Stable solar-driven water splitting by anodic ZnO nanotubular semiconducting photoanodes

Abstract

The development of high performance artificial photosynthetic devices, to store solar energy in chemical bonds, requires the existence of stable light-absorbing electrodes for both the oxidative and reductive half-reactions. The development of such systems has been hindered in part by the lack of semiconducting photoanodes that are stable under the water spitting conditions. We demonstrate, for the first time, the synthesis of ZnO nanotubes via anodization of Zn foil and their use as photoanodes in photoelectrochemical hydrogen generation systems. Structural, optical and photoelectrochemical measurements showed the superiority of the nanotubular structure over the nanowire and hierarchical counterparts.

---

Zinc oxide (ZnO) is an extraordinarily versatile material with wide use in sensing,¹ catalysis,² solar cells³ and nanogenerators.⁴ It has a wide band gap (3.37 eV), allowing absorption in the UV region, a large exciton binding energy (60 meV),⁵ and a high electron mobility (205–1000 cm² V⁻¹ s⁻¹).⁶ Therefore, many reports have been devoted toward the fabrication of ZnO nanostructures with different shapes and sizes.^7,8 However, the fabrication processes are usually very complicated, expensive and irreproducible. The nanotubular morphology is one of the most interesting nanostructures of ZnO for many applications. Nanotubes can be grown on a variety of different substrates using various techniques including chemical vapor deposition (CVD), atomic layer deposition (ALD), hydrothermal growth and sol–gel chemistry. However, these growth techniques typically suffer one of two potential problems: (a) limiting the choice of substrate because of the high growth temperatures, and/or (b) low growth rate requiring reaction times on the scale of hours or even days to achieve a reasonable yield.^5–10 To this end, electrochemical anodization is a facile method that has been used to fabricate a variety of nanostructures.^11–13 The anodization of zinc has received relatively little attention compared to other metals, which could be attributed to the instability of zinc oxide in the acidic electrolytes commonly used during anodization.⁸ On the other hand, nanostructures including nanostripes, nanowires, nanodots and nanoflowers have all been produced using electrolytes ranging from basic NaOH solutions to highly acidic HF solutions.⁹ Hu et al.¹¹ were able to produce randomly oriented ZnO nanowires with aspect ratios exceeding 1000 via anodization in aqueous KHCO₃ solutions. Through optimized anodization conditions, Allam and co-workers¹² were able to fabricate vertically-oriented ZnO nanowires. Despite those great efforts, it is still challenging to produce stable thin films of ZnO nanotubes via anodization of Zn foil. Herein, we demonstrate, for the first time, the possibility to produce highly stable ZnO nanotubes via optimized anodization of Zn foil in 50 mM KHCO₃ solutions and tested their photocatalytic activity to split water. The detailed experimental set-up and anodization conditions are available in the ESI.† In order to investigate the morphology–activity relationship, three sets of samples were fabricated and tested, namely ZnO nanowires, hierarchical and nanotubes (NTs).

Fig. 1 shows FESEM images of the fabricated nanostructures: well-aligned nanowires with diameters of 69 ± 5 nm, hierarchical nanowires with horn shaped ZnO crystals (nanoparticles decorated-nanowires), and hexagonal ZnO nanotubes with diameters of 130 ± 5 nm and average wall thickness of 20 ± 5 nm. More images can be seen in Fig. S1–S4 in the ESI.† During anodization, two competing processes resulted in the formation of ZnO nanostructures: the dissolution of zinc foil (Zn²⁺) and the simultaneous oxidation to ZnO. Tuning the anodization parameters allowed control of the two processes, leading to the presence or absence of nanostructures and, in the former case, control over their morphology and properties.

Fig. 1 FESEM images of ZnO (a) nanowires, (b) hierarchical nanowires (nanoparticles decorated on nanowires), (c) top-view nanotubes, and (d) side-view nanotubes.

The general reaction scheme for the anodization of zinc can be represented as follows:¹⁴

At the cathode: 2H⁺ + 2e⁻ → H_2(g) (1)

At the anode: Zn → Zn²⁺ + 2e⁻ (2)

Zn²⁺ + 2OH⁻ → Zn(OH)₂ (3)

Zn(OH)₂ → ZnO + H₂O (4)

Dissolution: ZnO + HX* → ZnX* + H₂O (5)

where X* in the dissolution process refers to HCO₃⁻ anion from KHCO₃. This can be confirmed via the corresponding current–time behavior, see Fig. 2a, recorded during the anodization of Zn. Initially, a compact oxide layer is formed through hydrolysis of Zn, see eqn (1)–(3). This oxide layer leads to a dramatic decrease in the anodization current due to its poor electrical conductivity. After that, ZnO starts to dissolve, eqn (5), leading to the observed slight increase in current with time. This can be explained on the basis of the high field model (HFM)¹⁵ and its modified form.¹⁶ Under sufficient applied voltage magnitude, the electric field will be strong enough to migrate the Zn ions into the electrolyte leaving behind the observed structures. According to the HFM,^30,31 the current under high field conditions during the formation of the oxide layer takes the simple form:¹⁷

i = αe^βF (6)

where i represents the current, F is the electric field strength (V cm⁻¹), α is the jump probability of a cation interstitial given by eqn (7), and β is calculated for Zn by eqn (8):

Fig. 2 (a) Current–time response during the anodization of Zn foil to form ZnO nanostructures and (b) XRD pattern of the as-anodized and oxygen-annealed ZnO films.

In the above equations e is the electron charge, k is Boltzmann's constant, T is the absolute temperature, a is one half of the jump distance, φ₀ is the minimum potential barrier, and τ is the time of vibration. Note here that the charge associated with formation of a Zn²⁺ cation is ne where n = 2. The electric field linearly decreases with oxide layer thickness, while from eqn (6) the current will decay exponentially with film thickness.

The growth of the hierarchical nanowires was shown to occur via the branching of a single large, hexagonal wire into several smaller nanowires, with diameters similar to those at the start of the nanowire growth (Fig. S2†). These smaller nanowires, together with the underlying wires, continue to grow to form very thick nanowire films (>100 μm). As the thicknesses exceeds 50 μm, however, complex 3D networks of nanowires were typically formed rather than aligned nanowires,⁸ see ESI† for more details. Also, the two step anodization at 5 and 10 V resulted in nanowire/nanotube and nanotube morphologies, respectively (Fig. S3 and S4†). The exact reasons for both the formation of layered films of different sized nanowires and the formation of hierarchical structures are currently unknown. However, this could be related to the increasing distance from the source of Zn²⁺ ions to the nanowire tip, thereby reducing the concentration of Zn²⁺ ions available for reaction. It might also be attributed to a decreasing concentration of the already dilute bicarbonate species present in the electrolyte as a function of time.⁹ It is worthy to mention that leaving the nanowires film in a sealed vial containing deionized water for 24 h, followed by annealing for 1 h at 250 °C with heating and cooling rates of 1 °C min⁻¹, resulted in the formation of hierarchical nanostructures. The soaking of ZnO nanowires in aqueous medium is thought to enrich the hydroxyl group content on the surface, which might facilitate the partial dissolution of Zn²⁺ ions and forms Zn(OH)₂/[Zn(OH)₄]²⁻ and subsequently transforms to ZnO during the thermal treatment.

The co-existence of nanotubes and nanowires in the film indicate a unique mechanism for anodic nanotube growth. It does not appear fully consistent with the existing prominent models such as the stress-induced material flow for the anodic porous structure formation. To throw light into the nanotube formation mechanism, experiments were conducted on zinc foils anodized at different durations. The sample obtained after anodization for 10 s showed shallow etched pits from where needles started growing out. A sea urchin-like growth pattern appeared after 30 s. The needles were transformed into hexagonal rods of about 2 μm in one minute of anodization. At this time, pits started appearing on the flat top surface of the wire, which grew deeper with a concurrent increase in the nanowire length, resulting in more or less hexagonally shaped tubular structure. It was also obvious that a minimum diameter (∼100 nm) was required for the rods to have tubular conversion. A similar size effect was reported in single crystal zinc rods formed by other techniques when converted later to tubes by etching in acidic or alkaline media.

It can be concluded that anodization for short time at low potential leads to the formation of tubes, while increasing both time and potential resulted in the formation of nanowires. Table 1 summaries the fabrication conditions and the obtained morphologies.

Table 1 Summary of the experimental conditions and the resulted nanostructures

Electrolyte composition	Voltage [V]	Time [min]	Post treatment	Obtained structure
50 mM KHCO3	10	30	Annealing	Nanowires
50 mM KHCO3	10	30	Water treatment + annealing	Hierarchical
50 mM KHCO3	10 then 5	30 then 10	Annealing	Nanowires/nanotubes
50 mM KHCO3 + 30 mM Na2CO3	10	10	Annealing	Nanotubes

---

X-ray diffraction (XRD) analysis was performed for as-anodized and oxygen-annealed samples to elucidate their crystal structure, Fig. 2b. The XRD pattern of the as-anodized nanotubes showed a complex set of sharp peaks that can be assigned to the metal zinc foil (JCPDS 04-0831), as well as several broad peaks corresponding to the hexagonal wurtzite ZnO (JCPDS 36-1451). Upon annealing, the intensity of the ZnO diffraction peaks associated with (100), (002), (101), (012), (110), (013), (200), and (202) lattice planes drastically increased, indicating the formation of polycrystalline wurtzite ZnO structures. Note that all tested morphologies showed similar diffraction patterns, with the nanotubes showing the most intense peaks.

The high resolution transmission electron microscopy (HRTEM) investigation, Fig. 3, further confirms the formation of polycrystalline one-dimensional ZnO nanotubes, which is consistent with the XRD results. The spotty ring in the selected area electron diffraction (SAED) pattern shows six diffraction rings corresponding to (100), (002), (102), (110), (103), and (112) lattice planes of ZnO. The lattice fringes spectra showed a spacing of 4.1 Å, characteristic of ZnO.

Fig. 3 HRTEM images of the zinc oxide nanotubes film. The inset shows the SAED.

To elucidate the composition of the fabricated nanotubes, X-ray photoelectron spectroscopy (XPS) analysis was performed with the spectra charge referenced to O 1s at 532 eV, Fig. 4. The spectra showed well-resolved Zn 2p and O 1s peaks that are characteristic of ZnO. While the peak observed at 531 eV is characteristic of Zn–O bonding, the shoulder extending to 532.2 eV can be related to the O–H bonding. Doublet Zn peaks were observed at 1022.19 and 1045.14 eV, with a spin orbit splitting of ∼23 eV correspond to Zn²⁺.^12–14,18

Fig. 4 High resolution XPS spectra of the zinc oxide nanotubes film.

Raman-scattering is another effective technique to investigate the crystallinity and the vibrational properties of materials as the Raman signals are very sensitive to the crystal structure and defects. Wurtzite ZnO, which belongs to the P6₃mc space group, would possess six Raman active vibration modes:¹⁹

T = A₁ + 2B₁ + E₁ + 2E₂ (9)

Both A₁ and E₁ modes are polar and split into transverse optical (A₁TO and E₁TO) and longitudinal optical (A₁LO and E₁LO) components.²⁰ B₁ is silent and E₂ has two modes of low and high frequency phonons (E₂ low and E₂ high), which are associated with the vibration of the Zn sub-lattice and oxygen atom.²⁰ The peaks appeared at 208, 438 and 574 cm⁻¹, Fig. 5a, can be assigned to the E₂ low, E₂ high and A₁ LO, respectively.²¹ The Raman peak at 438 cm⁻¹ may be attributed to the non-polar optical phonon mode of ZnO (E₂ mode), characteristic of the hexagonal wurtzite phase.²² The peak at 578 cm⁻¹ corresponds to the LO mode with A₁ symmetry.²² Besides these first-order Raman modes, the E(LO) Raman peak can be assigned to the second-order Raman mode due to the vertical alignment of the nanotubes.²³

Fig. 5 (a) Raman spectra, (b) normalized absorption, (c) photoelectrochemical properties, and (d) the current transients of the fabricated ZnO nanostructures.

Fig. 5b shows the typical UV-vis diffuse reflectance spectra (DRS) of the fabricated ZnO nanostructured materials.¹⁴ Note that all films showed similar absorption spectra with a small red shift in the absorption edge from nanowires (3.31 eV) to the hierarchical structure (3.23 eV), nanowires/nanotubes structure (3.2 eV) and the nanotubes (3.15 eV). The differences in the bandgaps are within experimental error and therefore can be attributed to scattering effects due to the different film morphologies.¹⁴

The photoelectrochemical activity test for water photoelectrolysis using the grown ZnO nanostructured films was carried out in a 0.5 M Na₂SO₄ electrolyte solution using a typical three-electrode electrochemical cell, where the annealed ZnO nanostructured film is used as the working electrode, Pt foil as the counter electrode, and calomel as the reference electrode connected to Biologic SP200 potentiostat. Previous reports showed that ZnO nanostructures undergo photocorrosion upon their use in water splitting cells, due to the formation of a precipitate on the surface that hinders their proper irradiation.²⁴ Under AM 1.5G illumination (100 mW cm⁻²), Fig. 5c, the ZnO nanowires showed the lowest photocurrent (0.25 mA cm⁻² at 0.25 V_SCE), among the tested nanostructured films, probably due to the increase in the photogenerated electron pathway, which makes them more prone to recombination.²⁵ The hierarchal ZnO films showed higher photocurrent (0.31 mA cm⁻² at 0.25 V_SCE), which can be attributed to the higher surface area. The ZnO nanotubes showed the highest photocurrent (0.52 mA cm⁻² at 0.25 V_SCE), which is double that of the ZnO nanowires. The ZnO nanowires/nanotubes electrode showed higher photocurrent than the nanowires and hierarchal ZnO counterparts, which can be related to the presence of nanotubes that facilitate charge separation and collection. The superior performance of the nanotubes can be attributed to the higher surface area of the nanotubes and both better light absorption and reaction kinetics.²⁵ Note that the dark scans showed almost negligible current densities in the range of 10 μA cm⁻². The tested nanostructured electrodes show n-type behavior, i.e., positive photocurrents at anodic potentials.²⁶ To the best of our knowledge, this is the first time to report photocurrent of anodic zinc oxide nanotubes fabricated by anodization of zinc foil. This photocurrent is more than double that reported for both Al-doped ZnO nanorods^27,28 and pristine ZnO NWs.²⁹

To assess the stability of the fabricated ZnO nanostructures, the transient photocurrent (J–t) tests were carried out under light on/off conditions in 0.5 M Na₂SO₄ solutions at a constant external bias of 0.25 V_SCE, Fig. 5d. The photocurrent of the tested anodic films decays sharply under light-off conditions, indicating good carrier transport properties. The photocurrent was always the same over the entire duration of the test indicating the high stability of the tested electrodes. However, upon long time illumination (25 h), both nanotubes and hierarchical nanowires remain stable, while the nanowires deteriorate after the first two hours, see Fig. 6.

Fig. 6 Long term current transients of the fabricated ZnO nanostructures.

In order to understand the charge carriers lifetime, the photocurrent decay and response behavior were further examined. According to eqn (10) and (11), the decay time constant (τ_d) and rise time constant (τ_r) can be calculated:^30–32

I = I₀e^−t/τ_d (10)

I = I₀(1 − e^−t/τ_r) (11)

where t is the time, I is photocurrent, and I₀ is the initial photocurrent. The estimated τ_d values of the ZnO nanowires, hierarchical and nanowires/nanotubes and nanotubes are determined to be 3.49, 3.025, 2.832 and 2.619 s, respectively. On the other hand, τ_r values of the ZnO nanowires, hierarchical, nanowires/nanotubes and nanotubes are determined to be 1.84, 1.57, 1.43 and 1.31 s, respectively. The significant decrease in decay and rise time constants could be attributed to the increased carrier concentration resulted from the higher surface area of the nanotube, as well as significant conductivity enhancement.^28,33,34 The phonon lifetime (τ) can be derived from the Raman spectra via the energy-time uncertainty relation:³⁵

1/τ = ΔE/h = 2πcr (12)

where ΔE is the uncertainty in the energy of the phonon mode, h is Planck's constant, c is the speed of light, and r is the FWHM of the Raman peak in cm⁻¹. The calculated phonon lifetimes for the nanotubes, nanowires/nanotubes, hierarchical and nanowires are 0.825, 0.828, 0.831 and 0.834 ps, respectively. Note that the lower the life time, the higher the photocurrent, which is in agreement with the results shown in Fig. 5c and d, making ZnO nanotubes the best candidate as photoanode.

In summary, a new synthetic approach to produce zinc oxide nanotubes with pore diameter (130 ± 5 nm) and wall thickness of (20 ± 5 nm) via potentiostatic anodization in aqueous electrolytes is presented, for the first time. Chemical and structural characterization confirmed the formation of hexagonal wurtzite ZnO nanotubes, with band gap energy of 3.15 eV. The photoelectrochemical characterization indicated that ZnO nanotubes exhibited superior photocurrent density (0.52 mA cm² at 0.25 V_SCE) and stability. The observed high photocurrent is likely due to the high surface area, better light absorption, and reaction kinetics.

Acknowledgements

We would like to acknowledge the financial support of this work by The American University in Cairo, Grant #SSE – PHYS – NA – RSG – FY16 – FY17 – Contract No. 7.

Notes and references

1. R. A. Talib, M. J. Abdullah, H. S. Al-Salman, S. M. Mohammad and N. K. Allam, Mater. Chem. Phys., 2016, 181, 7–11.
2. M. D. Hernandez-Alonso, F. Fresno, S. Suarez and J. M. Coronado, Energy Environ. Sci., 2009, 2, 1231–1257.
3. Q. F. Zhang, C. S. Dandeneau, X. Y. Zhou and G. Z. Cao, Adv. Mater., 2009, 21, 4087–4108.
4. B. Kumar and S.-W. Kim, Nano Energy, 2012, 1, 342–355.
5. D. C. Look, D. C. Reynolds, J. R. Sizelove, R. L. Jones, C. W. Litton, G. Cantwell and W. C. Harsch, Solid State Commun., 1998, 105, 399–401.
6. C. Klingshirn, J. Fallert, H. Zhou, J. Sartor, C. Thiele, F. Maier-Flaig, D. Schneider and H. Kalt, Phys. Status Solidi B, 2010, 247, 1424–1447.
7. E. Galoppini, J. Rochford, H. Chen, G. Saraf, Y. Lu, A. Hagfeldt and G. Boschloo, J. Phys. Chem. B, 2006, 110, 16159–16161.
8. S. G. Kumara and K. S. R. K. Rao, RSC Adv., 2015, 5, 3306–3351.
9. D. O. Miles, P. J. Cameron and D. Mattia, J. Mater. Chem. A, 2015, 3, 17569–17577.
10. S. J. Kim and J. Choi, Electrochem. Commun., 2008, 10, 175–179.
11. Z. Hu, Q. Chen, Z. Li, Y. Yu and L.-M. Peng, J. Phys. Chem. C, 2010, 114, 881–889.
12. N. Samir, D. S. Eissa and N. K. Allam, Mater. Lett., 2014, 137, 45–48.
13. R. Al-Gaashani, S. Radiman, A. R. Daud, N. Tabet and Y. Al-Douri, Ceram. Int., 2013, 39, 2283–2292.
14. S. H. Huang, Z. Chen, X. C. Shen, Z. Q. Zhu and K. Yu, Solid State Commun., 2008, 145, 418–422.
15. N. Cabrera and N. F. Mott, Rep. Prog. Phys., 1949, 12, 163.
16. D. D. Macdonald, J. Electrochem. Soc., 1993, 140, L27–L30.
17. N. K. Allam and C. A. Grimes, J. Phys. Chem. C, 2007, 111, 13028–13032.
18. A. Ramirez-Canon, D. O. Miles, P. J. Cameron and D. Mattia, RSC Adv., 2013, 3, 25323–25330.
19. M.-C. Huang, W.-S. Chang, J.-C. Lin, Y.-H. Chang and C.-C. Wu, J. Alloys Compd., 2015, 636, 176–182.
20. H. N. Dinh, Photoelectrochemical water splitting: standards experimental methods and protocols, Spinger, 2003.
21. M.-I. Lee, et al., Thin Solid Films, 2014, 570, 516–526.
22. N. K. Hassan, M. R. Hashim and N. K. Allam, Chem. Phys. Lett., 2012, 549, 62–66.
23. R. Zhang, P.-G. Yin, N. Wang and L. Guo, Solid State Sci., 2009, 11, 865–869.
24. R. Jothilakshmi, V. Ramakrishnan, R. Thangavel, J. Kumar, A. Sarua and M. Kuball, J. Raman Spectrosc., 2009, 40, 556–561.
25. R. Sanchez-Tovar, et al., RSC Adv., 2016, 6, 30425–30435.
26. A. M. Mohamed, A. S. Aljaber, S. Y. AlQaradawi and N. K. Allam, Chem. Commun., 2015, 51, 12617–12620.
27. A. M. Mohamed, S. A. Shaban, H. A. El Sayed, B. E. Alanadouli and N. K. Allam, Int. J. Hydrogen Energy, 2016, 41, 866–872.
28. M.-C. Huang, et al., Appl. Surf. Sci., 2016, 360, 442–450.
29. C.-H. Hsu and D.-H. Chen, Int. J. Hydrogen Energy, 2011, 36, 15538–15547.
30. X. Li, et al., J. Phys. Chem. C, 2016, 120, 4183–4188.
31. B. S. Shaheen, A. M. Hafez, B. Murali, A. R. Kirmani, A. Amassian, O. F. Mohammed and N. K. Allam, Sol. Energy Mater. Sol. Cells, 2016, 151, 149–153.
32. S. Mridha, M. Nandi, A. Bhaumik and D. Basak, Nanotechnology, 2008, 19, 275705.
33. J. H. Kim, T. K. Yun, J. Y. Bae and K. S. Ahn, Jpn. J. Appl. Phys., 2009, 48, 120204.
34. M. Y. Lu, M. P. Lu, Y. A. Chung, M. J. Chen, Z. L. Wang and L. J. Chen, J. Phys. Chem. C, 2009, 113, 12878e82.
35. S. S. Mali, C. A. Betty, P. N. Bhosale and P. S. Patil, CrystEngComm, 2011, 13, 6349–6351.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra18747a

---