Sensors and Actuators: A. Physical 334 (2022) 113334 Contents lists available at ScienceDirect Sensors and Actuators: A. Physical journal homepage: www.elsevier.com/locate/sna Fluid compressional properties sensing at microscale using a longitudinal bulk acoustic wave transducer operated in a pulse-echo scheme ]] ]]]]]] ]] ⁎ Jesus Yanez, Arantxa Uranga, Nuria Barniol Department of Electronics Engineering, Universitat Autònoma de Barcelona, 08193 Bellaterra, Spain a r t i cl e i nfo a bstr ac t Article history: Acoustic devices have been widely used as smart chemical and biochemical sensors since they are sensitive Received 16 August 2021 to mechanical, chemical, optical or electrical perturbations on their surfaces; making them a reliable option Received in revised form 25 November 2021 for noninvasive detection of changes in physical properties of liquid samples for real-time applications. Here Accepted 21 December 2021 we present a longitudinal acoustic wave device for study of compressional properties of liquids in micro­ Available online 23 December 2021 fluidic systems, with the particularity of pulse-echo mode of operation. We have studied at a microscale the interaction between longitudinal acoustic waves and the compressional properties of liquid samples, in­ Keywords: Piezoelectric terrogating the fluids with short pulses of ultrasound at GHz, finding a direct relationship between the AlN magnitude of the bulk modulus or the specific acoustic impedance of liquids and the amplitude of the Bulk acoustic wave output voltage produced by acoustic echoes received by the aluminum nitride transducer. Analytical ex­ Ultrasonic microsensors pressions and FEM simulations support the detection mechanism, while applications such as classification HBAR of liquids and detection of concentration change in solutions experimentally demonstrate the method. This Pulse-echo contribution overcomes current restrictions of film acoustic resonators such as fragility of operation in Time-of-flight liquid environments, high manufacturing cost or limitations regarding narrow microchannels; offering an Microfluidics alternative to applications that demand ultra-low consumption, miniaturization, versatility (it offers multi- frequency operation in 1 – 10 GHz range) and ease of readout (peak voltage). © 2022 The Authors. Published by Elsevier B.V. CC_BY_4.0 1. Introduction (QCM), the surface acoustic devices (SAW) or the bulk acoustic re­ sonators (BAR) have demonstrated their potential use as biosensors Body fluids such as blood or urine are an attractive source of because they are sensitive to mechanical, chemical or electrical information for disease detection [1–4], and one of the techniques perturbations on their surfaces [10,11], these microsensors offer high for track their condition is by monitoring their mechanical proper­ resolution sensing of temperature, moisture, strain, pressure, shock, ties [5–7], among which are its compressional properties. Com­ acceleration, flow, viscosity, and magnetic or electric fields, while pressional information is an interesting parameter in analysis since, offering a low cost solution with simple operation and readout in­ at microscales, it is closely related to the chemical bond strength tegration. Despite the relative maturity of acoustic wave technolo­ between atoms, ions, or molecules, and therefore can be used in gies such as QCM or SAW, thin-film bulk acoustic resonators (FBAR) analyzing human physiological fluids [8,9]. For instance a change in are extensively studied since they are several orders of magnitude the mechanical properties of blood can be potentially related with smaller than the former, allowing reduction of manufacturing cost cardiovascular dysfunctions, providing a fast and simple method for and miniaturization, which is an advantage for its integration in Lab- rapid and early test of a possible disease as it is already explained in on-a-Chip devices (LoC) [12]. reference [5]. Among the existent microscale biosensing technolo­ FBARs consist of a piezoelectric layer sandwiched between two gies for liquid phase measurement, the acoustic wave methods are thin conductive electrodes; when the thickness extensional mode of popular since they are noninvasive, passive and offer high sensi­ the active layer is excited it generates and traps acoustic energy tivity; acoustic transducers such as the quartz crystal microbalance within its thickness. The resonance frequency of such devices, ty­ pically in the GHz order, depends on both the thickness and sound velocity of longitudinal waves of the piezoelectric film [13], offering higher sensitivity and increased resolution compared to other ⁎ Correspondence to: Department of Electronics Engineering, Escola d'Enginyeria, acoustic resonators; however, these thin released structures suffer Building Q. Office: QC-3045, 08193 Bellaterra, Spain. from low yield due to residual stress and their use as biosensors for E-mail address: nuria.barniol@uab.cat (N. Barniol). https://doi.org/10.1016/j.sna.2021.113334 0924-4247/© 2022 The Authors. Published by Elsevier B.V. CC_BY_4.0 J. Yanez, A. Uranga and N. Barniol Sensors and Actuators: A. Physical 334 (2022) 113334 liquid media is limited because tend to be brittle (trench-type and the air-gap-FBARs) [14]. Energy confinement in the thin-film pie­ zoelectrical layer can be also achieved using Bragg-mirror-type FBARs, known as solid mounted resonators (SMR) [15]; such trans­ ducers offer improved mechanical robustness but require multiple precisely controlled deposition steps for the acoustic mirrors, raising the cost of the device and causing its performance to be dependent on manufacturing defects [16]. A similar approach that addresses the aforementioned disadvantages are the High-Overtone Bulk Acoustic Resonators (HBARs) used as sensors [17]; in this case the piezo­ electric and its electrodes are deposited directly on an acoustic substrate offering both, a robust structure and simpler manu­ facturing. Unlike FBAR or SMR devices, HBAR structures allow leakage of acoustic energy into the wafer, generating stationary bulk acoustic waves in the substrate thickness, therefore the free face of the wafer (the opposite of where the piezoelectric layer and its Fig. 1. A 3D transparent slice representing the concept of the sensor: The lower face electrodes were patterned) can be used as a detecting surface totally of the wafer carries the transducer that emits and receives pulses of bulk waves free of electronics [18], which is an attractive feature for micro­ through the thickness of the wafer while the upper face remains free of electronics fluidics carrying conductive or non-conductive liquids. and cables and is used as a detection surface carrying liquid samples. The zoomed Here we present an HBAR microstructure, operating in a pulse- section is a 2D representation of the stack of materials that make up the transducer; the piezoelectric material is AlN, sandwiched between conductive materials for ex­ echo scheme, for studying compressional properties of liquids in citation of thickness extensional modes. Thicknesses and materials correspond to the real-time applications. Rather than the elaborated mechanisms for PiezoMUMPs multi-user process. signal control and measurement used in acoustic resonators (which consist of detecting changes in amplitude and frequency shift of resonance peaks in the impedance curve caused by modifying the The loss moduli K” and G” are respectively related to compres­ acoustic load coupled to the resonator [9,18–22]), this pulse-echo sional and shear viscosities as: µB = K”/ω and µ = G”/ω [52,53], where scheme only requires measurement of the peak voltage produced by µB is the real part of the bulk viscosity and µ is the real part of the an acoustic echo within a determined time-of-flight (ToF) [23,24]. shear viscosity. Thus, the dynamic modulus of a liquid subjected to a The transducer presented here allows multiple operating fre­ longitudinal compressive force: quencies in the 1–10 GHz range, offering a low-power solution for 4 monitoring compressional properties of fluids in microchannels M* ( ) = K + i µB + µ where the active detection area and depth of penetration of ultra­ 3 (2) sound in liquid samples are critical. In the case of liquids, the real part of the bulk modulus K’ is re­ lated to the speed of sound c as K’ = ρc2. Therefore, substituting Eq. 2 2. Principle of operation and FEM simulations in Eq. 1, the complex acoustic impedance of a liquid subjected to a pressure wave is [9,26]: 2.1. Detection mechanism and system description 4 In this work, the microscale measurement of compressional Zliquid = Zreal 1+i µB + µ c2 3 properties of liquids is based on the interaction between pulses of longitudinal acoustic waves traveling across the solid sensor to the Zreal = c (3) liquid sample and the opposition exerted by such fluid to acoustic flow (its acoustic impedance). The complex acoustic impedance of The latter indicates that compressional information such as the materials Z* is related to density ρ and the dynamic complex mod­ Bulk Modulus and the specific Acoustic Impedance can be obtained ulus M* (ω) as [25]: by interrogating a liquid sample with pulses of longitudinal waves produced by a transducer. To ease the interaction between liquid Z* = M* ( ) (1) samples and the transducer at microscales we use the sensor de­ With ω = 2πf as the angular frequency. For longitudinal waves, M* scribed by Fig. 1, where one face of the wafer carries an AlN piezo­ is a function of the complex compressibility modulus (Bulk mod­ electric layer emitting and receiving ultrasound pulses and the ulus) K* and the shear modulus G*[25]: opposite face, which is a bare silicon surface with no electronics or wires, is enabled to implement microchannels carrying either con­ 4 ductive or non-conductive liquids. Materials and thicknesses in Fig. 1 M* ( ) = K* ( ) + G* ( ) 3 correspond to the PiezoMUMPs multi-user process [27], by The complex modulus K* is stated as K* = K’ + iK” [51], where K’ is MEMSCAP. the storage modulus representing the elastic part and K” is the loss Rather than operate this HBAR as a resonator, we drive its pie­ modulus, representing the viscous part. Similarly for the shear zoelectric element with a short RF pulse to excite thickness exten­ modulus G* we have G* = G’ + iG”, with G’ as the shear storage sional modes. In consequence a short ultrasound pulse of bulk waves modulus and G” as the shear loss modulus. For Newtonian media the travels across the thickness of the substrate and it interacts with the shear modulus G* can be reduced to its imaginary part G* ≈ iG” [52], liquid sample in the opposite face of the wafer. The Time-of-Flight so that the dynamic modulus of the liquid for longitudinal waves is: (ToF) needed for the ultrasound pulse is ToF = hsubs/csubs = 48 ns; considering hsubs = 411 µm as the total thickness of the stack and csubs 4 M* ( ) = K + iK + i G = 8457 m/s as the reported sound velocity of longitudinal waves in 3 silicon. The acoustic impedance discontinuity at the solid-liquid 2 J. Yanez, A. Uranga and N. Barniol Sensors and Actuators: A. Physical 334 (2022) 113334 Fig. 2. Acoustic pressure distribution at the detection surface for transducers of different width (50 µm, 70 µm and 90 µm). a) Cross-section schematic of the sensor. b) Lobe-width defined by the main-lobe pressure as a function of transducer width. c) Normalized pressure showing the size of pressure lobes at the detection surface. Computed at 1 GHz. interface causes a part of the pulse to be reflected back towards its = 0.5 µm) [29]. The configuration of materials surrounding the pie­ emitting piezoelectric element, which provides a method to de­ zoelectric layer (zoomed section in Fig. 1), leads to additional mod­ termine Zliquid by measuring the amount of acoustic energy reflected. ulation of the frequency comb, causing appearance of multiple If diffraction losses, damping and scattering are negligible, the resonating modes with increased electromechanical coupling in the amount of energy reflected in at the detection surface is [28]: 1–10 GHz range [34]. Therefore, if we set 1 GHz as the lowest oper­ ating frequency and the detection surface is 411 µm far from the |Zliquid Zsubstrate | | |= transducer, it is found by wsinθ = λ that a value of w between 80 and Zliquid + Zsubstrate (4) 90 µm corresponds to the narrowest transducer that can be used avoiding diffraction at that operating frequency, i.e., the width of the With Zsubstrate as the acoustic impedance of the substrate (silicon, main lobe at the detection surface will be similar to the width of the in our device). In this approach, the acoustic energy reflected to­ emitting transducer (Fig. 2b). Fig. 2c show the normalized pressure wards its source is converted back to an electrical signal by the in­ distribution at the detection surface caused by transducers of dif­ verse piezoelectric effect. The amplitude of the output voltage is ferent widths operating at 1 GHz; this 2D pressure distribution in the then proportional to the specific acoustic impedance Zliquid of each far field was analytically obtained from (5) [30], plotting only dis­ liquid. tribution along θ. kcua w 2 sin( ) sin( ) [j ( t kr)] 2.2. Microsystem design and acoustic pressure distribution at the p (r , , , t) = j e detection surface 2r (5) = (kwsin sin )/2 In this scheme the same single-element transducer is responsible for both transmission and reception of ultrasound pulses interacting = (kwsin cos )/2 with the material to be studied; therefore, it is convenient to mini­ mize diffraction of the acoustic energy during its travel across the Where k is the wavenumber, c is the sound velocity in silicon, ua is thickness of the wafer. For the far field case, the width of the ul­ the transducer vibration velocity, ρ is the medium density, w is the trasound pulse at the detection surface could be estimated with the width of the squared transducer, ω is the angular frequency and r is condition wsinθ = λ, where w is the width of the transducer, θ the the distance from the transducer to the point to evaluate P(θ). angle formed between the central axis of Fig. 2a and the point in the Through time-dependent Finite Element (FEM) simulations detection interface where the minima of the main lobe will appear, (COMSOL) we verify the acoustic pressure distribution at the sensing and λ is the wavelength of the longitudinal ultrasound wave tra­ surface. For t = 50 ns (Fig. 3a) the acoustic pulse has already traveled veling in the substrate, which is related to the operating frequency through the thickness of the substrate and has reached the detection fop as λ = csubs/fop. An HBAR structure consisting of a 0.5 µm-thick AlN surface. This pressure distribution plot confirms low diffraction of piezoelectric layer over a 400 µm-thick silicon exhibits multiple re­ the acoustic pulse emitted by a squared transducer of width w sonant modes separated by Δf = 11 MHz (Δf = csubs / [2 * hsubs]) in a = 90 µm and the expected ToF = 48 ns. The pressure distribution plot comb profile, modulated by the piezoelectric resonance f0 at in Fig. 3b exemplifies the detection mechanism at t = 72 ns, when 10.9 GHz (f0 = cpiezo / [2 * hpiezo], where cpiezo = 10,900 m/s, and hpiezo part of the acoustic energy has been transferred to the liquid media 3 J. Yanez, A. Uranga and N. Barniol Sensors and Actuators: A. Physical 334 (2022) 113334 Fig. 3. Time dependent simulation of the evolution of the ultrasound pulse emitted by the w = 90 µm transducer working at 1 GHz showing low diffraction of the acoustic energy: a) Transmitting pulse, b) Echo pulse. c) Time response plot of output voltage signal produced by the received acoustic echo at t > 90 ns (corresponding to ToFx2). Fig. 4. a) Optical image of a manufactured 1-D linear array consisting of three HBAR structures, each square transducer is 92 µm wide, only the central element is connected to the drive & read circuit. b) Cross section profile of the stack of materials used to fabricate the HBAR structure (Cross section representation is not to scale and it is flipped with respect to Fig. 1 to correspond with Fig. 4a, where the top layer is the metal PAD). 4 J. Yanez, A. Uranga and N. Barniol Sensors and Actuators: A. Physical 334 (2022) 113334 Table 1 Properties of the materials forming the stack of the manufactured device. Quantities correspond to default values from COMSOL library and their respective acoustic im­ pedance is obtained by the approximation Z ≈ ρcmat, with cmat as the sound velocity of each material. Material h [μm] ρ [kg/m3] cmat [m/s] Z ≈ ρcmat [Rayl] Air 1.14 349.1 398 Al 1 2700 6450 17.415E+6 AlN 0.5 3300 10954 36.148E+6 Doped Si 10 ± 1 2329 8457 19.696E+6 SiO2 1 2200 5848 12.866E+6 Si < 100 > 400 ± 5 2329 8457 19.696E+6 and it continues its flight in the direction indicated by the pink arrow, meanwhile, the complementary portion of the emitted acoustic energy was reflected back towards the emitter (red arrow). The voltage plot in Fig. 3c (FEM simulation) shows the output signal produced by the received acoustic echo at t > 90 ns; the amplitude of the output voltage will be proportional to Zliquid according to Eq. 4. 3. Fabrication and operating frequencies 3.1. Fabrication This sensor was manufactured using the PiezoMUMPs multi-user process from MEMSCAP [27]. The transducer consists of a 0.5 µm thick AlN layer patterned over a Silicon-On-Insulator wafer (SOI), and its metal electrode for drive and read-out purposes. The side of the wafer carrying the AlN layer is 10 ± 1 µm of silicon superficially Fig. 6. a) 2D model for frequency-domain FEM simulations, illustrating the confine­ doped, used here as reference to ground, allowing the piezoelectric ment of acoustic energy caused by a continuous low-power RF source in an HBAR to operate in its thickness extensional mode [31]. The remaining with dimensions corresponding to materials and thicknesses of the manufactured device. b) Input electrical impedance plot as result of frequency domain simulations thicknesses are 1 µm of SiO2 and 400 ± 5 µm of crystalline silicon, in the 0.8–2 GHz region; the configuration of layers and materials constituting our which is our lossless propagation medium for the ultrasound pulses. bulk acoustic transducer cause modulation of the frequency comb. This process leads us to the 1-D linear array of HBARs shown in Fig. 4a; according to the purposes of this document, only the central transducer (the device under test, DUT) is connected to the drive & 3.2. Operating frequency read circuitry. Fig. 4b is a cross section representation of the stack of materials used to fabricate the HBAR structure, this cross section The resonating frequency of a thin piezoelectric layer operating representation is not to scale. on its thickness expansion mode is given by fr = cpiezo/2hpiezo, with cpiezo as the sound velocity of longitudinal waves in the piezoelectric material and hpiezo as its thickness [31]. When this film is deposited on an acoustic substrate, bulk acoustic waves are generated and keep trapped in the thickness of the whole stack. If the substrate thick­ ness hsubs > > hpiezo, then the fundamental mode of the structure is f0 ≈ csubs/2hsubs (with csubs as the sound velocity of the substrate). The electrical response of such structure is a comb formed by the mul­ tiple harmonics of the fundamental mode of the structure, separated by Δf0 [32]. For this two-layer device the amplitude of the comb of substrate modes is modulated by the resonance of the piezoelectric. Table 1 lists material properties and thicknesses hmat of materials forming the fabricated device. According to values of Table 1, the magnitude of the electrical response of our device is a comb of resonances separated by Δf0 = 11 MHz, modulated by the piezoelectric resonance at fr = 10.9 GHz. The above allows formation of stationary bulk acoustic waves of λ < 10 µm for operating frequencies above 1 GHz, therefore, the in­ fluence of the thickness of the different materials surrounding the AlN layer is not negligible [31,33]. Fig. 5 shows thicknesses and acoustic impedance of the different materials surrounding the pie­ zoelectric; since the acoustic impedance of these solid materials is Fig. 5. Thickness and acoustic impedance of materials surrounding the piezoelectric layer. the same order of magnitude as that of the AlN film, it is expected 5 J. Yanez, A. Uranga and N. Barniol Sensors and Actuators: A. Physical 334 (2022) 113334 Fig. 8. Optical image of the experimental setup. interact with the excitation/read circuit. The schematic in Fig. 7a describes the electronic components used to drive and read the sensor. It starts with a Voltage Controlled Oscillator (VCO) ZX95–1750 W-S+ generating a 1.12 GHz continuous RF signal (6 dBm), then a Cortex-M7 microcontroller activate both the SPST Fig. 7. a) Schematic of the experimental set-up. b) Characteristics of the excitation RF Switch (ADG901) and the SPDT switch (ADG918) allowing to pass a pulse. short RF pulse of duration τ = 25 ns as the shown in Fig. 7b. Once the ultrasound pulse has been emitted the microcontroller returns the SPDT switch to its original position allowing the voltage signal that the acoustic energy generated by exciting the thickness exten­ produced by the reflected echoes to pass through the Variable Gain sion mode of the piezoelectric will be transferred to these materials Amplifier (VGA) ZFL-1200 G+ (gain 25.8 dB @1.1 GHz), the output causing multiple acoustic reflections. signal is retrieved by the 50 Ω input of the oscilloscope Rohde& In Fig. 5 the Al electrode is twice as thick as the piezoelectric and Schwarz RTO2064. Since one face of the chip carries the transducer both materials have similar acoustic impedance, therefore, the me­ and the detection surface is at the opposite face, the GCPW is flipped chanical energy produced by the piezoelectric is immediately upside down as it features a hole exposing the detection surface transferred to this thicker layer and then is reflected back towards (bare silicon) to place liquid samples. Samples are delivered with a the piezoelectric (red arrow) because of the huge acoustic im­ transfer pipette (sample ~ 50 µL) and confined with an o-ring (Fig. 8). pedance difference with ‘air’ in the lower boundary. In the opposite As speed of sound in materials changes with temperature, side, partial reflection of the longitudinal wave is induced by the leading to deviations of acoustic impedance, we take measurements mismatch between acoustic impedances of the doped silicon and of the reflection coefficient (silicon/liquid interface) at the same buried oxide. To investigate the influence of these reflections in the room temperature for all liquid samples. The following experiments transfer function of the device we use the 2-D Frequency-domain were performed to verify sensitivity to both Bulk Modulus and FEM simulation of Fig. 6a. The impedance plot in Fig. 6b compares Acoustic Impedance of tested materials and to determine how they the magnitude of the input electrical impedance Zin for a transducer are related to the amplitude of the output voltage signals: a) Ability as manufactured (black trace) respect to the transfer function of a to detect the presence of liquid on the detection surface and feasi­ similar device but electrodes of negligible thickness (magenta), for bility of application as a biometric fingerprint sensor, b) the frequency range 0.8–2 GHz. Identification and classification of liquids by their specific acoustic As seen in the electrical impedance plot (Fig. 6b), the config­ impedance, and c) Detection of concentration change in solutions uration of layers and materials constituting this bulk acoustic (with different mixtures of glycerol and water). Table 2 lists prop­ transducer cause additional modulation of the frequency comb. erties of the liquids used in the experimental part and their expected These frequency spots with higher amplitude modes also offer a reflection coefficient Γ calculated by Eq. (4). Due to the relative high significant increase in the effective electromechanical coupling k2eff viscosity of glycerol and the interaction with pressure acoustic [34], improving transduction of our pulse-echo devices and pro­ waves at GHz, the frequency-dependent value of its acoustic im­ viding them with multi-frequency capability. For the sake of sim­ pedance (Eq. 3) is also included. plicity, here we have used the peak of transduction located at 1.1 GHz for the experimental part. 4.1. Presence of liquids in microchannels and impedance imaging (fingerprint sensor) 4. Experimental part and discussion To test the capability of this all-solid sensor to identify if a mi­ The device under test (DUT) is the squared transducer with side crochannel is filled with liquid or not, we record the output voltage w = 92 µm shown in Fig. 4a, this device is wire bonded to a 50 Ω produced by echoes reflected from the detection surface for both, Grounded Coplanar Waveguide (GCPW) with an SMA connector to when there is liquid on the detection surface and when it is free. 6 J. Yanez, A. Uranga and N. Barniol Sensors and Actuators: A. Physical 334 (2022) 113334 Table 2 Properties of liquids analyzed (values obtained from literature [35–46]). Material Density Acoustic velocity Viscosity Bulk modulus Acoustic impedance Γ 31 °C ρ [kg/m3] c [m/s] μ K’ = ρc2 Z ≈ ρc [Rayl] (Eq.4) [cP] [GPa] Air 1.14(35) 349.1(36) 0.018(35) 1.3893E-04 398 0.9999 Ethanol 780.1(37) 1105.7(38) 1.06(39) 0.9537 0.8625E+06 0.9161 FC-70 1924(40) 680(41) 3.36(40) 0.8896 1.3083E+06 0.8754 DI Water 995.6(42) 1511.4(43) 0.79(44) 2.2743 1.5047E+06 0.8580 Glycerol 1254(45) 1902(46) 552.1(45) 4.5364 2.3851E+06 0.7839 Acoustic impedance Z(ω),Eq. 3 [Rayl] Glycerol* 1254 1902 552.1 3.2239E+06 0.7186 @ 1.1 GHz Fig. 9. Envelope of five consecutive echoes for the case of "air" on the detection Fig. 10. Reflection coefficient obtained from measurements and related to the specific surface (black) and envelope for the case of "water" (blue). acoustic impedance of each substance. Due to the viscosity of glycerol (> 500 cP), a frequency-dependent acoustic impedance value (Eq. 3) has also been added. Fig. 9 presents the envelope of output signal recorded by the oscil­ loscope, up to t = 550 ns, after applying one input RF pulse for the coefficient differences in Table 2, demonstrating ease of detection as case of both, no material filling the o-ring, and filled with deionized long as the 'air' signal is used as a reference. water. Since acoustic impedance of skin is close to that of water, an Fig. 9 showed a first echo rising up to 38 dB above the noise floor array of these transducers can be used for imaging purposes (ul­ and the time-of-flight is according with the expected value ToF(x2) trasonic fingerprint sensors) by recording the difference in output ≈ 96 ns (for a round trip across the thickness of the wafer). This echo voltage between zones of the sensing surface in contact with air is followed by multiple remaining echoes of lower amplitude with (fingerprint valleys), and zones in contact with skin (fingerprint same ToF since the side of the wafer carrying the transducer is also ridges) [47]. As seen in Fig. 9, each pulse reflection (round trip) in­ in contact with air, causing total reflection of the acoustic energy and creases the difference between the envelope caused by 'water' and therefore interaction with the material to be studied more than once the caused by 'air' (because acoustic energy leakage into the liquid while the pulse gradually fades with time. Considering only the first on each interaction), therefore, enhanced contrast can be obtained at echo we notice that the amplitude of the envelope corresponding to the cost of increasing acquisition time. We use the fourth echo the case of distilled water (blue trace) drops close to 16% respect to (t = 400 ns, SNR ~ 20 dB) to indicate contrast between echoes re­ the maximum amplitude obtained when the detection surface is flected by water (acting as a fingerprint ridge) respect to echoes cleared (air, black trace), which agrees to the theoretical reflection reflected by the cleared sensor (acting as a fingerprint valley) in the Table 3 Comparison against a similar transducer implemented as an imaging element (fingerprint sensor). [47], 2019 This transducer Technology AlN MEMS + 0.18 µm CMOS (monolithic) AlN MEMS + Off the shelf electronics Mode of operation Pulse-echo Pulse-echo Elements required 2 (Transmitter + Receiver) 1 (Same element for Tx & Rx) Drive voltage > 1V < 0.4 V Frequency 1.69 GHz 1.1 GHz (but possible for higher frequencies, up to 10 GHz Contrast 2.3: 1 2: 1 Acquisition time 344 ns 400 ns (Second echo) (Fourth echo) 7 J. Yanez, A. Uranga and N. Barniol Sensors and Actuators: A. Physical 334 (2022) 113334 Fig. 11. a) Measured reflection coefficient as a function of glycerol content, using the first echo to detect changes in concentration. b) Measurements using the second echo. Fig. 12. Measured reflection coefficient of each glycerol concentration (0–60%), as a function of a) Reported Acoustic impedance b) Bulk modulus. comparison of Table 3. Note that in this work only one single ele­ ment is used both for Tx and Rx. According to the diffraction condition wsinθ = λ, our approach Table 4 with multi-frequency capability (up to 10 GHz) would allow single Comparison to a reported resonating sensor of liquid bulk-modulus at microscales. element transducers of width w = 25 µm sending low diffraction [9], 2019 This transducer pulses of ultrasound to scan the detection surface with a lateral Acoustic device type SMR HBAR resolution of 30 µm, improving current ultrasonic sensors while of­ (FBAR) fering very low drive voltage and low electrical input impedance Mode of operation Resonating Pulse-echo (~150 Ω for such parameters). Sensitive area 0.01 mm2 0.008 mm2 Drive voltage 10 dBm 4 dBm (continuous) (25 ns pulse) LAW frequency 2.5 GHz 1 – 10 GHz range 4.2. Identification and classification of liquids by their specific acoustic Liquid operation Yes Yes impedance by passivation layer native Parameter to detect Change of resonant Voltage amplitude frequency Liquid substances in Table 2 exhibit dissimilar densities and Bulk modulus sensitivity 5.83 MHz / GPa 6.89% / GPa * longitudinal sound velocities, which offer opportunity of classifica­ * Percentage change of reflection coefficient, taking water measurement as reference. tion according to their degree of mismatch respect to the char­ acteristic acoustic impedance of the silicon substrate transporting the pulses of longitudinal waves emitted by the piezoelectric ele­ while Vliq is voltage generated in the case of each liquid (which ment. Fig. 10 relates the values in Table 2 for reflection coefficient Γ depends on the amount of acoustic energy absorbed by each liquid). As seen in Fig. 10, the reflection coefficient obtained from the with their respective acoustic impedance Z (reference values). In the case of experimental data, each reflection coefficient is obtained measured output voltage of each liquid shows a linear relationship from the Vliq/Vref relation, where Vref is the voltage generated by with the reference values for acoustic impedance. It is appreciated received echoes when the detection surface is clear ('air' causes total major dispersion for the measurement "glycerol", verifying that it reflection of the acoustic pulse and therefore a reference value), should not be correlated neither with the acoustic impedance value 8 J. Yanez, A. Uranga and N. Barniol Sensors and Actuators: A. Physical 334 (2022) 113334 calculated from the approximation Z ≈ ρc nor with the FEM result approaches, this bulk acoustic structure is easy and inexpensive to based on a simple linear elastic fluid model, but with the value manufacture, providing superior mechanical robustness and a sen­ obtained from Eq. 3, since the high viscosity of this material makes sing surface free of electronics where microchannels can be con­ the frequency dependences relevant [48–50]. structed carrying even conductive liquids. This device provides an alternative for noninvasive study of liquid samples at microscales; 4.3. Detection of concentration change in solutions envisioned for real-time applications where low-power consump­ tion, easy readout method and size matters. We prepare glycerol/water mixtures (percentage by weight) to determine the ability to identify changes in concentration, demon­ Author contributions strate correlation between the output voltage of the sensor and bulk modulus K, and confirm lack of sensitivity to viscosity when Jesus Yanez: Conceived and designed the analysis, Collected the studying biological fluids, where viscosity rarely exceeds 10 cP [48]. data, Performed the analysis, Wrote the paper. Arantxa Uranga: Each experimental point in Fig. 11a is the average value of five Contributed data or analysis tools, Other contribution (Revision). measurements performed at each concentration value (first echo), Nuria Barniol: Conceived and designed the analysis, Contributed normalized to ‘water’ (0% glycerol). Measurements were performed data or analysis tools, Wrote the paper, Other contribution at 31 °C (room temperature), so that the blue reference indicates the (Revision). analytical value of the viscosity of each solution at that temperature [45]. Fig. 11b shows measurements of the second echo. Declaration of Competing Interest Fig. 11a shows that the level of acoustic energy absorbed by the liquid sample increases linearly (R2 > 0.97) as the concentration of The authors declare that they have no known competing fi­ glycerol in the mixture increases up to 80% (viscosity = 32.7 cP, at nancial interests or personal relationships that could have appeared 31 °C [45]). From that point, experimental data show slight sensi­ to influence the work reported in this paper. tivity to viscosity, when it increases its value above 32.7 cP (blue trace). If detection with the second echo is performed, the same Acknowledgments behavior is observed, but for this case the pulse has interacted twice with the material to be detected, leading to a steeper calibration This research was partially funded by the Spanish Ministry of curve. Fig. 12a relates experimental measurements of Fig. 11b with Science and Innovation, with project PID2019-108270RB-I00. their respective Acoustic Impedance value, obtained from literature [49], up to 60%. The values for Bulk Modulus in Fig. 12b were ob­ References tained from the same acoustic impedance data, using Eq. 3. Fig. 12 indicates that the output voltage decreases 1.8% for each [1] I.J. Miller, S.R. Peters, K.A. Overmyer, B.R. Paulson, M.S. Westphall, J.J. Coon, Real- time health monitoring through urine metabolomics, npj Digit. Med. 2 (2019) 100 kRayl increment in the acoustic impedance Z, and it decreases 109, https://doi.org/10.1038/s41746-019-0185-y 6.9% per GPa in the case of changes in Bulk modulus K. This sensor [2] J.J. Coon, P. Zürbig, M. Dakna, A.F. Dominiczak, S. Decramer, D. Fliser, shows high linearity for the measurement of these parameters M. Frommberger, I. Golovko, D.M. Good, S. Herget-Rosenthal, J. Jankowski, B.A. Julian, M. Kellmann, W. Kolch, Z. Massy, J. Novak, K. Rossing, J.P. Schanstra, (R2 > 0.99, for both Z and K) as long as the viscosity of the solution E. Schiffer, D. Theodorescu, R. Vanholder, E.M. Weissinger, H. Mischak, does not exceed 10 cP as shown in Fig. 10 and Fig. 11, while further P. Schmitt-Kopplin, CE-MS analysis of the human urinary proteome for bio­ adjustments should be performed with liquid samples of much marker discovery and disease diagnostics, Prot. Clin. Appl. 2 (2008) 964–973, https://doi.org/10.1002/prca.200800024 higher viscosity, such as pure glycerin, with a value above 500 cP at [3] M. Li, Urine reflection of changes in blood, in: Y. Gao (Ed.), Urine Proteomics in 31 °C. This sensor boasts application for biosensing in liquid media, Kidney Disease Biomarker Discovery, Advances in Experimental Medicine and where viscosity rarely exceeds 10 cP. Studies on compressional Biology, 845 Springer, 2015, pp. 13–19, https://doi.org/10.1007/978-94-017- 9523-4_2 properties of liquids are mainly tackled by large-scale systems [4] Wu Jianqiang, Youhe Gao, Physiological conditions can be reflected in human making it difficult to find direct competitors to compare; Table 4 urine proteome and metabolome, Expert Rev. Proteom. 12 (2015) 623–636, compares our pulse-echo microsensor to a recently reported https://doi.org/10.1586/14789450.2015.1094380 acoustic resonator based on SMR technology. [5] Hye Jin Kim, Kim Jinsik, Zandieh Omid, Myung-Sic Chae, Tae Song Kim, Jeong Hoon Lee, Jung Ho Park, Seonghwan Kim, Kyo Seon Hwang, Piezoelectric Despite sharing some characteristics such as similar dimensions layer embedded-microdiaphragm sensors for the determination of blood visc­ of the active area or wavelength, our HBAR-based microsensor offers osity and density, Appl. Phys. Lett. 105 (2014) 153504, https://doi.org/10.1063/1. lower drive voltage with higher sensitivity, simpler manufacture and 4898637 [6] K. Roy, H. Gupta, V. Shastri, A. Dangi, A. Jeyaseelan, S. Dutta, R. Pratap, Fluid increased reliability respect to FBAR/SMR devices. In addition, its density sensing using piezoelectric micromachined ultrasound transducers, IEEE multi-frequency capabilities allow controlling penetration in liquids Sens. J. 20 (2020) 6802–6809, https://doi.org/10.1109/JSEN.2019.2936469 avoiding unwanted reflections in narrow channels or due to low [7] X. Lv, Y. Luo, Y. Zheng, C. Dong and M. Deng, The Body Fluid Density Sensor Based on Ultrasonic Wave, 2006 IEEE International Conference on Information sample volumes, which can also be discarded considering only sig­ Acquisition, (2006), 727–731, https://doi.org/10.1109/ICIA.2006.305818. nals received at the expected ToF. A very controllable RF pulse with [8] Siyuan Zhang, Huiling Li, Huaiyong Li, Shihong Zhou, Xueqiang Cao, Calculation of finite length is chosen to avoid additional and non-desired echoes. the bulk modulus of simple and complex crystals with the chemical bond method, J. Phys. Chem. B 111 (2007) 1304–1309, https://doi.org/10.1021/jp0651539 Moreover, the pulse-echo mode of operation offers a much simpler [9] Yao Lu, Menglun Zhang, Hongxiang Zhang, Yuan Jiang, Hao Zhang, Wei Pang, drive/lecture method, where only output voltage amplitude should Microfluidic bulk-modulus measurement by a nanowavelength longitudinal- be recorded. acoustic-wave microsensor in the nonreflective regime, Phys. Rev. Appl. 11 (2019) 044091, https://doi.org/10.1103/PhysRevApplied.11.044091 [10] J.C. Andle, J.F. Vetelino, Acoustic wave biosensors, Sens. Actuators A Phys. 44 5. Conclusion (1994) 167–176, https://doi.org/10.1016/0924-4247(94)00801-9 [11] Moussa Hoummady, Andrew Campitelli, Wojtek Wlodarski, Acoustic wave In this work we have presented a microsensor to study com­ sensors: design, sensing mechanisms and applications, Smart Mater. Struct. 6 (1997) 647–657, https://doi.org/10.1088/0964-1726/6/6/001 pressional properties of liquids in microfluidic systems by using longitudinal acoustic waves in a pulse-echo regime. Unlike FBAR 9 J. Yanez, A. Uranga and N. Barniol Sensors and Actuators: A. Physical 334 (2022) 113334 [12] A. Mujahid, A. Afzal, F.L. Dickert, An overview of high frequency acoustic sensors- Symposium on Applications of Ferroelectrics (IFCS-ISAF), 2020, 1–5. https://doi. QCMs, SAWs and FBARs-chemical and biochemical applications, Sensors 19 org/10.1109/IFCS-ISAF41089.2020.9234913. (2019) 4395, https://doi.org/10.3390/s19204395 [34] J. Yanez, A. Uranga and N. Barniol, Multi-frequency thin film HBAR microsensor [13] Harmeet Bhugra, Gianluca Piazza, Piezoelectric MEMS Resonators, Springer, for acoustic impedance sensing over the GHz range, 21st International 2017, https://doi.org/10.1007/978-3-319-28688-4 Conference on Solid-state sensors, Actuators and Microsystems (Transducers), [14] Y.Q. Fu, J.K. Luo, N.T. Nguyen, A.J. Walton, A.J. Flewitt, X.T. Zu, Y. Li, G. McHale, 2021, 1347–1350, https://doi.org/10.1109/Transducers50396.2021.9495750. A. Matthews, E. Iborra, H. Du, W.I. Milne, Advances in piezoelectric thin films for [35] J. Kestin, J.H. Whitelaw, The viscosity of dry and humid air, Int. J. Heat Mass acoustic biosensors, acoustofluidics and lab-on-chip applications, Prog. Mater. Transf. 7 (1964) 1245–1255, https://doi.org/10.1016/0017-9310(64)90066-3 Sci. 89 (2017) 31–91, https://doi.org/10.1016/j.pmatsci.2017.04.006 [36] Cramer Owen, The variation of the specific heat ratio and the speed of sound in [15] S. Jose, A.B. M. Jansman and R.J. E. Hueting, A design procedure for an acoustic air with temperature, pressure, humidity, and CO2 concentration, J. Acoust. Soc. mirror providing dual reflection of longitudinal and shear waves in Solidly Am. 93 (1993) 2510–2516, https://doi.org/10.1121/1.405827 Mounted BAW Resonators (SMRs), 2009 IEEE International Ultrasonics [37] T.S. Khasanshin, A.A. Aleksandrov, Thermodynamic properties of ethanol at at­ Symposium, (2009), 2111–2114, https://doi.org/10.1109/ULTSYM.2009.5442065. mospheric pressure, J. Eng. Phys. 47 (1984) 1046–1052, https://doi.org/10.1007/ [16] L. García-Gancedo, J. Pedrós, E. Iborra, M. Clement, X.B. Zhao, J. Olivares, BF00873717 J. Capilla, J.K. Luo, J.R. Lu, W.I. Milne, A.J. Flewitt, Direct comparison of the [38] J. Tong, M.J.W. Povey, X. Zou, B. Ward, C.P. Oates, Speed of sound and density of gravimetric responsivities of ZnO-based FBARs and SMRs, Sens. Actuators B ethanol-water mixture across the temperature range 10 to 50 degrees Celsius, J. Chem. 183 (2013) 136–143, https://doi.org/10.1016/j.snb.2013.03.085 Phys. Conf. Ser. 279 (2011), https://doi.org/10.1088/1742-6596/279/1/012023 [17] T. Baron, E. Lebrasseur, F. Bassignot, G. Martin, V. Pétrini, S. Ballandras, High- [39] I.S. Khattab, F. Bandarkar, M.A.A. Fakhree, A. Jouyban, Density, viscosity, and overtone bulk acoustic resonator, in: Marco G. Beghi (Ed.), Modeling and surface tension of water+ethanol mixtures from 293 to 323K, Korean J. Chem. Measurement Methods for Acoustic Waves and for Acoustic Microdevices, Eng. 29 (2012) 812–817, https://doi.org/10.1007/s11814-011-0239-6 IntechOpen, 2013, , https://doi.org/10.5772/56175 [40] 3M Fluorinert Electronic Liquid FC-70, 〈https://www.3m.com/3M/en_US/p/d/ [18] Y. Yamakawa, K. Sano, R. Karasawa and T. Yanagitani, Broadband frequency b40045189/〉, 2021, (accessed 2 July 2021). viscosity measurement using low TCF shear mode resonators consisting of C- [41] L.M. Krutyanskiy, F. Zoueshtiagh, P. Pernod, P. Shirkovskiy, A.P. Brysev, Separation axis tilted scaln thin film on thick at-cut quartz plate, 19th International of two fractions of immiscible liquids by ultrasound in microgravity, Phys. Wave Conference on Solid-State Sensors, Actuators and Microsystems Phenom. 25 (2017) 151–155, https://doi.org/10.3103/S1541308X17020133 (TRANSDUCERS), 2017, 2135–2138, https://doi.org/10.1109/TRANSDUCERS.2017. [42] J.B. Patterson, E.C. Morris, Measurement of absolute water density, 1 °C to 40 °C, 7994497. Metrologia 31 (1994) 277–288, https://doi.org/10.1088/0026-1394/31/4/001 [19] W. Xu, J. Appel, J. Chae, Real-time monitoring of whole blood coagulation using a [43] Nykolai Bilaniuk, George S.K. Wong, Speed of sound in pure water as a function microfabricated contour-mode film bulk acoustic resonator, J. of temperature, J. Acoust. Soc. Am. 93 (1993) 1609–1612, https://doi.org/10.1121/ Microelectromech. Syst. 21 (2012) 302–307, https://doi.org/10.1109/JMEMS.2011. 1.406819 2179011 [44] Joseph Kestin, Mordechai Sokolov, A.Wakeham William, Viscosity of liquid water [20] A.J. Flewitt, J.K. Luo, Y.Q. Fu, L. Garcia-Gancedo, X.Y. Du, J.R. Lu, X.B. Zhao, in the range −8 °C to 150 °C, J. Phys. Chem. Ref. Data 7 (1978) 941–948, https:// E. Iborra, M. Ramos, W.I. Milne, ZnO based SAW and FBAR devices for bio-sensing doi.org/10.1063/1.555581 applications, J. Non-Newton. Fluid Mech. 222 (2015) 209–216, https://doi.org/10. [45] Cheng Nian-Sheng, Formula for the viscosity of a glycerol−water mixture, Ind. 1016/j.jnnfm.2014.12.002 Eng. Chem. Res. 47 (2008) 3285–3288, https://doi.org/10.1021/ie071349z [21] L.M. Dorojkine, V.V. Volkov, V.S. Doroshenko, A.A. Lavrenov, D.A. Mourashov, [46] Ana P.V. Egas, Nieves M.C. Talavera-Prieto, Abel G.M. Ferreira, Jaime B. Santos, I.A. Rozanov, Thin-film piezoelectric acoustic sensors. Application to the detec­ M.ário J. Santos, Zaida L. Almeida, Isabel M.A. Fonseca, Speed of sound and de­ tion of hydrocarbons, Sens. Actuators B Chem. 44 (1997) 488–494, https://doi. rived thermodynamic properties of glycerol, J. Chem. Thermodyn. 156 (2021) org/10.1016/S0925-4005(97)00260-8 106367, https://doi.org/10.1016/j.jct.2020.106367 [22] G.D. Mansfeld, Theory of high overtone bulk acoustic wave resonator as a gas [47] M. Abdelmejeed, J. Kuo, A. Ravi and A. Lal, CMOS Controlled Ghz Ultrasonic sensor, 13th International Conference on Microwaves, Radar and Wireless Impedance Imager, 20th International Conference on Solid-State Sensors, Communications, 2, 2000, 469–472, https://doi.org/10.1109/MIKON.2000. Actuators and Microsystems & Eurosensors XXXIII (TRANSDUCERS & 913971. EUROSENSORS XXXIII), 2019, 57–60, https://doi.org/10.1109/TRANSDUCERS. [23] J.C. Kuo, J.T. Hoople, M. Abdelmejeed, M. Abdel-moneum and A. Lal, 64-Pixel 2019.8808493. solid state CMOS compatible ultrasonic fingerprint reader, IEEE 30th [48] C.édric Ayela, Liviu Nicu, Micromachined piezoelectric membranes with high International Conference on Micro Electro Mechanical Systems (MEMS), 2017, nominal quality factors in newtonian liquid media: a Lamb’s model validation at 9–12, https://doi.org/10.1109/MEMSYS.2017.7863326. the microscale, Sens. Actuators B Chem. 123 (2007) 860–868, https://doi.org/10. [24] M. Abdelmejeed et al., Monolithic 180nm CMOS Controlled GHz Ultrasonic 1016/j.snb.2006.10.048 Impedance Sensing and Imaging, 2019 IEEE International Electron Devices [49] H. Antlinger, S. Clara, R. Beigelbeck, S. Cerimovic, F. Keplinger, B. Jakoby, Sensing Meeting (IEDM), 2019, 34.3.1–34.3.4., https://doi.org/10.1109/IEDM19573.2019. the characteristic acoustic impedance of a fluid utilizing acoustic pressure 8993623. waves, Sens. Actuators A Phys. 186 (2012) 94–99, https://doi.org/10.1016/j.sna. [25] C. Verdier, P.Y. Longin, M. Piau, Dynamic shear and compressional behavior of 2012.02.050 polydimethylsiloxanes: Ultrasonic and low frequency characterization, Rheol. [50] S. Cerimovic, R. Beigelbeck, H. Antlinger, J. Schalko, B. Jakoby, F. Keplinger, Acta 37 (1998) 234–244, https://doi.org/10.1007/s003970050111 Sensing viscosity and density of glycerol–water mixtures utilizing a suspended [26] Y. Lu, H. Zhang, Y. Jiang, H. Zhang, W. Pang and M. Zhang, Bulk modulus mea­ plate MEMS resonator, Microsyst. Technol. 18 (2012) 1045–1056, https://doi.org/ surement of microfluidics by longitudinal bulk acoustic wave sensors, 2018 IEEE 10.1007/s00542-012-1437-9 Micro Electro Mechanical Systems (MEMS), (2018), 854–857, https://doi.org/10. [51] Marc Meyers and Krishan Chawla. Mechanical Behavior of Materials, 2nd 1109/MEMSYS.2018.8346690. Edition, Cambridge University Press, 2009. [27] PiezoMUMPs Reference Material, Design rules, 〈http://www.memscap.com/ [52] C. Verdier, M. Piau, Acoustic wave propagation in two-phase viscoelastic fluids: products/mumps/piezomumps/reference-material〉, 2012, (accessed 2 July the case of polymer emulsions, J. Acoust. Soc. Am. 101 (4) (1997) 1868–1876, 2021). https://doi.org/10.1121/1.418238 [28] Lawrence E. Kinsler, Austin R. Frey, Alan B. Coppens, James V. Sanders, [53] D.J. Dos Santos, D.J. Carastan, L.B. Tavares, G.F. Batalha, Polymeric Materials Fundamentals of Acoustics, 4th edition, John Wiley & Sons Inc, 2000. Characterization and Modeling, Comprehensive Materials Processing, 2 (2014), [29] J. Yanez, F. Torres, A. Uranga and N. Barniol, A feasibility study of AlN ultrasonic pp. 37–63, https://doi.org/10.1016/B978-0-08-096532-1.00205-3 transducers fabrication using the multi-user PiezoMUMPs process for finger­ print scanning at GHz range, 15th Conference on Ph.D Research in Jesus Yanez received both the B.Eng. degree in telecommunication and electronics Microelectronics and Electronics (PRIME), 2019, 293–296. https://doi.org/10. engineering and the M.Sc. degree in Computer engineering from the National 1109/PRIME.2019.8787767. Polytechnic Institute, México, in 2008 and 2014 respectively. He is currently pursuing [30] G. Fan, J. Li, C. Wang, Design and analysis of MEMS linear phased array, the Ph.D. degree in electronics engineering at Universitat Autònoma de Barcelona, Micromachines 7 (2016) 8, https://doi.org/10.3390/mi7010008 Spain. His research interest includes piezoelectric acoustic transducers for ultrasonic [31] J.F. Rosenbaum, Bulk Acoustic Wave Theory and Devices, Artech House, 1988. sensor applications. [32] S. Ballandras, T. Baron, E. Lebrasseur, G. Martin, D. Gachon, A. Reinhardt, P. Lassagne, J. Friedt, L. Chommeloux, D. RabusHigh, High Overtone Bulk Acoustic Arantxa Uranga received the degrees in physics and electronics engineering from Resonators: application to resonators, filters and sensors, Acoustics 2012 (2012) Valladolid University, Spain, in 1994 and 1996, respectively, and the Ph.D. degree in 3111–3117. electronics engineering from the Universitat Autònoma de Barcelona, Spain, in 2001. [33] J. Yanez, E. Ledesma, A. Uranga and N. Barniol, Improved Electromechanical Since 1996, she has been with the Department of Electronics Engineering, Universitat Transduction for PiezoMUMPS HBAR Impedance Sensors, 2020 Joint Conference Autònoma de Barcelona, where she is currently an Associate Professor. Her research of the IEEE International Frequency Control Symposium and International interests include the design and implementation of CMOS integrated circuitry in­ tended to sense the movement of MEMS-NEMS resonators; the implementation of 10 J. Yanez, A. Uranga and N. Barniol Sensors and Actuators: A. Physical 334 (2022) 113334 self-sustained oscillators based on MEMS, and the implementation of CMOS circuitry resonators and their integration within CMOS technologies focused on the reduction to excite and sense piezoelectric micromachined ultrasound transducers (PMUTs). of dimensions towards nanoelectromechanical devices with optimized CMOS con­ ditioning circuitry. She has coauthored more than 100 research articles and 200 peer- Núria Barniol received the Ph.D. degree in physics from the Universitat Autònoma de reviewed conferences. Her research interests include the study of novel piezo­ Barcelona (UAB), Spain, in 1992. She is a Full Professor with the Department of electrical micro/nanometric ultrasonic transducers, their integration with CMOS to­ Electronics Engineering, UAB. She has been working for the last 20 years with MEMS ward efficient biometrics systems and exploitation of non-linear NEMS resonators. 11