Copper(II) oxide is a p-type semiconductor possessing a narrow bandgap, along with many beneficial electrical, optical, and magnetic properties. Particularly at the nanoscale, these properties set themselves apart and have sparked growing attention on CuO. As a result, CuO has become the focus of both fundamental research and efforts aimed at its practical application across various areas. Nowadays, CuO thin films showing strong hydrophobic properties are tested, among others, in the context of self-cleaning , anti-corrosion , and antibacterial coatings. Like other CuO nanostructures, thin films also show potential for applications in photovoltaic cells , lithium-ion batteries , supercapacitors , gas sensors , and biosensors . Furthermore, the literature reports their utility in photodetectors , memory structures , and p-type channels in transistors .

Simultaneously, ongoing efforts are consistently directed towards developing deposition techniques and functionalization of CuO films in order to modify their properties for specific applications. Despite the existence of numerous methods for thin film growth based on various physical and chemical phenomena, it remains acknowledged that these techniques do not yield ideal outcomes in terms of applications in any area . Therefore, comprehending the potential and limitations of these different methods becomes important for their continued enhancement and optimization. Such endeavors might consequently lead to the development of novel techniques for producing films tailored to specific parameters and purposes. The preparation of CuO thin films using various techniques is reported in the literature. These encompass methods such as molecular beam epitaxy , direct current magnetron sputtering , and pulsed laser deposition . Alternative approaches involve techniques such as chemical vapor deposition and atomic layer deposition . CuO thin films can also be grown from liquid phases, employing techniques such as chemical bath deposition , successive ionic layer adsorption and reaction (SILAR) , sol–gel processes (using dip and spin coating) , as well as spray pyrolysis . This work focuses on hydrothermally grown CuO films. There are only a few reports in the literature indicating the use of this method for the growth of CuO films. An example of a multistep method of producing CuO films has been described in . It combines the hydrothermal process, annealing, and spin coating. Conversely, direct thin film growth through the hydrothermal approach is outlined in . In a different study , CuO films were grown on glass substrates coated with CuHNO4 through 4, 8, and 12 h long hydrothermal treatments of copper hydroxide nitrate suspension at 200 °C. Another example is the work , outlining the fabrication of structured CuO films on fluorine-doped tin oxide from a solution devoid of additional substances.

Post-processing plays a significant role in the field of nanomaterials fabrication. Metal oxide thin films often undergo annealing. This practice aims to improve the properties of these layers and to attain the desired parameters in the final product. Controlled thermal treatment serves various purposes, such as enabling crystal structure relaxation, defect reduction, and enhancement of the films’ crystalline arrangement. Furthermore, it can lead to improved electrical conductivity and changes in optical attributes.

The CuO films that are the subject of this work were grown using a novel approach to the hydrothermal method. They are produced from aqueous copper(II) acetate solutions heated to approximately 95 °C. Notably, a key distinction of these growth processes is their execution in an open system, accomplishing synthesis within significantly shorter durations (from 48 s up to 6 min) than those reported in the aforementioned literature . Mild growth conditions enhance the appeal of the technology by fostering energy efficiency, cost-effectiveness, and safety. It aligns with the concept of rapid hydrothermal reactions, which is one of the trends in the modern advancement of the method. The work carried out in this area is aimed at obtaining specific materials in the shortest possible time. Taking into account the limited electrical stability of the so-prepared CuO films, their properties were modified by thermal post-processing. The resultant procedure of sequential hydrothermal processes and rapid thermal annealing (HT+RTA) allows for the control of the physical properties of CuO films.

The analyzed films were fabricated on n-type silicon (100) substrates (Siegert Wafer). These substrates were covered with a nucleation layer in the form of gold nanoislands sputtered using a turbomolecular-pumped coater Quorum Q150T ES. CuO films were obtained according to the following procedure. First, an aqueous solution of copper(II) acetate (Chempur) with a concentration of Cu(II) ions of 1 mM was prepared. The mixture was then precipitated by adding NaOH until a pH of 6.5 was attained. Subsequently, the so-prepared solution along with the substrate was placed in a reaction vessel and uniformly heated utilizing an induction cooker (heating power = 3.5 kW) up to a boiling temperature of ca. 95 °C and boiled for 30 s. After this, the sample was removed from the vessel and cleaned in an ultrasonic cleaner in isopropanol for 30 s and subsequently dried with nitrogen. The resultant as-grown film underwent rapid thermal annealing (RTA, AccuThermo AW610, Allwin21) at 450 °C for 5 min. The heating rate was 45 °C/s. This thermal processing took place in an O2/N2 atmosphere at a 1:1 ratio. The complete sequence of the hydrothermal growth process and rapid thermal annealing steps is referred to as the HT+RTA cycle. Samples that underwent this procedure are denoted as “1×”, “2×”, and “3×”, depending on the number of HT+RTA cycles. It is worth highlighting that re-executing the hydrothermal processes does not necessitate any supplementary preparation of samples, for example, surface re-nucleation.

CuO thin films were characterized using an analytical scanning electron microscope (SEM) Hitachi SU-70 with the X-ray microanalysis system (EDX) enabling composition analysis (Thermo Fisher Pathfinder system with a silicon drift detector). Additionally, in the study, a scanning probe microscope (Dimension Icon, Bruker) was used, which allowed for the investigation of both topography and electrical properties of the films. Surface topography analysis was performed by utilizing an atomic force microscopy (AFM) operating in Peak Force Tapping mode. The surface was scanned at a resolution of 1024 × 1024 measurement points using a silicon nitride probe, ScanAsyst-AIR (Bruker). Scanning capacitance microscopy (SCM) measurements were conducted in contact mode using a silicon probe coated with a PtIr layer, SCM-PIT-V2 (Bruker). Capacitance measurements were taken with VAC = 2 V and VDC = 1 V applied. The carrier distribution maps at a resolution of 256 × 256 pixels presented in the paper were derived from the “SCM data” channel. Contact potential difference (VCPD) measurements were carried out using Kelvin probe force microscopy (KPFM) in amplitude modulation mode, also employing SCM-PIT-V2 probes from Bruker. These measurements were performed under ambient conditions, which may have introduced surface states or absorbed water molecules on the surface of the investigated samples. Consequently, the obtained VCPD values and the resulting work function values are susceptible to additional errors, including those arising from the screening effect . Despite the measurement conditions, it was still possible to qualitatively analyze changes in surface potential and work function based on the sample preparation method. The AFM, SCM, and KPFM data were analyzed using the Nanoscope Analysis 3.0 software (Bruker). The CuO films also underwent structural analysis using a high-resolution X-ray diffractometer X’Pert Pro MRD (Panalytical) equipped with a Cu anode (λ = 1.54060 Å).

X-ray photoelectron spectroscopy (XPS) measurements were conducted utilizing a Scienta R4000 hemispherical analyzer with a pass energy of 200 eV and monochromatic Al Kα (1486.7 eV) excitation (Scienta MX-650) operating at 150 W. The energy scale was calibrated by setting the C–C bond at 285 eV. Analysis of the spectra was performed using the commercial CASA XPS software package (Casa Software Ltd., version 2.3.17) with Shirley background. The Gaussian–Lorentzian functional GL(30) was applied during the simulations. Raman scattering measurements were performed under ambient conditions and room temperature using a T64000 Horiba Jobin-Yvon spectrometer configured in a backscattering geometry with a 1800 gr/mm grating and a microscope objective of 100× magnification. The spectral resolution was of the order of 0.5 cm−1. A 532 nm semiconductor laser was used to illuminate the samples. The measurements were performed without detection of polarization of the scattered light. A liquid nitrogen-cooled multichannel silicon CCD camera was used as a detector.

Figure 1: SEM images of the films surface prior to (left) and after exposure of the indicated area to an electron beam for 30 s (right).

Simultaneously, exposing the samples to thermal processing results in a reduction of the content of carbon atoms. On this basis, it is postulated that, under the influence of temperature, organic compounds are released from the material, leading to the formation of discontinuities within the layers. Therefore, the HT+RTA sequencing procedure was developed.

Table 1: Summary of the average thicknesses of the analyzed samples and their elemental composition.

Figure 2: The C 1s peaks in high-resolution XPS spectra of as-grown films (left) and processed using the HT+RTA procedure (2×) (right).

Figure 3: SEM and AFM images of the surface of the as-grown, 1×, 2×, and 3× samples, as well as topography profiles along the marked lines (every 500 nm) presented at the same scale.

Table 2: The analyzed parameters of the surface of the as-grown and sequenced films determined from AFM measurements.

Based on the surface parameters determined from AFM data, that is, root mean square (Rq) and arithmetic average (Ra) of height deviations, it was noticed that the surface roughness progressively increases with each HT+RTA cycle. The most significant increase was observed in films that underwent three sequential cycles of HT+RTA, primarily because of the presence of numerous discontinuities. Furthermore, the HT+RTA sequencing results in a variation of height distributions, as indicated by the skewness (Rsk) and kurtosis (Rku). The most homogeneous surface coverage is observed for the as-grown and 2× samples. However, the 3× sample is the only one characterized by negative skewness, which, together with the high value of kurtosis, indicates that the surface is slightly dominated by steep valleys. The Rq/Ra ratio of approx. 1.25 implies that, despite those discontinuities, the height distributions do not show large deviations from the normal distribution . In conclusion, the 2× sample exhibits the most favorable surface parameters among all samples treated with the HT+RTA procedure. It demonstrates homogeneity without a notable dominance of hills or valleys.

Figure 4: (A) Diffractograms and (B) variations in crystallite sizes and dislocation density determined for as-grown and HT+RTA films.

Three different copper oxide phases can be obtained, namely, cupric CuO, cuprous Cu2O and the intermediate phase paramelaconite Cu4O3. The aforementioned phases of copper oxide have different physical and electrical properties, different colors, and crystal structures . By examining the Raman spectra of copper oxide compounds, phase as well as chemical composition can be identified. Each form of copper oxide presents a different Raman spectrum.

Figure 5: Raman spectra of as-grown (B), 2× (C), and 3× (D) CuO/Si structures along with that of a reference Si sample (A) for comparison.

Comparing the Raman spectra of CuO/Si samples with the spectrum of the Si reference sample, one can notice a dominating phonon mode at a frequency of 521 cm−1 originating from the silicon substrate . The remaining Raman lines, which were fitted with Lorentz functions, reveal mode frequencies in the ranges of 296–300, 345–346, and 624–633 cm−1. These values correspond to the following vibrations in the CuO layer: CuO Ag ≈ 300 cm−1, CuO ≈ 345 cm−1, and CuO ≈ 630 cm−1. Their values are similar to the CuO-like phonon mode frequencies that have been documented by other authors . Thus, the obtained results confirm the cupric phase of the copper oxide layer. After comparing the Raman spectra of all CuO/Si structures, it can be noticed that the 2× and 3× samples exhibit the best crystal quality among the studied structures because they have well-resolved and the most exposed CuO-like Ag, , and phonon modes in their Raman spectra.

The surface potential (i.e., contact potential difference, VCPD) and work function (ϕ) variations resulting from the HT+RTA sequencing were examined using KPFM. First, following the procedure described in , the work function of the probe was determined (ϕtip ≈ 4.95 eV), which is consistent with the literature . Then, the relation ϕCuO = ϕtip − eVCPD was used to calculate the work function of the CuO films.

The work function represents the sum of chemical potential and surface dipoles, which are sensitive to numerous factors. In addition to the stoichiometry of the material and the presence of impurities, their value may also be affected by, among others, surface roughness, exposed crystal plane, and exposure to radiation or gasses that may be absorbed at the surface . The measured VCPD may considerably deviate from the real work function values as some of these factors influence the surface studied under ambient conditions. Nevertheless, because constant conditions were maintained for each of the analyzed samples, it is possible to determine the general trend of parameter changes with subsequent HT+RTA cycles.

Figure 6: (A) Exemplary maps (1 × 1 μm2) of the contact potential difference (VCPD) for each sample. (B) Summary of the average values of VCPD and ϕ. (C) Distributions of the work function values.

The change in contact potential difference provides valuable information about the direction and degree of band bending resulting from surface states. The previously discussed results do not indicate significant differences in the structural properties and elemental composition among the 1×, 2×, and 3× samples. Hence, it is believed that the most probable cause for the increasing work function value is the annealing atmosphere of the samples (O2/N2, 1:1) . The presence of oxygen in the annealing atmosphere fills the oxygen vacancies present in the material, consequently leading to an elevation in the work function value. The combination of O2 and N2 was chosen over pure oxygen because of cost-effectiveness. At the same time, efforts were made to ensure that the selected annealing atmosphere was closer to the composition of air. Nevertheless, additional investigation is required to validate this hypothesis.

Figure 7: The dC/dV signal maps of the analyzed samples with the SCM data and height profiles recorded along the marked lines.

The search for new growth methods of nanomaterials and the continuous enhancement of existing techniques hold significant importance, particularly within the realm of producing materials for specific applications. CuO thin films produced through hydrothermal processes were initially targeted for electronic applications. However, the low electrical stability of the as-grown films, attributed to the presence of organic residues within the samples, necessitated the development of innovative solutions. These solutions aimed to lower the organic compound content while concurrently enabling modifications to the resultant material’s properties.

As elucidated in this study, the sequential execution of hydrothermal growth processes followed by rapid thermal annealing has a positive influence on material characteristics. Analysis of the films’ elemental composition reveals that repeating HT+RTA cycles leads to a significant reduction in carbon compound content within the produced CuO thin films. Consequently, it has a favorable impact on both electrical and thermal stability. Moreover, as evidenced by XRD and Raman spectroscopy analyses, the sequencing enhances the crystal quality of the films, which remain free from other copper compound phases regardless of the preparation procedure. The absence of notable changes in crystallographic structure also implies that the organic compounds present in the films after hydrothermal processes do not integrate into the CuO crystal lattice. Rather, they become entrapped between grains during the dynamic film formation process. This phenomenon likely accounts for their subsequent release from the layers at relatively moderate temperatures, resulting in the formation of discontinuities along the grain boundaries.

Furthermore, an interesting observation pertains to alterations in the surface potential of the films, directly impacting their work function. This modulation of work function holds immense significance, particularly in the context of electronic applications. Subsequent investigations will delve into comprehending the mechanism underlying the change in the work function value. Particularly, the impact of oxygen concentration in the annealing atmosphere on this parameter will be studied in the future. Another noteworthy observation is the change in the carrier distribution, as depicted by SCM. While the singly and doubly sequenced films exhibit minor changes, the 3× sample exhibits distinct regions of reduced carrier concentration at grain boundaries and surfaces of individual grains. Further research is required to gain insight into the phenomena at grain boundaries, aiming to provide a precise explanation of the factors contributing to such a rapid change in the distribution of carriers in CuO films.