Cu Direct Nanopatterning Using Solid‐State Electrochemical Dissolution at the Anode/Polymer Electrolyte Membrane Interface

Patterned copper (Cu) has applications in various electronic devices such as metal interconnects. Electrochemical approaches to patterning, such as electrochemical machining and electroplating, are promising methods for the direct fabrication of patterns on Cu surfaces. However, owing to the use of electrolytes, this technique suffers from issues such as low patterning resolution and significant environmental impact. Herein, a novel direct electrochemical patterning technique for the anodic dissolution of Cu using a polymer electrolyte membrane (PEM) stamp with a patterned surface instead of a liquid electrolyte, is proposed. In this technique, an electrochemical reaction selectively ionizes the Cu surface in contact with the PEM stamp, enabling the fabrication of patterns on Cu surfaces with a resolution of several hundred nanometers under a low‐temperature air atmosphere. Overall, the proposed all‐solid‐state electrochemical patterning technique employing PEM stamps offers a facile, environmentally friendly, and direct process that does not use resists or harsh chemicals, making it a cost‐effective approach that can improve processing efficiency.


Introduction
[3] Materials with patterned surfaces have applications in environmental engineering, [4] biomedical engineering, [5] electronic devices, [6] automotive, and aeronautics. [3]Owing to their high electrical conductivity and DOI: 10.1002/admi.202300896high resistance to electromigration, [7,8] patterned copper (Cu) is extensively used in the fabrication of semiconductor devices, particularly metal interconnects.Other applications of patterned Cu include current collectors to improve the performance of lithium-ion batteries [9] and substrates for the direct growth of graphene nanostructures. [10,11]arious patterning techniques, such as photolithography [7,12,13] and imprint lithography [14] have been extensively investigated.Photolithography has been widely employed for semiconductor device fabrication; however, the process, which employs photo-resist, is complex and requires expensive equipment, which increases the processing costs.Imprint lithography is a promising next-generation process that enables low-cost, high-throughput, and highresolution patterning.[16][17][18] Direct nanopatterning on metallic materials via imprint lithography using extremely high pressure has been reported; however, high pressure shortens the lifetime of the mold. [16]Laser beam machining can efficiently fabricate patterns directly on material surfaces. [19]However, challenges persist in achieving micro-nanopatterns because of the limitations imposed by the focused spot size of the laser beam.
To fabricate micro/nanopatterns directly at a low cost, direct patterning techniques based on electrochemical approaches, such as electrochemical micromachining (ECM) [20][21][22] and electrolytic plating, [23] have been proposed.ECM involves transferring the patterned surface of the cathode tool onto the material surface via anodic dissolution at the interface between the anode and electrolyte.However, fabricating high-resolution patterns via ECM requires an interelectrode gap of 5-15 μm. [21]Thus, the removal of process products like sludge is difficult, and the consequent accumulation of sludge deteriorates the processing characteristics, thereby preventing high-resolution patterning.Furthermore, the use of electrolytes increases the environmental impact of the waste liquid treatment.Hsu et al. [24][25][26][27][28][29] proposed a solid-state superionic stamping (S4) method, in which material surfaces can be directly patterned via anodic dissolution using a stamp with patterns comprising a superionic conductor, such as Ag 2 S and Cu 2 S, instead of liquid electrolytes.Anodic dissolution is confined to the physical contact area between the patterned superionic conductor and the material (anode), enabling high reproducibility and high-resolution pattern transfer.Moreover, this technique has been used for fabricating patterns with sub-hundred-nanometer resolution on silver or Cu films.However, some superionic conductors only function at high temperatures, and the fabrication of nanopatterned stamps is challenging because of their brittle nature.Owing to the high mechanical hardness of superionic conductors, achieving high throughput and large area patterning using roll-to-roll and roll-to-plate processes via the S4 process is difficult. [30]n our previous study, we developed a liquid electrolyte-free electrochemical oxidation method using a polymer electrolyte membrane (PEM) and demonstrated the surface oxidation of gallium nitride (GaN), [31] polishing with the surface oxidation of silicon carbide (SiC), [32,33] and patterned oxide film formation on silicon (Si) [34] and titanium (Ti). [35]This approach involved electrochemical treatment performed using an anode/PEM/cathode sandwich electrochemical system.The PEM employed in this patterning technique is a flexible solid electrolyte, rendering the technique liquid-free and capable of gently modifying the material surface at room temperature and atmospheric pressure.While our previous studies focused on anodic oxidation between the PEM and the material, anodic dissolution at the interface be-tween the anode and PEM has not been reported.We first discovered the ionization of Cu via anodic dissolution at the interface between the PEM and anode.Therefore, we propose a direct patterning technique based on the anodic dissolution of the Cu surface using a patterned PEM.As shown in Figure 1, patterning was performed using an electrochemical system where the PEM was sandwiched between Cu (anode) and a cathode.The PEM can transport ions through nanoscale water channels within the membrane, effectively functioning as solid-state electrolytes. [36,37] PEM stamp with a pattern on its surface was used, and the electrochemical reaction proceeded only at the contact area of the PEM stamp, thereby transferring the pattern of the PEM stamp onto the Cu surface.
In this study, the electrochemical anodic dissolution characteristics of Cu surfaces were investigated.The patterning technique is discussed based on the results obtained from electrochemical measurements, X-ray analysis of the chemical state of the Cu surface, and the achieved patterning resolution.Finally, the application of the roll-to-plate process is demonstrated.

Investigation of Cu Ionization at the PEM/Cu Interface
Cu ionization at the interface between the PEM and Cu surfaces and the transfer of the resulting Cu ion through the PEM to the cathode (illustrated in Figure 1a-1) were verified through simple solid-state electrolysis using a nonpatterned PEM.The cross-sectional scanning electron microscopy (SEM) image of the PEM after electrolysis (Figure 2a) reveals the formation of a thin film with a thickness of several micrometers on the PEM surface in contact with the cathode during electrolysis.The energydispersive X-ray spectroscopy (EDX) elemental maps shown in Figure 2b-1,-2 clearly demonstrate the deposition of a Cu thin film on the PEM surface, whereas fluorine was uniformly distributed inside the PEM.The EDX spectra (Figure 2b-3) show weak Cu peaks, even inside the PEM (indicated as point 2 in Figure 2a), along with clear fluorine and sulfur peaks.The fluorine and sulfur peaks were respectively derived from the C─F bonds and sulfonic group (─SO 3 H) of perfluorosulfonic acid (PFSA), which is a component of the PEM used in this study. [36,37]he presence of the Cu peak in the PEM suggests that Cu ions, formed via anodic dissolution at the interface between the Cu anode and the PEM, were transferred through the PEM from the anode to the cathode.The Cu ions were subsequently reduced at the interface between the cathode and PEM, resulting in the formation of a Cu thin film on the PEM.A water adsorption layer was believed to be present at the interface between the material (anode) and the PEM, which was likely highly acidic owing to the presence of sulfonic acid groups on the PEM.The Pourbaix diagram (potential-pH equilibrium diagram [38] ) indicates that Cu is ionized when a positive potential is applied in the strongly acidic region, suggesting that Cu was removed at the interface between Cu and the PEM.In contrast, the Pourbaix diagram indicates that the Si and Ti surfaces are oxidized in an acidic atmosphere.Thus, as reported in our previous study, the patterned oxide was formed on the Si and Ti surfaces, [34,35] and not on the Cu surface.

Electrochemical Etching Characteristics at the PEM/Cu Interface
The electrochemical etching characteristics at the interface between the PEM and Cu surfaces were investigated using waterswollen and dried PEM.Water-swollen PEM was prepared by immersing the PEM sample in deionized water for 10 min; dried PEM was prepared by storing it in an atmosphere with a relative humidity of 20-30% without immersion in deionized water.As shown in Figure 3a, during electrolysis using dried PEM, the current density increased with voltage at the initial stage of the linear sweep voltammetry (LSV) curve as the voltage was increased to less than ≈5 V.This indicates that the electrolysis is based on a charge-transfer-limited process in which the rate of Cu 2+ formation at the PEM/Cu interface limits the electrochemical reaction.However, the current density saturated at voltages greater than 5 V, suggesting that the kinetics of the electrolysis changed to a diffusion-limited process.Owing to the significantly lower ionic conductivity of the dried PEM compared to that of a conventional liquid electrolyte, the Cu 2+ ions generated by electrolysis from the Cu anode accumulate in the electrical double layer of the PEM, which counteracts the potential applied at the PEM/Cu interface.In contrast, the LSV curve obtained using the water-swollen PEM exhibited a clear peak at the initial stage of electrolysis at ≈1.2 V, which can be attributed to water electrolysis.A sufficient amount of water is contained in the water-swollen PEM compared to the dried PEM; therefore, oxygen gas evolution via water electrolysis at the PEM/Cu interface is believed to be dominant at lower applied voltages: As observed with the dried PEM, the current density increased with an applied voltage of less than 5 V when using the waterswollen PEM.However, the current density when using the water-swollen PEM was higher than that of the dried PEM because of the high ionic conductivity of the water-swollen PEM, [37] resulting in a higher diffusion rate of the formed Cu 2+ ions.Subsequently, the current density gradually decreased with increasing voltage.The water contained in the PEM was consumed as electrolysis proceeded, according to Equation 1, and the ionic conductivity of the water-swollen PEM decreased, thereby decreasing the current density at an applied voltage exceeding 5 V. Cyclic voltammetry (CV) was conducted using the dried PEM at a scan speed of 50 mV s −1 , and the results are shown in Figure 3b.No obvious peaks were observed in the positive bias region of the LSV curve (Figure 3a), and the current density increased with voltage.Meanwhile, a reduction current can be observed in the negative bias region, suggesting that the Cu 2+ ions in the PEM were reduced at the PEM/Cu interface.The chemical state of the Cu surface treated with water-swollen and dried PEM was characterized using X-ray photoelectron spectroscopy (XPS).[41] These peaks provide evidence of a metallic Cu (or Cu 2 O) layer on the sample surface.[41] This suggests that an oxide (CuO) or hydroxide (Cu(OH) 2 ) layer was formed on the surface treated with dried PEM.In electrolysis using water-swollen PEM, a strongly acidic thin water layer is believed to be formed between the PEM and Cu surfaces because of the presence of sulfonic (─SO 3 H) groups terminated on the PEM.According to the Pourbaix diagram, under such strongly acidic conditions, the ionization of Cu to generate Cu 2+ at the electrolyte/anode interface was dominant, and thus, no obvious CuO oxide layer was observed on the surface treated with the water-swollen PEM. [38]Even if an oxide layer is formed on the Cu surface, the strong acidic layer can remove the oxide layer. [42]In contrast, owing to the insufficient acidic water layer formed between the dried PEM and Cu interfaces, the oxide layer cannot be removed from the surface, resulting in the CuO oxide layer remaining after electrolysis using the dried PEM.
As shown in Figure 4a,c, although the untreated surface does not exhibit an obvious peak, a clear F 1s peak can be observed on the surface treated with the dried PEM, which can be attributed to C─F or O─C─F (690 eV) bonds. [43]These components were derived from PEM (PFSA), suggesting that PEM stamps were decomposed by strong oxidizing agents such as hydroxyl radicals (OH•) and atomic oxygen (O•), which are potentially generated through the electrolysis at the PEM/Cu interface, [31,32] and the resulting PFSA components from PEM stamps adhered to the Cu surfaces.The intensity of the F 1s peak from the surface treated with the water-swollen PEM was lower than that treated with the dried PEM, which was mainly due to the indirect contact between the water-swollen PEM and the Cu surfaces owing to the presence of the water layer at the interface.For the C 1s spectra (Figure 4d), a fluoride-related peak (CFx at ≈292 eV) [43] can be observed on the surface treated by the dried PEM, along with the organic contaminants (C─C at a BE of ≈284 eV). [44]The C─F component was also derived from PEM, suggesting that a large amount of PEM adhered to the oxidized Cu surface treated with dried PEM.For the O 1s spectra shown in Figure 4e, peaks related to the organic contaminants (C═O at 531.4 eV) [45] can be observed for all samples; however, the surface treated by the dried PEM exhibits only a broad peak at a higher BE, which can be assigned to C─O (533 eV). [46,47]urther investigation of the chemical states of the Cu surfaces was conducted using X-ray absorption fine structure (XAFS).As shown in the Cu L-edge spectra (Figure 4f), the untreated Cu surface and the surface treated with the water-swollen PEM exhibit a peak at a photon energy of ≈934 eV, whereas the spectrum from the surface treated with the dried PEM shows a peak at a lower photon energy of ≈931 eV.According to Greiner et al., [48] the peaks at the lower (931 eV) and higher (934 eV) photon energies were assigned to CuO and Cu 2 O (or metallic Cu), respectively.In the O K-edge spectra (Figure 4g), a sharp peak at ≈533 eV can be observed for the untreated surface and the surface treated with the water-swollen PEM, which is associated with Cu 2 O.In contrast, the spectra for the Cu surface treated with the dried PEM exhibits a slight peak at ≈530 eV and broad peaks at ≈537 eV, which can be assigned to CuO. [49] These results indicate that a CuO layer formed on the sample surface treated with dried PEM, whereas the surface treated with water-swollen PEM comprised Cu 2 O or metallic Cu, which aligns with the XPS analysis.As shown in Figure 4h, although a higher intensity was obtained for the dried PEM, the shape of the F K-edge spectra of the treated Cu surfaces is similar to that obtained for the PEM, clearly indicating that the PEM adhered to the treated Cu samples. [50]he effects of water-swollen or dried PEM stamps on the patterning performance of Cu were investigated for the fabrication of micropatterns on Cu surfaces.As shown in the chronopotentiometry (CP) curves (Figure 3c), the voltage during electrolysis using the water-swollen PEM remained constant for ≈120 s and then increased drastically from 2 to 10 V. Note that the electrolysis was intentionally stopped after the voltage reached 10.8 V, and the duration is defined as T e , as shown in Figure 3c.The sharp increase in the bias voltage at the final stage of electrolysis was believed to be due to the partial exposure of the dielectric glass substrate on the sample surface caused by the electrochemical removal of the Cu thin film.The bias voltage of the dried PEM was ≈2.5 times higher than that of the water-swollen PEM during the initial stage of electrolysis.As reported previously, the electrical conductivity of PEM significantly depends on the water content in the membrane; the higher the water content in the PEM, the higher the electrical conductivity. [37]The lack of water channels in the dried PEM resulted in a higher electrical resistance of the electrochemical system compared to that of the water-swollen PEM.The voltage sharply increased at ≈45 s.The electrolysis time T e of the dried PEM was 0.36 times shorter than that of the water-swollen PEM, which suggests that excessive electrochem-ical etching by water-swollen PEM occurred outside of the contact area between the PEM stamp and Cu surface.The optical microscopy and SEM images of the Cu surface treated with the water-swollen PEM (Figure 3e-1,f-1) show that the Cu surface was partly removed, which resulted in the nonuniform pattern formation of the Cu film.Furthermore, as indicated by the red arrows in Figure 3f-1, the loss of the formed pattern owing to excessive etching of the Cu film can be partly observed on the surface.In contrast, the Cu surface treated with the dried PEM exhibits a uniform pattern in the contact area between the Cu and PEM stamp surfaces, and the pattern structure on the PEM stamp was accurately transferred over a large area (Figure 3e-2,f-2).The lowaccuracy pattern formation with excessive removal of the Cu surface caused by the water-swollen PEM can be attributed to water electrolysis, which causes undesired electrochemical etching of the Cu film, even at the non-contact area between the PEM stamp and Cu surfaces.Furthermore, the evolution of oxygen gas between the PEM stamp and Cu surfaces via water electrolysis hindered the electrochemical reaction, resulting in the nonuniform pattern formation.
Microcompression testing (MCT) was performed on the waterswollen and dried PEM.As shown in Figure 3d, the compression deformation of the water-swollen PEM was significantly larger than that of the dried PEM.Such low mechanical strength of the water-swollen PEM can cause significant deformation of the microstructures formed on the PEM stamps, which can deteriorate the accuracy of the pattern transfer on the Cu surface.The results obtained suggest that dried PEMs are superior to water-swollen PEMs for use as stamps in electrochemical imprinting.

Micro-Nanopatterning of Cu Surface Using PEM Stamp
We investigated the effects of the applied pressure between the PEM stamp and the Cu sample on the patterning performance of the Cu thin-film surface.As shown in Figure 5a, the CP curves exhibit different characteristics depending on the applied pressure, even under a constant electrolytic current density (1 mA cm −2 ).Higher applied pressure resulted in a lower bias voltage during the initial stage of electrolysis.The electrolytic time, T e , increased as the applied pressure increased.The optical microscopy images in Figure S1 (Supporting Information) clearly show an increase in the processing area with the applied pressure.This can be attributed to the enlargement of the contact area between the PEM stamp and the Cu surface at a higher contact pressure.An increase in the contact area decreased the electrical resistance of the electrochemical system, thereby reducing the electrolytic voltage at a high contact pressure (Figure 5a).An enlarged contact area results in a lower current density, which decreases the etching rate of the Cu film (as discussed below).Consequently, a longer T e is required to etch the Cu thin film to expose the glass substrate under higher contact pressure.As indicated by the red arrows in Figure S1a-1 (Supporting Information), several circular unprocessed regions with diameters of several hundreds of microns can be observed on the Cu surface under a processing pressure of 3.8 MPa.This originates from the gas bubbles existing between the PEM stamp and the Cu surface, which hinder electrolysis at the PEM/Cu interface, as discussed previously.Although circular features can be observed on the Cu surfaces treated under pressures of 10 MPa (Figure S1a-2, Supporting Information) and 17 MPa (Figure S1a-3, Supporting Information), their diameters decreased with increasing contact pressure because of the breakup of gas bubbles caused by the high pressure.As shown in Figure 5b-1, the patterned structure on the PEM stamp was not accurately transferred because of insufficient contact between the PEM stamp and the Cu surface under a lower contact pressure (3.8 MPa).In contrast, a high contact pressure (17 MPa) produced a rough pattern structure on the Cu surface, with the Cu film exhibiting a large and nonuniform removal width (Figure 5b-3).Such an inaccurate pattern formation can be attributed to the elastic deformation of the PEM stamp caused by the high pressure applied to the stamp.As shown in Figure 5b-2, a precise and uniform pattern transfer on the Cu surface was achieved at an intermediate contact pressure (10 MPa).
Next, the effect of the electrolytic current density on the etching rate was investigated.As shown in Figure 5c, T e decreased with increasing electrolytic current density.The etch rate R etch of the Cu film was estimated as R etch (nm s −1 ) = t (nm)/T e (s), where t is the initial thickness of the Cu film and T e is the electrolysis time determined by the CP curve, as described previously.The ideal etching rate was calculated using Faraday's law of electrolysis as follows: where w (g) is the mass of the material etched by the electrolysis, M is the molar mass of the material (63.546 g mol −1 for Cu), Q (C) is the electric charge, n is the valence (2 for the chemical reaction of Cu), F is Faraday's constant (96 485 C mol −1 ).For the CP mode (constant current mode), Q can be expressed as Q = It, where I (A) is the current, and t (s) is the electrolysis time.The volume V (cm 3 ) etched by electrolysis can be expressed using the density of the material  (8.96 g cm −3 for Cu): The etch depth of material d (cm) can be obtained as follows: where S (0.25 cm 2 ) is the area of the PEM stamp and j (A cm −2 ) is the current density (j = I/S).The ideal etching rate (R i etch ) was determined as follows: Finally, the ideal etching rate can be defined as a function of the current density j: The Faradaic efficiency  was obtained by dividing the experimental etching rate R etch by R i etch .As shown in Figure 5d, the etching rate increases linearly with an increase in the electrolytic current density.However, the Faradaic efficiency decreased as the current density increased from 82% (j = 0.2 mA cm −2 ) to 57% at (j = 4 mA cm −2 ).This can be attributed to the side reaction, that is, the evolution of oxygen gas at the PEM/Cu interface via water electrolysis, according to Equation 1, which is dominant under high electrolytic current-density conditions.
Subsequently, the effect of the electrolytic time on the pattern structure was investigated.In this study, the pattern depths of the Cu surfaces treated for different electrolytic times were compared under a uniform electrolytic current density.The atomic force microscopy (AFM) images shown in Figure 5e-1,f show that a shallow pattern with a depth of ≈20 nm was obtained on the Cu surface after 10 s of electrolysis.In contrast, as shown in Figure 5e-2,f, a clear pattern with a depth of ≈50 nm was obtained on the Cu surface after a longer electrolytic time.The thickness of the Cu film was ≈50 nm; therefore, the surface of the SiO 2 substrate was exposed at the bottom of the trench.Unlike conventional ECM using a liquid electrolyte, the pattern width and pitch are nearly independent of the depth of the pattern, which is evident when the Cu surface is etched at the solid interface between the PEM stamp and the Cu surface.
To demonstrate the applicability of our technique for fabricating various types of patterns, the same procedure was used to fabricate line and space (L&S), shark skin, and micropillar array patterns on Cu surfaces using PEM stamps.As shown in Figure 1c, L&S (Figure 1c-3), shark skin (Figure 1c-1), and micropillar array (Figure 1c-2) patterned Cu structures were achieved with pattern resolutions ranging from several hundreds of nanometers to several micrometers.These results show that the anodic dissolution proceeds only at the contact area between the PEM stamp and Cu, thus allowing the patterned structure of the PEM stamp to be transferred to the Cu surface with high resolution.In addition, the PEM stamp can be used to form the pattern on Cu surfaces for multiple times (Figure S3, Supporting Information).However, as shown in Figure S3c (Supporting Information), the fifth use of the PEM stamp caused a slightly nonuniform pattern formation, which can be attributed to the decomposition of PEM stamps during the electrolysis, as discussed in Section 2.2.
Finally, to demonstrate the applicability of our technique for fabricating patterns over a large area, a pattern was formed on the Cu thin-film surface using the roll-to-plate method.In this method, an automated stage was used to move the material and a rotating roll electrode (cathode) wound with the PEM stamp was used for the processing, as illustrated in Figure 6a.The PEM stamp was attached to the roll using electroconductive doublesided adhesive tape.As shown in Figure 6b, the treated Cu thin film surface exhibited structural color over a wide area of 50 × 30 mm 2 , which was attributed to the uniformly arranged microstructures.The SEM image (Figure 6c) shows that a lattice pattern with a pitch of ≈2 μm was formed on the Cu thin film surface.These results suggest that roll-to-plate patterning of the Cu thin-film surface is also feasible.Therefore, by exploiting the flexibility of the PEM, the proposed patterning process applies to large-area processing methods such as roll-to-plate.

Conclusion
This study proposes a novel electrochemical approach for directly forming patterns on Cu surfaces.Using a PEM stamp with a pattern, the Cu surface in contact with the PEM stamp was selectively anodically dissolved by the electrochemical system of Cu(anode)/PEM/cathode without the use of a liquid electrolyte, and a pattern with a resolution of several hundred nanometers was formed.Analysis of the PEM cross-sections after use for patterning demonstrated the ionization of Cu and ion transport from the anode to the cathode.Dried PEMs are suitable because they have a minimal effect on water electrolysis and help achieve high patterning resolution.The adhesion of PEM components on the Cu surface and the lifetime of the PEM stamp remains an issue for industrial applications.Therefore, it is necessary to clarify the optimum conditions, such as electrolytic conditions that do not cause the decomposition of PEM components, in the future.Process conditions are also crucial for improving pattern resolution.The proposed patterning technique is applicable to roll-to-plate processes and enables large-area patterning.Furthermore, the proposed technique, involving the anodic dissolution of the Cu surface using a PEM, is an environmentally friendly and cost-effective approach as it provides a facile direct imprinting process without the need for resists or harsh chemicals.Moreover, this technique can improve processing efficiency as micro/nanopatterns can be formed on the Cu surface in a single step, without the need for multiple complex processes.

Experimental Section
Preparation of PEM Stamps: The PEM used in this study was a PFSA membrane (Aquivion, E98-09S, SOLVAY) with a thickness of 90 μm.The PEM stamps were prepared via hot embossing, [51,52] in which the patterned structures from commercially available master molds (DTM2-2, Kyodo International Inc. and SharkletShield Germ Barrier Panels 5″ × 7″ Pack of 10, Sharklet Technologies, Inc.) were transferred onto the PEM surface by pressing the PEM onto the mold surface under the pressure of 5 MPa at a temperature of 135 °C.After pressing for 6 min, the PEM and mold were cooled to 80 °C while keeping the pressure applied.Finally, the PEM was released from the mold to obtain the PEM stamp.As shown in Figure 1b, PEM stamps with L&S (Figure 1b-3), shark skin (Figure 1b-1), and micropillar array (Figure 1b-2) structures, with pattern resolutions ranging from several hundreds of nanometers to several micrometers, were achieved.
Pattern Formation on the Cu Surface Via the Direct Electrochemical Imprinting Process: The Cu sample prepared using vacuum evaporation on a glass substrate with a thickness of ≈50 nm, was subjected to liquid-free electrolysis with the prepared PEM stamp.As illustrated in Figure 1a-1, the PEM stamp was sandwiched between the Cu sample (anode) and cathode to form an electrochemical system.The electrolytic current and bias voltage were controlled and monitored using a potentiostat (HZ-7000, Hokuto Denko Corp.) at room temperature (≈25 °C).Owing to the absence of a liquid electrolyte in the electrochemical system, a reference electrode was not used.A conductive rubber sheet was used as the cathode to allow uniform contact between the PEM stamp and the Cu sample.Dried and water-swollen PEM stamps were used for electrolysis, and their patterning characteristics were compared.The effects of the electrolytic current density and applied pressure between the PEM stamps and the Cu sample on pattern formation were investigated.
Characterization: The chemical composition of the PEM after electrolysis was investigated via elemental mapping using SEM (SU6600, Hitachi High-Tech Corp.) and EDX (INCA X-act, Oxford Instruments).To investigate the electrochemical characteristics during electrolysis, LSV, CV, and CP were performed using a potentiostat.Electrochemical experiments were performed at room temperature and atmosphere using a twoelectrode electrochemical system consisting of an anode/PEM/cathode sandwich, with Cu as the anode (working electrode) and Pt as the cathode (which serves both the counter and the reference electrode).The voltage range was 0-20 V for LSV and −5 to 5 V for CV, and the scanning speed was 50 mV s −1 for both.CP was performed at a constant current density of 1 mA cm −2 .The chemical states of the treated Cu surfaces were characterized using XPS (SCIENTA SES2002, Scienta Omicron) with Al K source (1486.6 eV) and XAFS.The XAFS measurements of the Cu surfaces were performed at BL-8 of the SR Center, Ritsumeikan University, Japan, equipped with a grazing incidence monochromator with a varied line-spacing plane grating.The XAFS spectra of the samples were collected using the partial electron yield (PEY) through a micro-channel plate detector with retarding grids.The PEY spectra contained information near the surface (less than 1 nm).The measurements were performed at room temperature in an ultrahigh vacuum of ≈1 × 10 −7 Pa.The incident angle of the synchrotron radiation light with respect to the surface normal was set to 0°.The surface structures of the PEM stamp and treated Cu surface were characterized using optical microscopy and SEM.MCT (MCT-510, Shimadzu Corp.) was conducted to analyze the mechanical deformation of the PEM stamp.Furthermore, the patterned Cu surfaces were evaluated using an AFM technique (L-trace II, SII Nanotechnology, Inc.) over areas of 30 × 30 μm 2 to analyze the microscopic surface topography and roughness.

Figure 1 .
Figure 1.Schematics of the direct electrochemical imprinting process.a-1) The Cu (anode)/polymer electrolyte membrane (PEM)/cathode electrochemical system.a-2) Selectively etching Cu surfaces in contact with the patterned PEM (PEM stamp).a-3) Pattern structures are formed on the Cu surface without using a liquid electrolyte.Scanning electron microscopy (SEM) images of the PEM stamp containing b-1) shark skin, b-2) micropillar array, and b-3) line and space (L&S).The inset shows the photograph of the PEM stamp.SEM images of the Cu surface treated via the direct electrochemical imprinting process using a c-1) shark skin, c-2) micropillar array, c-3) L&S patterned PEM stamp.The inset shows a photograph of the pattern of Cu.

Figure 2 .
Figure 2. a) SEM image of the PEM cross-section after the direct electrochemical imprinting process.b) Energy-dispersive X-ray spectroscopy (EDX) chemical analysis of the PEM cross-section after electrolysis: b-1) Cu, b-2) F elemental maps, and b-3) EDX spectra emitted from points 1 and 2 in (a).

Figure 3 .
Figure 3. a-c) Electrochemical analysis results of PEM/Cu etching: (a) Linear sweep voltammetry (LSV) results for water-swollen and dried PEM, and (b) Cyclic voltammetry (CV) results under dried PEM conditions.Effect of water-swollen and dried PEM.(c) Chronopotentiometry (CP) curves during electrolysis.The electrolysis was intentionally stopped after a voltage of 10.8 V was reached, and the corresponding duration is defined as T e .Microcompression testing (MCT).d) Comparison of the water-swollen PEM and dried PEM.Optical microscopy images of the Cu surface treated by the direct electrochemical imprinting process with e-1) water-swollen PEM and e-2) dried PEM.f-1, f-2) SEM images of (e-1) and (e-2), respectively.

Figure 4 .
Figure 4. a-e) X-ray photoelectron spectroscopy (XPS) spectra emitted from Cu surfaces treated under different processing conditions (comparison of the untreated surfaces, treated with water-swollen PEM, and treated with dried PEM): (a) wide scan, (b) Cu 2p, (c) F 1s, (d) C 1s, and (e) O 1s spectra.f-h) X-ray absorption fine structure (XAFS) spectra from Cu surfaces treated under different processing conditions (comparison of the untreated surfaces, treated with water-swollen PEM, and treated with dried PEM): (f) Cu L-edge, (g) O K-edge, and (h) F K-edge.

Figure 5 .
Figure 5. a, b) Effect of the processing pressure on the patterning performance of Cu.(a) CP curves and (b) optical microscopy images of Cu surface patterned with pressures of b-1) 3.8, b-2) 10, and b-3) 17 MPa.The inset depicts a magnified image.c, d) Effect of the current density of patterning on the etching rate of Cu.(c) CP curves and (d) etching rate and Faradaic efficiency as a function of current density.e, f) Effect of the processing time on the height of pattern formation.Atomic force microscopy (AFM) observation of the treated Cu surface: (e) 30 × 30 μm 2 surface images after processing for (e-1) 10 s and (e-2) 95 s.(f) Cross-sectional profile along the A-A′ line in (e-1) and (e-2).

Figure 6 .
Figure 6.Demonstration of large area pattern fabrication using roll-to-plate process.a) Schematic of the roll-to-plate process.b) Photograph of the Cu surface showing the structural color in the treated area.c) SEM image of the pattern formed on the Cu surface.