2019 年 60 巻 4 号 p. 531-537
In this study, pH measurement was performed in a thin electrolyte droplet with a thickness <1000 µm by the measurement of the equilibrium electrode potential of an Sb/SbxOy electrode used as a pH sensor. The equilibrium potential of the Sb/SbxOy electrode was evaluated by using the Kelvin probe (KP) technique. To investigate the potential response of the Sb electrode in a thin electrolyte droplet, the dependency of the Volta potential difference between the Sb and a gold wire as a KP on electrolyte droplet thickness was measured. The Volta potential difference had a linear response with respect to the buffer solution pH, independent of the droplet thickness. This result indicates that the KP technique, combined with an Sb electrode, is sensitive to the pH of a thin electrolyte droplet of thickness ≥50 µm. This pH measurement technique was also applied to measure pH in a corrosion model of steel. The corrosion model consisted of two steel plates in the same plane as the anode and cathode, with a constant current between them. During the corrosion process, the pH value decreased from 6 to 5 near the anode and increased from 6 to 12 at the cathode. The changes in pH measured in the thin electrolyte droplet were in good agreement with the color changes of the solution containing pH indicators.
Fig. 7 Changes in (a) the applied current density between the steels, (b) the potentials of the anode and cathode sides of the steel, and (c) the pH near the steel.
When steel is used in atmospheric environments, a water film forms on its surface due to rain and/or dew condensation, and then a corrosion reaction occurs. If the condensation droplets have a neutral pH, the steel dissolves producing Fe2+ ions, and the dissolved oxygen and water molecules are reduced resulting in OH− ion formation. Additionally, the dissolved Fe2+ ions are oxidized to Fe3+ ions in the droplet containing dissolved oxygen, and the Fe3+ ions interact with the OH− ions or water molecules. Consequently, these reactions may lead to localized changes in the pH on the surface of the corroding steel.1)
Whitman et al. investigated the effect of pH on the corrosion rate of steel;2) they clarified that a decrease in pH accelerates the corrosion rate of the steel, whereas an increase in pH moderates the corrosion process. In atmospheric corrosive environments, one can expect significant changes in pH during corrosion, as the electrolyte solution volume is extremely small. Therefore, it is important to know the pH in a thin electrolyte solution during atmospheric corrosion of metallic materials such as steel.
With regard to pH measurements during the corrosion process of metallic materials, numerous studies have been performed on bulk electrolyte solutions with glass-type pH electrodes. In contrast, few studies have examined pH changes in thin electrolyte solutions. This may be related to the difficulty on the setup of the glass-type pH electrodes into the thin electrolyte solutions. Other probes or techniques such as fluorescence probes,3,4) scanning chemical microscopy (SCHEM),5–7) and metal/metal oxide electrodes8,9) have been applied to pH measurements of thin electrolyte solutions. However, these techniques are unable to resolve the difficult on the pH measurement in the thin electrolyte solutions. Thus, a different approach is required to measure the pH of these solutions.
Some metal/metal oxide electrodes, such as antimony (Sb)10–14) tungsten (W),8,15–17) and iridium (Ir),9,18) show equilibrium electrode potential that depends on the pH of the aqueous solution. Therefore, these metals can be used as pH sensors; the corresponding mechanism is described by the Nernst equation. In addition, the electrode potential of some metallic electrodes can be expected to respond to pH even when the electrolyte layer is thin. The Kelvin probe (KP) technique uses a contactless reference electrode,19–21) this technique should be useful for the pH measurement in a thin electrolyte layer because it does not change the thickness of the electrolyte layer nor the solution chemistry. Therefore, we hypothesized that the KP technique could be applied to pH measurements of thin electrolyte solutions during steel corrosion by measuring the Volta potential difference between a metal/metal oxide electrode and a KP.
In this study, we investigated whether the KP technique could be combined with an Sb electrode for pH measurements of a thin electrolyte droplet. Subsequently, this system was used to estimate the pH near a simulated anode and cathode during the steel corrosion process in a thin electrolyte droplet.
The material used as the pH-sensing electrode was a 5 mm × 5 mm Sb plate, with a thickness of 1 mm (99.999%, Kojundo Chemical Laboratory Co., Ltd., Japan). Figure 1 shows a drawing of the sample of the Sb electrode, which was embedded in an epoxy resin after being connected to a lead wire. The sample was ground up to P1200 grit with waterproof abrasive papers and then degreased ultrasonically in ethanol. The surface area exposed to the aqueous solution was held constant at 0.20 cm2, and the sample surface, except for the exposed area, was covered with Teflon tape (NITOFLON No. 903UL., Nitto Denko Corporation, Japan).
Schematic drawing of the sample used for pH measurements in a thin electrolyte droplet.
Figure 2(a) shows a schematic drawing of the potential measurement system of the Sb electrode. The sample was set horizontally below a KP. The KP was a 1-mm-diameter gold (Au) wire (99.5%, Nilaco Corporation, Japan) that was vibrated, along with a bimorph-type piezoactuator (LPD3713X, Nihon Ceratec Co., Ltd., Japan). In this study, a small alternating current (AC) voltage signal from a lock-in amplifier (5610B, NF Corporation, Japan) was amplified by a piezodriver (As-904-150B, NF Corporation, Japan). The amplified AC voltage signal was input to the piezoactuator to vibrate the KP. The vibration frequency was 227 Hz, and the vibration amplitude was estimated to be 50 µm, based on the characteristics of the piezoactuator.
Schematic drawing of the experimental setup for (a) the potential measurement of an Sb electrode and (b) the measurement of electrolyte thickness.
In this study, the Volta potential difference between the Sb electrode and the KP was measured using a parasitic capacitance method,22,23) which is an improved null method. The AC current flowing in the circuit between the Sb electrode and the KP was measured with a low-current amplifier (T-IVA001H, Turtle Industry Co., Ltd., Japan). The amplitude of the AC current was calculated against the direct current (DC) bias voltage between the sample and the KP. The DC bias voltages were applied to the circuit with a data acquisition (DAQ) device (USB-6215, National Instruments Corporation, USA). A personal computer with a custom-designed program written in LabVIEW (National Instruments Corporation, USA) was used to control the motion of a z-stage (MM-60Z, Chuo Precision Industrial Co., Ltd., Japan), to operate the DAQ device and record the AC signal outputs from the lock-in amplifier.
Figure 2(b) shows a schematic diagram of the experimental setup to measure the height of the solution (i.e., droplet thickness) on the sample. This setup was similar to that reported by Stratmann24) and Nishikata:25) the thickness of the droplet was determined by measuring the potential difference between a 0.5-mm-diamater nickel (Ni) needle (99.99%, Nilaco Corporation, Japan) and the sample. The height of the Ni needle was controlled with a manual z-stage, and the potential difference was monitored with a digital voltmeter (2000 Multimeter, Keithley Instruments, Inc., USA). When the Ni needle was positioned above the thin electrolyte droplet on the sample and gradually lowered, a certain potential difference was measured when the tip of the needle touched the surface of the droplet. When the needle was lowered further and the tip of the needle touched the sample surface, the potential difference had a value of zero. Therefore, the thickness of the thin electrolyte droplet was measured from the difference in the position of the Ni needle; the measurement accuracy of this procedure was <10 µm.25)
The sample was placed in contact with commercial buffer solutions: potassium hydrogen phthalate (pH 4.01); a mixture of potassium dihydrogen phosphate and disodium hydrogen phosphate (pH 6.86); sodium tetraborate (pH 9.18); and a mixture of sodium hydrogen carbonate and sodium carbonate (pH 10.01) (Kanto Chemical Co., Inc., Japan). The volume of the solutions was 15 µL, which corresponds to an electrolyte droplet thickness of 1000 µm. The Volta potential difference measurement of the Sb electrode and the determination of the droplet thickness on the Sb electrode, as mentioned above, were carried out during the drying of the electrolyte droplet of various pH solutions. All measurements were conducted under an ambient laboratory atmosphere of 298 K and a relative humidity of 40–60%.
2.2 Measurement of pH during steel corrosion 2.2.1 Sample preparationCarbon steel sheet, whose chemical composition is shown in Table 1, was used as the corrosion material. The steel sheet was cut into small coupons: (10 mmL × 10 mmW × 5 mmT) and (10 mmL × 5 mmW × 5 mmT); a 1-mm-diameter hole was drilled in the center of each coupon. The coupon was then coated with electrodeposited paint. The paint thickness was about 20 µm. Sb wire was prepared by melting and extending Sb shot (1–3 mmϕ, 6N, NewMet Koch, UK) in a heated borosilicate glass tube. The diameter of the Sb wire extended was ca. 0.8 mm.
As shown in Fig. 3, a pair of steel plates and a pair of Sb wires were embedded in an epoxy resin together. The gap size between the steel plates was about 0.3 mm. The Sb wires were placed at the center of the drilled-hole machined in the steels. The Sb wires were electrically insulated from the steels. Samples were ground up to P1200 grit with waterproof SiC papers and then degreased in ethanol before the experiments. The exposed surface area was fixed at 3.8 cm2, and the other surface was covered with a tape. The ratio of the surface area of steels was ca. 1:2.
Schematic drawing of the sample used as a simulated anode and cathode system of steel corrosion by a thin electrolyte droplet.
Figure 4 shows the experimental setup for the pH measurement of steel corrosion. The as-prepared sample was placed just below the KP and a 380-µL droplet of 0.1-mol/dm3 NaCl solution, which was prepared with analytical grade chemicals (Kanto Chemical Co., Inc., Japan) and Milli-Q water (18 MΩ cm), was put on the sample surface. The initial droplet thickness was ∼1000 µm.
Schematic drawing of the experimental setup for pH measurements during the galvanic corrosion test for simulated steel corrosion.
After droplet formation, a constant current was applied between the steel plates, such that the steel shown on the right side of Fig. 3 worked as an anode and that on the left as a cathode. The actual values of the current densities applied are listed in Table 2. For the initial 30 min after droplet formation, current was not applied to the sample, providing free-corrosion conditions for the steel surfaces. After the initial 30-min period, a constant current was applied and increased step-wise every 15 min.
During the free corrosion and polarization conditions, the Volta potential difference between the steel of the anodic side and the KP was measured using the same system as described in Section 2.1. For these measurements, the Au wire vibrated at a frequency of 237 Hz and an amplitude of 25 µm. The potential difference between the anodic and cathodic steel plates was measured simultaneously. The pH values of the electrolyte solution at the anode and cathode were calculated using the Volta potential difference of the Sb wire. A multi-channel potentiostat/galvanostat (PS-08, Toho Technical Research Co., Ltd., Japan) was used to apply galvanostatic polarization and to measure the potential difference between the electrodes. The analog signal outputs from the potentiostat/galvanostat were measured with a digital voltmeter and recorded by a personal computer controlled by a custom-designed LabVIEW program.
To correlate Volta potential difference measured by the KP with the corrosion potential of the steel, the Volta potential difference and the corrosion potential for various metals under a droplet of 1.0-mol/dm3 Na2SO4 solution (thickness: 1 mm) were measured with the KP and a commercial Ag/AgCl electrode in a saturated KCl solution (SSE; ESSE = +0.197 V vs. SHE at 25°C), respectively.19) The Volta potential difference measured against the KP was then converted to the corrosion potential measured with the SSE.
Figure 5 shows the Volta potential difference of the Sb electrode measured in a droplet of various pH standard solutions as a function of the droplet thickness. The Volta potential difference in each pH standard solution was independent of the droplet thickness in the range of 50–1000 µm. Additionally, the Volta potential difference of the Sb electrode against the KP decreases as the droplet’s pH increased.
Volta potential difference of the Sb electrode against the KP in various pH buffer solutions for different droplet thicknesses.
The average value of the Volta potential difference from each pH standard solution was obtained. The Volta potential difference was independent of the droplet thickness in each pH standard solution. Figure 6 shows the average value of the Volta potential difference of the Sb electrode as a function of the pH of the standard solutions. As shown in the figure, the average Volta potential difference has a linear relationship with respect to the pH. The relationship between the Volta potential difference, E, and pH can be expressed using a least squares approximation as follows:
| \begin{equation} \text{$E$/V vs. KP} = -0.031\text{pH}-0.219 \end{equation} | (1) |
| \begin{equation} \text{$x$Sb} + \text{$y$H$_{2}$O}\rightleftarrows\text{Sb$_{x}$O$_{y}$} + \text{$2y$e$^{-}$} + \text{$2y$H$^{+}$} \end{equation} | (2) |
| \begin{equation} E = E^{\circ}-2.303\frac{RT}{F}\text{pH} \end{equation} | (3) |
Relationship between the Volta potential difference of the Sb electrode and droplet pH.
As described in eq. (1), the slope of the Volta potential difference was −31 mV/pH in this study. The slope value was smaller than that for the theoretical value obtained using the Nernst’s equation. However, the Volta potential difference for the Sb electrode used in this study indicated good linearity as a function of pH, i.e., a sub-Nernstian response. Other researchers have reported that some pH-sensing metal/metal oxide electrodes show sub-Nernstian responses with smaller slopes, e.g. −35 to −45 mV/pH for W,8) −42 mV/pH for Sb,10) and −55 mV/pH for Ir.14) Some have attempted to explain this effect in terms of the interference by the reduction reaction of dissolved oxygen,13–15) and the effect of the surface area.16,17) Although there was a slight difference in the slope between the obtained value and the theoretical value, as shown in Fig. 3, the Sb electrode demonstrated good sensitivity as a pH sensor in solutions with thicknesses ≥50 µm. Furthermore, it can be concluded that solution pH can be obtained using the Sb electrode/KP configuration.
3.2 pH estimation on a corroding steel in an NaCl dropletThe KP technique with an Sb electrode as a pH sensor was applied to measure local pH on a corroding steel surface in a thin electrolyte layer. As shown in Fig. 4, current was applied between two steels to mimic anodic and cathodic regions. Figure 7 shows changes in the current density, the potential of each steel, and pH evaluated from the Volta potential difference of the Sb electrode during the simulated steel corrosion. As shown in the figure, a 380-µL droplet of 0.1-mol/dm3 NaCl solution was placed on the sample’s surface at 0 min. For the initial 30 min, both the anode and cathode were under the free corrosion conditions, meaning that no current was applied between the anode and cathode. The corrosion potentials of the anode and cathode were monitored during the steel corrosion under a thin electrolyte droplet. Simultaneously, the pH values near each steel plate were measured with the Sb electrode. As shown in the figure, the corrosion potentials for the anode and cathode were −0.4 V vs. SSE when a droplet was placed on the sample. The corrosion potential for both steels gradually decreased to ca. −0.5 V vs. SSE. At this time, the solution pH near the Sb electrodes was estimated to be in the neutral region and remained nearly constant over the immersion time.
Changes in (a) the applied current density between the steels, (b) the potentials of the anode and cathode sides of the steel, and (c) the pH near the steel.
After 30 min of immersion, a constant current was applied between the anode and cathode. The current increased step-wise every 15 min. The potential of the steel at the anode reached a constant reading of −0.5 V vs. SSE, independent of the applied current density. The pH near the anode gradually decreased to ca. 5 after 150 min of immersion. In contrast, as the applied current increased, the cathode potential shifted gradually in the negative direction. When the applied current density exceeded −20 µA/cm2, the cathode potential shifted significantly in the negative direction from ca. −0.6 V. The negative potential shifts from −0.6 V is related to the polarization behavior on the cathode (the cathodic polarization curve will appear later in Fig. 9). The diffusion-limiting current of the reduction reaction of dissolved oxygen (ORR) was observed in the potential region less noble than −0.6 V; in the diffusion-limiting current region of ORR the cathodic potential can shift significantly in the negative direction even with a small increase in the applied current. Regarding the pH near the steel of the cathode, when the current was applied to the cathode, the pH began to increase and continued to increase with increasing the applied current. The pH reached a constant value at ca. 12 when the applied current density reached ca. −20 µA/cm2; this current was almost the same value as the diffusion-limiting current density of the ORR. The pH did not change very much with increasing the applied current; this is because the ORR occurs under the diffusion-limiting condition in the potential region less noble than ca. −0.9 V.
Our results showed that the pH measurement decreased to ca. 5 near the steel of the anode and increased to ca. 12 near the steel at the cathode during the steel corrosion process. To validate the pH measurement, a galvanostatic polarization test was performed in a droplet of 0.1-mol/dm3 NaCl containing a pH indicator. The pH was estimated from the change in color of the solution during the galvanostatic polarization. The pH indicator used in this test was Takemura’s pH indicator (Takemura Denki Seisakusho, Japan); this indicator can be used to resolve pH in the range of 4 to 10.
Figure 8 shows the change in the applied current density, surface images of the steel, and solution color. The steel shown on the right acted as the anode and on the left as the cathode. In the initial 30 min, the pH changes that occurred during this time period were caused by free corrosion of the steel, as current was not applied. Figure 8(a) shows the surface image of the steel plates taken at 30 min after droplet formation; pits were evident in the steel surfaces, as indicated by the arrows. This suggests that localized corrosion occurred on the steel plates in the droplet, with the pits acting as anodic sites and the other surfaces as cathodic sites.26) Additionally, the pH distribution in the droplet vicinity of the pits showed a pH of 5.5 to 7. At some distance away from the pits, the pH was over 10. Thus, the pH distribution was attributed to localized corrosion on the steel. This result is consistent with the findings of Hirohata et al., who measured the local pH changes from steel corrosion under a modified droplet.27) According to the results of pH measurement in their study, the pH during the steel corrosion was estimated to be in the range of 6 to 13.
Applied current density on the anode and cathode sides of steel during galvanostatic polarization and photographs of the samples in the solution taken after (a) 30, (b) 60, (c) 120, and (d) 150 min.
The galvanostatic polarization was conducted by applying a constant current density between the steels, such that the anode and cathode could be separated. As shown in Figs. 8(b)–(d), the low pH region expanded over time on the anode. The pH distribution around the pits (Figs. 8(a) and (b)) disappeared with increasing the current density since anodic reaction was enhanced on the whole surface of the anode. In the final stage of galvanostatic polarization, the pH over the whole surface of the anode was estimated as 5.5 at point (d) in Fig. 8 based on the change in the solution color, as the applied current density approached 105.7 µA/cm2. In contrast, the pH near the anode, which was evaluated by the measurement of the Volta potential difference of the Sb electrode, decreased to ca. 5 as the anode was polarized at the current density of 105.7 µA/cm2. Thus, these pH results are in good agreement.
The color distribution at the cathode and the color of the solution changed over time. The pH in the vicinity of localized corrosion during the free corrosion period (on the cathode side) was estimated to be 5.5, and the pH distant from the localized corrosion region was 10. Additionally, the pH near the Sb electrode on the cathode increased to over 10 and maintained a constant value based on the solution color results. The pH distribution observed around the pits on the cathode during the free corrosion (Fig. 8(a)) disappeared with increasing the cathodic polarization since the ORR was enhanced on the whole surface of the cathode. In contrast, the pH measurement using the KP system showed that the pH near the steel of the cathode side increased gradually from 7, finally reaching a stable value of 12. Therefore, the pH change induced during the simulated steel corrosion, as measured by the KP system, agreed with that measured with the pH indicator. Thus, it can be concluded that the pH measurement system using the KP technique/Sb electrode combination can be applied to pH measurements of thin electrolyte droplets on a corroding steel.
4.2 pH change during steel corrosion under an NaCl dropletThe corrosion reaction of steel in a neutral NaCl solution is a combination of the anodic dissolution reaction of iron (Fe) and the cathodic reduction reaction of dissolved oxygen (ORR) as follows:
| \begin{equation} \text{Fe}\rightarrow\text{Fe$^{2+}$} + \text{2e$^{-}$} \end{equation} | (4) |
| \begin{equation} \text{O$_{2}$} + \text{2H$_{2}$O} + \text{4e$^{-}$}\rightarrow\text{4OH$^{-}$} \end{equation} | (5) |
| \begin{equation} \text{2H$_{2}$O} + \text{2e$^{-}$}\rightarrow \text{H$_{2}$} + \text{2OH$^{-}$} \end{equation} | (6) |
Plots of the applied current densities on the anode and cathode sides of steel as a function of the measured potential during the galvanostatic polarization test in 0.1-mol/dm3 NaCl.
As described in eqs. (4) and (5), during the galvanostatic polarization test for steel, Fe2+ ions and OH− ions are generated in the solution. Additionally, dissolved Fe2+ ions are immediately oxidized to Fe3+ ions by dissolved oxygen in the electrolyte solution, and Fe3+ ions react with OH− ions to form Fe(OH)3 via a hydrolysis reaction depending on the solution pH as follows:
| \begin{equation} \text{4Fe$^{2+}$} + \text{O$_{2}$} + \text{2H$_{2}$O} + \text{8OH$^{-}$}\rightarrow \text{4Fe(OH)$_{3}$} \end{equation} | (7) |
As shown in Fig. 7, when the steel was polarized to a potential less noble than −0.6 V, the pH over the cathode side of the steel reached to ca. 12. Numerous experimental and theoretical studies have been performed to investigate the pH on the surface where ORR occurs under the diffusion-limiting condition. For example, Hirohata et al. found that the surface pH over the cathode site reached ca. 13 during corrosion in a thin electrolyte droplet.27) Ogle et al. also showed that the surface pH rose to 11.2 over the steel surface in 0.5-mol/dm3 NaCl.28) Engell et al. calculated the theoretical pH in the solution over the cathode under the diffusion-limiting condition of ORR by accounting for a flux balance of reactants and products, and found that the surface pH increased to 10.9.29) These findings are in good agreement with the pH results obtained by the KP technique in this study over the cathode side of the steel.
Over the anode steel, the pH decreased slightly from its original value, as shown in Fig. 7. This is attributed to the hydrolysis reaction of dissolved Fe3+ ions to form Fe(OH)3 by eq. (8).30) During the galvanostatic polarization for the anode at a current density of 100 µA/cm2, the concentration of dissolved Fe2+ ions can be estimated as ca. 4.5 × 10−3 mol/dm3, assuming that current efficiency for the dissolution of Fe2+ ions is unity and Fe2+ ions prevail in the solution over the anode. However, the dissolved Fe2+ ions can be transformed to Fe(OH)3 in the presence of dissolved oxygen, according to eq. (7). Finally, a steady state is formed by the dissolution equilibrium reaction of eq. (8), in which the Fe3+ ion concentration reaches 1.0 × 10−10 mol/dm3 which is calculated by the solubility product constant of Fe(OH)3 at 25°C, 2.8 × 10−39.31) Using eq. (9), the equilibrium pH was calculated to be about 4.9.
| \begin{equation} \text{Fe(OH)$_{3}$}\rightleftarrows \text{Fe$^{3+}$} + \text{3OH$^{-}$} \end{equation} | (8) |
| \begin{equation} \text{pH} = 1.61-\frac{1}{3}\log [\text{Fe$^{3+}$}] \end{equation} | (9) |
The pH over the surface of corroding steel was investigated using the combination of an Sb electrode as a pH sensor and a KP in a thin electrolyte droplet of 0.1-mol/dm3 NaCl. Our findings are summarized below.
This work was supported by JSPS KAKENHI Grant number JP15K14144.
