Project topics and materials
Get Complete Project Material File(s) Now! »
Structure and characteristics of high frequency piezoelectric transducers
There are essentially two kinds of structure configuration (Figure 2.2.2) to be studied for piezoelectric transducers of frequency superior to 1 OOMHz (half wavelength layer thickness in the order of severa! to severa! tenths of micrometers ).
One is the backing structure (Figure 2.2.2 (a)) and the other is the buffered structure (Figure 2.2.2 (b)). Both have a multi-layered structure. Their characteristics and performances in terms of the electrical impedance, frequency transfer function and impulse response, insertion loss, can be simulated by Mason or KLM modelling.
Transducers made of commonly used piezoelectric materials such as ZnO, LiNb03 and PZT-4 are considered here and their performances will be compared and analyzed.
The material constants employed in simulation are resumed in Table A.2.1.3 in Annex 2.1.
Before simulation, sorne well-known criteria and basic phenomena in the design of piezoelectric transducers need to be recalled. An optimized transducer is desired to have a large bandwidth and consequently short impulsion duration in transmission ultrasonic bearn because it determines the axial resolution of imaging; a high sensibility (small insertion loss) to have high signal-to-noise ratio in ultrasonic testing and imaging; matched input electrical impedance and matched transmission mechanical impedance that condition the electromechanical conversion e:fficiency of the transducer. Simultaneous improvement of all these characteristics is usually di:fficult to be fulfilled and trade-offs must be taken, among which an optimization procedure is necessarily employed.
In ali the following simulation, if not specifically indicated, the internai impedance of the excitation source is considered to be 50 Q and the input impedance of the reception circuit is supposed to be a real resistance of 1 MQ. The frequency functions are calculated in a range of frequency from zero to two times of resonant frequency and their impulsion response corresponds to that with limited band source excitation. All the transducers have a section area of A= wide*length = 40 Jlm*2 mm
and the normalized frequency f 0 is referred to as 100 MHz.
It can be stated that the baking or front charge of a piezo-transducer as resonator will enlarge its bandwidth and reduce its resonant frequency (this can be characterized by the Q factor that is defined as the ratio of resonant frequency to the bandwidth). If the material dissipation is not considered, the real part of the electrical impedance corresponds to the energy radiation in the backing and transmission media.
Conceming the material ‘s piezoelectric characteristics, the PZT-4 bas the most high electromechanical coupling coefficient and dielectric constant. This permits the PZT-4 transducer to have an electrical impedance near 50 n (about 43 !l) at its resonant frequency (100 MHz) with the given transducer dimension. The insertion loss (sensibility) at the resonant frequency is about 55 dB for LiNb03, 60 dB for PZT-4 and 75 dB for ZnO respectively.
Figure 2.2.4 is the simulation results for silicon buffered (0.5 mm in length) and water front charged piezoelectric transducers (without backing) of Znü, LiNb03 and PZT-4. In such transducer configuration, comb like frequency functions are observed due to the high thickness of the silicon buffer forming the standing waves in the buffer.
In the time impulsion response, multiple reflections between the transducer and the silicon-water interface occur. As can be observed, in the high order reflections, there is the dispersion of impulsions that is originated from the double-reflections by the two faces of the transducer layer at the one si de of the buffer.
The influence of other parameters such as electrode, bonding layer, on the transducer performances, usually neglected with low frequency transducer, become important at high frequencies. Figure 2.2.5 gives the simulation by taking into account these layers with a silicon baking transducer of the structure: silicon (substrate) /gold-electrode (0.5 J..Lm) /indium-bonding (1.5 J..Lm) /piezo-layer/gold-electrode (0.5 J.tm) /water. Compared the results (solid lines) to those obtained without (dash-point lines) the integration of these layers: the main difference is a reduction in their resonant frequency and a slight increment in their electrical impedance. It should be noted that for high frequency transducers (> 100 MHz), the material dissipation factor needs usually to be taken into consideration.
Electrical impedance and electrical matching
For a piezoelectric transducer, its sensibility or the capacity to transform electrical energy to acoustic one (or inverse! y) depends one part on the electromechanical coup ling coefficient of the piezoelectric material used, and another part on how the electrical impedance is matched between the transducer and the externat driving source (generator or pulser).
The electrical impedance of an ultrasonic transducer consists of a capacitance – due to the two metallic electrode layers ( clamped capacity) in series (KLM) or ]t»Co in parallel (Mason) with another equivalent complex impedance coupled from the acoustic load. The energy dissipation on the real part of this latter impedance corresponds to the acoustic energy transmission into the radiation medium and the beat dissipation in the material itself if it is considered. For an ideal electrical matching, the internai impedance of the generator should equal to the transducer input impedance. But in practice this is not feasible for a source generator to have its internai impedance adapted identically to a complex impedance of charged transducer.
Fig. 2.2.6 and Fig. 2.2. 7 are respectively the simulation results obtained for three transducers of ZnO, LiNb03 and PZT-4 at fo = 100 MHz with and without such inductance impedance compensation. Compared Figures 2.2.6 with 2.2.7, it can be observed that the real part of the electrical impedance remains strictly unchanged. The transducer electrical impedance becomes real near the resonance. For ZnO transducer, the return insertion loss is significantly reduced near the resonant frequency which means an increment in the sensitivity, but with also a reduction in its bandwidth leading the impulsion response less damped. For LiNb03 transducer, the return insertion loss is improved and the bandwidth also enlarged, but few changes in damping. For PZT-4 one, there is no gain in the insertion loss with slight improvement in the bandwidth, keeping the best impulsion response damping. So the conclusion is that for sensitivity improvement the inductance compensation has the most significant effect on the ZnO transducer which has worst performance in term of the sensitivity.
Another advantage of inductance compensation is the suppression of high order harmonie generations.
Figure 2.2.6: Electrical impedance, frequency transfer function and impulsion response for silicon backing and water front charged piezoelectric transducers of Znü, LiNb03 and PZT-4 at Jo = 100 MHz with the presence of electrodes (0.5 Jlm gold layers)and bonding layer (1.5 Jlm indium layer) and with the inductance compensation in series of L0 = 1 /(21(0 ) 2 C0 .
In the simulation results given in Figs. 2.2.6 and 2.2.7, the frequency functions are calculated up to four times of the free resonant frequency 4f0 instead of 2fo. It shows that there is an important third order harmonie resonance (Fig. 2.2. 7) in the frequency responses, (much higher order resonances exist if we extend the frequency rang), and these peaks almost disappear for the compensated transducers (Fig. 2.2.6).
In the Fourier inverse impulsion time responses, corresponding to those obtained with an exciting source of at least 4fo bandwidth, high frequency component will be superposed in the case without the inductance insertion (Fig. 2.2. 7) .
Table of contents :
Abstract
Résumé
Acknowledgments
General introduction of the thesis
Chapter 1 : Introduction
1.1 Ultrasonic imaging and its development
1.1.1 Development of ultrasonic phased arrays
1.1.2 Ultrasonic imaging
1.2 High-frequency ultrasonic imaging
1.2.1 Imaging resolution
1.2.2 Acoustic microscopy
1.2.3 Development ofhigh-frequency ultrasonic phased arrays
1.3 Theory of ultrasonic phased array
1.3 .1 Princip le of ultrasonic phased array
1.3.2 Sorne particularities ofhigh frequency arrays
1.4 Different techniques for the fabrication ofultrasonic arrays
1.4.1 Conventional arrays fabrication technologies
1.4.2 High-frequency arrays based on MEMS technologies
1.4.3 cMUTpMUTandmMUT
1.4.4 Laser technique
1.5 Summary and outline of the thesis
Chapter 2: Acoustic Waves in Elastic Media and Fundamentals of Piezoelectric Ultrasonic Transducers
2.1 Waves propagation in fluid, solid and piezoelectric materials
2.1.1 Waves equation in fluids
2.1.2 Waves equation in solid materials
2.1.3 Waves equation in piezoelectric materials
2.1.4 Numerical solution of Christoffel equation for analysis of material elasticity and piezoelectricity
2.2 Modelling and characteristics of piezoelectric transducers
2.2.1 KLM mode1
2.2.2 Structure and characteristics of high frequency piezoelectric transducers
2.2.3 Electrical impedance and electrical matching
2.2.4 Acoustic impedance and mechanical matching
2.3 Summary
Chapter 3: Finite Element Modeling of Piezoelectric Ultrasonic Array Transducers
3.1 Finite element formalism using COMSOL
3.1.1 PDE and COMSOL formalism
3.1.2 Sources of excitation
3.1.3 Boundary conditions
3 .1.4 Simulation examples
3.2 Simulation ofbuffered transducer arrays
3.2.1 2-D simulation in x-y plane (simulation of array transducer with a buffer layer)
3.2.2 2-D simulation in z-y plane (simulation of single element transducer with a focal lens)
3.3 Summary
Chapter 4: LiNb03-based array fabrication and characterization
4.1 Simulation ofbacking transducer arrays
4.1.1 Mode! design
4.1.2 Bearn behaviour
4.2 Investigation of etching LiNb03 36°N-cut arrays
4.2.1 Focusing Ion Bearn (FIB)
4.2.2 Inductively Coupled Plasma (ICP) Reactive Ion Etching (RIE)
4.3 Fabrication ofLiNb03 36°N-cut transducer array on silicon
4.3.1 Process flow
4.3.2 Fabricated array transducers
4.3.3 Discussion offabricated patterns
4.4 Characterization ofLiNb03 36°N-cut transducer array
4.4.1 Measurement method
4.4.2 Measured results
4.5 Summary
Chapter 5: Znü-based array fabrication and characterization
5.1 Fabrication of ZnO-based transducer array using wet etching method
5 .1.1 Geometry design
5.1.2 Fabrication ofthick-film ZnO transducer arrays
5.1.3 Discussion offabricated patterns
5.1.4 Characterization ofthick-film ZnO transducer arrays
5.1.5 Summary
5.2 A novel method for fabrication of piezoelectric transducer arrays on Si substrate
5.2.1 Fabrication ofthick-film ZnO transducer arrays
5.2.2 Discussion offabricated patterns
5.2.3 Characterization ofthick-film ZnO transducer arrays
5.2.4 Improvement ofsputtering process with rotation 90°
5.2.5 Discussion offabricated patterns using rotation method
5.2.6 Summary
Conclusions and prospects
Annex 1: Calculation of phase delay
Annex 2.1: Parameters of the used materials
Annex 2.2: Material tensor transformations
Annex 2.3: Multiple quarter wavelength mechanical matching
References
