Growth mechanism and formaldehyde sensing of mixed phase cobalt oxide nanowalls Authors: Sushil Barala, Sri Aurobindo Panda, Arnab Hazra, Subhashis Gangopadhyay Published: 2025 ABSTRACT Surface morphology of transition metal oxide (TMO)-based semiconductor nanostructures plays a crucial role in determining their device performance, gas sensing abilities, and catalytic properties. Additionally, chemical stoichiometry of oxide materials significantly influences their structural, electronic, and optical characteristics. In this study, we present the successful fabrication of two-dimensional (2D) porous nanowalls (67 nm thick) of mixed cobalt oxide (CoO + Co3O4) phases using a controlled thermal oxidation process, which are later used for selective formaldehyde sensing. High-purity cobalt (Co) films were deposited on glass substrates via vacuum assisted electron beam evaporation and subsequently oxidized in air ambient at temperatures up to 800 °C. Below 400 °C, the CoO phase stabilizes the highly porous 2D nanowalls morphology, while at higher temperatures Co3O4 phase promotes the homogeneous nanograin morphology. EDX, XPS and HRTEM analyses collectively confirm the formation of a CoO (core) / Co3O4 (shell) nanowalls structure. These porous nanowalls of mixed oxide phase significantly impacted on the chemiresistive gas sensing ability of cobalt oxide thin films. Mixed phase nanowalls exhibited significantly higher sensitivity (~ 60%) to formaldehyde vapour, compared to the nanograin Co3O4 films. Furthermore, these unique mixed oxide nanowalls displayed an excellent selectivity towards low concentration formaldehyde vapour against other volatile organic compounds (VOCs). All these findings provide valuable insights into the phase stabilization and growth mechanisms of CoOx nanowalls, highlighting their potential for high-performance gas sensing applications.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-025-29321-7. Introduction Transition metal oxides have attracted considerable attention due to their diverse electronic, optical, and catalytic properties, making them highly promising for various technological applications. The ability to modulate surface morphology and stoichiometry at the nanoscale is critical for optimizing these properties to meet specific technological demands. The surface morphology, which includes features such as porosity, grain size, and dimensionality, plays a key role in governing material-environment interactions, influencing charge transport, gas adsorption, and catalytic efficiency. By systematically controlling growth parameters, a wide scale surface morphologies of TMO-based semiconductor nanostructures such as nanorods, nanowalls, nanoflowers, and nanograins can be achieved1–3, each exhibiting their unique functional characteristics. Furthermore, the stoichiometry is a crucial determinant of their phase composition and defect states, directly impacting electrical conductivity, optical bandgap, and sensing capabilities4. Therefore, a deeper understanding of morphological evolution and precise stoichiometric control is essential for advancing the performance and applicability of these nanostructures in next-generation devices. Among diverse range of TMOs, cobalt oxide-based nanostructures have emerged as a highly promising class of materials due to their exceptional electrochemical, optical, and catalytic properties. Cobalt oxide primarily exists in two stable phases such as rock salt CoO and spinel Co3O4, both of which exhibit p-type conductivity. In the spinel structure (Fd-3 m space group), oxygen ions (O²⁻) occupy a face-centered cubic (FCC) lattice, while cobalt ions (Co3+ and Co2+) are positioned in octahedral (16d) and interstitial tetrahedral (8a) sites, respectively5. Additionally, the spinel structure is characterized by two direct energy bandgaps of approximately 1.5 eV and 1.9 eV, making it highly suitable for optoelectronic and catalytic applications6,7. The coexistence of mixed-valence states (CoO and Co3O4) endows cobalt oxide with excellent redox activity, making it a promising material for energy storage, gas sensing, and catalytic applications. In gas sensing, CoOx nanostructures demonstrate outstanding sensitivity and selectivity towards VOCs and toxic gases, owing to their high surface reactivity and efficient charge transport properties8. The ability to tailor these nanostructures in terms of morphology and phase composition further enhances their functional performance in various technological applications such as electrochromic displays9. As a result, ongoing research is directed towards optimizing synthesis methods and structural modifications to fully exploit their potential for future energy and environmental applications. Cobalt oxide nanostructures are commonly synthesized through various chemical growth techniques, such as hydrothermal synthesis8, spray pyrolysis9, sol-gel10, chemical vapour deposition11. By controlling precursor chemistry, reaction conditions, and surfactant-assisted growth, various nanostructures such as nanorods, nanowires, nanosheets, nanospheres, and hierarchical architectures are obtained12. Among these, hydrothermal method is most widely employed which can facilitate an anisotropic growth, enabling the formation of well-defined 1D and 2D nanostructures with enhanced electrochemical properties13. Similarly, sol-gel synthesis offers the advantage of producing porous nanostructures with high surface areas, making them particularly suitable for catalytic and sensor applications14,15. However, these chemical synthesis methods often require surfactants, organic solvents, and post-synthesis treatments, necessitating additional purification steps. Moreover, they frequently suffer from low material purity, reduced physical and thermal stability, and weak adhesion to substrates, which can hinder their practical applications. Table 1A collection of various reported CoOx nanostructures grown using different synthesis methods.S. No.Cobalt oxideMorphologySynthesismethodApplicationRef.1CoONanoflowersAACVDElectrocatalytic 16 2Co3O4NanorodsHydrothermalCO oxidation 12 3Co3O4MicrospheresSolvothermalPhotocatalytic 17 4Co3O4FlowerlikeMicrowaverefluxElectrocatalytic 3 5Co3O4NanorodsHydrothermalGas sensing 15 6Co3O4NanosheetsHydrothermaldecompositionSupercapacitor 13 7Co3O4NanosheetsHydrothermalGas sensing 15 8Co3O4NanobeltThermaldecompositionCatalyst 18 9Co3O4NanocubeHydrothermalGas sensing 19 10Co3O4NanograinsThermaldecompositionElectrocatalytic 14 11Co3O4NanowallsThermaloxidationField Emission 20 12Co3O4NanowireLow-pressurethermaloxidationField Emission 21 13Co3O4NanowallsO2 plasmatreatmentField Emission 21 14Co3O4NanoflowersHydrothermalSupercapacitor 22 15CoO + Co3O4NanowallsThermaloxidationGas sensingThis work Table 1A collection of various reported CoOx nanostructures grown using different synthesis methods.S. No.Cobalt oxideMorphologySynthesismethodApplicationRef.1CoONanoflowersAACVDElectrocatalytic 16 2Co3O4NanorodsHydrothermalCO oxidation 12 3Co3O4MicrospheresSolvothermalPhotocatalytic 17 4Co3O4FlowerlikeMicrowaverefluxElectrocatalytic 3 5Co3O4NanorodsHydrothermalGas sensing 15 6Co3O4NanosheetsHydrothermaldecompositionSupercapacitor 13 7Co3O4NanosheetsHydrothermalGas sensing 15 8Co3O4NanobeltThermaldecompositionCatalyst 18 9Co3O4NanocubeHydrothermalGas sensing 19 10Co3O4NanograinsThermaldecompositionElectrocatalytic 14 11Co3O4NanowallsThermaloxidationField Emission 20 12Co3O4NanowireLow-pressurethermaloxidationField Emission 21 13Co3O4NanowallsO2 plasmatreatmentField Emission 21 14Co3O4NanoflowersHydrothermalSupercapacitor 22 15CoO + Co3O4NanowallsThermaloxidationGas sensingThis work A collection of various reported CoOx nanostructures grown using different synthesis methods. Microwave reflux Hydrothermal decomposition Thermal decomposition Thermal decomposition Thermal oxidation Low-pressure thermal oxidation O2 plasma treatment Thermal oxidation In contrast, physical growth methods serve as an effective alternative, enabling the direct fabrication of thin cobalt oxide films with precisely defined nanoscale morphology, controlled chemical composition, and minimal contamination, eliminating the need for surfactants or chemical precursors22. Specifically, physical growth techniques offer superior control over the stoichiometry, purity, and crystallinity of as-grown cobalt oxide nanostructures. Additionally, they provide greater precision in regulating film thickness, grain size, and surface morphology, which are critical for applications requiring well-defined structural and electronic properties. Physical growth methods are also compatible with large-scale and industrial fabrication processes, offering a more environmentally sustainable alternative to chemical synthesis by minimizing the use of hazardous reagents. Recent studies on pulsed laser deposition (PLD), magnetron sputtering, and molecular beam epitaxy (MBE) have demonstrated the homogeneous formation of CoOₓ thin films, nanowires, and nanoparticles with superior structural integrity15,17–19. However, these techniques often require expensive instrumentation and specialized technical expertise, which can limit their accessibility. Consequently, controlled thermal oxidation of thin Co films presents a simple and cost-effective approach within the framework of physical vapor deposition (PVD) for the growth of cobalt oxide nanostructures with tailored surface morphology. Furthermore, this simple technique enables the development of cobalt oxide nanostructures with controlled defect densities and optimized surface properties. By capitalizing on the advantages of thermal oxidation process, controlled formation of cobalt oxide nanostructures can be utilized in gas sensing applications. A collection of various reported cobalt oxides nanostructured morphologies using different growth methods is summarized in Table 1. In general, metal oxide-based nanostructures play a vital role in gas sensing applications due to their high surface-to-volume ratio, tunable electronic properties, and rapid gas adsorption/desorption kinetics, enabling high sensitivity, selectivity, and fast response times for environmental monitoring and industrial safety. Additionally, they offer advantages such as ease of fabrication, structural stability, and enhanced reproducibility. However, the performance of chemiresistive gas sensors is strongly influenced by their surface morphology, which governs the interaction between gas molecules and the sensing material. More precisely, the morphology of metal oxides dictates the availability of active sites for gas interaction, electron transport pathways, and overall sensor performance. Furthermore, the stoichiometry of metal oxides plays a crucial role in determining their electronic structure, oxygen vacancy concentration, and defect states, key factors in the gas adsorption and charge transfer processes. Therefore, by modulating the nanostructure surface morphology and chemical stoichiometry, it is possible to significantly enhance the sensing performance of chemiresistive gas sensors, including sensitivity, selectivity, response/recovery time, and operating temperature. Among n-type metal oxides, ZnO23–25, WO326, SnO227, TiO228, Fe2O329 and In2O330 are widely employed for gas sensing applications, while Co3O431, NiO32, and CuO33 have been explored as p-type oxides. Generally, n-type oxides exhibit higher sensitivity, and a broader range of nanoscale surface morphologies compared to p-type materials. However, their performance is often hindered by poor humidity tolerance, a limitation that can be mitigated to some extent by incorporating p-type oxides, which are less sensitive to ambient moisture34,35. In this regard, cobalt oxide has emerged as a promising alternative, offering advantages such as variable oxidation states and diverse crystal structures. Its high surface-to-volume ratio exhibits excellent surface redox chemistry and strong interactions with ambient gas molecules, making them ideal candidates for gas and VOC sensing. Various techniques have been employed to enhance the sensing performance of cobalt oxide-based chemiresistive gas sensors. Modulating surface morphology at the nanoscale can significantly improve the sensitivity of Co3O4 gas sensors by increasing the number of adsorption sites36. In addition, heterostructures and nanocomposites play a critical role in improving sensitivity and selectivity by providing more active sites, optimizing charge transfer, and modulating the conduction channel at the interface. Core-shell architectures can further enhance gas diffusion and provide a larger active surface area for adsorption37. Defect state engineering also plays a crucial role in improving gas sensor performance38. Introducing oxygen vacancies or modifying defect density generates active sites that facilitate stronger adsorption and reaction with gas molecules39. The synergistic effect of defect states and heterojunctions also enhances the sensitivity, selectivity, and lower detection limits of Co3O4-based sensors40. Formaldehyde is a major indoor and outdoor air pollutant with carcinogenic properties, emitted from various sources such as wood-based materials, textiles, insulation, combustion processes, and atmospheric reactions of VOCs. It is widely used in adhesives, coatings, disinfectants, and preservatives, posing significant health risks. The World Health Organization (WHO) classifies formaldehyde as a Group 1 human carcinogen, recommending indoor air concentrations not to exceed 0.08 ppm41. Occupational exposure limits set by OSHA include a long-term exposure limit (LTEL) of 0.75 ppm over an 8-hour TWA and a short-term exposure limit (STEL) of 2 ppm for 15 min. Acute exposure can cause irritation of the eyes, nose, and throat, while chronic exposure may lead to respiratory conditions, allergic reactions, and increased cancer risk. Concentrations above 100 ppm are considered immediately dangerous to life and health (IDLH). Given its widespread use and associated hazards, early detection of low-concentration formaldehyde vapor is crucial for ensuring public health and safety. In this study, mixed-phase cobalt oxide nanowalls were synthesized using a simple physical growth approach, involving e-beam evaporation of cobalt thin films followed by controlled thermal oxidation. The growth mechanism of these 2D oxide nanowalls was investigated, and their potential for chemiresistive gas sensing was explored for the selective detection of low-concentration formaldehyde vapor. Mixed oxide phase offers enhanced stability and improved functionalities compared to their single-phase counterparts. The porous morphology of the cobalt oxide nanowalls provides a significantly enlarged active surface area, enhancing gas adsorption and diffusion. Additionally, the grain boundaries and defects associated with the mixed oxide phases act as crucial sites for surface redox reactions, thereby improving sensor sensitivity. Therefore, the ability to precisely modulate both surface morphology and stoichiometry in CoOx nanostructures opens new avenues for optimizing its performance in next-generation gas sensors, photocatalysts, and electronic devices. Experimental details Cobalt oxide thin-film-based gas sensors were fabricated through three-step process involving (i) cobalt thin film deposition, (ii) thermal oxidation, and (iii) sensor fabrication. The deposition and oxidation approach used here builds upon our previously reported method42, with the present study focusing on growth mechanism analysis and gas sensing behavior. High-purity cobalt (99.99%, Alfa Aesar) shots were used as the source material for thin-film deposition. Prior to deposition, glass and quartz substrates were sequentially cleaned ultrasonically in acetone, isopropanol, and deionized water for 15 min each, followed by drying under nitrogen gas. Cobalt films were deposited using an electron beam evaporation system (Hind High Vacuum) operated at a base pressure below 10⁻⁵ mbar. The accelerating potential of the electron beam was maintained at ~ 4.5 kV, while the deposition rate was precisely controlled by adjusting the emission current (~ mA). The film thickness was monitored in situ using a quartz crystal microbalance (QCM), with a deposition rate of ~ 1.2 Å/s, resulting in cobalt films of ~ 200–250 nm thickness. Afterwards, deposited cobalt films were thermally oxidized in a programmable muffle furnace (Tempcon Instruments) in ambient air over a temperature range up to 800 °C. The heating and cooling rates were fixed at 5 °C/min to minimize thermal stress. All oxidation was performed for 5 h with samples placed at the center of the furnace to ensure uniform heating. X-ray diffraction (XRD) was carried out on a Rigaku Miniflex-II diffractometer using Cu Kα radiation (λ = 1.5406 Å) to identify crystalline phases and estimate crystallite size, dislocation density, and lattice strain of the cobalt oxide thin films. Surface morphology was analyzed using field emission scanning electron microscopy (FESEM, FEI Apreo LoVac) to study topographical features and the effect of oxidation temperature on morphological evolution. Optical absorption spectra were recorded with a UV–Vis spectrophotometer (Shimadzu UV-2600) to evaluate the energy bandgap. Raman spectroscopy (HORIBA LabRAM HR) with a 633 nm He–Ne laser was used for phase identification, lattice vibration analysis, and structural characterization. The chemical composition and oxidation states were examined by X-ray photoelectron spectroscopy (XPS, Thermo Fisher Scientific K-Alpha) equipped with a monochromatic Al Kα source, with measurements performed under ultra-high vacuum (≤ 10⁻⁹ mbar) for high surface sensitivity. For all high resolution XPS spectra analysis, a Smart background subtraction (combination of Linear, Shirley and Tougaard) method and Powel fitting algorithm using Gaussian and Lorentzian sum are used till a convergence of 10− 5 achieved. Electrical behavior was assessed through I–V measurements at room temperature using a Physical Quantity Measurement System (PQMS, Quazar Tech) in a two-probe configuration. For IV measurements and gas sensing studies, 100 nm thick gold interdigitated electrodes were fabricated on top of the oxide films via thermal evaporation using a molybdenum finger mask (thickness ~ 0.5 mm, spacing ~ 0.5 mm). The mask was affixed directly onto the CoOx sensing film using copper tape to ensure intimate contact and prevent gaps. Gold was deposited under vacuum producing well-defined interdigitated electrodes and integrating oxide films into chemiresistive sensor devices. Sensing studies were conducted under static vapor conditions in a sealed chamber of ~ 650 mL, placed inside in a heating mantle equipped with a temperature controller. Target vapors of formaldehyde, acetone, benzene, toluene, xylene, and methanol were generated by injecting calibrated volumes of liquid precursors with a microsyringe (Hamilton 705RN, 50 µL) and allowing complete evaporation. The vapor concentration C (ppm) was calculated using the relation C = 2.46 × (V1D/VM) × 103, where ‘V1’ is the injected liquid volume, ‘D’ its density (g mL–1), ‘M’ is the molecular weight (g mol⁻¹), and ‘V’ is the chamber volume (L). Sensor resistance was continuously monitored using a digital multimeter (UNI-T UT61E+) linked to computer-controlled data acquisition software. Initially, dry (synthetic) air was introduced into the chamber, and the sensor responses to different VOCs were recorded. During recovery, synthetic air (450 sccm) was flushed through the chamber using a mass flow controller (ALICAT Scientific). The sensitivity (S) of the gas sensor was measured using the following formula. 1\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\text{S }}\left( \% \right){\text{ }}=~\frac{{\left| {{R_{Air}} - {R_{Voc}}} \right|}}{{{R_{Air}}}} \times 100$$\end{document} where RAir = Resistance of film in ambient air and RVoc = Resistance of the film after insertion of VOC. Response and recovery times were extracted from the 90% change in resistance values. However, for humidity effect on sensing studies, humid air (450 sccm) was generated by passing the same dry stream through a bubbler containing deionized water at controlled temperature, and the relative humidity (RH) inside the chamber was monitored in real time using a calibrated hygrometer. By increasing the bubbler temperature, different RH levels were achieved, enabling systematic evaluation of the device performance under varying humidity. Results and discussions Structural characterization using X-ray diffraction (XRD) Figure 1(a) presents the XRD pattern of cobalt thin films oxidized at different temperatures for a duration of five hours under ambient air conditions. The as-deposited cobalt (Co) film exhibits a strong diffraction peak corresponding to the (110) plane of metallic Co phase at 75.94°, along with three other peaks at 41.69°, 44.76°, and 47.57°, corresponding to the (100), (002), and (101) planes, respectively. These peak positions align well with the JCPDS card number 00-005-0727. It has been observed that the thermal oxidation of the Co film is significantly restricted at lower temperatures and successfully initiates only above below 200 °C, mainly due to a limited thermal diffusion and chemical reactivity. After thermal oxidation of Co films at 400 °C for five hours, the XRD pattern reveal a mixed oxide phase comprising CoO and Co3O4. The CoO phase is identified by diffraction peaks at 36.49°, 42.38°, 61.49°, and 73.67°, corresponding to the (111), (200), (220), and (311) planes, respectively [JCPDS card number 00-048-1719]. Meanwhile, the Co3O4 phase exhibits diffraction peaks at 19.00°, 31.27°, 36.85°, 38.54°, 44.81°, 55.65°, 59.36°, and 65.24°, corresponding to the (111), (220), (311), (222), (400), (422), (511), and (440) planes, respectively [JCPDS card number 00-042-1467]. Notably, the (200) and (111) planes of CoO, as well as the (220) and (311) planes of Co3O4, dominate the diffraction pattern. Following oxidation at 500 °C, a significant change is observed in the XRD pattern, as the CoO phase completely disappears, leaving only the Co3O4 phase with dominant (220) and (311) diffraction planes. Oxidation at even higher temperatures (up to 800 °C) does not result in significant changes to the XRD pattern. After oxidation at 600 °C, the (311) peak of Co3O4 becomes increasingly dominant over the (220) peak. Figure 1(b) and (c) present high-resolution spectra of the (311) and (220) diffraction peaks of Co3O4, respectively. A gradual increase in the intensity of the Co3O4 (311) peak is clearly observed in Fig. 1(b). Additionally, the presence of a mixed oxide phase in the film oxidized at 400 °C is further confirmed by the appearance of the (111) diffraction peak of CoO. A significant increase in the intensity of the Co3O4 (220) peak is also noted for samples oxidized between 400 °C and 500 °C, as illustrated in. Fig. 1XRD pattern of cobalt oxide thin films (a) oxidized at 400 °C, 500 °C and 600 °C for 5 h. Narrow scan for (b) (220) plane and (c) (311) plane of cobalt oxide samples. (d) Oxidation temperature dependent variation of crystallite size, number of crystallites, dislocation density and lattice strain. XRD pattern of cobalt oxide thin films (a) oxidized at 400 °C, 500 °C and 600 °C for 5 h. Narrow scan for (b) (220) plane and (c) (311) plane of cobalt oxide samples. (d) Oxidation temperature dependent variation of crystallite size, number of crystallites, dislocation density and lattice strain. The Debye–Scherrer formula (Eq. 2) was employed to calculate the crystallite size (D) of cobalt oxide nanograins. Additionally, to determine the number of crystallites per unit area (N), dislocation density (δ), and lattice strain (ε) of the oxide samples, the following expressions (Eqs. 3, 4, and 5) were used, respectively43. 2\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$D=\frac{{{\text{k\boldsymbol{\uplambda}}}}}{{{\mathrm{\boldsymbol{\upbeta}}}\cos {\mathrm{\boldsymbol{\uptheta}}}}}$$\end{document} 3\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$N=\frac{{\mathrm{t}}}{{{{\mathrm{D}}^3}}}$$\end{document} 4\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\delta =\frac{1}{{{{\mathrm{D}}^2}}}$$\end{document} 5\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\varepsilon =\frac{{\mathrm{\boldsymbol{\upbeta}}}}{{4\tan {\mathrm{\boldsymbol{\uptheta}}}}}$$\end{document} Here, k is an empirically determined constant with a value of 0.94, λ represents the wavelength of the X-ray source (Cu Kα = 1.5406 Å), β denotes the full width at half maximum (FWHM) of the diffraction peak, θ corresponds to the angular position of the diffraction peak, and t represents the oxide film thickness. The values of FWHM (β), number of crystallites per unit area (N), dislocation density (δ) and lattice strain (ε) all decreased, whereas crystallite size (D) is increased with an increase in oxidation temperature (Fig. 1d). It was calculated that the average crystallite size increased from 29.4 nm to 40.5 nm as the oxidation temperature increased from 400 °C to 600 °C. Additionally, the average lattice micro-strain decreased from 0.0046% to 0.0033% over the same temperature range. The higher dislocation density and lattice strain observed for the 400 °C sample indicate a greater density of defects and grain boundaries, which can serve as active sites for oxygen adsorption, therefore enhancing the gas-sensing performance. Additionally, the higher crystallite number density contributes to an increased surface area, facilitating improved interaction with target gas molecules. For better understanding, all the parameters, along with their calculated values for different oxidation temperatures, are systematically tabulated in Table 2. Table 2The average crystallite size (D), number of crystallites per unit area (N), dislocation density (δ) and lattice strain (ε) of Cobalt oxide samples oxidized at different temperatures (400–600 °C).TOxidation (°C)t (nm)(hkl)2θ (°)β (°)D (nm)N x 10− 3(nm− 2)δ x 10− 3(nm− 2)ε x 10− 3R.T.(Co)225110(Co)75.94 (Co)0.4141 (Co)25.42 (Co)13.703(Co)1.5479(Co)2.3152 (Co)400 °C (CoO + Co3O4)380 (CoO)200 (CoO)42.38 (CoO)0.2937 (CoO)30.30 (CoO)13.660(CoO)1.0892(CoO)3.3056 (CoO)380 (Co3O4)220 (Co3O4)31.27 (Co3O4)0.4288 (Co3O4)20.09 (Co3O4)46.839 (Co3O4)2.4767 (Co3O4)6.6854 (Co3O4)500 °C (Co3O4)40022031.270.288129.9114.9541.11804.4917600 °C (Co3O4)44022031.270.164552.383.0620.36452.5647 Table 2The average crystallite size (D), number of crystallites per unit area (N), dislocation density (δ) and lattice strain (ε) of Cobalt oxide samples oxidized at different temperatures (400–600 °C).TOxidation (°C)t (nm)(hkl)2θ (°)β (°)D (nm)N x 10− 3(nm− 2)δ x 10− 3(nm− 2)ε x 10− 3R.T.(Co)225110(Co)75.94 (Co)0.4141 (Co)25.42 (Co)13.703(Co)1.5479(Co)2.3152 (Co)400 °C (CoO + Co3O4)380 (CoO)200 (CoO)42.38 (CoO)0.2937 (CoO)30.30 (CoO)13.660(CoO)1.0892(CoO)3.3056 (CoO)380 (Co3O4)220 (Co3O4)31.27 (Co3O4)0.4288 (Co3O4)20.09 (Co3O4)46.839 (Co3O4)2.4767 (Co3O4)6.6854 (Co3O4)500 °C (Co3O4)40022031.270.288129.9114.9541.11804.4917600 °C (Co3O4)44022031.270.164552.383.0620.36452.5647 The average crystallite size (D), number of crystallites per unit area (N), dislocation density (δ) and lattice strain (ε) of Cobalt oxide samples oxidized at different temperatures (400–600 °C). 110 (Co) 13.703 (Co) 1.5479 (Co) 400 °C (CoO + Co3O4) 13.660 (CoO) 1.0892 (CoO) 46.839 (Co3O4) 2.4767 (Co3O4) Fig. 2Surface morphology of cobalt oxide thin films oxidized at: (a) RT, (b) 350 °C, [(c) and (d)] 400 °C, (e) 500 °C, (f) 800 °C, for 5 h. Insets show the closer views. Surface morphology of cobalt oxide thin films oxidized at: (a) RT, (b) 350 °C, [(c) and (d)] 400 °C, (e) 500 °C, (f) 800 °C, for 5 h. Insets show the closer views. Morphological analysis using FESEM Figure 2 presents the FESEM images of cobalt oxide thin films grown at different oxidation temperatures. The as-deposited Co film exhibits a uniformly grown and smooth surface morphology (Fig. 2a). These metal films remain thermally stable in their metallic state up to 200 °C. Although oxidation initiates above 200 °C, the process is extremely slow, with a significant degree of oxidation observed only beyond 300 °C. Upon oxidation at 350 °C, the smooth metallic surface transforms into granular oxide nanostructures of grain boundary network morphology (Fig. 2(b)). Subsequently, vertically grown nanowalls start to emerge on these initial grain boundary networks. The FESEM images of the 2D nanowalls formed at 400 °C are shown in Fig. 2(c). A closer view of highly porous nanowalls morphology is shown in Fig. 2(d). Inset confirms that nanowalls are reasonably thin and appear with a thickness of about 67 nm. However, oxidation above 400 °C drastically alters the surface morphology of cobalt oxide films. The vertically grown 2D nanowalls completely disappear, giving way to a homogeneously distributed granular nanoparticle morphology across the surface (Fig. 2(e)-(f)). Further increases in oxidation temperature do not result in significant morphological changes, except for a noticeable increase in the size of the nanoparticles. Fig. 3Schematic illustration of CoOx (CoO + Co3O4) nanowalls formation at 400 °C. (a) Co film with grain boundaries, (b) grain boundary induced initial CoO nucleation, (c) surface energy driven vertical growth of CoO nanowalls, (d) Co3O4 phase formation on outer surface of CoO nanowalls, (e) side-view (top) and cross-sectional view (bottom) of CoOx nanowalls. Schematic illustration of CoOx (CoO + Co3O4) nanowalls formation at 400 °C. (a) Co film with grain boundaries, (b) grain boundary induced initial CoO nucleation, (c) surface energy driven vertical growth of CoO nanowalls, (d) Co3O4 phase formation on outer surface of CoO nanowalls, (e) side-view (top) and cross-sectional view (bottom) of CoOx nanowalls. A schematic representation of the CoOx nanowall formation mechanism is illustrated in Fig. 3. The formation of the initial oxide layer with a grain boundary network (Fig. 3(a)) can be attributed to defect-induced nucleation of cobalt oxide. The grain boundaries of the Co film can serve as nucleation centers for the initial CoO phase at lower oxidation temperatures (Fig. 3(b)). The vertical growth mode of oxide nanowalls is most likely supported by the presence of the CoO phase within the oxide film (Fig. 3(c)). It is assumed that the sidewalls of these vertically oriented nanostructures are predominantly composed of the (100) plane, which has a significantly lower surface energy compared to the (110) and (111) planes44. This is further supported by the XRD analysis, where the (200) diffraction peak of CoO dominates over other CoO peaks. With increasing oxidation temperature, thermal diffusion and the chemical reactivity of oxide species are enhanced, leading to the formation of the Co3O4 phase (Fig. 3(d)). Although the Co3O4 phase is energetically favourable at this thermodynamic equilibrium, it is kinetically restricted due to a slower reaction rate at this oxidation temperature. During the thermal oxidation process, an increased concentration of O2 at the film surface effectively initiates the formation of Co3O4 at the outer surface of the nanowalls. However, due to limited thermal diffusion at 400 °C, the Co3O4 phase is primarily restricted to the outer surface, while the CoO phase is retained in the inner core layer of the nanowalls. As a result, a nanowall morphology with a mixed oxide phase of CoO and Co3O4 is formed (Fig. 3(e)). These findings are highly consistent with our XRD analysis, where a mixed oxide phase is clearly observed for the nanowalls. As the oxidation temperature exceeds 400 °C, the vertically grown nanowall morphology turns into a homogeneous distribution of nanograins composed of Co3O4 (Fig. 2(e)). This finding further supports our earlier assumption of CoO phase assisted vertical growth mechanism. XRD analysis also confirms the complete disappearance of this phase. The homogeneous growth of Co3O4 nanograins is due to faster kinetics of oxidizing species and rapid oxidation facilitated by increased thermal diffusion of Co and O atoms. At this stage, O₂ species which overcome the diffusion barrier, leading to the complete transformation of the CoO phase into the Co3O4 phase. With a further increase in oxidation temperature beyond 500 °C, the grain size of Co3O4 nanoparticles increases (Fig. 2(f)). The enhanced thermal diffusion of oxide species at higher temperatures promotes the coalescence of smaller grains into larger ones. This phenomenon leads to a gradual decrease in the number of crystallites per unit area. Additionally, the dislocation density and lattice strain both decrease at higher oxidation temperatures. This trend indicates a significant improvement in the crystalline quality of the oxide film (Fig. 1(d)). Fig. 4CoOx (CoO + Co3O4) nanowall imaging using EDS elemental mapping of (a) Both Co Kα and O Kα, (b) Co Kα and (c) O Kα signals. (d) FESEM image and (e) corresponding elemental line profile. CoOx (CoO + Co3O4) nanowall imaging using EDS elemental mapping of (a) Both Co Kα and O Kα, (b) Co Kα and (c) O Kα signals. (d) FESEM image and (e) corresponding elemental line profile. Further analysis of a mixed cobalt oxide (CoOx) nanowall through FESEM reveals a complex, multi-phase oxide structure as shown in Fig. 4. FESEM imaging was performed with a tilted view (45°) to highlight the right sidewall of the nanowall surface (marked with white dotted line). Energy-dispersive X-ray spectroscopy (EDX) elemental mapping of a tilted nanowall, shown in Fig. 4(a), confirms the overall distribution of cobalt (Co) and oxygen (O) throughout the structure. However, extended acquisition EDX mapping performed on a single nanowall surface revealed a weak modulation in elemental distribution, indicating a non-uniform composition. A darker contrast for Co (Fig. 4(b)) and a brighter contrast for O (Fig. 4(c)) was observed along the right sidewall of the nanowall (area within the dotted line), suggesting an oxygen-rich outer surface of the nanowall. For further confirmation, a line profile scan across the nanowall is also performed as shown in Fig. 4(d-e). For better comparison, the Co-Kα intensity was normalized with respect to the O-Kα line. The higher cobalt content at the top surface and its gradual reduction toward the sidewall strongly supports the hypothesis of the formation of the CoO (core) / Co3O4 (shell) nanostructure. These findings collectively suggest that the outer surface of the nanowall is oxygen-rich, while the core has a higher concentration of cobalt. This elemental segregation is characteristic of a core-shell structure. These findings are very much in line with our proposed structural model (Fig. 3) that the core is composed of a cobalt-rich CoO phase, while the outer shell consists of a more oxygen-rich Co3O4 phase. This unique structural and chemical heterogeneity is a direct consequence of the formation of mixed phase nanowall structures, where the vertical growth is primarily stabilized by the CoO core layer. Although EDX is not an efficient technique for any quantitative analysis of chemical compositions, XPS and HRTEM analysis of these CoOx nanowalls have also been performed for further understanding of these core-shell growth mechanisms. Structural characterization using TEM High-resolution transmission electron microscopy (HRTEM) and selected area electron diffraction (SAED) analyses of mixed-phase CoOx sample (oxidized at 400 °C for 5 h) were performed on to gain deeper insight into the structural characteristics and phase distribution of the mixed oxide nanowalls at high spatial resolution. As shown in Fig. 5, the HRTEM images and SAED patterns confirm the coexistence of both Co3O4 and CoO phases within the oxide nanowall45,46. The low-magnification TEM image (Fig. 5(a)) reveals the nanowall morphology, consistent with the earlier SEM observations. A representative HRTEM image of the mixed oxide region is presented in Fig. 5(b), where lattice fringes corresponding to the Co3O4 spinel structure (highlighted in red) and the CoO rock-salt structure (highlighted in yellow) can be clearly identified. The heterophase interface separating Co3O4 and CoO is delineated by the white dotted line. The measured interplanar spacings of 0.47 nm and 0.25 nm are assigned to the (111) planes of Co3O4 and CoO, respectively. The insets of Fig. 5(b) show the corresponding fast Fourier transform (FFT) patterns, further confirming the spinel Co3O4 and rock-salt CoO phases. Fig. 5(a) Low-resolution TEM micrograph of the mixed-phase CoOₓ (CoO + Co3O4) nanowall film oxidized at 400 °C for 5 h. (b) HRTEM image showing lattice fringes corresponding to the Co3O4 (111) (red) and CoO (111) (yellow) phases. Insets show the FFT patterns of the corresponding selected areas (c) SAED pattern confirming coexistence of both oxide phases. (a) Low-resolution TEM micrograph of the mixed-phase CoOₓ (CoO + Co3O4) nanowall film oxidized at 400 °C for 5 h. (b) HRTEM image showing lattice fringes corresponding to the Co3O4 (111) (red) and CoO (111) (yellow) phases. Insets show the FFT patterns of the corresponding selected areas (c) SAED pattern confirming coexistence of both oxide phases. The SAED pattern shown in Fig. 5(c) also supports the coexistence of these mixed oxide phases. For clarity, diffraction planes associated with Co3O4 and CoO are highlighted by red and yellow semi-circles, respectively. The innermost diffraction ring corresponds to the superposition of the closely placed (311) plane of Co3O4 and the (111) plane of CoO. Similarly, the second diffraction ring is attributed to the closely placed the Co3O4 (400) and CoO (200) diffraction plane. The third ring corresponds to the (220) plane of CoO. Additional diffraction rings corresponding to the (531) and (533) planes of Co3O4 and the (222) plane of CoO are also clearly visible in the outer region of Fig. 5(c). These results collectively confirm the presence of Co3O4/CoO mixed oxide phases within the nanowall structure and appear with well-defined crystallographic features. HRTEM and SAED analyses confirm the CoO (core) / Co3O4 (shell) configuration within the mixed oxide nanowalls. The images reveal lattice fringes of Co3O4 (111) and CoO (111), separated by a distinct heterophase interface. Since the Co3O4 (311) and CoO (111) planes have nearly identical interplanar spacings (~ 0.25 nm), some ambiguity may arise in HRTEM interpretation. However, the distinct FFT patterns (insets of Fig. 5(b)) from the respective lattice regions match well with reported data47,48, clearly distinguishing the Co3O4 and CoO phases and confirming their coexistence. Moreover, the absence of the rotational offset between the Co3O4 (111) and (311) planes (~ 29.5°) further support assigning the 0.25 nm spacing to the CoO (111) plane. Electrical characterization Figure 6(a) presents the I-V characteristics of cobalt films oxidized at temperatures ranging from 400 °C to 600 °C for 5 h. The inset provides a comparative study of sheet resistance as a function of oxidation temperature. With increasing oxidation temperature, the surface electrical conductivity of the oxide films increases, and a slight nonlinear I-V response is observed. However, all oxide films exhibit a very low conduction current (µA), which is dominated by the reverse saturation current of a back-to-back connected Schottky diodes formed between the Au electrodes and the oxide film. Consequently, the conduction current can be considered as restricted to flow in blocking/leakage mode. The highest sheet resistance is observed for CoOₓ films oxidized at 400 °C. At this temperature, the oxide layer consists of a mixed phase of CoO and Co3O4. The coexistence of these phases results in a high density of grain boundaries and defect states, which significantly hinder charge carrier mobility and reduce surface electrical conductivity. A notable increase in surface electrical conductivity occurs when the oxidation temperature is raised to 500 °C. At this stage, the mixed-phase oxide film transforms into a single-phase Co3O4 structure, significantly reducing the density of grain boundaries. The reduction in grain boundaries minimizes charge carrier scattering, thereby enhancing electrical conductivity. In addition, Co3O4 acts as a band semiconductor, whereas CoO behaves as a charge transfer insulator49. Further increasing the oxidation temperature to 600 °C results in an additional decrease in sheet resistance. This trend is primarily attributed to the formation of larger Co3O4 grains due to the coalescence of smaller grains, reducing the total number of grain boundaries50,51. Moreover, higher oxidation temperatures facilitate active oxygen diffusion, which enhances the concentration of cobalt vacancies, thereby increasing hole carrier density within the p-type oxide layer. These findings are well aligned with the XRD and following XPS analyses, further confirming that higher oxidation temperatures improve the electrical conductivity of the cobalt oxide thin films. Optical characterization The optical properties of cobalt oxide nanowalls (400 °C) were investigated using UV-Vis spectroscopy, as shown in Fig. 6(b). The inset presents a comparative analysis of the Tauc plot for CoOₓ films grown at oxidation temperatures ranging from 400 °C to 600 °C. The UV-Vis spectrophotometer used in this study is equipped with two light sources: (i) a tungsten lamp for the visible region and (ii) a deuterium lamp for the ultraviolet (UV) region. This setup ensures comprehensive spectral coverage across the UV-Vis range. The absorbance spectra of the oxide films were recorded, and the optical energy bandgap (Eg) was determined using the Tauc plot method, which is defined by Eq. (6): 6\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$~ah\upsilon \,=\,A{(h\upsilon \, - \,{E_g})^n}$$\end{document} where, α represents the absorption coefficient, Eg is the optical energy bandgap, hν is the photon energy, and n is a constant. The value of n was taken as ½ for direct band transitions and 2 for indirect band transitions. Our experimental data fit well with n = ½, confirming a direct optical bandgap for cobalt oxide films. The inset of Fig. 6(b) presents the Tauc plot of (αhν)² versus hν, revealing two distinct absorption bands associated with ligand-metal charge transfer processes52. Two direct optical bandgaps were observed for CoOX nanowalls: (i) near-infrared (NIR) region (E1∼1.55 eV), attributed to electronic transitions from O2– to Co3+ states and (ii) visible spectrum region (E2∼1.9 eV), related to charge transfer between O2– and Co3+ states42. The primary energy bandgap (1.9 eV) of Co3O4 nanostructures results from the interaction between the 3d orbitals of Co²⁺ (conduction band) and the 2p orbitals of O2– (valence band), whereas the electronic transition from the valence band to the Co3+ atoms associated sub-band accounts for the lower energy gap (1.55 eV). These findings align well with previously reported values53–57. Furthermore, a decrease in optical bandgap with increasing oxidation temperature is also observed (inset of Fig. 6(b)). This trend can be attributed to the increase in average particle size due to island coalescence at higher oxidation temperatures58. Fig. 6Comparative study of cobalt oxide thin films oxidized at temperatures between 400–600 °C for 5 h: (a) I-V analysis, (b) UV-Vis spectroscopy, and (c)-(d) Raman spectroscopy. Comparative study of cobalt oxide thin films oxidized at temperatures between 400–600 °C for 5 h: (a) I-V analysis, (b) UV-Vis spectroscopy, and (c)-(d) Raman spectroscopy. Raman spectroscopy Figure 6(c) presents the Raman spectrum of cobalt oxide nanowalls (400 °C), revealing five Raman-active modes at ~ 192 cm–1, 473 cm–1, 515 cm–1, 612 cm–1, and 679 cm–1 corresponding to F2g, Eg, F2g, F2g, and A1g modes, respectively, confirming the presence of the Co3O4 phase. CoO crystals possess a rock salt structure with a centrosymmetric lattice, where first-order phonon Raman scattering is forbidden by selection rules59. The fundamental F2g vibration mode is characterized by a two-fold degenerate irreducible representation. For CoOx nanowalls, the A1g mode (~ 679 cm–1) dominates over the other peaks. Figure 6(d) displays a comparative analysis of the higher resolution A1g spectra for cobalt oxide films oxidized between 400 and 600 °C. With an increase in oxidation temperature, a shift in the A1g peak position towards a higher wave number, along with an increase in its intensity, has been observed. This shift can be attributed to the presence of tensile strain within the crystal structure at lower oxidation temperatures (400 °C), which increases the bond length, thereby reducing the Raman wave number. In addition, the asymmetric growth of nanowall morphology can also influence the Raman shift. Finally, enhancement of Raman peak intensity at higher oxidation temperature indicates better crystalline quality of the film. The Raman data align well with the XRD findings, further validating the structural transformations with temperature variations55. Fig. 7XPS binding energy spectra of cobalt oxide nanowall grown at 400°: (a) survey scan, (b) high-resolution Co2p and (c) high-resolution O1s. Oxidation temperature dependent comparative study of: (d) Co²⁺ and Co3⁺ states and (e) O1 and O2 peaks. XPS binding energy spectra of cobalt oxide nanowall grown at 400°: (a) survey scan, (b) high-resolution Co2p and (c) high-resolution O1s. Oxidation temperature dependent comparative study of: (d) Co²⁺ and Co3⁺ states and (e) O1 and O2 peaks. X-ray photoelectron spectroscopy (XPS) Figure 7(a) presents the XPS survey spectrum of cobalt oxide nanowalls grown at 400 °C. The spectrum appears with core level binding energy (BE) and Auger peaks corresponding to cobalt, oxygen, and carbon (surface contamination). High-resolution BE spectra of Co2p and O1s are depicted in Fig. 7(b) and (c), respectively. The core-level Co2p spectrum exhibits a spin-orbit splitting of ~ 15.6 eV, with Co2p₃/₂ at 779.8 eV and Co2p₁/₂ at 795.4 eV, along with shake-up satellite peaks at 785.6 eV and 805 eV, characteristic of the Co3O4 phase60. The core level C1s BE peak at 285.0 eV is used as a reference for all XPS analysis. The Co2p peaks were further deconvoluted into Co2+ (779.4 eV and 795.1 eV) and Co3+ (780.7 eV and 796.5 eV) oxidation states for a quantitative assessment, as shown in Fig. 7(b). Similarly, the O1s BE peak was deconvoluted into two sub-peaks: O1 (529.3 eV) and O2 (531.4 eV). The O1 peak corresponds to oxygen atoms bonded within the Co3O4 crystal, whereas the O2 peak is associated with physisorbed oxygen or defect-induced states. Detailed analysis of XPS BE spectra of cobalt oxide films grown at oxidation temperatures of 400 °C to 600 °C are already reported in our previous work42. A comparative analysis of Co oxidation states is shown in Fig. 7(d). With increasing oxidation temperature, the intensity of Co3+ states rise, whereas Co2+ states diminish. Concurrently, the relative intensity of O1 states increases compared to O2 states, as shown in Fig. 7(e). These observations indicate the transition from the CoO phase to Co3O4 above 400 °C, along with oxygen enrichment and a reduction in crystal defect states within the oxide films60. Moreover, these results are very much complementary to our earlier XRD, I-V and Raman analysis. Fig. 8High resolution XPS binding energy spectra of deconvoluted Co 2p3/2 peaks of cobalt oxide nanowall grown at 400 °C. High resolution XPS binding energy spectra of deconvoluted Co 2p3/2 peaks of cobalt oxide nanowall grown at 400 °C. To validate the proposed core-shell structure of mixed CoOx nanowalls, XPS depth profile analysis on the nanowalls film surface was also performed. An Ar+ beam of 1000 eV energy was used to etch the CoOx surface and the etching rate was estimated to be 0.27 nm/s. High-resolution binding energy spectra of the Co 2p3/2​ core level peaks were recorded at various etching depths as shown in Fig. 8. The analysis revealed a significant change in the oxidation states of cobalt as a function of depth. All Co 2p3/2​ BE spectra are deconvoluted into three peaks such as Co+ 2 (779.4 eV), Co+ 3 (780.7 eV) and shake-up satellite peaks (785.6 eV). The signal corresponding to the Co3+ oxidation state, which is dominant in the outer Co3​O4​ shell, progressively decreased. Relative comparison of these integral peak intensities is summarized in Table 3. Concurrently, the relative intensity of the satellite peak characteristic of the Co2+ oxidation state, a hallmark of the CoO phase, was observed to increase with successive etching cycles (Figure SI 1). In addition, a comparative study of Co2p satellite peaks centered at 785.6 eV (represent Co+ 2 oxidation state) is also presented in Table SI 1. This systematic change in the Co2+/Co3+ ratio provides direct chemical evidence that the outer layer of the nanowall is rich in the Co3​O4​ phase, while the inner core is predominantly composed of the CoO phase. This observation strongly supports the hypothesis of a unique core-shell structure, with a CoO core surrounded by a Co3​O4​ shell, formed during the growth of the mixed-oxide nanowalls. Therefore, this depth-profile analysis provides direct chemical evidence supporting the formation of a CoO (core) / Co3O4 (shell) nanostructure. Table 3Comparative study Co+ 2 and Co+ 3 oxidation States of CoOx nanowalls film with depth profile.Depth (nm)Atomic Co3+ (%)Atomic Co2+ (%)0 (Top surface)45.3154.6913.543.3556.652738.6961.3140.535.7164.295434.4165.59 Table 3Comparative study Co+ 2 and Co+ 3 oxidation States of CoOx nanowalls film with depth profile.Depth (nm)Atomic Co3+ (%)Atomic Co2+ (%)0 (Top surface)45.3154.6913.543.3556.652738.6961.3140.535.7164.295434.4165.59 Comparative study Co+ 2 and Co+ 3 oxidation States of CoOx nanowalls film with depth profile. VOC sensing studies Figure 9(a) illustrates the temperature-dependent sensitivity measurement for 10 ppm of formaldehyde vapour of CoOx samples grown at 400 °C, 500 °C, and 600 °C. All three samples demonstrated excellent sensitivity, however, the cobalt oxide nanowalls grown at 400 °C exhibited the highest response. The optimal operating temperature for the nanowall-based sensor was estimated to be 150 °C, where it achieved a sensitivity of 60%. Due to its high sensitivity, the CoOx nanowall sensor was further investigated for transient response studies at 150 °C. Figure 9(b) presents its response to low concentrations of formaldehyde vapour, ranging from 1 ppm to 50 ppm. The sensor was able to detect formaldehyde vapour concentrations well below 1 ppm and reached saturation above 50 ppm, confirming its high sensitivity at the lower concentration range (inset of Fig. 9(b)). A comparative study of formaldehyde sensing using different TMO based gas sensors is summarized in Table 4. Although the response values reported in the literature are calculated using different methods and are not directly comparable, our CoOx nanowalls sensor demonstrates clear advantages of operating at a much lower temperature (150 °C) with reasonably faster response/recovery times and higher sensitivity to low-concentration formaldehyde vapour. These attributes highlight its practical applicability for reliable VOC detection. Fig. 9(a) Operating temperature dependent sensitivities of CoOx films grown at 400 °C, 500 °C, and 600 °C. b) Transient response of nanowall sensor (400 °C) for low concentrations formaldehyde (c) Schematic illustration of CoOx film-based gas sensor with gold finger electrodes, interacting with gas molecules, (d) Reproducibility of nanowall (400 °C) senor towards 10 ppm formaldehyde vapour. (e) Typical response curve of formaldehyde vapour and selective sensing of formaldehyde vapour against other VOCs (10 ppm each). (f) Schematic representation of charge transport and band bending across nanowalls junction. (a) Operating temperature dependent sensitivities of CoOx films grown at 400 °C, 500 °C, and 600 °C. b) Transient response of nanowall sensor (400 °C) for low concentrations formaldehyde (c) Schematic illustration of CoOx film-based gas sensor with gold finger electrodes, interacting with gas molecules, (d) Reproducibility of nanowall (400 °C) senor towards 10 ppm formaldehyde vapour. (e) Typical response curve of formaldehyde vapour and selective sensing of formaldehyde vapour against other VOCs (10 ppm each). (f) Schematic representation of charge transport and band bending across nanowalls junction. A schematic representation of the sensor device and its interaction with the target gas molecules, along with the CoOx nanowalls surface morphology are depicted in Fig. 9(c). The stability of this nanowall-based sensor was also tested over multiple sensing cycles, as shown in Fig. 9(d). The sensor exhibited excellent reproducibility with high thermal and physical stability, attributed to its robust structure and strong adhesion to the substrate. Its consistent performance over repeated cycles highlights its potential for long-term applications without significant degradation in accuracy or reliability. Figure 9(e) depicts the typical response curve of formaldehyde vapour and selectivity study of the CoOx nanowall sensor for formaldehyde vapour against other VOCs such as methanol, acetone, benzene, toluene, and xylene, all tested at a concentration of 10 ppm. The sensor displayed clear selectivity for formaldehyde with a selectivity of 60%, a response time of 100 s, and a recovery time of 332 s. High selectivity of nanowall sensor is crucial for distinguishing low concentrations of formaldehyde in mixed environments, addressing one of the most challenging aspects of gas sensor. Table 4Comparative study of formaldehyde sensing using TMO based gas sensors.SensingmaterialMorphologySynthesismethodWorking temp (°C)Conc. (ppm)ResponseResp/Rec. time (s)Ref.Ni/SnO2NanoparticlesHydrothermal10025001298a360/600 61 Fe2O3-In2O3NanotubesElectrospinning25010033a5/25 62 SnO2NanoflowersHydrothermal30010034.6a64/10 63 CeO2NanosheetsHydrothermal30020030a10/20 64 ZnONanoplatesHydrothermal25010016.8a80/60 65 ZnOMicroflowersHydrothermal2751009.6a16/28 66 NiO/ZnOMicroflowersHydrothermal20010026.2a18/30 66 CuOMicrospheresHydrothermal308032a301/705 67 SnO2NanosheetsHydrothermal200200207a30/57 68 NiONanograinsSputtering340204.3a420/1800 69 Co3O4NanowallsPVD1501060b100/332This work Table 4Comparative study of formaldehyde sensing using TMO based gas sensors.SensingmaterialMorphologySynthesismethodWorking temp (°C)Conc. (ppm)ResponseResp/Rec. time (s)Ref.Ni/SnO2NanoparticlesHydrothermal10025001298a360/600 61 Fe2O3-In2O3NanotubesElectrospinning25010033a5/25 62 SnO2NanoflowersHydrothermal30010034.6a64/10 63 CeO2NanosheetsHydrothermal30020030a10/20 64 ZnONanoplatesHydrothermal25010016.8a80/60 65 ZnOMicroflowersHydrothermal2751009.6a16/28 66 NiO/ZnOMicroflowersHydrothermal20010026.2a18/30 66 CuOMicrospheresHydrothermal308032a301/705 67 SnO2NanosheetsHydrothermal200200207a30/57 68 NiONanograinsSputtering340204.3a420/1800 69 Co3O4NanowallsPVD1501060b100/332This work Comparative study of formaldehyde sensing using TMO based gas sensors. The efficiency of a chemiresistive gas sensor is primarily influenced by the morphology and stoichiometry of the oxide film surface, which directly affects the adsorption-desorption process of gas molecules and surface charge transport mechanisms, ultimately impacting sensitivity, selectivity, and operating temperature. Vertically grown CoOx nanowalls (oxidized at 400 °C) exhibit a highly porous morphology with a significantly larger surface-to-volume ratio compared to the nanograins morphology of Co3O4 nanograin films oxidized at 500 °C and 600 °C. These effects facilitate an enhanced gas adsorption and diffusion of the nanowalls film. Additionally, the mixed-phase composition (CoO and Co3O4) of these nanowalls introduces numerous grain boundaries and defect states, which act as active sites for electron/hole trapping, modulating charge transport and improving reaction kinetics for formaldehyde vapor sensing. Furthermore, specific crystal facets in nanowall films may also enhance formaldehyde adsorption and reactivity, leading to superior sensor performance. These structural advantages collectively result in higher sensitivity and improved selectivity for formaldehyde detection at lower operating temperature, making CoOx nanowalls superior over Co3O4 nanograin. Apart from the morphological influences, stoichiometry of the cobalt oxide materials can significantly impact on the electronic structure, oxygen vacancy concentration, and defect states. The mixed oxide phase of CoOx nanowalls is expected to have more defects due to limited thermal diffusion of the oxidizing species at 400 °C. These parameters significantly enhance the gas adsorption process and charge transfer mechanism and hence the formaldehyde sensing efficiency. However, Co3O4 nanograins grown at higher temperature have a better crystal stoichiometry with a reduced defect density which effectively restricts the sensing efficiency. The charge transfer mechanism through a cobalt oxide nanowall junction in air and formaldehyde vapour is depicted in Fig. 9(f). Due to the p-type conductivity of cobalt oxide, conduction current occurred mainly by the hole transport. Here, the conduction mechanism is considered separately in air and in formaldehyde exposure. In air, adsorbed oxygen molecules on cobalt oxide surface behave like an electron acceptor (O−) and create a hole accumulation layer at the outer surface of the nanowalls. At this temperature the adsorbed oxygen molecules on the oxide surface are mainly O− form, as shown in Eq. (7). 7\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{\mathrm{O}}_{\mathrm{2}}}^{ - }\left( {{\mathrm{ads}}} \right)\,+\,{{\mathrm{e}}^ - } \to {\text{ 2 }}{{\mathrm{O}}^ - }\left( {{\mathrm{ads}}} \right){\text{ }}\left( {{\mathrm{1}}00^\circ {\mathrm{C}}\,<\,{\mathrm{T}}\,<\,{\mathrm{3}}00^\circ {\mathrm{C}}} \right)$$\end{document} The corresponding energy band diagram shows the band bending and the created built-in-potential (qVb−air) near the junction. The removal of adsorbed surface oxygen due to the reactions with formaldehyde vapour results in released back of the trapped electrons to the CoOx surface. As a result, lowering in band bending and built-in-potential (qVb−VOC) happened, as shown in Fig. 9(f). Therefore, the conduction current (I) through the CoOx nanowall junction can be written as shown in Eq. (8). 8\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${I_{air/VOC}}={I_0}\exp \left[ {\frac{{q{V_{b\left( {{\mathrm{air}}} \right)/\left( {{\mathrm{VOC}}} \right)}}}}{{kT}}} \right]$$\end{document} where, I0 is the current in flat band condition (i.e. no built-in-potential), Iair and IVOC are the current in air and in VOC ambience, kB is the Boltzmann’s constant, and T is the temperature70. In addition, the coexistence of CoO (Co+ 2) and Co3O4 (Co+ 3 and Co+ 2) phases form internal heterojunctions that modulate charge transport and create additional active sites for gas adsorption. Furthermore, the Co+ 2/Co+ 3 redox couple facilitates the catalytic oxidation of formaldehyde by promoting electron exchange and accelerating surface reactions. We have also discussed the influence of oxygen vacancies, which are more abundant at lower oxidation temperatures (400 °C) in the mixed phase nanowalls and diminish at higher temperatures, correlating directly with the higher sensitivity of the 400 °C samples. These interfacial and redox effects, combined with the nanowall morphology, explain the superior sensing behavior beyond the simplified band bending model. Apart from the sensitivity and selectivity, humidity tolerance and baseline stability over long-time operation are also very important features for any gas sensors, as they can directly impact on the reliability and commercial viability of the sensors (Fig. 10). Humidity tolerance is very critical for the deployment of gas sensors in real-world conditions where the concentration of water vapor can vary widely and interfere with the sensor’s response. Similarly, a stable baseline ensures accurate and repeatable measurements, preventing false alarms or missed detections. The gas sensing properties of the Co​Ox​ nanowall films were evaluated, with a particular focus on their performance in humid environments. To generate humid air, the dry air stream was passed through a bubbler containing deionized water, maintained at a controlled temperature. The humidified air was then directed into the testing chamber, and the relative humidity (RH) was monitored in real time using a calibrated hygrometer, placed directly inside the sensing chamber. By increasing the temperature of the water in the bubbler, higher RH levels were achieved. This procedure allowed systematic evaluation of the sensor performance under different humidity conditions (Fig. 10(a)). The as-prepared films exhibited stable chemiresistive behavior with a high degree of humidity tolerance. Specifically, the response showed a small variation of approximately 5% when the relative humidity increased from dry air to 80%. This remarkable humidity tolerance is a significant advantage for practical sensor applications, where ambient humidity can severely affect sensor accuracy and stability. The observed stability in a humid atmosphere can be attributed to the unique core-shell nanostructure of the nanowalls. It is assumed that the outer Co3​O4​ shell, possibly due to its nanowalls morphology along with specific surface chemistry exhibits a hydrophobic nature71. This effectively mitigates the interference from water molecules, preserving the primary gas-sensing reactions at the active sites. Beyond humidity tolerance, the long-term thermal and physical stability of the films were also assessed, as shown in Fig. 10(b). The nanowall films exhibited excellent thermal stability, with the sensor’s baseline resistance remaining stable over repeated temperature cycles from room temperature up to its optimal operating temperature. Furthermore, FESEM analysis after prolonged operation and thermal cycling confirmed that the unique nanowall morphology was physically preserved, with no signs of agglomeration or structural degradation (Fig. SI 2). This high degree of long-term stability in both electrical response and physical structure confirms the robustness of the Co​Ox​ nanowalls. All these findings demonstrate the potential of these core-shell cobalt oxide nanowalls for developing reliable and robust gas sensors for real-world applications. Fig. 10(a) Sensitivity of the sensor to formaldehyde vapour as a function of relative humidity (0–80% RH) under both dry and humid conditions. (b) Long-term (45 days) baseline stability and thermal endurance of the CoOx nanowalls sensor over repeated sensing cycles. (a) Sensitivity of the sensor to formaldehyde vapour as a function of relative humidity (0–80% RH) under both dry and humid conditions. (b) Long-term (45 days) baseline stability and thermal endurance of the CoOx nanowalls sensor over repeated sensing cycles. Conclusions Vertically grown 2D CoOx nanowalls (67 nm thick) were successfully synthesized through a controlled thermal oxidation of thin Co films at 400 °C. The lower oxidation temperature (< 400 °C) promoted CoO phase formation, which eventually stabilizes the vertical growth morphology of cobalt oxide nanowalls due to the lower surface free energy of the (100) crystal plane of CoO. At higher oxidation temperatures, the Co3O4 phase became dominant, resulting in a homogeneous nanograin morphology. Comprehensive structural and surface characterizations confirmed that these nanowalls exhibit a core (CoO) - shell (Co3O4) mixed oxide phase with higher defect densities and grain boundaries, which play a crucial role in enhancing their gas-sensing performance. The CoOx nanowall-based gas sensor demonstrated very high formaldehyde vapor sensing capabilities, including high sensitivity (detecting formaldehyde vapor below 1 ppm), excellent reproducibility (stable performance over multiple sensing cycles), superior selectivity (effective discrimination against other VOCs), fast response and recovery (ensuring real-time detection), significantly high humidity tolerance and structural robustness (high durability due to the physical growth route). These findings highlight the potential of CoOx nanowalls as efficient and reliable sensing material for detecting low-concentration formaldehyde vapor and other hazardous VOCs in industrial and environmental applications. Moreover, stable chemiresistive behavior of Co​Ox​ nanowall films, highlighting their tolerance to humidity and long-term thermal and physical stability, is achieved which are crucial for the reliability and commercial viability of gas sensors. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1