2025 年 93 巻 2 号 p. 027009
Interdigitated microarray electrodes were developed based on boron-doped amorphous carbon (B-a-C-IDA) that shows extremely higher overpotential for H2 and O2 evolution. Highly sensitive microanalysis of redox analytes with standard potentials higher than O2 and H2 evolution was achieved by B-a-C-IDA. The amplification of the oxidation current derived from redox cycling was observed at the generator electrode of B-a-C-IDA in the measurement with applying the reduction potential to the collector electrode (dual mode measurement). For Ce3+/4+ with a standard potential of 1.6 V, the amperometric current was amplified 180-fold by applying potentials of 1.7 and 0.8 V to collector and generator electrodes in chronoamperometry (CA) measurement at dual mode. Theoretical detection limit (S/N = 3) for Ce3+ was 0.13 µM. It was two orders of magnitude better than that at B-a-C plate electrode.
The factors controlling amplification at B-a-C-IDA with varying gap values were investigated using redox analytes with different types of reactions, diffusion coefficients, and electron transfer rate constants. In addition, the amplification mechanism at B-a-C-IDA electrodes, the contributions of diffusion coefficients, and electron transfer rate constants to the amplification factor were clarified. The method to estimate the amplification factor with diffusion coefficients and electron transfer rate constants when using unmeasured redox analytes was established. The amplification factor when using radioactive analyte UO22+ at B-a-C-IDA with gap size of 2.0 µm was estimated to be 43.2. B-a-C-IDA is expected to enable highly sensitive electroanalysis (low detection limit 0.62 µM).
With increasing awareness of environmental issues, there is an increasing need to measure the concentrations of heavy metal ions in the environment. In recent years, cobalt, manganese, and cerium have been widely used in mass-produced products such as lithium-ion batteries. These heavy metals are included as metal ions in mine drainage and factory wastewater from refineries and manufacturing plants where heavy metals are used. When industrial wastewater flow into rivers, concentrations of heavy metal ions in rivers increase. This may result in heavy metal ions being ingested by humans through crops produced in the surrounding area. Excessive intake of metal ions has been reported to cause various diseases, including neurotoxicity. Water Pollution Control Law stipulates that the concentrations of Co2+, Mn2+, and Ce3+ in wastewater must be kept below several micro-molar.1,2 Generally, redox species of metal ions are quantitatively analyzed using photochemical method. Analyzing samples, however, requires long time and is not suitable for environmental monitoring.
Electrochemical sensing, which enables rapid quantitative analysis, seems to be a promising method. The standard potentials of Co3+, Mn3+, and Ce4+ are higher than that of O2 evolution.
| \begin{equation} \text{Co}^{3 + } + \text{e}^{ - } \rightleftarrows \text{Co}^{2 + }\ (1.83\,\text{V vs. SHE}) \end{equation} | (1) |
| \begin{equation} \text{Mn}^{3 + } + \text{e}^{ - } \rightleftarrows \text{Mn}^{2 + }\ (1.50\,\text{V vs. SHE}) \end{equation} | (2) |
| \begin{equation} \text{Ce}^{4 + } + \text{e}^{ - } \rightleftarrows \text{Ce}^{3 + }\ (1.61\,\text{V vs. SHE}) \end{equation} | (3) |
At conventional electrodes such as Pt or glassy carbon (GC), redox reactions of the above analytes cannot be observed due to the interference of O2 evolution.
In addition, there is an increasing need to measure concentrations of radioactive substances in spent nuclear fuel pools and coolants at nuclear power plants.3–5 Electrochemical sensing is attracting attention because it can monitor the environment without the sophisticated instruments or the pretreatment by technical staffs.6 Radioactive water in spent nuclear fuel pools is known to contain ionized radioactive substances such as 137Cs, 238U, 239Pu, and 60Co. The redox reactions of these ionized radioactive substances occur at the standard potentials shown below.
| \begin{equation} \text{Cs}^{ + } + \text{e}^{ - } \rightleftarrows \text{Cs}\ ( - 2.92\,\text{V vs. SHE}) \end{equation} | (4) |
| \begin{equation} \text{U}^{4 + } + \text{e}^{ - } \rightleftarrows \text{U}^{3 + }\ ( - 0.61\,\text{V vs. SHE}) \end{equation} | (5) |
| \begin{equation} \text{UO}_{2}{}^{2 + } + \text{e}^{ - } \rightleftarrows \text{UO}_{2}{}^{ + }\ ( - 0.338\,\text{V vs. SHE}) \end{equation} | (6) |
| \begin{equation} \text{Pu}^{3 + } \rightleftarrows \text{Pu}^{4 + } + \text{e}^{ - }\ (0.46\,\text{V vs. SHE}) \end{equation} | (7) |
The reduction reactions of Cs+/0, U4+/3+, and UO22+/+ occur at potentials lower than the standard potential of H2 evolution (0 V vs. SHE). The oxidation reaction of Co2+/3+ occurs at potentials higher than the standard potential for O2 evolution (1.23 V vs. SHE). The redox reactions of these radioactive analytes cannot be observed at conventional electrodes such as Pt and GC due to the interference of H2 and O2 evolution.
In the case of 238U, which is known to be the most abundant radioactive substance contained in spent nuclear fuel pools, its concentration in normal (non-polluted) environment is 0.3 µM when taking account of a half-life, and is calculated to be 1637 µmol L−1 (µM) in spent nuclear fuel pools. The discharge standard is set at 6.0 µM (= 0.04 Bq/mL) to avoid harm to human health.6 Hence, a wide dynamic range of more than four-orders of magnitude, from 0.3 µM to 1637 µM, is required for the electrochemical sensor of 238U.
In this study, boron-doped amorphous carbon (B-a-C) was selected as an electrode material.7 B-a-C has higher overpotential for O2 and H2 evolution. Redox reactions of analytes such as Ce3+ and UO22+/+ can be observed at B-a-C. The detection limit of Ce3+ (model analyte) at B-a-C thin film electrode (B-a-C-plate) was found to be 39 µM (>S/N = 3).7 It is two orders of magnitude higher than the regulated value mentioned above. In order to improve the detection limit, our research group tried to fabricate interdigitated microarray (IDA) electrodes of B-a-C (B-a-C-IDA) that seem to show one or two orders of magnitude higher sensitivity than that of B-a-C-plate. IDA electrode consists of a set of two comb-like microband array electrodes, generator and collector electrodes (the generator and the collector), facing each other. The potential of one microband pair (the collector) is set to the reduction potential of redox analyte, and that of the other pair (the generator) is swept toward the oxidation potential. The oxidation current at the generator is measured with applying the reduction potential to the collector (dual mode measurement) or without applying a potential to the collector (single mode measurement). In dual mode measurement, redox cycling occurs in the area where diffusion layers overlap between the collector and generator,8 and the concentration of reduction products increases near the generator. The oxidation current that the generator exhibits by the amplification derived from redox cycling becomes higher than that without applying the reduction potential to the collector.
In this study, highly sensitive detection of Ce3+ in several hundred nano-molar ranges was developed by fabricating B-a-C-IDA. The thickness of the diffusion can be controlled with potential values applied on the generator and collector. We tried to enhance the amplification of redox current of Ce3+ by optimizing the potential applied to the generator and collector.
In Japan, the use of radioactive materials such as UO22+/+ for electroanalytical research must be permitted by the government agency (the Nuclear Regulation Authority). It is not realistic to use radioactive materials for the electroanalytical measurements using B-a-C-IDA fabricated in this study to examine the detection limit. The design of IDA electrode, however, can be determined if its detection limit can be accurately predicted using the reaction parameters (e.g., diffusion coefficient and heterogeneous electron transfer rate constant). In this study, our group attempted to clarify the relation between the amplification factor (Fa) when using the redox analytes and the reaction parameters. We also aimed to devise a technique to predict the detection limit of the redox analyte of interest at IDA electrode.
B-a-C films were prepared using a radio frequency (r.f.) plasma enhanced CVD system (model BP-1, SAMCO Co., Ltd.). The detailed procedure was reported previously.7 B-a-C films were deposited on quartz substrates (2.5 cm × 3.0 cm) using trimethyl borate and n-hexane as a source material after ultrasonic cleaning in 2-propanol for 15 minutes. The vapor generated from a degassed mixed solution of n-hexane and trimethyl borate kept at room temperature was introduced into an evacuated reaction chamber at a flow rate of 5 mL min−1. B-a-C film with thickness of 0.7 µm was grown by applying r.f. output (175 W, 13.56 MHz) on the cathode where the quartz substrate on a quartz liner was placed. The deposition time was 40 minutes. The chamber pressure was adjusted at 40 Pa. The stage temperature was set at 260 °C.
The structure of the film is designed as follows; one set of B-a-C-IDA consists of 65 pairs of microbands with a length of 200 µm and bandwidth of 2.0 µm. B-a-C-IDA with different gaps were fabricated. The gaps between two sets of B-a-C-IDA were designed to be 20.0, 10.0, 5.0, and 2.0 µm. Micro patterning was carried out using the process combining metal deposition, lift-off technique, and O2 reactive ion etching (O2 RIE), as reported in literature.9 Negative photoresist (ZPN 1150, ZEON Co., Ltd.) was spin coated on B-a-C. Photoresist was exposed using the mask aligner (MJB3, KARL SUSS Co., Ltd.) and developed by agent (Tetramethylammonium hydroxide of 2.5 atom%). Aluminum thin layer was deposited on the photoresist pattern with a resistance heating deposition system (SVC-700TMSG, Sanyu Electron Co., Ltd.). Negative photoresist was removed by immersing in acetone for one hour and ultrasonic cleaning in acetone for 10 seconds. O2 RIE of B-a-C was carried out with plasma etching apparatus (Model BP-1, SAMCO Co., Ltd.) for 15 minutes. The operating pressure of the chamber was 10 Pa, and r.f. power was 150 W. Aluminum pattern on B-a-C-IDA was removed by immersing in 0.1 M sodium hydroxide solution for 10 minutes.
Raman spectra of the samples were measured using Raman microscope system (RPM-500, JASCO Co., Ltd.) with Ar+ laser (wave length 514.5 nm) as an excitation source. The geometries and cross section of B-a-C-IDA were examined by a field emission scanning electron microscope (SEM) (S-4700, Hitachi High-Technologies Co., Ltd.).
The electrochemical measurements were carried out in a single compartment using three-electrode glass cell with Ag|AgCl as a reference electrode and Pt wire as a counter electrode. The cyclic voltammetry (CV) at B-a-C-plate and B-a-C-IDA to measure the potential window and background current was performed using a digital electrochemical analyzer (Model HZ-3000, Hokuto Denko Co., Ltd.) in 0.1 mol L−1 (M) nitric acid solution at room temperature. The CVs when using redox analytes at B-a-C-IDA were measured in single mode and dual mode with a dual potentio/galvanostat (700D series, BAS Co., Ltd.) at room temperature. The CVs at B-a-C-IDA were measured in 0.1 M HNO3, 100 µM Ce(NO3)3·6H2O + 0.1 M HNO3, 1 mmol L−1 (mM) K3[Fe(CN)6] + 0.1 M Na2SO4, 1 mM Ru(NH3)6Cl3 + 0.1 M Na2SO4, 1 mM Dopamine + 0.1 M HClO4, 1 mM 4-methylcatechol (MC) + 0.1 Na2SO4, 1 mM hydroquinone + 0.1 M Na2SO4 and 1 mM methyl viologen dichloride (MV) + 0.1 M HClO4. In single mode measurement, the generator electrode was swept at a sweep rate of 5, 10, 50, 100, and 500 mV/s. In dual mode measurement, the potential of the collector was kept at the optimized potential to obtain maximum current at the generator. The generator was swept at a sweep rate of 10 mV/s. In chronoamperometry (CA) measurements of Ce3+/4+ in dual mode measurement, the potentials at the collector and generator were set at 1.7 and 0.8 V, respectively.
AC impedance measurements were carried out with amplitude of 10 mV in the frequency range from 100 kHz to 0.1 Hz using a frequency response analyzer (Solartron, Type 1287 and 1260). Before the measurement, the solution was Ar bubbled for 15 minutes to remove dissolved O2 thoroughly. All measurements were performed at room temperature (25 °C).
All chemicals were used without further purification. N-hexane, trimethyl borate, 2-propanol, sodium hydroxide, nitric acid, sulfuric acid, perchloric acid, sodium sulfate, Ce(NO3)3·6H2O, K3[Fe(CN)6], Ru(NH3)6Cl3, dopamine, 4-methylcatechol, hydroquinone, and methyl viologen dichloride were prepared by Wako chemical Co., Ltd.
Figures 1a–1d show SEM images of the prepared B-a-C-IDA with gap sizes of 20.0, 10.0, 5.0, or 2.0 µm.10 Table 1 summarizes the observed bandwidths and gaps of the four type B-a-C-IDA. One of the four B-a-C-IDA electrodes was designed to have a bandwidth and gap of 2.0 µm, respectively (B-a-C-IDA with 2.0 µm). The resulting B-a-C-IDA with 2.0 µm was measured and its bandwidth and gap were found to be 2.04 and 1.96 µm. Compared to the designed values, the measured width was increased by 0.04 µm and gap decreased by 0.04 µm. The gap values of B-a-C-IDA with 10.0, 5.0, and 2.0 µm were decreased by 0.02–0.05 µm compared to the designed values, as shown in Table 1.
SEM images of B-a-C-IDA with (a) 20 µm, (b) 10 µm, (c) 5 µm, and (d) 2 µm. (e) Cross sectional, (f) 45° view, and (g) magnified figure of B-a-C-IDA with 2 µm. Figures (d) to (g) are reprinted from Ref. 10. Copyrights for the figures belong to The Electrochemical Society (ECS), all rights reserved.
| Designed value | Measured value | ||
|---|---|---|---|
| Width (µm) | Gap (µm) | Width (µm) | Gap (µm) |
| 2.00 | 20.00 | 1.96 | 20.04 |
| 2.00 | 10.00 | 2.05 | 9.95 |
| 2.00 | 5.00 | 2.02 | 4.98 |
| 2.00 | 2.00 | 2.04 | 1.96 |
The average diffusion length during the diffusion of reactant from the generator to the collector has been reported to be expressed by the Eq. 8.11
| \begin{equation} \text{d} \approx \frac{w}{4} + g \end{equation} | (8) |
d is the average diffusion length, w is the width of the microband, and g is the gap between the microband electrodes. It has been reported that the smaller d is, the higher the amplification factor (Fa) is. The designed d value was 2.50 µm for B-a-C-IDA with 2.0 µm. The observed d value was 2.47 µm.10 The obtained IDA structure is expected to have Fa higher than that of the designed structure.
The film thickness of the comb parts after micro-fabrication of B-a-C-IDA with 2.0 µm was 0.71 µm (Figs. 1e–1g).10 At IDA electrodes, when the thickness of IDA and the width of microband are extremely low (up to 0 µm), the non-linear diffusion of reactant at unit IDA area reaches the maximum value and amperometric current per unit area for redox species also reaches the maximum.12 However, the thinner the film thickness of the comb parts is, the higher the resistance of the electrode is. When the resistance of the electrode becomes higher, all areas of the comb parts cannot be used as an electrode. In order to confirm whether all surfaces of the honeycomb part of IDA acts as an electrochemical electrode, the double layer charging current at B-a-C-IDA in CV in 0.1 M nitric acid was compared to that of B-a-C-plate. The currents before and after processing IDA were almost the same. It is, therefore, confirmed that the film thickness of B-a-C-IDA is sufficient, and all areas of the comb part can be used as the electrode.10
As shown in Fig. 1e (the cross-sections of B-a-C-IDA with 2.0 µm), the sidewall of IDA was perpendicular to the surface of quartz substrate.10 If the shape of cross-section is trapezoidal or inverted trapezoidal, the average gaps between the collector and generator become larger, resulting in a decrease in Fa. The shape of the cross-section of B-a-C-IDA was confirmed to be ideal as IDA electrodes.10
3.2 Condition of surfaceFigures 1f and 1g show SEM images of top and side surfaces of B-a-C-IDA with 2.0 µm.10 The top surface of B-a-C-IDA was flat and smooth without any pits, similar to the surface of B-a-C-plate. The top surface was protected from O2 plasma etching by an aluminum layer. On the surface of the side wall of B-a-C-IDA with 2.0 µm (Fig. 1g), many grooves vertical to the substrate were found.10 The surface of sidewall was rougher due to O2 plasma etching. It is considered that O2 plasma exposure introduced a high density of oxygen-containing functional groups onto the sidewall surface.10 The change in surface termination seems to induce a significant increase in the double layer charging current.7
3.3 Chemical composition of B-a-C-IDAThe extreme increase in sp2 carbon content in conductive amorphous carbon (a-C) causes a decrease in the overpotential for H2 and O2 evolution, resulting in a narrower potential range in aqueous media. The decrease in sp3 carbon content of a-C degrades the conductivity of impurity-doped a-C because the concentration of dopant atoms bonded to sp3 carbon decreases; in other words, the carbon structure and composition of B-a-C cause changes in the conductivity and potential window in aqueous media and influence the performance of electrochemical sensing using B-a-C-IDA. Therefore, we evaluated the carbon compositions of B-a-C before and after the fabrication of IDA.
Figures 2a and 2b show Raman spectra of B-a-C-plate and B-a-C-IDA with 2.0 µm. In both Raman spectra, G- and D-peak were observed around 1,578 and 1,355 cm−1, respectively.10 The ratio of sp2/(sp2 + sp3) can be estimated from the peak positions.13,14 The ratios of sp2/(sp2 + sp3) before and after the fabrication of IDA were 0.752 and 0.755, which were almost equal.10 B-a-C was heated to approximately 100 °C during aluminum deposition on photoresist pattern. The sp2 carbon content of carbon material such as a-C:H have been reported to be increased by the heat treatment under vacuum.15 It has also been reported that the increase in sp2 carbon content degraded the resistance to electrochemically induced corrosion at the B-a-C electrode.7 However, the sp2 carbon content was equal before and after the IDA fabrication, which means the resistance to electrochemically induced corrosion did not change.10
Raman spectra of (a) B-a-C-plate and (b) B-a-C-IDA with 2.0 µm. Reprinted from Ref. 10. Copyrights for the figures belong to ECS, all rights reserved.
In Fig. 2, full widths at half maximum of G- and D-peaks before the IDA fabrication process (IDA processing) were 116 and 368 cm−1, respectively. The values at B-a-C-IDA with 2.0 µm were 115 and 362 cm−1, respectively.10 Since the values before and after IDA processing were almost the same, amorphous structure of B-a-C did not change at B-a-C-IDA with 2.0 µm. The sp2/sp3 ratios and chemical structure of B-a-C-IDA with 2.0 µm are same as those of B-a-C-plate, and the carrier (hole) mobility, conductivity, and reactivity toward H2 and O2 evolution are not expected to change drastically.10
3.4 Electrochemical properties of B-a-C-IDA 3.4.1 Double layer charging currents at B-a-C-IDABoron or Nitrogen-doped amorphous carbon electrodes show a double-layer charging current that is an order of magnitude lower than those of glassy carbon and poly-crystalline Pt electrodes.7,16 This property enables highly sensitive amperometric detection of electroactive compounds because the lower background (double layer charging) currents improve the signal to background (S/B) ratio. In this section, the double layer charging current at B-a-C-IDA electrode was examined to confirm the change of the double layer charging current by microfabrication.
Figure 3a shows the cyclic voltammogram at B-a-C-plate and B-a-C-IDA with 2.0 µm electrodes in 0.1 M nitric acid.10 The double layer charging current at B-a-C-plate was 0.6 µA cm−2 at 0.4 V vs. Ag|AgCl at a sweep rate of 50 mV s−1. The current at B-a-C-IDA with 2.0 µm was 2.2 µA cm−2, approximately four times higher than that of B-a-C-plate.10 The increase in the current seems to have been caused by two factors; the first factor is that oxygen-containing functional groups are introduced onto the surface of B-a-C-IDA sidewalls by O2 plasma exposure. The oxygen-containing functional groups, such as quinone group and carbonyl group, on the electrode surface can be ionized and increase pseudo-capacitance in low pH solution.17 The second factor is the increase in surface roughness caused by O2 plasma etching.
Cyclic voltammograms in 0.1 M HNO3 at B-a-C-IDA (dashed line) and B-a-C-plate (solid line). CV measurement for (a) the double layer charging current (at 50 mV s−1) and (b) the working potential range (at 100 mV s−1). Reprinted from Ref. 10. Copyrights for the figures belong to ECS, all rights reserved.
The dependence of the density of oxygen-containing functional groups and surface roughness on O2 plasma exposure (30 Pa, 50 W) on the surface of B-a-C-plate electrode was examined. When the exposure time was changed from 0 to 5 minutes, the surfaces were flat and smooth, but the double layer charging current increased from 0.64 µA cm−2 to 1.46 µA cm−2 at a sweep rate of 50 mV s−1 (increased by approximately double after 5 min O2 plasma exposure) (Fig. S1 of the Supporting Information). In order to clarify the cause of the increase in the double layer charging current, the morphological changes of B-a-C-plate surface before and after O2 plasma exposure were observed using an optical microscope, and changes in the chemical compositions were examined by XPS measurement. Figure S2 of the Supporting Information shows micrographs of B-a-C-plate surfaces before and after 5 min of O2 plasma exposure. The results of XPS measurement were summarized in Fig. S3 and Table S7 of the Supporting Information. No evidence of major microstructural damage was found. The surface features after 5 min of O2 plasma exposure are the same as those before exposure. However, the results of XPS measurement showed that the O/C ratio increase from 0.01 to 0.16, indicating that the increase in double layer charging current was caused by the oxygen-containing surface functionalities introduced by O2 plasma exposure. C1s peaks of XPS spectra and the analysis results of waveform separation are shown in Fig. S3 of the Supporting Information. After O2 plasma exposure, the intensities of the peak at 286.5 eV assigned to carbon singly bonded to oxygen (C–O), such as hydroxyl group (C–OH) and ether structure (C–O–C), and the peak at 289.0 eV assigned to carbon doubly bonded to oxygen (C=O), such as carboxyl group (C–OOH) and carbonyl group (C=O), increased. Therefore, the carboxyl group (C–OOH) that contributes to both peaks is supposed to be introduced at high density by O2 plasma exposure.
The pH dependence of the double layer capacitance (Cdl) was examined by AC impedance measurement. The Cdl value calculated from impedance data at 10 Hz presents the pseudo-capacitance caused by the capacitance component from ionized surface functionalities. The Cdl value at 250 Hz shows the capacitance value at space-charge layer in the semiconductor electrode.7 Both Cdl values (at 250 and 10 Hz) of B-a-C-plate without O2 plasma exposure show the same levels (2.8 and 8.7 µF cm−2, 2.7 and 6.5 µF cm−2) at pH 1.0 and 13.0. After 5 min of O2 plasma exposure, the Cdl at 250 Hz did not change (3.5 µF cm−2, 3.2 µF cm−2), but the Cdl at 10 Hz increased from 9.2 µF cm−2 to 14.5 µF cm−2 with increasing pH (1.0 to 13.0) (Table S8 of the Supporting Information). These results indicate that the density of the surface functionalities ionized under basic conditions increased after O2 plasma exposure, and these surface functionalities are supposed to be carboxyl groups (C–OOH). This suggests that the carboxyl group (C–OOH) is oxygen-containing surface functionalities introduced by O2 plasma exposure and contributes to enhance the double layer charging current. When the O2 plasma exposure time was 20 min, the double layer charging current reached 2.4 µA cm−2 due to the increases in the oxygen-containing surface functionalities (O/C ratios = 0.28) and in surface roughness (Microstructural damages were found in the optical micrograph in Fig. S2 of the Supporting Information).
From these results, the increase in the double layer charging current of B-a-C-IDA with 2.0 µm may be caused by the increases in oxygen-containing functional groups and roughness at sidewall surface. This is because O2 plasma etching was carried out at extremely higher r.f. power (150 W) and longer exposure time (15 min) in IDA processing.10
3.4.2 Potential window of B-a-C-IDAIn this section, the reactivity toward H2 and O2 evolution at B-a-C-IDA electrodes is investigated to confirm the change in the potential window in aqueous media due to IDA processing and verify the effectiveness of B-a-C-IDA against redox analytes with higher standard potentials.
Figure 3b shows the cyclic voltammograms (CV) in 0.1 M nitric acid solution (working potential range measurement) at B-a-C-plate and B-a-C-IDA with 2.0 µm electrodes.10 In this study, the working potential range where H2 and O2 evolution do not occur is defined as the potential region where the anodic and cathodic currents are less than 200 µA cm−2. With the IDA fabrication, the onset potential where H2 evolution current reached 200 µA cm−2 (Ered) shifted from −1.09 V to −1.50 V. The onset potential where O2 evolution current reached −200 µA cm−2 (Eox) shifted from 1.85 V to 1.91 V.10 It has been reported that the onset potential for H2 evolution shifts positively with increasing boron concentration in B-a-C-plate electrode. Boron has been reported to act as a catalyst for H2 evolution during water discharge.7 From the results of XPS measurement at B-a-C-plate electrodes (Table S7 of the Supporting Information), the atom% of boron decreased from 2.2 to 0.6 after 20 min of O2 plasma exposure. The onset potential for H2 evolution was shifted to negative potential side after 20 min of O2 plasma exposure, as shown in Fig. S4 of the Supporting Information. The negative shift in the onset potential for H2 evolution is probably caused by the decrease in the boron concentration at the surface with the increase in oxygen-containing functional groups due to O2 plasma exposure.
XPS measurements were performed to determine the chemical composition and boron atom concentration of the comb part of B-a-C-IDA after processing into interdigitated microarray electrodes (B-a-C-IDA with 2 µm). The atom% of boron could not be accurately estimated from XPS data because the spot size of X ray source (50 µm) was larger than the width of IDA (2 µm) and the intensity of B1s peak was extremely low. Although C1s and O1s peaks could be obtained in the observed XPS spectra (atom% of C = 83.0, atom% of O = 17.0), B1s peak was at the same level as noise.
To determine the dependence of boron concentration on the reactivity toward H2 and O2 evolution, the current-potential behavior in negative and positive potential regions was examined by a series of polarization experiments performed at the fixed potentials (Fig. S5 of the Supporting Information). Preliminary values of the exchange current densities (i0) and the transfer coefficients (α) were estimated and summarized in Table S9 of the Supporting Information. In this study, the overpotential η = 0 for H2 evolution was set at 0 V vs. SHE (the standard potential of 2H+ +2e−→H2), and η = 0 for O2 evolution was set at 1.23 V vs. SHE (the standard potential of O2 + 4H+ + 4e− → 2H2O). The α values for H2 evolution were in the same range (0.21–0.23) with or without O2 plasma exposure. However, after 5 min of O2 plasma exposure, i0 values decreased to 1/3 at B-a-C-plate electrode. These results suggest that the activation energy for H2 evolution was increased because of the decrease in boron concentration, which acts as a catalyst for H2 evolution. The α and i0 values for O2 evolution were in the same range (0.17–0.19 and 68–106 × 10−7 A cm−2) regardless of the surface boron concentration. Eox of O2 evolution did not change before and after IDA processing. This suggests that the conductivity of B-a-C-IDA was not drastically changed by IDA fabrication, even though the surface boron concentration was decreased. It indicates that B-a-C-IDA electrodes can sensitively detect redox species with standard potentials higher than that of O2 evolution, such as Ce3+/4+ and Co2+/3+. Consequently, micro patterning can be applied to B-a-C while maintaining intrinsic electrochemical properties of B-a-C.
3.4.3 Electrochemical properties of B-a-C-IDA with 2.0 µm as a micro-band electrodeFigure 4a shows CVs for 100 µM Ce3+ at B-a-C-IDA with 2.0 µm and B-a-C-plate electrodes.10 Well-defined redox peaks were observed at both electrodes at a sweep rate of 500 mV s−1. The values of peak separation at B-a-C-IDA with 2.0 µm and B-a-C-plate were 114 and 101 mV, respectively, indicating the reversible response.10 The currents at oxidation or reduction peaks at B-a-C-IDA with 2.0 µm were proportional to the square root of the sweep rate in the range from 10 mV s−1 to 500 mV s−1 (Fig. 4b). The kinetics of the redox reaction of Ce3+/4+ at both electrodes was a diffusion-controlled reaction process. However, the oxidation peak current at B-a-C-IDA with 2.0 µm was 133 µA cm−2 at a sweep rate of 500 mV s−1, which was 3.5 times larger than that at B-a-C-plate (37.8 µA cm−2). The higher amperometric current observed at B-a-C-IDA was caused by non-linear diffusion of reactants to the microband surfaces of B-a-C-IDA.10
(a), (d) Cyclic voltammograms in single mode measurement in 100 µM Ce(NO3)3·6H2O + 0.1 M HNO3 at B-a-C-IDA with 2 µm (solid line) and B-a-C-plate (dotted line). The sweep rate was (a) 500 mV s−1 and (d) 5 mV s−1. (b) The dependence of the peak current densities at B-a-C-IDA with 2 µm on sweep rate1/2. (c) Cyclic voltammograms in single mode measurement in 100 µM Ce(NO3)3·6H2O + 0.1 M HNO3 at commercial Pt-IDA. The sweep rate was 500 mV s−1. Figures (a) and (d) are reprinted from Ref. 10. Copyrights for the figures belong to ECS, all rights reserved.
Figure 4c shows CVs for 100 µM Ce3+ at commercial polycrystalline Pt-IDA electrode at a sweep rate of 500 mV s−1. Redox peaks could not be observed due to O2 evolution. These results clearly demonstrate the ability of B-a-C-IDA electrode to quantitatively analyze the concentrations of redox species whose standard potentials are higher than that of O2 evolution.
Furthermore, quasi-steady-state waves at B-a-C-IDA with 2.0 µm were observed in the potential region between 1.3 and 1.5 V at a sweep rate lower than 5 mV s−1 (Fig. 4d).10 Re-increase in the oxidation current at 1.5 V in Fig. 4d was caused by O2 evolution at B-a-C-IDA with 2.0 µm. The same oxidation current was observed in CV measured in 0.1 M nitric acid. The reactants were supplied to the B-a-C-plate electrode surface by linear diffusion from a bulk solution. In CV measurement at B-a-C-plate, a peak was formed because reaction currents in the higher potential region decreased due to the decrease in the concentration gradient of reactant species. Since the reactant species are supplied to the microelectrode surface by non-linear diffusion from three-dimensional directions (hemispherical diffusion), the amount of the reactant species supplied per unit electrode area at the microelectrode is higher than that at flat plate electrode. The steady-state wave is observed at relatively higher sweep rates at the microelectrode.10 It has been reported that, in theoretical analysis, the relation of amperometric current at microband electrodes to the width of the microband and the diffusion coefficient was described by the following Eqs. 9 and 10.18
| \begin{equation} I = nFCDb\left(0.439p + 0.713p^{0.108} + \frac{0.614p}{1 + 10.9p^{2}} \right) = nFcDbx \end{equation} | (9) |
| \begin{equation} p = \sqrt{\frac{nFw^{2}\nu }{RTD}} = \sqrt{( nF/RT )\nu } \times \frac{1}{\sqrt{Dw^{ - 2}} } \end{equation} | (10) |
b and w are the length and the width of the microband electrodes, respectively. n is the number of electrons transferred, F is Faraday constant, C is the bulk concentration of the reduced species, D is the diffusion coefficient, R is the gas constant, T is the temperature of the solution, and ν is the sweep rate. The value p is the ratio of the amount of reduced species required for a potential sweep (rate ν) at the electrode surface to the amount of reduced species supplied by diffusion in a unit area (width w). When p is less than 0.1 (supplied amount is fully sufficient), x in Eq. 9 is approximately 1. Quasi-steady-state behavior is, therefore, observed in the voltammogram because Eq. 9 is approximated by Eq. 11.
| \begin{equation} I_{\textit{lim}} \cong nFc^{*}Db \end{equation} | (11) |
D of Ce3+/4+ is 5.40 × 10−6 cm2 s−1 and w is 2.04 µm at B-a-C-IDA with 2.0 µm. Using these values, the sweep rate estimated from p = 0.1 and Eq. 10 is 33.3 mV s−1. The quasi-steady-state current at B-a-C-IDA was observed at the sweep rate of 5 mV s−1 in Fig. 4d. The sweep rate (5 mV s−1) is close to the theoretical value (33.3 mV s−1). This indicates that, at B-a-C-IDA with 2.0 µm, the effect of higher reactant supply with non-linear diffusion is available.10 The oxidation currents to Ce3+ at B-a-C-IDA with 2.0 µm and B-a-C-plate are 22.7 and 4.3 µA cm−2 (at 5 mV s−1), respectively. The current to Ce3+ at B-a-C-IDA with 2.0 µm is 5.3 times higher than that at B-a-C-plate. The effect of non-linear diffusion becomes significant as the sweep rate decreases.10
3.4.4 Amplification of amperometric currents at B-a-C-IDA with 2.0 µm in dual mode measurementIn single mode measurement, the reduction peak current could not be obtained at a sweep rate of 5 mV s−1 due to the effect of non-linear diffusion. Dual mode measurement at B-a-C-IDA with 2.0 µm was carried out at a sweep rate of 10 mV s−1, at which a reduction peak current was clearly observed in single mode. Amplification factor (Fa) in dual mode is calculated as the ratio of the steady-state current in dual mode to the peak current in single mode.
Figure 5 shows the voltammograms of Ce3+/4+ at B-a-C-IDA with 2.0 µm in single and dual mode measurement.10 The collector potential was set at 1.0 (Fig. 5a) or 1.7 V (Fig. 5b), and the generator potential was swept from 0.8 V to 1.7 V at a sweep rate of 10 mV s−1. The Fa value obtained in Fig. 5a was the ratio of the oxidation current at the generator when the applied potential was 1.7 V in dual mode to the oxidation peak current at 1.42 V in single mode. It was estimated to be 9.36. Fa in Fig. 5b is the ratio of the reduction current at the generator when the applied potential was 0.8 V in dual mode to the reduction peak current at 1.3 V in single mode. It was found to be 68.9.10 It has been reported that Fa value becomes higher with increasing overlap of diffusion layer between the generator and collector.8 Fa increases as the diffusion layer widens with increasing amount of a product at each electrode. The amount of the product (Ce4+) generated by the constant potential (1.7 V) oxidation was higher than that by the potential sweep oxidation (10 mV s−1 at 1.7 V). The difference in the amount of product caused a difference in the width of the area where the diffusion layers overlap, which resulted in a difference in Fa.10 Clear steady-state behavior was not obtained in dual mode when the applied potential of the collector was set at 1.0 V in Fig. 5a. In Figs. 4a and 4d, an oxidation current for O2 evolution was observed in the potential region over 1.5 V. The voltammograms in Fig. 5a did not show steady-state waves due to O2 evolution.10
Cyclic voltammograms in single and dual mode measurement in 100 µM Ce(NO3)3·6H2O + 0.1 M HNO3 at the generator in single mode (solid line) and the generator (dotted line) and collector (dashed line) in dual mode. The potential applied to the collector was (a) 1.0 or (b) 1.7 V. The sweep rate was 10 mV s−1. Reprinted from Ref. 10. Copyrights for the figures belong to ECS, all rights reserved.
In order to maximize Fa values for Ce3+/4+, the potential applied to the collector was optimized using B-a-C-IDA with 2.0 µm. Figure 6a shows the relation between the steady-state current at the generator (at 0.8 V, 10 mV s−1) and the applied potential of the collector in the range of 1.42 V to 1.8 V.10 When the potential of the collector raised from 1.42 V to 1.8 V, the current at the generator increased by a factor of 1.7. In Fig. 6a, the slope of the reduction current at the generator becomes gentler when the applied potential exceeds 1.7 V. At this time, the reduction current is close to steady-state value (ca. 246 µA cm−2) because the formation of product (Ce4+) tends to converge to a constant value.10 Accordingly, the potential applied to the collector was set to 1.7 V. The response current for Ce3+/4+ in dual mode reached maximum value and was 68.9 times higher than that in single mode due to redox cycling.
(a) The reduction current measured at the generator (at 0.8 V) vs. the applied potential at the collector in dual mode (at 10 mV s−1). (b) Fa with applied potential of (●) 1.7 V or (○) 1.0 V to the collector vs. the sweep rate in the range from 10 mV s−1 to 500 mV s−1. Fa values obtained at the constant potential are shown at 0 mV s−1. Reprinted from Ref. 10. Copyrights for the figures belong to ECS, all rights reserved.
Figure 6b shows the dependence of Fa values for 100 µM of Ce3+ at the generator in dual mode on the sweep rate at B-a-C-IDA with 2.0 µm.10 The potentials applied to the collector were 1.0 and 1.7 V.
When the sweep rate of 10 mV s−1 was change to the constant potential (Fa values obtained at the constant potential are shown at 0 mV s−1 in Fig. 6b), Fa increased from 68.9 to the maximum value of 174 (the generator set at 0.8 V) when the applied potential of the collector was 1.7 V.10 Because the product (Ce3+) and the thickness of the diffusion layer increase, Fa increases by a factor of 2.1.
When the potential of the collector was set at 1.0 V, Fa at 10 mV s−1 was found to be approximately 9.36, as shown in Fig. 6b. This Fa value is close to the value reported at Pt-IDA electrodes with width/gap = 2 µm/2 µm (Fa = ca. 14).11 The ratio of the oxidation current at the generator (the sweep rate 10 mV s−1) when the potential of the collector was set at 1.0 V to the peak current at B-a-C plate electrode (the sweep rate 10 mV s−1) was 49.2. The oxidation currents obtained at B-a-C-plate and B-a-C-IDA with 2.0 µm were 5.67 and 29.8 µA cm−2 at 10 mV s−1 in single mode. The value at B-a-C-IDA with 2.0 µm was 5.25 times higher than that at B-a-C-plate because of the effect of semi-spherical diffusion of reactants, as shown in Fig. 4d. The oxidation current obtained at B-a-C-IDA with 2.0 µm was 279 µAcm−2 in dual mode at 10 mV s−1 and was 9.36 times higher than that in single mode. Hence, the contribution of redox cycling is highest when amplifying amperometric current. The overlap of the diffusion layer can be increased by reducing the gap of B-a-C-IDA. Fa can be further improved by narrowing the gap of B-a-C-IDA to less than 2 µm.
3.5 Amplification factor when using various redox species at B-a-C-IDA 3.5.1 Relation between amplification factor and type of electrochemical reactionIt has been reported that Fa increases with widening the overlapped region in diffusion layers formed on the surfaces of the collector and generator in previous publication.19 The diffusion layer becomes thick in proportion to the increase in the concentration of oxidant (or reductant) generated by the electron transfer reaction on the generator (collector). Therefore, when the electron transfer rate is higher, the concentration of electrochemically generated species becomes higher, resulting in a thicker diffusion layer and higher Fa. Fa is also expected to depend on the diffusion coefficient and the type of electrochemical reactions (inner-sphere or outer-sphere redox reaction) because redox cycling occurs through the back-and-forth movement of redox species, induced by diffusion between the reaction planes of the generator and collector. Hence, by clarifying the relation of Fa to the type of electron transfer reaction, the electron transfer rate constant, and the diffusion coefficient, Fa values when using a redox analyte (Fa with a redox analyte) can be estimated without actual measurement. If Fa can be estimated, it will be possible to calculate Fa with a redox analyte of the radioactive materials such as UO22+/+, and to verify the effectiveness of B-a-C-IDA as an electrode for quantitative analysis of the radioactive material without conducting experiments using substances that are harmful to the human body.
In this study, Fa values at B-a-C-IDA having various gap values (20.0, 10.0, 5.0, and 2.0 µm) were examined using redox analytes whose types of the electrochemical reactions, electron transfer rate constants, and diffusion coefficients were different. Table 2 shows the diffusion coefficients (D) (at 298 K) and the electron transfer rate constants (k0) of redox analytes used in this study. The diffusion coefficients (D) in Table 2 were the values measured by CV measurement at B-a-C plate electrode. The electron transfer rate constants (k0) in Table 2 were the values measured by AC impedance measurements at B-a-C plate electrode. The measured D and k0 value were close to the reported values.20–24
| Redox species | Diffusion coefficient [×10−6 cm2 s−1] |
Electron transfer rate constant [×10−3 cm s−1] |
|
|---|---|---|---|
| Inner-sphere redox analytes |
Dopamine | 4.78 | 0.0400 |
| Methylcatechol | 6.00 | 0.110 | |
| Hydroquinone | 2.70 | 0.0200 | |
| Outer-sphere redox analytes | [Ru(NH3)6]2+/3+ | 5.50 | 359 |
| [Fe(CN)6]3−/4− | 6.30 | 5.66 | |
| Ce3+/4+ | 5.40 | 0.0495 | |
Dopamine, methylcatechol, and hydroquinone were used as inner-sphere redox analytes (inner-SPRA) that involve an adsorption process on the electrode surface during the electron transfer reaction. As outer-sphere redox analytes (outer-SPRA) without adsorption in the electron transfer reaction, [Ru(NH3)6]2+/3+, [Fe(CN)6]3−/4−, and Ce3+/4+ were used. Figure 7 shows the voltammograms of 1 mM [Fe(CN)6]3−/4−, 1 mM [Ru(NH3)6]2+/3+, 1 mM dopamine, and 1 mM methylcatechol in single and dual mode measurement at B-a-C-IDA with 2.0 µm. The potentials of the collector were set at −0.6 V vs. Ag|AgCl for [Fe(CN)6]3−/4−, −0.77 V for [Ru(NH3)6]2+/3+, 0.8 V for dopamine, 0.9 V for methylcatechol, and 0.8 V for hydroquinone. The potentials of the generator electrodes were swept at a rate of 10 mV s−1. Figure 8 shows the relation between Fa with all redox analytes calculated from dual mode (Fig. 7) and the gap values (20.0, 10.0, 5.0, and 2.0 µm).
Cyclic voltammograms in single and dual mode measurement in (a) 1 mM K3[Fe(CN)6] + 0.1 M Na2SO4 (b) 1 mM Cl3[Ru(NH3)6] + 0.1 M Na2SO4 (c) 1 mM Dopamine + 0.1 M Na2SO4 (d) 1 mM Methylcatechol + 0.1 M Na2SO4 at the generator in single mode (solid line) and the generator (dotted line) and the collector (dashed line) in dual mode. The potential applied to the collector was (a) −0.6 V, (b) −0.77 V, (c) 0.8 V, and (d) 0.9 V. The sweep rate was 10 mV s−1.
Fa values with redox analytes at each B-a-C-IDA with various gap values. (○) [Fe(CN)6]3+/4+, (□) [Ru(NH3)6]2+/3+, (△) Ce3+/4+, (×) Dopamine, (●) Methylcatechol and (+) Hydroquinone.
At B-a-C-IDA with the widest gap size of 20.0 µm, Fa value with inner-SPRA (dopamine) was approximately 4.2 and Fa with outer-SPRA ([Fe(CN)6]3−/4−) was approximately 6.5. Although Fa with outer-SPRA resulted in 1.5 times higher than that with inner-SPRA, Fa values at IDA having widest gap (20.0 µm) were at the same level, regardless of the inner- or outer-SPRA, as shown in Fig. 8. At B-a-C-IDA with the smallest gap size of 2.0 µm, Fa values with inner-SPRA (dopamine and methylcatechol) was 39.8–43.6, and Fa with outer-SPRA ([Fe(CN)6]3−/4− and [Ru(NH3)6]2+/3+) was 87.9–95.0. As shown in Fig. 8, Fa values with outer-SPRA ([Fe(CN)6]3−/4 and [Ru(NH3)6]2+/3+) were found to be approximately two times higher than those for the inner-SPRA (dopamine and methylcatechol). This large difference between Fa values with inner- and outer-SPRA suggests different mechanism of amplification of the response current.
3.5.2 Amplification mechanism of inner-sphere redox analytesInner-SPRA require an adsorption on the electrode surface during the electron transfer reaction, and the reaction occurs on the electrode surface. The distance of diffusion necessary for redox cycling is exactly equal to the gap of IDA. Therefore, reducing the gap of IDA is expected to have the same effect as shortening the distance of diffusion for inner-SPRA. This shortening enhances the redox cycling. From Fig. 8, Fa with inner-SPRA increases in inverse proportion to the decrease in the gap of IDA. It was confirmed that inner-SPRA were electrochemically oxidized (reduced) on the surface of the generator and collector, and, by diffusion, move back and forth between the generator and collector during redox cycling. Fa values with dopamine, methylcatechol, and hydroquinone at B-a-C-IDA with 2.0 µm are 38.9, 43.6, and 15.5, respectively. These Fa values are linearly related with diffusion coefficients (D), as shown in Fig. 9. The slopes and intercepts of the approximate straight line of Fa values at B-a-C-IDA with different gap values are shown in Table 3. These results clearly indicate that the amplification of response current for inner-SPRA by redox cycling at B-a-C-IDA strongly depends on the diffusion coefficient of redox analyte.
The dependence of Fa values with inner-sphere redox analytes on diffusion coefficients at B-a-C-IDA with various gap values. The gap sizes are (a) 2 µm, (b) 5 µm, (c) 10 µm, and (d) 20 µm.
| B-a-C-IDA electrodes | Parameters | |
|---|---|---|
| Gap (µm) | Slope | Intercept |
| 20 | 6.65 × 105 | 1.26 |
| 10 | 2.29 × 106 | −1.88 |
| 5 | 5.43 × 106 | −4.60 |
| 2 | 8.51 × 106 | −7.58 |
Figure 10 shows the results of single (cyclic voltammetry) and dual mode in 1 mM methyl viologen (MV) dichloride + 0.1 M HClO4 at B-a-C-IDA with 2.0 µm. Sweep rate was 100 mV s−1 in single and dual mode. Well-defined redox peaks (MV0/+) were observed at single mode (Fig. 10). The values of peak separation at B-a-C-IDA were 130 mV. In dual mode (potential of the collector electrode; −1.3 V vs. Ag|AgCl), the current at oxidation peak potential (−0.9 V vs. Ag|AgCl) at the generator was at the same level as that in single mode (not amplified) and the reduction current at the collector was maintained at a constant value (−1200 µA cm−2). Amplification of the response current was not observed.
Cyclic voltammograms in single and dual mode measurement in 1 mM methyl viologen (MV) dichloride + 0.1 M HClO4 at the generator in single mode (solid line). The sweep rate was 100 mV s−1. Cyclic voltammograms of the generator (dotted line) and the collector (dashed line) in dual mode measurement. The potential applied to the collector was −1.3 V. The sweep rate was 100 mV s−1.
A previous publication25 has reported that, after electrochemical reduction from MV+ to MV0, many monolayers of MVabs0 molecules were adsorbed on the electrode surface. After the formation of monolayer, the potential was swept toward higher potential side, MVabs0 was oxidized to MV+ and diffused to the electrolyte (MVabs0 → MV+aq + e−). However, monolayer of MVabs0 was held on the electrode surface unless the potential was swept toward the oxidation peak (−1.13 V vs. Ag|AgCl). The results of dual mode of MV at B-a-C-IDA (Fig. 10) suggest that the redox reaction of MV0/+ occurred at the generator and MV0 generated at the collector was adsorbed on the collector surface and did not diffuse to the generator. It was clarified that, for inner-SPRA of which reaction products were stuck on the electrode surface, the amplification of the response current at IDA was not obtained.
3.5.4 Amplification mechanism of outer-sphere redox analytes 3.5.4.1 Factors of controlling amplification at B-a-C-IDA with gap size of 10 µmOuter-SPRA do not require adsorption on the electrode surface during the electron transfer reaction. The reaction is thought to occur at a position (reaction plane or reaction zone) 1.0 nm away from the electrode surface toward the bulk solution, although it depends on the type of outer-SPRA.26 The distance of diffusion necessary for redox cycling is equal to the distance between the reaction plane of the generator and that of the collector. This distance is slightly shorter than the gap value. It results in enhancing the redox cycling. The diffusion coefficient seems to contribute significantly to the observed currents in this study. This is because the diffusion was enhanced by the spherical diffusion at micro-electrodes, and the observed currents were close to the limiting current density. When the outer-SPRA have a higher electron transfer rate constant (k0), the amount of product produced by redox reaction at the reaction plane will be higher than that of the outer-SPRA with lower k0. The concentration gradients of products for the redox analytes with higher k0 become higher than that for redox analytes with lower k0. According to Fick’s law of diffusion, the flux of products is proportional to the concentration gradient and the diffusion coefficients. The outer-SPRA with higher k0 value causes a higher flux. It is assumed that the higher flux leads to an increase in the concentration gradient in a very local area or the increase in the diffusion layers overlap, and, furthermore, to the higher observed current densities and higher Fa values. Thus, there is some possibility that k0 indirectly contributes to Fa values.
Hence, we assumed that the parameters controlling Fa with outer-SPRA at B-a-C-IDA are the diffusion coefficient and the electron transfer rate constant. Fa with outer-SPRA was examined with varying k0 and D values ([Ru(NH3)6]2+/3+, [Fe(CN)6]3−/4−, Ce3+/4+) at B-a-C-IDA with the widest gap (10 µm). Figure 11 shows the relation between the observed Fa values and k0 and D values of outer-SPRA examined at B-a-C-IDA with 10 µm. The observed Fa values increased linearly with increasing D values, but were not clearly dependent on the k0 values (Fig. 11b). The results show that, with outer-SPRA at B-a-C-IDA with the wider gap (10 µm), the diffusion coefficients (diffusion rate) mainly contribute to Fa, similar to inner-SPRA, whereas the electron-transfer rates have a very small contribution to Fa.
The dependence of Fa values at B-a-C-IDA with 10 µm on (a) the diffusion coefficients and (b) the electron transfer rate constants.
Figure 12A shows Fa values measured and estimated (Fa (D)) by considering the contribution of D values of [Fe(CN)6]3−/4− and [Ru(NH3)6]2+/3+ at B-a-C-IDA with various gap values. Fa (D) values were calculated by substituting D values of [Fe(CN)6]3−/4− and [Ru(NH3)6]2+/3+ (6.30 × 10−6 cm2 s−1 and 5.50 × 10−6 cm2 s−1, respectively) into the equations of the approximate straight line in Fig. 9 (the relation between D and Fa at certain gap values) and Table 3. As shown in Fig. 12A, the observed Fa values are relatively higher than the estimated Fa (D) values. For [Fe(CN)6]3−/4− redox analyte, the observed Fa at B-a-C-IDA with gap size of 2.0 µm was 1.9 (87.9/46.0) times higher than the estimated. For [Ru(NH3)6]2+/3+, the observed Fa value was 2.4 (95.0/39.2) times higher than the estimated.
(A) The comparison of the measured Fa values and the estimated Fa(D) values for (a, b) [Fe(CN)6]3−/4− and (c, d) [Ru(NH3)6]2+/3+ at B-a-C-IDA with various gaps. (B) The dependence of Xg with various gap values on the electron transfer rate constants at B-a-C-IDA with (a) 2 µm, (b) 5 µm, and (c) 10 µm.
Figure 12B shows the relation of Xg with the electron transfer rate constant (k0), where Xg is defined as the ratios of the measured Fa to the estimated Fa (D). Xg values at B-a-C-IDA with certain gaps were proportional to the k0 value. The higher the k0 values of the redox analyte, the higher the Xg values tended to be. Table 4 shows the slopes and intercepts of the approximate straight line of Xg values at B-a-C-IDA with different gap values. In the case of outer-SPRA, the redox cycling (Fa) was Xg times higher than Fa (D) due to the effect of electron transfer rate.
| B-a-C-IDA electrodes | Parameters | |
|---|---|---|
| Gap (µm) | Slope | Intercept |
| 10 | 0.0423 | 1.22 |
| 5 | 0.0498 | 1.10 |
| 2 | 0.0706 | 1.95 |
The use of radioactive materials in electroanalytical research requires prior approval by Nuclear Regulation Authority Japan, a government agency. It is difficult to actually use the radioactive materials in the electroanalytical measurement using B-a-C-IDA fabricated in this study, and it is unclear whether the traces of radioactive materials can be quantitatively analyzed using B-a-C-IDA with gap size of 2.0 µm. We attempted to devise a simple method to estimate Fa value using the reaction parameter (type of reaction, electron transfer rate constant, and diffusion coefficient) in order to determine the geometric structure of B-a-C-IDA applicable to target redox species. The target redox species was UO22+/+, which consists of the radioactive material (238U). The possibility of quantitative electroanalysis of UO22+/+ at B-a-C-IDA was examined by estimating Fa values of Ce3+/4+, model analyte. Ce3+/4+ redox analyte has electrochemical reaction parameters (Table 2) that are similar to those of UO22+/+ (Table 5).27,28 Ce3+/4+ redox analyte has a higher standard potential (1.72 vs. SHE) and cannot be detected at conventional electrodes (e.g., polycrystalline Pt) as is the case for UO22+/+.
| Redox species | Diffusion coefficient [×106 cm2 s−1] |
Electron transfer rate constant [×10−3 cm s−1] |
|
|---|---|---|---|
| Outer-sphere redox analytes |
UO22+/+ | 3.41 | 2.74 |
First, we predicted Fa (D) values of Ce3+/4+ when only the diffusion coefficient contributed. Fa (D) values at B-a-C-IDA with different gap values were calculated by substituting D values (5.40 × 10−6 cm2 s−1) into linear approximate equation (Fa (D) = 8.51 × 106 × D − 7.58 at the gap of 2 µm) in Fig. 9 and Table 3 (the relation between D and Fa (D) at certain gap values). Since Ce3+/4+ is outer-SPRA, the effect of electron transfer rate constants (k0) should also be considered. Second, taking advantage of the proportional relationship between the electron transfer rate constant (k0) and Xg values as shown in Fig. 12B, the contribution of the electron transfer rate was introduced into the estimated Fa values (Fa (D, k0)) by multiplying Fa (D) by Xg. Xg values for certain redox analyte such as Ce3+/4+ were calculated by substituting k0 values (Ce3+/4+; 0.0495 × 10−3 cm s−1) into linear approximate equation (Xg = 0.0706 × ln(k0) + 1.95 at the gap of 2 µm) in Fig. 12B and Table 4 (the relation between Xg and k0 at certain gap values). The estimated Fa (D, k0) values of Ce3+/4+ and UO22+/+ were compared with the observed Fa values of Ce3+/4+ and shown in Fig. 13. The difference between the estimated Fa (D, k0) values of Ce3+/4+ and the observed values was quite small (<7.5 % of the observed values).
The comparison of (a) ○ the measured Fa values, (b) △ the Fa(D, k0) values for Ce3+/4+, and (c) □ the Fa(D, k0) values for UO22+/+.
The prediction method proposed in this study (using type of reaction, electron transfer rate constant and diffusion coefficient) was proven to be suitable for the evaluation of Fa value with redox analytes that are difficult to measure. The estimated Fa (D, k0) values of UO22+ at B-a-C-IDA with gap size of 2.0 µm were calculated to be 43.2 using D = 3.41 × 10−6 cm2 s−1 and k0 = 2.74 × 10−3 cm s−1, as shown in Fig. 13. It has been reported that the current density of 630 µA cm−2 was observed at the oxidation peak for 12.5 mM UO22+ on the boron-doped diamond electrode surface (sweep rate 10 mV s−1).27 The current density observed at B-a-C-IDA with gap size of 2 µm is expected to be 43.2 times of 630 µA cm−2 (27.2 mA cm−2). Since the background current density at B-a-C-IDA with 2.0 µm was N = 0.45 µA cm−2 at a sweep rate of 10 mV s−1, the current density of the response signal at the detection limit (S/N = 3) is 1.35 µA cm−2. Assuming that the current density is linear-proportional to UO22+ concentration, the UO22+ concentration corresponding to the detection limit (S/N = 3) is roughly estimated to be 0.62 µM. Therefore, B-a-C-IDA with 2.0 µm has a potential to quantitatively electroanalyze the concentration of UO22+ within the target range of 0.3 µM to 1637 µM.
3.5.6 Calibration curve of Ce3+The performance of B-a-C-IDA with 2.0 µm as an electrochemical sensor for Ce3+ analysis was evaluated. Figure 14 shows the plots of the response currents versus Ce3+ ion concentrations (calibration curve) in dual mode when the potentials of the collector and the generator were set at 1.7 and 0.8 V.10 The chronoamperograms in the dual mode measurements were shown in Fig. S6 of the Supporting Information. The observed current densities used in Fig. 14 were the average values of the measured currents during from 289 to 300 seconds after the start of chronoamperometry (CA) measurement. In all CA measurement for calibration curves, the maximum value of the observed currents varied from the average current densities during 2 seconds was 0.3 µA cm−2. The maximum value is added to Fig. 14 as the error bars.
Calibration curve for (a) the reduction current at 0.8 V and (b) the oxidation current at 1.7 V in CA measurement at dual mode vs. Ce3+ concentrations. Reprinted from Ref. 10. Copyrights for the figures belong to ECS, all rights reserved.
In Fig. 14a, the reduction current at the generator shows linear relation with Ce3+ concentrations in the range from 100 µM to 0.1 µM (linear dynamic range). The theoretical detection limit was approximately 0.13 µM at S/N = 3. In contrast, the linear dynamic range of Ce3+ concentrations was from 100 µM to 25 µM in the calibration curve (Fig. 14b) of the oxidation current at the collector. The theoretical detection limit was approximately 63 µM (S/N = 3).10 In Fig. 5a, the currents attributed to O2 evolution was observed at 1.7 V in single mode. Furthermore, in CV measurement with 0.1 M nitric acid solution, the current for O2 evolution was also observed at 1.7 V, and was 97.2 µA cm−2 (in Fig. 3b). The currents observed from 25 µM to 0 µM in Fig. 14b were close to the O2 evolution current (101.3–74.4 µA cm−2). Hence, the higher detection limit (low sensitivity) at the collector may be due to the interface of O2 evolution current.10 These results indicate that, by fabricating B-a-C-IDA with 2.0 µm, it is possible to realize a high sensitive electrochemical system, which has a higher overpotential than O2 and H2 evolution. The detection limit of the system for redox species is lower than that of B-a-C-plate (1/100).10
It is not clear why the observed currents are proportional to the concentration of redox analyte, but the structure of IDA and the diffusion layer formed on IDA seemingly contribute to the experimental result. IDA used in this study was composed of an array of cuboid electrodes with a width of 2 µm, a height of 0.71 µm and a length of 200 µm at 2 µm interval. One of the adjacent cuboid electrode pairs is the generator electrode and the other is the collector electrode. The facing side walls of the adjacent cuboid electrodes were a pair of electrodes, opposing to each other. In the space between the side walls, redox cycling occurs and the overlap of diffusion layers also forms. Meanwhile, the top surfaces of the cuboid electrodes were immersed in the electrolyte including redox analyte and did not face counter electrodes. The diffusion layer is so formed as to cover the entire area of IDA electrode. It is composed by alternating serial connections of the semi-infinite diffusion layer formed on the top surface of the cuboid electrodes and the diffusion layer formed between the side walls of a pair electrodes. When a constant potential is applied to a pair of the generator and collector to steady state, the diffusion layer between the side walls of a pair electrodes is considered to be influenced by the concentration of redox analyte in electrolyte through the semi-infinite diffusion layer formed on the top surface of the cuboid IDA electrodes. One of the reasons that the observed current depends on the concentration of redox analyte is assumed to be derived from the structure of diffusion layer formed on the IDA electrodes.
The structure of B-a-C-IDA with 2 µm after the CA measurements for the calibration curves of Ce3+/4+ (Fig. 14) was examined by SEM. The IDA structure after CA measurements was identical to that before measurements, without peeling off of IDA from the substrate or damaging. B-a-C-IDA electrode showed higher durability in electrochemical measurements.
Microanalytical method for the redox analyte (such as Ce3+/4+) with a standard potential higher (or lower) than O2 (or H2) evolution was established by realizing B-a-C-IDA electrodes. The amplification of the oxidation current derived from redox cycling was observed at the generator electrode of B-a-C-IDA in the measurement with applying the reduction potential to the collector electrode (dual mode measurement). The overlap of diffusion layer could be enhanced by optimizing the potentials applied to the collector and generator electrodes. As a result, Fa could be improved up to 180.
Fa values at B-a-C-IDA with various gap values were examined using redox analytes with different types of electrochemical reactions, diffusion coefficients, and electron transfer rate constants. By clarifying the mechanism of amplification at B-a-C-IDA electrodes and the contribution of the diffusion coefficients and electron transfer rate constants to Fa, we were able to establish the method to estimate Fa value with redox analytes using the diffusion coefficients and electron transfer rate constants. Fa with redox analytes harmful to human body at B-a-C-IDA can be estimated without actual measurement. For UO22+, Fa at B-a-C-IDA with gap size of 2.0 µm is estimated to be 43.2 when calculated using D = 3.41 × 10−6 cm2 s−1 and k0 = 2.74 × 10−3 cm s−1. The detection limit (S/N = 3) of UO22+ is roughly estimated to be 0.62 µM.
There are two points to be improved to achieve higher Fa; one is to reduce the thickness of the comb parts at B-a-C-IDA. At microband electrodes, non-linear diffusion of reaction per unit electrode area reaches a maximum value when the thickness of electrode is extremely thin compared to the width of microband. Amperometric current per unit area for redox species also reaches a maximum. The overlap of diffusion layers can be enhanced by increasing the amperometric current. It results in enhancing redox cycling and increasing Fa value. The thickness of B-a-C-IDA that allows the entire surface to be used as the electrode was 0.71 µm. Therefore, we are developing B-a-C with lower volume resistance to realize B-a-C-IDA sensor with reduced thickness. The other is to reduce the distance (gap) between the generator and collector electrodes. It is difficult to fabricate B-a-C-IDA with the gap narrower than 2.0 µm by micro patterning using photolithography method. Research to fabricate B-a-C-IDA with submicron gap using electron beam lithography is now in progress.
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.28148081.
Kensuke Honda: Conceptualization (Lead), Data curation (Lead), Formal analysis (Lead), Funding acquisition (Lead), Investigation (Lead), Writing – original draft (Lead)
Shinpei Ohtomo: Investigation (Supporting)
The authors declare no conflict of interest in the manuscript.
A part of this paper has been presented in Pacific Rim Meeting on Electrochemical and Solid-state Science (PRiME) 2016 (Presentation #M01-3787) and reported in ECS Trans., 75, 217 (2016).
K. Honda: ECSJ Active Member
