A nanostructured ZnO–ZnFe₂O₄ heterojunction for the visible light photoelectrochemical oxidation of water

Abstract

A heterojunction is successfully fabricated by decorating ZnO nanorods with water-soluble, mono-dispersed ZnFe₂O₄ particles. In comparison with bare ZnO, the enhanced visible light absorption and photocurrent (λ > 420 nm) for the photoanode of the ZnO–ZnFe₂O₄ composite is correlated with the visible light response of the ZnFe₂O₄ prepared and the rapid charge transfer within the heterojunction as evidenced by the electrochemical impedance spectra.

---

During the past decades, great efforts have been made to explore renewable energy due to the increasing consumption of fossil fuels, which will be depleted soon in the near future.^1–3 Among various approaches, photoelectrochemical water splitting is considered to be one of the most promising methods for producing H₂ fuel from solar energy and water since the initial work on TiO₂.⁴ As a key component in the photoelectrochemical cell, photo-electrodes have been the focus of investigation. Hence, numerous materials, especially inorganic semiconductors such as ZnO,^5–7 Fe₂O₃,^8,9 WO₃¹⁰ and BiVO₄,¹¹ have been investigated aiming at achieving a high efficiency of solar energy conversion. Among them, ZnO has been extensively studied due to its fascinating features such as comparable photocatalytic properties to TiO₂, nontoxicity, low cost and the large electron mobility and various morphologies that can be facilely achieved.^12–14 Recently, one-dimensional (1D) nanostructures, such as nanorods have been of special interest in the fabrication of photoelectrochemical (PEC) devices. Nanorods offer improved charge transport over zero-dimensional (0D) nanostructures while maintaining the high surface area desired for the increasing loading of the dye or quantum dots (QDs).^15,16

However, the main drawback of ZnO is the poor visible light absorption rooted in its wide bandgap of 3.2–3.4 eV. Considerable efforts have been made to improve its visible light absorption, including elemental doping, dye or quantum dot (QD) sensitization.^17,18 For example, N-doping achieved by ammonia annealing can significantly improve the absorption of ZnO in the visible light region.¹⁹ However, this is known to introduce a trap state in the ZnO, which acts as a recombination center for the photo-generated carriers. The results suggest that the enhanced photocurrent for hydrogen annealed ZnO is most likely due to the more efficient use of UV light as a result of the increased donor density via a self-doping process.²⁰ Its photoelectrochemical performance in the visible light range remains nearly unchanged. The quantum dot (such as CdS) sensitized ZnO can generate a high photocurrent, but sacrificial reagents such as S²⁻ are required because of the photo-corrosion of the quantum dots.²¹ Therefore, it is greatly desirable to explore a material which is capable of not only broadening the adsorption of ZnO towards the visible range but also stable enough under sun light illumination.

ZnFe₂O₄, as an n-type semiconductor with a band gap of 1.9–2.0 eV, is one of the most important spinel system materials.²² It has advantages of a low cost, is environment friendly, is photo-corrosion resistant and has a strong visible light absorption.²³ All of these make it a promising material for the application of photoelectrochemical water oxidation. Furthermore, ZnFe₂O₄ has more negative conduction and valance band edge than ZnO, so a heterojunction with rational band alignment can be formed once ZnFe₂O₄ is decorated onto ZnO. The heterojunction is expected to offer efficient charge separation. ZnFe₂O₄ can extend visible light absorption to at least 600 nm which corresponds to its band gap of 1.9–2.0 eV, and a visible light photoresponse can be therefore expected.

In this work, ZnFe₂O₄ nanoparticles with size between 5–8 nm are synthesized via a thermal decomposition of metal–surfactant complexes according to a previously reported method.²⁴ The particles are subsequently modified to be mono-dispersed in water by polyethylene glycol.²⁵ The heterojunction consisting of ZnO nanorods and ZnFe₂O₄ nanoparticles is prepared by a simple drop coating method on a transparent conducting substrate. The heterojunction is characterized by SEM, TEM, STEM elemental mapping, XRD and UV-vis absorption spectra. The photoelectrochemical performance of the ZnO–ZnFe₂O₄ composite under visible light (λ > 420 nm) is carefully investigated.

The FE-SEM images of the bare ZnO nanorods and the ZnO–ZnFe₂O₄ composite with different amounts of ZnFe₂O₄ deposition are shown in Fig. 1. As can be seen from Fig. 1(a) and (a′), the ZnO nanorods grow vertically on the substrate with a diameter of ∼50–80 nm and a length of ∼2.0 μm. The surface of the ZnO nanorods is smooth and without aggregation of the ZnO nanoparticles. After being coated with 30 μL ZnFe₂O₄, part of the gap between each rod is filled and the surface of the rods becomes rough as shown in Fig. 1(b) and (b′). When repeating the above coating three times, we can see from Fig. 1(c) and (c′) that the top surface of the ZnO nanorod arrays is completely covered by ∼100 nm thick ZnFe₂O₄. When 3 × 60 μL ZnFe₂O₄ is deposited, the layer of ZnFe₂O₄ on the ZnO nanorod film reaches ∼1.0 μm as shown in Fig. 1(d) and (d′).

Fig. 1 The FE-SEM images of the bare ZnO nanorods (a and a′), the ZnO–ZnFe₂O₄ composite with 30 μL ZnFe₂O₄ deposition (b and b′), 3 × 30 μL ZnFe₂O₄ (c and c′) and 3 × 60 μL ZnFe₂O₄ (d and d′).

The absorption spectra of the ZnO nanorods and the ZnO–ZnFe₂O₄ composite are shown in Fig. 2(a). It is seen that the bare ZnO nanorods exhibit an absorption cutoff edge at about 390 nm, which is consistent with the typical band gap of ZnO of 3.2 eV. After being decorated with ZnFe₂O₄, the absorption range extends significantly to the visible light region, reaching 600 nm. This confirms the success of our thoughts to utilize ZnFe₂O₄ to extend the absorption of ZnO towards the visible light direction. The decrease of the absorption in the UV region for the ZnO–ZnFe₂O₄ composite is observed. This differs from the p–n heterojunction of the ZnFe₂O₄–ZnO composite, in which the increase of the absorption intensity in the UV region is reported.²⁶ It is obvious that the absorption of the ZnO–ZnFe₂O₄ composite for visible light increases with the increased amount of ZnFe₂O₄. It reaches saturation until 3 × 60 μL of ZnFe₂O₄ is deposited. The XRD patterns of the ZnO nanorods and the ZnO–ZnFe₂O₄ composite are illustrated in Fig. 2(b). It can be seen that in addition to the dominating diffraction peak of (002) at 2θ = 34.4° for the ZnO nanorods, two weak diffraction peaks appear at 30.72 and 42.6° which can be assigned to the diffraction of the (220) and (400) peaks for spinel ZnFe₂O₄, respectively. The weak diffraction suggests the low crystallinity of ZnFe₂O₄.

Fig. 2 The UV-vis absorption spectra (a) and XRD patterns (b) of the samples. The inset in the absorption spectra is the photograph of the corresponding samples. The triangles in the XRD patterns represent the peaks of FTO.

It can be seen from the TEM image shown in Fig. 3(a) that the ZnFe₂O₄ has the nature of uniform particle size. The average size is about 6 ± 0.5 nm, which is judged from the counting of 100 particles. From the HR-TEM image in Fig. 3(b), the atomic lattice of the ZnFe₂O₄ can be clearly seen, indicating the crystallinity of ZnFe₂O₄. The d-spacings of 0.300, 0.483 and 0.256 nm can be assigned to the reflections of the (220), (111) and (311) facets for the cubic spinel phase, respectively. Fig. 3(c) shows the TEM image of the ZnO–ZnFe₂O₄ composite, from which we can see clearly that the ZnO nanorod is decorated by the ZnFe₂O₄ particles uniformly. The HR-TEM image shown in Fig. 3(d) reveals the direct contact of the ZnFe₂O₄ particle with the ZnO nanorod. The lattice fringe with a d-spacing of 0.483 nm is the (111) plane of ZnFe₂O₄. These results confirm the successful fabrication of the ZnO–ZnFe₂O₄ composite.

Fig. 3 The TEM images of the ZnFe₂O₄ nanoparticles (a and b) and the ZnO–ZnFe₂O₄ composite with 3 × 60 μL ZnFe₂O₄ deposition (c and d).

In Fig. 4, the dark field image of an individual ZnO nanorod–ZnFe₂O₄ (Fig. 4(a)) and the elemental maps of Zn–K/L, Fe–K/L and O–K (Fig. 4(b)–(f)) are illustrated. The location outlined by a square is the area for conducting the elemental map. In this technique, the relative amount and distribution of the element can be judged from the different color. Given that Fe only exists in ZnFe₂O₄, hence we can obtain the distribution of ZnFe₂O₄ from analyzing the elemental map data of Fe. From Fig. 4(d) and (e) we can see that the Fe distributes uniformly in the area centered by the ZnO nanorod, indicating the decoration of the ZnFe₂O₄ around the entire rod. In comparison with the Zn and O, the discontinued margin of Fe in the elemental map is observed. This is probably due to the slight aggregation of the ZnFe₂O₄ particles, which is consistent with the results gained from the TEM in Fig. 3(c).

Fig. 4 The STEM mapping of the ZnO–ZnFe₂O₄ composite with 3 × 60 μL ZnFe₂O₄ deposition.

The photocurrent measurements were conducted to investigate the photoelectrochemical properties of the photoanode fabricated from bare ZnO nanorods or the ZnO–ZnFe₂O₄ composite. All measurements were carried out under AM 1.5 G (100 mW cm⁻²) illumination coupled with a 420 nm light filter, in a three-electrode photoelectrochemical cell using a Pt foil as the counter electrode and a saturated calomel electrode as a reference in 0.1 M Na₂SO₄ electrolyte. Fig. 5(a) shows the photocurrent response of the samples in the wavelength longer than 420 nm. It can be seen that the bare ZnO nanorods have the minimum photocurrent in the visible light range, which increases remarkably with the deposition of ZnFe₂O₄. The results are in good agreement with the UV-vis spectra of the corresponding samples shown in Fig. 2(a), confirming that the photocurrent generates as a result of the band gap transition of ZnFe₂O_4. The fact that the sample with the highest absorption results is found in the greatest photocurrent indicates the efficient carrier separation within the thickness of the ZnFe₂O₄ tested. The results demonstrate that the ZnFe₂O₄ prepared in this work is photoactive under visible light illumination.

Fig. 5 The photocurrent response (a) and the i–t curve ((b), 0.6 V bias) in visible light (λ > 420 nm) in 0.1 M Na₂SO₄. (c) A simplified energy diagram of the heterojunction.

Additionally, as can be seen from the i–t curve in Fig. 5(b), a transient photocurrent appears when the light is switched on, implying the existence of the surface states in the ZnFe₂O₄. The deep energy level capture caused by the surface states leads to the recombination of the electron–hole pairs, and consequently an electron backward and transient photocurrent occurs. This phenomenon is also observed in the previously reported Fe₂O₃–ZnFe₂O₄ composite and can be suppressed by surface modification with Al₂O₃.²⁷ Fig. 5(c) exhibits the corresponding energy diagram of ZnO–ZnFe₂O₄. At equilibrium, the Fermi levels of ZnO and ZnFe₂O₄ will be aligned with the electrolyte solution. Most importantly, the conduction band edges of ZnFe₂O₄ are higher than those of ZnO, allowing an efficient transfer of photo-excited electrons from ZnFe₂O₄ to ZnO nanorods and finally to the FTO back contact. While the electrons are transported to the cathode for hydrogen generation in a photoelectrochemical cell device, the holes will be consumed for the oxidation of water at the ZnFe₂O₄ electrolyte interface.

To better understand the effect of the heterojunction on the series charge transfer in the photoelectrochemical process, the electrochemical impedance spectra (EIS) were measured under visible light illumination in 0.1 M Na₂SO₄. The EIS spectra were simulated by using an equivalent circuit R_s(CPE − R_p), where R_s is the ohmic contribution, CPE is the constant phase element that takes into account non-idealities in the capacitance of the Helmholtz layer, and R_p is the charge-transfer resistance. For comparison, the electrochemical impedance spectra of bare ZnFe₂O₄ film and the ZnO–ZnFe₂O₄ composite with the same ZnFe₂O₄ thickness of ∼1.0 μm are shown in Fig. 6. It can be clearly seen that the series charge transfer resistance of the bare ZnFe₂O₄ film is one order of magnitude higher than that of the ZnO–ZnFe₂O₄ composite, which are 266 600 and 21 763 Ω, respectively. The series resistances for ZnFe₂O₄ and the ZnO–ZnFe₂O₄ film are 67.4 and 37.0 Ω, respectively. These demonstrate that introducing ZnO can greatly promote the charge transfer rate of ZnFe₂O₄. We believe this is due to the beneficial effect of the heterojunction constructed by ZnO and ZnFe₂O₄ and the rational band alignment between them. Once the ZnO–ZnFe₂O₄ composite gets illuminated, the photo-excited electron–hole pairs near the interface of ZnO and ZnFe₂O₄ separate rapidly and flow in an opposite direction, resulting in a lower charge transfer resistance and consequently an enhanced photocurrent.

Fig. 6 The electrochemical impedance spectra of the ZnFe₂O₄ film and the ZnO–ZnFe₂O₄ heterojunction film with 3 × 60 μL ZnFe₂O₄ deposition, measured under visible light illumination and 0.8 V bias. The line is the value obtained by the simulation using an equivalent circuit R_s(CPE − R_p). The square symbol represent the experimental data.

In summary, the water-soluble, mono-dispersed ZnFe₂O₄ nanoparticles with an average size of 6 ± 0.5 nm are prepared. The ZnFe₂O₄ nanoparticles are loaded onto the ZnO nanorods for fabricating the ZnO–ZnFe₂O₄ heterojunction film. In comparison to the bare ZnO nanorods, the photocurrent is enhanced as a result of the visible light response of the ZnFe₂O₄ layer, and the improvement of the charge separation in the ZnO–ZnFe₂O₄ heterojunction is determined by the electrochemical impedance spectra.

Acknowledgements

This work was financially supported by the Science and Technology Support Program of Gansu Province (no. 1304GKCA038) and the Natural Science Foundation of Gansu Province in China (no. 1308RJYA043).

Notes and references

1. T. E. Mallouk, Nat. Chem., 2013, 5, 362–363.
2. J. A. Turner, Science, 1999, 285, 687–689.
3. A. Ramírez, P. Bogdanoff, D. Friedrich and S. Fiechter, Nano Energy, 2012, 1, 282–289.
4. A. Fujishia and K. Honda, Nature, 1972, 238, 37–38.
5. H. Q. Wang, L. C. Jia, P. Bogdanoff, S. Fiechter, H. Möhwald and D. Shchukin, Energy Environ. Sci., 2013, 6, 799–804.
6. X. J. Feng, L. Feng, M. H. Jin, J. Zhai, L. Jiang and D. B. Zhu, J. Am. Chem. Soc., 2004, 126, 62–63.
7. X. Y. Yang, A. Wolcott, G. M. Wang, A. Sobo, R. C. Fitzmorris, F. Qian, J. Z. Zhang and Y. Li, Nano Lett., 2009, 9, 2331–2336.
8. D. D. Qin, C. L. Tao, S. I. In, Z. Y. Yang, T. E. Mallouk, N. Z. Bao and C. A. Grimes, Energy Fuels, 2011, 25, 5257–5263.
9. P. P. Shangguan, S. P. Tong, H. L. Li and W. H. Leng, RSC Adv., 2013, 3, 10163–10167.
10. D. D. Qin, C. L. Tao, S. A. Friesen, T. H. Wang, O. K. Varghese, N. Z. Bao, Z. Y. Yang, T. E. Mallouk and C. A. Grimes, Chem. Commun., 2012, 48, 729–731.
11. J. Z. Su, L. J. Guo, S. Yoriya and C. A. Grimes, Cryst. Growth Des., 2010, 10, 856–861.
12. W. T. Sun, Y. Yu, H. Y. Pan, X. F. Gao, Q. Chen and L. M. Peng, J. Am. Chem. Soc., 2008, 130, 1124–1125.
13. H. Wang, Y. Bai, H. Zhang, Z. Zhang, J. Li and L. Guo, J. Phys. Chem. C, 2010, 114, 16451–16455.
14. M. Seol, H. Kim, W. Kim and K. Yong, Electrochem. Commun., 2010, 12, 1416–1418.
15. S. Emin, M. Fanetti, F. F. Abdi, D. Lisjak, M. Valant, V. D. R. Krol and B. Dam, ACS Appl. Mater. Interfaces, 2013, 5, 1113–1121.
16. X. N. Wang, H. J. Zhu, Y. M. Xu, H. Wang, Y. Tao, S. K. Hark, X. D. Xiao and Q. Li, ACS Nano, 2010, 4, 3302–3308.
17. S. Nagaya and H. Nishikiori, ACS Appl. Mater. Interfaces, 2013, 5, 8841–8844.
18. F. E. Sarac, C. Yilmaz, F. Y. Acar and U. Unal, RSC Adv., 2012, 2, 10182–10184.
19. Y. C. Qiu, K. Y. Yan, H. Deng and S. H. Yang, Nano Lett., 2012, 12, 407–413.
20. C. Z. Yao, B. H. Wei, H. X. Ma, H. Li, L. X. Meng, X. S. Zhang and Q. J. Gong, J. Power Sources, 2013, 237, 295–299.
21. X. P. Qi, G. W. She, Y. Y. Liu, L. X. Mu and W. S. Shi, Chem. Commun., 2012, 48, 242–244.
22. M. G. Zhao, J. Y. Huang, Y. Zhou, X. H. Pan, H. P. He, Z. Z. Ye and X. Q. Pan, Chem. Commun., 2013, 49, 7656–7658.
23. W. Q. Zhang, M. Wang, W. J. Zhao and B. Q. Wang, Dalton Trans., 2013, 42, 15464–15474.
24. C. W. Yao, Q. S. Zeng, G. F. Goya, T. Torres, J. F. Liu, H. P. Wu, M. Y. Ge, Y. W. Zeng, Y. W. Wang and J. Z. Jiang, J. Phys. Chem. C, 2007, 111, 12274–12278.
25. Z. C. Liu, B. Li, B. D. Wang, Z. Y. Yang, Q. Wang, T. R. Li, D. D. Qin, Y. Li, M. F. Wang and M. H. Yan, Dalton Trans., 2012, 41, 8723–8728.
26. Y. Y. Bu, Z. Y. Chen and W. B. Li, Dalton Trans., 2013, 42, 16272–16275.
27. K. J. McDonald and K. S. Choi, Chem. Mater., 2011, 23, 4863–4869.

---

Footnote

† Electronic supplementary information (ESI) available: other details are given in the supporting information. See DOI: 10.1039/c4ra00204k

---