High quality and wafer-scale cubic silicon carbide single crystals
Guobin Wang1,2†, Da Sheng1,2†, Yunfan Yang1,2, Hui Li1,2*, Congcong Chai1,2, Zhenkai Xie1,2, Wenjun Wang1,2, Jian-gang Guo1, Xiaolong Chen1,2*
1Beijing National Laboratory for Condensed Matter Physics, Institute of Physics, Chinese Academy of Sciences, Beijing, 100190, China
2University of Chinese Academy of Sciences, Beijing 100049, China
*Corresponding authors: lihui2021@iphy.ac.cn; chenx29@iphy.ac.cn
†These authors contributed equally to this work.
Keywords: wide band gap semiconductor, cubic SiC, high temperature solution growth, high temperature surface tension, solid-liquid interfacial energy
Abstract: Cubic silicon carbide (3C-SiC) has superior mobility and thermal conduction than that of widely applied hexagonal 4H-SiC. Moreover, much lower concentration of interfacial traps between insulating oxide gate and 3C-SiC helps fabricate reliable and long-life devices like metal-oxide-semiconductor field effect transistors (MOSFETs). However, the growth of high quality and wafer-scale 3C-SiC crystals has remained a big challenge up to now despite of decades-long efforts by researchers because of its easy transformation into other polytypes during growth, limiting the development of 3C-SiC based devices. Herein, we report that 3C-SiC can be made thermodynamically favored from nucleation to growth on a 4H-SiC substrate by top-seeded solution growth technique (TSSG), beyond what’s expected by classic nucleation theory. This enables the steady growth of high-quality and large-size 3C-SiC crystals (2~4-inch in diameter and 4.0~10.0 mm in thickness) sustainable. The as-grown 3C-SiC crystals are free of other polytypes and have high crystalline quality. Our findings broaden the mechanism of hetero-seed crystal growth and provide a feasible route to mass production of 3C-SiC crystals, offering new opportunities to develop power electronic devices potentially with better performances than those based on 4H-SiC.
1. Introduction
Silicon carbide (SiC) is an important semiconductor material for fabricating power electronic devices that exhibit higher switch frequency, lower energy loss and substantial reduction both in size and weight in comparison with its silicon (Si) based counterparts.[1-4] Currently, most devices, such as metal-oxide-semiconductor field effect transistors (MOSFETs), which are core devices used in electric vehicles, photovoltaic industry and other applications, are fabricated on a hexagonal polytype 4H-SiC because of its commercial availability.[5, 6]Cubic silicon carbide (3C-SiC), the only cubic polytype, has a moderate band gap of 2.36 eV at room-temperature, but a superior mobility and thermal conduction than that of 4H-SiC.[4, 7-12]Moreover, the much lower concentration of interfacial traps between insulating oxide gate and 3C-SiC helps fabricate reliable and long-life devices.[8-11, 13-15] The growth of 3C-SiC crystals, however, has remained a big challenge up to now despite of decades-long efforts by researchers because of its easy transformation into other polytypes during growth,[16-20]limiting the development of 3C-SiC based devices.
The physical-vapor-transport (PVT) method is the state-of-the-art technique for growing hexagonal 4H- and 6H-SiC crystals. This involves heating raw SiC powder above 2000 ℃ to produce gas species containing Si and carbon (C), which are then transported to a cold end where crystallization occurs on a seed crystal.[21, 22]This process, however, does not work well when it comes to grow cubic 3C-SiC as a higher Si/C ratio in gas species is required. A modified method, close-space PVT, by which a high enough Si/C ratio can be created by separation of raw SiC powder and seed 1~2 mm, allows to grow 3C-SiC.[23, 24] This is not a practical pathway to mass production considering the very limited thickness (<1 mm). Recently, a reduction in defects for 3C-crystals can be achieved by further PVT on a free-standing single crystal first prepared by chemical vapor deposition on Si substrates.[25]But the grown SiC boules’ thickness and the efficiency are still problematic towards mass production of wafers although in-situ switch between the two involved growth methods is feasible.[26] In addition, early attempts to grow 3C-SiC from high temperature melts are not successful either on 6H- or 4H-SiC seeds because these two polytype inclusions always coexist along the grown 3C-SiC.[19]Alternatively, 3C-SiC films are directly deposited on Si substrate then further process into devices on it. But the large lattice mismatch (~19%) and thermal expansion mismatch (~8%) between 3C-SiC and Si result in too high density of defects, significantly deteriorating the performances of devices.[8] Thus, available technique is highly desired to production of high quality and wafer-scale 3C-SiC single crystals.
Structurally, 3C-SiC differs from 4H-SiC in the stacking of identical Si-C bilayers.[27] In 3C-SiC, the bilayers are stacked as a crystallographic plane (111) in the sequence of ABC.[27] In contrast, in 4H-SiC, the bilayers are stacked as a (0001) plane in the sequence of ABCB.[27] The two stacking ways do not cause a significant difference in formation energy, typically a few meV per formula higher for 3C-SiC than for 4H-SiC at zero Kelvin.[27] At temperatures around 1727 ℃, the energy difference between 3C-SiC and 4H-SiC widens to about 5~10 meV per formula, enhancing the stability of 4H-SiC further. However, it is not clear why 3C-SiC is often found as inclusions in 4H-SiC films deposited at around 1650 ℃. Ramakers et al. [28] proposed that surface energy plays a crucial role in stabilizing 3C-SiC over 4H-SiC and 6H-SiC, as the former has surface energy that is 20~150 meV per SiC lower than the latter two. This also means that the 3C polytype may be energetically favored over a certain temperature range if surface energy contributes significantly to the change in the overall formation energy, which depends on different surface reconstruction configurations.
In this work, we smartly modified the surface tension of the melt via nitrogen (N) incorporation for the growth of 3C-SiC via TSSG, effectively adjusting the solid-liquid interfacial energy of SiC and melt. By decreasing the solid-liquid interfacial energy of 3C-SiC/melt and increasing that of 4H-SiC/melt, the nucleation and growth of 3C-SiC is more energetically favorable than that of 4H-SiC on a 4H-SiC substrate. In this way, high quality and wafer-scale 3C-SiC single crystals (2~4-inch in diameter and 4.0~10.0 mm in thickness) are successfully grown via TSSG. Our work provides a feasible route to mass production of 3C-SiC crystals and facilitates the development of new power electronic devices potentially with better performances that are widely used in electric vehicles and photovoltaic industry.
2. Results and Discussion
2.1. Considerations on Stabilizing and Growth of 3C- over 4H-SiC
We start off our exploration of growing 3C-SiC single crystals by employing the TSSG technique. Our strategy is based on two primary considerations. First, the interfacial energy between SiC and melts can be more easily adjusted through altering their chemical compositions in TSSG in comparison to PVT, in which only interface between SiC and gaseous phase exists. Liquid phases are generally thought to be has a more significant effect in changing the interfacial energy than gaseous phases do.[29] It is possible to achieve a lower enough interfacial energy for 3C- over 4H-SiC, which will prioritize the nucleation and subsequent growth for former, and suppress that for the latter. Second, 4H-SiC crystals larger than 4-inch have been successfully obtained by TSSG at 1700~1800 ℃.[30] In this work, we demonstrate that our strategy works well and bulky 3C-SiC crystals up to 4-inch in diameter and more than 4.0 mm in thickness are successfully grown.
Figure 1a shows the schematic setup for growing 3C-SiC by TSSG. Crucibles made from high purity graphite serve as both container and carbon source. Inside the crucible, a melt temperature gradient is set as 5~15 ℃/cm with a temperature of top melt at about 1850 ℃ by induction heating. The melt is usually composed of Cr, Ce and Si, which become a liquid above 1680 ℃ (Figure S1) and act a flux having a solubility of C depending on temperature and composition. Three basic steps are involved in the growth process: 1) the flux dissolves the crucible bottom and 10~15 at. % C enter the flux,[30] 2) thermal conventions convey these C atoms from the bottom to top, and 3) the C and Si atoms combine and crystallize onto the seed as SiC crystal where the temperature is several to a dozen of degrees lower, see Figure 1b. The stable growth of SiC crystal requires the C flow is at equilibrium among these three steps. In a typical run, we use commercial semi-insulating 4H-SiC (0001) wafer as seed crystal and the growth is performed under a mixed Ar/N2 gas.
For a typical vicinal (0001) surface, the Gibbs free energy change (\(\Delta G_{h\text{omo}}\)) for the formation of a two-dimensional 4H-SiC nucleus with a radius of \(r\) on a 4H-SiC step terrace is:
\(\Delta G_{h\text{omo}}=\pi r^{2}h\text{Δg}+2\pi rh\sigma_{4\text{Hside}}\)(1)
In comparison, if a two-dimensional 3C-SiC nucleus on a 4H-SiC step terrace, the change of Gibbs free energy (\(\Delta G_{h\text{etero}}\)) is:
\(\Delta G_{h\text{etero}}=\pi r^{2}h\text{Δg}+2\pi rh\sigma_{3\text{Cside}}+\pi r^{2}(\sigma_{3C/\text{melt}}-\sigma_{4H/\text{melt}})+{\pi r^{2}\sigma}_{3C/4H}\)(2)
where Δg is the Gibbs free energy change from liquid to solid per volume; \(\sigma_{4\text{Hside}}\), \(\sigma_{3\text{Cside}}\),\(\sigma_{4H/\text{melt}}\), \(\sigma_{3C/\text{melt}}\) the interfacial energies between lateral surfaces, (0001), (111) facets to melts for 4H- and 3C-SiC, respectively; \(\sigma_{3C/4H}\) the interfacial energy for (0001) and (111) crystallographic planes between the two polytypes;\(h\) the height of the nucleus.
It is reasonable to assume that\(\sigma_{4\text{Hside}}\approx\sigma_{3\text{Cside}}\) because these lateral surfaces form from stacking Si-C bilayers in a similar spacing but in a different sequence, their surface energies will approach equal if averaging the fluctuations of interactions at a macro-scale.\(\sigma_{3C/4H}\approx\ \)0 is a reasonable assumption because of the negligible lattice mismatch between 4H-(0001) and 3C-(111). Therefore, the \(\Delta G_{h\text{etero}}\) is always smaller than the\(\Delta G_{h\text{omo}}\) if the\(\sigma_{3C/\text{melt}}-\sigma_{4H/\text{melt}}<0\). This means that nucleation and crystal growth are favored for 3C- than for 4H- if the difference between\(\Delta G_{h\text{omo}}-\Delta G_{h\text{etero}}\) is large enough. It is expected that 3C- nucleation easily occurs on 4H- substrate and its step flow is faster than that for 4H-, leading to the total coverage of 3C- on 4H- substrate. Then the growth of 3C- will proceed steadily. Figure 1c schematically describes the possible route for the phase transition starting from preferential hetero-nucleation to subsequent growth for 3C-SiC single crystal on the condition that it has a lower enough interfacial energy with melts. In this study, it is found that the \(\sigma_{3C/\text{melt}}-\sigma_{4H/\text{melt}}\) is negative enough when N2 partial pressures (\(p_{N_{2}}\)) above the melt is above 15 kPa, justifying the above arguments and expectations. Figures 1d-f and Figures S2a, b show the photographs for 2~4-inch 3C-SiC crystal boules grown under \(p_{N_{2}}\)of 20 kPa, respectively. The thickness varies between 4.0~10.0 mm in an 84 h-long growth duration (Table 1). The growth rate is about 50~113 μm/h, a little bit lower than 150 μm/h for the PVT method.[21] The 1 mm thick wafers are black in color (Figure S2) because of high carrier density introduced by N-doping. It is green color under strong light (Figure 1g).[31]