a State Key Laboratory for Manufacturing Systems Engineering, the International Joint Laboratory for Micro/Nano Manufacturing and Measurement Technologies, Xi’an Jiaotong University, Xi’an 710049, China
b Xi’an Jiaotong University (Yantai) Research Institute for Intelligent Sensing Technology and System, Yantai 264003, China
c School of Instrument Science and Technology, Xi’an Jiaotong University, Xi’an 710049, China
d School of Mechanical Engineering, Xi’an Jiaotong University, Xi’an 710049, China
e Shandong Laboratory of Yantai Advanced Materials and Green Manufacturing, Yantai 265503, China
f Bio/CMOS Interfaces Laboratory, Institute of Electrical and Microengineering, School of Engineering, École Polytechnique F éd érale de Lausanne, Neuchâtel 2000, Switzerland
g College of Mechanical and Electrical Engineering, Shaanxi University of Science and Technology, Xi’an 710021, China
Gravimetric resonant-inspired biosensors have attracted increasing attention in industrial and point-of-care applications, enabling label-free detection of biomarkers such as DNA and antibodies. Capacitive micromachined ultrasonic transducers (CMUTs) are promising tools for developing miniaturized high-performance biosensing complementary metal-oxide-silicon (CMOS) platforms. However, their operability is limited by inefficient functionalization, aggregation, crosstalk in the buffer, and the requirement for an external high-voltage (HV) power supply. In this study, we aimed to propose a CMUTs-based resonant biosensor integrated with a CMOS front-end interface coupled with ethylene-glycol alkanethiols to detect single-stranded DNA oligonucleotides with large specificity. The topography of the functionalized surface was characterized by energy-dispersive X-ray microanalysis. Improved selectivity for on-chip hybridization was demonstrated by comparing complementary and non-complementary single-stranded DNA oligonucleotides using fluorescence imaging technology. The sensor array was further characterized using a five-element lumped equivalent model. The 4 mm2 application-specific integrated circuit chip was designed and developed through 0.18 μm HV bipolar-CMOS-double diffused metal-oxide-silicon (DMOS) technology (BCD) to generate on-chip 20 V HV boosting and to track feedback frequency under a standard 1.8 V supply, with a total power consumption of 3.8 mW in a continuous mode. The measured results indicated a detection sensitivity of 7.943 × 10−3 μmol∙L−1∙Hz−1 over a concentration range of 1 to 100 μmol∙L−1. In conclusion, the label-free biosensing of DNA under dry conditions was successfully demonstrated using a microfabricated CMUT array with a 2 MHz frequency on CMOS electronics with an internal HV supplier. Moreover, ethylene-glycol alkanethiols successfully deposited self-assembled monolayers on aluminum electrodes, which has never been attempted thus far on CMUTs, to enhance the selectivity of bio-functionalization. The findings of this study indicate the possibility of full-on-chip DNA biosensing with CMUTs.
Zhikang Li, Yihe Zhao, Gian Luca Barbruni, Jie Li, Zixuan Li, Jiawei Yuan, Ping Yang, Libo Zhao, Zhuangde Jiang, Sandro Carrara.
Development of an Integrated CMUTs-Based Resonant Biosensor for Label-Free Detection of DNA with Improved Selectivity by Ethylene-Glycol Alkanethiols.
Engineering, 2024, 41 (10) : 244-255 DOI:10.1016/j.eng.2023.12.015
Over the last few decades, biosensor technology, which relies on the detection of biomolecules and their specific interactions, has been investigated as an essential analytical tool for various applications, including environmental monitoring and medical care, with a special focus on diagnosing infectious diseases [1]. To date, the coronavirus disease (COVID-19) is clinically reported as a fast-spreading disease due to individual asymptomatic carriers of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) [2]. Therefore, new biosensing platforms that combine accuracy, rapid pathogen detection, and miniaturization are being developed. Target detection involves the identification of biomolecules, such as DNA, RNA, immunoglobulins, enzymes, and aptamers [3]. According to construction strategies, biosensors are mainly classified into two categories: label-based and label-free [4], [5]. Compared to classical label-based biosensing methodologies such as enzyme immunoassays, chemiluminescent assays, polymerase chain reactions, and mass spectrometry, which are commonly used in medical laboratories as infectious disease biomarkers, label-free techniques provide efficient and improved cost-effectiveness by avoiding tedious covalent signal-probe labeling processes and label interference with binding sites [6]. Label-free methods principally involve surface plasmon resonance (SPR) [7], surface-enhanced Raman scattering (SERS) [8], [9], interference-based techniques [10], electrochemical detection [11], capacitive complementary metal–oxide–silicon (CMOS) chips [12], metasurface-assisted terahertz (THz) [13], [14], and microresonators [15], [16]. Ghayoor et al. [17] designed an SPR sensor based on a novel Fibonacci structure made of thin nanostructured films and graphene, achieving an ultrahigh sensitivity of 192.19 (° )∙RIU−1 (RIU: refractive index units) to identify DNA-based infectious diseases, such as SARS-CoV-2. However, this method generally relies on precise instrument calibration and sophisticated equipment. Anwar et al. [18] used gold nanocubes (AuNCs) to enhance SERS signals to detect human immunodeficiency virus type 1 (HIV-1) DNA strands and discovered that Raman bands in the case of AuNC-coated HIV-1 DNA molecules were observed with enhanced intensity compared to silver nanocubes (AgNCs). However, inconsistencies in signal enhancement can occur because of variations in nanostructures. Lomae et al. [19] fabricated a paper-based electrochemical device to detect complementary DNA (cDNA) in which the electrodes were modified with a pyrrolidinyl peptide nucleic acid. A linear range from 0.1 to 200 nmol∙L−1 and a detection limit of 1.0 pmol∙L−1 were obtained after a 42 min performance with the whole assay. Lai et al. [20] presented a 16 × 16 CMOS capacitive sensor array based on a 0.35 μm CMOS process, applying small, interdigitated microelectrodes and shared switched-capacitor circuits to detect DNA molecules in a range of 10 amol∙L−1 to 100 pmol∙L−1 at diluted ionic strength with a capacitance sensitivity of (1fF = 10−15 F) at a clock frequency of 2 MHz. Li and Zhang. [21] investigated the vibrational coupling effects of THz irradiation on DNA replication and discovered that 53 THz irradiation reduced the denaturation temperature of DNA duplexes and the melting time by 3 °C and 20%, respectively. In addition, with advances in material chemistry, flexible electronic devices for label-free health monitoring have revolutionized diagnostics. These flexible sensors seamlessly conform to the body, offering real-time noninvasive monitoring of biomarkers and enhanced early disease detection [22], [23], [24], [25]. The challenges of flexible electronics in practical applications still exist, including material durability, performance stability under bending and stretching, and integration with power sources [26], [27]. With the technical development of microelectromechanical systems (MEMS) and CMOS, microresonator-based biosensors have emerged as high-potential biosensing platforms with high sensitivity and miniaturization.
Several types of microresonance biosensing devices with specific probes directly immobilized onto sensing surfaces have been proposed, including quartz crystal microbalances (QCMs) [15], [28], microcantilevers [29], as well as microacoustic devices such as surface acoustic wave (SAWs) devices [30], [31], film bulk acoustic resonators (FBARs) [32], and micromachined ultrasonic transducers (MUTs) [33], [34], [35], [36], [37]. The QCM comprises a specifically cut quartz crystal and upper and lower electrodes, which is advantageous for a simple configuration and high-quality factors (up to 104) [38]. However, it is incompatible with CMOS technology, and frequency drifts lead to a reduction in stability because of the high temperature coefficient of quartz crystals. Although microcantilevers are compatible with the CMOS process and can achieve convenient signal detection, their kHz-range resonant frequencies induce sensitivity limitations. SAW devices, composed of interdigital electrodes and piezoelectric substrates, operate at hundreds of MHz-range frequencies [30]. However, high frequencies usually affect stability because of the magnified effects of environmental factors. Furthermore, piezoelectric materials such as aluminum nitride (AlN) and zinc oxide (ZnO) are typically not budget-friendly for high-performance applications. Their piezoelectric performance is notably influenced by fabrication and polarization properties (d33: AlN = 5.1 pm∙V−1 [39], ZnO = 9.93 pm∙V−1 [40]). However, costs can vary based on factors such as quantity, purity, and customization, therefore, they are still valued for their unique characteristics in certain applications. Piezoelectric performance can be detected in the nanomolar range, which is a problem that has similarly been reported for FBAR, whose frequency approaches tens of GHz [32]. However, MUTs can be realized using two types of devices: piezoelectric MUTs (PMUTs) and capacitive MUTs (CMUTs). Both these methods are attractive for biomedical applications. However, compared to PMUTs devices, which have relatively low electromechanical conversion coefficients, CMUTs are more efficient for biosensing. They are typically fabricated with a backside vacuum cavity, which enables a high electromechanical conversion efficiency of up to 85% [41]. Additionally, CMUTs arrays usually comprise of hundreds of cells, thereby reducing their motional capabilities and thermal noise. Functionalization with sensitive materials benefits from such structures to achieve cell-to-cell uniformity, whereas selectivity remains limited owing to inaccurate biofunctionalization.
Previous studies have proposed various strategies for biochemical detection using CMUTs-based biosensors, most of which have contributed to the detection of environmental gases, including CO2 [42], humidity [43], [44], SO2 [45], and volatile organic compounds (VOCs) [46], [47]. The sensitivity of dimethyl methylphosphonate (DMMP) detection using CMUTs is almost at the parts per trillion (ppt) level [48]. Recently, researchers started using CMUTs as potential immunosensors to detect immunoglobulin G (IgG) over an extensive range of concentrations (10-500 ng∙mL−1) with a detectable mass sensitivity of 0.44 ag∙μm−2 [49] and low detection capabilities (10 pg∙mL−1-1 ng∙mL−1) of neuropeptides involved in the progression of Alzheimer’s disease [50]. However, external equipment is required typically owing to the supply of high voltages, thereby reducing convenience and miniaturization of the proposed devices. However, high-voltage (HV) power suppliers are always required to provide a direct-current (DC) bias voltage, whereas impedance analyzers measure frequency shifts during detection. Furthermore, to the best of our knowledge, only thiol-based functionalization methods have been developed for CMUT devices using Au electrodes or particles. Aluminum (Al) electrodes are commonly used in MEMS technology because of their high-cost effectiveness, reasonable electrical properties, and significant package compatibility. Moreover, Al is commonly used as the top metal layer in various CMOS technologies; therefore, it does not require postprocessing, which is necessary for CMOS chips with exposed gold electrodes. However, there is a lack of research on suitable biofunctionalization methods for CMUTs with Al electrodes, including physicochemical surface activation and bioreceptor immobilization, which are crucial and have a notable influence on the selectivity and sensitivity of the resulting biosensors [51]. The motion loss of the flexural vibration mode, aggregation phenomenon, and crosstalk from neighboring cells in buffer solutions make it challenging to precisely monitor the frequency shifts [50]. Therefore, it is necessary to develop reliable methods for functionalizing Al into CMUTs-based DNA biosensors.
This study proposes a CMUTs-based resonant biosensor integrated with a CMOS front-end interface to detect single-stranded DNA oligonucleotides. Our CMOS driving and processing blocks directly provide an on-chip HV-DC bias and enable the continuous monitoring of frequency shifts. Fig. 1 illustrates the detection principle, including using ethylene-glycol alkanethiols to prevent non-specific binding [52]. A CMUT array, consisting of hundreds of circular cells was fabricated using a low-temperature direct bonding technology. A lumped-element model (LEM) is presented to extract the electrical equivalent parameters, which are then used to properly design the application-specific integrated circuit (ASIC) interface in Taiwan Semiconductor Manufacturing Company Limited (TSMC; China) 0.18 μm HV bipolar-CMOS-DMOS (BCD) technology. A self-assembled monolayer (SAM) of using ethylene-glycol thiol-phosphate acid was used to functionalize the top Al electrodes of the CMUTs. This SAM ensures the adjustment of water molecules owing to the ethylene-glycol function. The assured adjustment of water molecules is the key to improving the selectivity of the biosensing surface obtained by integrating single-strand DNA (ssDNA) probes. The functionalization and hybridization processes are characterized by scanning electron microscopy (SEM) and fluorescence imaging microscopy. To avoid limitations from solution ions and neighborhood crosstalk, detection in the air was performed to improve the reliability of the measurement. Compared with measured results from the impedance analyzer, the proposed CMUTs-based biosensor has the ability to detect DNA molecules over a range of concentrations from 1 to 100 μmol∙L−1 with good sensitivity.
2. Materials and methods
2.1. CMUTs-based biosensors
When designing the CMUTs-based biosensors, a few typical characteristics were considered. For instance, CMUTs operate in the resonant mode by applying a DC bias and an alternating current (AC) voltage between the top and bottom electrodes. This provides an electrostatic force that causes the top-clamped membrane of the CMUTs to vibrate. Here, the CMUTs cell was regarded as a physical plate capacitor with nonlinear responses that contributed to substantial electrical characteristic changes such as the resonant frequency, impedance, and phase over the parallel-resonance region.
Therefore, the resonant frequency fmem of a single CMUT cell was estimated based on the following mechanical properties [35], [45]:where Rmem and tmem are the radius and thickness of the top membrane, respectively. E is the Young modulus, υ is the Poisson ratio, and ρ is the density. For silicon, E = 169 GPa, υ = 0.28, and ρ = 2332 kg∙m−3.
The collapse voltage Vcollapse of a CMUTs cell with no hydrostatic pressure applied to the top membrane was calculated as follows [53]:where d0 is the effective electrode distance, ε0 is the vacuum dielectric permittivity, and D is the flexural rigidity.
Two types of 2 MHz CMUT arrays composed of circular cells were designed and fabricated using low-temperature (350 °C) direct bonding technology described by a previous study [54], which is compatible with CMOS technology. The minimum cavity height, which was constrained by the fabrication process, was 400 nm. Based on Eqs. (2), (3), reducing the cavity height was beneficial for reducing the collapse voltage. Based on the above principles, the cavity height of the structure was designed to be 450 nm. The purpose of the SiO2 insulating cavity layer on the lower electrode is to prevent short circuits between the top and bottom electrodes. Considering the electric field breakdown strength, a thermal oxide process was employed for fabrication, and the designed thickness of the insulating layer was set to 100 nm. To ensure the uniformity of the top silicon film and the density of the insulating layer, the insulating layer thickness was designed to be 0.2 μm, utilizing the plasma-enhanced chemical vapor deposition (PECVD) process and the top SiO2 insulating layer was deposited with tetraethylorthosilicate (TEOS). The top Si membrane was fabricated using a silicon-on-insulator (SOI). To ensure bonding strength and prevent fracturing during the thinning process, the thickness of the top silicon film was set at 2 μm.
In this context, the CMUT cell radius was optimized by analyzing the trends of the collapse voltage and resonant frequency as they varied for different cell radii, as illustrated in Fig. 2. The cell radius was designed to fall within the range of 60 to 70 μm to accommodate the human safety threshold and substantial power consumption associated with high operating voltages, as the CMUTs must operate at voltages below 36 V and the resonant frequency must be high enough to maintain a high level of detection sensitivity. Therefore, fabricated 4.7 mm × 4.7 mm silicon chips were proposed, and these consisted of 30 × 30 circular sensory cells for CMUTs of 60 μm radius (type A) and 20 × 20 circular cells for CMUTs of 70 μm radius (type B). In the array design of CMUT-based biosensors, electrical parallelism increased the signal readout capability and motional impedance as the number of cells increased. Moreover, expanding the effective motional area helped mitigate the influence of parasitic effects, particularly the parasitic capacitance. Given that the fabrication area of each CMUT chip was limited to 4.7 μm × 4.7 μm, a sufficient edge distance (at least 200 μm between the outermost cell in the array and the scratch track), and an 8 μm edge-to-edge separation in silicon chips that contained 900 and 400 cells was achieved by designing the diameters of the top moveable membranes to 120 and 140 μm, respectively. The effectiveness of the mechanical and thermal noises benefited from such an array arrangement with a large number of cells because they were electrically connected in parallel. The detailed structural parameters are listed in Table 1.
According to Eq. (1), the resonant frequency shifts of the CMUTs exhibited a linear relationship with the mass change of the top membrane. Thus, by using a functionalized probe layer, a CMUTs-based device can detect target biomolecules. Differentiating Eq. (1), the mass sensitivity per unit area of the CMUTs-based biosensor Sm is defined by Eq. (4):where mmem is the mass of the top membrane of CMUTs, equaled ρAmemtmem, and the Amem is the area of the top membrane of CMUTs, equaling πR2mem.
Chip types A and B were subjected to Eq. (4) to determine the detection sensitivity and noise effectiveness, indicating that a low Sm value indicated high mass sensitivity and CMUTs with a low radius can obtain high mass sensitivity. Consequently, the theoretical mass sensitivity was estimated to be 4.043 ag∙Hz−1∙μm−2 for the designed CMUT structure with a radius of 60 μm, which was an excellent and valuable sensitivity. The electrical characteristics of the designed CMUT cell under different bias voltages were analyzed based on the parameters listed in Table S1 in Appendix A, as shown in Figs. S1 and S2 in Appendix A.
2.2. CMOS interface
The CMOS architecture of the ASIC interface for our CMUTs-based biosensor was fabricated with TSMC 0.18 μm HV BCD technology. Two core circuits: an HV charge pump and a feedback oscillator under a regular supply voltage of 1.8 V using Cadence IC6.1.7 (Cadence Design Systems Inc., USA) were integrated [55]. The topology of the HV charge pump was designed using HV p-channel metal-oxide-semiconductor (PMOS) transistors with an N+-buried layer to reduce the body effects of n-type metal-oxide-semiconductor (NMOS) transistors, resulting in an improved breakdown voltage of 50 V was achieved. Simultaneously, the designed structure was compact. A 50 MHz ring oscillator was buffered to provide digital clock signals for different pumping stages, control the output voltage ripples, and decrease the pumping periods.
A free-running oscillator with closed-loop feedback was investigated to determine the resonant frequency of a CMUTs-based biosensor composed of two-stage inverter-based amplifiers and high-pass filters [56]. To improve the open-loop gain, the inverters were operated in the on-off transition region with a resistive push–pull bias port for the first–stage inverter. This design ensured the maximum transconductance for analog signals and reduced the influence of process variation and transistor mismatch [35]. The open-loop gain AOP of the inverter was derived using Eq. (5):where gmN and gmP denote the transconductances of the NMOS and PMOS transistors, respectively; gdsN and gdsP denote the corresponding transconductances of the output resistance.
The introduction of CMUTs-based biosensors into feedback oscillators using a LEM with parasitic effects was established and a closed-loop topology was optimized to meet the Barkhausen criterion for the transfer function around the resonant frequency.
2.3. Experimental procedure
The experiments were performed at room temperature and pressure; SAM procedures were conducted under a fume hood due to the formation of toxic solvents. First, the CMUT surface was sequentially washed with acetone, absolute ethanol, methanol, and isopropanol for 15 min each to remove the protective glue from the top of the membrane. Deionized water (DI) was used for 5 min to clean the CMUTs surface, which was then dried with nitrogen (N2). By homogenizing with a shaker (MS 3 basic, IKA, Staufen, Germany) at 1000 r∙min−1 for 10 s, a 2 mmol∙L−1 SAM solution, with methanol as solvent, was then prepared with a mixture of 0.04 mmol∙L−1 (HO)2POS(CH2)11EG3OCH2COOH and 1.96 mmol∙L−1 (HO)2POS(CH2)11EG3OH. This mixture was included to address the issue that the DNA hybridization rate may diminish when the anchoring groups (COOH) are close to each other. Water molecules can form robust coordination bonds with the ethylene-glycol chains, resulting in a high degree of correlation between the solvated ions surrounding the ethylene-glycol chains. This coordination results in a densely packed crystalline helical conformation [57]. Micropipettes (Eppendorf SE, Germany; Joanlab, China) and an analytical balance (KERN ALJ160-4 NM, KERN, Germany) were used to control the concentration accurately during solution preparation. The electrodes were incubated overnight (16 h) with the SAM solution in 50 μL liquid phase deposited onto the CMUTs surface in dark conditions. A working temperature of 0 °C was used to ensure the quality of SAM formation by kinetic control, reducing thermal agitation, and minimizing desorption. At lower temperatures, the kinetic energy of the molecules was reduced, thus allowing them to combine and form ordered structures. Furthermore, low temperatures reduced the thermal agitation of molecules, allowing them to settle in an ordered arrangement, lower temperatures also minimize the desorption of molecules from the SAM, ensuring that the SAM remains intact and stable. After incubation, the electrodes were washed with methanol and DI water to remove unbound molecules, and then dried with N2. The ssDNA probe was diluted to 1 mmol∙L−1 with tris-ethylenediaminetetraacetic acid (TE) buffer (0.3 mol∙L−1 NaCl, pH 7.0), followed by homogenization using a shaker at 1000 r∙min−1 for 10 s. The pH of each solution was measured using a pH probe (Model 610, Orion, USA). The CMUTs surface with the SAM was incubated for 30 min in a 0.05 mol∙L−1N-hydroxysuccinimide (NHS) and 0.2 mol∙L−1N-(3-dimethylaminopropyl)-N'-ethylcarbodiimide hydrochloride (EDAC) water solution to ensure surface activation for efficient DNA anchoring [52]. ssDNA probes were processed for 5 min in a water bath at 75 °C, the probes were then immobilized onto the CMUTs surface and incubated for 2 h. The solution was then left cooled down to room temperature. The CMUTs surface was washed three times with TE buffer (0.3 mol∙L−1 NaCl, pH 7.0) to remove non-specific binding. Subsequently, it was washed with DI water to remove additional salts and dried with N2. Both complementary and non-complementary 30-mer (mer from the Greek meros) ssDNA were diluted in different solutions made by a series of concentrations from 1 to 100 μmol∙L−1 in TE buffer (0.3 mol∙L−1 NaCl, pH 7.0) and homogenized with a shaker. The 8 μL 30-mer ssDNA target solution was heated in a water bath at 75 °C and spread onto the CMUTs surface until it cooled down to room temperature (30 min). The devices were then washed using the above steps to remove non-specific targets and salt and finally dried with N2. The fluorescence imaging and frequency of the biofunctionalized CMUT surfaces were captured and measured, respectively, under dry conditions. Following another round of washing twice with heated TE buffer ((75 °C), once with DI water) it was dried with N2 and various concentrations of ssDNA oligonucleotides were repeatedly measured and characterized. All chemicals used were of analytical reagent grade, as shown in Table S2 in Appendix A.
3. Results and discussion
3.1. CMUTs characterization
Based on the results, chip type A achieved higher detection sensitivity and noise effectiveness than those of type B. The arrangement of the top electrodes of type A was reported in Fig. 3(a). Characterization of the collapse deformation of the CMUTs membrane and the statistics of the CMUTs structural morphology are shown in Fig. S3 and Table S3 in Appendix A, respectively. The silicon chips were bonded to the test printed circuit board (PCB) using a conductive epoxy and a wire bonder for the bottom and top electrodes, respectively. It was observed that the larger CMUTs array provided a flatter and more uniform Al surface for bio-functionalization. Furthermore, a 5 μL droplet was sufficient enough to cover the sensitive area. Fig. 3(b) showed a cross-sectional micrograph of a CMUT cell captured using SEM (Hitachi SU8010, Hitachi, Japan), indicating the structural distribution. A nine-point measurement procedure was employed to analyze the statistical values of the fabricated CMUT structure. Typically, the measured average thickness of the top electrodes was 412 nm with a maximum size error of 3%. The measured average vacuum cavity height was 462 nm with a relative error of 2.67%. Hence, the direct bonding technology ensured reliable fabrication. Fig. 3(c) showed the energy dispersive X-ray (EDX) analysis of the CMUTs under different electron high tension (EHT) from 5 to 15 kV using a SEM (Zeiss GeminiSEM 450, Zeiss, Germany). With an increase in EHT, the intensities of the Al, Si, and titanium (Ti) increased in the EDX analysis. This is consistent with the material distribution in the depth direction of the CMUTs. Concurrently, Ti and TiN were found to have a high EHT because they served as adhesion layers for the electrodes. In contrast, the intensities of elements including C, N, and O were observed without apparent changes, indicating that most of these elements were active on the surface. The electrical impedance and phase responses of the CMUTs were characterized using an impedance analyzer (Agilent 4294A, Agilent, USA), as shown in Figs. 3(d-i) and (d-ii), respectively. A wedge-and-ball bonder (HB10, TPT Wire Bonder GmbH & Co. KG, Germany) was used to wire bond the CMUTs chips. The CMUTs resonated around the parallel-resonance region with a series fs and parallel fp resonant frequencies. A five-element LEM was used to fit the measured characteristics in the resonance region [58]. As shown in the model in Fig. 3(e), Rp represents the series of parasitic resistances, C0 is the physical capacitance, including the parallel parasitic capacitances, and the resonant arm is composed of the equivalent motional loss R, stiffness C, and mass L. The CMUTs resonated at fp = 2.222 MHz with a quality factor of 70.41, which was calculated by Eq. (6):
Compared with the measured parallel resonant frequency of 2.229 MHz shown in Fig. 3(d), the relative error of the proposed LEM was calculated to be 0.314%. Moreover, Eq. (6) also revealed fp was advantageous as the working frequency for biosensing, partially because it yielded a higher quality factor and resonant frequency than fs, which reduced the noise caused by cell-to-cell nonuniformity. Furthermore, the relationship between the frequency and mass changes was calculated by differentiating Eq. (6)
Eq. (7) demonstrated that frequency changes in fp related to mass changes in mmem had a linear multiple relationship with changes in fs related to mmem when additional mass changes did not affect the stiffness and parallel capacitance. Therefore, fp was selected as the working frequency for label-free biosensing.
3.2. SAM
Figs. 4(a) and (b) represent micrographs before and after the SAM process, comparing the surface morphology of the CMUTs before and after functionalization with ethylene-glycol alkanethiols (also called thiol-phosphate acid). After washing and drying, the surface of the Al electrode was observed to be clean, which was necessary for the incubation of the SAM layer, as shown in Fig. 4(a-i). A magnified micrograph of Fig. 4(a-ii) Al electrodes still contained glue due to the microscale basin, which was difficult to remove. However, it did not the affect process of SAM as it was present in a non-sensitive region. However, as shown in Fig. 4(b-i), a better SAM was achieved as presented by good cell-to-cell uniformity. The SAM molecules were also deposited on the SiO2 surface [59], as shown in Fig. 4(b-ii). However, these regions did not contribute to the vibration and frequency changes. Fig. 4(b-iii) reported a magnified image of the morphology of the Al electrode after functionalization compared with Fig. 4(a-iii), resulting in a uniform covering over the metal grain boundary, which was a standard physical behavior.
The energy intensities of the three CMUT cells (Fig. 4(c)) demonstrated similar elemental distributions. Compared to the energy intensity map in Fig. 3(c), the intensities of carbon (C) of these three cells (cell-1 to cell-3 in Fig. 4(b-i) demonstrated an apparent increase and became larger than that of nitrogen element (N), indicating that the SAM was successfully deposited on the Al surface with good consistency. Fig. 4(d) depicts the chemical binding of thiol phosphate acid to the CMUT surface. For the Lewis acid surface, the coordination of phosphoryl oxygen to a Lewis acidic site, followed by heterocondensation with electrophilic phosphorus, contributes to the formation of a robust tridentate binding state [60], [61], as shown in Fig. 4(d-i). For a surface with a poor Lewis acid, a bidentate-bound state is established by heterocondensation with two surface hydroxyl groups and hydrogen bonding between the phosphoryl and surface hydroxyl groups [62], as shown in Fig. 4(d-ii). Although alumina is present on Al electrodes in natural environments, both can occur during the SAM deposition.
3.3. Fluorescence imaging
For the complementary ssDNA targets, the fluorescence intensity increased with increasing concentrations of ssDNA, as shown in Figs. 5(a-i)-5(a-v). Simultaneously, the fluorescence intensity was higher in the SiO2 regions, corresponding to the results to the findings after the SAM process, indicating that ethylene-glycol was deposited in the basin region among the CMUT cells. However, they cannot cause frequency shifts that react to mass changes in such regions, because they are not sensing functions. The CMUTs array with 70 μm radius cells were further exposed to non-complementary ssDNA targets to investigate improved selectivity. Fig. 5(a-vi) showed the fluorescence image obtained with 100 μmol∙L−1 non-complementary ssDNA and demonstrated that the fluorescence in this case was completely absent, but in the edge-region of the rings, particularly at the edge between CMUTs cell and bottom SiO2. This indicates that the non-complementary ssDNA had low efficiency in binding to the immobilized probes, demonstrating improved selectivity due to ethylene-glycol alkanethiols.
By isolating the fluorescent influence of the cell edges and improving effective fluorescent intensity, the averaged fluorescent area ratios were estimated by calculating the proportion of fluorescence in every CMUT cell by a defined 54 μm radius with a minimum color threshold of 20 for each fluorescent figure. The effective statistical area decreased for cells near the fluorescence edge. For complementary ssDNA, the area ratios and error bars were discovered to increase with the rising of concentrations of ssDNA, such as, 6.34% ± 2.67% with 1 μmol∙L−1 and 40.40% ± 12.51% for 100 μmol∙L−1, as shown in Fig. 5(b), and spots appeared to be an aggregation of ssDNA were observed under high concentrations. In the case of non-complementary ssDNA, the maximum area ratio approached 0.10% ± 0.05%, which was decreased by 99.75% compared with that of 100 μmol∙L−1 complementary ssDNA, as depicted in Fig. 5(c). Moreover, the substantial difference observed between the area ratios of cDNA with respect to those of the non-cDNA (e.g., 40.40% versus only 0.10%, respectively, at 100 μmol∙L−1) demonstrated excellent selectivity due to the ethylene-glycol functional groups.
3.4. CMOS test and frequency shifts
The CMOS interface was connected to the CMUTs sensor by two ports: One provided a DC bias voltage through resistor RHV, and the other was coupled to the feedback oscillator through capacitor COSC, as shown in Fig. 6(a). The ASIC interface was bonded to a test PCB designed by Altium Designer 2022 (Altium, Australia), as shown in Figs. 6(b-i) and 6(b-ii), which was packaged with a quad-flat no-lead package (QFN) with 64 pins (9 mm × 9 mm), as shown in Figs. 6(a-iii) and 6(a-iv). The bare chip occupied an area of 1.897 mm × 1.897 mm, as shown in Fig. 5(a-v). For the electrical test, power was provided by a supplier (Keysight E3646A, Keysight, USA) with a mixed-signal oscilloscope (MSO-S 804A, Keysight, USA) to output the signals and a universal frequency counter (53220A, Keysight, USA) to detect the frequency shifts of the CMUTs chip.
Fig. 6(c) showsed the measurement results for the startup voltage of the fabricated PMOS-based HV charge-pump circuit. Fig. 6(d) reported the transient voltage ripple for the HV output signals from 30 to 40 ms, and an average voltage output of 21.37 V was measured. To evaluate the stability of the on-chip HV charge pump, the transient signals from 15 to 45 ms were divided into six groups with a spacing of 5 ms to calculate the ripple level, as shown in Fig. 6(e). The maximum ripple voltages were observed to be 149.96 mV. The startup period for the oscillator outputs was less than 1 ms, as shown in Fig. 6(f). The steady-state waveforms were reported in Fig. 6(g), demonstrating that the system operated at a fundamental frequency of 2.146 MHz with a peak-to-peak amplitude of 1.891 V. Overlapping Allan deviation σy was used to estimate the sensor noise and short-term stability, which was calculated for multiples of averaging time τ0 combining m sequential measurements with a total number of samples N, as in Eq. (8) [63]:where is the average frequency of m samples, and i and j are integers. The minimum Allan deviation (1σ) of 0.27 kHz was observed at an averaging time of 0.1 s, which was computed from the data collected at a sampling rate of 200 Hz, as shown in Fig. 6(h). The system dissipates less than 3.8 mW. In particular, HV boosting and the oscillator consumed 3.6 mW and 183.7 μW, respectively.
Distinctive frequency shifts in the CMUTs-based biosensors were observed by varying the concentration of complementary ssDNA. Exact measurements were conducted using an external impedance analyzer (Fig. 6(i)) and internal ASIC oscillators (Fig. 6(j)). In particular, the resonant frequency decreased with an increase in mass-loading effects related to ssDNA concentration, as measured after washing under dry conditions. The resonant frequency returns with a slight shift after melting and drying (Fig. 6(i)). The average frequency shifts measured by the impedance analyzer were computed based on four measurements, resulting in good stability with low error bars, as shown in Fig. 6(k). Although the average frequency shifts measured using the ASIC interface exhibited large variations, similar shifting trends were observed for both methods. This frequency noise was related to the parasitic effects of the system. The detection sensitivity calculated using the frequency shifts measured by the impedance analyzer was 9.533 nmol∙L−1∙Hz−1 with a linearity of 97.94%. In contrast, it was 7.943 nmol∙L−1∙Hz−1 for the CMOS interface with a linearity of 97.45%, corresponding to a relative error of 16.68%. Simultaneously, these results demonstrated that an experimental mass detection sensitivity of 0.061 fg∙Hz−1∙μm−2 was achieved using the proposed CMUTs-based biosensor over the range of concentrations from 1 to 100 μmol∙L−1. The limit of detection (LOD) was calculated to be 0.635 μmol∙L−1. It was not surprising that the sensitivity was higher when using the integrated CMUTs-based biosensor with respect to the measurements required with the external impedance analyzer of the laboratory. Nevertheless, a detection sensitivity of 0.061 fg∙Hz−1∙μm−2 was achieved with the proposed CMUTs-based biosensor integrated with on-chip oscillators and the high-voltage supplier, was extremely good and a significant step ahead concerning the present literature.
3.5. Possibility of further improvements
Compared with the CMUTs-based biosensors previously presented in the literature, this work pioneered the investigation of the possibility of label-free detection of DNA with improved selectivity, owing to ethylene-glycol alkanethiol, as a new approach for the effective functionalization of Al electrodes for CMUTs arrays. Furthermore, the developed CMOS/ASIC interface with on-chip HV boosting and a local oscillator achieved extremely good detection sensitivity for a fully integrated CMOS solution with a relatively low working frequency (Table 2) [35], [49], [50], [65], [63].
However, the oscillator stability is often affected by Johnson noise, noise from circuit components, random vibrations, shot noise, and noise from power sources [66], [67]. In future, this issue will be investigated in depth to achieve higher-resolution sensing. Several other factors should also be considered. First, noise from circuit components can be reduced by improving the integration level to achieve CMUTs-on-ASIC devices through wafer-bonding integration and the through-silicon via (TSV) process [68]. Furthermore, for the CMUTs chip, improving the wafer-bonding reliability and reducing the residual stress can enhance the consistency of the cell-to-cell vibration [69]. In addition, large parallelism with more cells further reduces system noise owing to the high quality factor and low motional impedance [70], [71].
4. Conclusions
Here, a label-free biosensor for detecting DNA based on 2 MHz CMUTs bonded to a CMOS front-end interface operated in air and dry conditions was reported. To date, a new functionalization method based on SAMs of ethylene-glycol alkanethiols has been used to immobilize single-stranded DNA on the Al electrodes of the CMUTs array. The experimental results demonstrated the good consistency and uniformity of the SAM and improved selectivity owing to the presence of ethylene-glycol functional groups. An application-specific integrated CMOS circuit was developed to ensure an on-chip HV charge pump and feedback oscillator for a total power dissipation of less than 3.8 mW and an area of only 3.6 mm2. The oscillator demonstrated an experimental Allan deviation of 0.81 kHz (3σ) and successfully detected ssDNA oligonucleotides over a range of concentrations from 1 to 100 μmol∙L−1. The resulting mass detection sensitivity was up to 0.061 fg∙Hz−1∙μm−2 with detection limit equal to 6.243 μg∙mL−1. Thus, these investigations demonstrated that CMUTs are promising devices for detecting low DNA concentrations. In the future, a step forward in miniaturization and reliability will be achieved through the direct integration of CMUTs-on-ASICs.
Acknowledgments
This work was supported by the National Key Research and Development Program of China (2022YFB3205400), the National Natural Science Foundation of China (52275570), the Postdoctoral Innovation Talents Support Program (BX20230288), and the Postdoctoral Science Foundation of Shaanxi Province (2018BSHEDZZ08). We would like to thank the EPFL BioImaging & Optics Core Facility and in particular Arne Seitz and José Artacho for their assistance in fluorescence imaging.
Compliance with ethics guidelines
Zhikang Li, Yihe Zhao, Gian Luca Barbruni, Jie Li, Zixuan Li, Jiawei Yuan, Ping Yang, Libo Zhao, Zhuangde Jiang, and Sandro Carrara declare that they have no conflict of interest or financial conflicts to disclose.
J.Daniels, S.Wadekar, K.DeCubellis, G.W.Jackson, A.S.Chiu, Q.Pagneux, et al. A mask-based diagnostic platform for point-of-care screening of COVID-19. Biosens Bioelectron, 192 (2021), Article 113486.
[2]
J.Chen, C.Hu, L.Chen, L.Tang, Y.Zhu, X.Xu, et al. Clinical study of mesenchymal stem cell treatment for acute respiratory distress syndrome induced by epidemic influenza A (H7N9) infection: a hint for COVID-19 treatment. Engineering, 6 (10) (2020), pp. 1153-1161.
[3]
M.D.Baaske, M.R.Foreman, F.Vollmer. Single-molecule nucleic acid interactions monitored on a label-free microcavity biosensor platform. Nat Nanotechnol, 9 (11) (2014), pp. 933-939.
[4]
J.Mertens, C.Rogero, M.Calleja, D.Ramos, J.A.Martín-Gago, C.Briones, et al. Label-free detection of DNA hybridization based on hydration-induced tension in nucleic acid films. Nat Nanotechnol, 3 (2008), pp. 301-307.
[5]
H.Altug, S.H.Oh, S.A.Maier, J.Homola. Advances and applications of nanophotonic biosensors. Nat Nanotechnol, 17 (1) (2022), pp. 5-16.
[6]
Y.Du, B.Li, E. Wang. “Fitting” makes “Sensing” simple: label-free detection strategies based on nucleic acid aptamers. Acc Chem Res, 46 (2) (2013), pp. 203-213.
[7]
B.G.Andryukov, N.N.Besednova, R.V.Romashko, T.S.Zaporozhets, T.A.Efimov. Label-free biosensors for laboratory-based diagnostics of infections: current achievements and new trends. Biosensors, 10 (2) (2020), p. 11.
[8]
E.H.Koh, W.C.Lee, Y.J.Choi, J.I.Moon, J.Jang, S.G.Park, et al. A wearable surface-enhanced Raman scattering sensor for label-free molecular detection. ACS Appl Mater Interfaces, 13 (2) (2021), pp. 3024-3032.
[9]
B.Yin, W.K.H.Ho, Q.Zhang, C.Li, Y.Huang, J.Yan, et al. Magnetic-responsive surface-enhanced Raman scattering platform with tunable hot spot for ultrasensitive virus nucleic acid detection. ACS Appl Mater Interfaces, 14 (3) (2022), pp. 4714-4724.
[10]
J.Yoon, M.Shin, J.Y.Lee, S.N.Lee, J.H.Choi, J.W.Choi. RNA interference (RNAi)-based plasmonic nanomaterials for cancer diagnosis and therapy. J Control Release, 342 (2022), pp. 228-240.
[11]
X.Zhang, Y.Liu, S.Du, Y.Yin, L.Kong, Y.Chang, et al. Engineering a rolling-circle strand displacement amplification mediated label-free ultrasensitive electrochemical biosensing platform. Anal Chem, 93 (27) (2021), pp. 9568-9574.
[12]
C.Stagni, C.Guiducci, L.Benini, B.Ricco, S.Carrara, B.Samori, et al. CMOS DNA sensor array with integrated A/D conversion based on label-free capacitance measurement. IEEE J Solid-State Circuits, 41 (12) (2006), pp. 2956-2964.
[13]
J.Zhou, X.Zhao, G.Huang, X.Yang, Y.Zhang, X.Zhan, et al. Molecule-specific terahertz biosensors based on an aptamer hydrogel-functionalized metamaterial for sensitive assays in aqueous environments. ACS Sens, 6 (5) (2021), pp. 1884-1890.
[14]
R.Zhou, C.Wang, Y.Huang, K.Huang, Y.Wang, W.Xu, et al. Label-free terahertz microfluidic biosensor for sensitive DNA detection using graphene-metasurface hybrid structures. Biosens Bioelectron, 188 (2021), Article 113336.
[15]
K.Noi, K.Ikenaka, H.Mochizuki, Y.Goto, H.Ogi. Disaggregation behavior of amyloid β fibrils by anthocyanins studied by total-internal-reflection-fluorescence microscopy coupled with a wireless quartz-crystal microbalance biosensor. Anal Chem, 93 (32) (2021), pp. 11176-11183.
[16]
S.Yousuf, J.Kim, A.Orozaliev, M.S.Dahlem, Y.A.Song, J.Viegas. Label-free detection of morpholino-DNA hybridization using a silicon photon suspended slab micro-ring resonator. IEEE Photon J, 13 (4) (2021), pp. 1-9.
[17]
R.Ghayoor, S.Zangenehzadeh, A.Keshavarz. Design of high-sensitivity surface plasmon resonance sensor based on nanostructured thin films for effective detection of DNA hybridization. Plasmonics, 17 (4) (2022), pp. 1831-1841.
[18]
S.Anwar, M.B.Khawar, M.Ovais, A.Afzal, X.Zhang. Gold nanocubes based optical detection of HIV-1 DNA via surface enhanced Raman spectroscopy. J Pharm Biomed Anal, 226 (2023), Article 115242.
[19]
A.Lomae, P.Preechakasedkit, O.Hanpanich, T.Ozer, C.S.Henry, A.Maruyama, et al. Label free electrochemical DNA biosensor for COVID-19 diagnosis. Talanta, 253 (2023), Article 123992.
[20]
P.H.Lai, L.S.Tseng, C.M.Yang, M.S.C.Lu. Design and characterization of a 16×16 CMOS capacitive DNA sensor array. IEEE Sens J, 23 (8) (2023), pp. 8120-8127.
[21]
N.Li, F.Zhang. THz-PCR based on resonant coupling between middle infrared and DNA carbonyl vibrations. ACS Appl Mater Interfaces, 15 (6) (2023), pp. 8224-8231.
[22]
A.Salim, S.Lim. Recent advances in noninvasive flexible and wearable wireless biosensors. Biosens Bioelectron, 141 (2019), Article 111422.
[23]
G.Balakrishnan, J.Song, C.Mou, C.J.Bettinger. Recent progress in materials chemistry to advance flexible bioelectronics in medicine. Adv Mater, 34 (10) (2022), Article 2106787.
[24]
Y.Yang, W.Gao. Wearable and flexible electronics for continuous molecular monitoring. Chem Soc Rev, 48 (6) (2019), pp. 1465-1491.
[25]
Y.Ma, Y.Zhang, S.Cai, Z.Han, X.Liu, F.Wang, et al. Flexible hybrid electronics for digital healthcare. Adv Mater, 32 (15) (2020), Article 1902062.
[26]
J.Gao, K.Shang, Y.Ding, Z.Wen. Material and configuration design strategies towards flexible and wearable power supply devices: a review. J Mater Chem A, 9 (14) (2021), pp. 8950-8965.
[27]
Q.He, Y.Cheng, Y.Deng, F.Wen, Y.Lai, H.Li. Conductive hydrogel for flexible bioelectronic device: current progress and future perspective. Adv Funct Mater, 34 (1) (2023), Article 2308974.
[28]
P.J.Jandas, K.Prabakaran, J.Luo, M.G.Holaday. Effective utilization of quartz crystal microbalance as a tool for biosensing applications. Sens Actuator A Phys, 331 (2021), Article 113020.
[29]
R.M.R.Pinto, V.Chu, J.P.Conde. Label-free biosensing of DNA in microfluidics using amorphous silicon capacitive micro-cantilevers. IEEE Sens J, 20 (16) (2020), pp. 9018-9028.
[30]
J.Zhou, D.Zhang, Y.Liu, F.Zhuo, L.Qian, H.Li, et al. Record-breaking frequency of 44 GHz based on the higher order mode of surface acoustic waves with LiNbO3/SiO2/SiC heterostructures. Engineering, 20 (2023), pp. 112-119.
[31]
M.Bharati, L.Rana, R.Gupta, A.Sharma, P.K.Jha, M.Tomar. Realization of a DNA biosensor using inverted Lamb wave MEMS resonator based on ZnO/SiO2/Si/ZnO membrane. Anal Chim Acta, 1249 (2023), Article 340929.
[32]
Y.Zhang, J.Luo, A.J.Flewitt, Z.Cai, X.Zhao. Film bulk acoustic resonators (FBARs) as biosensors: a review. Biosens Bioelectron, 116 (2018), pp. 1-15.
[33]
F.Pop, B.Herrera, M.Rinaldi. Lithium niobate piezoelectric micromachined ultrasonic transducers for high data-rate intrabody communication. Nat Commun, 13 (1) (2022), p. 1782.
[34]
E.Şennik, F.Erden, N.Constantino, Y.Y.Oh, R.A.Dean, Ö.Oralkan. Electronic nose system based on a functionalized capacitive micromachined ultrasonic transducer (CMUT) array for selective detection of plant volatiles. Sens Actuators B Chem, 341 (2021), Article 130001.
[35]
S.Park, I.Yoon, S.Lee, H.Kim, J.W.Seo, Y.Y.Chung, et al. CMUT-based resonant gas sensor array for VOC detection with low operating voltage. Sens Actuators B Chem, 273 (2018), pp. 1556-1563.
D.Virzonis, G.Vanagas, A.Ramanaviciene, A.Makaraviciute, D.Barauskas, A.Ramanavicius, et al. Resonant gravimetric immunosensing based on capacitive micromachined ultrasound transducers. Mikrochim Acta, 181 (13-14) (2014), pp. 1749-1757.
[38]
W.Pang, H.Zhao, E.S.Kim, H.Zhang, H.Yu, X.Hu. Piezoelectric microelectromechanical resonant sensors for chemical and biological detection. Lab Chip, 12 (1) (2012), pp. 29-44.
[39]
M.H.Zhao, Z.L.Wang, S.X.Mao. Piezoelectric characterization of individual zinc oxide nanobelt probed by piezoresponse force microscope. Nano Lett, 4 (4) (2004), pp. 587-590.
[40]
I.L.Guy, S.Muensit, E.M.Goldys. Extensional piezoelectric coefficients of gallium nitride and aluminum nitride. Appl Phys Lett, 75 (26) (1999), pp. 4133-4135.
[41]
K.Brenner, A.S.Ergun, K.Firouzi, M.F.Rasmussen, Q.Stedman, B.Khuri-Yakub. Advances in capacitive micromachined ultrasonic transducers. Micromachines, 10 (2) (2019), p. 152.
[42]
D.Barauskas, S.J.Park, D.Pelenis, G.Vanagas, J.J.Lee, D.Viržonis, et al. CO2 and SO2 interactions with methylated poly(ethylenimine)-functionalized capacitive micromachined ultrasonic transducers (CMUTs): gas sensing and degradation mechanism. ACS Appl Mater Interfaces, 1 (7) (2019), pp. 1150-1161.
[43]
Z.Zheng, N.Kim, W.S.Wong, J.T.W.Yeow. Inkjet-printed CMUT humidity sensors with high sensitivity and low hysteresis. Sens Actuators B Chem, 327 (2021), Article 128920.
[44]
Z.Zheng, Y.Yao, Y.H.Sun, J.T.W.Yeow. Development of a highly sensitive humidity sensor based on the capacitive micromachined ultrasonic transducer. Sens Actuators B Chem, 286 (2019), pp. 39-45.
[45]
L.Zhao, Y.Zhao, Y.Xia, Z.Li, J.Li, J.Zhang, et al. A novel CMUT-based resonant biochemical sensor using electrospinning technology. IEEE Trans Ind Electron, 66 (9) (2019), pp. 7356-7365.
[46]
M.M.Mahmud, C.Seok, X.Wu, E.Şennik, A.Ö.Biliroğlu, O.J.Adelegan, et al. A low-power wearable E-nose system based on a capacitive micromachined ultrasonic transducer (CMUT) array for indoor VOC monitoring. IEEE Sens J, 21 (18) (2021), pp. 19684-19696.
[47]
C.Seok, M.M.Mahmud, M.Kumar, O.J.Adelegan, F.Y.Yamaner, Ö.Oralkan. A low-power wireless multichannel gas sensing system based on a capacitive micromachined ultrasonic transducer (CMUT) array. IEEE Internet Things J, 6 (1) (2019), pp. 831-843.
[48]
H.J.Lee, K.K.Park, M.Kupnik, Ö.Oralkan, B.T.Khuri-Yakub. Chemical vapor detection using a capacitive micromachined ultrasonic transducer. Anal Chem, 83 (24) (2011), pp. 9314-9320.
[49]
MahmudMM, ReeseH, JoshipuraA, SeokC, YamanerFY, DanieleM, et al. Gravimetric biosensor based on a capacitive micromachined ultrasonic transducer functionalized with peptide ligands. In: Proceedings of the 2017 IEEE Sensors; 2017 Oct 29-Nov 1; Glasgow, UK. Piscataway: IEEE; 2017. p. 1-3.
[50]
LeeS, EomG, YoonI, ParkS, KookG, KimMK, et al. Capacitivemicromachined ultrasonic transducer CMUT-based biosensor for detection of low concentration neuropeptide. Proceedingsof the 2018 40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society EMBC; 2018Jul 18-21; HonoluluHI, USA. Piscataway: IEEE; 2018. p. 2897-900.
[51]
S.Mariani, V.Robbiano, L.M.Strambini, A.Debrassi, G.Egri, L.Dähne, et al. Layer-by-layer biofunctionalization of nanostructured porous silicon for high-sensitivity and high-selectivity label-free affinity biosensing. Nat Commun, 9 (1) (2018), p. 5256.
[52]
S.Carrara, L.Benini, V.Bhalla, C.Stagni, A.Ferretti, A.Cavallini, et al. New insights for using self-assembly materials to improve the detection stability in label-free DNA-chip and immuno-sensors. Biosens Bioelectron, 24 (12) (2009), pp. 3425-3429.
[53]
Z.K.Li, L.B.Zhao, Z.D.Jiang, Z.Y.Ye, Y.L.Zhao. An improved method for the mechanical behavior analysis of electrostatically actuated microplates under uniform hydrostatic pressure. J Microelectromech Syst, 24 (2) (2015), pp. 474-485.
[54]
L.Zhao, J.Li, Z.Li, J.Zhang, Y.Zhao, J.Wang, et al. Fabrication of capacitive micromachined ultrasonic transducers with low-temperature direct wafer-bonding technology. Sens Actuators A Phys, 264 (2017), pp. 63-75.
[55]
ZhaoY, BarbruniGL, LiZ, ZhaoL, JiangZ, EnzC, et al. AnASIC interface for CMUTs-based biosensors with high voltage boosting and oscillator. USA. Proceedings of the 2022 IEEE International Symposium on Circuits and Systems (ISCAS); 2022 May 27-Jun 1; Austin, TX, Piscataway: IEEE; 2022. p. 1625-9.
[56]
B.Razavi. The transimpedance amplifier [a circuit for all seasons]. IEEE Solid-State Circuits Maga, 11 (1) (2019), pp. 10-97.
[57]
CarraraS. Co-design of a full bio/CMOS interface. In: CarraraS, editor. Bio/CMOSinterfaces and co-design. Cham: Springer; 2023. p. 431-53.
[58]
Y.Zhao, Z.Li, X.Wang, G.Luo, J.Li, Z.Li, et al. New insights for parasitic effects of label-free biosensors based on capacitive micromachined ultrasonic transducers. IEEE Sens J, 22 (21) (2022), pp. 20575-20584.
[59]
S.A.Paniagua, A.J.Giordano, O.L.Smith, S.Barlow, H.Li, N.R.Armstrong, et al. Phosphonic acids for interfacial engineering of transparent conductive oxides. Chem Rev, 116 (12) (2016), pp. 7117-7158.
[60]
P.J.Hotchkiss, S.C.Jones, S.A.Paniagua, A.Sharma, B.Kippelen, N.R.Armstrong, et al. The modification of indium tin oxide with phosphonic acids: mechanism of binding, tuning of surface properties, and potential for use in organic electronic applications. Acc Chem Res, 45 (3) (2012), pp. 337-346.
[61]
K.Nakamura, T.Takahashi, T.Hosomi, Y.Yamaguchi, W.Tanaka, J.Liu, et al. Surface dissociation effect on phosphonic acid self-assembled monolayer formation on ZnO nanowires. ACS Omega, 7 (1) (2022), pp. 1462-1467.
[62]
N.S.Tambe, B.Bhushan. Nanotribological characterization of self-assembled monolayers deposited on silicon and aluminium substrates. Nanotechnology, 16 (9) (2005), pp. 1549-1558.
[63]
D.W.Allan. Time and frequency (time-domain) characterization, estimation, and prediction of precision clocks and oscillators. IEEE Trans Ultrason Ferroelectr Freq Control, 34 (6) (1987), pp. 647-654.
[64]
S.Eaimkhong, M.Steiert, T.F.Harper, M.D.Cable, J.Gimzewski. Label-free biodetection using capacitive micromachined ultrasonic transducers (CMUTs) and its application for cardiovascular disease diagnostics. J Nanomed Nanotechnol, 3 (5) (2012), Article 100014.
[65]
H.J.Lee, K.K.Park, Ö.Oralkan, M.Kupnik, B.T.Khuri-Yakub. A multichannel oscillator for a resonant chemical sensor system. IEEE Trans Ind Electron, 61 (10) (2014), pp. 5632-5640.
B.Razavi. Design of analog CMOS integrated circuits. McGraw-Hill, New York (2000).
[68]
C.Seok, O.J.Adelegan, A.Ö.Biliroğlu, F.Y.Yamaner, Ö.Oralkan. A wearable ultrasonic neurostimulator—part II: a 2D CMUT phased array system with a flip-chip bonded ASIC. IEEE Trans Biomed Circuits Syst, 15 (4) (2021), pp. 705-718.
[69]
C.D.Herickhoff, R. vanSchaijk. CMUT technology developments. Z Med Phys, 33 (3) (2023), pp. 256-266.
[70]
LeeHJ, ParkKK, CristmanP, OralkanÖ, KupnikM, Khuri-YakubBT. The effect of parallelism of CMUT cells on phase noise for chem/bio sensor applications. In: Proceedings of 2008 IEEE Ultrasonics Symposium; 2008 Nov 2-5; Beijing, China. Piscataway: IEEE; 2008. p. 1951-4.
[71]
ParkKK, LeeHJ, CrismanP, KupnikM, OralkanÖ, Khuri-YakubBT. Optimum design of circular CMUT membranes for high quality factor in air. In: Proceedings of 2008 IEEE Ultrasonics Symposium; 2008 Nov 2-5; Beijing, China. Piscataway: IEEE; 2008. p. 504-7.
PDF
(4407KB)
