Acoustic radiation-free surface phononic crystal resonator for in-liquid low-noise gravimetric detection
Feng Gao, Amine Bermak, Sarah Benchabane, Laurent Robert, Abdelkrim Khelif
GGao et al.
Microsystems & Nanoengineering (2021) 7:8
Microsystems & Nanoengineering
A R T I C L E O p e n A c c e s s
Acoustic radiation-free surface phononic crystalresonator for in-liquid low-noise gravimetricdetection
Feng Gao , Amine Bermak , Sarah Benchabane , Laurent Robert and Abdelkrim Khelif Abstract
Acoustic wave resonators are promising candidates for gravimetric biosensing. However, they generally suffer fromstrong acoustic radiation in liquid, which limits their quality factor and increases their frequency noise. This articlepresents an acoustic radiation-free gravimetric biosensor based on a locally resonant surface phononic crystal (SPC)consisting of periodic high aspect ratio electrodes to address the above issue. The acoustic wave generated in the SPCis slower than the sound wave in water, hence it prevents acoustic propagation in the fl uid and results in energycon fi nement near the electrode surface. This energy con fi nement results in a signi fi cant quality factor improvementand reduces frequency noise. The proposed SPC resonator is numerically studied by fi nite element analysis andexperimentally implemented by an electroplating-based fabrication process. Experimental results show that the SPCresonator exhibits an in-liquid quality factor 15 times higher than a conventional Rayleigh wave resonator at a similaroperating frequency. The proposed radiation suppression method using SPC can also be applied in other types ofacoustic wave resonators. Thus, this method can serve as a general technique for boosting the in-liquid quality factorand sensing performance of many acoustic biosensors. Introduction
The rapid and decentralized detection of biomolecules hasbeen increasingly demanded for various applications, such asinfectious disease diagnosis and food safety tests. Thisdemand has been particularly evident during the recentoutbreak of the novel coronavirus (COVID-19), where thethroughput of time-consuming laboratory virus tests sig-ni fi cantly delayed the diagnosis of the disease. In recent years,various techniques have been developed to meet these risingneeds, which can be classi fi ed into four major categories:electrochemical, thermal, optical, and mass-sensitive bio-sensors . Electrochemical biosensors detect signal variationsin potential, current, or conductivity . Thermal biosensorsuse the biochemical-reaction-induced temperature variation as the signal . Optical biosensors detect the adsorption oremission of light at speci fi c wavelengths . Mass-sensitivebiosensors sense surface mass variation due to the binding ofanalytes to bioreceptors . Novel types of biosensors, such asphotothermal and photoelectrochemical – biosensors,have also been increasingly studied. These sensors use lightas the excitation source while implementing detection basedon temperature (photothermal) or electrical current (photo-electrochemical) variations. Among these different types,optical biosensors based on fl uorescent labels or the label-free surface plasmon resonance technique are the mostwidely used for protein and aptamer detection – . How-ever, due to the relatively complex setup of the opticaldetection scheme, optical sensors remain relatively costly andhard to miniaturize.Mass-sensitive biosensors based on acoustic wave resona-tors are competitive alternatives to optical biosensors .They can reach a higher level of integration at lower cost, asthey do not require the use of peripheral equipment, such as © The Author(s) 2021 OpenAccess
Thisarticleislicensedundera CreativeCommons Attribution 4.0 InternationalLicense,which permitsuse,sharing,adaptation,distribution and reproductionin any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate ifchanges were made. The images or other third party material in this article are included in the article ’ s Creative Commons license, unless indicated otherwise in a credit line to the material. Ifmaterial is not included in the article ’ s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtainpermission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Correspondence: Feng Gao ([email protected]) College of Science and Engineering, Hamad Bin Khalifa University, EducationCity, Doha, Qatar Institut FEMTO-ST, CNRS, Universite ́ de Bourgogne-Franche-Comte ́ , Besanc ̧ on,France () :,; () :,; () :,; () :,; xcitation light sources. As their operating principle relies onthe direct detection of an added mass to the surface of thesensor and on the direct measurement of the correspondingelectroacoustic response, the entire detection scheme isintrinsically embedded into the acoustic device, leading tosmall-sized, low-cost sensors that can be easily fi tted into asmall micro fl uidic chamber. In addition, as mass variation isa physical signal that exists in any type ofanalyte – bioreceptor binding reaction, mass-sensitive sensorscan be applied to all types of biomolecule detection, whileother biosensors are limited by their signal type and requirethe speci fi c design of the detection protocol. For example,potentiometric electrochemical biosensors require the ana-lyte binding reaction to generate a potential variation so thatit can be detected. Despite these advantages, acoustic waveresonators generally suffer from strong acoustic radiation inliquid, which decreases their quality factor (Q) and increasesthe signal noise. To maintain operation in liquid, acousticwave resonators for biosensing are usually designed tooperate in shear modes. When the movement of the shearwave is parallel to the liquid – solid interface, the mechanicalmotion transferred to the liquid is reduced compared tovertically polarized waves – . The acoustic radiation, how-ever, cannot be eliminated, as the horizontal friction betweenthe solid and water particles still conveys energy. The com-plete suppression of acoustic radiation can only be achievedby completely preventing wave propagation in the liquid.This scenario requires the velocity of the acoustic wavegenerated in the solid substrate to be lower than the soundvelocity in water, which does not occur in natural materials.The acoustic wave velocity in all piezoelectric substrates isindeed always higher than the sound velocity in waterbecause of the large elastic constants of solid materials .It was, however, reported that one-dimensional surfacephononic crystals (SPCs) made of periodic high aspect ratioelectrode strips can be used to signi fi cantly slow down Rayleigh waves and Lamb waves . In this paper, we exploitthis idea for the realization of SPC resonators operating inliquid. The proposed SPC resonator is theoretically studiedby the fi nite element method (FEM) and experimentallyimplemented by the classical lithography, electroplating, andmolding (Lithographie, Galvanoformung, Abformung(LIGA)) process. By incorporating SPC with interdigitatedtransducers (IDTs), we fi nd that the velocity of Rayleighwaves can be reduced to a value lower than the velocity ofthe sound in water. This successfully stops the propagation ofthe acoustic wave in water and eliminates acoustic radiation.Because of the complete suppression of radiation, the in-liquid Q factor of the resonator is improved by more than 15times compared to a conventional Rayleigh wave resonatorworking in the same frequency range. In addition to theslowing down of the phase velocity, the group velocity is alsofound to be reduced to almost zero, hence suppressingenergy propagation in the horizontal plane. This resultenables the use of zero or a small number of re fl ectors whenconstructing a resonator, therefore signi fi cantly reducing thesensor size and fabrication cost. Moreover, the high aspectratio electrodes constituting the SPC naturally lead to anincrease in the surface-to-volume ratio of the device andhence increase the mass sensitivity. The proposed acousticradiation suppression method can also be applied in othertypes of acoustic waves, which makes it a general techniquethat can signi fi cantly push forward the performance limit ofmany acoustic biosensors. Results
Design of the SPC resonator
Figure 1a shows a top-view diagram of the SPC resonator.IDTs consisting of 30 pairs of high aspect ratio electrodeslocated in the center of the device are deposited on a 128° Y-cut lithium niobate substrate. Nickel is chosen as the elec-trode material because high aspect ratio nickel electrodes can
Reflector ReflectorIDTs (cid:2)
LiNbO Ni LiNbO substrateNi electrodeConfined acoustic energy(red color)BioreceptorsTarget biomoleculesBiomolecules Water ba t ele Fig. 1 Diagram and sensing mechanism of the SPC resonator. a
Diagram of the SPC resonator. IDTs for wave stimulation are in the center, whilere fl ectors for resonance enhancement are on the two sides. b Diagram of the SPC used for biomolecule detection. The speci fi c binding betweenbioreceptors and target biomolecules induces mass loading to the device surface, which causes a resonance frequency shift. Gao et al.
Microsystems & Nanoengineering (2021) 7:8(2021) 7:8
Microsystems & Nanoengineering (2021) 7:8(2021) 7:8
Page 2 of 10 e reliably implemented using the LIGA process. When anelectrical fi eld is applied to the IDTs, mechanical deforma-tions are stimulated in the piezoelectric substrate, eventuallyforming phononic resonance. Both the electrode width andthe electrode spacing are equal to 2.5 µm, resulting in awavelength ( λ ) of 10 µm. The electrode height ( t ele ) is set to~7.5 µm. This value is much larger than the thickness ofconventional SAW resonators, which are usually one or twopercent of the wavelength . Because of the low groupvelocity in the horizontal plane, only 20 re fl ector strips areplaced on the two sides of the IDTs to enhance the reso-nance. This number is much smaller than that used inconventional SAW resonators, which usually require hun-dreds of re fl ector strips . The resonance frequency of thedevice is determined together by the electrode geometry,electrode periodicity, and materials of the electrode andpiezoelectric substrate.Figure 1b shows a diagram illustrating how the SPCresonator can be used for biosensing. To detect biomole-cules, bioreceptors are immobilized on the surface of theelectrodes. These bioreceptors can bind speci fi cally with theirdetection targets. Typical bioreceptors are DNA probes,antigens, and antibodies . They are widely used in thedetection of human immunoglobulins or pathogens, such asviruses and bacteria. Less speci fi c bioreceptors, such asfunctionalized nanoparticles and bioimprinted poly-mers , can be used for the detection of small biomoleculesthat can serve as biomarkers for early disease diagnosis. Oncethe detection targets bind with the bioreceptors, the additional mass attached to the device surface results in atwofold variation in the SPC resonance. First, the attachedmolecules increase the mass of the high aspect ratio elec-trodes, which reduces their mechanical resonance frequency.Second, the additional mass reduces the surface acousticwave velocity due to the classical mass loading effect alsoobserved in surface acoustic wave resonators . As the SPCresonance is a coupling between the mechanical resonance ofthe high aspect ratio electrodes and the surface acousticwave, the combination of the above two effects reduces theSPC resonance frequency. By building an oscillator-basedfrequency readout circuit with the SPC resonator , thesensor response can be converted to real-time digital signals. Slow acoustic wave in the SPC
If the acoustic wave velocity is reduced to a value lowerthan the sound velocity in water, its propagation in water isinhibited. This eliminates the acoustic radiation of the waveand thus enables high Q resonance in liquid. By exploitingthis principle, we used SPC made of periodic high aspectratio electrodes to slow down the surface wave on a lithiumniobate substrate. The SPC induces a hybridization betweenthe Rayleigh-type surface wave and the elastic resonance ofthe high aspect ratio electrodes. This hybridization results inthe occurrence of hybrid modes with velocities directlyconditioned by the geometrical characteristics of the elec-trodes. Figure 2a shows the simulated resonance frequency ofthe localized mode of the SPC resonator in air and water fordifferent electrode heights. It can be seen from the triangular- f = 124.2 MHz A c ou s t i c p r e s u rr e ( P a ) D i s p l a c e m en t ( μ m ) f = 270 MHz A c ou s t i c p r e s u rr e ( P a ) D i s p l a c e m en t ( µ m ) –3 ×10 ×10 P M L P M L W a t e r W a t e r L i N b O L i N b O N i N i N i N i ×10 –5 ba c μ m 50 μ m 3025201510503210–1–2–3–4765432100 6420–2–4–6–8 Fig. 2 Slow-down curve and radiation suppression of the SPC resonator. a
Resonance frequencies of the SPC resonator in relation to theelectrode height. The electrode height needs to be higher than 5.6 µm for the device to maintain resonance in liquid. b Mode shape (enlarged) andacoustic pressure distribution of the 2-µm electrode device. The alternating positive (red color) and negative (blue color) acoustic pressures showstrong acoustic radiation. c Mode shape (enlarged) and acoustic pressure distribution of the 7.6-µm electrode device. The acoustic energy is con fi nedin a small region near the electrodes. Gao et al.
Microsystems & Nanoengineering (2021) 7:8(2021) 7:8
Microsystems & Nanoengineering (2021) 7:8(2021) 7:8
Page 3 of 10 arked blue curve that the resonance frequency decreasescontinually with an increasing electrode height in air. How-ever, this mode cannot be observed in water for electrodeheights lower than 5.6 µm, for which the acoustic wavevelocity ( v ¼ λ ´ f ) on the substrate remains higher than thevelocity in the liquid. For electrode heights larger than5.6 µm, in-liquid resonance occurs, as seen from the circle-marked orange curve. This result is because the corre-sponding wave velocity in the piezoelectric substrate is lowerthan the sound velocity in water, which prevents acousticradiation in the liquid. The corresponding displacement fi elds in the substrate and acoustic pressure distribution inthe liquid were then simulated. The results obtained forelectrode heights of 2 and 7.6 are reported in Fig. 2b and c,respectively. In the case of the 2-µm electrodes, acoustic wavegeneration and radiation were simulated at 270 MHz. Asshown by the alternating positive (red part) and negative(blue part) acoustic pressures, the acoustic wave propagatesin water until it reaches the perfectly matched layer (PML)boundary, where the wave is absorbed by the PML. It shouldbe noted that 270 MHz is the resonance frequency of the 2-µm electrode device in air. This frequency is chosen becausethe device cannot resonate in water. In comparison, theacoustic pressure distribution of the SPC resonator with the7.6-µm electrode operating in water at its resonating fre-quency (124.2 MHz) is shown in Fig. 2c. It can be seen thatthe positive (red part) and negative (blue part) acousticpressures exist only in the region near the electrodes, whichmeans that the acoustic energy is con fi ned in this region andthe acoustic wave is not radiative. Figure 2c also reveals themode shape of the SPC resonators. The displacements of the electrodes are much larger than those observed at the pie-zoelectric substrate surface, which means that most of theelastic energy is stored in the electrodes. The transition fromsurface-con fi ned energy to electrode-surface hybridizedenergy reveals how the wave is transformed from a con-ventional Rayleigh SAW into an interfacial wave, where theelastic energy is distributed between the high aspect ratioelectrode and the near surface of the substrate. The move-ment of the electrodes is vertically polarized, as they move upand down in an oscillation cycle. The displacements of theelectrodes in a full cycle are shown in Fig. S1, SupplementaryInformation. Although the vertical movement transfers moremechanical energy to water compared to the horizontalmovement in shear waves, resonance in water still occurs.This behavior indirectly suggests the effectiveness of theacoustic radiation suppression technique.Dispersion curves providing information on the groupand the phase velocities constitute a very importantdesign tool for acoustic devices. Figure 3a shows twoindividual impedance curves of an SPC resonator with a7.6-µm thick electrode when k x equals zero and 0 : π = λ .By varying the wavevector k x in the fi rst Brillouin zone,the dispersion curves (Fig. 3b) of these resonance modesare obtained. As an example, the two impedance curves inFig. 3a yield the dispersion curve points marked by twovertical dashed black lines in Fig. 3b. Because of theelectrical periodicity of the electrodes, the dispersioncurves fold at k x ¼ π = λ , which means that its folded upperstrand is the dispersion curve of π = λ 80 90 100 110 120 130 140 150Frequency (MHz) I m pedan c e ( Ω ) k x (m/s) π / λ π / λ π / λ π / λ π / λ Fig. 3 Dispersion diagram of the SPC resonator. Impedance curves ( a ) and dispersion diagram ( b ) of the SPC resonator with the 7.6-µm electrode.The dispersion diagram is synthesized by the impedance curves at different wave vectors. The color represents the impedance magnitude. The red-and yellow-dashed lines are the dispersion curves of the vertically polarized and shear horizontal waves, respectively Gao et al. Microsystems & Nanoengineering (2021) 7:8(2021) 7:8 Microsystems & Nanoengineering (2021) 7:8(2021) 7:8 Page 4 of 10 ertically polarized wave, while the yellow-dashed line isthe dispersion curve of the shear horizontal wave. The reddot in Fig. 3b thus marks the base mode of the verticalpolarized phononic wave, where k x ¼ π = λ . It can be seenfrom the slope of the dispersion curve that the corre-sponding group velocity of this point is almost zero(slightly negative). As the group velocity represents thetransmission of the wave energy, this signi fi cantly reducedgroup velocity along the x direction means that very littleenergy propagates horizontally. Because of this, the use ofa large number of re fl ectors, as in conventional SAWresonators, is not necessary. This result enables a sig-ni fi cant reduction in the device size and saves cost. Electrode pro fi le and impedance of the SPC resonator A scanning electron microscopy (SEM) image of thefabricated device is shown in Fig. 4a. Both positive pho-toresist (pPR) and negative photoresist (nPR) were used aselectroplating molds to produce different electrode pro- fi les and investigate the corresponding impact on theresonator performance. The cross-sectional SEM imagesof the electrodes fabricated with the pPR and nPR areshown in Fig. 4b and c, respectively. The cross-sectionwas obtained by cutting with a focused ion beam, duringwhich platinum was deposited to protect the adjacent topsurface of the electrode. The platinum protection layerand the nickel electrode are marked by blue and redoverlays in one of the electrodes of each image, respec-tively. The pro fi le of the electrode fabricated by nPR is atrapezoid with a slightly shrunken bottom. In comparison,a reverted trapezoidal pro fi le is obtained with pPR.The impedance curves of the resonators fabricated withthese different photoresists are shown in Fig. 4d. The twoSPC resonators have similar impedance characteristics.The difference in resonance frequency is caused by boththe electrode pro fi le and electrode thickness differences ofthe devices. For comparison, the response of a conven-tional Rayleigh wave resonator is also shown. This reso-nator is fabricated on the same 128° Y-cut lithium niobatewafer with a one-step lift-off process. The wavelength andelectrode thickness are 30 µm and 200 nm, respectively.Its layout design is directly scaled up three times from theSPC resonator for a fair comparison. The impedancecurve of the conventional Rayleigh wave resonator ismuch fl atter than that of the SPC resonators, with a Qfactor in liquid that only reaches 3.4 due to its acousticradiation in liquid. In contrast, the Q factors of the SPCresonators in liquid are 50.4 and 44.5 for the pPR and nPRprocesses, respectively. The sharp resonance peaks meanthat the SPC resonators have better frequency resolutionwhile detecting the external mass loads. In practicalapplications, oscillators incorporating sensor resonatorsare usually built to produce frequency signals that can becounted by readout circuits. The phase noise of the oscillator is inversely proportional to the Q factor of theresonator and can be estimated by Leeson ’ s equation .Resonators with high quality factors will lead to lowerfrequency noise in the fi nal sensor oscillator. Mass sensitivity of the SPC resonator The overall performance of a gravimetric biosensor can bedescribed by its limit of detection (LOD), which is obtainedby LOD ¼ Sn f ð Þ where S and n f are the mass sensitivity and frequencynoise of the sensor, respectively. The frequency noise ofthe sensor is evaluated by its Q factor, as discussed in theprevious section. To evaluate the overall sensing perfor-mance, the mass sensitivity of the SPC biosensor alsoneeds to be characterized. The variation in the nickelelectrode height across the wafer is used here to calculatethe mass sensitivity, as a change in electrode height isequivalent to the variation in loaded mass on top of theelectrodes. Nevertheless, it should be noted that thissensitivity is an underestimate of the actual masssensitivity expected from an actual biosensing processwhere the mass loading would be homogeneous on theentire device surface rather than localized on the top ofthe electrode. The resonance frequencies of the SPCresonators operating in air and in water are shown in Fig.5a and b, respectively. The measurement data correspondto the triangular data points in the two fi gures. There aremore data points of the resonators fabricated with pPRthan for the devices fabricated with nPR because moreunits of the corresponding layout cells were arranged inthe positive fabrication mask than in the negativefabrication mask. Linear regression was used to fi nd therelations between the electrode height and resonancefrequency, which are shown by solid lines in the two fi gures. Using the actual electrode pro fi le shown in Fig.4b, c, the resonance frequencies of the devices were alsosimulated to validate the effectiveness of the FEM model.The corresponding simulated resonance frequencies in airand in water are shown by the dashed lines in Fig. 5a andb, respectively. Both of the simulated resonance frequen-cies agree well with the experimental data. The masssensitivity of the resonators to the loading on the top ofthe electrode is obtained from the slope of the fi tted curveby the following equation: S ¼ ρ Ni ´ dfdt ð Þ where ρ Ni and t are the density and height of the nickelelectrode, respectively. The mass sensitivity is a negativenumber, as the mass loading results in a decrease in the Gao et al. Microsystems & Nanoengineering (2021) 7:8(2021) 7:8 Microsystems & Nanoengineering (2021) 7:8(2021) 7:8 Page 5 of 10 esonance frequency. The mass sensitivities of theresonators in air ( S a ) and in water ( S w ) obtained fromthe linearly fi tted curve are shown in Table 1. Theircorresponding coef fi cients of determination ( R ) are alsolisted. All the R values are greater than 0.94, whichindicates a good match between the fi tted line andexperimental data. The data in Table 1 also reveal that thepPR SPC resonators have a mass sensitivity 30% higherthan the nPR SPC resonators, whether operating in air orwater. This result is mainly because the trapezoidal pro fi leof the nPR process makes the mass-loaded area (the topsurface) smaller than that obtained from the invertedtrapezoidal pro fi le of the pPR process. The masssensitivity of the resonator in air is only slightly higherthan the mass sensitivity of the resonator in water, whichmeans that the water loading has a negligible impact onthe mass sensitivity of the resonator. The simulated masssensitivities in air ( S a sim ) and in water ( S w sim ) are alsolisted in Table 1. The sensitivity modeling for the pPR SPC resonator agrees well with the experimental results,with discrepancies of 6.2 and 7.6% in air and water,respectively. The modeling for the nPR SPC resonators isless accurate, with simulated experiment discrepancies of26.7 and 21.7% in air and water, respectively. The lessaccurate modeling of the nPR SPC resonator is possiblydue to the discrepancy between the more distortedexperimental electrode pro fi le (Fig. 4c) and the simpli fi edtrapezoidal electrode pro fi le in the simulation. Never-theless, the FEM modeling results generally agree wellwith the experimental data.The mass sensitivity under homogeneous loading con-ditions, or full coverage mass sensitivity, was thenobtained by the same FEM simulation model. A polymerlayer (PMMA) covering the entire device surface wasadded as the mass loading layer. The device frequencyresponse as a function of added mass is simulated bychanging the PMMA layer density. The simulated unitcells for the pPR and nPR SPC resonators operating at Frequency (MHz) I m pedan c e ( (cid:6) ) Postive PR profile, t = 6.6 (cid:7) m HV5.00 kV curr43 pA detETD WD4.2 mm HV5.00 kV curr43 pA detTLD WD4.1 mmmag tilt0° 300 (cid:7) mInstitut FEMTO-ST200 ×HV5.00 kV curr43 pA detTLD WD4.1 mm mag tilt52° 5 (cid:7) mInstitut FEMTO-ST . mm ( cs ) . (cid:7) m ( cs ) (cid:7) mInstitut FEMTO-ST12 000 × Negative PR profile, t = 7.5 (cid:7) mConventional SAW (cid:2)(cid:8) = 30 (cid:7) m ac db Fig. 4 Device SEM images and impedance curves of the SPC resonator. a Top-view SEM image of the SPC resonator. Cross-sectional imagesshow the inverted trapezoid and trapezoid pro fi les of the high aspect ratio electrode fabricated with positive ( b ) and negative ( c ) photoresists,respectively. The red and blue overlays mark the nickel electrode and the titanium protection layer deposited during the cutting by an FIB,respectively. d Impedance magnitude curves of the SPC resonators and conventional Rayleigh SAW resonators in liquid. Gao et al. Microsystems & Nanoengineering (2021) 7:8(2021) 7:8 Microsystems & Nanoengineering (2021) 7:8(2021) 7:8 Page 6 of 10 20 MHz are shown in Fig. S3a, b, respectively. As listedin the last column of Table 1, the simulated full coveragemass sensitivity in water ( S w full ) is found to be − − ) for the pPR and nPR SPC resonators,respectively. The good agreement between the full cov-erage mass sensitivities of the pPR and nPR SPC resona-tors can be accounted for by their similar surface areasunder full coverage conditions. This result also revealsthat the electrode pro fi le does not affect the mass sensi-tivity of the SPC when the mass loading is homogenouson the entire surface. As expected, the mass sensitivity forfull surface coverage loading is signi fi cantly higher thanthe mass sensitivity for loading only on the top of theelectrode because of the much larger loading area. Forcomparison, the mass sensitivity of the widely used 36° YLiTaO -based SH-SAW sensing device, here working at120 MHz, is also simulated. The corresponding simula-tion unit cell is shown in Fig. S3c. The result shows a masssensitivity of 10.02 Hz/(ng/cm ), which agrees well withthe experimental results reported by Barie et al. . Their36° Y LiTaO SH-SAW operating at 380 MHz showed a25 kHz response to the adsorption of a bovine serumalbumin monolayer that has a surface mass of 200 ng/cm .Considering that the mass sensitivity is proportional to the square of the operating frequency, the equivalent masssensitivity at 120 MHz is 12.46 Hz/(ng/cm ), which issimilar to our simulated SH-SAW sensitivity. This com-parison shows that the proposed SPC biosensor canexhibit a mass sensitivity close to six times higher thanthat of conventional SH-SAW biosensors, which is due tothe higher surface-to-volume ratio of the high aspect ratioelectrodes. Discussion As mentioned in the “ Introduction ” section, the usualstrategy for in-liquid acoustic sensing relies on the use ofshear-polarized acoustic modes. The shear polarizationindeed minimizes mechanical motion transfer to water.However, it does not fully prevent acoustic radiation inwater: the wave velocity in the solid substrate remainshigher than the sound velocity in liquid and coupling withthe vertically polarized mode generally occurs. Thisradiation loss is one of the major limitations of thesedevices in achieving a high Q factor for in-liquid sensing.In contrast, the SPC resonator proposed in this workrelies on the complete suppression of acoustic radiation inwater; this suppression is accomplished by a signi fi cantslowing down of acoustic wave propagation in the solid R e s onan c e f r equen cy ( M H z ) pPR measurement in waterpPR simulation in waternPR measurement in waternPR simulation in water a b R e s onan c e f r equen cy ( M H z ) pPR measurement in airpPR simulation in airLinear fit of nPR data in airnPR simulation in airnPR measurement in airLinear fit of pPR data in air Linear fit of nPR data in waterLinear fit of pPR data in water Fig. 5 Mass sensitivities of the SPC resonators. Resonance frequencies of the two types of SPC resonators fabricated with different photoresistsoperating in air ( a ) and in water ( b ). The slope of the linear regression curve is proportional to the mass sensitivity of the resonators Table 1 Mass sensitivity of the SPC resonators. Pro fi le S a a R of S a S w R of S w S a sim S w sim S w full pPR − − − − − − − − − − a Unit of mass sensitivity: Hz/(ng/cm ). Gao et al. Microsystems & Nanoengineering (2021) 7:8(2021) 7:8 Microsystems & Nanoengineering (2021) 7:8(2021) 7:8 Page 7 of 10 ubstrate due to the hybridization mechanism betweenlocalized modes in the electrodes and the surface wave.Due to the complete suppression of acoustic radiation, theenergy leakage can be theoretically fully stopped, henceenabling high Q factor in-liquid sensing. In the presentwork, the Q factor of the SPC resonators is mostly limitedby the electroplating process. The sidewall surfaceroughness and the overall process inhomogeneity result instructural defects breaking the periodicity of the IDTs,which limits further improvement in the quality factor.Nevertheless, this technological shortcoming can beovercome, notably by using alternative thick- fi lmdeposition techniques, such as thermal evaporation andchemical vapor deposition. Another limitation of the SPCbiosensor compared to conventional SAW devices is thatit involves a more complex fabrication process. Althoughit is a trade-off to achieve better performance, the addedcost will not be signi fi cant in a complete sensing systemthat includes other components, such as micro fl uidicchannels and readout circuits.The proposed method for acoustic radiation suppres-sion is applied here to devices initially operating onelliptically polarized Rayleigh waves. This method can bemore generally transposed to any IDT-based acousticwave resonator, including SH-SAW or Love-waveacoustic resonators. The combination of shear modepolarization and acoustic radiation suppression has thepotential to signi fi cantly advance the performanceboundary of acoustic wave biosensors. Conclusion In summary, we proposed a gravimetric biosensor basedon an SPC resonator that can achieve acoustic radiation-free operation in liquid. The SPC induced a hybridizationbetween a Rayleigh-type surface wave and the elasticresonance of the high aspect ratio electrodes, which led tostrong con fi nement of the elastic energy at theelectrode – substrate interface. The acoustic wave velocitywas then reduced to be lower than the speed of sound inwater. This mechanism resulted in the complete sup-pression of acoustic radiation, leading to high Q reso-nance in liquid. This device principle was validated byFEM analysis and experimentally implemented throughthe fabrication of SPC resonators using the LIGA process.Q factors on the order of 50 and a mass sensitivity of12.92 Hz/(ng/cm ) were obtained by the pPR process. Acomparison with a Rayleigh SAW resonator revealed thatthe SPC could improve the Q factor of the resonator by 15times. The effect was demonstrated here for resonatorsoperating on vertically polarized waves, and it could alsobe combined with existing shear mode acoustic waveresonators. This proposed biosensor has the potential tosigni fi cantly advance the performance boundary of cur-rent acoustic wave biosensors. Materials and methods Numerical analysis Numerical analysis based on the FEM was used to studythe behavior of the SPC resonator. The analysis was per-formed using Comsol multiphysics software. The FEMmodel couples multiple physics, including solid mechan-ics, pressure acoustics, electrostatics, piezoelectric effects,and acoustic structure interactions. As a 3D numericalanalysis is computationally heavy, a simpli fi ed unit cell ofthe resonator (Fig. S2, Supplementary Information) wasadopted to reduce the computation time. The width of theunit cell was set to one wavelength with continuity-typeperiodic conditions for the mechanical, electrical, andacoustic components applied to the two sides. This setupwas equivalent to continuously repeating the unit cell andthus representing an in fi nite number of IDT pairs. Theactual simulation model was in 3D with a depth of 0.1 λ .Continuity-type periodic conditions were also applied tothe front and back of the model to equivalently extend thedepth to in fi nity, which effectively made the acousticaperture in fi nitely wide. The thickness of the substrate ( t sub ) was set to 5 λ , with the bottom 1 λ set as the PML. ThePML absorbed all the waves leaking into it, equivalentlymaking the substrate in fi nitely thick. A fi xed boundarycondition was applied to the bottom of the substrate,which represented the fi xation on the accommodatingpackage. The depth of the water ( d water ) was also set to 5 λ ,with the top 1 λ set as the PML to equivalently extend thewater depth to in fi nity. A free boundary condition wasapplied to the top of the water. The height and pro fi le ofthe electrodes were used as sweeping parameters in theanalysis to study the behavior of the SPC resonator. Theideal periodic conditions and PMLs ruled out the perfor-mance impact from parameters of the substrate thickness,acoustic aperture width, and IDT designs, which made theanalysis results only relevant to the pro fi le and dimensionsof the high aspect ratio electrodes. An alternative electricalpotential (1-V amplitude) was applied to the two metallicelectrodes as electrical stimulation.Fundamental fi eld parameters, such as the electrical fi elds, particle displacements, and acoustic pressures, wereobtained from numerical analysis. Other needed infor-mation, including electrical impedance and mode shape,was further derived from those fi eld parameters. Toobtain the dispersion diagram of localized modes in SPC,Floquet periodic conditions were applied to the left andright sides of the model so that impedance curves underdifferent wave vectors ( k x ) were computed. The variationin k x covered the fi rst Brillouin zone, i.e., from 0 to π = λ .This equivalently changed the periodic conditions on thetwo sides from continuity type ( k x ¼ 0) to antiperiodictype ( k x ¼ π = λ ). The fi nal dispersion diagram wasobtained by a color map representation of the impedancemagnitude in the k x (cid:2) f plane. Gao et al. Microsystems & Nanoengineering (2021) 7:8(2021) 7:8 Microsystems & Nanoengineering (2021) 7:8(2021) 7:8 Page 8 of 10 abrication process The device was fabricated on a 4-inch 128° Y-cutlithium niobate wafer. The wafer was initially cleaned witha piranha solution and deionized water to removepotential contaminants. Following cleaning, 20-nm tita-nium and 100-nm copper thin fi lms were deposited on thewafer surface. The copper layer was used as a seed layerfor subsequently electroplating nickel, while the titaniumlayer was used to promote adhesion between the copperlayer and lithium niobate substrate. Subsequently, pho-tolithography was used to create the mold for electro-plating. A thick pPR, AZ9260, and a thick nPR, AZ nLOF2070, were used to create different electrode pro fi les. Twophotolithography masks with opposite polarity weredesigned correspondingly. After the mold was created,7 – fi ne the electrically isolated IDT fi ngers.Subsequently, 200 nm of silicon dioxide was deposited onthe sensor surface using chemical vapor deposition. Thislayer was used as a passivation layer and could also pro-vide an interface for bioreceptor immobilization. Anotherphotolithography step was then performed to create thephotoresist etching mask for the area excluding the padsthat will be used for external electrical connection. Theremoval of silicon dioxide covering the contact pads wascompleted by another RIE process. The complete process fl ow is shown in Fig. S4. Sensor characterization The variation in the nickel fi lm thickness across thewafer was exploited for the calculation of the mass sen-sitivity of the sensor because the small height difference ofthe nickel electrodes across the substrate caused a naturalmass loading variation for the sensor, which changed itsresonance frequency. The thickness of the nickel layerswas measured by a pro fi lometer.The Q factor of the SPC resonator was used to evaluatethe frequency stability (frequency noise) of the sensor. Itwas calculated from its impedance by the followingequation: Q ¼ f ´ d ϕ df ð Þ where ƒ and ϕ are the operating frequency and phase ofthe electrical impedance of the device, respectively. Themeasurement of the impedance was completed by a vector network analyzer (VNA). Two RF probes wereused to connect the device to the VNA. To obtain theperformance data of the sensor in both air and water, itsimpedance was measured under both conditions. Acknowledgements This work is funded by NPRP grant no. NPRP10-0201-170315 from the QatarNational Research Fund (a member of Qatar Foundation). This work is alsosupported by the EIPHI Graduate School (contract “ ANR-17-EURE-0002 ” ) andthe French RENATECH network with its FEMTO-ST technological facility. The fi ndings herein re fl ect this work and are solely the responsibility of the authors. Author contributions F.G., A.B., and A.K. proposed the idea. F.G. and S.B. designed the fabricationprocess. F.G. and L.R. fabricated the device. F.G. performed the device test. F.G.,A.K., and S.B. drafted the manuscript. Con fl ict of interest The authors declare that they have no con fl ict of interest. Supplementary information accompanies this paper at https://doi.org/10.1038/s41378-020-00236-9. Received: 7 July 2020 Revised: 17 November 2020 Accepted: 13 December2020 References 1. Gaudin, V. Advances in biosensor development for the screening of antibioticresidues in food products of animal origin – a comprehensive review. Biosens.Bioelectron. , 363 – 377 (2017).2. Metkar, S. K. & Girigoswami, K. Diagnostic biosensors in medicine – a review. Biocatal. Agric. Biotechnol. , 271 – 283 (2019).3. Khan, M., Hasan, M., Hossain, S., Ahommed, M. & Daizy, M. Ultrasensitivedetection of pathogenic viruses with electrochemical biosensor: state of theart. Biosens. Bioelectron. , 112431 (2020).4. Casadio, S. et al. Development of a novel fl exible polymer-based biosensorplatform for the thermal detection of noradrenaline in aqueous solutions. Chem. Eng. J. , 459 – 468 (2017).5. Khansili, N., Rattu, G. & Krishna, P. M. Label-free optical biosensors for food andbiological sensor applications. Sens. Actuators B: Chem. , 35 – 49 (2018).6. Zhang, Y., Luo, J., Flewitt, A. J., Cai, Z. & Zhao, X. Film bulk acoustic resonators(FBARs) as biosensors: a review. Biosens. Bioelectron. , 1 – 15 (2018).7. Huang, L., Chen, J., Yu, Z. & Tang, D. Self-powered temperature sensor withseebeck effect transduction for photothermal – thermoelectric coupledimmunoassay. Anal. Chem. , 2809 – Anal. Chem. , 1470 – Trends Anal. Chem. , 115814 (2020).10. Shu, J. & Tang, D. Recent advances in photoelectrochemical sensing: fromengineered photoactive materials to sensing devices and detection modes. Anal. Chem. , 363 – 377 (2019).11. Fan, X. et al. Sensitive optical biosensors for unlabeled targets: a review. Anal.Chim. Acta , 8 – 26 (2008).12. Chen, T., Wang, X., Alizadeh, M. H. & Reinhard, B. M. Monitoring transientnanoparticle interactions with liposome-con fi ned plasmonic transducers. Microsyst. Nanoeng. , 16086 (2017).13. Oh, S. Y. et al. Development of gold nanoparticle-aptamer-based LSPR sensingchips for the rapid detection of Salmonella typhimurium in pork meat. Sci. Rep. , 1 – 10 (2017).14. Voiculescu, I. & Nordin, A. N. Acoustic wave based MEMS devices for bio-sensing applications. Biosens. Bioelectron. , 1 – fi lm for high-sensitivity detectionof a label-free endotoxin. Microsyst. Nanoeng. , 1 – 11 (2020). Gao et al. Microsystems & Nanoengineering (2021) 7:8(2021) 7:8 Microsystems & Nanoengineering (2021) 7:8(2021) 7:8 Page 9 of 10 6. Song, S. et al. Shear mode bulk acoustic resonator based on inclined c-AxisAlN fi lm for monitoring of human hemostatic parameters. Micromachines ,501 (2018).17. Ferreira, G. N., Da-Silva, A.-C. & Tomé, B. Acoustic wave biosensors: physicalmodels and biological applications of quartz crystal microbalance. TrendsBiotechnol. , 689 – 697 (2009).18. Länge, K., Rapp, B. E. & Rapp, M. Surface acoustic wave biosensors: a review. Anal. Bioanal. Chem. , 1509 – Multilayer Integrated Film Bulk Acoustic Resonators (Springer Science & Business Media, 2012).20. Laude, V., Robert, L., Daniau, W., Khelif, A. & Ballandras, S. Surface acoustic wavetrapping in a periodic array of mechanical resonators. Appl. Phys. Lett. ,083515 (2006).21. Gao, F. et al. Towards acoustic radiation free lamb wave resonators forhigh-resolution gravimetric biosensing. IEEE Sens J , 2725 – Surface Acoustic Wave Devices in Telecommunications (Springer, 2000).23. Morgan, D. Surface Acoustic Wave Filters: With Applications toElectronic Communications and Signal Processing (Academic Press,2010). 24. Vogiazi, V. et al. A comprehensive review: development of electrochemicalbiosensors for detection of cyanotoxins in freshwater. ACS Sens. , 1151 – fi cation of proteins usingnanoparticle – fl uorescent polymer ‘ chemical nose ’ sensors. Nat. Nanotechnol. ,318 (2007).26. Whitcombe, M. J. et al. The rational development of molecularly imprintedpolymer-based sensors for protein detection. Chem. Soc. Rev. , 1547 – ACS Sens. , 1171 – fi cation. Sens.Actuators B , 21 – 27 (2019).29. Gao, F., Boussaid, F., Xuan, W., Tsui, C.-Y. & Bermak, A. Dual transduction surfaceacoustic wave gas sensor for VOC discrimination. IEEE Electron. Device Lett. ,1920 – Proc. IEEE ,329 – 330 (1966).31. Barie, N. & Rapp, M. Covalent bound sensing layers on surface acoustic wave(SAW) biosensors. Biosens. Bioelectron. , 979 – 987 (2001). Gao et al. Microsystems & Nanoengineering (2021) 7:8(2021) 7:8