Paper Trail

cmd = pdftohtml -s -i apache/papers/15/paper15.pdf

ret = Page-1
Page-2
Page-3
Page-4
Page-5
Page-6
Page-7
Page-8
Page-9
Page-10
Page-11

numPages = 23

pageNum = 1

Design Criteria, Operating Conditions, and Nickel–Iron
Hydroxide Catalyst Materials for Selective Seawater
Electrolysis Fabio Dionigi, Tobias Reier, Zarina Pawolek, Manuel Gliech, and Peter Strasser*[a]
1. Introduction
Water splitting by electrolyzers or photoelectrochemical devi- ces provides a sustainable route to efficiently convert and store energy that comes from intermittent renewable resour- ces. The electrolysis of water produces hydrogen and oxygen gas (2H2O!2H2+O2) that can recombine in a fuel cell releas-
ing most of the stored energy and clean water as the only by- product.
Many electrocatalytic systems were investigated over the
past decades, mainly operating with electrolytes consisting of high purity distilled water to which acids, bases, or buffer sys- tems were added.[1] Only a few studies report investigations concerning the use of seawater in electrochemical[2] and pho- toelectrochemical[3] water splitting devices. The direct use of seawater instead of fresh water or distilled water offers tre- mendous advantages for implementation of electrolyzers and solar driven photoelectrochemical devices in areas where fresh water is scarcely available or the use of fresh water by the elec- trolyzer will constitute a competing drain of water from the local fresh water reserve. Seawater is available in sufficient quantities on the planet (~ 97% of the total water) and has a fairly homogeneous geographic distribution, clearly reducing
competition with the use of fresh water by other human activi- ties. Furthermore, arid zones will benefit from the combination of a seawater electrolyzer with a fuel cell, as this technology will not only provide a way to store energy in chemical fuel, but will also produce fresh drinking water from seawater.
Electrochemical water splitting is an energetically uphill pro-
cess involving the hydrogen evolution reaction (HER) at the cathode and the oxygen evolution reaction (OER) at the anode. High activity and faradaic selectivity is particularly de- manding for OER:
4 OH¢ ! 2 H2O þ O2 þ 4 e¢; E0 ¼ þ1:23 VSHE or E0 ¼ þ1:23 V¢0:059 pH
because of its extremely poor kinetics originating from the fact that OER is a multi-electron reaction (four electrons per oxygen molecule), requiring the removal of four protons and involving more than one intermediate.[4] As a result, an energy barrier is associated with the formation of every intermediate. The design of a single catalyst that minimizes all of these barri- ers is not an easy task.
Dealing with seawater in water splitting devices is challeng-
ing owing to the variety of dissolved ions that can affect the catalytic system. Their average molar concentration is ~0.599m, corresponding to an average global salinity of ~3.5%. Dissolved ions in the electrolyte may poison or acceler- ate degradation of the water splitting catalysts through the formation of soluble complexes at both cathode and anode. Even more compromising to the operation of a seawater elec-
Seawater is an abundant water resource on our planet and its direct electrolysis has the advantage that it would not com- pete with activities demanding fresh water. Oxygen selectivity is challenging when performing seawater electrolysis owing to competing chloride oxidation reactions. In this work we pro- pose a design criterion based on thermodynamic and kinetic considerations that identifies alkaline conditions as preferable to obtain high selectivity for the oxygen evolution reaction. The criterion states that catalysts sustaining the desired operat- ing current with an overpotential sess the best chance to achieve 100% oxygen/hydrogen selec- tivity. NiFe layered double hydroxide is shown to satisfy this criterion at pH 13 in seawater-mimicking electrolyte. The cata-
lyst was synthesized by a solvothermal method and the activi- ty, surface redox chemistry, and stability were tested electro- chemically in alkaline and near-neutral conditions (borate buffer at pH 9.2) and under both fresh seawater conditions. The Tafel slope at low current densities is not influenced by pH or presence of chloride. On the other hand, the addition of chloride ions has an influence in the temporal evolution of the nickel reduction peak and on both the activity and stability at high current densities at pH 9.2. Faradaic efficiency close to 100% under the operating conditions predicted by our design criteria was proven using in situ electrochemical mass spec- trometry.
[a] Dr. F. Dionigi, Dr. T. Reier, Z. Pawolek, M. Gliech, Prof.Dr. P. Strasser
The Electrochemical Energy, Catalysis, and Materials Science Laboratory
Department of Chemistry, Chemical Engineering Division
Technical University Berlin
10623 Berlin (Germany)
E-mail: pstrasser@tu-berlin.de
Supporting Information for this article can be found under http://
dx.doi.org/10.1002/cssc.201501581.
ChemSusChem 2016, 9, 962 – 972
Ó 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
962
Full Papers
DOI: 10.1002/cssc.201501581

pageNum = 2

trolyzer is the fact that anions, such as, chloride, give rise to undesired competing electrochemical reactions at the electro- lyzer anode, liberating undesired side products, such as, for in- stance, molecular chlorine or chlorinated oxidants.
However, despite the chlorine evolution reaction (ClER) com-
petition with OER in seawater electrolysis, it is worthy to men- tion that both anodic electrode processes have no doubt a high importance in technologies aimed at chemical energy conversion or industrial synthesis of chemicals.[5] Indeed, chlor- ine is a valuable intermediate product in industry and is pro- duced at large scale by electrolysis of brine. In the case of ClER, OER is now the undesired reaction owing to its role in ac- celerating degradation of the catalyst layer. Therefore, the search for more selective and stable materials, as well as deeper fundamental understanding of the mechanism behind improvements in selectivity, are of great interest.[6] According to data provided by the World Chlorine Council (WCC)[7] and the Center for European Policy Studies (CEPS),[8] the world chlorine demand is attested at ~60 million metric tonnes per year (2012), confirming the importance of this reaction. How- ever, we believe that OER is favorable with respect to ClER when seawater is employed in processes aimed to store energy from intermittent renewable sources for the following reasons. The first problem arises from the fact that chlorine is a difficult product to handle and transport. For this reason, in industry, chlorine is essentially always generated on site and on time at the production factory where it is utilized as raw material. Indeed the high transportation costs limit the amount of chlorine transported to 5–6% in Europe (Eurochlor 2012 annual review).[9] The generation of H2 by intermittent re-
newable energy will then have to match the local chlorine demand and the electrolyzer must be located at a factory where chlorine is directly utilized, limiting the applicability in specific areas. Furthermore, the projected H2 demand in
a future global scale hydrogen economy will greatly overcome the chlorine demand by orders of magnitude, ultimately making oxygen the best co-product in hydrogen generation processes for renewable energy storage application.
Selective hydrogen production by water splitting without
liberating unmanageable amounts of poisonous chlorine gas, therefore, requires extremely selective OER catalysts or precise- ly tuned operating conditions for selective OER. Little work was done in this area and this gap in our knowledge of the sci- entific basis of selective seawater electrolysis is what the pres- ent contribution addresses.
Among the catalysts proposed for OER in alkaline environ-
ments, where economically attractive non-noble metal-based materials can be used, NiFe mixed oxides and hydroxides were shown to possess relatively low overpotential and high stabili- ty.[10] Recently these materials were also successfully employed on photoanodes as protection layers and surface modifica- tions.[11] For seawater applications these materials are unex- plored to date. Therefore, we focused our study on these ma- terials.
In this work, we propose and utilize a general design criteri-
on for oxygen-selective seawater oxidation electrocatalysis. The criterion specifies the maximum allowed OER overpoten-
tial as a function of pH that ensures selective seawater splitting under suppression of any chlorine redox electrochemistry. Fol- lowing our selectivity criterion, we show that the activity, sta- bility, and selectivity of NiFe layered double hydroxide (LDH) are not compromised by the presence of chloride ions, while operating inside the design criterion. In contrast, under elec- trolysis conditions outside the selectivity criterion, we experi- mentally verify the predicted competition of chlorine redox chemistry associated with severe catalyst degradation. Taken together, our study demonstrates the scientific feasibility of se- lective operating conditions for seawater electrolysis using NiFe LDH catalysts; then shows possible technologically-viable direct seawater electrolysis.
2. Experimental 2.1. Synthesis of NiFe layered double hydroxide
NiFe LDH was synthesized by solvothermal method. First, 79.6 mg of nickel(II) acetate tetrahydrate [Ni(C2H3O2)2·4H2O]
and 25.8 mg of iron(III) nitrate nonahydrate [Fe(NO3)3·9H2O]
were hydrolyzed in ~2.4 and 1.6 mL of Millipore water. The starting molar ratio of Ni/Fe is 5. Then the two solutions were added to a mixed solution of 30 mL of water and 16 mL of an- hydrous N,N-dimethylformamide (DMF) directly in the glass liner of a stainless steel autoclave (Roth, 100 mL/100 bar Model I). After 5 min of ultrasonication, the solvothermal reaction was performed at 1308C for 16 h followed by a second solvother- mal treatment at 1708C for 2 h. Magnetic stirring was em- ployed only for the first 30 min of the low temperature step. At the end of the synthesis, the autoclave was let cooled down naturally. The obtained suspension was ultrasonicated briefly and divided into two. One half was mixed with 44.2 mg of carbon Vulcan powder (Cabot XC-72R) and ultrasonicated. After aging overnight, the suspension was washed with a cen- trifuge for two times in ethanol/water mixture and two times with pure water (8500 rpm, 10 min). The samples were then freeze dried overnight.
2.2. Ink preparation
For electrochemical measurements a catalyst ink was prepared. 5 mg of supported catalyst was weighted in a glass vial. Then 500 mL of MilliQ water, 750 mL of isopropanol and 5 mL of Nafion solution (5 wt%) were added. The solution was ultraso- nicated with a 1/8 in microtip sonifier for 30 min. 5 mL of ink were drop casted on a previously polished and cleaned glassy carbon (GC) electrode (5 mm of diameter) and dried in an oven at 608C for 7 min. The catalyst loading, including carbon Vulcan support, is about 0.1 mgcm¢2.
The GC disks were polished manually with a 1.0 and 0.05 mm
micropolish alumina suspension for ~3 min each before each catalyst coating. After polishing, the disks were cleaned three times by ultrasonication in water, acetone, and water and final- ly dried with a nitrogen flow.
ChemSusChem 2016, 9, 962 – 972
www.chemsuschem.org
Ó 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
963
Full Papers

pageNum = 3

2.3. Materials characterization
TEM images were acquired by a FEI TECNAI G2 20 S-TWIN transmission electron microscope with LaB6 cathode. The mi-
croscope operated at an accelerating voltage of 200 kV. Energy-dispersive X-ray spectroscopy (EDX) data were taken by an EDX detector. The catalyst powders were suspended in eth- anol; then a carbon coated copper grid (400 mesh, Plano) was immersed in the solution and dried in an oven at 60 8C. Induc- tively-coupled plasma optical emission spectroscopy (ICP-OES) data were obtained by a 715-ES-ICP analysis system (Varian). The phase of the as-synthesized NiFe LDH nanoplates was ex- amined by XRD. XRD patterns were collected using a D8 Ad- vance-Diffractometer (Bruker) equipped with a Lynx Eye Detec- tor and using a CuKa source.
2.4. Electrochemical measurements
Electrochemical experiments were performed in a three-com- partment 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 hy- drogen electrode (RHE, Gaskatel) were used as counter elec- trode and reference electrode, respectively. The counter elec- trode was placed in a compartment that was separated by a fine-porosity glass frit from the working electrode compart- ment and a Luggin capillary was used for the reference elec- trode. A titanium shaft was used to prevent corrosion in case of chlorine evolution.
The electrolytes were prepared with KOH pellets (semicon-
ductor grade, 99.99% trace metals basis, Aldrich), H3BO3
(Emsure, ACS, ISO, Reag. Ph Eur, Merck), NaCl (99.5 +%, ACS, Chempur), and MilliQ water. The borate buffer was prepared with H3BO3 and KOH pellets that were added to reach the de-
sired pH of ~ 9.2. All electrochemical measurements were car- ried out in N2-saturated and rotation rate of 1600 rpm and re-
peated at least 3 times. All the current density (J) values report- ed are normalized by the geometric area (0.196 cm2). Internal resistance (iR) correction was applied after the measurements by using the value of resistance obtained during electrochemi- cal impedance spectroscopy (EIS). All the potentials reported are iR-corrected, unless otherwise stated. Averaged values are reported in the supporting online information.
Cyclic voltammetry
Cyclic voltammetry (CV) was conducted at the sweep rate of 50 mV s¢1. The CVs were performed by cycling 50 times the (not iR-corrected) between 1 and 1.9 V versus RHE for both the two chloride free electrolytes, between 1 and 1.75 V vs RHE for the chloride containing electrolyte at pH 13 and between 1 and 1.8 V vs RHE for the borate-buffered chloride-containing electrolyte at pH 9.2. The different potential range was chosen to have similar iR-corrected potential range for the chloride- free and chloride-containing electrolyte at each pH (from 1 to ~1.75 V vs. RHE for both the borate buffer electrolytes at
pH 9.2 and from 1 to ~1.65 V vs. RHE for the two electrolytes at pH 13).
Linear sweep voltammetry
After the CV, linear sweep voltammetry (LSV) measurements were conducted by sweeping the potential (not iR corrected) from 1.2 to 1.9 V versus RHE at a scan rate of 10 mVs¢1.
Stability test
Stability tests were conducted by performing chronopotenti- ometry (CP) at constant current of 1.96 mA (J=10 mA cm¢2) for 2 h. A pretreatment consisting of 5 cycles was performed before each CP test. A volume of electrolyte of ~50 mL was used in the stability tests that were aimed at detecting possi- bly dissolved metals in the electrolyte after the test. The meas- urements that were aimed at electrolyte titration were con- ducted in a small three-compartment glass cell containing ~40 mL total electrolyte to concentrate possibly produced oxi- dized chlorine species. In case of borate buffer+ NaCl, the ex- periment was interrupted after the catalyst film breakdown.
2.5. Hypochlorite titration analysis
Iodide titration was performed immediately after the stability test. 20 mL of electrolyte were pipetted from the working elec- trode compartment to an Erlenmeyer flask. Then 15 mL of freshly prepared 0.5m KI solution was added under magnetic stirring. In case a color change was observed, a 0.01m thiosul- fate solution was added dropwise using a burette. When the color of the solution became a faint yellow, 1 mL of starch so- lution was added, turning the solution blue. The thiosulfate ad- dition was interrupted when the solution became transparent. The amount of mol of oxidized chloride species is calculated by first obtaining the mol of reacted thiosulfate by multiplying the volume difference in the burette by the concentration of thiosulfate solution. Then this value is divided by the volume of electrolyte pipetted for the titration and multiplied by the total volume of the electrolyte. For every stability test the pro- cedure was averaged over two titrations. The corresponding charge is obtained by multiplying the amount of mol by the Faraday constant and by 2, assumed as the number of elec- trons per oxidized chloride species. Then this value is divided by the total charge passed during the stability test to obtain the percentage of charge associated with the formation of oxi- dized chloride species.
2.6. Selectivity measurements with quadrupole mass spec- trometer
A two-compartment glass cell with the compartments separat- ed by an anion-exchange membrane (Fumapem FAA-3-PK-130 from Fumatech) was used for the selectivity measurements with a RDE (10 mm in diameter of GC) and a potentiostat (Bio- logic) at room temperature. The area of the membrane was ~4.9 cm2. The rotor and shaft were specifically made and as-
ChemSusChem 2016, 9, 962 – 972
www.chemsuschem.org
Ó 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
964
Full Papers

pageNum = 4

sembled to resist chlorine corrosion and be gas tight. A Pt- mesh and a saturated calomel electrode (SCE) were used as counter electrode and reference electrode, respectively. A Luggin capillary was used for the reference electrode and the distance between the working and the counter electrode was roughly 8 cm. The working electrode compartment has a glass outlet in the top part for gas product detection and an aper- ture (gas inlet) connected with a glass tube through which N2
was bubbled in the solution. The SCE was calibrated versus RHE at pH 13 and the potentials reported in the paper were normalized with respect to RHE by adding 1.014 and 0.790 V for the potential at pH 13 and pH 9.2 respectively. The total metal loading on the working electrode was 7.92 mgcm¢2, ob- tained by drop casting 20 mL of ink prepared with a catalyst supported powder with 7.8% weight metal loading. The total catalyst loading, including carbon Vulcan support, is about 0.1 mgcm¢2. Therefore, the same metal loading per area and catalyst loading per area were used, as in the other electro- chemical experiments, despite the larger electrode area.
A quadrupole mass spectrometer (QMS, Thermostar from
PfeifferVacuum) was used to detect evolved chlorine and to determine the selectivity towards OER. A capillary connected with the QMS was inserted in a septum of the glass cell. The QMS was calibrated by a gas mixture of chlorine (52.1 5 ppm), oxygen (152 2 ppm), and N2 from a pre-mixed bottle
(Linde). The partial pressures of the gas in the pre-mixed bottle are given by the supplying company. chlorine (m/z= 70), oxygen (m/z=32), N2 (m/z=28), carbon dioxide (m/z=44)
and water (m/z=18) were monitored with the QMS.
The electrochemical measurements were performed under
constant N2-bubbling, with a gas flow of 500 NmLmin¢1 set by
a mass flow controller (MFC). CV was carried out as pretreat- ment (5 cycles) in all measurements. After the pretreatment, the stability was analyzed by conducting a sequence of CP measurements with constant current steps of 15 min each. The quasi-stationary conditions that were investigated allow to ne- glect differences in time constants for the different gases that otherwise would be important for a correct quantification of the products. For the electrolytes at pH 13, the current was in- cremented in the following step: 1, 3, 5, 7, 10, and 20 mA. For the electrolytes at pH 9.2 lower currents were set owing to in- stability: 1, 2, 3, 4, and 5 mA. All reported potentials are iR cor- rected.
2.8. Faradaic efficiency calculation
For the determination of OER, the faradaic efficiency the mo- lecular oxygen concentration (detected by QMS) was normal- ized by the faradaic oxygen equivalent concentration (100% faradaic efficiency).
The molecular oxygen concentration in ppm was calculated
by multiplying the averaged oxygen QMS ion current after background subtraction and N2 normalization with the calibra-
tion factor, obtained by dividing the provided measured value of 152 ppm of oxygen and the measured oxygen QMS current measured during calibration after background subtraction and N2 normalization. All the oxygen QMS ion currents measured
in the experiments and during calibration were normalized by the N2 ion current (N2 normalization), used as internal standard
to take into account concentration changes owing to possible fluctuations in the flow or dilution of the gas by water vapor from the electrolyte. For example, assuming saturated vapor pressure of water (26 mbar, 228C) the partial pressure of N2
will decrease by ~2.6% with respect to the dry mixture. In ad- dition, owing to different electrolyte concentrations in the four electrolytes investigated, the water vapor pressure changes within ~3%. These changes do not affect our results because of our N2 normalization.
For the faradaic oxygen equivalent concentration, the flow
of N2 was first calculated in mmols¢1 by using the set flow
value (500 NmLmin¢1) and the ideal gas law at standard condi- tions (1 atm, 273 K). A flow of 0.372 mmol s¢1 was obtained. Fi- nally, the electrochemical current expressed in mA was divided by 4F, where F is the Faraday constant and further divided by the nitrogen flow in mmol s¢1. The result was then expressed in ppm by multiplying by 106.
The error associated with the faradaic efficiency is obtained
by considering the error on the oxygen ppm value provided by the gas bottle provider (1.3 %), error on the MFC measure- ment (2%), error on the determination of the QMS current in the calibration experiment (0.1 %), and standard error of the mean for the averaging of the oxygen QMS signal during the experiments.
2.9. Turnover frequency calculation of the catalysts
The turnover frequency (TOF) value is calculated from the equation:
TOF ¼
J
4Fm*
m* ¼ Lð
Niwt%
Niu þ
Fewt%
Feu Þ
where J is the current density at an overpotential of 0.3 V esti- mated from the LSV, F is the Faraday constant and m is the mol of metal per cm2. m* is obtained by multiplying the cata- lyst loading L (0.1 mg cm¢2) by the sum of the weight percent- age of nickel (Niwt%) and iron (Fewt%) divided by their respective
atomic mass (Niu and Feu, respectively). The weight percentage
of nickel and iron are obtained from ICP measurement. We notice that this TOF is a lower estimation, since it considers all the metal atoms on the GC as active sites.
3. Results and Discussion 3.1. A general design criterion for selective seawater splitting
Chlorochemistry in aqueous environments comprises a complex ensemble of possible reactions that depend on pH and con- centration of chloride ions. Figure 1 displays a computed Pour- baix diagram of aqueous chlorochemistry for the conditions relevant to electrochemical OER at room temperature and total
ChemSusChem 2016, 9, 962 – 972
www.chemsuschem.org
Ó 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
965
Full Papers

pageNum = 5

mass of chlorine species fixed at 0.5m. A description of the construction of the Pourbaix diagram for chlorine species is available in Ref. [12]. In acidic solutions, the ClER[13]
2 Cl¢ ! Cl2 þ 2 e¢
E0 ¼ þ1:36 VSHE
ð1Þ
can occur and does compete with the OER. The Pourbaix dia- gram shows that OER is thermodynamically favored over ClER. However, ClER is a two-electron reaction that involves only a single intermediate. As a consequence, ClER has much faster kinetics than OER and is the dominant anodic reaction in acidic electrolytes on many oxide catalysts.[5,6c,14]
In alkaline conditions, hypochlorite formation must be con-
sidered:[13]
Cl¢ þ 2 OH¢ ! ClO¢ þ H2O þ 2 e¢; E0 ¼ þ0:89 VSHE,pH 14 or E0 ¼ þ1:72 V¢0:059 pH
ð2Þ
This reaction is also a two-electron reaction, so it has a kinet-
ic advantage over OER. However, thermodynamics highly favor OER over hypochlorite formation. Furthermore, the standard electrode potential of hypochlorite formation, unlike ClER, is pH dependent and it follows the OER potential in the Pourbaix diagram (Figure 1). Therefore, the electrode potential differ- ence to OER is fixed at ~0.480 V.
If the electrocatalyst is operating at an overpotential (h)
lower than this value, hypochlorite formation cannot occur and so OER does not compete with a chlorine redox reaction with faster kinetics. A similar argument could be formulated for acid and ClER, but the difference between ClER and OER potentials in acid is smaller, making it much more challenging to reach high currents at an electrode potential where ClER is not yet thermodynamically allowed. Therefore, alkaline condi- tions seem preferable for seawater oxidation. Furthermore, non-noble metal-based catalysts that would degrade in acidic
can be used in alkaline conditions. Based on these considera- tions, we establish a general design criterion [Eq. (3)] for OER/ oxygen selective operation of noble-metal-free electrocatalysts operating at pH>7.5 in seawater electrolyte as the difference in the standard potentials (DE0) between the OER and the chlororeactions, such as Equation (2):
hOER 480 mV
at pH > 7:5
ð3Þ
The lower pH limit was taken at 7.5, the pKa of the hypo-
chlorous acid, below which the hypochlorous acid formation becomes dominant respect to the hypochlorite ion, and the difference of the undesired side reaction potential respect to the OER potential becomes slightly smaller. The design criteri- on states not the only, but the most favorable conditions to achieve high selective oxygen evolution from seawater oxida- tion.
The requirement to operate at hOER480 mV and at J=
10 mA cm¢2, often indicated as technological target for com- mercial integrated devices based on solar-driven photoelectro- chemical water splitting where the photoabsorber and elec- trode areas are identical, or at a higher densities closer to the state-of-the-art of industrial alkaline electrolyzers, is demanding for noble-metal-free materials.
3.2. Synthesis and Structure NiFe-layered double hydroxide catalysts
NiFe LDH catalysts are known to reach such low overpotentials for OER in chloride-free alkaline electrolyte.[10c] A solvothermal method involving a mixture of water and DMF was used to synthesize NiFe LDH.[10c] The detailed protocol for the synthesis of the catalysts is presented in the Supporting Information. The synthesized NiFe LDH catalyst presents hexagonal nano- plate morphology, typical of well crystallized LDH materials (Figure 2a–b).[15] The NiFe LDH nanoplates are decorated with smaller amorphous FeOx particles. It is known that iron that is
not incorporated in the NiFe LDH can form FeOx or FeOOH
nanoparticles or domains, both when the NiFe LDH is synthe- sized by a solvothermal/hydrothermal synthesis as well as by electrodeposition method.[10c,16] All XRD reflections are as- signed to a hydrotalcite structure, typical of NiFe LDH (Fig- ure 2c).[15] The solvothermal synthesized NiFe LDH presents long crystalline order, with narrow and well defined reflections. The main diffraction peak at the 2q angle of 11.38, labeled with the Miller indexes (003), corresponds to diffraction from planes along the stacking direction. Therefore, the d-spacing, 7.8 æ in this case, is a measure of the distance between the LDH layers. This value is compatible with NiFe LDH with inter- calated carbonate anions.[17] The crystallite size obtained from the (003) peak width could be used as a gross estimate of the nanoplates thickness, under the exclusion of vertically stacked multicrystalline domains.[10c] In our case, the Scherrer equation estimates a crystalline size of 15 1 nm [full-width-at-half maxi- mum (FWHM) =0.588]. The difference with the 5 nm reported by Gong et al. could be attributed to the slightly higher tem- perature used in this work.[10c] A 3D atomic model of the ex-
Figure 1. Pourbaix diagram for artificial seawater model. A chlorine system,
in the case of dissolved 0.5 m NaCl aqueous solution and no other chlorine
sources, with a total chlorine species (cT,Cl) of 0.5m. The electrode potential
for OER is also included (assuming oxygen partial pressure of
0.21 atm=0.021 MPa). Two red square points show the operating potentials
(vs. SHE) after 1 h constant current electrolysis (10 mAcm¢2) with NiFe LDH
catalyst in 0.1m KOH+ 0.5m NaCl (pH 13) and 0.3m borate buffer+0.5 m
NaCl (pH 9.2) electrolyte. The light blue box highlights our proposed design
criterion.
ChemSusChem 2016, 9, 962 – 972
www.chemsuschem.org
Ó 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
966
Full Papers

pageNum = 6

tended unit cell of NiFe LDH in the fully protonated form and with formula [Ni2+1¢xFe3+x (OH)2]x+(CO32¢)x/2·y(H2O) is presented
in Figure 2d. Layer of edge sharing [NiO6]/[FeO6] octahedra
stacks along the c-axis with OH groups on both sides and water and charge balancing anions (i.e., carbonate anions) in- tercalated between the layers. ICP-OES and EDX were used to analyze the composition of the catalyst. ICP showed an aver- age sample composition of 73 at% nickel and 27 at% iron, whereas the more local EDX analysis reveals that areas with few FeOx particles have a slightly lower iron content of
~13 at%. Therefore, we consider the value of 13 at% iron a more accurate estimate of the real iron content of the NiFe LDH nanoplates.
3.3. Electrochemical seawater splitting: pH and Cl¢ effects
The electrochemical measurements were performed using a RDE in a three-electrode system and three-compartment glass cell. An ink was prepared with NiFe LDH catalyst support- ed on carbon Vulcan (Cabot) and drop-casted on a GC elec- trode. We tested the catalysts in alkaline (0.1m KOH, pH 13) and near-neutral electrolyte (0.3 m borate buffer corrected with KOH to reach pH 9.2, the pKa), both in chloride free conditions
and with 0.5m of NaCl addition. In the text, the following short notations will be used for indicating the four electrolytes: 1) KOH, 2) borate buffer, 3) KOH +NaCl, and 4) borate buffer+ NaCl. Borate buffer is added as a proton-accepting support in near neutral pH, to contrast changes in local pH.[2b,18] The group of Nocera, and more recently other groups, also investi- gated OER activity and stability of electrodeposited nickel oxide catalyst in potassium–borate solutions, showing long-
term stability under this mild pH condition.[19,20] To avoid a dra- matic decrease in local pH at the anode,[21] seawater cannot be used without buffer additives. Even though carbonate and borate ions are present in seawater, their average concentra- tion is too low to sustain proton handling at high currents. In the Supporting Information, this is shown by solving calcula- tions based on general mass transfer equation theory for a simple model system consisting of a flat plane rotating disk electrode under stationary conditions (Figure S1 and S2 in the Supporting Information).[18,22]
Therefore, the choice of the two pH values and the support-
ing buffer was based on these reasons and on the previous works reported in the literature (performed in the absence of NaCl). The choice of utilizing borate buffer at pH 9.2 was also supported by a previous screening that we performed with borate (0.1m, pH 9.2), phosphate (0.1 m, pH 7), and carbonate (0.1 m, pH 8.6 and 10) buffers, which showed that higher stabil- ity was obtained with borate buffer.[23]
Figure 3 compares the electrochemical OER performance in
fresh and seawater conditions and two buffered pH conditions. After a voltammetric “break-in” treatment (50 cycles, see Fig- ure S3), LSV was recorded at a lower scan rate to evaluate the electrochemical activity (Figure 3a). All the anodic LSVs show an anodic wave, attributed to nickel(II/III) redox reaction, char- acteristic of nickel hydroxide systems. The change in oxidation state of nickel is associated with a loss of proton:
NiðOHÞ2 þ OH¢ ! NiOOH þ H2O þ e¢
ð4Þ
The molecular mechanism associated with the anodic wave
is likely more complex than the simple deprotonation shown in Equation (4) and co-involves exchange of ions and water be- tween the metal oxide layers.[24] At more anodic potentials an increase in J is observed and attributed to OER.
Figure 3 evidences a pronounced effect of the electrolyte pH
on the electrochemical water splitting performance of the cat- alysts. Both the redox wave and the OER occur at more posi- tive potentials (vs. RHE) by decreasing the pH. The OER over- potential increases ~110 mV (vs. RHE) at 1 mAcm¢2. This shift corresponds to a shift of ~ 87 mV per pH on SHE scale.
The shift of the oxidation peak, on the other hand, is
~85 mV per pH step in the SHE scale. Therefore both the oxi- dation wave and the OER show a similar super-Nernstian po- tential pH shift. A shift of 88 mV per pH was observed for the redox peaks assigned to the hydrous a-Ni(OH)2 phase, in con-
trast to a Nernstian shift of the redox peaks in case of anhy- drous b-Ni(OH)2.[25] This is consistent with the model of the
NiFe LDH as hydrous hydroxide with water intercalated be- tween the layers. A detailed analysis of the LDH redox chemis- try during break-in CVs is presented in Figure S3. Similar to what was observed in the LSV for the anodic wave, we ob- served a shift (in the RHE scale) towards higher potentials for both the anodic and cathodic wave in the lower pH cases. Second, the separation in potential between the redox peak maximum (for anodic) and minimum (for cathodic) increases at lower pH. This separation was evaluated for the 50th cycle in the RHE scale to be 119 9 mV for pH 13 and 18310 mV for
Figure 2. a) TEM image of hexagonal NiFe LDH nanoplates and smaller FeOx
particles. b) TEM image of a single hexagonal NiFe LDH nanoplate. c) XRD
pattern of NiFe LDH with insert showing the higher 2q angle range. d) 3D
structure model of the as prepared NiFe LDH with intercalated water and
carbonate ions.
ChemSusChem 2016, 9, 962 – 972
www.chemsuschem.org
Ó 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
967
Full Papers

pageNum = 7

pH 9.2. These observations seem to point to a less reversibility of the nickel redox couple at lower pH. And finally, the data show that the area of the peaks in pH 13 is larger than at pH 9.2, consistent with reduced nickel accessibility.
We now turn to the discussion of the chloride-containing
seawater conditions. Our data evidence a striking effect of the added NaCl on the temporal evolution of the metal redox peaks during potential “break-in” cycling (Figure S3a–b), that is particularly evident for pH 13. Here, the anodic peak in NaCl- free electrolyte is growing in intensity very slowly, whereas in presence of chloride and sodium ions, the metal redox peak reaches essentially the same intensity after only the 2nd cycle, as observed after the 50th cycle in NaCl-free conditions. Similar- ly, the cathodic peak grows slowly in chloride-free conditions, whereas, in presence of chloride and sodium ions, the redox peak starts out intense and very narrow and slowly widens at the 50th cycle. Despite different NaCl-dependent evolutions, the peaks for NaCl-containing and NaCl-free electrolyte reach similar shape and intensity at the 50th cycle, both at pH 13 and at pH 9.2 (Figure S3b,d). This observation suggests that the presence of chloride and sodium ions accelerates the electro-
chemical access to nickel redox centers that are electrochemi- cally active for the nickel(II/III) redox reaction and make the re- duction much more easier at the beginning of the cycling pro- cess.
The overpotentials at 10 mAcm¢2 and the Tafel slopes ob-
tained from the LSV experiments (Figure 3a) for the chloride- free and chloride-containing electrolytes are reported in Table 1. The experimentally derived TOF in 0.1m KOH electro- lyte at the overpotential of 300 mV amounted to ~ 0.03 s¢1. The TOF was calculated assuming all metal atoms represent active sites, that is, it represents a lower limit.
The chloride ions do not seem to adversely affect the cata-
lytic OER reactivity of the NiFe LDH catalysts at moderate J as can be seen from both the LSVs and the Tafel slopes in Fig- ure 3b. The values of the Tafel slopes at moderate current den- sities are similar in all cases. This may indicate a similar OER mechanism at either pH as well as in presence of chloride ions, even though attention must be paid in analyzing the Tafel slope absolute values owing to the complexity of multistep, multielectron OER.[10b] Interestingly, a similar Tafel slope in borate buffer, ~56 mV dec¢1, was recently obtained by Smith et al. with a NiFe electrochemically co-deposited film.[20] Notice that for both pH values, the electrode potentials during the LSVs are well inside the design criterion in the case of low J; therefore, no difference should be expected according to our model.
Surprisingly, in the case of pH 9.2, chloride ions appear to
boost catalytic OER activity at higher J. Whether that boost can be entirely attributed to molecular oxygen evolution or wheth- er by products are formed is unclear without a detailed discus- sion of chemical selectivity (vide infra). Generally, at the chosen pH values, chlorine should not be produced; nonetheless, this could happen if the local pH at the anode is strongly de- creased by an inefficient proton abstraction and proton trans- port. In this case, the local acidity will negatively affect both stability and selectivity of the non-noble metal catalyst. This is why sufficient diffusive and convective mass transport at the electrolyzer electrode is crucial to maintain constant pH opera- tion and selectivity.
Figure 3. a) Electrocatalytic OER activities of NiFe LDH nanoplates supported
on carbon, measured using LSV in four different electrolytes after CV “break-
in” (50 cycles). h of approximately 480 V, corresponding to the design criteria
limit, is marked by a dashed vertical line. b) Corresponding Tafel plot for low
J. Measurement conditions: room temperature, 1600 rpm, scan rate:
1 mV s¢1
. The total metal loading determined by ICP is 7.9 mgcm¢2. Electro-
lytes: (c) 0.1m KOH, pH 13; (a) 0.1m KOH+0.5 m NaCl, pH 13;
(c) 0.3m borate buffer, pH 9.2; and (a) 0.3 m borate buffer+ 0.5m NaCl,
pH 9.2.
Table 1. LSV overpotential and Tafel slopes with respect to the four elec-
trolyte conditions.[a]
Electrolyte
h[b] [mV]
Tafel slope [mVdec¢1]
KOH
3603
511
KOH+NaCl
3591
501
Borate buffer
52912
504
Borate buffer+NaCl
4904
513
[a] Electrolytes: 1) 0.1m KOH, pH 13; 2) 0.1 m KOH+0.5m NaCl, pH 13;
3) 0.3m borate buffer, pH 9.2; and 4) 0.3m borate buffer+0.5m NaCl,
pH 9.2. [b] At 10 mAcm¢2 measured during LSV (1 mVs¢1).
ChemSusChem 2016, 9, 962 – 972
www.chemsuschem.org
Ó 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
968
Full Papers

pageNum = 8

3.4. Faradaic selectivity measurements using in situ mass spectrometry
To demonstrate the validity of the design criterion and rule out the formation of molecular chlorine or hypo-chloric acid within 480 mV overpotential under alkaline conditions, we carried out chemical selectivity measurements combining in situ head- space mass spectrometry with constant-current electrochemi- cal measurements in a custom-made titanium-based gas-tight two-compartment RDE setup with an anion-exchange mem- brane (AEM) to minimize gas cross mixing and rotor corrosion by produced chlorine. Selectivity test measurements under acidic conditions where chlorine evolution is dominant (see Figure 1) confirmed the simultaneous detection capability of gaseous oxygen and chlorine (Figure S4). Detection limits were estimated to few ppb. To measure the faradaic efficiencies of the NiFe LDH sea water splitting catalysts under alkaline condi- tions, both oxygen (m/z=32) and chlorine (m/z= 70) were monitored, while the applied J was increased in CP steps of 15 min each. The QMS ionization currents for oxygen and chlorine are reported in Figure S5.
While the oxygen signal increased stepwise with increasing
current, no significant chlorine signal could be detected across the sampled current range. This means that under the current conditions the selectivity of the OER and molecular oxygen re- mained fairly high. At the same time, this measurement indi- rectly confirmed the good proton accepting efficiency of the borate buffer electrolyte preventing the local pH to drop to acidic pH where gaseous chlorine would evolve. We also esti- mated the faradaic efficiency for molecular oxygen by relating the ion currents (black trace, Figure S6) and faradaic currents (red steps, Figure S6) and deriving faradaic efficiencies. Figure 4 plots the faradaic efficiencies together with the ap- plied current densities over their corresponding time-averaged electrode potentials. The corresponding detailed time traces of the electrode potentials are shown in Figure S7. The faradaic efficiency towards the OER and molecular oxygen remained close to 100% (experimental error less than 5%) under both pH conditions in both fresh and sea water electrolyte condi- tions inside the overpotential design criterion. Trace amounts of hypochlorous acid at electrode potentials outside our design criterion—mostly below detection limits—suggested that the hypo chlorite formation process [see Eq. (2)], despite being a two-electron process, is likely subject to own kinetic overpotentials, limiting the accumulation of hypochlorite ions. In all, our observations are in excellent agreement with our earlier prediction that molecular chlorine cannot form in sea water electrolysis under alkaline conditions (see Figure 1). Our selectivity measurements demonstrate the high faradaic effi- ciency of the NiFe LDH materials for water oxidation in seawa- ter electrolyte while operating inside our general design criteri- on.
3.5. Long-term stability and degradation in- and out-side the OER selectivity range
To address the longer-term stability of the NiFe LDH catalyst, a 2 h test at fixed J of 10 mAcm¢2 was performed according to a recently proposed protocol.[10b] An initial activation proce- dure consisting of 5 CV cycles was adopted before the CP measurement. The potential recorded during the constant-cur- rent measurement is displayed in Figure 5. At pH 13, the cata- lyst showed a reasonably stable behavior in chloride-free con- ditions (black line), with an increase of the overpotential of merely ~ 0.04–0.06 V, similar in magnitude to values reported for electrodeposited NiFeOx.[10b] In the presence of dissolved
chloride ions (red line) the catalyst is able to operate well inside the critical overpotential limit of 0.480 V required for OER selectivity (horizontal dotted line). Despite the slow in- crease of potential, sustained selective seawater electrolysis at 10 mA cm¢2 is feasible in the selective region. No change in the averaged nickel/iron ratio (Ni/Fe=3.3) and no evidence for permanent incorporation of chloride anions were observed by EDX analysis after this experiment.
In contrast, the catalyst stability in the lower pH electrolyte
(“Borate Buffer” in Figure 5) was generally worse. Considering the relatively high current densities associated with fast hy-
Figure 4. Faradaic efficiency of NiFe LDH on carbon support for OER
(*, !,&,~) and current density (*, !, &,~) as a function of averaged
measured potential during constant current potentiometric steps of 15 min
each. h of approximately 480 V, corresponding to the design criteria limit, is
marked by a dotted vertical line. Electrolytes: (c) 0.1 m KOH, (c) borate
buffer, (a) 0.1 m KOH+NaCl, and (a) borate buffer+ NaCl. Measure-
ment conditions: room temperature, 1600 rpm, N2 bubbling. The total metal
loading on the working electrode is 7.9 mgcm¢2.
ChemSusChem 2016, 9, 962 – 972
www.chemsuschem.org
Ó 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
969
Full Papers

pageNum = 9

droxide removal during OER, the catalyst degradation in borate buffer could be explained by gradual support corrosion combined with catalyst dissolution owing to a reduced interfa- cial pH caused by a limited borate buffer capacity. In all elec- trolytes, catalyst degradation could be further enhanced by mechanical impact owing to the strong bubble formation. 
 A particularly severe corrosion of the NiFe LDH water split- 
 ting catalyst was observed in borate buffer in the presence of chloride (seawater conditions). Here, the catalyst film almost completely detached into small material flake, whereas the electrode potential sharply increased to 2.4 V after about 60 min. Figure 5 evidences that under this pH condition it is no longer possible to sustain J= 10 mA cm¢2 inside the oxygen-selectivity limit. As a result of this, hypochlorite, a strong oxidant, formed hypochlorite in a two-electron pro- cess [Figure 1 and Eq. (2)]. At this bulk pH, hypochlorite ions will be present even if gaseous chlorine should locally evolve at the electrode interface owing to local acidity (low local pH owing to hydroxide removal during OER and hypochlorite for- mation), because molecular chlorine would immediately disso- ciate into hypochlorite upon diffusion into the bulk. To quanti- fy the formation of hypochlorite we have develop and utilized a iodometric titration of the chloride-containing borate-buf- fered electrolyte of the working compartment. Our titration after the stability test confirmed the formation of oxidized chlorine species (HClO and OCl¢). In total, 2.5 mmol of hypo- chlorite were produced in the experiment shown in Figure 5, corresponding to ~5% of the total current passed (total charge passed ~9.878 C). Thus, the enhanced hypochlorite for- mation in borate buffer appears a likely cause of the reduced stability of the catalyst in borate buffer. We do note that some hypochlorite could have been produced at the bare GC sup- port electrode after catalyst detachment. On the other hand, no hypochlorite was detected after the stability test in seawa- ter at pH 13 in agreement with the predictions of Figure 1. 
 The chemical dissolution of a solid electrocatalyst by com- 
 plexation with charged or neutral species is another important aspect to discuss in the context of stability. Taking the solubili- ty product at 25 8C, Ksp=5.48 10¢16 of the NiFe LDH close to 
 that of Ni(OH)2,[13] the expected free equilibrium concentration 
 of Ni2+ ions in solution (Ni(OH)2$Ni2+ +2OH¢) is extremely 
 low ranging from 5.48Õ10¢14m at pH 13 to 1.38Õ10¢7m at pH 9.2. Free Ni2+ ions form octahedral aqueous complexes, the hexa-aqua nickel(II) ions, [Ni(H2O)6]2+. However, complexation 
 of nickel ions and chloride ions to form soluble nickel chlorides could drive up the dissolution of the NiFe LDH catalyst. Indeed, under seawater conditions, some of the water ligands can be replaced by chloride ligands. It has been shown that in acidic conditions and room temperature the hexa-aqua nickel is the dominant species and the octahedral [Ni(H2O)5Cl]+ is the 
 only significant chloro-complex formed.[26] Other complexa- tions, like the octahedral NiCl2(aq) or the tetrahedral NiCl3¢ and 
 NiCl42¢, become relevant only at higher temperatures (i.e., 
 100–2008C and higher) or in chloride concentrations higher than 0.5m. The logarithm of the formation constant K of [Ni(H2O)5Cl]+ complex is log(K)=¢0.42 at 258C ([Ni(H2O)6]2+ + 
 Cl¢$[Ni(H2O)5Cl]+ +H2O).[26] Therefore, the expected distribu- 
 tion of nickel species in 0.5m chloride comprises [Ni(H2O)5Cl]+ 
 at 15% and of [Ni(H2O)6]2+ at 75%. We checked for dissolution 
 of nickel and iron in the electrolytes after the stability meas- urements by ICP-OES measurements and we compared the re- sults with the fresh electrolytes. For all 8 catalyst samples no nickel ions could be detected or remained below our detection limit of 2 mgL¢1. Iron ions were detected in all 8 samples, with concentration fluctuating between about 2 and 7 mgL¢1 (ppb). This result indicates that iron impurities were present in the electrolytes before testing and their amount was not signifi- cantly affected after the 2 h electrolysis. Therefore, if nickel and iron dissolution is occurring is beyond our detection limit (2 and 1 ppb corresponding to a ratio of the ICP detection limit to the highest possible metal concentration if all NiFe LDH was dissolved of 4 and 6%, respectively). We notice that, despite the 2 h test providing a valid screening for the analyzed condi- tions, an extended protocol is necessary to check the NiFe LDH stability in operating conditions that more closely resem- ble that of a commercial device (i.e., 8 hday¢1 for a diurnal cycle or longer times for an industrial electrolyzer). 
 4. Conclusions We have analyzed the competition of oxygen and chloro-elec- trochemistry in the context of electrochemical hydrogen pro- duction by splitting of seawater. For the first time, a rigorous general design criterion for oxygen-selective seawater splitting was derived from thermodynamic and kinetic considerations. Figure 6 summarizes our results, showing alleviated conditions for selective oxygen evolution reaction (OER) in alkaline condi- tions. Validity of the selectivity criterion was demonstrated using a family of noble-metal-free NiFe layered double hydrox- ide (LDH) electrocatalysts operated in seawater. We conclude that, at pH 13, NiFe LDH nanoplates can safely operate as OER- selective electrocatalysts in seawater inside the selective over- 
 Figure 5. a) Electrocatalytic stability test of NiFe LDH on carbon support in 
 the four electrolytes measured by 2 h CP after 5 activation cycles. The over- 
 potential h ~480 V, corresponding to the design criterion limit, is marked by 
 a dotted horizontal line. Measurement conditions: Constant J=10 mAcm¢2, 
 1600 rpm. Electrolytes: (c) 0.1m KOH, pH 13; (a) 0.1 m KOH+0.5 m 
 NaCl, pH 13; (c) 0.3 m borate buffer, pH 9.2; and (a) 0.3 m borate buf- 
 fer +0.5m NaCl, pH 9.2. (d) The potential corresponding to bare GC. 
 ChemSusChem 2016, 9, 962 – 972 
 www.chemsuschem.org 
 Ó 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim 
 970 
 Full Papers

pageNum = 10

potential range ( geted current densities of 10 mAcm¢2. This is thanks to their very high activity and stability that makes competing chlorine reactions, such as the hypochlorite formation, thermodynami- cally unfeasible below 1.72 V versus the reversible hydrogen electrode (RHE). Selectivity experiments confirmed the absence of chlorine evolution and a faradaic efficiency of ~100% to- wards OER under these conditions. Seawater electrolysis with NiFe LDH at neutral pH is limited by the lower activity ob- served at this pH condition and a strong instability, despite the better stability at near-neutral pH in chloride-free borate buffer. Suppression of chloro-chemistry at technological cur- rent densities and near-neutral pH is much more difficult to achieve owing to the lower activity at these pHs. Here, high current densities, and associated low local pH, are a likely re- sulting from catalyst corrosion. Our data strongly suggest alka- line conditions for seawater oxidation and NiFe LDH as a candi- date seawater oxidation catalyst for photoelectrochemical de- vices and electrolyzers operating at moderate current densi- ties. 
 Realizing OER-selective seawater electrolysis under acidic 
 conditions where noble metal catalysts, such as Ir or Ru, are re- quired, constitutes a much more severe challenge, as the po- tential range with high chemical selectivity becomes very narrow (180–350 mV, see Figure 6) within which even the best performing IrOx catalysts are unable to achieve current densi- 
 ties near or beyond 10 mAcm¢2.[27] 
 In all, we are confident that this first-of-its-kind analysis of 
 the scientific basis of suitable operating conditions of seawater electrocatalysis will aid in the future design of selective seawa- ter electrolyzers and seawater splitting catalysts, which consti- tutes an important contribution to a future clean power and water supply infrastructure to arid geographical world areas with ocean access. 
 Acknowledgements 
 We acknowledge financial support by the Federal Ministry of Education and Research and Federal Ministry of Economy and Energy under the grant reference number 03SF0433A “MEOKATS”. We thank the center for electron microscopy at the TU Berlin (ZELMI) for help with the TEM analysis. 
 Keywords: electrocatalysis · nickel–iron hydroxide · oxygen evolution reaction · seawater · water splitting 
 [1] E. Fabbri, A. Habereder, K. Waltar, R. Kotz, T. J. Schmidt, Catal. Sci. Tech- 
 nol. 2014, 4, 3800 –3821. 
 [2] a) J. E. Bennett, Int. J. Hydrogen Energy 1980, 5, 401 –408 ; b) Y. Surendra- 
 nath, M. Dinca, D. G. Nocera, J. Am. Chem. Soc. 2009, 131, 2615 –2620 ; 
 c) H. K. Abdel-Aal, K. M. Zohdy, M. Abdel Kareem, Open Fuel Cells J. 
 2010, 3, 1–7; d) A. A. El-Moneim, N. Kumagai, K. Hashimoto, Mater. 
 Trans. 2009, 50, 1969– 1977 ; e) K. Fujimura, K. Izumiya, A. Kawashima, E. 
 Akiyama, H. Habazaki, N. Kumagai, K. Hashimoto, J. Appl. Electrochem. 
 1999, 29, 769 –775 ; f) N. Jiang, H. M. Meng, Surf. Coat. Technol. 2012, 
 206, 4362– 4367. 
 [3] a) W. J. Luo, Z. S. Yang, Z. S. Li, J. Y. Zhang, J. G. Liu, Z. Y. Zhao, Z. Q. 
 Wang, S. C. Yan, T. Yu, Z. G. Zou, Energy Environ. Sci. 2011, 4, 4046 – 
 4051 ; b) K. Maeda, H. Masuda, K. Domen, Catal. Today 2009, 147, 173 – 
 178 ; c) S. M. Ji, H. Jun, J. S. Jang, H. C. Son, P. H. Borse, J. S. Lee, J. Photo- 
 chem. Photobiol. A 2007, 189, 141 –144. 
 [4] a) H. Y. Su, Y. Gorlin, I. C. Man, F. Calle-Vallejo, J. K. Norskov, T. F. Jaramillo, 
 J. Rossmeisl, Phys. Chem. Chem. Phys. 2012, 14, 14010 –14022 ; b) W. T. 
 Hong, M. Risch, K. A. Stoerzinger, A. Grimaud, J. Suntivich, Y. Shao-Horn, 
 Energy Environ. Sci. 2015, 8, 1404 –1427. 
 [5] S. Trasatti, Electrochim. Acta 1984, 29, 1503 –1512 . 
 [6] a) A. R. Zeradjanin, N. Menzel, W. Schuhmann, P. Strasser, Phys. Chem. 
 Chem. Phys. 2014, 16, 13741– 13747 ; b) R. K. B. Karlsson, H. A. Hansen, T. 
 Bligaard, A. Cornell, L. G. M. Pettersson, Electrochim. Acta 2014, 146, 
 733– 740 ; c) K. S. Exner, J. Anton, T. Jacob, H. Over, Angew. Chem. Int. Ed. 
 2014, 53, 11032 – 11035 ; Angew. Chem. 2014, 126, 11212 –11215 ; d) V. 
 Petrykin, K. Macounova, O. A. Shlyakhtin, P. Krtil, Angew. Chem. Int. Ed. 
 2010, 49, 4813 –4815 ; Angew. Chem. 2010, 122, 4923 –4925 ; e) K. Ma- 
 counov, M. Makarova, J. Jirkovsky, J. Franc, P. Krtil, Electrochim. Acta 
 2008, 53, 6126 – 6134 ; f) T. Arikawa, Y. Murakami, Y. Takasu, J. Appl. Elec- 
 trochem. 1998, 28, 511–516. 
 [7] WorldChlorineCouncil. Available from http://www.worldchlorine.org/ 
 wp-content/themes/brickthemewp/pdfs/sustainablefuture.pdf. 
 [8] C. Egenhofer, L. Schrefler, V. Rizos, F. Infelise, G. Luchetta, F. Simonelli, W. 
 Stoefs, J. Timini, L. Colantoni, Centre for European Policy Studies, 2014, 
 p. 25. 
 [9] EuroChlor, 2012, p. 48. Available from http://www.eurochlor.org/media/ 
 63146/2012-annualreview-final.pdf. 
 [10] a) M. W. Louie, A. T. Bell, J. Am. Chem. Soc. 2013, 135, 12329 – 12337 ; 
 b) C. C. L. McCrory, S. H. Jung, J. C. Peters, T. F. Jaramillo, J. Am. Chem. 
 Soc. 2013, 135, 16977 – 16987 ; c) 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 –8455 ; d) L. Trotochaud, S. L. Young, J. K. 
 Ranney, S. W. Boettcher, J. Am. Chem. Soc. 2014, 136, 6744 –6753 . 
 [11] a) C. Du, X. G. Yang, M. T. Mayer, H. Hoyt, J. Xie, G. McMahon, G. Bischop- 
 ing, D. W. Wang, Angew. Chem. Int. Ed. 2013, 52, 12692 – 12695 ; Angew. 
 Chem. 2013, 125, 12924 – 12927 ; b) L. Wang, F. Dionigi, N. T. Nguyen, R. 
 Kirchgeorg, M. Gliech, S. Grigorescu, P. Strasser, P. Schmuki, Chem. Mater. 
 2015, 27, 2360 –2366 ; c) 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 –3461. 
 [12] J. C. Crittenden, R. Rhodes Trussell, D. W. Hand, K. J. Howe, G. Tchoba- 
 noglous, MWH’s Water Treatment: Principles and Design, 3rd ed., 2012. 
 [13] CRC Handbook of Chemistry and Physics, 96th ed. (Ed.: D. R. Lide), CRC 
 Press, New York, 2015. 
 [14] H. A. Hansen, I. C. Man, F. Studt, F. Abild-Pedersen, T. Bligaard, J. Ross- 
 meisl, Phys. Chem. Chem. Phys. 2010, 12, 283– 290. 
 Figure 6. Maximum allowed overpotential of OER electrolyzer catalysts to 
 ensure 100% selective water splitting. Values are obtained as difference be- 
 tween standard electrode potentials of the 3 relevant chloride oxidation re- 
 actions (chlorine, hypochlorous acid, and hypochlorite formation) and the 
 OER versus pH. The dahsed area corresponds to oxygen-selective overpoten- 
 tial–pH conditions where thermodynamics point to 100% selective oxygen 
 evolution. Above the limiting lines, chlorine-based reaction products are 
 thermodynamically feasible, yet may be limited by kinetic overpotentials. 
 ChemSusChem 2016, 9, 962 – 972 
 www.chemsuschem.org 
 Ó 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim 
 971 
 Full Papers

pageNum = 11

[15] 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– 363. 
 [16] 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 – 1313. 
 [17] F. Song, X. L. Hu, Nat. Commun. 2014, 5, 4477. [18] M. Auinger, I. Katsounaros, J. C. Meier, S. O. Klemm, P. U. Biedermann, 
 A. A. Topalov, M. Rohwerder, K. J. J. Mayrhofer, Phys. Chem. Chem. Phys. 2011, 13, 16384 – 16394. 
 [19] a) B. J. Trzes´niewski, O. Diaz-Morales, D. A. Vermaas, A. Longo, W. Bras, 
 M. T. M. Koper, W. A. Smith, J. Am. Chem. Soc. 2015, 137, 15112 –15121; b) M. Dinca, Y. Surendranath, D. G. Nocera, Proc. Natl. Acad. Sci. USA 2010, 107, 10337–10341 . 
 [20] A. M. Smith, L. Trotochaud, M. S. Burke, S. W. Boettcher, Chem. Commun. 
 2015, 51, 5261 –5263. 
 [21] A. T. Kuhn, C. Y. Chan, J. Appl. Electrochem. 1983, 13, 189 –207. [22] A. J. Bard, L. R. Faulkner, Electrochemical methods : fundamentals and ap- 
 plications, 2nd ed., Wiley, New York, 2001. 
 [23] M. J. Kenney, M. Gong, Y. G. Li, J. Z. Wu, J. Feng, M. Lanza, H. J. Dai, Sci- 
 ence 2013, 342, 836–840. 
 [24] R. L. Doyle, I. J. Godwin, M. P. Brandon, M. E. G. Lyons, Phys. Chem. 
 Chem. Phys. 2013, 15, 13737 –13783. 
 [25] M. E. G. Lyons, A. Cakara, P. O’Brien, I. Godwin, R. L. Doyle, Int. J. Electro- 
 chem. Soc. 2012, 7, 11768– 11795. 
 [26] W. H. Liu, A. Migdisov, A. Williams-Jones, Geochim. Cosmochim. Acta 
 2012, 94, 276– 290. 
 [27] a) H. N. Nong, H. S. Oh, T. Reier, E. Willinger, M. G. Willinger, V. Petkov, D. 
 Teschner, P. Strasser, Angew. Chem. Int. Ed. 2015, 54, 2975– 2979 ; Angew. 
 Chem. 2015, 127, 3018 –3022 ; b) H. S. Oh, H. N. Nong, T. Reier, M. 
 Gliech, P. Strasser, Chem. Sci. 2015, 6, 3321– 3328 ; c) T. Reier, M. Oezas- 
 lan, P. Strasser, ACS Catal. 2012, 2, 1765 –1772 ; d) T. Reier, D. Teschner, T. 
 Lunkenbein, A. Bergmann, S. Selve, R. Kraehnert, R. Schlçgl, P. Strasser, 
 J. Electrochem. Soc. 2014, 161, F876 – F882. 
 Received: November 26, 2015 
 Revised: January 29, 2016 
 Published online on March 24, 2016 
 ChemSusChem 2016, 9, 962 – 972 
 www.chemsuschem.org 
 Ó 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim 
 972 
 Full Papers

pageNum = 12

pageNum = 13

pageNum = 14

pageNum = 15

pageNum = 16

pageNum = 17

pageNum = 18

pageNum = 19

pageNum = 20

droxide removal during OER, the catalyst degradation in borate buffer could be explained by gradual support corrosion combined with catalyst dissolution owing to a reduced interfa- cial pH caused by a limited borate buffer capacity. In all elec- trolytes, catalyst degradation could be further enhanced by mechanical impact owing to the strong bubble formation. 
 A particularly severe corrosion of the NiFe LDH water split- 
 ting catalyst was observed in borate buffer in the presence of chloride (seawater conditions). Here, the catalyst film almost completely detached into small material flake, whereas the electrode potential sharply increased to 2.4 V after about 60 min. Figure 5 evidences that under this pH condition it is no longer possible to sustain J= 10 mA cm¢2 inside the oxygen-selectivity limit. As a result of this, hypochlorite, a strong oxidant, formed hypochlorite in a two-electron pro- cess [Figure 1 and Eq. (2)]. At this bulk pH, hypochlorite ions will be present even if gaseous chlorine should locally evolve at the electrode interface owing to local acidity (low local pH owing to hydroxide removal during OER and hypochlorite for- mation), because molecular chlorine would immediately disso- ciate into hypochlorite upon diffusion into the bulk. To quanti- fy the formation of hypochlorite we have develop and utilized a iodometric titration of the chloride-containing borate-buf- fered electrolyte of the working compartment. Our titration after the stability test confirmed the formation of oxidized chlorine species (HClO and OCl¢). In total, 2.5 mmol of hypo- chlorite were produced in the experiment shown in Figure 5, corresponding to ~5% of the total current passed (total charge passed ~9.878 C). Thus, the enhanced hypochlorite for- mation in borate buffer appears a likely cause of the reduced stability of the catalyst in borate buffer. We do note that some hypochlorite could have been produced at the bare GC sup- port electrode after catalyst detachment. On the other hand, no hypochlorite was detected after the stability test in seawa- ter at pH 13 in agreement with the predictions of Figure 1. 
 The chemical dissolution of a solid electrocatalyst by com- 
 plexation with charged or neutral species is another important aspect to discuss in the context of stability. Taking the solubili- ty product at 25 8C, Ksp=5.48 10¢16 of the NiFe LDH close to 
 that of Ni(OH)2,[13] the expected free equilibrium concentration 
 of Ni2+ ions in solution (Ni(OH)2$Ni2+ +2OH¢) is extremely 
 low ranging from 5.48Õ10¢14m at pH 13 to 1.38Õ10¢7m at pH 9.2. Free Ni2+ ions form octahedral aqueous complexes, the hexa-aqua nickel(II) ions, [Ni(H2O)6]2+. However, complexation 
 of nickel ions and chloride ions to form soluble nickel chlorides could drive up the dissolution of the NiFe LDH catalyst. Indeed, under seawater conditions, some of the water ligands can be replaced by chloride ligands. It has been shown that in acidic conditions and room temperature the hexa-aqua nickel is the dominant species and the octahedral [Ni(H2O)5Cl]+ is the 
 only significant chloro-complex formed.[26] Other complexa- tions, like the octahedral NiCl2(aq) or the tetrahedral NiCl3¢ and 
 NiCl42¢, become relevant only at higher temperatures (i.e., 
 100–2008C and higher) or in chloride concentrations higher than 0.5m. The logarithm of the formation constant K of [Ni(H2O)5Cl]+ complex is log(K)=¢0.42 at 258C ([Ni(H2O)6]2+ + 
 Cl¢$[Ni(H2O)5Cl]+ +H2O).[26] Therefore, the expected distribu- 
 tion of nickel species in 0.5m chloride comprises [Ni(H2O)5Cl]+ 
 at 15% and of [Ni(H2O)6]2+ at 75%. We checked for dissolution 
 of nickel and iron in the electrolytes after the stability meas- urements by ICP-OES measurements and we compared the re- sults with the fresh electrolytes. For all 8 catalyst samples no nickel ions could be detected or remained below our detection limit of 2 mgL¢1. Iron ions were detected in all 8 samples, with concentration fluctuating between about 2 and 7 mgL¢1 (ppb). This result indicates that iron impurities were present in the electrolytes before testing and their amount was not signifi- cantly affected after the 2 h electrolysis. Therefore, if nickel and iron dissolution is occurring is beyond our detection limit (2 and 1 ppb corresponding to a ratio of the ICP detection limit to the highest possible metal concentration if all NiFe LDH was dissolved of 4 and 6%, respectively). We notice that, despite the 2 h test providing a valid screening for the analyzed condi- tions, an extended protocol is necessary to check the NiFe LDH stability in operating conditions that more closely resem- ble that of a commercial device (i.e., 8 hday¢1 for a diurnal cycle or longer times for an industrial electrolyzer). 
 4. Conclusions We have analyzed the competition of oxygen and chloro-elec- trochemistry in the context of electrochemical hydrogen pro- duction by splitting of seawater. For the first time, a rigorous general design criterion for oxygen-selective seawater splitting was derived from thermodynamic and kinetic considerations. Figure 6 summarizes our results, showing alleviated conditions for selective oxygen evolution reaction (OER) in alkaline condi- tions. Validity of the selectivity criterion was demonstrated using a family of noble-metal-free NiFe layered double hydrox- ide (LDH) electrocatalysts operated in seawater. We conclude that, at pH 13, NiFe LDH nanoplates can safely operate as OER- selective electrocatalysts in seawater inside the selective over- 
 Figure 5. a) Electrocatalytic stability test of NiFe LDH on carbon support in 
 the four electrolytes measured by 2 h CP after 5 activation cycles. The over- 
 potential h ~480 V, corresponding to the design criterion limit, is marked by 
 a dotted horizontal line. Measurement conditions: Constant J=10 mAcm¢2, 
 1600 rpm. Electrolytes: (c) 0.1m KOH, pH 13; (a) 0.1 m KOH+0.5 m 
 NaCl, pH 13; (c) 0.3 m borate buffer, pH 9.2; and (a) 0.3 m borate buf- 
 fer +0.5m NaCl, pH 9.2. (d) The potential corresponding to bare GC. 
 ChemSusChem 2016, 9, 962 – 972 
 www.chemsuschem.org 
 Ó 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim 
 970 
 Full Papers

pageNum = 21

potential range ( geted current densities of 10 mAcm¢2. This is thanks to their very high activity and stability that makes competing chlorine reactions, such as the hypochlorite formation, thermodynami- cally unfeasible below 1.72 V versus the reversible hydrogen electrode (RHE). Selectivity experiments confirmed the absence of chlorine evolution and a faradaic efficiency of ~100% to- wards OER under these conditions. Seawater electrolysis with NiFe LDH at neutral pH is limited by the lower activity ob- served at this pH condition and a strong instability, despite the better stability at near-neutral pH in chloride-free borate buffer. Suppression of chloro-chemistry at technological cur- rent densities and near-neutral pH is much more difficult to achieve owing to the lower activity at these pHs. Here, high current densities, and associated low local pH, are a likely re- sulting from catalyst corrosion. Our data strongly suggest alka- line conditions for seawater oxidation and NiFe LDH as a candi- date seawater oxidation catalyst for photoelectrochemical de- vices and electrolyzers operating at moderate current densi- ties. 
 Realizing OER-selective seawater electrolysis under acidic 
 conditions where noble metal catalysts, such as Ir or Ru, are re- quired, constitutes a much more severe challenge, as the po- tential range with high chemical selectivity becomes very narrow (180–350 mV, see Figure 6) within which even the best performing IrOx catalysts are unable to achieve current densi- 
 ties near or beyond 10 mAcm¢2.[27] 
 In all, we are confident that this first-of-its-kind analysis of 
 the scientific basis of suitable operating conditions of seawater electrocatalysis will aid in the future design of selective seawa- ter electrolyzers and seawater splitting catalysts, which consti- tutes an important contribution to a future clean power and water supply infrastructure to arid geographical world areas with ocean access. 
 Acknowledgements 
 We acknowledge financial support by the Federal Ministry of Education and Research and Federal Ministry of Economy and Energy under the grant reference number 03SF0433A “MEOKATS”. We thank the center for electron microscopy at the TU Berlin (ZELMI) for help with the TEM analysis. 
 Keywords: electrocatalysis · nickel–iron hydroxide · oxygen evolution reaction · seawater · water splitting 
 [1] E. Fabbri, A. Habereder, K. Waltar, R. Kotz, T. J. Schmidt, Catal. Sci. Tech- 
 nol. 2014, 4, 3800 –3821. 
 [2] a) J. E. Bennett, Int. J. Hydrogen Energy 1980, 5, 401 –408 ; b) Y. Surendra- 
 nath, M. Dinca, D. G. Nocera, J. Am. Chem. Soc. 2009, 131, 2615 –2620 ; 
 c) H. K. Abdel-Aal, K. M. Zohdy, M. Abdel Kareem, Open Fuel Cells J. 
 2010, 3, 1–7; d) A. A. El-Moneim, N. Kumagai, K. Hashimoto, Mater. 
 Trans. 2009, 50, 1969– 1977 ; e) K. Fujimura, K. Izumiya, A. Kawashima, E. 
 Akiyama, H. Habazaki, N. Kumagai, K. Hashimoto, J. Appl. Electrochem. 
 1999, 29, 769 –775 ; f) N. Jiang, H. M. Meng, Surf. Coat. Technol. 2012, 
 206, 4362– 4367. 
 [3] a) W. J. Luo, Z. S. Yang, Z. S. Li, J. Y. Zhang, J. G. Liu, Z. Y. Zhao, Z. Q. 
 Wang, S. C. Yan, T. Yu, Z. G. Zou, Energy Environ. Sci. 2011, 4, 4046 – 
 4051 ; b) K. Maeda, H. Masuda, K. Domen, Catal. Today 2009, 147, 173 – 
 178 ; c) S. M. Ji, H. Jun, J. S. Jang, H. C. Son, P. H. Borse, J. S. Lee, J. Photo- 
 chem. Photobiol. A 2007, 189, 141 –144. 
 [4] a) H. Y. Su, Y. Gorlin, I. C. Man, F. Calle-Vallejo, J. K. Norskov, T. F. Jaramillo, 
 J. Rossmeisl, Phys. Chem. Chem. Phys. 2012, 14, 14010 –14022 ; b) W. T. 
 Hong, M. Risch, K. A. Stoerzinger, A. Grimaud, J. Suntivich, Y. Shao-Horn, 
 Energy Environ. Sci. 2015, 8, 1404 –1427. 
 [5] S. Trasatti, Electrochim. Acta 1984, 29, 1503 –1512 . 
 [6] a) A. R. Zeradjanin, N. Menzel, W. Schuhmann, P. Strasser, Phys. Chem. 
 Chem. Phys. 2014, 16, 13741– 13747 ; b) R. K. B. Karlsson, H. A. Hansen, T. 
 Bligaard, A. Cornell, L. G. M. Pettersson, Electrochim. Acta 2014, 146, 
 733– 740 ; c) K. S. Exner, J. Anton, T. Jacob, H. Over, Angew. Chem. Int. Ed. 
 2014, 53, 11032 – 11035 ; Angew. Chem. 2014, 126, 11212 –11215 ; d) V. 
 Petrykin, K. Macounova, O. A. Shlyakhtin, P. Krtil, Angew. Chem. Int. Ed. 
 2010, 49, 4813 –4815 ; Angew. Chem. 2010, 122, 4923 –4925 ; e) K. Ma- 
 counov, M. Makarova, J. Jirkovsky, J. Franc, P. Krtil, Electrochim. Acta 
 2008, 53, 6126 – 6134 ; f) T. Arikawa, Y. Murakami, Y. Takasu, J. Appl. Elec- 
 trochem. 1998, 28, 511–516. 
 [7] WorldChlorineCouncil. Available from http://www.worldchlorine.org/ 
 wp-content/themes/brickthemewp/pdfs/sustainablefuture.pdf. 
 [8] C. Egenhofer, L. Schrefler, V. Rizos, F. Infelise, G. Luchetta, F. Simonelli, W. 
 Stoefs, J. Timini, L. Colantoni, Centre for European Policy Studies, 2014, 
 p. 25. 
 [9] EuroChlor, 2012, p. 48. Available from http://www.eurochlor.org/media/ 
 63146/2012-annualreview-final.pdf. 
 [10] a) M. W. Louie, A. T. Bell, J. Am. Chem. Soc. 2013, 135, 12329 – 12337 ; 
 b) C. C. L. McCrory, S. H. Jung, J. C. Peters, T. F. Jaramillo, J. Am. Chem. 
 Soc. 2013, 135, 16977 – 16987 ; c) 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 –8455 ; d) L. Trotochaud, S. L. Young, J. K. 
 Ranney, S. W. Boettcher, J. Am. Chem. Soc. 2014, 136, 6744 –6753 . 
 [11] a) C. Du, X. G. Yang, M. T. Mayer, H. Hoyt, J. Xie, G. McMahon, G. Bischop- 
 ing, D. W. Wang, Angew. Chem. Int. Ed. 2013, 52, 12692 – 12695 ; Angew. 
 Chem. 2013, 125, 12924 – 12927 ; b) L. Wang, F. Dionigi, N. T. Nguyen, R. 
 Kirchgeorg, M. Gliech, S. Grigorescu, P. Strasser, P. Schmuki, Chem. Mater. 
 2015, 27, 2360 –2366 ; c) 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 –3461. 
 [12] J. C. Crittenden, R. Rhodes Trussell, D. W. Hand, K. J. Howe, G. Tchoba- 
 noglous, MWH’s Water Treatment: Principles and Design, 3rd ed., 2012. 
 [13] CRC Handbook of Chemistry and Physics, 96th ed. (Ed.: D. R. Lide), CRC 
 Press, New York, 2015. 
 [14] H. A. Hansen, I. C. Man, F. Studt, F. Abild-Pedersen, T. Bligaard, J. Ross- 
 meisl, Phys. Chem. Chem. Phys. 2010, 12, 283– 290. 
 Figure 6. Maximum allowed overpotential of OER electrolyzer catalysts to 
 ensure 100% selective water splitting. Values are obtained as difference be- 
 tween standard electrode potentials of the 3 relevant chloride oxidation re- 
 actions (chlorine, hypochlorous acid, and hypochlorite formation) and the 
 OER versus pH. The dahsed area corresponds to oxygen-selective overpoten- 
 tial–pH conditions where thermodynamics point to 100% selective oxygen 
 evolution. Above the limiting lines, chlorine-based reaction products are 
 thermodynamically feasible, yet may be limited by kinetic overpotentials. 
 ChemSusChem 2016, 9, 962 – 972 
 www.chemsuschem.org 
 Ó 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim 
 971 
 Full Papers

pageNum = 22

[15] 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– 363. 
 [16] 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 – 1313. 
 [17] F. Song, X. L. Hu, Nat. Commun. 2014, 5, 4477. [18] M. Auinger, I. Katsounaros, J. C. Meier, S. O. Klemm, P. U. Biedermann, 
 A. A. Topalov, M. Rohwerder, K. J. J. Mayrhofer, Phys. Chem. Chem. Phys. 2011, 13, 16384 – 16394. 
 [19] a) B. J. Trzes´niewski, O. Diaz-Morales, D. A. Vermaas, A. Longo, W. Bras, 
 M. T. M. Koper, W. A. Smith, J. Am. Chem. Soc. 2015, 137, 15112 –15121; b) M. Dinca, Y. Surendranath, D. G. Nocera, Proc. Natl. Acad. Sci. USA 2010, 107, 10337–10341 . 
 [20] A. M. Smith, L. Trotochaud, M. S. Burke, S. W. Boettcher, Chem. Commun. 
 2015, 51, 5261 –5263. 
 [21] A. T. Kuhn, C. Y. Chan, J. Appl. Electrochem. 1983, 13, 189 –207. [22] A. J. Bard, L. R. Faulkner, Electrochemical methods : fundamentals and ap- 
 plications, 2nd ed., Wiley, New York, 2001. 
 [23] M. J. Kenney, M. Gong, Y. G. Li, J. Z. Wu, J. Feng, M. Lanza, H. J. Dai, Sci- 
 ence 2013, 342, 836–840. 
 [24] R. L. Doyle, I. J. Godwin, M. P. Brandon, M. E. G. Lyons, Phys. Chem. 
 Chem. Phys. 2013, 15, 13737 –13783. 
 [25] M. E. G. Lyons, A. Cakara, P. O’Brien, I. Godwin, R. L. Doyle, Int. J. Electro- 
 chem. Soc. 2012, 7, 11768– 11795. 
 [26] W. H. Liu, A. Migdisov, A. Williams-Jones, Geochim. Cosmochim. Acta 
 2012, 94, 276– 290. 
 [27] a) H. N. Nong, H. S. Oh, T. Reier, E. Willinger, M. G. Willinger, V. Petkov, D. 
 Teschner, P. Strasser, Angew. Chem. Int. Ed. 2015, 54, 2975– 2979 ; Angew. 
 Chem. 2015, 127, 3018 –3022 ; b) H. S. Oh, H. N. Nong, T. Reier, M. 
 Gliech, P. Strasser, Chem. Sci. 2015, 6, 3321– 3328 ; c) T. Reier, M. Oezas- 
 lan, P. Strasser, ACS Catal. 2012, 2, 1765 –1772 ; d) T. Reier, D. Teschner, T. 
 Lunkenbein, A. Bergmann, S. Selve, R. Kraehnert, R. Schlçgl, P. Strasser, 
 J. Electrochem. Soc. 2014, 161, F876 – F882. 
 Received: November 26, 2015 
 Revised: January 29, 2016 
 Published online on March 24, 2016 
 ChemSusChem 2016, 9, 962 – 972 
 www.chemsuschem.org 
 Ó 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim 
 972 
 Full Papers

Home