Smartphone-Based NFC Potentiostat for Wireless Electrochemical Sensing

: Most electrochemical sensing requires affordable, portable and easy-to-use electrochemical devices for use in point-of-care testing and resource-limited settings. This work presents the design and evaluates the analytical performance of a near-ﬁeld communication (NFC) potentiostat, a ﬂat card-sized electrochemical device containing a microchip for electrical analysis and an NFC antenna for smartphone connection. The NFC interface is a wireless connection between the microchip and smartphone to simplify measuring units and make the potentiostat into a passive operated device, running without a battery. The proposed potentiostat can perform the common electrochemical techniques including cyclic voltammetry and chronoamperometry with a current range and voltage range of ± 20 µ A and ± 0.8 V. The performance of the NFC potentiostat is compared to a commercial benchtop potentiostat using ferricyanide as a standard solution. The results show that the NFC potentiostat is comparable to a commercial benchtop potentiostat for both cyclic voltammetry and chronoamperometry measurements. The application of the proposed potentiostat is demonstrated by measuring ascorbic acid concentration. As described, the NFC potentiostat, which is compatible with a smartphone, is low-cost, small in size and user-friendly. Thus, the device can be developed for on-site measurement to apply in various ﬁelds.


Introduction
Electrochemical sensing has been developed and is widely applied in various fields including food safety and control [1][2][3], environmental monitoring [4][5][6] and clinical diagnostics [7][8][9], since it offers many advantages such as high sensitivity and selectivity, rapid detection and small sample volume [10][11][12]. Typically, a potentiostat is an electronic instrument used to perform electroanalytical measurements by controlling the voltage difference and measuring the current flow through an electrochemical cell [13,14]. However, the commercially available potentiostats that are suitable for the laboratory are large and expensive, with a price of several thousand dollars [15][16][17], which largely limits their application for point-of-care testing and in resource-limited settings. For these reasons, an electrochemical workstation needs to be portable, low-cost and easy to use to meet the demand for on-site measurement.
In recent years, there has been growing research in the development of low-cost and portable potentiostats. Some studies have developed inexpensive and open-source potentiostats that complement commercial laboratory potentiostats [18][19][20][21], but the instrument requires a computer to operate the detection process, analyze the data and display the results. These devices limit the application for home-based testing and in resource-limited settings. Moreover, those devices cannot easily transmit the data to the cloud for storage

Chemicals and Materials
All chemicals used in this work were analytical grade and deionized water was used throughout the experiments. Potassium ferricyanide (K 3 [Fe(CN) 6 ]), potassium chloride (KCl) and ascorbic acid were purchased from Sigma Aldrich (St. Louis, MO, USA). All electrochemical measurements including cyclic voltammetry and chronoamperometry were performed on an NFC potentiostat developed by Silicon Craft Technology PLC (Bangkok, Thailand) and a Galaxy Nexus smartphone (Samsung, Seoul, South Korea) with Android operating system. A CH instrument (CHI 611E, Texas Instruments Inc., Austin, TX, USA) was used as a standard instrument. A commercial screen-printed carbon electrode (TE100, Zensor R&D Co., Ltd., Taichung City, Taiwan), consisting of a 3 mm diameter carbon working electrode, carbon counter electrode and Ag/AgCl reference electrode, was used to perform the experiment.

NFC Potentiostat Design
The NFC potentiostat is designed in the form of a compact credit card having dimensions of 5.5 cm × 8.6 cm, consisting of a planar antenna, NFC microchip (SIC4341) and connector for electrode interface. The dimension of the NFC potentiostat is designed to ensure stable operation during tapping the mobile, while leaving some space for electrode insertion and solution drop. The NFC microchip harvests energy from the magnetic field emitted from a smartphone or RFID reader through a loop antenna realized from 1 oz. (35 µm) copper trace on FR4-PCB. The printed circuit board was processed by PCBway (Shenzhen, China). The antenna consists of a 4-turn loop with a size of 4.5-by-3.0 cm maximum perimeter and 250 µm conductor width and spacing. The antenna has an inductance of 2.04 µH with a quality factor of 130, measured by LA-19-13-02 VNA from LA technique Ltd. (Surrey, UK) and couples to a microchip having a 50 pF internal capacitance. This creates a parallel resonance circuit at a peak resonance frequency of 15.7 MHz, based on the f r = 1/2π √ LC relationship. The resonance frequency of the antenna is set a bit higher than the operating frequency of 13.56 MHz in order to compensate coupling effects from near-field operation. The maximum operating read distance from the antenna is 2 cm, over which the potentiostat can still operate stably. The energy is regulated to the internal 1.8 V power supply for the internal potentiostat to cover a maximum operation range of ±20 µA, an analog-to-digital data converter, a digital signal processor to convert raw data into measurement current information and an NFC analog front-end communication circuit to communicate with the smartphone or RFID reader.

Android Application
A user-friendly SIC4341-POTEN Android mobile application was developed to operate the NFC potentiostat. The main screen shows the detection method as depicted in Figure 1a. After selecting the electrochemical method, the user can set the parameters of the experiment as shown in Figure 1b. Next, the software allows the user to drop the substrate. Then, the screen automatically displays the voltammogram as presented in Figure 1c. After finishing the process, users press the "next" icon. Data were obtained as illustrated in Figure 1d. Three options are then available, namely: back to the main screen, perform the electroanalytical experiment again and export the data. The resulting data are stored as ".txt" files and transferred to a computer for the interpretation steps by ES File Explorer. The resulting data enable upload through the cloud (e-mail, social media or Dropbox) for sharing storage or remote analysis of the data. Furthermore, the application can be adapted to suit different analyte detection by reporting the concentration of the sample calculated from the embedded standard calibration in the system. Appl. Sci. 2021, 11, x FOR PEER REVIEW 4 of 13

Performance Test of the NFC Potentiostat
To evaluate the performance of the NFC potentiostat, a commercial screen-printed carbon electrode consisting of a 3 mm diameter carbon working electrode, carbon counter electrode and an Ag/AgCl reference electrode was used to perform cyclic voltammetry and chronoamperometry. The experiments on the NFC potentiostat were compared to a commercial laboratory potentiostat (CHI 611E, Texas Instruments Inc., Austin, TX, USA) used as a standard instrument. For each experiment, 100 µL of potassium ferricyanide in 0.1 M KCl as a supporting electrolyte was applied on a screen-printed carbon electrode. The solution of potassium ferricyanide is a common analyte to test the performance of a potentiostat, because it exhibits a reversible electrochemical behavior and rapid one-electron process [45,46]. Cyclic voltammetry was carried out over the potential range from −300 to 600 mV at a scan rate of 5-200 mV/s. For chronoamperometric detection, the fixed potential of −100 mV was applied to the working electrode and the current period of 0-30 s was recorded.

Ascorbic Acid Measurement
To demonstrate the application of the NFC potentiostat, the detection limit of standard ascorbic acid was determined by cyclic voltammetry. In this experiment, different concentrations of ascorbic acid, ranging from 0.1-1 mg/mL, were drop-cast onto a commercial screen-printed carbon electrode. Cyclic voltammetry was performed over the potential range from 200 to 800 mV, and the current at 550 mV was used to determine the ascorbic acid concentration. Commercial drinking water was spiked with ascorbic acid into different concentrations. The concentration of ascorbic acid in spiked commercial drinking water was analyzed from the calibration curve.

NFC Potentiostat Operation
The main components of an NFC potentiostat are shown in Figure 2a, containing a loop antenna, NFC microchip (SIC4341) and connector for electrode interface designed on a printed circuit board with a credit card size. The details of the integrated circuit shown in Figure 2b include (1) an analog part to connect between radio signals and controller, (2) digital controller and memory unit to control the operation of integrated circuits to respond to data transmission that operates at 13.56 MHz in the high frequency (HF) radio

Performance Test of the NFC Potentiostat
To evaluate the performance of the NFC potentiostat, a commercial screen-printed carbon electrode consisting of a 3 mm diameter carbon working electrode, carbon counter electrode and an Ag/AgCl reference electrode was used to perform cyclic voltammetry and chronoamperometry. The experiments on the NFC potentiostat were compared to a commercial laboratory potentiostat (CHI 611E, Texas Instruments Inc., Austin, TX, USA) used as a standard instrument. For each experiment, 100 µL of potassium ferricyanide in 0.1 M KCl as a supporting electrolyte was applied on a screen-printed carbon electrode. The solution of potassium ferricyanide is a common analyte to test the performance of a potentiostat, because it exhibits a reversible electrochemical behavior and rapid oneelectron process [45,46]. Cyclic voltammetry was carried out over the potential range from −300 to 600 mV at a scan rate of 5-200 mV/s. For chronoamperometric detection, the fixed potential of −100 mV was applied to the working electrode and the current period of 0-30 s was recorded.

Ascorbic Acid Measurement
To demonstrate the application of the NFC potentiostat, the detection limit of standard ascorbic acid was determined by cyclic voltammetry. In this experiment, different concentrations of ascorbic acid, ranging from 0.1-1 mg/mL, were drop-cast onto a commercial screen-printed carbon electrode. Cyclic voltammetry was performed over the potential range from 200 to 800 mV, and the current at 550 mV was used to determine the ascorbic acid concentration. Commercial drinking water was spiked with ascorbic acid into different concentrations. The concentration of ascorbic acid in spiked commercial drinking water was analyzed from the calibration curve.

NFC Potentiostat Operation
The main components of an NFC potentiostat are shown in Figure 2a, containing a loop antenna, NFC microchip (SIC4341) and connector for electrode interface designed on a printed circuit board with a credit card size. The details of the integrated circuit shown in Figure 2b include (1) an analog part to connect between radio signals and controller, (2) digital controller and memory unit to control the operation of integrated circuits to respond to data transmission that operates at 13.56 MHz in the high frequency (HF) radio band and is compatible with ISO14443A, and (3) the chemical sensor processing to measure signal responses from sensors. DAC voltage. The amount of current from the sensor in the physical world is processed by the 10-bit analog-to-digital converter with a digital signal processor, and finally converted into digital data stored in internal memory, ready to be transmitted back to the NFC smartphone. The application on the NFC smartphone can periodically transmit commands to set the DACs, following biasing profiles such as cyclic voltammetry. The NFC smartphone can read the conversion result after each period to reconstruct a voltammetry waveform in the application graphic interface. As the timing can be set by the application, for example, for cyclic voltammetry, the sweep rate can be set on the application from 5 mV/s up to 1000 mV/s. The application provides a graphic user interface (GUI) to set the three electrode terminals freely and flexibly to support various electrode arrangements.

Cyclic Voltammetry
NFC potentiostat performance was compared to a commercial potentiostat (CHI 611E, Texas Instruments Inc., Austin, TX, USA), which is a benchtop potentiostat designed to enable various electrochemical techniques to be performed in the laboratory. The cyclic voltammograms of 0.1, 0.5 and 1 mM potassium ferricyanide in 0.1 M KCl at a scan rate of 100 mV/s performed on the NFC potentiostat and the commercial potentiostat are shown in Figure 3. The cyclic voltammograms showed a characteristic duck shape for a fully reversible one-electron reaction of ferricyanide. The results showed good agreement The operational steps of the NFC potentiostat are shown in the Supplementary Information Video S1. When an NFC smartphone approaches the NFC antenna, the NFC microchip is woken up by the radio frequency (RF) power from the induced magnetic field. If the incoming power is sufficiently high, indicated by its internal status, the NFC microchip starts initializing the system, activates the microchip peripheral devices and waits for a command from the NFC smartphone for further operations. The NFC microchip (SIC4341) consists of an internal potentiostat that is capable of supporting current measurement from chemical sensors at a maximum input current of ±20 µA. The biasing voltage between working electrode and reference electrode can be between −800 mV and 800 mV, generated from the differential voltage of dual digital-to-analog converters (DACs) having full scale voltage of 1.28 V with a step of 5 mV. The NFC smartphone application controls the biasing voltage across the sensor by transmitting a command to adjust the DAC voltage. The amount of current from the sensor in the physical world is processed by the 10-bit analog-to-digital converter with a digital signal processor, and finally converted into digital data stored in internal memory, ready to be transmitted back to the NFC smartphone. The application on the NFC smartphone can periodically transmit commands to set the DACs, following biasing profiles such as cyclic voltammetry. The NFC smartphone can read the conversion result after each period to reconstruct a voltammetry waveform in the application graphic interface. As the timing can be set by the application, for example, for cyclic voltammetry, the sweep rate can be set on the application from 5 mV/s up to 1000 mV/s. The application provides a graphic user interface (GUI) to set the three electrode terminals freely and flexibly to support various electrode arrangements.

Cyclic Voltammetry
NFC potentiostat performance was compared to a commercial potentiostat (CHI 611E, Texas Instruments Inc., Austin, TX, USA), which is a benchtop potentiostat designed to enable various electrochemical techniques to be performed in the laboratory. The cyclic voltammograms of 0.1, 0.5 and 1 mM potassium ferricyanide in 0.1 M KCl at a scan rate of 100 mV/s performed on the NFC potentiostat and the commercial potentiostat are shown in Figure 3. The cyclic voltammograms showed a characteristic duck shape for a fully reversible one-electron reaction of ferricyanide. The results showed good agreement of cyclic voltammograms between the two devices. The higher the concentration of ferricyanide, the greater the current signals observed. The average and relative standard deviation (RSD) values of cathodic peak current and anodic peak current obtained from both potentiostats are listed in Table 1. As shown in the table, the relative standard deviation was less than ±5% (n = 3), indicating that both potentiostats had a good reproducibility for conducting the experiments.
Appl. Sci. 2021, 11, x FOR PEER REVIEW 6 of 13 of cyclic voltammograms between the two devices. The higher the concentration of ferricyanide, the greater the current signals observed. The average and relative standard deviation (RSD) values of cathodic peak current and anodic peak current obtained from both potentiostats are listed in Table 1. As shown in the table, the relative standard deviation was less than ±5% (n = 3), indicating that both potentiostats had a good reproducibility for conducting the experiments.  To evaluate the performance of the potentiostats in cyclic voltammetry at various scan rates, cyclic voltammograms were recorded for 1 mM potassium ferricyanide containing 0.1 M KCl by sweeping the potential from −300 to 600 mV at various scan rates of 5-200 mV/s. Cyclic voltammograms produced on the NFC potentiostat are shown in Figure 4a. When the potential is applied at higher scan rates, higher peak currents are observed because of the decrease in the diffusion layer [47]. Figure 4b shows the theoretical linear relationship between the redox peak current of ferricyanide and the square root of the scan rate according to the Randles−Sevcik equation (Equation (1)), indicating that the reaction was a diffusion-controlled process [47,48]. Furthermore, the calibration plots obtained from the NFC potentiostat and the commercial potentiostat showed parallel results. A plot from the NFC potentiostat provided a linear slope for cathodic peak current of 1.242 A/(V/s) 1/2 (R 2 = 0.995) and an anodic peak current of −1.222 A/(V/s) 1/2 (R 2 = 0.996). In comparison, the linear slopes for cathodic and anodic peak currents recorded from the commercial potentiostat were 1.237 A/(V/s) 1/2 (R 2 = 0.999) and −1.242 A/(V/s) 1/2 (R 2 = 0.999), respectively.
where ip is the peak current (A), n is the number of electrons involved in the redox reaction, F is the Faraday's constant (96,485 C/mol), A is the surface area of the working elec-  To evaluate the performance of the potentiostats in cyclic voltammetry at various scan rates, cyclic voltammograms were recorded for 1 mM potassium ferricyanide containing 0.1 M KCl by sweeping the potential from −300 to 600 mV at various scan rates of 5-200 mV/s. Cyclic voltammograms produced on the NFC potentiostat are shown in Figure 4a. When the potential is applied at higher scan rates, higher peak currents are observed because of the decrease in the diffusion layer [47]. Figure 4b shows the theoretical linear relationship between the redox peak current of ferricyanide and the square root of the scan rate according to the Randles−Sevcik equation (Equation (1)), indicating that the reaction was a diffusion-controlled process [47,48]. Furthermore, the calibration plots obtained from the NFC potentiostat and the commercial potentiostat showed parallel results. A plot from the NFC potentiostat provided a linear slope for cathodic peak current of 1.242 A/(V/s) 1/2 (R 2 = 0.995) and an anodic peak current of −1.222 A/(V/s) 1/2 (R 2 = 0.996). In comparison, the linear slopes for cathodic and anodic peak currents recorded from the commercial potentiostat were 1.237 A/(V/s) 1/2 (R 2 = 0.999) and −1.242 A/(V/s) 1/2 (R 2 = 0.999), respectively. ip = 0.4463nFAC 0 nFD RT where ip is the peak current (A), n is the number of electrons involved in the redox reaction, F is the Faraday's constant (96,485 C/mol), A is the surface area of the working electrode (cm 2 ), C 0 is the concentration of the redox species (mol/cm 3 ), D is the diffusion coefficient of the redox species (cm 2 /s), R is the universal gas constant (8.314 J/mol K), T is the absolute temperature (K) and υ is the scan rate (V/s).
Appl. Sci. 2021, 11, x FOR PEER REVIEW 7 of 13 trode (cm 2 ), C 0 is the concentration of the redox species (mol/cm 3 ), D is the diffusion coefficient of the redox species (cm 2 /s), R is the universal gas constant (8.314 J/mol K), T is the absolute temperature (K) and υ is the scan rate (V/s).

Chronoamperometry
To evaluate the performance of the potentiostat in chronoamperometric detection, an experiment was performed with ferricyanide at concentrations ranging from 25-100 µM. For chronoamperometric detection, a fixed potential was applied at −100 mV and the measurement period was 30 s. Although the signal response from the NFC potentiostat showed current noise, as shown in Figure 5a, while the current signals from the commercial potentiostat were smooth and free of noise (Figure 5b), there was good agreement between data obtained by the NFC potentiostat and the commercial potentiostat at different concentrations of ferricyanide (Figure 6a). The final steady-state current after 25 s, calculated as the mean current due to the steady-state value, was proportional to the concentration of the analyte [49,50]. Figure 6b demonstrates the theoretically expected linear relationship between the current and concentration of ferricyanide. The calibration curve obtained from the commercial potentiostat corresponds to Equation (2) with the regression coefficient as R 2 = 0.999, y (I(µA)) = 0.0207 × x(µM(mol/L)) + 0.0303, while the calibration curve for the NFC potentiostat was obtained as the following Equation (3) with R 2 = 0.998, y (I(µA)) = 0.0212 × x(µM(mol/L)) + 0.0169.
The reliability of the potentiostat in detection of the analyte by chronoamperometry was calculated by comparing the slopes between our device and the standard device [51]. The result demonstrated that the reliability was greater than 97%. The average (AVG) and relative standard deviation (RSD) values of the current obtained from both potentiostats are shown in Table 2. The current values obtained from the commercial potentiostat were reproducible with RSD of ±2.55% and ±3.40% corresponding to chronoamperometric measurement for 25 × 10 −6 M and 1000 × 10 −6 M, respectively, while the RSD obtained from the NFC potentiostat gave values ±0.91% and ±7.74% for 25 × 10 −6 M and 1000 × 10 −6 M, respectively.

Chronoamperometry
To evaluate the performance of the potentiostat in chronoamperometric detection, an experiment was performed with ferricyanide at concentrations ranging from 25-100 µM. For chronoamperometric detection, a fixed potential was applied at −100 mV and the measurement period was 30 s. Although the signal response from the NFC potentiostat showed current noise, as shown in Figure 5a, while the current signals from the commercial potentiostat were smooth and free of noise (Figure 5b), there was good agreement between data obtained by the NFC potentiostat and the commercial potentiostat at different concentrations of ferricyanide (Figure 6a). The final steady-state current after 25 s, calculated as the mean current due to the steady-state value, was proportional to the concentration of the analyte [49,50]. Figure 6b demonstrates the theoretically expected linear relationship between the current and concentration of ferricyanide. The calibration curve obtained from the commercial potentiostat corresponds to Equation (2) with the regression coefficient as R 2 = 0.999, y (I(µA)) = 0.0207 × x(µM(mol/L)) + 0.0303, while the calibration curve for the NFC potentiostat was obtained as the following Equation (3) with R 2 = 0.998, y (I(µA)) = 0.0212 × x(µM(mol/L)) + 0.0169.
The reliability of the potentiostat in detection of the analyte by chronoamperometry was calculated by comparing the slopes between our device and the standard device [51]. The result demonstrated that the reliability was greater than 97%. The average (AVG) and relative standard deviation (RSD) values of the current obtained from both potentiostats are shown in Table 2. The current values obtained from the commercial potentiostat were reproducible with RSD of ±2.55% and ±3.40% corresponding to chronoamperometric measurement for 25 × 10 −6 M and 1000 × 10 −6 M, respectively, while the RSD obtained from the NFC potentiostat gave values ±0.91% and ±7.74% for 25 × 10 −6 M and 1000 × 10 −6 M, respectively.

Application of NFC Potentiostat
To evaluate the application of the NFC potentiostat, it was used to measure the ascorbic acid concentration and determine ascorbic acid in spiked commercial drinking water. The detection limit of standard ascorbic acid was determined by cyclic voltammetry over the potential range from 200 to 800 mV. Cyclic voltammograms of different concentrations of ascorbic acid (0.1-1 mg/mL) are shown in Figure 7a. The current at 550 mV corresponds to the oxidation of ascorbic acid [19,52]. Therefore, the current at 550 mV was used to plot the calibration curve as shown in Figure 7b. The linear regression equation of ascorbic

Application of NFC Potentiostat
To evaluate the application of the NFC potentiostat, it was used to measure the ascorbic acid concentration and determine ascorbic acid in spiked commercial drinking water. The detection limit of standard ascorbic acid was determined by cyclic voltammetry over the potential range from 200 to 800 mV. Cyclic voltammograms of different concentrations of ascorbic acid (0.1-1 mg/mL) are shown in Figure 7a. The current at 550 mV corresponds to the oxidation of ascorbic acid [19,52]. Therefore, the current at 550 mV was used to plot the calibration curve as shown in Figure 7b. The linear regression equation of ascorbic

Application of NFC Potentiostat
To evaluate the application of the NFC potentiostat, it was used to measure the ascorbic acid concentration and determine ascorbic acid in spiked commercial drinking water. The detection limit of standard ascorbic acid was determined by cyclic voltammetry over the potential range from 200 to 800 mV. Cyclic voltammograms of different concentrations of ascorbic acid (0.1-1 mg/mL) are shown in Figure 7a. The current at 550 mV corresponds to the oxidation of ascorbic acid [19,52]. Therefore, the current at 550 mV was used to plot the calibration curve as shown in Figure 7b. The linear regression equation of ascorbic acid measurement was y(µA) = 4.709x(mg/mL) + 1.198, with a correlation coefficient of 0.998. The limit of detection (LOD) was calculated from 3 s/m and the limit of quantitation (LOQ) was obtained from 10 s/m where s is the standard deviation of blank (n = 10) and s is the slope of the calibration curve. According to the equation, the limit of detection and the limit of quantitation were calculated to be 0.0024 mg/mL and 0.0080 mg/mL, respectively. King et al. reported that the detection limit for measuring ascorbic acid using the pencil electrode was 0.0326 mg/mL [52]. So, the limit of detection for ascorbic acid determination using a screen-printed carbon electrode in this study was lower than the previous work. The proposed NFC potentiostat was employed to determine ascorbic acid in spiked commercial drinking water. The results are shown in Table 3. The recoveries were 98.78% ± 4.74, 95.80% ± 2.87 and 92.69% ± 2.35 for spiking ascorbic acid at concentrations of 0.2, 0.4 and 0.6 mg/mL, respectively. The results indicated that the NFC potentiostat has acceptable accuracy for ascorbic acid measurement.
Appl. Sci. 2021, 11, x FOR PEER REVIEW 9 of 13 acid measurement was y(µA) = 4.709x(mg/mL) + 1.198, with a correlation coefficient of 0.998. The limit of detection (LOD) was calculated from 3 s/m and the limit of quantitation (LOQ) was obtained from 10 s/m where s is the standard deviation of blank (n = 10) and s is the slope of the calibration curve. According to the equation, the limit of detection and the limit of quantitation were calculated to be 0.0024 mg/mL and 0.0080 mg/mL, respectively. King et al. reported that the detection limit for measuring ascorbic acid using the pencil electrode was 0.0326 mg/mL [52]. So, the limit of detection for ascorbic acid determination using a screen-printed carbon electrode in this study was lower than the previous work. The proposed NFC potentiostat was employed to determine ascorbic acid in spiked commercial drinking water. The results are shown in Table 3. The recoveries were 98.78% ± 4.74, 95.80% ± 2.87 and 92.69% ± 2.35 for spiking ascorbic acid at concentrations of 0.2, 0.4 and 0.6 mg/mL, respectively. The results indicated that the NFC potentiostat has acceptable accuracy for ascorbic acid measurement. The calibration curve between current (µA) and concentration of ascorbic acid (mg/mL). The current observed at 550 mV was used to plot the calibration curve to determine the ascorbic acid concentration.

Discussion
This study demonstrated the design and performance of an NFC potentiostat (Silicon Craft Technology PLC, Bangkok, Thailand), which is a battery-free and flat card-sized device that interfaces with a smartphone through near-field communication. The device was designed for electrochemical measurement in point-of-care testing and resource-limited settings that do not require technical users. These potentiostats have the following aspects: (1) portability: the device is credit-card-sized (5.5 cm × 8.6 cm) and weighs only 8.07 g; (2) low cost: this potentiostat is designed to be used with a screen-printed electrode, which is low-cost, disposable and has ease of modification to detect various analytes [53,54]; (3) ease of use: the NFC potentiostat is operated via a smartphone application. The smartphone application allows the user to set the parameters of an electrochemical experiment and receive the resulting data; (4) wireless and cloud connectivity: the connection The calibration curve between current (µA) and concentration of ascorbic acid (mg/mL). The current observed at 550 mV was used to plot the calibration curve to determine the ascorbic acid concentration.

Discussion
This study demonstrated the design and performance of an NFC potentiostat (Silicon Craft Technology PLC, Bangkok, Thailand), which is a battery-free and flat card-sized device that interfaces with a smartphone through near-field communication. The device was designed for electrochemical measurement in point-of-care testing and resource-limited settings that do not require technical users. These potentiostats have the following aspects: (1) portability: the device is credit-card-sized (5.5 cm × 8.6 cm) and weighs only 8.07 g; (2) low cost: this potentiostat is designed to be used with a screen-printed electrode, which is low-cost, disposable and has ease of modification to detect various analytes [53,54]; (3) ease of use: the NFC potentiostat is operated via a smartphone application. The smartphone application allows the user to set the parameters of an electrochemical experiment and receive the resulting data; (4) wireless and cloud connectivity: the connection of smartphones facilitates sharing and storage of all information to the database in order to minimize the manual error and create big data for further analysis, such as machine learning or artificial intelligence.
The performance of the NFC potentiostat was evaluated by comparing it to a commercial benchtop potentiostat. Cyclic voltammetry and chronoamperometry were carried out in a potassium ferricyanide solution containing 0.1 M KCl. The results showed good agreement of cyclic voltammograms between the two devices, but current signals of our NFC potentiostat were lower than the commercial potentiostat. The deviation was from the difference of applied potential step voltage from the DAC and differences in sampling time and period designed for each potentiostat. For amperometric measurements, the slope of the current response obtained by the NFC potentiostat was similar to the commercial potentiostat with a reliability greater than 97%, although the signal from the NFC potentiostat was observed to have current noise, compared with the commercial potentiostat, which was free of noise. The noise was from a limitation of the analog-to-digital converter (ADC) resolution of the potentiostat, which is 12 bit and less than that of the commercial benchtop potentiostat. Note that the setting measurement range was ±20 µA, hence, the noise floor should be around 10 nA.
The first prototype of the NFC potentiostat can perform cyclic voltammetry and chronoamperometry with the input current ±20 µA and the biasing voltage at ±0.8 V. Although the benchtop potentiostats offer a wide operational window of current and voltage (±2 A, ±10 V), they are too expensive and large for on-site measurement. Therefore, an NFC potentiostat can be an alternative device for electrochemical measurement in pointof-care testing and resource-limited settings. In a future prototype, the operation ranges will be expanded and the system will be developed to support various electrochemical techniques such as linear sweep voltammetry, square wave voltammetry and differential pulse voltammetry.

Conclusions
This study demonstrated the design and performance of an NFC potentiostat developed by Silicon Craft Technology PLC (Bangkok, Thailand). The NFC potentiostat is a flat, card-sized, lightweight device and is operated using an NFC smartphone, so it is portable and easy to use. The performance of the NFC potentiostat exhibits a good agreement with a commercial benchtop potentiostat for both cyclic voltammetry and chronoamperometry. The application of this device was demonstrated through its use in ascorbic acid measurement. This platform is suitable for various analytes, depending on the design and modification of the electrodes used as a sensor. Therefore, an NFC potentiostat can be an alternative device for electrochemical measurement in point-of-care testing and resource-limited settings in many fields, including healthcare, agriculture and the environment.