Low-temperature synthesis and investigation into the formation mechanism of high quality Ni-Fe layered double hydroxides hexagonal platelets

This paper describes the wet-chemistry synthesis of highly crystalline hexagonal flakes of Ni-Fe layered double hydroxide (LDH) produced at temperature as low as 100 °C. The flakes with diameter in the range of 0.5–1.5 μm and the thickness between 15 and 20 nm were obtained by homogeneous precipitation method with the use of triethanolamine (TEA) and urea. By analyzing the intermediate products, it is suggested that, differently from previous reports, a thermodynamically metastable iron oxyhydroxide and Ni-TEA complex are firstly formed at room temperature. Subsequently, when the mixture is heated to 100 °C and the pH increases due to the thermal decomposition of urea, Ni2+ and Fe3+ are slowly released and then recombine, thus leading to formation of pure, highly-crystalline Ni-Fe LDH flakes. This material showed promising results as an electrocatalyst in oxygen evolution reaction (OER) providing an overpotential value of 0.36 V.

of Fe 2+ ions in the reaction mixture, together with other elements (such as Mn, C, S) used to make steel balls, might contaminate the LDH phase. Finally, the topochemical route relies on an initial precipitation of Fe 2+ and its subsequent oxidation to Fe 3+ using iodine 10 or anthraquinone-2-sulfonate 10,15 . This additional synthesis step led to the formation of regular hexagonal platelets 15 , however Fe 2+ ions and oxidizing agents are possible sources of contamination. Recently, in order to overcome these issues and produce regular platelets of Ni-Fe LDH, special attention has been placed on the use of capping agents, such as trisodium citrate or triethanolamine (TEA), which coordinate Fe 3+ , hence preventing the formation of its hydroxide while simultaneously facilitating the combination with Ni 2+ 11,12 .
The synthesis of Ni-Fe LDH with the use of TEA follows the so-called "atrane route" 9 . In a simple way, the formation of Fe-TEA complexes, which are inert to hydrolysis, prevents the precipitation of insoluble iron hydroxides. This inertness is controlled by pH and temperature, so when both parameters are increased 4 , Fe-TEA complex decomposes and Fe 3+ ions recombine with other metal ions present in the solution to form mixed-metal LDH 9 . However, the hydrolysis of this atrane complex takes place at temperature exceeding the boiling point of the solvent (typically 150 °C) 3,11 , which, in case of wet-chemistry synthesis, involves the use of autoclaves. In general, the scalability of the hydrothermal methods is limited by the size of a pressure vessel and the total synthesis costs are significantly higher, which results from special equipment and high temperature requirements 11 .
In this work, pure, highly crystalline Ni-Fe LDH hexagonal platelets were produced by homogeneous precipitation at a relatively low temperature of 100 °C. In contrast to previous reports 3,9 , we used a large excess of the capping agent in relation to Fe 3+ ions which caused the precipitation of a metastable solid iron(III) oxyhydroxide at room temperature, while nickel(II) remained in the solution coordinated by TEA molecules. Then, upon heating and pH increase, the reaction intermediates decompose and recombine forming Ni-Fe LDH hexagons. To the best of our knowledge, the transformation of iron(III) oxyhydroxide to LDH phase at the temperature as low as 100 °C has not been previously reported in literature. Moreover, this process was carried out in a conventional round-bottom flask, suggesting upscaling possibilities of the procedure. Preliminary electrochemical measurements suggest that this material might be potentially used as a catalyst for OER.

Results and Discussion
The morphological and structural characterization of Ni-Fe LDH hexagonal platelets. The morphological and structural characterization of the as-prepared LDH was carried by SEM, TEM, AFM, FT-IR and TGA analysis. SEM micrograph (Fig. 1a) revealed that the as synthesized sample is composed of flakes with a well-defined hexagonal shape and on the order of 1 µm in lateral dimension. The height profiles of the flakes, measured by AFM (Fig. 2), are in the range of 15 to 20 nm, which roughly corresponds to 18-25 layers.
The homogenous contrast observed in the TEM micrographs ( Fig. 1b) suggests that the flakes have a uniform thickness and the SAED pattern (Fig. 1c) shows a hexagonal symmetry (a = 3.08 Å, c = 23.55 Å), which is confirmed by XRD patterns (Fig. 1d). The absence of additional reflections in the SAED pattern indicates that the sample is solely composed by Ni-Fe LDH. This is evidenced by the (101), (011), (112 and (110) lattice planes (Fig. 1c).
The homogeneous distribution of the metal cations within the nanoflakes 15 was confirmed via SEM-EDX mapping ( Fig. 1e-h). Nickel, iron and oxygen are evenly distributed in the flakes surface, thus indicating a consistent substitution of Ni 2+ by Fe 3+ in the brucite-like layers of Ni(OH) 2 .
FT-IR analysis of the flakes (Fig. 3a) revealed the typical spectrum of LDH. Adsorbed water in the samples leads to the broad signal at 3450 cm −1, which is characteristic of the O-H stretching mode in brucite-like layers. , while the weak peaks at 1155 cm −1 and 1040 cm −1 (green labels) are characteristic of the C-N stretching mode in tertiary amines 20 , suggesting a small contamination with TEA.
LDH are known to thermally transform into mix-metal oxide and spinel phases 21 . The TGA curve (Fig. 3b) indicates that this process occurs in two steps 11 . Firstly, in the temperature range of 30 °C to 200 °C, the adsorbed and interlamellar water is removed, resulting in a 10% mass loss. Secondly, at temperatures of 200 to 800 °C, the weight loss is promoted by the dehydroxylation of the brucite-like layers and the decomposition of the counter anions 17 . The total mass loss of roughly 30% is consistent with other works reported on the thermal decomposition of LDH materials 11,14,15 . Finally, Ni 2+ /Fe 3+ ratio was calculated as 3.54 by atomic absorption spectroscopy which, in combination with TGA results, led to the formula Ni 0.78 Fe 0. 22 Electrocatalytic activity in OER. The performance of the Ni 0.78 Fe 0.22 (OH) 2 (CO 3 ) 0.11 •0.6H 2 O as an electrocatalyst for OER was tested by linear sweep voltammetry polarization (Fig. 4).
As expected the synthesized Ni-Fe LDH exhibits electrocatalytic activity towards OER, presenting an overpotential of 0.36 V, which is similar to other nanostructured Ni-Fe LDH materials 2,[22][23][24][25][26] . The overpotential of the support (Ni-foam) was also measured resulting in a value of 0.43 V. The cyclic voltammograms shown in the inset of Fig. 4 reveal clear anodic and cathodic peaks which are associated with the redox couple Ni II /Ni III . 2 This is a typical behaviour of Ni-Fe LDH and it has origin in the insertion/desertion of OH − ions from the LDH structure. A similar response was observed for the bare Ni foam (Fig. S1 in ESI †).
These preliminaries results indicate that the two-dimensional Ni-Fe LDH hexagons described in this paper might represent a viable option for OER applications. Therefore, further characterization and optimization of the LDH hexagons will be considered for future work.
Possible formation mechanism of Ni-Fe LDH platelets. Results presented so far indicate that homogeneous precipitation at low temperature is a very interesting technique for the synthesis of crystalline LDH hexagons with promising electrocatalytic properties. However, the processes behind this approach are not fully understood. Further experimental work allowed us to come forth with a hypothetical model for the LDH formation by homogeneous precipitation at low temperature. First, the addition of TEA to the reaction mixture drastically changes the pH of the starting solution (from 2.5 to 6.6). At these conditions, a gel-like brown precipitate  was formed at room temperature (Fig. 5a), which upon heating to 100 °C and pH increase transformed to Ni-Fe LDH hexagonal platelets. In order to try to clarify how the final product is produced we split the synthesis process between two steps: at room temperature and at 100 °C.
Step 1-Room Temperature. The formation of the iron oxyhydroxide precipitate at room temperature was investigated by UV-Vis measurements performed at 350 nm, which is linearly related to the Fe 3+ concentration (Ni 2+ has a negligible absorbance at that wavelength, Fig. S2 in ESI †). Before recording the UV-Vis spectra the reaction mixture was centrifuged (3000 rpm/10 min.) in order to remove the precipitate and monitor the Fe 3+ concentration in the solution. As presented in the Fig. S3 in ESI †, the formation of the precipitate has an induction period of 2 hours, at which the Fe 3+ concentration decreases slightly, followed by an exponential decay, which results in negligible concentration of Fe 3+ in solution after 9 hours, whilst maintaining a constant pH of 6.6. This  experiment led us to find out the minimum time required to precipitate all Fe 3+ from the starting solution, which is approximately 9 hours.
The precipitate formed after 24 hours of stirring was isolated, washed several times with water, dried at room temperature and analyzed by FT-IR, XRD and TEM (Fig. 5b-d).
In the FT-IR spectrum (Fig. 5b) a broad adsorption band with a maximum at 3270 cm −1 is attributed to υ O-H stretching mode, which is typical for hydroxides and also to structural/adsorbed water. The peak at 1640 cm −1 additionally confirms the presence of water in the precipitate. Furthermore, the peaks at 1470 cm −1 and 1360 cm −1 , assigned to the Fe-O and Fe-OH vibrations, respectively, suggest that the precipitate is most probably one of the forms of ferric oxyhydroxide 27,28 . The very strong absorption at wavenumber <1000 cm −1 can be explained by the deformation vibrations of surface O-H groups in ferric hydroxides 28 . The weak absorption bands at the region of 1200-1000 cm −1 might be related to the partial complexation of Fe 3+ by TEA molecules 29 .
The XRD pattern of the precipitate (Fig. 5c) shows that the material presents a low degree of crystallinity, as only two broad peaks at about 2.6 Å and 1.5 Å are detected, which are typical for iron oxyhydroxides previously reported [30][31][32] . Additional characterization by TEM and SAED (Fig. 5d) demonstrate that the material presents an irregular morphology and the particles are strongly agglomerated. The SAED (inset in Fig. 5d) highlights only two bright, diffused rings which positions fit well to the pattern obtained by XRD. TEM results are in a good agreement with those found in the literature suggesting that the precipitate formed at room temperature is most probably one of the forms of ferrihydrite 33 . This thermodynamically metastable Fe 3+ oxyhydroxides finally transform to more crystalline forms depending on the temperature and pH 34,35 . Considering that those parameters are both changing during the further thermal decomposition of urea, we hypothesize that, in presence of Ni 2+ and CO 3 2− , the ferrihydrite transforms into Ni-Fe LDH. Since most of the Fe 3+ precipitates as ferrihydrite, the remaining solution is expected to be composed of Ni 2+ , NO 3 − , TEA and urea. In order to understand the coordination of Ni 2+ , the reaction mixture was analyzed by NMR and UV-Vis spectroscopy and compared to the spectra of pure Ni(NO 3 ) 2 and TEA (Fig. 6).
Firstly, the reaction mixture was characterized by 1 H and 13 C NMR spectroscopy and compared to the spectrum of pure TEA (Fig. 6a,b). The changes in both 1 H and 13 C environment are clearly visible: 1 H NMR spectrum of pure TEA shows two triplets for the methylene groups bonded to N and O, respectively, which are observed to be broadened and shifted downfield in the reaction mixture due to the presence of Ni 2+ . This is in good agreement with the theoretical predictions, suggesting that the loss of electron-density upon coordination to the metal results in a shift to higher frequency of the protons adjacent to the ligand 36 . In contrast, in the 13 C NMR spectrum the signals are shifted upfield, which was also previously observed in metal-organic compounds 37 , which is an additional proof of the formation Ni 2+ -TEA complex. Several coordination compounds of Ni-TEA-NO 3 − have been described in the form of blue crystals 38,39 , which differs from the clear, pale green/blue solution obtained in the present study. This difference is most probably the result of an insufficient amount of TEA in relation to Ni 2+ , since the TEA:Ni 2+ ratio is 4:3 instead of 2:1 to form [Ni(TEA) 2 ](NO 3 ) 2 in a solid form. The UV-Vis absorption spectra of both Ni(NO 3 ) 2 and the reaction mixture (Fig. 6c) show two weak bands assigned to d-d transitions corresponding to an octahedral coordination geometry of Ni 2+ 38,40,41 . A sharp peak centred at 390 nm can be assigned to 3 A 2g → 3 T 1g (P) transition in both cases 38,40 , while the absorbance is higher in case of the reaction mixture, which probably is a result of metal coordination by TEA ligands. Moreover, the doublet with maxima at 656 nm and 726 nm ascribed to 3 A 2g → 3 T 1g (F) transition typical of nickel(II) aqua complex ([Ni(H 2 O) 6 ] 2+ ) 41,42 is slightly shifted towards lower wavelengths in the reaction mixture. Higher absorbance of the peak at 390 nm and a blue shift at 656-720 nm were previously observed by Agarwala et al. 42  Therefore, on the basis of UV-Vis and NMR spectroscopy, nickel(II) is most probably present in the reaction mixture as an octahedral Ni-TEA metal complex, however the chemical formula cannot be accurately defined, due to many possibilities of ligand coordination 42,43 , and most probably it consists as a mixture of different Ni-TEA derivatives in quick kinetic equilibrium.
Step 2-Heating to 100 °C. The heating of the reaction mixture (gel-like ferrihydrite and Ni 2+ -TEA complex) to 100 °C was monitored by UV-Vis measurements (Fig. 6d). Within the first 4 hours of heating the mixture, a dramatic raise of Fe 3+ concentration was observed, suggesting the decomposition of ferrihydrite. The Fe 3+ concentration augmented for a maximum of 8 hours and started decreasing until the end of the measurements. Nevertheless, after 48 hours of heating the reaction mixture, some unbounded Fe 3+ ions remained in the solution, as observed in Fig. 6d. This is consistent with the Ni 2+ /Fe 3+ ratio in the final product (3.54), which is slightly higher than initially used (3.00) suggesting that not all Fe 3+ was incorporated to the Ni-Fe LDH.
The evolution of the pH is also plotted in Fig. 6d. During the first 12 hours, a linear increase of the pH is observed due to the urea hydrolysis at 100 °C. After this time, the consumption of OH − by metal cations leads to a slower pH increase. Carbonate anions produced from urea decomposition intercalate between the layers to neutralize a positive charge generated by Fe 3+ substitution within Ni(OH) 2 .
The precipitates formed after 4, 12 and 24 hours of heating were analysed by SEM (Fig. S4 in ESI †), in order to investigate the platelets formation process. After 4 and 12 hours, hexagonal flakes with slightly jagged edges and diameters around 0.5 μm mixed with an amorphous precipitate (most probably, remaining ferrihydrite) were observed. Further heating for an additional 12 hours lead to the complete consumption of the amorphous precipitate and only hexagonal platelets remained. Nevertheless, the heating was continued for additional 24 hours, in order to maximize the incorporation of Fe 3+ to the LDH.

Conclusion
In summary, we obtained Ni-Fe LDH hexagonal platelets by a homogeneous precipitation method at 100 °C under atmospheric pressure. The flakes' diameter is in the range of 0.5-1.5 μm and thicknesses between 15 and 20 nm. This material is highly crystalline and the metal cations are homogeneously distributed within a single flake. We believe that those platelets are formed in a two step process, where firstly reaction intermediates (iron oxyhydroxide and Ni-TEA complex mixture) are formed at room temperature, and then, upon heating to 100 °C, Ni 2+ and Fe 3+ are released from their precursors and react with OH − and CO 3 2− (from urea hydrolysis) forming regular hexagons of Ni-Fe LDH. The preliminary results suggested that the obtained material showed favourable electrocatalytic activity towards OER, exhibiting an overpotential of 0.36 V vs RHE, which is consistent with literature reports.
We have demonstrated in this paper that high quality Fe 3+ -containing LDH platelets can be produced by a simple method at 100 °C in a conventional synthesis setup.

Methods
Chemicals. Nickel in 80 ml of deionized water to a final concentration of 7.5, 2.5 and 17.5 mM, respectively. Then, 0.8 mmol of TEA were added and the solution was stirred at room temperature for 24 hours, which led to the formation of gel-like brown precipitate. After that, the reaction mixture was transferred to 100 ml round-bottom flask and immersed in an oil bath previously heated to 100 °C under reflux conditions and heated for 48 hours. The flask was naturally cooled down to room temperature and the obtained precipitate was collected by centrifugation (3000 rpm/ 10 min) and washed with water several times. Equipment and characterization techniques. Scanning electron microscopy (SEM) images were acquired using a Zeiss Utra Plus (Carl Zeiss AG, Germany) operated at 2-3 kV, while energy dispersive X-ray spectroscopy (EDX) mapping was performed at 15 kV. Transmission electron microscopy (TEM) and selected area electron diffraction (SAED) were conducted using an FEI Titan (FEI, Oregon, USA) microscope operated at 80 keV. Atomic force microscopy (AFM) images were taken on an Asylum Research MFP 3D microscope working in a tapping mode. Fourier-transform infrared (FT-IR) spectra were recorded using a Perkin Elmer Spectrum 100 FT-IR spectrometer via attenuated total reflectance (ATR) method. Powder X-ray diffraction (XRD) was measured in a Bruker Advance Powder X-ray diffractometer equipped with a Mo-Kα emission source (λ = 0.7107 Å) in the Bragg-Brentano configuration. The metals content was measured with a Varian 55 Atomic Absorption spectrometer. Thermogravimetric analysis (TGA) was carried out in a Perkin Elmer Pyris 1 TGA (the samples were heated up to 850 °C at a rate of 10 °C min −1 ) under air atmosphere. Ultraviolet-visible (UV-Vis) spectroscopy measurements were conducted using a Biochrom Libra S22 UV/Vis Spectrophotometer. 1 H and 13 C nuclear magnetic resonance (NMR) analyses were performed using a Bruker Avance HD 400 NMR equipped with a BBFO probe. An ultrasonic spray deposition (USI Prism Ultracoat 300) was used to deposit the sample dispersion onto nickel foam substrates (0.25 cm 2 ) and prepare electrodes with an average mass load of 0.3 mg•cm −2 . Electrochemical measurements were conducted in a BioLogic VMP 300. A platinum wire and Ag/AgCl were used as counter and reference electrodes in 1 M KOH electrolyte solution, respectively. Both cyclic and linear sweep voltammetry curves were acquired in the range of 0 V to 0.6 V at a scan rate of 5 mV•s −1 against Ag/AgCl. All the potentials were calibrated with respect to a reversible hydrogen electrode (RHE) reference as follows: 2 V RHE = V Ag/ AgCl + V Ag/AgCl(vs RHE) + (0.059 × pH). All voltammetry curves shown are compensated for solution resistance.