Achieving broad availability of SARS-CoV-2 detections via smartphone-based analysis

Abstract

With the development of COVID-19, widely available tests are in great demand. Naked-eye SARS-CoV-2 test kits have recently been developed as home tests, but their sensitivity and accuracy are sometimes limited. Smartphones can convert various signals into digital information, potentially improving the sensitivity and accuracy of these home tests. Herein, we summarize smartphone-based detections for SARS-CoV-2. Optical detections of non-nucleic acids using various sensors and portable imaging systems, as well as nucleic acid analyses based on LAMP, CRISP, CATCH, and biosensors are discussed. Furthermore, different electrochemical detections were compared. We show results obtained using relatively complex equipment, complicated programming procedures, or custom smartphone apps, and describe methods for obtaining information with only simple setups and free software on smartphones. Then, the combined costs of typical smartphone-based detections are evaluated. Finally, the prospect of improving smartphone-based strategies to achieve broad availability of SARS-CoV-2 detection is proposed.

Abbreviations

SARS-CoV-2

Severe acute respiratory syndrome coronavirus 2

PCR

polymerase chain reaction

RT-PCR

real-time quantitative polymerase chain reaction

LFA

lateral flow assay

ELISA

enzyme-linked immunosorbent assay

RNA

ribonucleic acid

HRP

horseradish peroxidase

NPs

nanoparticles

TMB

3,3′,5,5′- tetramethylbenzidine

NLICS

nanozyme-linked immunochromatographic sensor

ICA

immunochromatography assay

PPA

positive percent agreement

T-line

Test line

SP

spike protein

CATCH

Catalytic amplification by a transition-state molecular switch

Au NPs

gold NPs

HFIS

high-throughput fiber integrated immunosensing system

PTEM

Polycarbonate track-etched membrane

SPR

surface plasmon resonance

SERS

surface-enhanced Raman scattering

SEIRA

surface-enhanced infrared absorption spectroscopy

SEF

surface-enhanced fluorescence

Sens

sensitivity

Spec

specificity

PPV

positive predictive value

NPV

negative predictive value

PS

positive

NG

negative (NG)

THz

terahertz

vp

virus particles

SARS

Severe acute respiratory syndrome

MERS

Middle East respiratory syndrome

VSV

Vesicular stomatitis viruses

MATLAB

MATrix LABoratory

LAMP

loop-mediated isothermal amplification

RT-LAMP

Reverse-transcription LAMP

RT-eLAMP

RT-LAMP onto a PoC platform

CMOS

complementary metal-oxide-semiconductor

ISFET

ion-sensitive field-effect transistors

RT-qPCR

RT-PCR using a real-time benchtop platform

RT-qLAMP

RT-LAMP assay used a real-time benchtop instrument

RT-eLAMP

The lab-on-chip (LoC) platform used a smartphone with a customized App to process the sensing data of RT-LAMP

AWS

Cloud Server

Cap-iLAMP

capture and improved LAMP

VTM

viral transport medium

IoT

internet of things

smaRT-LAMP

smartphone-based RT-LAMP

CRISPR

clustered regularly interspaced short palindromic repeat

crRNA

CRISPR RNA

RNP

nuclease-inactive ribonucleoprotein complex

HEPN

higher eukaryotic and prokaryotic nucleotide-binding domain

PLA

Black poly(lactic acid)

RT-RPA

reverse transcription recombinase polymerase amplification

DM

droplet magnetofluidics

TOPSE

True Outcome Predicted via Strip Evaluation

FnCas9

Cas9 ortholog from Francisella novicida

iSCAN

RT-LAMP-coupled CRISPR-Cas12 module for rapid sensitive detection of SARS-CoV-2

LCs

Liquid Crystals

μPADs

microfluidic paper-based analytical devices

CV

cyclic voltammetry

EIS

eLectrochemical impedance spectroscopy

R_CT

charge-transfer resistance

ePAD

electrochemical paper-based analytical device

OECT

Organic electrochemical transistors

MIP

Molecularly imprinted polymers

LSG

laser-scribed graphene

TB

toluidine blue

EAB

electrochemical aptamer

CNF

carbon nanofiber

RCA

rolling circle amplification

RBD

receptor-binding domain

ssDNA

single-strand DNA

LSG

laser-scribed graphene

ACE2

Angiotensin-Converting Enzyme 2

SCX8

p-sulfocalix [8]arene

SCX8-RGO

SCX8 functionalized graphene

TAMRA-FAM

TAMRA dye works as an internal standard, and FAM dye serves as a sensitive sensing agent. The TAMRA and FAM are orange-red-emitting and green fluorescent dyes used to label peptides

MECS

self-actuated molecular-electrochemical system

E-INAATs

electrochemical isothermal nucleic acid amplification tests. Nab, neutralizing antibody

PtNP

platinum nanoparticle

MSAA

Microbubbling SARS-CoV-2 Antigen Assay

ML

machine learning

QD-LFIA

quantum dot lateral flow immunoassay strip

IPCF

an optical sensor based on an imprinted photonic crystal film

Go Spectro

a device that turns a smartphone into an ultracompact and powerful light hand spectrometer

F-IPCF

ANTIBIDOY functionalized-IPCF

OPTIMA-dx

a mobile phone application developed by the related authors

DAMPR

DNAzyme reaction triggered by LAMP with clustered regularly interspaced short palindromic repeats (CRISPR)

ABTS

2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid)

1. Introduction

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is a biological hazard responsible for COVID-19. With the continued transmission of this virus around the world, there has been increasing demand for more convenient, sensitive, accurate, and inexpensive tests [1,2]. Currently, the SARS-CoV-2 analysis mainly depends on recognizing the following targets: nucleic acids [[3], [4], [5]], non-nucleic acids such as antigens [[6], [7], [8], [9]], antibodies [[10], [11], [12]], associated biomarkers [13,14], symptom-related parameters [15], or multiplex factors [16,17]. Among these strategies, the real-time quantitative polymerase chain reaction (PCR) (RT-PCR) relying on nucleic acid amplification and detection is the most widely used and accurate [18]. However, the PCR-based methods last for several hours and are dependent on professional equipment and operators [[19], [20], [21]]. Once large-scale tests are required, this method may be labor-intensive and require high costs [22,23]. In addition, many people may gather together to take these PCR tests for sample collections, which increases the risk of uninfected cases in susceptible environments [[24], [25], [26]].

Recently, convenient home tests mainly based on the analysis of the protein of the virus were investigated. The sandwich lateral flow assay (LFA) and immunoassay including gold-labeled strips [[27], [28], [29], [30]], chemiluminescence assay [31], fluorescence-based LFA [[32], [33], [34]], and enzyme-linked immunosorbent assay (ELISA) [35], etc., have been used. Normally, the antibodies or other biorecetors are modified on test devices that can recognize the targets and then show signal changes. These home tests can quickly display results that are visible to the naked eye within 15–20 min. Both the professional staff and normal people can operate [36]. However, according to some references, these commercial home tests have an accuracy of about 70%–90% [37,38], and may also have more “false positive/negative” probability compared to PCR tests [39,40]. Especially, with relatively few mounts of the virus load (e.g. the sample from a just infected person), these home tests may be even less sensitive and accurate.

Recently, some techniques were summarized for the diagnosis of COVID-19. For instance, Bukkitgar et al. presented electrochemical methods for the analysis of various viruses including SARS-CoV-2 with a focus on nanotechnology-based detections [41]. Suleman et al. reviewed the field-effect transistor, optical biosensors, RT-LAMP-based detections, and miscellaneous biosensors for the diagnosis of COVID-19 [42] with a focus on nucleic acid detections. Shetti et al. summarized early detection methods for SARS-CoV-2 analysis based on the sensing of targets including ribonucleic acid (RNA), spike protein (SP), and antibodies such as IgG and IgM, and primarily mention nanosensors [43]. In these earlier reviews, different nanomaterials and sensing strategies for the construction of electrochemical or optical biosensors were shown, such as magnetic nanoparticle carbon nanomaterial, gold silica, novel metallic nanomaterial, and quantum dot functionalized biosensors; The label-free, aptamer modification methods, and immune sensing systems were discussed. Because the research at the early stages was limited, the reported reviews mainly focused on the primary detection strategies or propose possible family sensing methods, but a further discussion of widely available tests was still lacking. For instance, smartphone-based SARS-CoV-2 detection strategies were mainly proposed as prospects.

Smartphone-based tests can convert signals into precise digital data that may show more sensitive and accurate readings than visual observations; smartphones can also control the progress of detections through a well-designed program; at the same time, due to smartphones are widely used, so it doesn't take much of a burden [[44], [45], [46], [47], [48], [49]]. Some smartphone-based tests have already shown great promise for the accurate and sensitive detections of various analytes [[50], [51], [52], [53]]. Until now, a few smartphone-based detections have been investigated for nucleic acid analysis. The non-nucleic acid detections depending on the recognization of antigens, antibodies, or virus particles using smartphones also access the accuracy of PCR tests [[54], [55], [56], [57]]. These methods have broad application potential and deserve in-depth discussion. Among the smartphone-based detection techniques, it is worth comparing which strategy will be more widely adopted and reproduced. Therefore, we comprehensively compare these reported smartphone-based detection tests and discuss their cost, sensitivity, accuracy, and possible availability (Fig. 1 ). Typical examples are employed to represent the mechanisms of different tests. The required accessories, sample collections, and the way smartphones acquire signals, etc., are exhibited to evaluate the feasibility of the methods. Both optical and electrochemical tests are deeply discussed. The nucleic acid and non-nucleic acid detections are compared respectively. Then, the prospect of further improving smartphone-based SARS-CoV-2 detection is proposed. This review will facilitate the development of smartphone-based tests and enable the facile analysis of SARS-CoV-2 more broadly.

Fig. 1.

Smartphone-based analysis of SARS-CoV-2: the red and green color indicates the relatively expensive and low-cost component of the test respectively; typical samplers, biorecognition, and transducer setup are exhibited; Mini devices and Mini instrument indicate the relatively complicated-designed or expensive instruments.

2. Smartphone-based optical analysis

Optical detections based on UV–visible absorption, phosphorescence, surface plasmon resonance, fluorescence, chemiluminescence, and surface-enhanced Raman scattering techniques are widely used to detect various analytes. Traditional optical detections require instrumental analysis, which needs a relatively high cost for the general public. For example, UV–Vis detections usually use UV–Vis spectrophotometers and microplate readers for a single sample and multiple sample detections respectively; Fluorescent detection of single analytes and multiple biological analytes can be performed with the aid of spectrofluorometers and flow cytometers; Other optical detections such as surface plasmon resonance (SPR)-, chemiluminescence-, and phosphorescence-based analysis also require relatively expensive instruments. On the other hand, signals such as absorbance, fluorescence, SPR change, etc., can be captured by smartphone cameras or portable smartphone attachments, whose information may be further analyzed by certain software. This process may replace traditional instrument detection and realize convenient home testing.

Nucleic acid detection is normally considered to be the most accurate and reliable method because other tests all tend to have more false results. For instance, the antigen and virus particle tests have relatively poor accuracy for low virus load samples. Meanwhile, the antibody levels of a person are uncertain resulting in a significant limitation of accurate diagnosis of SARS-CoV-2 infections. On the other hand, the analyses of antigens, antibodies, or viral particles have the advantages of rapidity, non-invasiveness, simplicity, and low cost. Both simple non-nucleic acid tests and nucleic acid analyses are important supplements to each other, which deserve further improvement.

2.1. Non-nucleic acid tests based on optical analysis

2.1.1. Indirect and direct data conversion by a smartphone

Some tests require simple optical signal capture equipment and then indirectly use a smartphone to calculate the test data. During the progress, an optical sensor is normally employed. Typically, some element is functionalized with certain receptors. It will induce color, fluorescence, or other optical signal change by recognizing the analytes. Through a smartphone, the information on the reaction sample can be obtained by analysis of image grayscale, absorbance, color RGB, intensity, etc. For instance, Nanozyme is an element that can catalyze the reaction of small molecules leading to a color change. Among the nanozymes, the horseradish peroxidase (HRP) mimicking nanoparticles (NPs) has been most widely used, which can catalyze the oxidation of H₂O₂ to ‧OH radical that further oxidize colorless 3,3′,5,5′- tetramethylbenzidine (TMB) to generate blue color ox-TMB. Liang et al. developed an HRP-like nanozyme-linked immunochromatographic sensor (NLICS) for the fast analysis (<1 h) of the antigen (nucleocapsid protein (NP)) of SARS-CoV-2 [58]. The system consists of a 3D printed U-shape immunochromatography assay (ICA) device ($1.50), an inexpensive photometer (<$10), and a smartphone with an app specially designed for this platform. During the test progress, NP interacted with the first specific monoclonal antibody (mAb1) which was coupled with the HPR-like nanozyme (Au@PtNPs) and sprayed on a conjugate pad of a testing strip. Then, Au@PtNPs-mAb1-NP migrated and conjugated to mAb2 immobilized on the T-line (Test line) of the test strip, forming Au@PtNPs-mAb1-NP-mAb2. This conjugate further catalyzed the TMB substrate solution to ox-TMB and exhibited blue color. A 450 nm laser (<$100) was used to generate blue light that could cross the reacted substrate. The light after filtration was then captured by the portable photometer. The absorbance value (OD450) and the corresponding concentration were demonstrated by the smartphone with the app. NLICS detected NP with a detection limit of 0.026 ng/mL and had a linear range between 0.05 and 1.6 ng/mL within 25 min. Of 21 COVID-19 cases, NLICS found 76.2% NP-positive clinical serum samples, while a commonly used enzyme-linked immunosorbent assay (ELISA) only found 47.6% NP-positive cases. Both NLICS and ELISA give 100% accuracy for the NP-negative cases of 80 healthy blood donor samples. Though the accuracy of this method for the detection of positive samples is expected to be improved, this work provides a promising strategy for designing reproducible optical sensing devices for the analysis of SARS-CoV-2 antigens.

Smartphone apps and sampling devices are difficult to prepare under conditions without enough facilities. Additional optical signal capture equipment may increase the complexity of the test. To solve this problem, some strategies directly use images taken by smartphone cameras to simulate corresponding data. For instance, Fabiani et al. fabricated a simply prepared paper-based immunoassay using 96-well wax-printed cardboard (<$1) for colorimetric sensing of SARS-CoV-2 spike protein (SP) (Fig. 2 ). Sandwich-like immune chains were supported with anti-mouse IgG-conjugated magnetic beads and SARS-CoV-2 spike antibody-HRP chimeric monoclonal antibody (MAb-HRP). This “sandwich” catalyze TMB substrates to show different shades of blue and color-to-intensity data were transformed using a smartphone in conjunction with a free app (Spotxel reader) [59]. Spotxel Reader enables the reading of multiple samples in array formats such as samples in commercial 96, 48, or other numbers of well plates or printed arrays. This method detected SP in saliva up to 10 μg/mL and a detection limit of 0.1 μg/mL within 30 min. By comparison with 12 nasopharyngeal swab samples from patients infected with the same Delta variant studied by RT-PCR, 100% accuracy was found. Compared to the analysis of nasopharyngeal swabs using RT-PCR, the cost of analyzing one patient's specimen is reduced from about $20 to $3. On the other hand, a paid premium version ($1195 for a Perpetual License) of the Spotex reader can directly plot standard concentration curves and calculate the concentration of SP in unknown samples. However, if the microwell contains standard samples, it may directly recognize SARS-CoV-2 as positive or negative based on the color comparison, which is completely free and has the potential to meet the needs of the general public. Some other free software such as the Color Picker has also been used to analyze SARS-CoV-2 sensing results [54,60]. The application of these simply fabricated devices and free programs on a smartphone will be of great importance to reduce the cost and facilitate broad availability.

Fig. 2.

Direct data conversion from a smartphone without the assistance of any instrument: A scheme of the setups of the smartphone-assisted optical-sensing devices for analysis of SARS-CoV-2 using a simple sampler and a free smartphone app (Spotxel Reader). Reprinted with permission from Ref. [59], Copyright 2022, Elsevier.

Various smartphone-based optical sensors have been reported to detect SARS-CoV-2 antigens and antibodies (Table 1 ). Some of these methods are near 100% agreement with the PCR assays for analysis of clinical samples [59]. Since these tests have advantages including simple setup and operation, low cost, excellent accuracy, and satisfactory precision, it can be expected further modification of these methods may provide wider availability.

Table 1.

Detection of SARS-CoV-2 antigens and antibodies by smartphone-based optical analysis.

Sensor	Samples	Mechanisms	Detection limit; time; sensitivity/accuracy	Data analysis	Ref.
NPs transfer biosensors	Virus in face masks	A polymer-modified filter paper stored antibody-decorated Au NPs for recognizing NP	NP (3 ng mL−1); <10 min; 96.2% sensitivity and 100% specificity	The smartphone camera with a commercial reader	[61]
NLICS	Clinical samples	LFA for color reaction on the test strip for recognizing NP	0.026 ng/mL NP; 25 min; 76.2% sensitivity and 95.1% accuracy	Smartphone with author-designed App and portable photometer	[58]
Colorimetric immunosensor	Saliva samples	Antibody conjugated magnetic beads to recognize SP and a 96-well wax-printed paper plate for color visualization	100 fg/mL SP, 1.6 × 101 PFU/mL SARS-CoV-2; 45 min; 100% accurate for 6 negative and 6 positive saliva samples	Smartphone with Spotxel free-charge app for image analysis	[59]
HFIS	Clinical serum samples	PTEM-coated microplate for the immunoassay and a sandwich recognition method for analysis of NP	NP (7.5 pg/mL); 45 min; 72% sensitivity and 95% accuracy	Optical fibers for light transmission and collection; Designed App for data processing	[62]
TEMFIS	Patient samples, vaccinees and healthy blood	TEM-microplate with optical fibers transmission immunosensing of Nab	Nab; 45 min; positivity (sensitivity) in 92.68% and 76% vaccinees, negativity (specificity) in 100%	Optical fibers for light transmission and collection; Designed App for data processing	[63]
MSAA	Clinical swab samples	Sandwich complexes formed between magnetic bead/NP/PtNP and bright field images of oxygen microbubbles generated through catalysis of H2O2 decomposition	NP (0.5 pg/mL); 30 min; PPA = 97%, 53%, 26%, 45 for symptom onset <7, 7–12, >12 days and Asymptomatic, 97% for negative cases	Computer vision image recognition and ML-based algorithms on smartphones	[64]
QD-LFIA	Human serum or whole blood samples	IgG or NAb could combine with RBD-His, reacting with QD@anti-His mAb, which migrates to T1 and T2 lines respectively	IgG; Nab 98.8% (80/81) and 88.9% (72/81) were positive for IgG and Nab of recovered patients; 90% (63/70) and 82.9% (58/70) were positive for IgG and Nab for 64 vaccinated people	Self-produced portable fluorescence real-time camera reader; Data transfer by WIFI to smartphone	[65]
IPCF	Atificial saliva	Label-free detection of SP was realized based on an antigen-antibody reaction	SP (429 fg/mL); <1 h; Not investigated	Go Spectro on a smartphone	[66]

Note: Au NPs, gold NPs; PFU, plaque-forming units; HFIS, high-throughput fiber integrated immunosensing system, which was constructed by a PTEM-based high-throughput immunoassay platform and handheld microplate reader connected with a bundle of optical fiber; PTEM, Polycarbonate track-etched membrane; TEMFIS, A track-etched membrane microplate and optical fibers transmitted immunosensing smartphone platform; Nab, neutralizing antibody; PtNP, platinum nanoparticle; MSAA, Microbubbling SARS-CoV-2 Antigen Assay; ML, machine learning; QD-LFIA, quantum dot lateral flow immunoassay strip; IPCF, an optical sensor based on an imprinted photonic crystal film; Go Spectro, a device that turns a smartphone into an ultracompact and powerful light hand spectrometer; F-IPCF, antibody functionalized-IPCF; PPA, positive percent agreement.

2.1.2. Portable smartphone imaging system

Microscopy provides an important strategy for observing microorganisms. Large microscopes are expensive and complex to operate. Imaging bioassays normally require expensive microscopes ($750 to over $89,000) for detection. This makes the broadly available detection of viruses difficult to achieve. On the other hand, some portable microscopes are developed for smartphones, which are cost-effective and easy to use. It is difficult to observe small-size viruses with ordinary portable microscopes. However, through a certain virus particle identification strategy, the real-time infection status of the virus can be observed. Many virus particles were carried by droplets and aerosols. Digital imaging systems can be developed to obtain information on droplet species [67]. In typical imaging progress, recognization elements with catalytic activity and optical properties are labeled and confined to tiny microreactors. When the target molecule appears in the droplet, the trapped droplet interacts with the recognition element and displays an optical signal change in the microreactor within a short time. Kim et al. combined a smartphone with a handheld microscope. A paper-based microfluidic chip was modified with antibody-conjugated submicron particles, which captured the airborne droplets of human saliva samples spiked with SARS-CoV-2 (Fig. 3 a) [68]. Based on antibody-antigen binding and subsequent particle aggregation, the capture-to-assay time was smaller than 30 min. Two sprays at a distance of 6 inches showed the most significant differences between virus samples and controls (Fig. 3b and c). A fan was set up since the virus capture required air circulation (Fig. 3d). The virus was observed by a smartphone-based fluorescence microscope by counting the immunoagglutinated particles on the paper chip and the data was simulated by MATrix LABoratory (MATLAB) (Fig. 3a). Besides the smartphone, this portable imaging system was simply produced using low-cost components including an LED, a 9-V battery, acrylic film, and a mini microscope attachment with a total cost of $46.60. The setup for fabricating this mini imaging system was simple and inexpensive, but the calculation software (MATLAB) is a computing platform that is normally used by professionals. A more complete integration program is expected to be developed for the users in the future. Breshears et al. employed a smartphone-based fluorescence microscope and paper microfluidic chips to construct particulometric SARS-CoV-2 assay for clinical saline gargle samples [69]. The unprocessed image indicated the virus infection conditions and free ImageJ was used to process the imaging data. The limit of detection was 10 ag/μL; for n = 27 clinical human samples, 13 of which were positive by RT-qPCR, the sensitivity, specificity, and accuracy were 100%, 86%, and 93%, respectively. The method is simple, sensitive, and accurate, and at a total cost of only $46.40 per device, These portable smartphone imaging systems have the potential for a wide range of applications.

Fig. 3.

Detection of SARS-CoV-2 from airborne droplets by a smartphone-based fluorescence microscope. (a) From left to right: The air collection and capture by spraying the spiked SARS-CoV-2 salvia samples; The antibody particles were modified on the chip; Smartphone-based fluorescence microscopic observation; The data analysis with a MATLAB script; (b) Two-times spraying results; (c) Five-times spraying results, and (d) comparison of two-times spraying with fans on and off. 4 times repeating pixel counts from 5 different images of a single channel for the control (0) and samples (600 pg/mL) were compared by column plots. Reprinted with permission from Ref. [68], Copyright 2022, Elsevier.

2.1.3. Plasmonic sensors for signal magnifications

Detections of SARS-COV-2 without target amplification may have poor sensitivity for analysis. On the other hand, certain nanotechnologies have significant signal-enhancing capabilities that may overcome this problem. Plasmons are generated when electromagnetic lightwave interacts with the free surface electrons on nanosized metals [70]. Surface plasmonic enhancements have been employed for amplifying the signal in biosensors [71]. This includes SPR [72,73], surface-enhanced Raman scattering (SERS) [74], surface-enhanced fluorescence (SEF), etc [75]. These surface plasmonic enhancement strategies have also been employed for SARS-CoV-2 detections in combination with smartphones. Noble metal NPs possess remarkable plasmonic properties [76,77], which can be modified with bioreceptors and show plasmonic resonance change after interacting with the antigen of SARS-CoV-2 [78]. For instance, Olalla Calvo-Lozano et al. employed an electron beam-deposition system and fabricated 1 nm of titanium (Ti) and 49 nm of gold (Au) sensor chips by metal evaporation [79]. With surface cleaning, modification biofunctionalization, immobilization, etc., a serological biosensor assay was prepared for analysis of multiantigen (RBD peptide and NP). The sensor surface was excited by a collimated halogen light source ($1000 - $3000), and the reflected light was collected and coupled to a CCD spectrometer ($2000 - $3000) via an optical fiber. The resonance peak position (Δλ) that indicated the interactions of the antigens could be tracked in real-time using custom readout software (Fig. 4 a), which might be set up on a smartphone in the future. This plasmonic sensor rapidly (<15 min) analyzed SARS-CoV-2 in clinical samples (n = 120) with detection limits in ng mL⁻¹ and showed sensitivity and specificity of 99% and 100% respectively (Fig. 4b). Ahmadiv et al. developed terahertz (THz) (Terahertz band range from 1 mm to 0.1 mm) plasmonic metasensors for the analysis of SARS-CoV-2 antigen [80]. A miniaturized plasmonic immunosensor was fabricated based on toroidal electrodynamics that could confine plasmonic modes in the THz frequencies. The designed metasurface with metallic unit cells was fabricated by photolithography technique and e-beam metallization. The toroidal dipole mode was excited by a quasi-infinite metasurface and an S1 protein antibody functionalized Au NPs. In the presence of S1 protein spiked samples, the resonance shifts could be induced and measured by a THz time-domain spectroscopy instrument (about $29,950). The detection limit was as low as ∼4.2 fM and the authors suggested that the smartphone-based operation of plasmonic metasensors may be published elsewhere for the diagnosis of SARS-CoV-2 infections. Huang et al. developed an SP-specific nanoplasmonic resonance sensor using a generic microplate reader with a specific nanostructure and proper antibody functionalization on the surface [81]. The nanoplasmonic assays were produced by replica molding progress. The tapered nanopillar arrays were fabricated by photolithography equipment ($8,730.99-$9,492.99) and plasma etching ($2000-$15000). After spraying with optical adhesive liquid, ultraviolet irradiation, nano Ti and gold Au deposition, etc., a sheet was prepared and glued to a 3D printed chip cartridge or an open-bottom 96-well plate ($3-$20). The plasmonic sensor was then fabricated and showed resonance change with the capture of the pseudovirus without the need for additional optics (Fig. 4c and d). A designed smartphone app directly calculated the SARS-CoV-2 pseudovirus in spiked samples in one step within 15 min from 0 to 6.0 × 10⁶ virus particles (vp)/mL with a quantification limit of about 4000 vp/mL SARS-CoV-2. These nanoplasmonic biosensing platforms may enable amplification-free, accurate, selective, and sensitive detections [[82], [83], [84]]. Normally, low-cost plasmon-enhanced substrates can be constructed by noble metal nanomaterials such as Au or Ag NPs, nanostars, and nanorods with facile synthetic approaches [85,86]. However, the photolithography instrument, nano metal evaporation setup ($10000-$20000), and plasma etching progress mentioned for fabrication of the sensor chips are not widely available. Therefore, whether low-cost plasma detection products can be produced at low cost will be an important factor in their availability.

Fig. 4.

(a) A two-antigen co-immobilized SPR sensor biochips for COVID-19 serology tests. (b) Sensor signal distribution, sensitivity (Sens), specificity (Spec), positive predictive value (PPV), negative predictive value (NPV), and threshold of 100 COVID-19 positive (PS) and 20 negative (NG) clinical samples; the right axis is the total immunoglobulins (Ig) concentration calculated based on the WHO standard; Reprinted with permission from Ref. [79], Copyright 2022, American Chemical Society. Nanoplasmonic sensor chips for analysis of SARS-CoV-2 pseudovirus. (c) Scheme of nanoplasmonic sensor chip cartridge for analysis of SARS-CoV-2 pseudovirus. (d) The modification of the proper antibody on the sensor chip cartridge for specific SP recognition. Reprinted with permission from Ref. [81], Copyright 2021, Elsevier.

2.2. Nucleic acid tests based on optical analysis

The samples for nucleic acid tests are frequently collected from a swab of a specimen from a patient's throat or nose and then the virus RNA was detected by RT-PCR after the pretreatment. One disadvantage of these tests is the requirement of a long time and relatively advanced laboratory conditions. However, several quick and simple optical detection strategies have recently emerged that are expected to be popularized in the field of nucleic acid tests.

2.2.1. Smartphone-based LAMP

The loop-mediated isothermal amplification (LAMP)-based method has been used for point of care (PoC) detection of various viruses in a short time. Normally, four to six different primers were designed for recognizing the corresponding segments on a target and the reaction was performed at a relatively low temperature (60–65 °C). The strand displacement activity of DNA polymerase facilitates the denaturation of DNA by heating unnecessarily. Some LAMP attachments tend to be simpler, less expensive, and smaller than the thermal cyclers required for PCR. Reverse-transcription LAMP (RT-LAMP) can take one-step RNA preparation in a thermos by noninvasive sample collection, and an optical signal change can be observed with the naked eye. The entire sample-to-result normally takes less than 60 min [[87], [88], [89], [90], [91]]. At present, LAMP commercial household SARS-CoV-2 test strips have been adopted and each test cost about $50. Meanwhile, convenient smartphone detection methods based on LAMP are being explored. For instance, Manzano et al. developed a rapid smartphone-based diagnostic test (<20 min) for the detection of RNA of SARS-CoV-2 in the extracted clinical samples based on RT-LAMP onto a PoC platform (RT-eLAMP) [92]. A kit consisting of a complementary metal-oxide-semiconductor (CMOS) ion-sensitive field-effect transistors (ISFET) microchip and a microfluidic module, which accommodates two wells, one for the sample and one for the control reaction was used for the sample test. After RNA was extracted, it was analyzed by different methods including RT-qPCR (RT-PCR using a real-time benchtop platform), RT-qLAMP (RT-LAMP assay used a real-time benchtop instrument), and RT-eLAMP (The lab-on-chip (LoC) platform used a smartphone with a customized App to process the sensing data. When the reaction was stopped, the fitted data was synchronized to Cloud Server (AWS) and the GPS location of the test sample could be shared on the data map. This RT-eLAMP method showed a sensitivity and specificity of 90.55% and 100% with a detection limit of 10 copies per reaction for screening 52 samples, including 34 positive and 18 negative isolates, which is comparable to RT-qPCR and RT-qLAMP.

Several smartphone-based LAMP methods have provided promising strategies for the detection of SARS-CoV-2 (Table 2 ). Most of these methods have advantages such as the short time, and extraction-free, which can be performed at lower temperatures compared to PCR. Some of these methods also achieve the sensitivity and accuracy of the PCR tests. However, the designed app, the sampling progress, and the primer design for LAMP have not been communalized. The high sensitivity and improper operation may lead to false positives, so strategies to further modify the smartphone-based LAMP analysis are still needed.

Table 2.

Detection of SARS-CoV-2 using smartphone-based LAMP.

LAMP	Detection Limit	Time	Devices	Target	Compared to Traditional PCR	Ref
PD-LAMP	35 × 104 vp/mL in saliva	35 min	A microfluidic chip using a portable heating unit	N gene and the ORF1ab gene	No cold storage or extraction, low cost, fast	[56]
RT-LAMP	50 RNA copies/μL in the VTM solution	30 min	Manufactured 3D cartridge	Orf 1 ab, S, and Orf 8	100% accurate, no sample/reagent mixing, amplification, or extraction, low cost, fast	[90]
RT-LAMP	2 × 101 genome copies/μL for nasopharyngeal swab samples	13–51 min	A portable IoT-based POC genetic analyzer	Three target genes (As1e, N, and E genes)	More specifically for SARS-CoV-2 from respiratory viruses, the pre-extraction is not needed but has lower sensitivity, low cost, fast	[93]
RT-LAMP	1 × 103 copies/μL in saliva samples	30 min	Simple tube	RNA	98.8% accurate, low cost, fast	[94]
RT-LAMP	5 copies/μL of the saliva sample	<45 min	Microfluidic Reagent Cartridge	RNA	Without commercial thermocyclers, faster and more sensitive than PCR; 100% agreement with PCR results for 2 clinical samples	[95]
RT-LAMP	0.5 copy/μL for cold chain fruits	15 min	Gel RT-LAMP system	Virus	No virus pre-lysisor, purification or RNA extraction is required, low cost	[96]
smaRT-LAMP	103 copies/mL in spiked saliva samples	25 min	A hot plate, cardboard box, and LED lights	2 nucleocapsid (N) and ORF1ab genes	100% accurate, low cost, fast	[97]
Cap-iLAMP	5-25 viral genome copies per μL in gargle lavage	<1 h	PCR tube	Orf1a and N gene	100% accurate for high viral load and 83% of all investigated positive samples, low cost, fast	[60]

Note: Cap-iLAMP, capture and improved loop-mediated isothermal amplification; VTM, viral transport medium; Internet of things (IoT) [98]-based diagnostic devices; smaRT-LAMP, smartphone-based RT-LAMP.

2.2.2. Smartphone-based CRISP-Cas-biosensing technologies

The CRISPR systems were first identified in archaea and bacteria that had immune functions to degrade foreign viruses and plasmids with the CRISPR-associated (Cas) enzyme. CRISPR/Cas-system could mediate genome editing for various applications [98]. CRISPR-Cas-based strategies have been developed to target nucleic acid by using CRISPR RNA (crRNA), which guides Cas enzymes to cleave the targets by hybridizing to complementary sequences. CRISPR with Cas9, Cas10, Cas12, Cas13, and the enzyme combined system (such as Cas13 and Cas12) have been involved to analyze the SARS-CoV-2 virus and its variants [[99], [100], [101], [102], [103], [104], [105]]. For instance, Cas13 could be complexed with an editable CRISPR RNA (crRNA), generating a nuclease-inactive ribonucleoprotein complex (RNP). When the RNP hybridizes to the complementary target RNA, it activates the HEPN (higher eukaryotic and prokaryotic nucleotide-binding domain) motif of Cas13a and then cleaves the surrounding ssRNA, causing the initially quenched fluorophore to fluoresce. CRISPR-Cas-based detections tend to be very sensitive, so in many cases, the virus can be directly detected without extracting RNA or nucleic acid amplification. Fozouni et al. developed an amplification-free CRISPR-Cas13a-based assay for the analysis of SARS-CoV-2 RNA from a nasal swab [106]. The assay accurately analyzed pre-extracted RNA from positive clinical samples within 5 min and achieved a sensitivity of 100 copies/mL within 30 min. The further combinations of two or three crRNAs significantly increased the sensitivity of detections to ∼30 copies/μL by activating more Cas13a per target RNA (Fig. 5 a), which targeted multiple regions of the viral RNA and thus enhanced the sensitivity. Furthermore, a low-cost laser illuminator ($10 - $70) was used as the excitation light source, and the fluorescence was detected by a smartphone with an inexpensive optical collector (Fig. 5b and c). The imaging system consists of a compact triplet lens and interference filter. The optics and lighting components were packaged in a custom cassette into which the sample chip could be placed for loading images. Automatic time-lapse imaging was enabled by a custom Android app and Bluetooth receiver. The device was placed in a constant temperature incubator at 37 °C facilitating the Cas13a reaction. Finally, the response curve was obtained and analyzed by MATLAB, and the concentration of SARS-CoV-2 RNA could be calculated. This strategy detected SARS-CoV-2 RNA using the three crRNA Cas13a assay and a genomic RNA isolated from supernatants of virus-infected Vero CCL-81 cells (Fig. 5d) of different dilutions within 30 s. The accuracies for 500 copies/μL, 200 copies/μL, and 50 copies/μL were 100%, 100%, and 50% respectively (Fig. 5e). This method also correctly identified all five SARS-CoV-2 positive patient RNA samples within 5 min (Ct values 14.37 to 22.13).

Fig. 5.

(a) Combining crRNAs and Cas13a for the detection of SARS-CoV-2. (a) Scheme of two different RNPs binding to the same SARS-CoV-2 RNA at different locations, which cleave RNA reporter and generate fluorescence signal. (b) Scheme of smartphone-based microscope for fluorescence detection (left). Picture of the device and sample image after running a Cas13a assay captured by the smartphone camera (right). (c) Results of fluorescence signal as a function of time from the Cas13a assay obtained from the smartphone-based device using 3 combined crRNAs (crRNA 2, crRNA 4, and crRNA 21) for the analysis of 2 different dilutions of SARS-CoV-2 viral RNA, which was isolated from infected Vero CCL-81 cells (500 and 200 copies/μL) and RNP alone. (d) The slope of the curve from (c) with ±95% confidence interval. (e) Detection accuracy of the Cas13a assay using genomic SARS-CoV-2 viral RNA. Reprinted with permission from Ref. [106], Copyright 2021, Elsevier.

Liang et al. found that CRISPR-Cas12 assay could detect the mutations such as K417 N/T, L452R/Q, N501Y, E484K/Q, D614G, and T478K) using mutation-specific crRNAs, to distinguish the variants of SARS-CoV-2. This method showed 100% concordance with the sequencing approach for the major SARS-CoV-2 variants with a detection limit of 10 copies/reaction [107], which achieved the same accuracy as PCR. The authors suggested a microfluidic chip and smartphone-based analysis would be designed to detect these mutations simultaneously. However, this method was adapted for emerging mutations that already have been conducted with SARS-CoV-2 nucleic acid amplification tests using extracting nucleic acid. Further development is still expected.

By combining LAMP technology with CRISPR technology and biosensing strategies, the sensitivity and accuracy of the test were further improved [108,109]. Song et al. reported a colorimetric sensor based on DNAzyme reaction triggered by LAMP with CRISPR-Cas9 (which is defined as DAMPR assay) for ultra-sensitively detecting SARS-CoV-2 and variant genes with a detection limit of 1.08 aM (10 copies/sample), 0.92 aM (9 copies/sample), and 1.37 aM (13 copies/sample) for ORF1, N, and S genes, respectively [110]. Black poly(lactic acid) (PLA) filament was used to print the housing of a dark room. A heating bed was designed to facilitate the RT-LAMP reaction evenly to the entire 96-well plate at 65 °C. The LED lights were attached to the dark room as the light source. A consistent focal distance of the smartphone to the system was controlled to maintain the nonuniformity of the lighting condition by a holder. This DAMPR fastly detected the SARS-CoV-2 within an hour and showed a clinical sensitivity and specificity of 100% of 136 clinical samples. It also successfully discriminated the D614G (variant-common), T478K (delta-specific), and A67V (omicron-specific) mutations of the SARS-CoV-2 S gene of 70 SARS-CoV-2 delta or omicron variant patients. CRISPR-Cas systems combined with different techniques and smartphones are promising for realizing accurate, sensitive, and general detection of SARS-CoV-2 (Table 3 ). Most of these detections read signals based on fluorescence or color change. Thus, cheap optical meters (less than $10) can be used to analyze the sensing results. On the other hand, the sampling process of CRISPR-Cas-biosensing technology is still relatively complicated for common people. Further development of easy-to-handle semi-finished sensors may show potential to facilitate public use.

Table 3.

Smartphone-based detection of SARS-CoV-2 RNA using CRISPR-Cas-biosensing strategies.

App	Mechanism	Accuracy	Detection Limit, Time	Target	Ref.
Free Color Picker App	The nucleic acids triggered CRISPR-Cas12a-based degradation of ssDNA that linked two AuNPs, generating the color of dispersed AuNPs	100% positive and 100% negative agreement with qPCR for 20 positive and 30 negative clinical swab samples	1 copy/μL, 90 min;	Virus RNA of clinical samples	[54]
Customized software for fluorescence measurement	RNA facilitated the CRISPR-Cas12a cleaved a fluorophore quencher-labeled nucleotide reporter, generating fluorescence	90% accuracy of 115 nasopharyngeal swab samples from individuals showing COVID-19-like symptoms	6.25 copies/μL, <1 h;	Viru RNA of Clinical samples	[111]
Designed app for fluorescence measurement	One-pot SHERLOCK reaction with an RNA paper-capture process	96% sensitivity and 95% specificity in clinical saliva samples	1000 copies/ml, 55 min	RNA of B.1.1.7, B.1.351, or P.1 variants	[112]
TOPSE smartphone app	A paper-strip-based platform using FnCas9 to cleave the ssDNA probe and generate fluorescence upon target binding	87% sensitivity and 97% specificity for clinical samples	Not mentioned, 75 min	S gene mutation N501Y for clinical samples	[113]
Customized App for fluorescence measurement	CRISPR-Cas12a-assisted RT-RPA fluorescence assay via magnetic-based nucleic acid concentration and transport	27 out of 27 agreements with RT-qPCR	1 genome equivalent/μL, <30 min	RNA from unprocessed clinical NP swab eluates	[114]
OPTIMA-dx app	RT-LAMP isothermal amplification was coupled with in vitro transcription and Cas13-based detection in one step	95% sensitivity and 100% specificity for clinical samples	10 copies/μL for a synthetic SARS-CoV-2 genome	RNA from clinical samples	[115]
Author-designed DAMPR app	DNAzyme reaction triggered by LAMP with the recognization of the target and induce a color change of ABTS and H2O2	Both sensitivity and specificity are 100% for 137 clinical samples; Effective for delta or omicron variant patients	ORF1 gene (1.08 aM), N gene (0.92 aM), S gene (1.37 aM), <1 h	SARS-CoV-2 S gene, ORF1 gene, N gene, and the mutations	[110]

Note: NP, nasopharyngeal; RT-RPA, reverse transcription recombinase polymerase amplification; DM, droplet magnetofluidics; TOPSE [116], True Outcome Predicted via Strip Evaluation; FnCas9, Cas9 ortholog from Francisella novicida; iSCAN, RT-LAMP-coupled CRISPR-Cas12 module for rapid, sensitive detection of SARS-CoV-2; gRNAs, guide RNAs; OPTIMA-dx, a smartphone application developed by the related authors; DAMPR, DNAzyme reaction triggered by LAMP with clustered regularly interspaced short palindromic repeats (CRISPR); ABTS, 2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid).

2.2.3. CATCH detections

A considerable number of people have samples with low viral loads [117], requiring extremely sensitive detection methods. However, sensitive methods normally require either nucleic acid amplification, RNA extraction, or complicated primer-designed progress, and these procedures are temporarily difficult to apply to the public. To address these issues, Noah R. Sundaah et al. investigated new nanotechnology for catalytic amplification of transition-state molecular switches (CATCH), which enable simple, accurate, and sensitive detection of RNA targets in SARS-CoV-2. A DNA-enzyme hybrid complex was used to form a molecular switch (Fig. 6 A). By adjusting the ratio of its components, the multicomponent molecular switch was fabricated into a highly reactive transition state that is easily activated upon sparse RNA target binding, thereby significantly turning on enzymatic activities, resulting in very sensitive fluorescent detection (<1 h at room temperature) without the requirement of PCR amplification and heating steps (Fig. 6B). This method recognized 100% positive (n = 24) and 92% negative (n-25) swab extracted RNA samples compared to the clinical RT-qPCR results. The CATCH approach had a very low detection limit (∼8 RNA copies/μl) and sensitivity, which correctly identified 93.34% (n = 15) and 100% positive (n = 9) heat-treated swab samples without extraction of RNA. The sampling progress could be performed in both high-throughput 96-well or portable microfluidic assays (Fig. 6C-D). Meanwhile, smartphone detection devices can be simply constructed with a light-emitting diode (LED) source, an optical filter, and a magnifying glass placed in front of the smartphone camera to improve image quality. The chemifluorescence can be readily detected by the smartphone (Fig. 6D-F) enabling direct detection of SARS-CoV-2. Without nucleic acid amplification, RNA extraction, specialized instrument, complicated primer design, and dedicated fluorescence probes, this method significantly reduces the complexity of nucleic acid detection and has great potential for further improvement.

Fig. 6.

(A) Scheme of the CATCH assay. The CATCH assay employs the specific binding of SARS-CoV-2 RNA to activate molecular switches. Each molecular switch consists of an inhibitory DNA complex that binds and inactivates the polymerase and Taq DNA polymerase by inhibitory and enhancer strands. The viral RNA target destabilizes the inhibitory complex and releases the active polymerase while hybridizing it to the enhancer strand. The molecular switches in different states of target responsiveness including closed, transition, and open (right) were prepared. The switch is completely inactive and less susceptible to activation by sparse RNA targets in the closed state while the switch is activated and largely unresponsive to the target in the open state. On the other hand, different forms of switches exist in a delicate balance that can easily be altered by trace amounts of RNA targets in the transition state, which exhibits sensitive responsiveness. (B) The CATCH assay utilizes an additional enzymatic cascade to transduce and amplify target-induced polymerase activity as a fluorescent readout, greatly enhancing the signal response for sensitive analysis of low-load clinical viral samples. (C) A 96-well format for high-throughput applications (top) and a miniaturized microfluidic device (bottom). (D) The photograph of the smartphone-based fluorescence analyzer. (E) Unfiltered and filtered fluorescence emission spectra of the LED source. (F) CATCH assay performance in microfluidic and plate formats showed good agreement with each other. Reprinted with permission from Ref. [118], Copyright 2021, The Authors.

2.2.4. Simple smartphone-based optical sensors

Recently, there have been facile optical sensors with very simple designs that change their optical signal when exposed to SARS-CoV-2 RNA. For instance, thermotropic liquid crystals (LCs) are very sensitive to phase transition. LCs-based sensors have been used to detect several analytes of the specific binding of molecules that cause transformations of LC and its color change. Compared to many other biosensors, LC-based sensing provides simpler, cost-effective, rapid, and selective detection of various targets [119]. Xu et al. reported the design of an LC-based smartphone analysis method for the detection of SARS-CoV-2 RNA (Fig. 7 A-B) [120]. A 2.5 × 2.5 cm optical cell-based kit as constructed in Fig. 7C was prepared by pairing a bare glass slide and a DMOAP-functionalized glass slide. An opening 2-mm-thick poly(dimethylsiloxane) (PDMS) spacer was used to space the two surfaces and allowed the analysis and injection of test samples. A partially self-assembled cationic surfactant monolayer was formed at the aqueous-LC interface, where LC reorientation enabled the adsorption of ssRNA and/or ssDNA at the interface. Then, a 15-mer ssDNA probe that contains a complementary sequence of the SARS-CoV-2 RNA was adsorbed onto the cationic surfactant-loaded aqueous-LC interface. The ordering transition in the LC surface has a close relationship with the targeted nucleotide sequence. A very low concentration (30 fM) of SARS-CoV-2 RNA selectively drove the ordering transition in the LC film and induced the color change, but a 3-base pair mismatch of SARS ssRNA insignificantly influenced the transition. This allowed the LC-based sensor to analyze SARS-CoV-2 sensitively and selectively, which can be viewed by a smartphone with an app (Fig. 7D) to enhance the accuracy of the color readout.

Fig. 7.

(A) Schematic illustration of the LC film to the adsorption of SARS-CoV-2 ssRNA (ssRNA_CoV). The inset shows the dynamic response of the optical micrographs (crossed polarizers) for the dodecyltrimethylammonium bromide (DTAB)/ssDNA probe-decorated E7 film before and after the adsorption of ssRNA_CoV. Scale bars, 100 μm. (B) The optical appearance of the test kit viewed under a lamp; (C) Photograph of the kit. (D) Test result readout by a smartphone App: The negative results for ssRNA_SARS <100 nM and positive results for ssRNA_CoV >30 fM indicate excellent selectivity of the test. Reprinted with permission from Ref. [120], Copyright 2020, Elsevier.

Zhao et al. developed a 3D-printed smartphone platform for the detection of SARS-CoV-2 RNA [121]. This probe was functionalized with orange-red emitting TAMRA and green-emitting FA dyes used as internal standard and sensing agents. Under 365 nm UV excitation, the emission intensity shows a ratiometric change, switching on and off at 580 nm and 518 nm, respectively. The color change from orange-red to green and the signal can be analyzed by a smartphone with an RGB system within 25 min. The detection limit of SARS-CoV-2 nucleic acid is 0.23 nM. These facile sensors are very convenient, simple in design, and easily repeatable. However, the verification of the sensitivity and accuracy for real sample analysis is expected.

3. Mini electrochemistry test platforms

Electrochemical detections tend to have higher sensitivity and selectivity than optical analysis, which has been used for the accurate analysis of both nucleic acids of SARS-CoV-2 and non-nucleic acids associated with the virus infections [[122], [123], [124], [125]]. A normal system includes three electrodes, an electrochemical cell, and an electrochemical workstation for reading the signals [126,127]. Different analytes can be selectively analyzed by modifying the relevant probes with different recognition agents. Nanomaterials and bioreceptors could be employed to modify the electrode for improving the detection limit, selectivity, and sensitivity [128]. Both the size of the whole system can be minimized for realizing PoC detections. For instance, the electrodes and electrochemical cells have been fabricated on microfluidic paper-based analytical devices (μPADs) which are mini-analytical devices based on cellulose materials [129]. μPADs have several advantages including low cost and easy fabrication based on established patterning methods [130,131]. Some common electrochemical techniques such as cyclic voltammetry (CV), electrochemical impedance spectroscopy (EIS), chronoamperometry, and electroluminescence were used to show the sensing signals using mini instruments [[132], [133], [134], [135]].

3.1. Smartphone-based electrochemical analysis of non-nucleic acids

Currently, a few electrochemical methods (Table 4 ) were combined with smartphone-based analysis for the detection of antigens, and antibodies. Among these electrochemical detection techniques, EIS depends on the characterization of the resistance of electrode surfaces and the transduction of biosensors [136,137], which have been used frequently for the analysis of viruses. For instance, Marcelo D.T.Torres et al. developed a real-time accurate portable impedimetric detection prototype 1.0 (RAPID 1.0) for analysis of SARS-CoV-2 antigen (Fig. 8 ). The screen-printing method and wax-printing were used to fabricate the electrodes and pattern the electrochemical cells on the phenolic paper circuit board ($40.00/m²) or filter paper ($0.50/m²). Only 10 μL of saliva samples were required and the result could be obtained within 4 min. The selective interaction between the bioreceptor on the electrode surface (such as Angiotensin-Converting Enzyme 2 (ACE2)) to the SARS-CoV-2 antigen (i.e., SP) causes a change in interfacial electron transfer and charge-transfer resistance (R_CT). Thus, the SP was analyzed based on the increased resistance to charge transfer of the redox probe and measured by EIS using Sensit Smart (PalmSens) potentiostats. The sensitivity and specificity of RAPID 1.0 for nasopharyngeal/oropharyngeal swabs and saliva samples were 85.3% and 100%, 100% and 86.5%, respectively.

Table 4.

Smartphone-based electrochemical analysis of non-nucleic acids.

Sensor	Samples	Mechanisms	Analyte (detection limit); time	Ref.
Portable three-in-one biosensor	Pseudovirus or spiked samples	Selective binding of SARS-CoV-2 biomarkers to surface-linked capture probes produces current changes	RNA S gene (100 pM in PBS), SP (100 pg mL−1 in serum), SP antibody (10 ng mL−1 in serum), about 2 h	[139]
OECT test platform	Spiked serum and saliva sample	SARS-CoV-2 IgG bonded with SP through antibody-antigen reaction	IgG (1 fM); ≤30 min	[140]
MIP-based sensor	Nasopharyngeal swab samples	Sensor chip - TFE - interfaced with MIP for recognizing NP	NP (15 fM); <1 h	[141]
EAB-based sensor	Serum and artificial saliva	The binding of aptamer-modified electrode induced conformational electrochemical signal change	SP (10 pM); <5 min	[142]
Cotton-Tipped Electrochemical Immunosensor	Spiked nasal samples	NP antibody was immobilized on CNF-modified screen-printed carbon electrodes for recognizing N antigen	NP (0.8 pg/mL); >20 min	[143]
EIS based detector	Human serological samples	A 16-well plate containing sensing electrodes pre-coated with RBD of SP, which recognized anti-SARS-CoV-2 monoclonal antibody CR3022	Spike Antibody (CR3022) (0.1 μg/mL); not mentioned	[144]
A label-free voltammetric-based immunosensor	Clinical Samples of nasopharyngeal swabs	Attachment of the anti-nucleocapsid antibody on Au NPs-modified electrodes for recognizing NP	NP (0.4 pg.mL−1); about 3 h	[145]
LSG-based electrochemical platform	Clinical nasopharyngeal swabs	LSG sensors are coupled with Au NPs as sensing platforms where ACE2 is chosen as a biorecognition unit to recognize SP.	5.14 ng/mL and 2.09 ng/mL for S1 and S2 protein; 1 min	[146]
Capillary-flow immunoassay device	Human blood samples	Anti-N antibody is detected using chronoamperometry in a sandwich assay setup	IgG (5 ng/mL); <20 min	[147]
Aptamer based biosensor	Spike RBD recombinant protein	Aptamer-based sensing platform for impedimetric analysis of SP	SP (66 pg/mL); <40 min	[148]

Note: OECT, Organic electrochemical transistors; MIP, Molecularly imprinted polymers; LSG, laser-scribed graphene; TB, toluidine blue; EAB, Electrochemical aptamer; CNF, carbon nanofiber; RCA, rolling circle amplification that is an isothermal amplification method; RBD, receptor-binding domain; AuNPs, gold NPs; ssDNA, single-strand DNA; LSG, laser-scribed graphene; ACE2; Angiotensin-Converting Enzyme 2; SCX8, p-sulfocalix [8]arene; SCX8-RGO, SCX8 functionalized graphene; TFE, thin film electrode; BioFET, electrical double layer (EDL)-gated field-effect transistor-based biosensor.

Fig. 8.

Smartphone-based detection of SARS-CoV-2 using RAPID 1.0: (a) Detection of SARS-CoV-2 antigen by RAPID 1.0 in neat saliva and nasopharyngeal/oropharyngeal (NP/OP) swab samples; three-electrode configuration cell and electrodes (CE, counter electrode; WE, working electrode; and RE, reference electrode) were screen-printed on a phenolic paper circuit board or filter paper with conductive carbon; The reference electrode was printed by Ag/AgCl inks. The WE was modified by glutaraldehyde and ACE2, and bovine serum albumin subsequently. A Nafion permeable membrane was used for chemical preconcentration of cation species and protecting the electrode's surface against biofouling with the biological sample matrix; (b) The comparison of cost and detection time comparison between RAPID 1.0 and typical FDA-approved tests. (c) Photo of smartphone-based detections; (d) Nyquist plots measuring by EIS were obtained for different concentrations from 1 pg mL⁻¹ to 100 ng mL⁻¹ of SARS-CoV-2 SP. The inset shows the calibration curve based on the normalized charge-transfer resistance (RCT) values as a function of SP recorded in triplicate. Reprinted with permission from Ref. [138], Copyright 2021, Elsevier.

3.2. Smartphone-based electrochemical analysis of nucleic acids

Recently, various electrochemical probes modified by nanomaterials have been used for nucleic acid detections for the diagnosis of SARS-CoV-2 infections (Table 5 ). For typical electrochemical nucleic acid sensors, capture probes are used to modify the working electrode, and complementary target nucleotides are hybridized into the sensing interface. When a target nucleic acid is present, it hybridizes with the recognition nucleotide on the electrochemical probe, causing an electrochemical signal change. Compared to optical detections, the electrochemical analysis of the virus may have many advantages. For instance, some electrochemical sensors can detect non-nucleic acids such as antigens and antibodies while detecting nucleic acids [139]; Some electrochemical analyses can recognize unamplified SARS-CoV-2 RNAs so fast and sensitive that 4 copies of the virus in 80 μL saliva can be detected within 1 min [149]; Some reported electrochemical probes even have higher sensitivity than the PCR test [150], but most of these approaches require further validation with clinical samples. Although these methods have been considered low-cost strategies, the frequently used signal output instrument (Sensi-smart) in combination with a smartphone is relatively expensive (about 15,000 US dollars or more) [[151], [152], [153]]. On the other hand, some USB-disk electrochemical workstations have recently been developed and achieved lower-cost analysis [154]. In future research, simpler and cheaper electrochemical signal reading methods may be of great help in the popularization of the detections [155,156].

Table 5.

Smartphone-based electrochemical analysis of Nucleic acids.

Sensor	Samples	Mechanisms	Analyte (detection limit); time	Ref.
MECS	Clinical samples	The configuration of the tentacles nearby changes with the recognization of the nucleic acid, pushing the signal change.	RNA (4 copies in 80 μL); 1 min	[149]
Portable three-in-one biosensor	Pseudovirus or spiked samples	Selective binding of SARS-CoV-2 biomarkers to surface-linked capture probes produces current changes	RNA (100 pM in PBS), SP (100 pg mL−1 in serum), SP antibody (10 ng mL−1 in serum); about 2 h	[139]
Supersandwich-type electrochemical biosensor	Artificial targets and clinical RNA samples	TB enrichment by SCX8-RGO for supersandwich-type recognition of SARS-CoV-2 RNA	RNA (200 copies/mL); about 2 h	[157]
Multiplex RCA-based sensor	Clinical samples	Multiplex RCA for the detection of the N and S genes of SARS-CoV-2	N and S genes (1 copy/μL); <2 h	[158]
Paper-based electrochemical sensor chip	Clinical Samples of nasopharyngeal swabs	Thiol-modified ssDNA-capped Au NPs on top of the gold electrode was designed to target two separate regions of the viral N-gene	RNA (6.9 copies/mL); <5 min	[159]
E-INAATs	Artificial swab samples	The sensing device is pH-sensitive and shows potentiometric change to the N gene of the virus	N gene (2 × 102 copies/test); 10 min	[160]

Note: OECT, Organic electrochemical transistors; MIP, Molecularly imprinted polymers; LSG, laser-scribed graphene; TB, toluidine blue; EAB, Electrochemical aptamer; CNF, carbon nanofiber; RCA, rolling circle amplification that is an isothermal amplification method; RBD, receptor-binding domain; AuNPs, gold NPs; ssDNA, single-strand DNA; LSG, laser-scribed graphene; ACE2; Angiotensin-Converting Enzyme 2; SCX8, p-sulfocalix [8]arene; SCX8-RGO, SCX8 functionalized graphene; TFE, thin film electrode; MECS, self-actuated molecular-electrochemical system; E-INAATs, Electrochemical isothermal nucleic acid amplification tests.

4. Comprehensive cost compare

Different strategies were used for the detection of SARS-CoV-2 with the aid of smartphones. From these methods, several typical strategies with detailed setup information were selected for price comparison (Table 6 ). Because these methods detect different targets, the sensitivity and accuracy are not comparable under the current conditions. As can be seen from Table 6, the optical detections based on NLICS and smartphone microscopes may be the most cost-effective and simplest to be designed and repeated by common people. The recognition of SARS-CoV-2 by both technologies is based on simple optical sensors. Compared with the acquisition of other signals, optical signals are currently more convenient to be obtained for personal use. These methods are not necessarily the most sensitive. If they can satisfy the accuracy for recognizing initially infected individuals, further development of these techniques will make an important contribution to the widespread availability.

Table 6.

Cost comparison of different smartphone-based detection techniques for analysis of SARS-CoV-2.

Devices	Attachment Cost ($)	Additional (Cost $)	Strategy	Test	Ref.
NLICS	1.5	Hand Photometer (Sanfu) (about 30)	Immunoreaction and enzyme-catalyzed substrate color reaction	NP	[58]
RAPID 1.0	4.67	Sensit Smart (PalmSens) (about 15 k)	Electrochemical reaction-induced EIS change	SP	[138]
miSHERLOCK	15	Author made app (unkonw)	CRISPR-based PoC diagnostic platform provides fluorescent visual output	Viral RNA	[161]
Smartphone microscope	46.4	Pocket Microscope （25）	Isolated and counted the immunoagglutinated particles on the paper chip.	Droplets/aerosols containing virus particle	[68]
Harmony COVID-19	300	Author made software	Wet RT-LAMP reactions and heater/reader operated by a cell phone	Viral RNA	[162]
Nanoplasmonic sensors	Not mentioned	Xlement SPR100 (>20 k)	The plasmon resonance wavelength and intensity change on the virus-capturing	Virus particles	[81]

Note: k, 1000 dollars.

5. Other simple sensors

A widely applicable method requires a simple way to read the detection signal. SARS-CoV-2 may cause different physical/chemical state changes for a sensor, and by constructing corresponding analytical devices, the virus can be analyzed. However, besides electrochemical detection and optical analysis of SARS-CoV-2, other smartphone-based sensing strategies are still limited. Common signal reading sources including sound, light, electricity, heat, force, chemistry, etc., have been developed for sensing various analytes. Accordingly, more types of simple sensors may be expected for effective SARS-CoV-2 detections.

Currently, in addition to electrochemical and optical analysis, several other assays have been investigated for the analysis of SARS-CoV-2. For instance, microdevices attached to a smartphone have been developed to detect viruses [17,18] using flow rate assays recently. The microfluidic chip flow sensor measures the flow rate of various liquids within microchannels in real time through contact with the analytes [163]. Akarapipad et al. fabricated a device combing a paper-based microfluidic chip, which was designed by SolidWorks 2020 and printed with a 3D and a wax printer (Fig. 9 ) [164]. 4 parallel channels were fabricated on a single chip for high throughput analysis. The outer green boxes and three red squares at the three corners of the chip allow for orientation identification and localization of the channel area with automated flow measurements. The surface tension and capillary flow velocity profile were changed by the particle-target immunoagglutination. The SARS-CoV-2 negative samples enable Ab-particles to take more time to reach constant velocity, while the positive samples facilitate Ab-particles to take less time to reach constant velocity due to immunoagglutination. An antibody-conjugated particle suspension and a smartphone were used to recognize and monitor the virus particles from saliva samples based on the flow rate change. The flow profile was videoed by the smartphone camera and extracted using an author-designed program (Python script) that automatically searched the channels and provided results. This method detected SARS-CoV-2 virus particles from 1% saliva samples and simulated saline gargle samples with a detection limit of 1 fg/μL and 10 fg/μL respectively in 16 min. The method does not require laboratory equipment, sample pretreatment, or complicated manipulations, making it easy to use widely. However, the accuracy was 89% for the analysis of relatively clean clinical saline gargle samples and showed some limitations in accurately analyzing turbid clinical samples. Some modifications such as the substrate functionalizations may enable more accurate detection of SARS-CoV-2 in the future at a low cost [165,166].

Fig. 9.

The paper-based microfluidic chip and flow profile assay for smartphone-based analysis of SARS-CoV-2 virus particles. (a) Paper microfluidic chip with green edges and three red squares for identification of chip regions in automated flow distance. (b) The chip lock and chip holder. (c) A chip holder was used to flatten the paper-based microfluidic chip. (d) 4 μL of the sample was loaded into the square inlet (top) of each channel and dried for 10 min. (e) 4 μL of antibody-conjugated (Ab) particles were loaded onto the chip. (f) A smartphone camera recorded liquid flow on a paper microfluidic chip and analyzed how particle immune agglutination affected velocity distribution and flow distance. (g) In the absence of the virus particle, singlet Ab-particles (green) diffused rapidly to the wetting front, reducing surface tension and flow rate. Nitrocellulose fibers are light orange and salivary proteins are dark orange. (i) In the presence of virus particles (blue), the occurrence of immune agglutination produced clusters of larger and heavier particles, leaving only a few singlet antibody Ab-particles to diffuse to the wetting front. Reprinted with permission from Ref. [164], Copyright 2022, Elsevier.

A field-effect transistor (FET)-based biosensor is gated by changes in the surface potential induced by the binding of the target. The sensing elements are normally immobilized on the sensing channels, which are connected to the source (S) and drain (D) electrodes. The electrical signals can be measured by an electrical signal reading instrument such as a power supply (E3631A, Agilent) [167]. Ban et al. developed a label-free, rapid (≤20 min), DNA aptamer-derivatized graphene field-effect transistor (GFET) to analyze SARS-CoV-2 [168]. Even the unprocessed intact virus and its variants were detected at levels as low as 7 to 10 viruses. A smartphone provided real-time device location identification. This method is effective for both early-stage viral infections and diseases with accessible biofluids and the device provided handheld wireless readout. Chen et al. developed a saliva-based antigen test using the electrical double layer (EDL)-gated field-effect transistor-based biosensor (BioFET) system. Disposable testing sticks were used for sample collections. The biosensor facilitates the changes in EDL capacitance that can be detected by the author-designed BioFET system in real-time using a Bluetooth-embedded reader on an iPhone. The detection limits of SARS-CoV-2 NP are 0.34 ng/mL in 1 × PBS and 0.14 ng/mL in artificial saliva for short time (60 min). Excellent selectivity against MERS-CoV, Influenza A virus, and Influenza B virus was exhibited. Portable electric signal readers have been designed by the authors, but their availability to the public is still unknown.

6. Future prospect

There are some challenges to the ease of operation, cost, sensitivity, and accuracy of the detection of SARS-CoV-2 using smartphone-based tests. The price and complexity of some smartphone-based detection methods still need to be further optimized. Hence, modified approaches are needed to be proposed as a perspective covering the challenge. For the improvement of the optical detections by smartphone-based method, rather than the amplification of the target such as nucleic acid, the amplification of the detection signal without complicated strategies may be an effective strategy to improve the sensitivity and accuracy. For instance, the signal-amplified nanomaterials or polymers might facilitate the fabrication of new nanoplasmonic sensors for more specific and sensitive detection of SARS-CoV-2 antigens; The latest molecular nanotechnology utilizes CATCH to transduce a more sensitive signal output through an enzymatic cascade reaction, bringing new hope for the simultaneous realization of simplicity and sensitivity in nucleic acid detection.

Electrochemical signals still require relatively expensive instrumentation to detect, but optoelectronic signals can be acquired inexpensively. Thus, the combination of electrochemical sensors and optical detections may have the potential to improve sensitivity and availability. In future studies for analysis of SARS-CoV-2 using smartphone-based detections, the combination of different strategies, probes or physical/chemical signal changes may achieve higher sensitivity, accuracy, lower cost, as well as easier operations. Once the system more simply, accurately, and inexpensively recognized the virus, the availability will be broadened.

7. Conclusions

Smartphone-based tests offer promising strategies for diagnosing SARS-CoV-2 more broadly than conventional testing strategies. After an overall comparison of optical and electrochemical analysis, non-nucleic acid, and nucleic acid detections, we summarize relatively inexpensive, sensitive, and accurate methods. So far, optical sensor-based detections are convenient and inexpensive for wide availability. However, the sensitivities and accuracies of these strategies are expected to be further improved. Among the optical detection methods, CRISPR-Cas and LAMP are sensitive and accurate for nucleic acid analysis, but the preprocessing progress is relatively complicated. The electrochemical detection methods tend to be sensitive without pre-extraction or amplification of the targets, but the strategies still require expensive instruments. Some other infrequently reported smartphone-based detections such as flow rate assays and liquid crystals-based optical sensors are still at the primary stages for the analysis of pseudovirus or spiked samples. With various options for the development of nucleic acid and non-nucleic acid tests based on smartphones, the general population may benefit from wide-available detections in the coming years.

Funding

This study was funded by the 2021 Scientific Research Funding Project of Liaoning Provincial Department of Education (No. LJKZ0818) and the “Double First-Class” project of Liaoning Province.

Author's contributions

Dan Li: Conceptualization, Investigation, Writing - original draft. Cai Sun: Scheme drawing. Xifan Mei: Review, Supervision. Liqun Yang: Review, Supervision.

Availability of data and materials

Any data related to this review are available from the corresponding author on reasonable request.

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.