Zhi Li1, Jincheng Zhuang1, Lan Chen2, Zhenyi Ni3, Chen Liu4, Li Wang1, Xun Xu1, Jiaou Wang4, Xiaodong Pi3, Xiaolin Wang1, Yi Du1, Kehui Wu2, Shi Xue Dou1. 1. Institute for Superconducting and Electronic Materials (ISEM), Australian Institute for Innovative Materials (AIIM), University of Wollongong , Wollongong, New South Wales 2525, Australia. 2. Institute of Physics, Chinese Academy of Sciences , Haidian District, Beijing 100080, China. 3. State Key Laboratory of Silicon Materials and Department of Materials Science and Engineering, Zhejiang University , Hangzhou 310027, China. 4. Beijing Synchrotron Radiation Facility, Institute of High Energy Physics, Chinese Academy of Sciences , Beijing 100049, People's Republic of China.
Abstract
Interlayer interactions perturb the electronic structure of two-dimensional materials and lead to new physical phenomena, such as van Hove singularities and Hofstadter's butterfly pattern. Silicene, the recently discovered two-dimensional form of silicon, is quite unique, in that silicon atoms adopt competing sp(2) and sp(3) hybridization states leading to a low-buckled structure promising relatively strong interlayer interaction. In multilayer silicene, the stacking order provides an important yet rarely explored degree of freedom for tuning its electronic structures through manipulating interlayer coupling. Here, we report the emergence of van Hove singularities in the multilayer silicene created by an interlayer rotation. We demonstrate that even a large-angle rotation (>20°) between stacked silicene layers can generate a Moiré pattern and van Hove singularities due to the strong interlayer coupling in multilayer silicene. Our study suggests an intriguing method for expanding the tunability of the electronic structure for electronic applications in this two-dimensional material.
Interlayer interactions perturb the electronic structure of two-dimensional materials and lead to new physical phenomena, such as van Hove singularities and Hofstadter's butterfly pattern. Silicene, the recently discovered two-dimensional form of silicon, is quite unique, in that silicon atoms adopt competing sp(2) and sp(3) hybridization states leading to a low-buckled structure promising relatively strong interlayer interaction. In multilayer silicene, the stacking order provides an important yet rarely explored degree of freedom for tuning its electronic structures through manipulating interlayer coupling. Here, we report the emergence of van Hove singularities in the multilayer silicene created by an interlayer rotation. We demonstrate that even a large-angle rotation (>20°) between stacked silicene layers can generate a Moiré pattern and van Hove singularities due to the strong interlayer coupling in multilayer silicene. Our study suggests an intriguing method for expanding the tunability of the electronic structure for electronic applications in this two-dimensional material.
Low-energy electronic properties
of few-layer two-dimensional (2D)
materials are known to be strongly dependent on the stacking arrangement.[1−3] Twisted bilayers in Dirac Fermion systems, e.g., graphene,[4] are readily observed, which induce a crossover
of Dirac cones that is attributed to rotation of the Brillouin zone
(BZ). With interlayer coupling which ensures electron hopping between
adjacent layers, the emergence of two saddle points in the band structure
due to the overlaid Dirac cones gives rise to logarithmic van Hove
singularities (vHs) in the density of states (DOS). When the vHs is
close to the Fermi level (EF), its magnified
DOS results in electronic instability and consequently causes new
phases of matter with desirable properties, for example, superconductivity,
magnetism, and density waves.Silicene,[5−12] a silicon-based Dirac Fermion material, has attracted extensive
interest since its discovery due to its massless Dirac Fermion characteristics,[7,10] strong spin–orbital coupling (SOC),[13] and its great potential in electronic applications.[8] The successful fabrication[6,14−16] of silicene subverts the conventional wisdom on hybridization by
proving that silicon atoms can form an sp2-sp3 hybridized state and crystallize into a 2D structure. Although recent
scanning tunneling microscopy (STM) and Raman studies have demonstrated
that the sp3 component is much relaxed by the low-buckled
structure in silicene,[7,17] relatively strong interlayer
coupling is still expected in multilayer silicene in contrast to the
other 2D layered materials, such as graphene and boron nitride. How
this strong interlayer interaction perturbs the electronic structure
of multilayer silicene, and consequently, leads to new physical phenomena
has been rarely studied to date.In this work, we report that
√3 × √3 multilayer
silicene consists of the intrinsic 1 × 1 honeycomb structure
with a strong interlayer coupling. The Moiré pattern and vHs
were generated by an interlayer twisting in multilayer silicene with
a rotation angle of 21.8°. Our results show that a Moiré
superlattice gives rise to a periodic potential, which can modulate
the electronic dispersion at vHs.All silicene films investigated
in this work were grown in a preparation
chamber with base pressure ∼5 × 10–11 Torr in a commercial low-temperature scanning tunneling microscopy
system (LT-STM, Unisoku Co.). Clean Ag(111) substrates were prepared
by argon ion sputtering followed by annealing to 820 K for several
cycles. The silicene monolayers were then grown on the Ag(111) surfaces
by evaporation of silicon atoms from a heated silicon wafer with the
substrate kept at 470 K. The detailed growth method has been reported
elsewhere.[17−20] The differential conductance, dI/dV, spectra were acquired by using a standard lock-in technique with
a 10 mV modulation at 613 Hz. All the measurements were carried out
in an ultrahigh vacuum (UHV) at 77 K. Angle-resolved photoemission
spectroscopy (ARPES) characterizations were performed at photoelectron
spectroscopy station, Institute of High Energy Physics, Chinese Academy
of Sciences. A monochromatized He–I light source (21.2 eV)
was used for the band dispersion measurements.Density functional
theory (DFT) calculations were performed by
using the Vienna ab initio simulation package (VASP)
with the projector-augmented wave (PAW) method.[21,22] The Perdew–Burke–Ernzerhof (PBE) correlation exchange
functional at the GGA level was adopted. A Γ-centered (11 ×
11 × 1) Monkhorst–Pack k-point grid was adopted in structural
relaxation and energy calculations. All the calculations were carried
out until the change in energy and the force on each atom were less
than 1 × 10–6 eV per cell and 1 × 10–5 eV/Å, respectively.To investigate the
electronic properties of multilayer silicene,
we prepared a sample of multilayer silicene (thickness >5 monolayers)
on a Ag(111) surface by molecular beam epitaxial (MBE). Figure a presents an STM image indicating
the morphology of the multilayer silicene on the Ag(111) surface.
A well-defined √3 × √3 honeycomb structure is demonstrated
in the top layer of silicene, as shown in Figure b. Nevertheless, the √3 × √3
honeycomb lattice only consists of the upper buckled silicon atoms,
and the detailed structure of the lower silicon atoms is missing.
The higher resolution STM images are required to resolve all silicon
atoms in silicene lattice. The formation of the “supersharp”
tip composed by p-wave molecule is used as the critical technique
for the resolution enhancement of STM images, because spatial derivatives
of the sample wave functions will be resolved by the p-wave tip state.[23,24] Recently, by applying this technique on the p-wave atom Si, subatomic
resolution of Si(111)-(7 × 7) surface is achieved by a Si-terminated
tip.[25] In our experiment, in order to obtain
the Si-terminated tip, we use an extremely small tip–sample
distance to introduce a controllable contact between tip and silicene
surface during scanning. A new buckled honeycomb arrangement with
smaller lattice parameter (a ≈ 0.38 nm), which
is close to the periodicity of 1 × 1 silicene, is explored and
shown in Figure c
after the formation of Si-terminated tip. Furthermore, the Si atoms
in one √3 × √3 unit cell have three different apparent
heights, as shown in Figure d, which is identical to the predicted ABA̅ model for
1 × 1 silicene in previous work.[7] Thus, Figure c displays the integrated
atomic arrangement of 1 × 1 silicene, where the lower-buckled
silicon atoms were also resolved. It should be noted that √3
× √3 Ag on Si(111) substrate is isostructural to the √3
× √3 silicene superstructure, leading to the debate on
the presence or absence of √3 × √3 silicene phase.[26] Resolved lower-buckled silicon atoms in silicene
provide a solid evidence to clarify this debate. The honeycomb-chained
triangle (HCT) model[27] and inequivalent
triangle (IET) model[28] were used to describe
the atomic structure of the Si(111)-√3 × √3 Ag
surface, where each protrusion in STM images is contributed by one
Ag trimer. However, these Ag trimmers do not exist on √3 ×
√3 silicene from atomic resolution STM image (Figure c), excluding the possibility
of Ag-terminated Si(111) reconstruction. More results of the distinct
features between √3 × √3 silicene and √3
× √3 Ag reconstruction are identified by STS, insitu Raman spectra and surface reactivity
investigation (see Supporting Information).
Figure 1
√3 × √3 silicene film and 1 × 1 honeycomb
structure. (a) Typical STM topographic image (50 nm × 50 nm, V = −1 V, I = 5 nA) of the multilayer
silicene film. (b) STM image of √3 × √3 silicene
structure (4 nm × 4 nm, V = −3 mV, I = 4 nA). Only the topmost atoms in the buckled silicene
layer are observed, which are arranged in a honeycomb structure. The
dashed black rhombus marks a √3 × √3 unit cell.
(c) High-resolution STM image reveals that the √3 × √3
silicene is constructed from the 1 × 1 honeycomb structures (as
indicated by the yellow honeycomb, 1 × 1 HC). The black dashed
line represents the √3 × √3 unit cell (√3
× √3 UC). The ABA̅ buckled structure is reflected
by the brightness of the Si atoms. The yellow, blue, and red balls
labeled in the √3 × √3 silicene unit cell denote
the top, middle, and bottom silicon atoms, respectively (2 nm ×
2 nm, V = −3 mV, I = 4 nA).
(d) Height profile corresponding to the red dashed line in c. The
inset is a side view of the ABA̅ buckled structure corresponding
to the height profile.
√3 × √3 silicene film and 1 × 1 honeycomb
structure. (a) Typical STM topographic image (50 nm × 50 nm, V = −1 V, I = 5 nA) of the multilayer
silicene film. (b) STM image of √3 × √3 silicene
structure (4 nm × 4 nm, V = −3 mV, I = 4 nA). Only the topmost atoms in the buckled silicene
layer are observed, which are arranged in a honeycomb structure. The
dashed black rhombus marks a √3 × √3 unit cell.
(c) High-resolution STM image reveals that the √3 × √3
silicene is constructed from the 1 × 1 honeycomb structures (as
indicated by the yellow honeycomb, 1 × 1 HC). The black dashed
line represents the √3 × √3 unit cell (√3
× √3 UC). The ABA̅ buckled structure is reflected
by the brightness of the Si atoms. The yellow, blue, and red balls
labeled in the √3 × √3 silicene unit cell denote
the top, middle, and bottom silicon atoms, respectively (2 nm ×
2 nm, V = −3 mV, I = 4 nA).
(d) Height profile corresponding to the red dashed line in c. The
inset is a side view of the ABA̅ buckled structure corresponding
to the height profile.In addition to the √3 × √3 structure,
we also
observed a pattern with a large period of about 1.7 nm on the top
layer of the silicene film shown in Figure c, which has never been found in a single
layer of silicene. Considering that the √3 × √3
lattice is still resolved in STM image (Figure e), we infer that this larger pattern is
likely to be a Moiré pattern due to the lattice mismatch between
the top layer and the underlying layer of silicene. If interlayer
twisting induces the Moiré pattern, the relationship between
the twisted lattices and the Moiré pattern can be described
by the Moiré equations: D = a/[2 sin(θ/2)],[29] where D is the periodicity of the Moiré pattern, a is the lattice constant of √3 × √3 silicene,
which is 0.64 nm, and θ is the interlayer rotation angle in
twisted √3 × √3 silicene layers. Taking account
of the Moiré periodicity of 1.7 nm, the value of angle θ
is estimated to be around 21.8°. Furthermore, twisted angle θ
could be related to the misorientation angle between the √3
× √3 silicene lattice and the Moiré superlattice
ϕ as ϕ = 30° – θ/2. By using the value
of θ = 21.8° obtained above, ϕ could be estimated
to be around 19.1°, which is consistent with the experimental
observation as shown in Figure d. Therefore, this angle confirms that the observed Moiré
pattern on multilayer silicene originates from interlayer twisting.
It is worth mentioning that 21.8° is the only observed twisting
angle in our STM measurement, which may be the most energy favorable
twisting angle in multilayer silicene. Interestingly, the protuberance
and valley areas in the Moiré pattern can swap their positions
under STM scanning, as demonstrated in Figure f,g. The abrupt change in the protuberance
and valley areas might result from the dynamic flip-flop motion (see Supporting Information), which is possibly induced
by a structural buckling transition in √3 × √3
silicene.[30]
Figure 2
Moiré pattern
in twisted √3 × √3 multilayer
silicene. (a) Morphology of multilayer silicene (30 nm × 30 nm, V = −1.5 V, I = 100 pA). Regions
1 and 2 indicate a Moiré region and a √3 × √3
region, respectively. (b) Schematic diagram of Moiré pattern
induced by interlayer twisting in multilayer silicene with a rotation
angle (θ) of 21.8°. (c) Enlarged STM image of Region 1
in (a) shows a Moiré superlattice with a lattice constant of
1.7 nm. The Moiré unit cell is outlined by the black rhombus
(6 nm × 6 nm, V = −0.5 V, I = 100 pA). (d) STM image indicates a misorientation angle (ϕ)
of 19.1° between the √3 × √3 and the Moiré
lattices. (e) Enlarged STM image of Region 2 in (a) demonstrates a
√3 × √3 structure in adjacent layers without interlayer
rotation. The unit cell of √3 × √3 silicene is
outlined in green. (f) and (g) Flip-flop behavior of the Moiré
pattern in a continuous STM scan, which demonstrates the swapping
of protuberances and valleys in the Moiré superlattice, as
marked by the arrows (20 nm × 20 nm, V = −0.5
V, I = 100 pA).
Moiré pattern
in twisted √3 × √3 multilayer
silicene. (a) Morphology of multilayer silicene (30 nm × 30 nm, V = −1.5 V, I = 100 pA). Regions
1 and 2 indicate a Moiré region and a √3 × √3
region, respectively. (b) Schematic diagram of Moiré pattern
induced by interlayer twisting in multilayer silicene with a rotation
angle (θ) of 21.8°. (c) Enlarged STM image of Region 1
in (a) shows a Moiré superlattice with a lattice constant of
1.7 nm. The Moiré unit cell is outlined by the black rhombus
(6 nm × 6 nm, V = −0.5 V, I = 100 pA). (d) STM image indicates a misorientation angle (ϕ)
of 19.1° between the √3 × √3 and the Moiré
lattices. (e) Enlarged STM image of Region 2 in (a) demonstrates a
√3 × √3 structure in adjacent layers without interlayer
rotation. The unit cell of √3 × √3 silicene is
outlined in green. (f) and (g) Flip-flop behavior of the Moiré
pattern in a continuous STM scan, which demonstrates the swapping
of protuberances and valleys in the Moiré superlattice, as
marked by the arrows (20 nm × 20 nm, V = −0.5
V, I = 100 pA).Besides the Moiré pattern in the lattice structure,
the
twisting structure of multilayer silicene also leads to periodical
variations of DOS in real space mapping. Figure shows experimental and theoretically calculated
DOS distribution at different energies for the Moiré pattern
of twisted silicene multilayer. It is clear that at an energy lower
than the Fermi level (Figure a,e), a hexagonal pattern of potential occurs. The centers
of the hexagons show the highest density of states, while the edges
show the lowest. However, the distribution of DOS is inverted at a
energy higher than the Fermi level (e.g., Figure c,d,g,h). This is consistent with DOS variation
of graphene on boron nitride (BN),[29] which
also shows invert contrast between positive and negative bias mapping.
Near the Fermi level (Figure b,f), a combined feature of hexagonal and honeycomb lattices
indicates a transition between these two types of DOS distribution,
in contrast to random distributed electron and hole puddles of graphene
on SiO2,[31] where many charged
impurities are located at interface between graphene and SiO2. Exotic electronic properties can be induced by interlayer rotation
in 2D layered materials, for instance, van Hove singularities (vHs)
in twisted graphene layers[1] and a new set
of Dirac Fermions in single layer graphene on BN.[32]Figure shows scanning tunneling spectra (STS) collected on the Moiré
pattern in the twisted silicene layers. As shown in Figure b, the bump at −1.0
V in STS is attributed to a flat band.[33] The Dirac point at −0.1 V and flat band are observed in both
Moiré and √3 × √3 silicene regions. Two
prominent peaks at −1.2 and 0.75 V can be seen in the STS spectrum
taken from the Moiré regions, but are absent in the √3
× √3 silicene regions without interlayer twisting. The
positions of these two peaks are slightly deviated from symmetry with
respect to the Dirac point (−0.1 V), which is most likely due
to tip induced band bending.[34,35] We attribute these
peaks to interlayer-rotation-induced vHs in the DOS. The mechanism
of interlayer-twisting-induced vHs is illustrated in Figure c,d. In twisted silicene layers,
the Dirac cones corresponding to each layer are centered at different
points in reciprocal space (indicated as K1 and K2) due to rotation of the Brillouin
zone (BZ) with the same twisting angle (θ). Owing to interlayer
electron hopping, the overlapping Dirac cones generate two saddle
points that are symmetric with respect to the Dirac point. As a result,
they give rise to two vHs peaks in the DOS. Figure e is a line profile consisting of 50 spectra
collected along the dashed arrow in Figure a, which shows the spatial distribution of
the vHs. It shows that the vHs peaks at 0.75 V and −1.2 V in
the twisted silicene can be effectively modulated by the periodic
Moiré potential. The asymmetric peak intensity may result from
the influence of the third layer of silicene, similar to the case
of multilayer graphene.[1]
Figure 3
Experimental STS mappings
and theoretical images of DOS for moiré
pattern of twisted silicene multilayer at different energies. (a,
e) −0.5 V, (b, f) −0.1 V, (c, g) 0.4 V, (d, h) 0.5 V.
The size of all images is 6 nm. (a–d), Experimental STS mappings
for moiré pattern at different bias voltages. STS mapping below
and above Fermi surface show hexagonal and honeycomb lattice, respectively.
(e–h) Corresponding theoretical calculated DOS distributions.
The scales bars in all images are 2 nm.
Figure 4
Van Hove singularities in DOS originate from interlayer twisting
in multilayer silicene. (a) STM image of Moiré pattern (V = 1 V, I = 100 pA). (b) STS carried out
in the Moiré region (Figure c; set point: V = 1 V, I = 100 pA) and in the √3 × √3 silicene region
(Figure e; set point: V = 1 V, I = 100 pA). (c) and (d) Illustrations
of vHs that originates from twisted silicene layers. The interlayer
twisting with the angle θ leads to a rotation of the BZ with
the same angle. The Dirac cones corresponding to K1 and K2 are offset to each
other by ΔK. The shifted Dirac cones create
saddle points that induce two vHs peaks in the DOS. (e) 50 spectra
collected along the black arrow in (a) (from bottom to top; set point: V = 1 V, I = 100 pA).
Experimental STS mappings
and theoretical images of DOS for moiré
pattern of twisted silicene multilayer at different energies. (a,
e) −0.5 V, (b, f) −0.1 V, (c, g) 0.4 V, (d, h) 0.5 V.
The size of all images is 6 nm. (a–d), Experimental STS mappings
for moiré pattern at different bias voltages. STS mapping below
and above Fermi surface show hexagonal and honeycomb lattice, respectively.
(e–h) Corresponding theoretical calculated DOS distributions.
The scales bars in all images are 2 nm.Van Hove singularities in DOS originate from interlayer twisting
in multilayer silicene. (a) STM image of Moiré pattern (V = 1 V, I = 100 pA). (b) STS carried out
in the Moiré region (Figure c; set point: V = 1 V, I = 100 pA) and in the √3 × √3 silicene region
(Figure e; set point: V = 1 V, I = 100 pA). (c) and (d) Illustrations
of vHs that originates from twisted silicene layers. The interlayer
twisting with the angle θ leads to a rotation of the BZ with
the same angle. The Dirac cones corresponding to K1 and K2 are offset to each
other by ΔK. The shifted Dirac cones create
saddle points that induce two vHs peaks in the DOS. (e) 50 spectra
collected along the black arrow in (a) (from bottom to top; set point: V = 1 V, I = 100 pA).In principle, it is well-known that the interlayer coupling
parameter, tθ, depends on the three-dimensional
(3D)
separation parameter[36]R = (r2 + d⊥2)1/2, where r is the spatial
separation projected onto the plane, and d⊥ is the interlayer distance. The interlayer coupling is, therefore,
modulated by the interlayer rotation, since R increases
with increasing interlayer rotation angle (θ). Experimentally,
the interlayer coupling strength can be calculated[2] by tθ = (ℏvFΔK – ΔEvHs)/2. Here, vF is the Fermi
velocity, ΔK is the difference between the
positions of the Dirac cones in reciprocal space, and ΔEvHs is the energy difference between the Dirac
point and the vHs. Taking twisted graphene as an example, with small
interlayer rotation angles (<5°), tθ is about 0.108 eV,[1,2] and consequently, the vHs is preserved.
The interlayer coupling breaks down (tθ ≈ 0 eV) when θ is greater than 15° in twisted
graphene, and thus, the vHs vanishes.[37,38] In the case
of twisted silicene, the observed vHs corresponds to a large interlayer
rotation angle θ = 21.8°. It is worth noting that 21.8°
is the largest rotation angle between two honeycomb lattices that
can produce a commensurate superlattice.[1] Considering that vF is about 5.0 ×
105 ms–1 in twisted multilayer silicene
from ARPES measurements (see Supporting Information), which also show vHs, tθ is calculated
as 0.182 eV in the twisted multilayer silicene sample, which is even
greater than it is in twisted graphene with θ < 5°.
Hence, it reflects the strong interlayer interaction in multilayer
silicene. As observed in STM (Figure ), the buckled ABA̅ structure indicates that
silicon atoms take sp2–sp3 mixed states
in multilayer silicene. The partial sp3 components enhance
electron hopping between adjacent layers, in comparison to the case
of pure sp2 states, which leads to a stronger interlayer
interaction in silicene in contrast to graphene. The energy difference
between vHs is 1.95 eV, which indicates that the optical absorption
of visible light would be achieved in twisted silicene multilayers.
It paves a way to develop ultrathin optical devices by using this
2D Dirac-Fermion material.In conclusion, we have observed the
Moiré pattern and vHs
in twisted multilayer silicene. The existence of the 1 × 1 low-buckled
ABA̅ structure has been confirmed. The silicon sp2-sp3 mixed hybridization states lead to a robust interlayer
interaction in multilayer silicene, which is much stronger than the
interlayer interaction in graphene. It ensures electron hopping between
the twisted silicene layers, even with a large interlayer rotation
angle. The experimental observations suggest a possible way to engineer
electronic properties in multilayer silicene by interlayer twisting.
Authors: A Luican; Guohong Li; A Reina; J Kong; R R Nair; K S Novoselov; A K Geim; E Y Andrei Journal: Phys Rev Lett Date: 2011-03-21 Impact factor: 9.161
Authors: Régis Decker; Yang Wang; Victor W Brar; William Regan; Hsin-Zon Tsai; Qiong Wu; William Gannett; Alex Zettl; Michael F Crommie Journal: Nano Lett Date: 2011-05-09 Impact factor: 11.189
Authors: Patrick Vogt; Paola De Padova; Claudio Quaresima; Jose Avila; Emmanouil Frantzeskakis; Maria Carmen Asensio; Andrea Resta; Bénédicte Ealet; Guy Le Lay Journal: Phys Rev Lett Date: 2012-04-12 Impact factor: 9.161