sensors Article

Design and Validation of a 150 MHz HFFQCM Sensor for Bio-Sensing Applications Román Fernández 1, *, Pablo García 1 , María García 1 , José V. García 1 Antonio Arnau 2 1 2

*

ID

, Yolanda Jiménez 2 and

Advanced Wave Sensors S. L., Algepser 24, 46988 Paterna, Spain; [email protected] (P.G.); [email protected] (M.G.); [email protected] (J.V.G.) Centro de Investigación e Innovación en Bioingeniería, Universitat Politècnica de València, Camí de Vera S/N, 46022 Valencia, Spain; [email protected] (Y.J.); [email protected] (A.A.) Correspondence: [email protected]; Tel.: +34-675-401-871

Received: 21 July 2017; Accepted: 5 September 2017; Published: 8 September 2017

Abstract: Acoustic wave resonators have become suitable devices for a broad range of sensing applications due to their sensitivity, low cost, and integration capability, which are all factors that meet the requirements for the resonators to be used as sensing elements for portable point of care (PoC) platforms. In this work, the design, characterization, and validation of a 150 MHz high fundamental frequency quartz crystal microbalance (HFF-QCM) sensor for bio-sensing applications are introduced. Finite element method (FEM) simulations of the proposed design are in good agreement with the electrical characterization of the manufactured resonators. The sensor is also validated for bio-sensing applications. For this purpose, a specific sensor cell was designed and manufactured that addresses the critical requirements associated with this type of sensor and application. Due to the small sensing area and the sensor’s fragility, these requirements include a low-volume flow chamber in the nanoliter range, and a system approach that provides the appropriate pressure control for assuring liquid confinement while maintaining the integrity of the sensor with a good base line stability and easy sensor replacement. The sensor characteristics make it suitable for consideration as the elemental part of a sensor matrix in a multichannel platform for point of care applications. Keywords: HFF-QCM (high fundamental frequency quartz crystal microbalance); finite element method (FEM); flow cell; biosensor; PoC (point of care); MQCM (monolithic quartz crystal microbalance)

1. Introduction Point of care (PoC) devices are very promising diagnostic tools that can operate in different environments, from clinical labs to small-size hospitals or even at patient’s house, due to their unique characteristics. PoC platforms are designed to provide fast assay results, be portable and compact, and operate in a simple and automatic way. Such systems can be operated by non-specialists and provide valuable information about a large variety of diseases, e.g., cancer, cardiovascular, and infectious diseases [1]. The widespread application of PoC devices in common clinic practice will increase the prospects for more efficient patient monitoring during treatment, or even make available early diagnostic mechanisms for disease prevention in high-risk populations. PoC platforms are typically based on genomics and/or proteomics for the detection stage, and require low cost, fast, sensitive, reliable and small-size transduction technologies in order to be successfully implemented [2]. Acoustic wave sensors are perfectly suited to fulfill these requirements. Health care, food security, and environmental monitoring fields are all market applications in which acoustic sensing has shown

Sensors 2017, 17, 2057; doi:10.3390/s17092057

www.mdpi.com/journal/sensors

Sensors 2017, 17, 2057

2 of 13

its potential. Acoustic sensing has taken advantage of the progress made in the last decades in piezoelectric resonators for radio frequency (RF) telecommunication technologies. The piezoelectric elements used in radars, cellular phones, or electronic watches for the implementation of filters, oscillators, etc., have been applied to these types of sensors. Among the different approaches for the design of acoustic wave sensors, QCM (quartz crystal microbalance) is probably the most used sensing technology. Its operation is based on the so-called gravimetric technique [3], which relates resonance frequency shifts of the resonator to mass changes on the sensor surface. It has opened a great deal of applications in bio-chemical sensing: immunoassays, protein adsorption, DNA hybridization, etc. [4,5]. The main advantages of the QCM technique include: direct label-free detection, real-time non-invasive approach, low cost, and the ability to detect viscoelastic and conformational changes of the physical layer close to the sensor surface [6]. In the QCM technique, selectivity is provided by the appropriate sensor surface functionalization. Although QCM techniques have improved in robustness and reliability, and allow measuring molecular interactions in real time, the improvement of the sensitivity and the possibility of multi-analysis and integration capabilities can be considered remaining challenges. These characteristics are highly demanded in the field of point of care diagnosis. MQCM (Monolithic QCM) devices [7] have emerged as very promising acoustic wave sensors that could overcome the current limitations of the QCM technique in bio-sensing applications. They consist of the integration of many QCM resonators in the same quartz substrate in an array configuration. Through the years, several MQCMs have been developed for advanced sensing applications, including electronic nose, electronic tongue, and biosensor arrays [8–14]. The use of MQCM technology leads to relevant benefits including multi-analysis capabilities, lower cost per sensor unit, less sample/reagent consumption, faster sensing response, and shorter detection assay time. All of these advantages arise from having a large number of independent sensing spots in a smaller surface [7,13]. If high sensitivity and multi-analysis detection are simultaneously required, a feasible solution goes through a system based on a monolithic array made up of high fundamental frequency QCM (HFF-QCM) devices. The HFF-QCM principle of operation [15,16] is based on the reduction of the plate thickness of a classical QCM, which increases its resonance frequency. According to Sauerbrey´s relation [3], mass sensitivity is directly proportional to the square of its resonance frequency. HFF-QCM devices have been applied both in gas [17–19] and liquid media [20]. Improvements in the LoD (limit of detection) of around two orders of magnitude with respect to classical QCM have been reported using this acoustic sensor technology in immunoassays [21]. Furthermore, the sensing surface of HFF-QCM devices must be reduced with increasing frequency [22], leading thus to devices with small footprints that are easier to be integrated in an array. The main drawbacks of HFF-QCM concept are related to lower mechanical stability and difficult handling [23]. A suitable “inverted mesa” design [20,24,25] can be used to overcome these issues by thinning only a small area of the sensor, while the surroundings of the device provide a supporting frame with mechanical strength. Other acoustic technologies e.g., surface acoustic wave (SAW) or thin film bulk acoustic resonators (FBAR), have the capability to be designed in an array configuration as well. From these technologies, SAW devices and in particular Love-mode acoustic sensors are typically used in liquid medium applications because of their expected higher sensitivity [26]. However, a similar limit of detection has been obtained using HFF-QCM sensors for the same application and under the same experimental and instrumental conditions [27]. The size of a Love wave sensor is bigger than the corresponding HFF-QCM device for the same substrate material and the same operation frequency. Furthermore, the fabrication of Love wave sensors is more complicated and expensive mainly due to the more complex geometry and the necessity of implementing inter-digital transducers (IDTs). For the reasons exposed above, HFF-QCM sensors based on inverted mesa technology have been selected to constitute the “basic building unit” of a novel acoustic wave array. Obviously, the complexity of the design of HFF-QCM devices increases when a number of HFF-QCM sensors

Sensors 2017, 17, 2057

3 of 13

must be integrated together on the same substrate, and in a reduced area, in an array configuration. In this case, the sensor’s basic design requirements include: small size, high quality factor (considering the high operation frequency) and to avoid or reduce as much as possible the presence of inharmonic modes in the vicinity of the fundamental mode. On the other hand, a minimum thickness of the electrodes is necessary in order to assure good electrical conductivity and produce the ‘energy trapping’ effect [28,29]. This effect confines the thickness shear vibrations under the electrode area, thus improving the quality factor of the sensor and reducing the interferences among resonators in an array configuration [30,31]. Moreover, the combination of an excessive thick electrode together with a large electrode area could lead to the presence of spurious overtones [22]. Usually, the ‘plate-back’ equation is used in common quartz crystal resonator design (5–10 MHz) to properly determine the optimal electrode geometry and thus avoid the presence of these unwanted inharmonic modes [22]. This equation provides very small thicknesses when used for high frequency sensor design (150 MHz), which would lead to low conductivity issues in the sensor electrodes. For example, for a square 150 MHz resonator, with a lateral dimension of 0.55 mm, a maximum thickness of 0.75 nm of gold electrode would be required to avoid the presence of spurious modes according to the ‘plate-back’ equation. Thus, it is necessary to find a compromise solution to obtain a good electrical conductivity, a proper ‘energy trapping’ effect, and a sufficient separation between the fundamental thickness shear mode and the first spurious mode in order to minimize its interference during the operation of the sensor in liquid medium. In this sense, Equation (1) proposed by Sheahan [32] is suitable to estimate the resonance frequency of the different vibration modes of the device described in this work: √ q ω1mn = (π/ ρ) C66 /ly2 + m2 (C11 + π 2 C66 /12)/lx2 + n2 C55 /lz2 , (1) where ω1mn is the angular frequency of the different m and n modes, C66 , C55 , C11 are the corresponding AT-cut quartz elastic constants, ρ is the quartz density, ly the dimension in the thickness direction, and lx and lz are the lateral dimensions of the resonator. The presence of inharmonic modes in the proximity of the fundamental mode can be minimized by reducing these lateral dimensions. Simulation tools based on numerical methods are commonly used to support the design stage because of its complexity. Additionally, HFF-QCM sensors are usually employed under flow conditions. In such cases, a suitable flow measurement cell is a basic requirement. Different designs have been proposed in the literature for gas [33] and liquid [23] flows. In the concrete case of bio-sensing applications, most of the measurements are carried out in liquid medium. Then, the cell must confine the liquid over the sensor’s surface while isolating the electrical connections [23]. The key design prerequisites are: low distortion of the original response of the sensor, and reasonable baseline stability. In the case of HFF-QCM devices and due to their special fragility and small sensing area, a very small flow chamber volume and a design approach that assures sensor mechanical integrity and easy replacement are critical additional requirements. Finally, the biocompatibility and chemical resistance of the materials of the flow module that are in contact with the liquid medium must be taken into account. As a first step to developing a MQCM device, a single HFF-QCM sensor operating a 150 MHz has been designed and manufactured. A flow cell module has also been designed and manufactured to test the sensor in bio-sensing applications. Finally, the sensor and the cell have been validated through a simple molecular interaction experiment using Protein A and IgG antibody. 2. Materials and Methods 2.1. Sensor Numerical Model A numerical model of the sensor based on the finite element method (FEM) has been developed to simulate and optimize the electromechanical response of the acoustic wave microsensor prior to fabrication. A commercial software package (ANSYS, Canonsburg, PA, USA) has been used. In order to consider the piezoelectric nature of the AT-cut quartz wafer used as a substrate of the sensor,

2.1. Sensor Numerical Model A numerical model of the sensor based on the finite element method (FEM) has been developed to simulate and optimize the electromechanical response of the acoustic wave microsensor prior to fabrication. A2057 commercial software package (ANSYS, Canonsburg, PA, USA) has been used. In order Sensors 2017, 17, 4 of 13 to consider the piezoelectric nature of the AT-cut quartz wafer used as a substrate of the sensor, a multi-physics model has been built. Electric field and mechanical stress/strain have been calculated a multi-physics model of hasfreedom been built. Electric field and nodes mechanical stress/strain have been calculated in the model. Degrees (DOF) of the model include temperature, electric potential, in the model. Degrees of freedom (DOF) of the model nodes include temperature, electric potential, and 3D displacement. Values for the physical properties of AT-cut quartz, which are described and 3D displacement. forAppendix the physical properties of AT-cut quartz,starting which are elsewhere [22,31,34], areValues listed in A. These properties are obtained fromdescribed those of elsewhere [22,31,34], are listed in Appendix A. These properties are obtained starting from those of the the ‘mother’ quartz through rotation transformation formulas [35]. ‘mother’ quartz through rotation transformation [35]. A sensitivity analysis has been carried out toformulas study the performance of different finite elements A sensitivity analysis has been carried out to study the performance ofindifferent finite direction elements shapes and mesh configurations. A minimum number of elements is needed the thickness shapes and mesh configurations. A minimum number of elements is needed in the thickness direction to properly capture the characteristics of the resonance; 10 element layers have been found that model to properly capture the characteristics of the resonance; 10 element layers have been found the thickness shear mode with suitable resolution at a reasonable computational cost (time that and model the thickness shear mode with suitable resolution at atoreasonable computational and memory). ANSYS Element ‘SOLID226’ has been chosen model the device. Thiscost is a(time coupled memory). ANSYS Element ‘SOLID226’ has been chosen to model the device. This is a coupled element, element, and its nodes have four degrees of freedom: mechanical displacements in the X-, Y-, and Zand itsand nodes have potential. four degrees of freedom: in thewhere X-, Y-, the andelectric Z-axes; and and axes; electric Therefore, it ismechanical suitable todisplacements model problems electric potential. Therefore, it is suitable to model problems where the electric and mechanical fields mechanical fields are interacting. are interacting. The excitation voltage applied to the top electrode has been chosen to be 0.5 V, while a boundary The excitation appliedtotothe thebottom top electrode hasDamping been chosen to be 0.5been V, while condition of 0 V hasvoltage been applied electrode. factor has set toa5boundary × 10−5. condition of 0 V has been applied to the bottom electrode. Damping factor has been set to 5 × 10−5 . The ‘energy trapping’ effect has been modeled using an indirect method following the Mindlin The ‘energy effect hastheory, been modeled using anof indirect method to following theresponse Mindlin thin plate theory.trapping’ According to this the contribution the electrodes the sensor thin plate theory. According to this theory, the contribution of the electrodes to the sensor response would be equivalent to a proper increment of the density of the quartz crystal in the electroded would beThis equivalent to aisproper of the density the quartz crystal in the electroded regions. regions. increment calledincrement ‘mass loading’, and theof effective density predicted by this theory is This increment is called ‘mass loading’, and the effective density predicted by this theory is described described elsewhere [31,36,37]. elsewhere [31,36,37]. In Figure 1, a numerical model of the 150 MHz resonator is shown. In blue, the bare AT-cut In Figure numerical model the 150where MHz resonator shown. In blue,electrodes the bare AT-cut quartz quartz region 1,isa represented. The ofregions only theistop or bottom appear are region is represented. The regions where only the top or bottom electrodes appear are indicated in red. indicated in red. Finally, the regions where the top and bottom electrodes overlap are depicted in Finally, the regions where the top and bottom electrodes overlap are depicted in purple. purple.

Figure 1. Numerical model of the AWS high fundamental frequency quartz crystal microbalance Figure 1. Numerical model of the AWS high fundamental frequency quartz crystal microbalance (HFF-QCM) 150 MHz sensor. (HFF-QCM) 150 MHz sensor.

2.2. Chemicals 2.2. Chemicals Phosphate-buffered saline (PBS) tablets (dissolved in deionized water: 0.01 M phosphate buffer, Phosphate-buffered saline (PBS) tablets (dissolved in deionized water: 0.01 M phosphate 0.0027 M potassium chloride and 0.137 M sodium chloride, pH 7.4, at 25 °C), Protein A from buffer, 0.0027 M potassium chloride and 0.137 M sodium chloride, pH 7.4, at 25 ◦ C), Protein A Staphylococcus aureus, and IgG from human serum were purchased from Sigma-Aldrich (Sigmafrom Staphylococcus aureus, and IgG from human serum were purchased from Sigma-Aldrich Aldrich Química S.L.U., Madrid, Spain). (Sigma-Aldrich Química S.L.U., Madrid, Spain). 2.3. Instruments and Devices The resonator based on inverted mesa technology has been manufactured by using wet etching and metal sputtering technologies [16]. The final sensor includes the resonator mounted over a custom designed PCB (printed circuit board) in order to improve the device handling and robustness. PCB is

Sensors 2017, 17, 2057

5 of 13

2.3. Instruments and Devices The resonator based on inverted mesa technology has been manufactured by using wet etching 5 of 13 and metal sputtering technologies [16]. The final sensor includes the resonator mounted over a custom designed PCB (printed circuit board) in order to improve the device handling and robustness. PCB is rectangular, with dimensions mm × 14Itsmm. Its thickness 1.55Inmm. In Figure 2, a picture rectangular, with dimensions 17.5 mm17.5 × 14 mm. thickness is 1.55 is mm. Figure 2, a picture of the of the sensor CW 150 Advanced MHz, Advanced Wave Sensors S. L., Paterna, Valencia, Spain) is shown. sensor (AWS (AWS CW 150 MHz, Wave Sensors S. L., Paterna, Valencia, Spain) is shown.

Sensors 2017, 17, 2057

HFF-QCMwith with a fundamental resonance frequency of 150 MHz manufactured by Figure 2.2.HFF-QCM a fundamental resonance frequency of 150 MHz manufactured by Advanced Advanced Wave Wave Sensors S. L.Sensors S. L.

2.4. Electrical of the the Sensor Sensor 2.4. Electrical Characterization Characterization of The real real and and imaginary imaginary parts parts of of the the electrical electrical admittance admittance spectrum spectrum of of the the sensor sensor have have been been The measured using usingaanetwork networkanalyzer analyzer DG8SAQ VNWA 3 (SDR-Kits, Melksham, Wiltshire, The measured DG8SAQ VNWA 3 (SDR-Kits, Melksham, Wiltshire, UK). UK). The AWS AWS A20 platform (Advanced Wave Sensors S. L.) has been used to characterize the frequency and A20 platform (Advanced Wave Sensors S. L.) has been used to characterize the frequency and damping damping shifts of the sensor in response, in during real time, the experiments in flow shifts of the sensor response, real time, the during experiments carried outcarried in flowout conditions. conditions. The AWS F20 platform Wave (Advanced Wave S. used L.) has useda uniform to generate a The AWS F20 platform (Advanced Sensors S. L.)Sensors has been to been generate flow uniform flow through the sensor cell. The AWSuite software interface (Advanced Wave Sensors S. through the sensor cell. The AWSuite software interface (Advanced Wave Sensors S. L.) has been used L.)control has been used to control boththe platforms during the and to register and process to both platforms during experiments and to experiments register and process the acquired data. the acquired data. 2.5. Flow Cell 2.5. Flow Cell A flow cell has been designed and manufactured. The sensor cell includes two main assemblies. A flow is cell has been designed and manufactured. cellthe includes twointerface main assemblies. The sensor placed on the bottom assembly that, in The turn,sensor acts as electrical between Theflow sensor placed the bottom assembly that, in turn, acts of as aluminum the electrical interface between the the cellisand the on measurement system. The body is made (Aluminum 5083 H111, flow cell and thePaterna, measurement system. The is made aluminum (Aluminum 5083 H111, Broncesval S. A., Valencia, Spain) andbody includes fourof gold spring contacts that support the Broncesval S. A.,the Paterna, Valencia, andexternal includesconnectors four gold of spring contacts support the sensor, assuring electrical contactSpain) with the the cell. Thesethat spring contacts sensor, assuring the contact with the external connectors of the theoverpressures cell. These spring contacts allow the sensor to beelectrical maintained in a floating-like state, which controls on the sensor, allowassuring the sensor to be maintained in a floating-like state, controls the stability. overpressures on the thus its integrity and improving its performance in which terms of noise and The electrical sensor, thus of assuring integrity and improving its performance in terms of noise andA20 stability. The connections the cellits have been specifically designed to be compatible with the AWS platform electrical connections connection diagram. of the cell have been specifically designed to be compatible with the AWS A20 platform diagram. The connection upper assembly includes a transparent central core fabricated in PSU (Polysulfone, The upper includes a transparent central core in PSU (Polysulfone, Ensinger Ensinger S. A., assembly La Llagosta, Barcelona, Spain), where a fabricated custom OESTEMER (OSTEMER Flex, S. A., La Labs Llagosta, Barcelona, Sweden) Spain), where custom OESTEMER Mercene Labs Mercene AB., Stockholm, gasketahas been embedded to (OSTEMER create a flowFlex, chamber of 0.53 µL AB., Stockholm, gasket has been embedded to create awith flowa chamber 0.53 termination μL over the over the sensor. Sweden) The fluidic channels are built in this structure standard of fitting sensor. fluidic channels are to built this structure standard termination in order in orderThe to be easily connected theinexternal tubing.with Thea PSU core fitting is mechanically attached toto a be easily connected to the external tubing. The PSU core is La mechanically attached to a ring-shaped ring-shaped PEEK (Polyether-ether-ketone, Ensinger S. A., Llagosta, Barcelona, Spain) part that PEEK (Polyether-ether-ketone, Ensinger S. the A., cell. La Llagosta, Barcelona, Spain)screws, part that provides provides a quick and user-friendly lock of This design avoids using thus assuringa and pressure user-friendly of the and cell. aThis avoids using thus assuring a constant aquick constant over lock the sensor highdesign repetitiveness in its screws, response after cyclic open–close pressure overAll thethe sensor high in its response after cyclic procedures. partsand of athe cellrepetitiveness have been fabricated in a vertical CNCopen–close (computerprocedures. numerical All the parts of thecenter cell have been fabricated in a vertical CNC (computer control) control) machining (Chevalier 1418VMC-Plus, Falcon Machine Tools Co.numerical Ltd., Chang Hua, Taiwan). In Figure 3a, the 3D design of the cell and a section detail of the upper part assembly is shown. In Figure 3b, a section detail of the complete cell assembly is represented where the flow chamber is represented in brown, the sensor PCB in green, and gold contacts in yellow. In Figure 4a,b, the upper

Sensors 2017, 17, 2057 Sensors 2017, 17, 2057

6 of 13 6 of 13

machining center (Chevalier 1418VMC-Plus, Falcon Machine Tools Co. Ltd., Chang Hua, Taiwan). In machining center (Chevalier Tools Co.assembly Ltd., Chang Hua, Taiwan). In Figure the17, 3D design of the 1418VMC-Plus, cell and a sectionFalcon detailMachine of the upper part is shown. In Figure Sensors3a, 2017, 2057 6 of 13 Figure 3a, the 3D design of the cell and a section detail of the upper part assembly is shown. In Figure 3b, a section detail of the complete cell assembly is represented where the flow chamber is 3b, a section detail the of sensor the complete cell assembly represented where the flow represented in brown, PCB in green, and gold is contacts in yellow. In Figure 4a,b, chamber the upperis part of the cell once assembled in the platform and a detailed view of the OSTEMER gasket with represented in brown, the sensor PCB in green, gold contacts in yellow. In Figure 4a,b, the upper part of the cell once assembled in the platform and a detailed view of the OSTEMER gasket with thethe flow chamber, including the flow terminal connections for the fittings, are respectively shown. partchamber, of the cellincluding once assembled the platform and a detailed view ofare therespectively OSTEMER gasket flow the flowinterminal connections for the fittings, shown.with the flow chamber, including the flow terminal connections for the fittings, are respectively shown.

(a) (a)

(b) (b)

Figure 3. (a) Cell 3D view with a cross section detail of the upper part and (b) a section detail of the Figure 3. (a) Cell 3D view with a cross section detail of the upper part and (b) a section detail of the Figure 3.cell (a)assembly Cell 3D view with crosschamber section detail of the upper part and a section detail of the complete where theaflow is represented in brown, the(b) sensor printed circuit complete cell assembly where the flow chamber is represented in brown, the sensor printed circuit complete cell assembly where the flow chamber is represented in brown, the sensor printed circuit board (PCB) in green, and gold contacts in yellow. board (PCB) in green, and gold contacts in yellow. board (PCB) in green, and gold contacts in yellow.

(a) (a)

(b) (b)

Figure 4. Pictures of (a) the measurement cell once assembled in the A20 platform, and (b) detailed Figure 4. (a) measurement cell assembled in the platform, and detailed view of the OSTEMER the flow chamber, terminal connections the Figure 4. Pictures Pictures of of gasket (a) the the with measurement cell once onceincluding assembledthe inflow the A20 A20 platform, and (b) (b)for detailed view of the OSTEMER gasket with the flow chamber, including the flow terminal connections for fittings. view of the OSTEMER gasket with the flow chamber, including the flow terminal connectionsthe for fittings. the fittings.

2.6. Sensor Cleaning/Preparation 2.6. Sensor Cleaning/Preparation 2.6.The Sensor Cleaning/Preparation sensor surface was pretreated for 10 min with UV/ozone cleaner (BioForce Nanosciences The sensor surface was with pretreated for 10 min with UV/ozone cleaner (BioForce Nanosciences Inc., Chicago, IL, rinsed 99% ethanol, rinsed with bi-distilled water, dried with Nitrogen The sensorUSA), surface was pretreated for 10 min with UV/ozone cleaner (BioForce Nanosciences Inc., Chicago, IL, USA), rinsed with 99% ethanol, rinsed with bi-distilled water, dried with Nitrogen gas, and treated IL, for USA), a second timewith with99% UV/ozone 10 min. Inc., Chicago, rinsed ethanol,for rinsed with bi-distilled water, dried with Nitrogen gas, and treated for a second time with UV/ozone for 10 min. gas, and treated for a second time with UV/ozone for 10 min. 2.7. Protein A Adsorption over Gold Sensor Surface and IgG Binding to the Protein A 2.7. 2.7. Protein Protein A A Adsorption Adsorption over over Gold Gold Sensor Sensor Surface Surface and and IgG IgG Binding Binding to to the the Protein Protein A A The cleaned sensor was placed in the cell and exposed to PBS at a flow rate of 50 μL/min until The cleaned placed in and at flow rate of 50 µL/min μL/min until baseline stabilization. 250 was μL of Protein A atcell a concentration ofPBS 50 μg/mL in rate PBS of flowed over the The cleaned sensor sensor was placed in the the cell and exposed exposed to to PBS at aa flow 50 until baseline stabilization. 250 μL of Protein A at a concentration of 50 μg/mL in PBS flowed over exposed sensor surface at theµLabove flow rate. solutions of µg/mL IgG from serum at the athe baseline stabilization. 250 of Protein A at Different a concentration of 50 in human PBS flowed over exposed sensor surface at the above flow rate. Different solutions of IgG from human serum at concentration range of 0.01–10 PBSrate. wereDifferent added, followed (PBS)human rinse, for 5 min exposed sensor surface at theμg/mL aboveinflow solutionsbyofbuffer IgG from serum at aa concentration range 0.01–10 each at the above flowof rate. concentration range of 0.01–10 μg/mL µg/mLininPBS PBSwere wereadded, added,followed followedby bybuffer buffer(PBS) (PBS)rinse, rinse, for for 55 min min each at the above flow rate. each at the above flow rate.

3. Results and Discussion 3.1. Simulation a result SensorsAs 2017, 17, 2057 of

the design process, an AT-cut quartz wafer with a thickness of 66 microns7 of and 13 lateral dimensions of 5 mm × 7 mm has been selected to be the sensor substrate. This thickness assures enough mechanical robustness, while the lateral dimensions are large enough to allow easy handling 3. and but Discussion of Results the device, small enough to be cost-optimal from a manufacturing point of view. A square geometry has been proposed for the sensor to simplify the design of the fluidics and the routing of 3.1. Simulation the electric contacts in the future array. Then, a combination of a gold layer of 60 nm in thickness and As a Chromium result of the design process, AT-cut quartzhave wafer with a thickness of 66 microns and an initial adhesion layer of 5an nm in thickness been selected to implement the sensor lateral dimensions of 5 mm × 7 mm has been selected to be the sensor substrate. This thickness assures electrode and achieve a good electrical conductivity. This value is lower than the thicknesses used by enough mechanical robustness, while the lateral dimensions are large enough allow easy handling other authors [15,16], with the objective of avoiding unwanted spuriousto modes close to the 2 inverted of the device,mode. but small enough to be cost-optimal from a manufacturing view. A square fundamental Later, a squared 0.762 × 0.762 mm mesa regionpoint of 10 of microns’ thickness geometry has been sensor to simplifythe theSauerbrey design of approximation, the fluidics anditthe routing of has been built in theproposed center offor thethe substrate. Applying is possible to the electric contacts in theof future array.mesa Then, a combination of a goldwith layeran ofAu 60 nm in thickness estimate that 10 microns inverted thickness in combination 60 nm/Cr 5 nm and an initial Chromium adhesion layer of 5 nm in thickness been selected to implement the electrode produces a fundamental resonance frequency near have the objective frequency of 150 MHz. sensor electrode and achieve a good electrical conductivity. This value is lower than the thicknesses Finally, applying the Sheahan Equation (1) constrained by a minimum frequency spacing of 400 kHz used by other authors [15,16], with thethe objective of avoiding unwanted spurious close the between the fundamental mode and first inharmonic, a layout of 0.55 mm × modes 0.55 mm hastobeen fundamental squared × 0.762 mm2 inverted selected to bemode. placedLater, in thea center of0.762 the inverted-mesa region. mesa region of 10 microns’ thickness has been built in the center the substrate. Applying theby Sauerbrey approximation, it is Simulations carried outofusing a FEM model defined the geometry described in thepossible former to estimate predicted that 10 microns of inverted mesa thickness with an Au 60spacing nm/Cr of 5 nm paragraph a fundamental resonant frequencyin ofcombination 151 MHz, and a frequency 430 electrode produces a fundamental resonance frequency near the objective frequency of 150 MHz. kHz between the fundamental mode and the first inharmonic. This spacing in frequency is large Finally, the Sheahan while Equation (1) constrained by ainminimum frequency of 400 kHz enoughapplying to avoid disturbances the sensor is operating liquid medium. The spacing X-axis components between the fundamental modeare and the first inharmonic, a layout of 0.55 mm × in 0.55 has of been of displacement in both modes shown in Figure 5. A maximum displacement themm X-axis 6.4 selected to be placed in the center of the inverted-mesa region. nm occurs at the surface of the electroded area when the fundamental mode is excited. Simulations using a FEM model by the geometry described in the former Mechanical carried energy out in the sensor while thedefined fundamental mode is being excited has been paragraph predictedthe a fundamental resonant frequency of model. 151 MHz, and a frequency spacingthat of calculated through results provided by the numerical Simulation results indicate 430 kHz between the fundamental mode and the first inharmonic. This spacing in frequency is large 99.992% of this energy is confined in the electroded region, while the rest is distributed all over the enough to avoid disturbances while sensortrapping’ is operating in liquid medium. components AT-quartz substrate, as predicted bythe ‘energy theory [28,29]. Thus, a The veryX-axis low crosstalk level of displacement in both modes are shown in Figure 5. A maximum displacement in the X-axis of should be expected when this sensor design is employed to implement an array of acoustic wave 6.4 nm occurs at the surface of the electroded area when the fundamental mode is excited. resonators.

(a)

(b)

Figure 5. X-axis displacement patterns for the (a) fundamental thickness shear mode; and (b) first Figure 5. X-axis displacement patterns for the (a) fundamental thickness shear mode; and (b) first inharmonic mode. inharmonic mode.

3.2. Sensor Characterization Mechanical energy in the sensor while the fundamental mode is being excited has been calculated Thethe electrical admittance of the implemented been that acquired. through results provided by spectrum the numerical model. Simulationsensor resultshas indicate 99.992%The of fundamental and the first inharmonic modes have been included in the frequency range interest. this energy is confined in the electroded region, while the rest is distributed all over the of AT-quartz The average fundamental frequency of a batch of [28,29]. 10 manufactured sensors in the airlevel is 149,811,740 substrate, as predicted by ‘energy trapping’ theory Thus, a very low crosstalk should be Hz, with a standard maximum deviation smaller than ±250 kHz. The experimental spectrum of the expected when this sensor design is employed to implement an array of acoustic wave resonators. sensor admittance measured in air has been compared with the admittance calculated in the 3.2. Sensor Characterization The electrical admittance spectrum of the implemented sensor has been acquired. The fundamental and the first inharmonic modes have been included in the frequency range of interest. The average fundamental frequency of a batch of 10 manufactured sensors in the air is 149,811,740 Hz, with a

Sensors 2017, 17, 2057

8 of 13

standard deviation smaller than ±250 kHz. The experimental spectrum of the sensor Sensors 2017,maximum 17, 2057 8 of 13 admittance measured in air has been compared with the admittance calculated in the simulations. In Figure 6, aInreal part6,ofathe admittance the 150 MHz shown. The black simulations. Figure realelectrical part of the electrical of admittance of theresonator 150 MHzisresonator is shown. points represent simulation points at different while the red linethe represents The black pointsthe represent the simulation points atexcitation differentfrequencies, excitation frequencies, while red line the measured using a by network Figure 7, same is included represents theresponse measuredbyresponse using aanalyzer. networkIn analyzer. Inthe Figure 7,information the same information is for the imaginary part of the admittance spectrum of the resonator. The frequency axis has included for the imaginary part of the admittance spectrum of the resonator. The frequency axisbeen has normalized in bothinfigures for a better of the vibration modes. Bymodes. doing so, tolerances been normalized both figures for acomparison better comparison of the vibration By the doing so, the in the fabrication process that can slightly affect the thickness the inverted section, thus tolerances in the fabrication process that can slightly affect the of thickness of themesa inverted mesaand section, the resonant frequency, are considered. Results show a very good agreement between andexperimental thus the experimental resonant frequency, are considered. Results show a very good agreement experimental characterization and numerical predictions. This fact supports different between experimental characterization and numerical predictions. This factthe supports theboundary different conditions, meshing options, assumptions e.g., “mass e.g., loading” [31,36,37] boundary conditions, meshingand options, and assumptions “masstheory loading” theoryconsidered [31,36,37] during the development of the numerical The main difference the experimental considered during the development of model. the numerical model. The between main difference between and the numerical data is the offset found in the imaginary part of the admittance. This effect is due to the experimental and numerical data is the offset found in the imaginary part of the admittance. This parasitic capacitance present in the realpresent sensor,in and notsensor, taken into account in the is effect is due to the parasitic capacitance theisreal and is not taken intomodel. accountThis in the explained by is theexplained unavoidable parasitic capacitance, in parallel with the by the model. This by the unavoidable parasitic capacitance, in sensor, parallelintroduced with the sensor, sensor cell. by Nevertheless, this Nevertheless, fact does not invalidate the not model, and it will be considered in the introduced the sensor cell. this fact does invalidate the model, and it will be following considereddesign in thesteps. following design steps.

Figure 6.6.Comparison Comparison of real the part real ofpart of the electrical admittance (G) in measured in air of a Figure of the the electrical admittance (G) measured air of a manufactured manufactured AWS HFF-QCM and the response bymodel. the numerical model. AWS HFF-QCM sensor, and the sensor, response provided by theprovided numerical

Figure 7. 7. Comparison of the imaginary part of the electrical electrical admittance (B) measured in air of of the the Figure manufactured AWS AWSHFF-QCM HFF-QCMsensor, sensor,and andthe theresponse responseprovided providedby bythe thenumerical numericalmodel. model. manufactured

The quality factor (Q) of the sensor has been characterized in air and bi-distilled water conditions using the network analyzer described in Section 2. A quality factor of 9500 has been measured in air, while a Q of 600 has been obtained in water.

Sensors 2017, 17, 2057

Sensors 2017, 17, 2057

9 of 13

9 of 13

The quality factor (Q) of the sensor has been characterized in air and bi-distilled water conditions using the network described in Section A quality factorreported of 9500 has been measured in et air,al. Q factors of analyzer other HFF-QCM sensors have2.been previously in the literature. Lin while Q of 600Qhas been of obtained water. [38] areported factors 38,300 in and 1170 in air and water, respectively, for a 30 MHz HFF-QCM, Q factors of other HFF-QCM sensors have been previously reported in the Lin et al.As [38] and Rabe et al. [25] informed about Q factors of 50,000 and 1110 for their 50literature. MHz HFF-QCM. it is reported Q factors of 38,300 and 1170 in air and water, respectively, for a 30 MHz HFF-QCM, well-known, the quality factor of the quartz resonator is inversely proportional to its resonant and Rabe et [39], al. [25] about factors of 50,000 and 1110 for their 50 work MHzhas HFF-QCM. it frequency so informed it is expected thatQthe 150 MHz resonator presented in this a lower QAs factor is when well-known, the quality factor of the quartz resonator is inversely proportional to its resonant compared with other resonators with lower resonant frequencies. It is worth noting that while frequency [39],insoair it is that the 150 MHz resonatorfour presented intimes this work a lower Q factor the Q factor ofexpected our 150 MHz resonator is between and five lowerhas that those presented when compared with other resonators with lower resonant frequencies. It is worth noting that by Lin et al. and Rabe et al., the Q factor in water is only reduced by a factor of two. This factwhile can be the Q factor if inthe air penetration of our 150 MHz resonator is between four five times loweristhat those presented explained depth of the acoustic wave in and the liquid medium considered [4,40]. As bythe Linresonance et al. andfrequency Rabe et al., the Q factor water is only reduced by a factor of two.with Thisthe factenergy can becomes higher,inthe penetration depth decreases together belosses explained if the penetration depth of the acoustic wave in the liquid medium is considered [4,40]. due to the liquid. As the resonance frequency becomes higher, the penetration depth decreases together with the energy losses due toValidation the liquid. 3.3. Sensor

An experiment 3.3. Sensor Validation consisting of Protein A adsorption to the sensor gold surface and a subsequent IgG binding to the Protein A has been conducted to validate the performance of the HFF-QCM device An experiment consistingThe of Protein A adsorption to thefrequency sensor gold surface and ashifts subsequent in bio-sensing applications. real-time sensor resonant and resistance acquired IgG binding to the Protein A has been conducted to validate the performance of the HFF-QCM device during the experiment are shown in in bio-sensing applications. The real-time sensor resonant frequency and resistance shifts acquired Figures 8 and 9 respectively. Resistance is commonly used in QCM experiments to evaluate the during experiment aresensor. shownAt in the Figures 8 andof9 the respectively. Resistance is commonly in energythe dissipation in the beginning experiment, absolute resistance andused resonant QCM experiments to evaluate the energy dissipation in the sensor. At the beginning of the experiment, frequency were 582 Ω and 149,510,403 Hz, respectively. Periodic peaks that can be observed in the absolute resistance andsyringe resonant frequency 582 andover 149,510,403 Hz,surface. respectively. Periodic graphs are due to the pump used towere create theΩflow the sensor The syringe must peaks that can be observed in the this graphs dueprocess to the syringe pump to create the flow over is be re-filled every 5 min. During veryare short (2 s), the flow used rate over the sensor surface the sensor The syringe must re-filled aevery 5 min. During very short process (2 s), zero, andsurface. the pressure changes. Thisbeproduces change in the sensorthis conditions, and the baseline the flow rate overtwo theconsecutive sensor surface is zero,ofand the pressure changes. This produces a change in the changes. First, injections protein A (50 μg/mL) have been conducted to make sure sensor conditions, and the baseline changes. First, two consecutive injections of protein A (50 µg/mL) that the sensor surface is well covered. In response to the protein A injections, a change both in have been conducted to make sure that the sensor surface shift is well to the A frequency and resistance is observed. A total negative ofcovered. −9380 HzIninresponse frequency andprotein a positive injections, change both frequency resistance is observed. A total negative shift of −9380 Hz in change ina resistance of in 10.5 Ω were and measured. frequency and a positive change in resistance of 10.5 Ω were measured.

Figure 8. 8. Frequency shift measured during the adsorption ofof Protein AA over Figure Frequency shift measured during the adsorption Protein overthe thesurface surfaceofofthe thesensor, sensor, and thethe subsequent injection of of 0.10.1 µg/mL, and subsequent injection μg/mL,1 1µg/mL μg/mLand and10 10µg/mL μg/mL of ofIgG IgGin inthe thesensor sensorcell. cell.

Sensors 2017, 17, 2057 Sensors 2017, 17, 2057

10 of 13 10 of 13

Figure 9. 9. Resistance Resistance shift shift measured measured during during the the adsorption adsorption of of Protein Protein A A over over the Figure the surface surface of of the the sensor, sensor, and the subsequent injection of 0.1 μg/mL, 1 μg/mL and 10 μg/mL of IgG in the sensor cell. and the subsequent injection of 0.1 µg/mL, µg/mL and 10 µg/mL of IgG in the sensor cell.

Once Protein Protein A A has has been been absorbed absorbed over over the the surface surface of of the several injections injections with with different different Once the sensor, sensor, several concentration of IgG have been carried out. The response of the sensor to injections to 0.1 μg/mL, concentration of IgG have been carried out. The response of the sensor to injections to 0.1 µg/mL,1 of − −582 Hz, − −5585 and − −28,630 Hz, 1μg/mL, µg/mL,and and1010μg/mL µg/mLofofIgG IgGprovided provided frequency frequency shifts shifts of 582 Hz, 5585 Hz, Hz, and 28,630 Hz, respectively,asasdepicted depicted Figure 8. No detectable changes are observed in resistance when 0.1 respectively, in in Figure 8. No detectable changes are observed in resistance when 0.1 µg/mL μg/mL sample is injected. An increase in resistance of 7.5 Ω is detected when 1 μg/mL IgG is injected, sample is injected. An increase in resistance of 7.5 Ω is detected when 1 µg/mL IgG is injected, while an while an increment 4.8 Ω is registered when 10 IgG are Baseline injected. remains Baselinestable remains stable increment of 4.8 Ω isofregistered when 10 µg/mL ofμg/mL IgG areofinjected. once the once thevolume samplehas volume has passedthe through sample passed through sensor.the sensor. Sensor validation validation results results prove prove that that this this device device can can be be employed employed to to monitor monitor biomolecular biomolecular Sensor interactions carried carried out out in in liquid liquid medium. medium. A stable baseline baseline both both in in dissipation dissipation and and in in resonant resonant interactions A stable frequency is observed when buffer is passed through the sensor cell. Once Protein A is injected into frequency is observed when buffer is passed through the sensor cell. Once Protein A is injected into the flow cell, a negative shift in frequency and a positive shift in resistance are obtained, which is the flow cell, a negative shift in frequency and a positive shift in resistance are obtained, which is consistent with with aa typical typical QCM QCM protein protein adsorption adsorption experiment experiment [41]. [41]. The The measured measured frequency frequency shift shift consistent produced by the combined adsorption of Protein A and the subsequent injection of IgG has a value produced by the combined adsorption of Protein A and the subsequent injection of IgG has a value of of51,000 −51,000 as expected, much higher experiments described the literature, which − Hz,Hz, as expected, much higher thanthan otherother experiments described in the in literature, which report a frequency of around Hzconventional using conventional 9 MHzsensors QCM sensors fact areport frequency shift of shift around −200 Hz−200 using 9 MHz QCM [42,43]. [42,43]. This factThis reveals reveals the enhanced sensitivity of the HFF-QCM device. the enhanced sensitivity of the HFF-QCM device. 4. Conclusions Conclusions 4. A HFF-QCM HFF-QCM sensor sensor based based on on inverted-mesa inverted-mesa technology technology has has been been designed designed to to be be the the ‘basic ‘basic A buildingunit’ unit’of of a future MQCM device. basic requirements design requirements for this concrete building a future MQCM device. SensorSensor basic design for this concrete application application included small size, good quality factor (considering the high operation frequency), and included small size, good quality factor (considering the high operation frequency), and limitation limitation of the presence ofovertones inharmonic overtones the fundamental vicinity of the fundamental of the presence of inharmonic in the vicinity in of the mode. Followingmode. these Following these specifications, a theoretical study out has to been carried out togeometry define anfor initial geometry specifications, a theoretical study has been carried define an initial the resonator. for thea resonator. a numerical FEMdeveloped model hastobeen developed to estimate the Then, numericalThen, FEM model has been estimate the characteristics of characteristics such a sensor. of HFF-QCM such a sensor. A HFF-QCM basedhas on this has beenand implemented and characterized. A device based on device this design beendesign implemented characterized. Comparison Comparison between experimental and numerical results showed a very good correlation. between experimental and numerical results showed a very good correlation. After the the implementation implementationand andvalidation validationofofthe theAWS AWSHFF-QCM HFF-QCM sensor 150 MHz, this sensor After sensor ofof 150 MHz, this sensor is is going used “basic element” of an array. Thus, a novel acoustic wave array be made going to to be be used as as thethe “basic element” of an array. Thus, a novel acoustic wave array willwill be made up upaby a matrix thisoftype of sensors. Of course, it is necessary spacing between by matrix of thisoftype sensors. Of course, it is necessary to design to thedesign spacingthe between resonators resonators in Z-axes the X- and Z-axes so that the crosstalk betweenresonators neighbor is resonators is minimized. New in the X- and so that the crosstalk between neighbor minimized. New numerical numerical FEM models on the one presented in this papertowill to find an solution optimal FEM models based on thebased one presented in this paper will be used findbe anused optimal design design prior to array This work will be presentedarticles. in forthcoming articles. prior tosolution array fabrication. This fabrication. work will be presented in forthcoming

Sensors 2017, 17, 2057

11 of 13

Acknowledgments: This work was funded by the European Commission Horizon 2020 Programme under Grant Agreement number ICT-28-2015/687785-LIQBIOPSENS (Reliable Liquid Biopsy technology for early detection of colorectal cancer). Author Contributions: Román Fernández analyzed the experimental data and wrote the paper; Pablo García designed and implemented the measurement cell; María García designed and performed the validation experiments; José V. García designed the support PCB and characterized the sensor prototypes; Yolanda Jiménez and Antonio Arnau conceived the paper and developed the sensor numerical model. Conflicts of Interest: The authors declare no conflict of interest.

Appendix A Table A1. Values of the piezoelectric, dielectric and elastic constants of the AT-cut Quartz. 

39.21  0 0

Dielectric  S Matrix ε

0 39.82 0.86

 0 0.86 ·10−12 40.42

 Elastic  EMatrix  c

Piezoelectric Matrix [e]

86.74 −8.25 27.15 −3.66 0  −8.25 129.77 −7.42 5.7 0   27.15 −7.42 102.83 9.92 0   −3.66 5.7 9.92 38.61 0   0 0 0 0 68.81 0 0 0 0 2.53  0.171 −0.152 −0.0187 0  0 0 0 0.108 0 0 0 −0.0761

0 0 0 0 2.53 29.01

     ·109   

 0 −0.095  0.067

Table A2. Quartz and gold densities. ρ gold

19,300

kg/m3

ρ quartz

2650

kg/m3

References 1.

2. 3. 4. 5. 6. 7. 8. 9. 10.

Soper, S.A.; Brown, K.; Ellington, A.; Frazier, B.; Garcia-Manero, G.; Gau, V.; Gutman, S.I.; Hayes, D.F.; Korte, B.; Landers, J.L. Point-of-care biosensor systems for cancer diagnostics/prognostics. Biosens. Bioelectron. 2006, 21, 1932–1942. [CrossRef] [PubMed] Gubala, V.; Harris, L.F.; Ricco, A.J.; Tan, M.X.; Williams, D.E. Point of care diagnostics: Status and future. Anal. Chem. 2011, 84, 487–515. [CrossRef] [PubMed] Sauerbrey, G. Use of vibrating quartz for thin film weighing and microweighing. Z. Phys. 1959, 155, 206–222. [CrossRef] Arnau, A. Piezoelectric Transducers and Applications, 2nd ed.; Springer: Berlin, Germany, 2008. Steinem, C.; Janshoff, A. Piezoelectric Sensors; Springer Science & Business Media: Berlin, Germany, 2007; Volume 5. Tsortos, A.; Papadakis, G.; Gizeli, E. Shear acoustic wave biosensor for detecting DNA intrinsic viscosity and conformation: A study with qcm-d. Biosens. Bioelectron. 2008, 24, 836–841. [CrossRef] [PubMed] Tuantranont, A.; Wisitsora-At, A.; Sritongkham, P.; Jaruwongrungsee, K. A review of monolithic multichannel quartz crystal microbalance: A review. Anal. Chim. Acta 2011, 687, 114–128. [CrossRef] [PubMed] Tao, W.; Lin, P.; Ai, Y.; Wang, H.; Ke, S.; Zeng, X. Multichannel quartz crystal microbalance array: Fabrication, evaluation, application in biomarker detection. Anal. Biochem. 2016, 494, 85–92. [CrossRef] [PubMed] Rabe, J.; Seidemann, V.; Buettgenbach, S. Monolithic fabrication of wireless miniaturized quartz crystal microbalance (qcm-r) arrays and their application for biochemical sensors. Sens. Mater. 2003, 15, 381–391. Vig, J.R.; Filler, R.L.; Kim, Y. Microresonator sensor arrays. In Proceedings of the 1995 IEEE International Frequency Control Symposium (49th Annual Symposium), San Francisco, CA, USA, 31 May–2 June 1995; pp. 852–869.

Sensors 2017, 17, 2057

11. 12.

13. 14. 15.

16.

17. 18.

19.

20. 21.

22. 23.

24. 25.

26. 27.

28.

29. 30. 31. 32.

12 of 13

Abe, T.; Esashi, M. One-chip multichannel quartz crystal microbalance (qcm) fabricated by deep rie. Sens. Actuators A Phys. 2000, 82, 139–143. [CrossRef] Jaruwongrungsee, K.; Waiwijit, U.; Wisitsoraat, A.; Sangworasil, M.; Pintavirooj, C.; Tuantranont, A. Real-time multianalyte biosensors based on interference-free multichannel monolithic quartz crystal microbalance. Biosens. Bioelectron. 2015, 67, 576–581. [CrossRef] [PubMed] Hung, V.N.; Abe, T.; Minh, P.N.; Esashi, M. Miniaturized, highly sensitive single-chip multichannel quartz-crystal microbalance. Appl. Phys. Lett. 2002, 81, 5069–5071. [CrossRef] Kao, P.; Doerner, S.; Schneider, T.; Allara, D.; Hauptmann, P.; Tadigadapa, S. A micromachined quartz resonator array for biosensing applications. J. Microelectromech. Syst. 2009, 18, 522–530. Liang, J.; Huang, J.; Zhang, T.; Zhang, J.; Li, X.; Ueda, T. An experimental study on fabricating an inverted mesa-type quartz crystal resonator using a cheap wet etching process. Sensors 2013, 13, 12140–12148. [CrossRef] [PubMed] Zimmermann, B.; Lucklum, R.; Hauptmann, P.; Rabe, J.; Büttgenbach, S. Electrical characterisation of high-frequency thickness-shear-mode resonators by impedance analysis. Sens. Actuators B Chem. 2001, 76, 47–57. [CrossRef] Lubczyk, D.; Siering, C.; Lörgen, J.; Shifrina, Z.B.; Müllen, K.; Waldvogel, S.R. Simple and sensitive online detection of triacetone triperoxide explosive. Sens. Actuators B Chem. 2010, 143, 561–566. [CrossRef] Brutschy, M.; Schneider, M.W.; Mastalerz, M.; Waldvogel, S.R. Porous organic cage compounds as highly potent affinity materials for sensing by quartz crystal microbalances. Adv. Mater. 2012, 24, 6049–6052. [CrossRef] [PubMed] Brutschy, M.; Schneider, M.W.; Mastalerz, M.; Waldvogel, S.R. Direct gravimetric sensing of gbl by a molecular recognition process in organic cage compounds. Chem. Commun. 2013, 49, 8398–8400. [CrossRef] [PubMed] Uttenthaler, E.; Schräml, M.; Mandel, J.; Drost, S. Ultrasensitive quartz crystal microbalance sensors for detection of m13-phages in liquids. Biosens. Bioelectron. 2001, 16, 735–743. [CrossRef] March, C.; García, J.V.; Sánchez, Á.; Arnau, A.; Jiménez, Y.; García, P.; Manclús, J.J.; Montoya, Á. High-frequency phase shift measurement greatly enhances the sensitivity of qcm immunosensors. Biosens. Bioelectron. 2015, 65, 1–8. [CrossRef] [PubMed] Bottom, V.E. Introduction to Quartz Crystal Unit Design; Van Nostrand Reinhold: Chichester, UK, 1982. Sagmeister, B.P.; Graz, I.M.; Schwödiauer, R.; Gruber, H.; Bauer, S. User-friendly, miniature biosensor flow cell for fragile high fundamental frequency quartz crystal resonators. Biosens. Bioelectron. 2009, 24, 2643–2648. [CrossRef] [PubMed] Abe, T.; Hung, V.N.; Esashi, M. Inverted mesa-type quartz crystal resonators fabricated by deep-reactive ion etching. IEEE Trans. Ultrason. Ferroelectr. Freq. Control 2006, 53, 1234. [CrossRef] [PubMed] Rabe, J.; Buttgenbach, S.; Zimmermann, B.; Hauptmann, P. Design, manufacturing, and characterization of high-frequency thickness-shear mode resonators. In Proceedings of the 2000 IEEE/EIA International Frequency Control Symposium and Exhibition, Kansas City, MO, USA, 9 June 2000; pp. 106–112. Rocha-Gaso, M.-I.; March-Iborra, C.; Montoya-Baides, Á.; Arnau-Vives, A. Surface generated acoustic wave biosensors for the detection of pathogens: A review. Sensors 2009, 9, 5740–5769. [CrossRef] [PubMed] García, J.; Rocha, M.; March, C.; García, P.; Francis, L.; Montoya, A.; Arnau, A.; Jimenez, Y. Love mode surface acoustic wave and high fundamental frequency quartz crystal microbalance immunosensors for the detection of carbaryl pesticide. Proc. Eng. 2014, 87, 759–762. [CrossRef] Shockley, W.; Curran, D.R.; Koneval, D.J. Energy trapping and related studies of multiple electrode filter crystals. In Proceedings of the 17th Annual Symposium on Frequency Control, Atlantic City, NJ, USA, 27–29 May 1963; pp. 88–126. Shockley, W.; Curran, D.; Koneval, D. Trapped-energy modes in quartz filter crystals. J. Acoust. Soc. Am. 1967, 41, 981–993. [CrossRef] Shen, F.; Lu, P.; O’Shea, S.; Lee, K. Frequency coupling and energy trapping in mesa-shaped multichannel quartz crystal microbalances. Sens. Actuators A Phys. 2004, 111, 180–187. [CrossRef] Beaver, W. Analysis of elastically coupled piezoelectric resonators. J. Acoust. Soc. Am. 1968, 43, 972–981. [CrossRef] Sheahan, D. An improved resonance equation for at-cut quartz crystals. Proc. IEEE 1970, 58, 260–261. [CrossRef]

Sensors 2017, 17, 2057

33. 34. 35. 36.

37. 38. 39. 40.

41. 42. 43.

13 of 13

Wessels, A.; Klöckner, B.; Siering, C.; Waldvogel, S.R. Practical strategies for stable operation of hff-qcm in continuous air flow. Sensors 2013, 13, 12012–12029. [CrossRef] [PubMed] Lu, F.; Lee, H.; Lu, P.; Lim, S. Finite element analysis of interference for the laterally coupled quartz crystal microbalances. Sens. Actuators A Phys. 2005, 119, 90–99. [CrossRef] Cady, W.G. Piezoelectricity: An Introduction to the Theory and Application of Electromechanical Phenomena in Crystals; McGraw-Hill: New York, NY, USA, 1946. Pao, S.; Chaos, M.; Lam, C.; Chang, P. An efficient numerical method in calculating the electrical impedance different modes of at-cut quartz crystal resonator. In Proceedings of the 2004 IEEE International Frequency Control Symposium and Exposition, Montreal, QC, Canada, 23–27 August 2004; pp. 396–400. Mindlin, R.D.; Lee, P. Thickness-shear and flexural vibrations of partially plated, crystal plates. Int. J. Solids Struct. 1966, 2, 125–139. [CrossRef] Lin, Z.; Yip, C.M.; Joseph, I.S.; Ward, M.D. Operation of an ultrasensitive 30-mhz quartz crystal microbalance in liquids. Anal. Chem. 1993, 65, 1546–1551. [CrossRef] Ballato, A.; Gualtieri, J.G. Advances in high-q piezoelectric resonator materials and devices. IEEE Trans. Ultrason. Ferroelectr. Freq. Control 1994, 41, 834–844. [CrossRef] [PubMed] Montagut, Y.; García, J.; Jiménez, Y.; March, C.; Montoya, A.; Arnau, A. Frequency-shift vs. phase-shift characterization of in-liquid quartz crystal microbalance applications. Rev. Sci. Instrum. 2011, 82, 064702. [CrossRef] [PubMed] Dixon, M.C. Quartz crystal microbalance with dissipation monitoring: Enabling real-time characterization of biological materials and their interactions. J. Biomol. Tech. JBT 2008, 19, 151. [PubMed] Li, J.; Wu, Z.-Y.; Xiao, L.-T.; Zeng, G.-M.; Huang, G.-H.; Shen, G.-L.; Yu, R.-Q. A novel piezoelectric biosensor for the detection of phytohormone β-indole acetic acid. Anal. Sci. 2002, 18, 403–407. [CrossRef] [PubMed] Kengne-Momo, R.; Jeyachandran, Y.; Assaf, A.; Esnault, C.; Daniel, P.; Pilard, J.F.; Durand, M.; Lagarde, F.; Dongo, E.; Thouand, G. A simple method of surface functionalisation for immuno-specific immobilisation of proteins. Anal. Bioanal. Chem. 2010, 398, 1249–1255. [CrossRef] [PubMed] © 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).