Recent Advancements on the CVD of Graphene on Copper from Ethanol Vapor

Chemical vapor deposition (CVD) on metal foils is regarded as the most effective method to produce large-area graphene with properties in line with the requirements of an extensive range of electronic and optical applications. For the CVD of graphene, ethanol is a versatile carbon source alternative to the widely used methane. In this review we report on the current progress in this field showing how the CVD parameters can be modulated to gain full microstructural control on graphene grown on Cu foils. Using ethanol vapor, graphene can be grown as a continuous film with mono- or multi-layer thickness, and also in the form of isolated crystals on pre-oxidized Cu substrates. Overall, ethanol-CVD allows a controllable degree of crystallinity and tunable electrical/optical characteristics in the grown samples. In turn, this control translates into a superior versatility for device design and related applications.


Introduction
In hindsight, the early observations of atomically thin graphite crystals grown onto metal surfaces exposed to hydrocarbons at high temperature probably did not gain due attention from the scientific world [1]. Carbon-based materials and nanomaterials have since gained considerable interest both for research and industrial applications, and they are now at the core of technological innovation in many fields. Graphene, the basal plane of graphite, was first isolated as two-dimensional material in 2004 [2]. In the following years, graphene has captured the attention of researchers thanks to its electronic and physical properties, such as an extraordinarily high charge carrier mobility, mechanical strength, and chemical stability [2][3][4][5][6]. Graphene was originally obtained in the form of individual and small samples via the "scotch-tape" exfoliation of graphite [5]; nowadays, several advanced techniques have been devised to produce it in large quantities and with reproducible characteristics. In addition to mechanical and chemical exfoliation [2,5,7,8], and to epitaxial growth on SiC [8][9][10][11], chemical vapor deposition (CVD) emerged as the most reliable and cost-effective approach to make graphene in a form applicable to electronics [10,[12][13][14][15]. Thermal CVD offers the possibility to tune the growth process by exploring extensive combinations of parameters, such as substrate composition, structure and surface; precursors, gas type, pressure and flow; temperatures; cooling and heating rates; etc. Suitable combinations of parameters can be designed to target the production of graphene with specific properties (e.g., number of layers, domain size, etc.), over large areas, and in manners compatible with the direct fabrication of vertical and lateral heterostructures [16][17][18]. In specific cases, direct CVD growth C 2020, 6, 14 3 of 13 described by one Lorentzian component but with a larger linewidth (FWHM~45-60 cm −1 ) and with a frequency upshifted with respect to the 1L graphene [45,46].
Here we provide a concise review of the most important aspects of the graphene growth by ethanol-based CVD on Cu. We offer a summary of the effect of the CVD parameters on the structural properties of graphene, by investigating in detail Raman spectroscopy data. We complement our analysis with data from other analytic techniques, such as optical microscopy, Raman spectroscopy, atomic force microscopy and scanning electron microscopy (AFM, SEM). This review paper is organized in two main sections. The first section concerns the growth of continuous graphene films with quality comparable to those made from methane-CVD, focusing on the role of process temperature, gas flow, and time. In this context, one of the key features of the ethanol-CVD, the possibility of using ultra-fast growth process, will be presented and described. In the second section, the focus will shift to the growth of single-crystal graphene grains, with lateral size in the mm-range and beyond. Here, a comparison will be made between results obtained with methane and ethanol, to provide a perspective on a topic of interest for several nanotechnology applications.

Tuning of Growth Parameters
The first attempts to grow graphene by ethanol-CVD explored temperature ranges lower than those usually reported for methane-CVD (<1000 • C) [39,40,49], in order to lower the overall process costs. Miyata et al. synthesized high-quality monolayer graphene from ethanol on Ni foils at 900 • C and atmospheric pressure [49]. They introduced a "flash cooling" to reduce the duration of the standard cooling process and avoid the decomposition of the synthesized graphene [49].
Following these preliminary results, other researchers explored wider temperature ranges (650-850 • C) (Figure 1a) [40]. Guermoune et al. grew graphene from several liquid alcohols, such as methanol, ethanol and 1-propanol. Their goal was to improve the alcohol-based growth methods and obtain large-area, continuous graphene films with quality comparable to that obtained from methane-CVD [40]. By monitoring the D band in the Raman spectrum of the grown samples [44,45,47], the group found that the improvement in quality (in terms of crystallinity) strongly depended on the temperature: increasing the temperature from 650 to 850 • C, the graphene crystal domain size kept increasing in size, up to 168 nm ( Figure 1a) [40].
C 2019, 5, x FOR PEER REVIEW 3 of 12 layers, the 2D-peak is again described by one Lorentzian component but with a larger linewidth (FWHM~45-60 cm −1 ) and with a frequency upshifted with respect to the 1L graphene [45,46]. Here we provide a concise review of the most important aspects of the graphene growth by ethanol-based CVD on Cu. We offer a summary of the effect of the CVD parameters on the structural properties of graphene, by investigating in detail Raman spectroscopy data. We complement our analysis with data from other analytic techniques, such as optical microscopy, Raman spectroscopy, atomic force microscopy and scanning electron microscopy (AFM, SEM). This review paper is organized in two main sections. The first section concerns the growth of continuous graphene films with quality comparable to those made from methane-CVD, focusing on the role of process temperature, gas flow, and time. In this context, one of the key features of the ethanol-CVD, the possibility of using ultra-fast growth process, will be presented and described. In the second section, the focus will shift to the growth of single-crystal graphene grains, with lateral size in the mm-range and beyond. Here, a comparison will be made between results obtained with methane and ethanol, to provide a perspective on a topic of interest for several nanotechnology applications.

Tuning of Growth Parameters
The first attempts to grow graphene by ethanol-CVD explored temperature ranges lower than those usually reported for methane-CVD (<1000 °C) [39,40,49], in order to lower the overall process costs. Miyata et al. synthesized high-quality monolayer graphene from ethanol on Ni foils at 900 °C and atmospheric pressure [49]. They introduced a "flash cooling" to reduce the duration of the standard cooling process and avoid the decomposition of the synthesized graphene [49].
Following these preliminary results, other researchers explored wider temperature ranges (650-850 °C) (Figure 1a) [40]. Guermoune et al. grew graphene from several liquid alcohols, such as methanol, ethanol and 1-propanol. Their goal was to improve the alcohol-based growth methods and obtain large-area, continuous graphene films with quality comparable to that obtained from methane-CVD [40]. By monitoring the D band in the Raman spectrum of the grown samples [44,45,47], the group found that the improvement in quality (in terms of crystallinity) strongly depended on the temperature: increasing the temperature from 650 to 850 °C, the graphene crystal domain size kept increasing in size, up to 168 nm ( Figure 1a) [40].  Temperatures below 800 • C were also probed to tune the electronic properties of graphene and use it as functional material in optoelectronics [51]. Dong et al. investigated the CVD of ethanol at atmospheric pressure and 900 • C (for 30 min) [39]. The authors grew a continuous monolayer graphene film with overall low defect density. Some small areas of the film showed Raman spectra representative of bi-and multi-layer graphene, due to the occurrence of secondary nucleation during the growth (Figure 1c) [39]. Higher ethanol-CVD temperatures (930-1070 • C) were investigated and optimized in several studies [52][53][54]. Our group previously studied the effect of process duration (t) and H 2 flow rate (Φ H2 -used as co-carrier gas with Ar) [52]. At 1070 • C, a process time equal to that set by Dong et al. [39] (30 min) was reported to be enough to obtain multi-layer graphene with the highest crystalline quality (Figure 1d) [52]. Multi-layer graphene could be obtained thanks to a high H 2 flow rate (100 sccm): this indicates that hydrogen has a crucial role in the process, which lies in balancing and counteracting the etching effect of water molecules coming from the decomposition of ethanol [52]. Meanwhile, a proper amount of hydrogen (10 sccm) and high temperature (1070 • C) were needed to obtain mono-and bi-layer graphene with high crystalline quality and large crystallite size (280 nm) in only 10 min (Figure 1e) [52].
Despite the initial effort to lower the growth temperature using alcohols rather than methane, the aforementioned studies proved that high temperatures and a substantial flow of hydrogen are key requirements for the growth of graphene with high crystallinity by ethanol-CVD. This seems to be in part related to the lower mobility of carbon atoms on the metal surface when using ethanol instead of methane, which would explain the need for a higher growth temperature [40,52]. Taken together, those results pointed out that ethanol has the potential to replace methane as gas precursor in the CVD growth of continuous graphene films, as long as process temperature above 1000 • C are set (Table 1).

Ultra-Fast Growth
In order to make the CVD growth suitable for an industrial production, it is important to reduce the process duration as much as possible while maintaining a high material quality [55]. This both increases the production throughput and reduces the thermal loading on the growth substrates. Usually, the growth by methane-CVD requires rather long processes (20-30 min under atmospheric pressure, C 2020, 6, 14 5 of 13 30 min under low pressure) [25,34]. Despite this, the possibility to decrease growth time was carefully evaluated in the past [33,56]: as an example of a short process with methane, Li et al. reported the growth of monolayer graphene in 2 min, by using a CH 4 flow of 7 sccm (at 6 Pa) at 1035 • C with a partial coverage of Cu substrate [56]. A complete Cu surface coverage was obtained in 3 min by increasing the CH4 flow to 35 sccm [56]. Since then, the mechanism and growth rate of graphene on copper foils have been widely investigated in an extended range of conditions and for various precursors [39,40,[57][58][59][60]. As for the CVD of ethanol, Zhao et al. reported a self-limiting growth from ethanol occurring after 5 min at 1000 • C; however this was achieved with a Cu foil enclosure specifically designed to reduce the gas partial pressure at the catalyst surface (and consequently the nucleation rate) [61]. The graphene grown on the internal surface of the Cu foil is uniform (Figure 2a), monolayer and with high quality, as confirmed by Raman spectroscopy (Figure 2b, I 2D /I G = 2.0, I D /I G ∼0.05) [61]. Our group explored the initial growth stages of graphene on Cu from ethanol, demonstrating that graphene films can fully cover the copper catalyst substrate in 20 s and less [38]. In this study, the ethanol partial pressure was ~2 Pa and the flow 0.1 sccm, one order of magnitude less than those typical for methane [62], demonstrating that the growth from ethanol is generally quicker than that from methane. However, these continuous films are not entirely monolayer (Figure 2c) but appear polycrystalline and with a high number of wrinkles where, regardless of the short growth time, some  Our group explored the initial growth stages of graphene on Cu from ethanol, demonstrating that graphene films can fully cover the copper catalyst substrate in 20 s and less [38]. In this study, the ethanol partial pressure was~2 Pa and the flow 0.1 sccm, one order of magnitude less than those typical for methane [62], demonstrating that the growth from ethanol is generally quicker than that from methane. However, these continuous films are not entirely monolayer (Figure 2c) but appear polycrystalline and with a high number of wrinkles where, regardless of the short growth time, some secondary nucleation took place [38]. The films also show a residual defect-related D peak (Figure 2d, blue curve), implying that the fast Cu coverage occurs at the expense of the crystalline quality of the film. The graphene thickness could be reduced by throttling down the precursor gas feed. When the ethanol flow rate was decreased from the initial 0.1 to 2.5 × 10 −2 sccm and the total pressure from 400 to 65 Pa (Figure 2d, red curve), a continuous monolayer film could be grown which showed a minor residual D peak (this peak can be further reduced only using longer growth processes).

Single-Crystal Graphene Grains: A Comparison between Methane and Ethanol Precursors
CVD graphene is typically poly-crystalline, and the presence of grain boundaries-along with other structural defects-strongly degrades its physical properties, such as charge mobility, thermal conductivity and even mechanical strength [63][64][65]. Many researchers explored the possibility of growing isolated, single-crystal graphene grains (Table 2) [49,56,[66][67][68][69][70][71][72][73]. To this end, the first requirement is a significant decrease in nucleation density, in order to delay the coalescence of the grains into a film and allow them to grow bigger. Isolated graphene grains can be designed for specific uses in microelectronics; in this context, the ability to grow oriented grains in a specific pattern would be crucial for their application. The nucleation density of graphene on Cu obtained by methane-CVD was initially in the order of~10 6 nuclei/cm 2 [66]. Many efforts were made to decrease such a high nucleation density, by carefully modulating the CVD parameters [28,56,69,[74][75][76][77]. With the same objective, the Cu substrates were pre-treated by thermal annealing [66,78,79], electrochemical polishing [80], and pre-oxidation [34,78,[81][82][83]. Initial efforts (on un-treated Cu substrates) enabled the growth of mm-size graphene grains [33,66,69]. Li et al. grew graphene single crystals with lateral size up to 0.5 mm inside Cu-foil enclosures (Figure 3a) [34]. A small methane flow (less than 1 sccm) and partial pressure (50 mTorr) allowed the growth of large graphene grains with 'dendritic' edges. The Raman spectra recorded at two different spots (Figure 3g), one within the film and the other close to the dendrite edge are reported in Figure 3g. The electron mobility in these isolated grains that was higher than 4000 cm 2 V −1 s −1 [34]. The search for larger grain sizes pushed researchers to further optimize the methane-CVD process: Tour's group, for example, grew mm-sized single-crystals (~4.5 mm 2 ) [74]. The growth mechanism was optimized to increase hexagonal-shaped domains from~15 µm to~2.3 mm (Figure 3b). Figure 3h shows the Raman spectra recorded at six different spots in Figure 3b: the D peaks are in the noise level indicating samples with very high quality. The authors also fabricated Hall-bar FETs (on Si/SiO 2 ) with the grown graphene, which showed a hole mobility of~11,000 cm 2 V −1 s −1 at room temperature [74]. However, until 2012, a wide variation in domain size, shape, and film quality was recorded from lab to lab, suggesting that crucial growth parameters remained uncontrolled. For this reason, Ruoff's group demonstrated the possibility to suppress graphene nucleation and promote the growth of ultra large single-crystal graphene domains by pre-oxidizing the Cu foils [81]. The group systematically compared nucleation density, growth rate, and size of graphene grains grown on oxygen-treated Cu foils and produced for the first time cm-sized graphene single crystals (Figure 3c) (decreasing the nucleation density below~1 nuclei/cm 2 , i.e., 6 orders of magnitudes less than the initial values) [81]. Since 2013, many groups realized that involving oxygen in the CVD reaction can efficiently reduce the nucleation density, and several procedures were developed to exploit this effect [78,81]. min) and a low ethanol flow of 1.5 × 10 −3 sccm, the grain size scales up 500 μm and more (for 60 min at 400 Pa, Figure 3e). The nucleation density in this case is less than 3 nuclei mm −2 . These large graphene grains are monolayer with high quality (Figure 3i), and showed a field-effect mobility up to 1355 cm 2 /(V s) [84]. These results, obtained on flat Cu surfaces without the need of any Cu structuring (e.g., folding, sealing, using enclosures, etc.), are significant progress towards a prospective industrial production.  The growth of similar individual grains with large size by ethanol-CVD initially lagged behind. In 2015, Chen et al. reported the growth of isolated hexagonal graphene grains by ethanol-CVD on pre-oxidized Cu foils folded in enclosure [67]. The authors grew single-crystal grains as large as 5 mm at 1065 • C (Figure 3f) (with a nucleation density of~0.1 mm −2 ) with the following conditions: 90-min Cu foil pre-oxidation in air, and CVD with ethanol partial pressure of 0.03 Pa, low H 2 concentration C 2020, 6, 14 9 of 13 (3%), total pressure of 300 Pa. They also demonstrated that a higher oxidation temperature with a shorter duration (300 • C, 40 min), or a lower temperature with a longer duration (240 • C, 2 h) could result in similar nucleation density [67]. Without the need of folded enclosures, our group recently improved the ethanol-CVD growth of single crystals larger than 500 µm on pre-oxidized Cu substrates [84]. This work showed that without Cu pre-oxidation the high nucleation density usually prevents the growth of graphene grains larger than 1-3 µm. However, when using a combination of Cu-preoxidation (60-150min), and a CVD process with a short Ar pre-annealing (1 min) and a low ethanol flow of 1.5 × 10 −3 sccm, the grain size scales up 500 µm and more (for 60 min at 400 Pa, Figure 3e). The nucleation density in this case is less than 3 nuclei mm −2 . These large graphene grains are monolayer with high quality (Figure 3i), and showed a field-effect mobility up to 1355 cm 2 /(V s) [84]. These results, obtained on flat Cu surfaces without the need of any Cu structuring (e.g., folding, sealing, using enclosures, etc.), are significant progress towards a prospective industrial production.

Conclusions
In the present paper, we offer a concise summary of the state of the art on the growth of graphene on Cu by ethanol-CVD, focusing on both continuous films and isolated single-crystal grains. The reported results were compared to those commonly obtained by methane-CVD. Similarly to methane-based CVD, the quality of graphene films grown by ethanol-CVD was found to generally increase with temperature; but, as a marked difference, the required time to grow fully continuous films (i.e., entirely covering the Cu substrate surface) could be significantly reduced (down to a few seconds). Reducing the process duration can be key to a cost-effective industrial production of graphene, and ethanol thus seems to be an ideal carbon precursor in this respect. Concerning the ethanol-CVD growth of isolated single-crystal graphene grains, vital both for the production of seamless continuous graphene films and for the development of microelectronic applications, significant advancements have been reported since 2013, when the fundamental role of oxygen to suppress graphene nucleation was discovered. Using pre-oxidized Cu foils folded in enclosures, the size of isolated graphene grains was scaled up to cm-size. Importantly, the growth of high-quality mm-sized graphene grains was also achieved on flat Cu foils, removing the need of the undesired enclosure approach towards an easier production.