Modelowanie Nanostruktur
Semester Zimowy 2011/2012 Wykład
Jacek A. Majewski
Chair of Condensed Matter Physics Institute of Theoretical Physics Faculty of Physics, University of Warsaw
E-mail: Jacek.Majewski@fuw.edu.pl
Email: jacek.majewski@fuw.edu.pl
Optical Properties of Semiconductor Quantum Structures
Electrodynamics of heterostructures Light propagation perpendicular to the layers Propagation of light along the layers Light absorption in quantum wells Interband transitions
Intraband (intersubband) transitions
Modelowanie Nanostruktur, 2011/2012
Jacek A. Majewski
Wykład 10 – 13 XII 2011
OPTICAL PHENOMENA – BULK SYSTEMS
Optical phenomena
These processes can be described as
the action of a high-frequency macroscopic field on a solid
the interaction between the elementary excitation of the solid and the quanta of the electromagnetic field – the photons
The macroscopic interaction of electromagnetic fields with matter is already contained in the Maxwell equations
Transverse, electromagnetic waves
LIGHT
DEBH
1 Material equations For nonmagnetic solids
The material constant then contains all information about the interaction of the fields with the matter
Absorption Reflection Dispersion
DE Is this equation valid for high-frequency fields?
Optical phenomena
These processes can be described as
the action of a high-frequency macroscopic field on a solid
the interaction between the elementary excitation of the solid and the quanta of the electromagnetic field – the photons
The macroscopic interaction of electromagnetic fields with matter is already contained in the Maxwell equations
Transverse, electromagnetic waves
LICHT
DEBH
1 Material equations For nonmagnetic solids
The material constant then contains all information about the interaction of the fields with the matter
Absorption Reflection Dispersion
DE Is this equation valid for high-frequency fields?
Absorption Coefficient & Measurement of Optical Constants
| |2
I EB E
( ) 0 z
I z I e 2 4
c
dI I
dz The intensity of light is proportional to
with Absorption
coefficient
One can deduce that
I
satisfies the equationPhysical implification: Absorption coefficient is the amount of radiation energy absorbed by a unit volume of solid under unit radiation intensity (photon flux energy) in unit time
Empirical formula for semiconductors: n E4 Gap77 n[2.3 4.6] Energy gap in eV
for T. S. Moss, “Optical Properties of Semiconductors”,
Butterworths Scientific (1959)
The optical constants are often determined by measuring i) Transmition
ii) Reflection
Semiconductor
I0
Ir t
' I (1 ) 0
It R I
0
IrRI
n ( ) ( )
Electron-Photon Interaction – Theory of Direct Transitions
The simplest interaction process – the absorption of a photon by an electron The electron changes its energy and momentum by the energy and momentum of the photon absorbed
Our aim is to calculate the absorption coefficient (or ) due to such elementary process
2Absorption coefficient = the energy absorbed (per unit volume & time) the incident energy (per unit volume & time) ( )
W -the number of photons absorbed per unit volume per unit time -the energy absorbed
( ) W
( , ) 0 exp[ ( )] . .
A r t A e i q r t c c
We represent the incident light by its vector potential
The incident energy flux = uc
uv n 1 2 2
( )
u 8 E H
( ) W ucn
v c
n
Absorption Coefficient and Number of Absorbed Photons
( , ) ( , ) 1 A( , )E r t r t r t
c t
B r t( , ) A r t( , )curlA r t( , ) Coulomb gauge:
( , )r t 0 divA r t( , )00 exp[ ( )] . .
E i A e i q r t c c c
* 2 2
2 0
4 4
u E E A
c
1 2 02 2 1 02
4 4
uc A n A
n n c c
n2
n n n
2 0
( ) 4 c ( )
W n A
2 2
2 0
( ) cn ( ) 4 c ( )
W A
Interaction of light with solids - QM description
To determine we start from the Schrödinger equation W( )1 (ˆ )2 ( ) ( , ) ( , )
2
e r t
p A V r r t i
m c t
pˆ i
ˆ ˆ0 ˆ ' HH H
2
0 1 ˆ
ˆ ( )
H 2 p V r
m ˆ ' e ˆ
H A p
mc
ˆ ˆ
p A A p idivA
Since divA r t( , )0 (chosen gauge)
ˆ ˆ
p A A p
A2- small in comparison to A pˆ
0 0 0
Hˆ i t
Hˆ0n( , )k r E kn( )n( , )k r nk n( , )k r
( ) n( , ) exp[ n( ) ]
nk
t a k t i E k t nk
We now expand in terms of the solutions of the unperturbed problem
Bloch states
|| exp[i m( ) ] '
E k t mk Schr. Eq.
Quantum mechanical calculation of absorption coefficient
1 ˆ
( ', ) ( , ) exp [ ( ') ( )] ' '
m n m n
nk
a k t a k t i E k E k t mk H nk
i
( , 0) 1
a kj a kn( , 0)0
(0) jk
( , , ', '; , )
W j k j k t
j k' '2
|aj'( ', ) |k t
(1)
' '
1 ˆ
( ', ) exp [ j( ') j( )] ' ' '
j
a k t i E k E k t j k H jk
i
2
2 ' 0
1 ˆ
( , , ', '; , ) exp [ ( ') ( )] ' ' ' ' '
t
j j
W j k j k t
i E k E k t j k H jk dt Let us assume: and for all otherIt means, at time t = 0 the electron is described by (is in state ) jk
The probability that at time t it will be in state is then equal to
In the first approximation
Next we examine the matrix element j k H' ' ˆ' jk
Electron-Photon Interaction – Theory of Direct Transitions
The simplest interaction process – the absorption of a photon by an electron
Absorption coefficient = the energy absorbed (per unit volume & time) the incident energy (per unit volume & time) ( )
W -the number of photons absorbed per unit volume per unit time -the energy absorbed
( ) W
The total number of transitions per unit volume and time.
3 3 '
1 2
( ) ( , ', ; , )
jj (2 ) BZ
W d k W j j k t
t
2 0
( ) 4 c ( )
W n A
2 2
2 0
( ) cn ( ) 4 c ( )
W A
Direct Transitions in Semiconductors
are particularly important if the valence band maximum and conduction band minimum lie at the same
k
.cb vb
Such direct transitions are the lowest transition of the absorption spectrum and determine the shape of the absorption edge.
We assume that we can describe the extrema by an effective mass
2 2 2 2 2 2
' * *
'
( ) ( )
2 2 2
j j Gap Gap
j j comb
k k k
E k E k E E
m m
The reduced effective mass of electron and hole
* * * *
'
1 1 1 1 1
comb mj mj mc mv
3 / 2
1/ 2
' 2 2
2
( ) 1 ( )
2
comb
jj Gap
g E
( ) ( EGap)1/ 2 The combined density of states
Direct Optical Transitions: Absorption & Emission
Probability of photon absorption
Probability of photon emission
( , , ; , ) | if( ) |2 ( f( ) i( ) ) W i f k t t M k E k E k
( , , ; , ) | if( ) |2 ( f( ) i( ) ) W i f k t t M k E k E k
( ) 21 [ i(1 f) i(1 f)]
kfi
W W f f W f f
t
Number of absorbed photons2( ) W( )
( ) 2 | vc( ) |2 ( c( ) v( ) )[ ( v( )) ( c( ))]
kcv
W
M k E k E k f E k f E k ( ) 2 | vc( ) |2 ( c( ) v( ) )[ v(1 c) c(1 v)]kcv
W
M k E k E k f f f fDirect Optical Transitions: Absorption & Emission
1 2
| vc( ) | ( c( ) v( ) )
W M k E k E k
t
2
2
2
1 | ( ) | ( ( ) ( ) )
| ( ) | ( [ ( ) ( ) ])
| ( ) | ( ( ) ( ) ) 1
cv v c
vc c v
vc c v
W M k E k E k
t
M k E k E k
M k E k E k W
t
Absorption
(Stimulated) Emission
cb vb
cb vb
f c i v
i c f v
Excitons – General Theory
cb vb
e
h
We consider an insulating crystal at
T = 0 K
Ground State:
Valence band – fully occupied
Conduction band – completely empty Excited State:
Valence band – electron
vk
anihilated (hole created) Conduction band – electronck
createdCoulomb interaction between electron & hole
When this interaction is strong enough, electron & hole bind together
Electron + hole pair = EXCITON
Optical spectra of a semiconductor near the fundamental edge
A b s o rp ti o n c o e ffi c ie n t
Photon Energy
Coulomb enhancement
1s
2s
Energy Gap
Red line – with electron-hole interaction (excitonic effects) Black line – without electron-hole interaction
( EGap)
NANOSTRUCTURES – QUANTUM WELLS
Electrodynamics in Homogeneous Systems
( )
0 0
( , ) cos( ) i q r t . .
E r t eE q r t eE e c c
2 2
0
1 1
4 r ( ) 8 r
w E t E
2 0
1
8 r
I E
In the simplest homogeneous case
The energy of the wave can be characterized by the density of the electromagnetic energy
The intensity of the wave = the energy flux through the unit area perpendicular to the wave vector
q
Optical properties of Quantum Structures
Specific features of optical processes originate from two basic physical peculiarities
Spatial nonuniformity causes specific characteristics
of the interaction of light with matter, including light propagation, absorption, etc..
Electrons in quantum structures have energy spectra different from these of electrons in bulk materials.
Electrodynamics of Heterostructures
E
GapA B A
In heterostructures, both the refractive index
and the band gap, i.e., the fundamental edge of absorption,
vary spacially.
This changes the light propagation and character of the interaction of light with matter.
For example, a layer with a larger refractive index causes Partial reflection of electromagnetic waves propagating through the layer,
localization of electromagnetic modes propagating along the layer.
The electromagnetic fields (modes) in quantum structures are substantially different from plane waves.
Spatial modulation of the bandgap leads to nonuniform absorption and emission of light.
Electrodynamics of Heterostructures
Sizes of quantum structures are always much less than the wavelength of the light in the spectral region of interest
100 200
L A 1000 A
TWO CASES of light interaction with a quantum-well layer
Light propagation
perpendicularly to the layers
Propagation along the layers
( ) E z
z
( ) E z
z
Electrodynamics of Heterostructures Light propagation perpendicularly to the layers
Electromagnetic field depends on the
z
coordinate only1 *
( , ) [ ( ) ( ) ]
2
i t i t
E z t F z e F z e ( , ) 1[ ( ) . .]
2
A z t A z ei t c c
( ) ( )
F z A z c
in out out
I I
I ( )
E z
z
Heterostructure embedded in the optical resonator
Lz
Introduction of a local absorption coefficient is meaningless in this case.
Loss or gain of the light energy can be characterized as the change of the light intensity after the light passes through the layer
to the initial intensity
Electrodynamics of Heterostructures Light propagation perpendicularly to the layers
Another useful characteristic of the interaction of light with matter is the decay (gain) of the energy of standing waves.
The standing waves are formed by the optical resonator in which a heterostructure is embedded.
The decrement (increment) of the mode is
1 dN R N dt N
N – the number of photons of the mode under considerations R – total rate of photon absorption (emission)
2 ,
| | ˆ' | | ( f i )[ ( i) ( f)]
i f
R
i H f E E f E f E ˆ ' e ˆH A p
mc
Electrodynamics of Heterostructures Light propagation perpendicularly to the layers
2 2
,
ˆ
| ( ) | | | z| | ( f i )[ ( i) ( f)]
i f
R A z
i p f E E f E f Ez
L
Since , we can rewrite
R
in the form- the position of the center of the quantum-well layer The number of photons N of the fixed mode can be calculated as the total energy of this mode divided by
2 3 2
1 ( ) | ( ) | ( ) | ( ) |
8 r 8 r
N z F z d r S z F z dz
2 2
2 ,
| ( ) | 1 | |ˆ | | ( )[ ( ) ( )]
( ) | ( ) | i f z f i i f
r
F z i p f E E f E f E
z F z dzS
R
N
z r
L
c
Electrodynamics of Heterostructures Light propagation perpendicularly to the layers
The simplest case of a narrow well, the changes in the refractive index can be neglected and the standing wave is
( ) 0cos cos
E z eF qz
tF z ( ) eF
0cos qz e
- light polarizationq
– a discrete numberz
q n
L
2 2
,
1 ˆ
cos w | | z| | ( f i )[ ( i) ( f)]
z i f
qz i p f E E f E f E
SL
depends on the position of the quantum-well layer!Electrodynamics of Heterostructures Propagation of light along the layers
( ) E z
z
L
z( ) cos( )
E eF z q r t
q
lies in thex,y
planeA narrow-bandgap layer localizes the electromagnetic modes.
The amplitudes of these modes depend on z and decay far away from the layer.
These modes are called the waveguide modes.
Form factor of the mode
For the waveguide mode, one can define the electromagnetic energy per unit area of the waveguide plane as
2 2
1 1
( ) ( ) ( ) ( )
4 r 8 r
w z E z dz z F z dz
The number of photons of this mode is NwwS/( )
Electrodynamics of Heterostructures Propagation of light along the layers
Waveguides modes contd.
One can introduce the total intensity of the waveguide mode:
I wc
effeff / r
c c - is the group velocity of the mode under consideration Now one can define the gain (absorption) coefficient
for the waveguide mode
1 dI
I dx
2 2
2 2
,
4 | ( ) | 1 | | ˆ| | ( )[ ( ) ( )]
( ) | ( ) | r
f i i f
i f r
F z L
i e p f E E f E f E
c z F z dzSL
2 2
| ( ) | ( ) | ( ) |
r
F z L z F z dz
Optical confinement factor
characterizes a portion of the light energy accumulated within the active layer where the photo-transitions take place
1 is always smaller than 1,
The better the optical confinement, the larger the light absorption or gain.
Light Absorption by Confined Electrons
EGap
W B
Ec 1
Ec 2
Ec
Ev 1
Ev 2
Ev
k//
B
A three-layered semiconductor structure (QW) of type I
Band diagram
Two subbands are shown for both bands.
Parabolic dispersion relations are assumed for both the conduction and the simple valence bands.
The first obvious conclusion can be drawn, i.e., that there must be a shift of the interband spectra toward a short-wavelength region
(1) 1
1 1
0
QW Bulk
c v g Gap
E E E E
Light Absorption in Quantum Wells
2 ,
| | ˆ | | (
f i)[ (
i) (
f)]
i f
R i e p f E E f E f E
Bulk: i vk f ck
// //
// 0
( ) 1 ( ) v ( )
v
ik r
v vn
vk n
i r u r e z
S
// //
// 0
( ) 1 ( ) c ( )
c
ik r
c cn
ck n
f r u r e z
S
, , //v
i v n k f c n k, , //c QWs:
- periodic parts of Bloch functions for - two-dimensional (in-plane) wave vectors - the envelope wave functions of the quantized transverse motion
&
0( ) 0( )
v c
u r u r
&
//v //c
k k
0 0
k
&
( ) ( )
vn z cn z
S
- the area of the quantum well layerLight Absorption by Confined Electrons
0
( ) ( )
i i
i u r F r f u
f0( ) r F
f( ) r
Periodic, strongly oscillating functions
Momentum Matrix Element
Smooth functions
0 0 0 0 0
ˆ i i ˆ f f i i f( ˆ f ) f ( ˆ f)
i e p f u F e p u F u F F e pu u e pF
3 * *
0 0
3 * *
0 0
ˆ [ ( ) ˆ ( )][ ( ) ( )]
[ ( ) ( )][ ( ) ˆ ( )]
i f i f
i f i f
i e p f d r u r e pu r F r F r d r u r u r F r e pF r
3 * 3 *
0 0 0 0
ˆ i ˆ f i( ) f( ) i f i( ) ˆ f( )
i e p f u e p u
d rF r F r u u
d rF r e pF r bulkpif
if3 * 3 *
ˆ
bulkif i( )
f( )
if i( ) ˆ
f( ) i e p f p d rF r F r d rF r e pF r
Light Absorption in Quantum Wells Interband Optical Transitions
3 *
ˆ vcbulk i ( ) f( ) i e p f p
d rF r F r For interband transitions:
// // // //
// // //
3 * 3 *
( )
2 *
//
( ) ( ) ( ) ( )
( ) ( )
v c
c v
ik r ik r
i f vn cm
i k k r
vn cm
d rF r F r d re z e z
d r e dz z z
//v, //c
k k
vn cm
// , //
ˆ
v c
bulk
vc k k vn cm
i e p f p
//v //c
k k
The selection rule for two-dimensional wave vectors
This selection rule differs from the selection rule for bulk crystals as a result of the lack of translational symmetry in the z-direction.
Compared with the bulk case, a new factor appears in the matrix element, namely, the overlap integral of the envelope functions from different bands.
Light Absorption in Quantum Wells
Using the matrix elements , one can calculate all characteristics, for geometries presented previously.
i e p fˆ , , ,
// //
// //
2 2
,
, ,
// //
| | | |
( ( ) ( ) )[ ( ) ( )]
v c
v c
bulk
vc vn cm k k
n m k k
cm c vn v vn cm
p
E k E k f E f E
2 2 2
// // //
,
| vcbulk| | vn cm | ( cm( ) vn( ) )[ ( vn) ( cm)]
n m
p d k E k E k f E f E
The energy-conservation law following from the function
// //
( ) ( )
cm vn
E k E k
shows that photo-transitions can involve different subbands from both the valence and the conduction bands.
Light Absorption in Quantum Wells Case of parabolic subbands
2 2
0 //
// *
( )
cm cm 2
e
E k E k m
// 0 2 //2
( ) *
vn vn 2
h
E k E k m
2 2
0 0 //
// // * *
2 2
0 0 //
1 1
( ) ( )
2
2
cm vn cm vn
e h
cm vn
E k E k E E k
m m
E E k
* *
1 1 1
e h
m m
0 0
// 2
2 ( cm vn)
k E E
0 0
cm vn
E E
// //
( ) ( )
cm vn
E k E k
The magnitude of the two-dimensional wave vector corresponding to vertical transitions between
(vn)
and(cm)
subbands Transitions are possible only whenLight Absorption in Quantum Wells Case of parabolic subbands
2
2 2 2 2
// // // // 2 // //
( ( ) ( ) ) 2 ( )
cm vn 2
E k E k k k k k
2 2 // // // //
// //
//
( ) ( )
( )
2
k k k k
k k
k
2 2 0 0
2 ,
// //
| | | | ( )
[ ( ( )) ( ( ))]
bulk
vc vn cm cm vn
n m
vn cm
p E E
L
f E k f E k
0 0
( ) 2 ( )
opt
nm Ecm Evn
The optical density of states
are the step-like functions, which are consistent with the results on the density of states of two-dimensional electrons (holes).
Integration
d k2 //opt( )
nm
Light Absorption in Quantum Wells Case of parabolic subbands
The function also has a steplike shape
( )
Bulk
300 A QW 100 A QW
0 100 200
Absorption coefficient
Photon energy above bandgap [meV]
Light Absorption in Quantum Wells Case of parabolic subbands
( )
can be represented as sum over all pairs of the subbands involved in the photo-transitions
2 2
// //
( , , ) | | | | ( )
[ ( ( )) ( ( ))]
bulk opt
sub vc vn cm nm
vn cm
n m const p
f E k f E k
,
( ) sub( , , )
n m
n mEach contribution is proportional to the overlap integral of the envelope functions
( , , )
sub n m
&
( ) ( )
vn z cm z
These overlap integrals result in new selection rules
Light Absorption in Quantum Wells Selection Rules for Interband Transitions
VB 1 2 3
1 2 3 CB
1-1 3-3 2-1 2-3
1
v 2
v 3
v 1
c 2
c 3
c
Envelope functions
|
vn cn|
2 1
“allowed” transitions for| vn cm |2 1 nm “forbidden” transitions
The quantization of the electron energy leads to major changes in the selection rules and in the intensity of the band-to-band transitions
Intraband Transitions in Quantum Structures
We start by recalling that in an ideal bulk crystal, intraband photo-transitions are impossible because of the energy- and the momentum-conservation laws
&
( ) ( ') ' 0
E k E k k k q
cannot be satisfied simultaneously.
Intraband photo-transitions in a bulk can be induced only by phonons, impurities, and other crystal imperfections.
In contrast to bulk materials, intraband photo-transitions occur in semiconductor heterostructures.
Intraband Transitions in Quantum Structures
c1 c2 c3 CB
Quantum Well
3 * 3 *
0 0 0 0
ˆ i ˆ f i( ) f( ) i f i( ) ˆ f( )
i e p f u e p u
d rF r F r u u
d rF r e pF r , , //i c n k f c n k, ', //'
// //
// 0
( ) 1 c ( ) ik r cn( )
i ck n r u r e z
S
0 0 1
c c
u u
0 ˆ 0 0
c c
u e p u
'
// // // //
3 *
ˆ cn( ) ik r ˆ ik r cn'( ) i e p f
d r z e e pe zIntraband Transitions in Quantum Wells
'
// // //
'
// // //
( )
' ' 2 *
// '
( )
2 *
// '
ˆ ( ) ( ) ( )
( )ˆ ( )
i k k r
x x y y cn cn
i k k r
z cn z cn
i e p f e k e k d r e dz z z
e d r e dz z p z
for
*
( ) '( ) 0 '
cn cn
dz z z nn
'
// // //
' // //
( )
2
// ,
i k k r
d r e k k
' // //
' *
// // , '
ˆ ' z cn( )ˆz cn( )
cnk e p cn k
k ke
dz
z p
z'
// //
k k
If the vector of the polarization of light lies in plane of the quantum well layer (
x,y
– plane), the matrix element is zero and photo-transitions are impossible.e
If the vectror has a
z
-component ( ), intersubband processes take place.e 0
' ' '
// //
' '
// // // //
' '
' '
' '
ˆ ( ) ( ˆ ˆ ˆ ) ( )
( ) ( ) ˆ ( )
x ik yy
ik r ik x
cn x x y y z z cn
ik r ik r
x x x x cn z z cn
e pe z e p e p e p e e z
e k e k e z e e p z