comsol multiphysicsと実験を併用した薄膜製造プロセスの解析 · 2012-02-20 · 1...

1

COMSOL Multiphysicsと実験を併用した薄膜製造プロセスの解析

COMSOL Conference 2011，December 2, 2011 @ Akihabara

霜垣幸浩東京大学マテリアル工学専攻

〒113-8656 文京区本郷７－３－１

E-mail [email protected]://www.dpe.mm.t.u-tokyo.ac.jp

2

Outline

IntroductionMechanism of Metal Organic Chemical Vapor DepositionSelective Area Growth (SAG)

GaAs-SAGLinear kinetic analysisNon-Linear kinetic analysisDoping Effects

InP, InAs, InAsP, GaAsP, InGaAsPKinetics of InP/InAs and InAsP/GaAsP SAGEstimation of InGaAsP PL wavelength distribution

Conclusion

3

Precursors: Tertiarybutylarsenide (TBAs)Tertiarybutylphosphide(TBP) Trimethylgallium (TMGa)TrimethylIndium (TMIn)

V/III >> 1

III-V compound semiconductorMOCVD Process

4

MOCVD Reaction Mechanism

(CH3)3CPH2

(CH3)3In

Gas-phase decomposition: kg

Mass-transport

Surface reaction: ks

InP epitaxial film

Gas-flow and Gas-phase diffusion: D/

5

Gas-Phase reaction rate constant

Flow cracking reactor and FT-IR gas analysis

……GaAs substrate

Selective area growth (SAG) MOCVD

SiO2 mask SiO2 mask

1Vertical vapor phase diffusion

=0

Lateral vapor phase diffusion

2 2

3Surface migration

MASK MASK

Analysis of surface reaction kinetics(wide stripe SAG)

SAG

For monolithic integration of OEICs(one step fabrication)

Film thickness & composition can be locally controlled by mask.

7

Monolithic integration of multiple EgMulti-Quantum Wells by SAG

ACTIVE Devices PASSIVE Devices EV

EC

E

SAGPlanar area

8

Outline

IntroductionMechanism of Metal Organic Chemical Vapor DepositionSelective Area Growth (SAG)


InP, InAs, InGaAs, InGaP, InGaAsPKinetics of InP/InAs and InAsP/GaAsP SAGEstimation of InGaAsP PL wavelength distribution

Conclusion

9

105010 50 380 m

200 m

C = C0

x

z

Laplace’s Equation

GRE

sss Ckr

D/ks

ks: surface reaction rate constant

. . . SAG

Planar

RG R ER

Growth-Rate-Enhancement

COMSOL Simulation& Fitting

5020

50

3802000

x

z

Growth area

m

mask

SAG Analysis method by Simulation

10

Estimation of Diffusion Coefficient

Diffusion coefficient can be estimated from the slope of the growth rate profile.

We can also use Chapman-Enskog equation to estimate the diffusion coefficient.

0.1

1

10

100

0 50 100 150 200 250

D=0.01D=0.0075D=0.0050D=0.0040D=0.0020D=0.0010Growth Rate(nm/min)

InP

grow

th ra

te

[nm

/min

]

x [mm]

00

0 0.025 0.05 0.075 0.1 0.125D (m2/ s)

2

4

6

8

10

12

14

16

18

20

slop

e(m

-1)

InPk : A = 1.86E15 (1/s), Ea = 186 (kJ/mol)ks : A = 5E5 (m/s), Ea = 80 (kJ/mol)Experiment

：４．９３

D：0.003

Diffusion limited Regime

Reaction limited regime

11

Estimation of Surface Reaction Rate Constant, ks by SAG

200 m

mask

growth area 380 m

DC 0 0 nCD

ssCknCD

On masksOn films

1

1.2

1.4

1.6

1.8

2

100 150 200 250 300 350 400

Gro

wth

-Rat

e E

nhan

cem

ent [

-]

Position [m]

InPGaAs

Fitting D/ks is determined

D can be estimated

ks is obtained!

Growth Rate distribution analysis

12

Temperature Dependency of GR and ks700 650 600 550

5

10

15

temperature [oC]

G.R

. [nm/m

in]

625

in the planar area Growth ratein planar area

mass transport limited growth

1.00 1.05 1.10 1.15 1.20 1.25

10

30

50

70 5pTMGa= 0.0029 mbarpTBAs= 0.04 mbar

just Ea2= 87.9 kJ/mol

2o off Ea2= 76.0 kJ/mol



k s [m/s

]

1000/T [K]

]

ks-GaAsEa1 < 4 kJ/mol

Step flow growth mode

2D nucleation dominated

13

GaAs Surface Structure

c(44)

AsGa

2(24)

As surface coverage :

Ref. Q.Fu, JCG 225 (2001) 405

1.0 ML 0.75 MLLow Temperature High Temperature

14

Effect of PTMGa on SAG Profile

0 50 100 150 200 250 300 3501.0

1.1

1.2

1.3

1.4 experimental: pTMGa=8.3x10-4mbar

pTMGa=2.9x10-3mbar

pTMGa=5.8x10-3mbar

linear simulation kl

s=23.1 m/s

kls=17.5 m/s

kls=11.4 m/s

Position (m)

Gro

wth

rate

enh

ance

men

t (-)

@575ºC

15

PTMGa Dependency of ks

0.000 0.001 0.002 0.003 0.004 0.005 0.0065

10

15

20

25

30

35

40

45 just (100) 2o off 5o off 15o off

kl s (m/s

)

pTMGa (mbar)

432

5

1

575oCpTBAs = 0.04 mbar

0 5 10 15Growth Rate (nm/min)

sss Ckr

@575ºC

Ga

Gans

KCKCkr

1

MMGaAsH

AsGa

Langmuir-Hinshelwoodadsorption isotherm

Growth of GaAs

AsGanskr

Ga

GaGa KC

KC

1

V/III >> 1

No competition for the adsorption processes between Ga and As.

TBAsTMGa

MMGa + AsH → GaAs + by-products

17

mask

0

zCD0

zCD

0

xC0

xC

H

380 m

02 CDLaplace’s equation

spns

spsp

KrKkr

DHr

CC

0

sp

spns

spsp

planar KCKCk

rH

CCDJ

10

sp

s

sns

sp

s

rKCKCk

rrGRE

1

Cs : COMSOLrsp: experiment

ksn, K

s

sns

KCKCk

zCD

1

Non-Linear Simulation

18

0 50 100 150 200 250 300 3501.0

1.1

1.2

1.3

1.4

)(non-linear simulation kn

s=3.6x10-5mol/m2/s

K=6.9x105m3/mol

Just (100)

experimental: pTMGa=8.3x10-4mbar

pTMGa=2.9x10-3mbar

pTMGa=5.8x10-3mbar

Position (m)

Gro

wth

rate

enh

ance

men

t (-)

At 575 ºC

Non-Linear Simulation Results

0 50 100 150 200 250 300 3501.0

1.1

1.2

1.3

1.4

kns=3.57x10-5mol/m2/s

K=6.88x105m3/mol condition 1 (just)

X (m)

Gro

wth

rate

enh

ance

men

t

0 50 100 150 200 250 300 3501.0

1.1

1.2

1.3

1.4



Gro

wth

rate

enh

ance

men

t

X (m)0 50 100 150 200 250 300 350

1.0

1.1

1.2

1.3

1.4



X (m)

Gro

wth

rate

enh

ance

men

t

0 50 100 150 200 250 300 3501.0

1.1

1.2

1.3

1.4



Gro

wth

rate

enh

ance

men

t

X (m) 0 50 100 150 200 250 300 350

1.0

1.1

1.2

1.3

1.4



Gro

wth

rate

enh

ance

men

t

X (m)

0 50 100 150 200 250 300 3501.0

1.1

1.2

1.3

1.4


K=8.46x105m3/mol condition 1 (2off)

Gro

wth

rate

enh

ance

men

t

X (m)0 50 100 150 200 250 300 350

1.0

1.1

1.2

1.3

1.4



X (m)

Gro

wth

rate

enh

ance

men

t

0 50 100 150 200 250 300 3501.0

1.1

1.2

1.3

1.4



X (m)

Gro

wth

rate

enh

ance

men

t

0 50 100 150 200 250 300 3501.0

1.1

1.2

1.3

1.4



X (m)

Gro

wth

rate

enh

ance

men

t

0 50 100 150 200 250 300 3501.0

1.1

1.2

1.3

1.4



Gro

wth

rate

enh

ance

men

t

X (m)

0 50 100 150 200 250 300 3501.0

1.1

1.2

1.3

1.4



Gro

wth

rate

enh

ance

men

t

X (m)0 50 100 150 200 250 300 350

1.0

1.1

1.2

1.3

1.4



Gro

wth

rate

enh

ance

men

t

X (m)0 50 100 150 200 250 300 350

1.0

1.1

1.2

1.3

1.4 kn

s=3.12x10-5mol/m2/s


Gro

wth

rate

enh

ance

men

t

X (m)0 50 100 150 200 250 300 350

1.0

1.1

1.2

1.3

1.4



Gro

wth

rate

enh

ance

men

t

X (m)0 50 100 150 200 250 300 350

1.0

1.1

1.2

1.3

1.4



Gro

wth

rate

enh

ance

men

t

X (m)

0 50 100 150 200 250 300 3501.0

1.1

1.2

1.3

1.4 kn

s=3.28x10-5mol/m2/s


Gro

wth

rate

enh

ance

men

t

X (m)0 50 100 150 200 250 300 350

1.0

1.1

1.2

1.3

1.4 kn

s=3.28x10-5mol/m2/s


Gro

wth

rate

enh

ance

men

t

X (m)0 50 100 150 200 250 300 350

1.0

1.1

1.2

1.3

1.4



Gro

wth

rate

enh

ance

men

t

X (m)0 50 100 150 200 250 300 350

1.0

1.1

1.2

1.3

1.4



Gro

wth

rate

enh

ance

men

t

X (m)0 50 100 150 200 250 300 350

1.0

1.1

1.2

1.3

1.4



Gro

wth

rate

enh

ance

men

t

X (m)

5o off ksn=3.1x10-5mol/m2/s, K=10x105m3/mol

15o off ksn=3.3x10-5mol/m2/s, K=12x105m3/mol

2o off ksn=3.4x10-5mol/m2/s, K=8.5x105m3/mol

just ksn=3.6x10-5mol/m2/s, K=6.9x105m3/mol pTMGa

-2 0 2 4 6 8 10 12 14 160

1

2

3

4

5

6

7

8

0

2

4

6

8

10

12

14

kns

kn s (10-5

mol

/m2 /s)

misorientation angle, (o)

K

K (1

05 m3 /m

ol)

d

d

d

a

kkK ka: independent of d.

kd: ~ d

Kkd ddesorbedeasily kd

Kkd

d

edgesteptomoveeasily

kd

ksn

reactivity ofGa-species withAs-species onthe surface

Lifetime (on GaAs (100)) : 0.30 sec

21

surface coverage (Ga )

pTMGa(10-3 mbar) 0.83 1.7 2.9 4.1 5.8

Just 0.08 0.15 0.23 0.34 0.45

2o off 0.10 0.18 0.26 0.39 0.51

5o off 0.12 0.21 0.30 0.43 0.55

15o off 0.13 0.24 0.34 0.47 0.59

Surface coverage

Ga

GaGa KC

KC

1

Estimation of Surface Coverage

@575ºC

22

Outline

IntroductionMechanism of Metal Organic Chemical Vapor DepositionSelective Area Growth


InP, InAs, InGaAs, InGaP, InGaAsPKinetics of InP/InAs and InAsP/GaAsP SAGEstimation of InGaAsP PL wavelength distribution

Conclusion

23

Ga

In

In(1-x)GaxAsyP(1-y) (x)

simulation

0 1 0 0 0 2 0 0 0 3 0 0 0 4 0 0 0 5 0 0 00 .0 0

0 .0 1

0 .0 2

0 .0 3

0 .0 4

0 .0 5

0 .0 6

0 .0 7

Con

cent

raio

n

p o s itio n (m )

D /k s= 1 9 m (In ) D /k s= 1 5 0 m (G a )

Concentration close to surface(G) GRE of In and Ga precursorsG = (1-x0)GIn + x0 GGax = x0 GGa / {(1-x0)GIn + x0 GGa}

Estimation of PL Wavelength of InGaAsP

24

1.55m PL Wavelength Estimation

0 1000 2000 3000 4000 5000

1200

1300

1400

1500

1600

Barrier

MQW

Well

PL

Pea

k W

avel

engt

h (n

m)

Position (m)

InP buffer

SCH Q1.2Q1.2 barrierQ1.5 well

5set MQWSCH Q1.2

InP substrate

InP cap

Structure

0 48000 48000 4800

Measured and simulated photoluminescence (PL) peak wavelength

25

Mask design for transition region

-300 -200 -100 0 100 200 30013801400142014401460148015001520154015601580

conventional (calc.) taper (calc.) conventional (exp.) taper (exp.)

PL p

eak

wav

elen

gth

(nm

)

position (m)

50m20m

120m5m

D/ks(Ga) = 180mD/ks(In) = 30m

遷移領域Transition region

26

Summary

SAG-MOCVD is a powerful tool to fabricate OEICs and is also effective to extract true surface kinetics during MOCVD.GaAs-MOCVD process was examined by SAG analysis.

Below 600ºC, surface kinetics shows non-linear behavior.Surface reaction rate constant of adsorbed species was constant

against offset angle, while adsorption equilibrium constant has a offset angle dependency.

S/Zn doping shows little or no effect on surface kinetics.

InGaAsP PL wavelength was well predicted by SAG simulation based on the obtained kinetics.Mask design for OEICs is possible based on kinetic data

base and kinetic simulation.

comsol multiphysicsと実験を併用した 薄膜製造プロセスの解析 · 2012-02-20 · 1...

Documents

comsol multiphysicsと実験を併用した薄膜製造プロセスの解析 · 2012-02-20 · 1...