Characterisation of a Wireless Passive Surface Acoustic Wave Strain Sensor Rory Stoney

Department of Mechanical and Manufacturing Engineering, Trinity College Dublin

Abstract Wirelessly interrogated surface acoustic wave (SAW) sensors can operate passively and in harsh conditions reliably. The interrogation process of the signal measurement chain provides the power to excite the sensor and hence establish the sensors response in a passive manner. Development of wireless passive SAW sensors allows investigation of their potential to quantify physical measurands such as strain in aggressive environments. Wireless SAW strain measurement is shown to be advantageous over wired strain measurement alternatives where instrumentation at the sensor site can be prohibitive to strain measurement technology integration in certain application areas. In this paper, a one port SAW resonator (SAWR) strain sensor is presented and is shown to exhibit highly sensitive and linear responses to applied strain. The principle operation of the Trinity SAW is outlined and the processes involving the instrumentation of the packaged devices are discussed. Wireless strain measurement is demonstrated using a packaged SAW device that is calibrated using the standardised strain testing system developed in-house. The system facilitates a standardised and repeatable method for testing. This allows for both individual sensor performance and detailed comparative analysis between test sensors. A range of strain level performance assessments are presented and cross sensitivity at elevated temperatures is demonstrated and discussed. The performance of the SAW technology is shown to have equivalent sensitivity performance when compared to an industry standard strain gauge device. The work presented in this paper demonstrates the potential for SAW strain sensors to be integrated in real engineering applications such as process monitoring and tool condition monitoring in the future.

Introduction Traditionally surface acoustic wave (SAW) sensors have been associated with signal processing applications. In recent years SAW sensors have been developed to provide passive and wireless measurement of physical measurands such as strain, temperature and pressure in many engineering applications [1, 2]. SAW devices are highly attractive due to their high performance, small size and good reproducibility [1, 3]. SAW sensors are also easy to fabricate, small, inexpensive and respond to physical measurands with high dynamic measurement ranges [4]. SAW sensors have been shown to operate in very harsh environments [4], such as high voltage power grid switch gear systems [5], disc brakes [6], tyres [7], rotating machinery [8], electric motors [2] and inside jet engines [9]. SAW based strain sensors are demonstrated as a potential detection method for broken or loose fittings in aircraft wing structures [10]. Recently wireless SAW sensors and interrogation systems have become more readily available [11-13]. As a result SAW sensors have now been integrated into a variety of physical sensing applications. A SAW sensor has been successfully used to measure blood pressure using a strain sensitive structure inside a pressure measurement device [3]. Torque measurement is a favoured application for SAW strain sensors applied to transmission axles [14-16]. SAW strain based sensors have been used extensively by Transense [17] in automotive applications such as a contactless torque measurement device developed for use in automotive drivetrain systems and power steering applications [18] and an in-car tyre pressure monitoring system [7, 19]. Senseor currently offer a range of SAW sensors for strain temperature and pressure measurement [11]. While SAW devices have been shown to operate successfully as physical sensors their operation tends to be very application specific. Commercial applications are emerging but the performance of the SAW sensors and the added complexity of their use in strain measurement applications has been company specific know how.

Principle of passive wireless SAW strain sensor application Sensor operating principle The SAW sensor described in this paper is a one port SAW resonator (SAWR). The sensor is fabricated on AT-X quartz due to its constant strain sensitivity performance at elevated temperatures [20]. In a piezoelectric material, such as AT-X quartz, there is an electric field

associated with a propagating SAW on the quartz surface. Metal electrodes specifically located on the surface can detect this electric field [21]. The electrode setup, known as an interdigital transducer (IDT) as shown in Figure 1(a), when manufactured on the surface of a polished piezoelectric substrate, can generate, detect and influence the propagating SAW waves [1, 2]. The IDT therefore electrically excites a wave which is contained by the two reflective grating strips. Figure 1(a) shows the layout of the deposited aluminum structures on the SAWR surface. The magnitude of the SAWR response is maximized when an RF voltage is applied to the IDT at a frequency ƒc which is related to the surface wave velocity vs and the electrode pitch p by equation 1 [3, 6]: ƒc = vs/2p

(1)

Figure 1(b) demonstrates that applying strain biases will deform the surface features, thus changing the electrode pitch. The substrate material is exposed to stresses as a result of the applied strain changing the surface acoustic wave velocities [22]. The result is a SAWR sensors resonant frequency that changes linearly to the applied strain.

Figure 1: SAW structure schematic and sensor deformation description

Passive wireless operating principle Wireless SAWR interrogation is carried out using an off the shelf wireless interrogator. A pulsed interrogation method [23-25] is used to wirelessly monitor the resonant response of the SAW device. The important advantage is the ability of the sensor to store energy. The sensor can then operate in isolation without any active part, i.e. without any power supply or oscillators [25]. A pulsed interrogation procedure is initiated with a RF burst from the interrogator at a specific frequency and over a given time period. The energy is transmitted via induction through the antenna connected to the sensor therefore exciting the IDT on the quartz surface (shown in figure 2) and will cause the sensor to “store” energy if the excitation is at the resonant frequency. Subsequent removal of the RF burst, where the interrogator goes into “listening-mode” causes the SAW generated to be reconverted back into an RF response signal when the propagating decaying waves contact the IDT. A typical excitation-response profile is shown in Figure 2.

Figure 2: A Typical excitation-response profile from a SAW resonator [24]

Analysis of the decaying oscillations in the RF signal enables the interrogation device to evaluate the resonant frequency. The decaying oscillations will be strongest at the resonant frequency of the sensor and therefore the signal processing in the interrogation device can be programmed to find the resonant frequency. In this manner the sensor resonant frequency can be measured wirelessly and passively.

SAW Instrumentation Sensor Package Description The SAWR sensors are located on a diced quartz wafer and are 5mm x 5mm x 0.3 mm in dimension. The general performance of the SAW devices is summarised in figure 3(a). The devices are first packaged using off the shelf packages [26]. The package construction is a layered structure consisting of a kovar base material (ASTM-F15) and alumina/glass sidewalls.

(a)

(b)

(c)

Figure 3: Packaged SAW device prior to final sealing

The SAW device is adhered to the inside of the package base using an off the shelf strain gauge epoxy adhesive. The SAW electrical connections are applied using an ultrasonic wire bonding machine to connect the bond pads to the inside of the package I/O terminals. The package is then hermetically sealed with a kovar lid and tested using a network analyser to investigate the sensor S11 response. A PCB is currently used with an SMA connector to connect the network analyser to the sensor. The I/O pins are soldered to the PCB. Figure 3(c) shows the finally instrumented setup with the PCB and SMA connector to facilitate the interrogation of the sensor. The packaged SAWR is then adhered to the final strain substrate with the epoxy adhesive. The manufacturing process chain to instrument a SAWR sensor for strain measurement is carried out in house independent of any external collaboration. A successful sensor manufacture results in a S11 plot as shown in figure 4(a) on the network analyser and has a distinct shift in the peak response when strain is applied to the substrate the sensor is mounted on. The required response parameters are summarised in Figure 4(b)

(a)

(b)

Figure 4: Sensor S11 response and required response characteristics for strain measurement

Packaging Considerations The SAW die are very sensitive to contamination and therefore must be handled with extreme care during the assembly of the packaging. Adhesive quantities need to be monitored such that sufficient adhesive is used without contaminating the surface of the SAW die. Clamping is required to maintain a uniform and sufficiently thin bondline between the kovar and quartz substrate. Cure cycles are carried out as per the adhesive manufacturer’s specifications. Due to the thermal coefficient mismatch between the kovar, quartz

and steel there is a residual stress

generated by the adhesive bond and this creates a frequency offset in the SAW response. During the testing this offset, while present, did not shift the sensor outside of the 433MHz ISM band. While the packaging is demonstrated to work there are still improvements needed to standardise the instrumentation procedure and optimise the sensor’s performance.

SAW Performance Evaluation

Sensor Characterisation System The packaged sealed sensors are setup on the dedicated calibration system. A strain plate has been designed with a strain optimised gauge area used for the sensor characterisation. Vishay half bridge strain gauges with one gauge active in the strain field are used as a reference technology for performance validation. The strain gauges are instrumented and calibrated with an NI9237 data acquisition module. The fully instrumented gauge area is shown in figure 5 with two packaged SAW sensors and two foil strain gauges.

Figure 5: Instrumented strain calibration beam with SAW and strain gauge sensors

The strain plate is setup in the dedicated calibration system to implement the testing procedures using custom software. Figure 6 shows the system and its components. The actuator, SAW interrogation unit, strain gauge DAQ and temperature controller have been integrated into the rig control software. Strain levels, calibration cycles, test temperature and sensor responses are all set and acquired in a standardised automated manner.

Figure 6: dedicated calibration System for strain response characterisation

Sensor Performance The sensor range is assessed based on a direct comparison to the Vishay strain gauge instrumented as per the manufacturer’s recommendations. The sensitivities at very low and very high strains are investigated and shown in figure 7. The SAW sensor is shown to have a higher signal to noise ratio than the strain gauge at strain steps as low as 2με. The SAW response shows good correlation to the strain gauge at the very high strains where the maximum strain level was ~950με

Figure 7: SAW sensitivity assessment

A calibration data set is shown in figure 8. The test implements a cycle of compression and tension loading conditions of ± 400 με. The test is then repeated at steady state elevated temperatures and the parameters summarised in table 1. The sensor response is highly linear with very high linear correlation coefficients. Hysteresis and sensitivities have been shown to be robust and repeatable. The sensor was further cycled for ~13,500 calibration runs, at room temperature, to investigate the long term performance. Negligible decreases in the sensitivity were observed and the full scale frequency response and hysteresis levels remained constant. Temp 20˚C 30˚C 40˚C 50˚C 60 ˚C 70˚C 80˚C

Ssen -477.94 ±12.45 -478.03± 13.68 -479.75±16.74 -479.04±15.68 -477.46±16.84 -477.11±12.12 -474.87±23.8

R² 0.99996 0.99995 0.99993 0.9999 0.99982 0.99976 0.99969

Table 1: Calibration Results [27]

FS Hysteresis % 0.64 0.65 0.608 0.63 0.75 0.95 1.072

Figure 8: Calibration data set with a zoomed profile demonstrating the cross sensitivity to temperature

SAW Evaluation This paper has demonstrated the SAW technology operation and has explained how the strain measurement process can be implemented wirelessly and passively. However, as with conventional strain gauges, the final performance of the installed gauge is heavily dependent on the methods adopted during final instrumentation. It is therefore very important to demonstrate the performance of the finally instrumented sensor. The manufacturing process required to instrument a SAW device for strain measurement is shown to work and to provide repeatable results. This process is critical to the sensor performance and is significantly more difficult to implement when compared to a standard strain gauge. Given the sensor is a stiff section of quartz the surface and bonding conditions are critical for successful strain transfer to the sensor. The Sensor performance is shown to exhibit impressive and competitive sensitivities at very high (950με) and very low strains (2με). The cross sensitivity to temperature is shown in Figure 8 and can be accounted for. Future work involving two sensors and implementing a

differential measurement response variable to applied strain will reduce the cross sensitivity to temperature and reduce RF interference. Conclusion The wireless passive operation of the SAW strain technology is a key performance advantage for future applications of the technology. Given the work done to date, the SAW strain sensor manufacturing process is in a position to be used in an application specific area where real strains can be measured and used as a monitoring tool directly in real time. Further development in the understanding of the sensor bonding and the strain transfer to the sensor is required as well as more extensive long term testing of the sensor performance. References

[1]

[2]

[3]

[4] [5]

[6]

[7]

[8]

Bulst W E, Fischerauer G, and Reindl L, "State of the art in wireless sensing with surface acoustic waves," in Industrial Electronics Society, 1998. IECON '98. Proceedings of the 24th Annual Conference of the IEEE, 1998, pp. 2391-2396 vol.4. Scholl G, Schmidt F, Ostertag T, Reindl L, Scherr H, and Wolff U, "Wireless passive SAW sensor systems for industrial and domestic applications," in Frequency Control Symposium, 1998. Proceedings of the 1998 IEEE International, 1998, pp. 595-601. Ye X, Fang L, Liang B, Wang Q, Wang X, He L, Bei W, and Ko W H. Studies of a high-sensitive surface acoustic wave sensor for passive wireless blood pressure measurement. Sensors and Actuators A: Physical, vol. 169, pp. 74-82, 2011. Wolff U, Dickert F L, Fischerauer G, Greibl W, and Ruppel C C W. SAW sensors for harsh environments. Sensors Journal, IEEE, vol. 1, pp. 4-13, 2001. Andle J C, Sabah S, Stevens D S, Jumani S J, Baier M, Wall B W A, Martens T, and Gruenwald(4) R. Temperature Monitoring System Using Passive Wireless Sensors for Switchgear and Power Grid Asset Management. 2010. Pohl A and Seifert F. Wirelessly interrogable surface acoustic wave sensors for vehicular applications. Instrumentation and Measurement, IEEE Transactions on, vol. 46, pp. 1031-1038, 1997. Dixon B, Kalinin V, Beckley J, and Lohr R, "A Second Generation In-Car Tire Pressure Monitoring System Based on Wireless Passive SAW Sensors," in International Frequency Control Symposium and Exposition, 2006 IEEE, 2006, pp. 374-380. Marc. Loschonsky. Temperature and stress measurements on rotating machinery for industrial applicaions. Restricted access at - "Wireless SAW Sensor Symposium 2010", 12/11/2010 2010.

[9]

M. Pereira da Cunha A C, P.M. Duvulis, S. Moulzolf,, T. Moonlight R B, G. Bernhardt, D, Frankel, R.J. Lad, T., and Pollard D M. Wireless Interrogation of SAW Sensors in Operating Jet Engines at Extreme Temperatures. Restricted access at - "Wireless SAW Sensor Symposium 2010"

[10]

Wilson W, Rogge M, Malocha D, Fisher B, and Atkinson G. Fastener Failure Detection using a Surface Acoustic Wave Strain Sensor. Sensors Journal, IEEE, vol. PP, pp. 1-1, 2011. (09/01/2012). SENSeOR Transceivers. Available: http://www.senseor.com/transceivers.html (09/01/2012). SENGENUITY Sensor Engine Technology Available: http://www.sengenuity.com/tech_ref/TempTrackr_Wireless_Multipoint_System_ Kit-6.pdf (09/01/2012). Transense Technologies plc "Reader electronics". Available: http://www.transense.co.uk/technologies/reader-electronics Wolff U, Schmidt F, Scholl G, and Magori V, "Radio accessible SAW sensors for non-contact measurement of torque and temperature," in Ultrasonics Symposium, 1996. Proceedings., 1996 IEEE, 1996, pp. 359-362 vol.1. Kalinin V, Lohr R, and Leigh A, "Development of a calibration procedure for contactless torque and temperature sensors based on SAW resonators," in Ultrasonics Symposium, 2008. IUS 2008. IEEE, 2008, pp. 1865-1868. Beckley J, Kalinin V, Lee M, and Voliansky K, "Non-contact torque sensors based on SAW resonators," in Frequency Control Symposium and PDA Exhibition, 2002. IEEE International, 2002, pp. 202-213. (09/01/2012). Transense Technologies plc. Available: http://www.transense.co.uk/ Kalinin, V. Lohr, R. Leigh, and A. Bown G, "Application of Passive SAW Resonant Sensors to Contactless Measurement of the Output Engine Torque in Passenger Cars," in Frequency Control Symposium, 2007 Joint with the 21st European Frequency and Time Forum. IEEE International, 2007, pp. 499-504. Kalinin V. Wireless Resonant SAW Sensors for Automotive Applications. Restricted access at - "Wireless SAW Sensor Symposium 2010", 2010. Donohoe B, Geraghty D, and O'Donnell G E. Wireless Calibration of a Surface Acoustic Wave Resonator as a Strain Sensor. Sensors Journal, IEEE, vol. 11, pp. 1026-1032, 2011. Gordon S K and Matthews H. Signal processing in acoustic surface-wave devices. Spectrum, IEEE, vol. 8, pp. 22-35, 1971. Shvetsov A, Zhgoon S, Lonsdale A, and Sandacci S, "Deformation sensitive cuts of quartz for torque sensor," in Ultrasonics Symposium (IUS), 2010 IEEE, 2010, pp. 1250-1253. Reindl L, Scholl G, Ostertag T, Scherr H, Wolff U, and Schmidt F. Theory and application of passive SAW radio transponders as sensors. Ultrasonics,

[11] [12]

[13] [14]

[15]

[16]

[17] [18]

[19] [20]

[21] [22]

[23]

[24]

[25]

[26] [27]

Ferroelectrics and Frequency Control, IEEE Transactions on, vol. 45, pp. 12811292, 1998. Kalinin V, Bown G, Beckley J, and Lohr R, "Pulsed interrogation of the SAW torque sensor for electrical power assisted steering," in Ultrasonics Symposium, 2004 IEEE, 2004, pp. 1577-1580 Vol.3. Grossman R, Michel J, Sachs T, and Schrufer E, "Measurement of mechanical, quantities using quartz sensors," in European Frequency and Time Forum, 1996. EFTF 96., Tenth (IEE Conf. Publ. 418), 1996, pp. 376-381. Mini-Systems I. (01/06/2011). Electronic Packages. Available: http://www.minisystemsinc.com/msipkg/epd_pg.asp Donohoe B, Geraghty D, O’Donnell G E, and Stoney R. Packaging considerations for a Surface Acoustic Wave Strain Sensor. JOURNAL OF IEEE SENSORS, 2011.