Size and Density of Graphene Domains Grown with Different Annealing Times

Abstract

Single crystals of hexagonal graphenes were successfully grown on Cu foils using the atmospheric pressure chemical vapor deposition (CVD) method. We investigated the effects of reaction parameters, such as the growth temperature and annealing time, on the size, coverage, and density of graphene domains grown over Cu foil. The mean size of the graphene domains increased significantly with increases in both the growth temperature and annealing time, and similar phenomena were observed in graphene domains grown by low pressure CVD over Cu foil. From the comparison of micro Raman spectroscopy in the graphene films grown with different annealing times, we found that the nucleation and growth of the domains were strongly dependent on the annealing time and growth temperature. Therefore, we confirmed that when reaction time was same, the number of layers and the degree of defects in the synthesized graphene films both decreased as the annealing time increased.

Introduction

Graphene is a two-dimensional carbon material whose structure is a one-atom-thick planar sheet of sp2-bonded carbon atoms densely packed in a honeycomb crystal lattice. It has drawn significant attention because of its remarkable structural and electrical properties.1 Its extremely high electron mobility and tunable band gap make graphene potentially useful for innovative approaches to electronics.2-5 Although mechanical exfoliation of graphite and decomposition of SiC surfaces by thermal treatment have been the main methods for producing graphene,67 these methods have limits to the quality and scale of the produced graphene films. Solution-phase89 and solvothermal syntheses of graphene1011 have provided major improvements in terms of processing, although for device fabrication, a reproducible method such as chemical vapor deposition (CVD) growth yielding high-quality films with controlled thicknesses is required.

The CVD method is a powerful and cost-effective technique for the production of large-scale, high-quality graphene films. Recently, the CVD growth of graphene on metal catalysts such as Ir,12 Ru,13 Ni,14-16 and Cu17 foils has been reported. In spite of the complexity of CVD procedures involving different catalysts, carbon sources, and other variables, the physical principles underlying this method are relatively simple. It is widely accepted that CVD mainly involves a surface catalytic reaction1718 or bulk carbon precipitation onto the surface during cooling1516 for catalysts with low- and high-carbon solubilities, respectively. In both cases, graphene nucleation on the catalyst surface is a critical step in the growth process. Various factors affect the initiation of the graphene nucleation process, including the surface microstructure of the metal catalyst,1819 carbon source,20 carbon segregation from metal-carbon melts21 and reaction parameters during CVD growth.22–24

In this research, we synthesized hexagonal graphene flakes on Cu foils by CVD and controlled the coverage, density, and size of the graphene domains by changing the reaction parameters. It is important to control these graphene growth parameters during synthesis in order to achieve tunable properties and optimized device performance.

 

Experimental

Materials. The Copper foil was 0.025 mm (0.001in) thick, 99.999% purity. This was purchased from Alfa Aesar. And, the 300 nm SiO2/Si substrate was purchased from Hitsan.

Measurements. The scanning electron microscope (SEM) images of graphene flakes were obtained from a FE-SEM (Jeol 7600F) at an acceleration voltage of 10 to 20 kV, which was performed on the as-synthesized product on substrates. We obtained the Raman spectrum of graphene samples with a homemade micro-Raman spectroscopy system. In micro-Raman spectroscopy, the 514.5 nm line of an Ar ion laser was used as the excitation source with a power of ~1 mW. The heating effect can be neglected at this power range. The laser beam was focused onto the graphene sample by a 50 × microscope objective lens (0.8 N.A.), and the scattered light was collected in the backscattering geometry. The collected scattered light was dispersed by a Shamrock SR 303i spectrometer (1200 grooves/mm) and was detected with a CCD detector.

 

Results and Discussion

Synthesis and Characterization of Graphene Flakes.We synthesized graphene flakes by Cu-catalyzed CVD under atmospheric pressure using a gas mixture of Ar, H2, and CH4, as shown in Figure 1(a). Cu foils were first thoroughly cleaned with acetone, methanol and deionized (DI) water. Then, they were loaded into a 1 inch quartz tube inside a horizontal furnace (Lindberg/Blue M, Thermo Scientific). The tube was evacuated to a vacuum of 3 mTorr, which was maintained for 10 min. Then, the tube was brought back to atmospheric pressure by introducing mixed Ar and H2 gas. This process was repeated several times. The sample was then heated to 1050 ℃ over 1 h with 300 sccm and 20 sccm of Ar and H2, respectively. Afterward, for the thermal treatment of the Cu foil, the temperature in the tube was maintained for another 1 h for the annealing process. Then, Ar gas was shut off and 300 sccm of diluted methane (50 ppm in Ar) was allowed to flow into the tube along with 20 sccm of H2 for graphene growth. The growth was terminated by quenching the quartz tube under ambient pressure. Then, the as-grown graphene flakes were transferred onto 300 nm SiO2/Si substrates by wet-etching the underlying Cu foils.

Figure 1.(a) Schematic illustration of CVD apparatus for graphene synthesis. (b) Experimental diagram for growth of graphene including their four stages such as ramping, annealing, growth and cooling process; TG is growth temperature.

The effects of the reaction parameters such as the growth temperature and annealing time on the size, coverage, and density of the graphene domains were investigated for atmospheric pressure CVD (APCVD) growth of graphene on Cu foil. The purpose of this study is to find out how the growth parameters affect the size and density of graphene domains. The growth temperature (TG), annealing time (tA) and the growth time (tG) were the parameters investigated. The whole scheme is illustrated in Figure 1(b). Generally, the annealing process is used to improve the quality of the Cu surface, resulting in bigger graphene domains. Furthermore, the growth temperature significantly impacts the size of the graphene domains and the nucleation density. In addition, the cooling rate has a crucial role in suppressing the formation of multiple layers and affects the efficiency of the separation of graphene layers from the substrate in the later process. To investigate initial graphene flakes, graphene synthesis was conducted for 15 min and stopped before a continuous film was formed.

Raman spectroscopy has been a useful tool for investigation of graphitic materials such as fullerenes, carbon nanotubes (CNTs), and graphite.25-27 The number and orientation of graphene layers can be readily identified by their elastic and inelastic light scattering, as measured by Raman28-33 and Rayleigh spectroscopy.3435 Raman spectroscopy also allows the doping, defects, and strain to be measured.283136-39 To characterize graphene flakes, graphene was transferred using a PMMA-assisted wet-transfer method onto a 300 nm SiO2/Si wafer after growth for Raman spectroscopy. A thin layer of PMMA (MicroChem 950 PMMA C3, 3% in chlorobenzene) was spin-coated on the as-grown products at 3000 rpm for 1 min. Since both Cu surfaces were exposed to CH4, graphene was grown on both sides of the Cu foil. Excess PMMA on the back side of the PMMA-coated graphene was removed by acetone. Subsequently, the sample was placed in 0.1 M aqueous (NH4)2SO4 to etch off the Cu foil. Generally, the etching process runs overnight. After the Cu foil was completely etched away, graphene with a PMMA coating was scooped out from the solution and transferred it into DI water for 10 min three times to wash away the remaining etchants. Then, the SiO2/Si substrate was dipped into water, and the film was picked up. Finally, the PMMA was removed with acetone, and the sample was rinsed several times with DI water.

Domain Sizes of Graphene Grown at Different Growth Temperatures. We synthesized graphene flakes at different growth temperatures in order to analyze the influence of the growth temperature on the sizes of the graphene domains. Graphene flakes synthesized at different reaction temperature are shown in Figure 2; (a) 1030, (b) 1050, and (c) 1070 ℃. The growth time was 15 min for all samples.

In general, the nucleation and growth of the graphene domains on a Cu foil is initiated by the supersaturation of active carbon species produced by decomposition of carbon sources such as methane. Since the decomposition of carbon sources can occur at high temperatures, this process strongly depends on the growth temperature. In particular, the sample grown at 1030 ℃ has no graphene domains on the Cu surface (Figure 2(a)). On the other hand, once the supersaturation reaches a certain critical point at a high temperature, nuclei start to form on the Cu catalyst and grow into graphene domains that can continuously expand until they merge together with adjacent domains to fully cover the Cu surface (Figure 2(b) and 2(c)). The increase in the average size of the monolayer domains with increasing growth temperature can be attributed to the increase in the rate of graphene growth due to the higher supersaturation of thermally decomposed carbon species. The decrease in the domain density with increasing temperature can also be completely explained by the nucleation and growth process. Generally, the initial nucleation of graphene takes place more frequently at step edges, folds, or other imperfections on the Cu foil to form domains. As a result, the presence and the number of monolayer domains strongly depends on the condition of the Cu foil surface. Because the number of step edges, folds, or other impurities is reduced at higher temperatures, the smaller number of defects on the Cu surface results in a lower density of nucleation sites and fewer domains. In addition, the lower domain density at higher growth temperatures is attributed to the Ostwald ripening process during the initial growth stage.4 The flattening and smoothing of the Cu surface makes the diffusion length of the surface-bounded carbon species improve notably as compared to that on a rough surface. The longer diffusion length of carbon species increases the tendency toward Ostwald ripening and therefore reduces the number of nuclei. Both these factors lead to a reduction in the graphene domain density and an increase in the domain size at higher temperatures.

Figure 2.SEM images of graphene domains grown on Cu foil for 15 min using the APCVD method with different growth temperature, (a) 1030 ℃, (b) 1050 ℃, and (c) 1070 ℃, respectively. Asproduced graphene flakes are transferred on SiO2/Si substrates prior to SEM measurements and all scale bars are 10 μm.

Domain Density of Graphene Grown with Different Annealing Times. We synthesized graphene flakes with a variety of annealing times to confirm whether the annealing time can affect the coverage of the obtained graphene flakes. Figure 3 shows SEM images of graphene flakes synthesized with annealing times of (a) 30 min, (b) 60 min, and (c) 120 min. The growth time was 15 min for all samples.

It is well known that substrate annealing is advantageous for reducing the amount of volatile impurities, contaminants, and defects on a Cu surface, and thus leads to the inhibition of graphene nucleation. The influence of the annealing time on both the graphene nucleation density and the domain size was studied. The SEM images illustrate that the graphene domain density decreases for prolonged annealing times.40 This indicates that a longer annealing time reduces the nucleation density. In accordance with this tendency, the domain size increases for prolonged annealing times.3 For these reasons, when 120 min of annealing time was used in the synthesis, the grown graphene domains are more like 2D films even for the same growth time. In other words, the coverage of graphene flakes on the Cu foil improved remarkably after prolonged annealing compared to graphene flakes synthesized with short annealing times (Figure 3(c)).

Figure 3.SEM images of graphene domains grown on Cu foil for 15 min using the APCVD method with different annealing time, (a) 30 min, (b) 60 min, and (c) 120 min, respectively. As-produced graphene flakes are transferred on SiO2/Si substrates prior to SEM measurements and all scale bars are 10 μm.

Figure 4.SEM images of graphene domains and films grown on Cu foil using the LPCVD method with different annealing time, (a, d) 30 min, (b, e) 60 min, and (c, f) 120 min, respectively, and different growth time, (a-c) 10 min, and (d-f) 60 min, respectively. As-produced graphene flakes are transferred on SiO2/Si substrates prior to SEM measurements and all scale bars are 10 μm.

This result matches well with those obtained for the graphene grown at low pressure. Figure 4 shows graphene domains and films produced using the low-pressure CVD (LPCVD) method with different annealing times: (a, d) 0 min, (b, e) 60 min, and (c, f) 120 min. For 10 min of growth time, the size of the graphene domains increased as the annealing time was prolonged (Figure 4(a), 4(b), and 4(c)). This phenomenon suggests that the nucleation and growth of the domains were strongly dependent on the annealing time. For this reason, the mean size of the domains increased for prolonged annealing times at both low pressure and atmospheric pressure.

Characterization of Graphene Films Grown with Different Annealing Times. After 60 min of growth time, we investigated the number of layers and the amount of defects in graphene films using micro-Raman analysis. Figure 5 shows the Raman spectra (excited by 514.5 nm laser) of graphene films grown using the LPCVD method with annealing times of with annealing time, (a) 0 min, (b) 60 min, and (c) 120 min. The D, G, and 2D bands are indicated in the figure, and all graphene films were transferred onto SiO2/Si substrates before analysis.

We identified the number of layers and the amount of defects in graphene films by carefully examining the I2D/IG ratio and the intensity of the D peak, respectively.1740-45 In Figure 5(a), the graphene film grown without an annealing process showed a large D peak (~1350 cm−1) and a relatively low I2D/IG ratio (~1.5), indicating that it has many defects and consists of bilayer graphene, since the I2D/IG ratio is less than 2 but more than 1. Meanwhile, the graphene films grown with 60 min and 120 min of annealing time showed small D peaks and I2D/IG ratios of 2.5 and 2.7, respectively (Figure 5(b) and 5(c)). Because of the longer annealing time, these as-synthesized graphene films have few defects and consist of a monolayer, since the I2D/IG ratio is larger than 2.

The Raman spectra of each film discussed above show that the defect-related peak is considerably activated for graphene films grown with insufficient annealing processes. Surprisingly, it was found that the graphene films grown with different annealing times have different film qualities. Thus, we confirmed that increasing the annealing time results in fewer defects in large-area monolayer graphene films.

Figure 5.Raman spectra (excited by 514.5 nm laser) taken in graphene films grown for 60 min with different annealing time, (a) 0 min, (b) 60 min, and (c) 120 min, respectively, using the LPCVD method. As-produced graphene films are transferred on SiO2/Si substrates prior to Raman analysis and D (~1350 cm−1), G (~1580 cm−1), and 2D (~2680 cm−1) bands are indicated in the figure.

 

Conclusion

Single crystals of hexagonal graphenes were successfully grown on Cu foils using the APCVD method. We found that the nucleation and growth of the domains were strongly dependent on the annealing time and growth temperature. The mean size of the domains increased for prolonged annealing times and higher growth temperatures, while the density of the domains decreased with increasing growth temperature. Graphene domains grown on Cu foils using LPCVD showed similar effects of the annealing time. The mean size of the domains increased and the resulting large-area monolayer graphene film had fewer defects for longer annealing times. The high quality of the graphene films grown using APCVD was confirmed by micro-Raman spectroscopy. Therefore, we confirmed that when reaction time was same, the number of layers and the degree of defects in the synthesized graphene films both decreased as the annealing time increased.