nife‐based (oxy)hydroxide catalysts for oxygen evolution ......nife-based (oxy)hydroxide catalysts...

Rev

iew

© 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim (1 of 20) 1600621wileyonlinelibrary.com

NiFe-Based (Oxy)hydroxide Catalysts for Oxygen Evolution Reaction in Non-Acidic Electrolytes

Fabio Dionigi and Peter Strasser*

Dr. F. Dionigi, Prof. Dr. P. StrasserThe Electrochemical EnergyCatalysis, and Materials Science LaboratoryDepartment of ChemistryChemical Engineering DivisionTechnical University Berlin10623 Berlin, GermanyE-mail: [email protected]. Dr. P. StrasserErtl Center for Electrochemistry and CatalysisGwangju Institute of Science and TechnologyGwangju 500–712, South Korea

DOI: 10.1002/aenm.201600621

The efficiency of the electrolysis process depends on several factors, among which the sluggish kinetic of water oxidation to oxygen, presents a considerable challenge. This reaction is central in natural photo-synthesis, where it occurs at the oxygen-evolving center (OEC).[1,2] This catalytic unit is driven by the light absorbed by photosystem II (PSII) and oxidizes water into oxygen and reducing equivalents.[1,3,4] The OEC shows high efficiencies for water oxidation and an excellent estimated turn-over frequency.[4,5] Achieving such desir-able performances outside the protein environment that surrounds the OEC is difficult, due to the harsh oxidizing con-ditions and the complicated mechanism of the oxygen evolution reaction (OER) which makes the optimization of the cata-lyst problematic (see section 6 for further discussion).[6] Several catalysts have been

proposed to reduce the overpotential losses for OER and inves-tigated in different pH conditions, from acid to alkaline and in neutral pHs with buffer addition.[5,7,8] Among them, transition-metal oxides and hydroxides have proven so far to be the class of materials that show the best performances for OER. For example, IrO2 is often considered the benchmark catalyst for OER thanks to its superior stability and excellent activity both in acid and basic pH.[8] However, IrO2 still shows a considerable OER overpotential when compared to the overpotential required by the best catalysts to achieve similar currents for the other half cell reaction, the hydrogen evolution. Furthermore, due to the Ir limited resources and its high price several researchers are looking for strategies to reduce its amount without losing performances (alloy, core/shell particles)[9] or alternative cata-lysts based on earth abundant elements.[5,7,8] In alkaline and near neutral electrolytes, oxides and hydroxides of late first row transition metals (Mn, Fe, Co, Ni) show promising perfor-mances and so they are the subject of deep investigation.[6,10,11] In particular, hydroxides incorporating both Ni and Fe together are reported to show the lowest overpotential for OER in alka-line conditions (pH 13 and 14) and be active also at near neu-tral pH in borate buffer.[12–16] In addition, it has recently been shown that they can be employed as precursors to derive other NiFe-based oxides,[17] nitrides[18] and sulfides[19] with interesting electrocatalytic performances for hydrogen evolution reaction or overall water splitting. Thanks to their high OER activity, NiFe-based hydroxides are investigated also as surface modification of photoanodes for photoelectrochemical water splitting[20–22] and for seawater oxidation.[16] Corrigan firstly investigated their

NiFe-based (oxy)hydroxides are highly active catalysts for the oxygen evolu-tion reaction in alkaline electrolyte solutions. These catalysts can be synthe-sized in different ways leading to nanomaterials and thin films with distinct morphologies, stoichiometries and long-range order. Notably, their structure evolves under oxygen evolution operating conditions with respect to the as-synthesized state. Therefore, many researchers have dedicated their efforts on the identification of the catalytic active sites employing in operando exper-imental methods and theoretical calculations. These investigations are pivotal to rationally design materials with outstanding performances that will consti-tute the anodes of practical commercial alkaline electrolyzers. The family of NiFe-based oxyhydroxide catalysts reported in recent years is addressed and the actual state of the research with special focus on the understanding of the oxygen-evolution-reaction active sites and phase is described. Finally, an overview on the proposed oxygen-evolution-reaction mechanisms occurring on NiFe-based oxyhydroxide electrocatalysts is provided.

1. Introduction

The transition to a society where large part of the world energy demand is provided by renewable energy requires storage solutions due to the intermittency nature of renewable energy resources, either wind or sunlight. When the demand matches the supply from these sources, the energy can be captured and directly converted into electrical energy for example by solar cells or wind turbines. But excess electricity that is not imme-diately utilized should be stored to compensate future tempo-rary decrease of energy supply. A promising strategy to address this issue consists in storing the excess of electrical energy in the form of chemical energy by splitting the water in hydrogen and oxygen. Following this process, electrical energy can be re-obtained when needed by, for example, fuel cell technologies.

Adv. Energy Mater. 2016, 6, 1600621

www.MaterialsViews.comwww.advenergymat.de

http://doi.wiley.com/10.1002/aenm.201600621

Rev

iew

© 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim1600621 (2 of 20) wileyonlinelibrary.com

catalytic activity for the OER in the 1980s.[23–25] In more recent years, Gong et al. presented in their review the early discovery and recent progress on NiFe-based OER catalysts comprising oxides, metal derived oxides and (oxy)hydroxides.[26] In addi-tion, for the special class of layered double hydroxide (LDH) materials, a comparison of recent LDH based OER catalysts, including NiFe LDH, has been presented by Shao et al.[27] Recently, the understanding of NiFe-based (oxy)hydroxide under OER condition has improved considerably in large parts thanks to the availability of advanced in operando characteriza-tion techniques available and density functional theory calcula-tions. Therefore, we will focus our review on (oxy)hydroxides containing both Ni and Fe and the recent insights obtained in the last five years about the active state of these catalysts during the oxygen evolution reaction (Figure 1).

2. The Structure of NiFe-Layered Double Hydroxide

2.1. Crystal Structure and Phase Transformations of Ni(OH)2

The structure and surface electrochemical redox behavior of Ni-Fe mixed hydroxides can be conveniently understood by introducing Fe-free Nickel hydroxides. Nickel hydroxide, Ni(OH)2, where Ni is in his most common oxidation state of +2, is a stable compound in alkaline conditions in the pH range between ≈9 and ≈13 when considering the potential window of hydrogen-reduction/water-oxidation according to Pourbaix diagram and consistent with recent density func-tional theory calculations by Zeng et al.[28] Oxidation and deprotonation of Ni(OH)2 leads to compounds of the type NiOxHy, where the O/H ratio and consequently the average Ni oxidation state can in general assume several values. For example, full oxidation of Ni from +2 to +3, can be described by the redox transformation:

Ni OH OH NiOOH + H O + e2 2–( ) + →−

where in NiOOH the Ni oxidation state is +3. Bode et al. pro-posed four phases as limiting cases and models for divalent and trivalent Ni hydroxides/oxyhydroxides obtained by elec-trodeposition (Figure 2).[29] Even though developed for Fe-free Ni hydroxides, this model continues to be an important basis for understanding the electrochemical behavior of NiFe-based (oxy)hydroxides. The transformations among the phases are described in the model, usually referred to as Bode's diagram. This was analyzed extensively in many other works, including a perspective by Doyle et al. and a recent work by Klaus et al.[30,31] All the four phases are organized in layers of edge sharing octahedra with Ni metal center and O/OH groups at the ver-tices pointing out on both side of the layer along the stacking direction (crystallographic c-axis). The α-Ni(OH)2 and γ-NiOOH phases, where the average Ni oxidation state is ≈2 and ≈3.5–3.66 respectively, are hydrated phases, with water intercalated in the layers. Typically, the α-Ni(OH)2 phase is reported as a disor-dered phase, while the γ-NiOOH phase has the characteristic to incorporate metal cations between the layers. The incor-poration of cations has been proposed from the increase in

mass observed in electrochemical quartz crystal microbalance (EQCM) measurements.[32] On the contrary, the β-Ni(OH)2 and β-NiOOH phases are not hydrated and do not incorporates ions between the layers. Their average Ni oxidation state is ≈2 and ≈3, respectively. The distance between the layers changes signif-icantly between the water intercalated phases (α and γ) and the phases that are not hydrated (the two β). In addition to these four phases, other crystal structures with different interlayer distances were observed and identified as interstratified mixed α and β phases where partial water intercalation results in vari-able distance between the layers.[33]

The structure of β-Ni(OH)2 is the most understood. Figure 3a shows a drawing of the structure produced by using the software Vesta.

[34] In the β-Ni(OH)2, the layers are stacked on top of each following a ABAB oxygen stacking. The upward and downward pointy OH groups of each layer are aligned with the respective OH groups of the adjacent layers above and below. In this case the oxygen of an OH group occupies the corner of a tetrahedral structure in the interlayer gap with three oxygen atoms of the adjacent layer occupying the other

Peter Strasser is the chaired professor of “Electrochem-istry and Electrocatalysis” in the Chemical Engineering Division of the Department of Chemistry at the Technical University Berlin. He was Assistant Professor at the Department of Chemical and Biomolecular Engineering at the University of Houston, after he served as Senior

Member of staff at Symyx Technologies, Inc. He earned his PhD in Physical Chemistry and Electrochemistry from the ‘Fritz-Haber-Institute’ of the Max-Planck-Society in Berlin under the direction of Gerhard Ertl. He studied chemistry at Stanford University, USA, the University of Tuebingen, Germany, and the University of Pisa, Italy.

Fabio Dionigi completed his PhD in Physics in 2012 at Danmarks Tekniske Universitet (DTU), Denmark. Then he continued in the same university as a postdoc at DTU Nanotech, Department of Micro- and Nanotechnology, till November 2013. Since then he has joined “The electro-chemical energy, catalysis,

and materials science group” in the Department of chem-ical engineering at Technische Universität Berlin, Germany. His work is focused on photocatalytic water splitting and more recently on non-noble metal based catalysts for the oxygen evolution reaction.



Rev

iew


three corners (T1 oxygen stacking).[35] Brucite-like β-Ni(OH)2 has a trigonal crystal symmetry with hexagonal lattice and space group P-3m1. The layers in β-Ni(OH)2 are held together by van der Waals forces and no hydrogen bonds are observed, while hydrogen bonds are observed for the other phases.[36] The struc-ture of the oxidized phases, β-NiOOH and γ-NiOOH, are sub-ject to debate. Van der Ven et al. proposed for the β-NiOOH a trigonal prismatic arrangement of the oxygen ions between

two adjacent layer (P3 oxygen stacking) in contrast with the tetrahedral arrangement of the β-Ni(OH)2 (Figure 3).[35] In this configuration hydrogen bonds can form connecting the layers. Following this work, Casas-Cabanas et al. performed a high-res-olution transmission electron microscopy (HRTEM) study on a chemically oxidized β-NiOOH.[37] They observed doubling of the c-axis respect to β-Ni(OH)2 and confirmed an irreversible for-mation of slightly misoriented crystalline domains.[38] Based on their observation, they proposed that layer shearing is occurring in the transition to β-NiOOH that this is resulting in a structure with ABCA oxygen stacking and a C2/m space group. The mono-clinic structure was also considered by Kazimirov et al., who proposed two models for the β-NiOOH, with P3m1 space group and monoclinic Cm group, respectively.[39] On the other hand, the ABCA stacking was predicted to be unstable by Li et al.[40] In addition, they proposed formation of tunnel structures at the edge of the domains using genetic algorithm based calculations and suggested that disproportionation of half of the Ni3+ into Ni2+ and Ni4+ could occur in β-NiOOH. Recently, Tkalych et al. proposed based on density functional theory calculations that β-NiOOH has a structure with a staggered arrangement of pro-tons (protons are present on both sides of each layer) and that the Jahn-Teller distortion of the low spin Ni3+ centered octa-hedral might be the explanation for two different Ni-O bond distances experimentally observed for this phase.[41,42] In con-clusion, there is not a general agreement on the structure of β-NiOOH. Similarly controversial is the structure of γ-NiOOH. For example, Van der Ven et al. proposed a monoclinic unit cell with C2/m space group or a structure with R-3m space group symmetry as possible candidates.[35] In both cases the oxygen



Figure 1. Scheme illustrating the active research areas and techniques in the field of NiFe-based (oxy)hydroxide catalysts for the oxygen evolution reaction in non-acidic electrolytes.

Figure 2. Bode’s diagram. The interlayer distances and average Ni oxi-dation states are shown for the four phases. Reproduced with permis-sion.[30,29,32] Copyright 2013, Royal Society of Chemistry.

Rev

iew


stacking is of P3 type, forming trigonal prismatic structure in the interlayer gap. Contrasting experimental results based on γ-NiOOH obtained by chemical methods from NaNiO2 precur-sors support either one or the other structure.[43,44]

Among the four phases described by Bode et al., two distinct redox transformations are obtained by potential cycling across the Ni(II)/(III,IV) redox potential: the charge/discharge cycle between α-Ni(OH)2 and γ-NiOOH (α/γ) and between β-Ni(OH)2 and β-NiOOH (β/β). The oxidation of α to γ is reported to occur at lower potential than the one from β(II) to β(III).[30,31] Departures from these two distinct transformations have been reported by aging and overcharging. Aging, especially in elec-trolyte with high alkaline pH values, has been found to allow the transformation of α-Ni(OH)2 to β-Ni(OH)2. Increasing the potential above the charging of β-Ni(OH)2 to β-NiOOH leads to the transformation of β-NiOOH to γ-NiOOH (overcharging). This model is broadly accepted and has been recently demon-strated to hold in the case of electrodeposited Ni(OH)2 on Au electrode that was aged in rigorous Fe-free purified KOH elec-trolyte.[31] In contrast, it was found that aging in unpurified electrolyte (containing traces of Fe) leads to Fe incorporation and inhibitions of the two transformations caused by aging and overcharging. In addition to this effect, the incorporation of Fe has been shown to improve dramatically the activity of Ni(OH)2 and the surface redox chemistry.[25,45]

2.2. Crystal Structure of NiFe-Layered Double Hydroxide

Incorporation of Fe3+ in the neutral layers of Ni(OH)2 leads to positive charged layers that are charge-compensated by anion intercalation in the interlayer region. For example, carbonate

anions will be present in half the amount of the incorporated Fe3+ due to their double negative charge. Water also intercalates in the structure forming hydrogen bonds with the hydroxide groups. The compound can be expressed with the formula

Ni Fe OH (A ) y H O12 3

2 / 2( ) ( ) ⋅−+ + + −

x x

x nx n

where x is the molar ratio of Fe that is incorporated, A is the anion, n is the charge of the anions and y the amount of water that is intercalated.

The crystal structure formed is the same as in the mineral hydrotalcite and is often denoted in the literature as layered double hydroxide (LDH). Figure 3 shows a comparison between the crystal structure of β-Ni(OH)2, that is isostructural with brucite, Mg(OH)2, and the structure of NiFe LDH, isostruc-tural with hydrotalcite, where Al3+ substitute Mg2+ in layers. For NiFe LDH, Mössbauer spectroscopy[24] and extended X-ray absorption fine structure (EXAFS)[46] indicate that Fe is indeed substitutionally incorporated in the Ni host lattice, due to strong correlation between Ni and Fe features. The maximum amount of Fe that can be incorporated while preserving the LDH structure is not clear. The comparison with other hydro-talcite-like compounds suggests that the Fe content of ≈33%, corresponding to Ni/Fe ratio of 2:1, is the highest limit, after which the unavoidable formation of near neighbor Fe octahe-dral becomes the basis for the formation of FeOxHy phases.[47] For this specific ratio between the divalent and trivalent cations, direct nearest neighboring of the trivalent cation, i.e., Fe3+, is avoided for a perfect orderly distributed LDH, as discussed by Ma et al. for CoFe LDH.[48] A previous work by Sideris et al. demonstrated ordering of the cations in a honeycomb arrange-ment for Mg2Al1 LDH therefore excluding random cations



Figure 3. As-prepared crystal structures. The unit cell (black lines) and layers with four metal atoms each are shown for the fully protonated form of (a) Brucite-like β-Ni(OH)2 and (b) hydrotalcite-like NiFe LDH with intercalated water and carbonate anions (randomly distributed). The respective oxygen stacking across the interlayer region is shown in (c) and (d). In (e) the top-view along the c-axis of NiFe LDH shows three kinds of OH sites.

Rev

iew


arrangement, with only a minor presence of misplaced cationic defects, that are OH surrounded by three Mg cations, of <3%.[49] Cadars et al. later confirmed the presence of this small amount of defects and found a corresponding low amount of trivalent Al-Al clustering.[50] Recently, Abellán et al. also excluded a com-pletely random distribution of cations in NiFe LDH that was synthesized by homogeneous precipitation and found evidence of increasing Fe-Fe clustering with increasing iron content.[51] The arrangement of the cations is important because it deter-mines the kind of hydroxide (or oxygen) sites that might act as the active site for the oxygen evolution reaction. In LDH struc-ture we can distinguish 3 kinds of hydroxides sites, bounded respectively to 3, 2 and 1 metal centers and denoted as μ3, μ2 and μ1 (Figure 3e). We notice that the μ3 site is present only at the central area of the layers, while the μ2 and μ1 are located on the edges. Considering two different metal cations, Fe and Ni, we obtain a maximum of four sites of the kind μ3 (Mg3-OH, Fe3-OH, Mg2Fe-OH, MgFe2-OH), three sites of μ2 (Mg2-OH, Fe2-OH, MgFe-OH) and two μ1 sites (Mg-OH, Fe-OH), if we exclude influence of superior order coordination shells. There-fore the absence (or presence) of Fe-Fe neighbor atoms has a dramatic effect on the kind of hydroxide sites allowed. In gen-eral, increasing Fe content in the catalyst preparation often leads to formation of FeOxHy domains or small FeOxHy nano-particles that are difficult to detect by XRD methods available in standard laboratories. These phases are experimentally observed to start forming at Fe content lower than 33% and shown to cause decrease in OER activity.[31,46] The different values estimated from the experiments for the solubility limit of Fe in the Ni(OH)2 host might imply its dependence on the preparation method.[31] Besides cations arrangement, the incor-poration of Fe and the concomitant intercalation of water and anions also affects the stacking of the layers respect to the bru-cite-like Ni(OH)2.[47,52] Every layer of the NiFe LDH is shifted in the plane that is normal to the stacking direction respect to adjacent layers in a way that the oxygen atoms facing the inter-layer gallery from two neighboring layers occupy the corners of a trigonal prismatic structure connecting the two layers.

This is in contrast with the tetrahedral structure of the brucite β-Ni(OH)2. In this trigonal configuration the hydroxyl groups of adjacent layers that are pointing in the same interlayer gal-lery are aligned on top of each other and the unit cell contains atoms from 3 layers. This particular stacking is called polytype 3R and the layers are translationally equivalent (in contrast to 2H polytype that consist in a two-layer structure obtained by rotation of 180 degree of each layer and have space group sym-metry P63/mmc).[52] It has been noted that the O-H bond are not necessary perfectly aligned with the c-axis, but are slightly tilted.[52] The stacking configuration of the LDH is thought to be originated by the necessity to accommodate anions and water and optimize the hydrogen bonding of the OH groups of the layers with the intercalated water. In this configuration the oxygen of the intercalated water as well as the one of the car-bonate anions is envisaged to be randomly distributed between the 6 equivalent positions around the OH group alignment axis. This stacking configuration results in a different unit cell respect to brucite and labeling of the X-ray diffraction (XRD) reflections between the two materials. More specifically, hydro-talcite-like NiFe LDH has trigonal crystal symmetry with rhom-bohedral lattice (3R polytype) and space group R-3m. Due to the 3-layer structure of the LDH, the reflection corresponding to the separation of adjacent layers and occurring at low 2θ angles in XRD pattern is indexed as (003) (Figure 4a), while (001) is used for brucite. The separation between the layers is affected by the nature of the intercalated anions which can widely vary including organic and inorganic anions.[53–56] The anions can be chemically exchanged[51,53,57,58] and in particular carbonate, CO3

2−, was found to be not easily anion exchanged due to stronger electrostatic interaction with the layers respect to other anions and so it is considered one of the most stable anion for NiFe LDH.[59] A method based on a acid-salt mixed solution was developed by Iyi et al.[59] to anion exchange carbonate interca-lated LDH and further successfully used also for NiFe LDH. The anion exchange of NiFe LDH with anions of larger size than car-bonate (i.e., ClO4

− or NO3−) weakens the interlayer interaction

and allows for delamination of the LDH into single layers via



Figure 4. Structure characterization and morphology of NiFe LDH and NiFe oxyhydroxides obtained from different synthesis. a) X-ray diffraction pat-terns of well crystallized NiFe LDH with intercalated CO3

2– (pattern (a)), Cl– (pattern (b)) and ClO4– (pattern (c)). Reproduced with permission.[58] Copy-

right 2008, American Chemical Society. b) TEM image of hexagonal NiFe LDH nanoplates; Reproduced with permission.[58] Copyright 2008, American Chemical Society. c) SEM image of NiFe LDH flowers; Reproduced with permission.[67] Copyright 2009, IOP Publishing. d) SEM image of NiFe flakes grown on Ni foam. Reproduced with permission.[70] Copyright 2016, Royal Society of Chemistry. (e) SEM image of a cathodically electrodeposited NiFe (oxy)hydroxide film. Reproduced with permission.[45] Copyright 2014, American Chemical Society.

Rev

iew


a liquid phase exfoliation process typically performed in forma-mide.[51,57] Another proposed method for delamination of NiFe LDH is based on intercalation of aminoundecanoic acid and on following protonation of the amine group in acid aqueous envi-ronment to produce electrostatic repulsion between the layers and the cationic amino group of the intercalated molecules.[56]

As in the case of α-Ni(OH)2 and β-Ni(OH)2, the (electro)chemical oxidation of NiFe LDH is expected to have influence on the crystal and electronic structure. The determination of the structure obtained by increasing the potential more anodi-cally than the Ni redox transition and under OER conditions is not easy due to the instability of the Ni3+ species and the requirement of in situ and in operando methods. This topic will be discussed in section 5.

3. Synthesis of NiFe-Based Oxy(hydroxide) Catalysts

NiFe LDH or more generally NiFe-based (oxy)hydroxides can be synthesized with several methods. Different methods produce materials that vary largely in crystal sizes, long-range order and morphology and might present concomitant undesired phases and different electrochemical behavior.

3.1. Co-Precipitation at Constant pH

One of the most used methods consists in the co-precipitation of the metal precursors. In particular, in the “co-precipitation at constant pH”, an aqueous solution of Ni(II) and Fe(III) metal salts is added to a second water solution that is kept under stir-ring and at a certain mild basic pH (for example 9 or 11) by addition of a base or carbonate (i.e., KOH, Na2CO3,..).[60,61] A variation of this method consists in quickly mixing a solution of the metal salts and the desired alkali solution.[55] Increasing the temperature (i.e., to 80–120 °C) by a subsequent hydro-thermal treatment favors crystallization of the materials.[55,62] The mechanism of formation was studied by Gregoire et al. by titrate an aqueous solution of the two metal precursors.[63] They proposed that slow hydrolysis of the Ni-Fe salts dissolved in solution starts with precipitation of ferric (Fe3+) oxyhydroxide, followed by adsorption of cationic complexes like [Ni(OH)(H2O)5]+ that further evolve in molecular clusters by increasing pH. Ultimately, the LDH is formed by iron diffusion from the ferric oxyhydroxide phase into the α-Ni(OH)2 phase and anion intercalation. The particles formed by this method often show nanoplate morphology with the dimension along the layers stacking direction much shorter than the inplane dimensions. Samples with low crystallinity appear to have nanoplates more corrugated while high crystallinity samples have large and flat nanoplates that have the tendency to stack on top of each other as playing cards on a table.

3.2. Homogeneous Precipitation

In the homogeneous precipitation method, often referred to as solvothermal method or hydrothermal method, a hydrolysis agent is mixed with water and the metal salts precursors and

by progressive hydrolysis provides a weak base (amines or ammonia) that is responsible for the precipitation of of NiFe LDH. Urea, dimethylformamide (DMF) and hexamethylenete-tramine (HMT) are typical hydrolysis agents used.[16,51,58,64–68] This method requires a certain temperature to start the hydrol-ysis process and is performed in autoclaves. Well crystallized NiFe LDH nanoplates are obtained with this process, showing a hexagonal shape that reflects the layer symmetry (Figure 4b). In addition to this standard morphology, a flower-like morphology has also been obtained using ethylene glycol (EG) as chelating agent and urea (Figure 4c).[67] Other chelating agents have also been proposed. For example, trisodium citrate (TSC) has been proposed as chelating agent during urea hydrolysis in order to reduce the amount of amorphous FeOxHy particles.[58,69] Triethanolamine (TEA) is reported as another chelating agent in combination with a Fe2+ salt or a Fe3+ salt to better match the precipitation of Fe and Ni hydroxides and avoid nonsoluble metallic oxide.[51,65] The homogeneous precipitation method allows to grow the NiFe LDH catalyst on different substrates. For example, Liu et al. were able to grow highly oriented flakes arrays of NiFe LDH in vertical contact with the Ni foam substrate using a homogeneous precipitation method based on urea and assisted with NH4F as directing agent (Figure 4d).[70] Regarding photoelectrochemical applications, Wang et al. showed that homogeneous precipitation using DMF as hydrolysis agent is suitable for direct growth of NiFe LDH catalyst on Ta3N5 photo-anode.[22] More recently, a similar method using urea applied to Ta3N5 photocatalyst powder provided also enhanced photocata-lytic OER activity in the presence of an electron scavenger.[71]

3.3. Cathodic and Anodic Electrodeposition

Another common method consists in the electrodeposition of NiFe hydroxides thin films. The cathodic electrodeposition at constant current is the most used method and results in the formation on the cathode of NiFe-based hydroxides upon reduc-tion reaction in aqueous solution that contains the metal pre-cursors.[45,72–75] The pH of the deposition bath is either adjusted to acid values (i.e., pH 2) or slightly acid (pH between 2 and 6) or left unadjusted. Basic pH values are not used in order to avoid unintentional precipitation of Ni hydroxide particles in solution. A proposed formation mechanism is based on cata-lyst precipitation due to local basic pH created upon reduc-tion reaction of the anions (nitrate, sulfate), even though other reductions (nickel and iron plating, hydrogen evolution) are thermodynamically possible and difficult to distinguish.[72] The film obtained often shows a rough surface (Figure 4e) and is XRD amorphous or poorly crystallized. However, Trotochaud et al. showed that it is possible to increase the crystallinity and form NiFe LDH by aging in alkaline solution a cathodically electrodeposited NiFe oxyhydroxide film.[45] A recent variation of the cathodic electrodeposition method consists in using pulsed instead of continuous electrodeposition.[74] This method is particularly effective for thicker films and is able to produce denser films that are better electrically connected and have a more homogeneous Fe distribution. In contrast to the vast use of cathodic electrodeposition, the anodic variant is more rarely used. Recently, an anodic deposition method has been



Rev

iew


developed by Morales-Guio et al. that allowed the deposition of Fe rich FeNiOx catalyst on FTO and Au electrode.[21] The exact structure and phase of the FeNiOx catalyst is still under inves-tigation. The method was also applied by the authors using a hematite photoanode in replacement of FTO with deposition performed under irradiation. Finally, in contrast to the majority of the electrodeposition methods that are carried out in aqueous solvent, a method to electrodeposit NiFe alloy in an ionic liquid has been recently reported by Lo et al.[76] In this work, the film was deposited at constant potential on a Cu substrate and later converted in hydroxides by cycling in 0.1 m NaOH.

3.4. Alternative Synthetic Routes: Phase Transformations and Fe Impurity Uptake

In this section we include alternative methods of preparation of NiFe oxyhydroxydes that often involve a transformation from a preliminary solid-state phase containing Ni, Fe or both. Indeed, besides direct formation methods of mixed NiFe hydroxides explained so far, other indirect routes are possible: phase trans-formation by either electrochemical conditioning or chimie douche, Fe impurity uptake using a Ni oxide/hydroxide elec-trode and pulsed-laser ablation in liquid (PLAL). For example, Trotochaud et al. reported that thin film of thermal oxide of NiO transforms from a compact rocksalt structure to the more open and electrochemically accessible NiFe LDH structure upon sev-eral hours of galvanostatic conditioning at anodic current in 1 m KOH that contains Fe impurity.[77] Consistently, it has been shown that Fe ions in solution easily incorporate in Ni hydrox-ides compounds, and therefore NiFe-based oxyhydroxide can be prepared by electrochemical cycling or aging of Ni hydroxide electrode in iron containing electrolyte.[25,31,45] Completely dif-ferent is the route proposed in a previous study by Demour-gues-Guerlou et al. who reported a chimie douche method to obtain NiFe LDH by first preparation of iron substituted sodium nickelate, followed by oxidizing hydrolysis and a final chemical reduction.[78] Another recently reported method consists in

pulsed laser ablation in liquids.[79] In this method only one of the metal ion precursors is dispersed in the liquid while the second metal is introduced in suspensions as metal powder. A pulsed laser beam is focused on the solution and the nanopar-ticles are formed by cooling of the plasma formed around the ablation targets (the metal particles) and the surrounding solu-tion and then separated from the metal powder by a magnet and centrifuged. Recently, a new method was proposed by Bates et al. to prepare a NiFe catalyst (called Ni-Fe mixed metal oxide) supported on carbon and Raney Ni.[80] The supported catalyst was prepared by impregnation method involving chemical reduction to a mixed metal alloy obtained by NaBH4 addition and in presence of aniline as capping agent to control the mor-phology. The catalyst is further converted in hydroxide by elec-trochemical conditioning in 0.1 m KOH.[80] Similarly, metallic film obtained by sputter deposition can also be transformed in NiFe hydroxides by electrochemical conditioning in 0.1 m KOH.[46] Finally, an alternative method to electrodeposition has been proposed by Smith et al. to prepare amorphous NiFeOx.[81] This method, called photochemical metal-organic deposi-tion (PMOD), consists in the UV photolysis of organic groups of metal-organic complexes previously dissolved in hexanes and spin coated on a substrate. The obtained thin film is then annealed at low temperature in a furnace (i.e., 100 °C).

4. Electrochemistry in Alkaline Conditions

4.1. Surface Redox Chemistry

The surface redox chemistry of NiFe LDH is studied with cyclic voltammetry. Cyclic voltammograms are usually reported between 1–1.2 V vs RHE and 1.7 V vs RHE. No feature in the voltammograms is associated with change in oxidation state of Fe. On the other hand, an anodic and a cathodic wave is observed and assigned to oxidation of Ni(II) to an higher oxi-dation state (III, IV) and to the reverse reduction to Ni(II), respectively (Figure 5a). This change in oxidation state is



Figure 5. Effect on surface redox chemistry and OER activity of Fe content in NiFe (oxy)hydroxide. a) CV scans in 1 m KOH electrolyte solutions for electrodeposited Fe-free Ni (green) and NiFe oxyhydroxides film with different Fe content (blue, orange, red). Reproduced with permission.[45] Copyright 2014, American Chemical Society. b) Comparison of reported OER overpotential trends as a function of Fe content in NiFe (oxy)hydroxide catalysts. Measurements obtained in different conditions and overpotentials evaluated at different current densities. Therefore, the overpotential difference is plotted with respect to the minimum overpotential reported in each study. The numbers were manually extracted from the published graphs if not stated in tables. Data in black reproduced with permission.[25] Copyright 1987, The Electrochemical Society. Data in red and blue reproduced with permission.[72] Copyright 2013, American Chemical Society. Data in pink reproduced with permission.[87] Copyright 2011. The Royal Society of Chemistry. Data in green reproduced with permission.[46] Copyright 2015, American Chemical Society.

Rev

iew


accompanied by (de)protonation, similarly to the Ni(OH)2/NiOOH redox process in Fe free samples. At higher poten-tial than the Ni(II) oxidation wave an increase in current is observed and assigned to OER. Both redox waves shift anodi-cally and decrease in magnitude with increasing Fe content (Figure 5a).[72] At certain Fe contents the anodic wave disap-pears due to overlapping with the OER current. Friebel et al. proposed an explanation for this anodic shift based on the results of their density functional theory calculation.[46] The presence of Fe decreases the affinity to oxygen of the Ni surface sites and so rises the potential for the Fe:Ni(OH)2/Fe:NiOOH transformation to higher values. Chen et al. rationalized this shift in terms of destabilization of Ni3+ species in the presence of Fe3+ in the second coordination sphere.[65] The Fe3+ species makes the bridging oxygen/hydroxide ligands less electron-donating towards the Ni d-orbitals and so the higher formal oxidation of Ni is destabilized and the Ni2+/3+ redox potential increases. In general the complete understanding of the redox behavior is complicated due to possible appearance of multiple peaks and the fact that the anodic and cathodic waves usually evolve upon cycling, both in terms of magnitude as well as shift in potential. Multiple anodic waves could be assigned to dif-ferent Ni sites, while growing and shift of the waves might be due to more accessibility of the “bulk” sites that could be due to the hydrated nature of this material.[77] Formation of electri-cally disconnected domains upon cycling due to the different conductivity of the oxidized and reduced phase is also thought to affect the intensity of the redox peaks.[74] The redox behavior is still under debate also for the case of Fe free Ni(OH)2.[82] For example, in the Fe free Ni(OH)2 additional peaks are visible in the CV in the potential range between the main anodic wave and the OER onset potential that are not easy to assign.[25,31,45] Efforts in explaining the redox behavior of rigorous Fe free Ni(OH)2, will probably help the understanding of the redox fea-tures in NiFe LDH.

4.2. OER Activity: Overpotential, Turnover Frequency and Tafel Slope

NiFe LDH and amorphous NiFe (oxy)hydroxides catalysts show high activity for OER. In this paragraph, we first give an over-view of both the most used activity metrics and the related values reported in the literature of recent years. At the end of the paragraph, we discuss in more details the findings related to the various parameters that affect the activity of these cata-lysts. Among the wide range of possibilities for activity metrics, we have found that the overpotentials at fixed geometric cur-rent densities, Tafel slopes and turnover frequencies (TOF) at fixed overpotentials are the most common and are normally used for expressing the catalytic activity of these materials (Table 1). In particular, the overpotential at the current density of 10 mA cm–2 is often reported, due to its relevance for inte-grated photoelectrochemical water splitting devices operating at 1 sun where the photoabsorber is covered by the catalyst.[8] Higher current densities ≥ 500 mA cm–2 and closer to indus-trial electrolyzer conditions are more seldom considered even though equally relevant.[83] The overview in Table 1 shows that in 0.1 m KOH electrolyte the overpotential at 10 mA cm–2

varies in general for the best catalysts on flat substrate between values slightly below 300 mV vs RHE and ≈350 mV, apart some exceptions. The overpotential is further decreased by using porous substrate, like Nickel foam, of ≈50 mV vs RHE respect to typically reported and optimized values for flat sub-strates. In 1 m KOH electrolyte the overpotential at 10 mA cm–2 is lower than in 0.1 m KOH for similar catalysts tested in both electrolytes and varies typically between ≈260 mV vs RHE and ≈340 mV vs RHE, with further improvement of the low limit value of ≈40–60 mV with porous substrates (Ni foam, carbon fiber paper). As mentioned in the introduction, these values are among the highest ever reported for OER and show that these catalysts can achieve high current densities at moderate over-potentials. In this metric, there is no indication of how active a specific active site is, since it is not normalized by amount of material or type of sites. To address this important question, the TOF is often used, being a relevant metric for the comparison of intrinsic activity. However, the determination of the TOF per active sites for NiFe (oxy)hydroxides is controversial, mainly due to the unclear nature of the active sites (i.e., Ni or Fe, bridge μ2-OH or on top μ1-OH) and lack of reliable methods to determine them. In general, the TOF calculated by averaging over the total number of moles of metal ions (i.e., determined by inductively coupled spectroscopy) is the more practical and widely accepted method, even though it can only provide an underestimation of the real TOF. Therefore we consider this definition in our review and reported the TOF calculated at the overpotential of 300 mV in Table 1. It has been noted that higher TOF are calculated with low catalyst loading,[21] so the loading should always be indicated. The TOF vary widely among the catalysts. The best TOF approach ≈0.5 s−1 in 0.1 m KOH and ≈1.5 s−1 in 1 m KOH, even so the reported values span the same range for both electrolytes. At the present understanding of the catalyst, further discussion and comparison of the TOF is particularly difficult, in part for the limited data and the above mentioned problems related to the real active sites. In addition to the TOF, another figure of metric for assessing the intrinsic activity is the specific current density, where the current is nor-malized by the active surface area. This figure is particularly relevant for nonflat electrodes. However, several complications arise in determining the electrochemically active surface area via standard methods. Most of these complications originate from the differences between the more investigated thermal metal oxides, that are conductive and compact, and these hydrated materials, that are poor conductors in the reduced form and have bulk sites that are accessible to intercalated ions and possibly OH−.[72,74,84] Therefore calculations and compari-sons among different catalysts of electrochemically active sur-face area (ECSA) and specific current densities must be done with care. Recently, Batchellor and Boettcher have shown that a meaningful estimation of the ECSA is possible by imped-ance spectroscopy of NiFe oxyhydroxide catalyst in the oxidized phase, that is when is conductive.[74] Due to the complications mentioned, we have chosen to not report and discuss specific current densities. Finally, the analysis of Tafel plots is often car-ried out both as a comparative metric of activity as well as for elucidating rate limiting steps and the reaction mechanism. In particular, the exchange current density extrapolated from Tafel plots in conjunction with the Butler-Volmer equation is in



Rev

iew


general not reported for the OER reaction, due to the reaction complexity and large errors associated with their estimation.[8] Tafel slopes on the other hand are usually estimated from Tafel plots and reported. Despite they can provide a comparison of surface kinetics and can be very useful for elucidating reaction mechanisms, their analysis is particularly difficult in the case of multiple electron-proton transfer reactions, as it is the case for OER. For example, often a potential dependency of the Tafel slope is observed resulting in two or more “Tafel slopes” with a

non-linear transition between them.[8] In addition, the applica-tion of Butler–Volmer equation in conjunction with Tafel equa-tion is questionable in the case of OER.[85] Therefore we limit ourselves to report absolute values of Tafel slopes as metric to describe electrocatalytic kinetics in term of the dependence of the current on potential (Table 1). The Tafel slopes reported in the literature for NiFe (oxy)hydroxides typically vary between 25 and 60 mV dec–1 and are strongly dependent on catalyst prepa-ration and overpotential range analyzed.[72]



Table 1. Comparison of catalyst activity from works published in recent years. Methods: e.d. = electrodeposition, co-p. = co-precipitation, s.t. = solvothermal, h.p. (solvent) = homogeneous precipitation, sol. c. + an. = solution casting plus annealing, Fe i. = iron incorporation by aging, exf. = exfoliation, PLAL = pulsed laser ablation in liquid, red. = reduction. Substrate: Au (gold), GC (glassy carbon), CFP (carbon fiber paper), NF (Ni foam), HOPG (Highly oriented pyrolytic graphite). * = manually extrapolated from reported figures. ** = calculated by us. (?) = values with which we are not completely confident.

Catalyst Synthesis Method

Ni:Fe ratio

Substrate Catalyst loading [μg cm–2]/Thin film

thickness [nm]

Electrolyte η [mVRHE] @ Jgeo

= 10 mA cm–2

TOF [s–1] @ η = 0.3 V

(total metals)

Tafel slope [mV dec–1]

Ref. (year)

Ni-Fe film e.d. 60:40 Au ≈26.74** / 70 0.1 m KOH 280 ≈0.3** 40 [72] (2013)

NiFe LDH-NO3– co-p. 3:1 GC 100 0.1 m KOH 275* – 33.6 [55] (2016)

NiFeOx /C

(Vulcan XC-72R)

s.t. 62:38 GC 8 (only metals) 0.1 m KOH 290* 0.45 – [15] (2015)

NiFe LDH/oGSH

hybrid

h. p. (urea) 2.9:1 GC 250 0.1 m KOH 350 – 54 [108] (2015)

RGO-Ni-Fe LDH

composite

h. p. (DMF) 5:1 (?) GC 200 0.1 m KOH 250 – 33 [103] (2015)

NiFe LDH/C

(Vulcan XC-72R)

h. p. (DMF) 87:13 GC 100 (7.9 only

metals)

0.1 m KOH 360 0.03 51 [16] (2016)

– – – – – – – – –

NiFe LDH/CNT

hybrid

h. p. (DMF) 5:1 CFP 250 0.1 m KOH 308 – 35 [66] (2013)

NiFe/NF e.d. 3:1 NF 32 (?) 0.1 m KOH 240 – 33 [83] (2016)

3D NiFe LDH h. p. (urea) 3:1 NF 1000 0.1 m KOH 250* – 50 [64] (2014)

– – – – – – – – –

NiFeOx film sol. c.+ an. 9:1 Au/Ti QCM

crystal

1.17 / 2 1 m KOH 336 0.21 30 [77] (2013)

Fe:Ni(OH)2 film Fe i. 88-83:12-17 Au … / 30 1 m KOH 280 1.5 40 [31] (2015)

NiFe LDH h. p. (urea +

TEA)

3:1 GC 70 1 m KOH 347 0.01 67 [57] (2014)

NiFe LDH

exfoliated

exf. from h. p.

(urea + TEA)

3:1 (?) GC 70 1 m KOH 302 0.05 40 [57] (2014)

NiFe LDH co-p. + son. 80:20 GC 50 1 m NaOH 260 – 21 [107] (2016)

NiFe LDH-MoO42–

/C (Ketjen black)h. p. (urea) +

Molybdate salt

9:1 GC 280 1 m KOH 280 – 40 [116] (2015)

Ni–Fe/3D-ErGO e.d. 4:1 Au 140 1 m KOH 259 – 33 [104] (2015)

Ni-Fe e.d. 4:1 Au 140 (?) 1 m KOH 331 – 58 [104] (2015)

NiFe LDH PLAL 78:22 HOPG ≈444 ** 1 m KOH 280 – 47.6 [79] (2014)

– – – – – – – – –

NiFe LDH/CNT

hybrid

h. p. (DMF) 5:1 CFP 250 1 m KOH 247 0.56 31 [66] (2013)

NiFe/NF e.d. 3:1 NF 32 (?) 1 m KOH 215 – 28 [83] (2015)

FeNi LDH h. p. (urea+TSC) 3:1 (?) NF 250 1 m KOH 232 0.028 48 [100] (2014)

FeNi-rGO LDH

hybridh. p. (urea+TSC)

+ exf.+GO+red.

3:1 (?) NF 250 1 m KOH 206 0.987 39 [100] (2014)

Rev

iew


The extraordinary activity of NiFe (oxy)hydroxides is dependent on many parameters. The most common factors that influence the activity of NiFe (oxy)hydroxide and have been studied in the literature are: a) Fe content, b) the pH of the elec-trolyte, c) different substrate, d) different size/crystallinity, e) the nature of intercalated anions and f) the incorporation of a third metal.

4.2.1. Fe Content

Activity trends as a function of iron content have been investi-gated with the aim of determining the amount of incorporated iron that maximizes the activity.[25,46,72,86] Louie et al. found that the activity reaches a broad maximum across Fe compositions in the range of Fe content ≈15%–50% for electrodeposited films evaluated in 0.1 m KOH.[72] This is in agreement with Corrigan et al. who found that the lowest overpotential at 50% Fe content was not much different than the value at 10%.[25] On the other hand, Li et al. reported the optimum composition at a value ≈10% Fe for electrodeposited NiFe-based film.[87] Klaus et al. also measured the maximum OER activity at moderate Fe con-tent of ≈12–17% in Fe incorporated Ni oxyhydroxides film on Au substrate.[31] The identification of the optimal composition is difficult, probably due to formation of inactive Fe oxides/oxy-hydroxide phases starting at a Fe content that is dependent on the preparation method. However, it is generally accepted that the OER overpotential decreases dramatically for increase in Fe content from 0 to 10%, reaches a minimum at some concentra-tions in the range between 10% and 50% and further increases for Fe content higher than 50% (Figure 5b). Fe-free Ni hydrox-ides and Fe (oxy)hydroxides correspond to the two limiting cases of the trend, respectively on the lower and upper end of Fe content. Ni hydroxides have been extensively studied as cata-lysts for OER and many Ni hydroxide based catalysts have been reported to exhibit superior OER activity in alkaline and near neutral borate buffered electrolyte.[14,30,42,88–91] On the other hand, recent studies[13,45,46] in agreement with the pioneering work by Corrigan[25] showed that pure Ni hydroxides are poor catalysts when tested in specifically purified Fe-free electrolyte. It was found that incorporation of Fe impurities occurs during electrochemical experiments and increases the catalyst activity. Interestingly, despite the presence of Fe has a beneficial posi-tive impact on the OER activity,[45,84,92] pure Fe oxyhydroxide electrodes were reported to catalyze the OER with high over-potentials in alkaline conditions.[91,93] Recent studies addressing the intrinsic activity of these catalysts pointed out that the poor electrical conductivity at low overpotentials affects strongly the OER performance of Fe oxyhydroxide films, resulting in apparent low activity in the case of thick films.[94,95] On the con-trary, thin films showed good OER activity and high turnover frequencies[94,96] and in particular, thin films with low iron loading were reported to be highly efficient and promising OER catalysts in neutral electrolytes.[97]

4.2.2. pH of the Electrolyte

Another parameter affecting the OER activity is pH. It has been shown that the OER activity and the redox features shift

as a function of pH with super-nernstian behavior on the NHE scale.[16,66] Therefore is important to specify at what pH is the activity reported, even when the RHE scale is used. Both pH 13 and pH 14 are commonly used alkaline pH conditions so we will report the activity for these two pH values.

4.2.3. Substrate

In addition, the substrate has been found to have an effect on the activity. This interaction has been studied by many authors by using different substrates for growing NiFe oxyhydroxide thin films or different supports for NiFe LDH nanoparticles. In the case of thin films, in particular the effect of noble metal substrates respect to glassy carbon (GC) was investigated, fol-lowing previous works on Ni(OH)2.[89,90,98] Trotochaud et al. compared the dependence of the TOF on the thickness of a thin film of NiFe oxyhydroxide and found that is stronger on Au substrate than on GC.[45] Furthermore, the TOF values for the catalyst on Au were higher than the one on GC, especially for low thickness. Not a big difference in TOF was observed when a Pt substrate was used instead of Au.[94] In the case of Au, the possibility of formation of mixed Au-Ni(Fe) (oxy)hydrox-ides has been considered.[31,45,89] Interestingly the Au substrate has been found to affect not only the TOF but also the tran-sition of NiFe hydroxide to NiFe oxyhydroxide (see later sec-tions 5.3 and 5.5).[31,99] The interaction of NiFe LDH nanoplates and nanocarbon based materials has been explored in order to improve the electrical conductivity and the weak interaction of the catalyst with the conductive support. The reported materials vary from hybrid catalysts consisting of NiFe-based nanoplates combined with or supported on graphene with various degree of oxidation,[100–105] carbon nanoparticles (i.e., Vulcan XC-72R or Ketjen Black),[15,80] carbon quantum dots[106] and carbon nanotubes (Figure 6).[66,107–109] In particular, Gong et al. showed that direct nucleation and growth of NiFe LDH nanoplates on carbon nanotubes resulted in catalyst with higher activity than NiFe LDH mixed with carbon black or with the carbon nano-tubes.[66] This improvement was investigated by performing x-ray absorption at the carbon k-edge which showed evidence for strong interaction between the NiFe LDH and the carbon nanotubes. Finally, Bates et al. discussed charge transfer resist-ance in the case of Raney-Ni support by electrochemical imped-ance spectroscopy analysis and found conductivity comparable with standard Ir black catalyst.[80]

4.2.4. Size and Crystallinity

The influence of crystallinity on the activity has not been exten-sively studied. Trotochaud et al. showed that the activity does not change in a detectable amount between a poorly crystal-lized NiFe (oxy)hydroxide film and the same film after the long-range order has been improved by aging.[45] On the other hand, Xu et al. compared NiFe nanoplates after hydrothermal aging at different temperatures and found a clear trend where the low crystallinity samples were the most active.[55] However, we notice that in this comparison it was not possible to distin-guish between crystallinity effects and size effects since both



Rev

iew


improves with aging temperature, as shown by the authors. The important of size effect is also discussed by Song et al.[57] Song et al. compared NiFe LDH nanoplates with the same material after exfoliation and found that the latter performed better both in terms of overpotential and Tafel slope.[57] The authors proposed that the origin of this improvement might be attributed to the increase in the number of active sites at the edge of the nanosheets due to their concomitant fragmentation and smaller size occurring during exfoliation and to enhanced electronic conductivity.

4.2.5 Nature of Intercalated Anions

The influence of the intercalated anion on the OER activity has been investigated by Xu et al.[55] Comparing NiFe LDH intercalated with NO3

− and CO32− the author found a supe-

rior activity for the NO3− intercalated compounds. The inter-

calated NO3− anion increases the interlayer distance respect

to CO32– and it is easier to exchange by OH−, that is the main

reactant for OER in alkaline. Reduced charge transfer resist-ance was also found for the NO3

−. Therefore the authors attributed the enhanced activity to these effects. An investiga-tion of the intersheets population of anions in the active state of the catalyst under OER and during potential cycling will be important for further understanding of the anions effect. Recently, Hunter et al. analyzed the effect of interlayer anions in NiFe LDH in detail and in strict carbonate-free conditions obtaining new insights.[110] The authors synthesized NiFe-based materials with a wide variety of intercalated anions (NO3

−, BF4−, Cl−, ClO4

−, CO32−, C2O4

2−, F−, I−, PO43−, SO4

2−) either by anion exchange of nitrate intercalated NiFe LDH or

directly by PLAL. The authors confirmed the high affinity of NiFe LDH for the carbonate anion and found that carbonate substituted all the other anions after immersion in 1.0 m KOH solution in ambient air. Therefore the activity was tested in CO2-free atmosphere (glove box) and carbonate free solu-tion (saturated by barium hydroxide to precipitate carbonate as BaCO3). In these conditions the authors found no correla-tion of the OER activity with basal spacing (layer-to-layer dis-tance) but with the pKa of the conjugate acid of the interlayer anions. The authors proposed that his trend suggests a base-assisted deprotonation step with important implication in the OER mechanism.

4.2.6. Incorporation of a Third Metal

Incorporation of a third metal has also been proposed as a strategy to improve the activity of NiFe LDH by affecting the conductivity and the structure of the binary catalyst. For example, ternary NiFeMn LDH[111] and NiFeCo[80] mixed metal oxide film exhibited enhanced OER activity than reference NiFe LDH catalysts, supporting the idea that this approach is effective in tuning the catalyst activity towards higher value. Interestingly, ternary AlFeNi amorphous oxides were found to exhibit very low Tafel slopes (typically lower than 20 mV dec−1).[112] This result suggests that Al, despite its oxide being inert for OER, has a beneficial role when incorporated in amorphous NiFe-based catalysts. Other recent screening studies[113] investigating the introduction of foreign metals in NiFe-based catalysts have been reviewed by Gong et al.[26] These studies also reported high OER activity for ternary AlFeNi oxide catalysts.



Figure 6. NiFe LDH hybridized with nanocarbon materials. a) NiFe LDH nanoplates on graphene nanosheets. Reproduced with permission.[103] Copy-right 2016, Elsevier. b) Alternately stacked NiFe LDH nanosheets with graphene. Reproduced with permission.[102] Copyright 2013, American Chemical Society. c) NiFe LDH nanoplates on a 3D graphene structured electrode. Reproduced with permission.[104] Copyright 2015, Royal Society of Chemistry. d) Carbon quantum dot/NiFe LDH composite. Reproduced with permission.[106] Copyright 2014, American Chemical Society. e) NiFe LDH nanoplates on carbon nanotubes. Reproduced with permission.[56] Copyright 2015, American Chemical Society.

Rev

iew


4.3. Stability of NiFe LDH Catalysts During OER

The long term stability of a catalyst is amongst the paramount requirements for practical water splitting device. Assessing catalyst stability during long term operation conditions (tens of hours, days and years), variable load conditions simu-lating switch on/off processes and at high current densities > 500 mA cm–2 in the case of industrial electrolyzer conditions demands substantial efforts. Some protocols have been pro-posed to monitor the catalysts stability in milder conditions and shorter operating time frames. A useful protocol for screening stability among several catalysts consists in running the anodes at a current density of 10 mA cm–2 for 2 hours while recording the potential changes.[8] This method is particularly valid in discarding unstable catalysts, however of course is not meant to substitute long term stability tests. A more rigorous testing protocol for evaluate short term stability has been proposed combining the monitoring of mass losses to the previously mentioned protocol.[114] The methods proposed for checking corrosion are electrochemical quartz crystal microbalance (EQCM) and inductively coupled plasma mass spectroscopy (ICP-MS) applied to the electrolyte. NiFe LDH particles and oxyhydroxides films generally show good stability in short term tests in alkaline conditions. The long term stability has also been addressed and in particular Ni foam is often reported as the best substrate for these conditions.[115] Particularly long term conditions were investigated by Gong et al. with NiFe LDH nanoplates on Ni foam (≈5 mg cm–2 loading). The sta-bility test was performed in 1 m KOH at the constant current density of 20 mA cm2–2 and oxygen evolution was observed for more than 4 days.[66] From the reported graph we can estimate a decrease of ≈30 mV from the minimum potential to the final measured potential. Other works also report good stability for operating conditions of 10 hours or longer at current densities > 10 mA cm–2.[57,64,70,83,101,102,107,115,116]

Bubbles evolve at the electrodes at these high current den-sities and are often blocking active sites resulting in “bubble overpotentials” and apparent deactivation. Lu et al. reported that this problem is mitigated when macroporous Ni foam sub-strates are used instead of flat GC electrodes.[83] In the latter case removal of bubbles completely “reactivated” the catalyst. A reactivation of the catalyst was also observed by Luo et al.[117] The authors investigated the stability of NiFe LDH on Ni foam in 1 m NaOH in a two electrode configuration using two elec-trodes of the same kind for cathode and anode. A small deg-radation in the current density was observed after 10 hours of operation at the applied potential of 1.8 V. The current density was however recovered after a resting period of 8 hours, as shown by the authors in the next two tests of 10 hours each.

5. Advanced In Situ Characterization

5.1. Electrical Conductivity

Electrical conductivity is different among metal hydroxides and oxyhydroxides and might influence substantially the reported activity and the comparison amongst different catalysts.[94] In situ electrical conductivity experiments were performed by

Trotochaud et al. using interdigitated microelectrodes and cathodically electrodeposited NiFe (oxy)hydroxide.[45] The cata-lyst is deposited on top of two working electrodes forming a bridge between them. The experiment shows that the “effec-tive” conductivity is low for the reduced Ni and reduced NiFe containing films, independent of the Fe content. In all cases the conductivity increases when the potential passes the Ni2+/3+ oxi-dation potential (Figure 7a). This change in conductivity could be associated to a transition from semiconductor to half-metal, that was suggested by Li et al.,[118] and it is concomitant with a change in color of the film.[12,74] Finally, it was found that the incorporation of Fe increases the conductivity of the oxidized NiFe-based film respect to Fe free NiOOH, even though this increase is not enough to explain the activity enhancement.

5.2. X-ray Absorption Spectroscopy (XAS)

In situ and in operando X-ray near edge absorption (XANES) and complementary X-ray absorption fine structure (XAFS) are employed to reveal the nature of the active sites, metal oxidation states and local coordination structure of the catalyst during oxygen evolution reaction. Friebel et al. investigated in oper-ando Ni and Fe K-edge XANES and EXAFS of electrodeposited Ni-Fe catalysts on Au film over a wide range of Fe% in 0.1 m KOH.[46] Complementary EXAFS was performed also for 25% Fe containing NiFe film (Ni:Fe = 3:1) obtained by sputter depo-sition that was later oxidized by electrochemical conditioning in 0.1 m KOH. XAS shows a clear shift of the pre-edge peak and the main absorption edge in Ni K-edge when the potential is increased from 1.12 V above 1.52 V (Figure 7b). The shift further reaches a saturation value for higher potential (up to a measured potential of 1.92 V). The potential at which this shift occurs corresponds well with the Ni(II)/Ni(III) redox potential. In the same potential range no significant shift is observed for the pre-edge and edge in the Fe K-edge. Complementary XAFS analysis of both Ni and Fe K-edge of the oxidized/sputter-depos-ited sample confirm potential induced contraction of Ni-O and Fe-O distances as well as of the nearest metal-metal distances (Table 2). The correlation of bond length contractions in the Ni and Fe K-edge EXAFS and the further appearance of similar multiple scattering features in both EXAFS spectra confirm that Fe is substituting Ni and it is not interstitial. EXAFS of the elec-trodeposited samples that have different Ni/Fe ratio, show that the Ni-O bond length at rest (1.12 V) and during OER (1.62 V) is not dependent on the Fe content. Conversely, the Fe-O bond length during OER depends on Fe content. This could be due to formation of a secondary FeOOH phase and overlapping of the signal from the Fe species in the two phases. For high ratio of Ni:Fe (low Fe content, < 25%), where the FeOOH is negligible, the study shows that Fe-O bond distance in NiFe LDH during OER is smaller than in γ-FeOOH at the same potential and that this is almost identical to Ni-O, confirming the substitution of Fe in the NiOOH lattice.

In conclusion, the XANES and EXAFS observation of the Ni K-edge are associated with increased oxidation state of Ni and structural change from a α-Ni(OH)2-like to a γ-NiOOH-like phase. The average oxidation state of Ni and Fe under OER con-dition was found to be +3.6 and +3 respectively, so the authors



Rev

iew


exclude the formation of Fe4+ species. In a following study Bates et al. also analyzed the in operando XAS of Ni K-edge of Ni-Fe mixed metal oxide (Ni:Fe ratio = 9:1) in 0.1 m KOH. The NiFe catalyst was supported on carbon and obtained by electrochemical conditioning of metal NiFe catalyst formed by chemical reduction of metal nitride precursors.[80] They also observed shift of the Ni K-edge when the potential reached 1.5 V vs RHE, which stabilizes at higher potentials of 1.55 and 1.6 V. Both Ni K-edge and Ni-O and Ni-M bond distances match well with data of Friebel et al.[46] A slightly different XANES spectra was measured by Wang et al. who studied in oper-ando XAS of Ni and Fe K-edge of NiFe hydroxide catalyst Ni:Fe ratio = 9:1) electrodeposited on carbon paper in 1 m NaOH.[119] They observed a larger shift of the Fe K-edge of 2.8 eV when the potential is increased to ≈1.72 V vs RHE (reported 650 mV vs SCE) compared to the small shift of 1.3 eV observed by Friebel et al. at similar potentials and considered by the latter

authors negligible.[46] The authors attributed the measured shift to oxidation of Fe from the +3 value observed for the as-prepared catalyst to values close to +4 under OER conditions at the reported potentials. The result was compared with pre-vious finding by Balasubramanian et al., which reported a shift of 2.3 eV and also suggest predominance of Fe(IV) in the oxi-dized NiFe oxyhydroxide state.[120] However, the potential range for OER condition was not examined. In contrast to different degree of oxidation of Fe species found in the literature, the analysis of the Ni K-edge confirmed the previously observed shift of the edge with increasing potential. In particular the author observed that the oxidation state of +3 is reached at ≈1.45 V vs RHE and that by increasing potential the average oxidation gradually changed till reaching ≈ +3.6 at potential of 1.72 V vs RHE. Ex situ L-edge measurements were carried out in support to the in operando K-edge for both Ni and Fe. In the ex situ Fe L-edge no shifts were observed. This difference



Figure 7. Examples of in situ and in operando characterization. a) Electrical conductivity for NiFe (oxy)hydroxide films (points with solid lines) and oxidation wave (dotted lines). Reproduced with permission.[45] Copyright 2014, American Chemical Society. b) Ni K-edge XANES spectra of NiFe (oxy)hydroxide film showing pre-edge (left) and main adsorption edge (right) at different applied potentials. Reproduced with permission.[46] Copyright 2015, American Chemical Society. c) Raman spectroscopy of NiFe-LDH film at 1.6 V vs RHE applied potential (OER, red) and at 1.2 V vs RHE (without OER operation, black). Reproduced with permission.[64] Copyright 2014, Royal Society of Chemistry. d) Mössbauer spectra of NiFe LDH at resting state (open circuit, left) and at 1.76 V vs RHE (OER conditions, right). Adapted with permission.[65] Copyright 2015, American Chemical Society.

Table 2. Comparison between bond distances of as prepared and under OER condition of NiFe hydroxide catalyst (Ni:Fe = 75:25) sputter deposited on Au film and electrochemically oxidized. Reproduced with permission.[46] Copyright 2015, American Chemical Society.

Ni-O Ni-Ni Fe-O Fe-Fe

Bond length (Å) @ E = 1.12 V (≈as prepared) 2.06 (±0.01) 3.10 (±0.01) 2.01 (±0.01) 3.10 (±0.01)

Bond length (Å) @ E = 1.62 V (OER) 1.89 (±0.02) 2.82 (±0.01) 1.90 (±0.02) 2.84 (±0.02)

Rev

iew


respect to the in operando K-edge is explained as a reduction of the Fe(IV) back to Fe(III) after removal of the sample from the electrolyte for the ex situ measurements. The authors note how the oxidation state is sensitive to exposure to ambient environ-ment and so in situ measurements are necessary. Despite the absence of energy shift, other features in the XANES spectra of the Fe L-edge are in agreement with the higher oxidation state of Fe during OER. Finally, the ex situ XANES spectra of the Ni L-edge confirmed the potential driven oxidation of Ni observed in the in operando Ni K-edge. In a recent study, Görlin et al. reported contrasting results respect to previous studies detecting high-valent Ni ions under OER conditions.[121] The authors analyzed NiFeOx catalysts with different Fe con-tent with quasi in situ X-ray absorption spectroscopy after been freeze-quenched under applied potential in 0.1 M KOH and OER conditions. In particular, they found that at Fe levels above 4%, the oxidation state of Ni atoms were stabilized in +2 under OER conditions, in contrast with higher oxidation observed with Fe-free NiOOH. The authors suggested in reconciliation with previous studies that the ratio between the rate of metal reduction occurring under OER and the rate of metal oxidation is governing the population of high-valent Ni centers and so affecting their detection.

5.3. Raman Spectroscopy

In situ Raman spectroscopy provides a powerful tool to inves-tigate the structure and the potential dependent transforma-tions of Ni(OH)2 and NiFe (oxy)hydroxides. Notably, Lu et al. performed in situ Raman spectroscopy of NiFe LDH catalyst grown on Ni foam by urea based solvothermal synthesis.[64] The experiment were conducted in 0.1 m KOH at 1.2 V vs RHE (resting state) and 1.6 V vs RHE (OER condition) and explored the 300–800 cm−1 Raman shift region, where Ni-O and Ni-OH stretching mode are observed in Ni(OH)2 compounds.[12] At 1.2 V only a broad band at 528 cm−1 was observed which disap-peared under OER (1.6 V vs RHE) in concomitance with the appearance of two new bands at 476 and 556 cm−1 (Figure 7c). These bands are typical of β-NiOOH and γ-NiOOH phases and so the authors attributed their formation to the transition to a Ni(Fe)OOH-like phase. The authors compared the inten-sity ratio of the two bands (I476/I556) with the ratio of the bands observed at the same Raman shift values for Ni(OH)2 under OER. The bands for the oxidized NiFe LDH appeared broader, weaker and having lower I476/I556 ratio respect to NiOOH. This was considered an indication of a more disorder local struc-ture around Ni-O. This structural change induced by Fe is in agreement with the previous observation by Louie et al. that the intensity of the band at 555 cm−1 is growing with increasing Fe content in oxidized NiFe oxyhydroxide films that were prepared by electrodeposition on Au substrate.[72] Similar effect was also reported by Klaus et al. for a Ni(OH)2 film aged in unpurified 0.1 m KOH that was obtained by electrodeposition on a rough-ened Au electrode.[31] The aging process leads to Fe incorpora-tion, and under OER the samples shows two features at 476 and 555 cm−1, that were broader and with a lower I480/I560 ratio than the same peaks observed in Ni(OH)2 that was aged in purified electrolyte (where Fe incorporation is negligible). The authors

also noticed a Au induced substrate interaction affecting the transformation of Ni(Fe)(OH)2 to Ni(Fe)OOH. This transition begins 0.1–0.2 V below the electrochemical Ni(II)/Ni(III) oxi-dation observed in the linear potential scan and the potential at which this transformation start is consistent with the onset potential of Au oxidation.

5.4. Mössbauer Spectroscopy

Mössbauer spectroscopy has been used to elucidate the elec-tronic structure, d-electron configuration of the metals in NiFe (oxy)hydroxides and discuss possible distortions of their ligand sphere. Recently, Chen et al. performed in operando Mössbauer spectroscopy on NiFe LDH catalyst (Ni:Fe = 3:1) prepared by homogeneous precipitation using urea on carbon paper in 0.1 m KOH.[65] Their study develops on previous works that were done in ex situ conditions or not at potential high enough for sub-stantial OER activity.[24,122] The previous studies proposed the presence of Fe(IV) at potential higher than the Ni(II)/Ni(III,IV) redox transition. Consistently, Chen et al. observed a doublet in the Mössbauer spectra with isomer shift δ ≈0.34 mm s–1 and quadrupole splitting Δ ≈ 0.46 mm s–1 at open-circuit con-ditions, confirming the presence of high spin Fe3+ species when the potential is not applied (Figure 7d). Upon increasing the potential to OER conditions 1.62 and 1.76 V, the authors observed the appearance of a shoulder in the Mössbauer spectra interpreted as oxidation of Fe3+ to Fe4+, and corresponding to ≈12% and ≈21% of the total iron species (Figure 7d). Lowering the potential did not result into an immediate disappearance of the shoulder (reduction of the Fe4+). However, the disap-peared when the potential was kept at open-circuit potential for enough time, indicating that the oxidation of Fe(III) to Fe(IV) was somewhat reversible. These species were not detected in Ni-free FeOOH under OER conditions. Therefore the authors conclude that the formation of the Fe(IV) is due to stabiliza-tion of the NiOOH host lattice. The authors also speculated that these detected Fe(IV) species are located at an internal position in the layer and fully surrounded by Ni, since Fe(IV) species at the edges, corners or defects are not expected to be detected in the measurement conditions. On the other hand, these latter undetected Fe(IV) species, if present, are proposed by the authors as possible active sites.

5.5. Ambient-Pressure XPS

Surface characterization in liquid electrolyte solution and under operating conditions is of foremost importance for electro-catalysts and extremely challenging. Recently, Ali-Löytty et al. performed ambient pressure x-ray photoelectron spectroscopy (APXPS) on a 7 nm thick film of NiFe oxyhydroxide that was electrodeposited on Au electrode.[99] The authors used tender X-rays ( at photon energy of 4020 eV) and the “dip and pull” method,[123] that allowed to measure in operando conditions with a ≈18–30 nm thin film of 0.1 m KOH electrolyte on the electrode and water pressure of 18 Torr in the chamber. The sample was measured as prepared (at 9 Torr water exposure, without electrode immersion in the electrolyte) and at fixed



Rev

iew


potential (with the thin film electrolyte), after electrochemical conditioning by potential cycling. The authors analyzed the Ni 2p3/2 and the Fe 2p3/2 spectra and detected metallic Ni0 and Fe0 species in the as prepared sample and their irreversible oxida-tion after conditioning. These species were codeposited during the preparation of the NiFe oxyhydroxide catalyst by cathodic electrodeposition. The authors also observed an increase in the relative intensity of the M-O compared to M-OH peak in the O 1s XP spectra when they increased the potential from 0 to 0.3 V vs Ag/AgCl (≈1.264 V vs RHE), a potential range where sur-face Au oxide forms but the transition from Ni(OH)2 to NiOOH should not occur. The authors could not find further evidences of occurrence of Au oxidation and so proposed another expla-nation for the change in the O 1s spectrum. They suggested that this change is associated to a partial deprotonation of Ni(OH)2 occurring at lower potential than the oxidation of the Ni(II) to Ni(III). This process forms a negative [Ni(OH)2–xOx]x− layer that is stabilized by the electrostatic interaction with the gold surface. Therefore in this mechanism the transition from Ni(OH)2 to NiOOH mediated by a gold surface is seen as a reaction that does not occur with a concerted proton-electron transfer (CPTE), but the proton transfer precedes the electron transfer.[124] This explanation is in agreement with the observa-tion by Klaus et al. of a NiOOH feature appearing at a potential ≈0.1–0.2 V lower than the Ni(II) to Ni(III) oxidation in the in situ Raman measurements of an Fe incorporated Ni(OH)2 film on Au electrode.[31] Unfortunately, as Ali-Löytty et al. noticed, it was not possible to support this explanation by assignment of Ni oxidation state from the analysis of the Ni 2p3/2, since the sensitivity was not enough to distinguish between Ni(II)(OH)2 and γ-Ni(III/IV)OOH.

5.6. In Situ Differential Electrochemical Mass Spectroscopy

In situ differential electrochemical mass spectroscopy (DEMS) has been used to estimate OER Faradaic efficiencies of NiFe (oxy)hydroxides and to identify volatile species generated in elec-trolytes with different pH or containing different ions.[15,61,121] For example, Görlin et al. investigated changes in Faradaic efficiencies for OER with NiFeOx/C tested in 0.1 m KOH after cylcing the catalyst in 0.1 m phosphate buffer (pH 7) or 0.1 m borate buffer (pH 9.2).[15] In both cases, the Faradic efficiencies after cycling at lower pH were similar respect to the measure-ments before the cycling protocol. However, the Faradic effi-ciency was significantly lower during the cycling at pH 7 due to carbon corrosion, as confirmed by monitoring the CO2 signal. In a further study, the authors used in situ DEMS-based Fara-daic charge analysis to investigate the oxidation change of Ni in NiFe oxyhydroxides with different Fe content by distinguishing between the charge associated with oxygen evolution and with Ni oxidation, respectively.[121]

6. Proposed Mechanisms for OER

The mechanism for OER on oxide surfaces is still under debate and investigation. Several reaction mechanisms have been proposed in the literature for the OER on oxides[10,125–128] in

acid and alkaline electrolyte solution and reviewed recently by Hong et al. and Fabbri et al.[5,7] Important differences among the mechanisms include the reaction intermediates, the number of metal sites directly involved in the reaction per gen-erated oxygen molecule (only one or multiple sites) and the nature of the proton-electron steps involved, that can consist in a decoupled or concerted proton-electron transfer (CPET).

Selloni and Li conducted a theoretical study on the OER mechanism for higher index surfaces of γ-NiOOH and β-NiOOH, considering both pure and Fe doped phases.[118] The higher Miller-index surfaces are expected to be more active, in analogy to what have been predicted for CoOOH.[129] The model for the structures was based on previous theoretical and experi-mental works for pure NiOOH (see section 2.1) and monoclinic surface cells were used to model the surfaces.[35,37,43] By using spin-polarized density functional theory the authors found that Fe replacing surface Ni sites results in lower OER overpoten-tial than the one calculated for the respective pure NiOOH phase in both cases. In particular they found that Fe doped β-NiOOH(01-15) has a lower overpotential than the Fe doped γ-NiOOH(101) and the OER mechanism on the two surfaces is slightly different (Figure 8a,b). The mechanisms were inves-tigated by a recursive method based on addition of water and stepwise simultaneous removing of a proton and an electron resulting in four CPET steps. In both mechanisms, an O-O bond forms between an O atom and a surface lattice three-fold coordinated oxygen. On the other hand, the rate determining step was found to be the first deprotonation for Fe doped β-NiOOH(01-15) and the second deprotonation for Fe doped γ-NiOOH(101).

A different mechanism has been proposed by Bell and co-workers based on four CPET, the intermediates O*, OH*, OOH* and one active metal site per oxygen (Figure 8c).[46] The mechanism was derived from the comparison between XAS, OER activity measurements and density functional theory cal-culations with Hubbard U framework (DFT-U). The calcula-tions were performed for pure and Fe doped γ-NiOOH and for pure and Ni doped γ-FeOOH. Due to insufficient experimental data for the structure of γ-NiOOH, this phase was simulated by β-NiOOHx = 0.5 with Ni oxidation state of +3.5, so a half-depro-tonated β-NiOOH. The authors refer to the work by Van der Ven et al. for the choice of the structure.[35] The (01-12) surface has been chosen as example for high index surface of the Fe doped γ-NiOOH. These surfaces have μ1 (“metal on top”) and μ2 (“metal-metal bridge”) sites and are considered to be more active than the natural (0001) facet (where μ3 sites are dominant). The theoretical overpotential is obtained from comparison of the cal-culated Gibbs free energy differences of the reaction intermedi-ates at each step. Pure γ-NiOOH is shown to adsorb all the OER intermediates too weakly, while pure γ-FeOOH too strongly. The lower overpotential was found for Fe sites that are surrounded by Ni next-nearest neighbors in either γ-NiOOH or γ-FeOOH. This site shows near optimal binding energies for the OER intermedi-ates and the formation of OOH* was found to be the rate lim-iting step (RLS). The oxidation of O* to OOH* has also been pre-dicted to be the rate-determining step for NiFe LDH by DFT cal-culations reported in a recent work by Duan and co-workers.[75] Interestingly, the calculations by Bell and co-workers also show that subsurface Fe in γ-NiOOH increases the overpotential



Rev

iew


at a Ni surface site by further weakening the O* intermediate binding energy on Ni site, whose formation in this case is the RLS. Therefore the surrounding Ni next-nearest neighbors increase the activity of Fe sites, that becomes the most active and are consequently considered the active sites, while (subsurface) Fe actually decrease the activity of the Ni sites. Recent TOF cal-culation on a per-Fe site basis by Klaus et al. supports the predic-tion that Fe in the NiOOH host is the OER active sites.[31]

Calle-Vallejo and co-workers also performed DFT calcula-tions aimed at understanding the effect of transition-metal doping on the activity of Ni-based double hydroxides in alkaline conditions.[60] The authors considered a modified rocksalt NiO structure where the (001) surface was partially protonated (one hydrogen per two oxygen atoms) and doped with the second metal (50% surface Ni substitution) in order to model the sur-face as NiMOOH. This choice was based on the consideration



Figure 8. OER mechanisms proposed for NiFe (oxy)hydroxides (a–c) and activation steps proposed for Ni(OH)2 (d,e). (a,b) OER pathway proposed by Selloni and Li for Fe: β-NiOOH(01-15) (a) and Fe doped γ-NiOOH(101) (b). Nickel (blue), iron (gray), oxygen (red) and hydrogen (white) atoms are shown and the O–O bond is marked by a black ellipse. Molecular oxygen is liberated after the last step. Reproduced with permission.[118] Copyright 2014, American Chemical Society. c) OER mechanism proposed by Bell and co-workers for Fe-doped γ-NiOOH(01-12). Nickel (green), iron (yellow), oxygen (red) and hydrogen (white) atoms are shown. Reproduced with permission.[46] Copyright 2015, American Chemical Society. d) Two-electron two-proton PCET activation process leading to highly active oxyl radicals proposed by Nocera and co-workers for Ni(OH)2 electrodeposited in borate buffer. Reproduced with permission.[126] Copyright 2013, American Chemical Society. e) Chemical deprotonation step proposed by Koper and co-workers for Ni(OH)2 toward negatively charged and highly active surface species. Reproduced with permission.[125] Copyright 2016, Royal Society of Chemistry.

Rev

iew


that the limited information available for the active surface structure under OER conditions make several choices (i.e., Fe doped γ-NiOOH) arbitrarily equivalent at the present under-standing of the structure provided that they take into account a partially deprotonated surface and octahedral metal centers. A similar mechanism to the one considered by Bell and co-workers with OH*, OOH* and O* intermediates was chosen in the calculations of the overpotentials for OER. The calculations shown that the effect of Fe on the Ni sites is to slightly increase its overpotential, while the effect of the NiO host lattice on Fe sites results in near optimal binding energies and consequently in a low overpotential for the OER on this site. Therefore, besides the different structure, the DFT calculations of Calle-Vallejo and co-workers are in agreement with the results of the DFT calculations by Bell and co-workers regarding the identifi-cation of Fe surrounded by Ni as the active sites for NiFe oxy-hydroxide.[46] The authors also noticed that coverage effects are important in the calculations and that due to the different RLS on Ni sites, where formation of O* (from OH*) is RLS, and on Fe sites, where formation of OOH* (from O*) is RLS, the sites should be covered respectively by OH* and O* under OER conditions.

These studies investigated the nature of the active sites of mixed NiFe (oxy)hydroxides based on considerations about the affinity of a surface site towards adsorbed OER intermediates and their relative stabilities derived from DFT predictions and in comparison with experimental results.

It is worth to compare the mechanisms that we just described for NiFe (oxy)hydroxides with the ones proposed for “pure” phase Ni (oxy)hydroxides (or unintentionally incorpo-rating Fe impurities). The following mechanisms are based on the interpretation of the results from in operando spectroscopy or kinetics investigations, where a parallel investigation by DFT calculations has not been reported yet.

Nocera and co-workers proposed a OER mechanism based on a detailed kinetics study using an anodized Ni(OH)2 film that was electrodeposited on FTO in borate electrolyte.[126] They proposed that the turnover-limiting-step (TLS) has a chemical nature and most likely does not involve deprotonation of the catalyst. This TLS, where O-O bond forms, is anticipated by proton coupled electron transfer steps producing highly active oxyl radicals (Figure 8d). The importance of deprotonation of Ni oxyhydroxide for the OER mechanism highlighted by Nocera and co-workers was later further investigated by Smith and co-workers and Koper and co-workers[12,125] The former authors analyzed the surface enhanced Raman spectra (SERS) of an electrodeposited Fe containing Ni(OH)2 on Au electrode in 0.1 m KOH in the region ca. 900–1150 cm−1.[12] The authors observed broad peaks after the transition α/γ that resulted to be pH dependent and attributed them to the formation of “active oxygen” that may exist as peroxide or superoxide, species, in agreement with a previous work by Merril et al.[130] These spe-cies is proposed to origin from deprotonation of NiOOH into Ni(III)OO−, that is negatively charged. Smith and co-workers suggested that this species can act as precursor for the OER and so as the active site for this reaction. Recently, Koper and co-workers further elaborated this mechanism based on meas-urements on electrodeposited Ni(OH)2 in rigorous Fe-free elec-trolyte.[125] They also confirmed for their catalyst the formation

and pH dependence of similar broad peaks in the SERS in the region ca. 800–1150 cm−1. The OER mechanism proposed by Koper and co-workers involves proton-electron decoupled transfer. One of the four concerted proton–electron transfer steps described in the “classical” OER mechanism[127] is decou-pled into two steps: a deprotonation step leading to negatively charged surface species and a following electron transfer step. The deprotonation step has a chemical nature and may occur if the pH of the electrolyte is higher than the pKa of the proton attached to the NiO(OH) (Figure 8e). Two possible deprotona-tion pathways are considered where, in both associated mecha-nisms, O2 forms in the last step, via decomposition of adsorbed OO−*.

7. Summary and Outlook

NiFe-based (oxy)hydroxides and layered double hydroxides have proven to be among the most active OER electrocatalysts in alkaline electrolyte solutions, competing and in many cases surpassing state of the art oxides. In the as-prepared state, the structure is quite well understood, especially for the more crys-talline catalysts. However, there are undeniable evidences of structural transformations induced by anodic electrical poten-tial and likely also by the oxygen evolution reaction. Therefore the relevant structure under OER operating conditions is dif-ferent than the as prepared one and must be determined using in operando methods combined with theoretical predictions and modeling. In this review, we have discussed the recent advance in the determination of the atomic and electronic structure of these excellent catalysts in their OER active state and the efforts towards the understanding of the OER mecha-nism at the molecular level. We believe that these studies set the basis for the rational development of even more active and stable electrocatalysts for this important reaction. Future works should address the fundamental aspects that are still controversial and provide new strategy that will lead to cata-lysts that evolve oxygen at high rates at lower overpotentials and are stable in the time scale of years. Among the fundamental questions that still remain open, the clear determination of the catalytically active sites allowing more accurate estima-tion of intrinsic activity is probably one of the most impor-tant. This task is strongly connected with the clarification of the role of Fe and Ni atoms in the OER mechanism, which in turn requires an undisputable experimental verification. Other challenges include the elucidation of the dependence of the Fe and Ni oxidation state on the OER at different Fe contents, the inclusion of surface charged species in a priori theoretical models, a detailed analysis of the electrical conductivity and a deeper understanding of the active structure providing refined models for DFT calculations. Finally, the establishment of long term stability tests in combination with a detailed analysis of the deactivation mechanisms is required for the application to industrial electrolyzer systems or photoelectrochemical devices. For the latter application, in particular the understanding of the solid-solid interface between the semiconductor and NiFe LDH electrocatalysts needs to be further investigated, including a deeper study of charge injection/transfer dynamics and inter-facial resistance.



Rev

iew

© 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim1600621 (18 of 20) wileyonlinelibrary.com Adv. Energy Mater. 2016, 6, 1600621


AcknowledgementsWe acknowledge financial support by the German Federal Ministry of Education and Research (Bundesministerium für Bildung und Forschung, BMBF) and the German Federal Ministry of Economy and Energy (Bundesministerium für Wirtschaft und Energie, BMWi) under the grant reference number 03SF0433A “MEOKATS”. Financial support by the German Research Foundation (DFG) through grant reference number STR 596/8-1 is grateful acknowledged. P.S. acknowledges partial support by German Federal Ministry of Education and Research (Bundesministerium für Bildung und Forschung, BMBF) under Grant #03SF0523 -“CO2EKAT”. We thank also Mikaela Görlin for the fruitful discussions in these years on NiFe oxides and hydroxides.

Received: March 22, 2016Revised: April 28, 2016

Published online: July 4, 2016

[1] J. Barber, P. D. Tran, J. R. Soc. Interface 2013, 10, 20120984.[2] Y. Umena, K. Kawakami, J. R. Shen, N. Kamiya, Nature 2011, 473,

55.[3] K. N. Ferreira, T. M. Iverson, K. Maghlaoui, J. Barber, S. Iwata,

Science 2004, 303, 1831.[4] H. Dau, C. Limberg, T. Reier, M. Risch, S. Roggan, P. Strasser,

ChemCatChem 2010, 2, 724.[5] W. T. Hong, M. Risch, K. A. Stoerzinger, A. Grimaud, J. Suntivich,

Y. Shao-Horn, Energy Environ. Sci. 2015, 8, 1404.[6] M. W. Kanan, D. G. Nocera, Science 2008, 321, 1072.[7] E. Fabbri, A. Habereder, K. Waltar, R. Kotz, T. J. Schmidt, Catal. Sci.

Technol. 2014, 4, 3800.[8] C. C. L. McCrory, S. H. Jung, J. C. Peters, T. F. Jaramillo, J. Am.

Chem. Soc. 2013, 135, 16977.[9] a) T. Reier, Z. Pawolek, S. Cherevko, M. Bruns, T. Jones,

D. Teschner, S. Selve, A. Bergmann, H. N. Nong, R. Schlogl, K. J. J. Mayrhofer, P. Strasser, J. Am. Chem. Soc. 2015, 137, 13031; b) H. N. Nong, L. Gan, E. Willinger, D. Teschner, P. Strasser, Chem. Sci. 2014, 5, 2955.

[10] a) M. Risch, F. Ringleb, M. Kohlhoff, P. Bogdanoff, P. Chernev, I. Zaharieva, H. Dau, Energy Environ. Sci. 2015, 8, 661; b) M. Zhang, M. de Respinis, H. Frei, Nat. Chem. 2014, 6, 362.

[11] a) A. Bergmann, E. Martinez-Moreno, D. Teschner, P. Chernev, M. Gliech, J. F. de Araujo, T. Reier, H. Dau, P. Strasser, Nat. Commun. 2015, 6, 8625; b) A. Bergmann, I. Zaharieva, H. Dau, P. Strasser, Energy Environ. Sci. 2013, 6, 2745; c) J. Suntivich, K. J. May, H. A. Gasteiger, J. B. Goodenough, Y. Shao-Horn, Science 2011, 334, 1383; d) Y. Gorlin, T. F. Jaramillo, J. Am. Chem. Soc.. 2010, 132, 13612; e) A. Grimaud, K. J. May, C. E. Carlton, Y. L. Lee, M. Risch, W. T. Hong, J. G. Zhou, Y. Shao-Horn, Nat. Commun. 2013, 4, 2439; f) P. W. Menezes, A. Indra, N. R. Sahraie, A. Bergmann, P. Strasser, M. Driess, ChemSusChem 2015, 8, 164; g) M. Gliech, A. Bergmann, C. Spöri, P. Strasser, J. Energy Chem. 2016, 25, 278; h) J. Landon, E. Demeter, N. Inoglu, C. Keturakis, I. E. Wachs, R. Vasic, A. I. Frenkel, J. R. Kitchin, ACS Catal. 2012, 2, 1793; i) Y. L. Zhu, W. Zhou, Z. G. Chen, Y. B. Chen, C. Su, M. O. Tade, Z. P. Shao, Angew. Chem. Int. Ed. 2015, 54, 3897; j) Y. L. Zhu, W. Zhou, Y. B. Chen, J. Yu, M. L. Liu, Z. P. Shao, Adv. Mater. 2015, 27, 7150; k) Y. Zhu, W. Zhou, J. Yu, Y. Chen, M. Liu, Z. Shao, Chem. Mater. 2016, 28, 1691; l) W. Zhou, M. W. Zhao, F. L. Liang, S. C. Smith, Z. H. Zhu, Mater. Horiz. 2015, 2, 495; m) R. C. Liu, F. L. Liang, W. Zhou, Y. S. Yang, Z. H. Zhu, Nano Energy 2015, 12, 115; n) K. Mette, A. Bergmann, J. P. Tessonnier, M. Havecker, L. D. Yao, T. Ressler, R. Schlogl, P. Strasser, M. Behrens, ChemCatChem 2012, 4, 851.

[12] B. J. Trzesniewski, O. Diaz-Morales, D. A. Vermaas, A. Longo, W. Bras, M. T. M. Koper, W. A. Smith, J. Am. Chem. Soc. 2015, 137, 15112.

[13] A. M. Smith, L. Trotochaud, M. S. Burke, S. W. Boettcher, Chem. Commun. 2015, 51, 5261.

[14] M. Dinca, Y. Surendranath, D. G. Nocera, Proc. Natl. Acad. Sci. USA 2010, 107, 10337.

[15] M. Gorlin, M. Gliech, J. F. de Araujo, S. Dresp, A. Bergmann, P. Strasser, Catal. Today 2016, 262, 65.

[16] F. Dionigi, T. Reier, Z. Pawolek, M. Gliech, P. Strasser, ChemSu-sChem 2016, DOI: 10.1002/cssc.201501581.

[17] J. Qi, W. Zhang, R. J. Xiang, K. Q. Liu, H. Y. Wang, M. X. Chen, Y. Z. Han, R. Cao, Adv. Sci. 2015, 2, 1500199.

[18] a) M. Jiang, Y. Li, Z. Lu, X. Sun, X. Duan, Inorg. Chem. Front. 2016, DOI: 10.1039/C5QI00232J.; b) X. Jia, Y. Zhao, G. Chen, L. Shang, R. Shi, X. Kang, G. I. N. Waterhouse, L.-Z. Wu, C.-H. Tung, T. Zhang, Adv. Energy Mater. 2016, DOI: 10.1002/aenm.201502585.

[19] X. Long, G. X. Li, Z. L. Wang, H. Y. Zhu, T. Zhang, S. Xiao, W. Y. Guo, S. H. Yang, J. Am. Chem. Soc. 2015, 137, 11900.

[20] a) B. Mei, A. A. Permyakova, R. Frydendal, D. Bae, T. Pedersen, P. Malacrida, O. Hansen, I. E. L. Stephens, P. C. K. Vesborg, B. Seger, I. Chorkendorff, J. Phys. Chem. Lett. 2014, 5, 3456; b) C. Du, X. G. Yang, M. T. Mayer, H. Hoyt, J. Xie, G. McMahon, G. Bischoping, D. W. Wang, Angew. Chem. Int. Ed. 2013, 52, 12692.

[21] C. G. Morales-Guio, M. T. Mayer, A. Yella, S. D. Tiley, M. Grätzel, X. L. Hu, J. Am. Chem. Soc. 2015, 137, 9927.

[22] L. Wang, F. Dionigi, N. T. Nguyen, R. Kirchgeorg, M. Gliech, S. Grigorescu, P. Strasser, P. Schmuki, Chem. Mater. 2015, 27, 2360.

[23] D. A. Corrigan, R. M. Bendert, J. Electrochem. Soc. 1989, 136, 723.[24] D. A. Corrigan, R. S. Conell, C. A. Fierro, D. A. Scherson, J. Phys.

Chem. 1987, 91, 5009.[25] D. A. Corrigan, J. Electrochem. Soc. 1987, 134, 377.[26] M. Gong, H. J. Dai, Nano Res. 2015, 8, 23.[27] M. F. Shao, R. K. Zhang, Z. H. Li, M. Wei, D. G. Evans, X. Duan,

Chem. Commun. 2015, 51, 15880.[28] Z. H. Zeng, M. K. Y. Chan, Z. J. Zhao, J. Kubal, D. X. Fan, J. Greeley,

J. Phys. Chem. C 2015, 119, 18177.[29] H. Bode, K. Dehmelt, J. Witte, Electrochim. Acta 1966, 11, 1079.[30] R. L. Doyle, I. J. Godwin, M. P. Brandon, M. E. G. Lyons, Phys.

Chem. Chem. Phys. 2013, 15, 13737.[31] S. Klaus, Y. Cai, M. W. Louie, L. Trotochaud, A. T. Bell, J. Phys.

Chem. C 2015, 119, 7243.[32] M. Wehrens-Dijksma, P. H. L. Notten, Electrochim. Acta 2006, 51,

3609.[33] D. S. Hall, D. J. Lockwood, C. Bock, B. R. MacDougall, Proc. – R.

Soc. Edinburgh, Sect. A: Math. Phys. Sci. 2015, 471, 20140792.[34] K. Momma, F. Izumi, J. Appl. Crystallogr. 2011, 44, 1272.[35] A. Van der Ven, D. Morgan, Y. S. Meng, G. Ceder, J. Electrochem.

Soc. 2006, 153, A210.[36] P. Oliva, J. Leonardi, J. F. Laurent, C. Delmas, J. J. Braconnier,

M. Figlarz, F. Fievet, A. Deguibert, J. Power Sources 1982, 8, 229.

[37] M. Casas-Cabanas, J. Canales-Vazquez, J. Rodriguez-Carvajal, M. R. Palacin, J. Am. Chem. Soc. 2007, 129, 5840.

[38] A. Delahayevidal, B. Beaudoin, M. Figlarz, React. Solids 1986, 2, 223.

[39] V. Y. Kazimirov, M. B. Smirnov, L. Bourgeois, L. Guerlou-Demourgues, L. Servant, A. M. Balagurov, I. Natkaniec, N. R. Khasanova, E. V. Antipov, Solid State Ionics 2010, 181, 1764.

[40] Y. F. Li, A. Selloni, J. Phys. Chem. Lett. 2014, 5, 3981.[41] a) A. J. Tkalych, K. Yu, E. A. Carter, J. Phys. Chem. C 2015, 119,

24315; b) K. I. Pandya, R. W. Hoffman, J. Mcbreen, W. E. Ogrady, J. Electrochem. Soc. 1990, 137, 383.

Rev

iew

© 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim (19 of 20) 1600621wileyonlinelibrary.comAdv. Energy Mater. 2016, 6, 1600621


[42] D. K. Bediako, B. Lassalle-Kaiser, Y. Surendranath, J. Yano, V. K. Yachandra, D. G. Nocera, J. Am. Chem. Soc. 2012, 134, 6801.

[43] X. Yang, K. Takada, M. Itose, Y. Ebina, R. Ma, K. Fukuda, T. Sasaki, Chem. Mater. 2008, 20, 479.

[44] a) C. Delmas, J. J. Braconnier, Y. Borthomieu, M. Figlarz, Solid State Ionics 1988, 28, 1132; b) H. Bartl, H. Bode, G. Sterr, J. Witte, Electrochim. Acta 1971, 16, 615.

[45] L. Trotochaud, S. L. Young, J. K. Ranney, S. W. Boettcher, J. Am. Chem. Soc. 2014, 136, 6744.

[46] D. Friebel, M. W. Louie, M. Bajdich, K. E. Sanwald, Y. Cai, A. M. Wise, M. J. Cheng, D. Sokaras, T. C. Weng, R. Alonso-Mori, R. C. Davis, J. R. Bargar, J. K. Norskov, A. Nilsson, A. T. Bell, J. Am. Chem. Soc. 2015, 137, 1305.

[47] F. Cavani, F. Trifiro, A. Vaccari, Catal. Today 1991, 11, 173.[48] R. Z. Ma, J. B. Liang, X. H. Liu, T. Sasaki, J. Am. Chem. Soc. 2012,

134, 19915.[49] P. J. Sideris, U. G. Nielsen, Z. H. Gan, C. P. Grey, Science 2008,

321, 113.[50] S. Cadars, G. Layrac, C. Gerardin, M. Deschamps, J. R. Yates,

D. Tichit, D. Massiot, Chem. Mater. 2011, 23, 2821.[51] G. Abellan, E. Coronado, C. Marti-Gastaldo, J. Waerenborgh,

A. Ribera, Inorg. Chem. 2013, 52, 10147.[52] S. J. Mills, A. G. Christy, J. M. R. Genin, T. Kameda, F. Colombo,

Mineral. Mag. 2012, 76, 1289.[53] G. Abellan, E. Coronado, C. J. Gomez-Garcia, C. Marti-Gastaldo,

A. Ribera, Polyhedron 2013, 52, 216.[54] a) E. Coronado, J. R. Galan-Mascaros, C. Marti-Gastaldo,

A. Ribera, E. Palacios, M. Castro, R. Burriel, Inorg. Chem. 2008, 47, 9103; b) E. Coronado, C. Marti-Gastaldo, E. Navarro-Moratalla, A. Ribera, Appl. Clay Sci. 2010, 48, 228; c) A. I. Khan, D. O'Hare, J. Mater. Chem. 2002, 12, 3191.

[55] Y. Q. Xu, Y. C. Hao, G. X. Zhang, Z. Y. Lu, S. Han, Y. P. Li, X. M. Sun, RSC Adv. 2015, 5, 55131.

[56] W. Y. Hou, L. P. Kang, R. G. Sun, Z. H. Liu, Colloids Surf., A 2008, 312, 92.

[57] F. Song, X. L. Hu, Nat. Commun. 2014, 5, 4477.[58] Y. F. Han, Z. H. Liu, Z. P. Yang, Z. L. Wang, X. H. Tang, T. Wang,

L. H. Fan, K. Ooi, Chem. Mater. 2008, 20, 360.[59] N. Iyi, T. Matsumoto, Y. Kaneko, K. Kitamura, Chem. Mater. 2004,

16, 2926.[60] O. Diaz-Morales, I. Ledezma-Yanez, M. T. M. Koper,

F. Calle-Vallejo, ACS Catal. 2015, 5, 5380.[61] M. A. Oliver-Tolentino, J. Vazquez-Samperio, A. Manzo-Robledo,

R. D. Gonzalez-Huerta, J. L. Flores-Moreno, D. Ramirez-Rosales, A. Guzman-Vargas, J. Phys. Chem. C 2014, 118, 22432.

[62] F. B. D. Saiah, B. L. Su, N. Bettahar, J. Hazard. Mater. 2009, 165, 206.

[63] B. Gregoire, C. Ruby, C. Carteret, Dalton Trans. 2013, 42, 15687.

[64] Z. Lu, W. W. Xu, W. Zhu, Q. Yang, X. D. Lei, J. F. Liu, Y. P. Li, X. M. Sun, X. Duan, Chem. Commun. 2014, 50, 6479.

[65] J. Y. C. Chen, L. N. Dang, H. F. Liang, W. L. Bi, J. B. Gerken, S. Jin, E. E. Alp, S. S. Stahl, J. Am. Chem. Soc. 2015, 137, 15090.

[66] M. Gong, Y. G. Li, H. L. Wang, Y. Y. Liang, J. Z. Wu, J. G. Zhou, J. Wang, T. Regier, F. Wei, H. J. Dai, J. Am. Chem. Soc. 2013, 135, 8452.

[67] T. Xiao, Y. W. Tang, Z. Y. Jia, D. W. Li, X. Y. Hu, B. H. Li, L. J. Luo, Nanotechnology 2009, 20, 475603.

[68] S. Dresp, F. Luo, R. Schmack, S. Kühl, M. Gliech, P. Strasser, Energy Environ. Sci. 2016, DOI: 10.1039/c6ee01046f.

[69] Y. F. Han, H. J. Li, X. R. Ma, Z. H. Liu, Solid State Sci. 2009, 11, 2149.

[70] X. G. Liu, X. Wang, X. T. Yuan, W. J. Dong, F. Q. Huang, J. Mater. Chem. A 2016, 4, 167.

[71] E. Nurlaela, T. Shinagawa, M. Qureshi, D. S. Dhawale, K. Takanabe, ACS Catal. 2016, 6, 1713.

[72] M. W. Louie, A. T. Bell, J. Am. Chem. Soc. 2013, 135, 12329.[73] a) A. Kleiman-Shwarsctein, Y. S. Hu, G. D. Stucky, E. W. McFarland,

Electrochem. Commun. 2009, 11, 1150; b) J. R. Swierk, S. Klaus, L. Trotochaud, A. T. Bell, T. D. Tilley, J. Phys. Chem. C 2015, 119, 19022.

[74] A. S. Batchellor, S. W. Boettcher, ACS Catal. 2015, 5, 6680.[75] Z. H. Li, M. F. Shao, H. L. An, Z. X. Wang, S. M. Xu, M. Wei,

D. G. Evans, X. Duan, Chem. Sci. 2015, 6, 6624.[76] N. C. Lo, P. C. Chung, W. J. Chuang, S. C. N. Hsu, I. W. Sun,

P. Y. Chen, J. Electrochem. Soc. 2016, 163, D9.[77] L. Trotochaud, J. K. Ranney, K. N. Williams, S. W. Boettcher, J. Am.

Chem. Soc. 2012, 134, 17253.[78] L. Demourguesguerlou, J. J. Braconnier, C. Delmas, J. Solid State

Chem. 1993, 104, 359.[79] B. M. Hunter, J. D. Blakemore, M. Deimund, H. B. Gray,

J. R. Winkler, A. M. Muller, J. Am. Chem. Soc. 2014, 136, 13118.[80] M. K. Bates, Q. Jia, H. Doan, W. Liang, S. Mukerjee, ACS Catal.

2016, 6, 155.[81] R. D. L. Smith, M. S. Prevot, R. D. Fagan, Z. P. Zhang,

P. A. Sedach, M. K. J. Siu, S. Trudel, C. P. Berlinguette, Science 2013, 340, 60.

[82] R. D. L. Smith, C. P. Berlinguette, J. Am. Chem. Soc. 2016, 138, 1561.

[83] X. Y. Lu, C. A. Zhao, Nat. Commun. 2015, 6, 6616.[84] M. S. Burke, M. G. Kast, L. Trotochaud, A. M. Smith,

S. W. Boettcher, J. Am. Chem. Soc. 2015, 137, 3638.[85] T. Shinagawa, A. T. Garcia-Esparza, K. Takanabe, Sci. Rep. 2015, 5,

13801.[86] a) E. L. Miller, R. E. Rocheleau, J. Electrochem. Soc. 1997,

144, 3072; b) E. Potvin, L. Brossard, Mater. Chem. Phys. 1992, 31, 311.

[87] X. H. Li, F. C. Walsh, D. Pletcher, Phys. Chem. Chem. Phys. 2011, 13, 1162.

[88] a) M. R. Gao, W. C. Sheng, Z. B. Zhuang, Q. R. Fang, S. Gu, J. Jiang, Y. S. Yan, J. Am. Chem. Soc. 2014, 136, 7077; b) A. Singh, S. L. Y. Chang, R. K. Hocking, U. Bach, L. Spiccia, Energy Environ. Sci. 2013, 6, 579.

[89] B. S. Yeo, A. T. Bell, J. Phys. Chem. C 2012, 116, 8394.[90] M. E. G. Lyons, A. Cakara, P. O’Brien, I. Godwin, R. L. Doyle, Int.

J. Electrochem. Sci. 2012, 7, 11768.[91] R. Subbaraman, D. Tripkovic, K. C. Chang, D. Strmcnik,

A. P. Paulikas, P. Hirunsit, M. Chan, J. Greeley, V. Stamenkovic, N. M. Markovic, Nat. Mater. 2012, 11, 550.

[92] S. Klaus, L. Trotochaud, M. J. Cheng, M. Head-Gordon, A. T. Bell, ChemElectroChem 2016, 3, 66.

[93] M. E. G. Lyons, M. P. Brandon, J. Electroanal. Chem. 2010, 641, 119.

[94] M. S. Burke, S. H. Zou, L. J. Enman, J. E. Kellon, C. A. Gabor, E. Pledger, S. W. Boettcher, J. Phys. Chem. Lett. 2015, 6, 3737.

[95] S. H. Zou, M. S. Burke, M. G. Kast, J. Fan, N. Danilovic, S. W. Boettcher, Chem. Mater. 2015, 27, 8011.

[96] a) W. D. Chemelewski, H. C. Lee, J. F. Lin, A. J. Bard, C. B. Mullins, J. Am. Chem. Soc. 2014, 136, 2843; b) K. Jun, Y. S. Lee, T. Buonassisi, J. M. Jacobson, Angew. Chem. Int. Ed. 2012, 51, 423.

[97] a) Y. Z. Wu, M. X. Chen, Y. Z. Han, H. X. Luo, X. J. Su, M. T. Zhang, X. H. Lin, J. L. Sun, L. Wang, L. Deng, W. Zhang, R. Cao, Angew. Chem. Int. Ed. 2015, 54, 4870; b) M. X. Chen, Y. Z. Wu, Y. Z. Han, X. H. Lin, J. L. Sun, W. Zhang, R. Cao, ACS Appl. Mater. Interfaces 2015, 7, 21852; c) J. A. Seabold, K. S. Choi, J. Am. Chem. Soc. 2012, 134, 2186.

Rev

iew

© 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim1600621 (20 of 20) wileyonlinelibrary.com Adv. Energy Mater. 2016, 6, 1600621


[98] R. Subbaraman, D. Tripkovic, D. Strmcnik, K. C. Chang, M. Uchimura, A. P. Paulikas, V. Stamenkovic, N. M. Markovic, Science 2011, 334, 1256.

[99] H. Ali-Löytty, M. W. Louie, M. R. Singh, L. Li, H. G. Sanchez Casalongue, H. Ogasawara, E. J. Crumlin, Z. Liu, A. T. Bell, A. Nilsson, D. Friebel, J. Phys. Chem. C 2016, 120, 2247.

[100] X. Long, J. K. Li, S. Xiao, K. Y. Yan, Z. L. Wang, H. N. Chen, S. H. Yang, Angew. Chem. Int. Ed. 2014, 53, 7584.

[101] Y. Y. Feng, H. J. Zhang, Y. Zhang, X. Li, Y. Wang, ACS Appl. Mater. Interfaces 2015, 7, 9203.

[102] W. Ma, R. Z. Ma, C. X. Wang, J. B. Liang, X. H. Liu, K. C. Zhou, T. Sasaki, ACS Nano 2015, 9, 1977.

[103] D. C. Xia, L. Zhou, S. Qiao, Y. L. Zhang, D. Tang, J. Liu, H. Huang, Y. Liu, Z. H. Kang, Mater. Res. Bull. 2016, 74, 441.

[104] X. W. Yu, M. Zhang, W. J. Yuan, G. Q. Shi, J. Mater. Chem. A 2015, 3, 6921.

[105] C. Tang, H. F. Wang, H. S. Wang, F. Wei, Q. Zhang, J. Mater. Chem. A 2016, 4, 3210.

[106] D. Tang, J. Liu, X. Y. Wu, R. H. Liu, X. Han, Y. Z. Han, H. Huang, Y. Liu, Z. H. Kang, ACS Appl. Mater. Interfaces 2014, 6, 7918.

[107] R. Chen, G. Sun, C. Yang, L. Zhang, J. Miao, H. Tao, H. Yang, J. Chen, P. Chen, B. Liu, Nanoscale Horiz. 2016, DOI: 10.1039/C5NH00082C.

[108] X. L. Zhu, C. Tang, H. F. Wang, Q. Zhang, C. H. Yang, F. Wei, J. Mater. Chem. A 2015, 3, 24540.

[109] Y. M. Li, H. Y. He, W. Fu, C. Z. Mu, X. Z. Tang, Z. Liu, D. Z. Chi, X. Hu, Chem. Commun. 2016, 52, 1439.

[110] B. M. Hunter, W. Hieringer, J. R. Winkler, H. B. Gray, A. M. Muller, Energy Environ. Sci. 2016, DOI: 10.1039/C6EE00377J.

[111] Z. Y. Lu, L. Qian, Y. Tian, Y. P. Li, X. M. Sun, X. Duan, Chem. Commun. 2016, 52, 908.

[112] C. J. Zhang, R. D. Fagan, R. D. L. Smith, S. A. Moore, C. P. Berlinguette, S. Trudel, J. Mater. Chem. A 2015, 3, 756.

[113] a) J. A. Haber, C. X. Xiang, D. Guevarra, S. H. Jung, J. Jin, J. M. Gregoire, ChemElectroChem 2014, 1, 524; b) J. B. Gerken, S. E. Shaner, R. C. Masse, N. J. Porubsky, S. S. Stahl, Energy Environ. Sci. 2014, 7, 2376; c) J. Y. C. Chen, J. T. Miller, J. B. Gerken, S. S. Stahl, Energy Environ. Sci. 2014, 7, 1382.

[114] R. Frydendal, E. A. Paoli, B. P. Knudsen, B. Wickman, P. Malacrida, I. E. L. Stephens, I. Chorkendorff, ChemElectroChem 2014, 1, 2075.

[115] F. J. Perez-Alonso, C. Adan, S. Rojas, M. A. Pena, J. L. G. Fierro, Int. J. Hydrogen Energy 2014, 39, 5204.

[116] N. Han, F. P. Zhao, Y. G. Li, J. Mater. Chem. A 2015, 3, 16348.[117] J. S. Luo, J. H. Im, M. T. Mayer, M. Schreier, M. K. Nazeeruddin,

N. G. Park, S. D. Tilley, H. J. Fan, M. Gratzel, Science 2014, 345, 1593.[118] Y. F. Li, A. Selloni, ACS Catal. 2014, 4, 1148.[119] D. N. Wang, J. G. Zhou, Y. F. Hu, J. L. Yang, N. Han, Y. G. Li,

T. K. Sham, J. Phys. Chem. C 2015, 119, 19573.[120] M. Balasubramanian, C. A. Melendres, S. Mini, J. Phys. Chem. B

2000, 104, 4300.[121] M. Gorlin, P. Chernev, J. F. De Araujo, T. Reier, S. Dresp, B. Paul,

R. Kraehnert, H. Dau, P. Strasser, J. Am. Chem. Soc. 2016, DOI: 10.1021/jacs.6b00332.

[122] a) L. Demourguesguerlou, L. Fournes, C. Delmas, J. Solid State Chem. 1995, 114, 6; b) P. Axmann, O. Glemser, J. Alloys Compd. 1997, 246, 232.

[123] S. Axnanda, E. J. Crumlin, B. H. Mao, S. Rani, R. Chang, P. G. Karlsson, M. O. M. Edwards, M. Lundqvist, R. Moberg, P. Ross, Z. Hussain, Z. Liu, Sci. Rep. 2015, 5, 9788.

[124] M. T. M. Koper, Phys. Chem. Chem. Phys. 2013, 15, 1399.[125] O. Diaz-Morales, D. Ferrus-Suspedra, M. T. M. Koper, Chem. Sci.

2016, DOI: 10.1039/C5SC04486C.[126] D. K. Bediako, Y. Surendranath, D. G. Nocera, J. Am. Chem. Soc.

2013, 135, 3662.[127] a) I. C. Man, H. Y. Su, F. Calle-Vallejo, H. A. Hansen, J. I. Martinez,

N. G. Inoglu, J. Kitchin, T. F. Jaramillo, J. K. Norskov, J. Rossmeisl, ChemCatChem 2011, 3, 1159; b) J. Rossmeisl, A. Logadottir, J. K. Norskov, Chem. Phys. 2005, 319, 178.

[128] a) J. O. Bockris, T. Otagawa, J. Phys. Chem. 1983, 87, 2960; b) N. B. Halck, V. Petrykin, P. Krtil, J. Rossmeisl, Phys. Chem. Chem. Phys. 2014, 16, 13682; c) L. A. DeFaria, J. F. C. Boodts, S. Trasatti, J. Appl. Electrochem. 1996, 26, 1195; d) G. Mattioli, P. Giannozzi, A. A. Bonapasta, L. Guidoni, J. Am. Chem. Soc. 2013, 135, 15353; e) C. P. Plaisance, R. A. van Santen, J. Am. Chem. Soc. 2015, 137, 14660.

[129] a) M. Bajdich, M. Garcia-Mota, A. Vojvodic, J. K. Norskov, A. T. Bell, J. Am. Chem. Soc. 2013, 135, 13521; b) J. Chen, A. Selloni, J. Phys. Chem. C 2013, 117, 20002.

[130] M. Merrill, M. Worsley, A. Wittstock, J. Biener, M. Stadermann, J. Electroanal. Chem. 2014, 717, 177.

nife‐based (oxy)hydroxide catalysts for oxygen evolution ......nife-based (oxy)hydroxide catalysts...

Documents