Direct growth of high crystallinity graphene from water-soluble polymer powders

Among current graphene preparation methods, chemical vapor deposition (CVD) is the most promising approach to obtain high quality large graphene films [1–4]. Gaseous carbon sources, such as methane [5–7] and ethylene [8, 9] have been widely used in CVD for graphene growth. Compared with gaseous hydrocarbons, liquid and solid carbon sources have many advantages, including but not limited to controlled in situ doping [10–13], reducing growth temperature [14–17], transfer-free growth on dielectric substrate [18–20], and directed growth of patterned graphene [21]. Various kinds of polymers have been employed as solid carbon sources for graphene growth [22–24]. Both copper and nickel have been taken as catalyst in CVD. Thick graphene can be formed by a solid-state transformation of carbon-containing sources on Ni [25, 26]. Compared with Ni, graphene grown on Cu foil is often single layer, due to the negligible solubility and diffusion of carbon atoms in bulk Cu. Cu cannot directly induce graphitization of solid carbon sources [27], which makes graphene growth from polymer on Cu more charismatic and complex.

Based on previous studies, the molecular structure of a solid carbon precursor has major influence on the growth kinetic of graphene on Cu. The growth mechanisms may mainly have two types: (i) polycyclic aromatic hydrocarbons (PAHs) diffuse and coalesce into graphene on the surface of Cu [14, 16, 28]; (ii) amorphous carbon or polymers without aromatic moieties such as polymethylmethacrylate (PMMA) decompose into gaseous hydrocarbons followed by gas re-adsorption and graphene nucleation on Cu [27, 29]. The transformation of PMMA to graphene at high temperature [30] is similar to the CVD process of gaseous hydrocarbons. A clear picture of graphene growth mechanism can guide an effective preparation process and facilitate the synthesis of graphene with high quality. However, there are limited reports on graphene growth on Cu foils from polymers, particularly for industrial widely used water-soluble polymers. In general, most of water-soluble polymers, such as polyvinyl alcohol (PVA), polyvinyl pyrrolidone (PVP) and polyethylene glycol (PEG), are inexpensive and environmentally friendly [31–33]. Although there are several reports on the synthesis of graphitic materials using PVA and PVP [34–36], a systematic study on converting water-soluble polymers to graphene on copper foil is still absent.

There are two commonly used methods for graphene growth from solid carbon sources: direct CVD (DCVD) and rapid thermal treatment (RTT). Previous study suggested that when a solid carbon source was pre-deposited on metal catalysts, RTT was needed to avoid excessive dissociation of carbon source during the heating up process for graphene growth [18, 27, 37, 38]. To systematically study the growth of graphene from three types of water-soluble polymers (PVA, PEG and PVP) on copper foils, we adopted both these two approaches in this work, as illustrated in figure 1. Identical amount of polymer was applied for DCVD and RTT growth, respectively. Results of continuous graphene film (by DCVD) and amorphous carbon fragments (by RTT) revealed a slow decomposition mechanism of water-soluble polymer on Cu thanks to the existence of metal-binding functional groups in the carbon sources [34]. Detailed characterizations on the crystallinity, composition, transmittance and electrical properties of graphene prepared from the water-soluble polymers and polyethylene (PE) have been carried out to elucidate the effect of oxygenated functional groups in carbon precursors on graphene growth. In previous studies, oxygen has been intentionally introduced for graphene growth [39, 40], and a moderate amount of oxygen could benefit graphene growth by promoting hydrocarbon dissociation [8, 41]. We found the presence of oxygen in carbon sources could benefit the growth of graphene in the same way. Moreover, PVP could also work as a nitrogen-containing precursor, leading to in situ doping of pyrrolic-N in as-grown graphene.

Zoom In Zoom Out Reset image size

The synthesis process was carried out in a thermal CVD furnace system with a quartz tube (1.7 m in length, 5 cm in diameter) at atmosphere pressure (figure S1 (stacks.iop.org/TDM/5/035001/mmedia)). Copper foil (25 µm in thickness, item No. 46365), PVA (item No. 41241) and PE (item No. 42607) were purchased from Alfa Aesar. PVP was purchased from Sinopharm Chemical Reagent. PEG 20000 was purchased from Peking Reagent.

DCVD process

RTT process

Transfer process

Characterizations

Recently, successful growth of graphene from trace amount of carbon sources by CVD has been reported [42, 43]. Therefore, in this work, to avoid contamination from unintentionally introduced carbon, we customized a thermal CVD furnace system operated at atmosphere pressure. Without deliberately introducing a carbon source, no graphene grew on copper foil using our equipment. For DCVD process, as shown in figure 1(a), solid polymers were decomposed at the gas inlet side of the system. For RTT process, in order to uniformly deposit water-soluble polymers on copper foil for RTT process, the polymers were dissolved in deionized water to form an aqueous solution (figure 1(b)). Organic solvents, which were usually needed to dissolve polymers such as PMMA [30], were intentionally avoided to ensure the purity of carbon sources.

When a copper foil was covered by a layer of graphene, it could be protected from oxidation at low temperature. However, naked copper foil couldn't survive from low temperature oxidation [44]. This feature could be used to visualize graphene on copper due to different optical contrast between graphene covered copper and naked copper after oxidation treatment [45, 46]. For products from PVA on Cu by different growth processes, after annealing in air at 140 °C for 30 min, there was notable color contrast, as shown in figure 2(a). The bright color of DCVD sample (figures 2(a)(iii)) appeared the same as that before annealing, suggesting the presence of a uniform layer of graphene film on the surface of copper. Two RTT samples (figures 2(a)(i) and (ii)) showed characteristic red color of copper oxide, indicating no graphene formation during the RTT process.

Zoom In Zoom Out Reset image size

The morphology characterization of as-prepared DCVD graphene from PVA (PVA-G) on SiO₂ (300 nm)/Si substrate (figures 2(b) and (c)) showed continuous film with smooth surface. The SEM image in figure 2(d) showed features of graphene wrinkling on copper foil. The neat graphene film transferred onto a standard TEM micro-grid was shown in figure 2(e). As shown in figure 2(f), the sharp peaks of the G (~1583 cm⁻¹) and 2D (~2690 cm⁻¹) band in Raman spectrum demonstrated successful growth of PVA-G by DCVD. In contrast, Raman spectra of the samples from PVA on Cu treated by RTT for both 0.5 h and 1 h indicated the carbon product was amorphous. The results were the same with PEG and PVP by RTT, as shown in figures S3–S5. The transformation of water-soluble polymers to amorphous carbon on Cu by RTT process was supposed due to the presence of metal-binding functional group [34] (see details in supporting information). Taking this insight into consideration, we determined that the commonly used pre-deposition of carbon source on copper foil should be avoided in future works if polymers with metal-binding functional groups (such as hydroxyl) were used to grow graphene.

We further investigated the DCVD process of graphene growth using PEG, PVP, and PE (PEG-G, PVP-G, and PE-G). Raman spectra of as-prepared graphene films were displayed in figure 3(a). The D peak (~1350 cm⁻¹) was a defect-activated peak and the I_D/I_G ratio increased with defects and disorder in graphene [47, 48]. Compared to graphene synthesized from water-soluble polymers, PE-G exhibited broader D and G bands, indicating a high degree of disorder in graphene [49]. The Raman spectrum of PVP-G showed extremely high I_D/I_G ratios, along with a high G' (~1620 cm⁻¹) band. This result suggested that the defects may due to heteroatom (N) doping in PVP-G, which will be echoed by XPS results later. Raman mappings were recorded from 5  ×  5 µm regions of graphene films converted from four different polymers. The D and G band intensity maps in figure S6 were consistent with the Raman spectra (figure 3(a)) observed on each sample, indicating the repeatability of the Raman observations. As shown in figure 3(b), the obvious color difference clearly demonstrated the scale of I_D/I_G ratio for graphene samples converted from different carbon sources. PVA-G and PEG-G films showed relatively low I_D/I_G ratio (figure 3(c)). The average I_D/I_G ratio of ~0.2 suggested the formation of low density of defect in PVA-G.

Zoom In Zoom Out Reset image size

XPS spectra (figure 4(a)) revealed the existence of carbon in pure graphene. Detailed spectra of C 1s peaks (figures 4(b) and S7) showed low concentration of oxygen-containing bands in graphene synthesized from all polymers. This result indicated that oxygenated functional groups in water-soluble polymers do not induced oxidation of as-grown graphene. Furthermore, according to the fitting results of C 1s peak (table S1), the concentration of sp² carbon in graphene grown from oxygen-containing carbon sources (PVA and PEG) was higher than that of PE-G. This was consistent with the Raman observation of higher disorder in PE-G. Nitrogen-containing precursors have been proven efficient in growth of in situ nitrogen-doped graphene by CVD [11, 12]. A 3.47% of N doping was detected in PVP-G by XPS. The N 1s spectrum could be split into two individual peaks at 400.5 and 406 eV corresponding to different bonding configurations, as shown in figure 4(c). The presence of peaks at 404–406 eV was attributed from adsorbed N species or oxidized N species [50–52]. The origin of peaks at 399.8–401 eV corresponded to pyrrolic-N [11, 53, 54]. Pyrrolic-N represented the nitrogen atom in a five-membered ring and contributed two electrons to the π system, thus forming relatively stable doping in graphene. According to the fitting results, pyrrolic-N contributed to 70% of the N-doped graphene, which should be related to the existence of pyrrolic-N in PVP [55].

Zoom In Zoom Out Reset image size

Graphene layers with a clear edge could be found in the HRTEM images of PVA-G and PEG-G (figure 4(d)). The disorder in PVP-G was understandable, because of heteroatom doping. The HRTEM image of PE-G showed relatively high disorder, which was consistent with aforementioned Raman and XPS results. Selected-area electron diffraction (SAED) patterns in figure 4(e) displayed the typical hexagonal crystalline structure of as prepared graphene films.

Successful growth of graphene from PVA, PEG and PVP by a DCVD process indicated that the previously mentioned decomposition growth mechanism for polymers without aromatic moieties on Cu foil was also applicable for water-soluble polymers. Based on our experimental results, the molecular structure of polymer chains strongly influenced the growth kinetics of graphene, leading to graphene films with different qualities. In addition, the relatively low quality of PE-G demonstrated that oxygen element in PVA and PEG might be crucial for graphene growth. Because when water-soluble polymers were used as carbon sources, there would be oxydant species in the reaction atmosphere. In previous reports, oxygen has been intentionally introduced to promote hydrocarbon dissociation [8, 41]. We reasoned that the presence of oxygen in water-soluble polymers benefited the growth of graphene in the same way, as shown in figure 5. Polymers decomposed into gaseous hydrocarbons, subsequently re-adsorbed on Cu foil and further nucleated and grew into graphene. For PVA and PEG, both polymers contained oxygen element, which promoted the growth of graphene. For nitrogen-containing PVP, the presence of oxygen facilitated the growth of nitrogen doped graphene.

Zoom In Zoom Out Reset image size

The optical transmittance and electrical properties of graphene films were summarized in table S2. As shown in figure 6(a), PVA-G exhibited high transmittance of 93.6% at 550 nm with low sheet resistance of 2.6 kΩ/sq. This is comparable to graphene grown from PMMA (97.1%, 1.2 kΩ/sq) [30], from toluene (97.33%, ~8 kΩ/sq) [15], or from exfoliated graphene (94.7%, 4.5 kΩ/sq) [21]. PEG-G shows a better electrical conductivity of 1.1 kΩ/sq with a weaker transmittance of 88.9%.

Zoom In Zoom Out Reset image size

Hall Effect is a useful means to evaluate the carrier density of graphene [56, 57]. A macroscopic graphene film (4  ×  8 mm) was tested, as shown in the inset of figure 6(b). A magnetic field was applied perpendicular to the graphene plane, with an applied current of 1 mA. The Hall resistance R_H as a function of applied magnetic field was displayed in figure 6(b). To eliminate the magnetoresistance and improve accuracy [56], Hall resistance was measured at different sample positions at magnetic field of 9 T (figure 6(c)). The carrier density was extracted from the sum of resistance measured at opposite magnetic field. The carrier density of PVA-G is 1.2  ×  10¹³ cm⁻², which is close to that of graphene grown by CH₄ in previous report [58]. The carrier density of PVP-G was increased to 2  ×  10¹³ cm⁻² due to the nitrogen doping.

In summary, large-area and high-quality graphene films have been synthesized on Cu foils from water-soluble polymers by DCVD process. On the contrary, only amorphous carbon forms on Cu foil by RTT, because of the metal-binding functional group in water-soluble polymers. The results suggest that polymers with analogous structures may not be able to directly convert to graphene on Cu foil. The effect of oxygen in solid carbon sources has been identified to facilitate the crystallization of graphene. Specific heteroatom-containing water-soluble polymers can induce corresponding doped graphene. These insights on the growth mechanism of graphene can help guiding the selection of polymers and design of synthesis process to achieve desirable graphene products.

This work was supported by the National Natural Science Foundation of China (51672150), Beijing Natural Science Foundation (2172027), Tsinghua University Initiative Scientific Research Program.

Further details of the CVD system, oxygen plasma treatments, Hall Effect measurement, additional characterizations of amorphous carbon converted from water-soluble polymers by RTT process, detailed analyses of XPS C1s spectra of as-prepared graphene are available online.