Transport Through Quantum Dots

Lecture Notes DPGSchool Electronic Nanostructures

Oct 812, 2001, Bad Honnef (Germany)
Dr. Tobias Brandes
February 16, 2007
CONTENTS
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Quantum Dots: Basic Classications . . . . . . . . . . . . . . . 1
1.1.1 Semiconductor Based Systems . . . . . . . . . . . . . . . 1
1.1.2 Metallic Systems . . . . . . . . . . . . . . . . . . . . . . 2
1.1.3 Transport vs. Optics . . . . . . . . . . . . . . . . . . . . 2
1.1.4 Nonlinear Transport vs. Linear Transport . . . . . . . . 2
1.1.5 Single Dots vs. Coupled Dots/Dots in more complicated
structures . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Basic Phenomena . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2.1 Orthodox and Classics: Coulomb Blockade (CB), CB
Oscillations and CB staircase . . . . . . . . . . . . . . . 2
1.2.2 The Dot as a Many-Body System (Articial Atom): Level
Structure, Selection Rules, Spin Eects . . . . . . . . . . 2
1.2.3 Kondo Eect: Many Body Theory at its best. . . . . . . 2
1.2.4 The Quantum World enters: Coherent Eects in Cou-
pled Dots, Articial Molecules . . . . . . . . . . . . . . . 2
1.2.5 Best Regards from Quantum Optics: Spontaneous Emis-
sion of Phonons . . . . . . . . . . . . . . . . . . . . . . . 2
1.2.6 Not done here . . . . . . . . . . . . . . . . . . . . . . . . 2
2. The Single Electron Transistor . . . . . . . . . . . . . . . . . . . . . . . 3
2.1 Basic Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.1.1 Charging Energy . . . . . . . . . . . . . . . . . . . . . . 3
2.1.2 Linear Transport V
L
V
R
0: Coulomb Blockade Os-
cillations . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.1.3 Linear Transport: Conductance Formula . . . . . . . . . 5
2.1.4 Experiments: Coulomb Blockade Oscillations . . . . . . . 5
2.1.5 NonLinear Transport: Stability Diagram. Coulomb
Blockade Staircase . . . . . . . . . . . . . . . . . . . . . 6
Contents ii
2.2 Nonlinear transport theory: Anderson Impurity Model . . . . . 6
2.2.1 Rate Equations . . . . . . . . . . . . . . . . . . . . . . . 6
2.2.2 Linear Transport Regime: Coulomb blockade oscilla-
tions and addition energies . . . . . . . . . . . . . . . . . 7
2.2.3 Linear Transport Regime: Peak Shape, Limit
b
, U 8
2.2.4 Nonlinear transport regime . . . . . . . . . . . . . . . . 9
2.2.5 Comparison to Experiments: Nonlinear Transport . . . 10
2.3 Symmetries, Many Body Eects, Spin . . . . . . . . . . . . . . . 11
2.3.1 Symmetries . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.3.2 Exact ManyBody Calculations, Spin Blockade . . . . . 12
3. The Kondo Eect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.1 Kondo Eect in Metals. Impurity Model . . . . . . . . . . . . . 13
3.2 Kondo Eect in Dots . . . . . . . . . . . . . . . . . . . . . . . . 14
3.3 Theory Tool: Greens Functions . . . . . . . . . . . . . . . . . 14
3.4 Unitary Limit at T = 0 : . . . . . . . . . . . . . . . . . . . . . . 15
3.5 Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.6 Further topics . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
4. Coupled Quantum Dots . . . . . . . . . . . . . . . . . . . . . . . . . . 18
4.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
4.1.1 Superposition . . . . . . . . . . . . . . . . . . . . . . . . 18
4.1.2 Entanglement . . . . . . . . . . . . . . . . . . . . . . . . 18
4.2 Theory Tool: Master Equations . . . . . . . . . . . . . . . . . . 19
4.3 Microwave Radiation . . . . . . . . . . . . . . . . . . . . . . . . 20
4.4 Spontaneous Emission of Phonons . . . . . . . . . . . . . . . . . 20
4.4.1 Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
4.5 Equations of Motion . . . . . . . . . . . . . . . . . . . . . . . . 21
4.6 Stationary Current . . . . . . . . . . . . . . . . . . . . . . . . . 24
4.6.1 Experiments in Double Quantum Dots: Spontaneous
Emission of Phonons . . . . . . . . . . . . . . . . . . . . 24
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
1. INTRODUCTION
1.1 Quantum Dots: Basic Classications
1.1.1 Semiconductor Based Systems
Electron density easily controllable. Lower density larger Fermi wavelength
of electrons. Electrons can be conned in small structures. Very high mobility
2DEG. Both lateral and vertical dots possible (pictures).
Vertical dots can have a high symmetry (circular pillar): analogy to real
atoms but instead of 3 dimensions only 2 dimensions. Number of electrons in
those dots very well dened (1 to 10 or more).
Lateral dots have a low symmetry: can be useful for the study of quantum
chaos. Easier to integrate into circuits.
Fig. 1.1: LEFT: Vertical quantum dot (picture from a re-
view article by Kouwenhoven, Austing, and Tarucha
[1]. RIGHT: lateral quantum dot (C. Marcus, Harvard;
http://marcuslab.harvard.edu/research.html).
1. Introduction 2
1.1.2 Metallic Systems
Disorder, mesoscopic uctuations due to disorder.
1.1.3 Transport vs. Optics
Here we only deal with transport, although optics appears at some points
(microwave radiation plus transport etc.)
1.1.4 Nonlinear Transport vs. Linear Transport
1.1.5 Single Dots vs. Coupled Dots/Dots in more complicated structures
1.2 Basic Phenomena
1.2.1 Orthodox and Classics: Coulomb Blockade (CB), CB Oscillations and CB
staircase
IV curve deviates from a straight line (Ohms law). Already known in the
50ies (Gorter 1951) for granular metallic systems. Experiments (60ies/70ies)
dicult because of varying grain sizes.
1.2.2 The Dot as a Many-Body System (Articial Atom): Level Structure,
Selection Rules, Spin Eects
Tarucha, Kouwenhoven. Periodic Table. Spin Blockade.
1.2.3 Kondo Eect: Many Body Theory at its best.
1.2.4 The Quantum World enters: Coherent Eects in Coupled Dots, Articial
Molecules
1.2.5 Best Regards from Quantum Optics: Spontaneous Emission of Phonons
1.2.6 Not done here
Single Electron Box: Interesting , also to discuss dissipation. Superconduc-
tivity. Statistical properties of Dots, uctuations due to disorder. Aharonov
Bohm Eect and Dots. Fano Eect in Dots Electron pumps, turnstiles, adi-
abatic and nonadiabatic pumping of electrons ... Further literature + infos
(hopefully in the very near future) on my homepage http://bursill.phy.umist.ac.uk/.
2. THE SINGLE ELECTRON
TRANSISTOR
2.1 Basic Properties
Device with two tunnel junctions in series. Island in between is called dot.
Literature: chapter 3 in [2], also Beenakker 1991 [3].
Energy scales:
a) internal: U := e
2
/2C charging energy . E: single particle level
spacing . k
B
T: Temperature.
b) external : V
transport
:=
L

R
dierence of chemical potential of left
and right lead. V
gate
gate voltage leads to shift of single particle levels.
Why must the temperature T be low to see charging eects: for k
B
T
U := e
2
/2C, Coulomb blockade can be overcome by thermal excitation such
that charging eects disappear. Example: capacitance of small device with
size 100nm yields typical C 10
16
F corresponding to 10K.
Why must the connection between the dot and the leads be weak (weak
tunnel barriers): Invoke uncertainty principle for total energy E
t
charge
should leak out through resistance R very slowly so that charge on dot is well
dened and can be counted classically (in terms of integer multiples of the
elementary electron charge e. Use E
t
t > and estimate
E
t
= U = e
2
/2C, t = RC R h/e
2
25.8k. (2.1)
2.1.1 Charging Energy
Orthodox theory of Coulomb blockade due to Likarev et al., developed in the
80ies.
In metallic systems, basically classical electrostatic phenomenom with the
discreteness of energy levels neglected.
2. The Single Electron Transistor 4
In semiconductor based systems, discreteness of levels (level separation E
becomes important. In that case, interplay between resonant tunneling and
Coulomb blockade.
We treat the case E = 0 rst: Charge Q = Ne on dot can only change
in integer steps. Electrostatic potential
ext
can be changed continuously.
CC
LL
CC
RR
CC
GG
VV
RR
VV
GG
VV
LL
Ne
Fig. 2.1: Scheme of single electron transistor (SET).
Potential
D
(Q) of dot (central region) with N electrons and charge Ne,
where e is the electron charge:

D
(Q) =
Ne
C
+
ext

ext
=
V
L
C
L
+ V
R
C
R
+ V
G
C
G
C
, C := C
L
+ C
R
+ C
G
(2.2)
From this one obtains the total electrostatic energy U(N) (ground state
energy of N electrons in the dot) as
U(N) =
_
Ne
0
dQ
D
(Q) =
(Ne)
2
2C
Ne
ext
. (2.3)
Tunneling process N N + 1: an electron tunnels into the dot.

dot
(N + 1) := U(N + 1) U(N) =
(N + 1/2)e
2
C
e
ext
(2.4)
is called the electrochemical potential of a dot with N electrons: dierence of
ground state energies.
2. The Single Electron Transistor 5
Condition for current to ow from a left lead to the island and then to the
right lead: To go from the left lead into a dot with N electrons, the electron
has to provide the energy U(N + 1) U(N). Clearly, this is only possible if

L
:= eV
L
> U(N +1) U(N). To go from the dot to the right lead, electron
leaves and has energy U(N + 1) U(N) which has to be above the Fermi
energy of the right lead, otherwise Pauli blocked.
eV
L
> U(N + 1) U(N) > eV
R

L
>
dot
(N + 1) >
R
. (2.5)
For xed V
L
and V
R
, depending on the gate voltage V
G
one is either in the
Coulomb blockade regime (no current ows), or one has a nite current.
2.1.2 Linear Transport V
L
V
R
0: Coulomb Blockade Oscillations
REVIEW for this part : Chapter 5 of [4], and Beenakker 1991 [3].
Varying the gate voltage V
G
leads to Coulomb blockade oscillations. In-
creasing the gate voltage will change the number of electrons one by one.
Small bias voltage V
L
V
R
: Energy degeneracy at
U(N + 1) U(N) = 0
ext
=
(N + 1/2)e
C
. (2.6)
Conductance peaks at these values of
ext
. Of course we are not satised with
that explanation: want to know exact form of the peaks (widths, height, as a
function of temperature, magnetic eld etc etc.).
2.1.3 Linear Transport: Conductance Formula
Always k
B
T h: no level broadening. Two regimes: a) k
B
T E classi-
cal regime. b) k
B
T E resonant tunneling regime. Explicite derivation
for b) below: only one level participates in transport.
Two dierent temperaturedependent curves in a) and b).
2.1.4 Experiments: Coulomb Blockade Oscillations
First observation of single charge tunneling in microfabricated samples by Ful-
ton and Dolan (1987 [5]).
Meirav, Kastner, Wind 1990. First interpretation as Wigner crystal.
Show curves from Meir, Wingreen, Lee.
2. The Single Electron Transistor 6
2.1.5 NonLinear Transport: Stability Diagram. Coulomb Blockade Staircase
Consider again condition for current owing through dot:
eV
L
> U(N + 1) U(N) > eV
R
, V
L
> V
R
(2.7)
Assume symmetric situation V
L
= V
R
= V/2. If V > 0 we must have
eV/2 > U(N + 1) U(N) > eV/2. If V < 0 we must have e|V |/2 >
U(N +1) U(N) > e|V |/2 (current ows in opposite direction). Therefore,
the condition for current to ow is
e|V |/2 > U(N + 1) U(N) > e|V |/2 (2.8)
e|V |C/2 > Q
G
/e (N + 1/2) > e|V |C/2, Q
G
:= C
G
V
G
+ V (C
L
C
R
)/2.
Dene y := eV C/2 and x := Q
G
/e. We have the inequality |y| > xN1/2 >
|y| or |x (N + 1/2)| < |y| as a condition for current to ow (shaded areas
in the Figure). Outside the shaded areas, i.e. within the Coulomb blockade
diamonds, no current ows due to Coulomb blockade. For such values of V
and V
G
, the dot has a xed and stable number of N electrons.
2.2 Nonlinear transport theory: Anderson Impurity Model
Literature: See Meir, Wingreen, Lee 1991 [6]; Weinmann, Hausler, Kramer
1996 [7].
A model that already comprises a lot of the physics: charging energy U,
level spacing E, nite temperatures k
B
T, nite voltages V
transport
= (
L

R
)/e.
Consider dot with two single particle states E
a
and E
b
with 0, 1, or 2
spinless electrons, attached to left and right leads. Dot can be in either of
four states 0, a, b, 2 with energy 0, E
a
, E
b
, and E
a
+ E
b
+ U, where U is the
(Hubbard like) interaction energy. The external voltage is directly controlled
by the gate voltage V
g
, Eq. 2.2, and shifts the energy levels,
a
:= E
a
,

b
:= E
b
.
2.2.1 Rate Equations
Denote rates for internal transitions between dot states due to tunneling as

0a
(rate for internal transition from state a with 1 electron to state 0 with
no electron), etc.
2. The Single Electron Transistor 7
Rate equation (master equation) for probabilities as a function of time:
p
a
=
a0
p
0
(
2a
+
0a
)p
a
+
a2
p
2
p
b
=
b0
p
0
(
2b
+
0b
)p
b
+
b2
p
2
p
2
=
2a
p
a
+
2b
p
b
(
a2
+
b2
)p
2
1 = p
0
+ p
a
+ p
b
+ p
2
. (2.9)
The rates are given by

a0
=
L
f
L
(
a
) +
R
f
R
(
a
),
b0
=
L
f
L
(
b
) +
R
f
R
(
b
) (2.10)

0a
=
L

f
L
(
a
) +
R

f
R
(
a
),
0b
=
L

f
L
(
b
) +
R

f
R
(
b
)

2a
=
L
f
L
(U +
b
) +
R
f
R
(U +
b
),
2b
=
L
f
L
(U +
a
) +
R
f
R
(U +
a
)

a2
=
L

f
L
(U +
b
) +
R

f
R
(U +
b
),
b2
=
L

f
L
(U +
a
) +
R

f
R
(U +
a
).
Here,
f
L/R
() := f(
L/R
), f() =
_
exp
_

k
B
T
_
+ 1
_
1

f() := 1 f(). (2.11)

The stationary current I through the left junction is obtained from the sta-
tionary solution of the master equation (stationary values p
0
, p
a
, p
b
, p
2
)
I = e
__

L
a0
+
L
b0

p
0
+
_

L
2a

L
0a

p
a
+
_

L
2b

L
0b

p
b

L
a2
+
L
b2

p
2
_
. (2.12)
It doesnt matter if we calculate the current through the left or the right
junction: both currents are the same.
It is very easy to solve the system of linear equations Eq. (2.9). The
important ingredients are the transition rates Eq. (2.10).
2.2.2 Linear Transport Regime: Coulomb blockade oscillations and addition
energies
For V
transport
:= (
L

R
)/e 0, the Coulomb blockade oscillations discussed
above are visible as two consecutive peaks at the chemical potentials

dot
(N = 1) =
a
0 =
a

dot
(N = 2) =
a
+
b
+ U
a
=
b
+ U. (2.13)
2. The Single Electron Transistor 8
0
0.5
1
1.5
2
0 0.5 1 1.5 2 2.5 3
c
u
r
r
e
n
t
,

e
l
e
c
t
r
o
n

n
u
m
b
e
r

U=1.0, T=0.01

L
=0.04,
R
=0
E
a
=1.0,E
b
=1.1
current
electron number
0
0.5
1
1.5
2
0 0.5 1 1.5 2 2.5 3
c
u
r
r
e
n
t
,

e
l
e
c
t
r
o
n

n
u
m
b
e
r

U=1.0, T=0.01

L
=0.4,
R
=0
E
a
=1.0,E
b
=1.1
current
electron number
Fig. 2.2: Current and number of (spinless) electrons in the linear (left) and the
nonlinear (right) transport regime through a quantum dot, modeled
as an Anderson impurity with two energy levels E
a
and E
b
. The
current is in units of e
L

R
/(
L
+
R
). The external voltage on
the dot is directly controlled by the gate voltage V
g
, Eq. 2.2, and shifts
the energy levels,
a
:= E
a
,
b
:= E
b
. The transport (bias)
voltage V
transport
:= (
L

R
)/e is kept xed here. The repulsion
energy of two electrons is U.
The distance between the peaks is given by the addition energies E(N) :=
(N) (N 1) ,
E(N = 1) :=
dot
(N = 1)
dot
(N = 0) =
a
E(N = 2) :=
dot
(N = 2)
dot
(N = 1) =
b

a
+ U. (2.14)
2.2.3 Linear Transport Regime: Peak Shape, Limit
b
, U
In this limit, only 0 and a are involved and we have
p
a
=
a0
p
0

0a
p
a
, 1 = p
0
+ p
a
(2.15)
In the stationary case, this becomes
p
0
=

0a

0a
+
a0
, p
a
=

a0

0a
+
a0
I = e
_

L
a0
p
0

L
0a
p
a
_
= e
1

L
+
R
_

L
a0

0a

L
0a

a0
_
= e

L

L
+
R
{f
L
(
a
) f
R
(
a
)} . (2.16)
2. The Single Electron Transistor 9
In the linear transport regime, this expression can be further simplied: With
f
L
(
a
) f
R
(
a
) = f(
a

L
) f(
a

R
) = f(
a

L
) f(
a

L
+ (
L

R
))
= (
L

R
)(f

(
a

L
)) + O(
L

R
)
2
, (2.17)
we nd
I = e(
L

R
)

L

L
+
R
1
4k
B
T coth
2
_
a
L
2k
B
T
_ + O(
L

R
)
2
. (2.18)
The linear conductance G is dened as the limit
L

R
:= ,
G := lim
(
L

R
)0
I
e(
L

R
)
, (2.19)
so that we obtain
G =

L

L
+
R
1
4k
B
T coth
2
_
a
2k
B
T
_. (2.20)
This describes a temperaturedependent resonance peak, when the gate volt-
age V
g
and with it , Eq. 2.2, and the energy
a
= E
a
is swept through the
chemical potential or, alternatively,
a
is kept xed and is swept through
(see Fig. 2.3).
EXERCISE: Repeat and check the steps that lead to the conductance formula
Eq. (2.20).
2.2.4 Nonlinear transport regime
Interpretation of the results from Fig. 2.2: The chemical potentials are

dot
(N = 1) =
a
0 =
a

dot
(N = 2) =
a
+
b
+ U
a
=
b
+ U. (2.21)
for the transition N N + 1 between groundstates, and

dot
(N = 1) =
b
0 =
b

dot
(N = 2) =
a
+
b
+ U
b
=
a
+ U. (2.22)
2. The Single Electron Transistor 10
0
0.01
0.02
0.03
0.04
0.05
0.06
0.4 0.6 0.8 1 1.2 1.4 1.6
c
u
r
r
e
n
t

U=1.0

L
=0.01,
R
=0
T=0.10
T=0.10
T=0.05
T=0.05
E = 0.1
0
0.5
1
1.5
2
0 0.5 1 1.5 2 2.5 3
c
u
r
r
e
n
t
V
transport
U=1.0, T=0.01
E
a
=1.0

R
=0, =0
E=0.1
E=0.5
Fig. 2.3: LEFT: Current in the linear transport regime through a quantum
dot (as above). Comparison is made between the exact solution of the
master equation(solid lines) and the analytical form Eq. 2.20 which
works well for k
B
T E, where
E
is the separation between the
two levels. As above, the Coulomb repulsion is U = 1 here, and
E = 0.5 for the two lower curves, where the agreement is perfect.
For a smaller level separation, E = 0.1 and T = 0.05 (upper curve),
the agreement with the analytical form (for T = 0.05) is not good.
RIGHT: Coulomb staircase , i.e. currentvoltage characteristics of
the same system for xed voltage and variable transport voltage
V
transport
:=
L

R
.
for the transitions N N + 1 between ground and excited states. Increasing
the gate voltage increases and the following enter into the transport window
[
R
,
L
]: rst,
dot
(N = 1) one transport channel. Then,

dot
(N = 1)
two channels, current even larger. For the next current peak one has to increase
until
dot
(N = 2) (not

dot
(N = 2)) enters the window: immediately two
transport channels (large current). Further increase of shifts

dot
(N = 2) out
of the window only one channel, further increase (
dot
(N = 2) <
R
) leads
to zero current. Somewhat tricky when visualised in energy level picture.
2.2.5 Comparison to Experiments: Nonlinear Transport
In semiconductor dots: Johnson, Kouwenhoven, de Jong, van der Vaart, Har-
mans, Foxon 1992 [8]. Weis, Haug, v. Klitzing, Ploog 1992, 1993 [9, 10].
See double Coulomb staircase: not only dot ground states but also excited
states.
2. The Single Electron Transistor 11
2.3 Symmetries, Many Body Eects, Spin
Beyond the constant interaction model. Anderson impurity model above:
addition energy for one electron is the sum of (constant) Coulomb energy U
and single particle energy dierence
b

a
= E. Not much information on
the dot contained in the addition energy spectrum.
2.3.1 Symmetries
Fig. 2.4: a) Current through a circular dot (pillar). The number of electrons
N in the dot is indicated. b) is a simple shell model picture, giving
rise to a periodic table of articial elements . This picture is taken
from a review article by Kouwenhoven, Austing, and Tarucha [1].
Literature: Kouwenhoven, Austing, Tarucha 2001 [1].
2d circular symmetric dots: single particle states (orbitals) are 2d harmonic
2. The Single Electron Transistor 12
oscillator states |nl, eigenstates of single particle potential
V (r) =
1
2
m

2
0
r
2
. (2.23)
Gives rise to a shell structure in the addition energies, similar to 3d real atoms
where electrons move in the spherically symmetric 1/r potential of the nucleus
(states |nlm).
Magic numbers N = 2, 6, 12, ...: more energy is needed to add an electron
to a dot with 2, 6, 12,... electrons. This shell structure is due to degeneracies in
single particle energy spectrum (FockDarwin, perpendicular magnetic eld
can be included): single particle eect.
Phenomenological model: three energy scales single particle energy, direct
Coulomb energy, and exchange energy for electrons with parallel spin. The
latter is an extension to include manybody eects beyond the simple charging
model. Seems to work surprisingly well.
Periodic Table of articial atoms can be created and controlled within
one and the same quantum dot. Application of a perpendicular magnetic eld
to resolve dierent energy scales.
2.3.2 Exact ManyBody Calculations, Spin Blockade
Literature: Weinmann et al. [11] . Spin eect in tunneling: tunneling of an
electron with total spin 1/2, spin projection 1/2 into a dot with many
electron state with total initial spin S
i
, total initial spin projection M
i
: nal
manyelectron state must have with one electron more (or less) and total
spin projection M
j
increased (or decreased) by 1/2. Tunnel rate must contain
ClebschGordan coecient
S
i
, M
i
;
1
2
,
1
2
|S
j
, M
j
(2.24)
If total spins of groundstates with successive electron numbers n n1 dier
by more than 1/2: Coulomb oscillation peak is suppressed (linear transport,
spinblockade type II). Not possible in one dimension (LiebMattis Theorem:
spin of ground state is always 0 or 1/2).
Additional eects in nonlinear transport: negative dierential conduc-
tance possible.
Experiments: H uttel, Qin, Holleitner, Blick, Neumaier, Weinmann, Eberl,
Kotthaus [12].
3. THE KONDO EFFECT
3.1 Kondo Eect in Metals. Impurity Model
Kondo 1964: Calculation (second order perturbation theory) of resistance
R(T) of a metal with magnetic impurities (localised spins, example: magnetic
alloy CuMn), increases logarithmically for small T, R(T) = R
0
+R
1
ln(k
B
T/D)+
R
2
(T/T
p
)
5
, R
1
< 0, R
0
residual resistance due to impurites, term T
5
phonon
contribution resistance minimum, known since the 30ies.
Literature here: Yosida chapter 16 [13]
Higher order perturbation theory: terms ln
n
(k
B
T/D) appear, can be summed
up as a geometric series with a result that diverges below the Kondo temper-
ature T
K
. Haldane 1978: within the Anderson impurity model (one electron
with spin up or spin down on a level
0
, broadened due to coupling to
Fermi sea, double occupancy prevented by Coulomb repulsion U)
T
K
=
1
2

U exp
_

0
(
0
+ U)
U
_
. (3.1)
Spin ip processes : consider
0
(Fermi energy of metal). Spin-up elec-
tron can tunnel out of the impurity into the metal to occupy a virtual state
(energetically forbidden, but allowed in QM due to energytime uncertainty
relation). Electron then replaces by metal electron with spin down spin
ip. Leads to extra Kondo resonance in the local impurity density of states
at the Fermi energy of the metal). This drastically changes (linear) transport
properties (which are determined by states at the Fermi surface). Important
property of that phenomenon: single parameter scaling, i.e.
R/R
0
= f(T/T
K
), (3.2)
i.e. the whole temperature dependence is governed by one universal function
f with the material parameters only entering into the value for the Kondo
temperature T
K
.
3. The Kondo Eect 14
3.2 Kondo Eect in Dots
Glazman, Raikh 1988 [14]; Ng, Lee 1988 [15]: Consider Anderson impurity
model as a very simplied model for a dot: one spin-degenerate single particle
level
0
with strong Coulomb repulsion U:
H =

;kL,R

k
c

k
c
k
+

0
c

+ Un

;kL,R
_
V
k
c

k
c

+ H.c.
_
. (3.3)
Linear transport Fermi level of reservoirs
L
=
R
= with
0
< < +U
at very low temperatures: there is one electron on the dot. The conductance
G, however, is not zero (as one would expect due to the Coulomb blockade),
but
G =
2e
2
h
4
L

R
(
L
+
R
)
2
f(T/T
K
) (3.4)
with f(0) = 1. Here,

L(R)
:=
L(R)
(),
L(R)
() := 2

kL(R)
|V
k
|
2
(
k
) (3.5)
is the tunnel rate between the dot and the left (right) lead. In particular,
for symmetric coupling to the leads,
L
=
R
, G = 2e
2
/h corresponding to
complete transmission.
What happens to the system at T = 0: new correlated many body ground
state with one electron on the impurity, forming a spin singlet with the elec-
trons of the leads.
3.3 Theory Tool: Greens Functions
Literature: Book by Haug/Jauho [16]; Meir, Wingreen 1992 [17], 1994 [18].
The stationary current I through the dot can be expressed with the help of
the local density of states

() on the dot:
I =
e

d [f
L
() f
R
()]

L
()
R
()

L
() +
R
()

()

() :=
1

ImG
r

(), G
r

(t) = i(t)c

(t)c

(0) +c

(0)c

(t). (3.6)
3. The Kondo Eect 15
The last line denes the retarded Greens function G
r
of the dot. It has to be
calculated in presence of both interaction (U) and the coupling to the leads
(tunnel matrix elements V
k
). Even for this simple model the Greens function
in general cant be calculated exactly.
3.4 Unitary Limit at T = 0 :
At zero temperature T = 0, it is possible to make an exact prediction based
for the conductance when
L

R
= and the Greens function has to
be evaluated at the Fermi energy . In fact, this case is analogous to the
situation of a magnetic impurity in a metal, and this analogy has been used by
Glazman/Raikh and Ng/Lee. Basically, at T = 0 and in equilibrium
L

R
,
there are not many possibilities for scattering. The self energy

of the dot
at the Fermi energy fullls
Im

( i) = (
R
+
L
)/2, (3.7)
i.e. is only determined by elastic scattering. As a result, it is possible to for-
mulate a relation between the scattering properties of the Anderson impurity
and its occupation number. This Friedel Sum Rule can be exploited [15, 19] in
the regime
0
< <
0
+ U with one electron on the impurity: It follows
G =
2e
2
h
4
L

R
(
L
+
R
)
2
sin
2
_

2
n
d
_
(3.8)
with n
d
= n

+n

= 1. In particular, for symmetric coupling to the leads,

L
=
R
, G = 2e
2
/h corresponding to complete transmission. Eq. (3.8) looks
quite noninteracting but the underlying physics isnt: What happens to the
system at T = 0, is the emergence of a new correlated many body ground state
with an odd number of electrons on the impurity, forming a spin singlet with
the electrons of the leads. This is the Fermi liquid xed point of the Kondo
problem. At very low but nite temperatures the prediction for the deviation
from this point was made by Nozi`eres back in the 70ies: in our context here,
it reads
G =
2e
2
h
4
L

R
(
L
+
R
)
2
_
1

2
T
2
T
2
K
_
, (3.9)
where the Kondo temperature T
K
was dened above.
3. The Kondo Eect 16
On the other hand, approaching from high temperatures T T
K
, per-
turbation theory for the Greens function yields logarithmic divergences of the
type ln(T/D) (D is a cuto due to, e.g., the nite bandwidth of the conduction
band). These logarithmic singularities get even worse in higher order pertur-
bation theory. There is, however, a remedy to systematically deal with this
problem: renormalization group theory . This allows to establish the connec-
tion between the high temperature T T
K
and low temperature T T
K
regime, yielding
G =
2e
2
h
4
L

R
(
L
+
R
)
2
f(T/T
K
) (3.10)
with f(0) = 1.
3.5 Experiments
Since 1998, a lot. to name a few: GoldhaberGordon, Gores, Kastner, Shtrik-
man, Mahalu, Meirav 1998 [20]; W. G. van der Wiel, S. De Francheschi, T.
Fujisawa, J. M. Elzerman, S. Tarucha, and L. P. Kouwenhoven 2000 [21]
unitary limit:
3.6 Further topics
Kondo eect in an integer spin quantum dot: spin S = 1 of triplet
state also screened by series of cotunneling events. Spin S = 1 triplet
can occur due to Hunds rule (exchange energy wins over single particle
spacing).
Experiment: Sasaki, De Franceschi, Elzerman, van der Wiel, Eto, Tarucha,
Kouwenhoven 2000 [22]: vertical rectangular pillar, number of electrons
N in dot known exactly. Conductance as a function of magnetic eld B
and gate voltage B: spintriplet to spinsinglet transition in the ground
state.
For absence of evenodd behaviour in Kondo eect see also Schmid, Weis,
Eberl, v. Klitzing 2000 [23]
Theory: Eto, Nazarov 2000,...
Nonlinear transport and Kondo eect. Dephasing and Kondo eect.
3. The Kondo Eect 17
Fig. 3.1: LEFT: Unitary limit of the Kondo eect in quantum dots: exper-
imental results by van der Wiel et al [21]. Conductance G in units
of e
2
/h as a function of the gate voltage for dierent temperatures
(T = 15 mK for the black trace, up to T = 800 mK for the red
trace. Inset: logarithmic Tdependence of Gfor one xed gate votage.
RIGHT: Conductance as a function of temperature T for three dif-
ferent gate voltages V
g
. Inset: scaling of the conductance to a single
curve as a function of T/T
K
, where T
K
is the Kondo temperature that
depends on V
g
.
Magnetic eld and Kondo eect; AB ring and Kondo eect.
Many level dot and Kondo eect; Kondo eect in coupled dots: not spin,
but pseudo spin (dot 1 or dot 2).
4. COUPLED QUANTUM DOTS
4.1 Motivation
Why coupled quantum dots: motivation from quantum logic gates, exploiting
two core quantum mechanical phenomena a) superposition b) entanglement.
4.1.1 Superposition
A (single) Qubit is a superposition (linear combination) | = a|0 + b|1 in
the 2d complex Hilbert space C
2
. Example: |0 = | , |1 = | (a spin 1/2).
At the moment, many if not perhaps most of the activities on solid state based
qubits are based on the spin. Other example: |0 = |L, |1 = |R, an electron
with xed spin in either a left (L) or a right (R) quantum dot (charge based
qubit).
4.1.2 Entanglement
Deepest dierence between classical physics and quantum physics. Mathemat-
ically again due to the linear structure of QM (superposition principle). In QM
we describe a system composed of more than one particle by tensor products
of Hilbert spaces and not direct products: Example: Hilbert space of 2 Qubit
states for two spins 1/2 basis of 1 singlet and 3 triplet states
|S
0
:=
1

2
(| | )
|T
1
:= |
|T
0
:=
1

2
(| +| )
|T
1
:= | . (4.1)
|T
0
and |S
0
are entangled: we cant say if the rst spin is up or down (same
for second spin). |T
1
and |T
1
are not entangled. Gets more complicated if
we have electrons with spin: indistinguishable particles.
4. Coupled Quantum Dots 19
Quantum computing : basically, it is sucient to have 1qubits and 2
qubits.
Experiments on entanglement are most advanced in quantum optics , ion
traps etc. Entanglement of N = 2, N = 4 ions realized in ion traps (Sackett
et al 2000) [24]. Keywords: nonclassical Schrodinger cat states, control of
entanglement. Aim: to do that in a solid state environment, where things can
be (hopefully) scaled up to large arrays of qubits.
4.2 Theory Tool: Master Equations
What has to go into a coupled quantum dot theory: 1. strong interactions 2.
coupling between the dots 3. coupling to the external world (leads) 4. coupling
to other degrees of freedom such as bosonic excitations, photons, phonons.
Noone has succeeded to do all of them, although general framework does
exist: Greens function formalism. In the following, we concentrate on point
4 (relation to decoherence). Treat coupling to the leads in lowest order per-
turbation theory (T T
K
, no Kondo physics). Here only simplest case of
two coupled quantum dots. Extension to many dots possible, see Wegewijs,
Nazarov 1999 [25].
Literature here: Stoof, Nazarov 1996 [26]; Gurvitz, Prager 1996 [27]; Gurvitz
1998 [28], Brandes [29, 30].
Coulomb eects also dealt with in extremely simplied manner: Assume
two dots coupled by a tunnel barrier. Coulomb charging energy U in small
dots typically U 1 meV V
transport
= (
L

R
)/e: assuming there are N
electrons in the left and M electrons in the right dot ground state |0 =
|N, M. Additional transport electron is either on the left or on the right dot
states |L = |N + 1, M,|R = |N, M + 1. Therefore, only three states
are involved! Nevertheless, including dissipation, microwave radiation etc still
complicated enough to be of interest.
Describe the space spanned by the two states L and R with the help of
(pseudo) spin Pauli matrices: Hamiltonian for this subspace is
H
0
=

2

z
+ T
c

x
, (4.2)
where
z
= |LL| |RR| and
x
= |LR| + |RL|. Here, =
L

R
is
the dierence of the ground state energies of |L and |R, and T
c
describes the
coupling between the two dots.
4. Coupled Quantum Dots 20
4.3 Microwave Radiation
Eigenvectors and energy eigenvalues of H
0
are
| =
1
N

[2T
c
|L + ()|R] , N

:=
_
4|T
c
|
2
+ ()
2

:=
1
2
, :=
_

2
+ 4|T
c
|
2
. (4.3)
Microwave experiment should be able to detect the squareroot form of : Ex-
periment by Oosterkamp, Fujisawa, van der Wiel, Ishibashi, Hijman, Tarucha,
Kouwenhoven [31].
4.4 Spontaneous Emission of Phonons
For metals: weak electronphonon coupling, electronphonon scattering rates
go to zero when the temperature T 0. Reason: at T = 0, no occupied states
above the Fermi level E
F
, all states below E
F
are occupied, scattering would
be only possible by absorption of phonons. Absorption is proportional to Bose
distribution n
B
() = [exp(/k
B
T) 1]
1
0 for > 0 and T 0.
Quantum dots interact with the semiconductor substrate phonons. In cou-
pled quantum dots (threestate basis as discussed above) there is no Fermi
distribution. For example, left dot can be higher in energy than right dot
even at zero temperature there is spontaneous emission.
4.4.1 Model
We describe the coupling to phonons by a Hamiltonian
H =

2

z
+ T
c

x
+
1
2

z
A +

Q
a

Q
a
Q
, A :=

Q
g
Q
_
a
Q
+ a

Q
_
, (4.4)
Here,
Q
are the frequencies of phonons, and the g
Q
denote interaction con-
stants. Although not exactly solvable, the model is quite well understood for
closed systems [32] (isolated dots with one additional electron). The coupling
to external leads oers the possibility to study its nonequilibrium properties,
such as the inelastic stationary current through the dots.
4. Coupled Quantum Dots 21
Fig. 4.1: Left: Stationary current through a microwave irradiated double
quantum dot (zero transport voltage) for dierent microwave fre-
quencies f as a function of the energy dierence (= E here)
in the experiment by Oosterkamp and co-workers [31]. Positive or
negative peaks occur whenever hf matches the energy dierence,
hf = =
_

2
+ 4T
2
c
, between bonding and anti-bonding state in
the double dot, cf. RIGHT: From the position of the current peaks,
the relation ( E) =
_
(hf)
2
4T
2
c
is tested for various interdot
coupling constants T
c
(denoted as T in the picture). The pictures are
taken from [31].
4.5 Equations of Motion
We describe the dynamics of the double dot by a reduced (eective) statistical
operator (t). The coupling to the reservoirs is described by the additional
4. Coupled Quantum Dots 22
Hamiltonian
H
res
+ H
V
, H
res
=

L
k
c

k
c
k
+

R
k
d

k
d
k
(4.5)
H
V
=

k
_
V
k
c

k
s
L
+ W
k
d

k
s
R
+ c.c.
_
, s
L
:= |0L|, s
R
:= |0R|.
(4.6)
We start from the Liouvillevon Neumann equation for the density operator
of the total system (DOT + PHONONS + RESERVOIRS),
d
dt
(t) = i[H, (t)]. (4.7)
The rst step: transformation into the interaction picture with respect to
H
0
:= H
B
+ H
res
,

O(t) = e
iH
0
t
Oe
iH
0
t
for any operator O. Write the total
Hamiltonian H = H + V , and Eq. (4.7) becomes
d
dt
(t) = i[

V (t), (t)]. (4.8)

Iteration of Eq. (4.8),
d
dt
(t) = i[

V (t), (0)]
_
t
0
dt

V (t), [

V (t), (t)]
_
. (4.9)
We dene the eective density operator for the double dot,
(t) = Tr
res,phonon
(t). (4.10)
Now, second order Born and Markov approximation in the perturbation V
yields [26, 30]

LL
(t) = iT
c
[
LR
(t)
RL
(t)] +
L
[1
LL
(t)
RR
(t)]

RR
(t) = iT
c
[
RL
(t)
LR
(t)]
R

RR
(t),
where the tunnel rates are

L
= 2

k
|V
k
|
2
(
L
k

L
)
R
= 2

k
|W
k
|
2
(
L
k

R
). (4.11)
4. Coupled Quantum Dots 23
and we allow for tunneling from a left reservoir at rate
L
into the left dot,
and from the right dot to the right reservoir at rate
R
.
For the remaining equation for the odiagonal element
LR
=

RL
, one
has to choose between perturbation theory in g
Q
(weak coupling, PER), or
in T
c
in a polarontransformed frame (strong coupling, POL) . In general, no
exact solution of the model is available: this is the case even for coupling to
one bosonic mode only (g
Q

Q,Q
0
, Rabi Hamiltonian).
For the spinboson problem with
R/L
= 0, it is wellknown that POL
is equivalent to a doublepath integral noninteracting blip approximation
(NIBA) that works well for zero bias = 0 but for = 0 does not coincide
with PER at small couplings and low temperatures. In the following, we
compare both approaches for
R/L
= 0 and nd nearly perfect agreement for
very large T
c
, a regime that has been tested experimentally recently [33].
The standard Born and Markov approximation with respect to A yields
d
dt

PER
LR
(t) = [i
p

R
/2]
LR
(t) + [iT
c

]
RR
(t) [iT
c

+
]
LL
(t).
Here, the rates are

p
:= 2
T
2
c

2
() coth (/2) ,

:=
T
c

2
() coth (/2)
T
c

2
()
() :=

Q
|g
Q
|
2
(
Q
). (4.12)
where :=
_

2
+ 4T
2
c
is the energy dierence of the hybridized levels, and
= 1/k
B
T the inverse phonon equilibrium bath temperature. Note that beside
the odiagonal decoherence rate
p
, there appear terms

in the diagonals
which below turn out to be important for the stationary current.
On the other hand, the polaron transformation [30] leads to an integral
equation

POL
LR
(t) =
_
t
0
dt

e
i(tt

)
_

R
2
C(t t

)
LR
(t

)
+ iT
c
(t

) {C(t t

)
LL
(t

) C

(t t

)
RR
(t

)}
_
,
where
C() := exp
_

_

0
d
()

2
[(1 cos t) coth(/2) + i sin t]
_
. (4.13)
4. Coupled Quantum Dots 24
4.6 Stationary Current
Calculated as I
stat
= e2T
c
Im
LR
(z = 0) from Laplace transforming the equa-
tions of motion;
I
stat
=
e
L

R
G
+

L
G

+ (
L
+
R
)G
+

R
. (4.14)
has to be used with either the perturbative result (small ep coupling) or the
polaron transformation (small interdot coupling),
G
(PER)

:= 2T
c
Im
iT
c

i
p

R
/2
(4.15)
or
G
(POL)
+
:= 2T
c
Im
iT
c
C

1 + (1/2)
R
C

, G
(POL)

:= 2T
c
Im
iT
c
C

1 + (1/2)
R
C

, (4.16)
where C

:=
_

0
dte
it
C(t). More details in [30, 34].
Most interesting regime at low temperatures: Experiment by Fujisawa et
al 1998 [33].
4.6.1 Experiments in Double Quantum Dots: Spontaneous Emission of Phonons
We now shortly describe a recent experimental realization of a twolevel system
in a semiconductor structure. This experiment has been performed in coupled
articial atoms, that is coupled quantum dots, by Fujisawa and coworkers at
the Technical University of Delft (Netherlands) in 1998.
The double quantum dot is realized in a 2DEG AlGaAsGaAs semiconduc-
tor heterostructure, see Fig.(4.2). Focused ion beam implanted inplane gates
dene a narrow channel of tunable width which connects source and drain (left
and right electron reservoir). On top of it, three Schottky gates dene tunable
tunnel barriers for electrons moving through the channel. By applying nega-
tive voltages to the left, central, and right Schottky gate, two quantum dots
(left L and right R) are dened which are coupled to each other and to the
source and to the drain. The tunneling of electrons through the structure is
large enough to detect current but small enough to have a welldened number
of electrons ( 15 and 25) on the left and the right dot, respectively. The
Coulomb charging energy ( 4 meV and 1 meV) for putting an additional
electron onto the dots is the largest energy scale, see Fig.(4.2).
4. Coupled Quantum Dots 25
Fig. 4.2: Left: Schematic diagram of a double gate single electron transistor
(double quantum dot) by Fujisawa and Tarucha (1997). The 2DEG is
located 100 nm below the surface of an AlGaAs/GaAs modulation
doped heterostructure with mobility 8 10
5
cm
2
(Vs)
1
and carrier
concentration 3 10
11
cm
2
at 1.6 K in the dark and ungated. Ga
focused ion beam implanted inplane gates and Schottky gates dene
the dot system. A double dot is formed by applying negative gate
voltages to the gates GL, GC, and GR. The structure can also be
used for singledot experiments, where negative voltages are applied
to GL and GC only.
Right: Top view of the double quantum dot. Transport of electrons is through
the narrow channel that connects source and drain. The gates themselves have
widths of 40 nm. The two quantum dots contain approximately 15 (Left, L)
and 25 (Right, R) electrons. The charging energies are 4 meV (L) and 1 meV
(R), the energy spacing for single particle states in both dots is approximately
0.5 meV (L) and 0.25 meV (R).
By tuning simultaneously the gate voltages of the left and the right gate
while keeping the central gate voltage constant, the double dot switches be-
tween the three states |0 = |N
L
, N
R
, |L = |N
L
+ 1, N
R
, and |R =
|N
L
, N
R
+ 1 with only one additional electron either in the left or in the
right dot (see the following section, where the model is explained in detail).
The main experimental trick is to keep the system within these states
and to change only the energy dierence =
L

R
of the dots without
changing too much, e.g., the barrier transmissions. The measured average
spacing between singleparticle states ( 0.5 and 0.25 meV) is still a large
energy scale compared to the scale on which is varied. The largest value of
4. Coupled Quantum Dots 26
is determined by the sourcedrain voltage which is around 0.14 meV. The
main ndings are the following:
1. At a low temperature of 23 mK, the stationary tunnel current I as a
function of shows a peak at = 0 with a broad shoulder for > 0 that
oscillates on a scale of 20 30eV, see Fig.(4.3).
2. For larger temperatures T, the current increases stronger on the ab-
sorption side < 0 than on the emission side. The data for larger T can
be reconstructed from the 23 mK data by multiplication with EinsteinBose
factors (the Planck radiation law) for emission and absorption.
3. The energy dependence of the current on the emission side is between
1/ and 1/
2
. For larger distance of the left and right barrier (600 nm on a
surface gate sample instead of 380 nm for a focused ion beam sample), the
period of the oscillations on the emission side appears to become shorter.
Those who are interested in more details: T. Fujisawa, T. H. Oosterkamp,
W. G. van der Wiel, B. W. Broer, R. Aguado, S. Tarucha, and L. P. Kouwen-
hoven, Science 282, 932 (1998).
4. Coupled Quantum Dots 27
0
0.5
1
1.5
2
2.5
-100 -50 0 50 100 150 200
C
u
r
r
e
n
t
(
p
A
)
(eV)

R
=
L
= 0.03 eV
T
c
= 3.0 eV
g = 0.02

c
=100 eV

d
= 10 eV
PER, T=300 mK
POL, T=300 mK
PER, T= 23 mK
POL, T= 23 mK
g=0
Fig. 4.3: Left: Current at temperature T = 23mK as a function of the energy
dierence =
L

R
in the experiment by Fujisawa and coworkers.
The total measured current is decomposed into an elastic and an
inelastic component. If the dierence between left and right dot
energies
L
and
R
is larger than the sourcedrainvoltage, tunneling
is no longer possible and the current drops to zero. The red circle
marks the region of spontaneous emission of phonons. Phonons are
the quanta of the lattice vibrations of the substrate (the whole semi-
conductor structure), in the same way as photons are the quanta of
the electromagnetic (light) eld. The spontaneous emission is char-
acterized by the large shoulder for > 0 with an oscillationlike
structure on top of it. Right: Stationary current through double
dot, calculated in perturbation theory (PER) or in the polaron ap-
proach (POL).
BIBLIOGRAPHY
[1] L. P. Kouwenhoven, D. G. Austing, and S. Tarucha, rep. Prog. Phys. 64,
701 (2001).
[2] T. Dittrich, P. Hanggi, G.L. Ingold, B. Kramer, G. Schon, W. Zwerger,
Quantum Transport and Dissipation (Wiley-VCH, Weinheim, 1997).
[3] C. W. J. Beenakker, Phys. Rev. B 44, 1646 (1991).
[4] (Ed.) H. Grabert and M. H. Devoret, Single Charge Tunneling, Vol. 294
of NATO ASI Series B (Plenum Press, New York, 1991).
[5] T. A. Fulton and G. J. Dolan, Phys. Rev. Lett. 59, 109 (1987).
[6] Y. Meir, N. S. Wingreen, and P. A. Lee, Phys. Rev. Lett. 66, 3048 (1991).
[7] D. Weinmann, W. Hausler, and B. Kramer, Ann. Phys (Leipzig) 5, 652
(1996).
[8] A. T. Johnson, L. P. Kouwenhoven, W. de Jong, N. C. van der Vaart, C.
J. P. M. Harmans, and C. T. Foxon, Phys. Rev. Lett. 69, 1592 (1992).
[9] J. Weis, R. J. Haug, K. v. Klitzing, and K. Ploog, Phys. Rev. B 46, 12837
(1992).
[10] J. Weis, R. J. Haug, K. v. Klitzing, and K. Ploog, Phys. Rev. Lett. 71,
4019 (1993).
[11] H. P. Wei, L. W. Engel, and D. C. Tsui, Phys. Rev. B 50, 14609 (1994).
[12] A. K H uttel, H. Qin, A. W. Holleitner, R. H. Blick, K. Neumaier, D. Wein-
mann, K. Eberl, J. P. Kotthaus, preprint cond-mat/, 0109104 (2001).
[13] K. Yosida, Theory of Magnetism, Vol. 122 of SolidState Series (Springer,
Berlin, 1996).
Bibliography 29
[14] L. I. Glazman and M. E. Raikh, JETP Lett. 47, 378 (1988).
[15] T. K. Ng and P. A. Lee, Phys. Rev. Lett. 61, 1768 (1988).
[16] H. Haug and A.-P. Jauho, Quantum Kinetics in Transport and Optics of
Semiconductors, Vol. 123 of Solid-State Sciences (Springer, Berlin, 1996).
[17] Y. Meir and N. S. Wingreen, Phys. Rev. Lett. 68, 2512 (1992).
[18] N. S. Wingreen, Y. Meir, Phys. Rev. B 49, 11040 (1994).
[19] A. Kawabata, in Transport Phenomena in Mesoscopic Systems, Vol. 109
of Springer Series in Solid State Sciences, edited by H. Fukuyama and T.
Ando (Springer-Verlag, Berlin, 1991), p. 53.
[20] D. Goldhaber-Gordon, J. Gores, M. A. Kastner, Hadas Shtrikman, D.
Mahalu, and U. Meirav, Phys. Rev. Lett. 81, 5225 (1998).
[21] W. G. van der Wiel, S. De Francheschi, T. Fujisawa, J. M. Elzerman, S.
Tarucha, and L. P. Kouwenhoven, Science 289, 2105 (2000).
[22] S. Sasaki, S. De Franceschi, J. M. Elzerman, W. G. van der Wiel, M. Eto,
S. Tarucha, L. P. Kouwenhoven, Nature 405, 764 (2000).
[23] J. Schmid, J. Weis, K. Eberl, and K. v. Klitzing, Phys. Rev. Lett. 84,
5824 (2000).
[24] C. A. Sackett et al., Nature 404, 256 (2000).
[25] M. R. Wegewijs, Yu. V. Nazarov, Phys. Rev. B 60, 14318 (1999).
[26] T. H. Stoof and Yu. V. Nazarov, Phys. Rev. B 53, 1050 (1996).
[27] S. A. Gurvitz and Ya. S. Prager, Phys. Rev. B 53, 15932 (1996).
[28] S. A. Gurvitz, Phys. Rev. B 57, 6602 (1998).
[29] T. Brandes, Habilitation Thesis (University of Hamburg, Germany), 2000;
see under http://bursill.phy.umist.ac.uk/research.html.
[30] T. Brandes and B. Kramer, Phys. Rev. Lett. 83, 3021 (1999).
[31] T. H. Oosterkamp, T. Fujisawa, W. G. van der Wiel, K. Ishibashi, R. V.
Hijman, S. Tarucha, and L. P. Kouwenhoven, Nature 395, 873 (1998).
Bibliography 30
[32] A. J. Leggett, S. Chakravarty, A. T. Dorsey and M. P. A. Fisher, A. Garg
and W. Zwerger, Review of Modern Physics 59, 1 (1987).
[33] T. Fujisawa, T. H. Oosterkamp, W. G. van der Wiel, B. W. Broer,
R. Aguado, S. Tarucha, and L. P. Kouwenhoven, Science 282, 932 (1998).
[34] T. Brandes, T. Vorrath, in Recent Progress in Many Body Physics, Ad-
vances in Quantum Many Body Theory, edited by R. Bishop, T. Brandes,
K. Gernoth, N. Walet, and Y. Xian (World Scientic, Singapore, 2001).
INDEX
addition energies, 8
AlGaAsGaAs, 24
Anderson Impurity Model, 6, 13
Born and Markov approximation, 22
charging energy, 3
ClebschGordan coecient, 12
conductance, 9
conductance peak, 9
Coulomb blockade, 3
Coulomb blockade diamonds, 6
Coulomb Blockade Oscillations, 5
Coulomb staircase, 10
Coupled Quantum Dots, 18
coupled quantum dots, 24
current oscillations, 26
density operator, 22
electrochemical potential, 4
Entanglement, 18
exchange energy, 12
Fermi liquid, 15
Friedel Sum Rule, 15
gate voltage, 25
Greens Functions, 14
Hubbard, 6
inelastic current, 27
Kondo Eect, 13
Kondo temperature, 13
Lateral dots, 1
level spacing, 3
Liouvillevon Neumann equation, 22
logarithmic singularities, 16
master equation, 7, 19
Microwave Radiation, 20
NIBA, 23
Nozi`eres, 15
orthodox theory, 3
periodic table, 11
Phonons, 20
polaron transformation, 23
Quantum computing, 19
quantum optics, 19
renormalization group theory, 16
retarded Greens function, 15
Schottky gates, 24
shell structure, 12
Spin Blockade, 12
Spin ip processes, 13
spin singlet, 14
spinboson problem, 23
Spontaneous Emission, 20
Superposition, 18
tunnel rates, 7
TwoLevel System, 19
Vertical dots, 1
31