Cu³⁺ Ion Evaluation and O^2- Vacancies Identification in CuO Nanofibers, by XPS Spectroscopy

1. Introduction

Copper oxides, Cu_xO (x = 1–4) [1], are among the most intensively studied advanced materials showing to their relevant optical, electrical, thermal, and magnetic properties. When engineered as nanostructures, these oxide semiconductors can exhibit size- and morphology-dependent behaviors that are not observed in their bulk counterparts [1]. In particular, cuprous oxide (Cu₂O, cuprite) and cupric oxide (CuO, tenorite) in nanofiber form are representative examples of such structure–property relationships [3].

As noted by Sarkar et al. [1] and Collins et al. [2], the presence of Cu³⁺ species and excess holes is correlated, and can contribute to the semiconducting character of CuO. In their study, heat treatments applied to commercial CuO powders, combined with X- Ray Photoelectron Spetroscopy, XPS, and EDP analyses, confirmed a microscopic amount of Cu³⁺ species and holes, which increased with temperature, thereby reinforcing the semiconducting response and increasing dielectric constant. Similar conclusions were reported by Raj et al. [3], who found that the enhanced conductivity of nanostructured CuO was associated with a higher Cu³⁺ concentration together with a small contribution from oxygen vacancies (V_O). XPS is widely used to characterize Cu oxidation states, and the Cu 2p photoemission region is particularly informative for identifying Cu³⁺-related components. However, Cu 2p spectra are also challenging to interpret because of their complex background, asymmetry, and pronounced multiplet structure, which yields a rich but intricate spectral envelope compared to many other transition-metal oxides [4,5,6,7,8,9]. The objective of this work was to evaluate the presence of Cu³⁺ species in onedimensional CuO nanostructures and to examine their relationship with oxygen-vacancyrelated features using XPS. The electrospun CuO nanofibers were subjected to two thermal treatments (600 °C and 700 °C, 4 h each) and compared with a bulk CuO reference. XPS analysis employs recently developed fitting approaches [10]. In addition, the chemical species were quantified and HRTEM observations revealed Cu³⁺-related superlattice features. Finally, the Rietveld refinement of the X- Ray Diffraction, XRD, data was used to confirm the crystalline phases and estimate the crystallite size. To meet this objective, we combined XPS with nanobeam diffraction (NBD) and Rietveld refinement methods, providing a comprehensive approach to understanding the interaction between Cu³⁺ species and oxygen vacancies in the nanostructures.

2. Experimental Procedure

Polymeric precursor fibers were prepared via electrospinning. First, an 8 wt% poly(vinyl alcohol) (PVA; Sigma-Aldrich, M_w ≈ 130,000, high purity) solution was prepared by dissolving 8 g of PVA in 92 mL of triple-distilled water (resistivity of 1.1 MΩ·cm; J. T. Baker) under magnetic stirring for 24 h at room temperature. A copper precursor solution was prepared by dissolving 1 g of copper(II) acetate (Sigma-Aldrich, 99%) in deionized water and stirring for 4 h at 50 °C. Next, 20 g of PVA solution was added to the copper precursor solution, and the resulting mixture was stirred for 24 h at 600 rpm until a homogeneous, transparent solution was obtained.

Ten milliliters of the final solution were loaded into a syringe mounted on a syringe pump as part of the electrospinning setup. A voltage of 8 kV was applied between the syringe needle and an aluminum foil-covered collector plate placed 20 cm from the needle tip.

Two calcination temperatures were selected based on simultaneous TGA/DSC measurements. The electrospun fibers were heat-treated in air at a heating rate of 10 °C·min⁻¹ to 600 °C and 700 °C, yielding the CuO600 and CuO700 samples, respectively. The purpose of these heat treatments was to convert the electrospun precursor fibers Page 5 of 20 into CuO semiconductor nanofibers and remove residual organic components (PVA and acetate-derived species)

3. Characterization

3.1. Structural Characterization by XRD

The crystalline structures of the synthesized CuO nanofibers were analyzed using a high-resolution X-ray diffractometer (PANalytical X’Pert PRO) equipped with an X’Celerator detector. Data were collected using Cu Kα radiation (λ = 1.5418 Å) over the 2θ range of 30–90°, with a step of 0.01° and a counting time of 0.01 s per step. A fine step size of 0.01° allowed the resolution of subtle Cu₂O reflections, enhancing the detection of minor phases and reinforcing the credibility of the results.

The monoclinic CuO structural parameters were refined by Rietveld analysis (leastsquares method) using FullProf Suite [11]. The diffraction peaks were modeled using a pseudo-Voigt profile function. The apparent crystallite size was estimated using the spherical harmonics approach and the Thompson–Cox–Hastings pseudo-Voigt function, including axial-divergence asymmetry and an instrumental resolution file [12].

3.2. Scanning Electron Microscopy (SEM)

Samples were collected by cutting approximately 1 cm² of the aluminum foil used as the collector during electrospinning. Scanning electron microscopy (SEM) micrographs were acquired using a Hitachi SU3500 microscope.

3.3. High-Resolution Transmission Electron Microscopy (HRTEM)

The specimens were prepared by dispersing the synthesized materials in isopropanol (Sigma-Aldrich, 99.8%) for 1 h. A droplet of the suspension was deposited onto 3 mm Ni grids and dried prior to analysis. HRTEM images were acquired using a JEOL JEM-2200FS microscope and processed using DigitalMicrograph [13].

3.4. X-Ray Photoelectron Spectroscopy (XPS)

XPS measurements were performed using a Thermo Scientific ESCALAB Xi instrument equipped with a monochromatic Al-Kα source (hν = 1486.7 eV). The acquisition parameters were as follows: energy step, 10 eV; energy resolution 0.1 eV, dwell time, 200 ms; 40 scans; takeoff angle, 90°. The CuO nanofibers were mounted on a conductive graphite tape attached to the sample holder. The samples were introduced into the load-lock chamber until a vacuum of approximately 10⁻⁶ Torr was reached, and measurements were carried out at a base pressure of ~10 ⁻⁶ Torr in the analysis chamber. The monochromator was positioned at 45° relative to the photoelectron-collection axis.

3.5. Data analysis

Peak fitting of the Cu 2p and O 1s regions was performed using the AAnalyzer^® software, employing a Shirley/SVSC-type background, a two-parameter Tougaard function, and slope correction. The chemical composition was determined using the geometrical topofactor approach (spherical model). Calculations were performed using the Cumpson core–shell model [14,15].

4. Results and Discussion

4.1. XRD Analysis

Structural characterization of the CuO600 and CuO700 samples was carried out by Rietveld refinement of the XRD patterns using FullProf software. The refined patterns are shown in Figure 1, where the shift in the peak positions (dashed line) indicates the changes in the crystalline structure. The calculated profiles (red line) were in good agreement with the experimental data, supporting the reliability of the refinement.

For the CuO600 sample, the refinement indicated a mixed-phase composition consisting of CuO with minor Cu₂O contributions (PDF 75–1531 and 80–1268) (Figure 1b). In contrast, the CuO700 sample exhibited a single CuO (tenorite) phase, consistent with PDF 80–1268.

The refinement results obtained using FullProf [11] are summarized in Table 1. The average crystallite sizes estimated for CuO600 and CuO700 were 31.06 ± 0.06 nm and 20.59 ± 0.06 nm, respectively.

4.2. SEM, HR-TEM, and NBD Analysis

The SEM micrograph in Figure 2a shows the as-electrospun polymeric precursor nanofibers (PVA containing copper(II) acetate), forming a randomly oriented fibrous network. Figure 2b presents the diameter distribution of the as-spun nanofibers. Prior to calcination, the precursor fibers consisted of a polymer/metal-salt composite; subsequent heat treatments at 600 °C and 700 °C (yielding CuO600 and CuO700, respectively) promoted the conversion of the precursor into CuO nanofibers through the removal of the organic matrix and decomposition of the acetate species.

The calcined products (CuO nanofibers) were examined by HRTEM and nano-beam electron diffraction (NBD) (Figure 3). The representative morphologies of CuO600 and CuO700 are shown in Figure 3a and Figure 3d, respectively. The corresponding NBD patterns (Figure 3b and Figure 4e) were indexed using CrysTBox [18], confirming the presence of the tenorite (CuO) phase in both samples, in agreement with the XRD results.

In Figure 3b and Figure 3e, the reflections obtained by NBD for CuO600 and CuO700 suggest the presence of Cu³⁺-related features (highlighted by the yellow circles). In line with previous reports, these features have been described as a superlattice-like contributions associated with additional ionic species. Such contributions are not resolved by XRD, likely because of their low concentration; they may instead reflect local charge density variations, defect-related signatures (e.g., vacancies), and/or minor phases present at trace levels [1,19].

4.3. XPS Analysis and Chemical Composition

The XPS results for CuO600 (Figure 6a,b) and CuO700 (Figure 6 c, d) indicate the presence of Cu³⁺-related contributions in the Cu 2p region (j = 3/2 and j = 1/2), evidenced by features at 936.81 eV and 937.10 eV, respectively. In contrast, these features were not observed for the bulk CuO reference (Figure 6 e, f). An additional contribution associated with the Cu 2p_3/2 shake-up satellite was detected in the CuO700 and reference spectra, but it was absent in the CuO600 spectrum. The origin of this satellite feature remains debated; it has been linked to chemisorption effects and/or an excess of copper species at the surface [1,20,21]. The absence of suppressed satellites in CuO600 suggests a reduced surface disorder rather than a direct compositional change? This alternative explanation invites further critical engagement with spectral data. Table 2 summarizes the peak areas obtained from the fitting of the Cu 2p and O 1s spectra. The table also lists the binding-energy assignments for the proposed chemical species in the Cu 2p region (Cu¹⁺, Cu²⁺, and Cu³⁺) and the O 1s region (O²⁻) for CuO600, CuO700, and bulk CuO. Only minor variations were observed when using Gaussian– Lorentzian (GL) line shapes (see Table 3), and the fitted parameters remained consistent with previously reported values; a detailed discussion of the fitting theory is beyond the scope of this work [8].

In Figure 4, the fitted peak areas are compared across the proposed chemical species, showing the relationship between the peak area and the Cu oxidation states (Cu¹⁺, Cu²⁺, and Cu³⁺) (Figure 5a). For the Cu¹⁺ component, CuO600 and the bulk reference exhibit similar areas (≈ 835.7 a.u.), whereas CuO700 shows a markedly higher value (≈ 19,000 a.u.). For the Cu²⁺ component, the CuO600 and CuO700 samples fall in the range of approximately 16,000–22,000 a.u., whereas the bulk sample shows a substantially lower area (≈ 5,035 a.u.). Finally, the Cu³⁺-related component exhibits comparable areas in CuO600 and CuO700 (≈ 11,587.6 a.u.) and was not observed in the bulk reference. Overall, this indicates a trend toward an oxygen-rich but Cu¹⁺-dominant chemical landscape, particularly in the CuO700 sample, where the presence of Cu¹⁺ was significantly more pronounced.

In Figure 4b, the O^2- anion contribution is compared with the previously discussed cationic components through the ratios O^2- :Cuⁱ (with $i=1^{+},2^{+},3^{+}$). The O^2- :Cu³⁺ ratio yielded values of approximately 2,000 a.u. for both CuO600 and CuO700. For O^2-:Cu¹⁺, the minimum value was observed for CuO700 (≈ 4,000 a.u.), whereas the bulk reference and CuO600 showed higher values of approximately 8,000 a.u. and 12,000 a.u., respectively. In addition, O^2-:Cu²⁺ for CuO600 and CuO700 lies in the range of approximately 4,000–6,000 a.u., whereas the bulk sample reaches a maximum of approximately 10,000 a.u.

Cu 2p XPS spectra and the spin–orbit splitting between the Cu 2p_3/2 and Cu 2p_1/2 components for CuO600 (a,b), CuO700 (c,d), and bulk reference samples (e,f), respectively.

On the other hand, the spin–orbit splitting between the Cu 2p XPS spectra and the spin– orbit splitting between the Cu 2p3/2 and Cu 2p_1/2 components for CuO600 (a,b), CuO700 (c,d), and the bulk reference sample (e,f). The Cu 2p_3/2 and Cu 2p_1/2 components are shown in Figure 5b, d, and f. A slight decrease in the binding energy was observed, which may be associated with changes in the electronic structure of the Cu 2p levels (with degeneracy) [22]. The splitting of the main Cu 2p peaks was estimated from the measured positions of the Cu²⁺ components and the corresponding Cu 2p_3/2 shake-up satellite features, yielding values in close agreement with those reported by Pauly, Tougaard, and Yubero [22].

Table 3 lists the binding energies obtained from the fitting of the main Cu 2p peaks of the CuO600, CuO700, and bulk samples. The corresponding fitting errors are 3.7242 × 10^-5 eV, 2.5698 × 10^-5 eV, and 3.05987 × 10^-5 eV, respectively.

The fit parameters associated with the SVCS Shirley and Tougaard backgrounds were consistent with previously reported values. [1,3,22,23,24] The Gaussian and Lorentzian broadening components exhibited values that were characteristic of nanostructured samples (Table 3). In contrast, the bulk reference sample showed no Lorentzian contribution, and the Gaussian broadening remains ≤ 3.0 eV, which is consistent with its larger (micrometric/macroscopic) structural scale.

Table 4 summarizes the background parameters used in the deconvolution of the Cu 2p spectra and their associations with the proposed chemical species. Background modelling was performed using the SVCS Shirley-type background together with a Tougaard background, which enabled the reliable assignment of the chemical components in the nanostructured samples. The background energy parameters obtained for each Cu 2p spectrum (Table 4) were within the ranges previously reported. [10,25,26,27,28] In addition, the degree of oxygen homogeneity was evaluated using the Shirley-type slope parameter. [5].

Figure 6 shows the O 1s spectra of each sample and their relationship with the Cu 2p chemical components. The observed asymmetry could be associated with oxygen coordination in the tenorite (CuO) structure. [29] Unlike previous studies that fitted O 1s using only Gaussian functions, [3,23,29] the O 1s envelopes were fitted using a combined Gaussian–Lorentzian (GL) line shape, yielding an optimized fit consistent with the spectral broadening behavior [5,23].

For CuO, the main lattice oxygen contribution (O2-) appears at 529.50 eV (CuO600), 529.30 eV (CuO700), and 529.76 ± 0.05 eV (bulk). Additional components at 531.15 eV (CuO600), 531.33 eV (CuO700), and 530.66 ± 0.05 eV (bulk) are commonly assigned to higher-binding-energy oxygen-related species, and have also been reported as signatures associated with oxygen-deficient regions (oxygen vacancies). [3,29] Finally, features at 533.13 eV (CuO600) and 532.84 ± 0.01 eV (bulk) are attributed to surface oxygen species. [3] In the bulk sample, an additional peak near ~536 eV is observed, which can be attributed to adventitious carbon/CO2-related species, likely arising from the carbon tape used during the XPS measurements.

The stoichiometries calculated using the topofactor method were, for CuO600: Cu_1.02O_1.16, Cu_2.22O_2.12, Cu_2.22O_3.03, for CuO₇₀₀ were Cu_1.02O_1.15, Cu_2.06O_2.03, Cu_2.05O_3.19 and for the bulk sample: Cu_0.97O_1.02, Cu_2.43O_2.01, with an estimated error of ~4%. These values are supported by the geometrical core–shell model (cylindrical geometry) proposed in the literature. [14,15,30]

Table 5 summarizes the chemical compositions derived from the primary XPS signals (Cu 2p and O 1s). The observed non-stoichiometric ratios may arise from excess oxygen, heat treatments carried out under non-controlled atmospheres, and/or structural defects, including oxygen vacancies. [23,32,33].

5. Conclusions

XRD analysis using the FullProf suite revealed that CuO600 consists of a mixture of Cu₂O (cuprite) and CuO (tenorite) phases forming nanofibers with an average crystallite size of 31.06 ± 0.06 nm, whereas CuO700 shows single-phase CuO (tenorite) nanofibers with an average crystallite size of 20.59 ± 0.06 nm. The NBD analysis of the CuO600 and CuO700 nanofibers revealed superlattice reflections associated with the presence of Cu³⁺. [1,19] The indexed diffraction patterns further corroborated the mixed-phase sample (Cu2O–cuprite and CuO–tenorite) for CuO600 and a single-phase CuO–tenorite sample for CuO700. XPS enabled the identification of the main oxidation states (Cu1+, Cu2+, Cu3+, and O2 −). In particular, the Cu 2p spectra support the presence of Cu3+ in both nanofiber samples, whereas this contribution was absent in the bulk CuO reference. The Cu 2p spin–orbit splitting values are consistent with those reported previously. [1,3,7,22,23,34,35] In Page 16 of 20 addition, the measured Cu 2p spin–orbit splitting (ΔE = E2p1/2–E2p3/2) was 19.79 eV for CuO600, 19.89 eV for CuO700, and 19.92 eV for bulk CuO, suggesting changes in spin– orbit coupling related to exchange interactions involving the 3d states. [25] The fitted spectra also indicate non-stoichiometry, likely arising from calcination under noncontrolled atmospheres, consistent with the compositional analysis. [3,14,15] Finally, the O 1s spectra show that CuO600 exhibits a more pronounced oxygen-vacancy related component (≈ 531.33 eV) and a higher Cu3+-related intensity than CuO700, which is in agreement with previous reports. [3,29] The bulk reference lacks the Cu³⁺ component near ~937 eV; however, it strongly contributes to the oxygen vacancy region.