ARTICLE
In-situ structure and catalytic mechanism of NiFe
and CoFe layered double hydroxides during
oxygen evolution
Fabio Dionigi1,9✉, Zhenhua Zeng 2,9✉, Ilya Sinev3,4, Thomas Merzdorf1, Siddharth Deshpande2,
Miguel Bernal Lopez3,4, Sebastian Kunze3,4, Ioannis Zegkinoglou 3,4, Hannes Sarodnik1, Dingxin Fan2,
Arno Bergmann1,4, Jakub Drnec 5, Jorge Ferreira de Araujo1, Manuel Gliech1, Detre Teschner6,7, Jing Zhu8,
Wei-Xue Li 8, Jeffrey Greeley2, Beatriz Roldan Cuenya 4✉& Peter Strasser 1✉
NiFe and CoFe (MFe) layered double hydroxides (LDHs) are among the most active elec-
trocatalysts for the alkaline oxygen evolution reaction (OER). Herein, we combine electro-
chemical measurements, operando X-ray scattering and absorption spectroscopy, and density
functional theory (DFT) calculations to elucidate the catalytically active phase, reaction
center and the OER mechanism. We provide the first direct atomic-scale evidence that, under
applied anodic potentials, MFe LDHs oxidize from as-prepared α-phases to activated γ-
phases. The OER-active γ-phases are characterized by about 8% contraction of the lattice
spacing and switching of the intercalated ions. DFT calculations reveal that the OER proceeds
via a Mars van Krevelen mechanism. The flexible electronic structure of the surface Fe sites,
and their synergy with nearest-neighbor M sites through formation of O-bridged Fe-M
reaction centers, stabilize OER intermediates that are unfavorable on pure M-M centers and
single Fe sites, fundamentally accounting for the high catalytic activity of MFe LDHs.
https://doi.org/10.1038/s41467-020-16237-1 OPEN
1The Electrochemical Energy, Catalysis, and Materials Science Laboratory, Department of Chemistry, Chemical Engineering Division, Technical University
Berlin, Strasse des 17. Juni 124, 10623 Berlin, Germany. 2Davidson School of Chemical Engineering, Purdue University, West Lafayette, Indiana 47907, USA.
3Department of Physics, Ruhr-University Bochum, Universitaetsstrasse 150, 44801 Bochum, Germany. 4Department of Interface Science, Fritz-Haber-
Institut der Max-Planck-Gesellschaft, Faradayweg 4 –6, 14195 Berlin, Germany. 5European Synchrotron Radiation Facility, ID 31 Beamline, BP 220, F-38043
Grenoble, France. 6Department of Inorganic Chemistry, Fritz-Haber-Institut der Max-Planck-Gesellschaft, Faradayweg 4–6, 14195 Berlin, Germany. 7Max
Planck Institute for Chemical Energy Conversion, Stiftstrasse 34-36, 45470 Mülheim an der Ruhr, Germany. 8CAS Excellence Center for Nanoscience, Hefei
National Laboratory for Physical Sciences at Microscale, School of Chemistry and Materials Science, University of Science and Technology of China, Hefei,
Anhui 230026, China.
9
These authors contributed equally: Fabio Dionigi, Zhenhua Zeng. ✉email: fabio.dionigi@tu-berlin.de;zeng46@purdue.edu;
NATURE COMMUNICATIONS | (2020) 11:2522 | https://doi.org/10.1038/s41467-020-16237-1 | www.nature.com/naturecommunications 1
1234567890():,;
Water splitting to generate O
2
and H
2
has been a major
focus of (photo)electrochemical energy storage and
conversion research, but fundamental and practical
challenges remain. In this process, O
2
generation at the anode
through the oxygen evolution reaction (OER), which is inherently
slower by over four orders of magnitude compared with H
2
generation, accounts for the majority of energy losses1. NiFe-
based layered hydroxides are the most active OER catalysts in
base and are the catalysts of choice for industrial water electro-
lysis2–10, whereas CoFe-based layered hydroxides have compar-
able performance7,8,10–12. Very recently, it has been found that
NiFe and CoFe (MFe) layered (oxy)hydroxides are also the
common active phases of other highly active OER catalysts,
including perovskite oxides13,14, spinel oxides15, phosphides16,
and potentially other Co- and Ni-based OER catalysts with Fe
incorporated intentionally or accidentally, such as carbides17,
nitrides18, sulfides19, and selenides20, which are prone to hydro-
lysis and oxidation under OER conditions13,15,16,21,22. Thus,
studying the reactive structures of the MFe layered double
hydroxides (LDHs) under in-situ conditions and the catalytic
mechanism can provide a thorough understanding of the
structure–property relationships of many related catalysts and
potentially lead to the design of new catalysts with further
improved performance.
In spite of previous reports on the ex-situ crystal structure of
the as-synthesized precursors of MFe LDH catalysts23–28 and in-
situ local structure based on X-ray absorption spectroscopy (XAS)
measurements3,4,12,29–32, little is known about the long-range
crystal structures of the catalytically active phase under OER
conditions. As a result, most proposals regarding the in-situ
crystal structures of NiFe and CoFe LDHs under OER conditions
are indirectly inferred from the crystal structures of the host Ni
and Co oxyhydroxides, respectively. More specifically, for NiFe
LDH, a γ-NiOOH-type phase, in which water and cations are
intercalated between layers28, has long been speculated4,5,24,25,33.
However, no direct evidence has been observed to confirm this
hypothesis, as previous in-situ structural studies could not pro-
vide the characteristic interlayer spacing that can be used to
differentiate between the γ-NiOOH-type phase and other com-
mon phases, such as the anhydrous β-NiOOH-type phase28. For
CoFe LDH, in analogy to NiFe LDH, a transformation to a γ-
NiOOH-type phase can be hypothesized under OER conditions.
However, there is no analogous γ-CoOOH phase with species
intercalated between layers; the other two known β-CoOOH and
CoO
2
phases show no intercalation34. As a consequence, a Fe-
doped β-CoOOH has been proposed as the active phase of CoFe
LDH under OER conditions11,12.
Density functional theory (DFT) calculations allow us to
examine all of the above hypotheses and to extract atomic-scale
details by screening suitable candidate phases and comparing
their relative stability with that of known phases. Although sig-
nificant efforts have been made, particularly on the modeling of
the electronic structure effects and catalytic mechanism of Ni-
based catalysts for OER4,10–12,33,35–40, such a screening and
comparison has not yet been rigorously carried out because of the
structural complexity of the active phases. Indeed, even the
atomic-scale structure of the γ-NiOOH phase itself is still
unclear4,33,38. The lack of these atomic-scale details has, in turn,
made it highly challenging to choose appropriate models for
DFT-based mechanistic studies4,38,41. Hence, a variety of struc-
tures have been employed in the modeling, including those that
resemble as-synthesized precursor phases37, NiO38, two-
dimensional single layer (oxy)hydroxides35,39,β-MOOH
analogs4,10–12,42, and γ-NiOOH analogs33,40,43–45 with or with-
out Fe dopants. Although significant efforts have been made to
explain the high activity of MFe LDHs, the diversity of studies
suggests that large uncertainties exist concerning the relation-
ship between the active site structure and the catalytic
mechanism. This is because the predicted activity of the catalysts
is highly sensitive to, and is an ensemble of, the geometrical
structure46,47 and electronic structure (oxidation state)48,49 of
the active site, as well as non-covalent interactions originating
from bulk crystal structure50,51, the steady state of the surface
configuration52,53, and the electronic structure methods used in
the calculations54–56. These uncertainties, resulting from an
incomplete consideration of this ensemble of factors, have hin-
dered the mechanistic understanding of the high activity of NiFe
andCoFeLDHsfortheOER,whichfurtherhampersthepre-
diction of new catalysts with improved performance.
Herein, we combine electrochemical measurements with
operando wide-angle X-ray scattering (WAXS) and XAS data, as
well as ab initio molecular dynamic simulations and a synergistic
DFT approach that was benchmarked specifically for the strongly
correlated Fe, Co, and Ni oxides and (oxy)hydroxides55,to
unravel and contrast the crystal structures and electrocatalytic
OER mechanisms of the active phases of NiFe and CoFe LDH
catalysts. We provide the first direct atomic-scale evidence that,
under OER conditions, both NiFe and CoFe LDHs transform
from the as-prepared α-phase to a deprotonated γ-phase. The
oxidative phase transitions are characterized by ~8% contractions
in both the in-plane lattice constant and the interlayer distance,
which are induced by the oxidation of Fe and M (Ni, Co), and by
the anion-to-cation switching of intercalated ions, respectively.
We then adopt the in-situ identified γ-phases to study the OER
mechanism through DFT-based calculations. The calculated
surface phase diagrams indicate that surface O sites are saturated
with H by forming bridge OH, and undercoordinated metal sites
are saturated with atop OH under OER conditions. These
structures, and the associated reaction free energies, suggest that
the OER proceeds via a Mars van Krevelen mechanism, starting
with the oxidation of bridge OH at the Fe-M reaction centers
(M =Ni or Co) to form O-bridged Fe-M moieties. The flexible
electronic structure of the Fe site and its synergy with the nearest-
neighbor M sites through the formation of the O-bridged Fe-M
reaction centers fundamentally accounts for the high OER activity
of MFe oxyhydroxides due to the stabilization of OER inter-
mediates that are unfavorable on pure M-M centers and single Fe
sites. Our combined operando experimental and DFT computa-
tional approach thus provides a consistent atomic-scale expla-
nation for the high OER activity of the MFe LDHs.
Results
Electrochemical oxygen evolution and surface redox chemistry.
We studied the redox chemistry of NiFe LDH and CoFe LDH (M:
Fe =~3:1) using cyclic and linear sweep voltammetry (CV and
LSV) and compared their OER performance with that of their Fe-
free hydroxide analogs, including β-Ni(OH)
2
and β-Co(OH)
2
.
LSV curves (Fig. 1a) indicated that OER overpotentials at 10 mA
cm−2are +348 mV and +404 mV for NiFe LDH and CoFe LDH,
respectively, which makes them among the most active electro-
catalysts in alkaline conditions. NiFe and CoFe LDHs also
exhibited substantially higher catalytic activity than the hydro-
xides containing only Ni and Co. For NiFe LDH, the over-
potential is 225 mV lower than that of NiOOH, whereas for CoFe
LDH, the corresponding overpotential is 64 mV lower than that
of CoOOH. We note that, although it is not an intrinsic metric,
the overpotential measured at 10 mA cm−2from LSV is a valid
practical parameter to compare the activity trends of the cata-
lysts57. This is confirmed by the good agreement with the trends
of the intrinsic activity extracted with two distinct methods (see
discussion in Supplementary Methods and Supplementary Figs. 1
ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-16237-1
2NATURE COMMUNICATIONS | (2020) 11:2522 | https://doi.org/10.1038/s41467-020-16237-1 | www.nature.com/naturecommunications
and 2). Also, the trend of our measurement is consistent with
what was reported for electrodeposited films of similar
composition7.
The CV curves in Fig. 1b indicate that both NiFe LDH and
CoFe LDH undergo redox transitions prior to (or slightly
overlapping with) the onset of the OER, as confirmed by
differential electrochemical mass spectrometry (DEMS) by
comparing the Faraday current and the mass spectrum current
related to mass m/z=32 (Fig. 1c, d). For NiFe LDH, the Ni(II)
oxidation peak at +1.47 V
RHE
(A′) overlaps with the OER onset
and with the corresponding reduction wave peaks at +1.35 V
RHE
(B′). For CoFe LDH, the main oxidation peak at +1.35 V
RHE
(A)
occurs clearly prior to any OER onset. More anodically, a second
and small oxidation shoulder at around +1.55 V
RHE
(C) overlaps
with the OER. The broad peaks B at +1.1 V
RHE
and D at +1.4 V
RHE
constitute the corresponding reduction waves on the cathodic
scan, respectively. These redox features, in turn, provide strong
evidence that the active phases for OER are not the as-synthesized
phases (characterized in the Supplementary Methods and Supple-
mentary Fig. 3).
Tracking structural transformations during activation.To
follow the phase transition of the catalysts from their as-
synthesized precursor state into the catalytically active states,
synchrotron-based operando WAXS analysis was employed. In-
situ WAXS measurements were taken in 0.1 M KOH, starting
from the resting state (+1V
RHE
) of the catalysts, followed by
stepping the applied potential up to +1.7 V
RHE
and then back
down to the resting state or even lower potentials. The scattering
pattern was measured at the end of each step (i.e., Supplementary
Fig. 4). The potential window ranges from values closely prior to
the anodic wave of M(II) oxidation, reaching into the OER region
and then reverting to low values to ensure the reduction to M(II).
We will initially focus on the evolution of the (003) diffraction
peak of the LDHs (Fig. 2a, b), which provides the characteristic
interlayer distance that is absent from XAS measurements and
that is central to differentiating the phases with and without
intercalation of water molecules and ions. For both MFe LDHs,
the evolution of the (003) peak indicates a contraction of the
interlayer distance in the anodic scan and a re-expansion in the
cathodic scan. The detailed interlayer distances obtained by
Rietveld refinement are shown in Fig. 2c, d (additional details in
Supplementary Figs. 5–9).
At the resting state and the potential before M(II) oxidation,
the measured interlayer distances are 7.8 Å and 7.7 Å for NiFe
and CoFe LDHs, respectively, which are typical for LDHs with
intercalated water molecules and carbonate anions between the
layers23–28. As these interlayer distances resemble that of α-Ni
(OH)
2
(~8 Å, as proposed in the Bode’s diagram28,58), we named
this phase the α-MFe LDH. As soon as the potential increased
above the M(II) oxidation potential, the (003) reflections shifted
to shorter interlayer distances and a shoulder (Supplementary
Fig. 9) started to develop at the interlayer distance of 7.2 Å and
7.1 Å for NiFe and CoFe LDHs, respectively. These interlayer
distances are much larger than those of the anhydrous β-NiOOH
(~4.8 Å)59 and β-CoOOH (~4.4 Å)60 phases but are close to that
of the hydrous γ-NiOOH phase (i.e., ~7 Å)28,59. In analogy to γ-
NiOOH and previous literature4,24,61, we refer to these new
phases as γ-MFe LDHs.
During the cathodic scan, the interlayer distances started to re-
expand as the reduction to M(II) occurred. However, the
processes depended sensitively on the nature of M. For NiFe
LDH, the shoulder at the interlayer distance of 7.2 Å (γ-phase)
disappeared at the resting state, and the peak restored to the
original value of 7.8 Å (α-phase), which indicates the reversibility
of the α-to-γtransformation. Differently, for the CoFe LDH
(Fig. 2b), the re-expansion to the original value (7.7 Å) is very
limited under the resting state (1 V
RHE
) and is still incomplete at
lower potentials (0.5 V
RHE
). The limited reversibility occurring in
CoFe LDH has also been observed during electrochemical
activation treatments (Supplementary Fig. 10) and verified by
ab
cd
Fig. 1 Surface chemistry and OER of NiFe and CoFe LDHs. a Linear sweep voltammetry of NiFe LDH (black), CoFe LDH (red), β-Ni(OH)
2
(blue), and β-Co
(OH)
2
(green) at a scan rate of 1 mV s-1 in purified 0.1 M KOH by RDE (1600 r.p.m.). Catalyst loading on GC electrodes: 0.1 mg cm−2.bStable curves
obtained in cyclic voltammetry of NiFe LDH (black) and CoFe LDH (red) in 0.1 M KOH in the grazing incident cell. Redox features are indicated with capital
letters. c,dDifferential electrochemical mass spectrometry (DEMS) of NiFe LDH (c) and CoFe LDH (d) during a linear sweep voltammetry (LSV) in 0.1 M
KOH. The faradaic current normalized by the geometric area is shown in red, whereas the mass spectrum current related to mass m/z=32 is shown
in blue.
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-16237-1 ARTICLE
NATURE COMMUNICATIONS | (2020) 11:2522 | https://doi.org/10.1038/s41467-020-16237-1 | www.nature.com/naturecommunications 3
ex-situ soft X-ray XAS (sXAS) (Supplementary Figs. 11 and 12).
In addition, Co-based hydroxides also have shown irreversible
behavior in the literature32.
After clarifying the catalytically active phases under OER
condition via the (003) reflection, we now turn to the (110)
reflection representing the in-plane lattice constants (Fig. 2e, f and
SupplementaryFig.13).Asthe(110)reflection is much weaker than
the (003), and broadened under OER conditions, extracting exact
lattice parameters is non-trivial. Nonetheless, our Rietveld refine-
ment revealed an unambiguous trend toward shorter metal–metal
distances, from ~3.1Å to ~2.85Å upon α-to-γphase transitions for
both NiFe and CoFe LDHs (Fig. 2g, h). This trend agrees well with
the contraction of the local metal-O and metal–metal distances in
previous in-situ EXAFS measurements3,4,12,29–32.Thus,thereare
contractions on both interlayer distances and in-plane bonds upon
the α-to-γphase transition.
We note that, similar to what has been observed in previous
measurements with in-situ XAS31,62 and Mössbauer spectroscopy
on NiFe-based oxyhydroxides63, only a fraction of MFe LDHs in
our operando WAXS measurements undergo phase transitions
under OER potentials, although the fraction is higher for CoFe
than NiFe LDH (Supplementary Fig. 9). The incomplete phase
transition is likely because some nanoplates in the catalyst film
are not electrochemically accessible, e.g., not in contact with the
1.5 VRHE
1.2 VRHE
(003)
1.65 VRHE
1.1 VRHE
1.35 VRHE 1.1 VRHE 1.0 VRHE
1.7 VRHE
0.5 VRHE
(003)
1.15 VRHE
6.5
24.0
α: 7.8 Å
α: 7.7 Å
γ: 7.2 Å
α: 3.11 Å α: 3.13 Å
γ: 2.83 Å γ: 2.85 Å
γ: 7.1 Å
23.4
22.8
22.2
Lattice parameter c (Å)
Interlayer distance d (Å)
21.6
21.0
3.1
Collapsed
film
technique
Regular
meas.
3.0
Lattice parameter a (Å)
2.9
2.8
3.1
3.0
Lattice parameter a (Å)
2.9
2.8
8.0
7.8
7.6
7.4
7.2
7.0
1.6 V,
50 min
1.6 V,
40 min
1.5 V,
40 min
50 CV
Wet
Dry
1.7 V
0.9 V
1.2 V
0.5 V
1.3 V
1.0 V
50 CV
Wet
Dry
Interlayer distance d (Å)
8.0
7.8
7.6
7.4
7.2
7.0
24.0
23.4
22.8
22.2
Lattice parameter c (Å)
21.6
21.0
1.3 1.4 1.5
d (Å)
(113) (110) (113) (110)
1.6 1.3 1.4 1.5
d (Å)
1.6
Dry
Wet
1.2 VRHE
1.5 VRHE
1.6 VRHE
1.1 VRHE
Dry
Wet
1.2 VRHE
1.5 VRHE
1.6 VRHE
1.1 VRHE
Dry
Wet
1 VRHE
1.3 VRHE
1.7 VRHE
0.9 VRHE
Dry
Wet
1 VRHE
1.3 VRHE
1.7 VRHE
0.9 VRHE
7.0 7.5 8.0
Interlayer distance (Å)
8.5 9.0 6.0 6.5 7.0 7.5 8.0
Interlayer distance (Å)
8.5 9.0
ab
cd
ef
gh
Fig. 2 The evolution of the interlayer spacing and the intralayer metal–metal distances of NiFe and CoFe LDHs from WAXS measurement.
a,bWaterfall plot of normalized and background-subtracted (003) peak obtained during in-situ WAXS in 0.1 M KOH and potential steps for NiFe LDH (a)
and CoFe LDH (b). c,dInterlayer distances for NiFe LDH (c) and CoFe LDH (d) obtained by by Rietveld refinement. Full and open symbols are used for
different phases. The error bars represent the SE provided by Topas. e,fIn-situ WAXS patterns for d-values close to the (110) peak of NiFe LDH (e) and
CoFe LDH (f) under various conditions. For NiFe LDH, the WAXS patterns at the reported potentials have been obtained by the collapsed film technique.
In e, the dashed arrows point to the feature associated to the γ-phase. g,hLattice parameter a, corresponding to the intralayer metal–metal distance in
NiFe LDH (g) and CoFe LDH (h) obtained by Rietveld refinement. Full and open symbols are used for different phases. Error bars represent SD provided by
Topas for the refined parameters.
ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-16237-1
4NATURE COMMUNICATIONS | (2020) 11:2522 | https://doi.org/10.1038/s41467-020-16237-1 | www.nature.com/naturecommunications
electrolyte or with the external electrical circuit (see Supplemen-
tary Information for detailed discussion and Supplementary
Figs. 14 and 15). This incompleteness makes the quantitative
interpretation of XAS data challenging, as the measured local
structures and electronic structures are weighted averages of the
two phases. Thus, the ensemble-averaged structural parameters
and the electronic structure do not necessarily reflect the actual
crystal structure parameters and the electronic structure of a
specific phase, but strongly depend on the ratio of the α-to-γ
phase transition. This issue is known for unsupported NiFe
(oxy)hydroxide nanocataylsts31, and confirmed by our operando
XAS (see Supplementary Methods, Supplementary Figs. 16–25,
and Supplementary Tables 1–4). Therefore, what sets the present
operando WAXS measurements apart from other ensemble-
averaging approaches is their ability to probe both intrinsic local
and longer-range geometric effects of specific phases, providing
essential information for the identification of the active phase
under OER which cannot be achieved by the experimental
techniques that solely provide average local structure information.
This intrinsic structural information can serve as the reference for
DFT calculations to study atomic-scale geometric structures and
the intrinsic electronic structure of γ-MFe LDHs, which can in
turn be further employed to study the catalytic mechanism for
OER. We note that, while the γphase is the focus of the study, a
consistent measurement of the αphase is important for
establishing a general picture regarding the completeness and
the reversibility of the phase transition.
Geometric and electronic structures from DFT calculations.
Following the order in the above experimental section, we begin
by discussing the as-prepared MFe phases (M:Fe =3:1, α-MFe
LDHs). DFT calculations indicate that α-MFe LDHs adopt the
structure of hydrotalcite (Mg
6
Al
2
CO
3
(OH)
16
·4H
2
O), which is the
archetypical LDH material with its characteristic three layer
rhombohedral structure. Hydrotalcite formation is favorable from
the component (oxy)hydroxides (FeOOH, M(OH)
2
), water (in
electrolyte), and CO
2
(in atmosphere) (Supplementary Figs. 26–
28 and Supplementary Table 5), which highlights the reliability
of the present calculations. In these M
6
Fe
2
CO
3
(OH)
16
·4H
2
O
structures, Fe3+ions are separated by M2+cations within the
layer, and the H
2
O and CO
3
2−ions are intercalated between
layers in a flat configuration, interconnected through hydrogen
bonds. The intercalated species are further connected with the
M
0.75
Fe
0.25
(OH)
2
sheets by accepting hydrogen bonds from the
OH terminations of the sheets (see Fig. 3). The calculated inter-
layer distances are 7.7 Å for both NiFe and CoFe LDHs, which is
fully consistent with the measured distances of 7.7 Å−7.8 Å. The
calculated in-plane lattice constants are 3.10 Å and 3.15 Å for
NiFe and CoFe LDHs, respectively, which also fully agree with the
measured WAXS values (3.11 Å and 3.13 Å, respectively; Fig. 2).
To identify the catalytically active phases of MFe LDHs under
OER conditions, we first calculated a series of structures and
configurations of γ-NiOOH with seven possible nominal
oxidation states of Ni, varying from 3+to 4+, and various
amounts of water molecules and ions intercalated through ab-
initio molecular dynamics (AIMD) simulations (see Fig. 3a, b).
We then used the most plausible γ-NiOOH as the basis to study
the possible configuration of γ-MFe LDHs. Among the structures
considered, a phase with 4 water molecules and 2K+cations
intercalated between M
6
Fe
2
O
16
layers is the most plausible phase
under OER conditions. This conclusion is suggested by the
favorable formation energies from its components ((hydroxy)
oxides, water, and cations in the electrolyte, see Supplementary
Table 5), and by the stability under OER conditions (see Fig. 3).
The interlayer distances and the in-plane lattice constants are
7.18 Å and 2.84 Å, respectively, for both γ-type NiFe and CoFe
LDH phases. These values are in excellent agreement with the
measured values during OER of the MFe γ-phases: 7.1-7.2 Å and
~2.85 Å, respectively. We note that the anhydrous phases with
similar overall oxidation state (M
0.75
Fe
0.25
OOH
0.25
and M
0.75
Fe
0.25
O
2
;
see Supplementary Table 6) exhibit a similar in-plane lattice
constant, yet the interlayer distance is ~4.6 Å. The similarity in the
in-plane lattice constants of these two distinct phases strongly
suggests that local metal–metal distance alone is insufficient to
accurately identify the crystal phase present under OER condition.
This fact underscores that for the present catalyst systems, the
operando scattering analysis is the best technique for identifying
the 3D structure of the catalytically active phases.
As described above, DFT calculations indicate that under OER
conditions, MFe LDHs transform from the as-synthesized phase
with the stoichiometry M
6
Fe
2
CO
3
(OH)
16
·4H
2
Ototheγ-phase
with the stoichiometry M
6
Fe
2
K
2
O
16
·4H
2
O. We note that,
consistent with previous measurement with Raman spectro-
scopy64, there are no hydroxyl groups in M
0.75
Fe
0.25
O
2
layers in
the γ-phase. The deprotonation of the hydroxyls of the α-phase, in
turn, breaks the hydrogen bonds that exist between them and the
carbonate anions and makes the intercalation of the latter highly
unfavorable. Thus, CO
3
2−ions are expelled and K+ions are
intercalated from the electrolyte during the α-γphase transition.
K+ions connect M
0.75
Fe
0.25
O
2
sheets by forming O-K-O ionic
bonds in the form of zigzag chains. The channels between the
zigzag K+chains are filled with water molecules to fully saturate
the remaining oxygen atoms in the M
0.75
Fe
0.25
O
2
layers through
the formation of O-HOH-O hydrogen bonds (see Fig. 3). Based on
the intrinsic magnetic moment39, M cations are in mixed 3+and
4+oxidation states (see Supplementary Figs. 29 and 30, and
Supplementary Table 5), which is consistent with the average
oxidation states in the range of 3.0–3.7 that have been reported in
the literature based on XAS measurements4,12,31,32,61,65.Itis
worth noting that, for the cases of incomplete phase transition, the
measured oxidation state is a weighted average of 2+,3+, and 4+.
For γ-NiFe LDH, consistent with a previous assignment based on
operando Mössbauer spectroscopy studies25,63,66, Fe cations are in
a4+oxidation state (see Supplementary Fig. 29 and Supplemen-
tary Table 5). We note that, in addition to Fe4+, higher Fe
oxidation states have already been reported in the literature67–69.
We will show in the reaction mechanism study below that the
flexible oxidation state of the Fe site, and its synergy with M sites,
are responsible for the high catalytic activity of MFe LDHs. Based
on the energetics (see Fig. 3), the formation probability of the γ-
phase that we screened, K
1/4
(H
2
O)
1/2
MO
2
, is over three orders of
magnitude higher than the γ-NiOOH analog K
1/3
(H
2
O)
2/3
MO
2
used in previous studies70. As the activity is sensitive to non-
covalent interactions induced by the bulk structure and the
electronic structure, in addition to the geometric structure and
electronic structure of the active site, the γ-MFe phase is used in
the study of the OER mechanism below. We note that, because of
the characteristic stoichiometry of the layer and atomic-scale
details of the intercalated species, the γphase cannot be obtained
by simply introducing various amounts of water molecules and
cations into the interlayer space of the β-MOOH analogs used in
the literature. Further, as we demonstrate below, correct
determination of the OER mechanism requires not only an
accurate treatment of the bulk catalyst structure, but also a
complete consideration of all key factors that have been missed in
previous models, including the geometry, oxidation states, and
adsorbate coverages on the catalyst surface.
The catalytic oxygen evolution reaction mechanism. Beginning
with the elucidated bulk structures described above, we evaluated
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-16237-1 ARTICLE
NATURE COMMUNICATIONS | (2020) 11:2522 | https://doi.org/10.1038/s41467-020-16237-1 | www.nature.com/naturecommunications 5
the steady state of the (01–10) surface of γ-NiOOH, γ-NiFe LDH,
and γ-CoFe LDH through surface phase diagrams, then calculated
the reaction free energy diagram of oxygen redox (4OH−+
*←→3OH−+OH*+e−
←→2OH−+O*+H
2
O+2e−
→OH−
+OOH*+H
2
O+3e−
→O
2
+2H
2
O+4e−+*) (see Fig. 4,Sup-
plementary Figs. 31–41, and Supplementary Tables 7–11). We
have focused on the reaction of the surface oxygen species and
neglected the potential involvement of lattice oxygen, because
recent isotope experiments suggest that the latter is not favorable
for these specific systems2. We selected the (01–10) surface,
because it belongs to the family of surfaces that are exposed at the
edge of catalyst sheets, and thus widely used to study the catalytic
activity of layered materials4,11,33,40,71.
The calculated surface phase diagrams indicate that, under
OER conditions, undercoordinated surface O sites are saturated
with H by forming bridge OH species, and undercoordinated
metal sites are saturated with atop OH when the surface is in
equilibrium with the electrolyte and in steady state (see Fig. 4).
Thus, we analyze the reaction free energy with a Mars van
Krevelen-type mechanism, for which the reactions start from the
deprotonation of the surface OH of the in-situ surface phase,
instead of starting from OH adsorption, as has generally been
assumed in many previous studies (to further motivate this
choice, see the comparison with the reaction free energies of the
conventional mechanism on two artificial surface models in the
Supplementary Materials). For the Mars van Krevelen mechan-
ism, we have found that the oxidation of two-metal coordinated
bridge OH moieties is more favorable than that of one-metal
coordinated atop OH due to the synergy of the two nearest-
neighbor metal sites in stabilizing the potential limiting OER
reaction intermediates (O*radials) by forming an O-bridged
reaction center (see Fig. 4and Supplementary Methods for
details). Thus, we have focused our discussion below on the
synergistic bridge OH oxidation pathway.
The highest reaction free energy barriers (ΔG
a
)onγ-NiOOH,
γ-CoFe, and γ-NiFe oxyhydroxide surfaces are 1.90 eV, 1.71 eV,
and 1.68 eV (see Fig. 4c and Supplementary Figs. 31, 33, and 38),
respectively, which implies overpotentials (η) of 0.67 V, 0.48 V,
and 0.45 V (η=(ΔG
a
−1.23 eV)/e). The calculated overpotentials
are semi-quantitatively consistent with the present measurements
at 10 mA cm−2, 0.57 V, 0.40 V, and 0.35 V, respectively, and with
general trends in the literature. For γ-NiOOH and γ-NiFe LDH,
OH*deprotonation during the OER cycle has the highest
free energy barrier, which forms the potential limiting step,
followed by OOH*deprotonation. On the γ-NiOOH surface,
bridge OH (Ni3+-OH-Ni4+) deprotonation at 1.90 V is accom-
panied by Ni3+oxidation to Ni4+, as characterized by the change
of Ni magnetic moment from 1 μ
B
to 0 μ
B
. On the other hand, on
0246810
–0.6
–0.5
–0.4
–0.3
–0.2
–0.1
0.0
dE (eV/Ni
8
O
16
K
2
(H
2
O)
4
)
AIMD (NVT) simulation time (ps)
1.0 1.2 1.4 1.6 1.8 2.0
–1.0
Δ
–0.8
–0.6
–0.4
–0.2
0.0
0.2
0.4
G(eV)
Potential (V vs RHE)
α-NiFe LDH
γ-NiFe LDH
23.0 Å
(7.7 Å/layer)
21.3 Å
(7.1 Å/layer)
3.1 Å 2.84 Å
α-NiFe LDH γ-NiFe LDH
cd e
HNi OKFe
b
a
0246810
–158
–157
–156
–155
–154
Total energy (eV)
AIMD simulation time (ps)
Heating
NVT simulation
Cooling
Structure optimization
Structure and cell optimization
0.25
0.20
0.15
0.10
NiOOH-(2 × 4)-nH2O
NiOOH-(2 × 4)CO3-nH2O
NiO2-(2 × 3)-2K-nH2O
NiO2-(2 × 4)-2K-nH2O
NiO2-(2 × 3)-1K-nH2O
NiO2-(2 × 4)-1K-nH2O
NiO2-(2 × 4)-nH2O
dG (eV/Ni)
0.05
0.00
–0.05 3.0 3.2 3.4
2H2O
8H2O
7H2O
6H2O
5H2O
4H2O
3H2O
3.6
Ni nominal oxidation state
3.8 4.0
Fig. 3 Screening process, structures, and stability (phase diagram) of NiFe LDH. a The relative energy of γ-NiOOH (Ni
8
O
16
K
2
·4H
2
O) at each picosecond
of the AIMD simulation, which is used to screen the most stable configuration (at the 5th ps) of this specific stoichiometry. The inset is the energy
evolution during the entire AIMD simulation. bFree energy of formation of a series of possible γ-NiOOH structures with various amounts of water and ions
intercalated between the NiOOH or NiO
2
layers. Each point is based on the most stable configuration of an AIMD simulation. For example, NiO
2
-(2 × 4)
−2K-4H
2
O is from the 5th ps simulation of Ni
8
O
16
K
2
·4H
2
O in A, which is then used to study the possible configuration of γ-NiFe LDH. cSide, top, and
bottom views of the α-NiFe LDH; dstability of α- and γ-NiFe LDH; eside, top, and bottom views of the γ-NiFe LDH. The structural parameters of α- and
γ-NiFe LDH are also given.
ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-16237-1
6NATURE COMMUNICATIONS | (2020) 11:2522 | https://doi.org/10.1038/s41467-020-16237-1 | www.nature.com/naturecommunications
the γ-NiFe LDH surface, bridge OH (Fe4+-OH-Ni3+) deprotona-
tion at 1.68 V is accompanied by Fe4+oxidation (as characterized
by the change of Fe magnetic moment, see Fig. 4), whereas the
oxidation state of Ni is constant. Clearly, it is more feasible for Fe
than Ni to be oxidized to a higher oxidation state, which stabilizes
O*intermediates at the Fe-Ni reaction center compared with that
at the Ni-Ni reaction center, and consequently lowers the
free energy barrier of OH*oxidation to O*, the potential limiting
step. Our calculations indicate that it is also the case for the other
γ-NiFe configurations with comparable energies that could co-
exist under the reaction conditions (see Supplementary Table 9).
It is worth noting that this stabilization effect is also valid and
even stronger on single Fe sites as compared with single Ni sites
(0.73 eV for the stabilization on single site vs. 0.22 eV for the
stabilization at the Fe-Ni center), as also observed in previous
studies (0.4–0.5 eV)33,43. However, there is a fundamental
difference between the synergistic stabilization through the Fe-
Ni reaction center and the stabilization through the single Fe site,
with the former being over five orders of magnitude more active
than the latter toward OER on NiFe LDH (see Supplementary
Figs. 33 and 34). Similar synergy of two nearest-neighbor metal
sites (reaction center) and flexibility of Fe site oxidation are also
found on γ-CoFe LDH, for which the stabilization effect is so
significant that bridge OH (Fe4+-OH-Co4+) deprotonation and
the accompanied Fe4+oxidation is not the potential limiting step
anymore. Instead, OOH*deprotonation to O
2
(g) +vacancy
(with a 1.7 eV free-energy barrier) becomes the potential limiting
step. On pure Co sites of γ-CoFe LDH, OOH*deprotonation is
also the potential limiting step but with higher overpotential (at
1.83 V) because of the more unfavorable O vacancy formation.
Therefore, in addition to the O*intermediate, the reaction center
also can stabilize O vacancies in CoFe LDH through synergy and
the flexible electronic structure of Fe. However, the stabilization
effect on the O vacancy formation does not seem large enough to
make γ-CoFe LDH more active than γ-NiFe LDH, whereas the
stabilization of O*by the introduction of Fe is beneficial in both
catalysts, resulting in a small difference in activity. As a
consequence, the overpotentials on NiFe and CoFe LDH are
only modestly (0.14–0.22 V) higher than the optimal over-
potential that is constrained by the scaling relationship (the
scaling relationship of OOH*intermediate and OH*intermedi-
ate, which is 2.95 eV in the present work, leads to an optimal
overpotential of 0.25 V)72. Such a constraint also implies that the
OER overpotential of LDHs can be modestly improved by further
stabilizing O*intermediates and surface O vacancies at the
reaction centers simultaneously, perhaps with a more redox-
flexible metal than Fe, or significantly improved by breaking the
OOH*vs. OH*scaling relationship.
Discussion
NiFe and CoFe LDHs are the archetypes of high-performing
electrocatalysts for oxygen evolution in alkaline conditions. In the
current work, we have identified the crystal structures of the
active phase and the reaction mechanism by combining operando
experiments, rigorous DFT calculations, and self-consistent
mechanistic studies. We have found that, under applied anodic
potentials, both NiFe and CoFe LDHs transform from the
0
1
2
3
4
5
2.95 eV
O
2
+ 2H
2
O + 4e
–
O* + H
2
O + 3OH
–
+ e
–
OOH* + H
2
O
+ 2OH
–
+ 2e
–
OH* + 4OH
–
γ
-NiOOH
γ
-NiFe LDH
γ
-CoFe LDH
U = 0V
RHE
Free energy (eV)
O
2
+ V + 2H
2
O + OH
–
+ 3e
–
1.90 eV
(
η
= 0.67
V)
1.68 eV
(
η
= 0.45 V)
1.71 eV
(
η
= 0.48 V)
0.6
0.5
0.4
0.3
1.0 1.2 1.4 1.6 1.8 2.0
0.0
0.4
0.8
1.2
1.6
Δ
G
O
-
Δ
G
OH
(eV)
ΔG
OH
(eV)
0.2
0.3
0.4
0.5
0.6
0.7
0.8
Overpotential (V)
NiFe
CoFe
Ni
cd
0.5 1.0 1.5 2.0 2.5
–0.4
–0.2
0.0
0.2
OHad
Had
ΔG(eV)
Potential (V vs RHE)
γ-NiOOH
γ-NiFe LDH
γ-CoFe LDH
OHad + Had
b
OHad + Had (OH*) OHad + Had (O*)
2.5
0.9
OHad + Had (OOH*)
2.0
0.9
2.0
0.9
OHad + Had (* + O2)
2.0
1.5
aHad OHad
“Pristine”
Fig. 4 OER mechanism on the γ-phase of MFe LDHs. a Structures of different surface phases and OER intermediates; adsorbates of surface phases are
highlighted by blue circles on the sides views, and OER intermediates are differentiated by colors (yellow instead of white for hydrogen and rose instead of
red for oxygen, respectively). A dashed rose circle indicates the formation of a surface O vacancy. The reaction centers are highlighted by large white
circles. The magnetic moments of Ni and Fe during OER are also given on the top views. bSurface phase diagram of of γ-NiOOH, γ-NiFe LDH, and γ-CoFe
LDH. The representative surface phases are given in a.cReaction free-energy diagrams for OER on γ-NiOOH, γ-NiFe LDH, and γ-CoFe LDH; the potential
limiting steps and the overpotentials are also given. dVolcano plot of the OER overpotential as a function of Gibbs free energies of the reaction
intermediates.
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-16237-1 ARTICLE
NATURE COMMUNICATIONS | (2020) 11:2522 | https://doi.org/10.1038/s41467-020-16237-1 | www.nature.com/naturecommunications 7
as-prepared α-phase to the active γ-phase. In comparison with
the as-prepared phase, with an interlayer distance of 7.7 Å and an
in-plane lattice constant of 3.1 Å, the catalytically active phases
are characterized by a compression of both lattice spacings to 7.1
Å and 2.8 Å, respectively. These values were extracted from
operando WAXS measurements and are also supported by DFT
calculations. Although the latter is induced by the oxidation of
both Fe(III) and M(II), the former is related to the swapping of
intercalated ions with K+, which is essential in identifying the
crystal structure of the active phases and cannot be accessed
experimentally with other local structure-based techniques. Thus,
the combination of DFT and operando WAXS confirms a long
speculated hypothesis regarding the crystal structure of NiFe
LDH under OER conditions and disprove previous assumptions
of the crystal structure of CoFe LDH, while, more importantly,
providing key atomic-scale details of the in-situ phases for the
study of the catalytic mechanism through DFT calculations. Our
calculations demonstrate that OER proceeds with a Mars van
Krevelen-type mechanism on these surfaces. The flexible elec-
tronic structure of the Fe sites and their synergy with the nearest-
neighbor M sites (M =Ni or Co) through forming O-bridged
Fe-M reaction centers stabilize OER intermediates that are
unfavorable on M-M centers and pure Fe sites. This synergistic
reaction center fundamentally accounts for the experimentally
observed low overpotentials of MFe for OER. The present study
suggests that doping oxides with additional redox-flexible metals
to form active reaction centers through the synergy with nearest-
neighbor metal sites constitutes a general design principle for the
synthesis of new OER catalysts design with improved catalytic
performance.
Methods
Synthesis. NiFe LDH (Ni : Fe =3.55:1) was synthesized by a previously reported
solvothermal route in an autoclave73. CoFe LDH (Co:Fe =3.33:1) was synthesized
by using co-precipitation followed by a solvothermal treatment in an autoclave. Ni
(OH)
2
was synthesized using a two-step synthesis consisting of a precipitation step
and a subsequent hydrothermal treatment. β-Co(OH)
2
was synthesized by a similar
process as that described by Ma et al.74 based on homogeneous precipitation.
Further details are available in the Supplementary Information.
RDE measurements and DEMS. RDE electrochemical experiments were per-
formed in a three-compartment glass cell with a rotating disk electrode (RDE,
5 mm in diameter of GC, Pine Instrument) and a potentiostat (Gamry) at room
temperature. A Pt-mesh and a Hydroflex reversible hydrogen electrode (RHE,
Gaskatel) were used as counter electrode and reference electrode, respectively. The
electrolytes were prepared with KOH pellets (semiconductor grade, 99.99% trace
metals basis, Aldrich) and MilliQ water, and were further purified75,76. The catalyst
was deposited on the GC by drop casting from an ink based on isopropanol/water
solution with Nafion as a binder. The catalyst loading was 0.1 mg cm−2. The
detailed protocol is provided in Supplementary Methods.
DEMS measurements were performed using dual thin-layer electrochemical
flow cell (see Supplementary Methods for details) with nitrogen-saturated
electrolyte 0.1 M KOH.
In-situ WAXS and Rietveld refinement. The electrodes used for in-situ WAXS
were prepared similarly as for the RDE measurements. A home-made grazing
incident cell (Supplementary Fig. 4) based on a thin-layer concept was used with a
polyether ether ketone foil covering the top part of the cell as X-Ray window77.
KOH (0.1 M) was used as electrolyte. In-situ WAXS experiments have been con-
ducted at the ID31 beamline of the European Synchrotron Radiation Facility
(Grenoble, France), using hard X-rays with a monochromatized beam (60–77 KeV).
The electrochemical protocol consisted in keeping the sample first at the potential of
1V
RHE
after electrolyte injection (wet condition), recording electrochemical impe-
dance spectroscopy, conducting an activation procedure by CV, and potential steps
of ~10 min for the regular measurements (40 min for collapsed film technique
explained in the in-situ WAXS section in the Supplementary Methods) from resting
state, before the M(II) oxidation (M =Ni or Co), to OER potentials and back in the
cathodic direction well below the reduction potential to M(II). The (003) and (110)
peaks were fitted by Pseudo-Voigt functions in the preliminary analysis, after
background subtraction. Rietveld refinement was performed on selected potentials.
The hydrotalcite structure with space group R-3m was used as a model for both the
LDH materials and for both the as-prepared and oxidized phases. For full details, see
the Supplementary Information.
Operando XAS.Operando XAS measurements were performed at the BL22
CLAESS beamline at ALBA light source (Barcelona, Spain) in fluorescence mode
using a silicon drift diode detector. A home-made electrochemical cell was
employed. A platinum mesh and leak-free Ag/AgCl electrode were used as counter
and reference electrode, respectively. The powder samples were deposited on
graphite paper discs (Toray Carbon Paper TP-060, Quintech) by filtration from a
slurry of the sample in ethanol containing Nafion (0.1 v/v %) as a binding agent.
The paper discs were mounted in the operando cell so that the unmodified side was
facing out, whereas the side containing the catalyst layer was in contact with the
electrolyte. The electrochemical conditions were identical to those described for in-
situ WAXS measurements.
DFT calculation parameters. Self-consistent, periodic DFT calculations were per-
formed with the projected augmented wave method, as implemented in the Vienna
Ab-initio Simulation Package. To generate highly accurate electrochemical stability
diagrams, we employ a recently developed approach55, which includes the use of a
Hubbard U term, a van der Waals functional (optPBE)78, and the use of a water-based
reference state for the calculations. U-values, which are applied to d-orbitals of Fe, Co,
and Ni are taken as 2.56, 3.50, and 5.20 eV, respectively. For cell shape and volume
relaxations of (hydroxy)oxide compounds, a cutoff energy of 500 eV is used for the
planewave expansion. For the calculations that do not involve cell optimization, a
cutoff energy of 400 eV is employed. Monkhorst–Pack k-point grids are used for
Brillouin zone integration. A (2 × 4 × 1) and a (2 × 4 × 3) k-point grid are employed for
α-andγ-phase of LDH with R3 and R1 symmetry, respectively. For the other bulk and
surface calculations, equivalent or denser k-point grids are utilized. An orthorhombic
box (14 × 15 × 16) Å3and a single k-point (0.25, 0.25, 0.25) for the Brillouin zone
sampling are used for gas phase species. The equilibrium geometries are obtained
when the maximum atomic forces are smaller than 0.01eV/Å and when a total energy
convergence of 10−5eV is achieved for the electronic self-consistent field loop. AIMD
simulations are performed at 400 K and quenched down to 0 K every 1 ps with a total
simulation time of 10 ps. To evaluate the solvation energy of OER intermediates (see
Supplementary Table 11), vacuum between the slab and the images is filled with liquid
water with a thickness that is equivalent to five water bilayers. Then AIMD simulations
are performed with the same protocols and time scales as that described above.
Data availability.
The data supporting the findings of this study are available within this Article and its
Supplementary Information files, or from the corresponding author upon reasonable
request. The Supplementary Information contains descriptions of methods, discussions
on physicochemical characterization of as-prepared MFe LDH, intermediate phases, size
of coherently scattering domains, operando XAS, ex-situ sXAS, and DFT calculation. It
also includes Supplementary Figs. 1–41 and Supplementary Tables 1–11.
Received: 30 March 2020; Accepted: 21 April 2020;
References
1. McCrory, C. C. L. et al. Benchmarking hydrogen evolving reaction and oxygen
evolving reaction electrocatalysts for solar water splitting devices. J. Am.
Chem. Soc. 137, 4347–4357 (2015).
2. Roy, C. et al. Impact of nanoparticle size and lattice oxygen on water oxidation
on NiFeOxHy. Nat. Catal. 1, 820–829 (2018).
3. Dresp, S. et al. Direct electrolytic splitting of seawater: activity, selectivity,
degradation, and recovery studied from the molecular catalyst structure to the
electrolyzer cell level. Adv. Energy Mater. 8, 1800338 (2018).
4. Friebel, D. et al. Identification of highly active Fe sites in (Ni,Fe)OOH
for electrocatalytic water splitting. J. Am. Chem. Soc. 137, 1305–1313 (2015).
5. Trotochaud, L., Young, S. L., Ranney, J. K. & Boettcher, S. W. Nickel–iron
oxyhydroxide oxygen-evolution electrocatalysts: the role of intentional and
incidental iron incorporation. J. Am. Chem. Soc. 136, 6744–6753 (2014).
6. Gong, M. et al. An advanced Ni-Fe layered double hydroxide electrocatalyst
for water oxidation. J. Am. Chem. Soc. 135, 8452–8455 (2013).
7. Burke, M. S. et al. Revised oxygen evolution reaction activity trends for first-
row transition-metal (Oxy)hydroxides in alkaline media. J. Phys. Chem. Lett.
6, 3737–3742 (2015).
8. Dionigi, F. & Strasser, P. NiFe-based (oxy)hydroxide catalysts for oxygen
evolution reaction in non-acidic electrolytes. Adv. Energy Mater. 6, 1600621
(2016).
9. Hoang, T. T. H. & Gewirth, A. A. High activity oxygen evolution reaction
catalysts from additive-controlled electrodeposited Ni and NiFe films. ACS
Catal. 6, 1159–1164 (2016).
ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-16237-1
8NATURE COMMUNICATIONS | (2020) 11:2522 | https://doi.org/10.1038/s41467-020-16237-1 | www.nature.com/naturecommunications
10. Stevens, M. B. et al. Ternary Ni-Co-Fe oxyhydroxide oxygen evolution
catalysts: Intrinsic activity trends, electrical conductivity, and electronic band
structure. Nano Res. 12, 2288–2295 (2019).
11. Zhang, B. et al. Homogeneously dispersed multimetal oxygen-evolving
catalysts. Science 352, 333–337 (2016).
12. Enman, L. J. et al. Operando X-ray absorption spectroscopy shows iron
oxidation is concurrent with oxygen evolution in cobalt–iron (oxy)hydroxide
electrocatalysts. Angew. Chem. Int. Ed. 57, 12840–12844 (2018).
13. Risch, M. et al. Structural changes of cobalt-based perovskites upon water
oxidation investigated by EXAFS. J. Phys. Chem. C. 117, 8628–8635 (2013).
14. Fabbri, E. et al. Dynamic surface self-reconstruction is the key of highly active
perovskite nano-electrocatalysts for water splitting. Nat. Mater. 16, 925
(2017).
15. Calvillo, L. et al. Insights into the durability of Co-Fe spinel oxygen evolution
electrocatalysts via operando studies of the catalyst structure. J. Mater. Chem.
A6, 7034–7041 (2018).
16. Zhang, B. W., Lui, Y. H., Zhou, L., Tang, X. H. & Hu, S. An alkaline electro-
activated Fe-Ni phosphide nanoparticle-stack array for high-performance
oxygen evolution under alkaline and neutral conditions. J. Mater. Chem. A 5,
13329–13335 (2017).
17. Kun, X. et al. Dual electrical‐behavior regulation on electrocatalysts realizing
enhanced electrochemical water oxidation. Adv. Mater. 28, 3326–3332 (2016).
18. Yongqi, Z. et al. Rapid synthesis of cobalt nitride nanowires: highly efficient
and low‐cost catalysts for oxygen evolution. Angew. Chem. Int. Ed. 55,
8670–8674 (2016).
19. Yu, C. et al. An electrocatalyst with anti-oxidized capability for overall water
splitting. Nano Res. 11, 3411–3418 (2018).
20. Zhang, J.-Y. et al. Rational design of cobalt–iron selenides for highly efficient
electrochemical water oxidation. ACS Appl. Mater. Interfaces 9, 33833–33840
(2017).
21. Fabbri, E. et al. Dynamic surface self-reconstruction is the key of highly active
perovskite nano-electrocatalysts for water splitting. Nat. Mater. 16, 925–92
(2017).
22. Zhuang, Z., Sheng, W. & Yan, Y. Synthesis of monodispere Au@Co3O4 core-
shell nanocrystals and their enhanced catalytic activity for oxygen evolution
reaction. Adv. Mater. 26, 3950–3955 (2014).
23. Ma, R., Liang, J., Liu, X. & Sasaki, T. General insights into structural evolution
of layered double hydroxide: underlying aspects in topochemical
transformation from brucite to layered double hydroxide. J. Am. Chem. Soc.
134, 19915–19921 (2012).
24. Demourguesguerlou, L., Braconnier, J. J. & Delmas, C. Iron-substituted nickel
oxyhydroxides and hydroxides obtained by Chimie-Douce. J. Solid State
Chem. 104, 359–367 (1993).
25. Demourguesguerlou, L., Fournes, L. & Delmas, C. On the iron oxidation-state
in the iron-substituted gamma-nickel oxyhydroxides. J. Solid State Chem. 114,
6–14 (1995).
26. Batchellor, A. S., Kwon, G., Laskowski, F. A. L., Tiede, D. M. & Boettcher, S.
W. Domain structures of Ni and NiFe (oxy)hydroxide oxygen-evolution
catalysts from X-ray pair distribution function analysis. J. Phys. Chem. C. 121,
25421–25429 (2017).
27. Hunter, B. M., Hieringer, W., Winkler, J. R., Gray, H. B. & Muller, A. M. Effect
of interlayer anions on [NiFe]-LDH nanosheet water oxidation activity.
Energy Environ. Sci. 9, 1734–1743 (2016).
28. Doyle, R. L., Godwin, I. J., Brandon, M. P. & Lyons, M. E. G. Redox and
electrochemical water splitting catalytic properties of hydrated metal oxide
modified electrodes. Phys. Chem. Chem. Phys. 15, 13737–13783 (2013).
29. Görlin, M. et al. Oxygen evolution reaction dynamics, Faradaic charge
efficiency, and the active metal redox states of Ni–Fe oxide water splitting
electrocatalysts. J. Am. Chem. Soc. 138, 5603–5614 (2016).
30. Bates, M. K., Jia, Q., Doan, H., Liang, W. & Mukerjee, S. Charge-transfer
effects in Ni–Fe and Ni–Fe–Co mixed-metal oxides for the alkaline oxygen
evolution reaction. ACS Catal. 6, 155–161 (2016).
31. Gorlin, M. et al. Tracking catalyst redox states and reaction dynamics in Ni-Fe
oxyhydroxide oxygen evolution reaction electrocatalysts: the role of catalyst
support and electrolyte pH. J. Am. Chem. Soc. 139, 2070–2082 (2017).
32. Smith, R. D. L. et al. Spectroscopic identification of active sites for the oxygen
evolution reaction on iron-cobalt oxides. Nat. Commun. 8, 2022 (2017).
33. Xiao, H., Shin, H. & Goddard, W. A. Synergy between Fe and Ni in the
optimal performance of (Ni,Fe)OOH catalysts for the oxygen evolution
reaction. Proc. Natl Acad. Sci. USA 115, 5872–5877 (2018).
34. Bajdich, M., Garcia-Mota, M., Vojvodic, A., Norskov, J. K. & Bell, A. T.
Theoretical investigation of the activity of cobalt oxides for the electrochemical
oxidation of water. J. Am. Chem. Soc. 135, 13521–13530 (2013).
35. Zaffran, J. & Toroker, M. C. Understanding the oxygen evolution reaction on
a two-dimensional NiO
2
catalyst. ChemElectroChem 4, 2764–2770 (2017).
36. Nagli, M. & Caspary Toroker, M. Communication: nickel hydroxide as an
exceptional deviation from the quantum size effect. J. Chem. Phys. 149, 141103
(2018).
37. Zhang, J. et al. Single-atom Au/NiFe layered double hydroxide electrocatalyst:
probing the origin of activity for oxygen evolution reaction. J. Am. Chem. Soc.
140, 3876–3879 (2018).
38. Diaz-Morales, O., Ledezma-Yanez, I., Koper, M. T. M. & Calle-Vallejo, F.
Guidelines for the rational design of Ni-based double hydroxide
electrocatalysts for the oxygen evolution reaction. Acs Catal. 5, 5380–5387
(2015).
39. Goldsmith, Z. K. et al. Characterization of NiFe oxyhydroxide electrocatalysts
by integrated electronic structure calculations and spectroelectrochemistry.
Proc. Natl Acad. Sci. USA 114, 3050–3055 (2017).
40. Li, Y. F. & Selloni, A. Mechanism and activity of water oxidation on selected
surfaces of pure and Fe-doped NiOx. Acs Catal. 4, 1148–1153 (2014).
41. Tripkovic, V., Hansen, H. A. & Vegge, T. From 3D to 2D Co and Ni
oxyhydroxide catalysts: elucidation of the active site and influence of doping
on the oxygen evolution activity. ACS Catal. 7, 8558–8571 (2017).
42. Martirez, J. M. P. & Carter, E. A. Unraveling oxygen evolution on iron-doped
beta-nickel oxyhydroxide: the key role of highly active molecular-like sites. J.
Am. Chem. Soc. 141, 693–705 (2019).
43. Shin, H., Xiao, H. & Goddard, W. A. In silico discovery of new dopants for Fe-
doped Ni oxyhydroxide (Ni1–xFexOOH) catalysts for oxygen evolution
reaction. J. Am. Chem. Soc. 140, 6745–6748 (2018).
44. Zaffran, J. et al. Influence of electrolyte cations on Ni(Fe)OOH catalyzed
oxygen evolution reaction. Chem. Mater. 29, 4761–4767 (2017).
45. Baker, J. G. et al. The role of aluminum in promoting Ni-Fe-OOH
electrocatalysts for the oxygen evolution reaction. ACS Appl Energ. Mater. 2,
3488–3499 (2019).
46. Liu, J.-X., Su, H.-Y., Sun, D.-P., Zhang, B.-Y. & Li, W.-X. Crystallographic
dependence of CO activation on cobalt catalysts: HCP versus FCC. J. Am.
Chem. Soc. 135, 16284–16287 (2013).
47. Li, H., Li, Y., Koper, M. T. M. & Calle-Vallejo, F. Bond-making and breaking
between carbon, nitrogen, and oxygen in electrocatalysis. J. Am. Chem. Soc.
136, 15694–15701 (2014).
48. Hammer, B. & Nørskov, J. K. in Advances in Catalysis 45, 71–129 (Academic
Press, Inc., 2000).
49. Lee, Y.-L., Kleis, J., Rossmeisl, J., Shao-Horn, Y. & Morgan, D. Prediction of
solid oxide fuel cell cathode activity with first-principles descriptors. Energy
Environ. Sci. 4, 3966–3970 (2011).
50. Strmcnik, D. et al. The role of non-covalent interactions in electrocatalytic
fuel-cell reactions on platinum. Nat. Chem. 1, 466–472 (2009).
51. Li, H., Xiao, J., Fu, Q. & Bao, X. Confined catalysis under two-dimensional
materials. Proc. Natl Acad. Sci. USA 114, 5930–5934 (2017).
52. Reuter, K., Frenkel, D. & Scheffler, M. The steady state of heterogeneous
catalysis, studied by first-principles statistical mechanics. Phys. Rev. Lett. 93,
116105 (2004).
53. Wang, S., Vorotnikov, V. & Vlachos, D. G. Coverage-induced conformational
effects on activity and selectivity: hydrogenation and decarbonylation of
furfural on Pd(111). ACS Catal. 5, 104–112 (2015).
54. Hensley, A. J. R. et al. DFT-based method for more accurate adsorption
energies: an adaptive sum of energies from RPBE and vdW density
functionals. J. Phys. Chem. C. 121, 4937–4945 (2017).
55. Zeng, Z. et al. Towards first principles-based prediction of highly accurate
electrochemical Pourbaix diagrams. J. Phys. Chem. C. 119, 18177–18187
(2015).
56. Zaffran, J. & Toroker, M. C. Benchmarking density functional theory based
methods to model NiOOH material properties: Hubbard and van der Waals
corrections vs hybrid functionals. J. Chem. Theory Comput. 12, 3807–3812
(2016).
57. Wei, C. & Xu, Z. J. The comprehensive understanding of as an evaluation
parameter for electrochemical water splitting. Small Methods 2, 1800168
(2018).
58. Bode, H., Dehmelt, K. & Witte, J. Zur kenntnis der nickelhydroxidelektrode—
I.Über das nickel (II)-hydroxidhydrat. Electrochim. Acta 11, 1079–1087
(1966).
59. Oskar, G. & Josef, E. Die Struktur höherer Nickelhydroxyde. Z. f.ür.
anorganische Chem. 261,43–51 (1950).
60. Delaplane, R. G., Ibers, J. A., Ferraro, J. R. & Rush, J. J. Diffraction and
spectroscopic studies of the cobaltic acid system HCoC
2
-DCoO
2
.J. Chem.
Phys. 50, 1920–1927 (1969).
61. Wang, D. et al. In situ X-ray absorption near-edge structure study of advanced
NiFe(OH)x electrocatalyst on carbon paper for water oxidation. J. Phys. Chem.
C. 119, 19573–19583 (2015).
62. Gonzalez-Flores, D. et al. Nickel-iron catalysts for electrochemical water
oxidation - redox synergism investigated by in situ X-ray spectroscopy
with millisecond time resolution. Sustain. Energ. Fuels 2, 1986–1994
(2018).
63. Chen, J. Y. C. et al. Operando analysis of NiFe and Fe oxyhydroxide
electrocatalysts for water oxidation: detection of Fe4+by Mössbauer
spectroscopy. J. Am. Chem. Soc. 137, 15090–15093 (2015).
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-16237-1 ARTICLE
NATURE COMMUNICATIONS | (2020) 11:2522 | https://doi.org/10.1038/s41467-020-16237-1 | www.nature.com/naturecommunications 9
64. Louie, M. W. & Bell, A. T. An investigation of thin-film Ni–Fe oxide catalysts
for the electrochemical evolution of oxygen. J. Am. Chem. Soc. 135,
12329–12337 (2013).
65. Risch, M. et al. Water oxidation by amorphous cobalt-based oxides: in situ
tracking of redox transitions and mode of catalysis. Energy Environ. Sci. 8,
661–674 (2015).
66. Corrigan, D. A. The catalysis of the oxygen evolution reaction by iron
impurities in thin film nickel oxide electrodes. J. Electrochem. Soc. 134,
377–384 (1987).
67. Berry, J. F. et al. An octahedral coordination complex of iron(VI). Science 312,
1937–1941 (2006).
68. Riedel, S. & Kaupp, M. The highest oxidation states of the transition metal
elements. Coord. Chem. Rev. 253, 606–624 (2009).
69. Hunter, B. M. et al. Trapping an iron(VI) water-splitting intermediate in
nonaqueous media. Joule 2, 747–763 (2018).
70. Van der Ven, A., Morgan, D., Meng, Y. S. & Ceder, G. Phase stability of
nickel hydroxides and oxyhydroxides. J. Electrochem. Soc. 153, A210–A215
(2006).
71. Wang, H. et al. Transition-metal doped edge sites in vertically aligned MoS2
catalysts for enhanced hydrogen evolution. Nano Res. 8, 566–575 (2015).
72. Man, I. C. et al. Universality in oxygen evolution electrocatalysis on oxide
surfaces. ChemCatChem 3, 1159–1165 (2011).
73. Dionigi, F., Reier, T., Pawolek, Z., Gliech, M. & Strasser, P. Design criteria,
operating conditions, and nickel-iron hydroxide catalyst materials for selective
seawater electrolysis. Chemsuschem 9, 962–972 (2016).
74. Ma, R. et al. Topochemical synthesis of monometallic (Co2+-Co3+) layered
double hydroxide and its exfoliation into positively charged Co(OH)2
nanosheets. Angew. Chem. Int. Ed. Engl. 47,86–89 (2008).
75. Trotochaud, L., Young, S. L., Ranney, J. K. & Boettcher, S. W. Nickel-iron
oxyhydroxide oxygen-evolution electrocatalysts: the role of intentional and
incidental iron incorporation. J. Am. Chem. Soc. 136, 6744–6753 (2014).
76. Burke, M. S., Kast, M. G., Trotochaud, L., Smith, A. M. & Boettcher, S. W.
Cobalt-iron (oxy)hydroxide oxygen evolution electrocatalysts: the role of
structure and composition on activity, stability, and mechanism. J. Am. Chem.
Soc. 137, 3638–3648 (2015).
77. Bergmann, A. et al. Reversible amorphization and the catalytically active state
of crystalline Co3O4 during oxygen evolution. Nat Commun 6, 8625 (2015).
78. Klimes, J., Bowler, D. R. & Michaelides, A. Van der Waals density functionals
applied to solids. Phys. Rev. B 83, 195131 (2011).
Acknowledgements
The WAXS experiments were performed on beamline ID31 at the European Synchrotron
Radiation Facility (ESRF), Grenoble, France. We thank ESRF and HZB Bessy II for
allocation of synchrotron radiation beamtime, and E. Hornberger and H. Schmies for
their help during beamtimes. We also thank Dr. M. Görlin for the scientific and helpful
discussions on NiFe (oxy)hydroxides. ZELMI of Technical University Berlin is
acknowledged for their support with TEM measurements. Help at the beamline from
Lukas Pielsticker (RUB) is appreciated, as well as the technical support from Dr Carlo
Marini and Dr Nitya Ramanan at CLAESS beamline of ALBA synchrotron during
the operando XAS measurements. The operando XAS work has funded by the European
Research Council under grant ERC-OPERANDOCAT (ERC-725915). S.K. acknowledges
funding from the IMPRS SurMat. This work was partially supported by the German
Research Foundation (DFG, Deutsche Forschungsgemeinschaf) in the frame of the col-
laborative research center/transregio TRR247 Heterogeneous Oxidation Catalysis in the
liquid Phase, project no. 388390466 and through grant reference number STR 596/8-1,
Bifunctional seawater electrolyzer, STR 596/12-1, catalyst-support interactions on the
activity and stability of water splitting catalysts, and by the Federal Ministry for economic
affairs and energy (Bundesministerium für Wirtschaft und Energie, BMWi) under grant
number 03EIV041F, MethFuel/ MethQuest. Work at Purdue was supported through the
Office of Science, Office of Basic Energy Sciences, Chemical, Biological, and Geosciences
Division under DE-SC0010379 (J.G.). Work at University of Science and Technology of
China was supported by the National Key R&D Program of China (2018YFA0208603)
and the Frontier Science Key Project of the Chinese Academy of Sciences (QYZDJ-
SSW-SLH054). F.D. and P.S. acknowledge partial funding by the Deutsche For-
schungsgemeinschaft (DFG, German Research Foundation) under Germany´s Excellence
Strategy –EXC 2008/1 –390540038.
Author contributions
F.D. conceived and performed the operando WAXS measurements at ESRF and the
analysis of the operando WAXS data including Rietveld refinement. Z.Z. conceived DFT
calculations. Z.Z. and D.F. performed DFT calculations of structure search. Z.Z., S.D. and
J.Z. performed the DFT calculations of OER mechanism. I.S. performed the operando
XAS experiments at ALBA, analyzed the corresponding data, and wrote part of the
manuscript. T.M. and H.S. synthesized all the samples and performed the RDE elec-
trochemical characterization. M.B.L., S.K. and I.Z. performed the operando XAS
experiments at ALBA. T.M., A.B. and J.D. performed the operando WAXS measurements
at ESRF. J.F.d.A. designed and performed the DEMS experiments. M.G. performed the
TEM. D.T. designed and performed the sXAS measurements at BESSY II and wrote part
of the manuscript. F.D., B.R.C., and P.S. designed the research and experiments and
wrote parts of the manuscript. Z.Z., W.-X.L. and J.G. designed the research and DFT
calculations, and wrote part of the manuscript. All authors discussed the results and
assisted during manuscript preparation.
Competing interests
The authors declare no competing interests.
Additional information
Supplementary information is available for this paper at https://doi.org/10.1038/s41467-
020-16237-1.
Correspondence and requests for materials should be addressed to F.D., Z.Z., B.R.C. or
P.S.
Peer review information Nautre Communications thanks the anonymous reviewers for
their contributions to the peer review of this work.
Reprints and permission information is available at http://www.nature.com/reprints
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons
Attribution 4.0 International License, which permits use, sharing,
adaptation, distribution and reproduction in any medium or format, as long as you give
appropriate credit to the original author(s) and the source, provide a link to the Creative
Commons license, and indicate if changes were made. The images or other third party
material in this article are included in the article’s Creative Commons license, unless
indicated otherwise in a credit line to the material. If material is not included in the
article’s Creative Commons license and your intended use is not permitted by statutory
regulation or exceeds the permitted use, you will need to obtain permission directly from
the copyright holder. To view a copy of this license, visit http://creativecommons.org/
licenses/by/4.0/.
© The Author(s) 2020
ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-16237-1
10 NATURE COMMUNICATIONS | (2020) 11:2522 | https://doi.org/10.1038/s41467-020-16237-1 | www.nature.com/naturecommunications
