Advertisement

AC Electrokinetics-Enhanced Capacitive Virus Detection

  • Cheng Cheng
  • Jayne WuEmail author
Living reference work entry
  • 17 Downloads

Abstract

The occurrence and spread of various serious viral outbreaks in pandemic dimension is one of the few major threats to public health and life worldwide nowadays. A disease diagnosis method that is sufficiently sensitive, rapid, and compact for point-of-care (POC) detection, and requires minimal sample pretreatment is highly desired. Hence, developing such a detection method for POC detection of viral disease is of the utmost significance for medical healthcare. Among different types of biosensors, capacitive biosensors, which fall into the category of electroanalytical biosensors, stand out and have shown great success due to their excellent performances and high potential to be developed as easy-to-handle devices. Capacitive sensors reported for virus detection are a type of surface-based affinity sensors. Specific binding reactions occurring on the sensor surface will cause changes in the dielectric properties or thickness of the interfacial layer at the electrolyte-electrode interface. By measuring the sensor’s capacitance, the sensor can convert the quantity of target analytes, such as antigen, or nucleic acids from virus, or antibody from patient serum, into readable outputs.

This chapter firstly provides some background information on the significance of capacitive virus detection and briefly reviews the current status that scientists and researchers have accomplished in developing POC tests of viral disease (section “Introduction”). Section “Capacitive Biosensors” briefly presents various receptors that can be utilized and the sensing mechanism. Section “AC Electrokinetic (ACEK) Enrichment” introduces alternating current electrokinetic (ACEK) effects that can be incorporated into capacitive sensing. Last, sections “Sensor Designs” and “Sensor Performances” present sensor electrode characterization and sensor performances, respectively. Sensor characterization methods include equivalent circuit extraction and fitting with respect to the electrode cell’s impedance spectrum, electrodes surface treatment, and data acquisition and analysis. The discussion of sensor performance includes the optimization of buffer solutions, electrical signals, and sensor specificity, sensitivities, etc.

Keywords

Capacitive biosensor AC electrokinetic Point-of-care Influenza A Human herpesvirus-1 Zika virus Virus Nucleic acids 

Introduction

Expansion of human activities and process of globalization speed up the pace of people’s life and interactions, as well as the spread of pathogens. Recent years have witnessed several serious viral outbreaks such as H5N1 bird flu, Zika virus, and Ebola virus. These recent outbreaks rapidly spread around the world to become pandemics with devastating effects on populations. An on-site screening tool for viral infections will be very helpful to control the spread of such viral outbreaks. Nowadays, biochemical diagnostics is either performed using sophisticated, expensive laboratory equipment capable of accurate measurement of complex biological interactions and constitutes, or by easy-to-handle, portable device for use by nonspecialists for decentralized, on site, or home analysis. The former are expensive and the latter are mass produced and inexpensive, and often referred to as biosensors. The best-known biosensor is glucose sensor. Professor Clark has been considered as the “father of biosensors” and the modern-day glucose sensor used daily by millions of diabetic patients is based on his research. The field of biosensor has attracted a great amount of research and development efforts since Professor Leland C. Clark’s monumental paper on the development of the first enzyme glucose oxidase biosensor in 1962 (Clark and Lyons 1963). Researchers and scientists from the fields of chemistry, physics, microbiology, and various disciplines of engineering have been deeply involved in this interdisciplinary field. There have been remarkable development and progress of sophisticated and accurate traditional laboratory-based biodetection instruments in the last three decades.

For virus detection and quantification, traditional and the most commonly used methods as of now include measurements of viral infectivity (such as viral plague assay), detection of viral proteins and nucleic acid (such as immunoblotting), and direct counting of virions (such as viral flow cytometry and transmission electron microscopy) (Pankaj 2013). For biosensors, based on the sensing targets, virus detection can be achieved by detecting the nucleic acid extracted from virions or detecting the virions themselves. For example, polymerase chain reaction (PCR) is one of the most widely used laboratory methods for detection of viral nucleic acids. This method is able to determine viral DNA inside virions. For RNA virus, such as Zika viruses, by converting RNA into DNA at the very first step, PCR can also be used to determine viral RNA. This method is known as reverse transcriptase PCR (RT-PCR). PCR has excellent sensitivity and specificity. However, this method needs several hours to yield results, with moderate requirements on sample treatment, facilities, and cost. An alternative category of sensing methods with less cost are immunoassays, which is based on highly specific antibody-antigen interaction, including immunoblotting (also known as western blot), immunoprecipitation (IP), enzyme-linked immunosorbent assays (ELISA), etc. Among them, ELISA is one of the most common test formats used for laboratory diagnosis of infections (Lequin 2005). Indirect detection of viral infection is often adopted, for example, through specific detection of antibodies in body fluids that are produced as a part of immune response, with the antigen of causative agent as a specific probe. Comparing with PCR, ELISA is not as sensitive as PCR, but the assay cost is much lower.

While being effective, the above traditional laboratory-based sensing methods for virus detection are limited by sample enrichment and purification required prior to analysis, expense, and time. These methods are not compatible with point-of-care (POC) settings due to their turnaround time, expenses, and labor-intensive sample preparation and handling process. Under the circumstances of serious viral outbreaks, a POC diagnostic device is desired that should be rapid, low-cost, sensitive, specific, and reproducible, with minimum need for sample preparation, compared to established traditional techniques.

Biosensors are considered as promising POC diagnostic devices. Recent advances and innovations in microfluidics, miniaturization, and enhanced signal detection technologies have fueled the development in this area. Among the different types of biosensors, label-free electrical biosensors have received particular attention owing to their properties of being fast, cheap, portable, miniaturized, and allowing direct and real-time monitoring of analytes. This review focuses on capacitive biosensors. Capacitive sensing is performed with AC signals, similar to electrochemical impedance spectroscopy (EIS) analysis, which is a well-established method for characterizing an electrolytic cell. Common to both capacitive sensing and EIS, their data interpretation is based on an equivalent circuit representing the electrode sensor cell. Such an equivalent circuit consists of a network of electrical circuit components, with each component accounting for a part or a process in an electrode/electrolyte cell. A capacitive sensor typically utilizes an array of interdigitated microelectrodes (IDE). Using a capacitive immunosensor as an example, before the tests, bioreceptors specific to the analyte target are immobilized on the surface of IDE. When an antibody-containing solution is loaded, the interaction between antibodies and antigens will cause a change in the impedance of the electrode/electrolyte system, which can be sensed by an external impedance readout system to realize the detection of the antibody–antigen interaction. Usually, in capacitive sensing, the change of electrical double layer capacitance (Cedl) is measured or monitored to indicate the biological reactions occurred on electrodes surface. Wang et al. measure the quantity of ΔCdl,Target, which is the electrical double layer capacitance change on IDE due to target West Nile virus DNA hybridization (Fig. 1a). The capacitive sensor is able to specifically detect as few as 20 complementary DNA target molecules in 30 min with an output of more than 70 nF in capacitance change. The magnitude of sensor output also displays good linearity with target DNA molecules concentration ranging from 20 to 2 million (Wang et al. 2017). Samanman et al. investigate and develop a highly sensitive capacitive biosensor for white spot syndrome virus (WSSV) detection in shrimp pond water. It adopts a flow injection capacitance measurement system, which consists of a measuring flow cell (10 μL) with three electrodes connecting to a potentiostat (Fig. 1b). When target sample (WSSV) is injected after running buffer, target molecules bind to the immobilized probes and lead to a decrease in Ctotal, which is the total capacitance measured at the working electrode/solution interface. Then the capacitance change (ΔC) from baseline to target sample that corresponds to the concentration of target sample is obtained. This sensor has a wide detection range from 1 to 1 × 105 copies/μL. The total assay time is 20–25 min plus another 25 min of regeneration time (Samanman et al. 2011). When detecting virion or nucleic acids extracted from virus, it is most likely that biosensors are challenged with insufficient detection limit, as viruses often present at low concentration and their concentration is further reduced after sample purification. As an alternative, virus detection can be achieved by indirect methods that measure other sensing targets such as virus-specific antibodies produced in blood as a part of immunoresponse. The level of specific serum antibody is usually significantly higher than that of virus antigen; however, the antibody concentration does not necessarily correlate with virion concentration or the progression of viral diseases. For instance, Afsahi et al. built a cost-effective and portable graphene-enabled field effect transistor (FET) by growing single layer graphene films with Ti/Pt leads on silicon wafer. By applying a source-drain voltage across the graphene channel and a gate voltage between the applied liquid and the drain electrode of the graphene, capacitance of the biosensor to the liquid can be measured (Fig. 1c). During the tests to detect Zika virus, binding reactions between Zika virus antibody (IgM) and the immobilized protein, anti-Zika NS1, alter the sensor capacitance, which can be sensed by a data readout system on personal computer (PC). This field effect biosensing (FEB) device shows a good selectivity and reproducibility with a detection limit of 450 pM in buffer solution (Afsahi et al. 2018). Wang et al. present a simple, robust capacitive biosensor using microwires coated with Zika or Chikungunya virus envelop antigen. This method is label-free and can reach rapid detection of ultra-low concentrations of virus antibodies (as low as 10 antibody molecules in a sample volume of 30 μL). This device immobilizes working (Au) and reference electrodes (Ag/AgCl) across the polydimethylsiloxane (PDMS) well, which is used to hold sample solution, on glass substrate. Then by extracting the equivalent circuit of the microwires, total capacitance illustrated in Fig. 1d can be measured and associated with biological reactions occurred on the microwires (Wang et al. 2019). Performances with regard to detection limit, assay time, and sensitivity of each sensor type are summarized in Table 1, including two reports on the detection of viral oligonucleotides and two reports on serum antibodies.
Fig. 1

(a) Equivalent circuit models on interdigitated electrode (IDE) sensors (Wang et al. 2017). (b) Schematic diagram of the flow injection capacitive biosensor system (Samanman et al. 2011). (c) Diagram of the sensor element of the graphene-enabled field effect transistor (FET) biosensor chip. Antibodies are immobilized on pristine graphene using a zero-length linker. Along with the PEG block, these antibodies form the dielectric in a liquid gated transistor with a graphene channel (Afsahi et al. 2018). (d) Schematic of capacitive immunosensor design and working principles. Working electrodes (Au microwire) surface chemistry and functionalized layers with the corresponding equivalent circuit and total capacitance equation (Wang et al. 2019). ((a) is originally published in: Wang et al. (2017); with kind permission of © Elsevier B.V. 2019. All Rights Reserved. (b) is originally published in: Samanman et al. (2011); with kind permission of © Elsevier B.V. 2019. All Rights Reserved. (c) is originally published in: Afsahi et al. (2018); with kind permission of © Elsevier B.V. 2019. All Rights Reserved. (d) is originally published in: Wang et al. (2019); with kind permission of © Elsevier B.V. 2019. All Rights Reserved)

Table 1

Performances of sensors for viral sequences and specific antibodies

Sensor type

Targets

Detection limit

Assay time

Sensitivity

References

IDE

ssDNA oligo

20 molecule/μL

30 min

12.21 nF/dec

Wang et al. (2017)

Flow injection

Virion (WSSV)

1 copy/μL

20–25 min

32 nF∙cm−2/dec

Samanman et al. (2011)

FET

Virus antibodies (ZIKA)

450 pM

5 min

Afsahi et al. (2018)

Microwire

Virus antibodies (ZIKA)

10 antibody molecules/30 μL

Several minutes

2.781 nF/dec

Wang et al. (2019)

While considerable research effort has been devoted to developing such biosensors, there are few successful POC devices being routinely used in real diagnostic applications at the bedside or in the clinic. POC diagnostic systems require the following critical attributes, namely sufficient sensitivity, robustness, simple test procedure, and short sample-to-result time. The obstacle to achieving rapid detection is the long diffusion time for the target bioparticles to reach the sensing site of a sensor. So accelerating the diffusion process has been an essential part of recently microfluidic study. Most of the reported biomolecular sensors work with heavily processed samples, requiring purification, pre-concentration, etc. in addition to sophisticated data processing and expensive equipment. Another challenge is specificity. A number of ultrasensitive affinity sensing methods have been developed, many based on nanotechnology. However, very few of the newer ultrasensitive methods have been evaluated with real patient samples, which is a key to establishing clinical sensitivity and selectivity.

Recently, a label-free capacitive sensing method based on AC electrokinetic (ACEK) effects is developed for virus detection (Cheng et al. 2017a, b, c). This method demonstrates good sensitivity, short response time, a simple operation on detecting virions (antigen) and nucleic acids such as DNA and RNA, making it highly suitable for POC detection and on-site monitoring. As illustrated in Fig. 2, the induced ACEK effects can facilitate the virus-antibody (Fig. 2a) or DNA-DNA probe (Fig. 2b) reactions by accelerating bioparticles movement toward sensing electrodes, leading to significant improvement in sensor responses. Virus RNA detection has similar mechanism to DNA detection with a capture probe targeting a specific region of the pathogen genome RNA. RNA detection is significantly more challenging than DNA detection due to the poor stability of RNA. An RNA assay needs to be completed quickly before RNAs degrade and break down.
Fig. 2

Schematic of ACEK capacitive sensing of (a) virus-antibody pair and (b) DNA-DNA probe pair. On the electrodes, analytes are attracted toward the electrodes surface by ACEK effects. Bindings between targets and receptors (virus-antibody and human herpesvirus-1, or HSV-1, DNA-probe) cause a change at the interface (Cint), which is detected electrically using the same ACEK signal. Other interferences and bioparticles in serum (such as protein, lipid) are nonspecific interferences (Cheng et al. 2017a, b). ((a) is originally published in: Cheng et al. (2017); with kind permission of © Springer Nature 2019. All Rights Reserved. (b) is originally published in: Cheng et al. (2017); with kind permission of © John Wiley & Sons, Inc. 2019. All Rights Reserved)

Capacitive Biosensors

Receptors

Capacitive sensors for direct virus detection can be divided into two main categories: immunosensors that detect the virions (Fig. 2a), and nucleic acid sensors that detect specific nucleic acid sequences (Fig. 2b) extracted from virions. The immunosensors in Table 1 are developed for the detection of specific antibody in serum, while here the immunosensors are for the detection of virions using specific antibody as the probe, as shown in Fig. 2a. Immunosensors for label-free measurements of various analyte have been studied and developed for many reasons. Among various detection schemes, such as optical, mass-sensitive, and electrochemical detection (Wang et al. 2017; Samanman et al. 2011; Cheng et al. 2017a), electrical immunosensors are expected to have better detection limits and less complicated instrumentation. This makes electrical immunosensors a good candidate for POC detection of targets such as virions. The same can be said for electrical nucleic acid sensors.

Nucleic acid (DNA or RNA) sensors utilize oligonucleotide primers such as human herpesvirus-1, or HSV-1, DNA probes shown in Fig. 2b as receptors in biosensing. They are short artificially synthesized nuclei sequences with high specificity. Once they reach the sensor electrodes surface, single strand target DNA (i.e., HSV-1 DNA) will hybridize with the immobilized HSV-1 DNA probes. Such reactions need to be sensed and transduced into singles that can be recorded and further quantitatively analyzed.

Currently, all nucleic acid biosensors suffer from lack of sensitivity to be used directly for pathogen detection. As a result, nucleic acid biosensors are often used after target amplification such as PCR or labeling by nanoparticles, molecular beacons to amplify the signal, and/or by incorporating an enrichment scheme such as electrophoretic preconcentration or magnetic beads for the target to reach detectable level. However, labeling requires multistep process, complicated preparation of functionalized beads, and oftentimes, careful design of receptor probes. Moreover, it should be noted that almost all the reported work was based on detecting ssDNA oligonucleotides or short DNA segments (20–300 bps) as targets. The reasons could be that shorter DNA has a higher diffusivity than longer DNA that improves hybridization rate, and amplicons from PCR process are fragmented DNA and most DNA sensors are developed to detect PCR products. As extra and further steps in sample processing are needed to release and obtain the desired short DNA segments from virus, the use of DNA sensors still faces great challenges in achieving point-of-care detection of clinical samples.

As with all surface-based biosensors, it is very important to design the sensor surface and assay protocols in such a way that it can ensure significantly higher specific binding reactions than nonspecific ones. As a result, for electrical sensors, especially capacitive biosensors, the immobilization of bioreceptors layer becomes very critical. Usually, a blocking reagent, illustrated as gray spheres in Fig. 2, is used to cover the bare part of electrodes surface. If the sensor surface is not completely covered and blocked, the open space can allow any analyte particles to deposit and cause false positive readings. Nevertheless, nonspecific binding is still challenging which makes it difficult to differentiate false positives from true positives when testing complex samples.

Capacitive Sensing Mechanism

Capacitive affinity sensors usually exploit the change in thickness or dielectric properties of the dielectric layer at the electrolyte–dielectric interface, which will lead to a change in the (interfacial) capacitance of the device during antibody–antigen interaction. The interfacial capacitor consists of a series connection of equivalent capacitors caused by electrical double layer (EDL) and macromolecule deposition. When a solid material is immersed into an electrolytic solution, the solid surface will acquire surface charges. To maintain charge neutrality, a thin layer of counter ions is formed at the solid/liquid interface to neutralize the surface charges at the solid surface, which is commonly known as the EDL. Electrically, EDL can be modeled as a capacitor. The layers of counter ions and surface charges are equivalent to the two plates in a capacitor, and the plate separation distance is the EDL thickness. When bioparticles deposited on sensor electrodes surface, the interfacial capacitance Cint illustrated in Fig. 2a will change due to the change in the thickness and surface area of Cint. This can then be utilized to indicate the deposition of bioparticles on sensor electrodes surface as well as to correlate with the bioparticle concentration in the testing sample fluid. The interfacial capacitance is a combination of electric double layer (EDL) and deposited bioparticles capacitance and is expressed in Eq. 8,
$$ {\mathrm{C}}_{\mathrm{int}}={\mathrm{A}}_{\mathrm{int}}/\left(\frac{1}{\upvarepsilon_{\mathrm{p}}}{\mathrm{d}}_{\mathrm{p}}+\frac{1}{\upvarepsilon_{\mathrm{s}}}{\mathrm{d}}_{\mathrm{edl}}\right) $$
(8)
where Aint is the surface area of the interfacial capacitor of the functionalized electrode, dp and dedl are the thickness of bioparticle immobilized and electric double layer formed on the electrodes surface respectively, εp is the bioparticle permittivity, and εs is the solution permittivity. When bioparticles are immobilized on the surface and binding reaction occurs, the interfacial capacitance will change. The change in Cint could be due to either an increase or a decrease, as conceptually shown in Fig. 3a, b. As a result of the binding reaction, targets (such as antibodies) are deposited onto the surface. The thickness of the dielectric layer could increase, which could cause a decrease in the interfacial capacitance. On the other hand, randomly deposited targets could cause an increase in the capacitor’s surface area due to extra topology introduced by the antibody, especially when the probe molecules (such as antigens) are spaced apart, leading to a higher interfacial capacitance, as shown in Fig. 3b. Both changes could occur during the binding. Often, one type of change dominates over the other, and then the detection of binding is possible.
Fig. 3

Two possible topology changes at the solid/fluid interface due to the protein binding reaction. (a) The thickness of the interfacial layer increases while its surface area decreases, Cint reduces as a result; and (b) when the increase in the surface area of Cint dominates over the changes in its thickness, Cint increases (Ab: antibody; Ag: antigen) (Cui et al. 2013b). (Originally published in: Cui et al. (2013); with kind permission of © Royal Society of Chemistry 2019. All Rights Reserved)

Generally, a decrease in Cint due to receptor-target binding (e.g. antigen–antibody binding or DNA-DNA probe hybridization) is commonly observed. In a diluted buffer solution, the EDL is relatively thick. As EDL envelops the antibodies on the electrodes, fine features on the scale of EDL thickness will be lost, and the Cint change will be dominated by an increase in its thickness, that is, Cint reduces. When EDL thickness is comparable to that of antibody topology, a positive change of Cint is possible, especially in a buffer solution of high ionic strength. Either an increase or a decrease in Cint can possibly result from antigen–antibody binding and DNA-DNA probe hybridization. As a matter of fact, an increase in Cint was consistently observed in influenza A virus detection (Cheng et al. 2017a, b) and HSV DNA detection (Cheng et al. 2017b), while a decrease in ZIKA virus RNA detection (Cheng et al. 2017c).

AC Electrokinetic (ACEK) Enrichment

As illustrated in Fig. 4, ACEK effects include dielectrophoresis (DEP), AC electroosmosis (ACEO), and AC electrothermal (ACET) effect (Ramos et al. 1998; Brown et al. 2000; Green et al. 2001). Dielectrophoresis, or DEP, refers to the interaction between a dipole moment on a particle and a nonuniform field (Wu 2008b). This technique has been studied in great details for controlled manipulation of particles, binary separation, and characterization of particles. The DEP velocity of a spherical particle can be described as follows (Castellanos et al. 2003), for spherical and cylindrical particle such as virions,
$$ {u}_{\mathrm{DEP}}=\frac{\varepsilon_m{a}^2}{6\eta}\operatorname{Re}\left[\frac{\varepsilon_p^{\ast }-{\varepsilon}_m^{\ast }}{\varepsilon_p^{\ast }+2{\varepsilon}_m^{\ast }}\right]\nabla {\left|E\right|}^2=\frac{\varepsilon_m{a}^2}{6\eta}\operatorname{Re}\left[{f}_{CM}\right]\nabla {\left|E\right|}^2 $$
(1)
for cylindrical particle such as nucleic acids,
$$ {u}_{\mathrm{DEP}}=\frac{r_{\mathrm{cross}}^2\mathit{\ln}\left(\frac{2l}{r_{\mathrm{cross}}}\right)}{18\eta }{\varepsilon}_m\operatorname{Re}\left[\frac{\varepsilon_p^{\ast }-{\varepsilon}_m^{\ast }}{\varepsilon_m^{\ast }}\right]\nabla {\left|E\right|}^2=\frac{r_{\mathrm{cross}}^2\mathit{\ln}\left(\frac{2l}{r_{\mathrm{cross}}}\right)}{18\eta }{\varepsilon}_m\operatorname{Re}\left[{f}_{CM}\right]\nabla {\left|E\right|}^2 $$
(2)
where εm is the medium permittivity, η is medium viscosity, a is the radius of the virions, rcross is the radius of nucleic acids cross section, l is the length of nucleic acids, and \( {\varepsilon}_p^{\ast } \) and \( {\varepsilon}_m^{\ast } \) are particle and medium complex permittivity, respectively. Complex permittivity is defined as \( {\varepsilon}^{\ast }=\varepsilon -j\frac{\sigma }{\omega \varepsilon} \) (σ: conductivity; ω angular frequency). fCM, a function of ω, is also known as Clausius–Mossotti factor. Therefore, the DEP velocity uDEP is frequency dependent. In the context of electrokinetic manipulation, the real part of the Clausius–Mossotti factor is a determining factor for the dielectrophoretic force on a particle. For Re[fCM] > 0 (or < 0), uDEP > 0 (or < 0) and positive (or negative) DEP will be applied on the particle. Since positive DEP force on a particle traps the particle at the surface of electrodes while negative DEP repels the particle away from the electrodes, in capacitive virus detection, positive DEP, as illustrated in Fig. 4a, is applied to accelerate the binding reaction between probes and target molecules. As is shown in Eq. (1) and (2), uDEP depends on the target particle volume. For nanoscale macromolecules, DEP may not be effective unless the molecules are located within a very short distance to the electrodes (<1 mm).
Fig. 4

ACEK effects consist of DEP, ACEO, and ACET effects. (a) Numerically simulates electric field distribution around a sphere particle in a nonuniform electric field. The particle will be more in the high field region, which is also known as pDEP (Wu 2008a). (b) Illustrates concept of generating ACEO net fluid transport by asymmetric polarization (Wu 2008a). (c) Indicates simulation of pumping and convection on sensor electrodes by ACET using Comsol® Multiphysics (Liu et al. 2011). (Originally published in: Liu et al. (2011); with kind permission of © 2019 Elsevier B.V. All Rights Reserved)

ACEO effect can induce microfluidic vortices above electrodes to transport target molecules to the electrode surface for binding (Green et al. 2000; Wu et al. 2005; Wu and Islam 2007), which improves the detection sensitivity and response time. Under the influence of an inhomogeneous AC electrical field, ACEO flow, shown in Fig. 4b, is caused by the movement of induced free ions in the double layer under the electrical field tangential to the electrode surface. Due to the viscosity between the free ions and fluid, the suspend fluid will be dragged by the motion of free ions to form microflows. ACEO fluid velocity is approximately given as
$$ {u}_{\mathrm{ACEO}}=-\frac{\varepsilon_m}{\eta}\bullet \varDelta \xi \bullet {E}_t $$
(3)
where εm and η are the permittivity and viscosity of the medium, Et is the component of the electric field strength tangential to the electrode surface, and Δξ is the voltage drop over the interfacial layer including the EDL and molecular deposition at the electrode surface (Castellanos et al. 2003). ACEO typically dominates at low ionic strengths, such as target samples diluted in water. But the flow velocity of ACEO has been observed to decrease significantly with increasing conductivity and eventually drop to zero above 0.085 S/m (Ramos et al. 1998). Hence for medical and biological applications that involve the use of solution with high conductivity, the ACEO flow will be negligible.
ACET effect arises from uneven Joule heating due to an electric current flowing through the fluid (Fig. 4c). Once the AC electric field is applied in the bulk solution, polarized particles would be separated and migrate, which generates the ionic current. ACET velocity can be expressed as follows (Green et al. 2001):
$$ {u}_{\mathrm{ACET}}=0.5\operatorname{Re}\left[\frac{\varepsilon_m\left(\alpha -\beta \right)}{\upsigma +\mathrm{j}\omega {\varepsilon}_m}\left(\nabla T\bullet E\right){E}^{\ast }-0.5{\varepsilon}_m\alpha {\left|E\right|}^2\nabla T\right]\bullet {l}^2/\eta $$
(4)
where εm and σ are the permittivity and conductivity of the medium, E is the electric filed strength, ω is the frequency of the applied excitation, η is medium viscosity, l is the characteristic length of the device, typically on the order of electrode spacing, T is the absolute temperature in Kelvin, \( \alpha ={\varepsilon_m}^{-1}\bullet \frac{\partial {\varepsilon}_m}{\partial T}\approx -0.004\ {\mathrm{K}}^{-1}, \) and \( \beta ={\upsigma}^{-1}\bullet \frac{\mathrm{\partial \upsigma }}{\partial T}\approx -0.02\ {\mathrm{K}}^{-1} \) for aqueous media.

Therefore, ACEK effects can occur when an inhomogeneous AC electric field is applied through microelectrodes to sample solution (Wu 2008a). Directed particle movement can be caused by DEP, and particle can also be carried by microflows such as ACEO or ACET flows (Cui et al. 2016) to reach the microelectrodes. For virus detection in biological matrix, aided by ACET, positive DEP (pDEP) dominates and is used for bioparticle enrichment around electrodes to accelerate the biological reactions between probe and target. ACEO is negligible due to the high conductivity of the biological matrix. In addition, as biological reactions between probes and target molecules happen, the interfacial capacitance (Cint) change caused by the binding process is measured by the same applied ACEK signal.

Sensor Designs

Equivalent Circuit Fitting of the Electrode Sensor in Sample Solution

An array of interdigitated electrodes, or IDE, can be fabricated using various materials such as aluminum, gold, copper, and irradiated polyimide. Regardless of its material, an IDE can be evaluated by extracting the equivalent circuit of the sensor cell shown in Fig. 2a. Figure 5a shows the fitting of aluminum IDE deposited on quartz substrate. This sensor cell is modified from AVX Corps’ PARS 433.92 Surface Acoustic Wave (SAW) chip shown in Fig. 5b. The equivalent circuit of sensor electrodes consists of the electrode’s self-resistance (Rwire), interfacial capacitance (Cint), charge transfer resistance (Rct), Warburg coefficient (ZW), fluid bulk resistance (Rs), and dielectric capacitance of the electrode cell (Cs). The Warburg (Zw) is added in series with Rct to represent the diffusion effect in EDL for a better fitting of the circuit at low frequency. By fitting the measured impedance spectrum with the equivalent circuit model shown in Fig. 2a, the values for each circuit element are found as follows: Rwire = 5 Ω, Cint = 40 nF, Rct = 10 Ω, ZW = 270 kΩ, Rs = 1.5 kΩ, and Cs = 3.2 pF.
Fig. 5

(a) Measured and fitted spectrum of extracted equivalent circuit of sensor electrodes cell. (b) Commercially available surface acoustic wave (SAW) electrode chip (Cheng et al. 2017a). (Originally published in: Cheng et al. (2017); with kind permission of © Springer Nature 2019. All Rights Reserved)

At the optimized experimental frequency, which is 100 kHz (Cheng et al. 2016, 2017b), the circuit is expected to be simplified into Cint in series with Rs only. In order to achieve that, XCint << XRct + Zw and XCs >> Rs are expected, where XCint, XCs, and XRct + Zw are the reactance of Cint, Cs, and Rct + ZW, respectively. According to the fitted parameters, at 100 kHz, their reactance is calculated as follows:
$$ {\boldsymbol{X}}_{{\boldsymbol{R}}_{\boldsymbol{ct}}+{\boldsymbol{Z}}_{\boldsymbol{W}}}={\boldsymbol{R}}_{\boldsymbol{ct}}+\sqrt{\mathbf{2}}\frac{{\boldsymbol{Z}}_{\boldsymbol{W}}}{\sqrt{\mathbf{2}\boldsymbol{\pi } \boldsymbol{f}}}=\mathbf{10}\ \boldsymbol{\Omega} +\sqrt{\mathbf{2}}\times \frac{\mathbf{2}\mathbf{70}\ \mathbf{k}\boldsymbol{\Omega }}{\sqrt{\mathbf{2}\boldsymbol{\pi } \times \mathbf{100}\ \mathbf{k}\mathbf{Hz}}}=\mathbf{491.71}\ \boldsymbol{\Omega} $$
(5)
$$ {\boldsymbol{X}}_{{\boldsymbol{C}}_{\mathbf{int}}}=\frac{\mathbf{1}}{\mathbf{2}\boldsymbol{\pi } \boldsymbol{f}{\boldsymbol{C}}_{\mathbf{int}}}=\frac{\mathbf{1}}{\mathbf{2}\boldsymbol{\pi } \times \mathbf{100}\ \mathbf{kHz}\times \mathbf{40}\ \boldsymbol{nF}}=\mathbf{39.79}\ \boldsymbol{\Omega} <<{\boldsymbol{X}}_{{\boldsymbol{R}}_{\boldsymbol{ct}}+{\boldsymbol{Z}}_{\boldsymbol{W}}} $$
(6)
$$ {\boldsymbol{X}}_{{\boldsymbol{C}}_{\boldsymbol{s}}}=\frac{\mathbf{1}}{\mathbf{2}\boldsymbol{\pi } \boldsymbol{f}{\boldsymbol{C}}_{\boldsymbol{s}}}=\frac{\mathbf{1}}{\mathbf{2}\boldsymbol{\pi } \times \mathbf{100}\ \mathbf{k}\mathbf{Hz}\times \mathbf{3.2}\ \boldsymbol{pF}}=\mathbf{497.36}\ \mathbf{k}\boldsymbol{\Omega } >>{\boldsymbol{R}}_{\boldsymbol{s}}\left(\mathbf{1.5}\ \mathbf{k}\boldsymbol{\Omega } \right) $$
(7)

It can be concluded that Cint and Rs dominate the impedance response for the frequency at 100 kHz. Therefore, the extracted equivalent circuit can be simplified to a series connection of Cint and Rs. Consequently, the measured capacitance at 100 kHz can be directly used to indicate the reaction occurred on the electrode surface, which greatly simplifies the process of interpreting experimental data.

Electrodes Surface Treatments

For the metal electrodes such as aluminum, gold, and copper electrodes, prior to incubation with linker and probe molecules, the microelectrode chips are thoroughly cleaned by washing with acetone, isopropyl alcohol, and deionized water, and then treated with ozone or plasma. The surface quality is closely monitored by measuring the Cint. As for the laser printed electrodes, due to the multiscale morphology of the irradiated polyimide surface, wetting properties of interfacial capacitance sensors have become particularly important. The static contact angle of untreated polyimide is found to be 79 ± 1° (Fig. 6a), consistent with previous reports (Least and Willis 2013). After laser ablation, the contact angle is increased to 99 ± 1° (Fig. 6b). Based on the microstructure characterization, this hydrophobicity can be attributed to both the carbonization and porous surface structure.
Fig. 6

A drop of deionized water (3 μL) on (a) pristine and (b) irradiated polyimide surfaces and (c) the irradiated polyimide surface after plasma treatment (Cheng et al. 2016). (Originally published in: Cheng et al. (2017); with kind permission of © American Chemical Society 2019. All Rights Reserved)

However, in most cases this hydrophobic character is not desirable for sensors, especially for detection in an aquatic medium. Plasma treatment is a good method for increasing the surface hydrophilicity by creating OH dangling bonds and enriching O ions on the surface without influencing the electrode microstructure characteristics. After 10 min of vacuum plasma treatment (PLASMA ETCH PE-50), the water drop completely infiltrates into the electrode surface (Fig. 6c), indicating a marked surface transition from hydrophobicity to hydrophilicity. The immobilization efficiency of sensors treated with plasma is improved since more bioreceptors attached to the electrode surface made the surface less likely to become saturated. Besides plasma treatment, ozone treatment can also improve the hydrophilicity of the electrode surface by enriching O ions on the surface.

The electrode’s functionalization after plasma treatment includes receptor incubation and uncovered surface blocking. As the incubation and blocking processes progressed, an increasing number of molecules become attached to the electrode surface. The quality of electrode functionalization is monitored by measuring the Cint (Cui et al. 2013a). During the incubation and blocking process, Cint values have reduced significantly, indicating adequate molecular immobilization on the electrode surface. Surface immobilization reduces the current flowing through the electrodes, effectively increasing the charge transfer resistance, Rct. Therefore, Rct is also measured by carrying out cyclic voltammetry between −0.60 and 0.6 V under the scan rate of 0.05 V/sec during the incubation and blocking. As shown in Fig. 7, Rct (reciprocal of the I-V curve’s slope) increases with time, indicating particles deposition on the sensor electrodes.
Fig. 7

Cyclic voltammetry characteristics of sensor’s (a) incubation and (b) blocking process

Measurements and Data Analysis

In ACEK capacitive sensing, Cint can effectively detect molecular deposition on the electrode surface with high sensitivity and specificity in a much quicker manner. Cint is found by measuring the sensor cell’s impedance at a fixed AC frequency and voltage continuously during the testing. The interfacial capacitance of the electrodes is sampled and recorded periodically by an Agilent 4294A impedance analyzer for 20 s.

Figure 8 shows the capacitance changes of normalized Cint with time for positive, negative, and control tests during influenza A virus detection. Normalized Cint is calculated with respect to the Cint of each sample at initial time point. Hence, problems with baseline drift or the need for a reference sensor can be avoided, greatly simplifying the detection procedure and instrumentation. Furthermore, it also relaxes the requirements on instrument precision and minimizes the effect of difference between sensors. Then the percentage change rate, dC/dt in %/min, of the measured capacitance is adopted as the readout of the sensor, indicating the binding reaction occurring on the electrodes surface. Least square linear fitting algorithm is performed to determine the capacitance change rate.
Fig. 8

Capacitance change of tests with negative (152.5 ng/mL influenza A virus sample on dummy electrodes), control (0.1 × PBS-T (phosphate buffered saline with tween 20), on functionalized electrodes) and positive tests (1.52 ng/mL influenza A virus on functionalized electrodes) (Cheng et al. 2017a). (Originally published in: Cheng et al. (2017); with kind permission of © Springer Nature 2019. All Rights Reserved)

Sensor Performances

Sensor Performance with Analytical Samples

Assays with Various Functionalization and Hybridization Buffers

Consideration for buffers selection includes whether the buffers will be suitable for electrodes surface functionalization, probe-target binding reaction, and the induction of ACEK effect during assay. In general, functionalization buffer plays a vital role in sensor performances such as the reliability and repeatability. This is because functionalization buffer with high ionic strength can screen the electrical charges of nucleic acids. Therefore, during probe immobilization process, using buffer with higher ionic strength can possibly reach higher coverage of probes on the electrode surface. However, nucleic acids will coil up if the functionalizing buffer contains too many ions, losing the ability to bind with other molecules. As for hybridization buffers, a lower ionic strength helps to linearize the target nucleic acids and expose the binding region. Additionally, due to its weaker electrostatic screening, nonspecific binding can also be reduced. Hence, ionic strength of the hybridization buffer and its electrostatic screening would significantly affect sensor’s sensitivity and specificity. These conclusions can be illustrated by the process of buffer selection for HSV-1 DNA detection.

In Fig. 9, both functionalization and hybridization buffers with different ionic strengths are tested in order to optimize the sensor performances. With HSV-1 DNA diluted in 0.5 × SSC (saline sodium citrate), 0.05 × PBS (phosphate buffered saline) as functionalization buffer gives sensor better performances than ultrapure water due to the screening effect brought by the ions in 0.05 × PBS. As for the hybridization buffer, usually higher fluid conductivity will facilitate DNA hybridization, so more DNA will hybrid if the process is reaction limited rather than transport limited. Since Fig. 9 shows that DNA samples diluted in 0.5 × SSC always show higher responses than those in 2 × SSC or 1 × SSC, it can be concluded that DNA hybridization in this assay is transport limited. Also, the travel of target DNA to the electrode slows down with increasing ion concentration from 0.5 × SSC and higher, which indicates that lower fluid conductivity can result in stronger positive DEP effect. In Eqs. (1) and (2), lower fluid conductivity gives greater Clausius–Mossotti factor value. Thus, bioparticles in low fluid conductivity will have high DEP velocity.
Fig. 9

Evaluation of sensor’s performances when probe is prepared in 0.05 × PBS and DNA samples in 0.5 × SSC (black), probe in 0.05 × PBS and DNA samples in 2 × SSC (red), probe in ultrapure water and DNA samples in 0.5 × SSC (blue), and probe in ultrapure water and DNA samples in 1 × SSC (green). Probe prepared in 0.05 × PBS and DNA samples in 0.5 × SSC (black) is the optimized with limit of detection (LOD) of 1 pg/mL (6.47 copies/μL or 10.7 aM) (Cheng et al. 2017b). (Originally published in: Cheng et al. (2017); with kind permission of © John Wiley & Sons, Inc. 2019. All Rights Reserved)

Based on the given results in Fig. 9, HSV-1 probe in 0.05 × PBS and HSV-1 DNA (target), HSV-2 (interference) in 0.5 × SSC are considered to be optimal with responses of −3.90 ± 0.52%/min (9 pg/mL), −6.92 ± 0.94%/min (90 pg/mL), and −9.72 ± 0.63%/min (900 pg/mL). The sensor’s LOD is defined as 3 standard deviations from the response of the background control (−0.22 ± 0.30%/min), so the cut-off d|C|/dt is calculated to be −1.12%/min, which corresponds to an HSV-1 DNA concentration of 0.986 pg/mL (6.38 copies/μL or 0.0106 fM). Tests of interference (HSV-2 DNA) also demonstrate a good specificity with a low response of −0.19 ± 0.60%/min at a concentration (5 ng/mL) 550 times higher than that of HSV-1.

Assays Under Various Applied AC Signal Frequency and Voltage

To elucidate the effects of ACEK mechanisms on detection, the first set of experiments is to find out the effect of AC frequency on the sensor response. Based on Eqs. (1) and (2), DEP effects are frequency-dependent. AC signals of various frequencies at 10 mV are used to measure the capacitance changes from influenza A virus samples at a concentration of 1.525 ng/mL. The measured capacitance change rates are given in Fig. 10. The response shows a bell-shape dependence on AC frequency, with its optimal frequency between 50 and 100 kHz, which indicates that the DEP is the dominant enrichment mechanism. Consequently, AC signal at 100 kHz is considered as the optimized frequency.
Fig. 10

Responses of 0.1 × PBS-T, 1.52 ng/mL influenza A virus on functionalized electrodes, and 152.5 ng/mL influenza A virus sample on dummy electrodes (a) when using 10 mV AC signal with its frequency varied from 20 to 200 kHz (Cheng et al. 2017a). (Originally published in: Cheng et al. (2017); with kind permission of © Springer Nature 2019. All Rights Reserved)

Next, AC voltages varying from 5 mV to 100 V are used to measure 1.52 ng/mL influenza A virus sample on functionalized electrodes. The background blank buffer, which is 0.1 × PBS-T, is also tested on the functionalized electrodes from 5 mV to 100 V as control. Negative control experiments with 152.5 ng/mL influenza A virus sample are measured under the same voltage conditions on dummy electrodes (electrodes without antibody). Experiments with each voltage are repeated three times.

As shown in Fig. 11a, responses of the 0.1 × PBS-T control samples on functionalized electrodes and 152.5 ng/mL influenza A virus sample on dummy electrodes remain quite small through the voltage range of 5–100 mV, with a limited response ranged from −0.14 to −0.24%/min and 1.02 to −0.01%/min. For tests on functionalized electrodes, due to DEP effect, the capacitive response decreases as the voltage increases from 5 to 100 mV, indicating that more binding takes place with higher AC voltage. When the voltage level is above 10 mV, the increase in sensor’s response becomes limited due to saturation of binding sites on the sensor. Therefore, 10 mV is chosen as the measuring voltage. At this voltage, DEP effect will be weak for particles smaller than virus such as protein to cause appreciable capacitance change, which therefore improved the sensor specificity in complex matrix. This can also be justified by the tests of HSV-1 virus DNA conducted under 10 and 25 mV in Fig. 11b. While the sensor yields higher outputs at 25 mV, the sensor also shows non-negligible responses (−1.87 ± 0.43%/min) to 5 ng/mL HSV-2 DNA. In contrast, the response of 5 ng/mL HSV-2 DNA is 1.21 ± 0.31%/min at 10 mV, which is considered to a negative response as it cannot be differentiated from that of the background. A good sensor requires the sensor to have large responses to target molecules with little to none responses non-targets. Test results of HSV-1 and HSV-2 indicate that using AC signal at 10 mV can achieve good specificity with only slight compromise on response (dC/dt values). Therefore, 10 mV is also considered to be superior to 25 mV in detection of HSV-1 virus DNA.
Fig. 11

Responses of 0.1 × PBS-T, 1.52 ng/mL influenza A virus on functionalized electrodes, and 152.5 ng/mL influenza A virus sample on dummy electrodes when using 100 kHz AC signal with its voltage varied from 5 to 100 mV (Cheng et al. 2017a). (b) Responses of 50, 500, and 5000 pg/mL HSV-1 virus DNA and 5000 pg/mL HSV-2 virus DNA (Cheng et al. 2017b). (Originally published in: Cheng et al. (2017); with kind permission of © Springer Nature 2019. All Rights Reserved)

Sensor Performance with Clinical Samples

Dilution Factor Optimization

It is common practice to dilute clinical samples in standard buffer. Due to the complexity of clinical samples, highly diluted samples can reduce nonspecific binding, which improves the selectivity of the sensor. With more dilution of clinical samples, chances of false positive results can be reduced. However, sensor’s sensitivity will also suffer since the concentration of target particles is reduced at the same time. The optimization dilution factor helps to decide which dilution can be used for the bind tests of unknown swab samples in the next step. In addition, clinical samples may contain background matrices that interfere with capacitance measurements.

For instance, in detection of influenza A virus, the clinical swab samples are in M4RT, which is a liquid medium commonly used in the transport of clinical specimens to the laboratory for qualitative microbiological procedures for viral and chlamydial agents. M4RT with no dilution can cause a decrease in capacitance (−0.48 ± 0.035%/min), but for M4RT with 1:1000 dilution or more in 0.1 × PBS-T its effect can be neglected (0.46 ± 0.45%/min at 1:1000 dilution). So, in Fig. 12, dilution factors higher than 1:1000 are studied. To find out the optimal dilution factor to test, two clinical nasal swab samples (one positive and one negative) at various dilution factors from 1:100,000 to 1:1000. Each sample is tested in triplicates, and each chip is tested with three dilutions in the sequence of 1:100,000, 1:10,000, and 1:1000.
Fig. 12

Responses of clinical negative and positive swab samples of dilution factor of 1:100,000, 1:10,000 and 1:1000 (Cheng et al. 2017a). (Originally published in: Cheng et al. (2017); with kind permission of © Springer Nature 2019. All Rights Reserved)

When testing nucleic acids in serum samples, the situation is a little more complicated. Nucleic acid samples need to be heated to 95 °C for linearization while protein sediment in serum has to be avoided. When testing HSV-1 DNA samples, DNA serum spiked samples are diluted into lysing solution first prior to heating to 95 °C for DNA denaturation. Lysing solution can break down peptide bonds, digest proteins in serum, and help reduce sediment that may affect DNA hybridization during tests. In Fig. 13, 900 pg/mL HSV-1 DNA in undiluted serum is 1:2, 1:2.5, 1:3.3, 1:5, 1:10, and 1:20 diluted into lysing solution, respectively, before denaturation (actual concentration in serum is 450, 360, 270, 180, 90 and 45 pg/mL). For dilution factor ranged from 1:2 to 1:20, sensors’ responses are 0.50 ± 0.19%/min, −1.38 ± 0.55%/min, −2.46 ± 0.15%/min, −3.437 ± 0.28%/min, −5.64 ± 0.59%/min, and − 6.34 ± 0.49%/min with a background response of −0.38 ± 0.20%/min. It can be concluded that serum spiked samples with high dilution will yield higher d|C|/dt responses. This is contrary to the intuition that higher target concentration will yield higher sensor response, which we attribute to the effect of complex matrix. Dilution factors 1:20 and 1:10 in Fig. 13 show the highest yet comparable responses. Therefore, dilution factors of 1:20 and 1:10 are used in the subsequent experiments of HSV-1 DNA spiked serum samples.
Fig. 13

Responses of DNA spiked serum diluted in lysing solution using different dilution factors (Cheng et al. 2017b). (Originally published in: Cheng et al. (2017); with kind permission of © John Wiley & Sons, Inc. 2019. All Rights Reserved)

Detection of influenza A Virions in Nasal Swab Samples

Blind tests for a panel of 20 nasal swab samples (10 positive, 10 negative) are conducted to detect influenza A virions. All samples are 1:100,000 diluted with 0.1× PBS-T. The threshold value is set at −0.40%/min, which is also the LOD from previous tests with spiked samples, meaning that samples with a response more negative than −0.40%/min will be considered as positive samples and others negative.

As shown in Fig. 14a, 9 out of 10 positive and 7 out of 10 negative samples are correctly identified by ACEK capacitive sensors. A negative sample with influenza B virus is also correctly identified. All these samples are verified by RT-qPCR, yielding a sensitivity of 90% and specificity of 70% for the panel. Figure 14b shows a positive correlation between the capacitance change rate and PCR cycles number. Weak positive samples are chosen for this set of experiments. Among all the detected positive samples, only the sample with the highest response can be detected by a commercial rapid influenza diagnostic tests (RIDTs), which corresponds to 22 PCR cycles. ACEK capacitive sensor can detect virus level corresponding to 35 PCR cycles. There is a false negative corresponding to 29 PCR cycles. This is possibly due to error during dilution or the binding site on the virus not being exposed.
Fig. 14

Comparison of results from ACEK capacitive sensors and those from commercial tests for a blind panel test of influenza virus A from nasal swabs. (a) Responses of all tested samples differentiated by the −0.40%·min1 cut-off line (blue) and (b) correlation between PCR cycles and responses of samples determined as positive by ACEK capacitive sensor in blind tests. The strongest positive sample is the limit of a commercial rapid influenza test (Cheng et al. 2017a). (Originally published in: Cheng et al. (2017); with kind permission of © Springer Nature 2019. All Rights Reserved)

Detection of HSV-1 DNA in Serum Samples

The logarithmic dependence of HSV-1 DNA concentrations on sensor response (d|C|/dt in %/min) is demonstrated in Fig. 15. The HSV-1 DNA concentrations that are shown are concentrations in neat serum, which are 90 pg/mL, 900 pg/mL, and 9 ng/mL. The two dilution factors of 1:20 and 1:10 obtained from previous optimization (Fig. 13) are adopted and yield similar performances. The LODs are calculated to be 19.46 pg/mL (125.98 copies/μL or 0.21 fM) and 29.73 pg/mL (192.46 copies/μL or 0.32 fM) for 1:20 and 1:10 dilution, respectively. Since samples with 1:20 dilution show slight advantages on sensitivity, LOD, and sensor readouts, it is considered to be the optimal dilution factor for HSV-1 DNA spiked serum sample tests.
Fig. 15

Dose response of HSV-1 DNA serum samples with 1:20 and 1:10 dilution (Cheng et al. 2017b). (Originally published in: Cheng et al. (2017); with kind permission of © John Wiley & Sons, Inc. 2019. All Rights Reserved)

Detection of Zika Virus RNA in Serum/Lysing Samples

Five concentrations of Zika virus RNA are spiked in the 1:1 mixture of 1% serum/lysing solution and are clearly differentiated from each other by testing on the functionalized sensor electrodes. As shown in Fig. 16, sensor responses show a clear logarithmic dependence on Zika virus RNA concentration over at least four orders of magnitude from 1.0 pg/mL to 10 ng/mL. The dependence of d|C|/dt on Zika virus RNA concentration can be approximated as y = 1.99369∙lgx-2.59808, where x is Zika virus RNA concentration in copies/μL and y is capacitance change rate, d|C|/dt, in %/min. The fitted line had a correlation coefficient of 0.97, and is used as the sensor standard curve. The LOD is defined as 3 standard deviations from the response of the background control. Because the background produced a response of −0.39 ± 0.72%/min, the cut-off d|C|/dt is calculated to be 1.786%/min, which corresponded to a Zika virus RNA concentration of 158.1 copies/μL, or 0.846 pg/mL. To exclude possible artifacts as causes of sensor response, control tests are conducted by applying Zika virus RNA samples of the same concentrations on dummy electrodes (without functionalization, and only blocked with 1.0 mM 6-mercaptohexanol in ultrapure water). As shown in Fig. 16, the responses from the dummy electrodes are very close to the responses of fluid background. This indicates that the responses from active sensors are indeed caused by the hybridization between the probe and the target RNA.
Fig. 16

Responses of nonspecific nucleic acid (HSV-1 and dengue) and virus (influenza A), and dose response of Zika virus RNA spiked in serum/lysing solution (Cheng et al. 2017c). (Originally published in: Cheng et al. (2017); with kind permission of © John Wiley & Sons, Inc. 2019. All Rights Reserved)

Conclusions

ACEK-enhanced capacitive biosensors incorporate sample targets enrichment by the ACEK effects with direct measurement of the fluid/electrode interfacial capacitance change in a single-step operation. Therefore, the operation complexity is heavily reduced, as well as the assay time. ACEK-enhanced capacitive sensing measures the interfacial capacitance continuously at an optimized fixed AC frequency and voltage for inducing the ACEK effect. The interfacial capacitance directly reflects the biological reactions on the sensor surface and therefore it allows for direct readout of interfacial capacitance change without complicated data interpretation. Table 2 summarizes the LOD and assay time of virus detections achieved by the presented ACEK-enhanced capacitive biosensors in this chapter, indicating ACEK-enhanced capacitive biosensors are ideal candidates for sensitive and rapid label-free virus detection.
Table 2

Summary of the sensor performances

Target

Matrix

LOD

Assay time

Influenza A virion

0.1 × PBS-T

0.25 pg/mL

30 s

gDNA/human herpesvirus 1

0.5 × SSC

0.986 pg/mL

30 s

gRNA/Zika

0.5 × SSC

1 pg/mL

30 s

References

  1. Afsahi S, Lerner MB, Goldstein JM et al (2018) Novel graphene-based biosensor for early detection of Zika virus infection. Biosens Bioelectron 100:85–88.  https://doi.org/10.1016/j.bios.2017.08.051CrossRefGoogle Scholar
  2. Brown ABD, Smith CG, Rennie AR (2000) Pumping of water with ac electric fields applied to asymmetric pairs of microelectrodes. Phys Rev E 63:016305–016312.  https://doi.org/10.1103/PhysRevE.63.016305CrossRefGoogle Scholar
  3. Castellanos A, Ramos A, González A et al (2003) Electrohydrodynamics and dielectrophoresis in microsystems: scaling laws. J Phys Appl Phys 36:2584–2597CrossRefGoogle Scholar
  4. Cheng C, Wang S, Wu J et al (2016) Bisphenol a sensors on polyimide fabricated by laser direct writing for Onsite River water monitoring at Attomolar concentration. ACS Appl Mater Interfaces 8:17784–17792.  https://doi.org/10.1021/acsami.6b03743CrossRefGoogle Scholar
  5. Cheng C, Cui H, Wu J, Eda S (2017a) A PCR-free point-of-care capacitive immunoassay for influenza a virus. Microchim Acta 184:1649–1657.  https://doi.org/10.1007/s00604-017-2140-4CrossRefGoogle Scholar
  6. Cheng C, Oueslati R, Wu J et al (2017b) Capacitive DNA sensor for rapid and sensitive detection of whole genome human herpesvirus-1 dsDNA in serum: nucleic acids. Electrophoresis 38:1617–1623.  https://doi.org/10.1002/elps.201700043CrossRefGoogle Scholar
  7. Cheng C, Wu J, Fikrig E et al (2017c) Unamplified RNA sensor for on-site screening of Zika virus disease in a limited resource setting. ChemElectroChem 4:485–489.  https://doi.org/10.1002/celc.201600831CrossRefGoogle Scholar
  8. Clark LC, Lyons C (1963) Electrode systems for continuous monitoring in cardiovascular surgery. Ann N Y Acad Sci 102:29–45.  https://doi.org/10.1111/j.1749-6632.1962.tb13623.xCrossRefGoogle Scholar
  9. Cui H, Cheng C, Wu J, Eda S (2013a) Rapid detection of progesterone by commercially available microelectrode chips. IEEE, pp 1–4Google Scholar
  10. Cui H, Li S, Yuan Q et al (2013b) An AC electrokinetic impedance immunosensor for rapid detection of tuberculosis. Analyst 138:7188–7196.  https://doi.org/10.1039/c3an01112gCrossRefGoogle Scholar
  11. Cui H, Cheng C, Lin X et al (2016) Rapid and sensitive detection of small biomolecule by capacitive sensing and low field AC electrothermal effect. Sensors Actuators B Chem 226:245–253.  https://doi.org/10.1016/j.snb.2015.11.129CrossRefGoogle Scholar
  12. Green NG, Ramos A, González A et al (2000) Fluid flow induced by nonuniform ac electric fields in electrolytes on microelectrodes. I. Experimental measurements. Phys Rev E 61:4011–4018.  https://doi.org/10.1103/PhysRevE.61.4011CrossRefGoogle Scholar
  13. Green NG, Ramos A, González A et al (2001) Electrothermally induced fluid flow on microelectrodes. J Electrost 53:71–87.  https://doi.org/10.1016/S0304-3886(01)00132-2CrossRefGoogle Scholar
  14. Least BT, Willis DA (2013) Modification of polyimide wetting properties by laser ablated conical microstructures. Appl Surf Sci 273:1–11.  https://doi.org/10.1016/j.apsusc.2012.12.141CrossRefGoogle Scholar
  15. Lequin RM (2005) Enzyme immunoassay (EIA)/enzyme-linked immunosorbent assay (ELISA). Clin Chem 51:2415–2418.  https://doi.org/10.1373/clinchem.2005.051532CrossRefGoogle Scholar
  16. Liu X, Yang K, Wadhwa A et al (2011) Development of an AC electrokinetics-based immunoassay system for on-site serodiagnosis of infectious diseases. Sensors Actuators Phys 171:406–413.  https://doi.org/10.1016/j.sna.2011.08.007CrossRefGoogle Scholar
  17. Pankaj K (2013) Methods for rapid virus identification and quantification. Mater Methods 3.  https://doi.org/10.13070/mm.en.3.207
  18. Ramos A, Morgan H, Green N, Castellanos A (1998) AC electrokinetics: a review of forces in microelectrode structures. J Phys Appl Phys 31:2338–2353CrossRefGoogle Scholar
  19. Samanman S, Kanatharana P, Chotigeat W et al (2011) Highly sensitive capacitive biosensor for detecting white spot syndrome virus in shrimp pond water. J Virol Methods 173:75–84.  https://doi.org/10.1016/j.jviromet.2011.01.010CrossRefGoogle Scholar
  20. Wang L, Veselinovic M, Yang L et al (2017) A sensitive DNA capacitive biosensor using interdigitated electrodes. Biosens Bioelectron 87:646–653.  https://doi.org/10.1016/j.bios.2016.09.006CrossRefGoogle Scholar
  21. Wang L, Filer JE, Lorenz MM et al (2019) An ultra-sensitive capacitive microwire sensor for pathogen-specific serum antibody responses. Biosens Bioelectron 131:46–52.  https://doi.org/10.1016/j.bios.2019.01.040CrossRefGoogle Scholar
  22. Wu J (2008a) Interactions of electrical fields with fluids: laboratory-on-a-chip applications. IET Nanobiotechnol 2:14–27.  https://doi.org/10.1049/iet-nbt:20070023CrossRefGoogle Scholar
  23. Wu J (2008b) Advances of LOC-based particle manipulation by AC electrical fields. Recent Pat Electr Eng 1:178–187.  https://doi.org/10.2174/1874476110801030178CrossRefGoogle Scholar
  24. Wu J, Islam N (2007) A simple method to integrate in situ nano-particle focusing with cantilever detection. IEEE Sensors J 7:957–958.  https://doi.org/10.1109/JSEN.2007.897361CrossRefGoogle Scholar
  25. Wu J, Ben Y, Battigelli D, Chang H-C (2005) Long-range AC electroosmotic trapping and detection of bioparticles. Ind Eng Chem Res 44:2815–2822.  https://doi.org/10.1021/ie049417uCrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media, LLC, part of Springer Nature 2020

Authors and Affiliations

  1. 1.School of Engineering and Computer ScienceMorehead State UniversityMoreheadUSA
  2. 2.Department of Electrical Engineering and Computer ScienceThe University of TennesseeKnoxvilleUSA

Personalised recommendations