Facile Fabrication of All-solid-state Ion-selective Electrodes by Laminating and Drop-casting for Multi-sensing

Abstract

In this study, we propose a facile fabrication process for all-solid-state ion-selective electrodes by laminating and drop-casting; these electrodes are suitable for applications in the environmental, agricultural, and medical fields, such as oral condition monitoring. To date, research has focused only on developing ion-selective electrodes for specific applications. However, an ion-selective electrode with compatibility for wide use has not been realized because specialized processing techniques and equipment are required. Our fabrication process achieved an ion-selective multi-sensor with a wireless system, which is crucial for overcoming the aforementioned challenge. The developed sensors exhibited sufficient sensitivity, repeatability, response time, and selectivity for medical, environmental, and agricultural applications.

1. Introduction

Monitoring the changes in ion concentration is significant for a wide range of areas, such as environmental,¹ agricultural,² industrial, and medical applications.³ This process can predict catastrophic shifts using sensor networks and emerging data technologies⁴ like the initiation of dental caries through elution of Ca²⁺ from the tooth surface^5,6 triggered by the collapse of homeostasis.⁷

Benchtop processes like atomic absorption spectroscopy, inductively coupled plasma optical emission spectrometry, and ion chromatography have been developed to quantify the various ion concentrations. These benchtops have high sensitivity and selectivity but lack portability. In contrast, ion-selective electrodes are cheap, compact, and portable sensors that operate on low energy without requiring a complex setup. Thus, they are used in wearable sensors that detect body fluids, including sweat,^8,9 urine,⁸ and tears.⁸ In particular, the electrodes that use ionophores as sensor elements are portable and hence suitable for use in wearable sensors. Additionally, the usage of various ionophores enables sensing of more than 60 different analytes (e.g., Mg²⁺, K⁺, Na⁺, Ca²⁺, Li⁺, F⁻, and Cl⁻),¹⁰ thereby covering a wide range of medical and environmental applications.

Despite the high demand for ion-selective electrodes for use in the aforementioned applications, an ion-selective electrode that can be used in all fields has not yet been developed. The electrodes that are produced in laboratories or industries are not compatible for use in a wide range of areas. We considered that the key to producing ion-selective electrodes that are widely applicable is the concept of “Do-It-Yourself (DIY).” This concept enables realization of custom-made ion-selective electrodes to adapt to different circumstances and makes them suitable for wearable ion sensors. For example, DIYbio¹¹ provides open access to biology and simple methods, including CRISPR-Cas9 for gene editing, which helps in spreading knowledge of biotechnology to the general public as well as expansion of its applications. Thus, design of a facile fabrication process of ion-selective electrodes without specialized processing techniques for realizing an open access technology will promote both versatile applicability of these electrodes and the era of the Internet of Things (IoT).

In the present study, we propose a simple procedure for all-solid-state ion-selective electrodes using a laminator and pipette available inhouse (as shown in Figs. 1a and 1b). We integrated Ca²⁺, Mg²⁺, K⁺, and Na⁺ selective ionophores as sensor elements in the electrodes for multi-sensing purposes. In addition, an open-source Arduino microcontroller with a Bluetooth module was assembled for system control, data processing, and wireless control (Fig. 1c). This enabled the generalization of these electrodes. Following the experimental setup, the sensor performance of the all-solid-state type ion-selective electrode was compared with that of a conventional inner solution type ion-selective electrode. Subsequently, the sensor performance, sensitivity, repeatability, and selectivity were investigated to discuss the versatile applications of the developed electrodes.

Figure 1.

(a) Deal drawing of an all-solid-state type electrode, (b) fabrication process of an all-solid-state electrode using laminating and drop-casting, and (c) conceptual images showing the application of ion-selective sensors with a wireless module produced by facile fabrication.

2. Experimental

2.1 Materials and reagents

Polyvinyl chloride (PVC), 2-Nitrophenyl octyl ether (NPOE), sodium tetrakis 3,5-bis(trifluoromethyl)phenyl borate (Na-TFPB), tetrahydrofuran (THF), Bis[(benzo-15-crown-5)-4-methyl]pimelate (Bis(benzo-15-crown-5)), and Bis[(12-crown-4)methyl]2-dodecyl-2-methylmalonate (Bis(12-crown-4)) were purchased from FUJIFILM Wako. Calcium ionophore II (ETH 129), magnesium ionophore I (ETH 1117), and poly(3,4-ethylenedioxythiophene)-poly(styrenesulfonate) (PEDOT:PSS; 1.0 wt% in H₂O) were purchased from Sigma-Aldrich. Laminate film was purchased from IRIS OHYAMA Inc. (100 µm, LZ-B620). Deionized water (18.2 MΩ cm) was used for all experiments (Millipore Milli-Q Direct MeRCK). Ag/AgCl reference electrode (3 mol L⁻¹ (M) NaCl solution) was purchased from BAS Inc. The chemical structures of ionophores are shown in Fig. S1.

2.2 Inner solution-type sensors fabrication

Ion-selective membrane solutions were prepared by dissolving ionophore, PVC, NPOE, and TFPB in THF. Bis(benzo-15-crown-5), Bis(12-crown-4), ETH 129, and ETH 1117 were added as ionophores. The compositions of ion-selective membrane cocktails and the conditions are listed in Table 1. The values were determined by an extensive literature survey of published research in the field.^12–15 Ion-selective membrane cocktails were poured onto a 26 mm diameter glass petri dish, and the amount of ion-selective membrane solutions was changed to control membrane thickness. The membrane solutions were dried at 25 °C in a closed box for 24 h, resulting in the formation of ion-selective membranes (3 mm diameter), which were subsequently cut out and peeled off. THF was used to glue the ion-selective membranes on a 2 mm diameter PVC pipe. A 3 M KCl solution was used as an inner solution. Ag/AgCl was employed as an electrode. All the procedures were performed at 25 °C.

Table 1. Composition of ion-selective membranes cocktail used for sensors.

Sensor	NPOE /mg	TFPB /mg	PVC /mg	Ionophore /mg	THF /mL	Conditioning solution	Amount of Cocktails drop-casted /mL cm−2	Ref.
Na+	250	2	100	Bis(12-crown-4) 10.4	1.5	4 M NaCl	1.7	12
Ca2+	93	1	47	ETH 129 1	1.5	0.01 M CaCl2	4.0	13
K+	99	9	51	Bis(benzo-15-crown-5) 9	3	0.01 M KCl	4.2	14
Mg2+	99	9	51	ETH 1117 9	3	0.01 M MgCl2	4.2	15

2.3 All-solid-state-type sensors fabrication

The ion-selective membrane solutions were the same as inner solution-types (Table 1). An Au electrode on PET film was used as a conductor. All-solid-state-type ion-selective electrodes were prepared, as shown in Figs. 1a and 1b. The PET film was first cleaned with isopropyl alcohol. A Ti/Au layer was deposited on the PET films (SVC-700TM, SANYUELECTRON, Ti thickness: 30 nm, Au Thickness, 50 nm). The deposited film was sealed with a PET film, including 3 mm diameter sensing area by lamination. We used PEDOT(PSS) as a conductive polymer. We drop-casted 2.5 µL PEDOT(PSS) onto the sensing area, and PEDOT(PSS) was cured at 140 °C for 5 min. Subsequently, ion-selective membrane solutions were drop-casted onto the sensing area. The compositions of ion-selective membrane solutions are listed in Table 1. The membrane solutions were dried at 25 °C for 24 h. A supplemental movie shows the fabrication process using a Ti film conductor.

2.4 Potentiometric measurements

The potentiometric measurements were performed at 25 °C using a 4-channel electrical potential measurement system consisting of a 16-bit analog to digital converter (ADS1115), a microcomputer board (Arduino Pro mini), and a wireless interface module (Xbee). Ag/AgCl (3.0 M NaCl) reference electrode was connected to the ground terminal such that potential between the four sensors and the reference electrodes could be monitored wirelessly. The measurement program was created using the Arduino Integrated Development Environment (IDE ver. 1.8.5.). The sampling rate was 1 s/data. All-solid-state-type and inner solution-type of K⁺, Na⁺, Ca²⁺, and Mg²⁺ sensors were conditioned for 2 h in KCl, NaCl, CaCl₂, and MgCl₂ solution (Table 1), respectively. The selectivity was evaluated using a separate solution method.¹³ The activity coefficient of ions was estimated by Debye–Hückel equation.¹⁶

3. Results and Discussion

3.1 Comparison between inner solution-type and all-solid-state-type sensors

The ion-selective electrode consists of two types of sensors, i.e., inner solution-type and all-solid-state-type sensors (Figs. 2a and 2b). Solution-type sensors have been extensively studied, and some of the inner solution-type sensors are commercially available. Accordingly, we utilized inner solution-type sensors to evaluate all-solid-state-type sensors. The inner solution type is typically bulky and fragile, which limits its application in terms of wearability, portability, and expendability in sensors.¹⁵ Therefore, all-solid-state electrodes have garnered attention due to their simplicity and compact structure.

Figure 2.

(a) All-solid-state electrode, (b) multi-ion sensor using Arduino microcontroller and Bluetooth modules, (c) potential shift based on ion concentration for K⁺ sensor, (d) potential shift based on ion concentration for Na⁺ sensor, (e) potential shift based on ion concentration for Ca²⁺ sensor, and (f) potential shift based on ion concentration for Mg²⁺ sensor. Error bars represent the standard error (N = 3). Red and black square plots show the all-solid state-type and solution-type electrodes, respectively.

We prepared an all-solid-state-type sensor (as presented in Figs. 1a and 1b). A conductor was laminated between the PET films that were used as insulators in our process. Conductive polymers and ionophores mixed with PVC were dropped on the area where the PET film did not cover the conductor. Selective ionophores were also incorporated into the electrodes. Note that lamination for insulation covering is a method that also involves the use of a non-conductive solvent such as gel nail for preparing rod-shaped ion-selective electrodes¹⁷ (Fig. S2). This electrode type can also measure several ion species by collecting ionophores (Fig. S3).

The aforementioned procedures ensure that the produced ion-selective electrodes are compatible for use in a wide range of fields. The ion-selective electrode’s size and shape are efficiently controlled by modifying the morphology of conductors, lamination films, and types of ionophores. Furthermore, a multi-ion sensing system using an open-source Arduino microcontroller with Bluetooth modules was developed to record the electrical signal from the ion-selective electrode. This apparatus, combined with the ion-selective electrodes, enabled parallel three-channel ion sensing and wireless data transfer (Fig. 2b). To verify the sensor performance of the all-solid-state type prepared through lamination, we compared the performance with that of an inner solution-type sensor. As displayed in Figs. 2c–2f, the sensors incorporating K⁺, Na⁺, Ca²⁺, and Mg²⁺ selective ionophores exhibited Nernstian behavior: the linear change based on the logarithm of the concentration. The variations in the absolute potentials of different sensors were resolved by one-point calibration. The open-circuit potentials of the sensors in 0.1 mM of each solution were set to zero. This zero point adjusting is similar to commercial pH sensing.¹⁷ All-solid-state sensors exhibited Nernstian behavior comparable to those of the inner solution. Miniaturization and robust structure are the key advantages of the all-solid-state sensors.¹⁸ In addition, decreasing the thickness of all-solid-state-type membranes is feasible because of the support of the PET substrate underneath the membrane, thereby expanding the detection limit.¹⁹

3.2 Comparison of measurable range of sensors and the ion concentration range for versatile fields

The ion concentration range of human body fluids (i.e., saliva, sweat, blood, and urine), soil, river, and sea with respect to the applications of ion sensors are shown in Fig. 3.^5,20–25 The detection limits of developed sensors were estimated by 3 sigma method from Figs. 2c–2f. The backgrounds were defined by slope lower than 20 mV/dec (between two points). The measurable ranges of all-solid-state sensors partially satisfied the ion concentration variation of K⁺, Na⁺, Ca²⁺, and Mg²⁺ in the applications mentioned previously (Fig. 3), suggesting the validity of all-solid-state sensors prepared through this facile process.

Figure 3.

Measurable range of the sensors (orange line) and ion concentration range of human body fluids (saliva,^5,20 sweat,²¹ blood,²⁰ and urine²²), soil,²³ river water,²⁴ and sea water.²⁵ Measurable range of the sensors (orange line) were estimated from detection limits of developed sensors.

Furthermore, sensitivity S was defined as the potential slope from the detection limit to 1 M (Figs. 2c–2f). We determined the S of the inner solution and all-solid-state sensors and relative standard deviation (RSD), as summarized in Table 2. The ideal sensitivity S of the sensors was described by the Nernst equation:   

\begin{equation} E = E_{0} + 2.303\frac{RT}{z_{i}F}\log a_{i} \end{equation} (1a)

  

\begin{equation} S \equiv 2.303\frac{RT}{z_{i}F} \end{equation} (1b)

where E₀ are the standard potential of measured potentials E with the ion activity a_i of the primary ion i, temperature T (25 °C). z_i is the valency of the primary ion i. In this equation, R and F represent the universal gas constant and the Faraday constant. Thus, the ideal Nernst slope for monovalent ions, K⁺ and Na⁺, is 59.2 mV/dec. In addition, the Nernst slope for divalent ions, Ca²⁺ and Mg²⁺, is 29.6 mV/dec. As shown in Table 2, the slopes of Na⁺, Ca²⁺ and Mg²⁺ sensors were close to the ideal Nernst slope. But the slope of K⁺ was 42.3 mV/dec was lower than the ideal slope. The low sensitivity of the K⁺ could be improved by using valinomycin as an ionophore.

Table 2. Sensitivity of K⁺, Na⁺, Ca²⁺, and Mg²⁺ sensors with inner solution and all-solid-state types. The values of S and coefficient of determination (R²) are estimated from the concentration range shown in Fig. 3 as each orange line (N = 3).

Ion	Inner solution type			All-solid-state
	S /mV decade−1	R2	RSD /%	S /mV decade−1	R2	RSD /%
K+	39.9	0.97	12.0	42.3	0.96	4.2
Na+	50.1	0.98	8.5	56.3	0.99	1.5
Ca2+	19.6	0.99	5.0	25.7	0.99	8.4
Mg2+	21.4	0.93	42.9	30.0	0.98	19.2

The noise level recorded by Arduino was less than 0.5 mV (standard deviation). Thus, the sensors with ion-selective electrodes possessed sufficient sensitivity and accuracy to quantify the ion concentration.

3.3 Evaluation of repeatability and response time

To investigate the repeatability and response of the all-solid-state sensor, concentrations were changed consecutively from high to low and then to high (0.25–2 mM). Figures 4a–4d illustrates the repeatability of the all-solid-state sensors, showing their dynamic response that corresponds to the consecutive ion concentration changes. The repeatability was based on the potential difference after consecutive changes from high to low and then to high concentrations (Fig. 4e). The values of repeatability for the K⁺, Na⁺, Ca²⁺, and Mg²⁺ sensors were within 5 mV (Fig. 4e). Moreover, the repeatability was affected by drift, as shown in Fig. S4, which was caused by the creation of a water layer that resulted from the water uptake in the membranes.²⁶ The calibration update algorithm has been developed for multi-sensor systems.²⁷ Our developed sensors possessed improved repeatability and accuracy for long-term monitoring. An example is the application of ion-selective electrodes in monitoring the temperature changes in the environment. The temperature shift also affects the potential difference according to the Nernst equation. The potential shift in this case was linear to the temperature change, as shown in Fig. S5. Thus, the potential shift caused by temperature change can be modified by integrating a thermometer.

Figure 4.

Responses of (a) K⁺, (b) Na⁺, (c) Ca²⁺, (d) Mg²⁺ sensors based on the target ion concentration, (e) repeatability of the sensors, and (f) response time of the sensors. Response time was evaluated after consecutive changes from low to high concentrations.

In addition, the sensor response time was investigated, as shown in Fig. 4f. The response time is the time required by the sensors to reach 90 % of their final values. The all-solid-state sensor quantified the ion concentration in less than 60 s. Thus, Fig. 4 shows good repeatability (i.e., higher than 11 %) and quick response (i.e., less than 60 s). This indicates the validity of our all-solid-state sensor in terms of these two parameters.

3.4 Evaluation of selectivity of the developed sensor

Selectivity is one of the most significant parameters. To evaluate the selectivity of the sensors, the selectivity coefficients of ion-selective electrodes were determined based on experimental data as follows:   

\begin{equation} \log K_{\text{A,B}}^{\text{pot}} = \frac{z_{\text{A}}F(E_{\text{B}} - E_{\text{A}})}{2.303RT} + \log \left(\frac{a_{\text{A}}}{a_{\text{B}}^{z_{\text{A}}/z_{\text{B}}}}\right) \end{equation} (2)

where E_A and E_B are measured potentials of target ion and interfering ion with ion activity a_A and a_B, respectively. Terms R, F, and T represent constants having standard meanings.^21,22 Results presented in Table 3 confirm that a majority of the all-solid-state sensors exhibited high selectivity to the target analyte ions. As listed in Table 3, Ca²⁺ sensor exhibited high selectivity, exceeding −7.0, thereby indicating that its sensitivity to Ca²⁺ was 10000000 times higher than those to K⁺, Mg⁺, and Na⁺. In contrast, Mg²⁺ sensor exhibited low sensitivity (1.2 ± 0.1) against Ca²⁺, indicating that Mg²⁺ sensor could not distinguish Mg²⁺ from Ca²⁺. Notably, our multi-ion sensing can complement the low selectivity of the Mg²⁺ sensor with Ca²⁺, whose selectivity is high against Mg²⁺. Thus, our multi-ion sensors with ion-selective electrodes prepared through a facile process can achieve higher selectivity than a single-ion sensor.

Table 3. Selectivity coefficients (K_A,B) of the K⁺, Na⁺, Ca²⁺, and Mg²⁺ sensors developed in this study.

Target ion	Interfering ions	Our sensors log KA,B
K+	Ca2+	−1.4
	Mg2+	−3.7 ± 0.1
	Na+	−1.6 ± 0.2
Na+	Ca2+	−2.9
	Mg2+	−3.5 ± 0.1
	K+	−2.0 ± 0.4
Ca2+	Mg2+	−7.4 ± 0.1
	K+	−7.1 ± 1.8
	Na+	−7.8 ± 16
Mg2+	Ca2+	1.2 ± 0.1
	K+	−1.4 ± 0.1
	Na+	−2.5 ± 0.1

4. Conclusions

In this study, we have proposed a simple process for developing an ion-selective multi-sensor with a wireless system. No existing ion-selective electrode can satisfy multiple applications because of low extensibility, and compatibility, even though various types of ion-selective electrodes have been extensively researched. The all-solid-state ion-selective electrode proposed herein can be prepared through laminating and pipetting. The proposed electrode exhibited sufficient sensitivity, repeatability, response time, and selectivity, implemented by connecting an open-source Arduino microcomputer with a precision analog to a digital converter and a wireless interface module. Based on the results, it is worth considering as an inhouse ion sensor to customize or optimize an ion-selective electrode for specific applications in several fields. Furthermore, the performance that is related to sensitivity, selectivity, and accuracy of the inhouse sensors can be enhanced by emerging algorithms,^27,28 which can result in the state-of-the-art ion-selective sensors being superseded.

Acknowledgments

A part of this study is based on the Cooperative Research Project of Research Center for Biomedical Engineering. This study received financial support from KAKENHI Grant-in-Aid for Scientific Research (C) grant number 19K10418, Adaptable and Seamless Technology transfer Program through Target-driven R&D (A-STEP) from Japan Science and Technology Agency (JST) Grant Number JPMJTR21RA.

Data Availability Statement

The data that support the findings of this study are openly available under the terms of the designated Creative Commons License in J-STAGE Data at https://doi.org/10.50892/data.electrochemistry.19920182.

CRediT Authorship Contribution Statement

Shingo Okubo: Investigation (Lead), Writing – original draft (Lead)

Yoshihisa Ozeki: Investigation (Lead), Writing – original draft (Lead)

Tetsuya Yamada: Project administration (Lead), Writing – original draft (Lead)

Kosuke Saito: Investigation (Lead)

Noboru Ishihara: Investigation (Lead), Methodology (Lead)

Yasuko Yanagida: Conceptualization (Lead), Funding acquisition (Lead), Project administration (Lead), Supervision (Lead)

Gen Mayanagi: Funding acquisition (Lead), Resources (Lead), Validation (Lead)

Jumpei Washio: Resources (Lead), Validation (Lead)

Nobuhiro Takahashi: Resources (Lead), Validation (Lead), Writing – review & editing (Lead)

Conflict of interest

The authors declare no competing financial interest.

Funding

Japan Society for the Promotion of Science: 19K10418

Adaptable and Seamless Technology Transfer Program through Target-Driven R and D: JPMJTR21RA

Footnotes

Y. Yanagida: ECSJ member

References

• 1)  R. De Marco, G. Clarke, and B. Pejcic, Electroanalysis, 19, 1987 (2007).
• 2)  L. Ruan, K. Wei, L. Wang, H. Cheng, F. Zhang, L. Wu, P. Bai, and C. Zhang, Sci. Rep., 6, 38370 (2016).
• 3)  T. Terse-Thakoor, M. Punjiya, Z. Matharu, B. Lyu, M. Ahmad, G. E. Giles, R. Owyeung, F. Alaimo, M. S. Baghini, T. T. Brunyé, and S. Sonkusale, npj Flex. Electron., 4, 18 (2020).
• 4)  J. Sun, S. L. Chen, S. X. Dai, R. G. Wang, N. Y. Li, X. Shen, X. Y. Zhou, C. F. Lu, X. J. Zheng, Z. M. Hu, Z. K. Zhang, J. Song, and Y. Xu, Plant Signal. Behav., 4, 261 (2009).
• 5)  G. Mayanagi, K. Igarashi, J. Washio, and N. Takahashi, Caries Res., 51, 160 (2017).
• 6)  G. Mayanagi, K. Igarashi, J. Washio, K. Nakajo, H. Domon-Tawaraya, and N. Takahashi, J. Dent. Res., 90, 1446 (2011).
• 7)  W. J. Cho, H. J. Kim, D. H. Jung, H. J. Han, and Y. Y. Cho, Sensors, 19, 5508 (2019).
• 8)  H. Y. Y. Nyein, W. Gao, Z. Shahpar, S. Emaminejad, S. Challa, K. Chen, H. M. Fahad, L. C. Tai, H. Ota, R. W. Davis, and A. Javey, ACS Nano, 10, 7216 (2016).
• 9)  W. Gao, S. Emaminejad, H. Y. Y. Nyein, S. Challa, K. Chen, A. Peck, H. M. Fahad, H. Ota, H. Shiraki, D. Kiriya, D. H. Lien, G. A. Brooks, R. W. Davis, and A. Javey, Nature, 529, 509 (2016).
• 10)  E. Bakker, P. Bühlmann, and E. Pretsch, Electroanalysis, 11, 915 (1999).
• 11)  T. Landrain, M. Meyer, A. M. Perez, and R. Sussan, Syst. Biol. Synth. Biol., 7, 115 (2013).
• 12)  Y. Shibutani, S. Mino, S. S. Long, T. Moriuchi-Kawakami, K. Yakabe, and T. Shono, Chem. Lett., 26, 49 (1997).
• 13)  E. Bakker, Anal. Chem., 69, 1061 (1997).
• 14)  Y. Yamashoji, M. Tanaka, S. Nagamune, M. Ouchi, T. Hakushi, and T. Shono, Anal. Sci., 7, 485 (1991).
• 15)  E. W. Nery, J. A. Guimarães, and L. T. Kubota, Electroanalysis, 27, 2357 (2015).
• 16)  R. G. Bates, Denki Kagaku, 46, 480 (1978).
• 17)  H. Freiser, J. Chem. Soc., Faraday Trans. 1, 82, 1217 (1986).
• 18)  J. Hu, A. Stein, and P. Bühlmann, TrAC, Trends Anal. Chem., 76, 102 (2016).
• 19)  T. Forrest, E. Zdrachek, and E. Bakker, Electroanalysis, 32, 799 (2020).
• 20)  M. Urbanowicz, A. Jasiński, M. Jasińska, K. Drucis, M. Ekman, A. Szarmach, R. Suchodolski, R. Pomećko, and M. Bocheńska, Electroanalysis, 29, 2232 (2017).
• 21)  B. L. Baker, T. C. Ungaro, A. K. Barnes, P. R. Nuccio, J. A. Reimel, and R. J. Stofan, Physiol. Rep., 2, e12007 (2014).
• 22)  F. Ghaderinezhad, H. C. Koydemir, D. Tseng, D. Karinca, K. Liang, A. Ozcan, and S. Tasoglu, Sci. Rep., 10, 13620 (2020).
• 23)  M. D. Jakovljević, N. M. Kostić, and S. B. Antić-Mladenović, Zb. Matice Srp. Prir. Nauke, 104, 11 (2003).
• 24)  A. K. Singh, G. C. Mondal, S. Kumar, T. B. Singh, B. K. Tewary, and A. Sinha, Environ. Geol., 54, 745 (2008).
• 25)  H. Sakihama, M. Ishiki, and A. Tokuyama, Atmos. Environ., 42, 2320 (2008).
• 26)  B. Hambly, M. Guzinski, B. Pendley, and E. Lindner, Electroanalysis, 32, 781 (2020).
• 27)  A. Rudnitskaya, A. M. S. Costa, and I. Delgadillo, Sens. Actuators, B, 238, 1181 (2017).
• 28)  A. D. Wilson and M. Baietto, Sensors, 11, 1105 (2011).