Plasma-Chemical Integration of Solution-Derived ATTT with 3D Carbon Nanowalls for Controllable WS2/WO x Hybrid Supercapacitor Electrodes Authors: Jinha Shin, Hyunjae Park, Soo Ouk Jang Published: 2026 ABSTRACT Two-dimensional (2D) tungsten disulfide (WS2) has attracted considerable interest as a pseudocapacitive material, yet its practical application is often hindered by limited electrical conductivity, complex fabrication routes, and the need for binders. Here, we report a plasma-chemical route for fabricating binder-free, composition-tuned WS2/WO x @CNW hybrid electrodes from solution-deposited ammonium tetrathiotungstate (ATTT) with a downstream hydrogen remote plasma process. By systematically varying the ATTT concentration and microwave power, the surface chemistry of the CNW scaffold is controllably routed into CNW-dominant, WS2-rich, S-oxide-rich, and WO x -dominant regimes while preserving the overall three-dimensional CNW framework, despite partial surface masking at high oxide coverage. Structural and chemical analyses reveal that incomplete plasma conversion at high ATTT and low power leaves sulfur-rich species that are prone to oxidation, whereas high power promotes sulfur-vacancy formation and W-rich surfaces that preferentially evolve into WO x upon air exposure. Electrochemical measurements demonstrate that introducing WS2/WO x functionalities markedly enhances areal capacitance relative to bare CNWs. The WO x -dominant electrode delivered the highest areal capacitance among the tested electrodes, 91.1 mF cm–2 at 0.1 V s–1 and 21.7 mF cm–2 at 1.0 V s–1, compared with 4.6 and 3.15 mF cm–2 for bare CNW, respectively, although this gain was accompanied by increased interfacial and transport resistance. Dunn and impedance analyses reveal that this improvement arises primarily from increased diffusion-controlled pseudocapacitive charge storage, accompanied by a trade-off in the form of higher interfacial and transport resistance. This work elucidates how precursor loading and hydrogen-radical flux govern competitive sulfur- versus tungsten-centered oxidation pathways and provides a practical plasma-chemical framework for engineering high-performance, binder-free WS2/WO x @CNW hybrid electrodes for supercapacitor applications. Introduction Electrochemical supercapacitors have attracted extensive attention as promising energy-storage devices bridging the gap between conventional dielectric capacitors and rechargeable batteries, owing to their high power density, long cycle life, and rapid charge–discharge capability. , In general, charge storage in supercapacitors is governed either by electrostatic ion adsorption at the electrode/electrolyte interface (electric double-layer capacitance, EDLC) or by fast, surface or near-surface redox reactions (pseudocapacitance). , Carbon-based nanomaterials with high specific surface area and excellent electrical conductivitysuch as activated carbons, graphene, carbon nanotubes, and vertically oriented carbon architectureshave been widely employed as EDLC-type electrodes or conductive scaffolds in hybrid devices. , Among them, carbon nanowalls (CNWs), consisting of vertically aligned few-layer graphene sheets, provide a three-dimensional (3D) open framework with short ion-diffusion paths, robust electrical pathways, and mechanical stability, making them particularly attractive as current collectors and hosts for pseudocapacitive coatings. , Transition-metal dichalcogenides (TMDs) and, in particular, tungsten disulfide (WS2) have emerged as promising pseudocapacitive materials because of their layered crystal structure, multiple accessible oxidation states, and abundant electrochemically active sites. − However, the practical electrochemical performance of WS2 is often limited by two intrinsic drawbacks. First, its relatively poor electrical conductivity can hinder rapid electron transport through the active layer, leading to increased polarization and degraded rate performance, especially at high scan rates or current densities. Second, the stacked van der Waals structure of WS2 nanosheets tends to undergo restacking or agglomeration during synthesis and electrode fabrication, which reduces the accessible surface area, blocks ion-transport pathways, and lowers the utilization of edge and near-surface redox-active sites. To address these limitations, a variety of WS2/carbon hybrid structures were developed. WS2/rGO composites use graphene sheets as conductive and mechanically compliant supports that improve electronic percolation while helping to disperse WS2 domains; WS2/CNT hybrids provide one-dimensional conductive pathways and nanoscale bridging networks that facilitate charge collection and suppress severe aggregation; WS2/carbon nanofiber systems offer self-supported fibrous frameworks with improved structural integrity and electrolyte accessibility; and WS2 embedded in three-dimensional porous or printed carbon architectures can more effectively preserve open ion-diffusion channels while exposing a larger fraction of electrochemically active WS2 surface. − ,,− As a result, these carbon-assisted architectures generally show improved capacitance retention, rate capability, and cycling stability relative to unsupported or densely restacked WS2. ,,− Nonetheless, many of these approaches still rely on multistep synthesis, high-temperature post-treatment, polymeric binders, or separate metallic current collectors, which complicate integration and scaling. ,,, Among these conductive carbon scaffolds, CNWs are particularly attractive because their vertically open 3D framework can simultaneously provide efficient electron pathways, minimize severe restacking of deposited WS2 domains, and preserve ion-accessible channels for pseudocapacitive charge storage. Solution-processed molecular precursors offer a scalable alternative to forming WS2 films and coatings. Ammonium tetrathiotungstate (ATTT, (NH4)2WS4) is one of the most widely used single-source precursors for WS2, enabling the fabrication of continuous few-layer WS2 films or nanostructures on various substrates via thermal decomposition or solvothermal treatment. − For instance, thermolysis of spin-coated ATTT layers under inert or sulfur-containing atmospheres can yield large-area WS2 thin films suitable for electronic applications, − while ATTT-based solution routes have also been employed for WS2 nanotubes, nanowires, and flakes. − However, these methods typically require relatively high annealing temperatures and long treatment times and have been mostly optimized for planar substrates. Their direct application to 3D carbon scaffolds such as CNWs, in the form of binder-free ATTT-derived WS2/WO x @CNW hybrid supercapacitor electrodes, remains scarcely explored. Plasma processing has recently emerged as a powerful tool for tailoring the surface chemistry, defect density, stoichiometry, and functionalization of 2D TMDs, offering degrees of control that are difficult to achieve with purely thermal methods. , Systematic studies have demonstrated that low-pressure plasmas can be used to introduce chalcogen vacancies, heteroatom dopants, controlled oxidation, or nanopores in WS2, thereby tuning its electronic, catalytic, and optical properties. − In particular, downstream (remote) hydrogen plasmas have been shown to clean monolayer WS2 surfaces and selectively remove contamination or weakly bound species while minimizing ion bombardment damage. Plasma-enhanced processes have also been used to engineer WS2 morphology, crystallinity, and stoichiometry by adjusting the composition of the coreactant plasma, for example, in plasma-enhanced atomic layer deposition (PEALD). Despite this progress, most plasma-based WS2 studies have focused on thin films for electronics or electrocatalysis and have employed pregrown WS2 layers on planar substrates. − To the best of our knowledge, there are very few reports that combine a solution-processed WS2 precursor such as ATTT with a hydrogen remote plasma to directly convert an ATTT precursor on a 3D carbon scaffold into a composition-tuned WS2/WO x hybrid supercapacitor electrode. Despite extensive studies on WS2/carbon hybrids, the direct fabrication of binder-free composition-tuned WS2/WO x hybrid electrodes on 3D carbon scaffolds using a simple, low-temperature route remains insufficiently explored. Here, we develop a plasma-chemical conversion strategy in which ATTT deposited on carbon nanowalls is transformed by downstream H2 remote plasma into composition-dependent WS2/WO x interfacial layers. The aim of this study is to determine how precursor loading and plasma power route the surface chemistry across WS2-rich, S-oxide-rich, and WO x -dominant regimes and how this compositional evolution governs electrochemical charge-storage behavior. By correlating structural and chemical characterization with electrochemical analysis, we identify the composition-dependent trade-off between capacitance enhancement and resistance buildup. Experimental Section ATTT-CNW supercapacitor electrodes were fabricated to enhance the electrochemical performance as electrochemical energy storage using ECR plasma without external substrate heating. Prior to plasma-chemical conversion of ATTT-coated CNW, CNW scaffolds were prepared as described previously. Ammonium tetrathiotungstate (ATTT; [[NH4]2WS4], Alfa aesar, 99.95%) was used as a single-source precursor for ATTT-derived WS2/WO x surface conversion. Dimethylformamide (DMF; HCON­[CH3]2, Alfa aesar, anhydrous) was used to prepare the ATTT solution. ATTT was dissolved in DMF to prepare solutions with concentrations of 0.05 and 0.1 M. The solutions were sonicated for 20 min to ensure complete dissolution. ATTT coating was performed using a spray-coating method, in which the ATTT solution was sprayed onto the CNW scaffold. The coated samples were dried at 80 °C for 10 min to remove the residual solvent. ATTT-coated CNW samples were converted to ATTT-CNW using a downstream remote hydrogen plasma configuration designed to prevent energetic ion bombardment as schematically illustrated in Figure . The plasma was generated in an ECR plasma source and separated from the sample region by a grounded metallic mesh. Hydrogen gas (H2, 99.999%) was introduced at 50 sccm to maintain a working pressure of 2 mTorr. The plasma was ignited at 0.5 and 1 kW and maintained for 10 min. The converted electrodes are denoted ATTT-CN­[xM/ykW], where x and y represent the ATTT concentration and microwave power, respectively. Schematic illustration of the fabrication of ATTT-carbon nanowall (ATTT-CNW) electrodes by a single-precursor, single-step hydrogen remote plasma process. The morphological properties were characterized using field emission scanning electron microscopy (FE-SEM) and Raman spectroscopy with an excitation wavelength of 532 nm. High-resolution X-ray photoelectron spectroscopy (XPS) analysis was conducted using an Al Kα X-ray source, and XPS spectra were deconvoluted and fitted using CasaXPS software for chemical-state analysis of W 4f and S 2p regions. Electrochemical measurements were conducted in a 1 M Na2SO4 electrolyte diluted in DI water. For the three-electrode system, Pt wire and Ag/AgCl were used as the counter and reference electrodes, respectively. Finally, CV measurements were conducted in the potential range of −1–0 V at various scan rates in the range of 100–1000 mV s–1, while EIS was conducted in the frequency range of 100 mHz–100 kHz. Results and Discussion First, vertically aligned carbon nanowalls (CNWs) grown on a conductive substrate were prepared as a 3D current collector. An ammonium tetrathiotungstate (ATTT) solution is then deposited onto the CNWs by spray-coating, forming a conformal precursor layer along the high-aspect-ratio walls. The ATTT-coated CNW substrate is then exposed to a remote/downstream H2 plasma generated in a quartz tube above a grounded metal mesh; the mesh effectively filters energetic ions and electrons so that mainly neutral hydrogen radicals diffuse through and interact with the precursor. During this remote plasma treatment, hydrogen radicals decompose ATTT and drive its conversion into composition-dependent WS2/WO x surface species on the CNW scaffold while preserving the underlying 3D carbon framework. Both the growth of CNWs and the subsequent hydrogen-radical treatment are carried out using an electron cyclotron resonance (ECR) plasma system, whose configuration and operating conditions have been described in detail in our previous work. , The resulting electrodes are binder-free WS2/WO x @CNW hybrids, in which a composition-dependent interfacial overlayer covers the CNW scaffold, providing abundant electrochemically active sites and continuous electronic pathways for supercapacitor operations. The bare CNW image (Figure a) shows a highly open, honeycomb-like network of vertically aligned carbon walls with uniform cell sizes on the order of a few hundred nanometers. The morphology of bare CNW provides an ion-accessible framework for electrolyte transport and the shortest transport pathways, but only the carbon surface contributes to the charge storage. Top-view SEM images of (a) bare CNW substrate, (b) ATTT-CNW [0.05 M/0.5 kW], (c) ATTT-CNW [0.1 M/1 kW], (d) ATTT-CNW [0.05 M/1 kW], and (e) ATTT-CNW [0.1 M/0.5 kW]. The images illustrate the gradual morphological evolution from an open, vertically aligned CNW network (a), to CNW-dominant structures where the walls are only thinly coated (b), more uniformly WS2-covered walls with preserved porosity (c), the CNW texture is largely masked by a dense and fine oxide nanowalls, film-like oxide layer (d), and ATTT-CNW [0.1 M/0.5 kW] showing rough and aggregated WS2/S-oxide flakes (e). Scale bars are indicated in each panel. In the ATTT-CNW [0.05 M/0.5 kW] sample (Figure b), thin WS2 (or WS x ) islands coat the sidewalls and edges in a discontinuous fashion, while the cell openings remain largely unobstructed. Such a morphology is consistent with relatively low ATTT loading and low microwave-power-assisted H-radical exposure, under which the precursor is only partially converted, adsorbed, and decomposed, resulting in a thin and discontinuous WS2/WS x O y -like overlayer that remains primarily on energetically favorable CNW sites. Under these conditions, the CNW remains the primary conductive framework and the ATTT–H· process adds only a limited amount of redox-active material. The ATTT-CNW [0.1 M/1 kW] image (Figure c) exhibits a fairly uniform thickness and continuous coverage of WS2-coated CNWs, with fine nanoscale roughness but without severe pore blocking. This morphology indicates high-ATTT loading combined with sufficient H-radical exposure to promote conversion toward a WS2-rich surface state. In this regime, the surface appears to be predominantly converted into a WS2-rich state, maximizing the density of electrochemically active WS2 edge and basal sites while maintaining accessible ion pathways along the CNW framework. In the ATTT-CNW [0.05 M/1 kW] sample, the original CNW cell pattern is largely masked by a dense, granular film that bridges neighboring walls and partially fills the pores (Figure d). Under low ATTT loading, the higher microwave power (i.e., higher H-radical flux) drives the system from a more WS2-rich interfacial state (state (c)) toward a W-rich/S-poor surface that readily evolves into a WO x -dominant overlayer (state (d), W-rich and S-poor), in line with reports that hydrogen treatments create sulfur vacancies and promote tungsten oxide formation in WS2-based materials. Figure (e) shows a surface covered with rough, aggregated, flake-like clusters. This morphology suggests that a relatively thick ATTT layer was deposited and only partially converted under a low-microwave-powered H remote plasma exposure. These morphology differences are expected to directly affect ion and electron transport before considering the detailed surface chemistry. Overall, the SEM results show a clear morphology-driven trade-off. Bare CNW and ATTT-CNW [0.05 M/0.5 kW] electrodes retain the most open framework and therefore the shortest ion/electron transport paths, but they contain only a limited amount of redox-active material. In contrast, ATTT-CNW [0.1 M/1 kW], [0.05 M/1 kW], and [0.1 M/0.5 kW] electrodes show progressively higher surface coverage and more continuous overlayers, which increase the amount of pseudocapacitive material but also increase the transport length and interfacial polarization. Therefore, the SEM results already suggest a morphology-driven trade-off in which capacitance should increase with surface coverage and oxidation, whereas the rate capability should deteriorate as the CNW surface becomes increasingly masked by a continuous overlayer. High-resolution X-ray photoelectron spectroscopy (XPS), including the W 4f and S 2p spectra shown in Figure a,b, classified the ATTT-CNW electrodes into four surface-chemistry regimes: CNW-dominant, WS2-rich, S-oxide-rich, and WO x -dominant, corresponding to ATTT-CNW [0.05 M/0.5 kW], ATTT-CNW [0.1 M/1 kW], ATTT-CNW [0.1 M/0.5 kW], and ATTT-CNW [0.05 M/1 kW], respectively. This classification is based on the relative fractions of W4+/W6+ species in the W 4f region and lattice S2–/SO x species in the S 2p region. The four regimes provide the compositional basis for interpreting the electrochemical behavior discussed below. High-resolution XPS analysis of W 4f and S 2p chemical states in ATTT-CNW electrodes: (a) W 4f and (b) S 2p spectra of ATTT-CNW. The WS2-rich electrode is characterized by a dominant W4+ doublet at 32.3–32.6 eV (W 4f7/2) and 34.5–34.8 eV (W 4f5/2) in the W 4f region, consistent with W4+ species in WS2, nonstoichiometric WS x , or WS x O y phases. In the S 2p region, it also shows a high fraction of S2– at 161.5–163.0 eV (S 2p3/2), assigned to sulfide sulfur in the WS2 lattice. Within the present series, this sample is therefore classified as WS2-rich because it combines a large S2– contribution with a relatively weak W6+ component. Together, these results indicate a predominantly WS2-rich interfacial layer on the CNW scaffold, accompanied by a minor WO x /WS x O y component. The S-oxide-rich electrode exhibits a similar overall W content, but a substantially larger S-oxide contribution at 168–170 eV in the S 2p region. This high-binding-energy feature is assigned to oxidized sulfur species (SO x ), such as sulfate- or oxysulfide-like surface components, indicating that the WS2 phase is partially covered or intermixed with oxidized sulfur/oxide species. Because the tungsten loading remains broadly comparable to that of the WS2-rich sample, this electrode serves as a useful reference for isolating the effect of sulfur oxidation on electrochemical behavior at similar W content. − Under lower tungsten loadings, the CNW-dominant electrode shows relatively weak W 4f and S 2p signals, suggesting that the highly conductive CNW network governs the electronic transport, while only a thin, discontinuous WS2/WS x O y layer decorates the surface. In the CNW-dominant sample, a weak additional S 2p component centered at ∼164.97 eV is observed. This higher-binding-energy feature is assigned to nonlattice sulfur associated with residual sulfur-rich precursor-derived WS x species, rather than to lattice S2– in crystalline WS2, based on prior XPS reports for sulfur-rich thiometallate-derived sulfide phases. , The feature was not clearly resolved in the other samples, which is consistent with different sulfur-conversion pathways under different precursor-loading and hydrogen-radical-flux conditions: sulfur is incorporated mainly into lattice WS2 in the WS2-rich sample, shifted toward SO x species in the S-oxide-rich sample, and depleted under the WO x -dominant condition. Accordingly, the ∼164.97 eV component is interpreted as a weak residual nonlattice sulfur signal specific to the low-ATTT/low-power CNW-dominant condition, where conversion and coverage remain limited. It is well-suited to evaluate how a limited amount of WS2/WS x O y decoration on a conductive scaffold affects the balance between electric double-layer and pseudocapacitive charge storage. In contrast, the WO x -dominant electrode shows pronounced W6+ features in the W 4f region, with peaks at 35.5–36.0 eV (W 4f7/2) and 37.6–38.0 eV (W 4f5/2), which are assigned to W6+ species in tungsten oxides such as WO3, along with high W and O atomic fractions. These features indicate that the CNW surface is covered by a relatively thick oxidized WS2/WS x O y overlayer and identify this sample as the highly oxidized limit within the present series. Remote/downstream H2 plasma is known to promote sulfur removal and sulfur-vacancy formation in TMDs, thereby generating under-sulfurized, W-rich reactive sites that can readily convert to W–O species upon subsequent exposure to oxygen or moisture, appearing as WO x in XPS. , Because tungsten oxides are well-known to provide pseudocapacitive charge storage through fast surface and near-surface cation insertion/extraction, − the WO x -dominant sample serves as a useful reference for evaluating how excessive oxidation and thick oxide-rich overlayers influence capacitance enhancement and kinetic limitation in the present ATTT-CNW system. Overall, the XPS results indicate that ATTT loading and microwave power are the key parameters controlling whether the surface evolves toward CNW-dominant, WS2-rich, S-oxide-rich, or WO x -dominant states. Under low-ATTT/high-power conditions, sulfur is removed more effectively during remote H2 plasma treatment, leaving W-rich surfaces that are more readily oxidized upon air exposure and thus shift toward a WO x -dominant composition. In contrast, under high-ATTT/low-power conditions, the radical flux appears insufficient for complete conversion of the precursor layer, allowing precursor-derived sulfur-containing residual species to persist; in the present XPS results, these conditions correlate with the emergence of a S-oxide-rich surface. , When precursor loading and radical supply are better balanced, the surface is driven toward a WS2-rich regime, whereas low-ATTT/low-power conditions mainly result in limited coverage and a CNW-dominant surface. This compositional evolution directly governs the electrochemical response: increasing WS2/WO x -derived redox-active phases enhances capacitance, but the development of more continuous and oxidized overlayers also increases interfacial and transport resistance, leading to the observed capacitance–resistance trade-off. In the present ATTT-CNW system, the relative fractions of WS2, WO x , and CNW are therefore expected to tune the ratio between the double-layer and pseudocapacitive components. In the following sections, the results of cyclic voltammetry and impedance spectroscopy analysis are directly compared with XPS-derived descriptors (such as S2–/S and W4+/W6+ ratios) to elucidate how the WS2-rich, S-oxide-rich, CNW-dominant, and WO x -dominant regimes govern the specific capacitance, rate capability, and charge-transfer resistance of ATTT-CNW supercapacitor electrodes. Raman spectroscopy (Figure ) was employed to characterize the structural fingerprints of the same four-regime classification obtained from XPS and to verify phase-/composition-related signatures of the hybrid surface. Using a 532 nm excitation, we monitored the characteristic WS2 vibrational modes and additionally screened for oxide-related features. Across the representative electrodes, the CNW-derived ID/IG values remained within a relatively narrow range (∼1.5–2.0), suggesting that the defect-rich graphitic CNW scaffold was broadly preserved after ATTT deposition and remote H2 plasma treatment. Therefore, the principal Raman differences discussed below are interpreted mainly in terms of WS2/WO x -related vibrational features and relative surface coverage, rather than a major change in the intrinsic disorder of the CNW scaffold. Raman spectra of representative ATTT-CNW electrodes. The corresponding carbon-band intensity ratios (I D/I G) were approximately 1.7 for WS2-rich, 1.5 for S-oxide-rich, 1.7 for CNW-dominant, and 2.0 for WO x -dominant, indicating that the CNW scaffold remained present in all samples, although its relative Raman contribution varied with surface coverage by WS2/WO x species. The spectrum of bare CNW is dominated by the characteristic sp2-carbon Raman bands: the D band (∼1350 cm–1), G band (∼1580–1605 cm–1), and a 2D band near ∼2680–2700 cm–1, consistent with an edge-/defect-rich graphitic carbon framework such as CNW. − For the WS2-rich sample, the spectrum shows clear WS2 vibrational modes: the in-plane E 2g 1 mode at 353 cm–1 and the out-of-plane A1g mode at 418 cm–1, with a separation Δω = A1g – E 2g 1 ≈ 65cm –1. This Δω value lies within the typical range reported for multilayer or bulk-like 2H-WS2, , corroborating the XPS conclusion that W is predominantly present as W4+ in a WS2 environment. In the low-frequency region (∼148 cm–1), a J1 mode is observed with an intensity ratio J1/A1g ≈ 1.1. This indicates the presence of some 1T’-like or defect-related domains but at a moderate level, i.e., the structure is still primarily 2H-WS2 with only localized distortions. , In the 500–800 cm–1 range, the WO x -related band (bridging W–O–W) at ≈ 529 cm–1 has a very small normalized intensity (WO x /A1g ≈ 0.09), consistent with a relatively thin or sparse WO x overlayer. This is in line with the XPS result for the WS2-rich group, where W6+(WO x ) contributes only weakly compared to W4+(WS2). , At higher wavenumbers, the carbon D and G bands associated with the CNW scaffold appear at ≈1352 cm–1 (D) and ≈ 1588 cm–1 (G), with D/G ≈ 1.7 and D/A1g ≈ 2.2. These values are typical of vertical graphene/carbon nanowalls with a significant density of defects and edges, indicating that a conductive CNW network underlies the WS2 layer. Overall, both Raman and XPS consistently support the interpretation that this group corresponds to a WS2-dominated surface with limited WO x and moderate structural defects, i.e., a WS2-rich/low-WO x ATTT-CNW electrode. As in the WS2-rich case, it was revealed that the E 2g 1 (≈ 352 cm–1) and A1g (≈ 414 cm–1) modes are clearly present for S-oxide-rich, as well, and the separation Δω ≈ 62 cm–1 still falls within the typical range for 2H-WS2. , However, compared to the WS2-rich sample, the slightly smaller Δω and broader peak shapes suggest increased interlayer interaction and/or lattice distortion. The most striking difference is the intensity of the low-frequency J1 mode. In S-oxide-rich, J1/A1g increases to ≈ 3.8, much higher than in the WS2-rich sample. Such strong J1 activity is commonly associated with increased 1T/1T’ phase fraction, sulfur vacancies, and oxygen substitution around W, which induce local octahedral coordination and a strongly perturbed vibrational landscape. , This matches well with the XPS observation that the S-oxide (SO x ) contribution grows substantially, while the WS2 lattice is still present, indicating coexistence of WS2 and oxidized sulfur environments. In the 500–800 cm–1 region, the W–O-related band intensity also increases (WO x /A1g ≈ 0.74), clearly higher than in WS2-rich. For the S-oxide-rich sample, the higher ATTT loading combined with insufficient H-radical fluence likely leaves sulfur-containing precursor-derived or WS x -like species, which are subsequently oxidized into SO x -containing surface components. The increased J1 intensity and W–O-related Raman features therefore indicate a defective, partially oxidized WS2/WS x O y interface rather than a purely W-rich oxide surface. , The smaller D/A1g ratio (≈ 1.1) implies that the WS2 layer is more continuous/thicker in S-oxide-rich, reducing the exposed CNW area and, at the same time, being partially oxidized to S-oxide and WO x . In summary, the Raman spectrum of S-oxide-rich reflects a ATTT-CNW electrode where WS2 crystallinity is preserved but S-oxide and 1T’-like defect domains are significantly increased, consistent with the XPS-defined S-oxide-rich group. The CNW-dominant sample still exhibits WS2-related modes (E 2g 1 ≈ 355 cm–1, A1g ≈ 421 cm–1, Δω ≈ 66 cm–1), but the most prominent feature is the overwhelming dominance of the carbon bands. The D and G bands are very strong, with D/A1g ≈ 6.0 and D/G ≈ 1.7. These values indicate that, in Raman scattering, the signal is dominated by the sp2– carbon framework of CNW, while WS2 contributes only a minor fraction of the overall scattering intensity. , The J1/A1g ratio (≈1.2) and WO x /A1g ratio (≈0.23) are intermediate between the WS2-rich and S-oxide-rich cases, suggesting moderate levels of defects and oxidation. In other words, WS2 is present but with low surface coverage, and much of the CNW surface remains exposed, with only sparsely nucleated WS2/WO x domains. Structurally, this indicates a hybrid electrode in which (i) the highly conductive CNW network governs electron transport, and (ii) localized WS2/WO x nanodomains provide additional faradaic/pseudocapacitive sites. , Thus, the Raman response of the CNW-dominant sample is fully consistent with the result of XPS: a ATTT-CNW electrode where the CNW scaffold dominates both composition and Raman intensity, and WS2/WO x acts as a low-coverage decoration on the carbon framework. The WO x -dominant sample is characterized by an exceptionally strong band in the 450–500 cm–1 region, with WO x /A1g ≈ 1.9, the highest among the four representatives. This band is assigned to W–O–W bending/stretching modes characteristic of tungsten oxide phases such as WO3 and substoichiometric WO3‑x , , directly matching the high W6+(WO x ) fraction found in the XPS analysis. The J1 mode is also present with J1/A1g ≈ 2.7, larger than in the WS2-rich and CNW-dominant samples but somewhat smaller than in the S-oxide-rich sample. This points to a mixed coordination environment around W, where WS2-like and WO x -like environments coexist, and local 1T’-like distortions persist within a predominantly oxidized matrix. The CNW-related D and G bands remain visible (D/G ≈ 2.0, D/A1g ≈ 3.3), but their relative contribution is reduced compared to the strong WO x and distorted WS2 peaks, confirming that a substantial portion of the CNW surface is covered by WO x /WS x O y interfacial layers. , When the XPS and Raman results are considered together, a clear composition–defect hierarchy emerges across the five electrodes. Bare CNW is governed by the sp2-carbon framework and therefore represents the EDLC baseline. CNW-dominant contains only a small amount of WS2/WO x decoration on a highly conductive carbon scaffold. WS2-rich is characterized by dominant W4+ and lattice S2– components together with clear E 2g 1/A1g modes, indicating that a relatively continuous WS2 phase has formed, while WO x remains minor. S-oxide-rich still preserves the WS2 lattice, but its larger SO x contribution, stronger J1 mode, and higher WO x /A1g ratio indicate a more defective and partially oxidized interfacial layer. Finally, WO x -dominant exhibits the highest W6+ contribution and the strongest WO x Raman signature, consistent with a thick W-rich oxidized overlayer. On this basis, the electrochemical response is expected to evolve from predominantly EDLC-type storage in bare CNW toward increasingly pseudocapacitive and increasingly diffusion-influenced charge storage in WS2-rich, S-oxide-rich, and WO x -dominant electrodes. Cyclic voltammetry analysis was conducted to compare the cyclic voltammograms of four representative ATTT-CNW electrodes (CNW-dominant, WS2-rich, S-oxide-rich, and WO x -dominant) and bare CNW at scan rates from 0.1 to 1.0 V s–1. For quantitative comparison, the apparent areal capacitance, C*­(v), was calculated from the CV curves as 1 C*(v)=12vAΔV∫V1V2|I(V)|dV where ΔV is the potential window (−1.0 to 0 V), v is the scan rate, and A is the geometric electrode area. − The overlaid CV curves at 0.1 and 1.0 V s–1 are shown in Figure a,b, respectively, and reveal a clear composition-dependent trade-off: areal capacitance increases in the order bare CNW < CNW-dominant < WS2-rich < S-oxide-rich < WO x -dominant, whereas rate retention decreases as the coating becomes more continuous and more oxidized. Thus, adding WS2/WO x -derived redox-active phases strongly enhances low-rate charge storage, but progressively increases kinetic limitation at a high scan rate. Overlaid CV comparison of bare CNW, CNW-dominant, WS2-rich, S-oxide, and WO x electrodes at (a) 0.1 V s–1 and (b) 1.0 V s–1 (1 M Na2SO4, 0 V to −1 V vs Ag/AgCl reference electrode). The bare CNW electrode exhibits a largely capacitive-type profile, whereas the CNW-dominant, WS2-rich, S-oxide-rich, and WO x -dominant electrodes show additional faradaic signatures (peak-like or broadened redox features), indicating increasing pseudocapacitive contributions associated with the composition-tuned surface chemistry. At 0.1 V s–1 scan rate, the measured specific capacitance is 4.6 mF cm–2 (bare CNW), 6.4 mF cm–2 (CNW-dominant), 40 mF cm–2 (WS2-rich), 52.7 mF cm–2 (S-oxide-rich), and 91 mF cm–2 (WO x -dominant). At 1.0 V s–1 (Figure (b)), the CV profiles emphasize rate-limiting factors such as polarization and ion-transport constraints. Compared with the low-scan regime, faradaic features broaden and shift, and distortion becomes more pronounced in electrodes where redox reactions are less kinetically accessible. Electrodes maintaining a more capacitive-like shape at a high scan rate indicate more favorable electronic/ionic percolation, whereas stronger distortion and area loss reflect increasing diffusion-controlled limitation and/or higher internal resistance. In the CV profiles of the ATTT-CNW, a broad hump centered at approximately −0.3 V is observed, which is absent in the bare CNW electrode. This feature becomes progressively more pronounced for samples with higher WO x content (e.g., S-oxide-rich and WO x -dominant groups), in good agreement with the XPS results showing an increased W6+ contribution associated with surface tungsten oxides and oxysulfide species. This systematic trend strongly suggests that the −0.3 V hump originates from WO x -related redox activity rather than from the carbon scaffold. Accordingly, the broad hump centered near −0.3 V is reasonably attributed to a WO x -derived mixed pseudocapacitive response, associated with reversible tungsten valence transitions (W6+ ↔ W5+, with partial W5+ ↔ W4+) coupled with the insertion/extraction of charge-compensating cations from the electrolyte. In neutral Na2SO4, Na+ (together with solvated protons or H3O+ near the interface) can be inserted into the near-surface region of WO x domains and W–O–S oxysulfide phases, giving rise to a broad, weakly resolved cathodic/anodic pair around −0.3 V. The coexistence of a quasi-rectangular capacitive background from CNW framework and this superimposed broad redox feature indicates mixed charge storage involving both capacitive surface-confined intercalation-type pseudocapacitive contribution The scan-rate dependence of this hump further supports a mixed-control behavior. At low scan rates, the −0.3 V feature appears relatively distinct, indicating that ion insertion into the WO x /oxysulfide layer proceeds sufficiently close to equilibrium. As the scan rate increases, the hump gradually broadens and partially merges into the capacitive envelope, implying the coexistence of fast surface redox and a diffusion-limited component within the nanostructured WO x domains. This behavior is consistent with prior reports on nanostructured WO3/WO x electrodes, where broad redox features in the −0.3–0.0 V region (vs Ag/AgCl) have been attributed to W-valence switching coupled with cation intercalation in hydrous or oxygen-deficient tungsten oxides. Taken together, these observations indicate that the −0.3 V hump in ATTT-CNW electrodes is best described as a WO x -origin mixed (surface + diffusion-controlled) pseudocapacitive feature. The presence of this WO x -related redox contribution provides an additional faradaic charge-storage channel on top of the double-layer capacitance of CNW and the WS2-derived pseudocapacitance, thereby contributing to the enhanced areal capacitance observed for the S-oxide-rich and WO x -dominant samples. At the same time, the broad and partially diffusion-limited nature of this feature explains the somewhat larger capacitance decay at high scan rates for WO x -dominant electrodes compared with WS2-rich compositions. Figure (a–e) presents scan-rate-dependent CV curves (0.1–1.0 V s–1) for each electrode. The corresponding areal capacitance values as a function of scan rate are summarized in Figure f. The evolution of CV shape with increasing scan rate provides mechanistic insight: electrodes dominated by capacitive storage tend to preserve a quasi-rectangular response, whereas electrodes with a stronger faradaic characteristic exhibit broadened/shifted redox features and increasing distortion at high scan rates. Scan-rate-dependent CVs of (a) bare CNW, (b) CNW-dominant, (c) WS2-rich, (d) S-oxide, and (e) WOx electrodes collected from 0.1 to 1.0 V s–1 (0.1 step). (f) Specific capacitance versus scan rate calculated by CV curves. Figure (a) shows the cyclic voltammograms of the bare CNW electrode collected within −1.0 to 0 V (vs Ag/AgCl) at scan rates from 0.1 to 1.0 V s–1. The CV profiles exhibit a well-preserved quasi-rectangular shape over the entire scan-rate range, with no discernible redox peaks, indicating that the charge storage of bare CNW is dominated by electric double-layer capacitance (EDLC) rather than Faradaic processes. The anodic/cathodic branches are nearly symmetric, suggesting reversible adsorption/desorption of ions and stable interfacial polarization within the chosen potential window. At 0.1 V s–1, the areal specific capacitance of bare CNW is 4.6 mF cm–2, serving as the baseline capacitive contribution of the carbon framework. Although the absolute capacitance is modest, the strong retention of the rectangular CV shape up to 1.0 V s–1 implies efficient charge transport through the conductive CNW network and rapid ion access to electrochemically active surface sites. Only a slight rounding near the vertex regions at higher scan rates is observed, which is typically attributable to ohmic resistance and finite ion-transport kinetics, yet the overall EDLC-like behavior remains robust. Consequently, bare CNW provides a stable, rate-tolerant reference electrode, enabling a clear assessment of additional pseudocapacitive contributions upon subsequent WS2/oxide incorporation in hybrid electrodes. As shown in Figure (b), CNW-dominant electrode, where XPS and Raman revealed a high sp2-carbon fraction, strong D/G bands, and only weak WS2/WO x features, consistent with a conductive CNW scaffold decorated with a low coverage of WS2/WO x nanodomains. Its CVs retain a quasi-rectangular shape even at 1.0 V s–1, with only a slight distortion, indicative of electric double-layer capacitive (EDLC) behavior with minor faradaic contribution. At 0.1 V s–1, CNW-dominant electrode shows C 0.1 * ≈ 6.4 mFcm–2, corresponding to ∼1.4× the capacitance of bare CNW, while at 1.0 V s–1, C 1.0 * ≈ 5.1 mFcm–2 with a rate retention C 1.0 */C 0.1 * ≈ 0.8. This combination of moderate capacitance enhancement and the best rate capability among the four representatives indicates that the charge storage is still largely governed by the high-conductivity CNW network, with only a limited number of additional redox-active sites. Such behavior is consistent with carbon-dominant EDLC electrodes where the introduction of a small amount of redox-active species modestly increases capacitance without severely penalizing kinetics. , In the context of structural analysis, the CNW-dominant sample serves as a kinetic “baseline” reflecting the intrinsic advantages of the CNW scaffold (low R ct, low diffusion impedance) with only weak pseudocapacitive contributions from WS2/WO x . In CV, the WS2-rich electrode exhibits more pronounced rounding at the edges of the voltammograms compared with the CNW-dominant sample, suggesting that faradaic processes associated with WS2 lattice redox are superimposed on the EDLC background. At 0.1 V s–1, the WS2-rich sample achieves C 0.1 * ≈ 40 mFcm–2, while C 1.0 * ≈ 13.3 mFcm–2 with a rate retention of ≈ 0.33. The approximately 2-fold increase in C 0.1 * relative to the CNW-dominant sample at the expense of somewhat reduced rate retention reflects the emergence of surface-confined pseudocapacitance associated with the WS2 layers. In TMD-based supercapacitor electrodes, such behavior is typically attributed to fast, near-surface redox reactions (e.g., reversible adsorption/intercalation of cations at edge and defect sites) that occur concomitantly with double-layer charging on the carbon scaffold. ,, Thus, the WS2-rich electrode can be understood as a mixed EDLC/pseudocapacitive system in which the CNW network ensures efficient electron transport and ion access, while the WS2 nanosheets provide additional faradaic storage without fully sacrificing the rate capability. A direct comparison of bare CNW, CNW-dominant, and WS2-rich electrodes clarifies how increasing WS2 coverage modifies the charge-storage behavior. Bare CNW shows a nearly ideal quasi-rectangular CV response over the full scan-rate range, consistent with EDLC-dominated storage, and its capacitance decreases only modestly from 4.6 to 3.15 mF cm–2 as the scan rate increases from 0.1 to 1.0 V s–1. The CNW-dominant electrode retains a largely capacitive CV shape and exhibits the best rate capability among the hybrid electrodes, with the capacitance changing from 6.4 to 5.10 mF cm–2, indicating that the exposed CNW network still governs charge transport, while only a limited number of additional redox-active sites are introduced. In contrast, the WS2-dominant electrode delivers a much larger capacitance, increasing to 40 mF cm–2 at 0.1 V s–1 but decreasing to 13.3 mF cm–2 at 1.0 V s–1, reflecting the emergence of substantial WS2-derived pseudocapacitive charge storage together with greater kinetic limitation at a high scan rate. Thus, increasing WS2 coverage markedly enhances low-rate charge storage, but this gain is accompanied by reduced rate retention as the response becomes increasingly influenced by faradaic processes. The CV curves of the S-oxide-rich electrode display more obvious deviations from the ideal rectangular shape than WS2-rich, including broad humps indicative of distributed redox processes. At 0.1 V s–1, the S-oxide-rich electrode reaches C 0.1 * ≈ 52.7 mFcm–2, which is higher than WS2-rich electrode, while its high-rate performance remains similar (C 1.0 * ≈ 16 mFcm–2, rate retention ≈ 0.30). This trend suggests that the introduction of S-oxide and 1T′-like defective domains increases the density of electrochemically active sites and enhances pseudocapacitive charge storage at low scan rates but does not drastically deteriorate the kinetics compared with the WS2-rich case. Such behavior is characteristic of defect-engineered pseudocapacitive materials, where structural disorder and surface oxidation introduce additional redox-active states and broaden the potential window for fast surface redox without transforming the system into a slow, diffusion-limited battery-type electrode. , Therefore, the S-oxide-rich electrode can be regarded as an optimized balance between high capacitance and still acceptable rate capability, supported by a chemically and structurally more complex WS2/S-oxide/WO x interface than in the WS2-rich electrode. For the WO x -dominant electrode, the CV curves show the strongest deviation from ideal rectangularity among all samples, with pronounced, broad redox features consistent with multielectron WO x redox processes. At 0.1 V s–1, the WO x -dominant electrode delivers the largest capacitance of all tested electrodes, C0.1 * ≈ 91 mFcm–2, corresponding to ≈ 20 × the bare CNW value. Even at 1.0 V s–1, it retains C 1.0 * ≈ 21.7 mFcm–2 with a rate retention of ≈0.24, comparable to the WS2-rich and S-oxide-rich groups despite the much thicker oxide layer. This combination of very high capacitance and moderate rate capability is characteristic of tungsten oxide-based pseudocapacitors, in which fast, surface-near insertion/extraction of protons or small cations into WO x tunnels or layers governs charge storage. The strongly enhanced capacitance of the WO x -dominant electrode relative to the WS2-rich electrode and S-oxide-rich electrode indicates that, in this system, WO x provides the dominant pseudocapacitive contribution, whereas WS2 and CNW mainly play the role of conductive support and structural scaffold. The fact that the rate retention remains comparable to the WS2-based electrodes suggests that the WO x formed here is not purely insulating but retains sufficient electronic/ionic conductivity to support rapid redox at the time scales probed by CV, in line with previous reports on nanostructured WO x supercapacitor electrodes. , A direct correlation between characterization and electrochemical performance can now be established for all five electrodes. Bare CNW, which shows the most open graphitic framework and no additional WS2/WO x signatures, exhibits the smallest capacitance (4.6 mF cm–2 at 0.1 V s–1; 3.15 mF cm–2 at 1.0 V s–1) and the most ideal EDLC-like response. CNW-dominant, where SEM/Raman/XPS indicate a thin and discontinuous WS2/WS x O y decoration on the CNW scaffold, shows only a modest capacitance increase (6.4 and 5.1 mF cm–2 at 0.1 and 1.0 V s–1, respectively) but the best rate retention among the hybrid electrodes because the exposed CNW network still dominates charge transport. WS2-rich, which combines continuous WS2 coverage with relatively low WO x content, delivers a much larger capacitance (40 and 13.3 mF cm–2) because additional WS2-related redox sites are introduced, while the open CNW framework remains accessible. S-oxide-rich further increases the capacitance (52.7 and 16.0 mF cm–2), consistent with the Raman/XPS evidence for defect-rich and partially oxidized WS2 domains that create additional electrochemically active states. WO x -dominant exhibits the highest capacitance (91 and 21.7 mF cm–2), in agreement with its high W6+ fraction and strong W–O vibrational signature, demonstrating that WO x -derived redox reactions provide the largest pseudocapacitive contribution in the present system. Thus, the electrochemical hierarchy of bare CNW < CNW-dominant < WS2-rich < S-oxide-rich < WO x -dominant is not arbitrary, but directly reflects the progressive increase in active WS2/WO x -related redox centers superimposed on the conductive CNW scaffold. The rate-dependent apparent capacitance C*­(v) extracted from cyclic voltammograms (Figure ) clearly differentiates the charge-storage behavior of bare CNW and the ATTT-CNW composite electrodes. At a slow scan rate of 0.1 V s–1, the specific capacitance increases monotonically from 4.6 mF cm–2 for bare CNW to 6.4, 40, 52.7, and 91 mF cm–2 for the CNW-dominant, WS2-rich, S-oxide-rich, and WO x -dominant electrodes, respectively. At 1.0 V s–1, the same series exhibits 3.15, 5.10, 13.3, 16.0, and 21.7 mF cm–2. Thus, the introduction of WS2 and its oxide derivatives onto CNW yields more than an order-of-magnitude increase in C* at a low scan rate, but a much more modest enhancement at a high scan rate. This behavior already suggests that a large fraction of the additional capacitance in ATTT-CNW composites is limited by diffusion-controlled processes. To elucidate the intrinsic charge-storage kinetics, Dunn’s method was applied to bare CNW and to four representative ATTT-CNW electrodes corresponding to the main structural groups identified by XPS/SEM. Within Dunn’s formalism, the current at a given potential (V) and scan rate (v) is decomposed into a surface-controlled term that scales with (v) and a diffusion-controlled term that scales with (v 1/2). , The current is written as 2 i(V)=k1(V)v+k2(V)v1/2 where (k 1(V)v) and (k 2(V)v 1/2) are potential-dependent contributions associated with surface-controlled and diffusion-controlled processes, respectively. Rearrangement gives 3 i(V)v1/2=k1(V)v1/2+k2(V) so that (k 1(V)) and (k 2(V)) are obtained by a linear regression of (i(V)/v 1/2) versus (v 1/2) over the set of scan rates. Once the two coefficients are known, the surface-controlled and diffusion-controlled current components at a given scan rate are reconstructed as 4 icap(V)=k1(V)v 5 idiff(V)=k2(V)v1/2 allowing k 1(V) and k 2(V) to be obtained from linear fitting at each potential. The corresponding current components were integrated over the potential window to obtain Q cap and Q diff, and the relative contributions were calculated as f cap = Q cap/(Q cap + Q diff) and f diff = 1 – f cap. The scan-rate-dependent surface-controlled fractions are summarized in Figure a, and the relative surface-controlled and diffusion-controlled contributions at 0.1 and 1.0 V s–1 are compared in Figure b. This methodology has been widely applied to distinguish double-layer-type and pseudocapacitive contributions in nanostructured transition-metal oxides and related materials. − Figure b shows stacked bar plots showing the relative surface-controlled and diffusion-controlled contributions (in %) obtained from Dunn’s analysis of the cathodic CV branches (−1.0–0.0 V) for bare CNW, CNW-dominant, WS2-rich, S-oxide-rich, and WO x -dominant electrodes at 0.1 and 1.0 V s–1. (a) Dunn analysis of surface-controlled fractions as a function of scan rate. (b) Relative surface-controlled and diffusion-controlled fractions from Dunn analysis at 0.1 and 1.0 V s–1. For bare CNW, the cathodic branch shows a surface-controlled fraction of approximately 69.8% at 0.1 V s–1, which increases to ∼88.0% at 1.0 V s–1. This trend confirms that bare CNW behaves as a typical EDLC-type electrode, with most of the charge stored via fast double-layer formation on accessible graphitic surfaces. Using the C*­(v) data from Figure , this corresponds to an absolute surface-controlled contribution of ∼3.21 mF cm–2 at 0.1 V s–1 and ∼2.77 mF cm–2 at 1.0 V s–1. The remaining ∼1.39 and 0.38 mF cm–2 at 0.1 and 1.0 V s–1, respectively, are assigned to diffusion-controlled processes, which likely arise from nonideal transport in pores and minor redox-active defect sites. In contrast, the CNW-dominant composite exhibits substantially lower surface-controlled fractions of 39.2% and 67.1% at 0.1 and 1.0 V s–1, respectively. Although its apparent capacitance is higher than that of bare CNW (4.6 → 6.4 mF cm–2 in the sample-to-sample comparison at 0.1 V s–1), only ∼2.51 mF cm–2 is surface-controlled and ∼3.89 mF cm–2 is diffusion-controlled. At 1.0 V s–1, the surface-controlled contribution grows to ∼3.42 mF cm–2 out of 5.10 mF cm–2, while the diffusion-controlled component decreases to ∼1.68 mF cm–2. These results indicate that the additional capacitance obtained in the CNW-dominant regime is largely associated with diffusion-limited processes, presumably due to the formation of WS2/oxide clusters that introduce redox-active sites but also increase transport path length, consistent with previous reports on pseudocapacitive transition-metal oxides. , The WS2-rich electrode exhibits a more pronounced diffusion-dominated character. At 0.1 V s–1, the surface-controlled fraction is only 12.4%, yielding a surface-controlled contribution of ∼4.34 mF cm–2 and a diffusion-controlled contribution of ∼30.66 mF cm–2 out of the total 40 mF cm–2. Even at 1.0 V s–1, the surface-controlled fraction reaches only 30.9%, corresponding to ∼4.11 mF cm–2 surface-controlled and ∼9.19 mF cm–2 diffusion-controlled capacitance. Notably, the absolute surface-controlled component of WS2-rich remains comparable to or slightly larger than that of bare CNW, whereas the large increase in C* relative to bare CNW is almost entirely due to diffusion-controlled storage. This strongly suggests that Na+ insertion/extraction into few-layer WS2 and associated interfacial redox reactions account for the majority of the capacity gain, analogous to pseudocapacitive behavior reported for layered oxides and MXenes. , The diffusion-dominated character becomes more pronounced in the S-oxide-rich electrode. Its surface-controlled fractions are 9.7% at 0.1 V s–1 and 25.4% at 1.0 V s–1. In absolute terms, only ∼5.11 mF cm–2 of the total 52.7 mF cm–2 at 0.1 V s–1 is surface-controlled, whereas ∼47.59 mF cm–2 is diffusion-controlled; at 1.0 V s–1, the corresponding values are ∼4.06 and 11.94 mF cm–2. These results indicate that the capacitance enhancement in the S-oxide-rich regime originates primarily from diffusion-influenced pseudocapacitive processes associated with the more defective and partially oxidized WS2/WO x interfacial structure, rather than from a large increase in the fast surface-controlled component. The WO x -dominant electrode shows the most extreme behavior, with the lowest surface-controlled fractions among all samples: 6.6% at 0.1 V s–1 and 18.3% at 1.0 V s–1. Using the measured capacitances from the CV analysis, these fractions correspond to an approximately surface-controlled contribution of ∼6.01 mF cm–2 and a diffusion-controlled contribution of ∼84.99 mF cm–2 at 0.1 V s–1, and to ∼3.97 and 17.73 mF cm–2, respectively, at 1.0 V s–1. This strongly diffusion-dominated response is consistent with the relatively thicker and more continuous WO x -rich overlayer inferred from SEM, Raman, and XPS. As the oxidized overlayer becomes more continuous, fast surface-confined charge storage is increasingly restricted to the outermost region, whereas a larger fraction of the stored charge requires near-surface cation insertion/extraction within the WO x -rich layer together with W-valence switching. This interpretation is consistent with the broad WO x -related CV feature, its scan-rate-dependent distortion, and the increased transport limitation indicated by the impedance results. , Taken together, Figures and reveal that the absolute surface-controlled contribution C cap at 0.1 V s–1 increases only modestly when going from bare CNW (∼3.21 mF cm–2) to the ATTT-CNW composites (∼2.5–6.0 mF cm–2), whereas the diffusion-controlled contribution C diff increases dramatically (from ∼1.39 mF cm–2 for bare CNW to ∼31–85 mF cm–2 for WS2-rich/S-oxide-rich/WO x -dominant). The enormous enhancement in C* observed for ATTT-CNW, particularly in the WO x -dominant regime, originates predominantly from diffusion-controlled pseudocapacitive/redox processes rather than from an increase in the surface-controlled (capacitive) component (i.e., EDLC plus fast surface-confined redox). This explains why the relative advantage in C* over bare CNW is much smaller at 1.0 V s–1 than at 0.1 V s–1: diffusion-controlled contributions are strongly suppressed at a high scan rate, whereas the intrinsic surface-controlled component remains in the range of 3–6 mF cm–2 across all samples. In summary, the Dunn analysis clarifies that the capacitance enhancement in the hybrid electrodes does not arise primarily from an increase in the fast surface-controlled component but from the emergence of a large diffusion-controlled pseudocapacitive contribution. Bare CNW shows the highest surface-controlled fraction (69.8% at 0.1 V s–1 and 88.0% at 1.0 V s–1), consistent with its EDLC-dominant carbon surface. In contrast, the hybrid electrodes display progressively lower surface-controlled fractions as the coating becomes more continuous and more oxidized: 39.2/67.1% for CNW-dominant, 12.4/30.9% for WS2-rich, 9.7/25.4% for S-oxide-rich, and 6.6/18.3% for WO x -dominant at 0.1/1.0 V s–1. Importantly, the absolute surface-controlled contribution remains within a relatively narrow range, whereas the diffusion-controlled contribution increases dramatically from bare CNWs to the WS2/WO x -containing electrodes. This means that the large increase in areal capacitance is mainly generated by ion-coupled redox processes within WS2-, defect-rich WS2/SOx-, and especially WO x -containing layers, rather than by a simple increase in the electrochemically accessible carbon surface area. In other words, SEM/Raman/XPS indicates why more active phases are created, and Dunn analysis quantitatively shows that these additional phases store charge predominantly through diffusion-influenced pseudocapacitive processes. Electrochemical impedance spectroscopy (EIS) was employed to compare the interfacial polarization and transport characteristics of five electrodes (bare CNW, CNW-dominant, WS2-rich, S-oxide-rich, and WO x -dominant), as shown by the full-range and magnified Nyquist plots in Figure a,b, respectively. All samples exhibited a comparable high-frequency intercept (Rs, ca. 4.13–4.18 Ω), indicating that the series contribution from the electrolyte, current collectors, and external contacts was largely invariant across the data set. EIS spectra of samples synthesized at different ATTT concentrations and microwave powers: CNW-dominant (0.05 M, 0.5 kW), WO x -dominant (0.05 M, 1 kW), S-oxide-rich (0.1 M, 0.5 kW), and WS2-rich (0.1 M, 1 kW). (a) Nyquist plots (−Z″ vs Z′) over the full frequency range and (b) magnified high-frequency region. Within the CNW-based series (bare CNW → CNW-dominant → WO x -dominant), the Nyquist response was dominated by a pronounced low-frequency capacitive tail, consistent with electric double-layer capacitor (EDLC)-like behavior in porous carbon architectures. , Notably, the WO x -dominant electrode showed the largest increase in the real component (Z′) toward low frequency, suggesting enhanced distributed resistance and/or diffusion-related limitations superimposed on capacitive storage, which is commonly observed when redox-active metal oxides contribute to charge storage but impose additional ionic/electronic transport constraints. In contrast, WS2-rich and S-oxide-rich electrodes displayed a clearer midfrequency depressed semicircle, implying a stronger interfacial polarization component (e.g., charge transfer and/or film resistance) relative to the CNW-dominant electrodes. The S-oxide-rich sample exhibited a larger semicircle than the WS2-rich one, consistent with an increased kinetic barrier attributable to oxidized surface species (e.g., sulfur oxide/oxide-containing layers) that can elevate interfacial resistance even when wettability is improved. Finally, because the low-frequency impedance magnitude differs substantially between the CNW family and the WS2/S-oxide group, quantitative comparisons of capacitance-related metrics should be made only after confirming identical frequency windows (especially the lowest frequency) and measurement conditions. The EIS results are fully consistent with the structure–performance correlation derived from SEM, Raman, XPS, CV, and Dunn analysis. Because Rs remains nearly unchanged for all electrodes, the performance difference does not originate from the electrolyte or external circuit. Instead, it is governed by the interfacial layer created on the CNW scaffold. The CNW-dominant electrode retains a mainly capacitive tail and low interfacial polarization because the thin and discontinuous coating does not significantly disturb the conductive carbon network. WS2-rich and especially S-oxide-rich show a clearer depressed semicircle, indicating that once the active layer becomes more continuous and more oxidized, charge-transfer and film-related resistance become more important. WO x -dominant, despite showing the highest capacitance, exhibits the largest increase in low-frequency Z′, confirming that the thick WO x -dominant overlayer introduces a substantial distributed resistance/diffusion penalty. Therefore, the present data set consistently indicates that increasing oxidation and coating continuity enhances pseudocapacitive charge storage but simultaneously increases interfacial and transport resistance. The Ragone plot derived from the CV results in the three-electrode configuration (Figure ) shows the expected trade-off, that is, a monotonic decrease in energy density with increasing power density for all electrodes, while the performance envelope is systematically shifted by phase/chemistry control. At a given power density, the energy density follows the order WO x -dominant > S-oxide-rich > WS2-rich > CNW-dominant > bare CNW. In the apparent per-electrode comparison derived from the three-electrode CV data, the WOx-dominant electrode retains the highest apparent energy at a given apparent power across the tested range. Ragone plot of the bare CNW, CNW-dominant, WS2-rich, S-oxide-rich, and WO x -dominant electrodes, comparing the energy–power trade-off across the tested operating range. Conclusion In this study, we developed a simple plasma-chemical conversion strategy to directly fabricate binder-free ATTT-carbon nanowall (ATTT-CNW) electrodes by combining solution deposition of ammonium tetrathiotungstate (ATTT) with downstream H2 remote plasma while preserving the three-dimensional conductive CNW scaffold. The primary objective was to clarify how precursor loading and plasma power govern the resulting WS2/WO x surface chemistry and, in turn, the electrochemical charge-storage behavior of the electrodes. We found that systematic variation of ATTT concentration and microwave power reproducibly routes the surface into CNW-dominant, WS2-rich, S-oxide-rich, and WO x -dominant regimes, as confirmed by correlated SEM, Raman, and XPS analyses. These compositionally distinct interfaces produced clearly different electrochemical responses: compared with bare CNW (4.6 mF cm–2 at 0.1 V s–1), the areal capacitance increased to 6.4, 40, 52.7, and 91.1 mF cm–2 for the CNW-dominant, WS2-rich, S-oxide-rich, and WO x -dominant electrodes, respectively, with the WO x -dominant electrode also retaining the highest capacitance at 1.0 V s–1 (21.7 mF cm–2). Dunn analysis further revealed that this performance enhancement arises mainly from increased diffusion-controlled pseudocapacitive charge storage rather than from a large increase in the fast surface-controlled component, whereas EIS showed that thicker and more oxidized interfacial layers simultaneously introduce higher interfacial and transport resistance. Therefore, the major finding of this work is that the electrochemical behavior of ATTT-CNW electrodes is governed not simply by the amount of deposited active material, but by the balance between the open CNW transport network, the fraction of WS2, the degree of defect/S-oxide formation, and the extent of WO x generation at the interface. These results provide a clear structure–chemistry–performance relationship and practical design guidance for engineering high-performance, binder-free WS2/WO x @CNW hybrid electrodes by remote plasma conversion.