Hierarchical micro-tile growth of monoclinic tungsten oxide nucleated on MWCNTs hexagonal skeletons: wide-potential solid-state supercapacitor with a mechanical bendable design

Abstract

Material mutualism in the growth process has facilitated the formation of monoclinic tungsten oxide (W₂₅O₇₃) on hexagonal multi-walled carbon nanotubes (MWCNTs), which act as self-sacrificing templates to produce a controlled micro-tile surface architecture through a simple chemical route. The monoclinic crystal structure of tungsten oxide (W₂₅O₇₃) was confirmed by using X-ray diffraction and high-resolution transmission electron microscopy. The three-electrode electrochemical analysis of the MWCNTs/W₂₅O₇₃ film revealed a high specific capacitance of 1301.46 F g⁻¹ (areal capacitance = 273.31 mF cm⁻²) at a scan rate of 1 mV s⁻¹ with a potential window of 0.94 V in 1 M LiClO₄. The designed flexible symmetric solid-state supercapacitor device displayed a remarkable voltage window of 1.86 V aided by Li⁺ ions embedded in the polyvinyl alcohol polymer matrix with a specific capacitance of 259.24 F g⁻¹ (areal capacitance = 54.44 mF cm⁻²) at 5 mV s⁻¹ scan rate. Even after 10 000 consecutive cycles, the device preserved 82.78% of its initial capacitance, confirming reliable long-term operational stability. An in-depth examination was conducted using power law calculations, diffusion and capacitive contributions, and electrochemical impedance spectroscopy analysis. The device exhibited an excellent mechanical flexibility of 99.62% at a bending angle of 170° along with practical viability in powering a small DC fan and LED panel, showcasing its potential for future advanced applications.

---

1. Introduction

Supercapacitors have gained prominence as next-generation energy storage systems, which is evident from their high power density, ultra-fast charge–discharge abilities, and prolonged cycling stability, making them well-suited for applications demanding rapid energy delivery.¹ Although lithium-ion batteries continue to dominate the commercial landscape due to their high energy density, their practical deployment is hindered by slow charging kinetics, performance degradation under high-current operations, and relatively limited cycle life.² On the contrary, supercapacitors demonstrate superior rate capability and long-term durability. Conventional devices that employ liquid electrolytes face critical challenges, including leakage, flammability, thermal instability, solvent decomposition, sealing inefficiencies, and potential chemical reactivities.² In addition, aqueous electrolytes are restricted to a voltage window of less than 1.23 V due to water dissociation, further limiting energy density and overall device reliability.³ To address these limitations, solid-state supercapacitors have emerged as leakage-free, environmentally friendly alternatives with enhanced chemical stability, longer shelf-life, and improved electrochemical performance. These devices offer enhanced safety, electrochemical robustness, and mechanical flexibility. The accelerating progress of wearable and portable electronic technologies, particularly in healthcare, has further intensified the demand for such devices.⁴

Surface morphology plays a decisive role in governing electrochemical performance. To improve the charge storage capability and charge–discharge behavior, strategies such as reducing the crystallite, grain or particle size, introducing structural defects, and increasing accessible surface area are often employed.⁵ Tailoring the morphology of zero-, one- and two-dimensional nanostructures offers a significant enhancement in surface area, thereby increasing the utilization of electroactive materials and boosting the specific capacitance in energy storage devices. Such nanostructures typically exhibit high porosity, efficient charge transport pathways, and superior ion diffusion, all of which contribute to improved electrochemical performance. Recent advancements emphasize the engineering of nanostructured electrode materials, such as nanoparticles, nanoflakes, nanowires, and nanosheets, which provide abundant electroactive sites, improved ion transport pathways, and enhanced electrochemical stability.⁶ Among them, 2D porous architectures stand out due to their unique characteristics, including accelerated ion diffusion, improved adsorption, and enhanced electroactive surface area, derived from their extended surface architecture.⁷

Tungsten oxide has recently been extensively investigated for a wide range of optical, memory storage, energy conversion, catalytic, sensor and energy storage applications owing to its superior physicochemical characteristics.⁸ Tungsten oxide is recognized as an environmentally benign material that is stable under physical, chemical and photochemical activities. It exhibits n-type semiconducting behavior with a band gap typically ranging between 2.5 and 3.0 eV.⁹ Moreover, it demonstrates remarkable structural versatility, existing in multiple polymorphs such as cubic, triclinic, monoclinic, orthorhombic, tetragonal, and hexagonal phases.¹⁰ These diverse crystal structures, along with variable stoichiometries, impart desirable attributes, including high electrical conductivity, large theoretical capacitance, excellent thermal stability, and small ionic radii. Collectively, these features render tungsten oxide one of the most promising pseudocapacitive electrode materials for next-generation electrochemical energy storage systems. One-dimensional MWCNTs have attracted significant attention owing to their unique hollow morphology, which provides a large interfacial area between electroactive sites and the electrolyte, while simultaneously facilitating efficient electron transport. Their nanostructured subunits also shorten the diffusion pathways for ions and electrons, thereby improving overall charge transfer kinetics.¹¹ MWCNTs primarily store charge via electric double-layer capacitance and act as highly conductive nano-networks that promote active nucleation sites, also imparting enhanced structural stability and improved physicochemical properties. Metal oxides, on the other hand, are known for their high specific capacitance due to rapid redox activities, but they generally suffer from poor cyclic stability. To overcome this limitation, metal oxides are often integrated with carbon-based materials such as MWCNTs to form composites that combine the merits of EDLC and faradaic charge storage. In such composites, MWCNTs serve as a robust nanomatrix that accelerates ion transport while providing strong mechanical support.¹²

Tungsten oxide has well-known non-stoichiometric compositions, such as WO_3−x, including distinct crystallographic phases like WO₃,¹³ W₂₅O₇₃,¹⁴ W₂₀O₅₈,¹⁵ W₅O₁₄,¹⁶ W₁₈O₄₉,¹⁷ and W₃₂O₈₄,¹⁸ which are collectively referred to as Magneli phases. However, stoichiometric WO₃ is a thermodynamically stable, where tungsten predominantly exists in the W⁶⁺ oxidation state. By engineering crystallographic defects in the form of oxygen vacancies, stoichiometric WO₃ can be transformed into oxygen-deficient WO_3−x (0 < x < 1).¹⁹ Such vacancy engineering is an effective strategy for improving the electrochemical performance, as oxygen vacancies create additional active sites and promote stronger adsorption of electrolyte ions, thereby facilitating enhanced redox activities.²⁰ Previous studies have indicated that when x ≤ 0.2, WO_3−x typically crystallizes into phases with the general formula W_nO_3n−2, or W_nO_3n−1. Among these, W_nO_3n−2 often adopts a monoclinic crystal structure with P2/c symmetry, containing a single W_nO_3n−2 moiety. In particular, the W_nO_3n−2 (W₂₅O₇₃) structure consists of WO₆ octahedra interconnected through edge sharing, with six such octahedra forming a characteristic structural motif. As the reduction level increases beyond x > 0.2, the crystal framework evolves further, incorporating both edge and face-sharing WO₆ octahedra. This rearrangement gives rise to pentagonal columnar units and hexagonal tunnel-like channels within the lattice. The oxygen deficiency present in WO_3−x is compensated by the formation of crystallographic shear (CS) planes, where some of the corner-sharing WO₆ octahedra become edge-sharing.²¹ Density functional theory investigations of W₂₅O₇₃ were carried out by Migas et al.²² The band structure revealed that its Fermi level partially overlaps with the conduction band, leading to division into occupied and unoccupied states, while simultaneously intersecting the bandgap region. Their calculations further estimated a remarkably high electron carrier concentration of 2.90 × 10²¹ cm⁻³.²²

Table 1 summarizes the key performance matrix from recent publications. Awasthi et al.²³ prepared WO₃/graphene quantum dots through electrochemical oxidation followed by microwave treatment for supercapacitor application. The prepared electrode exhibited the highest specific capacitance of 560 F g⁻¹ at scan rate of 10 mV s⁻¹, with a potential window of 0.85 V. Cubic WO₃ was synthesised by Hussain et al.²⁴ by a hydrothermal method, where they studied the effects of the crystal phase on charge storage kinetics. The tetragonal WO₃ exhibited the highest specific capacitance of 91 F g⁻¹, in comparison to monoclinic and mixed phase WO₃. Nishad et al.²⁵ studied the influence of the morphological transformation of WO_3−x·H₂O on electrochemical performance. Between the nanoflower and nanoribbon WO_3−x·H₂O, the nanoflower morphology exhibited the higher specific capacitance of 70 F g⁻¹, whereas the nanoribbon morphology exhibited 37 F g⁻¹ at 1 A g⁻¹ current density. They achieved an extended potential window of 1.5 V in 1 M H₂SO₄. The electrochemical synthesis method was effectively utilised for the synthesis of MoSe₂–PPD,²⁶ co-anodized Sr-PGC,²⁷ Ni(OH)₂/MWCNT,²⁸ and Pd–WO₃–PPD/GO/ACGS²⁹ for supercapacitor applications. MoSe₂–PPD was formed on a graphite sheet, which showed a high areal capacitance of 760.52 mF cm⁻² @ 6 mV s⁻¹, with excellent stability of 81.1% after 10 000 cycles.²⁶ Grape-like clusters of co-anodized Sr-PGC were utilised to fabricate a solid-state device with PVA–NaOH gel electrolyte, with an extended voltage window of 2.1 V.²⁷ Ni(OH)₂/MWCNT,²⁸ and Pd–WO₃–PPD/GO²⁹ were deposited on an anodized commercial graphite sheet for solid-state supercapacitor application, and exhibited areal capacitances of 115, and 177 mF cm⁻², respectively.

Table 1 Literature survey of tungsten oxides and the carbon composite material for supercapacitor application^a

Material	Method	Morphology	Substrate	Configuration	Electrolyte	Potential window (V)	Specific capacitance (F g^−1)	Energy density (Wh kg^−1)	Power density (W kg^−1)	Stability @ cycles	Ref.
a CBD: chemical bath deposition, SS: stainless steel, and PVA: polyvinyl alcohol.
WO3/graphene	Electrolysis	Quantum dots	Graphite sheet	Three-electrode	1 M H2SO4	0.85	560 @ 10 mV s^−1	50.57	—	—	23
WO3	Hydrothermal	Cubic	Glassy carbon	Three-electrode	0.5 M H2SO4	0.9	91 @ 0.5 A g^−1	—	—	99.6% @ 1000	24
WO3−x·H2O	Chemical	Nanoflower	Carbon paper	Liquid device	1 M H2SO4	1.5	40 @ 0.5 A g^−1	12.5	3784	90% @ 5000	25
Pt–WO3	DC-reactive magnetron sputtering	Nanoporous	SS	Three-electrode	1 M Na2SO4	1.0	559 @ 0.1 mA cm^−2	—	—	95.2% @ 5000	30
WO3/rGO	CBD	Nanoparticles	SS	Asymmetric device	PVA−H2SO4	1.0	175 @ 5 mV s^−1	19.1	430	81.3% @ 5000	31
MoSe2–PPD	Electrosynthesis	Nanocomposite	Graphene oxide sheet	Three-electrode	1 M Na2SO4	1.2	760.52 mF cm^−2 @ 6 mV s^−1	—	—	83.6% @ 10 000	26
Co-anodized Sr-PGC	Substitution reaction	Grape clusters	Graphite coin	Solid-state device	PVA−NaOH	2.1	177.5 mF cm^−2 @ 0.4 mA cm^−2	84.31 mWh cm^−2	7.34 W cm^−2	81.1% @ 5000	27
Ni(OH)2/MWCNT	Electrochemical method	Bulk structure	Anodized commercial graphite sheet	Solid-state device	PVA−NaOH	2.0	115 mF cm^−2 @ 1 mA cm^−2	—	—	72.18% @ 8000	28
Pd–WO3–PPD/GO/ACGS	Electrochemical method	Nanospheres	Anodized commercial graphite sheet	Solid-state device	PVA−H2SO4	1.3	177 mF cm^−2 @ 0.5 mA cm^−2	25.02 mWh cm^−2	3500 mW cm^−2	91% @ 7000	29
MWCNTs/W 25 O 73	CBD	Micro-tiles on a nanotube network	SS	Three-electrode	1 M LiClO 4	0.94 V	1301.46 @ 1 mV s ^−1	—	—	59.54% @ 5000	This work
MWCNTs/W 25 O 73	CBD	Micro-tiles on a nanotube network	SS	Solid-state device	PVA−LiClO 4	1.86 V	259.24 @ 5 mV s ^−1	10.10	601.03	82.78% @ 10 000	This work

---

Herein, we explore a two-step synthetic strategy to develop the MWCNTs–tungsten oxide composite (Fig. 1). The synthesis route involves an initial ‘dip and dry’ procedure to uniformly coat MWCNTs on stainless steel (SS), followed by the controlled growth of tungsten oxide through a facile, cost-efficient, and low-temperature chemical bath deposition (CBD) process. The obtained thin film was thoroughly characterized to establish the correlation between structural, morphological, and compositional features and electrochemical performance. Previous reports relied on expensive electrochemically active substrates such as nickel foam, carbon cloth, and graphite sheet, where the substrate contributed significant charge storage, which was often mistaken as active material properties. Herein, we demonstrate the successful integration of stainless steel as a low-cost and industry-scalable alternative substrate for fabricating supercapacitor devices. To further demonstrate its practical viability, a flexible symmetric solid-state device was assembled, which successfully powered an LED panel and a miniature DC fan.

Fig. 1 Schematic of the synthesis of the MWCNT–tungsten oxide composite.

2. Experimental

2.1. Coating of MWCNTs on stainless steel substrate

Multi-walled carbon nanotubes (MWCNTs) were surface-modified through chemical functionalization by refluxing in hydrogen peroxide (H₂O₂). This treatment enriched the nanotube surface with oxygen-containing functional groups, thereby improving their dispersibility in aqueous media and enhancing their ability to adsorb metal ions from the precursor solution, which subsequently act as nucleation centres for the film growth. MWCNTs (0.5 g) were refluxed in 200 mL of H₂O₂ at 90 °C for 48 hours, followed by filtering, thoroughly rinsing with double-distilled water (DDW), and drying in a hot-air oven at 90 °C for 12 hours.

A stable dispersion was prepared by suspending the functionalized MWCNTs (0.25 g) in DDW (50 mL) containing Triton X-100 (0.5 mL) surfactant, followed by probe sonication for 2 hours. Stainless steel substrates of 3.5 cm × 5 cm dimensions were mechanically polished with an emery paper, washed with Laboline, ultrasonicated for 15 min and subsequently coated with MWCNTs by immersing in the prepared dispersion for 10 s, followed by drying under an infrared lamp. These ‘dip and dry’ cycles were repeated 8 times to obtain a desired film thickness of MWCNTs over the SS substrate.

2.2. Deposition of tungsten oxide thin film

The use of a simple chemical approach, namely, chemical bath deposition (CBD), to deposit the tungsten oxide thin film is advantageous due to its low cost.³² Under constant stirring, sodium tungstate dihydrate (Na₂WO₄·2H₂O) (131.9 mg) and sodium chloride (NaCl) (46.7 mg) were dissolved in DDW (40 mL), followed by the dropwise addition of 1 mL of hydrochloric acid (37%), leading to the formation of a whitish yellow solution. Subsequently, oxalic acid (C₂H₂O₄, 252.1 mg) was introduced while maintaining the precursor at 60 °C, resulting in a clear, transparent solution. The well-prepared MWCNTs/SS thin films were immersed vertically in the precursor for 2, 3, and 4 hours and were designated as CWO2, CWO3, and CWO4, respectively. The WO₄²⁻ ions were adsorbed onto the functionalized MWCNTs, where active sites acted as nucleation centres for the growth of tungsten oxide micro-tiles.³³ After the deposition, the substrates were rinsed in DDW and dried at room temperature.

3. Results and discussion

3.1. Structural, chemical and morphological analysis

The X-ray diffraction (XRD) patterns of CWO2, CWO3, and CWO4 deposited on stainless steel are presented in Fig. 2(a). The peaks corresponding to tungsten oxide were matched with JCPDS card No. 71-0070 of W₂₅O₇₃ with a monoclinic crystal structure. At 2θ values of 12.78° and 16.46°, the MWCNTs–tungsten oxide composite displayed strong peaks that correspond to the (1 0 6), and (2 0 6) (h k l) planes of W₂₅O₇₃, respectively. The peak at the 2θ value of 25.64° corresponds to the MWCNTs' graphitic plane with (h k l) value of (0 0 2).³⁴ The SS substrate is the source of the peaks at 43.56°, 50.34°, and 74.5°.

Fig. 2 (a) XRD analysis of tungsten oxide deposited at different reaction times. (b) FTIR spectra of tungsten oxide deposited at different reaction times. (c) Wide scan XPS survey of CWO3 and narrow scans of (d) W 4f, (e) O 1s, and (f) C 1s.

Fig. 2(b) displays the Fourier transform infrared (FTIR) spectra of CWO2, CWO3, and CWO4. The spectra showed distinctive peaks with absorption bands in the 400–4000 cm⁻¹ range, indicating the existence of the MWCNT–tungsten oxide composite. Peaks corresponding to the O–W–O and W=O were found at 694.73 cm⁻¹, and 956.51 cm⁻¹, respectively.³⁵ The absorption peak at 1622.32 cm⁻¹ corresponds to the C=C stretching in MWCNTs.³⁶ The H–O–H bending mode seen at 1403.92 cm⁻¹ proved the presence of water molecules in the sample, while the broad peak in the range of 3000–3500 cm⁻¹ confirmed the presence of hydroxyl groups functionalized on MWCNTs.³⁶ The chemical states and elemental composition of MWCNTs–tungsten oxide thin films synthesized with a deposition duration of 3 hours (CWO3) were investigated using X-ray photoelectron spectroscopy (XPS). The wide scan spectrum displayed in Fig. 2(c) clearly indicates the presence of carbon, tungsten, and oxygen. The high-resolution spectrum of the W 4f is presented in Fig. 2(d), which exhibits two distinct peaks at binding energies of 35.14 and 37.26 eV, related to the tungsten 4f_7/2 and 4f_5/2 states, respectively. These features are characteristics of the W⁶⁺ oxidation states.³⁷ The presence of a small broad peak at 40.84 eV corresponds to the W 5p state, confirming the presence of a small amount of tungsten with the +4 oxidation state, which is understood from the stoichiometry of deposited tungsten oxide (W₂₅O₇₃), deviating from tungsten trioxide (W : O = 1 : 3).³⁸ Furthermore, the deconvolution of the O 1s spectrum (Fig. 2(e)) revealed multiple oxygen bonds; the peak at 529.86 eV is associated with the oxygen bond with tungsten, while the peak at 532.07 eV is due to the presence of hydroxyl groups attached to the surface of functionalized MWCNTs.³⁷Fig. 2(f) depicts the narrow-scan survey of C 1s. The peaks present at 284.45 and 284.86 eV are due to the presence of sp² and sp³ hybridized carbon in MWCNTs, respectively, while the peak at 283.83 eV arises from the carbon–oxygen bonding due to the surface functionalization of MWCNTs.³⁹

Fig. 3(a) displays Raman spectra of CWO2, CWO3, and CWO4. The Raman peaks around 1345.0 and 1576.26 cm⁻¹ corresponds to the D and G bands of MWCNTs.⁴⁰ The Raman spectra present a total of 6 peaks corresponding to different modes of vibrations caused by W and O bonds. The peaks present below 150 cm⁻¹ originate from W–O stretching. The peaks in the range of 150 to 275 cm⁻¹ indicate W–O deformation, whereas intense peaks in the range of 600 to 950 cm⁻¹ depict O–O deformations.⁴¹Fig. 3(b)–(d) depict the EDAX spectra of CWO2, CWO3, and CWO4, respectively; confirming the presence of C, W, and O. The CWO2, and CWO4 exhibited high atomic percentages of 77.51% and 67.73% of carbon, respectively, in comparison to W and O, indicating the non-uniform coverage of tungsten oxide on MWCNTs. The CWO3 film exhibited a uniform atomic contribution of C (57.19%), O (39.51%), and W (3.30%) due to the uniform coverage of tungsten oxide on MWCNTs. The atomic percentages of C, W, and O in CWO2, CWO3, and CWO4 thin films are tabulated in Table SI3.

Fig. 3 (a) Raman analysis of the tungsten oxide deposited at different reaction times. EDAX spectra of (b) CWO2, (c) CWO3, and (d) CWO4.

Fig. 4 displays the FE-SEM micrographs highlighting the surface architecture of MWCNTs, CWO2, CWO3, and CWO4 thin films. The MWCNTs showed networks of nanotubes as expected (Fig. 4(a and b)). All the prepared composite electrodes exhibited micro-tiles of tungsten oxide grown over the nanotube network of MWCNTs. The sample deposited for 2 hours (CWO2) showed partial coverage of tungsten oxide on MWCNTs, which is evident from Fig. 4(c and d). When the deposition duration was increased to 3 hours (CWO3), the uniform coverage of micro-tiles of tungsten oxide onto MWCNTs was clearly observed (Fig. 4(e and f)). A further increase in the deposition time to 4 hours (CWO4) resulted in the detachment of micro-tiles from the MWCNTs network due to overgrowth, which can be clearly seen through the reduced density of the micro-tiles of tungsten oxide on the MWCNTs network Fig. 4(g and h). Among the three samples, the film coated with 3 hours of growth (CWO3) was expected to demonstrate superior electrochemical performance since the surface morphology was the most favorable. The mean length of micro-tiles was evaluated to be 1057.55, 901.64, and 613.55 nm, for CWO2, CWO3, and CWO4, respectively (Fig. S4(a–c)). The 2-hour-deposited films showed a varied distribution of micro-tile length with less density. On increasing the reaction time from 2 to 3 hours, film micro-tiles of uniform length with uniform density were formed. Further increasing the reaction time to 4 hours resulted in smaller micro-tiles. This is mainly due to depletion in overgrown micro-tiles, and only the micro-tiles fused on MWCNTs were adhered to the film. Fig. S4(d–f) depicts the histograms of the micro-tile thicknesses of CWO2, CWO3, and CWO4, respectively. The increase in the reaction time from 2, 3, to 4 hours resulted in an increase in the thickness of the micro-tiles from 61.09, 87.33, and 99.15 nm, respectively. Among the three samples, the film coated with 3 hours of growth (CWO3) was expected to demonstrate superior electrochemical performance as the surface morphology was the most favourable.

Fig. 4 FE-SEM micrographs of (a and b) MWCNTs, (c and d) CWO2, (e and f) CWO3, and (g and h) CWO4.

To visually confirm the chemical composition, elemental mapping by EDAX was performed on CWO3. Fig. 5(a) depicts the scan region of CWO3. The corresponding elemental distribution maps (Fig. 5(b–d)) show the uniform homogeneous dispersion of C, W, and O throughout the film. The morphology and crystallographic features of CWO3 were well scrutinized through high-resolution transmission electron microscopy (HR-TEM), as illustrated in Fig. 5(e–g). Fig. 5(e) shows the micro-tiles grown on the nanotube network of MWCNTs. The atomic fringe pattern is displayed in Fig. 5(f). It clearly displays the interface of the tungsten oxide and MWCNTs. The d-spacing was measured and matched with JCPDS no. 71-0070 of W₂₅O₇₃, and MWCNTs. The monoclinic W₂₅O₇₃ displayed (0 1 1) and (1̄ 0 16) (h k l) crystallographic planes with d-spacings of 3.8, and 3.7 Å, respectively, while the d-spacing of 3.4 Å corresponds to MWCNTs.⁴² The SAED image of CWO3 exhibited two concentric circles of (5 1 0) (h k l) of W₂₅O₇₃ and MWCNTs with d-spacings of 2.0 Å, and 3.4 Å, respectively (Fig. 5(g)).⁴² The different structural and morphological characterizations are complementary to each other, confirming the growth of monoclinic W₂₅O₇₃ micro-tiles on the nanotube network of MWCNTs.

Fig. 5 Scanning electron microscopy coupled with energy-dispersive spectroscopy of CWO3; mapping results for (a) the scan region, (b) C, (c) W, and (d) O. (e) HRTEM images, (f) fringe pattern, and (g) SAED pattern of flakes of CWO3.

3.2. Electrochemical supercapacitive performance of the MWCNT–tungsten oxide electrode

The supercapacitor performance of the MWCNTs–tungsten oxide composite thin film and its feasibility for use in flexible solid-state devices were investigated in three-electrode configurations through systematic optimization and electrochemical analysis. The integration of advanced composite materials, combining the electric double-layer capacitor type MWCNTs with pseudocapacitive tungsten oxide, is expected to boost the overall supercapacitive behavior through synergetic interactions. To assess the influence of composite formation, the performance of the CWO3 electrode was compared with pristine tungsten oxide deposited for 3 hours (WO3), standalone MWCNTs, and SS substrate. The cyclic voltammetry (CV) profiles of SS, MWCNTs, WO3 (pristine tungsten oxide with deposition time of 3 hours), and CWO3 were recorded at a scan rate of 100 mV s⁻¹ in 1 M LiClO₄ electrolyte (Fig. 6(a)). The area under the CV plot of SS was negligible, leading to very low areal capacitance (1 mF cm⁻²). Fig. 6(b) illustrates the areal capacitance of MWCNTs, WO3, and CWO3, among which CWO3 exhibited the remarkable areal capacitance of 40.58 mF cm⁻², far exceeding those of WO3 (11.52 mF cm⁻²) and MWCNTs (5.2 mF cm⁻²). This remarkable improvement is attributed to the synergetic contribution of EDLC and pseudocapacitive mechanisms within the composite, enabling superior charge storage performance.⁴³ The optimal ratio of functional groups (–COOH, and –OH) on MWCNTs, and tungsten oxide is very important as it affects the conductivity of the electrode material. Multi-walled carbon nanotubes (MWCNTs) were surface-modified through chemical functionalization by refluxing in hydrogen peroxide (H₂O₂) for 36, 48, and 60 hours. The correlation between the density of the functional groups and the electrochemical performance of the composite material was studied in detail and provided in the SI 6. The 48-hour-functionalized MWCNTs–tungsten oxide composite showed better electrochemical performance. The reaction time was varied to 1, 2, 3, and 4 hours in order to identify the most favorable growth conditions for enhanced supercapacitor performance. Fig. 6(c) shows the CV curves of the MWCNTs–tungsten oxide composite deposited for different hours at a scan rate of 100 mV s⁻¹ in 1 M LiClO₄, within the potential window of −0.45 to 0.49 V. The results reveal that the deposition duration strongly influences both the morphology and surface coverage of the electrodes, impacting their capacitive performance. Among the deposited thin films (CWO2, CWO3, and CWO4), the film deposited for 3 hours (CWO3) demonstrated the best performance, delivering a specific capacitance of 193.26 F g⁻¹, and an areal capacitance of 40.58 mF cm⁻² (Fig. 6(d)). All three electrodes exhibited micro-tiles of tungsten oxide grown on nanotubes of MWCNTs. However, incomplete surface coverage was observed in the films deposited for 2 and 4 hours, which led to reduced capacitance. In contrast, the 3 hour-deposited films produced uniform distribution of micro-tiles across the MWCNTs, which account for the superior electrochemical characteristics of CWO3 compared with CWO2 and CWO4. The electrochemical behavior of the CWO3 electrode was examined through cyclic voltammetry performed over a wide range scan rates from 1 to 250 mV s⁻¹, as shown in Fig. 6(e). The obtained CV profile demonstrated a hybrid charge storage nature, combining pseudocapacitive and EDLC characteristics. This indicates that the energy storage in CWO3 is governed by both surface-controlled capacitive processes and faradaic redox reactions.⁴⁴ The CV profile clearly indicates a quasi-rectangular area with broad oxidation and reduction peaks. The CV curve shows electrochemical reversibility by retaining its shape at different scan rates. The specific capacitance of CWO3 increased from 131.52 F g⁻¹ (areal capacitance: 27.62 mF cm⁻²) at 250 mV s⁻¹, to 1301.46 F g⁻¹ (areal capacitance 273.31 mF cm⁻²) at 1 mV s⁻¹ (Fig. 6(f)). This enhancement is attributed to the extended residence time of electrolyte ions with the active electrode material at lower scan rates, which facilitates more efficient diffusion and stronger interfacial interactions with the electrode, thereby boosting the charge storage performance.⁴⁵

Fig. 6 (a) Cyclic voltammograms and (b) areal capacitance of SS, MWCNTs, WO3, and CWO3. (c) Cyclic voltammograms and (d) areal capacitance at different reaction times. (e) Cyclic voltammograms and (f) specific and areal capacitances at different scan rates of CWO3.

FE-SEM micrographs of tungsten oxide anchored on MWCNTs revealed a unique morphology consisting of micro-tiles embedded on nanotubes of MWCNTs. This architecture generates pathways that enhance Li⁺ ion transport during electrochemical surface capacitive and faradaic processes. The presence of broad redox peaks in CV profiles further confirms the faradaic contribution arising from electrolyte ion and electrode interactions. The charge storage mechanisms involve the combination of surface adsorption and surface redox activities, governed by the chemical affinity of hydrated Li⁺ ions for the electrode material. During the cathodic scan, Li⁺ ions interact with the CWO3 active material, inducing the partial reduction of W⁶⁺ to W⁵⁺. In the subsequent anodic process, W⁵⁺ is oxidised back to W⁶⁺, consistent with the following reactions:^33,46,47

W₂₅O₇₃ + Li⁺ + e⁻ → LiW₂₅O₇₃ (1)

LiW₂₅O₇₃ → W₂₅O₇₃ + Li⁺ + e⁻ (2)

The electrochemical response of the CWO3 electrode originates from a combination of faradaic and non-faradaic processes, corresponding to diffusion-controlled redox activities and surface capacitive mechanisms, respectively. To gain further insight into the charge storage kinetics, the power law relationship (described in the SI 7) was applied.⁴⁸ log(i) versus log(v) plots were generated at different potentials and linear fitting was used to extract b values (Fig. 7(a)). The potential-dependent variation of b values are represented in Fig. 7(b), yielding an average b value of 0.74, confirming the hybrid charge storage nature of the electrode. Specifically, the surface capacitive contribution is primarily attributed to the MWCNTs network, while diffusion-limited redox activities arise from tungsten oxide. To further distinguish qualitative contributions, the Trasatti method was employed (SI 8).⁴⁹ By plotting the total charge (Q_t) against the inverse square root of the scan rate, the surface-dependent capacitive component was estimated from the Y-intercept, as shown in Fig. 7(c). In this model, the surface capacitive contribution is considered to be constant across the scan rates, while the diffusion-controlled mechanisms increase substantially at lower scan rates. At 250 mV s⁻¹, the charge storage is predominantly governed by surface capacitive mechanisms, accounting for approximately 59.35% of the total charge. With a decrease in the scan rate, the diffusion-controlled mechanisms increased to 94.00% at 1 mV s⁻¹, reflecting dominant redox activities at lower scan rates (Fig. 7(d)). The CWO3 exhibited a hybrid nature of charge storage, which is in accordance with the power law analysis.

Fig. 7 (a) log(i) vs. log(v) plot, (b) b value at various potentials, (c) Q_tvs. v^−0.5, (d) capacitive contribution at different scan rates, (e) diffusion contribution to total charge stored in the CV plot for the 50 mV s⁻¹ scan rate, and (f) plausible band bending at MWCNTs and tungsten oxide during electrochemical performance.

The quantitative scrutiny of charge stored through surface capacitive and diffusion-controlled mechanisms in cyclic voltammetry were represented, employing the Dunn method (SI 9). Fig. 7(e) depicts the contribution of diffusion-controlled mechanisms (blue area) in cyclic voltammetry to total charge storage (pink area) at a 50 mV s⁻¹ scan rate. The flat band potentials of MWCNTs, and tungsten oxide were estimated using Mott–Schottky plots (SI 10).⁵⁰ Fig. S10 depicts the Mott–Schottky plots of MWCNTs and tungsten oxide respectively, with positive slope tangents indicating n-type semiconductivity. MWCNTs, and tungsten oxide showed E_fb of −0.44, and −0.32, respectively. Tungsten oxide exhibited a positive shift in comparison to MWCNTs, indicating that the Fermi level of MWCNTs is more negative than the Femi level of tungsten oxide.⁵¹ The literature reviewed experimental results have shown that W₂₅O₇₃ exhibited an optical band gap of 3.33 eV,⁵² whereas MWCNTs showed semiconducting behavior with an optical band gap of 1.4 eV.⁵³Fig. 7(f) depicts the plausible band bending at the interface of MWCNTs and tungsten oxide. The electrons from tungsten oxide can be easily transferred at the interface of MWCNTs and tungsten oxide due to the favorable band bending at the interface.⁵¹ In general, during charging, the electrons travel from the external circuit to the active electrode material, and the Li⁺ ions in the electrolyte intercalate and localize near the host material (CWO), as shown in Fig. 7(f), at the interface of the electrode and electrolyte.⁵⁴ During discharge, the electrons travel back to the load, and Li⁺ ions disperse back to the electrolyte.

The galvanostatic charge–discharge (GCD) profiles of SS, MWCNTs, pristine tungsten oxide, and MWCNTs/tungsten oxide obtained at a current density of 0.6 mA cm⁻² are presented in Fig. S11(a). The charging and discharging times of SS are negligible. The discharge time of the composite is significantly greater than that of the MWCNTs, and pristine tungsten oxide. The pristine tungsten oxide exhibited a large iR_drop, whereas in the composite, the electrode discharged gradually, showing better charge storage capabilities. Fig. S11(b) depicts the areal capacitances of SS, MWCNTs, pristine tungsten oxide, and MWCNTs/tungsten oxide at 0.6 mA cm⁻² current density. The areal capacitance of the composite (14.26 mF cm⁻²) is higher than those of SS, MWCNTs, and pristine tungsten oxide.

The GCD studies of the CWO3 electrode were carried out within a similar voltage range of −0.45 to 0.49 V to that of cyclic voltammetry. The GCD curves at different current densities are depicted in Fig. 8(a), where the CWO3 demonstrates a pronounced non-linear charge discharge profile. Such a profile signifies the faradaic contribution of the material, arising from reversible redox reactions that occur both on the external surface and within the interconnected porous network, along with charge storage at the interface of the electrode and electrolyte.⁵⁵ The discharge traces exhibited an initial potential drop, attributable to the intrinsic ohmic resistance of the electrode, followed by the curved region, which is characteristic of pseudocapacitive redox behavior. The similar charge discharge profile at different current densities further verified the excellent electrochemical reversibility of the CWO3 electrode. Fig. 8(b) illustrates the dependence of both the specific and areal capacitances on the current densities. At 0.6 mA cm⁻², the electrode achieved maximum specific and areal capacitance values of 68.87 F g⁻¹, and 14.46 mF cm⁻², respectively. The equivalent series resistance (ESR) was estimated from the slope of the iR_dropversus current density plot (Fig. 8(c)), yielding an ESR value of 81 Ω. This value suggests a better ability of the CWO3 electrode to hold the accumulated charges; however, the power delivery of the supercapacitor might be reduced.⁵⁶ To gain detailed insight into the electrical conductivity, charge transfer characteristics, ionic diffusion, and capacitive response of the CWO3 electrode, electrochemical impedance spectroscopy (EIS) was employed. The Nyquist profile illustrated in Fig. 8(d) presents real versus imaginary impedance components over the frequency domain of 100 mHz to 100 kHz. The experimental spectra were analysed using an equivalent circuit model fitted with z-SimpWin software (inset Fig. 8(d)). In this equivalent circuit, the solution resistance (R_S) corresponds to the combined contribution from the ionic resistance of the electrolyte and electrical resistance of the current collector. The charge transfer resistance (R_CT) reflects the impedance associated with the electrode–electrolyte interface, and faradaic reactions within the electrode.⁵⁷ The constant phase elements, CPE₁ and CPE₂, represent the non-ideal capacitive behavior originating from the surface heterogeneity and the interfacial faradaic processes.⁵⁸ The Warburg component (W) models ion diffusion from the electrolyte into the porous electrode structure, and leakage resistance (R_L) accounts for self-discharge losses.⁵⁹ The R_S value, obtained from the intercept of the real axis in the Nyquist spectrum, was 2.29 Ω cm², signifying a low interfacial resistance. The semicircle in the mid-frequency region corresponds to R_CT, measured to be 16.28 Ω cm², which is linked to electron transport limitations across the CWO3 framework. In this system, CPE plays a dominant role in governing the capacitive behavior, capturing both ideal and non-ideal contributions. Its impedance is defined by frequency-independent parameters T and n, along with angular frequency (ω), as expressed in the SI 12. The modelled n-values of CPE₁ (0.8) and CPE₂ (0.8) confirm the deviation from ideal capacitance and support the hybrid charge storage mechanism.⁶⁰ These findings are consistent with the results obtained from earlier analysis through cyclic voltammetry. The Warburg diffusion (σ_ω) was calculated using the graph of real impedance plotted against the inverse square root of angular frequency (ω) (Fig. 8(e)). The diffusion coefficient was estimated to be 4.29 × 10⁻⁹ cm² s⁻¹ (SI 13). The Bode phase angle plot (Fig. 8(f)) further illustrates the hybrid charge storage behavior. While an ideal capacitor exhibits a phase angle of −90°, supercapacitors typically show deviations due to concurrent faradaic and non-faradaic contributions. In this case, a phase angle of −79.18° in the low-frequency region was observed, reaffirming the hybrid charge storage performance of the CWO3 electrode. Moreover, the system demonstrated a short relaxation time (τ_O) of 0.63 s, indicating fast charge–discharge kinetics and efficient faradaic charge-transfer capabilities.⁶⁰ Fig. S14 depicts the electrochemical stability analysis of the CWO3 electrode for 5000 CV cycles. The electrode retained 59.54% of its initial capacitance after 5000 CV cycles. The capacitive retention gradually decreased due to degradation of the electrode material, along with structural collapse. To study the phase change in the crystal structure of W₂₅O₇₃, X-ray diffraction studies were conducted for CWO3 after the analysis of the CV stability for 5000 cycles. Fig. S15(a) shows the X-ray diffraction peaks of the CWO3 film after stability analysis in comparison to the CWO3 film before stability analysis. The CWO3 film, after stability analysis, did not show any changes in the (h k l) plane and crystal structure of W₂₅O₇₃. It also showed the MWCNTs graphitic plane of (0 0 2) at a 2θ value of 24.64°. However, the noticeable factor in CWO3 before and after stability films is the intensity of the diffracted peaks of W₂₅O₇₃ and MWCNTs in comparison to SS. Before stability analysis, the intensity of the diffracted peaks of W₂₅O₇₃ was dominant, and the intensity of the diffracted peaks of MWCNTs was comparable to that of SS. The intensities of the diffracted peaks of W₂₅O₇₃ and MWCNTs were significantly reduced in comparison to SS in the film after stability analysis. This clearly indicates material degradation when the CWO3 electrode is subjected to long term stability analysis. The FE-SEM micrographs of the CWO3 electrode after stability analysis for 5000 CV cycles clearly show the structural collapse and degradation of micro-tiles of W₂₅O₇₃, resulting in a gradual reduction in capacitive retention (Fig. S15b). The kinetics of electrolyte ions are very rapid in aqueous medium, which results in fast repeated ion insertion and extraction during the charging and discharging process.⁶¹ This creates internal strain, resulting in structural collapse. In hybrid composite materials, usually the carbon MWCNTs are mechanically flexible, whereas metal oxides are comparatively brittle, which causes a volume change mismatch in composite materials.^62,63 The FE-SEM micrographs clearly depict the structural collapse of W₂₅O₇₃ micro-tiles on the stable and flexible nanotube network of MWCNTs. The EDAX spectra of the CWO3 electrode before and after stability testing are presented in Fig. S15(c). Table SI16 represents the atomic percentage of C, W, and O, compared to the CWO3 film before and after stability studies. The atomic percentage of carbon increased from 57.19% to 64.37% after stability analysis, and the atomic percentage of W decreased from 1.61% to 0.18%, showing material degradation. The electrochemical impedance studies were carried out for the CWO3 electrode before and after stability. The Nyquist plot of the CWO3 electrode before and after stability are shown in Fig. S15(d). The Nyquist plot clearly indicates the increase in the resistive behavior of the CWO3 electrode after stability. The solution resistance (R_S) was found to be 2.29, and 2.9 Ω cm² for the CWO3 electrode before and after stability, respectively. The CWO3 electrode before and after stability exhibited charge transfer resistance (R_CT) values of 16.28 and 25.47 Ω cm², respectively. The reduction in the conductivity of the CWO3 electrode after stability analysis was well-interpreted by FE-SEM, and XRD analysis. The material degradation and structural collapse of the W₂₅O₇₃ micro-tile resulted in a reduced electrochemical performance after 5000 CV cycles.

Fig. 8 (a) GCD plots at different current densities, (b) specific and areal capacitance vs. current density, (c) iR_dropvs. current density, (d) Nyquist plot (inset: equivalent circuit), (e) Z_realvs. ω^−0.5, and (f) Bode plot of the CWO3 electrode.

3.3. Electrochemical performance of the flexible solid-state symmetric supercapacitor

Inspired by the outstanding electrochemical performance of the CWO3 electrode with its hierarchical nanostructure, a symmetrical flexible solid-state supercapacitor device was fabricated. The device was assembled using a PVA–LiClO₄ gel polymer electrolyte, where Li⁺ ions act as mobile charge carriers dispersed within the insulating polyvinyl alcohol framework, which simultaneously serves as a separator between two identical electrodes. Electrochemical testing of the as assembled cell was conducted by cyclic voltammetry at scan rates between 5 to 100 mV s⁻¹ within a broad operational potential window of 1.86 V (Fig. 9(a)). The identical CV profiles at all scan rates confirm stable and reversible faradaic processes, demonstrating the durability of the device under dynamic conditions. The quasi-rectangular loops with extended humps reveal the coexistence of surface capacitive and diffusion-controlled charge storage mechanisms. As depicted in Fig. 9(b), the device achieved a specific capacitance of 165.72 F g⁻¹ and an areal capacitance of 34.80 mF cm⁻² at a 100 mV s⁻¹ scan rate. When the scan rate decreased to 5 mV s⁻¹, the cell exhibited a remarkable increase in specific and areal capacitance, reaching 259.24 F g⁻¹ and 54.44 mF cm⁻², respectively. This improvement at slower scan rates reflects enhanced ion transport and deeper electrode electrolyte interaction, enabling more efficient charge storage.⁶⁴

Fig. 9 For flexible solid-state devices, (a) CV plots at different scan rates, (b) specific and areal capacitance vs. scan rate, (c) GCD plots at different currents, (d) specific and areal capacitance vs. current, (e) iR_dropvs. current, and (f) Coulombic efficiency at different current.

Fig. 9(c) presents the galvanostatic charge–discharge (GCD) curves of the flexible solid-state device, recorded over a voltage window of 0–1.86 V under different current loads from 2.5 to 5.0 mA. The non-ideal profiles highlight the coexistence of reversible faradaic reactions along with surface controlled electrochemical double layer capacitance. A distinct iR_drop appeared at the onset of each discharge cycle, arising from the internal resistance contributions of both the electrode and the gel polymer electrolyte. Following this, the curve displayed a non-linear region, characteristic of faradaic charge storage processes, while the quasi-linear segment corresponds to capacitive adsorption attributed to the MWCNTs framework.⁶⁵ At the lowest applied current of 2.5 mA, the cell delivered a maximum specific capacitance of 21.02 F g⁻¹ and an areal capacitance of 4.41 mF cm⁻². On increasing the current to 5 mA, these values declined to 19.04 F g⁻¹ and 4.0 mF cm⁻², respectively (Fig. 9(d)). The reduction in capacitance is attributed to insufficient ion diffusion at elevated currents. The iR_dropversus current plot is shown in Fig. 9(e) to calculate the ESR value. The ESR was calculated to be 49 Ω, indicating better charge storage. The coulombic efficiency of the flexible solid-state device was calculated at different currents (Fig. 9(f)). It represents the ratio of discharge time to charging time, providing insight into the electrochemical reversibility. The device exhibited coulombic efficiency in the range of 93% to 91% within the current range of 2.5 to 5 mA, indicating better electrochemical reversibility.⁶⁶ To further probe charge transport behavior, EIS was conducted, and the Nyquist response is shown in Fig. 10(a). The spectra were fitted with an equivalent circuit model as represented in the inset in Fig. 10(a). The fitted circuit revealed very low values for both solution resistance (R_S) and charge transfer resistance (R_CT) of 0.13 and 5.58 Ω, respectively, signifying efficient ion migration and electron conduction across the device. The constant phase element (CPE) yielded an empirical exponent (n) of 0.81, confirming a mixed charge storage mechanism involving both capacitive and faradaic contributions. Further evidence of this hybrid behavior is provided by the Bode phase angle plot (Fig. 10(b)), which shows a maximum phase angle of −55.98°, deviating from the ideal capacitive response but consistent with pseudocapacitive systems. Additionally, the device exhibited a short relaxation time (τ_O) of 0.12 s, reflecting its fast charge discharge response and efficient faradaic kinetics.⁶⁷ The solid-state gel electrolytes are electrochemically stable compared to aqueous electrolytes, as they suppress the active material dissolution or degradation.^68,69 The electrochemical endurance of the solid-state device was assessed through cyclic voltammetry, as illustrated in Fig. 10(c), and inset of Fig. 10(c). Even after 10 000 consecutive CV cycles, the device preserved 82.78% of its initial capacitance, confirming reliable long-term operational stability. The spontaneous reduction in the capacitance up to 1000 CV cycles can be attributed to the partial degradation or structural decomposition of the electrode material under repeated electrochemical stress. Further, the capacitive retention gradually reduced to 82.78% after 10 000 CV cycles. To evaluate the mechanical durability, the device was tested under various bending angles, as depicted in Fig. 10(d), with photographic insets highlighting its flexibility under mechanical deformation (inset Fig. 10(d)). CV analysis was conducted while bending the device at varied angles. Impressively, the device retained 100.5%, 100.6%, 100.63%, 100.38%, 100.09%, and 99.62% of its original capacitance, even at bending angles of 30°, 60°, 90°, 120°, 150°, and 170°, respectively. Fig. S16 represents the capacitive retention of the device at various bending angles after the device was subjected to 10 000 CV cycles. The device exhibited better mechanical stability, maintaining 82.78%, 82.33%, 81.76%, 81.00%, 81.76%, 80.57%, and 80.40% of its initial capacitance even at bending angles of 30°, 60°, 90°, 120°, 150°, and 170°, respectively. This reflects outstanding mechanical reliance and stable electrochemical behavior under physical stress. The Ragone plot (Fig. 10(e)) further demonstrated the energy-power characteristics of the CWO3-based solid-state supercapacitor across different currents. The device delivered a maximum energy density of 10.10 Wh kg⁻¹ at a power density of 293.74 W kg⁻¹, and a peak power density of 601.03 W kg⁻¹ at an energy density of 9.14 Wh kg⁻¹, highlighting its capability to effectively balance energy storage with rapid power delivery.

Fig. 10 (a) Nyquist plot (inset: equivalent circuit), (b) Bode plot, and (c) electrochemical stability after 10 000 CV cycles. (d) Mechanical stability up to a 170° bending angle for the flexible solid-state device, (e) Ragone plot and a practical demonstration of flexible solid-state device by (f) lighting a VNIT LED panel, and (g) powering a small DC fan.

Practical demonstrations of the device performance are shown in Fig. 10(f and g). A charging period of only 20 s was found to be sufficient to illuminate a VNIT LED panel consisting of 21 LEDs for more than 40 s. In addition, the device successfully powered a miniature DC fan for 10 s after just 10 s of charging, underscoring its strong potential for real-world deployment in flexible and portable energy storage applications.

4. Conclusion

The monoclinic tungsten oxide (W₂₃O₇₃) with a micro-tile architecture was successfully grown over a nanotube network of MWCNTs via an easy, cost-efficient, scalable, and low-temperature solution-based chemical bath deposition. To enhance the electrochemical performance, this work highlights the potential hybridization of tungsten oxide with the EDLC-based MWCNTs material. The micro-tile architecture of W₂₅O₇₃ on the MWCNTs network provided abundant electrochemical sites, facilitating efficient ion transport, and preserved mechanical robustness, leading to outstanding electrochemical performance. The electrode delivered a high specific capacitance of 1301.46 F g⁻¹ (areal 273.31 mF cm⁻²) at a scan rate of 1 mV s⁻¹ with a potential window of 0.94 V in a 1 M LiClO₄ aqueous electrolyte. A flexible solid-state device was assembled with a PVA–LiClO₄ gel matrix as the gel electrolyte, which delivered an extended voltage window of 1.86 V and achieved specific and areal capacitances of 259.24 F g⁻¹ and 54.44 mF cm⁻², respectively, at a 5 mV s⁻¹ scan rate. The device retained 82.78% capacitive retention even after 10 000 CV cycles, showing good electrochemical stability. The flexible solid-state device retained 99.62% of its initial capacitance even after bending at an angle of 170°, underscoring its excellent mechanical durability. In a practical demonstration, it was used to power an LED panel and a miniature DC fan, confirming its real-life application potential.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article have been included as part of the manuscript and supplementary information (SI). Supplementary information is available. See DOI: https://doi.org/10.1039/d5ta08483k.

Acknowledgements

T. K. S is grateful to DST (DST/INSPIRE/03/2022/000141) for providing the INSPIRE fellowship. Authors are thankful to DST-FIST (SR/FST/PSI/2017/5(C)) for providing the XRD facility. BRS acknowledges the support of SERB-CRG (CRG/2023/001184) and the DAE-BRNS Project [Ref. 59/14/07/2022-BRNS/34058].

References

1. K. Dissanayake and D. Kularatna-Abeywardana, J. Energy Storage, 2024, 96, 112563.
2. N. Vukajlović, D. Milićević, B. Dumnić and B. Popadić, J. Energy Storage, 2020, 31, 101603.
3. D. A. Vemaas, J. Mater. Chem. A, 2015, 3, 19556–19562.
4. X. Lu, M. Yu, G. Wang, Y. Tong and Y. Li, Energy Environ. Sci., 2014, 7, 2160–2181.
5. H. Zhang, L. Yang, X. Li, Y. Ping, J. Han, S. Chen and C. He, Dalton Trans., 2024, 53, 4680–4688.
6. A. Jeyaranjan, T. Selvan Sakthivel, M. Molinari, D. C. Sayle, S. Seal, A. Jeyaranjan, T. S. Sakthivel, S. Seal, M. Molinari and D. C. Sayle, Part. Part. Syst. Char., 2018, 35, 1800176.
7. M. Tomy, A. Ambika Rajappan, V. VM and X. Thankappan Suryabai, Energy Fuels, 2021, 35, 19881–19900.
8. H. Zheng, J. Z. Ou, M. S. Strano, R. B. Kaner, A. Mitchell and K. Kalantar-Zadeh, Adv. Funct. Mater., 2011, 21, 2175–2196.
9. K. Syrek, E. Wierzbicka, M. Zych, D. Piecha, M. Szczerba, M. Sołtys-Mróz, J. Kapusta-Kołodziej and G. D. Sulka, J. Photochem. Photobiol., C, 2025, 62, 100681.
10. M. Juelsholt, A. S. Anker, T. L. Christiansen, M. R. V. Jørgensen, I. Kantor, D. R. Sørensen and K. M. Ø. Jensen, Nanoscale, 2021, 20144–20156.
11. P. K. Basivi, S. Ramesh, V. Kakani, H. M. Yadav, C. Bathula, N. Afsar, A. Sivasamy, H. S. Kim, V. R. Pasupuleti and H. Lee, Sci. Rep., 2021, 11, 1–12.
12. J. Ji, S. Park and J. H. Choi, ACS Omega, 2023, 8, 16833–16841.
13. R. Kishore, X. Cao, X. Zhang and A. Bieberle-Hütter, Catal. Today, 2019, 321–322, 94–99.
14. J. F. Al-Sharab, R. K. Sadangi, V. Shukla, S. D. Tse and B. H. Kear, Cryst. Growth Des., 2009, 9, 4680–4684.
15. J. Q. Bai, M. Ma, H. Liu, Z. Qian, D. Liu, Y. Zhao, Y. Gao, J. Chen, M. Cai and S. Sun, ACS Catal., 2025, 15, 5086–5102.
16. A. Arfaoui, B. Ouni, S. Touihri and T. Mannoubi, Mater. Res. Bull., 2014, 60, 719–729.
17. K. Viswanathan, K. Brandt and E. Salje, J. Solid State Chem., 1981, 36, 45–51.
18. M. Juelsholt, A. S. Anker, T. L. Christiansen, M. R. V. Jørgensen, I. Kantor, D. R. Sørensen and K. M. Ø. Jensen, Nanoscale, 2021, 13, 20144–20156.
19. A. Nakrela, M. E. F. Nehal, A. Bouzidi, R. Miloua, M. Medles, M. Khadraoui and R. Desfeux, Ceram. Int., 2025, 51, 16997–17006.
20. T. H. Wondimu, A. W. Bayeh, D. M. Kabtamu, Q. Xu, P. Leung and A. A. Shah, Int. J. Hydrogen Energy, 2022, 47, 20378–20397.
21. L. Pirker, B. Višić, J. Kovač, S. D. Škapin and M. Remškar, Nanomaterials, 2021, 11, 1985.
22. V. L. Shaposhnikov and V. E. Borisenko, J. Appl. Phys., 2010, 108(9), 093714.
23. G. Awasthi, K. Santhy, N. Jamnapara, S. Agrawal, D. Mandal, M. Salot and S. K. Chaudhury, J. Mater. Sci.: Mater. Electron., 2024, 36(1), 56.
24. S. K. Hussain, M. S. Kim, R. Thota, S. W. Joo and J. H. Bang, Small Methods, 2025, 9, 2401854.
25. H. S. Nishad, S. P. Gupta, A. Ivaturi and P. S. Walke, Nanoscale, 2025, 17, 18571–18582.
26. N. Mostafazadeh, R. Dadashi, M. Faraji and M. Bahram, Fuel, 2026, 405, 136645.
27. M. Faraji, R. Dadashi and N. Mostafazadeh, J. Mater. Sci.: Mater. Electron., 2025, 36(3), 229.
28. R. Dadashi, M. Bahram and M. Faraji, J. Mater. Sci.: Mater. Electron., 2024, 35(12), 852.
29. P. Taghizadeh, R. Dadashi, M. Bahram and K. Farhadi, Diam. Relat. Mater., 2025, 158, 112620.
30. R. Rani, A. Kumar, M. Sharma, B. Mohan, R. Kumar, R. Chandra and V. K. Malik, J. Phys. Chem. Solids, 2025, 197, 112428.
31. S. B. Patil, R. P. Nikam, V. C. Lokhande, C. D. Lokhande and R. S. Patil, Adv. Compos. Hybrid Mater., 2025, 8(2), 175.
32. P. R. Nikam, P. K. Baviskar, J. V. Sali, K. V. Gurav, J. H. Kim and B. R. Sankapal, Ceram. Int., 2015, 41, 10394–10399.
33. P. A. Shinde, A. C. Lokhande, A. M. Patil and C. D. Lokhande, J. Alloys Compd., 2019, 770, 1130–1137.
34. X. He, X. Xu, G. Bo and Y. Yan, RSC Adv., 2020, 10, 2180–2190.
35. S. O. Ogungbesan, O. Ejeromedoghene, Y. Moglie, E. Buxaderas, B. Cui, R. A. Adedokun, M. Kalulu, M. A. Idowu, D. Díaz Díaz and G. Fu, New J. Chem., 2024, 48, 15428–15435.
36. D. Park, H. Ju, T. Oh and J. Kim, RSC Adv., 2018, 8, 8739–8746.
37. S. Y. Lee, G. Shim, J. Park and H. Seo, Phys. Chem. Chem. Phys., 2018, 20, 16932–16938.
38. X. Mao, Y. Xu, Q. Xue, W. Wang and D. Gao, Nanoscale Res. Lett., 2013, 8, 1–6.
39. L. Chen, Z. Xu, J. Li, B. Zhou, M. Shan, Y. Li, L. Liu, B. Li and J. Niu, RSC Adv., 2013, 4, 1025–1031.
40. Y. Zhang, Y. Li and P. Zhang, J. Mater. Sci., 2014, 49(9), 3469–3477.
41. S. K. Yadav, P. Yadav, A. K. Vishwakarma and L. Yadava, J. Porous Mater., 2025, 32(4), 1321–1331.
42. O. V. Kharissova and B. I. Kharisov, RSC Adv., 2014, 4, 30807–30815.
43. B. Chen, Y. Wang, C. Li, L. Fu, X. Liu, Y. Zhu, L. Zhang and Y. Wu, RSC Adv., 2017, 7, 25019–25024.
44. M. Javed, S. M. Abbas, S. Hussain, M. Siddiq, D. Han and L. Niu, Mater. Sci. Energy Technol., 2018, 1, 70–76.
45. W. T. Wahyuni, H. A. Rahman, S. Afifah, W. Anindya, R. A. Hidayat, M. Khalil, B. Fan and B. R. Putra, RSC Adv., 2024, 14, 24856–24873.
46. A. V. Salkar, A. P. Naik, V. S. Joshi, S. K. Haram and P. P. Morajkar, CrystEngComm, 2018, 20, 6683–6694.
47. P. A. Shinde, V. C. Lokhande, A. M. Patil, T. Ji and C. D. Lokhande, Int. J. Hydrogen Energy, 2018, 43, 2869–2880.
48. K. Chaudhary, S. Zulfiqar, K. M. Abualnaja, M. Shahid, H. M. Abo-Dief, M. F. Warsi and E. W. Cochran, Dalton Trans., 2024, 11147–11164.
49. S. Archana, P. C. Sharafudeen and P. Elumalai, Energy Technol., 2023, 11, 2201125.
50. S. Ravishankar, J. Bisquert and T. Kirchartz, Chem. Sci., 2022, 13, 4828–4837.
51. X. Zhou, Q. Gao, S. Yang and Y. Fang, Chin. J. Catal., 2020, 41, 62–71.
52. I. Moirangthem, S. H. Chen, B. S. Lou and J. W. Lee, Surf. Coat. Technol., 2022, 436, 128314.
53. U. Kumar, B. C. Yadav, T. Haldar, C. K. Dixit and P. K. Yadawa, J. Taiwan Inst. Chem. Eng., 2020, 113, 419–427.
54. M. J. Young, A. M. Holder and C. B. Musgrave, Adv. Funct. Mater., 2018, 28, 1803439.
55. S. T. Gunday, E. Cevik, I. Anil, O. Alagha, H. Sabit and A. Bozkurt, Macromol. Res., 2020, 28, 1074–1081.
56. T. K. Shivasharma, L. K. Bommineedi and B. R. Sankapal, Inorg. Chem. Commun., 2022, 135, 109083.
57. M. Y. Perdana, B. A. Johan, M. Abdallah, M. E. Hossain, M. A. Aziz, T. N. Baroud and Q. A. Drmosh, Chem. Rec., 2024, 24, e202400007.
58. W. A. El-Mehalmey, R. R. Haikal, A. M. Ali, A. B. Soliman, M. ElBahri and M.H. Alkordi, Mater. Adv., 2022, 3, 474–483.
59. W. Wang, S. Guo, I. Lee, K. Ahmed, J. Zhong, Z. Favors, F. Zaera, M. Ozkan and C. S. Ozkan, Sci. Rep., 2014, 4(1), 1–9.
60. A. C. Lazanas and M. I. Prodromidis, ACS Meas. Sci. Au, 2022, 3, 162–193.
61. D. P. Dubal, O. Ayyad, V. Ruiz and P. Gómez-Romero, Chem. Soc. Rev., 2015, 44, 1777–1790.
62. L. Zhang and X. S. Zhao, Chem. Soc. Rev., 2009, 38, 2520–2531.
63. G. Yu, L. Hu, N. Liu, H. Wang, M. Vosgueritchian, Y. Yang, Y. Cui and Z. Bao, Nano Lett., 2011, 11, 4438–4442.
64. S. Bhoyate, C. K. Ranaweera, C. Zhang, T. Morey, M. Hyatt, P. K. Kahol, M. Ghimire, S. R. Mishra, R. K. Gupta, S. Bhoyate, C. K. Ranaweera, C. Zhang, T. Morey, R. K. Gupta, M. Hyatt, P. K. Kahol, M. Ghimire and S. R. Mishra, Glob. Chall., 2017, 1, 1700063.
65. M. Z. Iqbal, M. M. Faisal and S. R. Ali, Int. J. Energy Res., 2021, 45, 1449–1479.
66. H. Jiang, C. Li, T. Sun and J. Ma, Nanoscale, 2012, 4, 807–812.
67. W. Si, C. Yan, Y. Chen, S. Oswald, L. Han and O. G. Schmidt, Energy Environ. Sci., 2013, 6, 3218–3223.
68. P. Kaur and K. Singh, Sustain. Energy Fuels, 2025, 9, 3981–3998.
69. D. Luo, L. Yang, Z. Geng, H. Liu, X. Zhong, Z. Zheng, Z. Hu, S. Lu, W. Deng, G. Zou, H. Hou and X. Ji, Adv. Energy Mater., 2025, e04151.

---