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]