Scalable fabrication and electrochemical characterization of binder-free Fe₃O₄@C films on a Cu current collector

Abstract

Here we present a scalable strategy for fabrication of nanostructured magnetite (Fe₃O₄) films directly on a copper (Cu) current collector as potential anodes for energy storage devices. Hydrothermal synthesis was employed in the first step to realize Fe₃O₄ films which were subsequently encapsulated in a carbon shell in a controlled way using chemical vapor deposition (CVD) techniques. Through systematic investigations, it was revealed that lower hydrothermal temperature (80 °C) results in rhombus-shaped nanoarchitectures which were primarily composed of α-FeOOH. At higher temperature (120 °C), three-dimensional (3D) superstructures comprising of Fe₃O₄/iron carbonate (FeCO₃) composite nanolayers were found. Furthermore, it was revealed that, irrespective of the composition of the hydrothermally synthesized film, each one converted upon calcination to cubic Fe₃O₄ without any significant changes in the morphology. Conformal carbon encapsulation allowed the formation of Fe₃O₄@C core–shell structures having tunable shell thickness and strong interfacial bonding (Fe–O–C). Although structural investigations showed improvement in crystallinity with carbon encapsulation by both CVD techniques, it was more pronounced in the case of microwave plasma CVD (MP-CVD), which also induced partial reduction of Fe₃O₄ to metallic iron (Fe). The films showed pseudocapacitive behavior in 1 M Na₂SO₄ during electrochemical evaluation, and stored charge at lower scan rates predominantly by a diffusion-controlled process, which decreased with increasing scan rate, leading to a dominant capacitive process for charge storage. The developed binder-free films have great potential as highly robust anodes for lithium and post-lithium ion batteries (LIBs), supercapacitors, and battery–supercapacitor hybrid devices.

---

Introduction

The demand for environmentally responsible and renewable energy has increased significantly, owing to the escalating demand for rapid electrification of transportation; consequently, the demand for high capacity energy storage systems for large scale applications has intensified during the last couple of years.^1–3 Presently, lithium ion batteries (LIBs) are the preferred energy storage devices owing to their high energy storage density and reliability.^3–5 However, the commercially available LIBs of today rely primarily upon electrode materials which suffer from intrinsically low theoretical specific capacities.⁵ For example, the main anode material, i.e., graphite, has a theoretical specific capacity of 372 mAh g⁻¹ which is already achieved.⁶ It is, therefore, unrealistic to look towards these materials for their implementation in LIBs having the potential to fully meet future requirements of the extended range and high power applications.^7–9 Owing to these facts, there is an urgent need to develop alternative anode materials having high specific capacities, low cost, environmental accountability and robustness compared to the conventional graphite-based anodes.¹⁰

Transition metal oxides (TMOs) feature nearly all these characteristics, and therefore have attracted significant research attention. While operating via the conversion reaction mechanism, majority of TMOs have the capability of delivering specific capacities many times higher than those of graphite-based anodes owing to their utilization of multiple oxidation states.^11,12 Among the TMOs, magnetite (Fe₃O₄) is particularly interesting as an electrode material for energy storage devices owing to its high theoretical capacity of ∼920 mAh g⁻¹, environmental benignity, abundance, and cost-effectiveness.¹³ Fe₃O₄ is highly versatile as an energy storage material, showing extended potential as an electrode material not only for LIBs but also in some of the emerging energy storage technologies like post-LIBs and supercapacitors.^14–17

Despite all these advantages, Fe₃O₄ anodes could not pave their way towards large scale implementation in practical energy storage devices owing to their fundamental challenges which are intrinsic to nearly all conversion-type electrode materials. Two of these challenges, i.e., extreme volume changes during conversion and low electrical conductivities, are particularly significant to be considered while designing and developing electrode materials. First, the volume change during metalation/demetalation is massive (>200%), leading to significant mechanical stress which causes active material pulverization, electrical isolation, and continuous fracturing of the solid electrolyte interphase (SEI) layer.^18–20 Second, the intrinsically low electrical conductivity of Fe₃O₄ is responsible for restricting the electron transport kinetics on one hand, and on the other hand, limiting the rate capability.²¹ Additionally, these issues are further aggravated in the traditional slurry based electrode fabrications by the polymer binders and conductive additives as they are responsible for introduction of inactive mass/volume on one hand, and on the other hand, rely on the weak van der Waals forces between the material and the current collector.^14,22,23

The issues related to the binders could be overcome by growing the active materials directly onto the current collector through an effective strategy to realize binder-free electrodes.^24–26 The architectures which are engineered this way help in ensuring robust electrical contact on one hand, and on the other hand, help in eliminating inactive weights of the resulting electrodes.²⁶ Additionally, if included during fabrication, the open porous structure accommodates effectively the strain produced during cycling.²⁶ This strategy has been previously demonstrated and reported, showing improved performance of hydrothermally grown Fe₃O₄ materials directly onto the substrate.^27–29 However, the intrinsic issue of conductivity and pulverization at the particle level could not be resolved through the binder-free approach alone. The remaining challenges could be effectively overcome by encapsulating the Fe₃O₄ particles of the directly deposited materials on the current collector within a conductive carbon shell.^30–32 A viably designed and robustly engineered carbon shell while acting as a physical buffer mitigates the volume changes during cycling on the one hand, and on the other hand, provides continuous conductive pathways.^33–35 Of the various techniques and approaches available and employed for this purpose, chemical vapor deposition (CVD) is quite a unique technique with numerous advantages while coating complex three dimensional (3D) architectures. For example, CVD provides uniform and conformal coatings without compromising the original morphology of the underlying particles as it involves a gas-phase process that can penetrate complex pores.^36–38

However, to the best of our knowledge, there is a lack of systematic investigations of binder-free Fe₃O₄ encapsulation in a carbon shell by CVD, and thus a large research and technology gap exists in this regard. Therefore, in our opinion, it is required to address this gap effectively through a viably designed and robustly executed fabrication strategy. Furthermore, the effect of CVD parameters on the phase, microstructure and morphology of the carbon encapsulated Fe₃O₄ architectures, and thus on their electrochemical performance, is not well understood. Therefore, a systematic probing of the influence of CVD parameters on binder-free Fe₃O₄ material properties and its electrochemical performance is critical.

Numerous synthesis methods have been explored for preparation of morphology-specific nanoarchitectures. However, realization of Fe₃O₄ directly on the current collector in the desired shape is quite challenging, if not impossible. Despite these challenges, the hydrothermal synthesis method has more pros than cons, including control over engineering of the desired architecture at the nanometer-level scale, for achieving this goal.³⁹ It is pertinent to mention here that the shape of nanoparticles is critical in dictating the functionalities of the materials.^40,41 For example, rhombus-shaped one-dimensional (1D) nanostructures showed higher electrochemical performance, which was attributed to their unique architecture and morphology.^42–44 These unique architectures, i.e., rhombus-shaped 1D nanorods for Fe₃O₄, have not been explored to date, to the best of our knowledge, as loose powder, let alone as binder-free electrode materials directly on the current collector.

Here, we present the fabrication of binder free Fe₃O₄ films directly on the current collector in the desired morphologies by the so-called hydrothermal synthesis, and carbon encapsulation of these films by two distinct CVD techniques including thermal CVD (T-CVD) using acetylene (C₂H₂) gas and microwave plasma enhanced CVD (MP-CVD) using methane (CH₄) gas as carbon sources to investigate the effect of carbon encapsulation on the structure and properties of the resulting nanomaterials. The aim was to demonstrate that the selected method of synthesis of Fe₃O₄, i.e., hydrothermal, has the potential to equip us with the ability to tailor the shape of the Fe₃O₄ nanoarchitectures as desired. Furthermore, the aim was to coat the binder-free Fe₃O₄ intricate films without compromising their morphology. Finally, it was aimed to systematically investigate the structure–property relationships by following the fabricated film electrochemically.

Experimental

Materials

All the materials used in this research were purchased from commercial suppliers and used as received unless mentioned otherwise. Copper (Cu) foil (>99.95%) of 0.1 mm thickness was purchased from Carl Roth GmbH and cleaned by ultrasonication for 10 minutes in distilled water (produced in our own laboratory facility) followed by ethanol (>96%, VWR Chemicals). The thoroughly cleaned Cu foil was dried overnight in a laboratory oven at 60 °C. Iron(II) sulfate heptahydrate (FeSO₄·7H₂O) (>99%, Sigma-Aldrich), Na₂SO₄ (>99%, Sigma-Aldrich), urea (99%, ThermoFischer) and ammonium fluoride (NH₄F) (∼96%, Alfa Aesar) were purchased from commercial suppliers and used as received.

Hydrothermal processing

Following the reported procedure,²⁶ the solution for hydrothermal processing of Cu foil was prepared by dissolving 5 mmol of FeSO₄·7H₂O, 15 mmol of urea and 10 mmol of NH₄F in 70 mL of distilled water by stirring at 500 rpm at room temperature. After 10 minutes of stirring, the solution was transferred into a PTFE-lined 125 mL stainless steel autoclave (acid digestion vessel of Parr Instrument (Deutschland) GmbH) along with thoroughly prepared Cu foil which was positioned at 45°. After tightly closing, the vessel was placed in a laboratory oven which was pre-heated to the desired temperature of 80 °C (120 °C), and maintained for 12 hours. Afterwards, the vessel was allowed to cool down to room temperature naturally. The Cu foil was retrieved and rinsed with ethanol and distilled water. The loose powder, which was precipitated in the vessel, was recovered by centrifugation at 6000 rpm and cleaned by washing with distilled water and ethanol. Both Cu foil and the powder were dried overnight at 60 °C in the laboratory oven, and then calcined at 450 °C under a nitrogen (N₂) stream (200 sccm) for 4 hours while the furnace was heated at a rate of 5 °C min⁻¹.

Carbon encapsulation

Carbon encapsulation of the resulting film particles was carried out by CVD techniques including thermal CVD (T-CVD) and MP-CVD. The post-calcination films were inserted into the T-CVD chamber which was then sealed tightly and flushed with dry N₂ for 30 minutes. Afterwards, the chamber was heated at a rate of 5 °C min⁻¹ up to 450 °C and then flushed with Ar gas. For coating, a mixture of Ar (80%) and acetylene (C₂H₂, Messer Industriegase GmbH, Germany) (20%) was allowed into the chamber for 30 minutes followed by cooling of the mixture naturally to room temperature. Likewise, MP-CVD was used for the carbon encapsulation of the film particles. The sample was placed in the MP-CVD chamber which was flushed for 10 minutes with hydrogen (H₂), flowing at a rate of 100 sccm and pressure of 15 Torr. Afterwards, CH₄ at 10 sccm flow rate and 19 Torr pressure was allowed to enter the chamber. The coating was continued for five minutes at a microwave power of 500 W during which time the temperature raised to ∼450 °C.

All the prepared samples are summarized in Table S1, showing the sample codes used in this manuscript and the synthesis parameters.

Characterization

The X-ray diffraction (XRD) data were collected by using a PANalytical Empyrean (XRD) diffractometer which was equipped with Cu Kα (λ = 1.5425 Å) radiation. The data were searched and matched using XPertPro software while Rietveld refinement was carried out with FullProf Suite. Raman spectra were measured using InVita Reflex (Renishaw) equipped with a laser (excitation λ = 656 nm) which was used for continuous perpendicular illumination of the sample. X-ray photoelectron spectroscopy (XPS) measurements were performed on an SSX-100 (XPS) spectrometer (Evans Analytical Group LLC) and the data were fitted using CasaXPS software. Scanning electron microscopy (SEM) images were acquired by using Zeiss Ultra 55 (FE-SEM). Energy-dispersive X-ray (EDX) spectroscopy was carried out with a FE-SEM coupled Thermo Scientific detector, and quantitative and qualitative elemental analysis spectra and mapping were acquired. Transmission electron microscopy (TEM) and scanning TEM (STEM) were performed over FEI Talos 200 FA.

Electrochemical measurements

Electrochemical measurements were performed in a conventional three-electrode cell by using a CHI660E Electrochemical Workstation. A 1 cm × 1 cm piece of the film was used as the working electrode. Platinum (Pt) wire was used as the counter electrode while Ag/AgCl was used as the reference electrode. All the measurements were made in a 1 M Na₂SO₄ solution at room temperature. Cyclic voltammetry (CV) measurements were conducted in the potential window of 0 V to −1.0 V at different scanning rates ranging from 5 mV s⁻¹ to 150 mV s⁻¹. For each sample, two cycles were recorded. Electrochemical impedance spectroscopy (EIS) measurements were carried out at open circuit potential in the frequency range of 10 kHz–0.01 Hz while keeping an AC perturbation amplitude of 5 mV.

Results and discussion

The samples were prepared with the aim to fabricate Fe₃O₄ films directly on the current collector, i.e., Cu, for application as a negative electrode (anode) in solid state batteries. As anticipated, the material deposits on the Cu substrate as a film, while the excess settles down as powder. The powder and the film were retrieved and characterized by powder XRD. The pre- and post-calcination XRD patterns of the powder received at the end of the hydrothermal process carried out at 80 °C are presented in Fig. 1(a), showing distinct diffraction peaks. However, the position and intensities of the peaks are different for the pre- and post-calcination powder, indicating that the crystalline phase and composition of the as-received powder have changed upon calcination. Peaks in the XRD pattern of pre-calcination powder were searched and matched using XpertPro and it was found that the powder is composed of iron oxy-hydroxide (FeOOH). The main intensity peaks match with the standard XRD pattern of β-FeOOH given by the powder diffraction file (PDF) [01-075-1594] in the inorganic crystal structure database (ICSD). A small fraction of peaks matching the standard XRD pattern of α-FeOOH (PDF: [00-017-0536]) was found, indicating that the FeOOH powder crystallizes in two phases in which β-FeOOH is the majority phase while α-FeOOH is the minority phase. The post-calcination XRD pattern shown in Fig. 1(a) was searched and the peaks were matched with the standard XRD pattern of cubic Fe₃O₄ given by PDF [01-076-1849]. From the XRD data it is obvious that FeOOH decomposes upon heating under the applied conditions and converts to Fe₃O₄ according to the following proposed decomposition reaction (eqn (1)):Furthermore, it is evident that the transformation of FeOOH to Fe₃O₄ completes under the applied conditions, as no other phase is shown by the XRD pattern of the post-calcination powder.

Fig. 1 Powder XRD patterns of the pre- and post-calcined powder obtained by hydrothermal processing at 80 °C (a) and 120 °C (c), and films deposited on the Cu substrate by hydrothermal processing at 80 °C (b) and 120 °C (d), and subsequently coated with carbon by thermal and microwave plasma CVD.

The XRD pattern of the film deposited directly on the Cu substrate through hydrothermal processing at 80 °C is presented in Fig. 1(b), showing distinct diffraction peaks at various 2θ. The intense peaks at 2θ of 43.29°, 50.42° and 74.11° correspond to the diffraction peaks of the Cu substrate as per matching with the standard XRD pattern given in ICSD by PDF [01-085-1326] and the uncoated Cu substrate's XRD pattern which is presented in Fig. S1. The rest of the diffraction peaks were searched and matched by XpertPro, and it was found that they match well with the standard XRD pattern of α-FeOOH which is given by PDF [01-081-0462] in ICSD. Contrary to the settled down powder which was mostly composed of β-FeOOH, the film is composed of α-FeOOH. Similar to the powder, the as-fabricated film by hydrothermal treatment at 80 °C converts to Fe₃O₄ which is confirmed by the XRD pattern presented in Fig. 1(b), showing intense diffraction peaks which match exactly with the standard PDF [01-076-1849] of cubic Fe₃O₄. The post-calcination film on the Cu substrate was subjected to carbon coating of the Fe₃O₄ material by thermal CVD, and the pertinent XRD pattern is presented in Fig. 1(b). In comparison to the pre-carbon-coated film, the XRD pattern of the post-carbon-coated film presents some visible changes. A hump superimposed with a diffraction peak of Fe₃O₄ appears in the 2θ range of 15° to 25°. Secondly, the intensity of Fe₃O₄ peaks decreases. However, the peaks upon search and match correspond to the reflection of the standard XRD pattern given by PDF [01-076-1849] of Fe₃O₄. This indicates that the film material is crystalline, and the phase and composition remained unchanged after coating with carbon. The changes including the appearance of the hump and the decrease in the intensity of Fe₃O₄ peaks could be attributed to the presence of amorphous carbon. The XRD pattern (Fig. 1(b)) of the film coated with carbon by MP-CVD tells a different story. Similar to the XRD pattern of the carbon-coated film by thermal CVD, the MP-CVD carbon-coated film shows a broad hump in the 2θ range of 15° to 30°. The search and match of the XRD pattern by XpertPro show that the basic crystalline structure remains unchanged, i.e., the film is composed of Fe₃O₄ as the pre-carbon-coated film was. However, contrary to the thermal CVD carbon-coated film, the intensity of Fe₃O₄ peaks increased significantly, indicating an improvement in the crystal quality. Furthermore, new peaks in addition to those of Fe₃O₄ and the Cu substrate appeared at 2θ of 44.63° and 64.87°. Upon search and match, it was found that these very peaks match with the standard XRD pattern of iron (Fe) given by PDF [01-087-0721] in the ICSD. It means that the Fe₃O₄ film upon subjecting to carbon coating by MP-CVD underwent reduction, though partially.

XRD patterns of the pre- and post-calcination powder sample retrieved from the autoclave which was heat treated at 120 °C for fabrication of the film on the Cu substrate are presented in Fig. 1(c). The diffractogram of the as-received powder sample shows distinct diffraction peaks which were searched and matched by XpertPro with the standard XRD pattern of cubic Fe₃O₄ given by PDF [01-076-1849] in ICSD. A second set of lower intensity reflections at 2θ of 24.62°, 31.97°, 38.09°, 46.15°, 50.99°, 52.79°, etc. was found besides the diffraction peaks of Fe₃O₄. After search and match by XpertPro with the standard XRD pattern given by PDF [00-003-0746], it was found that these peaks are due to the reflections from crystal planes of iron carbonate (FeCO₃). The post-calcination XRD pattern of the powder retrieved from the autoclave following the hydrothermal process carried out at 120 °C is presented in Fig. 1(c), showing distinct peaks which were searched and matched with the standard XRD pattern (PDF: [01-076-1849]) of cubic Fe₃O₄. This indicates that under the given calcination regime, the powder comprising of Fe₃O₄ and FeCO₃ was converted to a single phase material, i.e., Fe₃O₄, according to the following proposed decomposition reaction of FeCO₃ (eqn (2)):

Besides the peaks of the Fe₃O₄ phase, no other peaks are shown by the XRD pattern of the post-calcination powder sample.

The films fabricated on the Cu substrate by hydrothermal processing at 120 °C were characterized by powder XRD and the pertinent diffractograms are presented in Fig. 1(d). The as-fabricated film gives well established peaks in its XRD pattern for the Cu substrate and the coating on it. Upon search and match by XpertPro, the peaks other than those of the Cu substrate are the result of reflections from crystallographic planes of Fe₃O₄ and FeCO₃ as these peaks match very well with the standard XRD patterns of cubic Fe₃O₄ and rhombohedral FeCO₃ which are given, respectively, by PDF [01-076-1849] and [00-003-0746] in ICSD. Contrary to the powder sample, the film has a higher amount of FeCO₃. Upon calcination of the as-received film, the crystal quality improved significantly, which is evident from the intensity and narrowing of the peaks in the XRD pattern of the post-calcination film presented in Fig. 1(d). The peaks were searched and matched by XpertPro, and it was found that the film is composed of a single phase, i.e., Fe₃O₄, as the reflections match very well with the standard XRD pattern of Fe₃O₄ given by PDF [01-079-0418] in ICSD. The XRD pattern of the post-calcination films after coating with carbon by thermal CVD is presented in Fig. 1(d). The peaks match with a single phase, i.e., cubic Fe₃O₄, given by PDF [01-079-0418] from the ICSD. However, the peak intensities are reduced after coating with carbon, which is also shown by a broad hump in the XRD pattern because carbon is an amorphous material. The film coated with carbon by MP-CVD gives peaks which match with the standard XRD pattern in PDF [01-075-1372] of cubic Fe₃O₄. However, the diffractogram also gives peaks for a second phase which is matched with PDF [00-006-0696] in ICSD of cubic Fe. A hump superimposed with the peaks of the Fe₃O₄ phase is also given by the XRD pattern, indicating the presence of amorphous carbon.

Rietveld refinement of the samples (residual powder and films) was carried out by employing Rietveld's whole-profile fitting method using FullProf software,⁴⁵ and the pertinent XRD patterns with Rietveld refined data are presented in the supplementary information (SI). In the refined XRD patterns presented in the SI, the Bragg's peak positions are represented by vertical lines, the line of difference between the observed and the calculated intensities is presented at the bottom of the corresponding patterns, the observed intensity is represented by open circles and the refined/calculated intensity by Rietveld refinement is represented by a solid line superimposed over the measured XRD pattern. The reflection peaks are labelled with the corresponding lattice planes. The pseudo-Voigt analytical function given in Tables S2 and 5 was used for fitting the experimental diffraction profiles as it has certain advantages discussed elsewhere.⁴⁶ The refinement steps involve refinement of the parameters in the following order unless stated otherwise: scale factor, zero point of detector, background which was fixed by using linear interpolation between a set of points provided, lattice constants, atomic positions unless stated otherwise, peak shape and symmetry parameters, atom occupancies unless stated otherwise and overall temperature factor. The quality of refinement was quantified from the figures of merit, including R_p (pattern residual), R_wp (weighted pattern residual), R_exp (expected pattern residual), R_B (Bragg's factor), R_F (structural factor) and goodness of fit (χ²), for which numerous mathematical relations are reported in the published literature.⁴⁷ The values of the above mentioned reliability factors for all the samples are presented in Tables S2 and 5. Refinement was continued till achieving convergence and goodness of fit (GOF = R_wp/R_exp) close to 1. It is pertinent to mention here that each time the data were plotted and the convergence/fitting of the refined pattern with the measured one was visually observed.

The XRD pattern of the residual powder (labelled as FO80-P_(bc)) retrieved from the autoclave which was subjected to heating at 80 °C was refined and is presented in Fig. S2(a), showing a very good fitting. The refined parameters such as crystal axes and the crystal axial angles match those of space groups I4/m of the tetragonal phase (β-FeOOH: Akaganeite) and Pbnm of the orthorhombic phase (α-FeOOH: Goethite), indicating that the powder is composed of two different phases of FeOOH. However, β-FeOOH is the major constituent amounting to 98.07%, while α-FeOOH contributes only 1.93% to the powder diffraction pattern. The interplanar spacing, d, presented in Table S4 shows a decrease as compared to the standard sample reported in ICSD, indicating that the lattice is strained. Therefore, the Williamson–Hall (W–H) method was used to estimate the micro-strain and crystallite size by plotting βcosθ vs. 4sinθ, using the relation reported elsewhere.^48,49 W–H plots for β- and α-FeOOH are presented in Fig. S2 (b) and (c), respectively, and the pertinent crystallite size and micro-strain data are given in Table S2, showing crystallite sizes of 53 nm and 25 nm, respectively, for the β- and α-FeOOH phases. As shown by the W–H plots of the β- and α-FeOOH phases in Fig. 2 (b) and (c), respectively, the micro-strain (slope) has positive and negative values, indicating that the strain is tensile in the case of the β-FeOOH phase and compressive in the case of the α-FeOOH phase.^50,51 The Rietveld refined XRD pattern (labelled as FO80-P_(ac)) of the post-calcination powder retrieved from the autoclave which was processed at 80 °C is presented in Fig. S3(a). The refined XRD pattern shows that the film is composed of a single phase, i.e., cubic Fe₃O₄ in the Fd3̄m space group. The calculated lattice constants are slightly shorter than the standard reported in PDF [01-076-1849]. The effect is also observed on the interplanar spacings which are also slightly shorter than those of the standard. The average particle size and micro-strain were obtained from the W–H plot which is presented in Fig. S3 (b). The linear fit gives a straight line which is not perfect. This indicates that there is some dispersion in the size and micro-strain of the film particles. The average crystallite size determined from the W–H plot is ∼37.23 nm (Table S2). The micro-strain of 0.0012 (0.12%) is positive, indicating that the micro-strain of the particles is due to lattice expansion. Rietveld refined XRD patterns of the pre- and post-calcination, and carbon-coated films on the Cu substrate retrieved from the autoclave after hydrothermal processing at 80 °C are presented in the SI along with their pertinent data. The Rietveld refined XRD pattern of the as-prepared film (labelled as FO80-0) is presented in Fig. S4(a), which matches exactly with a single phase, i.e., α-FeOOH (space group: Pbnm), indicating that the film is composed of only one phase contrary to the residual powder which was mainly composed of β-FeOOH. Furthermore, the lattice constants and the interplanar spacing of α-FeOOH constituting the film are significantly increased as compared to α-FeOOH of the powder and the standard (Tables S2 and 4). The W–H plot in Fig. S4(b) shows higher linearity which indicates that the film particles are homogeneous compared to the particles of residual powder. The crystallite size of 153.83 nm was estimated from the W–H plot. A positive micro-strain of 0.0024 (0.24%) indicates that the film particles are under tensile strain. The Rietveld refined XRD pattern (Fig. S5) of the calcined film (labelled as FO80-1) matches with the standard XRD pattern of cubic Fe₃O₄ having space group Fd3̄m. The lattice constants of FO80-1 are smaller compared to those of the powder sample and the standard XRD pattern of Fe₃O₄ given by PDF [01-076-1849]. Therefore, it is pertinent to confirm that the calcined film (FO80-1) is Fe₃O₄ and not maghemite (γ-Fe₂O₃), as the two have nearly indistinguishable crystal structures with the main difference being their lattice constants, which are smaller (∼8.34 Å) for γ-Fe₂O₃ and relatively large (∼8.39 Å) for Fe₃O₄, whereas we observed ∼8.363 Å for FO80-1, which lies somewhere in between the two. Previously, it has been reported that γ-Fe₂O₃ gives two additional small peaks at 2θ of 23.77° and 26.10°,⁵² which we did not observe in the diffractograms of FO80-1. Additionally, Kim et al.⁵² comprehensively investigated the Fe₃O₄ and γ-Fe₂O₃ systems, and proposed a method involving the comparison of reflection from the (511) plane, which gives a doublet if a mixture of the two phases is present. As shown in Fig. S5(c), we could not observe a doublet for this very plane, instead a Voigt-fitted single peak around ∼57.1° was found, corresponding to the Fe₃O₄ phase.⁵² It is pertinent to mention here that the reflection from the (511) plane in the case of the γ-Fe₂O₃ phase emerges at 2θ > 57.3°.⁵² Furthermore, Schwaminger et al.⁵³ differentiated between the two phases using Raman spectroscopy, which showed peaks at ∼660 cm⁻¹ for Fe₃O₄ and 710 cm⁻¹ for γ-Fe₂O₃.⁵³ Here in the case of FO80-1, the Raman spectrum (Fig. S14(b)) gives a single peak at 651 cm⁻¹. Even upon deconvolution, the γ-Fe₂O₃ peak reported at 710 cm⁻¹ is not observed (Fig. S14(h)). The decrease in the lattice constant of FO80-1 as compared to that of the standard Fe₃O₄ could be attributed to the lattice strain and distortions which are responsible for contraction of the unit cell as reported previously,⁵⁴ and evident from the W–H plot in Fig. S5(b), which shows compressive strain. Furthermore, the W–H plot (Fig. S5(b)) is completely straight and linear, indicating that the crystallites and/or particles and the micro-strain are homogeneously distributed. The crystallite size estimated from the W–H plot is ∼14.66 nm and the micro-strain −0.0007 (0.07%). The XRD pattern of the thermal CVD carbon coated film (FO80-2) was subjected to Rietveld refinement and it was found that the basic composition of the film remains unchanged as it matches exactly with the Fd3̄m space group of Fe₃O₄ (Fig. S6(a)). The measured lattice constants and the interplanar spacings are in good agreement with those of the standard, but larger than those of the precursor. The W–H plot in Fig. S6(b) shows some deviation from the straight line, indicating that the homogeneous distribution of particles and micro-strain is somewhat disturbed. The estimated particle size from the W–H plot is 28.85 nm which is almost double the particle size of its precursor film. Although the micro-strain has a negative value (−0.0006) as of the precursor, it has increased slightly. The Rietveld refined XRD pattern of the MP-CVD carbon coated film (FO80-3) is presented in Fig. S7(a), showing that the film is composed of two phases, i.e., Fe₃O₄ (space group Fd3̄m) and pure iron (Fe), which crystallizes in space group Im3̄m. The lattice constants of Fe₃O₄ have increased slightly compared to its precursor but remained smaller than that of the standard given by PDF [01-076-1849]. Likewise, the lattice constants of cubic Fe are slightly smaller than that of standard Fe in PDF [01-087-0721]. The percent fraction of Fe in the film is less than 1%. The crystallite size from the W–H plot, in the case of the Fe₃O₄ phase, and from Scherrer's equation, in the case of Fe, was estimated as 22.35 nm and 24.00 nm, respectively. The W–H plot in Fig. S7(b) of the Fe₃O₄ phase is exactly straight indicating that the particles and micro-strains are homogeneously distributed. The negative value of micro-strain shows a compressive strain in the particles amounting to -0.0005 (0.05%). The negative and smaller value of micro-strain further indicates that the peak broadening is insignificantly affected by micro-strain.

Fig. 2 (a) Deconvoluted Raman spectra of sample FO80-2 (a), FO80-3 (b), FO120-2 (c) and FO120-3 (d) in the region of 850 cm⁻¹ to 1950 cm⁻¹, showing the D- and G-band of the carbon material.

The Rietveld refined XRD pattern of the sample (FO120-P_(bc)) received as a powder from the autoclave, which was processed at 120 °C, is presented in Fig. S8(a). The powder is composed of two phases, i.e., Fe₃O₄ in space group Fd3̄m and FeCO₃ in space group R3̄c. The percent fraction of the two phases is, respectively, 91.95% and 08.05%, indicating that the main phase is cubic Fe₃O₄. The interplanar spacings of the Fe₃O₄ phase lattice are increased slightly compared to that of the standard lattice in PDF [01-076-1849]. However, a significant deviation from the standard in the lattice constants of the FeCO₃ phase is observed. The W–H plots of the two phases, i.e., Fe₃O₄ and FeCO₃, are presented, respectively, in Fig. S8(b) and (c), showing a marked difference between them. As evident from the straight line, the particles of Fe₃O₄ are more homogeneously distributed than the particles of FeCO₃. Furthermore, the crystallites of the two phases are differently strained as evident from the positive slope (tensile strain) of Fe₃O₄ and the negative slope (compressive strain) of FeCO₃ W–H plots. The crystallite size estimated from W–H plots is 72.05 nm (Fe₃O₄) and 50.28 nm (FeCO₃). The post-calcination powder sample (FO120-P_(ac)) was Rietveld refined and it was found that the resulting material consists of a single phase, i.e., cubic Fe₃O₄ (space group: Fd3̄m), as shown by the Rietveld refined XRD pattern in Fig. S9(a). From Tables S5 and S7, it is evident that there is no significant change in lattice parameters upon calcination of the as-received powder from the autoclave at 120 °C. However, the crystallite size estimated from the W–H plot in Fig. S9(b) is significantly large, i.e., ∼114.03 nm compared to 72.05 nm of its precursor. The Rietveld refined XRD pattern of the pre-calcination film (FO120-0) achieved by hydrothermal processing at 120 °C (Fig. S10(a)) shows that the film is also composed of two phases like the residual powder. However, the percent fractions of the two phases, i.e., Fe₃O₄ and FeCO₃, in the film are quite different from those in the residual powder. Contrary to the 8.05% share of FeCO₃ in the composition of the powder, the film is composed of 35.17% of FeCO₃. The particles are under tensile strain and more homogeneously distributed as evident, respectively, from the positive slope (micro-strain) and straight lines of W–H plots of Fe₃O₄ and FeCO₃ given in Fig. S10 (b) and (c). The crystallite size estimated is 33.22 (Fe₃O₄) nm and 42.85 nm (FeCO₃). From Tables S5 and S7, it is evident that the lattice parameters deviate slightly from those of standard lattices. The post-calcination Rietveld refined XRD pattern of the film (FO120-1) given in Fig. S11(a) shows a single phase, i.e., cubic Fe₃O₄ (space group: Fd3̄m). The lattice constants of the cubic Fe₃O₄ phase decreased significantly compared to the lattice constants of the same phase in its precursor. However, the crystallite size determined from the W–H plot in Fig. S11(b) increased up to 46.42 nm. Furthermore, contrary to the precursor, the W–H plot deviates significantly from a straight line and gives a negative slope, indicating that the particles are less homogeneously distributed and under compressive strain. The post-calcination film upon coating with carbon by T-CVD (FO120-2) was subjected to Rietveld refinement and it was found that the refined XRD pattern matches very well with the cubic Fe₃O₄ (space group: Fd3̄m) as shown in Fig. S12(a). No significant change in its lattice parameters was observed compared to those of the precursor but the crystallites size is witnessed to be decreased to 37.91 nm from 46.42 nm of its precursor. This may be due to the so-called pore-wedge effect. Contrary to FO80-2, FO120-2 involves the evolution of CO₂ which is responsible for the creation of larger pores in the skeleton. During carbon coating by T-CVD, the C₂H₂ precursor molecules easily diffuse deep into these internal pores. The following carbon deposition then exerts compressive strain through mechanical pressure, which leads to the fragmentation of the Fe₃O₄ crystallites and thus a decrease in their size.⁵⁵ The W–H plot in Fig. S12(b), which presents a negative plot and increased deviation from straight line, further supports this explanation. The Rietveld refined XRD pattern of the film (FO120-3) obtained by coating the calcined film with MP-CVD is presented in Fig. S13(a), showing that the film is composed of two phases, i.e., cubic Fe₃O₄ (space group: Fd3̄m) and cubic Fe (space group: Im3̄m), with a percent fraction of 93.73% and 6.27%, respectively. The W–H plot in Fig. S13(b) for the cubic Fe₃O₄ phase is positive and shows lesser deviation from straight line compared to its precursor. This indicates that the Fe₃O₄ particles have tensile stress and more homogeneous distribution. Furthermore, the crystallite size (58.63 nm) of the Fe₃O₄ phase is far larger than that of its precursor (46.42 nm) and the T-CVD coated film (37.91 nm).

The films were characterized by Raman spectroscopy to get some insight into the nature of the film materials. The as-synthesized film (FO80-0) by hydrothermal processing at 80 °C showed an almost identical Raman spectrum as reported previously.^56,57 As shown in Fig. S14(a), the peaks are quite broad and thus overlapping, but their positions correspond to those of α-FeOOH (Goethite).^56,57 Upon calcination of FO80-0, the Raman spectrum Sof film FO80-1 (Fig. S14(b)) shows a peculiar peak of cubic Fe₃O₄ (magnetite) at 651 cm⁻¹ for the A_1g mode.^58–60 Besides, we can see another strong peak at 1306 cm⁻¹, which is attributed to the decomposition of Fe₃O₄ to Fe₂O₃ upon irradiation of the sample with strong laser light during Raman measurements.⁶¹ The Raman spectrum Sof the sample (FO80-2) obtained by coating the post-calcination film (FO80-1) with T-CVD is presented in Fig. S14(c). As one can see the Raman spectrum shows a peculiar peak for the magnetite (cubic Fe₃O₄) at 669 cm⁻¹ as was observed in the Raman spectrum of its precursor. Besides the magnetite peaks, the Raman spectrum of the film (FO80-2) shows two other distinct peaks at 1347 cm⁻¹ and 1581 cm⁻¹. These peaks are attributed to the disordered (D-band) and graphitic (G-band) bands, respectively, of carbon.⁶² These very bands were deconvoluted by fitting with the Lorentzian function as shown in Fig. 2(a), and it was found that the area under the D band is quite larger than the area under the G band, indicating that the carbon material is largely disordered. The ratio (I_D/I_G) of D and G bands was determined from the peak area and was found to be ∼1.21. The film (FO80-3) obtained from the same precursor film subjected to coating by MP-CVD gave Raman peaks for Fe₃O₄ and also for the carbon materials as shown in Fig. S14(d) and Fig. 2(b). However, the carbon peaks are superimposed by the dominant peaks of Fe₃O₄ in contrast to that of the T-CVD carbon coated film (FO80-2). The peaks in the region of 850 cm⁻¹ to 1950 cm⁻¹ were deconvoluted to extract the peaks of the carbon materials. As shown in Fig. 2(b), the peak of D-band is weaker than the overlapped peaks of Fe₂O₃, which come from decomposition of Fe₃O₄ under the Raman laser light, while that of G-band is further weaker.⁶² The ratio of I_D/I_G is ∼1.72, which is larger than that of the T-CVD coated Fe₂O₄ film (FO80-2). This indicates that the disordered carbon is larger in quantity as compared to the ordered carbon.⁶² As we would see in the following sections, the lower intensity of the carbon peaks in the presence of magnetite peaks could also be due to the thinner carbon coating around magnetite particles as compared to carbon coating produced by T-CVD. The post-calcination sample (FO120-1) which was obtained from the precursor film processed hydrothermally at 120 °C shows in its Raman spectrum in Fig. S14 (e) a peculiar peak of cubic Fe₃O₄ (magnetite) at 665 cm⁻¹. However, this very film upon coating with carbon by T-CVD shows in its Raman spectrum in Fig. S14(f) peaks besides those of magnetite (cubic Fe₃O₄). The peaks at 1315 cm⁻¹ and 1565 cm⁻¹ are the corresponding peaks of D- and G-band of carbon material. The two peaks were deconvoluted using the Lorentzian function and the ratio of I_D/I_G was determined to be ∼2.11, as shown in Fig. 2(c). The Raman spectrum in Fig. S14(g) for the sample FO120-3 (the film obtained by carbon coating with MP-CVD of the post-calcination film obtained at 120 °C by hydrothermal processing) gives the peculiar peaks of magnetite (cubic Fe₃O₄). Besides, peaks for carbon material appear at 1323 cm⁻¹ and 1538 cm⁻¹ due to the D- and G-band, respectively. The peak for D-band is quite clear, but the G-band peak could only be obtained through deconvolution of the spectrum employing the Lorentzian function. As presented in Fig. 2(d), the deconvoluted spectrum gives peaks for D- and G-band with a ratio of 2.13, indicating that the carbon material is for the most part a disordered one.⁶²

XPS was used to probe the elemental composition and valence states of the surfaces of pre- and post-calcination, and carbon coated post-calcination films. The as-prepared, by hydrothermal processing at 80 °C, film (FO80-0) shows a typical α-FeOOH (Goethite) XPS survey spectrum (Fig. S15(a)) with major peaks of Fe 2p, O 1s and Fe 3p, corresponding very well to the reported XPS spectra of α-FeOOH.^63,64 As shown in Fig. S15(a), the XPS spectrum of FO80-0 also shows peaks of carbon (C 1s) and fluorine (F 1s), which could be attributed to the presence of adsorbed organic species from solvents and NH₄F, respectively. As shown in Fig. 3(a), Fe 2p bands appear at a binding energy (BE) of 711.0 eV and 724.9 eV for 2p_3/2 and 2p_1/2, respectively.⁶⁵ Besides, two peaks at 719.7 eV and 732.4 eV appear, respectively, for the satellite peaks of Fe 2p_3/2 and Fe 2p_1/2.^63,66–68 Each of the Fe 2p peaks shows two sub-bands upon deconvolution at 710.7 eV and 712.8 eV, and 724.7 eV and 727.8 eV, respectively, for the Fe 2p_2/3 and Fe 2p_1/2 bands, indicating the 3+ (i.e., Fe³⁺) valence state of Fe. Materials having Fe in its Fe³⁺ valence state show this sort of typical bands for Fe 2p.^65,69,70 The O 1s spectrum of sample FO80-0 is given in Fig. 3(b), showing BE peaks at 529.8 eV, 531.2 eV, 532.2 eV and 533.4 eV, respectively, for the lattice oxygen (Fe–O), lattice hydroxyl group (Fe–OH), adsorbed hydroxyl group (Fe–OH_ads) and adsorbed water (H₂O).^71,72 The deconvoluted O 1s spectrum shown in Fig. 3(b) reveals that the contribution of lattice Fe–O and lattice Fe–OH is predominant. The presence of Fe–OH_ads could be justified by the fact that water (H₂O) after adsorption on Fe species decomposes to OH when subjected to heating. The films were prepared by processing the Fe-based material at elevated temperature (80 °C) in a tightly closed vessel (i.e., autoclave). Therefore, the adsorbed water would have been transformed upon the surface of the α-FeOOH film according to the following chemical reaction (eqn (3)), which was previously proposed based on the temperature programmed desorption (TPD) studies:⁷²

H₂O → OH_ads + H_ads (3)

Owing to this reaction on Fe, the contribution of Fe–OH_ads is increased than the contribution of adsorbed H₂O.⁷² The XPS survey spectrum of the post-calcination film (FO80-1) is presented in Fig. S15(b). By comparing it with its precursor's XPS survey spectrum, one can clearly see that the peaks of F are completely vanished, while the peaks of carbon decreased in intensity to be nearly insignificant. From these observations, it is obvious that the F and C peaks in its precursor's XPS spectrum are owing to the presence of adsorbed F and carbon species, which upon calcination were removed. The significant peaks shown in the XPS spectrum of FO80-1 are owing to Fe 2p, O 1s and Fe 3P. The Fe 2p spectrum is presented in Fig. 3(c), showing two peaks of Fe 2p_3/2 and Fe 2p_1/2 at 710.8 eV and 724.5 eV, respectively, resembling the reported Fe 2p spectrum of Fe₃O₄.⁷³ The ratio of 2p_3/2/2p_1/2 determined from the area under the peaks is approximately 1.65, which is closer to the expected value of 2. The Fe 2p_3/2 and Fe 2p_1/2 peaks were deconvoluted and the peaks for Fe²⁺ were found at 710.4 eV and 724.1 eV, and those for Fe³⁺ at 712.7 eV and 727.0 eV, which are in good agreement with the previously reported values of pure Fe₃O₄.⁷⁴ The Fe 2p_3/2 satellite peak shifts to a lower value on the one hand, and on the other hand is poorly resolved owing to the overlapping of satellite structures of the octahedral Fe²⁺ and Fe³⁺, a characteristic of Fe₃O₄.⁷⁴ The ratio of the Fe 2p_3/2 satellite peak over the main Fe 2p_3/2 is quite small, i.e., 0.09 (8.5% of the main peak), which is reaching above 0.45 in the case of iron oxide containing only Fe³⁺ (like γ-Fe₂O₃ and α-Fe₂O₃).⁷⁴ This indicates that the post-calcination film is primarily composed of Fe₃O₄. The O 1s spectrum in Fig. 3(d) shows an increased contribution of the lattice oxygen (i.e., Fe–O) as compared to its precursor. This is further indication of the fact that the precursor α-FeOOH transformed to Fe₃O₄ upon calcination under the applied heat treatment regime. The contribution from adsorbed H₂O is insignificant (the ratio of the main Fe–O peak to the H₂O peak is ∼0.04 compared to 0.10 in the case of the precursor). Similarly, the ratio of the main Fe–O peak to OH_ads is lowered to 0.19 from 0.49 of the precursor. On the other hand, the ratio of the main Fe–O peak to Fe–OH increased by ∼293%, confirming further that the main phase of the post-calcination film is Fe₃O₄. This is in good agreement with the XRD and Raman data presented above. The full survey spectrum of the carbon coated post-calcination film (FO80-2) is presented in Fig. S15(c), showing peaks for the core-shell electrons of Fe 2p, O 1s and C 1s besides the peaks for Fe 3p and Fe 3s. Contrary to the XPS spectrum of its precursor, the core level electron peak of C 1s is quite strong and prominent, indicating that the film is coated with the required element, i.e., carbon. The deconvoluted Fe 2p spectrum of FO80-2 is presented in Fig. 3(e), showing a BE peak at 711.1 eV and another one at 724.5 eV for the core electrons of Fe 2p_3/2 and Fe 2p_1/2, respectively.⁷⁵ The peaks were deconvoluted to 710.2 eV and 712.4 eV (Fe 2p_3/2), and 724.1 eV and 728.2 eV (Fe 2p_1/2). The Fe²⁺ to Fe³⁺ ratio was calculated to be 0.37 : 0.63, which slightly deviates from the 0.33 : 0.67 ratio of stoichiometric Fe₃O₄ expressed as FeO·Fe₂O₃.⁷⁵ This indicates that the resulting film contains nearly stoichiometric Fe₃O₄, corresponding the XRD and Raman data presented above. Furthermore, the single phase of iron oxide, i.e., Fe₃O₄, composing the film is evident from two other but important features of the Fe 2p spectrum. First, the satellite peak at 719.1 eV which is a trademark of pure Fe³⁺ based materials is completely absent in the XPS spectrum of Fe 2p of the film FO80-2. Second, the shoulder peak at 710.2 eV which appears for Fe²⁺ in Fe₃O₄ became quite prominent as compared to its precursor.⁷⁴ From these observations, it is quite clear that by treating the film by T-CVD, we not only incorporated the carbon coating around Fe₃O₄ but also improved the crystallinity of iron oxide to single phase Fe₃O₄. The O 1s spectrum of FO80-2 in Fig. 3(f) shows a dominant peak at 530.5 eV for the lattice oxygen (Fe–O). There are two other important peaks in the O 1s spectrum of the FO80-2 film, appearing at a BE of 531.7 eV and 532.9 eV, which are attributed to the core-shell electrons of carbon involved in bonding with the oxygen of Fe₃O₄ in Fe–O–C and C–O bonds, respectively.^28–30 It is worth mentioning here that the formation of the Fe–O–C bond indicates a strong interaction between the core (i.e., Fe₃O₄) and the shell (i.e., carbon). This interaction is particularly important for enhanced strength of encapsulation of the core in the shell, improving the conductivity by shedding of Fe₃O₄ from the carbon matrix on the one hand. On the other hand, it helps in stabilizing Fe₃O₄ by restraining its particle aggregation during electrochemical applications.⁷⁶ A peak appearing at 534.0 eV in the O 1s spectrum of FO80-2 shows an insignificant contribution to this very spectrum from the adsorbed water species. It is possible that an insignificant amount of water has been adsorbed during handling of the sample. The high resolution core level XPS spectrum of C 1s is presented in Fig. 3(g), showing a main peak at 284.8 eV which is attributed to graphitized carbon (sp²-C).^77–81 This indicates that most of the carbon atoms are in a conjugated hexagonal form.⁸⁰ The peak at 286.0 eV is attributed to C–O, indicating that the carbon shell is strongly bonded with the Fe₃O₄ core through Fe–O–C bonding as was also observed in the XPS spectrum of O 1s.^77–79 The tail of the peak at the higher BE side was also deconvoluted to small peaks appearing at 287.4 eV and 288.9 eV which correspond to C=O and O–C=O.^{77,78,80–82} The deconvoluted peak at 290.7 eV is attributed to π–π* of graphitic carbon.⁸² Similar to the FO80-2 film, the XPS spectrum in Fig. S15(d) of carbon coated Fe₃O₄ by MP-CVD shows three main peaks for the core electrons of Fe 2P, O 1s and C 1s. The high resolution XPS spectrum of Fe 2p is presented in Fig. 3(h), showing the usual two peaks of Fe 2p_3/2 and Fe 2p_1/2 at 710.7 eV and 724.2 eV, respectively. These very peaks were deconvoluted to 710.1 eV (Fe²⁺) and 712.2 eV (Fe³⁺), and 723.9 eV (Fe²⁺) and 727.8 eV (Fe³⁺). The satellite peak at 719.4 eV is negligible, indicating that the material is single phase pure Fe₃O₄. The O 1s spectrum of FO80-3 shows three peaks at 530.5 eV, 531.8 eV and 532.9 eV corresponding to lattice oxygen (Fe–O), oxygen bridging iron and carbon (Fe–O–C) and C–O, respectively.^78–80 Unlike its precursor and the T-CVD carbon coated Fe₃O₄ film (FO80-2), the high resolution C 1s spectrum of the MP-CVD coated Fe₃O₄ film (FO80-3) shows only three peaks at 284.7 eV, 286.2 eV and 288.24 eV, corresponding to C–C, C–O and C=O, respectively, of graphitic carbon, most probably of graphene.^82,83

Fig. 3 XPS spectra of (a) Fe 2p and (b) O 1s of FO80-0, (c) Fe 2p and (d) O 1s of FO80-1, (e) Fe 2p, (f) O 1s and (g) C 1s of FO80-2, and (h) Fe 2p, (i) O 1s and (j) C 1s of FO80-3.

The films prepared at 120 °C by the hydrothermal technique and followed by heat treatment to get the calcined film (FO120-1) which was then coated with carbon by T-CVD (FO120-2) and MP-CVD (FO120-3) were also studied by XPS analysis to get their elemental composition and valence states. The relevant XPS survey spectrum of the post-calcination film (FO120-1) is presented in S15(e) which shows core electron peaks of Fe 2p and O 1s along with an insignificant peak of C 1s, corresponding to the XPS spectrum of the post-calcination film (FO80-1). The high resolution Fe 2p XPS spectrum of FO120-1 is presented in Fig. 4(a) which shows two peaks at 711.0 eV and 724.6 eV for Fe 2p_3/2 and Fe 2p_1/2, respectively. Both the peaks were deconvoluted to get the contribution from Fe²⁺ and Fe³⁺ to each peak. Fe 2p_3/2 gives deconvoluted peaks at 710.4 eV and 712.6 eV for the contribution of Fe²⁺ and Fe³⁺, respectively. Likewise the Fe 2p_1/2 peak is deconvoluted to two peaks at 724.3 eV and 728.9 eV, respectively, for the contribution of Fe²⁺ and Fe³⁺. These values correspond to the post-calcination film (FO80-1) and the reported Fe₃O₄ materials, indicating that the film is composed of cubic Fe₃O₄ as revealed by XRD and Raman data presented in the preceding sections. Furthermore, similar to the post-calcination film FO80-1, the Fe 2p XPS spectrum of the post-calcination film FO120-1 also shows a clear shoulder at around 710.1 eV and a flattened satellite peak at 719.3 eV which is obviously insignificant. The appearance of a clear shoulder and the absence of the satellite peak are the peculiarity of Fe₃O₄.⁷⁴ This indicates that the film in FO120-1 is composed of single phase Fe₃O₄. Contrary to the as-retrieved films, the deconvoluted O 1s core electron XPS spectrum presented in Fig. 4(b) of FO120-1 shows the main peak at 530.7 eV, corresponding to the lattice oxygen, i.e., Fe–O. Similar to the post-calcination film FO80-1, the O 1s XPS spectrum of FO120-1 also shows peaks at 531.6 eV, 532.6 eV and 533.8 eV for the Fe–OH, Fe–OH_ads and H₂O oxygen, respectively. The adsorbed water contribution to the O 1s spectrum is nearly negligible, and could be due to the adsorbed water during handling of the sample. The XPS survey spectra of the post-calcination film coated with carbon by T-CVD (FO120-2) and MP-CVD (FO120-3) are presented in Fig. S15(f) and (g), respectively, showing peaks for the elements Fe, O and C. The high resolution Fe 2p XPS spectra presented in Fig. 4(c) and (f) are almost identical, and correspond to those of FO80-2 and FO80-3. This indicates the fact that the films are composed of Fe₃O₄. The high resolution O 1s XPS spectra presented in Fig. 4(d) and (g), respectively, of FO120-2 and FO120-3 are also in good agreement with the high resolution O 1s spectra of FO80-2 and FO80-3. As shown in the XPS survey spectra (Fig. S15), the T-CVD and MP-CVD carbon coated films, i.e., FO120-2 and FO120-3, respectively, show the C 1s core electron peak which is nearly negligible in their precursor film, i.e., FO120-1. Furthermore, the high resolution C 1s XPS spectra of FO120-2 and FO120-3 presented, respectively, in Fig. 4(e) and (h) are no different from the C 1s spectra of FO80-2 and FO80-3. This indicates that the Fe₃O₄ material of FO120-2 and FO120-3 films is successfully encapsulated in carbon materials.

Fig. 4 XPS spectra of (a) Fe 2p and (b) O 1s of FO120-1, (c) Fe 2p, (d) O 1s and (e) C 1s of FO80-2, (f) Fe 2p, (g) O 1s and (h) C 1s of FO80-3.

The structural morphology of the film surfaces was examined by FE-SEM and the pertinent FE-SEM images are presented in Fig. 5 and Fig. S16. As shown in the FE-SEM image in Fig. S16(a), the as-synthesized film (FO80-0) by hydrothermal processing at 80 °C gives densely populated nanorods (NRs) which are uniformly grown on the Cu substrate surface. The high resolution FE-SEM image of FO80-0 in Fig. S16(b) shows that the NRs have a rhombus shape, having sides of equal lengths and smooth surfaces. The average side length ranges between 205 nm and 400 nm, and the induced edge angle is 50° which is a peculiarity of the rhombus shape. Previously, an identical morphology was reported for Co₃O₄ when it was realized under nearly the same synthesis conditions as those applied here by us.^42,43 Upon calcination, the morphological structure of the film (FO80-1) does not vary significantly. As one can see in Fig. 5(a), the FE-SEM image of the FO80-1 film shows densely populated Cu-substrate surface with NR arrays. The high magnification FE-SEM image of FO80-1 in Fig. 5(b) shows that the basic rhombus shape of the NR arrays is intact after calcination. It is evident from the FE-SEM images of samples FO80-2 and FO80-3 in Fig. 5(c) and (e), respectively, that the morphological structure upon the Cu-substrate surface is more pronounced than in their precursor (i.e., FO80-0). However, in the case of both the carbon coated samples, i.e., FO80-2 and FO80-3, the basic rhombus structure of the NR arrays remained intact, which is evident from their high magnification FE-SEM images given, respectively, in Fig. 5(d) and (f). Although the basic rhombus structure was preserved upon carbon coatings, the surface of the NR arrays becomes coarse in the case of FO80-2, and coarse and porous in the case of FO80-3. The T-CVD carbon coated rhombus NR arrays also show an increase in the length of their edges and a change in induced edge angles. The average edge length of the rhombus NRs is ranging between 325 nm and 475 nm while the induced edge angle is between 45° and 63°. The significant increase in edge length and the change in edge angle could be attributed to the carbon coating which may have brought some deformation to the basic structure of the rhombus NR arrays. Contrarily, the MP-CVD carbon coated rhombus NR arrays show a slight increase in their edge lengths and angles, indicating that the carbon coating is thin and more uniform in the case of MP-CVD than the T-CVD.

Fig. 5 Low as well as high magnification FE-SEM images of FO80-1 ((a) and (b)), FO80-2 ((c) and (d)), FO80-3 ((e) and (f)), FO120-1 ((g) and (h)), FO120-2 ((i) and (j)), and FO120-3 ((k) and (l)).

The structural morphology of the as-synthesized film (FO120-0) at 120 °C for 12 hours is quite the same as that reported earlier.²⁶ However, contrary to the earlier cases, the morphology in this case is realized at higher hydrothermal temperature, i.e., at 120 °C. As shown in Fig. S16(c), the FE-SEM image of the sample FO120-0 shows nanolayers, which are assembled to form layer-by-layer three dimensional (3D) superstructures. The high magnification FE-SEM image of FO120-0 in Fig. S16(d) shows that the nanolayers are actually formed of self-assembled nanoparticles. Upon calcination, the basic morphology is retained as shown by the FE-SEM image of FO120-1 in Fig. 5(g). The high resolution FE-SEM image of the same film in Fig. 5(h) indicates very clearly the nanolayers which are formed of nanoparticles. The layered structure of the precursor is also preserved in the carbon coated films, i.e., FO120-2 and FO120-3 by T-CVD and MP-CVD, respectively. However, contrary to the precursor and also the post-calcination films, the carbon coated film shows porous morphology. Using an EDX analyzer coupled with a FE-SEM, the elemental composition of the films was also studied, and the pertinent data are presented in Fig. S17. The EDX spectra and the corresponding elemental mappings show all the essential elements of the films. It is pertinent to mention here that the EDX spectra of the precursor films, i.e., FO80-0 and FO120-0, also show peaks for fluorine (F₂) besides the peaks for integral elements of the films.

The TEM image of FO80-0 is given in Fig. S18(a), showing self-assembled NR arrays, confirming the morphology shown by FE-SEM images in Fig. S16(a) and (b). The high magnification TEM image which is given as the inset in Fig. S18(a) shows a single rod having a length of 1.96 µm and a width of 430 nm, indicating that the NR arrays have length in micrometers while the width is in nanometers. Furthermore, the high magnification TEM image which is given as the inset in Fig. S18(a) shows that the NRs are quite smooth, having no structures on their surfaces. The HR-TEM image of FO80-0 is presented in Fig. S18(b), showing clear lattice fringes of ∼0.273 nm which correspond to the interplanar distance of 0.269 nm of the lattice plane (130) of α-FeOOH.⁸⁴ The inset in Fig. S18(b) represents the fast Fourier transformed (FFT) pattern of the HR-TEM image. The clear pattern of dots shown by the FFT image indicates that the NR arrays are crystalline in nature, confirming the XRD results. The elemental composition of the NR arrays was also studied by EDX coupled with TEM. As shown in Fig. S18(c), the EDX spectrum of FO80-0 shows peaks for the integral elements, i.e., Fe and O of α-FeOOH. Besides, the peaks of C and F are also clearly shown as were shown in the EDX spectrum obtained using an EDX analyzer coupled with a SEM. The EDX line scan profile of a single NR was obtained along the line indicated in Fig. S18(d), showing the C pattern which is mainly due to the carbon film of the TEM grid. It is pertinent to mention here that the carbon film supported TEM grids were used to provide support to the nanoparticles during TEM analysis. Elemental mapping images in Fig. S18(e) show uniform distribution of the two essential elements, i.e., Fe and O, besides of F. However, in the elemental mapping image of C, the background carbon film is more visible than any carbon of the NR array. The TEM image of the post-calcination film, i.e., FO80-1, is given in Fig. S18(f), showing NR arrays having lengths and widths comparable to those of the precursor (FO80-0). The NRs have smooth surfaces as was the case with their precursor, showing no additional features and/or structures over their surfaces as evident from the high magnification TEM image presented as the inset in Fig. S18(f). The HR-TEM image presented in Fig. S18(g) shows clear fringes with spacings of 0.260 nm, 0.300 nm and 0.460 nm, respectively, corresponding to the interplanar spacings of lattice planes (311), (220) and (111) of cubic Fe₃O₄.⁸⁵ This confirms the XRD data which show that upon calcination, the α-FeOOH film completely transforms to Fe₃O₄ under the applied calcination regime. The crystalline nature of the film material of FO80-1 is further evident from the FFT image presented as the inset in Fig. S18(g), showing a clear pattern of dots. The EDX spectrum in Fig. S18(h) of FO80-1 shows only the essential elements, i.e., Fe and O, of Fe₃O₄, and contrary to the precursor, the F peak is missing in the EDX spectrum of the post-calcination film as was the case with the EDX spectrum obtained by using an EDX analyzer coupled with a SEM. However, the C peak could be seen and could be due to the carbon film of the TEM grid. The EDX line scan profile of a single NR was obtained along the line indicated in Fig. S18(i). As shown by the graph in Fig. S18(i), Fe and O are uniformly distributed over the entire NR. The insignificant profile of C could be due to the carbon film of the TEM grid. This is further evident from the EDX mapping images of FO80-1 presented in Fig. S18(j), showing uniform distribution of Fe and O over the entire NR. However, the EDX mapping of C is clearly shown to be due to the carbon film of the TEM grid.

Low magnification TEM images of the film, FO80-2, obtained after carbon coating of the post-calcination film (FO80-1) show NR arrays having lengths in micrometers while the diameter spans over a large range, confirming the SEM results. Upon scratching from the substrate and then dispersing in ethanol for casting on the TEM grid, the smaller width NR arrays mostly remained agglomerated as shown in Fig. S19(a) while the larger width NR arrays dispersed completely into individual NRs as shown in Fig. 6(a). Contrary to the precursor, surfaces of the NR arrays turned rough upon carbon coating in T-CVD. The STEM HAADF images presented as the inset in Fig. 6(a) and Fig. S19(a) show blind holes which seem to be formed by etching of the surface by the reactive gas. The high magnification TEM images presented in Fig. 6(b) and Fig. S19(b) show a central part (core) covered by a thin shell. The FFT performed over the central part (i.e., core) shows a clear pattern of dots while the overlayer (shell) shows no diffraction spots, as shown in the insets of Fig. 6(b) and Fig. S19(b). This indicates that the core is crystalline in nature while the shell is amorphous. HR-TEM images of NR arrays are presented in Fig. 6(c) and Fig. S19(c), showing well defined fringes at ∼0.498 nm and 0.300 nm for the interplanar spacings of lattice planes (111) and (220), respectively, of cubic Fe₃O₄.⁸⁵ EDX spectra measured over an agglomerate as well as standing alone NRs show peaks for the three elements, i.e., Fe, O and C, as shown in Fig. S19(d). As indicated in Fig. S19(e), the EDX line scan profile of a single NR shows variation in Fe, O and C distribution. While passing through the ends and the surface holes of the NR array, the intensity of the C line surges while that of Fe and O decreases, indicating that C surrounds the Fe₃O₄ particles. This fact is further evident from the EDX elemental mapping images in Fig. 6(d) and Fig. S19(f), which show that despite uniform distribution of the elements Fe, O and C, the intensity of C is stronger at the extremities and at the holes in the NR array. To discern the increased intensity at the extremities and the holes in the NR arrays, the scratched film was subjected to strong and prolonged ultrasonication and then drop cast over the TEM grid. The TEM image given in Fig. S19(g) shows that the holes worked as fissure lines, supporting disintegration of the NR arrays into non-uniform core–shell particles. The high magnification TEM image in Fig. S19(h) was subjected to FFT and it was found that the shell is amorphous as the FFT pattern is blurry, while the core is crystalline as the FFT pattern shows a clear pattern of dots. The HR-The TEM image in Fig. S19(i) shows that the core is primarily made up of Fe₃O₄ by giving clear fringes of 0.450 nm which correspond to the interplanar spacing of the lattice plane (111) of cubic Fe₃O₄. The MP-CVD carbon coated film (FO80-3) is no more different in its morphological structure from the T-CVD carbon coated film (FO80-2). The low and high magnification TEM images in Fig. 6(e) and (f), respectively, show NR arrays having a rough surface with holes as those of FO80-2. High magnification and HR-TEM images in Fig. 6(f) and (g), respectively, show that the NR arrays are consisting of crystalline particles of Fe₃O₄ surrounded by amorphous carbon as depicted by the FFT dot patterns and fringes, and the blurry FFT pattern, respectively. This is further supported by the EDX data and their line scan profile depicted in Fig. S20(c) and (d), respectively. Contrary to the precursor (FO80-1) as well as the T-CVD coated film (FO80-2), the EDX line scan profile shows decreased O intensity in comparison to the intensity of Fe. It is understandable as the MP-CVD carbon coated film contains metallic Fe besides Fe₃O₄, increasing the overall content of Fe in comparison to O. The presence of metallic Fe is well established from the powder XRD data presented in the preceding sections.

Fig. 6 Low (a) and high (b) magnification TEM images, HR-TEM image with FFT as the inset (c), and elemental mapping (d) of FO80-2, low (e) and high (f) magnification TEM images, HR-TEM image with FFT as the inset (g), and elemental mapping (h) of FO80-3, low (i) and high (j) magnification TEM images, HR-TEM image with FFT as the inset (k), and elemental mapping (l) of FO120-2, and low (m) and high (n) magnification TEM images, HR-TEM image with FFT as the inset (o), and elemental mapping (p) of FO120-3.

To get clear insight into the material's morphology, the T-CVD coated post-calcination film (FO120-2) obtained by hydrothermal processing at 120 °C was subjected to prolonged ultrasonication before drop casting on the TEM grid. In Fig. 6(i) is presented the TEM image which shows tiny particles arranged into a layer. The high magnification TEM image in Fig. 6(j) shows that the particles are consisting of a core (average size ∼73 nm) and a shell (average thickness ∼9.5 nm). The shell is amorphous as shown by the blurry FFT pattern presented as the inset in Fig. 6(j). The core on the other hand is crystalline and is supported by the very clear fringes of 0.300 nm and 0.253 nm corresponding to the interplanar spacings of lattice planes (220) and (311), respectively, of cubic Fe₃O₄ (Fig. 6(k)). Furthermore, crystallinity of the core is also evident from its FFT pattern comprising of clear dots as shown as the inset in Fig. 6(k). The core and shell structure of the particles is also evident from the elemental mapping images of Fe, O and C in Fig. 6(l). The EDX line scan profile along the line depicted in Fig. S20(e) shows a clear surge in carbon intensity at the extremities of the particles while at the center the Fe and O intensities are stronger. The morphology of the MP-CVD carbon coated film material is no different from that of the T-CVD carbon coated film material, which is evident from the TEM, HR-TEM and elemental mapping images presented in Fig. 6(m) and (n), (o) and (p), respectively. However, the EDX line scan profile along the line depicted in Fig. S20 (f) shows an increase in the intensity of Fe in comparison to the intensity of O exactly in the same fashion and for the same reason as was the case with MP-CVD carbon coated film FO80-3. Furthermore, the C intensity is increased at the extremities while it is almost uniform in the middle, indicating overlapping of the Fe₃O₄ particle by the carbon shell.

Electrochemical behavior of the samples was studied by cyclic voltammetry (CV) and the pertinent data are presented in Fig. 7. The voltammograms were measured while applying cut-off potentials of −1.0 V and 0 V at different scan rates (ν) ranging from 5 mV s⁻¹ up to 150 mV s⁻¹ in a 1 molar Na₂SO₄ solution. Cyclic voltammograms of the as-prepared Fe₃O₄ film (FO80-1) which was obtained by calcination of the as-received film from hydrothermal processing at 80 °C are presented in Fig. S21(a), showing nearly identical voltammograms to the previously reported ones.⁸⁶ Besides that Fe₃O₄ is generally considered a negative electrode material of pseudocapacitive nature, the nearly symmetrical nature of the curves with redox peaks supports the fact that the film possesses pseudocapacitive properties. The broad redox peaks could be due to the Fe^II/Fe^III redox reactions which may also be accompanied by the (de)intercalation of Na⁺ ion to maintain the local charge neutrality.^86,87 The phenomenon which was also previously observed for Fe₃O₄ and other charge storage materials is depicted by the following chemical reaction (Eqn 4), where x is the mole fraction of Na⁺ ions (de)intercalated:^86,88–90

Fe₃O₄ + xe⁻ + xNa⁺ → Na_xFe₃O₄ (4)

Furthermore, the current intensity increases with increasing ν but the shape remains almost unchanged even up to ν of 150 mV s⁻¹, indicating that the Fe₃O₄ film has good rate capability with rapid redox reaction occurring on the electrode.⁹¹ To understand the electrochemical kinetics of the Fe₃O₄ film, the current at fixed anodic and cathodic potentials of −0.243 V and −0.607 V, respectively, was plotted against the square root of varying ν. As shown in Fig. S21(b), the relationship between the current and the square root of ν is deviating from linear, indicating that the redox reaction is predominantly controlled by the capacitive process.⁹² To further understand the nature of the current, either capacitive or diffusion-controlled, CV data at various ν were analyzed by following the power law (eqn (5))⁹³

i = aν^b (5)

In this equation (eqn (5)), i is the current at a selected voltage V and ν is the scan rate while a and b are adjustable parameters. The contribution of current, either capacitive or diffusion-controlled, depends upon the b value, which is determined from the slope of the log(i) vs. log(ν) plot. When it is close to 1, the current is predominantly capacitive in nature, while when it is close to 0.5, the current is predominantly diffusion-controlled.⁹⁴ The value of b determined from the log(i) vs. log(ν) plot given in Fig. S21(c) is equal to ∼0.879, indicating that the contribution of current to the CV curve is predominantly from charge storage by the capacitive process. Furthermore, the percentage contribution to the CV curve from capacitive and diffusion-controlled processes was determined following the equation (eqn (6))^95,96

i(V) = k₁ν + k₂ν^1/2 (6)

where i(V) is the current at a given potential, and k₁ν and k₂ν^1/2 are, respectively, the capacitive-controlled (I_c) and diffusion-controlled (I_d) current fractions contributing to the total stored current. The equation could be further simplified by rearranging to the one shown below:where ν is the scan rate and k₁ and k₂ are adjustable parameters, which can be determined from the slope and intercept of the i(V)/ν^1/2 vs. ν^1/2 plot, respectively (Fig. S21(d)). The percentage capacitive and diffusion-controlled current contributions are presented in Fig. S21(e), showing on the one hand higher capacitive contributions as compared to the diffusion-controlled contributions. On the other hand, it shows increasing capacitive current contribution with increasing scan rate, indicating that the capacitive charge storage process is dominant in storing the total charge by the Fe₃O₄ film (FO80-1).

Fig. 7 CV curves of FO80-2 (a) and FO80-3 (f) at varying ν of 5 mV s⁻¹ to 150 mV s⁻¹ in 1 M Na₂SO₄ in the potential range of 1.0 V to 0.0 V, peak current (i_p) vs square root of scan rate (ν^1/2) of FO80-2 (b) and FO80-3 (g), log(i) vs. log(ν) plots for determination of b values of FO80-2 (c) and FO80-3 (h), contribution proportions of diffusion-controlled and capacitive to the total charge storage at different scan rates of FO80-2 (d) and FO80-3 (i), and separation of diffusion-controlled and capacitive current at 100 mV s⁻¹ of FO80-2 (e) and FO80-3 (j).

The post-calcination film after coating with carbon by T-CVD was also followed by CV for exploration of its electrochemical characteristics, and the pertinent CV curves at the cut-off voltage range of −1.0 V to 0.0 V at varying ν are presented in Fig. 7(a). As depicted by the CV scans in Fig. 7(a), the carbon-coated film (FO80-2) shows higher charge storage than its precursor (FO80-1), which is evident from the comparison of areas under the closed patterns. Furthermore, it is also evident from the specific capacitance values, which increased significantly upon carbon coating of the film, as shown by Fig. S22(a). The increase in specific capacitance could be attributed to enhanced electrical conductivity on the one hand and, on the other hand, reduced charge-transfer resistance due to the carbon coating. Similar to the precursor film (FO80-1), CV curves of FO80-2 (carbon-coated film) show clear but remarkably enhanced redox peaks, indicating that the film has strong pseudocapacitive and/or diffusion-controlled capacity.⁹⁷ This corresponds to the vast literature citing Fe₃O₄ as a negative electrode material of pseudocapacitive nature.⁹⁷ The increase in the strength of current–potential response giving wider CV curves of the FO80-2 film as compared to its precursor FO80-1 film could be attributed to the carbon coating, which increases the conductivity on the one hand, and on the other hand, contributes to charge storage through its own electrochemical double layer (EDL) process.⁹⁸ Furthermore, the multiple redox peaks appearing in the CV curves of FO80-2 are attributed to the fact that the film underwent a number of valence state transformations under the applied potential window. The plots in Fig. 7(b) of peak current (i_p) of the prominent oxidation and reduction peaks as a function of the square root of ν show a linear relationship, indicating that the charge storage is dominated by a diffusion-controlled process. The anodic and cathodic i_p – (scan rate) plots were linearly fitted, and the measured correlation coefficient (R²) values of 98.2% and 98.7%, respectively, indicate that the film shows battery behavior while storing charge dominantly by a diffusion-controlled process.⁹⁹ Fig. 7(c) of the prominent anodic and cathodic peaks also shows a linear relationship, giving slopes (= b values according to eqn (5)) of 0.695 and 0.681, respectively. This further supports the fact that the diffusion-controlled process is the dominant mechanism for charge storage. As shown in Fig. 7(d), the percentage contribution of the diffusion-controlled process towards the total charge storage is higher in the case of the carbon-coated film (FO80-2) compared to its precursor film (FO80-1). Even at 100 mV s⁻¹, the percent contribution from the diffusion controlled process is half of the total charge storage as shown in Fig. 7(e). The enhanced contribution of charge storage by the diffusion controlled process could be attributed to the carbon coating as well as to the morphology of the film. The conductive carbon coating provides highways for rapid and efficient transport of electrons and ions, thus supporting the diffusion controlled charge storage process.¹⁰⁰ Additionally, as one can see in the TEM and SEM images presented in the preceding sections, the particles of carbon-coated films possess a large number of tiny holes compared to the particles of the precursor which have smooth surfaces. These holes could allow the electrolyte to reach deep into the bulk of the film particles, ensuring a pseudocapacitive process on the surface as well as in the bulk. In fact nanoporous structures facilitate rapid diffusion of electrons and ions, thus enhancing the share of charge storage by the diffusion controlled process.¹⁰¹ This is further evident from the EIS Nyquist plot presented in Fig. S22(a), showing a small semi-circle at the higher frequency end but a strong and steeper oblique line at an angle of approximately 45°. The semicircle represents the resistance to charge transfer at the interface, while the oblique line is the Warburg resistance, which is resistance to the diffusion of ions.¹⁰² This indicates that mostly the charge transfer occurs through a diffusion-controlled process, and thus the contribution of diffusion-controlled charge storage is higher than the capacitive contribution. The post-calcination film (FO80-1) upon coating with carbon by MP-CVD shows, in principle, identical CV curves in the potential window of −1.0 V to 0.0 V at varying ν as depicted in Fig. 7(f). The enclosed area under CV curves is larger compared to that of its precursor (FO80-1) and smaller compared to that of the T-CVD carbon coated film (FO80-2). The electrochemical charge storage mechanism is dominated by the pseudocapacitive process which is evident from the peaks in its CV scans. It is further supported by the linear relationship of i_p–(ν)^1/2 curves presented in Fig. 7(g), giving R² values close to 1 for the anodic and cathodic peaks. Following eqn (5), b values were determined by plotting log(i_p) versus log(ν) as shown in Fig. 7(h). It is also evident from the b values that the charge storage mechanism is dominated by a diffusion-controlled process. The percentage charge storage contribution from either capacitive or diffusion-controlled to the total was determined by following Eqn 7. As presented in Fig. 7(i), the charge storage mechanism at lower scan rates is dominated by diffusion-controlled amounting to approximately 62% at 5 mV s⁻¹, decreasing with increasing scan rate. As shown in Fig. 7(j), the diffusion controlled charge storage share in the case of FO80-3 (MP-CVD carbon coated Fe₃O₄ film) drops to ∼26% as compared to that of the T-CVD coated Fe₃O₄ film (∼46%) but is still higher by ∼19% compared to that of FO80-1 (bare Fe₃O₄ film). These observations signify the impact of carbon coating and the nanoporosity upon the charge storage mechanism.

The CV curves of FO120-1 (the calcined film derived from the film prepared by hydrothermal processing at 120 °C) are presented in Fig. S23(a), showing nearly rectangular shapes with an oxidation peak which increases in intensity with the increasing scan rate of up to 40 mV s⁻¹. Afterwards, this very peak decreases in intensity and finally adopts a rectangular shape at 100 mV s⁻¹. This very peak also shifts towards right with increasing scan rate indicating irreversible oxidation of the iron species.¹⁰³ Contrary to the uncoated Fe₃O₄ film (FO120-1), the carbon coated, by either T-CVD or MP-CVD, films show nearly rectangular shapes of their CV curves, corresponding to the typical CV curves of carbon coated Fe₃O₄ materials reported earlier.^101,104,105 The R² and b values obtained from the graphs of i vs. (ν)^1/2 and log(i) vs. log(ν), respectively, lie in the range (see Fig. 8(b) and (c) and (g) and (h)), respectively, for FO120-2 and FO120-3, which supports major contribution from the diffusion-controlled process to the total charge. The contribution proportions based on Eqn 7 at different scan rates in Fig. 8(d) and (i), respectively, for the films FO120-2 and FO120-3 show that the diffusion-controlled process is dominant at lower scan rates. Even at 100 mV s⁻¹, the contribution from the diffusion-controlled process is almost half of the total charge stored.

Fig. 8 CV curves of FO120-2 (a) and FO120-3 (f) at varying ν of 5 mV s⁻¹ to 150 mV s⁻¹ in 1M Na₂SO₄ in the potential range of 1.0 V to 0.0 V, peak current (i_p) vs. square root of scan rate (ν^1/2) of FO120-2 (b) and FO120-3 (g), log(i) vs. log(ν) plots for determination of b values of FO120-2 (c) and FO120-3 (h), contribution proportions of diffusion-controlled and capacitive to the total charge storage at different scan rates of FO120-2 (d) and FO120-3 (i), and separation of diffusion-controlled and capacitive current at 100 mV s⁻¹ of FO120-2 (e) and FO120-3 (j).

Based on the presented results, the deposition of films has been depicted by an illustration given in Fig. 9. As reported elsewhere,¹⁰⁶ and observed here by us, the phase and morphological evolution of the films was found to be highly sensitive to the hydrothermal environment, which includes the autoclave pressure, hydrolysis kinetics of urea and substrate-directed nucleation. Urea undergoes competitive reactions, i.e., oxidative hydrolysis and carbonation, depending on the hydrothermal process temperature. These reactions are given by the following chemical equations (eqn (8)–(10)):^107,108

NH₃ + H₂O → NH₄⁺ + OH⁻ (9)

CO₂ + H₂O → H₂CO₃ → HCO₃⁻ + H⁺ → CO₃²⁻ + H⁺ (10)

As evident from these equations, urea has served as a pH-regulating agent at both the temperatures, i.e., 80 °C and 120 °C, by releasing hydroxyl (OH⁻) and carbonate (CO₃²⁻) ions. In addition to these ions, the possible formation of transient metal fluoride complexes such as [FeF_n]⁽²⁻ⁿ⁾ may have moderated the process by preventing the premature precipitation.⁴² Mostly derived from ionization of urea, all these ions have served as precursors for realization of either FeOOH (80 °C) or FeCO₃ (120 °C) after reacting with the iron ions (Fe²⁺) under the applied hydrothermal processing conditions. Initially, the Fe²⁺ ions partially oxidized to Fe(OH)₃ during the solution formation under ambient conditions (eqn (11)).²⁶

Fe²⁺ + O₂ + H₂O → Fe(OH)₃ (11)

Fe(OH)₃ has strong potential for anchoring, and thus adhered to the surface of the copper (Cu) substrate, serving as seeds for the nucleation process.⁴² As the hydrothermal processing at 80 °C proceeded, the reaction occurred between the leftover Fe²⁺ ions and OH⁻ ions, which were produced during urea hydrolysis, giving FeOOH (Eqn 12). Furthermore, Fe(OH)₃ produced in the previous step gets converted to FeOOH (eqn (13)).²⁶

4Fe²⁺ + 8OH⁻ + O₂ → 4FeOOH + 2H₂O (12)

Fe(OH)₃ → FeOOH + 2H₂O (13)

On the other hand, upon hydrothermal processing at 120 °C, besides the formation of Fe(OH)₃ and FeOOH, precipitation of Fe(CO₃) also occurred. It is pertinent to mention here that the formation of metal carbonates in hydrothermal processing at higher temperatures in the presence of urea is not without precedence.^44,107,108

Fig. 9 Illustration of the process for preparation of the resulting films.

An interesting situation was observed in the case of hydrothermal processing at 80 °C. At this temperature, α-FeOOH and β-FeOOH, respectively, were predominantly formed at the substrate and in the bulk solution as a result of oxidative hydrolysis of Fe²⁺ ions (eqn (12)). This indicates that the Cu foil surface, i.e., the substrate surface, supported the growth of the α-FeOOH phase, while a suitable environment was provided by the solution for the growth of the β-FeOOH phase. This is attributed to the difference in nucleation energies at the two different places, i.e., at the surface of the substrate (Cu foil) and in the solution. The heterogeneous surface provided by the Cu substrate stabilizes the thermodynamically favoured α-FeOOH phase (Goethite) through the potential lattice matching with the copper oxide surface layer which is inevitably formed during hydrothermal processing.¹⁰⁹ In fact, the preferential growth of the α-FeOOH phase on the substrate is due to the lower interfacial energy provided by the Cu substrate surface.¹⁰⁹ On the other hand, the solution contains F⁻ ions furnished by the NH₄F precursor. These ions incorporate into the β-FeOOH phase (Akaganeite), which has a tunnel-type structure, and thus stabilize it. Therefore, owing to the availability of a high local concentration of F⁻ ions in solution for formation of tunnel structures of FeOOH, the powder has a dominant β-FeOOH phase.¹¹⁰ It is pertinent to mention here that the achievement of a single phase film, i.e., α-FeOOH, is contrary to the mixed phases, i.e., α-FeOOH/β-FeOOH, reported by Cao et al.,²⁶ although the experimental conditions were exactly the same except the autoclave size. Besides, there is a marked difference in the morphology of the films. Instead of the 3D framework of layered nanosheets of Cao et al.,²⁶ we obtained 1D rhombus-shaped nanorods. The difference in morphology is then due to the autoclave size as all other hydrothermal processing parameters were the same.

Rhombohedra were commonly achieved owing to the oriented attachment growth.¹¹¹ It was observed that the realization of rhombohedra at the expense of other structures through a dissolution–reprecipitation process enhances with the increasing amount of ammonia (NH₃). By using a larger vessel for the same amount of solution, the filling ratio (f) of the autoclave decreased from 70% in the case of Cao et al.²⁶ to 56% in the present work, leading to the creation of lower internal pressure.¹¹² The reduced internal pressure then facilitated and accelerated the urea hydrolysis reaction (eqn (8)),¹¹³ producing an increased amount of NH₃, which then ultimately favoured the formation of anisotropic 1D rhombohedral structures. It is also pertinent to mention here that the role of F⁻ as a structure directing agent could not be ignored, as previously it was reported that F⁻ is critical in the realization of 1D rhombus nanorods upon diverse substrates at hydrothermal temperatures below 100 °C.^42–44,114

By increasing the temperature from 80 °C to 120 °C, the conditions resembling those of Cao et al.²⁶ may have been achieved even with the larger size autoclave, leading to the formation of the 3D framework of layered nanosheets. However, the increase in hydrothermal processing temperature also shifted the thermodynamic stability. As a result, a phase transition occurred from oxyhydroxides at 80 °C to a mixture of Fe₃O₄ and FeCO₃ at 120 °C.⁴⁴ In fact, urea hydrolysis was further accelerated by increasing the hydrothermal temperature from 80 °C to 120 °C which increased the concentration of CO₃²⁻ (eqn (10)), shifting the equilibrium toward the precipitation of FeCO₃ (eqn (14)).¹⁰⁸

Fe²⁺ + CO₃²⁻ → FeCO₃ (14)

Interestingly, the film contains significantly higher FeCO₃ content (∼35%) than the precipitated powder (∼8%). This highlights the role of the Cu substrate in stabilizing FeCO₃ nucleation under the applied hydrothermal conditions. Nonetheless, Fe₃O₄ is the dominant component in the film as well as in the precipitate. This is due to the reaction between FeOOH and Fe(OH)₂ suspension under decreasing concentration of CO₃²⁻.¹⁰⁸

2FeOOH + Fe(OH)₂ → Fe₃CO₄ + H₂O (15)

From these observations, the role of hydrothermal temperature in the phase and morphology of the nanostructures is obviously demonstrated.

Upon calcination at 450 °C, all the as-synthesized precursors underwent a topotactic transformation to Fe₃O₄. The fundamental morphologies, i.e., the rhombus-shaped 1D nanorods and the 3D framework of layered nanosheets, respectively, from the 80 °C and 120 °C synthesis were preserved. This indicates the robustness of the nanostructures. However, tiny holes appear upon calcination at 450 °C due to the decomposition and degassing of the volatile species.²⁶ These species are in fact the chemically bound water and CO₂ molecules, which evolve as a result of the decomposition of FeOOH and FeCO₃, respectively, upon heating.

As no reduction was induced by T-CVD, the Fe₃O₄ films were partially reduced to metallic iron (Fe) during carbon coating by MP-CVD. This was inevitable as the strong hydrogen plasma alongside the carbon precursor produces a more aggressive reducing environment.¹¹⁵ Despite the reduction induced by MP-CVD, the structural framework of the Fe₃O₄ nanostructures was robust enough to preserve the basic morphology. The preservation of morphology and tiny holes during the carbon coating by T- and MP-CVD is critical from the performance point of view. They maximize the surface area, ensuring high electrolyte accessibility in electrochemical applications. The effect of these factors has been clearly demonstrated by the synthesized films during their electrochemical characterization.

Conclusion

In conclusion, it is stated that carbon encapsulated Fe₃O₄ films as binder-free electrodes for potential application in energy storage devices were prepared directly on the Cu current collector by a scalable approach, involving hydrothermal growth in the first step, followed by calcination in the second step and finally carbon coating by CVD in the third step. The results showed that the hydrothermal parameter, i.e., temperature, has a marked influence on the phase and morphology of the films. It was found that the hydrothermal processing of the desired reactants at 80 °C gives rhombus-shaped α-FeOOH nanostructures. At a higher hydrothermal processing temperature of 120 °C, the phase and morphology of the films were entirely different, i.e., nanolayers of the Fe₃O₄/FeCO₃ composite arranged into 3D superstructures. The nanostructures were stable and robust enough that the basic morphology was retained after calcination and even after carbon coating by CVD techniques. Although the morphology was not affected by either calcination or carbon coating, the crystallographic phase changed from FeOOH to the desired phase, i.e., Fe₃O₄. Both the CVD techniques, i.e., T- and MP-CVD, gave well-defined Fe₃O₄@C core–shell nanoarchitectures having strong interfacial interactions through the formation of Fe–O–C bridges. The crystallinity was further improved by MP-CVD, which also induced partial reduction of Fe₃O₄ to Fe. The electrochemical characterization of the samples showed that the charge storage predominantly occurs by a diffusion-controlled process at lower scan rates, while by a capacitive process at higher scan rates. In addition to the improved charge storage, it was found that the charge transfer resistance is lower for the carbon-encapsulated binder-free electrodes. Furthermore, these electrodes have shown superior stability. The improvements in electrochemical characteristics were attributed to the synergistic benefits of the Cu substrate which provided direct electron pathways, nanostructuration which shortened the diffusion pathlength, and the carbon shells which provided a conductive protection. In a nutshell, it is concluded that the key limitations of Fe₃O₄ electrodes were addressed through a thoroughly explored strategy which includes control over film morphology, binder-free architecting, and carbon encapsulation. Therefore, we confidently say that the strategy offers a promising platform for realization of high rate and long-life anodes for next generation electrochemical energy storage devices, which include but not limited to batteries, supercapacitors and hybrid devices. Additionally, the rationally designed and viably realized strategy will pave the way for fabrication of metal-oxide-based electrodes which show improved performance on the one hand, and on the other hand, potential of scalability.

Author contributions

All the authors contributed to the development of the manuscript. H. Ullah and S. Mortazavi conducted the experiments and investigated the processes. H. Ullah, S. Mortazavi, and A. M. Flatae acquired the data using various techniques. They also analysed the data. H. Ullah prepared the initial draft of the manuscript. A. M. Khan, X. Jiang and all the above mentioned authors contributed to the reviewing, editing and approving of the manuscript for publication.

Conflicts of interest

There are no conflicts to declare.

Data availability

All the data supporting the findings of this manuscript are included in the manuscript and the associated supplementary information (SI) files. Supplementary information is available. See DOI: https://doi.org/10.1039/d6ya00012f.

Acknowledgements

H. Ullah acknowledges the financial support received from the European Union's Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No. 945422. We also acknowledge the German Research Foundation (DFG) (INST 221/118-1 FUGG) for financial support. The authors greatly acknowledge “Gerätezentrum für Mikro- und Nanoanalytik MNaF, Siegen University, Germany” for providing support for the analysis of samples by FE-SEM, TEM, XRD, XPS, etc.

References

1. N. Khan, S. Dilshad, R. Khalid, A. R. Kalair and N. Abas, Energy Storage, 2019, 1, e49.
2. G. Smdani, M. R. Islam, A. N. Ahmad Yahaya and S. I. Bin Safie, Energy Environ., 2023, 34, 1094–1141.
3. J. Lu, Z. Chen, F. Pan, Y. Cui and K. Amine, Electrochem. Energy Rev., 2018, 1, 35–53.
4. J. B. Goodenough and K.-S. Park, J. Am. Chem. Soc., 2013, 135, 1167–1176.
5. M. Li, J. Lu, Z. Chen and K. Amine, Adv. Mater., 2018, 30, 1800561.
6. L. Zhao, B. Ding, X.-Y. Qin, Z. Wang, W. Lv, Y.-B. He, Q.-H. Yang and F. Kang, Adv. Mater., 2022, 34, 2106704.
7. N. Nitta, F. Wu, J. T. Lee and G. Yushin, Mater. Today, 2015, 18, 252–264.
8. V. Patil, A. Patil, S.-J. Yoon and J.-W. Choi, Sci. Adv. Mater., 2018, 10, 507–512.
9. M. V. Reddy, G. V. Subba Rao and B. V. R. Chowdari, Chem. Rev., 2013, 113, 5364–5457.
10. M. M. Rahman, U. Nisar, A. Abouimrane, I. Belharouak and R. Amin, Electrochem. Energy Rev., 2025, 8, 14.
11. J. Zhu, Y. Ding, Z. Ma, W. Tang, X. Chen and Y. Lu, J. Electron. Mater., 2022, 51, 3391–3417.
12. C. Zhao, S. Yao, C. Li, Y. An, S. Zhao, X. Sun, K. Wang, X. Zhang and Y. Ma, Chem. Eng. J., 2024, 497, 154535.
13. H. Ren, H. Li, P. Barry, Z. Wang, A. R. Campos, E. S. Takeuchi, A. C. Marschilok, S. Yan, K. J. Takeuchi and E. Reichmanis, Chem. Mater., 2024, 36, 9299–9319.
14. J. Guan, Y. Chen, L. Cao, Y. Liu, P. Lian, Y. Gao and X. Shi, J. Power Sources, 2020, 469, 228307.
15. N. Manikandan, B. Lakshmi and S. Shivakumara, J. Solid State Electrochem., 2022, 26, 887–895.
16. J. Ma, Y. Xu, Y. Xu, Y. Zhang, M. Zhang, Z. Cao and C. Liu, J. Energy Storage, 2025, 108, 115164.
17. J. Tu, H. Tong, P. Wang, D. Wang, Y. Yang, X. Meng, L. Hu, H. Wang and Q. Chen, Small, 2023, 19, 2301606.
18. J. P. Pender, G. Jha, D. H. Youn, J. M. Ziegler, I. Andoni, E. J. Choi, A. Heller, B. S. Dunn, P. S. Weiss, R. M. Penner and C. B. Mullins, ACS Nano, 2020, 14, 1243–1295.
19. Z. Gao, S. Rao, J. Wang, D. Wang, T. Zhang, X. Feng, Y. Liu, J. Shi, Y. Xue, W. Li, L. Wang, C. Rong and Y. Chen, ChemSusChem, 2024, 17, e202400830.
20. D. Deng, Energy Sci. Eng., 2015, 3, 385–418.
21. X. Chen, X. Zhu, G. Cao, S. Zhang, Y. Mu, H. Ming and J. Qiu, Energy Fuels, 2021, 35, 1810–1819.
22. K. Shen, S. Zhai, S. Wang, Q. Ru, X. Hou, K. San Hui, K. Nam Hui and F. Chen, Batteries Supercaps, 2021, 4, 860–880.
23. Y. Kang, C. Deng, Y. Chen, X. Liu, Z. Liang, T. Li, Q. Hu and Y. Zhao, Nanoscale Res. Lett., 2020, 15, 112.
24. O. Gerard, A. Numan, S. Krishnan, M. Khalid, R. Subramaniam and R. Kasi, J. Energy Storage, 2022, 50, 104283.
25. Y. Yang, J. Li, D. Chen and J. Zhao, ACS Appl. Mater. Interfaces, 2016, 8, 26730–26739.
26. K. Cao, L. Jiao, H. Liu, Y. Liu, Y. Wang, Z. Guo and H. Yuan, Adv. Energy Mater., 2015, 5, 1401421.
27. J. Chen, K. Huang and S. Liu, Electrochim. Acta, 2009, 55, 1–5.
28. F. Bojabady, E. Kamali-Heidari and S. Sahebian, Mater. Chem. Phys., 2023, 305, 127828.
29. D. Q. Han, F. Yang, Y. Y. Gong and S. G. Yang, Surf. Eng., 2015, 31, 481–485.
30. C. He, S. Wu, N. Zhao, C. Shi, E. Liu and J. Li, ACS Nano, 2013, 7, 4459–4469.
31. L. Wang, F. Liu, A. Pal, Y. Ning, Z. Wang, B. Zhao, R. Bradley and W. Wu, Carbon, 2021, 179, 327–336.
32. J. Ryu, J. Park, W.-G. Lim and J. Hwang, Appl. Surf. Sci., 2024, 675, 160976.
33. Z. Liu, X.-Y. Yu and U. Paik, Adv. Energy Mater., 2016, 6, 1502318.
34. M. Li, H. Du, L. Kuai, K. Huang, Y. Xia and B. Geng, Angew. Chem., Int. Ed., 2017, 56, 12649–12653.
35. Z. Zeng, H. Zhao, J. Wang, P. Lv, T. Zhang and Q. Xia, J. Power Sources, 2014, 248, 15–21.
36. E. Ngondo, X. Fuku, M. A. Kebede and N. W. Hlongwa, ChemistrySelect, 2025, 10, e02667.
37. J. Wang, M. Gao, D. Wang, X. Li, Y. Dou, Y. Liu and H. Pan, J. Power Sources, 2015, 282, 257–264.
38. B. Wang, G. Wang, Z. Zheng, H. Wang, J. Bai and J. Bai, Electrochim. Acta, 2013, 106, 235–243.
39. A. T. Adeleye, K. I. John, P. G. Adeleye, A. A. Akande and O. O. Banjoko, J. Mater. Sci., 2021, 56, 18391–18416.
40. T. Sheng, Y.-F. Xu, Y.-X. Jiang, L. Huang, N. Tian, Z.-Y. Zhou, I. Broadwell and S.-G. Sun, Acc. Chem. Res., 2016, 49, 2569–2577.
41. S. E. F. Kleijn, S. C. S. Lai, M. T. M. Koper and P. R. Unwin, Angew. Chem., Int. Ed., 2014, 53, 3558–3586.
42. W. Mei, J. Huang, L. Zhu, Z. Ye, Y. Mai and J. Tu, J. Mater. Chem., 2012, 22, 9315–9321.
43. Z. Wen, L. Zhu, W. Mei, L. Hu, Y. Li, L. Sun, H. Cai and Z. Ye, Sens. Actuators, B, 2013, 186, 172–179.
44. Z. Wen, L. Zhu, W. Mei, Y. Li, L. Hu, L. Sun, W. Wan and Z. Ye, J. Mater. Chem. A, 2013, 1, 7511–7518.
45. J. Rodríguez-Carvajal, Phys. B, 1993, 192, 55–69.
46. L. B. McCusker, R. B. Von Dreele, D. E. Cox, D. Louër and P. Scardi, J. Appl. Crystallogr., 1999, 32, 36–50.
47. A. Kumar and D. Varshney, Ceram. Int., 2012, 38, 3935–3942.
48. S. Debnath, K. Deb, B. Saha and R. Das, J. Phys. Chem. Solids, 2019, 134, 105–114.
49. M. Kuru, T. Ş. Kuru, E. Karaca and S. Bağcı, J. Alloys Compd., 2020, 836, 155318.
50. K. R. Desai, S. T. Alone, S. R. Wadgane, S. E. Shirsath, K. M. Batoo, A. Imran, E. H. Raslan, M. Hadi, M. F. Ijaz and R. H. Kadam, Phys. B, 2021, 614, 413054.
51. N. S. Gonçalves, J. A. Carvalho, Z. M. Lima and J. M. Sasaki, Mater. Lett., 2012, 72, 36–38.
52. W. Kim, C.-Y. Suh, S.-W. Cho, K.-M. Roh, H. Kwon, K. Song and I.-J. Shon, Talanta, 2012, 94, 348–352.
53. S. P. Schwaminger, C. Syhr and S. Berensmeier, Crystals, 2020, 10, 214.
54. A. Gholizadeh, J. Am. Ceram. Soc., 2017, 100, 3577–3588.
55. I. Lungu, T. Potlog, M. Dobromir and S. Moldovan, ChemistrySelect, 2026, 11, e05915.
56. W. D. Chemelewski, H.-C. Lee, J.-F. Lin, A. J. Bard and C. B. Mullins, J. Am. Chem. Soc., 2014, 136, 2843–2850.
57. M. V. Abrashev, V. G. Ivanov, B. S. Stefanov, N. D. Todorov, J. Rosell and V. Skumryev, J. Appl. Phys., 2020, 127, 205108.
58. X. Wu, J. Tang, Y. Zhang and H. Wang, Mater. Sci. Eng. B, 2009, 157, 81–86.
59. S. Tiwari, R. Prakash, R. J. Choudhary and D. M. Phase, J. Phys. Appl. Phys., 2007, 40, 4943.
60. C. Ruby, B. Humbert and J. Fusy, Surf. Interface Anal., 2000, 29, 377–380.
61. Y. Liang and W. Lu, J. Mater. Sci.: Mater. Electron., 2020, 31, 17075–17083.
62. T. Muraliganth, A. V. Murugan and A. Manthiram, Chem. Commun., 2009, 7360–7362.
63. T. Wang, Z. Jiang, K. H. Chu, D. Wu, B. Wang, H. Sun, H. Y. Yip, T. An, H. Zhao and P. K. Wong, ChemSusChem, 2018, 11, 1365–1373.
64. W. Luo, C. Jiang, Y. Li, S. A. Shevlin, X. Han, K. Qiu, Y. Cheng, Z. Guo, W. Huang and J. Tang, J. Mater. Chem. A, 2017, 5, 2021–2028.
65. B. G. Kim, J. Park, W. Choi, D. S. Han, J. Kim and H. Park, Appl. Catal., B, 2021, 283, 119608.
66. T. Yamashita and P. Hayes, Appl. Surf. Sci., 2008, 254, 2441–2449.
67. C. Long, L. Jiang, T. Wei, J. Yan and Z. Fan, J. Mater. Chem. A, 2014, 2, 16678–16686.
68. M. C. Biesinger, B. P. Payne, A. P. Grosvenor, L. W. M. Lau, A. R. Gerson and R. S. C. Smart, Appl. Surf. Sci., 2011, 257, 2717–2730.
69. A. Bak, W. Choi and H. Park, Appl. Catal., B, 2011, 110, 207–215.
70. Z. Cai, D. Zhou, M. Wang, S.-M. Bak, Y. Wu, Z. Wu, Y. Tian, X. Xiong, Y. Li, W. Liu, S. Siahrostami, Y. Kuang, X.-Q. Yang, H. Duan, Z. Feng, H. Wang and X. Sun, Angew. Chem., Int. Ed., 2018, 57, 9392–9396.
71. R. Pai, A. Singh, S. Simotwo and V. Kalra, Adv. Eng. Mater., 2018, 20, 1701116.
72. A. P. Grosvenor, B. A. Kobe and N. S. McIntyre, Surf. Sci., 2004, 572, 217–227.
73. A. Barbieri, W. Weiss, M. A. Van Hove and G. A. Somorjai, Surf. Sci., 1994, 302, 259–279.
74. T. Fujii, F. M. F. de Groot, G. A. Sawatzky, F. C. Voogt, T. Hibma and K. Okada, Phys. Rev. B: Condens. Matter Mater. Phys., 1999, 59, 3195–3202.
75. W. Lu, Y. Shen, A. Xie and W. Zhang, J. Magn. Magn. Mater., 2010, 322, 1828–1833.
76. M. Khan, N. Shahzad, C. Xiong, T. K. Zhao, T. Li, F. Siddique, N. Ali, M. Shahzad, H. Ullah and S. A. Rakha, Diamond Relat. Mater., 2016, 61, 32–40.
77. Q. Wu, R. Zhao, X. Zhang, W. Li, R. Xu, G. Diao and M. Chen, J. Power Sources, 2017, 359, 7–16.
78. M. Yu, L. Sun and X. Ning, J. Alloys Compd., 2021, 878, 160359.
79. Y. Yan, X. Lu, Y. Li, J. Song, Q. Tian, L. Yang and Z. Sui, J. Alloys Compd., 2022, 899, 163342.
80. R. Kumar, R. K. Nekouei and V. Sahajwalla, Ceram. Int., 2025, 51, 12312–12320.
81. X. Hu and J. C. Yu, Chem. – Asian J., 2006, 1, 605–610.
82. X. Chen, X. Wang and D. Fang, Fullerenes, Nanotubes Carbon Nanostruct., 2020, 28, 1048–1058.
83. L. Li, A. Kovalchuk, H. Fei, Z. Peng, Y. Li, N. D. Kim, C. Xiang, Y. Yang, G. Ruan and J. M. Tour, Adv. Energy Mater., 2015, 5, 1500171.
84. J.-L. Hazemann, J. F. Bérar and A. Manceau, Materials Science Forum, Trans Tech Publications, Ltd., 1991, vol. 79, pp. 821–826.
85. A. A. Claassen, Proc. Phys. Soc., London, 1925, 38, 482.
86. A. Chouchaine, S. Kouass, F. Touati, N. Amdouni and H. Dhaouadi, J. Aust. Ceram. Soc., 2021, 57, 469–477.
87. C. D. Wessells, R. A. Huggins and Y. Cui, Nat. Commun., 2011, 2, 550.
88. K. Lu, D. Li, X. Gao, H. Dai, N. Wang and H. Ma, J. Mater. Chem. A, 2015, 3, 16013–16019.
89. Z. Chen, V. Augustyn, X. Jia, Q. Xiao, B. Dunn and Y. Lu, ACS Nano, 2012, 6, 4319–4327.
90. L. Su, J. Winnick and P. Kohl, J. Power Sources, 2001, 101, 226–230.
91. X. Tang, R. Jia, T. Zhai and H. Xia, ACS Appl. Mater. Interfaces, 2015, 7, 27518–27525.
92. J. Zhu, Q. Zhang, S. Yang, L. Chen, P. Zhao and Q. Yan, ChemElectroChem, 2021, 8, 4817–4825.
93. R. Li and J. Liu, Electrochim. Acta, 2014, 120, 52–56.
94. H. Yin, C. Song, Y. Wang, S. Li, M. Zeng, Z. Zhang, Z. Zhu and K. Yu, Electrochim. Acta, 2013, 111, 762–770.
95. J. Xie, Y. Zhu, N. Zhuang, H. Lei, W. Zhu, Y. Fu, M. S. Javed, J. Li and W. Mai, Nanoscale, 2018, 10, 17092–17098.
96. H. He, D. Huang, Y. Tang, Q. Wang, X. Ji, H. Wang and Z. Guo, Nano Energy, 2019, 57, 728–736.
97. J. Sun, P. Zan, X. Yang, L. Ye and L. Zhao, Electrochim. Acta, 2016, 215, 483–491.
98. M. Rezapour, Anal. Bioanal. Electrochem., 2018, 10, 450–464.
99. D. Muthu, T. Sadhasivam and T. H. Oh, Mater. Sci. Eng. B, 2024, 299, 116972.
100. N. Sinan and E. Unur, Mater. Chem. Phys., 2016, 183, 571–579.
101. Z. Cheng, T. Du, L. Zhou, H. Wang and L. Cao, J. Colloid Interface Sci., 2025, 700, 138347.
102. H. Wang, Y. Ma and W. Zhang, Nanomaterials, 2021, 11, 2203.
103. S.-Y. Wang, K.-C. Ho, S.-L. Kuo and N.-L. Wu, J. Electrochem. Soc., 2005, 153, A75.
104. J. Yu, B. Wang, Q. Lu, L. Xiao, X. Ma, Y. Feng and Y. Qian, Electrochim. Acta, 2022, 427, 140843.
105. G. Bharath, K. Rambabu, F. Banat, A. Hai, A. F. Arangadi and N. Ponpandian, Sci. Total Environ, 2019, 691, 713–726.
106. Q. Liu, D. Mao, C. Chang and F. Huang, J. Power Sources, 2007, 173, 538–544.
107. F. I. Saaid, M. F. Kasim, C.-C. Yang, A. Kumar, T.-Y. Tseng and T. Winie, J. Mater. Sci.: Mater. Electron., 2022, 33, 3648–3669.
108. T. Yang, Z. Huang, Y. Liu, M. Fang, X. Ouyang and M. Hu, Ceram. Int., 2014, 40, 11975–11983.
109. R. M. Cornell and U. Schwertmann, The Iron Oxides: Structure, Properties, Reactions, Occurrences and Uses, Wiley, 1st edn, 2003.
110. N. K. Chaudhari and J.-S. Yu, J. Phys. Chem. C, 2008, 112, 19957–19962.
111. M. Lin, H. R. Tan, J. P. Y. Tan and S. Bai, J. Phys. Chem. C, 2013, 117, 11242–11250.
112. K. Byrappa and M. Yoshimura, Handbook of Hydrothermal Technology, Elsevier Inc., 2nd edn, 2013.
113. F. Zhu, J. Gao, X. Chen, M. Tong, Y. Zhou and J. Lu, Ind. Eng. Chem. Res., 2015, 54, 9072–9080.
114. F. Xu, Y. Lu, L. Sun and L. Zhi, Chem. Commun., 2010, 46, 3191–3193.
115. T. Maki, H. Kura, H. Ishida, T. Kaneko, R. Hatakeyama, M. Takahashi and T. Ogawa, J. Phys.: Conf. Ser., 2011, 266, 012120.

---