2005-2006 International Distinguished Lecturer Programken/UFFC_IDL/SAW_Simulation.pdf · 2005-2006...

Sponsored byThe Institute of Electrical and Electronics Engineers (IEEE)Ultrasonics, Ferroelectrics and Frequency Control (UFFC) Society

U

2005-2006 International Distinguished Lecturer Program

Ken-ya HashimotoChiba University

Simulation of Surface Acoustic Wave Devices

Current Status and Future Prospects

Ken-ya HashimotoChiba University

[email protected]://www.em.eng.chiba-u.jp/~ken

SAW Device SimulationSAW Device Simulation! Background! COM analysis for rapid simulation! FEM/SDA analysis for parameter determination! Correction for SH-SAW simulation! Transverse-mode analysis! FEM/BEM analysis for precise simulation! Summary and remaining problems

SAW Device SimulationSAW Device Simulation! Background!!! COM analysis for rapid simulationCOM analysis for rapid simulationCOM analysis for rapid simulation!!! FEM/SDA analysis for parameter determinationFEM/SDA analysis for parameter determinationFEM/SDA analysis for parameter determination!!! Correction for SHCorrection for SHCorrection for SH---SAW simulationSAW simulationSAW simulation!!! TransverseTransverseTransverse---mode analysismode analysismode analysis!!! FEM/BEM analysis for precise simulationFEM/BEM analysis for precise simulationFEM/BEM analysis for precise simulation!!! Summary and remaining problemsSummary and remaining problemsSummary and remaining problems

Rayleigh-Type SAW

Propagation Direction

Depth

Surface Acoustic WaveSurface Acoustic WaveBulk Acoustic Wave (BAW):•Longitudinal Wave (Primary Wave)

•Transverse Wave (Share Wave, Secondary Wave)

Surface Acoustic Wave (SAW):

(Rayleigh SAW)Propagation of

Seismological Waves

Vout

Vin

Piezoelectric substrate

Interdigital Transducers (IDT)

SAW

! Mass-Production by Photolithography! Low-loss, Small-Size & Low Price! Operation in VHF-UHF Range

Line width λ/4=0.5µm (f=2GHz)

SAW Transversal FiltersSAW Transversal Filters

W

wL1=N1pI

wLT

IDT-1 IDT-2

x t

pI

(a) surface deformation (b) SAW propagation

τI

pI pI pI pI pI pI pI pI pI

L2=N2pI

Delta Function ModelDelta Function Model

2/1ˆ −= ii NN

Delta Function Model Analysis

Coupling-of-Modes Analysis

Reference: Influence of Internal Reflection

Surface Acoustic Wave (SAW) Resonator

inV


Interdigital Transducer (IDT)

Grating (Bragg) Reflectors

"It seems well-developed. Why further research necessary?"

System designers need devices with zero insertion loss, zero TCF, 100dB out-of-band rejection, 200%

bandwidth, one day delivery, and price minimization.

Why Fast & Precise Simulators Needed?Why Fast & Precise Simulators Needed?

! Improving Production Yields

Current Computers with Mass Memories are! Low Price, Maintenance Free, Space Saving! Working 365 days with 24 hours

! No Friction with Workers Union

! Better Understanding of Underlying Physics! Reducing Number of Trials

! Meaningful Optimal Design

What are Necessary for Simulators?What are Necessary for Simulators?

! Improving Production Yields

! Better Understanding of Underlying Physics! Reducing Number of Trials

! Meaningful Optimal Design

AccuracyAccuracySpeed

For Speedy Simulation

For Accurate Simulation

Behavior-Model Based Simulation

Full Wave Simulation

Practical Simulation ProcedurePractical Simulation Procedure

Pseudo 2D Simulation of Whole Structure

Pseudo 3D Simulation of Whole Structure

Simulation of Package Characteristics by Commercial Tools

Derivation of Simulation Parameters by 2D (X-Z)

Simulation

Derivation of Simulation Parameters by 2D (X-Y)

Simulation

3D Full Wave Simulation is not Realistic

Simulation Including Package Influence

1nH ( ≈ gold wire of 1mm length)

6Ω/mm @ 1GHz

1pF ( ≈ 1mm×1mm pad on LiNbO3 substrate)

150Ω/mm2 @ 1GHz

Influence of parasitic impedances is significant in GHz range!

SAW Device SimulationSAW Device Simulation!!! BackgroundBackgroundBackground! COM analysis for rapid simulation!!! FEM/SDA analysis for parameter determinationFEM/SDA analysis for parameter determinationFEM/SDA analysis for parameter determination!!! Correction for SHCorrection for SHCorrection for SH---SAW simulationSAW simulationSAW simulation!!! TransverseTransverseTransverse---mode analysismode analysismode analysis!!! FEM/BEM analysis for precise simulationFEM/BEM analysis for precise simulationFEM/BEM analysis for precise simulation!!! Summary and remaining problemsSummary and remaining problemsSummary and remaining problems

Pseudo 2D Simulation of SAWRPseudo 2D Simulation of SAWR

inV


Interdigital Transducer (IDT)

Grating (Bragg) Reflectors

! Combination of Periodic Structures! Wide Aperture

Accurate Simulation by 1-D Model

! Equivalent Circuit Analysis! Coupling-of-Mode (COM) Analysis! p-Matrix Analysis (Hybrid Scattering Matrix)

Modeling of Element (IDT)Modeling of Element (IDT)

! Number of Independent Parameters! Simple to Use! Easy to Extend! Many Articles

Compatible (Basically Equivalent)

U-

U+

V0

δI

U++δU+

U-+δU-

COM Equation for Periodic StructureCOM Equation for Periodic Structureβu: Wavenumber(Unperturbed)κ : Reflection Coef.ζ : Transduction Coef.θu=βu-2π/pI

pI: IDT Period

044 CVjUjUjxI

VjUjUjx

U

VjUjUjx

U

u

u

ωζζ

ζθκ

ζκθ

+−−=∂∂

−++=∂

∂

+−−=∂

∂

−+

−+−

−++ Propagation

Reflection

DetectionCapacitance

Excitation

00

00

)exp()exp()()exp()exp()(

VXjAXjAXUVXjAXjAXU

pp

pp

ξθθξθθ

+++−Γ=

++Γ+−=

−+−

−++

General Solution

where θp=(θu2-κ2)0.5, Γ0=(θp-θu)/κ, ξ=ζ/(θu+κ)

dXCVjUUjdXXXII

L

L

L

L

])(4[)(0

2/

2/

2/

2/

*0 ωζζ ++−=

∂∂= −+

+

−

+

−∫ ∫

Total Current Input

L: IDT Length

[P]

V0

U+(0)U-(0)

−−=

−

+

+

−

0332313

232221

131211

0

)()0(

44)()0(

VLU

U

PPPPPPPPP

ILU

U

U+(L)U-(L)

I0

P-Matrix Expression

Reflector ReflectorIDTGap(ϕ) Gap (ϕ)

Cascade-ConnectionCascade-Connection

VinIin

Short-Circuited Grating = Short-Circuited IDT

IDT Modeling Modeling of Whole Structure

Open-Circuited Grating = Open-Circuited IDT

Gap = IDT without reflection & Excitation

SAW Device SimulationSAW Device Simulation!!! BackgroundBackgroundBackground!!! COM analysis for rapid simulationCOM analysis for rapid simulationCOM analysis for rapid simulation! FEM/SDA analysis for parameter determination!!! Correction for SHCorrection for SHCorrection for SH---SAW simulationSAW simulationSAW simulation!!! TransverseTransverseTransverse---mode analysismode analysismode analysis!!! FEM/BEM analysis for precise simulationFEM/BEM analysis for precise simulationFEM/BEM analysis for precise simulation!!! Summary and remaining problemsSummary and remaining problemsSummary and remaining problems

Full Wave Analysis for Infinite Grating! Perturbation or Functional Method ⇔

inaccurate (?)

3800381038203830384038503860387038803890

0 0.05 0.1 0.15 0.2Relative Al Thickness h/λ

SAW

Vel

ocity

(m/s

)

SAW velocity on Al/128oYX-LiNbO3

Complex variation at very thin film thickness

No guarantee of rapid convergence for perturbation analysis


inaccurate (?)! Boundary Element Method (BEM) ⇔

complex for Green function calculation! Finite Element Method (FEM) ⇔

slow convergence (?)

11.021.041.061.081.1

1.121.141.16

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1Inverse of number of subdivision, 1/(N-1)

Nor

mal

ized

cap

acita

nce

Piece-wise expression

with x-0.5 dependence

BEM calculation of IDT capacitance

∑−

= −=

1

02)/2(1

)/2()(N

n

nn

wXwXAXq

Used function

Charge concentration at electrode edges ((((∝∝∝∝x-0.5))))

What should be used for

infinite case?

FEM Analysis Example


inaccurate (?)! Boundary Element Method (BEM) ⇔


slow convergence (?)

! Spectrum Domain Analysis (SDA) ⇔simple analysis, fast convergence but applicable structure limited

FEM/SDA

Blφφφφtekjær's Method for infinite IDTsBlφφφφtekjær's Method for infinite IDTs

∑

∑∞+

−∞=

+∞

−∞=

−−−−

=

−−−=

n GG

gn

n GG

Gn

wXnXjXB

Xe

wXnXjAXq

)2/cos()cos()]2/1(exp[)sgn(

)(

)2/cos()cos()]2/1(exp[)(

βββββ

β

Field Expression (Rigorous for electro-statics)

How shall we include mechanical effects?

For non-piezoelectric case, Bm=-jεAm

FEM/SDA AnalysisFEM/SDA Analysis

SAW Propagation Direction

Finite Element AnalysisFor Arbitrary Electrode Cross-Section (+ Analytic Solution not Available)

Piezoelectric Substrate

Metal Electrode

Flat Substrate Surface

Analytic Solution = Fast Analysis

Spectral Domain Analysis

Boundary Condition：Minimization of Radiated Power (Error) from Boundary

Characterization of Wave Properties in Infinite Structures

Dispersion Characteristics of Rayleigh-type SAW on 128-LN. Blue: FEMSDA, Red: COMVB=4,025 m/s (Slow-shear SSBW velocity)

3880

3900

3920

3940

3960

3980

4000

0.45 0.46 0.47 0.48 0.49 0.5 0.51 0.5200.10.20.30.40.50.60.7

Relative frequency, fp/VB

Phas

e ve

loci

ty (

m/s

)

Atte

nuat

ion

(dB

/λ)

OC

SC

h=2%λ

COM Parameter Determination by Numerical Fitting

00.10.20.30.40.50.60.7

0.47 0.48 0.49 0.5 0.51

Atte

nuat

ion

(dB

/λ)

OS

SS

3880390039203940396039804000

0.47 0.48 0.49 0.5 0.51Ph

ase

velo

city

(m/s

ec)


SS

OSOS

SS

Relative Frequency, fp/VB

h/p=0.01h/p=0.02h/p=0.03h/p=0.04

h/p=0.01h/p=0.02h/p=0.03h/p=0.04

Dispersion Relation vs. Al Thickness

0.02

0.03

0.04

0.05

0.06

0.07

0 0.01 0.02 0.03 0.04Al Thickness/Grating-Period

Rel

ativ

e C

OM

Par

amet

ers Ku

c-110κ12

κ11

2

Change in COM Parameters with Al Thickness

Cp

K Iu ω

ζπ 22 = : Electromechanical Coupling Factor

for Perturbed Modec=VB/Vref VB=4,025 m/s (Slow Shear SSBW)

-0.08

-0.06

-0.04

-0.02

-0.00

0.02

0.04

0.0 0.2 0.4 0.6 0.8 1.0R

efle

ctiv

ity,

r

Metallization ratio, w/p

short-circuited grating

open-circuited grating

h/p=0.06

Rayleigh-type SAW on 128oYX-LiNbO3Solid Lines: FEMSDA,　+×: Experiment

3.803.823.843.863.883.903.923.943.963.984.00

0.0 0.2 0.4 0.6 0.8 1.0

Effe

ctiv

e ve

loci

ty, V

e(k

m/s

)

Metallization ratio, w/p

open-circuited grating

short-circuited grating

h/p=0.06

Comparison with ExperimentComparison with Experiment

Correction of Simulation ParametersCorrection of Simulation Parameters

! Uncertainties in Substrate Material Constants (Supplier and Lot Dependent)

! Uncertainties in Film Material Constants (Fab. Process Dependent)

! Electrode Cross-Section (Fab. Process Dependent)

Although their Absolute Values may be Doubtful, Dependencies on Device Parameters might be held

SAW Device SimulationSAW Device Simulation!!! BackgroundBackgroundBackground!!! COM analysis for rapid simulationCOM analysis for rapid simulationCOM analysis for rapid simulation!!! FEM/SDA analysis for parameter determinationFEM/SDA analysis for parameter determinationFEM/SDA analysis for parameter determination! Correction for SH-SAW simulation!!! TransverseTransverseTransverse---mode analysismode analysismode analysis!!! FEM/BEM analysis for precise simulationFEM/BEM analysis for precise simulationFEM/BEM analysis for precise simulation!!! Summary and remaining problemsSummary and remaining problemsSummary and remaining problems

Propagation Direction

Depth

SH(Shear-Horizontal)-Type SAW

Properties of SH-type SAWsProperties of SH-type SAWsSimilar to Rayleigh-Type SAWs (SV+L) at First Look

•Influence of Bulk Wave Radiation

•Frequency Dependent Velocity

•Frequency Dependent Piezoelectricity

Incident SAWBragg-Reflected

SAW

Bragg-Reflected BAW

Back-Scattering to BAW

For SH-type SAW,Cutoff For BAW Back-Scattering≈SAW Resonance Frequency

Atte

nuat

ion

(dB

/ )

385039003950400040504100415042004250

0.4 0.44 0.48 0.52 0.56 0.60

0.5

1

1.5

2

2.5

3

Phas

e ve

loci

ty (

m/s

)


λ

Dispersion in phase velocity & attenuation of SH-type SAW on SC grating on 42-LT. Blue: calculated by FEMSDA, and red: calculated by conventional COM.

Stopband BAW backscattering

Frequency

Am

plitu

de SAW Response

BAW Radiation

Stored Energy(∝ Velocity Reduction)

BackscatteredBAW cutoff

Origin of dispersion near stopband

385039003950400040504100415042004250

0.4 0.44 0.48 0.52 0.56 0.6

OCSC

Phas

e ve

loci

ty (

m/s

)


Phase velocity of SH-type SAW for grating structure on 42-LT. Blue:calculated by FEMSDA, and red: calculated by Plessky's model.

generation detection

SAW radiationconductance + >0

generation detection

SH-typeSAW

BAWinternal reflection

BAW radiationconductance

Coupling between SAW and BAW radiation through internal reflection.

-15-10

-505

10152025

865 870 875 880 885 890 895

conductance

susceptance

Frequency (MHz)

Rel

ativ

e ad

mitt

ance

Y/G

0

Experimental (blue) and calculated (red) input admittance of finite synchronous resonator. (G0=20 mS)

Simulation of SAW Resonator for Ladder-Type Filters

Simulation of SAW Resonator for Ladder-Type Filters

-1.5

-1

-0.5

0

0.5

1

1.5

840 860 880 900 920 940 960

Rel

ativ

e ad

mitt

ance

Y/G

0

Frequency (MHz)

conductance

susceptance

Experimental (blue) and calculated (red) input admittance of finite synchronous resonator. (G0=20 mS)

Lateral Leakage BAW Radiation from Finite IDT

Slowness Surface of SH-type SAW on 36-LTSlowness Surface of SH-type SAW on 36-LT

! Beam-like Propagation (No Guided Transverse Mode)

! Power Leakage through Bus-Bar Regions

! Frequency Dependent Behavior

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0 0.1 0.2 0.3Sx sec/km

S yse

c/km

1800 1900 2000

-40

-20

0

Atte

nuat

ion

[dB]

Frequency [MHz]

Metallization-ratio

50%35%

1800 1900 2000

-40

-20

0

Atte

nuat

ion

[dB]

Frequency [MHz]

Metallization-ratio

50%35%

IDT

Busbar

BusbarTransverse Leakage

Reflector

Inte

nsity

[a.u

.] High

Low

IDT

Busbar


ReflectorIDT

Busbar


Reflector

Inte

nsity

[a.u

.] High

Low

Inte

nsity

[a.u

.] High

Low

Laser Probe Image Impact of Metallization Ratio to Transverse Leakage

Presented at 2003 IEEE FCS By Tstutsumi, et al. (Fujitsu Labs)

COM analysis of DMS FiltersCOM analysis of DMS Filters

-100

-80

-60

-40

-20

0

800 850 900 950 1000 1050 1100Frequency (MHz)

Scat

terin

g Pa

ram

eter

(dB

)

Experiment

Simulation

S11

S21

Overestimated Propagation Loss

! Influence of DiscontinuitiesNo Practical Theory ⇒ Phenomenological Model(additional phase shift)

! Influence of BAWsModeling has not been Established yet

Difficulties in Simulation of DMS FiltersDifficulties in Simulation of DMS Filters

Discontinuities

SAW Device SimulationSAW Device Simulation!!! BackgroundBackgroundBackground!!! COM analysis for rapid simulationCOM analysis for rapid simulationCOM analysis for rapid simulation!!! FEM/SDA analysis for parameter determinationFEM/SDA analysis for parameter determinationFEM/SDA analysis for parameter determination!!! Correction for SHCorrection for SHCorrection for SH---SAW simulationSAW simulationSAW simulation! Transverse-mode analysis!!! FEM/BEM analysis for precise simulationFEM/BEM analysis for precise simulationFEM/BEM analysis for precise simulation!!! Summary and remaining problemsSummary and remaining problemsSummary and remaining problems

Frequency

Adm

ittan

ce

ωr ωa

B

GIn-harmonic Resonance

3D Simulation of Whole Structure3D Simulation of Whole Structure

Origin of Transverse

Mode

Precision Stage

Optics

IF Detector

Stage Controller RF Source

LO Source

CCD camera

Data Logger

N Divider

10

100

1k

428 430 432 434 436 438Im

peda

nce

(Ω)

Frequency (MHz)

(a) (b) (c)

SAW Resonator

20 min. for 2,620×410 pixels

1-port SAW ResonatorPeriod: 7.18 µmSubstrate: ST-Cut QuartzElectrode: Al (140 nm)

Magnitude (433.00 MHz)



700 800 900 1000

0.3

0.2

0.1

0

-0.2

-0.1

0.5

0.4

Frequency [MHz]

Adm

ittan

ce [S

]

IDT Resonator on Cu-Grating/15oYX-LiNbO3IDT Resonator on Cu-Grating/15oYX-LiNbO3

G

B

1300Frequency, f [MHz]

700 800 900 1000 1100

0

10

20

30

50

Inse

rtion

loss

[dB

] min IL=0.5 dB

BW3dB=152MHz (15%)

40

1200

h/2p=7.2-9.3%

Wideband SAW filter on Cu/15oYX-LiNbO3 structureWideband SAW filter on Cu/15oYX-LiNbO3 structure

Presented at IEEE 2004 UFFC Conf.

Measured Field distribution by Laser Probe

340.30 MHz

373.75 MHz363.24 MHz350.00 MHz

Presented at 2005 IEEE Ultrasonics Symp.20min. for each image

Transverse Mode AnalysisTransverse Mode AnalysisIgnoring

Depth Dependence

Simplifying IDT Region as Uniform

Layer

wG

wB

wB

Region B

Region G

Region B

x

y

Mode Orthogonality + Completeness Enable us to Estimate Mode Excitation Efficiency

Res

onan

ce F

requ

ency

, fr[M

Hz]

Aperture, w/λ

Rayleigh mode

S0

S1

S2

S3

S4

S5

Calculated and Experimental Spectra of Transverse Modes

0 10 20 30

850

830

840

820

800

810

7905 15 25

Weighted Dummy Electrodes for Suppression of Transverse Modes

Scattering of Unnecessary Modes Through Coupling with Dummy Electrode Modes

1.3Frequency, f [GHz]

0.7 0.8 0.9 1 1.1

Inse

rtion

loss

[dB

]

0

10

20

30

40 1.2

W Weighted Dummy ElectrodesW/O Weighted Dummy Electrodes

1.3Frequency, f [GHz]

0.7 0.8 0.9 1 1.1

Inse

rtion

loss

[dB

]

0

10

20

30

40 1.2

SAW Filter Using Cu/15oYX-LiNbO3 Structure

SAW Device SimulationSAW Device Simulation!!! BackgroundBackgroundBackground!!! COM analysis for rapid simulationCOM analysis for rapid simulationCOM analysis for rapid simulation!!! FEM/SDA analysis for parameter determinationFEM/SDA analysis for parameter determinationFEM/SDA analysis for parameter determination!!! Correction for SHCorrection for SHCorrection for SH---SAW simulationSAW simulationSAW simulation!!! TransverseTransverseTransverse---mode analysismode analysismode analysis! FEM/BEM analysis for precise simulation!!! Summary and remaining problemsSummary and remaining problemsSummary and remaining problems

Full Wave Analysis for Finite Structure! Perturbation or Functional Method ⇔

inaccurate! Boundary Element Method (BEM) ⇔


slow conversion

Rigorous Analysis of arbitrary 2D structure

FEM/BEM

This simulation is time consuming but current computers are a little powerful

Structure of PitchPitch--Modulated IDT & ReflectorModulated IDT & ReflectorConventional Structure PitchPitch--Modulated StructureModulated Structure

IDT IDTREFECTOR

ModulatedModulated

ModulatedModulatedGap ControlledGap Controlled


Relative frequency

-100

-80

-60

-40

-20

0

-0.15 -0.1 -0.05 0 0.05 0.1 0.15Scat

terin

g co

effic

ient

S21

(dB

) ModulatedConstant

Designed Results

Wideband & Without gaps between IDTsPresented at IEEE 2002 Ultrasonics Symp.

Red line: Modulated-Pitch Structure (Lg=0λ）Blue line: Unmodulated-Pitch Structure (Lg=0.3λ）

COM Analysis (BAW Ignored)

∆IL=0.1dB

Remaining Loss (1-|S21|2-|S11|2)


∆∆∆∆IL=0.4dB

FEM/BEM Analysis (BAW Incl.)

Experimental Verification

∆IL=0.4dB

Red line: Modulated-Pitch StructureBlue line: Unmodulated-Pitch Structure (Lg=0.3λ）

Remaining Loss


DMS Filter Using Modulated StructureDMS Filter Using Modulated Structure

Courtesy of Fujitsu Media DevicesPresented at IEEE 2004 UFFC Conf.

SAW Device SimulationSAW Device Simulation!!! BackgroundBackgroundBackground!!! COM analysis for rapid simulationCOM analysis for rapid simulationCOM analysis for rapid simulation!!! FEM/SDA analysis for parameter determinationFEM/SDA analysis for parameter determinationFEM/SDA analysis for parameter determination!!! Correction for SHCorrection for SHCorrection for SH---SAW simulationSAW simulationSAW simulation!!! TransverseTransverseTransverse---mode analysismode analysismode analysis!!! FEM/BEM analysis for precise simulationFEM/BEM analysis for precise simulationFEM/BEM analysis for precise simulation! Summary and remaining problems

! Pseudo 2D Simulation of Whole Structure

! Derivation of Simulation Parameters by 2D (X-Z) Simulation

! Correction of Simulation Parameters! Simulation of SH-type SAW! Precise Simulation of Ladder-Type Filters! Transversal Mode Analysis! Package Modeling

What we can do ......

What we should do next?! Fast & Precise Simulation of DMS-Type Filters

(Including Frequency Dependent K2 and Attenuation)

! Pseudo 3D Simulation of Whole Structure (Including Lateral Leakage)

! Rapid Package Simulation for Optimal Design! Finding Remaining Loss Mechanism! Precise Simulation Tools for More Complicated

Structures

2005-2006 International Distinguished Lecturer Programken/UFFC_IDL/SAW_Simulation.pdf · 2005-2006...

Documents

Transcript of 2005-2006 International Distinguished Lecturer Programken/UFFC_IDL/SAW_Simulation.pdf · 2005-2006...