorenzini

• In the early 20th century :
• Particle theory of the matter:
• « works » even at microscpics scale
• But very complex at this scle  statistics mechancs
• Wave theory of ligth:

• Accepted since Fresnel work
• All known phenomena of light are interpreted
• Maxwell relation + radio waves (Hertz)  optical and electricity synthesis
Ph. Lorenzini

• New experiences / discoveries pose problems of interpretation:
Black body radiation
Inside the cavity, the electromagnetic field is

equivalent to a set of independant harmonic
oscillators. Their mean energy (assuming that the
energy can vary continuously) is given by :
/
kT
/
Black body: object which absorbs all

impinging radiation
Ph. Lorenzini

Black body radiation
M. Planck : He postulates that the energy exchange between
matter and radiation are not made continuously but by
indivisible and discrete quantities: quanta of energy. He shows
that the relationship between energy and frequency is given by:
En =n.h
M. Planck : energy and frequency are dependent  allowed
modes are more and more difficult to populate. In the cavity, (
the black body), atoms in the wall act as oscillators with a
frequency and the atom’s energy is given by:
En =n.h
∑ /

∑ / / 1
Ph. Lorenzini

• photo électric effect
The remarkable aspects of the

photoelectric effect when it was first
observed were:
1. The electrons were emitted immediately - no time
lag!
2. Increasing the intensity of the light increased the
number of photoelectrons, but not their
maximum kinetic energy!
3. Red light will not cause the ejection of electrons,
no matter what the intensity!
4. A weak violet light will eject only a few electrons,
but their maximum kinetic energies are greater
than those for intense light of longer
wavelengths!
A. Einstein : quantification of ligth ennergy (photon)  nh , ligth is like a

grain (ball) of light  corpuscular aspect !!
Ph. Lorenzini

Ph. Lorenzini

• Wave – particle duality: particle nature(photoelectric effect) of light is accepted
but its wave nature is also accepted (double slit experiences - Young)!
• The ligth is composed with a quatum, a quantic particle wich can not be
described only by the new mechanics: the wave mechanics is born.
• Dependint of experience, we can use the wave descripton of the light ( light
propoagation) or particle aspects (light – particle interaction).
Ph. Lorenzini

• Louis de Broglie: new concepts of waves of matter. (1924)
• Hypothesis: wave particle duality is a general propertie of microscopics objects
and the matter, as light, shows in the same time, wave property and particle
property.
• This hypothesis is confirmed few years later by the diffraction of electrons.
(1927).
Waves of matter de Broglie wavelength :
h h
 
mv p
Ph. Lorenzini
Wave mechanics
• The relation of de Erwin Schrödinger:
Ph. Lorenzini
Wave mechanics
• the relation of Schrödinger:
the motion of a particle is described by a wave function (r,t) and
the norm squared of this function represents the probability of
presence or density on a point r. the functions (r,t) obey to the
Schrödinger’s relation:
2  (r , t )
  (r , t )  V (r , t ) (r , t )  i
2m t
• (r,t) is the wave function (can be a complex quantity)

• V(r) is the potentiel energy applied to the particle
• m is the mass of the particle
 The relation of Schrödinger:

 Taking into account the quanta notion ( Planck) but also the wave-
particle duality ( de Broglie)
Ph. Lorenzini
Wave mechanics
• One particular point for this lecture is concerning steady state (stationary
state) where the potential energy V=V(r) applied to the particle is time
independent. We have to determine the time-dependent portion and the
position-dependent by using the technique of separation of variables. We
assume that we can write the wave function:
 (r , t )   (r )  (t )
• r) is only position dependent
• t) is only time dependent
• Susbtituting this form of the solution into Schrödinger’s wave equation, we

obtain (one dimension approximation):
 2 1  2 ( x) 1  (t )
  V ( x )  i
2m  ( x) x 2
 (t ) t
Ph. Lorenzini
Schrödinger equation:
 2 1  2 ( x) 1  (t )
  V ( x )  i
2m  ( x) x 2
 (t ) t
g(x) f(t)
• Since the left side is a function of position x only and the right side is a
function of time t only, the only solution is that the both terms are equal to a
constant. On the other hand, we see by dimensional analysis of the right side
that the constant is an energy:
E
1  (t ) i t
i E  (t )  Ae 
 (t ) t
• Avec E, total energy of the particle: E  

Ph. Lorenzini
Schrödinger equation:
• The time independent part of Schrödinger equation can be writen
as:
 2  2 ( x)
  V ( x) ( x)  E ( x)
2m x 2
 iE
• with :  ( x, t )   ( x) exp( t)


  ( x) dx  1
2
• Limits boudaries:

• Continuity of the wave function
• Continuity of the first derivative

Ph. Lorenzini
Schrödinger’s equation:
• At this stage, it is important to distinguish the

concept of wave state (plane wave) and
combination of wave states (wave packet)
Ph. Lorenzini
Plane wave/ wave packets

• Plane wave:
ikx  it
Ae e
• Wave packets:
k 0  k / 2
 e dk
ikx it
 ( x, t )  Ae
k 0  k / 2
• We will show that only the waves packet can be a good

representation of a particle
Ph. Lorenzini

• We suppose a particle are travelling in a constant potential
energy V(x) =V0 = cte. It is subjected to a force given by a
F = - grad V0 = 0. it means that its motion is uniform and its
energy E is a constant. We’ll now:
• Find simple solutions of the E.S (plane wave)

• Verify that these solutions have a physics reality
• If not, we’ll try a linear combination (wave packet) of
simple solutions (note that Equation of Schrödinger is a
Linear equation with )
Ph. Lorenzini

 2  2 ( x)
  V0 ( x)  E ( x)
2m x 2
E
 2 ( x) 2m( E  V0 )
  ( x)  0 , E  V0  0
x 2
 2
V0
One simple solution can be found (x):

 ( x)  Ae ikx k
2m( E  V0 )  2k 2
k E  V0
with  2 or 2m
The question is "Can we represent a particle by this function?". You

only have to calculate the probability of presence in space (or in 1D)
of this particle to answer this question
Ph. Lorenzini

• Plane wave:
 ( x, t )  Aeikx e  it
 ( x, t )   ( x, t ) ( x, t )   A  cste
2 * 2 2
Prob Expected result ( x = vp.t )

Of presence
|A|²
t0 t1 time
• the probability of presence is independent of x and time, which is far from being
"reasonable" from a physical point of view
Ph. Lorenzini

• So we will try a linear combination of plane waves centered
on k1 ie a wave packet from k1 –k/2 and k1 + k/2 . The
wave function can be written E
k1  k / 2
 Ae
ikx it E1
 ( x, t )  e dk V0
k1  k / 2
k
• We suppose A=cste (hyp) k1 –k/2 k1
• at t=0:
k1  k / 2 k / 2
 
ik1 x i ( k  k1 ) x i ( k  k1 ) x
 ( x,0)  Ae e dk  Ae ik1 x
e d (k  k1 )
k1  k / 2  k / 2
kx )
 
1 k / 2 sin(
 ( x,0)  Aeik1x . e i ( k  k1 ) x  Aeik1x k 2
ix
 k / 2
k
x
2
Ph. Lorenzini

• We calculate again the probability of presence at x:
sin 2
z k  x
 ( x,0)  A k avec z 
22 2
z 2
2
2 4
x1    x  2 x1 
1,0
k k
0,8
 x  k  4
2
0,6
[sin(z)/z]
0,4
0,2
By improving the law of amplitude (instead
of A=cte) , one would get the Heisenberg’s
equation:
x  k  2
0,0
-3 -2 -1 0 1 2 3
z ( in unity of  )
Ph. Lorenzini

• At t=t1:
k1  k / 2 k / 2
 
i k t1 ikx ik1 x i1t1 i ( k  k1 ) x i ( k 1 ) t1
 ( x, t1 )  Ae e dk  Ae e e e d (k  k1 )
k1  k / 2  k / 2
 d 
Taylor series : k  1   k  (k  k1 )  ...
 dk  k1
k / 2  d 
i ( k  k1 )  k  t1

 dk  k1
 ( x, t1 )  Aeik1x e i1t1 e i ( k  k1 ) x
e d (k  k1 )
 k / 2
k / 2
 d 
we take x'  x   k  t1   ( x, t1 )  Aeik1x e i1t1  ei ( k  k1 ) x d (k  k1 )
'
 dk  k1  k / 2
k ' k ' k '

x  x sin x
ik1 x i1t1 e 2
e 2
2 k ' k 
 ( x, t1 )  Ae e   Bk avec x  x t1
ix ' k ' 2 2 2
x
2
Ph. Lorenzini

• we get the same result when t = 0 but for x’ = 0, ie
 d 
x  t1
 dk  k1
dP
x0 (t=0) x1 (t=t1) x
The velocity of the particle = vg = group velocity which is calculated at the

center of the wave packet:
 d 
vg   
 dk  k1
Ph. Lorenzini
Phase velocity/ group velocity

• In the case of a plane wave, the  reasoning on the wave packet:
velocity is the propagation
velocity of planes in phase or
phase velocity :
 2k 2
 2 2  E  V0
v  (k    ) 2m
k  v T v
 dE   d   2 k1 p1
hk h         v g 
from de Broglie ( p    k )  dk  k1  dk  k1 m m
 2
p
E  We get v 
ω  E E ²k ² m
v     or E   V0
k p p k 2m

this term should not exist (!)
1 k V0 1 p V0 V =cst
 v     0  no action on the motion of the
2 m k 2 m p particle; absurd because v=p/m
Ph. Lorenzini
Wave Equation: applications of

Schrödinger’s equation
• Infinite potential well and width a:
• Classical representation of a bound particle.
 2 ( x) 2m 2mE
 ( E  V ( x )) ( x )  0 we set k 
Particule x 2 2 2
with energy
E  Region II:  (x)=Acos(kx)+Bsin(kx)
 Region I et III: (x)=0
 Continuity at x=0 and x=a
 A=0
 Bsin(ka)=0  k=n/a
 probability between 0 and a must be 1:
2 2
 B=  ou  i
a a
Ph. Lorenzini

• Infinite potential well and width a:
2  nx   2 n 2 2
 ( x)  sin  with n  1, 2, 3, ... E  En  with n  1, 2, 3, ...
a  a  2ma 2
Particule
with energy
E
Ph. Lorenzini

• The step potential function:
• Region I:
 21 ( x) 2m V(x)
 2 E1 ( x)  0 Incident
x 2
 Particules
• Region II (0<E<V0)
V0
2mE
 1 ( x)  A1e  B1e ikx ikx
avec k 
2 Region I Region II
 2 2 ( x) 2m
 2 (V0  E ) 2 ( x)  0
x 2
 x=0
x  x 2m(V0  E )
  2 ( x)  A2 e  B2 e avec  
2
Ph. Lorenzini

• The step potential function
V(x)
• A2 =0 (divergence) Incident
• Reflexion coefficient: Particules
Flux of reflected
particles (cm-2 s-1) 2
V0
vr B1 B1* vr B1
R 
vi A1 A1 vi A1
*
Region I Region II
• In region I, V=0  energy is only

x=0
cinetic energy
2m  1  mv 
k  mv ²  vi  vr  k
m
 2  
R 1
2
! x)≠0  particle(in contradiction with classical mechanics) can be present

Ph. Lorenzini

• The Potential barrier (E<V0):
2mE
1 ( x)  A1e  B1e
ikx ikx
with k 
2
2m
 2 ( x)  A2 e x  B2e  x with   2
(V0  E )

3 ( x)  A3eikx  X
B3e ikx
• Transmission coefficient :
vt A3 A3* A3 A3* 4 E (V0  E )

T
vi A1 A1
*
 
A1 A1 4 E (V0  E )  V0 sh 2m(V0  E ) a / 
* 2 2

0

It’s tunneling or tunnel effect (T~78% with 1 alectron, E=V0/2=1ev and a=1Å)
CHAPTER 4
Free electrons:
model of Sommerfeld
Ph. Lorenzini
Free electron
• Free electron energy potential is zero everywhere V(x)=0:
 2 ( x) 2m
 2 E ( x)  0
x 2

• Solution of this differential equation is given by:
 i 2mE   i 2mE 
 ( x)  A exp  x   B exp  x
     
• And :
E i   i 
 ( x, t )   ( x) exp(i t )  A exp  ( x 2mE  Et )  B exp  ( x 2mE  Et )
     
Ph. Lorenzini
Free electron
E i   i 
 ( x, t )   ( x) exp(i t )  A exp  ( x 2mE  Et )  B exp  ( x 2mE  Et )
     
+x -x
• wave propagation  a particle in the vacuum can be represented by a

wave: this is normal, we started with Schrödinger’s equation!
• We suppose for simplicity taht B=0  incident wave only in the positive x
• Wave vector: k  2 
• wavelength: h 2mE
• Energy: E   ² k ² 2m E(k) represents a parabole
• Probability density: cste =AA*  in accordance with Heisenberg (k known => x

unknown)  physics reality wuth wave packets
Ph. Lorenzini
Sommerfeld Model
• A. Sommerfeld considers one unidirectional crystal (1D) with length
L
• The electrons are bound to the crystal by the Coulomb attraction
forces.
• The electrons are trapped  they can not get out
crystal
V0
1 2 3
-L/2 0 +L/2
Ph. Lorenzini
Sommerfeld potential
• Simplification: we get closer to reality by considering that the wave
function associated with the electron cancels to the limits (± L / 2)
 electrons can not escape from the crystal
• steps:
• Find solutions of Schrodinger’s equation
• Keep only those that meet/fit conditions of continuity
• Built a wave packet for a better representation of reality
Ph. Lorenzini
• Solution for region 2 (idem free electron):
 2 ( x) 2m 2mE  2k 2
 2 E ( x)  0  ( x)  Aeikx k E
x 2
 2 2m
• Continuity conditions: E
• real: cancel in ± L/2
Aeikx  cc Aeikx  cc V0
2 2
cosinus
 
sinus
A cos(kx) A sin( kx)
• Cosinus solutions
L  2 k
cos k  0  k  (2 p  1)  k 
2 L L
• Sinus solutions Wave vector is quatified
L 2 2  Energy too !!
sin k 0k  p  k 
2 L L
Ph. Lorenzini
• :Born–von Karman boundary condition (BVK condition)
• The idea is to consider that at the scale of electronic wavelength, the
crystal size is infinite  we assimilates the one-dimensional crystal of
length L to a circle of perimeter L: we "forget" the boundary conditions
• The position at x and x+L is the same:
e ikx  e ik ( x  L )
e ikL  1  kL  p  2 We get the previous result : k
and E are again quantified
2
k  p
L
Ph. Lorenzini
• Sommerfeld Electron: E
• Suppose electron at k0: the order of
L is 108Å (1 cm)  we assume that
the wave packet with a width k can
be considered as a continuum
because: 2 2 
k    10 7 /A
L 108
• Electron velocity: k0 k
 d   2k 2 k0 –k/2
vg    and E    k0 +k/2
 dk  k0 2m
k 0 p 0
 vg  
m m
Ph. Lorenzini
Sommerfeld potential E
• Sommerfeld electron:
• Electron acceleration:
dE
+
-
Velec
k0 k
V(x) = -qVe k0 –k/2

k0 +k/2
The state (electron) gains energy:
 dE  1  dE 
 between t and t+dt =>dE=F.vg.dt dE  F  v g  dt    dk  F   dt  dE
 dk  k0   dk  k0
 IfiE changes, k must change  the
central value of k0 is changed (-).
dk0 dp0
F  
dt dt
Ph. Lorenzini
• Sommerfeld electron:
• Electron Accélération :
If we define:
F dv g
  and F   gradV0  Fext 1 1  d 2E   2k 2
m dt  2  2  because E   V0
(=0, V0=cste) -qEelec m   dk  k 2m
0
dv g  d  1  dE 
 qEelec and v g      
dt  dk  k0   dk  k0
dv g F 1 1  d 2E 
dv g dv g dk   avec   
  dt m m  2  dk 2  k
dt dk dt 0
1  d 2 E  dk 1  d 2 E  F
   2   2  2  F 
  dk  k dt   dk  k m
0 0
dk0
F 
dt
Ph. Lorenzini
dvg F 1 1  d 2E 
  avec  2  2 
dt m m   dk  k
0
• Nota: the relationship F = mg is always true. The calculated mass is the

mass of the particle only if that V = V0 = constant, ie that the resultant
forces are equal to the external forces. For a potential V (x) which is not
constant, the relationship is the same if we replace the mass by an
effective mass (m*) that will reflect a response different ( inertia different
due to potentiel energy) compare to a free particle (see below!)
Ph. Lorenzini
Wave theory applied to atoms: the Bohr – Sommerfeld

atom
• Atom with one eletron (Hydrogen):
• nucleus « heavy », positively charged (proton)
• It is immobile
• Electron is « ligth », negatively charged
• Coulomb attraction between electron and proton is given by:
• e : electronic charge e2
V (r )  
•   permitivity of free space 4 0 r
Leads to a 3D problem of Schrödinger equation

Ph. Lorenzini
Bohr – Sommerfeld atom

• 3D Schrödinger equation (sperical coordinates)
2m0
 2 (r ,  ,  )  2
( E  V (r )) (r ,  ,  )  0

• and:
1     1  2
 r²  
r ² r  r  r ² sin  r ²
2
1     2m0
 sin   2 ( E  V (r ))  0
r ² sin      
Ph. Lorenzini

• We apply the separation of variables technique:
 (r , ,  )  R(r ).( ). ( ) Function of 

• We get: only
sin 2    R  1  ² sin     
 r²    sin  
R r  r    2
    
2m0
 r ² sin ² 2 ( E  V )  0

• We may writte
1  2 
1 im
avec m  0,1,2,...
 m² e
  2 2
Be careful : m is not the mass, is a constant obtained via separation of variables technique
Ph. Lorenzini

• If we proceed the solving (irrelevant for this course) for the two other
function, we have to introduce two additional constants l and n. m, l
and n are known as quantum number:
• n= 1, 2, 3,…  principal quantum number
• l= n-1, n-2,…, 0  azimuthal quantum number
• |m|=l, l-1, …, 0  magnetic quantum number
• Main result:
• Each set of quantum number corresponds to a quantum state
which electron can occupy. We get the well known quantified
energy of hydrogen
m0 e 4 Ry 13,6
En     2   2 eV
(4 0  ) 2n
2 2 2
n n
http://scienceworld.wolfram.com/physics/HydrogenAtom.html
Ph. Lorenzini

• We can show that for the lowest energy state (n=1, l=0 m=0), the
wave function is given:
3/ 2  100  100
*
1 1 r
 100    e a0
  a0 
4 0  2 0
a0   0,529 A Bohr radius
m0 e 2
• Common writing:
• s state l=0
• p state l=1
• d state l=2
• Electronic states:
• 1s 2s 2p 3s 3p 3d …
Ph. Lorenzini
Sommerfeld model
• conclusion:
• Specific heat explained
• Ohm’s law explained (see futher)
• …
• We can’t explain the difference between the conductivity of an
insulator and a metal
 metal
 1032 !!!
 insulator
There must be "somewhere" in the model a rough

approximation!
CHAPTER 6
Electrons in periodical structure :
Bloch – Brillouin model
Ph. Lorenzini
plan
• Formation of energy bands
• Periodicity of the crystal
• Pseudo wave vector or crystal wave vector k
• Bloch waves
• Kronig-Penney model
• Energy band diagram in k space
Ph. Lorenzini
Formation of energy bands

• Qualitative model:
1 isolated
hydrogen
atom
2 adjacent
hydrogen
atom
Ph. Lorenzini

• :Qualitative model for a Silicon crystal
• Electronic structure (14 electrons) Overlapping
• 1s2 2s2 2p6 3s2 3p2
Splitting of energy levels
3p2
3s2
Ph. Lorenzini
Energy
band
3s2 3p2 Antibonding states (4N)
N atomes Conduction
3p2 band
Forbidden band
Valence
3s2 band
Bonding states (4N)

Ph. Lorenzini

• The value of the forbidden band depends of the equilibrium
distance between atoms: a0
Ph. Lorenzini

• Theoritical model using quantum mechanics and Schrödinger
eqaution.
• What we have to add /modify to be able to explain/understand the

notion of allowed and forbiden energy bands?
In fact, the crystaline potential where electron is moving is

not a constant but it’s a periodic function !
Ph. Lorenzini

a
The crystaline potential (energy) is

periodic, with a periode « a
V ( x)  V ( x  a)
(from Neaman)
Ph. Lorenzini
Bloch Brillouin model

• BVK condition: (allows you to ignore what happens at the end of the
sample with a length L)
• Number of Atoms?
• 1D:
1
N  10 7 cm 1 a
• 3 D: a
13 L=N x a
N  3  10 21 cm  23
a
V ( x  L)  V ( x)
 ( x  L)   ( x)
The crystal being periodic with a period a, the crystal properties are the same at x
and at x + a. It is the same for the wave function, and also for the density
probability for the electron.
Ph. Lorenzini

dP ( x)   ( x)
2
(1)
dP ( x  a )   ( x  a )
2
(2)
(1)  (2)   ( x) * ( x)   ( x  a ) * ( x  a )
• The two wave functions can be separated by a phase

factor:
 ( x  a )  ei ( x) ,  ( x  2a)  ei 2 ( x), ...

  ( x  Na )  e iN ( x)   ( x)
iN 2 2a 2
e 1    p p p a
N Na L
Ph. Lorenzini

• modification of writing : introduction of pseudo wave vector or crystal
vector K.
2
• We introduce K  p , so we get:
L
 ( x  a)  eiKa ( x)
• K is not a real wave vector, it only represents a phase shift between two
waves
• K is quantified. It takes N values separated by2 and has the same values
2 L
each
a
Ph. Lorenzini

• Bloch waves:
 ( x  a)  eiKa ( x)  eiK ( a  x ) e iKx ( x)

set u ( x)  e iKx ( x)   ( x)  u ( x)e iKx
and
u ( x)  e iKx ( x)  e iK ( x  a ) eiKa ( x)  e iK ( x  a ) ( x  a )
 u ( x)  u ( x  a )
• Stationnary wave functions for an electron in 1D periodic

potential must be of the form:
 ( x )  u ( x )e iKx
x with u ( x)  u ( x  a )
Ph. Lorenzini

• Difference plane wave/ Bloch’s wave:
• Plane wave:
2 2
 ( x)  Ae ikx
k and   x
 
• Bloch’s wave:
2
 ( x )  u ( x )e iKx
Kp and   Kx  arg(u ( x))
L
unknown
Ph. Lorenzini

• The Bloch’s function has to be injected into the Schrödinger’s equation.
 for Solving this equation we need to determine u(x) !
2 2
 (u ( x ) e ikx
)  V  u ( x ) e ikx
 E  u ( x ) e ikx
2m x 2
0
• To solve the problem we have to know perfectly the form of the potential
energy ( not trivial !). A simplified model could give us an idea
concerning energy band structure.
Ph. Lorenzini
Kronig – Penney model

Ph. Lorenzini

• We approximate the periodic potential by a simpler model.
III
• Zone I : 0<x<a V(x)=0
• Zone II: -b<x<0 V(x)=V0
Ph. Lorenzini

• Periodic potential Bloch waves as solutions
• The functions( and the derivatives) only differ by a factor of phase
in region II and III
 III ( x)   II ( x  a)e ika
• Bound electrons in crystal E<V0

• Potential V(x) exists everywhere (everywhere finite)  wave
function and its derivative have to be continue everywhere.
Ph. Lorenzini

• Solving the Schrödinger’ Equation:
• Region I (the well):
i x  i x 2mE
 I ( x)  Ae  Be avec  
• Region II (the barrier): 2
x x 2m(V0  E )
 II ( x)  Ce  De avec  
2
• Continuity of the wave function (and its first derivative) at a and -b: be
carreful of the phase shift at -b
Ph. Lorenzini

• We get now four homogeneous equations ( 2 for the wave function
continuity and 2 for the first derivative continuity) with 4 unknows
(A,B, C et D)
• Nontrivial solution ( ie A=B=C=D=0) if and only if the determinant

of the coefficients is zero
• We get (laborious, tedious but feasable !!)
2 2
cos k (a  b)  sin( a ) sinh(b)  cos( a ) cosh(b)
2
Reminder !! : k is a pseudo wave vector (crystal vector)!!
Ph. Lorenzini

• And now ?
2 2
cos k (a  b)  sin(  a ) sinh(b)  cos( a ) cosh(b)
2
• Graphics or digital solving to obtain a relation between E, V0 and k.

• To be more realistics and to get more easely a « visible » graphics
solution, let the potential barrier width b=>0 and the barrier height
V0=>∞ but with the product bxV0 which remains finite
sin a mV0ba
P  cos a  cos ka avec P 
a 2
Ph. Lorenzini

• Plot E(k)
sin a
P  cos  a  f ( a )  cos (ka)
a
f(a)
 2n
a
Ph. Lorenzini
Plot E(k) 
2mE
E
2 2
 2
2m
a
sin a
• ka=0  1 P  cos a
a
We deduce a value of  thus of
energy E1.
• ka=  a=
We deduce a value of  =a thus
of energy E2.
2mE2 

 2
a
Ph. Lorenzini
Energy bands
• As the partcle is propagating in a periodic potential (crystal potential), we have
to introduce the notion/concept of allowed energy bands, separated to each
other by forbidden energy band ( ) where we can not find particles.
 Particles (electrons) are

localised in allowed energy
band
 As energy increases, the
width of the allowed energy
band increases
 We can plot E(k) only in a
reduced zone k:
 2n periodicity
cos ka  cos(ka  2n )  cos(ka  2n )
Ph. Lorenzini
Reduced zone representation
(From Neaman)
Concept of direct band gap : when the minimum of the upper band and the
maximum of the lower band are localised on the same k value, we speak about
direct band gap. This arrive in the center of zone (k=0) and in the edge zone
(k=±/a ). Remind: this model is a 1D model, so a simpler model than the real
model, but it’s enough to introduce the energy band concept.
Ph. Lorenzini
E(k) near an extremum

dE
• Curve Sketching : ?
dk 0, 

a
• E and k are not directly related!
dE dE d sin a
 ??? butP  cos a  cos ka
dk d dk a
OK ! sin a
d[P  cos a ]
d [cos ka] d [cos ka] dk a
 
d dk d d
dE
We can show taht at center and at the edge of Brillouin zone, 0
dk 0, 

a
Ph. Lorenzini
Effective mass
• Near the bottom ( or top) of an extremum: (Taylor’s serie , second
order)
dE 1 d 2E
E k  E ( 0)  ( k  0)  ( k  0) 2
dk k 0 2! dk 2
k 0
=0
1 d 2E  k
2 2
1 1 d 2
E
E k  E( 0 )  k 2  E( 0 )  with  2
2 dk 2 k 0
2 m * m *  dk 2
(0)
Within a certain range around the band

extrema, we have the right to confuse the true
curve with a parable: it is the approximation of
the effective mass
Ph. Lorenzini
Effective mass
• The effective mass is a parameter that relates the quantum
mechanical results to the classical results).
• the idea is that when an electron is moving in a non constant

potential (ie periodic potential) its inertia is modified and for
continue to use the « classical » concept, we have to use this
modified mass:
dvg F 1 1  d 2E 
  avec  2  2 
dt m* m *   dk  k
0
Ph. Lorenzini
Sign of the effective mass

• Bottom :
 2k 2
E k  E( 0 ) 
2m1*
• Ek >E(0)=Eminimum
1st rule: the effective mass at the bottom band is a positive parameter
 top:
 2k 2
Ek  EMax 
2m2*
2nd rule: the effective mass at the top
band is a negative parameter
Ph. Lorenzini
Density of states ( mass effective

approximation
• currentcharges moving
• Need to count the number of charges in the bands
(conduction or valence)
• This number is a function of allowed energy states in the
bands
• Warning: we have to take into account Pauli exclusion
principle
Ph. Lorenzini
Density of states (mass effective

approximation)
• Before continuing:
• Concept of electron spin:
 (r , ,  )  R(r ).( ). ( )

 In addition to the three quantum numbers
(principal, magnetic and azimuthal), the electron
has an additional property which results by an
intrinsic angular quantized momentum (± ½) to
which is attributed a fourth quantum spin number,
s, independent of the other 3.
 Principe d’exclusion de Pauli (dans les solides):
we can not find two electrons in the solid with the
same energy state (same quantum numbers).
Ph. Lorenzini

approximation)
• Unidimensional crystal with a length L:
 allowed states are quantized and électrons

2
separated by k 
L
 near the minimum, on approxime E(k) by:
 2k 2
E  Emin 
2m *
 The quantum states density is the
number of modes (states, places) per
unit of energy and by unit length (L=1!): électrons
dN 2m * 1
g1D ( E )   ( E  Emin ) 2 J -1m 1
dE 
Ph. Lorenzini

approximation)
• Crystal 2D of dimensions Lx et Ly:
« Surface » of an (2 )
2
 Allowed states are allways quantized.
elementary state Lx Ly
 Energy at the edge of the band (
conduction or valence) is approximated
by:
 2k 2
E  Emin 
2m *
 The 2D states density is now given by:

dN m * -1 m * -1  2 No Energy
g2D (E)   Lx Ly 2 J ou 2 J m
dE   dependent!
Ph. Lorenzini

approximation)
• 3D crystal with Lx, Ly and Lz for the dimensions:
 Allowed states are still quantized. kz

 Energy on top and bottom is given by: (2 ) 3
2
 2
 k 2 Lx Ly Lz
Ek  Emin  (k x  k y  k z )  Emin 
2 2 2
2m * 2m * ky
 The density of states is given by ( with
Lx x Ly x Lz=1 ie per volume): kx
3/ 2
dN  2m* 
g 3 D(E)   4π  2  (E  Emin )1 / 2 J -1m -3 bottom
dE  h 
top
= (Emax  E)1 / 2
Ph. Lorenzini
3D states density
Emax
??
Emin
  g(E)
a a
The effective masses are not the same and the
approximation of m * only applies on band edge!
Ph. Lorenzini
Real crystal band structure!
Si
Eg
(0 0 0.85)
(111) (000) (001)

Ph. Lorenzini
Real crystal band structure!

Ge
Eg
GaAs
Ph. Lorenzini
synthesis
• Crystal  periodic potential
• introduction of energy band concept
• Introduction of forbidden energy band
Is it enough to explain the difference between metal

and insulator (or semiconductor)?
To be continued in the next chapter
CHAPITRE 7
Current in the solids:
Specific case of semiconductors
Ph. Lorenzini
Conduction current in the Bloch – Brillouin

crystal
• Density of current vector:
• Current density ? It’s a moving of electric charges per time and per
surface
j  q  ni vi  q  vi
i i
with ni number (1D density) of electrons with velocity vi
• At 1 dim:
j  q  vi
i
Ph. Lorenzini

crystal
• Contribution of the various bands to electric current
• Full band
 On each allowed values of k0 we can form a
E wave packet
 The band is full if all allowed states of k0 are
occupied (all energy levels are occupied).
 E(k) is even with k E(-k0)= E(+k0)  the
probability of find the electron at k0 or at –k0
is the same (see next chapter)
1  dE 
 The velocity vg    dk  of the particle is odd
k
with k vg(k)=-vg(-k)
0
E0
-k0 k0 j  q  vi  0 (the band is full)

i
Ph. Lorenzini

crystal
• Rule 1: the contribution to the current for a full band is allways zero.
• Rule 2: the contribution to the current for an empty band is of

course zero.
• Corollary : the electric current in only due to partially filled band

Ph. Lorenzini
Contibution with a partially filled band

• Perfect crystal, elec = 0 and uniform Temperature.
• Thermodynamical equilibrium
• Occupancy probabilities of the energy states k and -k are identical
j  q  vi  0 [v g (k)  vg (  k)]

i
• Not surprising results : no electric field  no conduction current!

Ph. Lorenzini
Contribution with a partially filled band

• Perfect crystal, elec ≠ 0 and uniform Temperature.
• At time = 0, we apply an electric field elec.
• A wave packet (in fact an electron) wich « had » before k0 undergoes
an acceleration acceleration
dk0
   q elec
• At time >0,
dt

k (t )  k0  q t
'
0

k
• All electrons undergo the same variation k

 k0   q   t ∀k0

Ph. Lorenzini

• Perfect crystal, elec ≠ 0 and uniform Temperature.
k k
t=0 t=t1>0
k0 q
k0  t1

Now all velocities are not
balaced/compensated! J (t1 )  0  0
Ph. Lorenzini

• Perfect crystal, elec ≠ 0 and uniform temperature.
k k
t2>t1 t3>t2
J (t 2 )  0  0 J (t3 )  0  0 !!!
Strange ! Electric field is constant and current is alternatively
changing (its sign), function of the time !!
Ph. Lorenzini

• in the Bloch – Brillouin model, we predict that a constant applied
field elec must generate an alternative electric current .
• La période de ce courant serait:
dk 2 1  2
T a
   q    q
dt a T q
• This result is not coherent with experimental results !!!!
• The crystal is not as perfect as it!!

Ph. Lorenzini
The real crystal in the Bloch-Brillouin

model
• Real crystal, elec ≠ 0 and uniform Temperature.
• The crystal has a number of defects main class:
• Chemical impureties
• Intrinsic point defects :
• Vacancy
• Interstitials
• Antisites
• Défauts étendus
• Grain boudaries
• Stacking defects
• Dislocations
• Clusters
• Lattice vibration/oscillation around equilibrium position (phonons)
Ph. Lorenzini
The real crystal in the Bloch-Brillouin model

• Real crystal: consequence on charge carrier:
• The BB potential is no more exactly periodic potential:
• B-B v1(x)=v1(x+a)
• Réel v(x)=v1(x) + v(x) random!
• Electron in this potential experiences a force (-gradv(x))
dk0
  q  grad (V ( x))  q  random force
dt
• This force gives random pulse(similar to forces during impacts with « impureties »)
• On average, every  (average time between subsequent collisions), the electron

changes direction and especially loses its kinetic energy  therefore acquires an
average "excess" wave vector :
velocity
velocity

k  k0  q 
'
0 t
time   time 
Ph. Lorenzini
The real crystal in the Bloch-Brillouin model
• Real crystal,  elec ≠ 0 and uniform Temperature.
k k
k0 k0 k’0
time=0,  elec=0 
time≠0 , elec≠0

k  k0  q 
'
0 t

Ph. Lorenzini
Real crystal in Bloch-Brillouin model

• Almost empty band:
• Energy and group velocity:
k
E k 0  Ec 
 2 k02
vg 
1  dE 

 
  k k0'  k0  q 
•  :
* 0
2m1*   dk  k0 m1 

j  q  v gi  -q k i 0
m1*
• ≠  :
i i
i

j  q  v gi  -q *
m1
' 
i k i  -q m*
1
x   elec
i ki  q m*  n
2

j  qµ n n elec with µ n  q 0
m *
1 It’s Ohm’s law
j   n elec Mobility (cm²/Vs)
Ph. Lorenzini
Real crystal in Bloch-Brillouin model
• Almost full band:

• trick: to calculate the electric current in an almost full band, we add the missing
electrons to fill the band and then we remove them: which will allow us to
introduce the concept of holes. E
• N : number of places (~ number of atoms)
• N-p: electrons on the band
• p: empty places (states) or number of states
Np N p
j  q  vgi   q  vgi  (q  vgi )
i 1 i 1 i 1 Positive
p charge!
 0  q  vgi
Full band i 1
Ph. Lorenzini
E
Real crystal in Bloch-Brillouin
model
k
• Almost full band:
• Rule: la contribution au courant d’une bande presque pleine est obtenue en

supposant que les états vides sont effectivement occupés par des porteurs de
charges positives +q: des trous. Il suffit de sommer sur le nombre total de trous
• Trous: paquet d’ondes de charges positives
Ph. Lorenzini
E (b)
Concept of the Hole (a)
• Drift current:
 Nota: to “extract" an electron in (a)

k
requires more energy than (b) a
hole in (a) is more energetic than in Eholes
(b).
 k
2 2
k 0
 E(e ) = -E(h )  Eholes   EV 
- +  vg 
holes
2m * m*
 elec=0  J  q   k0   0
 
m* 
 p   elec  q   elec
2
 elec≠0  J  q  *  0 k  q   * p
 m2 i 1    m2 
q
J  q  p  µ p   elec with µ p 
m2*
Ph. Lorenzini
Concept of hole : conduction

Ph. Lorenzini
Bipolar (electrons and holes) conduction

• Common case in semiconductors :
• In practice, in the semiconductor, there is an almost

empty band, called conduction band (CB), over a
nearly full band called the valence band (VB). They
are separated by a forbidden energy band ("gap or
bandgap"). The current is the sum of the two
components
J  qnµn  qpµ p  
BC BV
Ph. Lorenzini
Semiconductor or insulator?
• Same energy band diagram? The only difference is the value of the
bandgap. If Eg larger than 3.5/4 eV  insulator
Isolant , semiconducteur semiconducteur
T= 0 K T≠0K
Ph. Lorenzini
Metal ?
• T=0K ou T≠ 0K
even at 0K, a partially filled band exist.this band conducts. An

other way to get a metal is the existence of overllaping allowed
energy bands (semi metal)
Ph. Lorenzini
Summary:
• Metal:
• Very low resistivity
• When the temperature is lowering, the conductivity increases
• Insulator:
• No conduction even at high temperature
• Bandgap higher than 3.5 eV
• Semiconductor:
• Bandgap lower than 3 eV
• The conductivity is not a monotone function of the temperature.
CHAPITER 8
Statistical mechanics : Fermi – Dirac
function and Maxwell – Boltzmann
function
Ph. Lorenzini
The statistics
• Maxwell – Boltzmann
• Distinguishable particle
• Gas molecules at fairly low pressure
• Bose – Einstein
• Indistinguishable particle
• Several particles permitted in each quantum state
• Photons, Phonons
• Fermi – Dirac
• Indistinguishable particle
• Only one particule in each quantum state (Pauli principle
• Electron in the crystal, solids
In each statistics, particles are assumed to be noninteracting

Ph. Lorenzini
The statistics
• Solving methods:
n particles
Quantum states
ni
n2
n1
indistinguishables distinguishables
• Conditions :
• Number of particles is constant : n i n
n E
i
• Internal energy is constant : i i E
• Indistinguishables or distinguishables ? i
• One or more particles in each quantum state?

Ph. Lorenzini
Maxwell – Boltzmann
• Distinguishable
• Each state can be occupied by unlimited number of particles
• Equilibrium state is the one which is the most probable
 Consider an equilibrium state with n1 particules in c1, n2 in c2,

…. , ni in ci.
 What is the probability of this configuration?
n! n!
P 
n1!n2 !...ni !...  ni !
i
 This configuration is the equilibrium one if P or ln(P) is
maximum d(lnP)=0
 We then get
ni
f ( Ei )   e ( Ei  EF ) / kT ni: number of particles in Ei
gi : number of places (staes) on Ei
gi
Ph. Lorenzini
Distinguishable or not
(1 « ball » in each case  Pauli)
Distinguishable undistinguishable
gi !
( g i  N i )!
gi !
N i !( g i  N i )!
ni: number of particles in Ei

Ph. Lorenzini
Undistinguishable
« Bose » « Pauli »
( N i  g i  1)!
( g i  1)! N i !
gi !
N i !( g i  N i )!
ni: number of particles in Ei

Ph. Lorenzini
Fermi –Dirac statistics

• Same approach than before
gi!
Ei
N i !( g i  N i )!
• Energy level Ei, degenerated gi times ( it means gi « boxes » or
cells/ quantum states with the same energy Ei) and Ni particles to
distribute.
• Si maintenant , on a plusieurs (n) niveau d’énergie, le résultat est le
produit de toutes les distributions
Number of energy levels

n
gi!
P
i 1 N i !( g i  N i )!
Ph. Lorenzini

• Same approach than before
E4
gi!
E3
N i !( g i  N i )!
E2
Electrons density on
E1
energy level E
Number of energy levels
n
gi! n( E ) 1
P f (E)  
i 1 N i !( g i  N i )!
g (E)  E  EF 
1  exp 
 kT 
States density on
Fermi level
energy level E
Ph. Lorenzini

1
E  EF  f ( E )  1 f (E) 
at T= 0K:  E  EF 
1  exp 
E  EF  f ( E )  0  kT 
BC
EF
BV
Ph. Lorenzini
T = 0K
EF
T> 0K
Ph. Lorenzini
Temperature dependance
E1 E2 E3 E4 E5
EF
Ex: T=300K, if E – EF = 3kT, f(E) ?

Ph. Lorenzini
Approximation FD  MB ?
• The use of FD is not
always easy in calculus
• In which conditions can
we approximate FD by
MB ?
• Condition:
Si E - EF > 3 kT, 5% of error
 Rule: we can confound ( 5% error) Fermi – Dirac

function and Maxwell – Boltzmann function when | E -
EF |> 3kT. At room temperature (T=300K), kT=25 meV
 | E - EF |> 75 meV.
Ph. Lorenzini
Approximation FD  MB ?
E
EF -E EF +E
For E < EF, it’s easy to show that the unoccupation probility 1-f(E) is the
same at EF - E than the occupation probability f(E) at EF + E.
1  f ( E F  E )  f ( EF  E )
The point f(E) = 0,5 is the center of symmetry for the curve.
Ph. Lorenzini
E E E
EF
g(E) f(E) E dn(E)/dE

E E
EF
dn(E)/dE
g(E) f(E)
CHAPITER 10
The semiconductor in equilibrium
Ph. Lorenzini
The semiconductor in equilibrium

• equilibrium?
• No external forces:
• No applied voltage
• No magnetic field
• No gradient of temperature
Ph. Lorenzini
Charge carrier density in semiconductor.

• Question: how many electrons in CB and holes in VB in a
semiconductor at a given temperature ?
• answer : (see chap 9!)
• At an energy E: n( E )  g c ( E ) f ( E )
p ( E )  g v ( E )(1  f ( E ))
• In a band with a given width
Ec max
n0   g c ( E ) f ( E )dE
Ec min
Ev max
p0   g v ( E )(1  f ( E ))dE
Ev min
Ph. Lorenzini
Charge carrier density in semiconductor.

Ph. Lorenzini
n0 and p0 equations
• hypothesis: EF is 3kT (or more) below CB and is 3KT
above BV  we can use MB function.
• n0 relations:
E  EF
Ec max  4 (2m )* 3/ 2
( )
n0   g c ( E ) f ( E )dE   n
E  EC  e kT
dE
Ec min Ec min h3
3/ 2
  ( Ec  E F )   2m kT  *
n0  N C exp   avec N C  2  n
 kT   h 
2
Ph. Lorenzini
n0 and p0 equations
3/ 2
  ( Ec  E F )   2m kT *
n0  N C exp   with N C  2 
n
 kT   h
2

3/ 2
  ( E F  Ev )   2m kT  *
with NV  2  p
p0  NV exp  
 kT   h  2
 
Nc and Nv are the effective density of states function (~ density of

place) in the conduction band and valence band respectively. The
magnitde is around Nc ~ Nv ~ 2.5 x 1019 cm-3.
It is as if all the energy states in the CB were effectively squeezed
into a single energy level, EC or Ev
Ph. Lorenzini
Law of mass action

• Pratical rule: if we take the product of electrons density in the
CB and holes density in the VB, we obtain:
Ec  E v Eg
n  p  N c N v exp( )  N c N v exp( )  ni
2
kT kT
• Important remark: the expressions we have just established

for n and p have been assuming that EF is separated from
CB and VB by at least 3kT. This calculation does not assume
that n = p Rule n x p = ni2 will remain true even when the
SC will be doped, provided the condition on EF is valid.
Otherwise (exception) we say that the SC is degenerate.
Ph. Lorenzini
Fermi level in an intrinsic semiconductor

• Intrinsic semiconductor : free impureties semiconductor, ie only allowed
energy level are the CB and the VB.
• n electrons in the CB . These electrons come from VB. It leads to p
holes in the VB and of course:
 n  p  ni T
Ph. Lorenzini
Fermi level in an intrinsic semiconductor

• Postulat: n = p
  ( Ec  EFi )    ( EFi  Ev ) 
n0  N C exp    p0  NV exp  
 kT   kT 
 2 EFi  N C  ( Ec  Ev ) 
exp    exp  kT 
 kT  NV  
Ec  Ev 1 N V Ec  E v 3 mv* Ec  Ev
E Fi   kT ln   kT ln *   E Fi
2 2 NC 2 4 mc 2
~10 meV
Ph. Lorenzini
Bipolar conduction
• The presence of electrons

and holes causes bipolar Hole
conduction in the SC
 By adding a small amounts of
specific dopants, we can alter the
conduction process and we can Hole
favor conduction by electrons or
holes
E external Field
Ph. Lorenzini
Electrons in a diamond structure (ex: Si)

Ph. Lorenzini
Electrons in a diamond structure (ex: Si)

Ph. Lorenzini
Some semiconductors
• Data (300K):
mc* mv* NC NV Eg ni
m0 m0 (1019 cm-3) (1019 cm-3) (eV) (cm-3)
Si 1,06 0,59 2,7 1,1 1,12 1,5x1010
Ge 0,55 0,36 1 0,5 0,66 2,4x1013
GaAs 0,067 0,64 0,04 1,3 1,43 2x106
GaN 0,2 1,4 0,223 4,6 3,39
4H-SiC 1,69 2,49 2,86
InP 0,073 0,87 0,05 2 1,27
Ph. Lorenzini
ni as a function of
temperature
Eg
ni  N c N v exp( )
2kT
• In an intrinsic SC:
Eg
• ni is not a constant ni  N c N v exp(
2kT
)
• ni increases as:
• Eg decreases
• Temp increases
Ph. Lorenzini
Doping of semiconductors
• Exponential variation of the
intrinsic carrier density
• If ni>1015cm-3, the • Remark: n x p product
semiconductor is not is independant of Fermi
suitable for electronic
devices level (except Boltzmann
• We can decrease ni by approximation is valid)
using large band gap
material (SiC, GaN,  valid even the SC is
Diamond)
doped
Ph. Lorenzini
Semiconductor doping
• Allow to change and control electrical properties of semiconductor
• Doping impurities will change the relation n = p:
• Donor like impurety  n > p  n type

• Acceptor like impurety  p > n  p type
Depending of the nature of the chemical species (position in the

periodic table), the behavior of these impureties will favor n
type conduction or p type conduction.
Ph. Lorenzini
• Case of Silicon (col IV)
• If we introduce doping impureties from column V:
• For example N, As or P  5 valence electrons ie one extra electron
compared to the Silicon.
e-
P
P
neutral Ionised +
It looks like Hydrogen

e- +
atom
Ph. Lorenzini
• Column V atom : analogy with the hydrogen atom
• The extra electron does not belong to VB !
• The extra electron does not belong to CB !
• Where is its energetic position? Between the both of them !!
• Midified hydrogen model »:

• Electron is not moving in a free space
• Electron has an effective mass ( different of the free mass)
We will now calculate this energetic position

Ph. Lorenzini
Ionization energy of dopants.

1 q2
• Force F We have to replace:
4 0 vacuum r2
• vaccum by sc
1 q2 • m by m*.
• Potential V 
4 0 vacuum r
mq 4
• Energy En  
2n 2 (4 0 vacuum  ) 2 1 m*
R  Ry 2
*
 sc m0
y
• Rydberg mq 4
R  13,6eV
2(4 0 vacuum  ) 2
Ph. Lorenzini
Donnor level in the forbidden energy band

( gap).
• Application to Silicon ( Si~12)
• In our case, E is energy we have to give to electron to elevate it (the
extra electron) into the conduction band.
• E = EC – ED=Ry*/n²=Ry*
1 0,5  m0
R  13,6
*
y  40meV
(12) 2
m0
• Experimental measurements :
• P => 44 meV
• As => 49 meV
• Sb => 39 meV
• Bi => 67 meV
Ph. Lorenzini
Donor level in the forbidden energy band (

gap).
• Summary :
• When the extra electron is bonded to P atom  energy level
ED and its orbit are localized around the donor atom.
• When the electron moves to conduction band  its quantum
state changes ( quantum state describes by Bloch wave)) 
it’s moving in the crystal  it contributes to the conduction, it
is a free electron
 compounds SC
 GaAs or InP (III-V)
 VI substitutes to As => Se, Te
 IV substitutes to Ga => Si, C, Sn
 m*=0,05 m0  EC – ED = 5 meV !!
Ph. Lorenzini
Donor level in the forbidden energy band (

gap).
• we can create / elevate electrons in CB without any holes

in VB  n >> p.
• VB is full  no conduction allowed in this band.
• only charge carriers in CB (electrons) conduct  unipolar
conduction
Ph. Lorenzini
Acceptor level in the forbidden energy

band ( gap)..
• Case of Silicon (col IV)
• If we introduce doping impureties from column III:
• For ex B, Al ou Ga  3 valence electrons ie one fewer compared to Si.
B
P
+e
neutral Negatively
Ionised
One again looks like

e+ -
hydrogen atom
Ph. Lorenzini

band ( gap)..
Free quantum state: we can show that this energy level is very close to
the valence band. Same calculation than « donor » level.
Ph. Lorenzini
How to calculate the energy level Ed or Ea
• The problem "looks like" the

model of the hydrogen atom :
m0e 4 13.6
En   2 eV
2(4 0 ) 
2 2
n
• Introduction of modified
Rydberg:
 m*   0 
2
Ed  EC  13.6  
 m0   
Few examples
Ph. Lorenzini

band ( gap)..
p >> n
 We get :
13,6 mv*  B => 45 meV
Ea  EV  2  Al => 57 meV
 sc m0  Ga => 65 meV
 In => 160 meV
Ph. Lorenzini
Statistics for donor and acceptor levels

• Donor states:
• The writing has to be lightly modified:
1
f D (E) 
1 E  EF
1  exp( )
2 kT
• Factor ½ ?
• P : 3s²3p3  2 electrons s and 2 électrons p participate in
the bindind=> the extra electron (p state) are on the Ed level.
This electron has a particular spin value (up ou down).
• Once that an electron leaves this level, the box (the level) is
empty and can trap a up or down spin  the capture
mechanism is increased compared to the emission one
f D (E)  f (E)
Ph. Lorenzini

• Donor states:
• Electrons density trapped on ED ?
ND
nD  N D  f D ( E D ) 
1 ED  EF
1  exp( )
2 kT
• Or :
nd  N d  N d
Ph. Lorenzini

• Acceptor level:
• One can only capture the electron that has the right "spin" but we
can release any  f ( E )  f ( E )
A
1
f A (E) 
E  EF
1  2 exp( )
kT
1 g : degeneracy factor
f (E) 
E  EF 1: intrinsic
1  g exp( ) 2 ou 4: acceptor
kT ½: donor
Ph. Lorenzini

• Acceptor level:
• Electron density on EA
NA
nA  N A  f A ( E A ) 
E  EF
1  4 exp( A )
kT
• or:
NA 
p A  N A  nA  N A  N 
1 EF  E A
A
1  exp( )
4 kT
Ph. Lorenzini
Total ionization and carriers freeze out

• Neutrality relation:
• Positive charges = negative charges
 
 
e p  ND  e n  NA 

• Simplification : NA = 0 (n type)

n p  N D
  ( Ec  E F )    ( E F  Ev )  ND
N C exp   - NV exp   
 kT   kT  1  2 exp( ED  EF )
kT
Ph. Lorenzini
Fermi level EF(T)

• We “split” of the problem in the temperature range :
• ULT (ultra low temperature) : we neglect the holes ; OK but until which
temperature this assumption is correct?
Eg
ni  N D  kT 
2 ln[ NcNv 
1/ 2
/ ND ]
kT must be around Eg/10
  ( Ec  E F )  ND
N C exp   - 0
 kT E
 1  2 exp( D  E F
)
kT
Ph. Lorenzini
Fermi level EF(T)
 1   Nd ( Ec  Ed ) / kT   
1/ 2
E F  E D  kT ln   1  1  8 e  
 4   Nc   
• kT<<Ec‐Ed
Ec  Ed 1 Nd Ec  E d
EF   kT ln , E F (T  0)  E Fmax  N C  N D /2
2 2 2NC 2
• Ec‐Ed <kT<Eg
NC
( Taylor’s series of the square root) EF  E c  kT ln
ND
Ph. Lorenzini
EF function of the Température
Conduction Band
EC
ED
Fermi Level
EC  EV
2
Temperature
Ph. Lorenzini
And the carrier density ?

• LT - - - - - - - - -
Freeze out (« gel des porteurs »)
• T « middle range » - - - - - - - - -
Exhaustion/saturation of donors
• « High » Température
intrinsic
+++++++++++++++++
Ph. Lorenzini
Carrier density versus temperature
Complete ionization of
donors
3 régimes:
•Extrinsic (Freeze out)
•saturation
•Intrinsic
Ph. Lorenzini
Carriers distribution at equilibrium

 ( Ec  E f )   ( Ec  E fi )  ( E f  E fi ) 
n0  N c exp    N c exp  
 kT   kT 
 ( Ec  E fi )   ( E fi  E f ) 
n 0  N c exp   exp  
 kT   kT 
 ( E fi  E f ) 
n0  ni exp  
 kT 
 ( E f  E fi ) 
p0  ni exp  
 kT 
Ph. Lorenzini
Fermi level in intrinsic SC

Ph. Lorenzini
Fermi level in intrinsic SC

Ph. Lorenzini
SC doping: n type
Ph. Lorenzini
SC doping: p type
Ph. Lorenzini
h+ and e- concentration at equilibrium

Ph. Lorenzini
h+ and e- concentration at equilibrium

• Neutralité électrique:
n0  N A  p0  N D
• Ionisation totale:
n
2
n0  N A  p0  N D avec p0  i
n0
• Soit:
n 0  ( N D  N A )n0  n  0
2 2
i
• On obtient:
(ND  N A ) ( N D  N A )2
n0    ni2
2 4
Ph. Lorenzini
Ef – Efi difference
• Instead of expressing Ef according to Nc and Nv, we

can write :
 Nd  n type
E f  Ei  kT ln 
 ni  eFi
 Na 
Ei  E f  kT ln  p type
 ni 
Ph. Lorenzini
Ef – Efi difference
• We can then express the density of electrons and holes at
equilibrium by :
n  ni e ( E F  E Fi ) / kT
 ni e e Fi / kT
Boltzmann
approximation
p  ni e  ( E F  E Fi ) / kT
 ni e  e Fi / kT
with:
eFi  EF  EFi  0 N type
eFi  EF  EFi  0 P type

CHAPITER 11
Nonequilibrium semiconductor
Ph. Lorenzini
Plan:
• Recombination and generation processes
• Currents in semiconductors
• Current density relations
• Continuity Equations
• Debye length
• Poisson’s Equation
• Dielectric relaxation time constant
Ph. Lorenzini
Recombination and generation processes

• Mass action law:
• At equilibrium: np  ni 2
• Nonequilibrium state : onset of exces carrier generation and

recombination processes
• Carrier generation and recombination:
Unity [g]=[r]=s‐1cm‐3
• Net recombination rate:
g ' r '  g  g th  r '  g  r with r  r ' g th

external internal
Ph. Lorenzini
Recombination and generation processes
At equilibrium, g (generation rate)

must be equal to r recombination
rate.
If T=0K (and no light !), g = r = 0.

Ph. Lorenzini
Recombination processes: 2 « ways » are possible (1)
• Direct Electron – hole Recombinaison

• process is function of electrons and holes concentration:
p n
rp  rn 
p n
• Example: n type + light excitation in low injection regime

lumineuse en faible injection ( ie n  p  n0 )
p  p0  p n  n0  n  n0
• In low injection regime, majority caririer density is not
modified.
Ph. Lorenzini
• Recombination assisted by trap:

• In general, traps are localized in the mid gap
• Recombination can be expressed as:
1 np  ni2 Shockley‐Read
r
 m 2ni  p  n relation
where  m is characteristic of the trap
• If the both processes are present:
1 1 1
 
 m  n( p)
Ph. Lorenzini
• If intrinsic (or low doping level) semiconductor: we apply

Shockley‐Read relation
• If doped semiconductor (for ex n type):
p 1 1 1
rp  with  
  m p
• If region with no carrier (example: space charge region)
x i
np
1  n 2
ni
r r 0 Net generation rate.
p x
 m 2ni  x n 2 m Charge carriers are
created
Ph. Lorenzini
Light excitation
P type
Ph. Lorenzini
radiative and nonradiative recombination

Ph. Lorenzini
Surface effects
Ph. Lorenzini
Currents in semiconductor
• Drift current : an electric field E is present/applied
• If E=0, electrons velocity = thermal velocity (107 cm/s) but average

velocity is zero due to scattering with atoms and impureties.
• mean free path:
o
l  vth.  100 A
• Mean time between collisions
  0.1 ps
Ph. Lorenzini
drift velocity
Ramdom behavior of a hole without and with an electric field

Ph. Lorenzini
Drift current
• Drift current:
• Between two collisions, electrons are uniformly accelerated
• Acceleration:   qE / m *
• velocity: v  qE / m*   µE
• Mobility: µ  q / m * Si : 1500 cm2/Vs
GaAs: 8500 cm2/Vs
In0.53Ga0.47As:11000 cm2/Vs
Ph. Lorenzini
Drift current
• Drift current density:
• For electrons: J c n  nev n  neµn E
• For the holes: J c p   pev p  peµp E
• global: J c total  J n  J p  (neµn  peµp ) E

Ph. Lorenzini
Mobility versus temperature
Ph. Lorenzini
Mobility versus doping

concentration
Ph. Lorenzini
Saturation velocity
• saturation velocity of electrons (« overshoot »)
• Linear relationship velocity – electric field

only valid for:
• Low electric field
• Carrier in equilibrium with the lattice
• If not:
• velocity saturation for electric field above the critical
field
• velocity overshoot for multi-valley semiconductors.
• Balistic regime: for devices with low dimension
(below mean free path (0.1µm))
Ph. Lorenzini
Overshoot for multivaley semiconductors

Ph. Lorenzini
Currents in semiconductors
• Diffusion current:
• Origin: electrons (or holes) density gradient
• Diffusion from a region from high concentration toward a region of
low concentration.
• 1st law of Fick:
dn Number of electrons which diffuse

n   Dn
x
D
dx by time unit and volume unit (flow)
dp Number of holes which diffuse by
pD   D p
x
dx time unit and volume unit (flow)
Ph. Lorenzini
• Diffusion current : we sum up electrons and holes

contribution:
dn dp
J diff  e(n  p )  eDn
x
D
x
D
 eDp
dx dx
• Electron diffusion coefficient or constant
D   cm² / s
n, p
Ph. Lorenzini
• Total current density : we have 4 possible independant current:
J T  J cond  J diff  J n  J p
dn dp
J T  (neµn  peµp ) E  e( Dn  Dp )
dx dx
• D and μ express the ability of carriers to move. There is a
relationship between them : it’s the Einstein relation:
D kT

µ e
Ph. Lorenzini
Continuity equation – diffusion length
• G and R alter the distribution of

carriers and so the current.
dn( x, t )  J n ( x  x) J n ( x) 
Ax  A   RG
dt  e e 
dn( x, t ) dJ n ( x) x
Ax A RG
dt dx e
 We obtain the continuity equations for
electrons and holes:
dn( x, t ) 1 dJ n dp ( x, t ) 1 dJ p
  rn  g n   rp  g p
dt e dx dt e dx
Ph. Lorenzini
• Example: case where the current is exclusively due

to a diffusion process:
dn d n n  n0
2
 Dn 2 
J n (diff )  eDn
dn dt dx n
dx
dp dp d2 p p  p0
J p (diff )  eD p  Dp 2 
dx dt dx p
Ph. Lorenzini
• In steady state, the time

derivatives are zero:
d 2 (n  n0 ) n  n0 n  n0
 
dx 2 D n n L2n
d 2 ( p  p0 ) p  p0 p  p0
   Diffusion length : represents the
dx 2
D p p L2p mean distance an electron (a hole)
can travel before it recombines
L  D
• Solutions:
n n n
Lp  D p p with a hole (an electron)
 Ln or Lp >> VLSI devices
n( x)  (n( x)  n0 )  n(0)e  x / L n

Ph. Lorenzini
Poisson’s equation
• Comes from the first Maxwell equation. It etablishes a relation

between electric potential (V(x)) and charge carrier density ((x)):
d 2V dE  ( x)
   
dx 2 dx  sc
• In the SC, two types of charges (fixes and mobiles):
d 2V e
   p ( x )  n ( x )  N 
( x )  N 
A ( x)

dx 2
 sc D
Mobile Charges Fixe Charges
(electrons and holes ) (ionised dopants)
Ph. Lorenzini
Debye length
• If we write Poisson’s equation in an n type SC and if we express n

with respect of Fi :
d 2  Fi
dx 2
 
e
 sc
N 
d ( x )  ni e e Fi / kT

• If Nd(x) => Nd+Nd(x) , then Fi is modified by a quantity

Fi
If we remark that : V(x)=Fi cte
d 2 Fi e 2 N d e
    N d ( x)
 sc kT  sc
Fi
dx 2
Ph. Lorenzini
Debye length
• Physical meaning ?
• Solution of the differential equation :
x  sc kT
Fi  A exp avec LD 
LD e2 N D
• The energy bands can’t vary abruptly (and in the same
maner the carrier density), but need few LD (if Nd=1016 cm-3,
LD=0.04µm). In this region, an electric field appears.
Ph. Lorenzini
Dielectric relaxation time constant
• Comment évolue dans le temps la densité de

porteurs majoritaires ?
• Équation de continuité (R et G négligés):
n 1 J n E
 or J n  E  E /  n et  en /  sc
t e x x
d’où
n n
 Solution: n(t )  exp(t /  n sc )
t  n sc
   n sc Temps de relaxation diélectrique ( 10-12 s)

CHAP 12
The PN junction diode
Ph. Lorenzini
PN Homojunction
• Non linear device
• rectifier devices (composants redresseur)
• 2 devices reach the same results:
• PN Junction(this chapter)
• Schottky barrier or Metal / SC contact (next chapter)
Ph. Lorenzini
The Junction’s formation mechanism

•PN Junction at equilibrium
1st Step: diffusion mechanism
2nd Step: built in Electric Field

appears compensates diffusion
forces e-h pair’s
recombination
Flat Fermi level:
No current / thermal equilibrium
E int
Ph. Lorenzini
« built in potential VB i »
• Definition : Potential drop between N and P regions
Vbi  VN  VP
 dp ( x) 
Holes current equation: J P ( x)  e   P p ( x) E ( x)  D p  0
 dx 
p 1 dp ( x)  e dV ( x) 1 dp( x)
or E ( x)  or 
Dp p ( x ) dx kT dx p( x) dx
kT pp
Integrating from P to N region: Vbi  ln( )
e pn
kT N N
finally: VD  Vbi  ln ( A 2 D )
e ni
Ph. Lorenzini
•Field, potential and Space Charge width(1)

• Poisson’s equation:
d 2V ( x)  ( x)
 
dx 2  sc
 In N and P region:
d 2V (x) e
  ND 0  x  WN
dx 2
 sc
d 2V (x) e
  NA WP  x  0
dx 2
 sc
-WP -WN
Ph. Lorenzini
 Electric Field E(x)
eN D eN A
En ( x )   ( x W N) EP ( x)   ( x W P)
 sc  sc
 Continuity of Field on x=0:
N DW N  N AWP
eN DW N eN AW P
EM   
 sc  sc
-WP -WN
Ph. Lorenzini

 Built in potential V(x)
eN D
Vn ( x )   ( x W N ) 2  Vn
2 sc
eN A
V p ( x)  ( x W P) 2  V p
2 sc
 Depletion layer (ZCE)
eN DWn2 eN AW p
2
V (Wn )  V (W p )  Vd  
2 sc 2 sc
2 sc ND
W p (Vd )  Vd
e N A (N A  N D )
2 sc NA
Wn (Vd )  Vd
e ND (N A  ND )
2 sc N D  N A
W (Vd )  Vd -WP WN
e NAND
WARNING: WHEN A VOLTAGE V IS
APPLIED ON P SIDE, VBI HAVE TO
BE REPLACED BY VBI - V
Ph. Lorenzini
BIASED PN JUNCTION
• When a positive voltage is applied on p side, the

equilibrium is destroyed and a net current can flow
 simplifying assumptions :
 Depletion layer with no free carriers (e‐ and h+)
 Low injection
 Boltzmann’s approximation
 Drop voltage only in depletion layer
 No generation‐recombination mechanisms present
Ph. Lorenzini
BIASED PN JUNCTION
• Foward Biasing
• Positive voltage on P
• lowering of built in
potential
• Diffusion mechanism
dominates
• High current
Ph. Lorenzini
BIASED PN JUNCTION
Fdiff e-
• Forward biasing
• Lowering of built in Field due to
opposite external field
• Electrons injected from N to P
regions: minoritary carriers
injection
• High current due to full
« reservoir »
Fdiff h+
Ph. Lorenzini
Forward Biasing
Eext
Ph. Lorenzini
Jonction PN sous polarisation
Fconde‐
• Reverse Biasing
• Global Electric Field increases
(External Field added to built in
Field)
• Injection of electrons from P to
N and holes injection from N to
P: majority carriers injection
• Low current ( leak current) due
to empty « reservoir »
Fcond h+
Ph. Lorenzini
PN Junction under biasing
At equilibrium, null current  two components compensate between it. Taken
separately, the magnitude of these components 104 A / cm² (ie 1A for typical diode)
and at low injection I is of the order of few mA (max 10mA)
•Boltzmann’s Approximation: The Boltzmann approximation is to say that the resulting

current being small compared with the components of this current, we consider that we
are still in quasi equilibrium and therefore that the current's equation is still valid by
replacing Vbi by Vbi ‐VA :
 e dV ( x) 1 dp( x)

kT dx p( x) dx
Ph. Lorenzini
Density of carriers injected to the limits of depletion

layer
p (WN ) pn eVbi
• If Va=0  exp( )
pp pp kT
• If Va 0 p' (WN ) p'n

e(V  VA )
 exp( bi
p eV
)  n exp( A )
pp pp kT pp kT
eV A ni2 eV A eV A ni2 eV A
p ' n  p n exp( ) exp( ) n' p  n p exp( ) exp( )
kT ND kT kT NA kT
eVa
n * p p  p * nn  n exp(
'
p
'
n
2
i )
kT
Ph. Lorenzini
Holes density injected versus bias voltage Va
1017
1016
Na= 1E17 cm-3
1015 Vd=0.7 V
1014
1013
1012
P'(Wn) (cm-3)
1011
1010
109
108
107
106
105
104
0,0 0,1 0,2 0,3 0,4 0,5 0,6 0,7
Va (V)
Ph. Lorenzini
Minority carriers distribution in neutral region
• Due to gradient concentration,

carriers will diffuse and produce
diffusion current (no electric field
in neutral region !)
• Distribution is geometry
dependant
• Discrimatory parameter : length

diffusion LDn,p of electrons and
holes and neutral region widths
dn,p
-WP 0 WN
Ph. Lorenzini
Minority carriers distribution in neutral region
• Long regions ( d n , p  L p , n )  Short (narrow) regions (d n , p  L p ,n)

eV
eVa p n kTa
p ' ( x )  p n  p n (e kT
 1)e
(W N  x ) / L p p' ( x)  p n  (e  1)( x c  x)
dn
eVa
( x Wp ) / Ln n p eVkT
a
n' ( x )  n p  n p (e kT
 1)e n' ( x)  n p  ( e  1)( x 'c  x )
dp
 General case pn
eVa
  xc  x  
p' ( x)  p n  (e  1) sh 
kT

dn  L 
sh( ) p
Lp
np eVa  x  xc'  
n' ( x )  n p  (e kT  1) sh  
dp  Ln 
sh( )
Ln
Ph. Lorenzini
Minority currents in neutral region
Knowing minority distribution we are in position to calculate the

current wich is a diffusion current (very low field in neutral region):
dp ( x) dn( x)
J p ( x)  eD p J n ( x)  eDn
dx dx
 Hypothesis : no G‐R process in depletion layer (ZCE)
J (V )  J p ( W p )  J n (W p )  J p (Wn )  J n (W p )
 We get the classical and well known diode equation:

J (V )  J S (e eV / kT  1)
Js is the theorical saturation current or reverse current
Ph. Lorenzini
Minority currents in neutral region
 Short (Narrow) region
eni2 D P eni2 Dn
JS  
NDdn N Ad p
 Long region
eni2 DP eni2 Dn
JS  
N D LP N A Ln
 General case -WP 0 WN
eni2 DP eni2 Dn
JS  
dn dp
N D LP th( ) N A Ln th( )
LP Ln
Ph. Lorenzini
The real diode: generation‐recombinaison
mechanism in depletion layer
• The model is refined  we take into account G‐R process in
depletion layer
• Well understood mechanism (Shockley‐Read)
1pn  ni2
r
 2 ni  p  n
eVa
 We know that p (W N ) n(W N )  p (W P )n(W P )  ni2 exp( )
kT
 n 2
If we suppose np constant in depleted region and np >>
i
(in forward bias) , the rate r is max when n=p, and it can
be rewritten
ni  eVa 
rmax  exp 
2  2kT 
Ph. Lorenzini
The real diode: genération‐recombinaison
• Generation‐Recombinaison current in depletion
layer can be expressed as:
WN
J n (Wn )  J n ( W p )  J GR  e  rdx
WP
 pn  ni2
For reverse biasing ( ), we have a
n
negative rate ( ) . It means
r  
2
0 i
dans we
have a net generation process
 For forward bias, rmax=cte>0 and the current is
due to recombinations.
Ph. Lorenzini
The real diode: genération‐recombinaison
• Finally GR current present in depletion layer can be
expressed as:
 eVa  eni
J GR  J 0
GR  exp( )  1 J GR
0
 WT
 2kT  2
 If we take into account the diffusion current, we get:

 eV  0  eV 
J (Va )  J S  exp( a )  1  J GR  exp( a )  1
 kT   2kT 
 The expression above can be generalized by introducing
Ideality factor:
 eVa 
J (V )  J 0  exp( )  1
 nkT 
Ph. Lorenzini
Reverse biasing:
Junction breakdown
• Thermal Effect (Narrow bandgap)
• Zener Effect:
• direct flowing from VB to CB by
tunnel effect (0), if electric field tunn
Tunnel
above critical Field Ec el
• Avalanche Effect:
• before« tunneling », hot electrons
(accelerated electrons) excite by impact
Avalanche
ionisation electrons from VB to CB
(1,2,3) etc….
 .EC2
VBD 
2eN B
• « Punchtrough »
Ph. Lorenzini
Small signal model of the diode: capacitances
• Capacitance associated to charges

• 2 types of charges present in the junction
• Fixed charges (ionised dopants) in depletion layer
• Mobiles (e- et h+) injected when forward biasing
• 2 types of capacitance
• Junction (or Transition) Capacitance
• Charge Storage (or diffusion) Capacitance
Ph. Lorenzini
Junction capacitance
Simply associated to charges present in depletion layer
dQ
CT  C j  Q  eAN AWP  eAN DWN
dV
or:
A 2e N AND A
CT  C j  
2 (VD  VA ) ( N A  N D ) WT
Ph. Lorenzini
Charge Storage (diffusion) capacitance
• Reflects the delay between XC
the voltage and current QSp  A e( p ' ( x)  p n ) dx
WN
• Associated with charges
injected into the neutral Holes density in excess
regionsTraduit le retard entre present in N region
la tension et le courant
QSp   P J P  
 dn 1 
QSp  e( p ' (0)  p n ) LP coth( )  
QSn   n J n  LP dn 
 sh( ) 
 LP 
Ph. Lorenzini

Ph. Lorenzini

• We can transform the previous expression by:
 
 
QSp   J P (WN ) avec 
   P 1
1 
 dn 
 ch( ) 
 LP 
 Time expression can be simplified, depending of the
neutral geometry:
d n2
 t 
Narrow diode:  transit time
2 DP
    P lifetime
Long diode: 
Ph. Lorenzini
• The previous expression, valid in N region, can be

generalized in P region and we obtain for the whole diode:
QS  QSn  Q Sp   ( n ) J n ( W P )   ( p ) J p (W N )
dQS
If we use: CS  Cd 
dV
e
C S  Cd  C Sn  C Sp  K ( ( n ) J n   ( p ) J p )
kT
K : Geometry dependant Factor
(2/3  narrow)
(1/2  long)
Ph. Lorenzini
Equivalent circuit of foward diode
(from Neamen)
rd : diode resistance (dynamic resistance) given by the differential slope of

the I‐V characteristics
kT 1
rd 
e I
rs : serie resistance of neutral region n and p
Ph. Lorenzini
Large signal switching of diode
Wn
As long as the stored charge is positive
 forward bias diode  voltage across
diode is small (few 10 mV)
eVa
p 'n  pn  pn (e kT
 1)
 sd  Storage time ie p ' (WN )  pn

Ph. Lorenzini
Switching Junction PN
Ph. Lorenzini
Large signal switching of diode

• Storage time :The main problem in minoritary carriers
devices:
• Storage time:
 If If 
 sd   p ln(1  )  ln(1  )
 Im I f  I m 
• Rise (or fall) time :
 F   RC j  If
 f  2.3  avec  
 1   I f  Im
• Cj: mean value of capacitance between zero and –V2
Ph. Lorenzini
Tunnel Diode –Backward diode

(a)
(b)
 Va   
I t  I pe   exp1  Va 
V   V  (c)
 pe   pe 
 4a 2m * e  (d)
Tt  exp  b 
 3 
 
(e)
Ph. Lorenzini
PIN Diodes
• VLSI devices  high electric fields
• Avalanche issues and hot electrons effects
• Solutions:
• Reducing the electric field by increasing the space charge
• Incorporation of an "undoped" layer called Intrinsic hence the
name!
Ph. Lorenzini
PIN Diodes
P Region Depletion zone N Region
(x)
+ + + +
-Wp -- d +Wn x
--
Intrinsic
layer.
Ph. Lorenzini
PIN Diodes
Max electric field ( intrinsic zone)
eN A x p eN ( x  d )
Em    D n
 Si  Si
built in potentiel
Em Em
Vbi  ( x p  xn  d )  (Wd  d )
2 2
space charge width
2 Si ( N A  N D )
Wd  Vbi  d ²  W d20  d ²
e N AND
capacity electric field
CT Wd 0 1 Em Wd d² d
    1 2 
CT 0 Wd d² Em 0 Wd  Wd 0 W d 0 Wd 0
1 2
Wd 0

Semiconductors Physics PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Semiconductors Physics PDF

Uploaded by

Copyright:

Available Formats

SEMICONDUCTORS

1. Crystal structure and crystallography

• H. Mathieu, « Physique des semiconducteurs et des composants

• J. Singh, « physics of semiconductors and their

• D.A.Neamen, « semiconductor physics and devices: basic principles »,

• Chenming C. Hu , « Modern Semiconductor Devices for Integrated

• Cours de Physique des semiconducteurs, Pr. Rouzeyre, Université de

• McMurry and Fay, « Chemistry », Prentice Hall; 4th edition (April 7,

Crystal structure and crystallography

• Crystalline and amorphous states

• Geometry of the crystals

• Crystal diffraction - Reciprocal Lattice

Amorphous and crystal states

Amorphous and crystal states

Amorphous Polycrystalline Single Crystal

The idea : U crystal  U free  0

• Several parameters / considerations to take onto account:

• Keep the positively charged ions away from each other

• metal • Atoms try to have their last

• Van der Waals (noble gaz) or

Metallic bond: copper

• Need four more electrons to complete the valence energy

A significant difference between the

• Lower energy if the electrons

(D’après McMurry and Fay)

• the electron released from alcali meal (Na) is trapped by

Van der Waals bond

• The lattice represents a set of points in space which

Lattice Basis (base): Crystal structure

Notation to denote planes and points in a

• Nota : in case of negative number, we place a minus sign over h

Crystalline planes (h,k,l)

Using Miller index, we get : hx  ky  lz  N . If N=2,  plane is the second

Miller Indices (continued)

(From McMurry and Fay)

The 14 Bravais lattices in three dimensional system

Triclinic Monoclinic centered Orthorhombic centered

base centered faces centered

The 14 Bravais lattices in three dimensional system

Simple Cubic body centered face centered

Crystalline structures : multiplicity of the lattice

1 atom 2 atoms (From McMurry and Fay)

Crystalline structures : multiplicity of the lattice

4 atoms (from McMurry and Fay)

Crystalline structures : multiplicity of the lattice

(from McMurry and Fay)

How to define the primitive unit cell

Primitive unit cell

The body centered cubic (bcc)

Primitive basis vectors for a fcc: the resulting lattice is a rhombohedric

From C. Kittel, Dunod, 5°ed.

Atoms surface density

Silicon and its « diamond » structure

The reciprocal lattice

The reciprocal lattice: definition

The reciprocal lattice: definition

f ( x  l )  f ( x) avec l  l1a l1 entier

The reciprocal lattice: definition

The reciprocal lattice: definition

Reciprocal lattice building

Plane of the direct lattice

Properties of the reciproqual lattice

Properties of the reciproqual lattice