Clearing Up Discrepancies in 2D and 3D Nickel Molybdate Hydrate Structures

When electrocatalysts are prepared, modification of the morphology is a common strategy to enhance their electrocatalytic performance. In this work, we have examined and characterized nanorods (3D) and nanosheets (2D) of nickel molybdate hydrates, which previously have been treated as the same material with just a variation in morphology. We thoroughly investigated the materials and report that they contain fundamentally different compounds with different crystal structures, chemical compositions, and chemical stabilities. The 3D nanorod structure exhibits the chemical formula NiMoO4·0.6H2O and crystallizes in a triclinic system, whereas the 2D nanosheet structures can be rationalized with Ni3MoO5–0.5x(OH)x·(2.3 – 0.5x)H2O, with a mixed valence of both Ni and Mo, which enables a layered crystal structure. The difference in structure and composition is supported by X-ray photoelectron spectroscopy, ion beam analysis, thermogravimetric analysis, X-ray diffraction, electron diffraction, infrared spectroscopy, Raman spectroscopy, and magnetic measurements. The previously proposed crystal structure for the nickel molybdate hydrate nanorods from the literature needs to be reconsidered and is here refined by ab initio molecular dynamics on a quantum mechanical level using density functional theory calculations to reproduce the experimental findings. Because the material is frequently studied as an electrocatalyst or catalyst precursor and both structures can appear in the same synthesis, a clear distinction between the two compounds is necessary to assess the underlying structure-to-function relationship and targeted electrocatalytic properties.


Synthesis
Nickel molybdate hydrate nanorods (NMO-H2O-rods) were synthesized in a hydrothermal synthesis.In a representative synthesis 0.01 M (NH4)6Mo7O24 • 4H2O (AHM) and 0.07 M Ni(NO3)2 • 6H2O was dissolved in 60 mL deionized water (DI).These concentrations were chosen to provide a 1:1 molar ratio between molybdenum and nickel, as for the 2D nanosheet synthesis below.The solution was transferred into a 100 mL Teflon lined stainless steel autoclave and heated up in a muffle furnace to 150 °C with 2 °C min −1 and hold at this temperature for 6 h.After cooling to room temperature, the yellow precipitate was washed with DI three times followed by one time with ethanol, collected each time by centrifugation and finally dried at 60 °C.It should be noted that the reaction solution after the synthesis was still clear green.In order to remove lattice water, the dry precipitate was further hold at 200 °C for 8 h.
For nickel molybdate hydrate nanosheets (NMO-H2O-sheets) the synthesis route of Chen et al. was mostly followed. 1Herein 2 mmol Na2MoO4 • 2H2O and 2 mmol Ni(NO3)2 • 6H2O was dissolved in 60 mL (DI).Then 8 mmol urea was added.After stirring for 30 minutes the solution was transferred into a 100 mL Teflon lined stainless steel autoclave and heated up in a muffle furnace to 160 °C and hold for 8 h.Since no heating ramp was reported in the original paper, a rate of 10 °C min −1 was applied.After synthesis the solution was clear and colorless with bright green precipitate.Also, those precipitates were washed, collected and dried as the NMO-H2O-rods above.

Characterization
Scanning Electron Microscopy (SEM) and Energy Dispersive X-ray Spectroscopy (EDX).The nanostructures were analyzed with a high-resolution SEM (ZEISS 1530) with a Schottky field emission gun and acceleration voltages of 15 -20 kV.The secondary electrons were detected with an Inlens detectors.The software used was SmartSEM (Version 5.07).EDX were detected with an Oxford Instruments X-Max N detector.The software AZTec was used for acquiring and analyzing the detected X-rays.
X-Ray Photoelectron Spectroscopy (XPS).XPS analysis was performed with a Physical Electronics PHI Quantera II Scanning XPS Microprobe.Monochromatic Al Kα X-rays with 1486.6 eV and acceleration voltage of 15 kV were used.For all survey spectra the sample was illuminated with a 200 µm diameter X-ray beam with a power of 50 W.A pass energy of 224.00 eV with an acquisition time of 50 ms per step and a step size of 0.8 eV.For the high-resolution spectra, an X-ray beam diameter of 100 µm was used with a pass energy of 55.00 eV, an acquisition time of 50 ms per step and a step size of 0.1 eV.For the NMO-H2O-rods the acquisition on Ni, Mo, O and C was repeated 20, 10, 10 and 10 times, respectively.For the NMO-H2O-sheets the acquisition on Ni, Mo, O and C was repeated 10, 20, 10 and 10 times, respectively.All spectra were analyzed with CasaXPS software (Version 2.3.24). 2 For elemental analysis a Shirley background correction was used.Prior to analysis, all spectra were charge corrected versus adventitious carbon at 284.8 eV binding energy.
Powder X-ray Diffraction (PXRD).PXRD was done using a Bruker D8 Advance diffractometer with a Cu Kα radiation in a Bragg-Brentano geometry and a solid state rapid LynxEye detector.The software used was DIFFRACT plus software and the patterns acquired at room temperature between 5° -90° 2θ with at step size of 0.013° and a time per step of 4 seconds.The collected data were analyzed using HighScore Plus 3.0 software from PANalytical.
Transmission Electron Microscopy (TEM).TEM analysis was conducted with a JEOL JEM-2100F with a Schottky-type field emission gun and an acceleration voltage of 200 kV.Bright Field images and Selected Area Electron Diffraction patterns were taken with a Gatan Ultrascan 100 and Orius 200 D camera, respectively.The software for acquisition was Gatan Microscopy Suite while for analysis ImageJ (Version 1.53) was used.ImageJ was used for post processing the TEM images. 3or three-dimensional electron diffraction (3D ED) the single crystal samples were first separated by ultrasonication in ethanol.The solution was drop casted on a copper TEM grid and analyzed in a ThermoFisher Themis Z microscope operating with an acceleration voltage of 300 kV.The diffraction pattern was detected with a Gatan Oneview camera.A camera length of 360 mm and a spot size of 3 were used.A self-developed DigitalMicrograph plugin was used to control the data collection. 4Diffraction data were acquired by continuously rotating the sample stage with the crystal at a rate of 1.47 ° s −1 .The acquisition time was set to 0.125 s, meaning the individual diffraction images were integrated over 0.18 ° of reciprocal space.Several data sets from different single crystals were collected by 110 ° rotation angle.For the analysis the cRED processing software REDp was utilized to determine the unit cell parameters and the space group based on the symmetry and reflection conditions of the collected diffraction pattern. 5The data were processed with the crystallography software XDS. 6The obtained data sets were merged together according to their cross correlation in order to improve the level of completion and to obtain a single data set suitable for structure solution and refinement.For solving the structure according to their lowest cost function, the software SHELXT and SHELXLE were used. 7,8For illustration of the crystal structure VESTA Version 3.5.2was used. 9ro Field Cooled and Field Cooled (ZFC-FC) and Isothermal Field-dependent Magnetization Loops.Magnetization measurements were carried out using a MPMS XL SQUID magnetometer equipped with a superconducting magnet (up to ±50 kOe) from Quantum Design, Inc.The sample was immobilized in parafilm in a capsule to prevent any movement of the particles during the measurements.The thermal dependence of magnetization was measured according to the zerofield cooled (ZFC) and field cooled (FC) protocols.ZFC and FC measurements were carried out as follows: first, the sample was cooled down from 300 K to 5 K with no magnetic field applied.Then a static magnetic field of 100 Oe was applied and the magnetization was measured during warming up from 5 K to 300 K. Finally, the sample was cooled down to 5 K under the same magnetic field and the magnetization was measured during the cooling.Isothermal fielddependent magnetization loops were recorded at 5 K and 300 K by sweeping the field between −50 kOe and 50 kOe.The obtained magnetization values were normalized by the weight of the powder present in the sample (≈ 10 mg).
Rutherford Backscattering Spectrometry (RBS).For RBS a 2 MeV 4 He + beam was hitting the sample at an incident angle of 5 ° with respect to the surface normal.The sample was a pellet (d = 7 mm) of the of the NMO-H2O nanostructures powder compressed with approximately 2 t.A passivated implanted planar silicon (PIPS) energy detector for scattered particles was placed at a backscattering angle of 170°.Outgoing particle trajectories were inclined 5 ° with respect to the sample surface normal.The analysis was conducted with the software Simnra (Version 7.03). 10me-of-Flight Elastic Recoil Detection Analysis (ToF-ERDA).For ToF-ERDA a beam of 44 MeV 127 I 10+ ions were used to bombard the compressed pellet previously analyzed with RBS.The incident beam hit the sample at an angle of (23 ± 1)° to the sample surface.The detector was positioned at an angle of 45 ° with regard to the forward beam direction. 11For detection an electron mirror flight time detector was employed together with an energy resolving gas ionization chamber.For the data analysis the software Potku, developed at the University of Jyväskylä, was employed. 12ermogravimetric Analysis (TGA).TGA was conducted with a TA Instruments TGA Q500 instrument in air from room temperature to 600 °C with 2 °C min −1 on a platinum pan.Data were collected and analyzed with Universal Analysis 2000 (Version 4.5A).
Raman Spectroscopy.Most Raman spectra were acquired with Renishaw Reflex (Invia) Raman spectrometer.In this spectrometer a frequency doubled Nd:YAG 532 nm laser with a grating of 2400 lines mm −1 and a Renishaw streamline CCD 1024 chip detector were used.The software used for acquisition was WiRE 3.4.For the analysis with 785 nm laser excitation a Renishaw Qontor (Invia) Raman Spectrometer with the Renishaw HPNIR 785 semiconductor laser source, a grating of 1200 lines mm −1 and a Centrus detector was employed.The software used was WiRE 5.5.Prior to all experiments, the Raman spectrometer was calibrated with a silicon reference to (520.5 ± 0.2) cm −1 .All spectra were taken with a 20× magnification lens.

Attenuated Total Reflection Fourier-Transform Infrared Spectroscopy (ATR-FTIR). ATR-FTIR was
conducted with a Bruker Vertex 70v spectrometer with a A225/Q Platinum ATR diamond unit.For the background and sample 16 scans were acquired from 4500 cm −1 to 200 cm −1 with a resolution of 2 cm −1 .The software used for acquisition was Opus (Version 8.2.21).Intensity (a.u.)

Ni 2p
Ni   In the last column labeled "Corr.At.%", the likely oxygen contamination from the Leit tab was compensated, assuming all detected carbon originated from the Leit tab.Previous measurements have shown that without the Leit tab there is only a very small amount of adventitious carbon present, so the real oxygen content may be in between the presented ones.
Ion Beam Analysis (IBA) of NMO-H 2 O nanostructures  The H/Ni ratio shows a clear decay over the total ion fluence for NMO-H2O-sheets.The datapoints were fitted according to following expression 13 with () being the hydrogen to nickel ratio at the total ion fluence ,   and   the initial and final hydrogen to nickel ratio, respectively and  the decay constant.The compensation factor for the hydrogen content is calculated with 14 with  being the compensation factor for the hydrogen content and () ̅̅̅̅̅̅ the average hydrogen to nickel ratio over the whole ion fluence.The atomic concentrations detected by ToF-ERDA depth profile are corrected before the nickel to molybdenum ratio refinement by RBS.And hence The amount of reversibly bond water is calculated the following way: To estimate the amount of reversibly bond water two temperatures at the onset of the removal of crystal water are chosen from TGA.
Weight percent at 180 °C: 99.12 wt.% Weight percent at 200 °C: 98.87 wt.% Solving the equation (S6) for  RW gives the amount of reversibly bond water and subsequently of crystal water.To probe a possible removal of crystal water during the drying process at 200 °C for 8 hours TGA was conducted before and after drying.The percental loss of weight after removal of all water compared to before the removal of crystal water for the two NMO-H2O-rods samples is (4.69 ± 0.13) % and (4.71 ± 0.21) % for the dried and not dried sample, respectively.This indicates that no significant amount of crystal water was removed during the drying process.
All the calculations above assumed that the IBA concentration represents NMO-H2O with reversibly bond water.However, it is uncertain how much reversibly bond water was present during the analysis, hence another stoichiometry is calculated assuming no reversibly bond water in the IBA according to In which . % w.o.CW represents the measured weight percent after removal of crystal water (550 °C, 600 °C) and .% w.CW represents the measured weight percent with only crystal water (180 °C, 200 °C).
For NMO-H2O-sheets an additional contribution of the possible presence of hydroxide has to be added.
Assuming that hydroxides are present and during the TGA dehydrate according to: (S8) Hence only 0.5 of the original oxygen content from the hydroxide is present after heating.
The equation for the calculation of evaporated water is modified to: The percental loss of weight after removal of all water compared to before the removal of crystal water and hydroxides for the two NMO-H2O-sheets samples is (10.43 ± 0.07) % and (10.88 ± 0.12) % for the dried and not dried sample, respectively.This indicates that no significant amount of crystal water or hydroxides from the crystal were removed during the drying process also for this nanostructure.
In case the elemental concentration detected by IBA represents the material without reversibly bond water ( RW = 0), the following equation was utilized to calculate  O +  The inverse of the magnetic susceptibility χ −1 was plotted over the temperature.At high temperatures the magnetic susceptibility in the paramagnetic regime follows the characteristic  − behavior (with C as Curie constant, T as temperature and θ as Curie-Weiss temperature), known as the Curie-Weiss law.The linear fit in the range 25 K -180 K led to an effective magnetic momentum derived from the C as µeff = 3.3 µB, with µB as Bohr magneton, and a Curie-Weiss temperature of θ = -5 K and, indicating antiferromagnetic behavior.A fit between 210 K to 280 K resulted in a slightly different magnetic moment µeff = 3.6 µB, which could originate from a Ni 2+ in a tetrahedral ligand field, a magnetic anomaly, the presence of small impurities, or structural changes owing to the water coordination.The inverse of the magnetic susceptibility χ −1 was plotted over the temperature.The linear fit in the range 25 K -280 K led to an effective magnetic moment of µeff = 2.6 µB and a Curie-Weiss temperature of θ = 34 K, indicating ferromagnetic behavior, which likely agrees with the hysteresis loop at low temperature presented in the main work.

Density-Functional Theory (DFT) Calculations
With DFT-calculations the PDF 04-017-0338 crystal structure was optimized (as explained in the computational section).The global optimization leads to the crystallographic data below as the ground state structure.

Figure S8 .
Figure S8.Fit of the relative H/Ni yield over the total ion fluence.(a) NMO-H2O-rods and (b) NMO-H2Osheets.The graph for the sheets shows a clear ion beam induced hydrogen loss.

Figure S11 .
Figure S11.From Figure 4a derived inverse of the magnetic susceptibility χ −1 over temperature.

Figure S12 .
Figure S12.From Figure 4b derived inverse of the magnetic susceptibility χ −1 over temperature.

Figure S13 .Figure S14 .
Figure S13.PXRD data for clarification with the observed data (red circles), the calculated data after refinement with the Le Bail methods (black line), the difference between calculated and observed intensity (blue line), and the angles at which the PDF 04-017-0338 reports reflexes.

Figure S24 .
Figure S24.Comparison of experimental collected data of NMO-H2O-rods (yellow) and the simulated PXRD pattern of the optimized crystal structure (black).

Table S2 .
Calculated amount of total water and its contribution of crystal water and reversibly bond