Biosensors and Bioelectronics 220 (2023) 114861

Contents lists available at ScienceDirect

Biosensors and Bioelectronics

journal homepage: www.elsevier.com/locate/bios

A high-throughput fully automatic biosensing platform for efficient

COVID-19 detection
Guoguang Rong a, b, c, Yuqiao Zheng a, b, c, Xiangqing Li b, c, Mengzhun Guo d, e, f, Yi Su a, b, c,
Sumin Bian a, b, c, Bobo Dang d, e, f, Yin Chen g, Yanjun Zhang g, Linhai Shen h, Hui Jin h,
Renhong Yan d, e, Liaoyong Wen b, c, Peixi Zhu i, Mohamad Sawan a, b, c, *
a
CenBRAIN Neurotech, School of Engineering, Westlake University, 600 Dunyu Road, Xihu District, Hangzhou, Zhejiang, 310030, China
b
School of Engineering, Westlake University, 600 Dunyu Road, Xihu District, Hangzhou, Zhejiang, 310030, China
c
Institute of Advanced Study, Westlake Institute for Advanced Study, Hangzhou, Zhejiang, 310024, China
d
Key Laboratory of Structural Biology of Zhejiang Province, School of Life Sciences, Westlake, University, Hangzhou, Zhejiang, 310030, China
e
Center for Infectious Disease Research, Westlake Laboratory of Life Sciences and Biomedicine, Hangzhou, Zhejiang, 310030, China
f
Institute of Biology, Westlake Institute for Advanced Study, Hangzhou, Zhejiang, 310030, China
g
Zhejiang Provincial Center for Disease Control and Prevention, 3399 Binsheng Road, Hangzhou, Zhejiang, 310051, China
h
Hangzhou Center for Disease Control and Prevention, 568 Mingshi Road, Jianggan District, Hangzhou, Zhejiang, 310021, China
i
College of Pharmaceutical Sciences, Zhejiang University of Technology, Hangzhou, Zhejiang, 310014, China

A R T I C L E I N F O A B S T R A C T

Keywords: We propose a label-free biosensor based on a porous silicon resonant microcavity and localized surface plasmon
Biosensor resonance. The biosensor detects SARS-CoV-2 antigen based on engineered trimeric angiotensin converting
Localized surface plasmon resonance (LSPR) enzyme-2 binding protein, which is conserved across different variants. Robotic arms run the detection process
SARS-CoV-2
including sample loading, incubation, sensor surface rinsing, and optical measurements using a portable spec­
COVID-19
Automatic detection
trometer. Both the biosensor and the optical measurement system are readily scalable to accommodate testing a
High throughput wide range of sample numbers. The limit of detection is 100 TCID50/ml. The detection time is 5 min, and the
throughput of one single robotic site is up to 384 specimens in 30 min. The measurement interface requires little
training, has standard operation, and therefore is suitable for widespread use in rapid and onsite COVID-19
screening or surveillance.

1. Introduction
SARS-CoV-2 detection is critical (Larremore et al., 2021), and there
are several techniques available to detect virus in human samples in
analytical labs (Table 1) (Jayamohan et al., 2021). The nucleic acid
amplification test (NAAT) is a mainstream detection technique (Jaya­
mohan et al., 2021) based on the specific amplification of nucleic acid
sequences of target analytes such as the ribonucleic acid (RNA) sequence
of SARS-CoV-2. NAAT includes polymerase chain reaction (PCR) and
isothermal techniques. PCR is currently the most popular technique and
is a gold standard due to its high sensitivity and specificity (Zhu et al.,
2020). PCR requires thermal cycling in three temperature zones of 95◦ C,
55◦ C, and 72◦ C. PCR typically takes at least 4 h to obtain analytical
results due to this thermal cycling process as well as the requirement for
sample preparation such as nucleic acid extraction and purification; this
can be 24–48 h with logistics included. All NAAT techniques require a
clean laboratory environment and trained personnel, thus hindering
application for onsite rapid detection of SARS-CoV-2. Isothermal tech­
niques, such as recombinase polymerase amplification (RPA) (Behr­
mann et al., 2020) and loop-mediated isothermal amplification (LAMP)
(Haq et al., 2021) can amplify specific gene targets without thermal
cycling, and thus are much faster than PCR (ranging from 20 min to 1 h).
Antibody detection is faster, simpler, and identifies human immune
components specific to viral infection (Tantuoyir and Rezaei, 2021)
including in lateral flow assay (LFA) formats (Cavalera et al., 2021).
Unfortunately, antibodies lack the specificity of NAAT techniques. Viral
antigen detection is also possible with LFA (Grant et al., 2020), ELISA
(Kyosei et al., 2020), and chemiluminescence (Liu et al., 2021) tech­
niques, but again suffer sensitivity and detection limit challenges.
Indeed, there is still a lack of scalable, high-throughput, accurate, and

* Corresponding author. CenBRAIN Neurotech, School of Engineering, Westlake University, 600 Dunyu Road, Xihu District, Hangzhou, Zhejiang, 310030, China.
E-mail address: sawan@westlake.edu.cn (M. Sawan).

https://doi.org/10.1016/j.bios.2022.114861
Received 28 December 2021; Received in revised form 19 September 2022; Accepted 24 October 2022
Available online 3 November 2022
0956-5663/© 2022 Elsevier B.V. All rights reserved.
G. Rong et al. Biosensors and Bioelectronics 220 (2023) 114861

fully automatic onsite rapid detection techniques for SARS-CoV-2 (Lar­
remore et al., 2021; Mina and Andersen, 2021).
Biosensors are promising for onsite rapid detection of SARS-CoV-2
due to their high sensitivity, ease-of-operation, fast turnaround, and
integration with high-throughput applications. Examples include elec­
trochemical (Eissa and Zourob, 2021), electrical (Seo et al., 2020), and
optical biosensors (Funari et al., 2020a) for SARS-CoV-2 detection. Of
these, optical biosensors provide high sensitivity. In a recent study,
biolayer interferometry immunosorbent assay measuring optical signal
excited by SARS-CoV-2 antibody-antigen binding reaction has been re­
ported (Dzimianski et al., 2020). This biosensor can be applied to
existing BLI platforms for quantitive detection of SARS-CoV-2 antibody.
As a nanostructured material, porous silicon is desirable for high
performance optical biosensing due to its high surface area, ease of
fabrication and tunable optical properties. It has been used for the
sensitive detection of deoxyribonucleic acid (DNA) (Rong and Weiss,
2009), antibody/antigens (Wu et al., 2012a), and viruses (Rossi et al.,
2007). However, these biosensors typically require analytical targets to
diffuse through the nanopores of porous silicon materials to spatial re­
gion of confined field for high sensitivity detecton, leading to longer
incubation times and careful selection of pore sizes (Rong and Weiss,
2009). These biosensors typically have detection limits in the range of
nM to μM.
Surface Plasmon Polariton (SPP) is a collective oscillation of free
electrons in noble metals. It can be excited by photons in visbile spec­
trum. Many different kinds of plasmonic devices have been designed and
studied in detail (Wang et al., 2021), for various applications including
biosensing. Surface plasmon resonance (SPR) sensors uses electrical
field confined at the interface between metal and dielectric to interact
with biomolecules for biosensing applications. SPR sensors solve many
of these limitations mentioned before. Here, Au, Ag, and Pt can be used
to construct SPP-based biosensors such as SPR (Li et al., 2015) and
localized surface plasmon resonance (LSPR) biosensors (Csáki et al.,
2018). These SPP-based biosensors have high sensitivity towards mo­
lecular binding events near the metal surface because the surface has
strong field confinement resulting in strong field-biomolecule in­
teractions (Qu et al., 2020). In contrast to SPR with low nM detection
limits, LSPR-based biosensors can be as low as the fM range (Kaye et al.,
2017). Furthermore, SPP-based biosensors have biomolecular binding
events on the metal surface, thus eliminating the need for diffusion into
nanopores. Versus SPR biosensors, LSPR biosensors also have a field
confinement much closer to the metal-dielectric interface; thus, they are
more suitable for biomolecular detection (Mataji-Kojouri et al., 2020).
LSPR has been applied to diagnosing COVID-19 by Funari(Funari et al.,

Table 1
Current SARS-CoV-2 detection techniques in practice (FDA approved) and research, their estimated usage percentage, advantages, and disadvantages. Biosensors
include electrochemical (Eissa and Zourob, 2021), electrical (Seo et al., 2020), and optical (Funari et al., 2020a) transduction mechanisms. Examples of “Others” are
coughing sound analysis by AI (Laguarta et al., 2020) and waste water analysis to detect population infection before clinical diagnosis (Evangelista and Bryner, 2020).
Detection techniques Usage Limit of Detection Detection Advantage Disadvantage
percent time

NAAT (Jayamohan PCR (Zhu et al., 2020) 40% N/A N/A Sensitive, specific, Complex procedure, long turn-around
et al., 2021) accurate time, clean environment and training
required
Isothermal (Behrmann 6% 7.74 RNA copies/reaction 15–20 min Fast, simpler than PCR Sensitivity worse than PCR, limited
et al., 2020; Haq et al., 95% detection probability throughput, need for training
2021)
Isothermal (Behrmann N/A N/A
et al., 2020; Haq et al.,
2021)
CRISPR (Nouri et al., 3% N/A N/A Fast, on-site detection Complex sample preparation, limited
2021) is possible target gene regions, multiplexing
challenging
Sequencing (Lu et al., 2020) 5% N/A N/A Sensitive and specific, Complex procedure, high cost,
can detect virus training
mutants
Antigen (Grant et al., 2020; Kyosei et al., 2020; 15% 0.65 ng/ml (95% Confidence N/A Rapid, simple, onsite Limited sensitivity and specificity
Liu et al., 2021) Interval of 0.53–0.77 ng/ml) detection
Antigen (Grant et al., 2020; Kyosei et al., 2020; Approaches PCR assays’ ~10 min
Liu et al., 2021) sensitivity
Antigen (Grant et al., 2020; Kyosei et al., 2020; 360 TCID50/mL Within 16
Liu et al., 2021) min
Antibody (Cavalera et al., 2021; Speletas et al., 15% N/A N/A Rapid, simple, onsite Limited sensitivity and specificity
2020; Tantuoyir and Rezaei, 2021) detection
Antibody (Cavalera et al., 2021; Speletas et al., N/A N/A
2020; Tantuoyir and Rezaei, 2021)
Antibody (Cavalera et al., 2021; Speletas et al., N/A N/A
2020; Tantuoyir and Rezaei, 2021)
CT scan (Lee et al., 2020) 10% N/A N/A Evaluates impact on Complex procedure, high cost, non-
organs specific, low throughput, training
required
Biosensora (Eissa and Zourob 2021; Funari et al., 3% 0.8 pg/ml N/A Rapid, simple Throughput and onsite operation
2020a; Seo et al., 2020) detection needs improvement
Biosensora (Eissa and Zourob 2021; Funari et al., ~0.08 ng/ml ~30min
2020a; Seo et al., 2020)
Biosensora (Eissa and Zourob 2021; Funari et al., Culture medium: 16 pfu/ml; N/A
2020a; Seo et al., 2020) Clinical samples: 242 copies/
ml
Other techniquesa 3% N/A N/A
(Evangelista and Bryner 2020; Laguarta et al., N/A N/A
2020)

FDA: Food and Drug Administration; NAAT: nucleic acid amplification test; PCR: polymerase chain reaction; CRISPR: clustered regularly interspaced short palindromic
repeats; CT: computed tomography; N/A: not available.
a
For research use only.

2
G. Rong et al.
2020b), Behrouzi(Behrouzi and Lin, 2022), and Qiu (Qiu et al., 2020).
These LSPR biosensors are mainly based on reflection or transmission
interferometry of dielectric structures coated with noble metal thin film,
or reflection, transmission and absorption spectra of noble metal
nanoparticles.
Here, we describe a new biosensing platform that exploits the ad­
vantages of SPP and combines a porous silicon resonant microcavity
with the LSPR phenomenon to construct a highly sensitive biosensor for
rapid onsite detection. Porous silicon was used to construct a resonant
microcavity, and gold thin film was deposited on top of the microcavity
to form mode coupling between Tamm Plasmon Polariton (TPP) mode
and microcavity resonant mode. The porous silicon resonant micro­
cavity can be constructed relatively easily without complex thin film
deposition processes, and its resonant mode has strong field confinement
in the central defect layer (Wu et al., 2012a). TPP mode has field
confinement at the interface between metal and top Bragg mirror of the
microcavity (Kaliteevski et al., 2007). By coupling the resonant cavity
mode with TPP mode, the field confined in defect layer of a microcavity
can leak out and reinforce TPP field so that the total field in the vicinity
of the metal thin film can be enhanced. Moreover, this strong field ex­
cites LSPR resonances around the discontinuous gold thin film with
nanostructures, which can further enhance the interaction between
biomolecules and electrical field. Sensitive detection of biomolecules
can be obtained without requiring biomolecules to diffuse through
nanopores into the defect layer of microcavity. We propose a
SARS-CoV-2 antigen detection system consisting of optical biosensors
based on porous silicon, portable fiber spectrometers, and robotic arms
performing automatic operation to achieve all of these aims. Robots or
autonomous systems have been designed to realize automatic
SARS-CoV-2 detection using RT-PCR detection technique. An
open-source robot platform has been developed to detect SARS-CoV-2 of
96 samples in about 4 h (Villanueva-Canas et al., 2021). The system
involves 6 main steps, and 6 different equipment are needed to complete
each step. Another automatic system having better compactness (one
single floor-type equipment) and higher throughput has been proposed
that can carry out SARS-CoV-2 detection of 192 samples under 180 min
at detection limit of 100 copies per reaction (Lu et al., 2022). In this
work, we propose a desktop type automatic detection system that can
detect SARS-CoV-2 of 384 samples in around 30 min with detection limit
of 100 TCID50/ml. The detection system can be extended to include
antibody and nuclei acid detection and can thus be considered an
all-in-one detection tool.
2. Materials and methods
The proposed biosensor is intended for high-precision measurements
by reflection spectroscopy through a portable spectrometer. Using this
technique, viral particles do not need to diffuse into the nanopores as is
the case for most porous silicon-based biosensors. This proposed method
allows one to shorten the detection time. For this new LSPR resonant
microcavity biosensor, a salient resonant feature is available for better
resolution of small shifts. Thus, the new design can substantially
improve the sensitivity and the detection limit versus previously re­
ported LSPR biosensors. For high throughput screening, we fabricated a
sensor array in a standard 96-well format on a porous silicon wafer via a
fast and low-cost approach. The sensor array is readily scalable to satisfy
screening needs (frequency and population size) by simply exploiting
silicon wafers of larger diameters in fabricating the porous silicon. The
porous silicon wafer can also be cut into different sizes of sensor arrays
via a commercially available die saw.
The optical measurement system of the sensor array uses a portable
fiber spectrometer to measure the reflection spectrum. A robotic arm
automatically aligns each sensor spot with a fiber probe of the spec­
trometer to save the optical reflection data automatically. Another ro­
botic arm automatically accomplishes the common tasks encountered in
biosensing: loading of clinical specimens onto the sensor spot and
3
Biosensors and Bioelectronics 220 (2023) 114861
rinsing the sensor surface after incubation. The automatic optical mea­
surement system can achieve high throughput and scalability corre­
sponding to the sensor array design. Multiple spectrometers can be
stacked to achieve simultaneous multiple-spot measurement for a large
well array—this will reduce turnaround time. The fully automatic
design, from specimen loading to results output, can protect health
professionals from biosafety hazards. The biosensor and measurement
system proposed here may be a powerful tool to control the COVID-19
pandemic via efficient population screening (Larremore et al., 2021).
2.1. Biosensor fabrication
Previously, researchers have demonstrated optimization of the
number of periods in the Bragg mirror for excitation of TPP (Kumari
et al., 2018), and for coupling between TPP and resonant mode of a
microcavity(Lu et al., 2019). Basically, the Au thin film has an effect on
the quality factor of the TPP mode. The number of periods in the top
Bragg mirror of the porous silicon microcavity has an effect on quality
factor of the resonant microcavity mode, quality factor of the TPP mode,
and the coupling between TPP mode and resonant microcavity mode.
After considering these factors, we adopted a design of porous silicon
resonant microcavity previously reported (Wu et al., 2012b), and used
Au thin film with thickness of 30 nm.
Porous silicon was fabricated by electrochemical anodization of sil­
icon wafer (6-inch diameter, boron doped P-type, <100> crystal
orientation, 0.01 Ω cm resistivity) in 15% aqueous hydrofluoric (HF)
acid (mixing 75 ml 50% aqueous HF with 175 ml 95% ethanol). The
etching conditions have been reported previously (Wu et al., 2012b).
There are two different porous silicon layers in the resonant microcavity
design. The top Bragg mirror consists of six periods of alternating porous
silicon layers 1 and 2. The bottom Bragg mirror has the same structure as
the top Bragg mirror. A defect layer with the same optical parameter as
the second layer is embedded between the two Bragg mirrors. This
defect layer introduces a resonance peak in the total reflection band of
the Bragg mirror resulting in a resonant microcavity (Wu et al., 2012b).
The refractive indices of each porous silicon layer have been estimated
from its porosity via Bruggeman effective medium theory (Grigoriev
et al., 2020). The thickness of each porous silicon layer was estimated by
multiplying etching time with etching speed as characterized by scan­
ning electron microscopy (Rong et al., 2008a). The resulting porous
silicon has a circular area with a diameter of 11 cm as shown in Fig. 1a
and b. After anodization, the porous silicon wafer is thermally oxidized
in ambient air under 800◦ C for 30 min. This process converts the silicon
hydrogen bond (Si-H) to a silicon oxygen bond (Si-O) and thus stabilizes
the porous silicon in ambient air and makes the porous silicon hydro­
philic (Rong et al., 2008b).
The 5 nm Ti and 30 nm Au thin films were consecutively deposited
(Kim et al., 2007) on the porous silicon wafer via physical vapor depo­
sition (PVD, ZD-400 single chamber high vacuum resistive evaporator;
Shenyang Kecheng Vacuum Tech Co. Ltd). The Au thin film is also
porous because it conformally covers the porous silicon thin film. The Ti
thin film serves as a buffer between the porous silicon and the Au thin
film.
The biosensor must be biofunctionalized to specifically detect SARS-
CoV-2. First, the gold-coated porous silicon wafer was immersed in a
solution of 11-mercaptoundecanoic acid (20 mM) in ethanol (10 ml) and
incubated for 24 h (Ahmad and Moore, 2012; Yeom et al., 2011). After
the reaction, the surface of the chip was washed with ethanol and dried
with air. The functionalized gold-coated porous silicon wafer was then
immersed in 0.4 M 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide
hydrochloride (EDC) and a 0.1 M N-hydroxy succinimide (NHS) mixture
dissolved in 2-(N-morpholino) ethane sulfonic acid (MES) buffer for 15
min to activate the carboxylic groups of the gold-coated porous silicon
wafer. A polydimethylsiloxane (PDMS) well array was fabricated in
advance via a mask to form a sensor array on the porous silicon wafer.
The PDMS well array has a standard 96-well format with a well diameter
G. Rong et al.
Fig. 1. (a) Six-inch-diameter wafers; the fabricated porous silicon active area is
11 cm in diameter. (b) Polydimethylsiloxane (PDMS) well-based array on wafer
in 96-well format from which a PDMS 6 × 7 well array is cut to form a biosensor
chip. The zoom-in view of a single well shows that, within each well, a 2 × 2
spot array is defined through a robotic arm control program with each 0.5-mm
diameter spot measured by reflection spectroscopy. (c) Binding of SARS-CoV-2
by engineered trimeric angiotensin converting enzyme-2 (T-ACE2) immobilized
on biosensor surface. (d) The proposed sensing platform consists of a biosensor
array chip, fiber spectrometer, Y-shaped fiber, 8- or 12-headed pipette, as well
as two robotic arms for alignment/measurement and sample loading/rinsing.
Note that the halogen white light source is not shown in this figure, and that the
thicknesses of the different materials are not to scale.
of 6 mm and a well depth of 4 mm. The PDMS well array was prepared
on the silicon wafer in advance. When ready to use, the sample was cut
and peeled off the silicon wafer and placed on a functionalized porous
silicon wafer. Sealing between PDMS and the functionalized porous
silicon wafer with the activated carboxylic groups is waterproof; thus,
there is no leakage between the wells.
To detect SARS-CoV-2, we used engineered trimeric angiotensin
converting enzyme-2 (T-ACE2) protein as the probe (Fig. 1c). Engi­
neered T-ACE2 (0.2 mg/ml) has a much higher affinity towards the S-
protein of SARS-CoV-2 than monomeric ACE2 (Guo et al., 2021) and was
immobilized on the surface for 30 min by covalent binding to the acti­
vated carboxylic groups (Torrente-Rodríguez et al., 2020; Zang et al.,
2019). This was followed by immersion in blocking buffer (Superblock,
Sigma) containing bovine serum albumin (BSA) for 5 min to deactivate
the unreacted carboxylic groups. ACE2 also has enhanced affinity for
several variants including D614G (Zhou et al., 2021).
2.2. Biosensor measurement setup
The reflection spectrum of the LSPR resonant microcavity biosensor
was measured under wet conditions. Fig. 1d shows that the biosensor
was measured with a portable fiber spectrometer (PG2000-Pro, Idea
Optics, Shanghai, China). A Y-shaped optic fiber was used to guide the
incident white light toward the biosensor and collect the reflected light
from the biosensor. The Y-shaped fiber has one fiber in the center to
collect the reflected light and six fibers around the central fiber to pro­
vide the incident light. All seven fibers in the fiber bundle are 100 μm in
diameter. The distance from the fiber outer surface to the biosensor
surface is 3 mm, and the beam spot on the surface of the biosensor has a
4
Biosensors and Bioelectronics 220 (2023) 114861
diameter of 0.5 mm. The spot size is determined by the diameter of the
fiber providing incident light and the distance between the fiber outer
surface and the biosensor surface. A halogen white light source was used
to provide incident light at a perpendicular angle of incidence. The re­
flected light was collected and guided to a portable spectrometer for real
time spectra data display and saving. The wavelength range of the
spectrometer is 200–1100 nm; only data in the 600–1000 nm range is
used for analysis. The wavelength sampling rate is 0.48 nm.
2.3. Automatic and high throughput detection
A typical biosensing process consists of the following sequential
operation steps:
Step 1, buffer loading; Step 2, biosensor measurement; Step 3,
specimen loading; Step 4, biosensor rinsing; Step 5, buffer loading; and
Step 6, biosensor measurement.
These six steps can be grouped into three main modules, and these
modules can be combined in a mix-and-match way in the graphic pro­
gramming interface of the robot to fit the needs of various applications.
These modules can also be selectively activated or executed ad hoc by
clicking the software interface. Two robotic arms (Z-Arm 1632, HITBOT,
Shenzhen, China) were custom made and programmed to complete the
needed tasks either separately or in synchronization to complete an
integrated application. The following main modules were developed to
smoothly run the various tasks of the platform via the two robotic arms
programmed in the graphical interface. Fig. 1d shows the setup to
implement the following three main modules.
2.3.1. Module 1: specimen/buffer loading
This module uses pipettes on Robotic Arm 1 to automatically load
either phosphate buffer saline (PBS) buffer or clinical liquid specimens
onto the biosensor chip to cover the 6 mm diameter wells. Liquid sample
(4 μl) is needed to completely cover the surface. Accurate volume con­
trol (±10%) is achieved by controlling the pressure applied on the
pipette. After specimen loading, Robotic Arm 2 holds the biosensor chip
for 5 min (which is enough for binding of S-protein with T-ACE2);
optionally, the robotic arm can keep shaking the biosensor chip for a
better binding reaction. Note that the interaction between the virus and
the electric field in the vicinity of the metal thin film is instantaneous,
and there is no time effect on optical response of the biosensor. Buffer
loading does not need incubation or shaking, and its purpose is to make
the biosensor surface wet for optical spectral measurements. Simulta­
neous loading of multiple wells can be achieved with a multi-channel
pipette.
2.3.2. Module 2: biosensor measurements
This module uses Robotic Arm 2 for alignment of each spot within
each well with the fiber probe to take spectral measurements. The
repositioning error of the robotic arm is within 20 μm—this is much
smaller than the spot diameter of 0.5 mm. The current application has a
2 × 2 spot array in each well (Fig. 1b), which means four experiments for
each specimen (the location and number of spots in each well can also be
defined through graphical programming). LabVIEW is used to control
the spectrometer and is synchronized with Robotic Arm 2 so that the
whole process of aligning a spot and taking its measurement is complete
within 1 s. The Robotic Arm 2 moves to the next spot only after the
spectral data from the previous spot have been safely saved. Three
measurements are typically needed, and the average of the three mea­
surements are used to reduce the impact of alignment error. After all
measurements have been completed, the liquid clinical specimen or
buffer on the biosensor surface is removed by pipette on Robotic Arm 1
to prepare for the next operation.
2.3.3. Module 3: biosensor rinsing
This module uses a pipette on Robotic Arm 1 for rinsing the surface of
the biosensor with PBS. This includes applying PBS buffer on the surface
G. Rong et al.
and removing the majority of the buffer from the surface with a pipette.
Typically, 20 μl of PBS buffer is used to rinse the surface. The liquid
residue has no influence on the result since after rinsing, the biosensor
surface will be loaded with PBS buffer before reflection spectral mea­
surement. And shaking the biosensor is optional. Simultaneous rinsing
of multiple wells can be achieved by using a multi-channel pipette.
Repeated rinsing of the wells can help remove non-specific interferents
and can be selected through graphical programming.
2.4. S-ECD viral protein expression and purification
The extracellular domain (1-1208 amino acids) of S protein (S-ECD)
of SARS-CoV-2 (Genebank ID: QHD43416.1) was cloned into the pCAG
vector (Invitrogen) with a “GSAS” mutation at residues 682 to 685, two
proline substitutions at residues 986 and 987, and a C-terminal T4
fibritin trimerization motif followed by one Flag tag and Strep tag. This
construct will hereafter be referred to as S-ECD.
The purification process of S-ECD has been described previously (Chi
et al., 2020). Briefly, the recombinant protein was overexpressed using
HEK 293F mammalian cells (Invitrogen) at 37◦ C under 5% CO2 in a
Multitron-Pro shaker (Infors, 130 rpm). The secreted proteins were
purified with anti-FLAG M2 affinity resin (Sigma Aldrich). After loading
two times, the anti-FLAG M2 resin was washed with the wash buffer
containing 25 mM tris(hydroxymethyl)aminomethane (Tris, pH 8.0)
and 150 mM NaCl. The protein was eluted with the wash buffer plus 0.2
mg/ml flag peptide. The eluent of the peptidase domain (PD) was then
concentrated and subjected to size-exclusion chromatography (Super­
dex 200 Increase 10/300 GL, GE Healthcare) in buffer containing 25 mM
Tris (pH 8.0) and 150 mM NaCl. The peak fractions were collected for
further assays.
2.5. Inactivated SARS-CoV-2 virus solution preparation
Clinical specimens (throat swab) obtained from a SARS-CoV-2
infection case were inoculated into Vero cells at 37◦ C in a 5% CO2
incubator, and the cytopathic effect (CPE) was monitored daily. The
positive isolate was further inoculated into Vero cells to generate viral
stock. The viral stock was titrated and reached a titer of 106.3 TCID50/ml
(tissue culture infectious dose). Beta-propiolactone was added to the
suspension of virus stock at 0.05% and kept overnight at 4◦ C. The
inactivated virus solution was incubated at 37◦ C for 4 h for hydro­
lyzation of beta-propiolactone. The treated virus was subjected to three
consecutive passages on Vero cells to confirm the complete inactivation
of SARS-CoV-2. Inactivated SARS-CoV-2 solution was then serially
diluted to obtain concentrations from 1 to 106 TCID50/ml. Each dilution
was detected by one biosensor well. Each biosensor has four spots of
detection; thus, four experiments were done for each dilution.
2.6. Automatic data analysis
The biosensor reflection was measured before and after virus binding
on the biosensor surface. The reflection spectrum was analyzed by an in-
house developed Matlab program to identify the resonant valley posi­
tion. Virus binding on the biosensor surface can cause a red shift of the
resonant valley because virus binding adds material to the metal surface.
The amount of red shift of the resonant valley is the response of the
biosensor to virus binding and is proportional to the concentration of the
virus in the sample being tested.
Matlab software can automatically identify the resonant valley po­
sition for a given reflection spectrum. These tools first smooth the
original spectrometer data via a Gaussian window with a pre-selected
window size to remove noise and smooth the spectra curve. The Mat­
lab program then finds the peak and valley positions of a given full-
width-half-maximum (FWHM) in a given wavelength range. In addi­
tion to the resonant valley, five peaks and five valleys were identified,
and their respective wavelength positions were recorded. This analysis
5
Biosensors and Bioelectronics 220 (2023) 114861
was done on biosensor spectra before and after virus binding: Either the
red shift of the resonant valley alone or the total red shift of the five
peaks and five valleys can be used to monitor the response of the
biosensor. The latter method can help capture the red shifts due to virus
binding in all available wavelength ranges. While the resonant valley
can resolve small red shifts, longer wavelength features can provide
larger shifts (Wu et al., 2012a).
2.7. Environmental sample treatment
Environmental samples were collected using viral transport medium
(VTM) supplied by YOCON Biology Technology Company (Beijing,
China). For biosensor detection, samples were applied on the biosensor
surface directly without any pre-processing. For RT-PCR detection, viral
RNA was extracted using Qiagen RNeasy Mini Kit following the manu­
facturer’s instruction. Then RT-PCR process was carried out.
3. Results and discussion
This section describes biosensor surface characterization, optical
reflection spectroscopy, and detection of SARS-CoV-2 are presented.
3.1. Biosensor surface characterization
Fig. 2a shows top view scanning electron microscopy (SEM) images
of porous silicon coated with gold. Partial removal of the Au thin film
from the biosensor surface causes the nanoporous structure of both the
porous silicon and gold thin film to be observed. Fig. 2b shows a higher
resolution image of the gold surface using an atomic force microscope
(AFM). The deposited Au thin film on the porous silicon is formed by
clustering gold nanoparticles.
3.2. Optical reflection spectroscopy
The biosensor is measured by reflection spectroscopy via a portable
fiber spectrometer. Fig. 2c shows the reflection spectrum of the porous
silicon resonant microcavity after anodization, thermal oxidation, and
Au coating. The resonant feature remains obvious after each step of the
fabrication process. In addition to the resonant valley, five peaks and
five valleys were ordered from the resonant valley towards the longer
wavelength direction and were included in the calculation of total red
shift as the final response signal of the biosensor. The signal-to-noise
ratio of the spectrum is relative to background noise when the white
light source is shut off and there is only ambient light; this value is over
100 dB.
3.3. Detection of SARS-CoV-2 related targets
After biofunctionalization with T-ACE2, the biosensor is ready for
detection of SARS-CoV-2. We used four kinds of SARS-CoV-2-related
targets for detection: (1) S-ECD viral protein; (2) pseudovirus (Sino­
Biological, Catalog Number PSV001, 1010 copies/ml stock concentra­
tion); and (3) inactivated SARS-CoV-2 virus cultured from a positive
clinical specimen. (4) environmental samples.
To demonstrate the proof-of-principle operation, Fig. 3 shows
representative spectral shifts of the biosensor upon binding reaction
with specific (SARS-CoV-2) and non-specific (other viruses) targets. For
specific targets, the biosensor generates red shifts for all characteristic
peaks and valleys. These shifts are negligible for non-specific targets. In
addition, the extent of red shift correlates with the quantity of specific
targets present in an unknown specimen.
Fig. 4a shows the response of the biosensor (shift of resonant valley
only, as marked in Fig. 3) for detection of S-ECD protein (40 nM). The
biosensor has a specific response for S-ECD viral protein and a negligible
response for PBS buffer only. Fig. 4b shows the response of the biosensor
(shift of resonant valley only as marked in Fig. 3) as a function of
G. Rong et al.
pseudovirus concentration. The biosensor response increases with
increasing pseudovirus concentration, and the limit of detection is more
than 107 copies/ml. Three wells with one spot in each well were
measured for each kind of specimen. For all experiments shown in
Fig. 4a and b; three spectral measurements were taken both before and
after binding reaction with T-ACE2. The error bars in Fig. 4a and b
shows the standard error of responses in the three wells for the same
specimen also demonstrating the well-to-well variation.
To optimize the performances of the biosensor for inactivated SARS-
CoV-2 detection in BSL-2 lab, we adapted data processing of the
biosensor by including five peaks and five valleys as discussed in Section
2.6 and Section 3.2 (Fig. 2c). Fig. 4c shows the total red shift of the
biosensor as a function of inactivated SARS-CoV-2 virus concentration.
Four spots (2 × 2 array) within each well were measured for each
concentration, and three spectral measurements were taken for each
spot both before and after virus binding. The height of the red shifts
increases as virus concentration increases. The response for negative
controls is also shown including the culture medium and five clinical
throat swab specimens of patients infected by viruses other than SARS-
CoV-2 as diagnosed by PCR: human bocavirus, enterovirus, adenovirus,
parainfluenza virus Type 3 (PIV-3), and influenza A. The error bars in
Fig. 4c show the standard error of responses of four spots to detect one
kind of specimen in the same well, thus demonstrating the spot-to-spot
variations. The limit of detection is calculated as three times the stan­
dard deviation of the signal levels of all the negative control experi­
ments. As such, the LOD is close to but lower than the response signal
level of the biosensor at 100 TCID50/ml. Thus we can estimate the LOD
6
Biosensors and Bioelectronics 220 (2023) 114861
Fig. 2. (a) Scanning electron microscopy (SEM)
image of porous silicon coated with gold thin film.
The left half of the SEM shows the porous struc­
ture of porous silicon wherein the silicon skeleton
and air pores form the nanoscale structures; the
right half of the SEM shows the physical vapor
deposition (PVD)-deposited gold thin film with
nanoscale porous features. (b) Atomic force mi­
croscope (AFM) image of the Au-coated porous
silicon biosensor surface. The nanoscale features
of the gold thin film are obvious. The Au thin film
is porous on the nanoscale suggesting that it
supports the localized surface plasmon resonance
(LSPR) mode. (c) Reflection spectra of porous
silicon resonant microcavity in different stages of
the fabrication process: after anodization (black),
after thermal oxidation (red), and after Au coating
(blue). Five peaks (dotted light green circle) and
five valleys (solid dark green circle) of the spec­
trum after Au coating are marked. Either the
resonant valley alone (the leftmost valley, as
marked in the figure) or all five peaks and five
valleys of the porous silicon microcavity can be
used for biosensing response calculation.
of the biosensor to be 100 TCID50/ml, which is better than the state-of-
the-art antigen detection methods with limits of detection of several
hundred to thousands of TCID50/ml (Cerutti et al., 2020; Liu et al., 2021;
Porte et al., 2020; Scohy et al., 2020). The dynamic range lies in clini­
cally relevant SARS-CoV-2 concentrations in throat swab and saliva
specimens (Nouri et al., 2021).
Besides clinical throat swab specimens, we tested the system by
environmental samples. The first to twelfth specimens were sampled at
different locations in the vaccination hospital, where sample 3, 6 and 7
are strong positive, sample 1, 2, 4, 5 and 8 are weak positive, sample 9,
10, 11 and 12 are negative according to PCR results. The thirteenth and
fourteenth specimens are PBS buffer. Similar to the inactivated SARS-
CoV-2 virus experiment described in Fig. 4c, we measure a 2 × 2
array (four spots) in each biosensor chip, and do three rounds of mea­
surements to guarantee the detection accuracy. Fig. 5 shows the total
reflectance spectrum red shift of each specimen. It is obvious that bio­
sensors with strong positive samples loaded have larger amount of
redshift (more than 30 nm), while those loaded with weak negative
samples have relatively lower amount of redshift (more than 10 nm). In
addition, biosensors with negative samples and PBS buffers loaded are of
much lower amount of red shift (less than 5.5 nm) compared to those
loaded with positive samples. This experiment illustrates that our
biosensor has high accuracy in distinguishing environmentally positive
samples from negative samples. Moreover, the signal strength has pos­
itive relationship with the viral load in the sample.
G. Rong et al.
Fig. 3. Representative spectral shifts due to binding of (a) specific targets
(SARS-CoV-2) and (b) non-specific targets (other viruses). The red shifts of the
characteristic spectral peak and valley features are clearly visible for (a) posi­
tive results and negligible for (b) negative results. The circled feature is the
resonant valley.
3.4. Comparison of detection time
We also characterized the time needed to detect different numbers of
specimens with the proposed biosensor and its detection system (time
including sample loading, biosensor surface rinsing, and reflection
spectroscopy measurement) and versus PCR. The shortest turnaround
time with our approach is 5 min, and the time increases with the number
of specimens (Fig. 4d). Twenty min is needed for a single spectrometer
with single well detection and 96 specimens. Less than 1 h is needed for
384 specimens. Roughly 30 min is needed for 384 specimens when using
a double spectrometer for simultaneous two-well measurements. In PCR,
at least 240 min (4 h) is needed regardless of the number of specimens.
The throughput of the biosensor system is much higher than PCR and
can accommodate many specimens more easily than PCR for practical
applications. The detection time for large specimen numbers can be
further reduced, and thus the throughput can be enhanced by inte­
grating more spectrometers in a scalable system for multiple and
simultaneous well detection.
The biosensor array is readily scalable. To use smaller number of
wells per chip, the silicon wafer can be cut to achieve any array size such
as 1 × 6, 4 × 6, or 8 × 6. At the other extreme, scaling can be achieved
using silicon wafers with larger diameters such as 18 inches, which is
7
Biosensors and Bioelectronics 220 (2023) 114861
Fig. 4. (a) Detection of S-ECD viral protein [40 nM in phosphate buffer saline
(PBS) buffer] with control experiments in PBS buffer only. The error bars show
the standard error from three experiments with each experiment carried out in
one well; only one spot in each well is measured. The standard error also
demonstrates the well-to-well variations. (b) Detection of SARS-CoV-2 pseu­
dovirus at different concentrations with control experiment of dilution medium
only (Dulbecco’s modified eagle medium (DMEM), Thermo Fisher p/n
10569044). The error bars represent the standard error from three experiments
with each experiment carried out in one well; only one spot in each well is
measured. That is, the standard error demonstrates the well-to-well variations.
(c) Detection of different concentrations of inactivated SARS-CoV-2 virus. The
positive samples are a serial dilution of inactivated SARS-CoV-2 from 1 to 106
TCID50/ml. N1 is culture media that contains no virus. N2–N6 are five clinical
throat swab specimens from patients infected by human bocavirus, enterovirus,
adenovirus, PIV-3, and influenza A, respectively. The height of the vertical bars
represents the average total red-shifts of five peaks and five valleys from four
experiments with each experiment carried out in one well on four spots. The
error bars show standard errors from those four experiments and demonstrate
the spot-to-spot variation in the same well. The standard error comes from the
non-uniform binding of virus on the biosensor surface reflected by different red
shifts for the four spots in a well. The standard errors can be reduced by
measuring more spots in a well. The limit of detection serves as a cutoff
threshold to differentiate between positive and negative results and is estimated
to be 100 TCID50/ml; this is shown as a dashed line. (d) Comparison of
detection time between proposed biosensor and polymerase chain reaction
(PCR) technique for different numbers of specimens. The time for biosensing
includes sample loading, biosensor surface rinsing, and reflection spectroscopy
measurement. The time for PCR includes nucleic acid extraction, purification,
and amplification. Note that PCR requires at least 240 min regardless of the
number of samples.
G. Rong et al.
Fig. 5. Detection of different environmental samples. The right table shows the number, sampling location and the PCR test result of each sample. The figure shows
the total reflectance spectrum red shift of biosensor.
equivalent to 1000 sensor wells per wafer. In our current design of the
robotic system, up to four wafers can be mounted on a robotic arm. As a
result, up to 4000 clinical specimens can be measured in one batch. The
time to measure the optical signal of such a large well array can also be
reduced by using multiple fiber spectrometers with multiple fiber probes
collecting the optical spectra of multiple wells simultaneously.
A key feature of this work is that both the sensor array and the signal
measurement system are readily scalable to accommodate screening of a
wide range of sample numbers. We believe that this automatic, high
throughput, and scalable detection system could be a useful tool to assist
in curbing the SARS-CoV-2 epidemic. Research shows that detection
methods that are not as sensitive and specific as RT-PCR—when applied
with frequency and good coverage of population—can help prevent
epidemic spread without the need for vaccines (Larremore et al., 2021).
The system proposed here is well suited for such applications.
4. Conclusions
A localized surface plasmon resonance biosensor based on a porous
silicon microcavity has been proposed, fabricated, and demonstrated for
highly sensitivity detection of SARS-CoV-2 virus. The biosensor detects
the coronavirus via the binding protein T-ACE2. Unlike nucleic acid
amplification tests, the biosensor detects virus directly without the need
for complex and time-consuming procedures such as nucleic acid
extraction, purification, and amplification. As a result, the fastest
detection time is 5 min, which is much shorter than PCR tests that can
take 4 h. In addition, the biosensor has a sensitivity and limit of detec­
tion at least four-fold better than the state-of-the-art antigen detection
methods; it is fully automatic and high throughput.
Future work will characterize regeneration and multiplexing:
Detection of antibody, nucleic acid, and other viral protein (such as N
protein) can all be incorporated in the system. These additional detec­
tion targets offer many advantages: Antibody detection can help stratify
the stage of disease (i.e., early asymptomatic/pre-symptomatic stage
(isolation needed) or in the recovery/clearance stage (isolation not
needed)). This can help avoid unnecessary isolation or quarantine
measures. Nucleic acid detection and/or other viral protein detection
can help confirm positive diagnosis.
This work does have some limitations. First, the use of a protein
target rather than a nucleic acid biomarker implies that the sensitivity
and specificity will never be as high as PCR-based methods. Less
8
Biosensors and Bioelectronics 220 (2023) 114861
selective methods still have significant value for point-of-care testing
and surveillance. Second, the silicon wafer surface could be susceptive to
fouling and thus careful longitudinal quality control studies that focus
on the surface chemistry are needed. Third, the liquid handling equip­
ment currently makes this a benchtop-sized design. Future integration of
microfluidic components can lead to point-of-care designs.
Declaration of competing interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
the work reported in this paper.
Data availability
No data was used for the research described in the article.
Acknowledgements
This work was supported by the Leading Innovative and Entrepre­
neur Team Introduction Program of Zhejiang [grant number
2020R01005]; Westlake University [grant number 10318A992001];
Tencent Foundation [grant number XHTX202003001]; Research Pro­
gram on COVID-19 of Zhejiang University of Technology, Zhejiang Key
R&D Program [grant number 2021C03002]; and the Bright Dream Joint
Institute for Intelligent Robotics [grant number 10318H991901].
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.bios.2022.114861.
References
Ahmad, A., Moore, E., 2012. Electrochemical immunosensor modified with self-
assembled monolayer of 11-mercaptoundecanoic acid on gold electrodes for
detection of benzo a pyrene in water. Analyst 137 (24), 5839–5844.
Behrmann, O., Bachmann, I., Spiegel, M., Schramm, M., Abd El Wahed, A., Dobler, G.,
Dame, G., Hufert, F.T., 2020. Rapid detection of SARS-CoV-2 by low volume real-
time single tube reverse transcription recombinase polymerase amplification using
an exo probe with an internally linked quencher (Exo-IQ). Clin. Chem. 66 (8),
1047–1054.
G. Rong et al.
Behrouzi, K., Lin, L.W., 2022. Gold nanoparticle based plasmonic sensing for the
detection of SARS-CoV-2 nucleocapsid proteins. Biosens. Bioelectron. 195, 10.
Cavalera, S., Colitti, B., Rosati, S., Ferrara, G., Bertolotti, L., Nogarol, C., Guiotto, C.,
Cagnazzo, C., Denina, M., Fagioli, F., Di Nardo, F., Chiarello, M., Baggiani, C.,
Anfossi, L., 2021. A multi-target lateral flow immunoassay enabling the specific and
sensitive detection of total antibodies to SARS COV-2. Talanta 223, 121737.
Cerutti, F., Burdino, E., Milia, M.G., Allice, T., Gregori, G., Bruzzone, B., Ghisetti, V.,
2020. Urgent need of rapid tests for SARS CoV-2 antigen detection: evaluation of the
SD-Biosensor antigen test for SARS-CoV-2. J. Clin. Virol. 132, 104654.
Chi, X., Yan, R., Zhang, J., Zhang, G., Zhang, Y., Hao, M., Zhang, Z., Fan, P., Dong, Y.,
Yang, Y., Chen, Z., Guo, Y., Zhang, J., Li, Y., Song, X., Chen, Y., Xia, L., Fu, L.,
Hou, L., Xu, J., Yu, C., Li, J., Zhou, Q., Chen, W., 2020. A neutralizing human
antibody binds to the N-terminal domain of the spike protein of SARS-CoV-2. Science
369 (6504), 650–655.
Csáki, A., Stranik, O., Fritzsche, W., 2018. Localized surface plasmon resonance based
biosensing. Expert Rev. Mol. Diagn. 18 (3), 279–296.
Dzimianski, J.V., Lorig-Roach, N., O’Rourke, S.M., Alexander, D.L., Kimmey, J.M.,
DuBois, R.M., 2020. Rapid and sensitive detection of SARS-CoV-2 antibodies by
biolayer interferometry. Sci. Rep. 10 (1), 12.
Eissa, S., Zourob, M., 2021. Development of a low-cost cotton-tipped electrochemical
immunosensor for the detection of SARS-CoV-2. Anal. Chem. 93 (3), 1826–1833.
Evangelista, S., Bryner, A., 2020. COVID-19: Using Wastewater to Track the Pandemic.
Funari, R., Chu, K.-Y., Shen, A.Q., 2020a. Detection of antibodies against SARS-CoV-2
spike protein by gold nanospikes in an opto-microfluidic chip. Biosens. Bioelectron.
169, 8.
Funari, R., Chu, K.Y., Shen, A.Q., 2020b. Detection of antibodies against SARS-CoV-2
spike protein by gold nanospikes in an opto-microfluidic chip. Biosens. Bioelectron.
169, 8.
Grant, B.D., Anderson, C.E., Williford, J.R., Alonzo, L.F., Glukhova, V.A., Boyle, D.S.,
Weigl, B.H., Nichols, K.P., 2020. SARS-CoV-2 coronavirus nucleocapsid antigen-
detecting half-strip lateral flow assay toward the development of point of care tests
using commercially available reagents. Anal. Chem. 92 (16), 11305–11309.
Grigoriev, F.V., Sulimov, V.B., Tikhonravov, A.V., 2020. Combined modeling of the
optical anisotropy of porous thin films. Coatings 10 (6), 517.
Guo, L., Bi, W.W., Wang, X.L., Xu, W., Yan, R.H., Zhang, Y.Y., Zhao, K., Li, Y.N.,
Zhang, M.F., Cai, X., Jiang, S.B., Xie, Y.H., Zhou, Q., Lu, L., Dang, B.B., 2021.
Engineered trimeric ACE2 binds viral spike protein and locks it in "Three-up"
conformation to potently inhibit SARS-CoV-2 infection. Cell Res. 31 (1), 98–100.
Haq, F., Sharif, S., Khurshid, A., Ikram, A., Shabbir, I., Salman, M., Ahad, A., Rana, M.S.,
Raja, A., Badar, N., Tashkandi, H., Al Amri, T., Azhar, E.I., Almuhayawi, M.S.,
Harakeh, S., Malik, M.F.A., 2021. Reverse transcriptase loop-mediated isothermal
amplification (RT-LAMP)-based diagnosis: a potential alternative to quantitative
real-time PCR based detection of the novel SARS-COV-2 virus. Saudi J. Biol. Sci. 28
(1), 942–947.
Jayamohan, H., Lambert, C.J., Sant, H.J., Jafek, A., Patel, D., Feng, H.D., Beeman, M.,
Mahmood, T., Nze, U., Gale, B.K., 2021. SARS-CoV-2 pandemic: a review of
molecular diagnostic tools including sample collection and commercial response
with associated advantages and limitations. Anal. Bioanal. Chem. 413 (1), 49–71.
Kaliteevski, M., Iorsh, I., Brand, S., Abram, R.A., Chamberlain, J.M., Kavokin, A.V.,
Shelykh, I.A., 2007. Tamm plasmon-polaritons: possible electromagnetic states at
the interface of a metal and a dielectric Bragg mirror. Phys. Rev. B 76 (16).
Kaye, S., Zeng, Z., Sanders, M., Chittur, K., Koelle, P.M., Lindquist, R., Manne, U., Lin, Y.
B., Wei, J.J., 2017. Label-free detection of DNA hybridization with a compact LSPR-
based fiber-optic sensor. Analyst 142 (11), 1974–1981.
Kim, D.-K., Kerman, K., Saito, M., Sathuluri, R.R., Endo, T., Yamamura, S., Kwon, Y.S.,
Tamiya, E., 2007. Label-free DNA biosensor based on localized surface plasmon
resonance coupled with interferometry. Anal. Chem. 79 (5), 1855–1864.
Kumari, A., Kumar, S., Shukla, M.K., Kumar, G., Maji, P.S., Vijaya, R., Das, R., 2018.
Coupling to Tamm-plasmon-polaritons: dependence on structural parameters.
J. Phys. D Appl. Phys. 51 (25).
Kyosei, Y., Namba, M., Yamura, S., Takeuchi, R., Aoki, N., Nakaishi, K., Watabe, S.,
Ito, E., 2020. Proposal of de novo antigen test for COVID-19: ultrasensitive detection
of spike proteins of SARS-CoV-2. Diagnostics 10 (8), 594.
Laguarta, J., Hueto, F., Subirana, B., 2020. COVID-19 artificial intelligence diagnosis
using only cough recordings. IEEE Open J. Eng. Med. Biol. 1 (10), 275–281.
Larremore, D.B., Wilder, B., Lester, E., Shehata, S., Burke, J.M., Hay, J.A., Milind, T.,
Mina, M.J., Parker, R., 2021. Test sensitivity is secondary to frequency and
turnaround time for COVID-19 screening. Sci. Adv. 7 (1), eabd5393.
Lee, E.Y.P., Ng, M.Y., Khong, P.L., 2020. COVID-19 pneumonia: what has CT taught us?
Lancet Infect. Dis. 20 (4), 384–385.
Li, M., Cushing, S.K., Wu, N., 2015. Plasmon-enhanced optical sensors: a review. Analyst
140 (2), 386–406.
Liu, D., Ju, C.H., Han, C., Shi, R., Chen, X.H., Duan, D.M., Yan, J.H., Yan, X.Y., 2021.
Nanozyme chemiluminescence paper test for rapid and sensitive detection of SARS-
CoV-2 antigen. Biosens. Bioelectron. 173, 112817.
Lu, H., Li, Y.W., Jia, H., Li, Z.W., Ma, D., Zhao, J.L., 2019. Induced reflection in Tamm
plasmon systems. Opt Express 27 (4), 5383–5392.
Lu, J., Fan, W.H., Huang, Z.H., Fan, K., Dong, J.H., Qin, J.S., Luo, J.Z., Zhang, Z.Z.,
Sun, G.D., Duan, C.H., Pan, K.Y., Gu, W.S., Zhang, X., 2022. Automatic system for
Biosensors and Bioelectronics 220 (2023) 114861
high-throughput and high-sensitivity diagnosis of SARS-CoV-2. Bioproc. Biosyst.
Eng. 45 (3), 503–514.
Lu, R., Zhao, X., Li, J., Niu, P., Yang, B., Wu, H., Wang, W., Song, H., Huang, B., Zhu, N.,
Bi, Y., Ma, X., Zhan, F., Wang, L., Hu, T., Zhou, H., Hu, Z., Zhou, W., Zhao, L.,
Chen, J., Meng, Y., Wang, J., Lin, Y., Yuan, J., Xie, Z., Ma, J., Liu, W.J., Wang, D.,
Xu, W., Holmes, E.C., Gao, G.F., Wu, G., Chen, W., Shi, W., Tan, W., 2020. Genomic
characterisation and epidemiology of 2019 novel coronavirus: implications for virus
origins and receptor binding. Lancet 395 (10224), 565–574.
Mataji-Kojouri, A., Ozen, M.O., Shahabadi, M., Inci, F., Demirci, U., 2020. Entangled
nanoplasmonic cavities for estimating thickness of surface-adsorbed layers. ACS
Nano 14 (7), 8518–8527.
Mina, M.J., Andersen, K.G., 2021. COVID-19 testing: one size does not fit all. Science 371
(6525), 126–127.
Nouri, R., Tang, Z.F., Dong, M., Liu, T.Y., Kshirsagar, A., Guan, W.H., 2021. CRISPR-
based detection of SARS-CoV-2: a review from sample to result. Biosens. Bioelectron.
178, 113012.
Porte, L., Legarraga, P., Vollrath, V., Aguilera, X., Munita, J.M., Araos, R., Pizarro, G.,
Vial, P., Iruretagoyena, M., Dittrich, S., Weitzel, T., 2020. Evaluation of a novel
antigen-based rapid detection test for the diagnosis of SARS-CoV-2 in respiratory
samples. Int. J. Infect. Dis. 99, 328–333.
Qiu, G.G., Gai, Z.B., Tao, Y.L., Schmitt, J., Kullak-Ublick, G.A., Wang, J., 2020. Dual-
functional plasmonic photothermal biosensors for highly accurate severe acute
respiratory syndrome Coronavirus 2 detection. ACS Nano 14 (5), 5268–5277.
Qu, J.-H., Dillen, A., Saeys, W., Lammertyn, J., Spasic, D., 2020. Advancements in SPR
biosensing technology: an overview of recent trends in smart layers design,
multiplexing concepts, continuous monitoring and in vivo sensing. Anal. Chim. Acta
1104, 10–27.
Rong, G., Najmaie, A., Sipe, J.E., Weiss, S.M., 2008a. Nanoscale porous silicon
waveguide for label-free DNA sensing. Biosens. Bioelectron. 23 (10), 1572–1576.
Rong, G., Weiss, S.M., 2009. Biomolecule size-dependent sensitivity of porous silicon
sensors. Phys. Status Solidi A Appl. Mat. Sci. 206 (6), 1365–1368.
Rong, G.G., Ryckman, J.D., Mernaugh, R.L., Weiss, S.M., 2008b. Label-free porous silicon
membrane waveguide for DNA sensing. Appl. Phys. Lett. 93 (16), 161109.
Rossi, A.M., Wang, L., Reipa, V., Murphy, T.E., 2007. Porous silicon biosensor for
detection of viruses. Biosens. Bioelectron. 23 (5), 741–745.
Scohy, A., Anantharajah, A., Bodeus, M., Kabamba-Mukadi, B., Verroken, A., Rodriguez-
Villalobos, H., 2020. Low performance of rapid antigen detection test as frontline
testing for COVID-19 diagnosis. J. Clin. Virol. 129, 104455.
Seo, G., Lee, G., Kim, M.J., Baek, S.H., Choi, M., Ku, K.B., Lee, C.S., Jun, S., Park, D.,
Kim, H.G., Kim, S.J., Lee, J.O., Kim, B.T., Park, E.C., Kim, S.I., 2020. Correction to
rapid detection of COVID-19 causative virus (SARS-CoV-2) in human
nasopharyngeal swab specimens using field-effect transistor-based biosensor. ACS
Nano 14 (9), 12257–12258.
Speletas, M., Kyritsi, M.A., Vontas, A., Theodoridou, A., Chrysanthidis, T.,
Hatzianastasiou, S., Petinaki, E., Hadjichristodoulou, C., Grp, C.S., 2020. Evaluation
of two chemiluminescent and three ELISA immunoassays for the detection of SARS-
CoV-2 IgG antibodies: implications for disease diagnosis and patients’ management.
Front. Immunol. 11, 609242.
Tantuoyir, M.M., Rezaei, N., 2021. Serological tests for COVID-19: potential
opportunities. Cell Biol. Int. 45 (4), 740–748.
Torrente-Rodríguez, R.M., Lukas, H., Tu, J.B., Min, J.H., Yang, Y.R., Xu, C.H., Rossiter, H.
B., Gao, W., 2020. SARS-CoV-2 RapidPlex: a graphene-based multiplexed
telemedicine platform for rapid and low-cost COVID-19 diagnosis and monitoring.
Matter 3 (6), 1981–1998.
Villanueva-Canas, J.L., Gonzalez-Roca, E., Unanue, A.G., Titos, E., Yoldi, M.J.M.,
Gomez, A.V., Puig-Butille, J.A., 2021. Implementation of an open-source robotic
platform for SARS-CoV-2 testing by real-time RT-PCR. PLoS One 16 (7).
Wang, B.Q., Yu, P., Wang, W.H., Zhang, X.T., Kuo, H.C., Xu, H.X., Wang, Z.M.M., 2021.
High-Q plasmonic resonances: fundamentals and applications. Adv. Opt. Mater. 9
(7).
Wu, B., Rong, G.G., Zhao, J.W., Zhang, S.L., Zhu, Y.X., He, B.Y., 2012a. A nanoscale
porous silicon microcavity biosensor for novel label-free tuberculosis antigen-
antibody detection. Nano 7 (6), 1250049.
Wu, C., Rong, G., Xu, J., Pan, S., Zhu, Y., 2012b. Physical analysis of the response
properties of porous silicon microcavity biosensor. Phys. E (Amsterdam, Neth.) 44
(7–8), 1787–1791.
Yeom, S.-H., Kim, O.G., Kang, B.H., Kim, K.J., Yuan, H., Kwon, D.H., Kim, H.R., Kang, S.
W., 2011. Highly sensitive nano-porous lattice biosensor based on localized surface
plasmon resonance and interference. Opt Express 19 (23), 22882–22891.
Zang, F., Su, Z., Zhou, L., Konduru, K., Kaplan, G., Chou, S.Y., 2019. Ultrasensitive Ebola
virus antigen sensing via 3D nanoantenna arrays. Adv. Mater. 31 (30), e1902331.
Zhou, B., Thao, T.T.N., Hoffmann, D., Taddeo, A., Ebert, N., Labroussaa, F.,
Pohlmann, A., King, J., Steiner, S., Kelly, J.N., Portmann, J., Halwe, N.J., Ulrich, L.,
Trüeb, B.S., Fan, X., Hoffmann, B., Wang, L., Thomann, L., Lin, X., Stalder, H.,
Pozzi, B., de Brot, S., Jiang, N., Cui, D., Hossain, J., Wilson, M.M., Keller, M.W.,
Stark, T.J., Barnes, J.R., Dijkman, R., Jores, J., Benarafa, C., Wentworth, D.E.,
Thiel, V., Beer, M., 2021. SARS-CoV-2 spike D614G change enhances replication and
transmission. Nature 592 (7852), 122–127.
Zhu, H.L., Zhang, H.Q., Xu, Y., Lassakova, S., Korabecna, M., Neuzil, P., 2020. PCR past,
present and future. Biotechniques 69 (4), 317–325.