Content uploaded by Turgut Yilmaz
Author content
All content in this area was uploaded by Turgut Yilmaz on Jun 24, 2020
Content may be subject to copyright.
1
Emergent of the flat band and superstructures in the VSe2 / Bi2Se3 system
Turgut Yilmaz1*, Xiao Tong2, Zhongwei Dai2, Jerzy T. Sadowski2, Eike F. Schwier3, Kenya
Shimada3, Sooyeon Hwang2, Kim Kisslinger2, Konstantine Kaznatcheev1, Elio Vescovo1, and
Boris Sinkovic4
1National Synchrotron Light Source II, Brookhaven National Lab, Upton, New York 11973, USA
2Center for Functional Nanomaterials, Brookhaven National Lab, Upton, New York 11973, USA
3 Hiroshima Synchrotron Radiation Center, Hiroshima University, 2-313 Kagamiyama, Higashi Hiroshima 739-
0046, Japan
4 Department of Physics, University of Connecticut, Storrs, Connecticut 06269, USA
*tyilmaz@bnl.gov
Dispersionless flat bands are proposed to be a fundamental ingredient to achieve the various
sought after quantum states of matter including high-temperature superconductivity1-4 and
fractional quantum Hall effect5-6. Materials with such peculiar electronic states, however,
are very rare and often exhibit very complex band structures. Here, we report on the
emergence of a flat band with a possible insulating ground state in the sub-monolayer VSe2
/ Bi2Se3 heterostructure by means of angle-resolved photoemission spectroscopy and
scanning tunneling microscopy. The flat band is dispersionless along the kll and kz momenta,
filling the entire Brillouin zone, and it exhibits a complex circular dichroism signal reversing
the sign at several points of the Brillouin zone. These properties together with the presence
of a Moiré patterns in VSe2 suggest that the flat band is not a trivial disorder or confinement
effect and could even be topologically non-trivial. Another intriguing finding is that the flat
band does not modify the Dirac cone of Bi2Se3 around the Dirac point. Furthermore, we
found that the flat band and the Dirac surface states of Bi2Se3 have opposite energy shifts
2
with electron doping. This opens a novel way of controlling the spin texture of photocurrents
as well as the transport properties of the heterostructure. These features make this flat band
remarkably distinguishable from previous findings and our methodology can be applied to
other systems opening a promising pathway to realize strongly correlated quantum effects
in topological materials.
The physics of solids is determined by their energy band structures. Therefore, investigation and
controlling distinct electronic band dispersions are of great importance in condensed matter to
understand and discover new states of the matter. One of the exotic electronic states is a type of
flat band which is predicted to host high-temperature superconductivity1-4, fractional quantum Hall
effect (FQHE)5-6, and ferromagnetism7-9. In a superconductor, a flat band can boost the coupling
constant and the transition temperature (Tc) because of the enhanced density of states at the Fermi
level (EF)10-11. This was utilized to explain the unexpected superconductivity in rhombohedral
graphite and twisted graphene12-14. Other examples of the flat band materials are Kagome lattices
in which the flat band stems from the destructive quantum interference due to the frustrated lattice
geometry7-8. The Kagome-type flat band was reported in FeSn and Fe3Sn2 by angle-resolved
photoemission spectroscopy (ARPES) and in Co3Sn2S2 by scanning tunneling spectroscopy
(STS)8-10. However, the complexity of the electronic structure in photoemission data and the lack
of momentum resolution in STS make the observations elusive. Another system with the
dispersionless flat electronic states could be the Mott insulators in which the strong correlation
effect forms the flat Hubbard band with an insulating gap the EF as predicted in TaS2-xSex and
NbS2-xSex15-18. The strong correlation effect was also investigated in Bi2Se3 and Bi2Te2Se
topological insulators (TI) and a possible p-band Mott insulating state where the Hubbard bands
were predicted to exist in these samples19-21. The experimental signature of the Mott insulating
3
state was also captured in multilayer graphene Moiré superlattice indicating the strong correlation
effect in that system22. Computational efforts were also focused on designing the flat band
transition metal dichalcogenides (TMDs) Moiré superlattices which could support strongly
correlated physics at higher-temperatures due to the flat bands23-25. Thereby, the previous works
conclude that the flat band media could be fertile to many novel states of the matter. However,
apart from Mott insulators, the limited number of the materials with such non-trivial bands hinders
future studies.
Motivated by earlier studies, we investigate the surface electronic structure of VSe2 TMD grown
on the surface of Bi2Se3 TI and show the emergence of a flat band in the electronic states at low
VSe2 coverage level. This flat band covers the entire kx - ky plane of the Brillouin zone (BZ) and
it displays dispersionless along the kz direction as well. Furthermore, our circular dichroism
ARPES (CD-ARPES) measurements revealed that the CD signal of the flat band reverses the sign
at several points within the BZ. Another notable observation is that VSe2 growth and emergence
of the flat band do not reshape the Dirac cone of Bi2Se3 in the vicinity of the DP unlike the case of
transition metal doping which opens a large gap at the DP26. We also observed Moiré patterns in
VSe2 domains of monolayer (ML) thickness and stripe type patterns in bare Bi2Se3 through
scanning tunneling microscopy (STM). Along with these observations, electron doping impact on
the band structure of the system suggests that the flat band might be due to the Mott insulator like
interaction correlated with the charge density wave (CDW) phase. To further shed light on the
crystalline and chemical properties of the system, scanning transmission electron microscopy
(STEM), micro-spot low energy electron diffraction (µLEED), and x-ray photoemission
spectroscopy (XPS) experiments were conducted. Our results demonstrate rich physics and
suggest a large family of materials of possible emergent flat bands and thus will motivate future
4
studies on materials with superconductivity at high-temperatures, QSL states, or FQHE by using
our approach.
Structural characterization: Bi2Se3 and VSe2 crystals are layered materials with their atomic
stacking geometry shown in Fig. 1a. The layers in each compound are separated by so-called van
der Waals (vdW) gaps with weak covalent out-of-plane bonds connecting the layers. This is the
reason that VSe2 / Bi2Se3 heterostructure can be grown despite the large in-plane lattice mismatch
of around 20 % between the two materials27-28. Figs. 1b and 1c depict the relevant core-levels of
such structures formed by the growth of 0.3 ML and 3 ML VSe2 on 12 quintuple layer (QL) Bi2Se3.
Upon deposition of the VSe2, the spectral shape of the Bi 5d peaks of Bi2Se3 does not exhibit a
prominent modification indicating the absence of V metals at the interface and/or in the bulk (Fig.
1b)29. Compared with pristine Bi2Se3, the Se 3d peak, however, appears at 0.1 eV higher binding
energy upon VSe2 deposition. This can be better seen in the inset of Fig. 1b displaying the scaled
Se 3d peaks to the same height. The difference in binding energy could be due to the CDW phase
of VSe227. V 2p1/2 and 2p3/2 peaks shown in Fig. 1c are located at 513 eV and 520.6 eV binding
energies, being in agreement with recent reports28.
To further explore the system, we show a high-angle annular dark-field (HAADF)-STEM cross-
section image of 3 ML VSe2 / 12 QL Bi2Se3 heterostructure in Fig. 1d. Bi2Se3 and VSe2 exhibit
regular atomic layers with smooth interfaces and vdW gaps marked with red arrows in Fig. 1d. On
the other hand, the STEM data does not exhibit a clear interface spacing between the VSe2 and
Bi2Se3 which could strongly modify the local electronic structure. Furthermore, the STEM energy
dispersive X-ray spectroscopy elemental maps presented in Fig. S1 (supplementary materials)
shows the atomic distribution of Bi in the Bi2Se3 layers, V in VSe2 layers, and Se across the
heterostructure as expected.
5
Fig. 1e and 1f show the STM image of Bi2Se3 and VSe2 regions of a 0.3 ML VSe2 / 12 QL Bi2Se3
sample, respectively. The Bi2Se3 surface has a stripe-like pattern similar to Cs and Fe doped
Bi2Se330. The STM image of the VSe2 domains presented in Fig. 1f exhibits a Moiré pattern with
~ 2 nm x 2 nm superstructure. This differs from the previous studies conducted on VSe2 / graphene
in which such superstructure formation was absent27. Moiré pattern can be formed by a small misfit
between the in-plane lattice parameters of the film and the underlying material or the relative
rotation of two layers to each other, or both. By contrast, the lattice mismatch between the VSe2
and Bi2Se3 is quite large (about 20 %). Unfortunately, we cannot make a quantitative analysis for
precise determination of the in plane lattice parameters or the atomic displacement due to the
limitation in our STM data taken at room temperature experiment. But, similar Moiré pattern
formation is also observed on monolayer MoSe2 grown on a graphene substrate whose origin is
attributed to the lattice mismatch between the multiple unit cells of the two materials31. Thereby,
the Moiré pattern in VSe2 could be formed due to the small mismatch between four-five-unit cells
of Bi2Se3 (4aBS = 16.56 Å or 5aBS = 20.7 Å) and five-unit cells of VSe2 (5aVS=16.8 Å or 6aVS=20.16
Å) for the rotationally aligned lattice geometry. Alternatively, the Moiré pattern could form be
formed by the rotational misalignments of Bi2Se3 and VSe2 atomic lattices. Moreover, the details
of the STM data reveal that the layer heights are 6.8 Å for VSe2 and 9.6 Å Bi2Se3 (Fig. S2,
supplementary materials), which is in agreement with published results32.
ARPES electronic structure: To examine the band structure, the binding energy versus ky plots
are given for 12 QL Bi2Se3 in Fig. 2a and for 0.3 ML VSe2 / 12 QL Bi2Se3 in Fig. 2b. The Bi2Se3
sample exhibits the typical band structure with the linear Dirac surface states (DSSs) forming the
Dirac cone with the Dirac point (DP) at 0.36 eV below EF33. Upon deposition of 0.3 ML VSe2 on
Bi2Se3, a flat band with 0.47 eV binding energy and ~0.18 eV bandwidth emerges in the surface
6
electronic structure (Fig. 2b). The flatness of the band is well distinguished in the energy
distribution curves (EDCs) shown in Fig. 2c. The bulk bands and the DSS of Bi2Se3 strongly
disperse as a function of , while the flat band retains dispersionless across the
high
symmetry lines. VSe2 growth also induces the well-known M-state quantization34 of the bulk
valance band (BVB) of Bi2Se3 shown in Fig. 2b. Moreover, the flat band overlaps with the lower
branch of the Dirac cone in the vicinity of ky = 0 Å-1 without inducing a prominent change in its
spectral shape. This can be even better seen in the films with larger VSe2 coverage confirming that
the flat band, DSSs, and the dispersive V 3d state of VSe2 coexist in the surface electronic structure
(Fig. S3, supplementary materials). Also, ARPES date for the thicker VSe2 coverage suggests that
the flat band is localized on the bare Bi2Se3 and/or at the interface between VSe2 and Bi2Se3 ( Fig.
S4, supplementary materials). We should also note that 0.3 ML VSe2 / 12 QL Bi2Se3 sample shows
an insulating gap at the EF which will be discussed in the following sections of the text.
To further investigate the flat band, a kx – ky intensity plot at EF and the binding energy of the flat
band (EFB) are shown in Fig. 2d and in Fig. 2e, respectively. The Femi surface is dominated by a
circular counter of the DSSs centered at
point. The constant energy cut at EFB reveals another
remarkable feature, namely that the flat band reaches beyond the
and
points in the BZ. This
observation reveals that the flat band fills the entire BZs of Bi2Se3 and VSe2 as depicted by black
and red hexagons in Fig. 2e, respectively. Such electronic state spread over large momentum area
can significantly enhance the electronic correlation yielding quantum effects at very high-
temperatures. It is also worth noting that the LEED pattern of the sample shows rotationally
stretched diffraction spots along the rotational direction indicating the presence of the rotationally
misaligned VSe2 domains ( Fig. S2a, supplementary materials) with respect to each other and to
the Bi2Se3 substrate. The rotational misfit of ±3o estimated from µLEED pattern, however, is too
7
small for a band to span whole BZ and to induce a fully occupied constant energy counter in the
momentum map.
Photon energy-dependent ARPES: In ARPES experiments, changing the photon energy
corresponds to mapping the electronic states along the kz direction of the BZ. By recording the
electronic structure with a wide photon energy range, a k‖ vs. kz or binding energy vs. kz dispersions
can be also extracted. This method allows studying the energy bands along the kz (ou-of-plane)
direction to distinguish the non-dispersive states from the dispersive bulk bands. Such spectra
acquired at varying photon energies (40 eV to 70 eV) for 0.3 ML VSe2 / 12 QL Bi2Se3 sample are
given in Fig. 3. In the plot of kz versus ky dispersion at EF, DSSs marked with dashed red lines
exhibit no kz dependence (Fig. 3a). Similar spectrum at EFB given in Fig. 3b shows that a high
spectral intensity along the ky = 0 Å-1 originates from the bottom of the Dirac cone of Bi2Se3. Away
from the ky = 0 Å-1, the plot has non-vanishing spectral intensity contributed by the flat band. This
can be better seen in the momentum distribution curves (MDCs) obtained at various kz points (Fig.
3c) in which each spectrum exhibits always finite density of states along the ky momentum
direction. This implies the dispersionless nature of the flat band along the kz momentum direction.
To further validate this observation, we show the binding energy - kz plots along the ky = ± 0.3 Å-
1 directions in Fig. 3d and 3e, respectively. The plots clearly show that the flat band at 0.47 eV
binding energy is kz - independent confirming its non-bulk derived nature. We should also note
that the M-shape bulk band located at 1 eV binding energy exhibits a nearly non-dispersive feature
along the kz as shown in Fig. 3d and 3e. To further reveal the details of the flat band, Fig 3f depicts
the EDCs taken at different kz points. One can see that the EDC of the flat band does not exhibit a
kz dependent evolution in the binding energy and bandwidth providing a signature that the flat
band originates from a single band.
8
CD-ARPES: Circular dichroism (CD)-ARPES has gained great attention due to its feasibility to
investigate the helical spin-orbit texture in topological surface states35. The principle of the method
is the spectral weight differences in ARPES arising from the opposite helicity of the circularly
polarized light. CD-ARPES is then obtained from where
and are photoemission intensities for RCP and LCP lights, respectively. Thus, we have
recorded the band structure of 0.3 ML VSe2 / 12 QL Bi2Se3 sample with RCP and LCP, shown in
Fig. 4a and in Fig. 4b, respectively. The corresponding CD-ARPES is presented in Fig. 4c with
red (negative-CD) and blue (positive-CD) color representations. The bulk bands of Bi2Se3
dispersing below 0.8 eV binding energy show a strong DC as seen in Fig. 4a-4c. CD from the
DSSs of Bi2Se3 is also seen switching the spectral weight from the -ky to +ky regions when
changing the excitation from RCP to LCP. For clarity, the DC signal versus ky is plotted in Fig. 4d
at 0.1 eV binding energy in which the CD is positive for left and negative for the right side of the
Dirac cone, marked with vertical arrows. Further away from the ky = 0 Å-1, the plot in Fig. 4d still
shows non-zero CD. This is likely originating from the V 3d orbitals which dominate the EF density
of states in VSe227. Also, since the spectra are conducted with 50 eV photons corresponding to
nearly Z-point of the BZ, the BCBs of Bi2Se3 do not contribute to ARPES spectra in the vicinity
of the EF as seen in Fig. 2a.
To investigate the dichroism effect in the flat band, the CD at EFB is plotted as a function of ky in
Fig. 4e and it exhibits sign inversions at ky = 0 Å -1 and ky = ±0.5 Å -1 and the maxima at ky = ±0.25
Å -1. Similar to the DSSs, the CD in the flat band also exhibits helical-texture where opposite ky
momentums have opposite signs of the CD. Notably, zero CD signal is also observed as white
color in the CD-ARPES along the ky = 0 Å-1. This depicts the nodal line which was proposed to
be the characteristic feature of the 2D electronic structure36. In particular, the CD signal of the
9
DSSs depends on the incident photon energy assigning it to the final state effect in the
photoemission process35. This was discussed in Ref 36 with details where they propose the non-
trivial connection between the spin orbit texture and the CD signal. Thereby, the helical CD texture
and the nodal line band suggest that the flat band could be also topologically non-trivial.
Electron doping impact on the band structure: Tuning the chemical potential in the strongly
correlated systems induces a mass renormalization. This has been seen in superconductors as the
mass enhancement referring to the flattening of the bands37-38 and also exists in the CDW phase of
the excitonic insulators39. Also, this mechanism usually coexists with a metal-insulator transition
in superconductors37. Likewise, the dispersion of the flat band could be controlled through the
electron or hole doping. However, in our case, since the flat band is already dispersionless, so the
mass normalization could be expected. Namely, the flat band can gain weak dispersion upon the
tuning of the EF. To test this idea and to probe the possible unoccupied bands above the Fermi
level, we studied the band structure of the sample under the surface deposition of potassium (K)
where we found further spectroscopic anomalies in 0.3 ML VSe2 / 12 QL Bi2Se3 sample (Fig. 5).
The first notable observation is that K deposition induces a pair of quantum well states (QWS)
shown with black arrows in Fig. 5b. This proves the electron doping induced band bending in the
BVB of Bi2Se3. More interestingly, the leading edge shifts with K doping revealing an 0.08 eV
insulating gap at the EF (Fig. 5a-5b). The shift is more obvious in the integrated EDCs (Fig. 5c)
where sharp edges are observed at different energies before and after K deposition. This indicates
that the gap closing is not due to the broadening or the formation of an additional spectral feature
in the vicinity of the EF. EDCs along the kR and kL momentums where the DSSs cut the EF (Fig.
5a) also exhibit the same insulating gap as shown in Fig. 5e and 5f. This shows that not only VSe2
but also bare Bi2Se3 regions have an insulating gap at the EF since the size of x-ray beam spot on
10
the sample is considerably larger than the size of individual VSe2 domains. This argument is also
supported by the absence of the multiple Dirac cones which are expected to be located at the bare
and the VSe2 covered parts of Bi2Se3 layers. These realizations show that the flat band could be
originating from Bi2Se3 when modified by VSe2 sublayer. Another peculiar anomaly is the
difference of binding energy shifts of bulk bands, DSS, and the flat band, induced by the K
deposition. While Bi2Se3 bulk and Dirac states exhibit the expected electron doping, the flat band
shows negative electron compressibility. The flat band binding energy changes from 0.47 eV
before K doping to a lower value of 0.33 eV after K doping. On the other hand, the binding energy
of the DP shifts from 0.36 eV to 0.55 eV. Similar band bending towards higher binding energy is
also seen in the bulk bands of Bi2Se3 (Fig. 5a-5b). (Note that the binding energies are given with
respect to the highest occupied states in the ARPES map shown in Fig. 5b.) The existence of
opposite energy shifts of the flat band and Dirac bands opens up a pathway to tune both, the binding
energy of the DP as well as the flat band with respect to each other. If hole doping of Bi2Se3 i. e.
via Ca substitution40, was to be included, both DP and the flat band could be tuned to EF which
might yield exotic transport properties. Hence, we have demonstrated the presence of a possibly
unique tuning parameter for the transport properties of such low dimensional topological system.
We should also point out that the ARPES data of the K doped sample in Fig. 5b shows minute
dispersion of the flat band, an apparent decrease of its binding energy when approaching the
point. However, this dispersion is too small for a precise detection but might be related to the
modification of the CDW phase39. Our fittings of the EDCs at different ky momenta (not shown)
yield the energy dispersion to be about 35 meV between ky = 0.15 A-1 and ky = 0.6 A-1 momentum
points. This effect might be further enhanced by tuning the chemical potential of the system which
opens an opportunity for future studies. Nevertheless, the insulating gap and the mass
11
normalization with the electron doping, even if it is small, suggest that the flat band could be
related to strong electron correlation effects.
Discussion: Dispersionless flat bands can be explained by several mechanisms. V 3d derived
magnetic impurity bands are the first mechanism to be considered as the spectroscopic origin of
the flat band41. However, our core level spectra and STM data do not show any signature of isolated
atoms. Also, our photon energy dependent ARPES and the rather complex CD-ARPES spectra
indicate that the flat band should be considered to have non-trivial origin.
Another scenario to consider would be the existence of superlattices as seen in x silicene
superstructure by STS in which the local density of states forms the electronic Kagome lattice42.
The interface dislocation or the strain can also flatten the original bands by introducing pseudo-
magnetic field term to the Hamiltonian in Moiré superstructures4. However, a pronounced band
flattening in this scenario requires superstructure patterns with at least a few tens of nanometers
periodicity which is much larger than one observed in the present case. Furthermore, the flat bands
discussed within the superlattice frameworks are dispersionless only in the BZ of the
superstructure which is smaller than the BZ corresponding to the unitcell25, 42. This mechanism
gives rise to a dispersionless flat band within a small momentum window. On the other hand, the
flat band reported here does not shrink into the BZ of the Moiré or the stripe pattern superstructures
and it is robust against electron doping. Therefore, although we cannot entirely exclude that the
flat band electronic state in VSe2 / Bi2Se3 may be related to the atomic superstructures, there are
other possibilities to consider23-25, 42.
Dispersionless flat bands with an insulating gap at the EF indicate to strong electronic correlation.
This could be explained by considering the Mott insulator or charge-transfer insulating states
which are generally in proximity to CDW phases16. Indeed, the presence of corrugated bright and
12
dark Bi2Se3 domains in the STM image (Fig. 1e) supports the persistence of the CDW phase in
agreement with earlier studies43-44. In the Mott phase, a strong Coulomb repulsion splits d- or f-
bands into the lower (occupied) and upper (unoccupied) Hubbard bands (LHB and UHB) which
are separated with an insulating gap located in the vicinity of EF16. The insulating gap in the Mott
insulator correlated with the CDW phase strongly depends on the chemical potential of the
system15. Thus, the chemical potential could shrink the gap by shifting the LHB and UHB closer
to the EF and each other. In this scenario, the Hubbard bands could exhibit opposite energy shifts
than other bands upon doping15-16. Similarly, a large Hubbard interaction could induce an
antiferromagnetic order and open an energy gap at EF in a topological honeycomb lattice45.
Furthermore, p-type Hubbard bands with an insulating ground state were also predicted in Bi2Se3
and Bi2Te2Se upon enhancing the Coulomb repulsion19-21. The similarities between the Mott
insulators and the VSe2 / Bi2Se3 could suggest that the flat band in our sample could be formed by
the strong correlation effects.
On the other hand, up to the date, the surface electronic structure of TIs has been studied by ARPES
under the deposition of various magnetic and non-magnetic elements, and none of the approaches
revealed the flat band spectral feature46. This controversy could be related to the location of the
dopants in the crystal. The elemental dopants occupy the vdW gap or the sublattice sites in the
bulk47, while VSe2 domains are located on the surface of Bi2Se3. Therefore, the magnetic properties
of the VSe2 could also play a vital role in the formation of the distinct electronic states on the
surface of Bi2Se3 without modifying its bulk band topology48. Nevertheless, the details of the flat
band electronic structure with a realistic scenario are more complicated since non-trivial band
topology and the strong spin-orbit coupling are also expected to play a role in the evolution of the
electronic structure. The complete understanding of the physics behind the flat band and the
13
insulating gap requires a broad and elaborate theoretical and experimental efforts. As a first
example of the dispersionless electronic excitation in a topologically non-trivial band structure,
our results could open a new pathway in the critical field of experimental realization and control
of novel quantum effects and new states of the matters.
Methods:
Material preparation: Molecular beam epitaxial growth (MBE) technique was employed to grow
VSe2 / Bi2Se3 and Bi2Se3 samples in a custom ultra-high vacuum system located at the ESM
beamline of NSLS-II. 5N Se and Bi sources were evaporated from the ceramic crucibles while the
e-beam evaporation method was used for V (99.8% purity) source. All samples were grown on
Al2O3(0001) substrates at 255 oC. Before the growth, the substrates were first degassed at 550 oC
for three hours and flashed at 850 oC for 5 min. Sample thicknesses were estimated within a 15 %
error bar by using a quartz thickness monitor and x-ray photoemission spectroscopy. Samples for
ARPES and µLEED experiments were capped with 20 nm amorphous Se film before being
removed from the MBE chamber.
XPS: Core-levels were recorded with 1486.7 eV non-monochromated x-ray source at room
temperature with a vacuum interlocked MBE-photoemission system at the surface science
laboratory of the Department of Physics of the University of Connecticut. Core-level binding
energies were determined with an error of 0.1 eV.
ARPES: ARPES experiments were performed at 21-ID-1 ESM beamline of National Synchrotron
Light Source II (NSLS-II) by using a DA30 Scienta electron spectrometer. The pressure in the
photoemission chamber was 1x10-10 Torr and samples were kept at 15 K during the experiment by
a closed-cycle He cryostat. The energy resolution in the ARPES experiments was better than
14
15meV with a spot size of ~20 µm. Before the ARPES experiments, samples were annealed at 220
oC for 30 minutes to remove the Se capping layer. The angle between the light and the surface
normal of the sample is 55◦ at the normal emission during the ARPES experiments. The films were
grounded with a tantalum clip. A part of the ARPES experiments was conducted at the linear
undulator beamline at the Hiroshima Synchrotron Radiation Center (HiSOR) BL-1 [S1]. Photon
energy is converted to momentum space by using the free electron final state approximation
where is the free electron mass, is the kinetic energy of a
photoelectron, and is the inner potential which is 11.8 eV for Bi2Se3 [S2].
TEM and STM: HAADF-STEM images were acquired with Hitachi HD2700C dedicated STEM
with a probe Cs corrector operating at 200 kV at room temperature. Samples were prepared using
the in-situ lift-out method on the FEI Helios 600 Nanolab dual-beam FIB. Final milling was
completed at 2 keV. Scanning tunneling microscope (Omicron VT- STM -XA 650 ) experiments
were performed in an ultrahigh vacuum (UHV) system with a base pressure of 2 × 10−10 Torr at
room temperature. All the STM images were observed in constant current mode using Pt/ Ir tips.
All bias values in the text refer to the bias applied to the sample. The STM images were analyzed
using Gwyddion-2.55 software package. HAADF-STEM and STM experiments were conducted
at the Center for Functional Nanomaterials, Brookhaven National Laboratory. Samples for STM
were transferred with a vacuum suitcase.
LEED: µLEED experiment was performed at XPEEM/LEEM endstation of the ESM beamline
(21-ID-2).
Acknowledgment: This research used ESM (21-ID-1, 21-ID-2) beamline of the National
Synchrotron Light Source II, a U.S. Department of Energy (DOE) Office of Science User Facility
15
operated for the DOE Office of Science by Brookhaven National Laboratory under Contract No.
DE-SC0012704. This work also used the resources of the Center for Functional Nanomaterials,
Brookhaven National Laboratory, which is supported by the U.S. Department of Energy, Office
of Basic Energy Sciences, under Contract No. DE-SC0012704. ARPES experiments in Hiroshima
were performed with the approval of program advisory committee of HISOR) Proposal No.
19BG041). Author T. Y. would like to thank Prof. A. V. Balatsky for useful discussions.
Author Contributions: T.Y conceived and designed the experiments. T.Y prepared the samples
and performed the photoemission experiments with K. K., E. V, and B. S’s help. E. F. S. performed
the ARPES experiments at HISOR. X. T. conducted the STM measurements. Z.D. and J. T. S.
performed µLEED measurements. S. H. and K. K. performed HAADF-STEM experiments. T. Y
analyzed the experimental results and wrote the manuscript with contribution from E. F.S., B. S.,
K. K., and E. V.
Competing interests: The authors declare that they have no competing interests.
References
1. Heikkilä, T. & Volovik, G. E. Flat bands as a route to high-temperature superconductivity in
graphite, 123–144 (Esquinazi, P. (ed.), (Springer International Publishing AG Switzerland,
2016).
2. Cao, Y. et al. Unconventional superconductivity in magic-angle graphene
superlattices. Nature 556, 43–50 (2018).
3. Heikkilä, T. T., Kopnin, N. B. & Volovik, G. E. Flat bands in topological media. JETP Lett. 94,
233–239 (2011).
16
4. Tang, E. & Fu, L. Strain-induced partially flat band, helical snake states and interface
superconductivity in topological crystalline insulators. Nat. Phys. 10, 964969 (2014).
5. E. Tang, J. W. Mei, X. G. Wen, High-temperature fractional quantum Hall states. Phys. Rev.
Lett. 106, 236802 (2011).
6. Sun, K., Gu, Z., Katsura, H. & Das Sarma, S. Nearly flatbands with nontrivial topology. Phys.
Rev. Lett. 106, 236803 (2011).
7. Guo, H. M. & Franz, M. Topological insulator on the kagome lattice. Phys. Rev. B 80, 113102
(2009).
8. Lin, Z. et al. Flatbands and emergent ferromagnetic ordering in Fe3Sn2 kagome lattices. Phys.
Rev. Lett. 121, 96401 (2018).
9. Kang, M. et al. Dirac fermions and flat bands in the ideal Kagome metal FeSn. Nat. Mater. 19,
163–169 (2020).
10. Yin, J. X. et al. Negative flat band magnetism in a spin–orbit-coupled correlated kagome
magnet. Nat. Phys. 15, 443–448 (2019).
11. Kopnin, N. B., Heikkilä, T. T. & Volovik, G. E. High-temperature surface superconductivity
in topological flat-band systems. Phys. Rev. B 83, 220503(R) (2011).
12. Kopnin, N. B., Ijäs, M., Harju, A. & Heikkilä, T. T. High-temperature surface
superconductivity in rhombohedral graphite. Phys. Rev. B 87, 140503 (2013).
13. Marchenko, D. et al. Extremely flat band in bilayer graphene. Sci. Adv. 4, 11 (2018).
14. Volovik, G. E. Graphite, graphene and the flat band superconductivity. JETP Letters 107, 516–
517 (2018).
15. Qiao, S. et al. Mottness collapse in 1T-TaS2−xSex transition-metal dichalcogenide: an interplay
between localized and itinerant orbitals. Phys. Rev. X 7, 041054 (2017).
17
16. Ang, R. et al. Superconductivity and bandwidth-controlled Mott metal–insulator transition in
1T-TaS2−xSex. Phys. Rev. B 88, 115145 (2013).
17. Pasquier, D. & Yazyev, O.V. Charge density wave phase, Mottness, and ferromagnetism in
monolayer 1T−NbSe2. Phys. Rev. B 98, 045114 (2018).
18. Tresca, C. & Calandra, M Charge density wave and spin 1/2 insulating state in single layer 1T-
NbS2. 2D Materials 6, 035041 (2019).
19. Craco, L. & Leoni, S. Bulk quantum correlations and doping-induced nonmetallicity in the
Bi2Se3 topological insulator. Phys. Rev. B 85, 075114 (2012).
20. Liu, M. et al. Electron interaction-driven insulating ground state in Bi2Se3 topological
insulators in the two-dimensional limit. Phys. Rev. B 83, 165440 (2011).
21. Craco, L. & Leoni, S. Tunable Kondo-Mott physics in bulk Bi2Te2Se topological insulator.
Phys. Rev. B 85, 195124 (2012).
22. Chen, G. et al. Evidence of a gate-tunable Mott insulator in trilayer graphene moiré
superlattice. Nat. Phys. 15, 237–241 (2019).
23. Wu, F., Lovorn, T., Tutuc, E. & MacDonald, A. H. Hubbard model physics in transition metal
dichalcogenide moiré bands. Phys. Rev. Lett. 121, 026402 (2018).
24. Naik, M. H. & Jain, M. Ultraflatbands and shear solitons in moiré patterns of twisted bilayer
transition metal dichalcogenides. Phys. Rev. Lett. 121, 266401 (2018).
25. Bi, Z., Yuan, N. F. Q. & Fu, L. Designing flat bands by strain. Phys. Rev. B 100, 035448
(2019).
26. Yilmaz, T. et al. Distinct Effects of Cr Bulk Doping and Surface Deposition on the Chemical
Environment and Electronic Structure of the Topological Insulator Bi2Se3. Appl. Surf. Sci.
407, 371−378 (2017).
18
27. Chen, P. et al. Unique Gap Structure and Symmetry of the Charge Density Wave in Single-
Layer VSe2. Phys. Rev. Lett. 121, 196402 (2018).
28. Liu, Z. L. et al. Epitaxially grown monolayer VSe2 An air-stable magnetic two-dimensional
material with low work function at edges. Sci. Bull. 63, 419 (2018).
29. L.A. Walsh, L. A. et al. Interface Chemistry of Contact Metals and Ferromagnets on the
Topological Insulator Bi2Se3, Jour. of Phys. Chem. C 121, 23551 (2017).
30. Shokri, R. Coexistence of impurity-induced quasi-one-dimensional electronic structure and
topological surface states of Bi2Se3. J. Appl. Phys. 119, 085304 (2016).
31. Ugeda, M. M. et al. Giant bandgap renormalization and excitonic effects in a monolayer
transition metal dichalcogenide semiconductor. Nature Mater. 13, 1091–1095 (2014).
32. Duvjir, G. et al. Emergence of a Metal–Insulator Transition and High-Temperature Charge-
Density Waves in VSe2 at the Monolayer Limit. Nano Lett. 18, 5432–5438 (2018).
33. Xia, Y. et al. Observation of a large-gap topological-insulator class with a single Dirac cone
on the surface. Nat. Phys. 5, 398 (2009).
34. Bianchi M., Hatch R. C., Mi J., Iversen B. B.&Hofmann P. Simultaneous quantization of bulk
conduction and valence states through adsorption of nonmagnetic impurities on Bi2Se3. Phys.
Rev. Lett. 107, 086802 (2011).
35. Wang, Y. & Gedik, N. Circular dichroism in angle-resolved photoemission spectroscopy of
topological insulators. Phys. Status Solidi 7, 64–71 (2013).
36. Scholz, M. R. et al. Reversal of the circular dichroism in angle-resolved photoemission from
Bi2Te3. Phys. Rev. Lett. 110, 216801 (2013).
37. Wang, Z, Tsurkan, V., Schmidt, M., Loidl, A. & Deisenhofer, J. Tuning orbital-selective
correlation effects in superconducting Rb0.75Fe1.6Se2−zSz. Phys. Rev. B 93, 104522 (2016).
19
38. Yoshida, T. et al. Direct observation of the mass renormalization in SrVO3 by angle resolved
photoemission spectroscopy. Phys. Rev. Lett. 95, 146404 (2005).
39. Fukutani, K. et al. Electrical Tuning of the Excitonic Insulator Ground State of Ta2NiSe5. Rev.
Lett. 123, 206401 (2019).
40. Hor, Y. S. et al. p-type Bi2Se3 for topological insulator and low-temperature thermoelectric
applications. Phys. Rev. B 79, 195208 (2009).
41. Feng, L. et al. Flat bands near Fermi level of topological line defects on graphite. Appl. Phys.
Lett. 101, 113113 (2012).
42. Li, Z. et al. Realization of flat band with possible nontrivial topology in electronic Kagome
lattice. Sci. Adv. 4, eaau4511 (2018).
43. Chen, K. P., Chung, F. R., Wang, M. & Koski, K. J. Dual element intercalation into 2d layered
Bi2Se3 nanoribbons. J. Am. Chem. Soc. 137, 5431–5437 (2015).
44. Koski, K. J. et al. Chemical intercalation of zerovalent metals into 2D layered Bi2Se3
nanoribbons. JACS 134, 13773–13779 (2012).
45. Raghu, S., Qi, X. L., Honerkamp, C. & Zhang, S.-C. Topological Mott insulators. Phys. Rev.
Lett. 100, 156401 (2008).
46. Valla, T., Pan, Z. H., Gardner, D., Lee, Y. S. & Chu, S. Photoemission spectroscopy of
magnetic and nonmagnetic impurities on the surface of the Bi2Se3 topological insulator. Phys.
Rev. Lett. 108, 117601 (2012).
47. Abdalla, L. B., Seixas, L., Schmidt, T. M., Miwa, R. H. & Fazzio, A. Topological insulator
Bi2Se3(111) surface doped with transition metals: an ab-initio investigation. Phys. Rev. B 88,
045312 (2013).
20
48. Bonilla, M. et al. Strong room-temperature ferromagnetism in VSe2 monolayers on van der
Waals substrates. Nat. Nanotechnol. 13, 289–293 (2018).
21
Figures:
Fig. 1│Structural characterization. a, Top and side views of Bi2Se3 and VSe2 crystal structures.
Hexagonal BZs with the high-symmetry points are given in the lower part of a. b, Core-level
photoemission spectra of Bi 5d and Se 3d obtained from 12 QL Bi2Se3, 0.3 ML VSe2 / 12 QL
Bi2Se3, and 3 ML VSe2 / 12 QL Bi2Se3 samples. The inset in b shows the Se 3d peaks for pristine
and 3 ML VSe2 covered Bi2Se3 after scaling peaks to the same height for visual comparison. c, V
2p core levels for 0.3 ML VSe2 / 12 QL Bi2Se3 and 3 ML VSe2 / 12 QL Bi2Se3 samples. The atomic
stoichiometry of Se to V is computed to be 2 by using the peak areas and photoionization cross-
sections. d, HAADF-STEM cross-section image of 3 ML VSe2 / 12 QL Bi2Se3 heterostructure.
The color contrast in d is correlated with the atomic number (Z-contrast). e, f, Room temperature
STM images of Bi2Se3 (at sample bias 100 mV, set point 1 nA) and VSe2 (at sample bias 80 mV,
set point 1 nA) surfaces, respectively obtained from 0.3 ML VSe2 / 12 QL Bi2Se3. Yellow
parallelogram in f represents the Moiré unit cell.
Fig. 2 │ Emergence of the flat band revealed by ARPES. a-b, Experimental electronic structures
of 12 QL Bi2Se3 and 0.3 ML VSe2 / 12 QL Bi2Se3 samples, respectively. Spectra were collected
with 50 eV linear horizontal polarized lights. BVB refers to the bulk valance band of Bi2Se3. c,
Momentum integrated EDCs of 0.3 ML VSe2 / 12 QL Bi2Se3 acquired with 110 eV photon energy
along the
-
direction in the BZ. d-e, Constant energy counters at the EF and the EFB, respectively
for 0.3 ML VSe2 / 12 QL Bi2Se3. Red and black hexagons in e correspond to the BZ of VSe2 and
Bi2Se3, respectively. In-plane lattice parameters of 4.14 Å for Bi2Se3 and 3.356 Å for VSe2 were
employed to compute the BZs27, 31.
22
Fig. 3│Photon energy-dependent ARPES experiment. a-b, ky - kz dispersions at the EF and the
EFB, respectively. Dashed red lines in a mark the DSSs. c, MDCs at different kz points d-e, Binding
energy versus kz maps at ky = ± 0.3 Å, respectively. Dashed cyan colored lines in d and e represent
the modulation of the flat band along the kz direction. f, EDCs at various kz points to study the
spectral shape of the flat band. ARPES maps for the plots were conducted along the
-
direction
in the BZ.
Fig. 4│CD-ARPES. a-b, ARPES maps of 0.3 ML VSe2 / 12 QL Bi2Se3 sample recorded with
RCP and LCP lights, respectively. c, Computed CD-ARPES. d-e, CDs as a function of ky at 0.1
eV binding energy and the EFB, respectively.
Fig. 5│Electron doping impact on the band structure. a-b, ARPES maps of 0.3 ML VSe2 / 12
QL Bi2Se3 before (bare) and after K-doping, respectively. Spectra were obtained with 50 eV
photons at 15 K. c-f present integrated EDC within ky = ± 0.6 Å and EDCs along the k0, kR, and
kL momentums as marked in a. In a-b, dashed lines show the energy shifts of the different bands
as specified in the figures and the yellow solid line indicates the leading edge. In c-f, green, pink,
black, brown vertical lines represent the binding energy of the flat band, leading edge, bulk bands,
and the DP, respectively. The arrows in c-f point the direction of the energy shifts.
23
Fig. 1
Fig. 2
24
Fig. 3
25
Fig. 4
Fig. 5
