Phys 446 Solid State Physics Cyclotron resonance and Hall effect in semiconductors

Lecture 10 Nov 16 with both types of carriers

(Ch. 6.8-6.14) Two cyclotron frequencies:
eB eB
ωce = - for electrons ω ch = - for holes
me* m h*
Last time: Statistics of charge carriers in semiconductors.
Electrical conductivity. Mobility.
Cyclotron resonance is used to obtain information
on effective masses/shape of energy surfaces

Suppose the constant energy surface is an ellipsoid

Today: High electric field and hot electrons in revolution.
B is applied at some angle
Optical properties: absorption, photoconductivity,
luminescence The cyclotron frequency is
12
⎡ cos 2 θ sin 2 θ ⎤
ωc = eB ⎢ 2
+ ⎥ -depends on effective masses and angle θ
⎣⎢ mt mt ml ⎦⎥
measuring ωc at various angles gives the
Lecture 10 effective masses

High electric field and hot electrons

Lorentz field for electrons: E Le = 1 J e B 1
ne
for holes: E Lh =− JhB
pe J = nev = neµeE
In steady state, no Jy : Jy = neµ e E Le + peµ h E Lh + E H (σ e + σ h ) = 0
d E ⎛⎜ d E ⎞⎟ ⎛⎜ d E ⎞⎟
= +
(µ e J e − µ h J h )B + E H (neµ e + peµ h ) = 0 dt ⎜⎝ dt ⎟⎠ ⎜⎝ dt ⎟⎠
E L
E (Te ) − E (T )
nµ e pµ h = −eEv − =0
Je + Jh = J x Je = J x Jh = Jx τE
nµ e + pµ h nµ e + pµ h
τE – energy relaxation time
v – electron drift velocity
EH
(µ J − µ e J e )B
= h h EH =
pµ h 2 − nµ e 2
JxB
e(nµ e + pµ h ) e(nµ e + pµ h ) 2 3 3
E (Te ) = k BTe E (T ) = k BT
2 2
Electron drift velocity in Ge vs. electric
2 eτ E µ e 2
pµ h 2 − nµ e 2 - used to determine carrier ⇒ Te = T + E field for different crystallographic
⇒ RH = 3 kB orientations at 300 K (from Landolt-
e(nµ e + pµ h ) 2 concentration and mobility
Boernstein - A. Neukermans, G. S. Kino, Phys.
Rev. B 7 2693 (1973).
 Negative differential conductance and Gunn effect Optical absorption processes
Conduction band in GaAs
1. The fundamental absorption process
• In the lower Γ valley, electrons exhibit a small
effective mass and very high mobility, µ1.
• In the satellite L valley, electrons exhibit a large
effective mass and very low mobility, µ2.
• The two valleys are separated by a small energy
gap, ∆ E, of approximately 0.31 eV.

direct process at k = 0 (zone center)

powerful method to determine the
energy gap Eg

Direct and Indirect Gaps 2. Exciton absorption

Exciton absorption in Ge
(experiment)

fundamental absorption
(theory)
 3. Free carrier absorption 4. Absorption involving impurities

intraband transition – like in metals

can occur even when the photon energy
is below the bandgap
depends on free carrier concentration –
more significant in doped
semiconductors a) neutral donor → conduction band
b) valence band → neutral acceptor
c) valence band → ionized donor
ionized acceptor → conduction band
d) ionized acceptor → ionized donor

Photoconductivity Evaluate g per unit volume through absorption coefficient α and slab
thickness d:
Phenomenon in which a material becomes more conductive due to
the absorption of electromagnetic radiation

"dark" conductivity: σ 0 = e ( n0 µ e + p 0 µ h ) N(ω) – number of photons incident per unit time:

Light absorbed: electron-hole pairs created; Then

carrier concentrations increased by ∆n, ∆p ∆n = ∆p

new conductivity: σ = σ 0 + e∆n( µ e + µ h ) ∆σ σ − σ 0 e∆n( µ e + µ h ) ∆σ αI (ω )τ ' ( µ e + µ h )

= = Change in conductivity: =e
σ0 σ0 σ0 σ0 =ωσ 0
Two opposite processes affecting ∆n:
n − n0 ∆σ ∆σ
• generation of free carriers due to absorption, rate g dn
=g− ∝α and ∝ I (ω )
dt τ' σ0 σ0
• recombination; lifetime of carriers τ'
dn numerical estimate: if
In steady state =0 ⇒ ∆n = n - n0 = gτ'
dt (for Ge), get ∆n ~ 5×1014 cm-3
 Luminescence Summary
™Conductivity of semiconductors:
Radiative recombination of charge carriers
mobility:
Classification by excitation mechanisms:
- photoluminescence ™Cyclotron resonance is used to obtain information on effective masses.
- electroluminescence pµ h 2 − nµ e 2
RH =
- cathodoluminescence ™Hall coefficient: e(nµ e + pµ h ) 2 Hall measurements are used
to determine carrier concentration and mobility.
- thermoluminescence
™In high electric field, the carriers acquire significant energy and become
- chemiluminescence
"hot". This affects mobility and can cause current instabilities (e.g. Gunn
effect caused by negative differential conductivity due to inter-valley transfer)
Same physical processes as for absorption,
™Mechanisms of optical absorption and luminescence.
but in opposite direction
Fundamental absorption occurs above the bandgap.
™photoconductivity – increase of conductivity by generation of additional
carriers by electromagnetic radiation

Carrier Diffusion Electric field (V - potential)

In general, total current in a semiconductor involves both
electrons and holes (in the presence of both a concentration
gradient and an electric field):
By applying a field, all energies will be pushed up by the potential V:

The second term in the above equations is the diffusion current

(Fick’s law). It arises from non-uniform carrier density.

In one dimension, for the negative carrier:

At equilibrium, the drift and diffusion currents are equal:

⇒ can write
Have
Diffusion equation for one carrier type
⇒ ∂p
J p = −eD p + peµ p E (holes, one dimension)
∂x
Variation of p(x) in time is given by continuity equation:

∂p 1 ∂J p
= Gp −U p −
∂t e ∂x
Substitute this into the diffusion equation,
generation recombination flow

Assume there is no external excitation, i.e., Gp = 0.

Einstein
Get ⇒
relation ⎛ ∂p ⎞ p − p0 τ’p - lifetime
Recombination term: Up =⎜ ⎟ =−
⎝ ∂t ⎠ Re comb τ' p of holes
Then ∂p ∂2 p ∂ p − p0
Similarly, for holes = D p 2 − µ p ( pE ) − - Diffusion equation
∂t ∂x ∂x τ' p

∂p p1
1) Stationary solution for E = 0: =0
∂t
hole E≠0
∂ p p − p0
2 injection
Dp 2 − =0
∂x τ 'p E=0
(
− x D pτ ' p )
12 0 x
let p - p0 = p1. Then p1 = p − p0 = Ae LD

The excess concentration decays exponentially with x. 1) Stationary solution for a uniform field E ≠ 0:

(
The distance LD = D pτ ' p )
12
is called the diffusion length ∂ 2 p µ p E ∂p1
−
p
− 12 = 0 ⇒ p1 = Ae −γ x LD
12
∂x 2
D p ∂x LD
LD ⎛⎜ D ⎞⎟
Effective diffusion velocity: vD = = µ p ELD
τ ' p ⎜⎝ τ ' p ⎟⎠ s=
Where γ = 1+ s2 − s and
2Dp
12
⎛ D⎞
Diffusion current: J D = ep1vD = ep1 ⎜ ⎟
γ < 1 ⇒ effective diffusion length LD/γ is larger
⎜τ ' ⎟
⎝ p⎠
 Summary of the semiconductors section Summary of the semiconductors section
™Semiconductors are mostly covalent crystals; They are characterized ™In a doped semiconductor, many impurities form shallow hydrogen-
by moderate energy gap (~0.5 – 2.5 eV) between the valence and like levels close to the conductive band (donors) or valence band
conduction bands (acceptors), which are completely ionized at room T:
™When impurities are introduced, additional states are created in the n = Nd or p = N a

gap. Often these states are close to the bottom of the conduction
band or top of the valence band ™ Conductivity of semiconductors:
™Intrinsic carrier concentration: mobility:

= p = ni
™Magnetic field effects:
strongly depends on temperature.
Cyclotron resonance is used to obtain information on effective
™Fermi level position in intrinsic semiconductor: eB eB
masses. ω ce = - for electrons ω ch = - for holes
me* m h*
pµ h 2 − nµ e 2
™Hall coefficient: RH = Hall measurements are
e(nµ e + pµ h ) 2
used to determine carrier concentration and mobility.

™In high electric field, the carriers acquire significant energy and
become "hot". This affects mobility and can cause current
Basics of selected semiconductor devices:
instabilities (e.g. Gunn effect caused by negative differential
conductivity due to inter-valley transfer) p-n junctions.
™Mechanisms of optical absorption and luminescence: Bipolar transistors.
band-to-band
excitonic Tunnel diodes.

free carrier Semiconductor lasers

impurity-related
Fundamental absorption occurs above the bandgap.
™photoconductivity – increase of conductivity by generation of
additional carriers by electromagnetic radiation
™Diffusion. Basic relations are Fick's law and the Einstein relation
 p-n junction In equilibrium, at zero bias, the chemical potential has to be the
same at both sides: bending of the conduction and valance bands

Charge density near the junction is not uniform:

electrons (majority carriers) from the n-side
and holes (majority carriers) from the p-side
will migrate to the other side through the junction.

These migrated particles leave the ionized impurities behind:

a charged region is formed
The band edge shift across the junction is called the built in voltage Vbi.

Recall that for carrier concentration we had Nc N

32
µ n = µ p ⇒ Ecn − kT ln = Ecp − E g + kT ln v ⇒
32 ⎛ m kT ⎞ ND NA
⎛ m kT ⎞
n = 2⎜ e 2 ⎟ e (µ − Ec ) kT = N c e (µ − Ec ) kT N c = 2⎜ e 2 ⎟
⎝ 2π= ⎠ ⎝ 2π= ⎠
Nc N N N
32 32
Vbi = Ecp − Ecn = E g − kT ln − kT ln v = E g + kT ln A D
⎛ m kT ⎞ ⎛ m kT ⎞ ND NA Nc Nv
p = 2⎜ h 2 ⎟ e (Ev − µ ) kT = N v e (Ev − µ ) kT N v = 2⎜ h 2 ⎟
⎝ 2π= ⎠ ⎝ 2π= ⎠

For n-type semiconductor (n, ND >> p, NA) n = ND ⇒ Note that ⎛ m kT ⎞

32
⎛ mh kT ⎞
32
− E g kT − E g kT
ni2 = 4⎜ e 2 ⎟ ⎜ 2
⎟ e = Nc Nve
N ⎝ 2π= ⎠ ⎝ 2π= ⎠
µ n = Ec − kT ln c
ND

For p-type semiconductor (p, NA >> n, ND) p = NA ⇒ N AND

⇒ Vbi = kT ln
Nv ni2
µ p = Ev + kT ln
NA
Use the above to calculate Vbi
 Reverse bias Forward bias
draws electrons and holes away from the n-side and holes from the “pushes” electrons in the n-side and holes in the p-
p-side. side towards the junction. The depletion width will
become thinner → current flows
The depletion width grows and the junction resistance increases.

For most diodes:

I-V characteristics of a p-n junction Current equation in neutral region (i.e. away from the depletion
layer) is given by the continuity equations:
Assume positive V when it is forward bias:
∂n 1 ∂J n ∂p 1 ∂J p
= Gn − U n + = Gp −U p −
∂t e ∂x ∂t e ∂x

∂n ∂p
For steady case, = =0
∂t ∂t

Current equations (last lecture):

There are two currents from two types of majority carriers, jn and jp

E ~ 0 at area outside the depletion layer ⇒ mostly diffusion current

E ~ 0 outside the depletion layer ⇒
outside the depletion layer.
Diffusion current needs inhomogeneity in carrier density. This is
indeed the case because of recombination. Combine these with the continuity equations:

Can write (subscripts indicate

n and p sides, respectively):
Sufficient to solve only one of these equations, because
np = ni 2 e
(µ n − µ p ) k BT = n 2 e eV k BT ni2 = N c N v e
− E g k BT
i

Also, assume there is no external excitation, i.e., Gn = Gp = 0.

The second equation becomes (last lecture): ∂ 2 pn pn − pn0

Dp − =0
∂x 2
τ 'p
General solution:

pn = 
Ae
(
x D pτ ' p )1 2 + Be − x (D pτ ' p )1 2 + pn0
N
solution for homogeneous equation inhomogeneous part

Now consider the boundary condition in solving this equation.

In the depletion layer,

The current depends on x because of the recombination process.

The current through the diode depends on the geometry (e.g.
length) of the diode.
For simplicity, we can define the current as the current at the
depletion layer (x=xn) because the depletion layer is thin and there
is not too many recombination in this region, i.e. Lp>>xn+xp, and
similarly Ln>>xn+xp.
Current at the depletion layer boundary in the n-side:

As shown earlier, ⇒

The total current is Similarly calculate current at the depletion layer boundary in the p-side
Total current: Omar uses another (equivalent) form :

where
( j p )x= x =
eD p p n 0
[e eV k BT
−1] ( j n )x = x p
=
eDn n p 0
Ln
[e eV k BT
]
−1
n Lp

So the total current is

- saturation current
⎛ Dn n p 0 D p p n 0 ⎞
where the saturation current is j s = e⎜ + ⎟
⎜ Ln Lp ⎟
⎝ ⎠

or, since nn0 pn0 = np0 pp0 = ni2 ⎛ Dn Dp ⎞

j s = eni2 ⎜ + ⎟
⎜ Ln p p 0 L p nn 0 ⎟
⎝ ⎠
- saturation current in terms of majority career concentrations
− E g k BT js may be reduced by choosing larger
Since ni2 = N c N v e
bandgap material

Bipolar junction transistor

Power gain: dPl 2α 2 Rl I e
emitter circuit – forward biased =
collector circuit – reverse biased dPe 1 + ln( I c I e0 )kT e

Emitter current: If we take Ie = 10 mA, Ie0 = 10 µA, kT = 25 meV at 300 K, α ≈ 1,

and Rl = 2 kΩ,
I e = I e0 e eVe kT
get voltage and power gains ~800 and 200, respectively.
injects holes into the base (n-region)

Holes diffuse through the base; some of them decay. Fundamental limitation of bipolar junction transistor – low frequency –
determined by diffusion of holes (electrons in npn case) into the base
Collector current: I c = I c 0 + αI e (α - fraction of holes that survive)
The high-frequency limit beyond which the device cannot function
Ic0 is very small → ca write Ic ≈ α Ie properly, usually lies in the range of tens – hundreds of MHz
Voltage drop across the load is: Vl = α Rl Ie amplification Other types of transistors are needed for higher-frequency range
Voltage gain dVl dVl dI e αRl I e
= =
dVe dI e dVe kT e
 Tunnel diode (very high doping levels) Summary of the semiconductors section
™p-n junction: both electrons and holes diffuse across the junction –
potential barrier develops, called built in voltage Vbi :
N N
Vbi = kT ln A 2 D
ni
™The junction acts as rectifier. The current vs applied voltage V is
eV kT
a) equilibrium b) Reverse bias - c) Small forward c) Large forward I = I 0 (e eVe kT
− 1) Forward: I ≈ I 0 e e Reverse: I ≈ − I 0
large tunneling bias - some bias - no
current tunneling tunneling ™Bipolar junction transistor – two back to back junctions: emitter is
current current forward biased, collector is reverse biased
Works as amplifier: when a signal is applied at the emitter, a current
←I-V characteristics pulse passes through the base-collector circuit. The voltage gain is:
dVl dVl dI e αRl I e
tunneling process is very fast – can = =
dVe dI e dVe kT e
operate at high frequencies (e.g. 10 GHz)
™Tunnel diode is realized when the doping levels in a p-n junction are
very high, so the junction width is very small – tunneling occur.
A region of negative differential resistance exists in forward bias.

Gunn diode Gunn-Oscillation

Gunn effect (discovered by J.B Gunn in 1963): Assume that a small local perturbation
Above some critical voltage, corresponding to E-field of 2 - 4 kV/cm in the net charge arises at t = t0
(in GaAs), the current becomes an oscillating function of time. This results in non-uniform electrical
Cause for this behavior – field distribution
negative differential resistance The electrons at point A, experiencing
an electric field EL1, will now travel to
the anode with velocity v4.
Gunn "diode" is a The electrons at point B is subjected to
bulk device: an electrical field EH1. They will
therefore drift towards the anode with
velocity v2 < v4
sample of length L
cathode anode ⇒The initial charge perturbation will
Oscillation frequency is given therefore grow into a dipole domain,
by the transit time of electrons known as a Gunn domain.
through the device:
Gunn domains will grow while
vd
f0 = propagating towards the anode until a
L stable domain has been formed. www2.hlphys.uni-linz.ac.at/mmm/uebungen/gunn_web/gunn_effect.htm
 Field effect transistor Emission of Light by Semiconductor Diodes
metal In a forward-biased p-n junction
fabricated from a direct band gap
material, the recombination of the
electron-hole pairs injected into
the depletion region causes the
emission of electromagnetic
radiation - a light emitting diode

If mirrors are provided and the concentration of the electron hole pairs (called the
injection level) exceeds some critical value → a semiconductor laser
Edge-emitting laser vertical cavity surface-emitting laser
Body is commonly tied to ground (0 V)
• When the gate is at a low voltage
– p-type body is at low voltage
– Source-body and drain-body diodes are OFF (reverse bias)
– Depletion region between n and p bulk: no current can flow, transistor is OFF
• When the gate is at a high voltage
– Positive charge on gate of MOS capacitor
– Negative charge attracted to oxide in the body (under the gate)
– Inverts channel under the gate to n-type
– Now current can flow through this n-type channel between source and drain
– Transistor is ON