Modeling of pulsed Laser Thermal Annealing for ... - comsol.se · COMSOL Conference Europe 2012 – Milan October 11th 2012 Silicon Dopants Junction Depth Laser shot Melting Temperature

Confidential: modification, copy and distribution prohibited without prior written permission from Excico ® 1

Modeling of pulsed Laser Thermal Annealing for junction formation optimization and

process control

R. Negru, K. Huet, P. Ceccato, B. Godard

EXCICO, 14 rue Alexandre, 92230 Genneviliers, France

http://www.excico.com

Excerpt from the Proceedings of the 2012 COMSOL Conference in Milan


OUTLINE

Laser Thermal Annealing technology

Experiments

Model

Phase-Field approach

Dopant diffusion and segregation

Results

Conclusions

COMSOL Conference Europe 2012 – Milan October 11th 2012


OUTLINE


Experiments

Model



Results

Conclusions



Laser absorption

Melting and recrystallization

Low thermal budget process

High temperature localized in space and time

Shallow depth effect (<µm)

Ultrafast (<µs)

Pulsed excimer Laser Thermal Annealing (LTA)


Technique for junction formation in process fabrication of semiconductor devices

Silicon

Laser shot


Laser tool characteristics

High Energy Gas laser

XeCl excimer gas

UV 308nm wavelength

Pulsed

Challenge: process variability

Energy and pulse variations



Laser Thermal Annealing Process Parameters

Long pulse laser

Pulse Duration: ~150 ns

Uniform high Energy Density

Up to 3 J/cm2

Up to 2x2 cm2 area


LTA Pulse

Uniform ED


LTA process simulation

Linking tool parameters to process

Understanding process variability

Process variability and junction formation

Depend on

Laser Energy Density

Pulse shape

Other laser annealing process parameters

Melt Depth (ex-situ measurement)

Temperature (no direct measurement)


Silicon

Dopants

Junction Depth

Laser shot

Silicon

Melting Temperature

Ambient Temperature


LTA simulation: 2 steps

Laser parameters

Structure - Material properties - Geometric dimensions

Temperature profiles Melt dynamics


Process Window determination to avoid damage

Temperature profiles &

Melt dynamics

- Dopant profiles Dopant distribution

Junction formation

Thermal step

Diffusion step


OUTLINE


Experiments

Model



Results

Conclusions



1E+16

1E+17

1E+18

1E+19

1E+20

0 100 200 300 400

Co

nce

ntr

atio

n [

at/c

m3]

Depth [nm]

B30keV1e15 as-imp

1E+16

1E+17

1E+18

1E+19

1E+20

0 100 200 300 400

Co

nce

ntr

atio

n [

at/c

m3]

Depth [nm]

B30keV1e15 as-imp

SIMS ED2.2

SIMS ED2.4

SIMS ED2.8

SIMS ED3.2

Typical Boron profiles after LTA

Melt Depth estimation vs Energy Density Profiles not explained by simple diffusion (Fickian)


Secondary Ion Mass Spectroscopy (SIMS) error: ±5% in depth and ±10 in concentration

1E+16

1E+17

1E+18

1E+19

1E+20

0 100 200 300 400

Co

nce

ntr

atio

n [

at/c

m3]

Depth [nm]

B30keV1e15 as-imp

SIMS ED2.4

1E+16

1E+17

1E+18

1E+19

1E+20

0 100 200 300 400

Co

nce

ntr

atio

n [

at/c

m3]

Depth [nm]

B30keV1e15 as-imp

SIMS ED2.2

SIMS ED2.4

SIMS ED2.8

SIMS ED3.2

Pile-up Depletion

1E+16

1E+17

1E+18

1E+19

1E+20

0 100 200 300 400

Co

nce

ntr

atio

n [

at/c

m3]

Depth [nm]

B30keV1e15 as-imp

SIMS ED2.4

SIMS ED2.8

1E+16

1E+17

1E+18

1E+19

1E+20

0 100 200 300 400

Co

nce

ntr

atio

n [

at/c

m3]

Depth [nm]

B30keV1e15 as-imp

SIMS ED2.2

SIMS ED2.4

SIMS ED2.8

SIMS ED3.2

MD estimation

Box-like profile

1E+16

1E+17

1E+18

1E+19

1E+20

0 100 200 300 400

Co

nce

ntr

atio

n [

at/c

m3]

Depth [nm]

B30keV1e15 as-imp

SIMS ED2.2

SIMS ED2.4

SIMS ED2.8

SIMS ED3.2

1E+16

1E+17

1E+18

1E+19

1E+20

0 100 200 300 400

Co

nce

ntr

atio

n [

at/c

m3]

Depth [nm]

B30keV1e15 as-imp

SIMS ED2.2

SIMS ED2.4

SIMS ED2.8

SIMS ED3.2


OUTLINE


Experiments

Model



Results

Conclusions



Phase-Field model

Heat (T) and Phase () equations are connected by coupling terms

Formalism: -1 +1

Heat equation

Source equation

Phase change equation

Thermal and phase change simulation

[Karma and Rappel, PRE 1998] [La Magna et al., JAP 2004]

Pure liquid

Pure solid

x

n eRtPEDtxS )1()(),(

),(18

15

2)(

222 txSt

LTk

t

Tc

fus

p

Coupling with phase

22222 )1()()1(

M

fus

pTT

L

cW

t

Coupling with temperature


Liquid

Solid

Laser shot

Silicon


Dopant distribution simulation

Diffusion + Adsorption model

Boron adsorbed at Liquid/Solid interface

With

CB: Boron concentration (cm-3)

DB: Boron diffusion coefficient (cm2s-1, phase dependent)

Cequ: Equilibrium concentration (fit parameter)

𝜕𝐶𝐵𝜕𝑡

= ∇ 𝐷𝐵∇𝐶𝐵 − ∇ 𝐷𝐵𝐶𝐵𝐶𝑒𝑞𝑢

∇𝐶𝑒𝑞𝑢

Fickian diffusion Adsorption and segregation

[H. M. You, et al., JAP 1993] [M. Hackenberg, R. Negru, K. Huet, et al., IIT 2012]



OUTLINE


Experiments

Model



Results

Conclusions



0.0

0.3

0.5

0.8

1.0

0 100 200 300 400

No

rmal

ize

d P

uls

e [

a.u

.]

Time [ns]

Pulse PD135

PD

0.0

0.3

0.5

0.8

1.0

0 100 200 300 400

No

rmal

ize

d P

uls

e [

a.u

.]

Time [ns]

Pulse PD135

Pulse PD145

PD

0.0

0.3

0.5

0.8

1.0

0 100 200 300 400

No

rmal

ize

d P

uls

e [

a.u

.]

Time [ns]

Pulse PD135

Pulse PD145

Pulse PD155

PD

0

50

100

150

200

250

300

350

1 1.5 2 2.5 3 3.5

Me

lt D

ep

th [

nm

]

Energy Density [J/cm2]

Experience PD135

Model EXCICO PD135

0

50

100

150

200

250

300

350

1 1.5 2 2.5 3 3.5

Me

lt D

ep

th [

nm

]


Experience PD135

Model EXCICO PD135

Experience PD145

Model EXCICO PD145

0

50

100

150

200

250

300

350

1 1.5 2 2.5 3 3.5

Me

lt D

ep

th [

nm

]


Experience PD135

Model EXCICO PD135

Experience PD145

Model EXCICO PD145

Experience PD155

Model EXCICO PD155

0.0

0.3

0.5

0.8

1.0

0 100 200 300 400

No

rmal

ize

d P

uls

e [

a.u

.]

Time [ns]

Pulse PD135

Pulse PD145

Pulse PD155

PD

Thermal and phase change simulation

Model fit very well the experimental data

Accuracy: R2 > 95%


0

50

100

150

200

250

300

350

1 1.5 2 2.5 3 3.5

Me

lt D

ep

th [

nm

]


Experience PD135

Model EXCICO PD135

Experience PD145

Model EXCICO PD145

Experience PD155

Model EXCICO PD155


1E+17

1E+18

1E+19

1E+20

0 100 200 300 400

Co

nce

ntr

atio

n [

at/c

m3]

Depth [nm]

B 30keV 1e15cm^-2

PD=155nsAs-impantedED=2.5J/cm2

B 30keV 1e15cm-2

Dopant distribution simulation

Good simulation of LTA junction formation

Good agreement between simulation and experience

Model accuracy: R2 > 90%


Pile-up Depletion

1E+17

1E+18

1E+19

1E+20

0 100 200 300 400

Co

nce

ntr

atio

n [

at/c

m3]

Depth [nm]

B 30keV 1e15cm^-2


ED=2.8J/cm2

B 30keV 1e15cm-2

Box-like profile

1E+17

1E+18

1E+19

1E+20

0 100 200 300 400

Co

nce

ntr

atio

n [

at/c

m3]

Depth [nm]

B 30keV 1e15cm^-2


ED=2.5J/cm2

ED=2.8J/cm2

ED=3.2J/cm2

B 30keV 1e15cm-2


OUTLINE


Experiments

Model



Results

Conclusions



CONCLUSIONS

Objective

Linking tool parameters to process

Model validation

Conclusions

Good agreement between model and experiences in case of LTA time shift Melt Depth

Good simulation of LTA junction formation Diffusion & segregation

Perspectives

Tool for process integration

Extend to other dopants



THANK YOU FOR YOUR THE ATTENTION !


Documents

Modeling of pulsed Laser Thermal Annealing for ... - comsol.se · COMSOL Conference Europe 2012 – Milan October 11th 2012 Silicon Dopants Junction Depth Laser shot Melting Temperature