446
VOL. 3X. NO. S. SEPTEMBER 1991
IEEE TRANSACTIONS
ULTRASONICS.
ON
FERROELECTRICS,
AND
FREQUEKCY
CONTROL,
Characterization of Lead Zirconate Titanate Ceramics
for Use in Miniature High-Frequency (20-80 MHz)
Transducers
F. Stuart Foster, Member, IEEE, Linda K . Ryan and Daniel H. Turnbull
Abstract-The development of new endoscopic applications of
ultrasound imaging is critically dependent on the availability of
efficient broadband transducers with areas of 2 mm2 or less.
The material properties of PZT ceramics for operation in the
thickness mode at frequencies as high as 80 MHz are reported.
Each of the ceramics tested showed a reduction in k, with increasing frequency. In a fine grained PZT, values of k, as high
as 0.44 were measured at 80 MHz. The effects of grain size were
also evident in the measurement of frequency dependent mechanical losses. Experimental and theoretical analysis of I mm2
45-kIHz PZT transducer verified the validity of the properties
measurements and demonstrated excellent insertion loss and
bandwidth characteristics. The minimum insertion loss of
-17.5 dBisin good agreement with theory and is a marked
improvement over the performance of polymer devices. Details
on the fabrication and testing of high frequency ceramic transducers are described.
I. INTRODUCTION
ANY NEWclinicalapplicationsofB-modeultrasound imaging at frequencies greater than 20 MHz
are currently under development. These include systems
designed to image the anterior segmentof the eye [l]-[2],
the skin [3]-[4], the gastrointestinal tract
[5]-[6]and intravascular imaging of blood vessels [7]-[lo]. Such systems hold the promise of providing subsurface detailwith
resolution approaching that of optical microscopy that is
unavailable using any other imaging means.
Although a great deal of progress has been made in the
aboveareas, theincreasedtissuelossesassociated
with
higherultrasoundfrequenciesdoeslimitimagingdepth
and image quality. Thus, optimization of transducer sensitivity and focusing properties
are critical factors in the
continuingevolution of thesesystems. In the past,ceramic transducers, because of their superior sensitivity,
have dominated applications up to frequencies of approximately 25 MHz [ 5 ] - [ 8 ] ,while polymer transducers have
dominatedapplicationsatfrequenciesgreater
than 25
MHz [2]-[4]. In this paper we describe investigations of
M
Manuscript received October 25, 1990; revised and accepted March I I .
1991. This work Was supported in part by the Medical Research Council
of Canada. in part by The National Cancer Institute of Canada. and in part
by the Ontario Graduate Scholarship program.
The authors are with the University o f Toronto, Sunnybrook Health ScienceCentre,ReichnlannResearch
Building, 2075 BayviewAve.North
York. O K . M 4 N 3M5 Canada.
IEEE Log Number 9101 152.
ceramicdevices in thefrequencyrangegreaterthan
25
MHz. The objective of ourwork is to develop transducers
that are 1 ) less than 2 mm’ in area, 2) matched to 50 Q
over the bandwidth of the transducer and 3) capable of
insertion losses of less than 20 dB. Such devices would
be particularly valuable in high-frequency endoscopic applicationssuchasintravascularimaging.Webegin
by
discussing the relevant material properties such as grain
size,
dielectric
properties,
coupling
coefficients
and
losses.
The development of practical devices
is illustrated by
treatingthecase of a 1 mm’ 45-MHz, PZT transducer.
Experimental and theoretical analysesof electrical impedanceforthefreeairresonator,backedtransducer
and
backedandtunedtransducerarepresented.Excellent
agreementbetweenexperimentalandtheoreticallypredicted insertion loss and wave shape is obtained. The results suggest that ceramic transducersmay offer improved
performanceoverpolymerdevices
in the 25-80MHz
range, particularly when the area
of the device is small
( < 2 mm’).
11. MATERIALS
A N D METHODS
Three types of lead zirconate titanate (PZT) ceramics
were used in this study. The first material (EC-65) is a
PZT5A mixture manufactured by Edo Western Ltd. (Salt
LakeCity,UT).Thesecond
andthird PZTceramics
(F3195 and D3203) are manufactured
by Motorola (Albuquerque, NM) using a high density process. For simplicity, we refertothesematerials
as PZT,, PZT’ and
PZT, respectively. Poled samples
of each material were
lapped to thicknessescorresponding to half-wavelength
resonances ranging from 50 to 80 MHz. Electrodes consistingof 200-A chromium and 3000-A goldwere deposited on the front and back surfaces of the lapped samples
and
the
samples were
cut
to
the
final lateral
dimensions using a Disco Abrasive Systems (Salem, NH)
dicing saw. Examples of 40-pm-thick PZT, samples made
in this fashion are shown in Fig. 1. These elements have
dimensions of 2mm2, 1 mm’ and 0.5 mm’. In Fig. 2 ,
scanningelectronmicrographs
of the3PZTmaterials
showsignificantvariation
in grainstructure.Thegrain
sizes were 7.5 f 3 pm, 3.2 & 0.5 pm and 2.5 f 0.6 pm
for PZT,, PZT2, andPZT,respectively.Since,
in this
0885-301019 1!0900-0446$01 .OO ‘3 1991 IEEE
FOSTER
PI
C H A R A C T E R I Z A T I O NOF
pz.r
447
CERAMICS
Fig. l . 40 p m thick PZT samples with area dimensions o f 4 mm’, I mm’.
and 0.25 m m 2 from left to rlght.
study, we are considering plate thicknesses as small as 30
pm, it is expected that grain size should play an important
role in determining transducer performance.
A . Dielectric Constant and Dielectric Loss Tungent
Since a piezoelectric material behaves as a capacitor in
frequency ranges away from resonance, the magnitude of
the dielectric constant, E , may be calculated as
6
= d/(wlZlto4,
(1)
where d is thesamplethickness,
121 is themeasured
impedance magnitude, A is the sample area, and eo is the
permitivity of freespace (8.85 x IO-’’ F / m ) . All experimental impedance measurements’were made with an
HP4 191A Impedance Analyser usingan HP16902A spring
clip fixture.
A plot of dielectric constants computed via (1) is shown
in Fig. 3 for a 50-pm-thick sampleof PZT?. The clamped
constant E’ was determined by averaging over a range of
frequencies in the slowly varying region between the first
and third harmonics. Similarly, the free constant. tT, was
calculated by averaging over arange of frequencies below
the first harmonic. The ranges used to calculate E’ and eT
are indicated in Fig. 3 .
Calculation of the dielectricloss tangent tan (6,) is given
by
(2)
tan (6,) = I / I tan 81
where 8 is the phase of the measured impedance. Free and
clamped loss tangents were computed by averaging over
the same frequency ranges as the dielectric constants.
B. Mechanical Losses und Coupling Coeflcients
Thethicknessmodecouplingcoefficient,
k,, andthe
mechanical losses of the piezoelectric material are critical
factors in determining
transducer
performance.
The
method most often used to detemline k, [ 1 l ] requires only
the determination of the series and parallel resonance frequencies. It has been observed that, while having the advantage of speed and simplicity, this method can intro-
IEEETRANSACTIONS
448
0
ON ULTRASONICS,FERROELECTRICS.ANDFREQUENCYCONTROL,
50
100
Frequency (MHz)
150
~VXV
Fig. 3 . hleasurementa o f dielectric constants. The clamped constants t S is
determined i n theconstantregionbetweenthe
first andthirdresonances
whereas. the free constant. c ' is calculated in the indicated region below
the first harmonic. Note that t r is always greater than tS.
duce errors when applied to materials with high dielectric
andmechanical losses.Sincelossesare
likely to be an
importantfactor
in plateswhosethicknesses
are approaching the grain size, we have developed a method of
independently computing mechanical losses and
k, based
on the KLM transducer model [ 121. The KLM model [ 131
conveniently represents the transducer as an electrical port
coupled to an acoustic transmission line
by a frequency
dependenttransformer as shown in Fig. 4. The nomenclature used here is that of [ 131 and precise definitions of
materials parameters are given in [ 141. The impedance Z ,
measured at the electrical port of the transducer is simply
I
z,= + jx, + -71 Z u ( a ) ,
4JWco
(3)
where W = 27rf is the frequency in rad/s, Z , ( a ) is the
radiation impedance at the center of the acoustic transmission line and C, is the bulk capacitance o f the resonator. The turns ratio of the transformer and the extra reactance, X , , in the primary are given by
4
=
( 1 / 2 M ) COS ( w d / 2 ~ )
X,
=
Z,M' sin ( w d / c )
Fig. 4 . KLM model for thickness mode resonator.
It is possible to show [ 121 that for small a d , the ratio of
the imaginary, X , to real, R , components o f Z:, is given
by
X/R
= k33/(WZ,,).
The mechanical losses of the material have been incorporated into the KLM model
via an effective attenuation
coefficient. a, which is computed by the method o f Bui
et al. [ 151. This approach has the advantage that it is independent of k , . For an air resonator. the radiation impedance at the center of the transmission line is given by
(5)
where /3 = w / c . Subtracting the effect of the bulk capacitance from themeasuredelectricalimpedance,
Z,,, near
the resonance gives
z;= z,
-
I /jWC0.
-
1)
(7)
where CY is in Np/m.
Thecoupling coefficient k, was determinedfrom ( 3 ) .
resulting in the expression:
jx;+ ( l / Q ' " Z , ( a )
where X ; and
sions
=
4' areconstantsdescribed
(k,)'X; and
(9)
by theexpres-
1 /4*= (k,)2(l / 4 r 2 ) . (10)
(4)
where c = ( ~ : ? / p ) " ~ is the longitudinal velocity associated with the thickness mode resonance, Z,, = pcA is the
characteristic acoustic impedance of the material, and M
(z(,/~)(I
- e-'"-./pc')/(l + epad-ia[')
-(./ad>(f/J;,
where f is the frequency and J;, is the antiresonant frequency. The slope of an X / R versusf/j;, plot is thus inversely proportional to the attenuation coefficient. An exampleofsucha
plot for PZT2 with anantiresonant
frequency o f 48 MHz is given in Fig. 5 . The calculated
value of a in this case is 3.3 dB/mm. Itis possibleto
convert a to the more commonly reported mechanical
Q
via the expression
X,
Z,(CU)=
VOL. 3 8 , NO. S . SEPTEMBER 19YI
(6)
C. Electrical Impedance and Insertion Losses
The design of efficient high frequency transducers requires careful matching of the electrical impedance of the
transducer to that of the driving and receiving circuitry.
Knowledge of the electrical impedance and insertion loss
are therefore importantto characterize overthe bandwidth
of the transducer. The KLM model, shown in Fig. 4, with
appropriate acoustical impedances, ZHfor the backing and
Z,,, for the medium (wateror air) into which the transducer
is radiating.was used tocomputetheelectricalimpedance, Z,, for various transducer configurations.
Computation of the pulse-echo response due to
an arbitrary transmitsignal is accomplished by treatingthe
KLM circuit as a two-port network and propagating the
input waveformtoandfromthewater
load. Details on
this approach are provided by Sherar er al. [ 161 and Desilets et al. [ 171.
FOSTER et
U/.:
CHARACTERIZATION OF PZTCERAMICS
449
f/f n
Fig. 5 . A plot of X / R vs f/J,for PZTz isinverselyproportionaltothe
(38
attenuationcoefficient of thematerialattheantiresonantfrequency
MHz). This slope was calculated with
in the region 0.99 < f/J, < 1.01
cw Source
4
Fig. 6. Apparatus for measurement
of transducer insertion loss
TABLE 1
AVF.R4GE FREF,A N D DIELECTRICPROPERTIES M E A S U R PFOR
D
,
PZT
PZTz
PZT,
1264
1156
*+ 50 (1700)
(1800)
46
1834 f. 61 (3300)
969 f I O
925 f 32
1296 i: 85
THE
THREE
PZT MATERIALS
0.42 f. 0.007
0.066
* 0.007
0.090 * 0.014
0.126
0.043
k 0.030
0.053 k 0.026
f. 0.122
*For eT, the manufacturers specifications are given in parentheses
The pulse-echoinsertionlossofthetransducerwas
studied (5-80 MHz). Table I shows the results of the dimeasured using the apparatus shown in Fig. 6 . A quasi- electricconstantmeasurements.Notethatnone
of the
continuouswave input is coupledthrough a directional coefficients reported here are as high as the manufacturers
coupler tothe transducer that transmits an ultrasound pulse claim. This is probablybecausethemanufacturersbase
toward the glass slide reflector. The returning ultrasound
theirdielectricconstants on low frequency ( l kHz) capulse is detected by the same transducer and the resulting pacitivemeasurements.Thedielectricconstant
of the
electrical pulse is amplified and measured across a
5 0 4 PZT, material was significantly higher than those of the
load. The received power is compared to the power avail- PZT, andPZT,materials.However.thecorresponding
able to a 504 load by coupling the input pulse to an open dielectriclosstangentofthePZT,material
was higher
circuit ensuring complete reflection. The pulse then passes than the other materials.
through anattenuatorand
is measuredacrossthe
504
load. The losses due to attenuation in the water and reB. MechanicalLosses
flectionfromthe
glass slideweresubtractedfromthe
Losses in each of the ceramics are plotted in Fig. 7. At
measurements.
5 MHz, the measuredvalue of 0.23 dB/mm for PZT,
agrees well with measurementsreported by Ih andLee
111. RESULTS
[ 181. Thefrequencydependent
mechanicallosses with
A . Dielectric Properties
reference to losses at 1 MHz were 1.32 X IO-’ dB/(mm
The dielectric constants of the three PZT materials did
* MHz’.”), 2.67 X lo-’ dB/(mm
.
and 1.05
not appear to vary significantly over the frequency range
X IO-’ dB/(mm * MHz’ h 3 ) for PZT,, PZT2 and PZT,
450
VOL. 3 8 . NO S. SEPTEMBER 1991
IEEE TRANSACTIONS ON ULTRASONICS,FERROELECTRICS.ANDFREQUEKCYCONTROL.
I
I
1
-~
i
a=1.32 x 1 0 2 d B /
(rnmMHz’’*)
100
10
10
1
100
Frequency (MHz)
(h)
(a)
(C)
Fig. 7. Frequency dependence of mechanical losses for (a) PZT,, ( h ) PZT,, and (c) PZT,. Error bars represent standard &via
tions in results from adjacent samples.
0.E
1
3
1
1
8
’
1
I
8
I
I
I
0.4
e
Y
0.2
slop = -2.02 X ~o%Hz-’
Slope=-1.24x7U3MHi’ j
Slope = -098 X 1U3MHi’
Intercept = 0.48
0
I
30
I
I
60
1
,
90 0
30
8
1
,
60
I
90 0
,
30
respectively. The mechanical losses of the PZT2 material
appears to have the lowest dependence on frequency.
At
50 MHz, the corresponding mechanical
(),,S
(8) for the
three materials are 28. 57, and 44.
C. Thickness Mode Coupling Coeficients
Frequency dependent plots of k, for each of the materials are given in Fig. 8 . As a benchmark, the accepted
value of k, for PZT5A in the diagnostic frequency range
( - 5 MHz) is 0.49 [19]. The k, for PZT, ceramic (0.47)
agreed well with this value while PZT, exhibited a lower
k , of 0.44. PZT,, which exhibited a higher dielectric constant. had thehighest k, at 5 MHz (0.52). Each of the
materials exhibit a reduction in k, as frequency increases.
A linear least squares fit to the data shown in Fig. 8 resulted in slopesof -2.02 X
MHz-’, - 1.24 X lo-’
MHz-’ and -0.98
X 10-3 MHz-’ for P Z T , , PZT?and
PZT,respectively.PZT,appearstohavethehighest
thickness model coupling coefficient up to about 80 MHz.
D. MiniatureTransducerCharacterization
PZT, ceramic was used tofabricatea l mm X 1 mm
miniature transducer with a center frequency of 45 MHz.
In this device, a 1 mm’ of 48-pm-thick PZT, was backed
with a conductive epoxy mixture (Ablebond 16-1 , Ablestick Laboratories. Gardena, CA) with a measured acoust-
/
90
60
Resonant Frequency (MHz)
(a)
(h)
Fig. 8 . Frequency dependence of k , for (a) PZT,, (b) PZT,. and
/
’
1y
(C)
( c ) PZT,.
ical impedance of 4.3 X lo6 Ray1 and tuned with a shunt
inductor. The measured properties of this sample of PZT,
were: 6’ = 961. tan 8; = .044, k , = .40 and cx = 9.7
dB/mm. A photograph of the backed transducer is given
in Fig. 9. Fig. 10 shows the electrical impedance of this
device at various stages in its fabrication. Fig. 10(a) demonstratestheimpedance
plot of the free airresonator
showing parallel and series resonances
at 41.5 MHz and
45.5 MHz respectively. Addingthebackingsharply
reduces the magnitude of the resonance as shown in Fig.
10(b) while tuning of the backed device with a shunt inductor of 80 nH gives the characteristic double peaked
impedancemagnitudeplot.Goodagreement
of experimental measurements with thosepredicted by the KLM
model using the measured material properties for PZT, is
achieved in all three cases. Note that the tuned transducer
is reasonably well matched to 50 fl over thefrequency
range 30 to 60 MHz.
A plot of theinsertionloss for the tuned transducer,
measured as described in section 11, is shown in Fig. 1 1 .
The minimum insertion loss is - 17.5 dB at 46 MHz. This
compares well with the theoretical prediction of - 15 dB
at 49MHz.Experimental andtheoretical -6 dBbandwidths are also in good agreement (50% and 47% respectively).Thus, thepulsecharacteristicsshould
be well
suited for medical applications. A final comparison of the
experimentally measured and theoretically predicted two
FOSTER er
U/.:
CHARACTERIZATIONOF
P T 1 CERAMICS
45 l
Fig. 9. Photograph of a 1 mm x I mm, 45-MHz transducer.Theactive
PZT, component is 48 pm thick, while the conductive epoxy backing
is
approximately 1.5 mm thick.
a
a
0
v)
c
a
-90
0
(b)
-.._...______
____......
50
100 0
50
100
(al
0
7
50
100
(C)
Frequency (MHz)
Fig. 10. Theoretical(solidline)andexperimentallymeasured(dottedline)impedancemagnitudeandphasefor
resonator, (b) backed transducer ( a s hhown i n Fig. 9). and (c) tuned and backed transducer.
(a) treeair
way pulse response is given in Figure 12. Here a single
cycle of 42 MHz RF was used as the excitation function.
Again, theory and experiment are in good agreement.
IV. CONCLUSION
20
40
50
Frequency (MHz)
30
60
Fig. I I . Theoretical (solid line) and experimentally meaaurcd (dotted line)
insertion loss for the transducer shown in Fig. 9 .
Thedevelopment of new endoscopicapplications of
high frequency ultrasound is critically dependent on the
availability of efficient broadbandtransducerswithdimensions on the order of a few millimeters or less. Althoughconventionalceramicmaterialshavebeenused
successfully at frequencies as high as 30 MHz, very little
attention has been given to their use at higher frequencies.
A number of groups have studied the properties of sputtered ceramics [20] and sol-gel processes [21] in the fre-
452
IEEETRANSACTIONS
ON ULTRASONICS.FERROELECTRICS.
0.25
0
Time ( p )
Fig. 12. Experimentallymeasured (a) andtheoretical ( h ) pulseshapefor
the transducer \hewn in Fig. 9.
A N D FREQUENCYCONTROL,
VOL. 38. NO. S. SEPTEMBER 1991
80 MHz. Additional losses and reductions in piezoelectric
coefficients at frequencies greaterthan 50 MHz may result
from interactions at the level of the domain boundaries as
described by Arlt [ 2 3 ] .
The results of the materials properties of the three test
ceramicsindicatedthatthesematerialswere
well suited
for use in high frequency transducers. Experimental and
theoreticalanalysis of a I mm square, 45 MHzdevice
fabricated with the PZT, ceramic showed the validity of
the material property measurements and demonstrated excellent insertion loss and bandwidth characteristics. The
minimuminsertion loss of - 17.5 dB at 46 MHz is a
marked improvement over polymer devices with similar
dimensionswhichtypicallyexhibitinsertionlossesapproaching -50 dB [ 161. The use of lumpedelement
matchingnetworkscouldreducethisfiguretoapproximately -30 dB but with some loss in bandwidth. Similar
improvements, using matching networks and quarter wave
face plates, could potentially improve the performance of
the ceramic transducers to the level of
- 8 to - 12 dB,
depending on the frequency.
Finally, PZT ceramics, in spite of their relatively large
grainsize(upto
10 pm) are well adapted to ultrasound
transducer applications at frequencies as high as 80 MHz.
Their high dielectric and coupling constants providea significant advantage over other materials in the fabrication
of miniature devices.
ACKNOWLEDGMENT
quency range from approximately 100 MHz to 1 GHz but
The authors wish to thank Henry Yoshida of Hewlett
it does not appear that these methods are well adapted for
Packard
Laboratories for his expert preparationof the PZT
use at frequencies of less than 100 MHz. In the present
samples.
paper, we have studied the properties of three commercially available ferroelectric ceramics from the PZT famREFERENCES
ily for application at frequencies as high as 80 MHz. The
dielectric properties of the three ceramic materials did not
M . D. Sherdr. B. G . Slarkoski.
and
F. S . Foster, “ A 100MHz
appear to vary significantly over the frequency range from
B-scan ultrasound backscatter microscope,” L l l r n r ~ o r ~/mrsgirzg,
.
vol.
1 I. pp. 95-105. 1989.
5-80 MHz with PZT, and PZTz having somewhat lower
C . J . Pavlin. M . D. Sherar. and F. S . Foster, “Subsurface imaging
dielectric constants (969 and 925) than PZT, (1296). Each
of the eye by ultrasound backscatter microscopy,” Ophtl2trlnlolo::~.
of the ceramics tested showed reductionin k, as frequency
vol. 97, pp. 422-250, 1990.
F. K . Forster, J . E. Olerud, G. R. Pomajevich, A . W . Holmes, and
increased.The mostsignificantreductionoccurred
in
S . R . Sharar,”High-frequencyultrasonicimaglngandbackscatter
PZTI, which exhibited a loss in k, of 2.02 X lo-’ per
attenuationtechniquesfordetermination
o f thermal in.jury t o the
MHz. None the less, our results indicate that k, of at least
skin,“ l E E E Lllrrcrson. Syrup. Proc... IEEECat. No. 86CH1375-4.
IEEE, New York. 1986. pp. 957-962.
0.3 and as highas 0.44 (PZT3) can be achievedat 80
T . Yano, H. Fukukita, S . Uneoand A . Fukomoto, “40 MHzultraMHz. The mechanism for the observed reduction in k, is
sound diagnostic system for: Dermatologic Examination.”
I n : lEEE
not clear at the present time.
It is possible that heating
Ultrusorr. Sw7p. Proc.. IEEECat.
No. 87CH2492-7..IEEE. New
York, 1987. pp. 8 7 5 4 7 8 .
andmechanicalagitation
in the lappingprocedure reM . J . Wiersema. C. R. Redly. N . T. Sanghvi. R. Hawea. L. Wiersulted in some depoling of the samples as was postulated
sema and C . Aust. “Twenty-tive MHz gastrointestinal ultrasonograby Onoe et al. [22] in theirearlystudies of electromephy.“ f E E E L’lrrmon. Sxmp. P r o c . . IEEE Cat.No.89CH2791-3,
IEEE, New York. 1989. pp. 845-848.
chanical coupling coefficients. In the present situation, we
F. E. Silverstein. R . W . Martin. M . B . Kirnmey. G . C. Jiranek. D.
do not believe this to be the case since the smallest
reW . Franklin and A . Proctor. “Experimental evaluation of an endoduction in k, occurred in PZT3, whichhasa low Curie
I n ~ . i r r oand i n w‘lv canine studies.” G U T scopicultrasoundprobe:
trornrrrolog!. vol. 96. pp. 1058-1062, 1989.
point (260°C) compared to PZT, and PZTz (350°C). It
N . G . Pandian. “Intravascular and intracardiac ultrasound imaging:
seems likely that the grain size and pore structure are some
An old concept. now ontheroadtoreality.“
Circulu/iorz. vol. 88,
how responsible for the reduction of electromechanicalefpp. 1091-1094, 1989.
ficiency in very thin plates. Mechanical losses showed a
P. G . Yock. D. T. Linker. B. A . Angelsen,, “Twodimensional intravascular ultrasound: Technical development and initial clinical exmarkedfrequency dependence.Thisincrease
with freperience,” J . Am. Soc. Ec,/zo, vol. 2. pp. 296-304. 1989.
quency may be due to losses at the grain boundaries which
J. M . Hodgson. S . P. Graham. A . D. Savakus, S . G . Dame, D. N.
become an appreciable fraction of a wavelength in size at
Stephens, D. Brand.H.Sheehan,andM.
J . Eberle,“Clinicalper-
FOSTER e!
(l/,:
453
CHARACTERIZATION OF PZT CERAMICS
cutaneous imaging of coronary anatomy using an over-the-wire ultrasound catheter system,” Inr. J . Curd. I m r r x i r l g . vol. 4 , pp. 186-193,
1989.
J. M.Tobias, J . A . Mallery. J . Gessert. J . Griftith. D. Mahon. M.
L . Mcleay.M.McRae.andW.
L. Hener)..
Bessen,M.Moriuchi,
”Intervascular ultrasound cross-section imaging before and after balloon angioplasty in virro,” Cirr,u/arion. vol. 80. pp. 873-882. 1989.
IREStandards on PiezoelectricCrystals. Proc. IRE. vol. 49. 1961.
pp.I161-1l69.
S. Foster, “Determinatmn of
D. H. Turnbull. M. D. Sherar. and F.
electromechanical coupling coefficients i n transducer materials with
high mechanical losses,” IEEE Ulrruson. S w ~ p .pp.
. 631-634. 1988.
D. A . Leedom. R . Krimholtz, and G . L. Matthaei.“Equivalentcircuits for transducers having arbitrary even or odd symmetry piczoelectricexcitation,” IEEE Truns. Sotlics U/trcrsott. vol. SU-18, pp.
128-141,1971.
D.A.Berlincourt.D.
R . Curran,and H . Jaffe.“Piezoelectricand
piezomagnetic materials and their function
in transducers,“ i n Physicd Acoustics, Vol. 1 . Part A , W.P.Mason,Ed.
New York:AcademicPress.1964,pp.169-270.
L. N. Bui, J . Shaw. and L. T . Zitelli. “Study of acoustic wave resonance in piezoelectric PVFl f i l m . “ IEEE Trcrrls. Sonic7 L’/rrrrsorz..
vnl. 24, pp. 331-336. 1977.
high
M. D. Sherar and F. S . Foster. “The design and fabricatlon of
frequencyPoly(viny1idenefluoride)
tranducers,” Ulrrcrsnrl. Imu,qing. vol. I I , pp. 75-94.1989.
C . S . Desilets, J . D. Fraser and G . S . Kino, ”The design of efficient
broadband piezoelectric transducers.’’ /Et% 7 n r m Sonic,.\ Ulrrcr.sorz..
vol. SU-25. pp. 115-125 1978.
J . H . Ih and B. H. Lee. “Attenuation coefficient of a piezoelectric
E/ec,rro,z L e u . , vol. 2 2 ,
resonatoroperating in athicknessmode.”
pp.357-358.1986.
0 . E. Mattiat, Ed., Ulrrusot~icTrrrnrrlurer Muteria/s. KewYork:
Plenum,1971.
K . Screenivas and M. Sayer, “Characterization of Pb(Zr, Ti)O, thin
films deposited from multi-element metal targets,” Appl. Ph?.r., vol.
64,pp. 1484-1493, 1989.
G. Yi. Z . Wu. K. Screenivas and M. Sayer, “Ultrasonic experiments
with PZT thin films fabricated by sol gel processing.” Necrron. h r . ,
vol. 25, pp. 307-308 1989.
M .Onoe. H. F. Tiersten, and A. H. Meitzler. “Shift
i n the location
of resonant frequencies caused by large electromechanical coupling
in thlckness mode resonators.” J . AUJU.S~.
Soc. A m . , vol. 35, pp. 3642.1963.
G . Arlt.“Domal in contributions to piezoelectricity in ceramics,” in
Proc 1990. IEEE- Ultrason. S y p . pp. 733-741.
F. Stuart Foster was born in Montrcal. PQ. Canada on J u l y 29. 1951. Hereceivedthe
B.4.Sc.
dcgree in engineering ph}slc\ from the UniLersit)
of Britlsh Columbia. Vancouver, Canada. in 1974.
and the M.Sc. and Ph.D. degree5 in medical biophysic\ from the Univcr\ity o f Toronto. Toronto.
ON. Canada. i n 1977and 1980. respectively.
From 1980 to 199 I , Dr. Fohtcr was a senior cci~n Toentist with theOntarioCancerInstitute
ronto. Canada. He is presently a Senior Scientist
with Sunnybrook Health Science Centre. a Senior
Research Scholar of the Kational Cancer Institute
of Canada, and an A \ sociate Professor of Medical Biophysics at the
Unlvcrhity of Toronto. He
has been involved with the development of conical and annular array transducers. ultrasound backmitter microscopy, tissue characterization and more
recently. two-dimensional array technology and intravascular imaging. He
has twice won the Ultrasound i n Medicine and Biology Prize.
Dr. Foster is on the editorial board o f Ulfrtrsorfic,I n r r r g l f f g .
Linda K. K t a n was born in Montreal, PQ, Canada o n July 28. 1964. Shc received the B.%. degree i n phys~csin 1988. from McMaster Univcrsity. Hamilton, ON, Canada. Since January 1991
she has been enrolled
a\ a Master\ \tudent in thc
Dcpartnient o f Medical Biophyhics at the University 0 1 Toronto, Toronto, ON. Canada.
Between1988and
1991. Ms. Ryanwasemployedas an intermediatetechnician at the Ontario Cancer Institute, Toronto. where she was involved in thedevelopment
of transducersfor
intravascular imaging