Color of Copper/Copper oxide

Stochastic inhomogeneous oxidation is an inherent characteristic of copper (Cu), often hindering color tuning and bandgap engineering of oxides. Coherent control of the interface between metal and metal oxide remains unresolved. We demonstrate coherent propagation of an oxidation front in single-crystal Cu thin film to achieve a full-color spectrum for Cu by precisely controlling its oxide-layer thickness. Grain boundary-free and atomically flat films prepared by atomic-sputtering epitaxy allow tailoring of the oxide layer with an abrupt interface via heat treatment with a suppressed temperature gradient. Color tuning of nearly full-color RGB indices is realized by precise control of oxide-layer thickness; our samples covered ~50.4% of the sRGB color space. The color of copper/copper oxide is realized by the reconstruction of the quantitative yield color from oxide pigment (complex dielectric functions of Cu2O) and light-layer interference (reflectance spectra obtained from the Fresnel equations) to produce structural color. We further demonstrate laser-oxide lithography with micron-scale linewidth and depth through local phase transformation to oxides embedded in the metal, providing spacing necessary for semiconducting transport and optoelectronics functionality.


Introduction
Surface oxidation of copper (Cu), one of the oldest problems in metallurgy, occurs naturally when Cu is exposed to air. The oxidation depends on the imposed environmental conditions. [1][2][3] The tensor relationship between control parameters and oxidation has not been addressed thus far, because the propagation of oxidation occurs randomly at the surface with a high density of low coordinated surface sites (preferentially along grain boundaries). Oxidation is neither prevented nor systematically controlled along a uniaxial direction. Thus, systematic control over surface oxidation is necessary to take full advantage of the properties of metals.
Color modulation of metals has been attempted by exploitation of electrochromism, laser coloration using marking, piezochromism, and plasmonic effects. [4][5][6][7] Highly porous thin films on metal substrates with ultra-thin, finely tuned optical coatings offer color purity enhancement. [8][9][10] However, Cu and its alloys become tarnished and corrode under the ambient conditions often involved in antimicrobial applications on various touch surfaces in healthcare facilities. [11,12] Another strategy to obtain wide color selectivity in a metal film is the construction of sophisticated nanostructures to realize various colors by means of polarization conversion. [13] Despite numerous attempts to modulate color by oxidation and nanostructuring efforts, [14][15][16] the complexity associated with conversion of the Cu lattice into an oxide remains an obstacle for coherent control of the interface between metal and metal oxide, which is necessary to obtain a full, well-defined color spectrum.
In this report, we present a breakthrough in the surface oxidation of Cu, using a grain boundaryfree, ultra-flat single-crystal Cu thin film (SCCF) prepared by atomic-sputtering epitaxy (ASE).
Inhomogeneous oxidation in the SCCF was highly suppressed by introduction of a treatment to minimize the temperature gradient in the film, resulting in the production of a full-color spectrum by precise control of the oxide-layer thickness. This approach is further extended to localized oxidation by laser-oxide lithography for photonic-electronic applications. Figure 1a shows a schematic diagram of ASE, in which all internal electrical circuits are replaced by single-crystal Cu wires instead of commercial Cu wires; the vibration due to ambient noise is highly suppressed by an anti-vibration system (Experimental Section/Methods). However, atomic sputter epitaxy (ASE) aims to realize atomically flat surfaces by stacking atom by atom. Hence, even minute vibration could significantly disturb initial nucleation and lateral growth, especially the coherent coplanar merging of the nuclei. We verified that precise signal transduction and cancellation of electrical interference through the use of grain boundary-free wires were essential for acquisition of high-quality Cu thin films. Indeed, ASE resulted in grain boundary-free, single-

5
A uniform, controllable oxide layer was obtained from the SCCF in a separate heating furnace.
We annealed both SCCF and PCCF films at 330 °C for 1 min under a mixed-gas atmosphere of Ar (83%) and O 2 (17%). The SCCF exhibited an oxide layer with a highly crystalline Cu 2 O layer and nanometer-scale abrupt interfaces. Figure 1 shows cross-sectional high-resolution (scanning)  S3 and S4, Supporting information). We emphasize that coherent oxidation, with a highly crystalline Cu 2 O layer and nanometer-scale abrupt interfaces, is critical for realization of homogeneous, genuine color in Cu films. This is achieved by minimization of the temperature gradient in the Cu film during heat treatment. For this purpose, we designed a double heating system equipped with an interior preheating furnace, in which the temperature was controlled to within ± 0.1 °C (Experimental Section/Methods).
By dramatically improving the interface quality with controllable oxidation in the SCCF, we realized a wide-color spectrum through exclusive use of simple heat treatment via temperaturegradient minimization. Figure 2a shows a photograph of representative SCCF samples with 6 systematically controlled oxide-layer thicknesses. This vivid representation of color, which was not previously achieved, emphasizes "color" as an indicator of the surface/interface quality of the Cu 2 O/Cu heterostructure; thus, it is an indicator of controlled oxidation. The color wheel ( Figure   2b) shows representative colors of Cu films constructed from photographic images of actual samples ( Figure S5, Supporting Information). [18,19] The range of colors achieved by our samples (>300) is mapped as a Commission Internationale de L'Eclairage (CIE) xy chromaticity diagram ( Figure 2c). [20] Notably, colors from the SCCF cover ~50.4% of the area of the sRGB color space of digitally available colors represented by the grey triangle, which represents enhancement of >250% in color coverage relative to a recent study regarding plasmonic color generation (18.4%). [20] The photographic images were also obtained using near-normal geometry. The spectra of incident lights used for the images correspond to AM1.5. The sample colors change slightly according to the different angles of view ( Figure S6, Supporting Information).
The underlying mechanism for the emergence of various colors can be explained in terms of the multiple reflections that occur at the oxidized film surface and interface between Cu 2 O and Cu, [21,22] as shown schematically in Figure 3a. The reflected light depends strongly on the Cu 2 O thickness. The change in color of incident white light that occurs upon reflection, is determined by the dielectric functions of the material. First, the dielectric functions of our Cu 2 O layer are consistently obtained by spectroscopic ellipsometry measurements ( Figure S7 and S8, Supporting Information). [21][22][23][24][25] Based on the obtained dielectric functions of each layer, distinct reflectance spectra in the near-normal incident geometry are simulated for specific CuO/ Cu 2 O thicknesses, using the Fresnel equations. Figure 3b shows the simulated reflectance spectra (dotted lines. The The three samples in Figure 1b  We next investigated whether spatially confined control of oxidation of the SCCF could be realized using laser irradiation. A SCCF sample was mounted on a motorized stage and irradiated using a motorized shutter-equipped continuous laser (wavelength: 488 nm). Notably, absorbance of the SCCF was ~40% ( Figure S15, Supporting Information). Reflectance was subsequently measured by a color scientific complimentary metal-oxide-semiconductor camera. Precise control of both laser intensity and duration was required to balance photothermal heating and conductive cooling. The samples were irradiated with a focal spot of 2 µm in diameter (e -2 ) at varying irradiance of 5 to 3500 kJmm -2 , by modulating the pulse number N (Figure 4a). Immediately after irradiation, colorful multilayered concentric circles were visible in the SCCF, with diameters ranging from 1.7 to 4.5 µm. The circular color pattern exhibited a smooth transition from the center to the surrounding area, suggesting that the depth profile of oxidation was also gradual 9 under irradiation of 2500 to 3500 kJmm -2 on the SCCF; notably, only 5 kJmm -2 was required for the PCCF. Importantly, the laser dose required to create an oxidation pattern of similar size in the SCCF was ~80-fold greater than that of the PCCF (Figure 4b). The heat-affected zone (HAZ) in the PCCF was ~10-fold larger than that of the SCCF for a similar size of oxidation pattern. These results can be explained by the degree of defects or grain boundaries, which allow for more rapid and longer propagation of oxidation in the PCCF, compared to the SCCF.

Experimental Section/Methods
Preparation of oxidised Cu thin film: Atomic sputtering epitaxy (ASE) was adopted to grow a pristine single-crystal Cu thin film (SCCF), which is an improved method achieved by modifying the technical limits of a conventional sputtering system. ASE deposition enhances the quality of the metal film and supplies nearly defect-free and grain boundary-free SCCFs. The system was improved by minimization of signal noise originating from grain boundary scattering of electrons in conductors through replacement of the conventional configuration with a single-crystal Cu wiring network. [27][28][29] Mechanical noises from other equipment, including motors and pumps, also contribute to rough surfaces and defect formation in the films ; thus, in our set-up, any vibration caused by ambient noise was minimized to the fullest extent using an anti-vibration system. Antivibration techniques during bulk single-crystal growth have been previously introduced. [ previous study]. [17] Oxidation of the Cu film using the aforementioned SCCF was controlled by the following parameters: treatment temperature, treatment time, oxygen partial pressure, and pre-treatment Cu film thickness. The most important factor for homogeneous color is the ability to eliminate the temperature gradient in the sample during the oxidation process. In particular, a heating furnace was designed, which was equipped with a preheating chamber connected to a gas inlet. The angles of 60°, 70°, and 80°. To obtain the layer structure and dielectric functions of oxidized Cu thin films, we performed optical analyses step-by-step using WVASE software. Figure S7 shows the experimental spectroscopic ellipsometry data and standard fitting procedure. In all procedures, we fitted Ψ(ω) and Δ(ω), which are the ellipsometric angles obtained from the intensity and phase ratio of the reflectance for s-and p-polarized light, respectively, to minimize the mean squared error (MSE). A lower value of MSE means that the fitted spectra are closer to the experimental spectra, indicating higher fitting quality. We began with the known reference complex (real and imaginary) dielectric functions (ε 1 (ω) and ε 2 (ω)) of Cu 2 O and Cu and fitted the thickness of the Cu 2 O layer ( Figure S7a); the initial dielectric functions of the layers were adapted from Palik. [32] 13 A thin CuO layer and surface roughness layer were added to improved further the fitting ( Figure   S7b and S7c) respectively. The layer structure was thus surface roughness/CuO/Cu 2 O/Cu. Finally, we optimized all parameters together to obtain the result and the actual complex dielectric functions of Cu 2 O. The final fitted results of the different samples are shown in Figure S8.
Consistent fitting of Ψ(ω) and Δ(ω) for the different samples provided reliable ε 1 (ω) and ε 2 (ω) values for Cu 2 O ( Figure S8d). Furthermore, the reflectance spectra of oxidized Cu thin films were simulated using the above model layer structure. Change in the reflectance spectra was intuitively understood from the oscillation of color shown in Figure S9d Figure S4 shows the strain relaxation behavior of the oxide layer via geometrical phase analysis of the cross-sectional interface structure at the atomic scale. The interface is found to be atomically sharp (abrupt) and layer mismatch is noticeably observed at the step edge of the Cu surface. Considering the orientational relationship, the misfit strain (δ) was calculated as ~17%.
In this highly lattice mismatched system, it is expected that the in-plane lattice strain is majorly     The temperature dependent oxidation behavior of a Cu thin film is rather different from that of the bulk crystal or polycrystalline thin film. Thermogravimetric-analysis (TGA) measured at the heating rate of 20 ℃/min ( Figure S11) shows that the SCCF thermal mass abruptly at ~350 °C due to oxidation, whereas the PCCF thermal mass increased rather gradually from 220 °C. This result supports the idea that the oxidation behavior of SCCF changes abruptly at 350 ℃, in which corresponds to the drastic change of Cu 2 O thickness noted by the reviewer. Fitting was carried out using library Cu 2 O, with a surface roughness of 1.5 nm. Figure S14. Reproducibility of the color from SCCF. The SCCF sample that heat treated at 250 ℃, for 2 min for as grown, 2 month old, 6 month old samples.

36
Coloration of the SCCF was relatively reproducible. Control of the oxide layer thickness was carried out with an accuracy of 2-3 nm, and this tight control resulted in good color reproducibility. Sometimes, depending on the degree of surface roughness of the initial Cu samples, a slight color variation was evident. However, using the same oxidation temperature and time resulted in the same color within the mean variation of the RGB values of ~ 2.8% (± 3.6/255) and the colors obtained from high-quality Cu films displayed consistent color within ~ 2.3% (± 2.9/255). The colors obtained from 2-and 6-month-old and as-grown SCCF samples are presented in Figure S14. While the value of R was relatively invariant, those of G and B changed slightly more than red. After irradiating with a laser intensity of 100 kW mm -2 for 1 min during a fixed-loop scan (200 m s -1 speed, 2 Hz repetition rate), the surface morphology was not appreciably altered. The treated area is marked by a white box in Figure S16a. The line profile below the image indicates a few nanometers of thickening after irradiation. The surface morphology change became more apparent after increasing the irradiation time to 2.5 min. We also observed that the surface height increased with increasing irradiation time ( Figure S16b) but the width of the irradiance trace was well maintained. This result supports the idea that coherent oxidation propagation in the depth direction of an SCCF can be achieved by precise control of various laser irradiance parameters, such as intensity, pulse width, repetition rate, treatment duration, wavelength, and beam profile.