Rapid diagnosis of COVID-19 using disposal paper capacitive sensor

Abstract

Accurate and rapid detection and isolation become indispensable to restrict the spread of COVID-19. Since the start of COVID-19 pandemic in December 2019, many indisposal diagnostic tools are being developed incessantly. Out of all presently used tools, the gold standard rRT- PCR tool having very high sensitivity and specificity is a time consuming complicated molecular technique having requirements of special expensive equipment. Here, the main focus of this work is to develop rapid disposal paper capacitance sensor having simple and easy detection. We discovered a strong interaction between limonin and Spike-glycoprotein of SARS-COV-2 in comparison to its interaction with other similar viruses such as HCOV-OC43, HCOV-NL63, HCOV-HKU1, Influenza B and A viruses. The antibody free capacitive sensor having comb electrode structure was fabricated on whatman paper with drop coating of limonin (extracted using green method from pomelo seeds) and calibrated with known swab samples. The Blind test with unknown swab samples shows high sensitivity of 91.5% and high specificity of 88.37%. Requiring low sample volume and detection time and using biodegradable materials in the sensor fabrication assure the potential application as a point of care disposal diagnostic tool.

Graphical abstract

1. Introduction

SARS-COV-2 is a positive single stranded RNA virus reported first in Wuhan, China in December 2019 [1]. It causes COVID-19 disease transmitted rapidly from human to human with close interactions via respiration droplets of cough and sneezing and surface contact having undamaged viral particles (named as virion) [2]. Till date, the world has been suffering from deadly COVID-19 with more than 64.4 lakhs death and more than 59.2 cores of infection exhibiting a major disruption in global business and economy, education and societal relationships. In general, one of ways to break the chain of infection of COVID-19 is detection and then isolation and quarantine [3]. Fast and accurate diagnosis is crucial to deal with the problem of high speed spread of COVID-19 infection. There are many ways being developed for reliable and accurate test of SARS-COV-2 infection such as real-time Reverse transcription-polymerase chain reaction (rRT-PCR) [2] Rapid Antigen Test (RAT) [4], Antibody Test [5], CT image screening [6,7], ACE₂ based capacitive sensor [8], FET based biosensor [9] etc. Among the above diagnosis techniques, viral gene detection through rRT-PCR and human antibody detection though Enzyme linked immunosorbent assays(ELISA) are used medical practices approved by Medical councils such as Food and Drug Administration (FDA) [10], National Health Service (NHS) [11], Indian Council of Medical Research (ICMR) [12] etc. Although rRT-PCR is highly specific to the detection of SARS-COV-2, it requires multiple steps along with the several costly reagents, enzyme and special machines to amplify the gene sequence and measure through florescent light [13]. Despite high accuracy and sensitivity, the mutation capability of COVID -19 virus produces new variants with new gene sequence making it more challenging for identifying new variants infection with rRT-PCR kit. Moreover, time consuming diagnosis of rRT-PCR makes delay in isolation [14].

The Antibody IgM and IgD are generated by immune system in response to COVID-19 antigen in blood and detected through enzyme linked immunosorbent type assays (ELISA) [5] but antibody test is not able to detect initial infection of COVID-19 due to more time required to develop antibody in human body after infection. The rapid antigen test based on lateral flow assays performs rapid detection of COVID-19 [15] but it exhibits false positive with other pneumonia viruses such as HCOV OC-43, HCOV NL-63, HCOV-HKU1,HCOV-229E, Influenza A and Influenza B [[16], [17], [18]].

Actually COVID-19 virus taxonomically belonging to β linkage of Beta corona virus genes has complete 30 kb genome consisting 2/3 encoded with polyprotein and 1/3 of genes encoded with surface (spike) glycoprotein (S), envelop protein(E), membrane protein(M) and nucleocapsid(N) proteins [19]. Out of the proteins of COVID-19, the spike protein made of S-Glycoprotein has binding capability with human angiotensin converting enzyme 2 (ACE₂) [[20], [21]] opening the development of biosensor using ACE₂ for efficient and rapid virus detection. In this direction, interdigital electrodes play a pivotal role by using as a capacitance or impedance biosensor for chemical and biochemical sensing [22,23]. The IDE biosensors have already been used for detection SARS-COV-2 antibodies [23,24], but the use of ACE2 and graphene in biosensor make the device expensive and not disposal. In this direction, phytochemicals extracted through green synthesis from citrus fruits are pivotal to make the device disposal due to having chemical interaction between S-glycoprotein and photochemicasn obtained by using molecular docking between S-protein of SARS-COV-2 with plant extract phytochemicals [25]. Their potential use of phytochemicals such as hesperidin, emodin, Chrysin, Luteolin, Genistein are already established in drug design [26]. Main challenge for the use of the materials in the diagnosis of the disease is selectivity.

Here, we have presented limonin extracted from pomelo seeds as a sensing material for detection of COVID-19 (Fig. 1 a). Molecular docking was computed to find the interaction between limonin and S-protein of SARS-COV-2 in comparison to its interaction with other viruses (supplementary Table-1S). The comb structure based electrode was formed on whatman filter paper and finally limonin solution was deposited on electrode surface to fabricate paper capacitance sensor (Fig-1b, c). The Paper capacitor was calibrated with known swab samples of COVID-19 +ve, COVID-19-ve, influenza A and Influenza B with respect to virus transfer medium using experimental set up (Fig. 1d, e and f).

Fig-1.

Synthesis of limonin, fabrication of paper capacitance with limonin as sensing layer and its measurement set up: a. Fabrication steps of extraction of limonin from pomelo seeds b. Fabrication steps for paper capacitor sensor: Comb structure on 3 x2 (cm)² whatman filter, drop coating of 4 μM Limonin –PVA solution on comb structure of electrode, drying at 40 °C for 2–3 h c. 3D structure of paper capacitive sensor d. Measurement set up of capacitance with Ardino e. formation of interior and edge capacitance with applying voltage from Ardino uno board f. equivalent circuit having total capacitance C_total and internal resistance (R_I) of Ardino uno board.

2. Sensing principle of device

The paper capacitance sensing device contains coupled comb electrode structure with the fingers of width (w) formed on 3 × 2 cm² flexible whatman filter paper capacitor (Supplementary Fig. S1) and the gap (a) between the finger electrode is filled with sensing material (Supplementary Fig-S2a). Considering eight fingers in comb electrode structure, two edge electrode capacitance and five interior capacitance are formed between finger electrodes (Supplementary Fig. S2b). After applying + V and -V potential at adjacent fingers of electrode, there are generation of traversing electric fields penetrating through sensing material and paper substrate and fringing fields travelling from one finger to other finger through air medium(supplementary Fig-S2c). The total capacitance contributed by the above fields are the function of the dielectric constant of sensing material and thickness (w) of the electrode and gap (a) between two adjacent fingers. The designed values of w and a are obtained as ∼1.5 mm (from Supplementary Fig. S2e) and ∼1.5 mm (from Supplementary Fig. S2f) respectively.

3. Result and discussion

The extraction process of limonin from pomelo seed (collected from Tezpur, Assam India) was made mainly by using Sexhlet instrument and maintaining pH value of 7 during the process (Fig. 1a). For a proof of extraction of limonin, the FTIR spectra of the sample was carried out. Fig. 2 a shows broad absorption peak at 3450 cm⁻¹ indicating OH⁻ stretching vibration group whereas the weak absorption peaks at 2920 cm⁻¹ and 1050 cm⁻¹ indicates the occurrence of β-furan ring in the sample [27] and the peak at 1395 cm⁻¹ reveals the presence of methyl groups in the sample [28]. Further, the absorption peak at 1650 cm⁻¹ attributes to the C=C radicals²⁸. The results of FTIR absorption peaks confirm the consistency with chemical structure of limonin because of the presence of the above chemical groups in the sample [27,28].

Fig-2.

Characterization of extracted limonin sample and fabricated sensor: a. FTIR spectra of extracted limonin b. Standard HPLC obtained using standard solution (made with 99% pure limonin procured from Sigma) c. HPLC obtained using extracted limonin solution using reverse phase C18 column at 207 nm, injection volume = 20 μL, mobile phase of tetrahydrofuran, acetonitrile and DI water(17.5:17.5:65), flow rate 0.5 ml/min d. EDX spectra of paper substrate before deposition of limonin e. EDX spectra of substrate of the sensor after deposition of limonin f. SEM image of paper substrate before deposition of limonin (g) SEM image of deposited area showing limonin fiber distributed over PVA.

For further analysis of purity of limonin extraction, High performance liquid chromatography (HPLC) was conducted [29]. Standard limonin solution was made as a reference for purity analysis of extracted limonin sample (supplementary section S2). HPLC of both standard limonin solution and extracted limonin were carried out by HPLC set up. Fig. 2b and c shows the HPLC analysis results of standard and extracted limonin sample respectively indicating major peak at same retention time of 5.6 min. The identical retention time of the peak of both solutions confirm limonin extraction. The purity of extracted limonin sample is estimated from HPLC curves as ∼98.6% (Supplementary section S2) which is more than previous work [30].

For the evidence of fabrication of capacitive sensor, the Energy dispersive X-ray (EDX) spectral analysis was performed. The EDX spectra of paper substrate having Ag ink printed comb finger in Fig. 2d shows higher weight percentage of Ag in comparison to other elements C, O an N confirming the fabrication of Ag coated finger on paper surface due to having no limonin. The EDX spectra of paper capacitive sensor having deposition of limonin (Fig. 2e) represent the increase of weight percentage of C, O and N (35.36% of C, 28.48 of N, 35.18% of O, 0.24% of Ag and 0.74% of Na) in comparison to Ag due to deposition of limonin in the sensing region of paper substrate. The presence of higher amount of C, N and O in comparison to negligible amount of Ag confirm the disposal of the paper capacitive sensor in soil after its use [31]. The Scanning Electron microscope (SEM) image of the paper substrate in Fig. 2f reveals the presence of the porosities on the surface making the absorption of limonin solution (Fig-2f). The SEM image (Supplementary Fig. S3) shows the finger thickness a ∼1500 μm and gap between finger W ∼ 1500 μm. The image of the sensing region shows the formation of limonin fiber of average diameter ∼490 nm distributed over PVA surface matching well with the previous work [32].

The capacitance of paper capacitor was measured by experimental setup in Fig. 1e with charging and discharging through RC circuit having capacitance C_total and internal resistance (R_I) of Aridino Uno (Fig. 1f). Before the exposure of swab samples the interior capacitance and edge capacitance are C_I and C_E respectively (Fig-3a ). During exposure of swab sample infected with SARS-COV-2 virus having spike glycoprotein (S-protein) over limonin coated surface of sensor, it forms limonin-S-protein complex via the chemical bonding between limonin and S-protein (Supplementary Table 1S). As result, after the exposure of the swab samples, the values of interior and edge capacitance become C_I ^ʹ and C_E ^ʹ respectively (Fig. 3b). The autodocking computation also confirms the formation of S-protein limonin complex via hydrogen bonding of limonin with TRPC: 423 amino acid residue of S-protein and p-alkyl bonding of limonin with ILEC-428 amino acid residue and bump with HOHL: C:31 amino acid residue (Supplementary Fig. 4s). For experimental verification, we have performed FTIR spectroscopy after interacting S-glycoprotein sample with limonin indicating broad peaks of OH⁻¹ in comparison to OH⁻¹ peak at 3450 cm⁻¹ obtained with limonin sample showing evidence of hydrogen bond (Fig. 3c). The peak at 1750 cm⁻¹ confirms presence of NH groups showing interaction between limonin and S-glycoprotein [33].

Fig-3.

Interaction of S-glycoprotein of SAR-COV-2 with limonin: (a) S-glycoprotein interacting with limonin deposited on comb electrode surface (b) Formation of S-glycoprotein-limonin complex after interaction (c) FTIR of S-glycoproteinand limonin after interaction.

For confirmation of chemical interaction between limonin sensing layer and spike protein of SARS-COV-2, we have made three standard solution by mixing S-glycoprotein concentration 2, 5 and 9 ng/μL in de-ionized water having pH value of 7.2 solution. Fig. 4 a shows the transient analysis of capacitance ($C_{total}$) of paper capacitive sensor exposed with standard solution of 2, 5 and 9 ng/μL. The $C_{total}$ increase with time and gets almost saturated at the time of 183 s. It is seen from Fig. 4a that saturated values of capacitance ($C_{total}^S$) increases with increase of S-glycoprotein concentration indicating increase of chemical interactions with increase concentration and the values of $C_{total}^S$ is more than that for COVID -ve swab samples and almost close to that for COVID + ve swab samples confirming chemical interaction with S-glycoprotein of Spike protein of SARS-COV-2. Further, the calibration of paper capacitive sensor was carried out by using known thirty-one COVID-19 +ve, known twenty seven COVID-19 -ve samples, 15 Influenza A and 15 Influenza B swab samples. Fig. 4b shows the transient analysis of capacitance ($C_{total}$) of paper capacitive sensor exposed to viral transfer medium (VTM), swab samples of SARS-COV-2 of viral load of CT = 16 and viral load of CT = 32, (where CT is cyclic threshold value representing viral load [13], Influenza A and Influenza B. The $C_{total}$ also gets almost saturated at same time of 182 s. The saturated value of sensor capacitance or SARS-COV-2 of CT value 16 and CT value 32 are obtained as ∼370 μF and 320 μF respectively at detection time of 183 s. The $C_{total}^S$ value for other COVID-19 +ve samples is in the range of 303 μF–414 μF (Supplementary Table-2S). The COVID-19 negative sample shows $C_{total}^S$ ∼234 μF almost at same detection time of 182s (Supplementary Table 3S) The $C_{total}^S$ for influenza A (id 2512) and Influenza B(id 2636) swab samples are 211 μF and 239 μF respectively almost at same detection time. We have observed $C_{total}^S$ in the range of 175 μF - 211 μF for other Influenza A samples (Supplementary Table 4S) and 148 μF - 239 μF for Influenza B swab samples (Supplementary Table 5S). The lower $C_{total}^S$ values of influenza A and influenza B samples in comparison to that of COVID-19 samples indicating high selectivity. Further, Fig. 4c shows capacitance transient analysis of five COVID-19 +ve samples of different CT values ranging from 17 to 32 and one COVID-19 -ve sample with reference to VTM sample for further verification of detection time. Almost all COVID-19 positive samples show the saturation of capacitance almost at same time of 182 s (approximately). But the saturated value of capacitance of the sensor with VTM sample is observed to be at 250 s. This is due to having very less interaction of VTM with limonin (Supplementary Table 6S). From the analysis, considering VTM and COVID-19 –Ve samples are considered as a control samples and $C_{total}^S$ >300 μF indicates COVID-19 positive and $C_{total}^S$ <300 μF shows COVID-19 negative samples

Fig-4.

Transient analysis of sensor (a) time variation of capacitance after exposure of limonin surface of sensor with standard S-protein solution of concentration of 2,5 and 9 ng/μL (b)time variation of capacitance after exposure of limonin surface of sensor with Influenza A infected swab sample, Influenza B infected swab sample Detection time ∼ 175ted swab sample with high CT value of 32 and low CT value = 16 and virus transfer medium (VTM) sample (c) Time variation of capacitance after exposure of SARS- COV 2 infected swab sample having CT values of 16, 18, 21, 24,28,32 (d) variation of $C_{total}^S$ with SARS –COV 2 virus load of swab samples (viral load is represented in terms of CT value).

Further, we have tested our paper capacitive sensor by 90 unknown COVID-19 samples through blind test (Supplementary Table 7S and .8S). The $C_{total}^S$ values of COVID-19 samples depend on CT value representing viral load. The $C_{total}^S$ value tends to decrease with increase of CT value due to low viral load at higher CT values and it almost remain constant for CT > 22 (Fig. 4d). At low viral load, there is less number of limonin- S-protien interaction leading to formation of less number of -limonin-S-protien complex at the sensing surface of paper capacitive sensor. The $C_{total}^S$ values for all the samples are estimated by taking standard deviation from five repeated measurement of capacitance transient curves (example calculation for COVID positive, negartive, Influenza A and B samples in Supplementary Table 9S).

We have also determined true positive, true negative, false positive and false negative swab samples with respect to rRT-PCR testing of samples (Supplementary Table-7S and Table-8S). The sensitivity, specificity, positive and negative predictive values are estimated [34] as a/(a+c)] × 100 = 91.5%, [d/(b + d)] × 100 = 88.37%,[a/(a+b)] × 100 = 89.58% and [d/(c + d)] × 100 = 90.5% respectively (Supplementary Fig. 4S, section S7) where a = number of True Positive COVID sample = 43, d = Number of True Negative COVID sample = 38, b = Number of False Positive = 5 and c = Number of True Positive COVID sample = 4.

The sample volume required for testing of swab samples using paper capacitor is obtained as ∼10 μL which is lowest among other COVID-19 kits (Supplementary Table 10S). In comparison to rRT-PCR COVID kit (used as a gold standard [2,35]), our paper capacitance sensor takes maximum 3 min to test. Although the detection time of RAT kit is slightly lower than that of paper capacitor kit, the sensitivity is lower than that of our paper capacitive sensor kit [4]. The ACE₂ COVID-19 capacitor shows high specificity and sensitivity but the detection time is higher than that of paper capacitor kit [8]. The Antibody based test kit having higher sensitivity and specificity fails to test initial stage of COVID-19 infection. Moreover, both RAT and antibody kit may exhibit false COVID-19 +ve due to formation of similar type of antibodies (IgG, IgM and IgA) after infection of HCOV OC-43, HCOV NL63, HCOV HKU1, Influenza B and Influenza A viruses [18,36,37].

4. Conclusion

Here, an antibody-free and highly disposal capacitive sensor was developed for rapid diagnosis of COVID-19. The limonin was chosen as a sensing material extracted from pomelo citrus seeds using green synthesis due to having its string interaction with S-protien of SARS-COV-2 in comparison to other similar pneumonia viruses. The capacitive sensor based on comb electrode structure was made on degradable whatman paper with drop coating of limonin and calibrated with known swab samples of COVID-19 +ve, COVID-19-ve, influenza A and Influenza B with respect to VTM. Owing to high sensitivity and high specificity, the reported sensor opens an avenue to realize rapid and disposal diagnostic tool for COVID-19.

5. Material and methods

5.1. Materials

Poly Vinyl Alcohol (PVA) was procured from Alfa Aesar whereas Standard Limonin and 3:1 dichloromethane and S-glycoprotein were procured from Sigma –Aldrich. All were of analytical grade and were used without further treatments. The isopropanol, NaOH and whatman filter paper grade 1 were collected from M/S Merck India. Pomelo fruits are purchased from local garden of Tezpur, Assam. Viral transfer medium (VTM) and swab samples was supplied from Regional Medical Research Centre, ICMR, N.E. Region, Dibrugarh, Assam, India. The conducting silver ink pen was procured from M/S Sparkfun.

5.2. Selection of limonin

Using Autodock software, the interaction between limonin and spike protein of virus HCOV OC-43(60 HW), HCOV NL63(2IEQ), HCOV HKU1(3D23), SARS-COV2(2GHV), Influenza B(2RFU) and Influenza A (6SPF) are computed through molecular docking(Supplementary Table 1S) showing strong interaction between spike S protein (2GHG) and limonin with binding affinity −9.0 kcal/mol. The strong interaction of spike S protein (2GHG) and limonin leads selective interaction of limonin with S-glycoprotein of COVID-19.

5.3. Limonin extraction from Pomelo citrus fruit and characterization

Limonin was extracted from seeds of pomelo citrus fruits by maintaining pH value of 7–7.3 (Fig. 1a). The seeds were collected from pomelo fruits, dried at sunlight for 2–3 days and grinded by grinder machine. Initially 10 g of grinding powder was defatted using Soxhlet apparatus with hexane. Then, 80 ml acetone was mixed with the powder and stirred (400 rpm) at 80 °C for 4 h. The filtrate was extracted with Whatman filter paper washed with ethanol. The purification of filtrate sample was made by using isopropanol and NaOH solution for dissolving unwanted residue. Further the crystallization of sample was carried out by dissolving in 3:1 dichloromethane and isopropanol solution and kept for 2 h at 4 °C.Finally the extracted sample was dried in oven at 60 °C for 4 h to obtain the extracted limonin powder sample.

FTIRs of powder sample and standard limonin sample were carried out by using Fourier Transform Infrared Spectroscopy (IMPACT 410, NICOLET, USA). The High Performance Liquid Chromatography (HPLC) of extracted limonin sample and standard limonin sample were performed by HPLC system (WATERS CORPORATION, USA model 2489).

6. Sensor design

Considering sensor structure (Supplementary Fig-S1), the total capacitance contributed by the above fields are the function of the dielectric constant of sensing material and thickness of the sensing layer and gap between two adjacent finger and derived as

C_Total = L(N-3)C_I+2C_E

where C_I is the initial capacitance of unit cell and C_E is the edge capacitance (function of thickness,(t) of sensing layer) which is neglected with respect to C_I as C_I ≫ C_E (Supplementary S3). For the design we have taken 4 fingers in each electrode of comb structure (N = 8). Considering the variation of C_Total with the dielectric with dielectric constant (ϵ_m) of sensing material, w is obtained as 1.5 mm (Supplementary Fig. S2e) and a is found to be 1.5 mm (Supplementary Fig. S2f).

7. Fabrication of sensor and its characterization

The comb structure having 4 fingers in each electrode was drawn on 3 × 2 cm² whatman paper (Fig. 1b) by the silver conducting pen. The limonin was mixed with PVA using ultra-sonication for 20–30 min to form the gel form of the sample. Then the sample was deposited in between the finger electrode structure with drop coating technique (Fig. 1b).The fabricated device in Fig. 1c was kept in an oven at 40 °C temperature for 2 h. The EDX spectra and SEM of whatman filter paper and sensing region of the sensor was carried out by Scanning Electron Microscope (SEM) Make: JEOL, JAPAN Model: JSM 6390LV for compositional and morphological analysis of the fabricated device.

8. Experimental procedure

The connecting pads of 2 × 2 (mm)² made on Whatman paper was connected to analog pin and ground of Arduino Uno for measurement (Fig. 1d). The capacitance changes after exposing to swab samples collected from patients in Regional VRDL, Regional Medical Research Centre, ICMR, N.E. Region, Dibrugarh, Assam, India. We have calibrated experimental setup with more than known 90 samples of Corona infected patient, influenza A, influenza B and viral transport media.

The comb structure paper capacitor provide two edge capacitor C_E between the end finger electrodes at two ends of comb structure and five interior capacitors of capacitance C_I (Fig. 1e). The equivalent circuit of equivalent capacitance C_Total and internal resistance (R_I) of Arduino Uno provides changing with application voltage V connected in series (Fig. 1f) The C_Total is obtained as 5LC_I+2C_E, where L = overlapping length of electrodes. The transient analysis of C_Total was performed after exposure of the swab samples to the paper capacitive sensor by using Arduino Uno through charging of paper capacitor.

Data material availability

All data, code and materials used in the analysis are available to the authors for the purposes of reproducing or extending the analysis.

CRediT authorship contribution statement

Partha P. Sahu: Conceptualization, Supervision, Formal analysis, Writing – original draft. Geetartha Sarma: Conceptualization, Methodology, Validation, Formal analysis, Writing – review & editing. Satyajit Das: Methodology, Investigation, Writing – review & editing. Biswajyoti Borkakoty: Supervision, Investigation, Formal analysis.

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.

Handling Editor: Professor Chuck Henry

Appendix A. Supplementary data

The following is the Supplementary data to this article.

Data availability

Data will be made available on request.