Graphene nanowalls grown on copper mesh

Graphene nanowalls (GNWs) can be described as extended nanosheets of graphitic carbon where the basal planes are perpendicular to a substrate. Generally, existing techniques to grow films of GNWs are based on plasma-enhanced chemical vapor deposition (PECVD) and the use of diverse substrate materials (Cu, Ni, C, etc) shaped as foils or filaments. Usually, patterned films rely on substrates priorly modified by costly cleanroom procedures. Hence, we report here the characterization, transfer and application of wafer-scale patterned GNWs films that were grown on Cu meshes using low-power direct-current PECVD. Reaching wall heights of ∼300 nm, mats of vertically-aligned carbon nanosheets covered square centimeter wire meshes substrates, replicating well the thread dimensions and the tens of micrometer-wide openings of the meshes. Contrastingly, the same growth conditions applied to Cu foils resulted in limited carbon deposition, mostly confined to the substrate edges. Based on the wet transfer procedure turbostratic and graphitic carbon domains co-exist in the GNWs microstructure. Interestingly, these nanoscaled patterned films were quite hydrophobic, being able to reverse the wetting behavior of SiO2 surfaces. Finally, we show that the GNWs can also be used as the active material for C-on-Cu anodes of Li-ion battery systems.

For the most part, graphene films are produced with the basal plane parallel to a substrate.However, in recent years, the study of nanoscaled graphitic carbon sheets with basal planes oriented vertically to a surface, also called graphene nanowalls (GNW s ) or vertical graphene (VG), has drawn increased attention from the community [12][13][14][15][16][17].The films of GNW s commonly aggregate large numbers of intertwined sheets, where each nanowall is composed of one to a few tens of wavy graphene layers that extend laterally for hundreds of micrometers [15,18,19].Given their orientation, both the bottom and top surfaces of the vertically-aligned graphene stack are exposed which, added to its cross-sectional surface (i.e. the graphene edges), results in large areas with a considerable number of reactive sites [20][21][22].Hence, it is understandable that GNW s offer excellent performance in various types of sensors (gas, bio, molecular, etc) [22][23][24][25].In Nanotechnology Nanotechnology 35 (2024) 085602 (11pp) https://doi.org/10.1088/1361-6528/ad0a0d* Author to whom any correspondence should be addressed.
Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence.Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.addition, the films of GNW s are mechanically and chemically resilient enabling their use in devices subjected to continuous stress (e.g. in flexible supercapacitors) [26][27][28].Other applications described in the literature include flat panel displays and light sources, due to the high density of sites for electron field emission [29].Presently, the key driver to develop GNW s is their role as anode materials in intercalation-type batteries.This is reasoned by the graphene basal planes being correctly oriented for a facile flow and uptake of charged ions (such as lithium) from the electrolyte to the anode [18,21,[30][31][32].
The production of GNW s films rely on plasma-enhanced chemical vapor deposition (PECVD) methods, where carboncontaining gases act as precursors [33][34][35].In addition to the variable parameters generally associated with the chemical vapor deposition (CVD) process (carbon feedstock, gas flows/ratio/pressure, reaction temperatureK), the plasma parameter space needs to be fully explored in PECVD systems to optimize VG production.Specific characteristics ranging from AC (Alternating Current) to DC (Direct Current) plasmas [21,36], power ratings from hundreds to thousands of watts [37,38] and frequencies windows from kHz (radio) to GHz (microwaves) [39,40] have been investigated to optimize the production of GNW s films.While such a rich landscape of process parameters implies a myriad of ways to run the deposition reaction, procedures that reduce cost and complexity should be favoured.In this respect, a DC plasma activated at low powers (few hundred watts) with the reactor running at reduced temperatures (less than 700 °C), represents a fairly balanced option.While some teams have claimed to work under the aforementioned conditions [28,34,41], the sample scale remains small in these work, and did not present reproducible wafer-scale production.Furthermore, even with applied power ratings pushed to more than 1 kW, the height of their GNW s films remained below one hundred nanometers.
A critical component of the GNW s growth by PECVD is the substrate selected.So far, along with transition metal foils (Cu, NiK), quartz and SiO 2 plates have been the preferred choice [42][43][44][45][46]. Carbon filaments have also shown some promising results, particularly when using a roll-to-roll approach to enable continuous scaled-up production of GNW s [47,48].Though the variety of substrate materials is clear, they all have in common the morphology, i.e. they are either shaped as foils/plates or filaments [33,[48][49][50].Irregular substrates such as patterned surfaces have seldom been explored.In one example, GNW s were grown on a patterned gold film and then integrated in field-effect transistors that acted as sensors for gas molecules [51].These nanowalls can also grow on patterned carbon nanotubes and applied in electrodes for micro-supercapacitors [12].The caveat with patterned substrates is the added cost to prepare them and the likelihood of a lower production yield and film uniformity.Taken together, a PECVD process that enables the production of high yield, uniform GNW s films at large scale (>2″ wafer), all with a relatively low cost/complexity and making use of trivial, but diversely-shaped substrates, would be very desirable.
The present paper describes a straightforward PECVD method for the large-scale growth of GNW s films on meshed copper substrates.The films can be transferred onto various surfaces/substrates without disturbance to their morphology or structural properties.

Growth and transfer of the graphene nanowalls
The GNW s were grown using a cold-wall DC-PECVD reactor (Black Magic Pro 4', Aixtron GmbH).Its chamber is characterized by the presence of a showerhead (figure S1(a)), a set-up that enables a more uniform mix and delivery of gases to the substrates (which are placed underneath).
High-purity methane (CH 4 ) was used as the carbon precursor, with hydrogen (H 2 ) and argon (Ar) acting as the etching agent and carrier gas, respectively.Square-shaped Cu foils and meshes (Alfa Aesar, 50 μm thick, Puratronic, 99.9999%), with equal lateral dimensions, were employed as wafer-scaled catalytic substrates.Ranging from 1 to 40 cm2 , the foils were continuous while the meshes had an array of 50 μm wide square-shaped cavities and a thread made of 0.11 mm diameter copper wires (figure 1).Although similar results were obtained whatever the substrate size used, only the 3 cm 2 samples are described because these were more convenient to illustrate the energy storage application.Before placing them in the reaction chamber, all substrates were cleaned ultrasonically by sequential 15 min immersion steps, in acetone, isopropanol and deionized (DI) water, and dried with a forced nitrogen flow.
In a typical growth run, a clean Cu substrate was first placed on the bottom graphite heater (figure S1).After purging the chamber with Ar and vacuum (10 −3 mbar), the substrate was quickly heated to 800 °C for 5 min, under a mixed flow of Ar at 160 standard cubic centimeters per minute, (sccm) and H 2 at 200 sccm.This heating (or annealing) step serves in cleaning the substrate from residuals, removing copper oxide particles, reorganizing the copper grains and smoothening the surface for graphene nucleation and growth [52,53].As schematized in figure S2, the temperature was then dropped to 600 °C and methane (CH 4 ) was introduced, at a rate of 60 sccm, to the Ar/H 2 gas mixture.Also, the DC plasma was activated at 150 W (voltage of 800 V) for 20 min.These parameters were selected after several preliminary optimization experiments to maximize the growth uniformity over the mesh.The effects of these process parameters on GNWs can be investigated which may lead to enhanced GNWs growth [23,47,54].The total pressure in the chamber was kept at 6 mbar.Next, the flow of reactive gases was stopped and the temperature in the chamber brought down to 200 °C, under 1000 sccm of Ar.Past this point, the chamber was vented, and cooled down to room temperature soon after.
To transfer a GNW s film from the Cu foil/mesh to a target surface (e.g.SiO 2 /Si wafers), the catalytic metal substrate had to be eliminated.The method followed was similar to that developed for nanorange graphite films, previously reported by our group [55,56].Briefly, the Cu was etched overnight with a FeCl 3 solution, which was then carefully diluted, and extracted, until it had been completely replaced by DI water (figure S2).The carbon films did not crumple and were left floating on the water, ready to be picked up using the target surface.

Characterization of the graphene nanowalls
The GNW s films were analyzed with a set of techniques that unveiled their morphological, structural and chemical characteristics.Upon taking the samples from the PECVD reactor, their general macroscopic appearance was recorded with a digital photographic camera.The topography and surface roughness of the films were studied using non-contact atomic force microscopy (AFM, Dimension Icon, Bruker).A closer look at the sample morphology, in plan-view and crosssection, was carried out with scanning electron microscopy (SEM, FEI NovaNano 450, operated at 5 kV).Double Cs corrected transmission electron microscopy (TEM), at Ther-moFisher TitanThemis 60-300 (300 kV, equipped with a high-brilliance field emission gun and a Wien-type monochromator) is used for the microstructures study of the nanowalls.Plan-view TEM sample is acquired by etching the Cu substrate with FeCl 3 solution, washing it with H 2 O, and then transferring it on lacey-carbon Au TEM grids.
Additional structural information was gathered with Raman spectroscopy.The optical images and spectra were recorded with a confocal micro-Raman workstation (Alpha300RA, WITec GmbH) operated at a laser wavelength of 532 nm and low excitation power (25%, to avoid heatinduced effects on the carbon film).The Raman spectral maps were acquired with 200 point-spectra per line over areas of hundreds of μm 2 .
The surface chemistry of the films was analyzed with x-ray photoelectron spectroscopy (XPS, Amicus Kratos Analytical), after transferring the samples to SiO 2 /Si wafers.
The contact angle measurements were made with KRÜSS Drop Shape Analyzer 100, with the contact angle fitted using the ellipse tangent method.

Results and discussion
The growth of GNW s films with PECVD is strongly dependent on the size and geometry of the substrate.Here, Cu was used in the form of a mesh or a foil (figure 1).Operating the reactor with a DC plasma of 150 W (under the conditions described in the Experimental section), the entire Cu mesh was covered with carbon (outset of figure 1).By contrast, on the metal foil, the GNWs deposition was limited to its edges (away from these, some residual carbon was observed but this was later identified as predominantly amorphous).Thus, experimental parameters including electric field intensity in DC plasmas and chamber temperature are insufficient to homogeneously grow GNW s on Cu foils.The outcome is remarkably different but understandable.In the meshed substrates, there is a higher density of metallic edges which tend to concentrate electric fields.These are instrumental in promoting and directing the growth of the graphene nanosheets.The same field concentration happens in the foils but it is restricted to the edges of the slab.In fact, this edge effect and its directional capability, were previously explored in the growth of aligned carbon nanotubes [38,57,58].
The visual inspection of the Cu substrates, after the PECVD procedure, identified the deposition of carbon on the bottom side of the two Cu substrates.However, to understand if these films were structured as GNW s , a number of characterization tools had to be used.In the following, the analysis of the Cu foils was focused on the edges since GNWs were not observed in the planar surface while, in the mesh, several locations of GNWs covering the entire surface were probed to ensure consistency.
The topography AFM images in figure 2 show shapes that resemble entangled filaments in the two samples.From the literature, these geometries are what one would expect for GNW s films, as they refer to the top of the curved walls [13].Still, the dimensions and density of these features are clearly different in the mesh and foil samples.To quantify the regularity of the walls' height, profiles (about 2 μm long) were recorded along the red lines marked in figures 2(a) and (c).At the sub-micrometer scale, the surface roughness of the GNW s films was notably smaller in the Cu foil since the walls are shorter and closer to each other.However, when considering longer scales (2 μm), there was more irregularity in this sample.The AFM data is corroborated by plan-view images taken with SEM (figure 3).Similar 'filamentous' shapes are seen in figures 3(a) and (c), with those in the Cu foil edges being shorter and denser.As opposed to the topography analysis, surface undulations are not perceptible in plan-view SEM.Figures 3(b) and (d) show the cross-section of the mesh and foil films, respectively.Interestingly, the top surface appears to be more regular for the mesh, which contradicts the AFM analysis.Besides issues regarding sample preparation and homogeneity, the density of carbon across the film thickness may explain this.In the mesh sample, the film has the appearance of a solid block, denoting a higher mass of deposited carbon.This would significantly lower the transparency and hamper the qualitative comparison of surface roughness using the two techniques.Overall, the average height of the GNW s films was similar in the two samples and within the range of 270-300 nm.
The Cu mesh offers homogeneous growth over the surface, which means that the walls are better structured standing vertically on the surface and growing uniformly.However, on the Cu foil substrate, the electric field is much denser at the edges (a few micrometers from the borders), in limited areas of the substrate.The higher plasma energy is expected to increase the substrate temperature with enhanced decomposition and ionization rates of the precursor gas molecules generating condensed gas ions in this region.This leads to increased GNWs density affecting the direction of the walls and altering their verticality on the surface.In Cu foils, the walls are randomly oriented from the surface while in Cu meshes the walls are more uniform.
In fact, the formation of the crystalline microstructure starts from the interface, where initial horizontal graphene layers are formed gradually at an early stage of the growth [27].These domains are developed separately on the surface and are often collapsed with borders mismatch to form nucleation sites.Further growth at these coalescence points with the presence of defects, layers deflection and edges mismatches will lead to the vertical bending of one domain and initiate the vertical nanowalls growth.Hence, strengthening the electric field will intensify the decomposition rate of the hydrocarbon gases, enhances the growth rate and increase the walls density in the film [59].The electric field is much higher in Cu foil sharp edges compared to the curved wires surfaces in mesh where the main difference in walls density is observed for Cu foil and meshed substrates.
The AFM and SEM data sets are consistent and identified the presence of the nanowalls, albeit with different characteristics in each sample.For instance, thicker and longer walls are present in the mesh sample.In any case, we can assign the black deposits in the substrates of figure 1 to structured GNW s films.
At this point, an overview of the samples' morphologies was obtained.To understand better these materials, a more localized structural analysis is required.In this respect, TEM images and diffraction patterns could shed light on phenomena such as variations in wall thickness, their separation, shape and the type of carbon that constitutes them.Fortunately, given the transfer process developed, it was possible to extract the GNW s films and mount them on TEM grids.In figures 4(a) and (b), low magnification plan-view TEM images are seen for the films originating from the Cu mesh and foil, respectively.It is apparent that the walls in the mesh result in higher mass-thickness contrast than those of the foil.Interestingly, the corresponding diffraction pattern (inset) unveils that GNW s grow in some orientational preference ordering in which the 0002 planes appear as opposing arcs (not a ring) compared to the foil.For the foil´s film, the walls are shorter and there is no preferred orientation (the inset diffraction pattern shows a more uniform brightness of the 0002 ring).At high magnification (figures 4(c) and (d)), it was possible to distinguish well, and measure, the cross-section of the walls.As previously observed, the walls in the mesh film are thicker.
A more detailed analysis of the wall thickness in the mesh sample was performed.Figure 5(a) shows how the thickness changes within the same wall.If we consider that we are viewing the wall from above, then this tapering would occur sideways.The number of layers is higher at the center of the wall and decreases gradually towards the edge.It is likely that this applies equally if we considered the 'root-to-top' orientation (though this would require highly resolved images of the film's cross-section).The histogram in figure 5(b) (built from sampling more than 100 different walls and wall sections), shows that the thickness ranges from 5 to 70 nm, with the statistical mode located at 14 nm.Similar to other reports in the literature [13,60], it was possible to identify few-layer graphene sections at the wall's edges (though these are more challenging to image in plan-view conditions).
One advantage of plan-view TEM images is the possibility to undertake structural studies using either diffraction patterns or HRTEM image processing tools.Given the different ways that carbon can arrange as a solid, plan-view HRTEM images and corresponding fast Fourier transform (FFT) patterns were used to study local variations in structure and lattice ordering of the mesh GNW s film (figure 6).Two different structures were identified: graphitic carbon (GC) and turbostratic carbon (TC).The FFT of the entire figure 6  with slightly different diameters, both attributed to the (0002) d-spacing of hexagonal graphite.In an FFT, a perfect (0002) graphene layer should appear as a single pair of spots.However, due to the non-uniform orientation distribution of the nanowalls and the number of stacked graphene layers, the spots will merge into arcs or, eventually, rings.The augmented view and FFT of two selected areas in figure 6(a) (marked with dashed squares) are presented in figures 6(c)-(d).
From these results, it is clear that GC and TC co-exist in the films.The GC corresponds to a well-graphitized carbon section, confirmed not just by the HRTEM image but also by the narrow d 0002 peak in the FFT (the latter suggesting less lattice strain in the GC sections).By contrast, the other selected area must have more defects, as evidenced by the broader and longer 0002 arcs.Thus, the GC areas have smaller spacing d 0002 (3.38 Å) than the TC ones (3.44 Å).Note that larger d 0002 for TC, ranging from 3.44 to 3.52 Å, were observed.In fact, these d 0002 are highly correlated with the section of the wall probed, whether near an edge or closer to the wall center or its root.
In addition to the localized structural data provided by TEM, Raman spectroscopy allows access to similar information for larger sampling areas.The study of GNW s films by Raman has been reported in the literature and some features are known to be typical such as the presence of a D′ peak and the overtone D+D′ [43]. Figure 7 shows the Raman spectra for the as-produced GNW s films, i.e. in the presence of the Cu mesh or Cu foil.The lower frequency peak refers to the D band (at 1350 cm −1 ), originating from a vibrational mode symmetry break that is associated with lattice vacancies, grain boundaries and heteroatoms in graphene sheets.Next, coupled to the graphite peak or G band (at 1580 cm −1 ), a shoulder is present that relates to the D′ (at 1620 cm −1 ).Further to this, three secondorder peaks are observed at 2700 cm −1 (2D), 2930 cm −1 (D +D′) and 3240 cm −1 (2D′).The presence of the D′ and its overtones are evidence of GNW s films.Considering the I D /I G ratio, which is used to estimate the crystallinity of graphene nanosheets [43], this was 2.14 for both samples.However, the peaks were generally more intense and resolved in the mesh sample, the main difference being the 2D.In fact, the I 2D /I G ratio is higher in the mesh film (approaching 1), which implies more structured and graphitized carbon layers in this substrate [61].Moreover, the proportion of the GC component in the walls of the mesh film is likely higher than the TC component.For the foil sample, this ratio is expectedly smaller, bearing a higher proportion of TC.
To understand further the microstructural characteristics and uniformity of the GNW s mesh films, Raman mapping was performed for the VG sheets grown on Cu mesh (area highlighted by the green square in figure 8(a)).The intensity of the G, 2D and D+D′ peaks was tracked and measured at a rate of 200 point spectra per line.Upon using an energy window at the locations of those peaks, the respective maps were extracted (figures 8(b)-(d)).The spatial correlation of the three peaks confirms the wider-area growth of the nanowalls, these being of similar thickness and structure.Overall, the structural analysis confirmed the presence of the GNW s in the two samples.Under the same growth conditions, the yield differed considerably, being higher in the Cu mesh samples.Although with similar heights, the films differed in thickness as well as length and orientation of the nanowalls.Critically, variations in the degree of crystalline order were identified.In fact, it was possible to isolate neighboring areas where the predominant type of carbon changed from graphitic to turbostratic.
Wet chemical transfer processes employed to transfer the carbon films grown on catalytic substrates may, unintentionally, introduce contaminants in the samples [62].In the present work, the chemical wet transfer of the GNW s film was realized without polymer support.Similar polymer support free transfer of graphite thin films (∼100 nm) reported by us recently [55].The Cu grid etching was performed in a FeCl 3 /DI water solution (10 ml/ 100 ml mixture) (figure S3).To assert the purity of the transferred samples and their structural stability, XPS analysis was carried out to explore the chemical composition and quality of the GNW s grown on Cu mesh and transferred into Si/SO 2 wafer.The survey XPS spectra shown in figure 9(a) refer to the GNW s films after being transferred onto Si/SO 2 wafers.Only three elements were identified, namely Si, C and O, attributed to the wafer and the carbon film.To assert that the film's structure was not modified during this process, a high-resolution XPS spectrum was acquired for the mesh GNW s (figure 9(b)).The deconvoluted C 1s peak indicates the presence of two types of carbon bonds, C=C (sp 2 ) at 284.7 eV and C-C (sp 3 ) at 285.6 eV.In addition, a residual of C-O bounds (286.9 eV) was seen.The same spectrum was observed before the transfer process.Therefore, XPS technique analysis indicates that the transfer process did not lead to sample contamination at the surface of the sample.However, a deeper analysis of the film quality by electron energy loss spectroscopy (EELS) technique shows small traces of Fe contaminants interspersed at the bottom of the GNW s .This will be addressed in detail in a follow-up study.
The wettability of carbon films is an important parameter in electrochemical applications, particularly when these films are used as electrodes in cells operated with a liquid electrolyte.GNW s films are known to be hydrophobic [21].Still, this may be manipulated since there is a correlation with surface roughness [63].For instance, Li et al reported that the roughness of GNW s films could be increased with longer growth times, plasma power or processing temperatures, thereby enhancing their hydrophobicity [23].These authors measured a water contact angle of 84°for a film produced with a plasma of 250 W [23].
In our study, the GNW s films grown on mesh, at the relatively low power of 150 W, exhibited a water contact angle of 118°(figures 10(a) and (c)).Interestingly, this hydrophobic behavior took place despite the array of holes in the films and the presence of the hydrophilic SiO 2 surface (with a contact angle of 37°) underneath the carbon film.Thus, despite the open mesh and being only 300 nm thick, the GNW s film can act as a hydrophobic coating.
The electrochemical performance of the meshed GNW s film was tested in Li-ion batteries as illustrated in the supplementary file (figure S4).The preliminary results demonstrate possible use of the as-grown GNWs film in Li-ion batteries that might lead to low-cost and enhanced energy storage systems.

Conclusions
In this study, a growth method of graphene nanowalls films on meshed copper substrates was developed through the lowtemperature PECVD technique.The meshed metal substrates enhance the growth of GNW s in PECVD and offer the possibility of controlled film morphologies.This production route is less expensive and simpler to fabricate than the current methods based on patterning catalytic substrates in clean room processes.For the same experimental processing parameters, meshed substrates offer the possibility of a homogeneous growth of GNW s at smoother conditions compared to metal foils, which require higher temperature, plasma power and growth durations.Meshed films offer increased surface roughness, and thicker nanowalls with both graphitic and turbostratic structures, that lead to enhanced hydrophobicity.It has been shown that these films are promising and could be used in many fields such as electrochemical and energy storage applications.Further parametric study would lead to enhanced growth parameters and optimized films structures to be more suitable for other applications such as gas sensing and miniaturized electronic devices.

Figure 1 .
Figure 1.Photographs of the GNW s grown uniformly on the Cu mesh and the edges of the 3 cm 2 Cu foil substrates (a).SEM images showing the morphology of the Cu mesh used for the growth and a portion of the grown GNW s highlighted with the red boxes (b).

Figure 2 .
Figure 2. AFM topography of the grown GNW s on Cu mesh (a)-(b) and foil (c)-(d) and the corresponding cross-section profile indicated by the red line respectively.

Figure 3 .
Figure 3. Plan and side views SEM images of the GNW s film grown on Cu mesh (a)-(b) and foil (c)-(d) under the same conditions.

Figure 5 .
Figure 5.The high-resolution TEM image of the GNW s grown on Cu mesh (a) with the distribution of the walls stacking thickness (b).The average of the GNW s thickness is ranging between 5 and 20 nm.

Figure 6 .
Figure 6.High-resolution TEM image of GNW s grown on Cu mesh (a) and corresponding FFT showing two rings of (0002) GC and (0002) TC (b) that correspond to the two tapered structure types of carbon highlighted in the dashed squares in (a), indicating both graphitic (c) and turbostratic structures (e).The local HRTEM and FFT of (c)-(d) GC and (e)-(f) TC are shown.The interplanar distances for GC TC are 3.38 Å and 3.44 Å, respectively.

Figure 7 .
Figure 7. Raman spectra of the GNW s grown on Cu mesh and foil at the same conditions.

Figure 8 .
Figure 8. Raman mapping of G, 2D and D+D′ peaks intensity obtained from the green box area highlighted in the optical image shown in (a).

Figure 9 .
Figure 9. XPS characterization of the GNW s grown on Cu mesh and foil transferred to Si/SiO 2 wafer (a) and typical high-resolution C 1s spectra of the GNW s (b).

Figure 10 .
Figure 10.The water contact angle obtained at the surface of the GNW s on SiO 2 /Si wafer presented in (a) without presence of the GNW s (b) and with GNW s film (c).