.

BAND STRUCTURE

Outline of the Chapter

Introduction Energy Bands - Introductory Remarks Formation of Energy Bands Band Diagrams Semiconductors Carrier Concentration MOSFET
Band Structure 1

.
BAND STRUCTURE

Introduction
We know: We will know: Electrons behavior in single atoms

Electrons behavior in solids (nature of electrical properties of solids) - Bohr model of single atom; - bonding models; - crystalline structure; - Pauli exclusion principle.
Band Structure 2

What to recall:

. Introduction As a first step towards understanding band structure and energy bands, lets first define an energy unit that is often used in semiconductor physics. Electron Volts (eV) In semiconductors, energies are usually very small and we often use the unit of electron-volts (eV) instead of Joules. 1 eV = Energy gained by an electron accelerated by a 1V potential difference. We can therefore write, 1 eV = 1.6 x 10-19 J
Band Structure 3

. Introduction To understand energy levels, lets start by considering, as a simple example, the interaction between visible light and diamond crystal: Diamond is transparent to light

When we say light, we usually mean visible light. But, to be more specific, this means photons of particular energy. (visible light: 400 - 700 nm) EPHOTON = hf = hc/
If we consider 500nm photons,

EPHOTON = (6.6 x 10-34 x 3 x 108)/(500 x 10-9) = 4 x 10-19 J = 2.5 eV

Band Structure 4

. Introduction So we can now say that diamond does not absorb photons with energies corresponding to visible light. When photons get absorbed in a crystal, since the nucleus of the atoms is very small compared to the overall diameter, we would expect the photon to be absorbed by transferring energy to electrons. Electrons that are tightly bound to the nucleus cannot readily change states so the photon energy is mostly given to the outer electrons. Since the outer electrons are also involved with bonding, photons are only absorbed if they have enough energy to break the bonds.
Band Structure 5

. Introduction A very simplistic picture of the variation of light absorption with photon energy might look like this:

% absorbed

EBOND
Band Structure

EPHOT
6

. Introduction We can now say that an electron can be moved between two states, bound and loose loose bound For diamond, EBOND 7eV EBOND

E = hc/ 180nm

NOTE: The source of energy to provide EBOND does not have to be optical, it could also be thermal , or electrical.
Band Structure 7

. Introduction Once the electron is freed from the bond, it could move if a voltage was applied. The diamond could then conduct electricity. If we try to obtain EBOND by thermal excitation, the energy available is of the order of kT, where, k : Boltzmanns Constant 1.38 x 10-23 J/K T : Absolute temperature This gives kT = 25.9 meV at room temperature. Since this is much less than EBOND of diamond, we can assume that diamond is a good insulator.
Band Structure 8

. Energy Bands - Introductory Remarks We can now imagine that we can create a loose electron by supplying enough energy to move an electron from one energy level to a higher energy level. Since optical absorption occurs for a large range of photon energies above the critical value, the upper level must in fact be a band of energy. Any of the photon energy left over after breaking the bond gives the loose electron some potential or kinetic energy. Heat
Band Structure 9

. Energy Bands - Introductory Remarks A simple picture of the allowed energies of the electrons can be put up as follows: EC EG EV This is called the band diagram of a material The forbidden region between the bands - the energy required to free the electron from the bond is formally called the band gap, EG of the material.
Band Structure 10

. Energy Bands - Introductory Remarks Since we presume that the free electrons can move can contribute to electrical conduction - the upper band is known as the conduction band, EC. The lower band is called the valence band, EV, because it arises from the valence (outer) electrons of the atom. So we are now in a position to formally state the differences between insulators, semiconductors, and conductors....
Band Structure 11

. Energy Bands - Introductory Remarks The terms insulator, semiconductor, and conductor are based on everyday experience. The energy to get an electron into the conduction band is usually gained from heat, (thermal energy). An insulator is a material for which the bandgap is so large compared with kT that there are almost no thermally generated carriers. Example: diamond, EG = 7eV A semiconductor is a material for which the bandgap is medium. Example: Si ( 1.1 eV), Ge (0.7 eV), GaAs (1.4 eV)
Band Structure 12

. Energy Bands - Introductory Remarks And a conductor has no bandgap. A metal is not bonded covalently and we dont think of it as having a bandgap at all! Until now, we have been looking at a simplified treatment of energy bands.

NOW, lets determine the origin of these energy bands, and see how these bands are formed.
Band Structure 13

. Formation of Energy Bands In a diamond crystal, atoms have covalent bonds, i.e., they share electrons. In covalent bonds, electrons interact due to electrostatic (Coulomb) forces. According to Pauli exclusion principle, no two electrons in an interacting system may occupy same quantum state. A solid state is an interacting system!
Band Structure 14

. Formation of Energy Bands The strength of the interactions depends on the electron separations, which are not constant. Therefore, the energy levels associated with these bonds broadens into a band. (all the levels for all the atoms are slightly different, and the number of electrons is huge, therefore we can consider all levels as a band). Obviously, the amount of interaction between electrons in different atoms depends on how close the atoms are.
Band Structure 15

. Formation of Energy Bands rA H eHYDROGEN MOLECULE

rB

eH

1s(rA)

1s(rB) r

d= When two hydrogen atoms are infinitely separated, each atom has 1s electron with the same wavefunction and the same set of quantum numbers.
Band Structure 16

. Formation of Energy Bands = 1s(rA)+1s(rB)

HYDROGEN MOLECULE

r = 1s(rA)-1s(rB)
Band Structure

As two atoms approach each other, their wavefunctions overlap, and two new molecular wavefunctions with different energies and hence different quantum numbers emerge. One is the sum of atomic wavefunctions, another is their difference.
17

. Formation of Energy Bands E

Bonding energy HYDROGEN MOLECULE

E*(r) E(r) 1s E1s

E(a)

Interatomic distance, r

As r decreases, the energy of passes through minimum at r = a. Therefore, this state is energetically more favorable than two isolated H atoms, and both electrons occupy this state. The wavefunction corresponding to the lowest energy level is called the bonding orbital, and is the antibonding orbital.
Band Structure 18

. Formation of Energy Bands

HYDROGEN MOLECULE

E E1s E E1s

When the two electrons in the H2 molecule occupy the bonding orbital, the negative potential energy arising from the electronproton attractions is stronger than the positive potential energy arising from electron-electron and proton-proton repulsions.

Band Structure

19

. Formation of Energy Bands

Diamond crystal

When interatomic distance decreases so that electrons interact, and their energy levels broadens (splits) into bands.

2s (2N states) 1s (2N states)

(Assume N atoms in our crystal.)

Atomic separation
Band Structure 20

Energy

2p (6N states)

. Formation of Energy Bands

Diamond crystal

Conduction band
forbidden gap E
G

2p+2s (8N states)

4N 4N

6N 2N Valence band

2s (2N states)

1s (2N states) Atomic separation

Diamond crystal spacing
Band Structure 21

Energy

2p (6N states)

. Formation of Energy Bands s-orbital

Diamond crystal

p-orbitals

In the single carbon atom, there are two 2s and two 2p electrons. The 2s and 2p energy levels are close, so when C atoms approach each other, the interaction results in the four orbitals mixing together to form new hybrid orbitals, which are directed in tetrahedral directions and have the same energy. This process is called sp3 hybridization.
Band Structure 22

sp3 hybrid orbitals

. Formation of Energy Bands

Diamond crystal

When we further move atoms close to each other, two hybrid orbitals of neighbor C atoms overlap and form bonding and antibonding molecular orbitals. In single carbon atom, the valence shell is half-filled. Therefore, all electrons will occupy the bonding orbital. In diamond crystal, bonding and antibonding orbitals split and form valence band and conduction band, respectively.
Band Structure 23

. Formation of Energy Bands Conduction band

forbidden gap E
G

2p+2s (8N states)

4N 4N

6N 2N Valence band

2s (2N states)

E E

sp3 (8N states) Atomic separation

Band Structure

2p (6N states) 2s (2N states)

Diamond crystal spacing

24

Energy

2p (6N states)

. Formation of Energy Bands In isolated carbon atom, there are 4 electrons in the outer electron shell (2s2 + 2p2). Therefore, in our diamond crystal there are 4N electrons in both valence band (4N states available) and conduction band (4N states available). At T=0K, all electrons will occupy states with lowest energy, therefore, they fill up the valence band, and conduction band will be empty of electrons. Fermi energy (EF) = energy level corresponding to highest filled electron state at 0K. To conduct electric current, there must be vacant states in the band.
Band Structure 25

Band Diagrams Band Diagram: Diamond EG - forbidden gap EC - conduction energy EC EV EV - valence energy EG Thermal energy: of the order of kT. At 300K, ETh=25.9 meV. Photons: EPHOTON = hf = hc/ = 1.8 - 3.1 eV
at = 400-700 nm (visible light).

For diamond : EG = 7 eV.

Band Structure 26

. Band Diagrams Band Diagram : Metals

Conduction band

EF
Valence band

EF

Partially filled band (e.g., copper)

Band Structure

Bands overlap (e.g.,magnesium)

27

. Band Diagrams Band Diagram : Semiconductors, Insulators

Conduction band

EG

EG

Valence band

Semiconductor : EG < 2eV

Insulator : EG > 2eV Filled valence band, empty conduction band; separated by forbidden gap.
Band Structure 28

. Semiconductors Electrons and Holes When the electron is excited to the conduction band, it leaves a vacant state. This state is called a hole. Like electrons in the conduction band, holes in the valence band are mobile and act like positively charged particles. However, holes are NOT particles. Hole is just the absence of an electron.
29

EC EV EG

hole

Band Structure

. Semiconductors Electrons and Holes Therefore, electrical conduction in a semiconductor includes movement of both electrons in conduction band and holes in valence band: electron and hole currents.

eis equal to:

h+

Band Structure

30

. Semiconductors Intrinsic Semiconductor A pure semiconductor is known as intrinsic material. (i.e. no foreign atoms are present) Its properties are intrinsic to that material. One very important property of an intrinsic semiconductor is that electrons and holes can only be created in pairs: electron-hole pairs Therefore in an intrinsic semiconductor, the electrons and holes are present in equal numbers.
Band Structure 31

. Semiconductors Intrinsic Semiconductor - carrier generation eh+

At T=0K, the conduction band is empty, the valence band is full. At T>0K, electrons are excited thermally across the band gap to the conduction band, leaving holes in the valence band. It is called electron-hole pair generation. Why electrons are being thermally excited, though thermal energy is much smaller than EG ? (25.9 meV vs 1.11 eV in case of Si)?
Band Structure 32

EC EV

EG

. Semiconductors Intrinsic Semiconductor - carrier generation Due to thermal energy, the atoms in the crystal are constantly vibrating, therefore, they deform interatomic bonds. Thus, in some regions, at some moments, bonds may be overstretched, and the bond energy may be smaller than thermal energy.
Band Structure 33

h+ e-

. Semiconductors Intrinsic Semiconductor - carrier generation In intrinsic semiconductor, the electron concentration in the conduction band, n (electrons per cm3 ) is always equal to the hole concentration in the valence band, p (holes per cm3 ) :

n = p = ni
ni = intrinsic carrier concentration The process opposite to the electron-hole pair generation is also important to maintain the steadystate balance. This process is called recombination.
Band Structure 34

. Semiconductors When an electron in the Conduction Band meets a hole in Valence Band (which means a vacant state EC of lower energy), electron occupies this state, falling from CB down to VB. This EV process is called recombination. The excess energy is released either in form of light photon or as heat.
Band Structure 35

Recombination

CB EG VB

. Semiconductors Recombination

The rate of recombination, ri , is proportional to equilibrium concentration of both electrons and holes:

ri = n0 p0 = ni2
: Constant n0 : Equilibrium electron concentration p0 : Equilibrium hole concentration
Band Structure

e-

h+

36

. Semiconductors Recombination

In steady state, the recombination rate must be equal to the rate of generation (gi) of electron-hole pairs. i.e.

ri = gi

in steady state.

Recombination is a very important mechanism in many devices, including anything with a p-n junction in it. Recombination phenomenon may take many forms and is used in electronics, like light emitting diodes (LEDs), where light emission is the result of the recombination. Band Structure 37

Semiconductors

Conduction

Conduction takes place only when electron-hole pairs are created. The energy for this creation may be: Thermal (lattice vibrations), or Optical (light photons) (photo-conductivity). At room temperature, the number of electron-hole pairs, (ni) in Si is about 1.51010 cm-3. But there are >1022 silicon atoms/cm3. Intrinsic silicon is not a good conductor at room temperature unless it is illuminated.
Band Structure 38

Semiconductors

Conduction

This will be evident if we look at the resistivity values: Metal : = in the range of 2 x 10-6 -cm, whereas Silicon : 2.3 x 105 -cm
(Diamond : 1 x 1016 -cm)

( = resistivity, and R = .l/A) To control the resistivity (i.e. to control the conductivity) of semiconductors such as silicon, we use a process called doping ..
Band Structure 39

. Semiconductors Doping, Extrinsic Semiconductor

Doping is the creation of carriers in semiconductors by introducing impurities into the crystal. By this way we gain extra carriers (electrons and/or holes) at room temperature, thus changing the conductivity. Such a doped semiconductor is referred to as extrinsic. In extrinsic semiconductor, we can obtain predominant concentration of either electrons or holes by controling the nature and amount of impurities.
Predominant electron concentration : n-type semiconductor Predominant hole concentration : p-type semiconductor
Band Structure 40

. Semiconductors n-type doping

A semiconductor with an excess of electrons is known as n-type semiconductor. n-type Si is obtained by adding into a Si crystal small amounts of a group V elements, such as phosphorus (P), arsenic (As) or antimony (Sb). Each of these elements has 5 valence electrons. However, in Si lattice these atoms cannot form 5 bonds since most of the atoms are fourfold bonded Si. Therefore, one of electrons is still weakly bonded to impurity atom.
Band Structure

ephosphorus
41

. Semiconductors n-type doping Phosphorus, arsenic or antimony atoms are referred to as donors in silicon since they donate an extra electron. ED EC EV ED EC EV

T > 50K T = 0K The energy needed to excite electrons is very small (30-60 meV), therefore, at temperatures above 50K most of the donors will be ionized (positively charged). At T = 0K, therefore there is NO electrons in the conduction band. Band Structure 42

. Semiconductors p-type doping

A semiconductor with an excess of holes is called p-type. p-type Si is obtained by adding into a Si crystal small amounts of a group III elements, such as boron (B), aluminium (Al), gallium (Ga) or indium (In). Each of these elements has 3 valence electrons. Therefore, in fourfold bonded Si lattice one of bonds will miss an electron. As we learned, the vacant electron state is a hole. The nearby electron from Si-Si bond can tunnel into this state, thus, moving hole away from the B atom, which turns into negative Band Structure ion.

boron

43

. Semiconductors p-type doping

Due to Coulomb interaction, hole gets attracted by the negative charge of B ion and take an orbit around it. The binding energy is about 30-60 meV.
EC EA EV EA EC EV

T = 0K

T > 50K

Boron, aluminium, gallium and indium atoms are referred to as acceptors in silicon since they accept an extra electron.
Band Structure 44

. Semiconductors Doping : summary

Typical dopant concentration is about 1016 - 1018 cm-3 because Material electrons holes at room temperature almost all n-type majority minority dopants are ionized.
p-type minority majority
(i.e. dopant concentration = carrier concentration) Note that this is 6 - 8 orders of magnitude more than that in intrinsic material. (1.5 1010 cm-3.)

This type of doping is called substitutional because dopant atoms are substitutes for Si atoms.
Band Structure 45

. Carrier Concentration Density of States

How to calculate the number of electrons and holes available for conduction? We need to know: 1. 2. The number of states available at a particular energy that can be occupied by electrons; The fraction of these available states that is in fact occupied at a particular temperature.

From that, we can calculate carrier concentration:

n0 =
f(E) - Fermi function.

Ec

f(E)N(E)dE,

N(E) - density of states;

Band Structure 46

. Carrier Concentration Density of States

Since conduction and valence bands consist many closely spaced energy levels, and because each level can have a number of electrons in it, we consider the number of available states per unit volume - the density of states (DOS). The DOS can be calculated quantum-mechanically. The expressions for the valence and conduction bands are:

8 2 NV ( E ) = h3 8 2 NC ( E ) = h3

* mp * mn

( ) ( )

3/2

(EV

E)

1/2

for E < EV for E > EC

47

3/2

(E EC )1/2

Band Structure

. Carrier Concentration Density of States

E EC N(E)=0 EV N(EC ) = N(EV ) = 0 N(E), cm-3 eV-1

Band Structure

48

. Carrier Concentration Effective Mass

Suppose we have electrons in the conduction band. If we apply electric field to a material, electrons gain energy from the field and move across. The driving force here is: Fext = e , where is electric field. Then the acceleration of electrons should be given by: a = Fext / me , where me is the electron mass in vacuum. However, in the solid, electron interacts with the crystal lattice atoms and thus experience internal forces Fint . The potential energy of the electron changes with distance. We must write then: a = (Fext + Fint ) / me Fortunately, we can simplify this equation ...
Band Structure 49

. Carrier Concentration Effective Mass

In a crystalline solid, atoms are positioned periodically, and thus the variation of Fint and of electron potential energy is also periodical. It has been found that electron acceleration is still valid in the crystal just after little modification: a cryst = Fext / me* where, me* is the effective mass of the electron. Effective mass depends on the material. For Si : mn* = 0.3 m0 , mp* = 0.6 m0 For GaAs : mn* = 0.067 m0 , mp* = 0.48 m0 m0 is the mass of electron in the vacuum : 9.11 x 10-31 kg.
Band Structure 50

. Carrier Concentration Electrons in solids obey Fermi-Dirac statistics. The distribution of electrons over a range of allowed energy levels at thermal equilibrium is described by f(E) . Fermi Function

1 f (E ) = E EF 1 + exp kT

This is Fermi-Dirac distribution function that gives us the probability that an available energy state at E will be occupied by an electron at temperature T. At E=EF, f(E) = . Therefore, a state at Fermi level has a probability of of being occupied by an electron.
Band Structure 51

. Carrier Concentration T2 >T1 f(E) 1 0 EF T1 T2 T=0K Fermi Function The shape of f (E ) depends on the temperature: At T=0K , it is rectangular: f (E) = 1 at E < EF f (E) = 0 at E > EF At T >0K , some probability exists for states above the E Fermi level to be filled.

The relative position of f (E) is governed by the value of EF .

Band Structure 52

. Carrier Concentration Fermi Function

Remember:
The Fermi function f(E) gives us the probability that an available energy state at E will be occupied by an electron at temperature T. Therefore: The probability that an available energy state is occupied by a hole will be:

1- f(E).
Band Structure 53

. Carrier Concentration Fermi Function in intrinsic material

In intrinsic material, EF lies at the middle of the band gap Eg because the concentration of holes in VB is equal to concentration of electrons in CB. E at E = EC : f(E) 0 (but small) CB at E = EV : 1 - f(E) 0 (but small) EC EF VB EV

f(E)

Band Structure

54

. Carrier Concentration Fermi Function in n-type material

In n-type material, EF moves towards EC because the concentration of electrons in CB is higher than the concentration of holes in VB. E at E = EC : at E = EV : f(E) gets larger 1 - f(E) 0 CB EC EF EV

f(E)

Band Structure

VB

55

. Carrier Concentration Fermi Function in p-type material

In p-type material, EF moves towards EV because the concentration of holes in VB is higher than the concentration of electrons in CB. E at E = EC : at E = EV : f(E) 0 1 - f(E) gets larger CB EC EF f(E) 1

Band Structure

VB

EV
56

. Carrier Concentration Equilibrium carrier concentrations

As we already know, the electron concentration in the CB is:

n0 =

EC

f (E ) N (E )dE

For equilibrium conditions, we can replace all distributed energy states in the CB by the effective density of states NC , located at the energy EC . This gives:

n0 = N C f (EC )
Band Structure

57

. Carrier Concentration Equilibrium carrier concentrations Assuming that EF lies at least several kT below EC , (i.e. EC - EF > 3kT), f (EC) is expressed as:

f (EC )= 1+ e
Therefore, we can write: where NC is a constant:

1
EC E F kT

E EF C kT

n0 = N C e

E EF C kT
3

2 m kT NC = 2 h
* n 2
Band Structure

58

. Carrier Concentration Equilibrium carrier concentrations

Similarly, the concentration of holes in the VB is given by,

p0 = NV [1 f (EV )]
Here, NV is the effective density of states in the valence band.

Band Structure

59

. Carrier Concentration Equilibrium carrier concentrations Again, assuming that EF is larger than EV by several kT, (i.e. EF - EV > 3kT):

1 f (EV )= 1 1+ e
Therefore, we can write: where NV is a constant:

1
EV E F kT

E EV F kT

p0 = NV e

E EV F kT
* p 2 3 2

Band Structure

2 m kT NV = 2 h

60

. Carrier Concentration Equilibrium carrier concentrations

n0 = N C e

E EF C kT

p0 = NV e

E EV F kT

What follows from these equations? if T increases, then n0 and p0 go up dramatically; if EF moves up towards EC, then n0 increases and p0 decreases; if EF moves down towards EV, then n0 decreases and p0 increases. (These equations are valid for both intrinsic and doped materials provided thermal equilibrium is maintained.)
Band Structure 61

. Carrier Concentration Equilibrium carrier concentrations

E EV i kT

In case of intrinsic material:

ni = N C e

E Ei C kT

pi = NV e

where ni, pi are intrinsic carrier concentrations and Ei is the Fermi level for intrinsic material;

EC EV Ei 2
Hence we get:

ni pi = n0 p0 = N C NV e
Band Structure

Eg kT
62

. Carrier Concentration Equilibrium carrier concentrations

In intrinsic material, ni = pi , thus:

ni = N C NV e 2 kT
And:

Eg

n0 p0 = n

2 i

This very useful relation is called the mass action law. Another way of writing the same thing:

n0 = ni e

E F Ei kT

and:
Band Structure

p0 = ni e

Ei E F kT
63

. Carrier Concentration Conductivity & Mobility

Electrons and holes that are able to move can contribute to electrical conductivity. The current depends on: carrier concentration (n, p) carrier speed ( vn , vp ) carrier charge (q or e) Therefore, current density (A/cm2 ) can be written the following way:

J n = nev n

and J p = pqv p
Band Structure 64

. Carrier Concentration Conductivity & Mobility

At low electric field, the carrier velocity is proportional to the field ( ):

v =
The proportionality constant is called the mobility.

vn = -n vp = p
The total current density is: J = q(nn+ pp)

= = /.
Band Structure

is called the conductivity.

For Si, n= 1350 cm2/Vs p = 480 cm2/Vs

65

. Carrier Concentration Temperature dependence This is not quite obvious as using the equations we derived before because: ni is a strong function of T:

2 kT ni (T )= 2 2 h

(m

* n

* p

Eg 4

e 2 kT

also, Eg is mildly T-dependent

Band Structure

66

. Carrier Concentration Temperature dependence

A good description of the variation of n0 with T is given by this figure: 1017 n0 (cm -3 ) 1015 1013 1011
0 2 (high T)

Ni >> Nd at high T

dopants are ionized; n0 Nd = const. dopants release electrons at T 100K

12 (low T)

ni
4

Nd = 1015 cm-3
6 8 10

1000/ T (K-1)

Temperature dependence of carrier concentration has 3 regions.

Band Structure 67

. Carrier Concentration Compensation doping

Above we considered cases of either n- or p- doping (n0 Nd or p0 Na, respectively). However, semiconductor may contain both donors and acceptors.
EC ED Ei EA EV

EF

ND > NA NC = ND - NA
(here, the subscript C stands for compensation)

This is compensation doping.

Band Structure

68

. Carrier Concentration Charge Neutrality

In complex case as compensation doping, the concentrations of electrons, holes, donors and acceptors can be obtained from space charge neutrality law. The material must remain electrically neutral overall. Therefore, p +N +=n +N 0 d 0 a

The consequence of this is that a material doped equally with donors and acceptors becomes intrinsic again: n0 = p0 = ni because EF = Ei = EG / 2
Band Structure 69

. Carrier Concentration Complete ionization At room temperature, it is almost always assumed that all dopants are ionized, i.e., n0 = ND+ = ND .

Taking into account that EC - ED is only about 50 meV, the probability of that level still being occupied by an electron is extremely low. (the donor atom will easily loose the electron into conduction band)

Band Structure

70

.
Direct and Indirect Bandgap Semiconductors

Dielectrics and semiconductors behave essentially the same way, and as you know, the only difference is the bandgap width. Photons with the energy exceeding Eg are absorbed by giving their energy to electron-hole pairs. They may or may not re-emit the light during the recombination depending on whether the gap is direct or indirect. In direct bandgap semiconductors, electron simply drops from the bottom of EG CB to the top of VB, and excess energy is emitted as a photon (GaAs is an example).
Band Structure 71

EC EV

.
Direct and Indirect Bandgap Semiconductors

In indirect bandgap semiconductors, direct electronhole recombination does not take place. Recombination process occurs in two stages via recombination centers (usually defects) in the bandgap: electron falls from the bottom of CB to the defect level, and then further down to the top of VB. Electron energy is therefore lost in two portions by the emission of phonons (lattice vibrations). This process is also called non-radiative recombination. (Si and Ge are the examples of indirect bandgap semiconductors).

Band Structure

72

.
Direct and Indirect Bandgap Semiconductors

Hence, the absorption coefficient of germanium looks like this: Semiconductors can look metallic if visible photons are all reflected (like Ge) but those with smaller Eg look colored (like CdS because it absorbs only photons with 540 nm).
(mm-1)

104 103 102 10 0.4

0.8 1.2 1.6 Wavelength (m)

This picture is valid for pure materials but impurities can add some extra absorption features...
Band Structure 73

.
Direct and Indirect Bandgap Semiconductors

Impurities create local levels in the bandgap and thus allow transitions with energies smaller than Eg. Recombination can be either radiative (produces photon) or non-radiative (produces phonon) depending on the transition probabilities.

EC hf1 hf2 EV

CB
phonons

EG VB

p-n diodes usually contain a small amount of impurity to promote recombination since Si is indirect material.
Band Structure 74

.
MOSFET

Metal-Oxide-Semiconductor Field Effect Transistor (MOSFET)

MOSFETs make up the vast majority of electronic devices: Pentium contains more than 3 million of them. Cross-section: Gate length: 0.35 m Metal (or poly-Si) oxide (SiO2) Equivalent circuit:
n gate length p-substrate
Band Structure 75

.
MOSFET

Metal-Oxide-Semiconductor Field Effect Transistor (MOSFET) No current flows from the source to the drain because of back-to-back diodes. To make channel conductive, gate electrode is used. There is an insulator between the metal and semiconductor, therefore, electric field builds across oxide layer if potential difference is applied.

Band Structure

76

.
MOSFET

Consider a parallel-plate capacitor: +++++++++++++++ +Q d - - - - - - - - - - - - - - - -Q C = 0 A /d, where 0 - dielectric constant, A - electrode area, d - electrode separation. The charge accumulates on the surface of the plates because either extra electrons or too few electrons are presented close to the plate surface.
Band Structure 77

Q=CV, where C = capacitance.

.
MOSFET

In case of metal plates, charge cannot penetrate too deep into the metal. Consider A=1 cm2, d=1 m, V=10 V. C = 0A/d = 8.8510141/10-4 = 8.8510-10 F; Q = CV = 8.8510-9 Coulombs; Q = nq, thus n = 8.8510-9/1.610-19 = 5.51010 electrons. Assume that typical metal has about 31023 atoms/cm3 at the surface and that one electron per atom can be displaced; then: n = 31023 A l (l - charge layer penetration depth) l = 1.810-13 cm.
Band Structure 78

.
MOSFET

Taking a lightly doped substrate with p = 1015 cm-3, we obtain: l = 5.51010/ 1015 = 5.510-5 cm = m. That means, we can create a space charge layer in the semiconductor under the gate in MOSFET! The charge polarity depends on the gate voltage polarity. Three operation modes can be distinguished: 1) Accumulation mode 2) Depletion mode 3) Inversion mode
Band Structure 79

.
MOSFET

MOSFET: Operation Modes +++++++++

1) Accumulation mode. Negative voltage applied to the gate increases the number of holes at the interface. 2) Depletion mode. A small positive voltage on the gate repels holes. 3) Inversion mode. A large positive voltage on the gate attracts electrons to the interface making it locally n-type!
Band Structure

p-type substrate

+
+ + + +

p-type substrate

++
----------

p-type substrate
80

.
MOSFET

MOSFET: Band diagram

Band diagrams are commonly used to describe work mechanisms of semiconductor devices. Evacuum The work function
ECi

ECs
EFm

Eis EFs EVs

is the same for each material, therefore, EVACUUM -EF= const. The substrate is p-type, therefore, Fermi level is close to VB.

metal

EVi

oxide

semiconductor
Band Structure 81

.
MOSFET

Band diagram Evacuum q q EC EF EV

- electron affinity; q - work function (energy needed to free electron from the solid). What if we bring two materials into contact?
Band Structure 82

.
MOSFET

Band diagram: Metal-metal contact Mo 2 EF2 EF1 EF1 2=4.2 eV

Band Structure

Pt

1 -2=e

EF2 =1.16 V

1=5.36 eV

83

.
MOSFET

Band diagram: Electric Field Effect

Electric field linearly EF0 drops across the solid (in the materials with uniform properties). The band diagram bends V (x) because electron potential V energy E = -eV.

collision collision

EF0-eV

x
Band Structure 84

.
MOSFET

When a voltage is applied: - accumulation mode

ECi EFm qV EVi Eis ECs EFs

MOSFET: Band diagram - inversion mode

ECi

EVs

qV EFm

ECs E EFs is EVs

EF is now closer to E V , that means, semiconductor is locally more ptype, there are more holes near insulator/semiconductor interface.
Band Structure

EVi

Near insulator/semiconductor interface: Ei < EF i.e., n-type (inversion)

85