A Thick-Film Acoustic Wave Sensor
N. M. White
(c) University of Southampton 1994
Contents
- Introduction
- The acoustic wave sensor
- Circuit design
- Experimental results and discussion
- Conclusion
- References
Acoustic wave sensors offer advantages of high sensitivity, good linearity,
low hysteresis and wide versatility. Their main use is in the higher cost,
precision areas. Low-cost, high volume applications are generally
disregarded. Thick-film technology is robust, compact and inexpensive.
Although commercial thick-film sensors are available, the number of
applications has been limited by the lack of specific sensor pastes within
the market place. A desirable exercise is to try to merge the advantages of
acoustic wave sensors together with thick-film technology to provide a
powerful and economic strategy for future sensor development.
Brignell et al. [1] noted that there were three main areas to which
thick-film technology could contribute to future sensor development:
Firstly, the provision of associated electronic circuitry which can be
mounted within the sensor housing. Secondly, the technology allows the
creation of support structures, such as electrode patterns, upon which
sensing materials can be deposited. The final area concerns the use of the
thick-film material itself as the primary sensing element.
This paper describes a sensor utilising all of the above aspects. The
primary sensor material is a screen printable thick-film piezoelectric
paste, similar to that reported by Baudry in 1987 [2] and Morten et
al. in
1991 [3]. It comprises a lead zirconate titanate (PZT) powder, a lead
borosilicate glass frit and an organic carrier which serves to give the
paste the required viscosity for screen printing. The processing can be
carried out using conventional thick-film equipment.
Fundamental material studies on the thick-film piezoceramic have revealed
that it possesses similar properties to the bulk PZT, the exception being a
reduction in the value of the relative permittivity. This is thought to
arise as a result of the morphology of the film being different to that of
the bulk.
Figure 1 shows the configuration of the acoustic wave delay line sensor.
The piezoelectric film and the silver/palladium electrodes were deposited
onto a 96% alumina substrate (thickness 625
m) using conventional
thick-film processing techniques. Minimum line and space dimensions of
around 200m are achievable with this technology. Hence the maximum
frequency is limited to 20MHz, assuming that the velocity of the acoustic
wave is limited to about 8
10
m/s. An accurate figure for this velocity is
not yet available as we have yet to establish firmly the exact nature of
the wave. Also, as we are dealing with a composite structure of alumina and
the new PZT thick-film, such figures are not readily available in the
literature.
Figure 1: Thick-film acoustic wave sensor.
The principle of operation of the device can be understood by considering an
alternating voltage applied to the input comb structure. The material
between the fingers of the interdigitated electrode pattern distorts because
of the piezoelectric effect. This periodic deformation gives rise to an
acoustic wave propagating both towards the second electrode pattern, and in
the other direction where it can be damped at the edge of the substrate to
prevent it interfering with the preferred wave. By virtue of the reverse
piezoelectric effect, the acoustic wave can be detected at the other end of
the substrate. Typical dimensions of the substrate are 50mm long by 10mm
wide. Experiments have shown that there is significant acoustic attenuation
within the piezoelectric thick-film. One consequence of this is that the
effect of unwanted reflections from the edge of the substrate is eliminated.
Figure 2 shows how the sensor is used as a delay line oscillator. Two
acoustic wave devices are depicted; an active device which responds to the
measurand, and a passive device for temperature compensation. For the
purpose of testing, we simulated the passive device with a HP 8656 signal
generator. The delay line oscillator is excited at its resonant frequency
and the output of the device is fed into a high-gain amplifier. A phase
shifting network closes the loop. The latter is adjustable and allows
resonance to occur for slightly different sample dimensions.
Figure 2: Delay line oscillator circuit.
The loop amplifier comprises three stages: the input, gain and output stage.
The input stage is essentially a shunt-voltage feedback amplifier acting as
a current-to-voltage converter. It was designed to have a bandwidth of 20MHz,
which is greater than the resonant frequency of a typical sensor (8MHz).
The measured value of the capacitance of the sensor was 64pF, and so
in order to reduce signal degradation the input impedance was designed to be
low, around 4.5
at 8MHz. The gain stage is based on a cascode
arrangement giving a high gain-bandwidth product. Analysis of the circuit
revealed a bandwidth of 60MHz and a mid-band gain of 250. The output stage
drives the acoustic wave sensor. The circuit used is a two-stage,
DC-coupled common emitter amplifier with series-voltage feedback. An emitter
follower circuit could not be used as the output impedance tends to be
inductive in nature and because this is driving the sensor directly (which
is capacitive in nature) there will be a tendency to oscillate. The overall
design of the amplifier is given in Figure 3.
Figure 3: Schematic of loop amplifier.
The main objective is to measure the frequency shift which occurs in
response to the measurand. A single sideband suppressed carrier modulator
was adopted to mix the two signals from the active and passive devices. The
two signals differ in frequency by a few kHz. Following the mixer, the
signal is filtered with a sixth-order bandpass filter and finally fed in a
phase locked loop (PLL) to generate a DC output voltage. The complete
circuit together with the sensor has been designed to fit onto a 50mm by
50mm standard alumina tile. The circuit layout was achieved using a package
called ICEBOX developed within the department. The following results were
achieved using a thick-film sensor with the electronic circuitry built on a
prototype board. The full hybrid version is currently being fabricated.
Initial experiments were undertaken to observe the frequency response of the
sensor alone. A number of sample devices having different electrode spacings
and number of fingers were observed. The results indicated that the
thick-film devices could generally be modelled using a similar approach to
that used with surface acoustic wave (SAW) sensors. For example, increasing
the number of fingers gives a sharper Q-factor and the shape of the
frequency response is a sinc function, as predicted using a delta-function
model [4]. However, one noticeable point was the occurrence of a double-peak
exhibited by some of the devices. This is believed to arise because of
reflections of the acoustic wave between adjacent fingers in the
interdigitated transducer pattern. This effect was eliminated using a
double-fingered electrode pattern. The reflected wave is then transposed to
a frequency which is twice that of the resonant frequency of the sensor
(i.e. greater than 16MHz). The second peak can then easily be removed by
filtering.
A test jig was constructed to allow the samples to be used as
freely-supported beams. A mass was applied in the centre of the beam and the
associated surface strain increased the acoustic path length, and hence
reduced the resonant frequency of the sensor. Figure 4 shows how the
frequency changes in response to an applied mass. The graph shows a linear
characteristic, which is to be expected for the relatively small force
applied to the substrate.
Figure 4: Frequency change against applied mass.
The device exhibits a suitable response for use in physical measurements,
particularly in areas such as torque, pressure and low-magnitude force
measurement. Its potential use in other areas such as chemical sensors has
yet to be established. Indeed, this programme of work is still in its
infancy as there are many fundamental issues still to be addressed. In
particular, optimising the electrode geometries and investigating the exact
nature of the acoustic wave. Many possibilities exist including; surface
waves, bulk waves, Lamb waves, Love waves, deep bulk acoustic waves (DBAW)
and Stonely waves, to name but a few. Further improvements to the design of
the loop amplifier are currently being made to allow a wider range of
devices with different resonant frequencies to be accommodated. The present
amplifier is optimised for a frequency of 8MHz.
The effect of temperature upon the acoustic wave sensors has been observed.
Not too surprisingly they have a large temperature coefficient. As the
temperature increases, the electrode separation also increases thereby
causing a shift in resonant frequency. A second effect is that the size of
the acoustic gap also changes with temperature. The overall temperature
coefficient is around -100ppm/
C. For a resonant frequency of 8MHz this
equates to a frequency shift of about 800 Hz/C.
Clearly, the thick-film acoustic wave device could be used directly as a
temperature sensor. For other measurands such as force, pressure,
acceleration etc., the response to temperature is undesirable. The
principles of intelligent sensors [5] may therefore need to be adopted.
Techniques such as compensation by design symmetry and monitored
compensation (or sensor-within-a-sensor) can be used to minimise the problem.
The area of thick-film sensors is expanding considerably. The major
attractions of these devices are their low-cost, robust and miniaturised
nature. We have used the technology to produce an acoustic wave sensor, a
device which has traditionally only been found in higher cost, precision
applications. The versatility of such a device has the potential for
expanding the technology into new sensor fields. It is predicted that
low-cost acoustic wave sensors will find use in many application areas,
particularly where disposable devices are required.
- Brignell, J. E., White, N. M. and Cranny, A. W. J.:`Sensor applications
of thick-film technology' , IEE Proc. Pt. I, 1988, 135, No. 4, pp
77-84.
- Baudry, H.: `Screen printing piezoelectric devices', Proc. 6th
Europ. Microel. Conf., Bournemouth, 1987, pp 456-463.
- Morten, B., De Cicco, G., Gandolfi, A., and Tonelli, C.:`Advances in
ferroelectric thick-film materials and sensors', Proc. 8th Europ. Microel.
Conf., Rotterdam, 1991, pp 392-399.
- Tancrell, R. H. and Holland, M. G.:`Acoustic surface wave filters'
Proc IEEE, 1971, 59, pp 393-409.
- Brignell, J. E. and White, N. M.: `Intelligent sensor
systems', IOP
Publishing, Bristol, 1994, ISBN 07503 0297 6.