Boosting Water Oxidation through In Situ Electroconversion of Manganese Gallide: An Intermetallic Precursor Approach

Abstract For the first time, the manganese gallide (MnGa4) served as an intermetallic precursor, which upon in situ electroconversion in alkaline media produced high‐performance and long‐term‐stable MnOx‐based electrocatalysts for water oxidation. Unexpectedly, its electrocorrosion (with the concomitant loss of Ga) leads simultaneously to three crystalline types of MnOx minerals with distinct structures and induced defects: birnessite δ‐MnO2, feitknechtite β‐MnOOH, and hausmannite α‐Mn3O4. The abundance and intrinsic stabilization of MnIII/MnIV active sites in the three MnOx phases explains the superior efficiency and durability of the system for electrocatalytic water oxidation. After electrophoretic deposition of the MnGa4 precursor on conductive nickel foam (NF), a low overpotential of 291 mV, comparable to that of precious‐metal‐based catalysts, could be achieved at a current density of 10 mA cm−2 with a durability of more than five days.


Instrumentations
Powder X-ray diffraction (PXRD) patterns were measured on a Bruker AXS D8 advanced automatic diffractometer equipped with a position-sensitive detector (PSD) and curved germanium (111) primary monochromator using Cu Kα radiation (λ = 1.5418 Å). The inductively coupled plasma atomic emission spectroscopy (ICP-AES) was conducted on a Thermo Jarrell Ash Trace Scan analyzer. The materials were digested in aqua regia HCl: HNO3 3:1 v/v (nitric acid, SUPRA-Qualität ROTIPURAN® Supra 69% and hydrochloric acid, SUPRA-Qualität ROTIPURAN® Supra 30%) and the average of three reproducible independent experiments has been presented (the electrolytes after electrochemistry were used directly). The digestion volume (2.5 mL) was diluted with Milli-Q water up to 15 mL. Calibration curves were prepared for both cobalt and phosphorus with concentrations between 1 mgL -1 and 100 mgL -1 from standard solutions (1000 mgL -1 Single-Element ICP-Standard Solution ROTI®STAR).
Fourier-transform infrared spectroscopy (FTIR) was studied using a BIORAD FTS 6000 FTIR spectrometer under attenuated total reflection (ATR) conditions. The data were recorded in the range of 500-4000 cm -1 with an average of 32 scans at 4 cm -1 resolution.
To gather information on the morphology, and the surface characteristics, the scanning electron microscopy (SEM) was conducted on an LEO DSM 982 microscope integrated with EDX (EDAX, Apollo XPP). Data handling and analyses were achieved with the software package EDAX.
The microstructure investigations of the materials were revealed by transmission electron microscopy (TEM) which was investigated on an FEI Tecnai G2 20 S-TWIN transmission electron microscope (FEI Company, Eindhoven, Netherlands) and JEOL 2100 electron microscope equipped with a LaB6 source at 200 kV acceleration voltage. For the investigation of the films after electrocatalysis, the films were scraped from the electrode substrate and transferred onto a carbon-coated copper grid. EDX analyses were achieved with an EDAX r-TEM SUTW detector (Si (Li) detector), and the images were recorded with a GATAN MS794 P CCD camera. The SEM and TEM experiments were conducted partially at the Zentrum für Elektronenmikroskopie (ZELMI) of the TU Berlin.
Gas chromatography was used to calculate the Faradaic efficiency (FE) of oxygen evolution reaction (OER) that was performed in a closed (gas-tight) electrochemical cell. An Agilent 7890A gas chromatograph (GC) was used to determine the oxygen content in the headspace of the electrochemical cell. The GC was furnished with a carboxen-1000 column and a thermal conductivity detector (TCD). The carrier gas was argon (Ar).
A Signature Pro4 System measured the resistivity with Keithley 2400 source-measure unit (SP4-40045TBY) using a four-point probe resistivity technique. The spacing between tungsten carbide tips was 1.016 mm with a radius of 0.245 mm, and a spring pressure was 45 grams. The materials were electrophoretically deposited on electrodes to estimate the specific resistivity of each synthesized material, and the average results are presented.

X-ray photoelectron spectroscopy (XPS)
The XPS measurements were carried out on a Kratos Axis Ultra X-ray photoelectron spectrometer (Kratos Analytical Ltd., Manchester, U.K.) using an Al Kα monochromatic radiation source (1486.7 eV) with 90° takeoff angle (normal to analyzer). The vacuum pressure in the analyzing chamber was kept at 2 × 10 −9 Torr. The XPS spectra were collected for C 1s, O 1s, Mn 2p, Mn 3p and Ga 3p levels with pass energy 20 eV and step 0.1 eV. The binding energies were calibrated relative to the C 1s peak energy position as 285.0 eV. Data analyses were carried out using Casa XPS (Casa Software Ltd.) and the Vision data processing program (Kratos Analytical Ltd.).

X-ray absorption spectroscopy (XAS)
Mn and Ga K-edge X-ray absorption near edge structure (XANES) spectra were collected using a von Hamos laboratory XAFS spectrometer, [1] which operates in transmission mode. As the source, an air-cooled microfocus X-ray tube with a power of 30 W, a spot size of 60 μm and molybdenum as the anode material was used. A cylindrically bent highly annealed pyrolytic graphite (HAPG) crystal with a size of 5 × 5 cm 2 and a bending radius of 15 cm was applied as a dispersive element and detection of the X-rays was accomplished using an indirectly detecting X-ray CCD camera of the type Andor Newton A-DY940P-FO-CSI with a spatial resolution of 50 μm and a size of 1 × 0.25 inch 2 . With this setup, a spectral resolving power of E/ΔE = 2000 can be achieved in an energy range from 4 keV up to 12 keV. The ground powder samples of as-prepared MnGa4 and Mn, as well as other reference materials MnO, MnO2, and Mn2O3, have been applied on scotch tape and stacked to adopt the adequate thickness for the transmission mode measurements. [2] Also, the as-deposited films of MnGa4 were also measured to have a fair comparison to that of ground powder. Similarly, the MnGa4 films after the electrochemical OER experiments were measured, and the shifts in energy concerning as-prepared materials and reference materials were noted. The absorption spectra were normalized and analyzed using the XAS evaluation software ATHENA.47. [3]

Synthesis of intermetallic MnGa4
The polycrystalline sample of MnGa4 was prepared by annealing the stoichiometric mixture of Mn and Ga in an evacuated quartz ampule. The synthetic conditions were chosen based on the reported phase diagram. [4] The ampule was heated in a programmable furnace to 900 ˚C, annealed at this temperature for four days to ensure homogeneity of the mixture, cooled at the rate of 20 ˚C/h to 380 ˚C and annealed at 380 ˚C for ten days. Then, the material was thoroughly ground and annealed at 380 ˚C for another ten days to produce pure MnGa4 phase. [5] 0.5 g of bis(cyclopentadienyl)manganese(II) was weighed in a glove box and carefully transferred under protective atmosphere into argon flushed tube furnace. The furnace was then flushed with argon for 2 h and then switched to oxygen atmosphere during annealing. The temperature was increased with 6 ˚C/min to 700 °C and maintained at that temperature for 12 h followed by cooling down naturally to room temperature. The product consisted of small particles and dark grey colored.

Synthesis of Mn2O3-SSP [6]
The preparation of Mn2O3-SSP was carried out in two steps. In the first step, micro-emulsions containing cetyltrimethylammonium bromide (CTAB, 2.0 g) as a surfactant, 1-hexanol (20 mL) as co-surfactant and hexane (35 ml) as the lipophilic phase were prepared and were mixed separately with an aqueous solution of 0.1 M manganese acetate and ammonium oxalate. Both micro-emulsions were mixed slowly and stirred overnight. The white precipitate then obtained was centrifuged and washed with 1:1 mixture of chloroform and methanol (200 mL) and subsequently dried at 60 ˚C for 12 hours. In the next step, the manganese oxalate precursor was heated in dry synthetic air (20% O2, 80% N2) at 400 ˚C for 8 hours (2 ˚C/min) to form monophasic Mn2O3.

Electrophoretic deposition (EPD) on NF, FTO, and CC
The investigated materials were deposited on both, NF, FTO and CC, electrophoretically, by applying a potential difference of 10 V in a mixture of iodine and acetone on a 1 × 1 cm 2 area. The detailed mechanism involving electrophoretic deposition has been described elsewhere. [7] The electric charge on the catalyst in acetone is insufficient for EPD as very small amounts of free ions exist in acetone, and therefore, large potentials are required for EPD. [7] When iodine is used as the dispersant, it can react with acetone through the keto-enol tautomerism to produce protons as per the following equation.
Thus formed protons are adsorbed on the surface of the suspended particles by making them positively charged. The applied electric field induces the positively charged particles to migrate towards and deposit on the cathode. [8] In order to have a better deposition on the electrode substrates, the large MnGa4 crystals were grounded for 15 min to reduce that reduced the particle size. Typically, 30 mg of the catalyst powder was suspended in 10 ml acetone, and 3 mg of iodine was then added. This solution was agitated in an ultrasonic bath for 30 min. Before EPD, the empty electrodes were weighed using an analytical balance, and the weights were noted. The EPD was conducted at various potentials different time intervals, and the thin uniform films were only achieved by applying a potential at -10 V for 2 minutes with stirring the solution continuously at room temperature. After each EPD, the increase in weight of the electrodes was monitored carefully. The catalyst loading on each NF, CC, and FTO was ~2, ~1.1 and ~0.7 mgcm −2 , respectively. The mass loading was reproducible within the margins of an experimental error (±0.05 mg).

Electrochemical measurements
A typical electrocatalytic run was carried out in a standard three-electrode (working, counter and reference) electrochemical cell in 1 M aqueous KOH with a potentiostat (SP-200, BioLogic Science Instruments) controlled by the EC-Lab v10.20 software package. The electrodes (NF/FTO/CC) with samples deposited served as the working electrodes, Pt wire (0.5 mm diameter × 230 mm length; A-002234, BioLogic) as a counter and Hg/HgO as the reference electrode (CH Instruments, Inc.). Cyclic voltammetry (CV) and linear sweep voltammetry (LSV) were carried out with an applied iR compensation of 85%. The potentials presented in this work were referenced to the reversible hydrogen electrode (RHE) through calibration, and in 1 M aqueous KOH, E(RHE) = E(Hg/HgO) + 0.098 V + (0.059 × pH) V. The chronoamperometric measurements were performed in 1 M aqueous KOH at selected constant potentials with respect to RHE. The Tafel slope was calculated according to Tafel equation η = blog j + a, where η is overpotential (V), j is the current density (mA cm -2 ), and b is the Tafel slope (mV dec -1 ). [9] The electrochemically active surface area (ECSA) of both MnGa4 and Mn was obtained by determining their double layer capacitances (Cdl) from the CV (cycled between 0.925 and 1.025 V vs. RHE) at a potential range, where no apparent faradaic process occurred. [10] The anodic charging currents measured at 0.975 V vs. RHE were plotted as a function of the scan rate and from the slope, and the double layer capacitance Cdl was attained. [8] The ECSA of the catalysts were then calculated using the equation ECSA = Cdl / Cs where Cs can be defined as the specific capacitance of the material per unit area under identical electrolyte conditions and a specific capacitance Cs of 1.7 mF cm -2 was utilized for NF substrate in 1 M aqueous KOH solution which is based on the literature reported values. [7d,11] One of the important measurements, the electrochemical impedance spectroscopy (EIS) were recorded at 1.5 V vs. RHE to obtain the Nyquist plots. [12] The amplitude of the sinusoidal wave was examined in a frequency range of 100 kHz to 1 mHz. All impedance spectra were fitted using an equivalent RC circuit model. The charge transfer resistance (Rct) was then obtained from the diameter of the semicircle in the Nyquist plots. Figure S1. The PXRD pattern of (a) as-synthesized and (b) electrophoretically deposited intermetallic MnGa4 film. The powder pattern of MnGa4 matches well with reported cubic MnGa4 (Im-3m, Nr. 229) with lattice parameter a = 5.5961(6) Å, V = 175.06(2) Å 3 and Z = 2. [4,13] The phase confirmation and elemental composition were further identified by HRTEM, EDX, ICP-AES, and XPS studies. The FTO powder pattern is shown in red.    Table S2).  Figure S5. The SEM image (a) and the EDX mapping (b-d) which was carried out on the grounded crystals of MnGa4 to ensure the phase purity of the material. Similar to crystals, the grounded MnGa4 particles exhibited a homogenous distribution of manganese and gallium within the structure without any oxygen (<1) content. Detailed atomic % of the distribution of the elements obtained by EDX is listed in Table S1.    Table S1.    . [16] Similarly, in the case of manganese-based materials, the oxidation state of Mn can also be deduced from their Mn 3p spectra, and it is typical to measure ~47.5 eV for Mn II , ~48.5 eV for Mn III and ~50 eV for Mn IV . [16c,17] Therefore, the resulted binding energy value of 50.2 eV in Mn 3p spectra could be assigned to Mn IV (b). In the case of Ga 2p, the binding energy of 1116.4 eV attained for Ga 2p3/2 is very similar to the binding energy of elemental Ga (1116.4 eV). [18] The second peak observed at the binding energy of 1118.2 eV could be corroborated with Ga bonded to oxo-species, confirming the unavoidable surface oxidation of the intermetallic phase, which is a common phenomenon and often observed in chalcogenides, pnictogenide or intermetallic compounds. [5,7d,19] The surface oxidation was also confirmed from O1s XPS spectra where two deconvoluted peaks (I and II) at 531.3 eV and 533.8 eV can directly be ascribed to surface hydroxylation and water adsorbed onto the surface. [5,7d,19b-e,20] Table S1.     (I, II and III) where the small peak at ~529 is due to the formation of oxide species while the peaks at 530.8 eV and a broad peak at 531.8 eV can be ascribed to hydroxylation. [5,7d,19b-e,20]

Calculation of Faradaic efficiency
The Faradaic efficiency (FE) of MnGa4 in 1M KOH towards oxygen evolution reaction was measured with MnGa4 on nickel foam in a closed electrochemical cell. The cell and the electrolyte were first degassed with Argon for 30 min under stirring. Afterward, the constant current density of 10 mAcm -2 was applied for a certain period. At the end of electrolysis, the gaseous samples were taken out of the headspace by a gas-tight syringe and analyzed by a GC calibrated for O2. Every injection step was repeated at least three times, and the average value is presented.

The Faradaic efficiency (FE) is calculated based on:
VO2 is the evolved volume of oxygen, F is the Faraday constant (96485.33289 C/mol), Vm is the molar volume of the gas, j is the current density (10 mAcm -2 ), and t is the period of electrolysis.    Although the manganese (b) and oxygen (d) were homogeneously distributed within the particles, the gallium (c) atoms mostly disappeared from the structure indicating heavy corrosion of the particles under OER conditions. Detailed atomic % of the distribution of the elements obtained by EDX is listed in Table S1.   Figure S40. The FTIR-spectra of MnGa4 and the films after OER CA. After the OER, a broad band was observed at ~3330 cm −1 which is due to stretching vibrations of interlayer water molecules whereas the band at ~1640 cm −1 could be assigned to the bending vibration of H2O and structural OH groups. [26] The bands between 800 and 1400 cm −1 (1372, 996 and 857 cm −1 ) are typically assigned to the bending vibrations of -OH groups bound with Mn atoms. [27] Bands around ~700 cm −1 and lower are characteristic bands of manganese oxides. [26b,28]  Similarly, as shown for as prepared MnGa4, the Mn 3p spectra showed a broad peak at a binding energy value of 50.2 eV that could also be assigned to Mn IV (b). [16c,17] In the case of Ga 2p, the peaks responsible for Ga were absent, which demonstrates the massive loss of Ga from the surface of MnGa4 under in situ OER conditions with the transformation of the initial structure. O1s XPS spectra further confirmed this. The O 1s spectrum was deconvoluted into three (I, II and III) peaks, the first at ~ 529 is due to the formation of oxide phase whereas the peak at the binding energy of 531.5 eV and 532.8 eV can directly be ascribed to hydroxylated (-OH/-OOH) and adsorbed water onto the surface. [5,7d,19b-e,20] Figure S35) the transformation was somewhat faster. At higher currents, a complete conversion of MnGa4 into δ-MnO2 was attained. This is a strong indication that the conversion of MnGa4 takes via the -Mn3O4 through -MnOOH and finally forming most stable and active δ-MnO2, and the conversion rate is dependent on the applied potentials.  The spectra exhibit drastic change than the original MnGa4 structure. The mapping showed a slight loss of gallium (c) from the structure with the inclusion of oxygen (d) in the structure. This indicates under the prolonged electrolysis, the loss of Ga from the materials takes place, which is likely to go deeper beyond the particle surface and completely transforms the original MnGa4 structure to manganese hydroxides. Detailed atomic % of the distribution of the elements obtained by EDX is listed in Table S1.     No nickel contained particles were observed, which was also confirmed by the EDX mapping (see Figure S51). The SAED pattern indicated crystalline particles with diffraction rings associated with δ-MnO2 (green indices), MnOOH (orange indices) and α-Mn3O4 (blue indices), respectively.   Table 4.    The spectra exhibit a change in the structure where the manganese (b) and oxygen (c) were homogeneously distributed within the particles. Detailed atomic % of the distribution of the elements obtained by EDX is listed in Table S1.    The Mn 3p spectra were deconvoluted into two peaks. The peak at the binding energy of ~48.5 eV corresponds to the oxidation state of Mn III while the peak at 50 eV is due to the Mn IV . [16c,17] The O1s XPS spectra could be deconvoluted into three peaks (I, II and III) where the small peak at 529.5 is due to the formation of manganese oxide (MnOx) whereas the peaks at 530.8 eV and a broad peak at ~533 eV can be ascribed to hydroxylation (-OH/-OOH) and adsorbed water onto the surface. [5,7d,19b-e,20] Figure S64. The crystal structure of (a) (K-) birnessite δ-MnO2, (b) feitknechtite -MnOOH and (c) hausmannite -Mn3O4. The birnessite δ-MnO2 (a) crystallizes in the monoclinic system (space group A2/m) and contain corner-sharing Mn octahedral layers (chartreuse octahedra) with a typical interlayer distance of ~7 Å occupied by interstitial disordered water/cationic sites (red and gray spheres). [10a,29] Some of the Mn cations within the MnO2 layer are reduced from Mn IV to Mn III . The Mn III ions are situated above or below the interlayer bonded through cornersharing bridges. [29a,30] The feitknechtite -MnOOH (b) belongs to the hexagonal system (space group P-3m) and has a layered structure where Mn is trivalent (blue octahedra), and one-half of the O atoms are replaced by hydroxyl anions (purple spheres). [31] The hausmannite -Mn3O4 is a tetragonal (space group I41/amd) spinel with the general formula AB2O4 (Mn II Mn2 III O4) structure in which the Mn atoms are placed among the tetrahedral (A) and octahedral (B) sites. [6] The octahedral sites (green octahedra) are close to each other sharing edges, but tetrahedral sites (blue octahedra) share only corners with the octahedral sites. It has been previously shown that -MnOOH is a metastable phase and is an intermediate phase between -Mn3O4 and δ-MnO2. [32] Recent computational studies have also confirmed the formation of the nascent δ-MnO2 layer produced in situ from spinel -Mn3O4 under electrochemical conditions. [30c] Therefore, from our experimental findings, it is reasonable to expect that the MnGa4 phase was electroconverted first to -Mn3O4 and then to δ-MnO2 through the intermediate -MnOOH. Although it is expected δ-MnO2 to be the active phase for durable water oxidation, however, the contribution of -Mn3O4 and -MnOOH for water oxidation during transformation cannot be ruled out completely.