1. Introduction
Alkaline water electrolysis is an effective method for the production of high purity hydrogen without the emission of environmental contaminants
[1,
2]. A dimensionally stable anode (DSA) is widely used for alkaline water electrolysis
[3,
4]. However, the fabrication of a DSA is complicated and expensive because of the use of noble metal oxides, such as IrO
2 and RuO
2. Many approaches have reduced the required amount of noble metals and simplified the manufacturing process of DSA
[5-
9]. Ti metal substrate can be covered with a thin oxide film consisting of amorphous or low crystalline TiO
2. TiO
2 is a non-toxic, chemically inert, and corrosion-resistant material
[6]. TiO
2 has been extensively studied under ultraviolet light for photocatalysis because it has semiconductor properties (anatase: 3.2 eV bandgap). Reduced titania (TiO
2-x), which contains oxygen vacancies (V
O-Ti
3+), has recently been reported to exhibit improved photocatalytic activity under visible light. Black titania nanotubes with a substantial amount of Ti
3+ were prepared by melted aluminum reduction and applied as the anode for photoelectrochemical water splitting
[10]. Schumuki et al. reported that anatase black TiO
2 nanotubes with a stable Ti
3+ defect-structure (formed by high-pressure H
2 treatment) exhibited a high photocatalytic behavior for cathodic hydrogen evolution
[11].
The low electrical conductivity (104-107 Ω·cm) and electrocatalytic activity of TiO2 prohibit its use as an electrocatalyst without a co-catalyst. Ti3+ may control the conductivity of TiO2, and the thermally reduced anatase TiO2 has a conductivity of 10-1 Ω·cm. Ti4+ inside the TiO2 lattice can be reduced to Ti3+ when a cathodic potential is applied, and the cathodic coloration of TiO2 is the result of the double injection of a monovalent ion and an electron into the oxide matrix,
where M is the monovalent ion
[12]. Yoon et al. reported the preparation of a blue TiO
2 nanotube array (NTA) by cathodic polarization and applied it to the anodic generation of chlorine and the hydroxyl radical
[8]. In line with the effect of the Ti
3+ defect-structure on the electrocatalytic behavior of TiO
2, the increase of the TiO
2 surface area is a critical issue because it determines the extent of the reaction. Therefore, either TiO
2 nanoparticles or nanotubes could be applied as the anode for alkaline water electrolysis.
In this work, a one-dimensional TiO2 nanotube was electrochemically fabricated and applied to the oxygen evolution reaction (OER) in alkaline water electrolysis. The durability of the TiO2 nanotube structure and its effect on the OER were investigated with electrochemical methods. Surface analysis was performed using scanning electron microscopy and X-ray photoelectron spectroscopy.
2. Experiment
Ti foil (99.7 %, 0.127 mm thickness, Sigma Aldrich) was degreased by ultra-sonication in a solution of acetone, ethanol, and deionized (DI) water for 30 min. The degreased Ti foil was anodized in an electrolyte (0.2 wt% NH4F, 2.5 wt% DI water in ethylene glycol). An amorphous TiO2 NTA was fabricated by the application of a constant voltage at room temperature using a DC power supply (HP/Agilent 6035A). The electrochromism of the TiO2 NTA was tested in a 0.1 M KH2PO4 buffer solution at pH 7.2 with cyclic voltammetry (CV). The blue TiO2 NTA was obtained by cathodic polarization at -3 mA/cm2 for 90 s. Pt wire was used as a counter electrode and Ag/AgCl/sat. KCl was used as a reference electrode.
The crystallinity of the TiO2 NTA changed from amorphous to anatase by annealing at 450 ℃ for 1 h. The surface characterization of the TiO2 NTA was investigated with a scanning electron microscope (SEM, S-4300, HITACHI), X-ray diffractometer (XRD, DMAX 2200, RIGAKU), and X-ray photoelectron spectroscope (XPS, K-Alpha, Thermo Scientific).
The electrochemical activity of the TiO2 NTA as an anode in 1 M KOH was analyzed with a potentiostat/galvanostat (Autolab, PGSTAT301N, METROHM) in a three-electrode system. Pt wire and Hg/HgO/1 M NaOH were used as a counter electrode and reference electrode in 1 M KOH solution, respectively.
3. Results and Discussion
Fig. 1 shows the effect of the heat treatment on the crystallinity of the electrochemically fabricated anodized TiO
2 NTA. After the heat treatment at 450 ℃ for 1 h, the amorphous TiO
2 changed to the gray crystalline anatase phase.
Fig. 2 shows the CVs of the amorphous and anatase TiO
2 NTA in a 0.1 M KH
2PO
4 buffer solution at pH 7.2. The potential was anodically scanned from 0.0 V to 3.0 V, reversed to -2.0 V, and repeated. Both specimens changed to the blue TiO
2 at around -1.5 V by the double injection reaction (1). During the repetitive CV, the amorphous TiO
2 underwent reversible electrochromism reactions (i.e., a cathodic coloration and anodic bleaching, as shown in
Fig. 2a) but did not show any anodic reactivity above 1.5 V. After the first cycling for the anatase TiO
2, the anodic current started to flow above 1.5 V (due to the reaction: 2H
2O→O
2 + 4H
+ + 4e
-) and continued until the end of the CV (
Fig. 2b), which is in contrast to the amorphous TiO
2 NTA. During the first CV between -1.5 V and 0.0 V, the anatase TiO
2 NTA experienced a cathodic color change to dark blue caused by reaction (1), but bleaching did not occur despite the anodic current flow, which suggests an irreversible electrochromism. The blue anatase TiO
2 NTA persisted during the entire cycling in
Fig. 2b and its catalytic activity for the anodic reaction was investigated. The blue anatase TiO
2 NTA was cathodically fabricated at -3 mA/cm
2 for 90 s in a pH 7.2 buffer solution and applied as the anode for the OER in 1 M KOH.
Fig. 3 indicates that the blue anatase TiO
2 NTA has an excellent catalytic activity for the OER (4OH
-→O
2 + 2H
2O + 4e
-), but the anatase TiO
2 NTA did not participate in the reaction. The XPS spectra in
Fig. 4 show that the Ti
3+ peaks (457.7 eV, 462.9 eV) decreased and the Ti
4+ peaks (459.5 eV, 464.7 eV) increased after the repeated 100 cycles of linear sweep voltammetry (LSV) for the OER. The Ti
3+ in the blue anatase TiO
2 NTA, which can be K
xTi
3+O
2, generated by cathodic polarization is responsible for the OER. However, the oxidation of K
xTi
3+O
2 to TiO
2 (Ti
3+→Ti
4+) during the OER, may cause the loss of its catalytic activity as the anodic OER continues.
Fig. 1.
XRD spectra of the anodized TiO2 nanotube array; (a) amorphous (b) anatase (after 1 h of annealing at 450 ℃).
Fig. 2.
Cyclic voltammograms of the (a) amorphous and (b) anatase TiO2 nanotube arrays in a buffer solution (0.1M KH2PO4, pH 7.2 with NaOH). Potential sweep rate: 40 mV/s.
Fig. 3.
Comparison of the oxygen evolution reaction activity of the blue and anatase TiO2 nanotube arrays in 1 M KOH solution. The potential was linearly swept from 1.5 V to 2.5 V at 20 mV/s. Anodization conditions: 45 V for 20 h.
Fig. 4.
XPS spectra of the blue anatase TiO2 nanotube array before and after 100 repetitive linear sweep voltammetries for the oxygen evolution reaction. Potential sweep rate: 20 mV/s.
Fig. 5 presents the surface morphology of the anatase TiO
2 NTA prepared at 45 V of anodization voltage with reaction times of 4 h, 8 h, and 16 h.
Fig. 6 indicates that the length (
l) and diameter (D) of the TiO
2 nanotubes increased with the anodization time. Since the inner surface area of the 1-D nanotube is expressed as N·π·D·
l (where N is the number density of the nanotubes), the surface area for the OER increases with the anodization time. The LSV in
Fig. 7a coincides with the expected surface area increase, i.e., the specimen prepared at 45 V over 16 h shows the highest catalytic activity for the OER. However, the OER activity decreased with each LSV iteration, and the color of the blue anatase TiO
2 NTA gradually tarnished because of the oxidation of Ti
3+ to Ti
4+, as suggested by the XPS spectra in
Fig. 4.
Fig. 5.
Morphology of the TiO2 nanotube array (NTA) with respect to anodization time. Top view of the TiO2 NTA prepared at 45 V for (a) 4 h, (b) 8 h, and (c) 16 h. Insets are cross-sectional views. Anodization solution: 0.2 wt% NH4F, 2.5 wt% DI water with ethylene glycol.
Fig. 6.
Average diameter and length of the TiO2 nanotube array with different anodization times. Anodization voltage: 45 V, anodization solution: 0.2 wt% NH4F, 2.5 wt% DI water with ethylene glycol.
Fig. 7.
Effect of anodization time on the oxygen evolution reaction activity and durability of the blue anatase TiO2 nanotube array in 1 M KOH solution. The anodization voltage, 45 V, in/i1, is defined as the ratio of the anodic current density of the nth cycle and 1st cycle from the linear sweep voltammetry at 2.5 V.
Fig. 7b shows that the extent of the catalytic activity degradation,
in/i1, increases with the increase in the number of LSV cycles, where
in/i1 is defined as the ratio of the anodic current density of the n
th cycle and 1
st cycle at 2.5 V. Irrespective of the anodization time, 35-40 % of the initial activity of the blue anatase TiO
2 NTA was lost after 10 LSV cycles. To investigate the effect of the different TiO
2 NTA fabrication methods on the catalytic activity for the OER, the anodization voltage was varied from 35 V to 45 V and 55 V with the same anodization time of 16 h.
Fig. 8 indicates that, as the anodization voltage increased from 35 V to 55 V, the tube diameter increased by ~25 %, but the tube length increased to nearly 4 times its initial length (from 6 μm to 22 μm).
Fig. 9 (in agreement with the data in
Figs. 6 and
7) suggests that the blue anatase TiO
2 NTA with a large surface area has high OER activities and their activity losses reach 35-42 % of the initial activities.
Fig. 8.
Average diameter and length of the TiO2 nanotube array with different anodization voltages. Anodization time: 16 h, anodization solution: 0.2 wt% NH4F, 2.5 wt% DI water with ethylene glycol.
Fig. 9.
Effect of the anodization voltage on the oxygen evolution reaction activity and durability of the blue anatase TiO2 nanotube array in 1 M KOH solution. Anodization time: 16 h. in/i1 is defined as the ratio of the anodic current density of nth cycle and 1st cycle from the linear sweep voltammetry at 2.5 V.
To overcome the catalytic activity degradation, the tarnished blue TiO
2 NTA was regenerated by cathodic polarization after repeated LSVs (-3 mA/cm
2 for 90 s in a pH 7.2 buffer solution). Regeneration occurred after every 10 LSV cycles.
Fig. 10 shows the effect of regeneration on the OER, where
is the current density at 2.5 V after 10 cycles. Although the catalytic activity of the blue anatase TiO
2 NTA for the OER was not completely recovered during the regeneration, its degradation rate was significantly reduced. This suggests that the blue anatase TiO
2 NTA has good catalytic activity for the OER and its degradation can be recovered by an intermittent cathodic polarization.
Fig. 10.
Effects of regeneration on the durability of the blue anatase TiO2. Blue anatase TiO2 was regenerated in a pH 7.2 KH2PO4 buffer solution with -3 mA/cm2 for 90 s after every 10 cycles of linear sweep voltammetry. is the current density at 2.5 V after the initial 10 cycles of linear sweep voltammetry. in is defined as the anodic current density of nth cycle at 2.5 V.