Silane-catalysed fast growth of large single-crystalline graphene on hexagonal boron nitride

The direct growth of high-quality, large single-crystalline domains of graphene on a dielectric substrate is of vital importance for applications in electronics and optoelectronics. Traditionally, graphene domains grown on dielectrics are typically only ~1 μm with a growth rate of ~1 nm min−1 or less, the main reason is the lack of a catalyst. Here we show that silane, serving as a gaseous catalyst, is able to boost the graphene growth rate to ~1 μm min−1, thereby promoting graphene domains up to 20 μm in size to be synthesized via chemical vapour deposition (CVD) on hexagonal boron nitride (h-BN). Hall measurements show that the mobility of the sample reaches 20,000 cm2 V−1 s−1 at room temperature, which is among the best for CVD-grown graphene. Combining the advantages of both catalytic CVD and the ultra-flat dielectric substrate, gaseous catalyst-assisted CVD paves the way for synthesizing high-quality graphene for device applications while avoiding the transfer process.

These results indicate that the graphene grown on h-BN is almost Si/Ge-free.  nm. An objective lens of 100× magnification and a 0.95 numerical aperture (NA) was used, producing a laser spot that was ~0.5 µm in diameter. The laser power was kept less than 1 mW on the sample surface to avoid laser-induced heating. The excitation-laser-energy-dependent-Raman spectroscopy was used to support the conclusion that the shoulder peak located at 1565 cm -1 in precisely aligned graphene domain is a TO phonon originated from the intervalley umklapp scattering activated by graphene/h-BN super-lattice.

Supplementary Note 1 | the effect of solid catalyst
To qualitatively understand the role of silicon (germanium) atom in the growth of graphene, solid silicon, solid germanium and their alloy were placed near h-BN flakes, respectively. The Ge/Si alloy consists of 30% Si in molar ratio. We found that the presence of solid silicon/germanium/their alloy can effectively improve growth rate of graphene when heating them to 1280 or higher. It indicates that silicon/germanium vapor plays a role of catalyst in graphene growth on h-BN. The results illuminate us that the saline/germane may be effective catalysts for graphene growth on h-BN.

Supplementary Note 2 | Key parameters for the graphene growth
The graphene growth was influenced mainly by the gaseous catalyst and growth temperature: 1. Argon/silane mixture flow rate: a C 2 H 2 flow and a mixture of silane/argon (the mole ratio of silane to argon was 5%) were introduced into the system for the graphene growth.
High flow rate of the argon/silane mixture resulted in a rapid growth rate. Some growth results with different argon/silane mixture flow rate are summarized in Supplementary 3. Too much silane could lead to the formation of SiC as shown in Supplementary Fig. 3.

Supplementary Note 3 | C 2 H 2 as a precursor.
It is worthy to note that the above picture hold valid only when C 2 H 2 is used as a precursor. If we use CH 4 as a carbon precursor, the atmospheric composition in the reaction chamber will be very different. We experimentally verified that the catalytic effect is not obvious, if using CH 4 . The catalytic effects due to silicon or germanium atoms are further validated by heating solid silicon or Si/Ge alloys to different temperatures at different pressures, thus producing different vapor pressures of Si and Ge atoms.

Simulation of the growth behaviors at zigzag edges.
To get a more detailed understanding on the growth mechanism at zigzag edges, we performed a DFT calculation and the results are shown in Supplementary Fig. 5. We assumed that the growth frontier is passivated by hydrogen atoms, and the C 2 H 2 molecule is the carbon feedstock. In the non-catalyst situation ( Supplementary Fig. 5a), the C 2 H 2 molecules need three steps to incorporate to the graphene edge to form a new carbon ring.
The energy barriers for the three steps are 5.69eV, 3.33eV and 5.69eV, respectively. After the new carbon ring is formed, the growth frontier steps forward and form a same structure as that before the C 2 H 2 molecule integrated in the carbon ring. In the silicon catalyst situation (Supplementary Fig. 5b), we start with one silicon atom being absorbed on the growth frontier, in the growth circle such silicon atom serves as a bridge between carbon dimer and graphene, which lowers the energy needed for dehydrogenation of C 2 H 2 and formation of carbon-carbon bonds to 2.77eV, 2.15eV and 0.5eV. The much lowered energy barriers account for the greatly enhanced growth rate.

Classification of graphene domains grown on the surface of h-BN.
Actually, we scanned the surface of h-BN flakes and measured the topography of the Although the investigations are actually tedious and time-consuming, it is worthy to do such an investigation to know the uniformity of the graphene domain. Supplementary Fig.   7 gives more examples for the explanation of the type survey.

Transport measurement
Electronic transport measurements in a Hall bar configuration were also carried out to characterize the graphene single crystal grown on h-BN. The gate voltage (V g ) dependence of the longitudinal resistance (R xx ) at different temperature is plotted in Fig.   5a. The main peak in R xx at V g = -5 V, represents graphene's main neutrality point. Two satellite peaks symmetrically appear on both sides of the main neutrality point. As the temperature decreases, these two satellite peaks, become more obvious. The satellite peak on the hole side appears much stronger than that on the electron side. The satellite peaks in transport properties are believed to be related to the spectral reconstruction in graphene brought to contact with h-BN. The superlattice potential induced by the h-BN results in the appearance of secondary Dirac points (SDP) in graphene's energy dispersion. They can be observed as the satellite peaks in the transport measurement after the Fermi energy is tuned to reach the reconstructed part of the spectrum. The resistances (R) at the Dirac Point (DP) and satellite peak at the hole branch as a function of T are plotted in the inset of Fig.5a, both of them exhibit very weak temperature dependence, the results are different from earlier reports. One possible reason is that the commensurate state is suppressed. 24 As the satellite resistance peaks in the transport data originates from the moiré pattern, the wavelength of moiré pattern can be estimated by measuring the relative position of the satellite resistance peaks. As the energy separation between the DP and the secondary  Fig. 5a, the electrical field mobility at 300 K is about 17,000 cm 2 V -1 s -1 , which is extracted from a density independent mobility model The color plots of the R xx and R xy as a function of both gate voltage and magnetic field are shown in Fig.5c and 5d, respectively. The standard quantum Hall effect (QHE) for graphene, is observed with valleys in R xx (Fig.5c) and plateau in R xy (Fig. 5d)