Self‐Oxidation Resistance of the Curved Surface of Achromatic Copper

Copper surfaces that exhibit a wide range of achromatic colors while still metallic have not been studied, despite advancements in antireflection coatings. A series of achromatic copper films grown with [111] preferred orientation by depositing 3D porous nanostructures is introduced via coherent/incoherent atomic sputtering epitaxy. The porous copper nanostructures self‐regulate the giant oxidation resistance by constructing a curved surface that generates a series of monoatomic steps, followed by shrinkage of the lattice spacing of one or two surface layers. First‐principles calculations confirm that these structural components cooperatively increase the energy barrier against oxygen penetration. The achromaticity of the single‐crystalline porous copper films is systematically tuned by geometrical parameters such as pore size distribution and 3D linkage. The optimized achromatic copper films with high oxidation resistance show an unusual switching effect between superhydrophilicity and superhydrophobicity. The tailored 3D porous nanostructures can be a candidate material for numerous applications, such as antireflection coatings, microfluidic devices, droplet tweezers, and reversible wettability switches.


Introduction
The unique reflectance of the metal due to free electron motion provides an aesthetic color on its surface.Tailoring the metal surface is key to improving optical, chemical, and biological applications, and not just high-reflectivity surfaces but also antireflective surfaces [1] are important for a variety of applications, such as lowreflection films with antiglare properties, antistatic performance, and higher contrast for optical amplifiers, lasers, and solar cells. [2,3]Antireflection (AR) coatings are used to suppress reflection from a material surface by laminating layers with different refractive indices on the front surface to minimize reflection via destructive interference. [4]Various types of AR films demonstrated their potential for antireflection coatings. [5]Recently, versatile topdown technologies, including femtosecond laser ablation, hole-mask colloidal lithography, and dealloying, [6] have been introduced for AR surface nanostructuring.Controlling nanostructures on surfaces using femtosecond laser ablation provides an opportunity to tailor the AR and achromaticity or the surface properties, and has been developed to optimize geometrical light trapping and enhance effective medium effects reducing the surface reflection, simultaneously, [7] whereas other top-down techniques require several complex procedures and chemicals. [8]For purely geometry-based AR metal films with controllable surface structures, considerable effort has been devoted to establishing a reliable methodology for fabricating large-scale AR surfaces during direct film growth.Oblique angle deposition of metal films is a straightforward technique for constructing nanostructures on AR surfaces. [9]However, the unavoidable polycrystalline nature and surface oxidation of AR metal films can limit the useful properties originating from their single-crystalline counterparts.
Here, we introduce an innovative growth approach that provides single-crystalline nanostructures that have a very rough surface and trap light.Single-crystal nanostructures are composed of curved surfaces but manifest strong oxidation resistance by themselves.Initiated with single-crystalline copper film (SCCF) deposition by coherent atomic sputtering epitaxy (ASE) in a previous report, [10] achromatic copper thin films (ACFs) were then deposited by subsequent incoherent ASE in a manner that preserved the crystallinity of the Cu lattice.Interestingly, the 3D porous nanostructures show strong oxidation resistance despite their rough surfaces.This exceptional chemical stability originates from the characteristic surface structure, which mainly consists of monoatomic step edges and lattice contraction associated with the outermost surface at inflection points of the 3D body.Moreover, we demonstrate that the achromaticity of copper (Cu) films can be tuned by simply controlling geometrical parameters such as pore size and 3D networking.

Fabrication of ACFs
A two-step growth method was employed with an ASE system (see the Experimental Section).An SCCF with a thickness of ≈100 nm was first grown on a 2-inch sapphire (0001) substrate by ASE, [10,11] which allows the growth of nearly defect-free and grain-boundary-free single crystals at heteroepitaxy by controlling the deposition at a single atom-level precision.An incoherent ASE was then introduced to fabricate porous nanostructures by slightly perturbing the coherent ASE condition.A rocking rotation of the growth stage at a low speed sets the ASE condition for the growth of the SCCF to deviate slightly, consequently allowing the growth of porous Cu nanostructures on the surface (Figure 1).Incoherent growth proceeded in such a way that epi-taxial growth conditions were maintained, at least for each nanostructure.A variety of ACFs are displayed in terms of achromaticity and lightness, along with RGB indices, on 2-inch sapphire substrates (Figure 1a).Note that the lightest sample on the top right in the chart is a single-crystalline copper thin film without porous nanostructure and is included for comparison.With the incoherent ASE conditions achieved by changing the rotation speed and radiofrequency (RF) power, the average RGB values of the obtained ACFs varied from 130 to nearly 0. We established processing tables under different growth parameters such as the rotation per minute (RPM) of the substrate, working pressure and RF power (Figure S1, Supporting Information).The light ACF had an average RGB value above 100 (RGB: 119, 103, 105) (Figure 1a, bottom right), but that of the dark ACF decreased to close to zero (Figure 1a, bottom left).In our study, a wide range of achromaticity values, including black metal, [12] was achieved with the combined coherent/incoherent growth process.The front side (Figure 1b, top) exhibits the darkest achromatic intensity (black), with an RGB value of (0, 0, 0) for the ACF marked by the red circle in Figure 1a, and the reverse side seen through the sapphire substrate (Figure 1b, bottom) has a typical bright brown color in SCCF, [11] suggesting the growth of the thin film consists of two stages.

Microstructures of ACFs
It is interesting to note from the X-ray diffraction (XRD) pattern of the ACF sample (marked by the red circle in Figure 1a) that the ACF has a metallic fcc Cu structure preferentially oriented along the [111] direction (Figure 1c).The Cu(111) peak intensity was higher than the Al 2 O 3 substrate (0006) peak intensity with a fullwidth-at-half-maximum of 0.18°, still exhibiting high crystallinity in the ACF.To check whether oxidation on the ACF surface influences the achromaticity, we closely examined the Cu(111) XRD profile replotted on a logarithmic scale.A very small peak corresponding to the Cu 2 O phase [13] is observed (the filled triangle in the inset of Figure 1c).This implies that the nature of Cu metal is well preserved in the ACFs without the formation of noticeable Cu oxide.However, the observed surface was completely rough, with a roughness of a few hundred nanometers.Cross-sectional (Figure 1d, top) and top-view (Figure 1d, bottom) scanning electron microscopy (SEM) images of the ACF show a columnar nanostructure of irregular porous dendrites grown above the Cu(111) film.Apparently, all Cu blobs with an average size of 100 ± 20 nm were arbitrarily agglomerated and preferentially grown toward the [111] direction, with porous nanocavities in between, resulting in the formation of a complex dendrite structure.However, the XRD results with only the Cu(111) peak seem to be far from the conventional prediction of a polycrystalline nanostructure.This indicates that the dendrite structure predominantly selects Cu(111) as the growth front (Figure 1e, left), regardless of their irregular shape evolution.This growth characteristic was preserved until the incoherent ASE growth was terminated, in contrast to the general consequence of a typical polycrystalline dendrite structure (Figure 1e, right).
The porous structure evolution of the ACFs was observed in the annular dark field (ADF) scanning transmission electron microscopy (STEM) images of the four ACFs with different colors: brown (Br), gray (G), dark gray (DG), and black (B) (Figure S2a-d, Supporting Information).The morphology gradually evolved from a vertically aligned columnar structure anchored on a 100 nm thick Cu(111) film to an irregular porous nanostructure.The orientation-dependent growth of nanostructures was further explored by atomic-scale structural analysis of the dense films near the interface and porous nanostructures (Figure S2e-l, Supporting Information).The results show that the orientation relationship between the dense Cu films and the Al 2 O 3 substrate is (111) Cu [110] Cu // (0001) Al 2 O 3 [1 100] Al 2 O 3 .Interestingly, unusual crystallographic features were observed between the Cu film and the porous Cu nanostructures.The (111) plane stacking was preserved even after the growth of the porous nanostructure, whereas individual grains were relatively rotated along the [111] axis.This ordered (111) stacking explains the intense and sharp Cu(111) peak observed in the XRD analysis of the Cu nanostructure.
Despite these geometrically complex structures, the sheet resistance of the ACFs increased marginally from 0.046 to 0.34 Ω sq −1 (Figure S3, column I, Supporting Information).These two values are equivalent to those of an 800 nm thick Cu film and a 100-nm thick Cu film, [14] which confirms the intact electrical and geometrical connectivity of the porous nanostructures.Based on the microstructure analysis, the achromaticity of the ACFs was strongly correlated with the 3D porous nanostructures.Therefore, pore size, distribution, and 3D linkage are the determining factors for tailoring achromaticity (Figure S3, columns II and III, Supporting Information).Regardless of the crystal surface morphology and degree of achromaticity, all the samples were well aligned along the (111) plane.(Figure S3, column IV, Supporting Information).

3D Nanostructure Analysis of ACFs
The achromaticity of Cu films is ascribed to microstructures with nanopores that trap incident light. [2,15]Thus, the lightness or darkness is expected to depend on the size of the nanopores and their connectivity.To reveal the microstructural origin of the achromaticity, we characterized four representative porous Cu thin films of different colors (marked by blue boxes in Figure 1a).As geometrical factors play an important role in the realization of ACFs, we performed 3D electron tomography based on the focused ion beam (FIB)-SEM serial section technique [16] to The total porosity increased with the change from Br-to B-ACFs, and a porosity of ≈30% was required to obtain the B-ACF with the lowest lightness.In addition, the average pore size was slightly reduced to ≈22 nm in B-ACF, starting from Br-ACF (30 nm approximately) (Figure 2f).
However, such a large porosity is not exclusively responsible for achromaticity because other geometrical factors, such as the average size of the pores and their linkages within the 3D volume, may be more sensitive in influencing light scattering; that is, longer connectivity results in more scattering.To resolve the network of the pores, we analyzed the 3D pore linkage, that is, the distribution of the linearly connected lengths between pores. [17]he reconstructed volumes, including the pore linkage and its size distribution, are represented by two G-and B-ACFs (see Column III).All the results for the four samples, together with the statistical deviations, are shown in Figure 2g.The pore linkage notably increased from 64 nm for the Br-ACF to 112 nm for the B-ACF (approximately) with a broadened length distribution.This result suggests that the average pore linkage lengthens as the level of lightness of the ACF decreases.This pore network analysis provides a valuable design rule for the pore distribution and its linkage to tune the light-scattering efficiency of porous Cu thin films.

Microscopic Origin of strong oxidation resistance at the Curved Surface of ACFs
We further examined the probable influence of oxides on the surface or at the interfaces by conducting chemical analyses based on STEM-energy-dispersive X-ray spectroscopy (EDX) and electron energy loss spectroscopy (EELS). [18]The interfaces between the Cu film/Al 2 O 3 substrate and Cu film/porous Cu nanostructure revealed no noticeable oxide formation, as confirmed by the elemental maps of Cu (green) and O (red) (Figure 3a, middle), although a small Cu 2 O signal was detected by the related XRD peak at the log scale (Figure 1c inset).The full-scale EDX data for the four ACF samples is provided in Figure S2a-d in the Supporting Information.Additional STEM-EELS data for the O K and Cu L 2,3 edges also show typical spectral features without noticeable oxides (Figure S4, Supporting Information).Considering that the high detection sensitivity on element was maintained below 1 at% level in our large-area EDX detector system, [19] the surface oxidation per se being too insignificant to influence the optical properties of pure Cu is reasonable.The close-packed Cu(111) surface is superior to the other Cu surfaces in terms of oxidation insensitivity. [10,18,19]Indeed, the stability of the oxidation resistance of the ACF with [111] growth orientation was revealed to remain unchanged after one-month of exposure to the air (Figure S5, Supporting Information).Macroscale chem-ical characterization based on X-ray photoelectron spectroscopy consistently corroborated the strong oxidation resistance of the ACFs (Figure S6, Supporting Information).However, considering the geometry of the Cu dendrite structure enveloped by a curved surface (Figure 3a, leftmost), the benefit of improving the oxidation resistance of the Cu(111) surface seems limited.Despite this, comprehensive tests confirmed the relatively superior oxidation resistance of the ACFs compared to the polycrystalline Cu film under various conditions, including long-term exposure to room-temperature air, high-temperature treatment, high humidity, and a harsh chemical environment (Figure S7, Supporting information).
To understand the sinuous dendrite structure that preserves the strong oxidation resistance, the structural integrity of the Cu nanostructure was first examined by atomic-resolution ADF-STEM imaging.Figure 3b,c shows the representative atomic structures of a Cu nanograin chosen among the Cu dendrites.The out-of-plane A-B-C stacking of the (111) planes was nearly free of structural disorder for any arbitrarily chosen Cu nanograin.This explains why the (111) peak intensity was strong in the XRD data and was distinct from the typical polycrystalline Cu dendrite structure.Examining the atomic array at the surface, we note that a series of monoatomic step edges characteristically construct nanoscale curved surfaces (see the dotted guidelines in Figure 3b,c), that is, the crystallographic planes at the surface are composed of Cu(111) and Cu(001) planes.Our previous report showed that monoatomic step edges are impervious to oxygen, analogous to flat surfaces. [10]Oxygen penetration at monoatomic step edges is an endothermic reaction, whereas oxygen penetration at multiatomic step edges, such as a two-atomic step edge or a three-atomic step edge, is an exothermic reaction.The (111) faceted surfaces around Cu nanograin can partially explain the observed effect of enhanced oxidation resistance from that point of view.However, as shown in the area marked with the yellow square in Figure 3c, a two-atomic step edge or a three-atomic step edge was also observed intermittently at points where the (111) and (001) planes crossed over.Despite the fact that these regions are prone to oxidation, no clear evidence of oxidation was observed within the resolving capability of STEM imaging.
To understand the lattice strain effect on the strong oxidation resistance of the ACF with the preferred orientation, [20] we measured the projected atomic distance (PAD) from the atomicresolution STEM images.For pure Cu lattice aligned to [110] direction, the PAD is measured to be 2.23 Å (d PAD ). Figure 3d shows the PAD maps for the left, right, and top parts of the chosen Cu nanograin and the histogram of the statistical distributions of all measured PADs.The PADs within one or two surface layers is 2.03 Å, which is contracted by ≈20 pm on average compared with the PADs measured from the inner Cu grain.Considering the lattice constants of the Cu 2 O phase, the PAD of the surface is expected to be larger than that of the inner Cu nanograin if the surface part is oxidized. [10]However, the PAD measurements showed that the surface lattices were substantially contracted.This counterintuitive lattice contraction on the outermost surface around the Cu nanograin, as well as a series of monoatomic step edges, leads to improved oxidation resistance.
To determine the physical link between the lattice contraction at the surface and the phenomenologically observed enhanced oxidation resistance, we now turn to first-principles calculations based on density functional theory (DFT).Considering that a monoatom step edge has a high oxidation resistance comparable to that of a perfectly flat surface, [10] we focused on the enhanced oxidation resistance of multiatom steps observed in the curved surfaces of ACF.We constructed a two-atom step edge on a simulated curved surface of an ACF to investigate the infiltration of oxygen atoms.The left panel in Figure 3e shows that oxygen penetration into the two-atom step edge on a simulated curved surface of the ACF is an endothermic reaction (ΔE = 0.99 eV) with a higher activation energy of E a = 1.91 eV.The two-atom step edge on the flat surface of an SCCF is an exothermic reaction (ΔE = −0.90eV) with a lower activation energy (E a = 0.97 eV).This enhanced oxidation resistance is attributed to the contraction of the interlayer distance in the out-of-plane direction at the step edges of the curved surfaces of the ACF, which is consistent with the experimental observations of a several percentage contractions in the interlayer distance, as shown in the right panel of Figure 3e.Such contracted surface layers are ultimately caused by a large difference in the surface tension of the two adjoining surfaces at the step edges.The Cu(111) surface has the lowest surface energy of 1.34 J m −2 among the Cu surfaces, while the Cu(001) surface has a relatively high surface energy of 1.47 J m −2 .Therefore, the dendrites of the ACF are primarily composed of Cu(111) surfaces connected by monoatom step edges, as shown in Figure 3.When Cu(001) surfaces occur at multiatom step edges, the contraction of bond lengths in the [001] direction effectively reduces the area (see Figure 3e, right panel), thereby lowering the total energy.

High Infrared (IR) Reflectance and Colossal Wettability Switching of ACFs
The spectral response of reflectance can be engineered with achromaticity, as shown in Figure 4a.The pure SCCF prepared by ASE (denoted by 1) shows a typically shiny brown color with a high reflectance of >90% at wavelengths above 550 nm, but a low reflectance of <60% at wavelengths below 550 nm and further <40% in the UV range.As the achromaticity increases, the ACFs denoted by 2-9 show a slowly decreasing reflectance in the IR region and a rapidly decreasing reflectance in the visible region.The reflectance of the B-ACF (denoted by 10) is much lower than that of the others over the entire range of the spectral response, that is, below 10% at wavelengths below 1200 nm.The variety of ACFs allows us to further apply them as high-IR reflectors for solar savings and cooling effects.
Porous and grooved nanostructures, as observed in ACFs, can be used to create superhydrophobic or superhydrophilic surfaces.Surface wettability is primarily determined by the surface topography, such as the effective surface area and porosity, which can be explained by the Wenzel and/or Cassie-Baxter models. [21]ost studies have been performed with polymers because their surface energies are relatively low, [22] which is beneficial for hydrophobicity.The surface energy of metals is generally high, which is often detrimental to achieving hydrophobic surfaces. [23]oreover, a metal surface can absorb hydrocarbons or oxygen species and be stabilized, thus becoming less hydrophilic or even hydrophobic. [24]Hydrophobicity can be induced or further enhanced by the formation of nanostructures.Figure 4b shows the contact angles (CAs) of the ACFs.As the achromaticity increased or lightness decreased, the CA increased, corresponding to greater hydrophobicity.Such high hydrophobicity has been observed in the cotton-like structures of Cu nanocrystals. [25]Since the CA of the pristine SCCF was measured to be ≈73°(i.e., hydrophilic), the Cassie-Baxter model seems to account for the hydrophobicity of the ACFs.By considering the Cassie-Baxter model, the fractions of air pockets in the vicinity of the surface due to the porous structure were estimated to be 51%, 65%, 72%, and 79% for Br-, G-, DG-, and B-ACFs, respectively.This result is consistent with the results estimated from the 3D electron tomography of the ACFs (Figure 3).
Figure 4c shows the remarkable wettability conversion of the ACFs induced by UV irradiation (particularly for the B-ACF from 135°to 11°) (Figure S8, Supporting Information).This wettability conversion can be attributed to the following reasons: (i) hydroxyl group functionalization, (ii) removal of hydrocarbons or physisorbed oxygen, and (iii) creation of oxygen vacancies in the topmost surface layer by UV irradiation. [24,26]In cross-sectional STEM measurements of achromatic Cu before and after UV irradiation, any structural and chemical changes were minimal (Figure S9, Supporting Information).The surface states converted by UV irradiation incorporate water droplets inside the pores, rendering the ACFs hydrophilic.Nevertheless, we cannot exclude the possibility that the overall wetting behavior is governed by the Wenzel model if the ACFs are intrinsically hydrophobic (assuming a perfectly flat surface).We estimated the ratio of the actual surface area to the projected flat surface area from the reconstructed surface topographies of the four ACFs; two examples for Br-and B-ACFs are displayed as insets in Figure 4b).The ratios of Br-, G-, DG-, and B-ACFs were determined to be 2.2, 3.6, 4.7, and 6.2, respectively.Therefore, the Wenzel model seems to be applicable for explaining the wetting behavior of ACFs.The superhydrophobic and superhydrophilic states (transformed by UV irradiation) of B-ACF were very stable and were maintained over 24 h (Figure 4d).
For comparison, we examined the wettability of an SCCF and a Cu 2 O thin film compared to ACFs, and the effect of UV irradiation and subsequent air exposure (Figure S10, Supporting Information).UV irradiation leads to hydrophilicity in both SC-CFs and Cu 2 O, but the wetting states are not stable and eventually quickly return to their original states.Surprisingly, the surface properties of ACFs were reversibly switched between superhydrophobic and superhydrophilic by alternating UV irradiation and heat treatment under vacuum (Figure 4e).The highly stable and reversible wettability uniquely observed in ACFs indicates that the surface states of ACFs are fundamentally different from those of SCCF and Cu 2 O.This result is consistent with the results obtained from the 3D nanostructure analysis and theoretical calculation above (Figure 3), which reveal that the single-crystalline dendrite consisting of the succession of monoatomic steps and the surface with contracted interlayer distance is energetically unlike the oxidation-vulnerable rough surfaces (Figure 3e).We speculate that the curved surfaces of the ACFs provide stable sites for anchoring hydrophilic functional groups or forming defects and structurally protect them from their surroundings.The switchable superhydrophobic and superhydrophilic states attainable on a single surface of the ACFs could be considered promising for a wide range of applications in open microfluidic devices, [27] droplet pumps, [28] droplet tweezers, [29] droplet microarrays/libraries, [30] etc.-simple patterning techniques for selective UV irradiation can further enable superhydrophobic/superhydrophilic patterning.

Conclusion
We have realized a series of ACFs with a wide range of achromaticity and self-oxidation resistance.Spectral achromaticity is tailored via a two-step coherent/incoherent film growth based on the ASE technique.The key factors of achromaticity are the porosity, pore size, and pore linkage in the films, which can be explained based on 3D electron tomography.Microscopy combined with spectroscopy demonstrated that the structural and electrical attributes of pure Cu were retained in the ACFs.More importantly, the ACFs exhibit strong oxidation resistance by selfregulation, which is attributed to the crystallographically preferred Cu(111) orientation at the atomic scale.The high oxidation resistance of the curved surface is understood by the presence of a succession of monoatomic step edges and contraction of the interlayer distance in the outermost atoms on the surface.While we demonstrated the probable applications of Cu black metal and wettability switching in our study, the applications of ACFs could be further extended to devices including AR coatings for solar savings and cooling effects, microfluidic devices, and porosity switching systems.

Experimental Section
Growth of ACFs and Structural Characterization: The precondition for the growth of an ACF is ASE, where the growth process occurs by the deposition of single atoms.Otherwise, a porous nanostructure aligned along the (111) and achromatic colors will not be realized. [11]ASE is distinct from other sputtering methods in terms of these three factors.The polycrystalline target should be replaced with a single-crystal target, the mechanical noise should be almost completely removed, and all the internal electrical networks of the equipment should be replaced with singlecrystal wiring.ASE initially operates under optimum conditions to produce defect-free single-crystal thin films of up to 100 nm, then slightly perturbs the system to change the growth process incoherently.However, the partial growth of each region remained coherent.A slight perturbation for incoherent growth is achieved primarily by applying friction to the rotating device of the ASE or by changing the rotational speed.
ACFs of various colors were grown on sapphire (0001) substrates by RF sputtering using a Cu target with a purity of 99.99%.The achromatic intensity, that is, the lightness of the ACFs, is sensitive to the growth conditions, especially to the RPM at low working pressures.The initial pressure in RF sputtering was 2.3 × 10 −3 Pa.To tune the achromaticity and lightness of the achromaticity depending on the growth conditions, the working pressure was controlled from 7.2 × 10 −1 to 1.6 Pa by injecting argon gas (99.999%).Most nanostructure films grown at a working pressure as high as 1.6 Pa have an achromatic or dark color (average of RGB indices < 45), although the darkest black Cu was grown at 0.72 Pa with 30 RPM and 40 W power.The RF power and rotation rate were controlled from 20 to 60 W and from 0 to 60 RPM, respectively.The substrate temperature was maintained at 190 °C during deposition, and the thickness of the film was controlled by the deposition time.The substrate holder was located 10 cm vertically above the single-crystal Cu target (25.4mm in diameter), which was tilted 30°from the vertical line.This tilt angle is far smaller than that for oblique (tilt angle > 60°) or glancing (tilt angle > 80°) angle depositions.The final film thickness was measured by atomic force microscopy (AFM).
Structural and Optical Characterization: The crystal structures of the grown Cu films were identified by XRD measurements conducted on a PANalytical X'Pert Pro instrument with a Cu target (Cu K,  = 0.154 nm) at a tube voltage of 40 keV and a current of 30 mA.Typical symmetric -2 scans were taken between 20°and 80°, with a step size of 0.0167°a nd a dwell time of 0.5 s per point in all cases.The surface morphologies of the Cu films were investigated using field-emission SEM (FE-SEM; S-4700, Hitachi) and AFM (XE-100, Park System).The optical diffusive reflectance was measured with a UV-vis-NIR spectrophotometer (JASCO V-770) equipped with an integrating sphere in the wavelength range of 200-2500 nm.A white BaSO 4 plate was used as the reference.All measurements were performed at room temperature.
Calibration of Color and RGB Value: The colors of the pictures taken with cameras are affected by the cameras, lenses, and lighting.It is necessary to calibrate the color and tone for an accurate color without relying on cameras.In this study, a color reference card, ColorChecker Passport 2 from X-rite, was used, which is the standard for the photography industry.A color profile was created using the ColorChecker Camera Calibration software after taking pictures of achromatic Cu with a reference card.The RGB value of the ACF was determined using Adobe Photoshop software after color calibration was performed using an appropriate color profile for each ACF image.
Cross-Sectional STEM Analysis and Serial Section Electron Tomography: To investigate the cross-sectional structures of the ACFs, the ADF imaging mode of an aberration-corrected STEM instrument (JEM-ARM200CF, JEOL) operating at 200 kV was employed.The angle range of the ADF detector was set to 45-175 mrad, and the semi-convergence angle of the condenser lens was ≈24 mrad.A home-built Python code for an intensityfinding algorithm using the center of mass and 2D Gaussian fitting methods was used to measure the PAD of Cu atoms in the ADF-STEM images of Cu nanograins in ACF.In combination with STEM imaging, elemental mapping of the Cu films grown on Al 2 O 3 (0001) substrates was performed in the same STEM imaging mode using an EDX spectrometer (JED-2300T, JEOL) with a dual-type silicon drift detector and a large effective solid angle (≈1.2 sr).Cross-sectional TEM sampling was conducted using the Ga ion milling and slicing method in an FIB scanning electron microscope (Helios NanoLab 450, Thermo Fisher Scientific).Low-energy Ar ion beam milling at 700 V for 10 min was consecutively performed as a postsurface treatment to remove the damaged surface layer that usually forms during heavy Ga-ion beam milling.Another FIB-SEM instrument (Auriga CrossBeam Workstation, Carl Zeiss) was used to obtain serial section images of the four ACFs for electron tomography analysis.The interval thickness of the FIB serial sectioning was set to 20 nm, and the total number of section images was 92 at the same magnification of 90k× as the initial image dataset.The FIB-SEM-based serial section imaging method collects a series of 2D SEM images during continuous FIB sectioning, which can be aligned to construct a 3D volume for a sample.Furthermore, it provides superior spatial resolution beyond that of X-ray computed tomography, which is essential for the 3D study of nanometric structural changes.This artificial volume reconstruction technique can thus be effectively used for visualizing and analyzing the 3D nature of porous structures with an internal distribution of closed pores and hidden linkages of open pores that cannot be measured by simple 2D microscopy imaging.All acquired images were processed by wavelet-Fourier filtering to remove image-curtaining artefacts due to the vertical milling of the Gaion beam across the interface sample.A commercial 3D software package (Dragonfly ver.2020.1,Object Research Systems (ORS) Inc.), stacking and alignment of the whole section images for each sample were carried out to reconstruct a 3D data volume, and the reconstructed tomographic data were further processed for quantitative image segmentation analysis to measure geometrical parameters, such as total porosity, pore size distribution, and pore-to-pore linkage.To visualize the 3D network structure of pores in each sample, the pore network modelling software OpenPNM package was used, [17] which assists in the modelling and simulation of pore networks, written in Python.To parameterize the mesoscopic surface structures of the ACFs, the surface topographies were used of the samples as shown in Figure 4b (insets).The volumes, including the surface topographies, were extracted from the tomographic volume data, with the definition of an artificial rectangular volume containing rough exposed surfaces with open pores.As the porosity of the ACF increased, the surface roughness was expected to also increase, resulting in a larger exposed surface area.B-ACF has the largest surface area because it has the roughest surface structure with the largest density of open pores.The surface areas (S) of the four ACFs relative to the flat surface, from Br to B, were estimated to be 2.2, 3.6, 4.7, and 6.2.
CA Measurement: To investigate the surface wettability of ACFs, CA measurements were performed (SmartDrop, Femtobiomed, Korea).A water droplet (3 μL) was dropped on the films, and the CA was measured within five seconds.The analysis of CAs from the captured images was performed using package software.UV ( = 184.9and 253.7 nm) irradiation was applied using a UVC-30 (Jaesung Engineering Co.).By considering the surface morphology of ACFs, the Cassie-Baxter and Wenzel models were used to analyze the CAs of ACFs.The Cassie-Baxter model considers two interfaces for a flat surface.For the solid and air pocket interfaces, the apparent CA ( c ) in this model is given by cos c = f (cos + 1) − 1, where f is the fraction of the solid interface (thus, (1 − f )∕100 is the estimated fraction of air pockets in the main text) and  is the CA of a liquid on a perfectly smooth and homogenous solid surface.The Wenzel model considers the increase in surface area that interacts with a liquid drop.This gives cos  c = r cos , where r is the roughness factor, which is the ratio of the actual surface area to the projected surface area.
Theoretical Calculations: All ab initio total energy calculations and geometry optimizations were performed using DFT in the generalized gradient approximation (GGA) Perdew-Burke-Ernzerhof functional [31] and the projected augmented plane-wave method, [32] as implemented by Kresse et al. [33] The Cu substrate was represented by slabs of six layers, with a theoretical equilibrium lattice constant.A vacuum length of 15 Å was used, and the bottom two layers of the slab were fixed at their bulk positions.The electron wavefunctions were expanded in a plane-wave basis set with a cutoff energy of 400 eV.The nudged elastic band method [34] was used to calculate activation energies.The activation energy of oxygen penetration is calculated from the relative total energies in reference to the energy of the configuration when the oxygen atom is adsorbed at the hollow site on the side face of the step edge.
Statistical Analysis: The measured values of geometric factors such as pore size from 3D tomographic volumes of each sample were extracted as Excel datasheets from the Dragonfly software package, and the statistical interpretations were performed using Origin software package.Likewise, all statistical analyses of the STEM image-based atomic distance measurements, contact angle measurements, XPS, and XRD data were performed using the Origin software with the number data.

Figure 1 .
Figure 1.Control of achromaticity of porous nanostructure films.a) Color chart of nanostructure films in terms of achromaticity and lightness along with average RGB values.Note that bright speckles seen on the edges of wafers are shadow marks from mechanical pins holding the wafers during RF sputtering.The Cu films marked with blue boxes are chosen for the structural and chemical characterizations hereafter.b) Optical photographs of the front and reverse sides of the darkest film (marked with the red circle in (a)).c) XRD analysis of the porous Cu film.The inset shows the XRD profile replotted on a logarithmic scale to represent weak peaks.The open inverted triangles (▽), filled red diamonds (◆), and filled triangles (▲) indicate the peaks of the Al 2 O 3 , Cu, and Cu 2 O phases, respectively.d) Side (upper panel) and top (lower panel) views of the darkest ACF observed by SEM.e) Schematic illustration of dendrites grown with single-crystalline structure (left) and polycrystalline structure (right) for comparison.

Figure 2 .
Figure 2. Electron tomography of ACFs displayed with different colors.a-d) Reconstructed 3D volumes of the four ACFs presented at different magnifications (columns I to II).The square size in column II is 100 × 100 nm 2 .3D linkages of pores in two Cu nanostructures (labelled as G and B) are represented in the dataset of column III.Examples of 3D connectivity of a pore inside each reconstructed volume are displayed in circles at the bottom right.Note that the different false colors and the sizes of spheres in the reconstructed volume data denote the number of connected pores and the diameter of measured pores, respectively.The pore diameter is defined by the largest diameter of an artificial sphere filled inside the real pore.e) Plot of the total porosities of the four samples.f,g) Statistics of the pore size and pore connection length (linkage) of each ACF derived from pore network analysis.

Figure 3 .
Figure 3. Surface structure analysis and origin of strong oxidation resistance in ACF.a) (left) Low-magnification ADF-STEM image of a top part of B-ACF, (middle) composite elemental maps of Cu (Cu K = 8.04 keV, green) and oxygen (O K = 0.525 keV, red) for B-ACF sample, and (right) ADF-STEM image of the interface region between Cu film and Al 2 O 3 substrate.The white dashed square in the elemental map of middle panel denotes the imaging region.b,c) (left) ADF-STEM images of parts (denoted by L and R in (a)) of Cu nanograin depicting atomic steps on the surface and (right) superposition of atomic model with A-B-C planar stacking on each structure image.d) Projected atomic distance (PAD) maps for left (L), right (R), and top (T) parts of chosen Cu nanograin (denoted in a) and histogram of the measured PADs.e) DFT results.(left) Relative total energy profile of the O atom penetrating from outside into inside of biatomic step edge of a curved surface (violet open square □), compared with the biatomic step edge of a flat surface (black open square □); (right) Compression of the interlayer distances in (001) direction near the biatom step edge of curved surface of ACF.Blue spheres represent Cu atoms in bulk and dark blue spheres represent Cu atoms in the steps.

Figure 4 .
Figure 4. Reflectivity and wettability switching of the ACFs.a) Reflectance spectra of the ACFs for the visible and IR ranges.The reflectance is sensitive to the achromaticity of the ACFs.The red line (denoted by 1) is the reflectance of the normal Cu film.The reflectance of the ACFs (denoted by 2-9) decreases as the achromaticity increases.The smallest reflectance is 2.5% at 550 nm, with a value of 19.5% at 1800 nm, for the darkest ACF (denoted by 10) with RGB indices of (0, 0, 0).b) Contact angles (CAs) of the four ACFs.c) Change in CAs of the ACFs with UV irradiation and subsequent air exposure as the function of duration time.d) Change in the CA of the B-ACF induced by UV irradiation and subsequent air exposure for 24 and 48 h (purple line).The CAs of the B-ACF without UV irradiation (i.e., only time-dependent) are included for comparison (orange line).e) Change in the CA of the B-ACF induced by repetitive UV irradiation (UV) and heat treatment (T) under vacuum (0.1 atm).The numbers denote the temperature and time for each heat treatment.