Optimizing Piezoelectric Crystal Preload in Ultrasonic Transducers
Dominick A. DeAngelis
And
Gary W. Schulze
Transducer
Device
Kulicke & Soffa’s Flagship
Semiconductor
Wire Bonding Machine
38th UIA Symposium, Vancouver 3/23/09
Wire Bonding in Action with Fine
Gold Wire (20 Wires/Sec)
Ultrasonic Transducer Used For
Wire Bonding Machine
-1-
OUTLINE
 Motivation for the Work
 Specific Transducer Application
 Research Summary
 Experimental Methods & Metrics
 Electromechanical Coupling
 Equivalent Circuits
 Bode Plot & Admittance Loop
 Experimental Results
 Conclusions
 References
 Questions
38th UIA Symposium, Vancouver 3/23/09
-2-
MOTIVATION FOR THE WORK
 Preload is Required to Maintain the Piezo Stack in Compression
 Most Often Integral Piezo Crystals Used as Force Sensor for Preload
 A Press Fixture w/ Charge Amp Measures Voltage vs. Force
 Method Inaccurate Since Crystal Properties Can Vary ±20% Lot-to-Lot
 Vendor Preload Guidelines Based on Static Crystal Testing Only
 Designers Often Struggle to Determine/Control Optimal Preload Level
 Transducer Based Methods and Metrics are Needed
38th UIA Symposium, Vancouver 3/23/09
-3-
MOTIVATION CON’T
Why is Controlling Preload So Important?
 Inadequate Preload Results in Dynamic Gapping at Interfaces
 Dynamic Gapping Manifests as Higher Impedance, Heating, Etc.
 Inadequate Preload Results in Prolonged Stress-Aging Affects
 Excessive Preload Results in Pronounced Depolling
 Excessive Preload Produces Unstable Impedance and Aging with Use
38th UIA Symposium, Vancouver 3/23/09
-4-
SPECIFIC TRANSDUCER APPLICATION
 K&S is the Leading MFG of Semiconductor Wire Bonding Equipment
 Transducer Delivers Energy to a Capillary Tool for Welding Tiny Gold Wires
 Patented Single Piece “Unibody” Transducer Design Ideal for Preload Study
 Portability Across 100’s of Machines Required for Same Customer Device
Crystals
Transducer
Body
On
Machine
Transduce
r
1¢ US
Device
Capillary
Tool
PZT8 Piezoelectric Crystals
38th UIA Symposium, Vancouver 3/23/09
Transducer Specs
500 mA Max Current
120 kHz Operating Mode
80 Ohm Max Impedance
Operation 40 Bonds/Sec
Bond Duration ~10mSec
PZT8 Crystals (4X)
Capillary
Actual Wire Bonds From
of a Multi-Tier Package
Typical Capillary Tool with Wire
Compared to Sewing Needle
-5-
RESEARCH SUMMARY
 Five Transducers Built at Various Preload Levels from 4.5 ksi to 18 ksi
 All 5X Transducers Bodies and 20X Crystals from Same Production Lot
 Experimental Methods were Both Crystal and Transducer Based
 Transducers Subjected to Stabilizing Heat-Treatment & Cycling After Build
Probe
Tweezers
Contact at
Dynamic
Node
1¢ US
Crystal Based Bode Plot Technique
38th UIA Symposium, Vancouver 3/23/09
The Five Transducers Build for the Study
-6-
EXPERIMENTAL METHODS & METRICS
 Bode Plots, Cap, DF of Individual Crystals Before Build
 Calibrated Stack Voltage vs. Force Prior to Assembly
 Bode Plots, Cap, DF of Transducer After Build and Heat-Treat
 Laser Vibrometer Gain Measurements After Heat-Treatment & Cycling
Charge
Amp
Tinius
Olsen
Network
Analyzer
Crystals
Scanning
Laser
Vibrometer
38th UIA Symposium, Vancouver 3/23/09
Crystal Voltage vs. Force Calibration
Fixture (Before Assembly)
-7-
METHODS & METRICS CON’T
 Results from Voltage vs. Force Calibration Fixture
 Piezoelectric Constant d33 Increases with Applied Stress
Charge Amp Voltage & d33 Versus Applied Stress
5
550
Voltage
d33
+
4
Charge Amp Voltage (V)
515
480
2F
PZT
Vout
-
3.5
445
3
410
Effective d33 Change
Including Non-uniform
Loading Effects,
Flatness, etc.
2.5
2
375
Voltage Recorded for
Desired Preload Prior
to Assembly Using
Calibration Fixture.
The Calibrated
Voltage is Then Used
for Transducer Build
1.5
1
0.5
0
0
2.5
5
7.5
10
12.5
Stress (ksi)
th
Charge Density/Stress (m/C * 10^12)
4.5
38 UIA Symposium, Vancouver 3/23/09
15
17.5
20
22.5
340
305
270
235
200
25
-8-
METHODS & METRICS CON’T
 Crystal Laser Vibrometry and Finite Element Analysis
 Measured Out-of-Plane Motion From In-plane
Structural Vibration (Poisson Effect)
Single Crystal Vibrometer
Setup (3X Mirrors)
Long Edge
Crystal Face
Example of weakly
coupled mode
Short Edge
1.E+02
Finite Element Model
Velocity
1.E+00
1.E-02
1.E-04
1.E-06
Excitation and
Velocity Direction
1.E-08
100
150
200
Mechanical Admittance Plot
38th UIA Symposium, Vancouver 3/23/09
250
300
350
400
Frequency, kHz
-9-
450
500
METHODS & METRICS CON’T
 Bode Plot and Corresponding Finite Element Admittance
 Single Crystal Analysis Demonstrates Short and Open Circuit Properties
Typical Bode Plot of Admittance Vs. Frequency of Single PZT8 Crystal
Bode Plot of Admittance Vs. Frequency for Single PZT8 Crystal
1
200
80
Magnitude (mS)
0.01
 40
E
Y 33
YD33
 160
Catalog
7.4
11.8
 280
Correlated
8.9
13.7
 400
10e-12
Pa
 520
0.001
Phase (Degrees)
0.1
Elastic Moduli
 640
 760
0.0001
Magnitude
Phase
0.00001
100
 880
150
200
Short Circuit
Properties
1.E+02
1.E+00
Velocity
250
300
350
400
 1000
500
450
Frequency (kHz)
Open Circuit
Properties
1.E-02
1.E-04
1.E-06
1.E-08
100
150
200
250
Calculated Mechanical Admittance
38th UIA Symposium, Vancouver 3/23/09
300
Frequency, kHz
350
400
Anticipated
Phase450
Window
(fa – fr)
-10-
500
ELECTROMECHANICAL COUPLING k33
 k33 Varies Proportionally to d33, But Inversely to Square Root of Permittivity
 k33 Varies Proportionally to Delta Between Open and Short Circuit Moduli
Deflection Based on Short
E
Circuit Modulus Y33
Apply Load F with
Electrodes Shorted
Relieve Load F with
Electrodes Open
Deflection Based on Open
D
Circuit Modulus Y33
Y33D >Y33E
F
F
F
+
Shorted
+
-
Crystal Returns to
Original State After
Charge Leakage
Polled
1
-
F
+
+
+
1
-
2
-
-
F
U mechanical 
F
1
1
2
K E12  K D 1   2 
2
2
K D  Y33D
Area
Thickness
K E  Y33E
Area
Thickness
Load
1
U electrical  CV 2
2
KD
KE
2
1
T
C   33
Area
Thickness
V
F
d 33 K D
k33 
U electrical
Electrical Energy Stored

Mechanical Energy Supplied
U mechanical
Deflection
Net Mechanical
Energy Supplied
k33  d33Y33D
Proportional
38th UIA Symposium, Vancouver 3/23/09
Y33D  Y33E
T
Y33E  33
k33  .63
 PZT8
Inversely
-11-
EQUIVALENT CIRCUITS
 For fs Series Resonance ipll  imo and ipll >> ico (Via Short Circuit Modulus)
 For fp Parallel Resonance ico  -imo and ipll << imo (Via Open Circuit Modulus)
 Spacing of fp - fs Proportional to Delta Between Y33D  Y33E (And k33 Too)
Parasitic
Capacitance
i pll
PLL Drive
V (t )
38th UIA Symposium, Vancouver 3/23/09
Imaginary ( j )
Piezo Motor
Z mo  R
Real
θ
imo
ic0
C0
R
Z C0 
Z pll
L
fs 
C
fp 
-12-
1
2 LC
1
2
C0  C
LC0C
1
2 f mo C0
ADMITTANCE (Y) LOOP
Imaginary Axis
Gmax
2
f1
Susceptance B (mS)
Y  G  jB 
Admittance
1
Z
Mechanical
“Q”
f  Hz 
Qm 
fs
2 f s L
1


2 f s CR
f 2  f1
R
Electrical
“Q”
Qe 
Bs
 2 f s C0 R
Gmax
Susceptance
fY max
Conductance
Z  R  jX 
Impedance
V 1

I Y
Reactance
k33  1 
Resistance
f s2
C

2
fp
C0  C
Zero Phase and Max Current at
Piezo Motor for Series Mode
2 f s C0  Bs
C0 
fs 
1
2 LC
B
f r2
C1  s
2
2
fa  fr
2 f s
Magnitude Minimum of
Electrical Anti-resonance
Max Z, Min Current
1
fp 
2
Magnitude Peak of
Electrical Resonance,
Min Z, Max Current
C0  C
LC0C
fY min
R
1
Gmax
L
Qm R
2 f s
C
1
Qm R 2 f s
fp
Zero Phase at Piezo
Motor for Parallel Mode
fa
Zero Phase, Electrical
Anti-resonance
38th UIA Symposium, Vancouver 3/23/09
fr
f2
Gmax 
1
R
Real Axis
Conductance G (mS)
Zero Phase, Electrical
Resonance
fY min  f p  f a  f 2  f r  f s  fY max  f1
-13-
EXPERIMENTAL RESULTS
 Typical Single Free-Free Crystal Bode Plot Results Prior to Build
 k33 Proportional to Phase Window fa –fr (Similar for All Crystals)
 Dissipation Factor DF was .004 for All Crystals (Tan)
Bode Plot of Admittance Vs. Frequency for Single PZT8 Crystal
0.1
100
Short Circuit
Series Mode
50
fa  fr
0.001
fr
0
fa
 50
0.0001
Open Circuit
Parallel
Mode
Magnitude
Phase
0.00001
390
395
400
405
410
415
420
425
430
435
440
445
450
455
 100
460
Frequency (kHz)
38th UIA Symposium, Vancouver 3/23/09
-14-
Phase (Degrees)
Magnitude (mS)
0.01
EXPERIMENTAL RESULTS CON’T
 Transducer Bode Plot Results After Stabilizing Heat-Treatment
 k33 Proportional to Phase Window fa –fr
Bode Plot of Admitance Vs. Frequency for Various Piezo Stack Preload
0.06
0.05
Magnitude (mS)
0.045
0.04
0.035
0.03
50
fr
fa  fr
0
fa
 50
Minimum
Impedance
 100
Max
Phase
Window
0.025
 150
0.02
 200
0.015
0.01
 250
0.005
0
114
115
116
117
118
119
120
121
122
123
124
Frequency (kHz)
38th UIA Symposium, Vancouver 3/23/09
-15-
 300
125
Phase (Degrees)
0.055
100
4.5 ksi
9 ksi
11.5 ksi
14 ksi
18 ksi
4.5 ksi
9 ksi
11.5 ksi
14 ksi
18 ksi
EXPERIMENTAL RESULTS CON’T
 Transducer Admittance Loop Bode Results After Stabilizing HT
 Find Preload that Maximizes k33 and Qm, but Minimizes Qe
Bode Plot of Admittance (Y) Loop for Various Piezo Stack Preloads
0.02
4.5 ksi
9 ksi
11.5 ksi
14 ksi
18 ksi
Imaginary Axis - Susceptance B (mS)
0.016
0.012
0.008
Maximize k33
Maximize Qm
Minimize Qe
0.004
0
Increasing
C0
 0.004
Equivalent Circuit Properties
Preload
(ksi)
4.5
9
11.5
14
18
 0.008
 0.012
 0.016
 0.02
0
0.004
0.008
0.012
0.016
0.02
0.024
0.028
C0
(pF)
1232
1410
1861
1922
1926
0.032
C
(pF)
59
68
70
59
21
L
(mH)
31
26
25
29
81
0.036
R
(ohms)
66
33
34
37
84
fs
(kHz)
117.678
119.988
121.065
122.370
123.192
fa
(kHz)
120.433
122.857
123.313
124.241
123.878
fr
(kHz)
117.671
119.985
121.055
122.367
123.220
f1
(kHz)
117.510
119.885
120.957
122.267
123.110
0.04
Real Axis - Conductance G (mS)
38th UIA Symposium, Vancouver 3/23/09
-16-
f2
(kHz)
117.846
120.091
121.173
122.473
123.275
fp
(kHz)
120.444
122.548
123.334
124.234
123.861
k 33
Qm
Qe
0.2130
0.2145
0.1905
0.1730
0.1029
350
582
560
594
747
0.060
0.035
0.047
0.055
0.125
EXPERIMENTAL RESULTS CON’T
 Tool Displacement to Current Gain Results After Stabilizing HT & Cycling
Capillary Tool Displacement Versus Current Using Laser Vibrometer
( For Various Piezo Stack Preloads, After Stabilizing Heat Treatment)
3.25
y = 0.00590x - 0.02298
y = 0.00541x - 0.01278
3.00
2.75
y = 0.00574x + 0.00069
y = 0.00609x - 0.02058
y = 0.00942x - 0.02936
2.50
Displacement Magnitude (Microns)
High Gain
Indicates
Depolling
2.25
2.00
Slope is
Effective
Gain
1.75
1.50
1.25
1.00
0.75
0.50
0.25
0.00
0
50
100
150
200
250
300
350
400
450
Capillary Tool Displacement Versus Current Using Laser Vibrometer
( For Various Piezo Stack Preloads, After Stabilizing Heat Treatment and Burn-in Cycling)
500
3.25
y = 0.00606x - 0.02738
y = 0.00543x - 0.01178
y = 0.00585x - 0.01384
y = 0.00597x - 0.00437
3.00
Constant Current Pk-Pk Probe (mA)
Stack Preload 9 ksi
Stack Preload 11.5 ksi
Stack Preload 14 ksi
Stack Preload 18 ksi
2.50
Displacement Magnitude (Microns)
Stack Preload 4.5 ksi
2.75
2.25
y = 0.00942x - 0.00246
2.00
1.75
1.50
1.25
1.00
0.75
0.50
0.25
0.00
0
50
100
150
200
250
300
350
400
450
Constant Current Pk-Pk Probe (mA)
Stack Preload 4.5 ksi
38th UIA Symposium, Vancouver 3/23/09
Stack Preload 9 ksi
Stack Preload 11.5 ksi
-17-
Stack Preload 14 ksi
Stack Preload 18 ksi
500
EXPERIMENTAL RESULTS CON’T
 Impedance Vs. Current Results After Stabilizing Heat-Treatment & Cycling
Impedance Versus Current For Various Piezo Stack Preloads
(After Stabilizing Heat-Treatment)
140
Unstable
120
Impedance (Ohms)
100
80
60
40
Lowest Overall
Impedance and
Change with Current
20
Impedance Versus Current For Various Piezo Stack Preloads
(After Stabilizing Heat-Treatment and Burn-in Cycling)
140
0
0
50
100
150
200
250
300
350
400
450
500
120
Stack Preload 4.5 ksi
Stack Preload 9 ksi
Stack Preload 11.5 ksi
Stack Preload 14 ksi
Stack Preload 18 ksi
Impedance (Ohms)
100
Constant Current Pk-Pk Probe (mA)
80
60
40
20
0
0
50
100
150
200
250
300
350
400
450
Constant Current Pk-Pk Probe (mA)
Stack Preload 4.5 ksi
38th UIA Symposium, Vancouver 3/23/09
Stack Preload 9 ksi
Stack Preload 11.5 ksi
-18-
Stack Preload 14 ksi
Stack Preload 18 ksi
500
EXPERIMENTAL RESULTS CON’T
 Frequency Vs. Current Results After Stabilizing Heat-Treatment & Cycling
Frequency Versus Current For Various Piezo Stack Preloads
(After Stabilizing Heat-Treatment)
124
123
Frequency (kHz)
122
121
120
119
Higher Negative Slope
Indicates Dynamic Gapping
and Insufficient Preload
118
117
116
Frequency Versus Current For Various Piezo Stack Preloads
(After Stabilizing Heat-Treatment and Burn-in Cycling)
124
0
50
100
150
200
250
300
350
400
450
500
123
Constant Current Pk-Pk Probe (mA)
Stack Preload 9 ksi
Stack Preload 11.5 ksi
Stack Preload 14 ksi
122
Stack Preload 18 ksi
Frequency (kHz)
Stack Preload 4.5 ksi
121
120
119
118
117
116
0
50
100
150
200
250
300
350
400
450
Constant Current Pk-Pk Probe (mA)
Stack Preload 4.5 ksi
38th UIA Symposium, Vancouver 3/23/09
Stack Preload 9 ksi
Stack Preload 11.5 ksi
-19-
Stack Preload 14 ksi
Stack Preload 18 ksi
500
CONCLUSIONS
 Optimal Preload Range for PZT8 Found From 9 to 14 ksi (11.5 ksi Best)
 Design Methodology Presented for Optimizing Transducer Preload
 Insufficient Preload Causes High Impedance and Dynamic Gapping
 Excessive Preload Causes Severe Depolling and High Impedance
 Stress Related Aging Affects Minimized at Higher Preloads
 New Preload Method Using Calibrated Stack Voltages Presented
 Bode Plot Method for Individual Crystals Presented
38th UIA Symposium, Vancouver 3/23/09
-20-
REFERENCES
 Stansfield D., Underwater Electroacoustic Transducers, Peninsula
Publishing, CA, 1991
 Wilson, O.B., Introduction to Theory and Design of Sonar Transducers,
Peninsula Publishing, CA, 1991
 Sherman, C.H., Butler, J.L., Transducers and Arrays for Underwater
Sound, Springer, NY, 2007
38th UIA Symposium, Vancouver 3/23/09
-21-
QUESTIONS
38th UIA Symposium, Vancouver 3/23/09
-22-