Report

Photochemical Route for Accessing Amorphous Metal Oxide Materials for Water Oxidation Catalysis

See allHide authors and affiliations

Science  05 Apr 2013:
Vol. 340, Issue 6128, pp. 60-63
DOI: 10.1126/science.1233638

Amorphous and More Active

The electrochemical generation of hydrogen from water could help in the storage of energy generated by renewable resources at off-peak times. However, catalysts for the slow step of this reaction, the oxygen evolution reaction (OER), are based on oxides of noble metals (iridium and ruthenium) that have limited abundance. A strategy for improving the performance of earth-abundant elements is to explore mixed-metal oxides and to prepare these as amorphous phases. Smith et al. (p. 60, published online 28 March) developed a general method for preparing amorphous oxides, based on photodecomposition of organometallic precursors. Amorphous mixed-metal oxides of iron, nickel, and cobalt were more active than comparable crystalline materials and provided OER performance comparable to noble metal oxides.

Abstract

Large-scale electrolysis of water for hydrogen generation requires better catalysts to lower the kinetic barriers associated with the oxygen evolution reaction (OER). Although most OER catalysts are based on crystalline mixed-metal oxides, high activities can also be achieved with amorphous phases. Methods for producing amorphous materials, however, are not typically amenable to mixed-metal compositions. We demonstrate that a low-temperature process, photochemical metal-organic deposition, can produce amorphous (mixed) metal oxide films for OER catalysis. The films contain a homogeneous distribution of metals with compositions that can be accurately controlled. The catalytic properties of amorphous iron oxide prepared with this technique are superior to those of hematite, whereas the catalytic properties of a-Fe100-y-zCoyNizOx are comparable to those of noble metal oxide catalysts currently used in commercial electrolyzers.

The scalable storage of renewable energy by means of converting water to hydrogen fuels (H2) electrochemically hinges on fundamental improvements in catalytic materials. A large overpotential (η) is usually required to produce H2 at a practical rate, which is primarily the result of slow oxygen (O2) evolution kinetics. Despite recent advances in the development of heterogeneous catalysts to negotiate the OER (Eq. 1; is the standard potential and NHE is the normal hydrogen electrode) (14), substantial market penetration by commercial electrolyzers has been hindered by the absence of inexpensive catalytic materials that exhibit high current densities (j) (>0.5 A cm−2) at low η (<0.3 V) over prolonged time periods.

2H2O(l) → O2(g) + 4H+(aq) + 4e E° (O2/H2O) = 1.23 V versus NHE (1)

Metal oxides are the most durable and active water oxidation catalysts (1, 2). Although IrO2 and RuO2 are among the best OER catalysts (5, 6), a myriad of perovskite (3, 7) and spinel (8) solids have proven to be competent catalysts. In recent years, amorphous metal oxides have also been demonstrated to be excellent OER catalysts (4, 9), including when integrated with photoactive electrodes (1012). Although an increasing number of amorphous metal oxide catalysts have been reported (4, 9, 1317), most have been produced by electrodeposition techniques—a methodology that does not necessarily translate to every metal. Moreover, amorphous film deposition can be sensitive to the voltage protocol, thereby rendering few examples of amorphous OER catalysts composed of multiple metal identities. Considering that the vast majority of crystalline catalysts consist of more than a single metallic element, there are anticipated benefits to developing amorphous metal oxide catalysts of more complex metal compositions. Following this line of inquiry, we show herein that photochemical metal-organic deposition (PMOD) (18) is a facile technique for preparing amorphous phases of (mixed) metal oxides that generate high OER electrocatalytic activity. This proof-of-principle investigation of amorphous iron oxide (a-Fe2O3) films and mixed-metal compositions (a-Fe100-y-zCoyNizOx) demonstrates the broad applicability of this methodology for accessing new OER catalysts.

Iron(III) 2-ethylhexanoate was used as a PMOD precursor for making a-Fe2O3 thin films (19). Thin films of optical quality were prepared by spin coating a hexane solution of the precursor onto fluorine-doped tin oxide (FTO) (for electrochemical and spectroscopic characterization) or silicon (for electron microscopy and elemental analyses). These films were then irradiated with 185- and 254-nm light until the vibrational stretching frequencies of the 2-ethylhexanoate ligands could no longer be detected by infrared spectroscopy (fig. S1) (20). The resultant a-Fe2O3 films were annealed at temperatures (Tanneal) up to 600°C. The thicknesses of the a-Fe2O3 films were measured by means of cross-sectional scanning electron microsocopy (SEM) imaging (Fig. 1). For Tanneal ≤ 200°C, films had a thickness of 150 to 200 nm; the thickness gradually decreased at progressively higher Tanneal values, reaching 100 to 150 nm at 500° to 600°C. Smooth, homogeneous films with only minor microscale defects were observed for Tanneal ≤ 200°C; higher Tanneal led to a greater degree of film cracking and an increase in apparent porosity (fig. S2) (20). The greater degree of defects presumably arises upon contraction of the film when transitioning from an amorphous to a crystalline structure.

Fig. 1

SEM cross-section and top-down surface images of a-Fe2O3 films prepared by PMOD, followed by a 1-hour annealing step in air at Tanneal = 100°, 400°, or 600°C.

X-ray diffraction (XRD) (Fig. 2 and fig. S3) (20) revealed no evidence for a crystalline phase at Tanneal < 500°C. Films heated to higher temperatures produced diffraction peaks congruent with those of hematite (α-Fe2O3). These results are consistent with an extended x-ray absorption fine structure analysis of as-prepared a-Fe2O3 films (19). Ultraviolet-visible (UV-vis) diffuse reflectance spectra recorded on films subjected to Tanneal ≥ 500°C revealed optical profiles resembling that of hematite (fig. S4) (20, 21). Films treated at Tanneal ≤ 200°C rendered spectral features distinctive from that of hematite. Optical data for films annealed at 300° and 400°C contained features that appear to be a hybrid of the two phases. These collective data show that although the structure of the amorphous material is not sufficiently coherent to produce observable Bragg reflections until Tanneal ≥ 500°C, the UV-vis data indicate that hematite may begin to form at lower temperatures. Hence, the electrochemical data presented below are on films annealed at 100°C to ensure that the catalytic activity originates exclusively from a-Fe2O3.

Fig. 2

XRD powder patterns acquired on as-prepared (i.e., no annealing step) and annealed (Tanneal indicated) Fe2O3 films. Bragg reflections for hematite are observed for films annealed at Tanneal > 500°C). Patterns for hematite [Joint Committee on Powder Diffraction Standards (JCPDS) card 33-664] and SnO2 (JCPDS card 41-1445) are also shown.

The electrochemical behavior of a-Fe2O3 and hematite is shown in Fig. 3, A and B. An oxidative sweep of the crystalline film revealed a small rise in current at 1.50 V versus RHE (reversible hydrogen electrode) that was maintained until a sharp increase in current at ~1.62 V, corresponding to catalytic water oxidation. The cyclic voltammogram for a-Fe2O3 was also featureless until the onset of catalytic water oxidation at ~1.50 V. The catalytic wave was independently confirmed to be associated with water oxidation on the basis of O2 evolution experiments (fig. S5) (20). Steady-state current densities (j) as a function of η were recorded to probe the kinetics of the reactions. Once a threshold potential was reached, log(j) exhibited a sharp increase and a linear dependence on η where the current density became limited by electron-transfer kinetics at the electrode surface. This experiment reveals useful electrode kinetic metrics, including the onset of linearity (Ecat) and the Tafel slope (Fig. 3B and fig. S6) (20). Ecat was observed at 1.55 ± 0.02 V (η = 0.32 V) for a-Fe2O3, corresponding to a 60-mV improvement in η relative to that of hematite. Tafel slopes of ~40 mV dec−1 were observed for both materials, and a current density of 0.5 mA cm−2 was reached at 1.63 ± 0.02 V for a-Fe2O3 (compared with 1.70 ± 0.06 V for hematite), thus demonstrating the superior electrocatalytic activity of the amorphous phase relative to the crystalline phase.

Fig. 3

Electrochemical behavior for oxide thin films prepared from 15% w/w precursor solutions by the PMOD technique, followed by annealing at 100°C. (A) Cyclic voltammograms for films of a-Fe2O3 (blue) and hematite (gray), and a blank fluorine-doped tin oxide (FTO) substrate. (B) Tafel plot showing the higher catalytic activity of a-Fe2O3 relative to hematite. Comparison of electrochemical behavior for thin films of a-Fe2O3, a-Fe50Ni50Ox, a-Fe50Co50Ox, and a-Fe33Co33Ni33Ox films: (C) cyclic voltammograms; (D) Tafel plots; (E) onset potentials (Ecat) and potentials required to reach j = 0.5 mA cm−2; and (F) Tafel slopes for the various catalyst films. Electrochemistry conditions: counterelectrode = Pt mesh; reference electrode = Ag/AgCl, KCl(sat’d); scan rate = 10 mV s−1; electrolyte = 0.1 M KOH(aq); current densities were corrected for uncompensated resistance. Dashed red lines correspond to the thermodynamic potential for water oxidation (26). Error bars indicate the SD between multiple electrodes (three minimum).

We next used PMOD to fabricate amorphous mixed-metal oxides of iron, nickel, and cobalt (19, 22) in pursuit of improved catalysts. Precursor solutions were prepared with the appropriate stoichiometries to produce a-Fe50Co50Ox, a-Fe50Ni50Ox, and a-Fe33Co33Ni33Ox upon photolysis (table S1) (20); a-CoOx and a-NiOx films were also prepared by PMOD to serve as benchmark materials. The UV-vis diffuse reflection spectra of the mixed-metal films (fig. S7) (20) reveal poorly resolved absorption features that do not match those of the known crystalline phases of the Ni or Co oxides (2325). The SEM images of the mixed-metal oxide films (fig. S8) (20) revealed a smooth surface akin to that observed for a-Fe2O3 annealed at low temperatures; the a-CoOx and a-NiOx films revealed less uniform film morphologies. Energy dispersive x-ray spectroscopy verified the elemental composition of the films at six distinct points of each sample to be within 2 atomic % (table S2) (20). The experimentally determined compositions of each of the three mixed-metal oxide films were in excellent agreement with the corresponding solution stoichiometries. These collective features indicate that the films consist of a homogeneous distribution of metal ions throughout the solid. Unlike what may arise with conventional metal oxide formation schemes (e.g., thermal decomposition, coprecipitation), the low-temperature PMOD process does not appear to lead to phase segregation.

Both a-CoOx and a-NiOx were, as expected, better OER catalysts than a-Fe2O3 (Fig. 3, C to F). Cyclic voltammetry (Fig. 3C) revealed additional electron-transfer processes that occur before water oxidation for all three mixed-metal oxide films. The current spike indicating catalytic water oxidation began between 1.41 and 1.44 V (η = 0.18 to 0.21 V) for the three materials. Moreover, Tafel slopes between 24 and 33 mV dec−1 enabled these materials to reach 0.5 mA cm−2 at potentials as low as 1.48 V (η = 0.25 V). Steady-state electrochemistry measurements on the mixed-metal films highlight an improvement in kinetics of water oxidation compared to those displayed by each of the monometallic amorphous phases (Fig. 3D). Although the onsets of linearity were similar for a-NiOx, a-CoOx, and each of the mixed-metal oxides, the mixed-metal compositions containing Fe were characterized by a lower Tafel slope and are therefore more efficient electrocatalysts at higher current densities (table 1) (26). The stabilities of the films, which are inherently sensitive to film composition, are also reasonably high at a current density of 1 mA cm−2 (e.g., a mere 6- and 30-mV increase in electrode potential was required to maintain a constant current density for a-NiOx and a-Fe2O3, respectively; fig. S9) (20). Each mixed-metal composition in this first generation of OER catalysts produced by PMOD exhibits catalytic parameters that approach those of the most active catalysts in the literature (Table 1). Given the broad applicability of this approach and the acute stoichiometric control of the metal compositions, we contend that the PMOD technique opens an entirely new parameter space for discovery and optimization of heterogeneous electrocatalysts.

Table 1

Comparison of catalytic parameters of amorphous and crystalline metal oxide OER catalysts.

View this table:

Supplementary Materials

www.sciencemag.org/cgi/content/full/science.1233638/DC1

Materials and Methods

Figs. S1 to S9

Tables S1 and S2

References and Notes

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20. Supporting online information.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26. The higher precatalytic current densities for a-CoOx relative to the mixed-metal oxides are ascribed to disparate geometric surface areas; SEM images indicate a higher relative surface roughness for a-CoOx.
  27. 27. Fabricated by PMOD and reported in this Report.
  28. 28.
  29. 29.
  30. 30.
  31. Acknowledgments: We thank Natural Sciences and Engineering Research Council of Canada, Mitacs, and FireWater Fuel Corp. for financial support. C.P.B. also thanks Canada Research Chairs and the Alfred P. Sloan Foundation for support. This research used facilities funded by the University of Calgary and the Canadian Foundation for Innovation. We thank R. Marr for running XRD experiments. S.T. and C.P.B. proposed the concept, designed the experiments, and supervised the project. R.D.L.S. and M.S.P. carried out electrochemical, structural, and optical characterization. Z.Z., P.A.S., and M.K.J.S. conducted preliminary experiments. R.D.F. carried out microscopy experiments. The authors declare competing financial interests: intellectual property pertaining to the technology described in this Report is protected by patent application PCT/CA2012/050609.
View Abstract

Navigate This Article