Nanomaterials Characterization by low- temperature ...web.nano.cnr.it/heun/wp-content/uploads/2013/07/... · Nanomaterials Characterization by low-temperature Scanning Probe Microscopy

SGM Group

Stefan Heun

NEST, Istituto Nanoscienze-CNR and Scuola Normale Superiore Piazza San Silvestro 12, 56127 Pisa, Italy

e-mail: [email protected]

Nanomaterials Characterization by low-temperature Scanning Probe Microscopy

SGM Group

Stefan Heun

NEST, Istituto Nanoscienze-CNR and Scuola Normale Superiore Piazza San Silvestro 12, 56127 Pisa, Italy

e-mail: [email protected]

Inner structure of edge channels in the quantum Hall regime probed by scanning gate microscopy

Outline

• Introduction to quantum Hall physics

• Basics of Scanning Gate Microscopy

• Inner structure of edge channels in the

quantum Hall regime by SGM

Outline





Further reading:

Klaus von Klitzing: The quantized Hall effect, Nobel lecture, 1985 (http://nobelprize.org/nobel_prizes/physics/laureates/1985/klitzing.pdf)

R. J. Haug: Edge-state transport and its experimental consequences in high magnetic fields, Semicond. Sci. Technol. 8 (1993) 131.

http://nobelprize.org/nobel_prizes/physics/laureates/1985/klitzing.pdf

2 Dimensional Electron System

see also: Horst L. Stoermer, Nobel Lecture, December 8, 1998

Hall bar geometry

K. von Klitzing et al.: Physik Journal 4 (2005) No.6 p.37.

Measurement of longitudinal and transversal resistance

Drude model

vmBvEe

dt

vdmF

(=0; steady state)

zz

x

y

xy

x

xxx

Bne

B

j

E

constj

E

.

1

0

yxcy

xycx

jjE

jjE

0

0

Evnejm

ne

m

eBzc

,,

2

0with

Transverse current is zero:

0yj It follows:

Ashcroft / Mermin: Solid State Physics.

Hall effect


Drude model:

Rxy = proportional to B

Rxx = costant in B

Quantum Hall Effect


Behaviour observed at

low temperature in high

mobility samples in high

B ( ):

Plateaux in Rxy

Minima in Rxx

Quantum Hall Effect:

deviations from Drude

model observed around

B = nh / ne

n: filling factor

1c

The Nobel prize in Physics

1985: integer QHE 1998: fractional QHE

http://nobelprize.org/

Landau Levels

Single particle Hamiltonian for an electron in an external magnetic field:

Landau-Gauge: with

For the wavefunction can be written as:

It follows:

rVAepm

H

2

2

1

0,,0 BxA

zeBB

xeyxyiky ,

xxxmm

pxE c

x

2

0

22

2

1

2

0rV

R. J. Haug, Semicond. Sci. Technol. 8 (1993) 131.

Landau Levels

This is the equation of a harmonic oscillator with angular frequency

center of cyclotron orbit at and magnetic length

Drift velocity

meBc

2

0 Bykx

)(

25

TB

nm

eBB

xxxmm

pxE c

x

2

0

22

2

1

2

2

1 jE cj

Landau Levels 0 yjy dkdEv


Landau Levels

Degeneracy of each

Landau Level:

The number of filled

Landau Levels for a given

carrier concentration ns is

then:

n: filling factor


Bh

eBn deg

eB

hn

n

n ss deg

n

Landau Levels

At the edges of the sample, a boundary confinement can be added by a small

V(x) which does not depend on y.

221 Bycyj kVjkE

B. I. Halperin, Phys. Rev. B 25, 2185 (1982).

Introduction to edge channels

x y x

see also: R. J. Haug, Semicond. Sci. Technol. 8 (1993) 131.


x y x



x y

edge: 1D “one-way” conductor, skipping orbits

bulk: insulator x



x y

edge: 1D “one-way” conductor, skipping orbits

bulk: insulator

Quantization of energy levels (Landau levels)

Gap for excitation in the bulk

Transport via edge states

x


Edge states

compressibility:

incompressible:

sch n

sch n


incompressibilities

Example Semiconductor

n

VB

CB

n*

n* independent of B

Edge states

D. B. Chklovskii et al.: PRB 46 (1992) 4026.

compressibility:

incompressible:

sch n

sch n

Edge states


compressibility:

incompressible:

sch n

sch n

SGM Group

Non-interacting vs. interacting picture


• The self consistent potential due to e-e interactions modifies the edge structure •For any realistic potential the density goes smoothly to zero. •Alternating compressible and incompressible stripes arise at the sample edge

Incompressible stripes: •The electron density is constant •The potential has a jump

Compressible stripes: •The electron density has a jump •The potential is constant

SGM Group

Non-interacting vs. interacting picture


• The self consistent potential due to e-e interactions modifies the edge structure •For any realistic potential the density goes smoothly to zero. •Alternating compressible and incompressible stripes arise at the sample edge

Incompressible stripes: •The electron density is constant •The potential has a jump

Compressible stripes: •The electron density has a jump •The potential is constant

Each of these edge channels carries M. Büttiker, Phys. Rev. B 38, 9375 (1988)

VheheI 2

How to probe edge channels?

S. Roddaro et al., Phys. Rev. Lett. 90 (2003) 046805.






V

t

r 0

V

I(V)

Vg


Quantum Point

Contact (QPC)

V

t

r 0

V

I(V)

Vg


we induce backscattering

by reducing this distance

Current conservation:

t + r = 1


Hot topics

• The quantum Hall Effect as an Electrical

Resistance Standard

RK = 25812.807572Ω

Hot topics


Resistance Standard

• Quantum Hall Interferometry

Y. Ji et al., Nature 422 (2003) 415.

Hot topics


Resistance Standard


• Quantum Computing

http://stationq.ucsb.edu/research.html

Hot topics


Resistance Standard


• Quantum Computing

• Basic Research

Outline




quantum Hall regime by SGM Further reading:

M. A. Topinka, B. J. LeRoy, S. E. J. Shaw, E. J. Heller, R. M. Westervelt, K. D. Maranowski, and A. C. Gossard: Imaging coherent electron flow from a quantum point contact, Science 289 (2000) 2323.

M. A. Topinka, B. J. LeRoy, R. M. Westervelt, S. E. J. Shaw, R. Fleischmann, E. J. Heller, K. D. Maranowski, and A. C. Gossard: Coherent branched flow in a two-dimensional electron gas, Nature 410 (2001) 183.

Scanning Gate Microscopy

AFM with conductive tip

Tip at negatively bias

(local gate - locally

depletes the 2DEG), no

current flows

SGM performed in

constant height mode

(10-50 nm above

surface), no strain

M. A. Topinka et al.:

Science 289 (2000) 2323.

Scanning Gate Microscopy

AFM with conductive tip

Tip at negatively bias

(local gate - locally

depletes the 2DEG), no

current flows

SGM performed in

constant height mode

(10-50 nm above

surface), no strain

A. Iagallo et al.:

Nano Reserach,

in press.

Coherent branched flow of

electrons

M. A. Topinka et al., Nature 410 (2001) 183.

G: 0.00e2/h 0.25e2/h

Tip-induced backscattering

N. Paradiso et al., Physica E 42 (2010) 1038.

Tip-induced backscattering


Coherent branched flow of

electrons

M. A. Topinka et al., Nature 410 (2001) 183.

G: 0.00e2/h 0.25e2/h

source-drain current

SGM Group

The SGM @NEST lab in Pisa

• AFM non-optical detection scheme (tuning fork)

• With vibration and noise isolation system

• 3He insert (cold finger base temp. :300 mK)

• 9 T cryomagnet

Tip at negative bias (moveable gate locally depletes the 2DEG)

M. A. Topinka et al.: Science 289 (2000) 2323. Pioneering work by:

SGM performed in constant height mode (10-50 nm above surface), no strain

Setup:

Oct 2007: First cooldown

AFM Head

• Positioner (range in xyz > 6 mm) allow to locate features on the sample with µm precision

• Scan range:

• 42µm x 42µm x 2µm @ RT

• 8.5µm x 8.5µm x 1.4µm @ 300 mK

• Temperature measurement (RuO2) close to sample

• Drift < 1 nm / h

• Temperature stability Delta T / T < 5% for hours, even at max. B field

• Noise in z at 300 mK: 1nm

Tuning Fork

Sample holder for transport

measurements

Base mounted on AFM scanner Contact via pogo pins

Chip carrier

holds sample

Sample holder mounted on scanner

Tip – sample geometry

Outline





Further reading:

N. Paradiso et al.: Spatially resolved analysis of edge-channel equilibration in quantum Hall circuits, Phys. Rev. B 83 (2011) 155305.

N. Paradisoet al.: Selective control of edge-channel trajectories by scanning gate microscopy, Physica E 42 (2010) 1038.

Hall-bar samples

1900 m x 300 m


Hall-bar samples


Hall-bar samples


Hall-bar samples


Hall-bar samples


SGM Group

Conductance quantization in QPCs

1D confinement

In 1D systems the current is carried by a finite number of modes (arising from confined subbands) . Each mode contributes two quantum of conductance.

H. van Houten and C. Beenakker: Physics Today, July 1996, p. 22

2e2/h

SGM Group


1D confinement


2e2/h

First we fix the mode number (QPC setpoint), then we start scanning the biased tip at a fixed height.

SGM Group


1D confinement


First we fix the mode number (QPC setpoint), then we start scanning the biased tip at a fixed height.

source-drain current 2DEG

The split gates define a constriction by depleting the 2DEG underneath

The biased tip creates a depletion spot that we use to backscatter the electrons passing through the constriction

SGM Group

QPC at 3rd plateau 3rd plateau

600nm600nm

�5.50

�

0.00

G = 0

G = 6 e2/h

600nm

SGM Group

QPC at 2nd plateau 2nd plateau

�5.50

�

0.00

G = 0

G = 6 e2/h

600nm

SGM Group

QPC at 1st plateau 1st plateau

�5.50

�

0.00

G = 0

G = 6 e2/h

SGM Group

Branched flow and interference fringes

400nm 100nm

Fringe periodicity: lF/2=20 nm

• QPC conductance G = 6 e2/h (3rd plateau) • Tip voltage Vtip = -5 V, height htip = 10 nm • see also M. A. Topinka et al., Nature 410 (2001) 183.

SGM in the Quantum Hall regime

Tip voltage Vtip = -5 V, height = 30 nm


Magnetoresistance


SGM Group

Selective control of edge channel trajectories by SGM

• Bulk filling factor n=4

• B = 3.04 T

• 2 spin-degenerate edge channels

• gate-region filling factors g1 = g2 = 0

�2.90

�

0.00

4.0 e2/h

0.0 e2/h

4.0 e2/h

0.0 e2/h

600nm


SGM technique: we select individual channels from the edge of a quantized 2DEG, we send them to the constriction and make them backscatter with the biased SGM tip.

SGM Group

How we probe incompressible stripes

-100 0 100 200 300 400 500 600 700 800

0

1

2

3

4

co

nd

uc

tan

ce

(e

2/h

)

tip position (nm)

tip position

Self-consistent potential

ħωc

Landau levels inside the constriction

tip induced potential

SGM Group


-100 0 100 200 300 400 500 600 700 800

0

1

2

3

4

co

nd

uc

tan

ce

(e

2/h

)

tip position (nm)

tip position

backscattering

SGM Group


-100 0 100 200 300 400 500 600 700 800

0

1

2

3

4

co

nd

uc

tan

ce

(e

2/h

)

tip position (nm)

tip position

SGM Group


-100 0 100 200 300 400 500 600 700 800

0

1

2

3

4

co

nd

uc

tan

ce

(e

2/h

)

tip position (nm)

tip position

backscattering

Energy gap: ħω=5.7 meV Plateau width: 60 nm Incompr. stripe width: ≈30nm

Can we exploit the non-trivial edge structure?

The picture of a QH device emerging from the experiments

A “bus” of compressible and incompressible channels

Is it possible to study and image the

microscopic details of the inter-channel

interactions?

SGM Group

SGM Group

Studying inter-channel equilibration

d

Two co-propagating edge channels originating from two ohmic contacts at different potential

n=2

g=1

SGM Group

Studying inter-channel equilibration

devices with fixed interaction length d: elusive determination of the microscopic details of

the equilibration mechanisms

d

n=2

g=1

SGM Group

The oppurtunity of the Scanning Gate Microscopy

Our technique allows to selectively control the channel trajectory

Our idea: exploit the mobile depletion spot induced by the SGM to continuously tune d

nbulk = 4 two spin degenerate edges

SGM tip

n=0

n=2

n=4 N. Paradiso et al., PRB 83 (2011) 115305.

SGM Group




SGM tip

n=0

n=2

n=4 N. Paradiso et al., PRB 83 (2011) 115305.


SGM Group




SGM tip

n=0

n=2

n=4

Each channel carries 2e2/h units of conductance

N. Paradiso et al., PRB 83 (2011) 115305.


SGM Group





SGM Group





No eq. Full eq.

A 2 e2/h 1 e2/h

B 0 e2/h 1 e2/h

SGM Group

Experimental setup

SEM micrograph of the device

source bias V=VAC + VDC

tip

Scheme of the electronic setup

edge-selector gate

reflected component

transmitted component

SGM Group

Imaging the inter-channel equilibration

Source bias: VAC=50V, VDC=0mV

The profiles of GB(d) along the trajectory show a strict dependance on the local details

Imaging the edge channel equilibration

1

0

[e2/h]

SGM Group

Imaging the inter-channel equilibration

The profiles of GB(d) along the trajectory show a strict dependance on the local details

We can directly image the potential induced by the most important defects by means of a

scan at zero magnetic field

correlation found

Source bias: VAC=50V, VDC=0mV

Imaging the edge channel equilibration

SGM scan at zero magnetic field

1

0

[e2/h]

-1 0 1 2 3 4 5 6 7

-0.05

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

dif

fere

nti

al c

on

du

cta

nc

e G

B (

e2/h

)

position (m)

Experimental data

Tight binding simulations

SGM Group

Tight binding simulations

Pictorial model for the disorder potential

tip potential

“big impurities” potential

background potential

scattering centers

SGM map of the inter-channel equilibration in another device

1

0

[e2/h]

Simulations made by the theoretical group of Scuola Normale Superiore (Pisa, Italy) D. Venturelli, F. Taddei, V. Giovannetti and R.Fazio


SGM Group

Nonlinear regime

The backscattered current is a function of the local imbalance V(x) that depends on the specific scattering process.

SGM Group

Two mechanisms for the inter-channel scattering

For low bias the only relevant mechanism is the elastic scattering induced by impurities, which determines an ohmic behavior (linear I-V)

At high bias (Δμ≈ħωc) vertical transition with photon emission are enabled (threshold and saturation)

SGM Group

Impact of the electron heating

Electron heating due to injection of hot carriers:

The relaxation of hot carriers induces a dramatic temperature increase. This is why the transition is smoothened and the threshold voltage reduced for high d

SGM Group

Conclusions

We explore the use of Scanning Gate Microscopy to selectively control edge channel trajectories

We built size-tunable QH circuits to directly image the equilibration between imbalanced co-propagating edges

The comparison with the SGM scan at zero magnetic field reveals a correlation between the local potential and steps in the GB(d) curve

Shift of the threshold voltage for the onset of photon emission is explained by a simple model for the electron heating

Control of the edge channel trajectory allows us to study their structure

Coworkers

R. Fazio V. Giovannetti F. Taddei D. Venturelli

N. Paradiso S. Roddaro G. Biasiol (IOM) L. Sorba F. Beltram

Funding

Documents

Nanomaterials Characterization by low- temperature ...web.nano.cnr.it/heun/wp-content/uploads/2013/07/... · Nanomaterials Characterization by low-temperature Scanning Probe Microscopy