Prediction and reconstruction of Scanning Kelvin Probe Microscope

drain semiconductor x. eV n, p, R. Vs. Vd. Vg. ∆Vsd. V(x) p n n' p'. W. R gate dielectric source drain semiconductor .... Dago de Leeuw. Data, discussions, samples.
2MB taille 5 téléchargements 204 vues
Prediction and reconstruction of Scanning Kelvin Probe Microscope measurements on Organic Ambipolar Field Effect Transistors

Dimitri Charrier

Plan • Organic Ambipolar Field Effect Transistor • Recombination width • Langevin recombination • Experiments

• Simple prediction/reconstruction model • Step edge model • Prediction • Reconstruction

/ applied physics department

22-9-2010

PAGE 1

Organic Ambipolar Field Effect Transistor organic semiconductor drain (Au) source (Au) dielectric (SiO2) gate (Si-n doped)

source-drain current on

p

n

off unipolar 0 V ++++++++++++++++++++ 1 V _________________ -5V

gate voltage 0V

ambipolar 0 V+++++++++++ _ _ _ _ _ _ _- 10 V _ _ _ _ _ _ _ _ +++++++++++ -5V

Potential applications: organic laser if good performances. Important parameters: carrier densities n p, width W. / applied physics department

22-9-2010

PAGE 2

n, p, R

p’

R

n’ n n, p, rec

p

1E-3

0

1E-5

-2 n p rec ( = 3 nm)

1E-7

-4

1E-9

-6

V [eV]

Langevin recombination

1E-11 -8

W eV

1E-13 0

Vd Vg Vs

2

4

6

8

10

x [m]

Vsd

1

V(x)

rec (1 nm) rec (3nm) rec (10 nm)

0.01

rec

x semiconductor

source

drain

dielectric gate

1E-4

1E-6

1E-8

Wtheoretical

4.34d   20  200nm 

/ applied physics department

8.0

8.5

x [m]

M. Kemerink et al, Appl. Phys. Lett. 93, 033312 (2008)

22-9-2010

PAGE 3

lever

Reported experimental results

tip

active layer source

channel dielectric

drain

gate

Optical technique

Electrostatic technique Scanning Kelvin Probe Microscope (SKPM) NiDT Should see a potential drop at x0

Confocal microscope / High fields PPV

J.S. Swensen et al, J. Appl. Phys. 102, 013103 (2007)

E.C.P. Smits et al, Phys. Rev. B 76, 125202 (2007)

/ applied physics department

Wexperimental ~ 2 µm

22-9-2010

PAGE 4

SKPM Response for FET Theoretical predictions (drift) from Smits = input of SKPM modeling

Assumption: W = “0” nm

SKPM experiments + 3D modeling

‘real’ W < 0.5 micron One curve = 13 hours !

Note: We checked that the SKPM probe influence only few % the source drain current. / applied physics department

22-9-2010

PAGE 5

Simple prediction/reconstruction real signal

measured response

xref

yref h

real signal

measured response

xexp

yexp yref ( x)  href ( x)  xref ( x ) F ( y )  F ( h) F ( x )

h = Apex(x,y,z) + Cone(x,y,z) + Lever(x,y,z) = electrostatic convolution Hypothesis: one single reference measurement contains all electrostatic interactions / applied physics department

22-9-2010

PAGE 6

Step edge convolution → impulse response properties

surface potential (mV)

200 150 100

0 25 50 100 150 200 500 1000

Ti

Si-n doped

prediction:

 ) yexp  F 1 ( X expYstep 200 150 100 50 0

50 0 -50 -10 -8

F (h)  Ystep

Au Ti SiO2

inversion:

xexp -0.5

-6

-4

-2

0

2

4

0.0

0.5

6

8

Yexp F ( )  Ystep 1

10

position ( m)

/ applied physics department

22-9-2010

PAGE 7

Prediction from step edge response real signal

measured response

xexp

yexp Wpredicted < 700 nm

W = “0” nm

theoretical signal 3D modeling prediction experiment

surface potential (V)

10 8 6 4 2 0 -2

0

2

4

6

8

10

12

position ( m) / applied physics department

22-9-2010

PAGE 8

Reconstruction from step edge response Winversed – Wtheoretical = 0.5 – 0.5 = 0 µm real signal

measured response

xexp

yexp Wexperiment ~ 2 µm

Winversed = 0.5 µm

Wtheoretical 

4.34d  0.5m 100

6 4

0

5

dV/dx (normalized)

theoretical signal calculated with β = 100 m reconstruction experiment

surface potential (V)

8

dV/dx (normalized)

10

10

2 0 -2

0

2

4

6

2 8

4

6 10

8 12

position ( m) / applied physics department

22-9-2010

PAGE 9

Summary • Electrostatic tip-electrodes convolution leads to amplitude loss of measured surface potential with SKPM. Good agreement between experiments and 3D modeling. • Prediction and Reconstruction methods successfully working using the step edge response tool.

surface potential (V)

10 8 6 4 2 0 -2

/ applied physics department

2

6 4

4

6

8

10

12

position ( m)

0

5

dV/dx (normalized)

surface potential (V)

8

• W recombination: • theoretical (Langevin) ~ 200 nm • experimental SKPM response − raw ~ 2 µm − difference with model (β=100) 0 µm

0

10

dV/dx (normalized)

• A higher resolution of SKPM is reached with the step edge response tool.

10

2 0 -2

0

2

4

6

2 8

4

6 10

position ( m)

22-9-2010

PAGE 10

8 12

Acknowledgments

Molecular Materials and Nanosystems Group Martijn Kemerink René Janssen Simon Mathijssen

Clean room facilities

Data, discussions, samples

Barry Smalbrugge Tjibbe de Vries Erik-Jan Geluk

Edsger Smits Reinder Coehoorn Dago de Leeuw

/ applied physics department

22-9-2010

PAGE 11