valley polarization in stacked mos induced by …eeeweba.ntu.edu.sg/bktay/pub/672.pdf · valley...
TRANSCRIPT
Valley polarization in stacked MoS2 induced by circularlypolarized light
Juan Xia1, Xingli Wang2, Beng Kang Tay2,3, Shoushun Chen4, Zheng Liu2,3,5, Jiaxu Yan1,6 (), and
Zexiang Shen1,3,7 ()
1 Division of Physics and Applied Physics, School of Physical and Mathematical Sciences, Nanyang Technological University, Singapore
637371, Singapore 2 NOVITAS, Nanoelectronics Centre of Excellence, School of Electrical and Electronic Engineering, Nanyang Technological University,
Singapore 639798, Singapore 3 CINTRA CNRS/NTU/THALES, UMI 3288, Research Techno Plaza, Singapore 637553, Singapore 4 School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore 639798, Singapore 5 Center for Programmable Materials, School of Materials Science and Engineering, Nanyang Technological University, Singapore
639798, Singapore 6 Institute of Advanced Materials (IAM), Jiangsu National Synergistic Innovation Center for Advanced Materials (SICAM), Nanjing Tech
University (NanjingTech), 30 South Puzhu Road, Nanjing 211816, China 7 Centre for Disruptive Photonic Technologies, Nanyang Technological University, Singapore 637371, Singapore
Received: 10 August 2016
Revised: 10 October 2016
Accepted: 12 October 2016
© Tsinghua University Press
and Springer-Verlag Berlin
Heidelberg 2016
KEYWORDS
circularly polarized
photoluminescence,
first-principles
calculations,
molybdenum disulfide,
ultra-low-frequency
Raman spectroscopy,
valley polarization
ABSTRACT
Manipulation of valley pseudospins is crucial for future valleytronics. The
emerging transition metal dichalcogenides (TMDs) provide new possibilities for
exploring the interplay among the quantum degrees of freedom, including real
spin, valley pseudospin, and layer pseudospin. For example, spin–valley coupling
results in valley-dependent circular dichroism in which electrons with particular
spin (up or down) can be selectively excited by chiral optical pumping in
monolayer TMDs, whereas in few-layer TMDs, the interlayer hopping further
affects the spin–valley coupling. In addition to valley and layer pseudospins, here
we propose a new degree of freedom—stacking pseudospin—and demonstrate
new phenomena correlated to this new stacking freedom that otherwise require
the application of external electrical or magnetic field. We investigated all possible
stacking configurations of chemical-vapor-deposition-grown trilayer MoS2 (AAA,
ABB, AAB, ABA, and 3R). Although the AAA, ABA, 3R stackings possess a sole
peak with lower degree of valley polarization than that in monolayer samples,
the AAB (ABB) stackings exhibit two distinct peaks, one similar to that observed
in monolayer MoS2 and an additional unpolarized peak at lower energy. Our
findings provide a more complete understanding of valley quantum control for
future valleytronics.
Nano Research
DOI 10.1007/s12274-016-1329-x
Address correspondence to Jiaxu Yan, [email protected]; Zexiang Shen, [email protected]
| www.editorialmanager.com/nare/default.asp
2 Nano Res.
1 Introduction
In analogy to spintronics [1, 2], valleytronics is proposed
to encode bits of information by manipulation of valleys,
the degenerate extrema of energy dispersion in the
momentum space [3–9]. Recently, transition metal
dichalcogenides (TMDs) have captured centralized
attention because of their intriguing physics associated
with the valley degree of freedom, which implies the
great potential for valley quantum control [10–17]. In
monolayer TMDs, by breaking the crystal inversion
symmetry, the valley pseudospins can be distinguished
by opposite signs at the corners (K and K’) of the
hexagonal Brillouin zone. The giant spin–orbit coupling
(SOC) separates the valence bands at each valley into
spin-up and spin-down states. Such inherent spin–
valley coupling can significantly protect the valley
polarization as the intervalley scattering requires a
simultaneous spin flip. Only carriers with selective
spin can be emitted under the valley optical selection
rule, which has been widely demonstrated by circularly
polarized excitation [10, 11, 13].
Beyond monolayer TMDs, bilayer TMDs introduce
another intrinsic degree of freedom called the layer
pseudospin [17–22]. The layer pseudospin refers to
the location of the carriers, either in the upper or lower
layer. The natural AB-stacked bilayer TMDs, in which
the lower layer is rotated by 180° with respect to
the upper layer, possess crystal inversion symmetry.
Consequently, the layer rotation switches two valleys
but leaves the spin unchanged, which results in a
sign change for the spin–valley coupling from layer
to layer. In theory, the valley-dependent effect should
vanish in the presence of inversion symmetry [14, 23].
Notably, nontrivial circular photoluminescence (PL)
polarization degrees have been observed in bilayer
MoX2 [11, 22] and WX2 [19, 20]. Contrary to early
beliefs, breaking the inversion symmetry by applying
an electric field in AB-bilayer samples is not required.
The possible origin remains in debate. For bilayer WSe2
and WS2, the giant SOC (up to ~0.5 eV) suppresses the
interlayer hopping, thus leading to the localized spin
states within either the upper or lower layer related
to the valley index that results in a nontrivial valley
polarization value [18–21]. This scenario appears
vague for bilayer MoS2 with a moderate SOC value of
approximately 0.1 eV. Zunger et al. [24, 25] claimed
that the circular polarization is intrinsic in centro-
symmetric layer stacks composed of individual
noncentrosymmetric layers.
The interplay among spin, valley pseudospins, and
layer pseudospins opens an unprecedented platform
toward manipulation of quantum states. Both spins
and valleys have magnetic moments and can be
manipulated by magnetic and optical means [26–29].
Layer pseudospin is associated with electric polarization
and can be tuned by an electric field [22]. Numerous
manipulations of degrees of freedom have been reported
to exploit this quantum control in monolayer and
bilayer TMDs. Along with the layer number, the stacking
configurations are also involved [30–32]. Iwasa et al.
[30] reported robust valley polarization in 3R phase
MoS2 using circularly polarized PL spectroscopy. Wu
et al. [31] examined several artificially stacked MoS2
bilayers by folding exfoliated monolayers and observed
different optical nonlinearities. On one hand, the
limited number of stacking patterns studied in the
literature (only natural 2H/3R phases and artificially
stacked MoS2 have been studied) prevents a good
understanding of stacking pseudospins. On the other
hand, the non-oriented folding may yield uncontrolled
interlayer coupling degradations.
In this work, we report on stacking-dependent
valley polarization in high-oriented chemical vapor
deposition (CVD)-grown trilayer samples. Using ultra-
low frequency (ULF) Raman spectroscopy, we observe
two distinct stacking configurations within one special
AAB-stacked MoS2 sample—one half AAB-stacking
and the other half 3R-stacking—with a shear (S) mode
at 16 cm–1 discovered first. Further utilizing circularly
polarized PL measurements, we report two exciton
peaks with distinct valley polarization behaviors in
AAB (ABB) stacking: The high-energy peak exhibits
an anomalously robust valley polarization (~90%),
and the low-energy peak becomes valley polarization
degrading. Combining this investigation with the ab
initio calculations, we uncover a physical picture of the
stacking-modulated spin–valley coupling in few-layer
MoS2. By extending our circularly polarized PL studies
to other stackings, this mechanism was further verified
in all stacked few-layer MoS2 systems. Our work
provides another possible freedom-controlled approach
to desirably manipulate the valley quantum states.
www.theNanoResearch.com∣www.Springer.com/journal/12274 | Nano Research
3 Nano Res.
2 Results and discussion
Figure 1 shows an AAB-stacked trilayer MoS2, where
we can clearly determine the layer number from both
the optical contrast image (Fig. 1(a)) and Raman/PL
intensity mappings (Fig. S1 in the Electronic Sup-
plementary Material (ESM)). The integrated Raman
intensity mapping of the A1g modes reveals the directly
proportional relationship between the Raman intensity
and layer number, and the PL intensity mapping clearly
demonstrates that the direct bandgap monolayer
emits more strongly. The PL intensity weakens with
increasing thickness of the MoS2 sample, in agreement
with the literature. The ULF Raman spectra provide
more information about different atomic arrangements
and interlayer coupling effects within the MoS2
nanoflake shown in Fig. 1(a), which cannot be clearly
observed in the high-frequency Raman and PL
spectrum. In Fig. 1(b), the two peaks labeled S and B
correspond to the shear and layer breathing modes,
involving the rigid motions of adjacent layers that
are parallel and perpendicular to the atomic planes,
respectively. This assignment is confirmed by the
polarized ULF Raman spectra, as clearly indicated in
Fig. S2 (in the ESM). In general, the S modes appear
in both ( ) ( )Z xx Z Π and ( ) ( )Z xy Z configurations,
whereas the B modes exist only in the ( ) ( )Z xx Z Π
configuration and not the ( ) ( )Z xy Z configuration.
Furthermore, the higher intensity ratio of the B mode
to S mode identifies the stacking configuration of the
bilayer region as AA-stacking, which is consistent with
the distinct features of AB-stacked and AA-stacked
bilayers, as shown in Fig. S1 (in the ESM). The features
of the ULF Raman intensity mapping of the bilayer
region using the B mode (Fig. S2 in the ESM) are
also consistent with those of a typical AA-stacking
spectrum.
For trilayer MoS2, two branches of the S modes exist,
i.e., the low and high modes of shear vibration (LS
and HS modes), as shown in Fig. 1(g). The LS mode
represents interlayer vibration in which the first
and third layer move in opposite directions while the
middle layer remains static; the HS mode represents
the motion in which the first and third layer move
together in the same direction while the middle layer
moves in the opposite direction. We identified the
Figure 1 Structure characterizations of AAB-stacked trilayer MoS2. (a) Optical contrast image of AAB-stacked trilayer MoS2. (b) ULFRaman spectra of the bilayer region. (c) and (d) ULF Raman mapping of the LS mode (c) and HS & B mode (d). (e) and (f) PolarizedULF Raman spectra of the left (e) and right (f) part of the trilayer sample. The yellow strips indicate the integrated peaks in (c) and (d),and the blue and orange curves represent the ( ) ( )Z xx Z Π and ( ) ( )Z xy Z polarized configuration, respectively. (g) and (h) Vibrational schematics of LS/HS and B modes in 3R-stacked (g) and AAB-stacked (h) trilayer MoS2. The numerical values were obtained from both experiments and calculations.
| www.editorialmanager.com/nare/default.asp
4 Nano Res.
distinct stacking configurations of the trilayer region
using ULF Raman spectroscopy (Figs. 1(c) and 1(d)).
For the left part of the trilayer region, the ULF Raman
spectra show two sharp peaks at 16 and 28 cm–1
(Fig. 1(e)), and both modes (16 and 28 cm–1) exist
under ( ) ( )Z xx Z Π configuration; however, only the low
frequency mode at 16 cm–1 is present in the ( ) ( )Z xy Z
configuration. The simulation identifies the former
as the LS mode and the latter as B modes, which
are assigned as the 3R phase. For the right part of
the trilayer region, the peak at 28 cm–1 represents the
accidental degeneracy of both the S and B modes,
which can also be distinguished using polarized ULF
Raman spectroscopy (Fig. 1(f)). The S mode in the
merged peak of the right region at approximately
28 cm–1 corresponds to the HS mode. The sole peak in
the right region is present under both ( ) ( )Z xx Z Π and
( ) ( )Z xy Z configurations except for a relatively large
intensity decrease up to 70% under the ( ) ( )Z xy Z
configuration. We can thus identify the stacking
configuration in the left region as AAB stacking.
Therefore, combining these results with our simulation
work, we can conclude that the HS (LS) mode exclusively
exists in the AAB (3R) stacking order. This situation
is similar to the lowest S mode observed in ABC-
stacked graphene [33, 34]. Unlike the S modes, only
one B mode centered at approximately 28 cm–1 is
present in all the trilayer samples of different stacking
configurations, indicating that the B mode is not
sensitive to stacking order. The distinct stacking-
dependent behaviors of the Raman S mode between
3R and AAB stackings can be easily understood through
a classical physical picture [33], as discussed in Fig. S7
(in the ESM).
Next, we discuss the stacking-modulated valley
polarization in trilayer MoS2 based on circularly
polarized PL measurements. Figures 2(a)–2(d) show
the valley and spin polarizations for different layer
regions of the trilayer MoS2 under σ– excitation with
left-hand (σ– red) and right-hand (σ+ black) detection
at 77 K. The degree of circular polarization η is defined
as ( ) / ( )I I I I
, where I denotes the
intensity of the detected I emission. The data indicate
high circular valley polarization in the monolayer
(~80%), AA-stacked bilayer (~80%), and 3R trilayer (60%)
MoS2. This observation demonstrates that valley optical
selection rules in 3R-phase MoS2 samples are consistent
with previous reports [30]. With the lack of inversion
symmetry in monolayer MoS2, the extraordinary spin
and valley contrasting physics emerge. For AA-stacked
bilayer MoS2, the lack of inversion symmetry still holds,
in which the spin and valley degrees of freedom are
locked in the AA bilayer similar to two individual
Figure 2 Valley and spin polarizations in 3R-stacked and AAB-stacked trilayer MoS2. (a)–(d) CP-PL spectra of the monolayer (a), bilayer (b), and left (c) and right (d) part of the trilayer regions in the AAB-stacked trilayer MoS2.
www.theNanoResearch.com∣www.Springer.com/journal/12274 | Nano Research
5 Nano Res.
monolayers. A similar situation persists in the trilayer
3R-stacked region, yielding strong circular dichroism.
Notably, there are two peaks under σ– detection for
the right part with ABB stacking. The energies of the
corresponding two excitons are 1.91 eV (high energy,
HE) and 1.86 eV (low energy, LE). Moreover, the HE
peak displays a strong PL polarization (~90%) under
σ– and σ+ detection, whereas the LE peak shows almost
no spin-valley selectivity. We attribute the peaks located
at ~1.76 eV to radiative recombination of bound
excitons [35], i.e., neutral excitons bound to defects,
consistent with the previously reported defect-activated
peaks at ~1.78 eV in monolayer MoS2.
To reveal the different dichroism behaviors, we
calculated the band structures of the 3R-stacked (left)
and AAB-stacked (right) trilayer MoS2 considering
SOC, as shown in Fig. S9 (in the ESM). The spin-
polarized band structures around the K valley are
enlarged. Because of different interlayer coupling and
charge locking effects, the top valence band of the 3R
phase almost exclusively originates from the spin-up
electrons in the upper layer, whereas for the AAB
stacking, the top two valence bands are the mixed
contributions of the upper and middle layer. The
different PL polarization behaviors arise from the
intrinsic stacking-modulated Bloch electrons at K(K’)
valleys. Here, we focus on the spin textures at the
band edge, which are marked with yellow rectangular
blocks. The splittings between the conduction bands
(Δc1,Δc2) and valence bands (Δv1,Δv2) are induced by
interlayer coupling effects. Figure 3(a) shows the energy
level diagram of 3R-stacked trilayer MoS2. The three
layers are coupled with each other by reduced interlayer
coupling in the 3R phase, resulting in nontrivial
energy splittings to form a “triple monolayer”. The
A-exciton transitions can only be excited by σ– polarized
light from the K’ valley, and the Bloch wave functions
are localized in the upper layer. Because of the large
energy splittings (~0.1 eV), the interlayer hopping
is effectively suppressed, and high circular valley
polarization persists similar to the monolayer case. In
the AAB-stacked trilayer, however, the energy levels
can be treated as a mix of coupled AB bilayers and
decoupled AA bilayers, as demonstrated in Fig. 3(b).
Figure 3 (a) and (b) Schematic illustration of circularly polarized-light-induced spin and valley polarizations in 3R-stacked (a) and AAB-stacked (b) trilayer MoS2 under σ − excitation. The spin-up and spin-down bands are marked in red and blue, respectively. The orange (grey) balls/circles denote resonantly excited (after relaxation) electron–hole pairs under a 633-nm laser. The dominant spin relaxation pathway following the intra-valley spin-flip of the photo-excited electron accompanied by dissipation of energy into the environment is depicted by the solid green arrow, and the ω1 emission at –K valley with weak cross section that needs the spin-flip and energy cost is indicated by the dotted green arrow.
| www.editorialmanager.com/nare/default.asp
6 Nano Res.
Under σ– excitation, the two detected σ– signals, ω1 and
ω2, correspond to the electron–hole recombinations in
the upper layer at the K valley and the middle layer
at the K’ valley. At the K’ valley, the spin can flip
from the upper layer to the middle layer through the
interlayer coupling by dissipation of Δv1 (~20 meV).
In this process, σ+ PL will be emitted at ω2, which
contributes to the detected σ+ peak in Fig. 2(d). However,
σ– at ω1 must gain an energy cost of ~50 meV (Δv1 for
an electron and Δc1 for a hole) to achieve the spin-flip
from the middle layer to the upper layer at the K
valley, which is greater than the thermal fluctuation
kT. Therefore, the degree of circular polarization η for
ω1 emission is larger than that for ω1 emission.
We note that our observation about stacking-
modulated spin and valley polarization shares some
similarities with gate-induced peak splitting in bilayer
WSe2 [19]. In that case, emissions from the upper and
lower layers at ω1 and ω2 arise from the difference
between the conduction and valence band energy shifts
(Δc and Δv, respectively) induced by an external
electric field, i.e., Δc – Δv. A high voltage is needed to
make this splitting larger than the spectral linewidth
(20-meV splitting under 150 V). Our peak splitting
value almost equals ~50 meV, i.e., Δc + Δv, which has
been certainly observed in experiments. Moreover, the
degrees of polarization for the two peaks behave in
opposite directions: η at ω1 is 90%, whereas η at ω2 is
nearly zero.
Figures 4(a)–4(d) present the circularly polarized
PL spectra of four types of trilayer MoS2 with ABB, AAB,
AAA, and ABA stacking, respectively. Note that ABB
and AAB stacking are equivalent by spatial inversion.
Hence, we observe almost the same PL polarization
features for the ABB stacking as for the ABB stacking,
that is, prominent PL polarization for the HE exciton
and almost no spin–valley selectivity for the LE exciton
(Figs. 4(a) and 4(b)). For the AAA and ABA stackings,
there is only one peak situated at 1.91 eV under both
σ– and σ+ detection, similar to the results for the 3R
stacking. We calculated both charge density distribution
diagrams (Figs. 4(e) and 4(f)) and band structures with
spin electronic states (Figs. S9–S11 in the ESM) for all
four stackings. Because of SOC and the strong interlayer
locking effect, the charges in the top valence band for
the AAA stacking almost exclusively originate from
the middle layer, which is similar to the findings for
3R stacking. However, for ABA stacking, because of
effective interlayer coupling, there is almost no band
splitting among the layers only with a large SOC gap
(Δc1 = Δc2 = 0, Δv1 = Δv2 = 0). Therefore, there is only
one HE peak (1.91 eV) for ABA stacking.
Figure 4 Valley and spin polarizations in ABB-, AAB-, AAA-, and ABA-stacked trilayer MoS2. (a)–(d) Circularly polarized PL spectra from ABB- (a), AAB- (b), AAA- (c), and ABA-stacked (d) trilayer MoS2. (e) and (f) Charge density distribution for AAA- (e) and ABA-stacked (f) trilayers. Because of the weak interlayer coupling, the charges in the top valence band for AAA stacking almostexclusively originate from the middle layer, which shares some similarity with monolayer MoS2. For the ABA stacking case, the charges are contributed from three coupled layers because of the effective interlayer coupling.
www.theNanoResearch.com∣www.Springer.com/journal/12274 | Nano Research
7 Nano Res.
3 Conclusions
In summary, we demonstrated new coupling
phenomena connecting real spin with stacking
pseudospin in trilayer MoS2. Using ULF Raman
spectroscopy, we identified two distinct stacking
configurations within the trilayer region of the MoS2
sample; that is, the left half of the sample exhibits 3R
stacking, and the right half exhibits AAB stacking.
The 3R phase in trilayer CVD-grown samples was
identified for the first time and was characterized by
a unique Raman peak at 16 cm–1 representing the LS
mode among layers. Combining this investigation
with our simulation work, we concluded that the HS
(LS) mode exclusively exists in the AAB (3R) stacking
order. In polarization-resolved PL measurements, we
observed the distinct behaviors of valley polarization
in different structural configurations without any
external field. Similar to monolayer and AA-stacked
bilayer MoS2, the trilayer 3R-stacked region (left)
yields a strong circular dichroism. Notably, the AAB-
stacked trilayer region (right) exhibits two peaks
under σ– detection with exciton energies at 1.91 eV
(HE, ~90%) and 1.86 eV (LE, ~0%), respectively. We
then extended our circularly polarized PL study to four
other types of trilayer MoS2 with ABB, AAB, AAA,
and ABA stacking. All the ABB and AAB stackings
showed a prominent PL polarization for the HE exciton
but almost no spin–valley selectivity for the LE exciton.
Similar to 3R stacking, AAA and ABA stacking result
in only one peak situated at 1.91 eV under both σ–
and σ+ detection. Our calculated band structures with
spin electronic states and charge density distribution
diagrams for all four stackings are in good agreement
with our experimental results. Therefore, such stacking-
modulated coupling between spin and pseudospin
broadens the understanding of valley quantum control
for future valleytronics.
4 Materials and methods
4.1 Sample preparation
A special-stacked AAB trilayer MoS2 nanoflakes with
a size of tens of micrometers was synthesized using
vapor solid deposition technology, in which 1 g MoS2
was placed in a ceramic boat and then placed at the
center of a 1-inch quartz tube with a piece of Si/SiO2
wafer acting as a substrate. The system was pumped
down to 30 Torr before being heated to 900 °C in
35 min; the pressure remained constant at 30 Torr for
another 10 min and was then increased to 760 Torr in
1 min and remained constant for another 20 min. After
the reaction, the growth system was cooled down to
room temperature naturally. Finally, the few-layered
MoS2 films consisting of 1–5 layers were produced on
the Si/SiO2 substrate. For the growth process, 20 sccm
Ar was used as the carrier gas.
4.2 Raman spectra and polarization-resolved PL
measurements
We used a Witec CRM200 backscattering Raman system
equipped with a solid-state yttrium aluminum garnet
laser at 2.33 eV with an appropriate power to avoid
sample heating (below 1 mW). The size of the laser
beam was approximately 250 nm and was focused on
the sample using an objective (Olympus, 100×/NA0.95).
For the ultra-low-frequency Raman spectra, we used
two BragGrate Notch Filters (BNF) centered at 532 nm
with bandwidths as narrow as 5 cm–1 and a large OD
(> 4) to achieve the low-frequency region. In addition,
a linear polarizer with different collection angles was
used to realize the parallel ( ) ( )Z xx Z Π and perpen-
dicular ( ) ( )Z xy Z backscattering configurations. For
the circularly polarized PL studies, an extremely
low laser power of ~60 μW was used before a long-
working-distance objective (Nikon, 50×/NA0.45, WD
13.8 mm) to avoid possible sample heating, and the
sample was maintained at 77 K. The excitation laser
with linear polarization, either horizontal or vertical,
was transformed into a corresponding circularly
polarized wave, either left-hand (σ–) or right-hand
(σ+), using an achromatic quarter-wave plate (Thorlabs,
AQWP05M-600). The emitted PL (σ– or σ+) passed
through the same quarter-wave plate followed by a
linear polarizer, and finally, the circularly polarized PL
was converted into a linearly polarized wave, which
was finally collected by a Si charge-coupled device.
4.3 Ab initio calculations
The first-principles calculations to compute the
| www.editorialmanager.com/nare/default.asp
8 Nano Res.
electronic structures were performed with the plane
wave basis set as implemented in the Quantum
ESPRESSO program package [36]. The norm-conserving
pseudopotentials [37], including spin-orbit coupling
effects, were adopted, and the cutoff energy for the
basis set was 140 Ry. The exchange correlation potential
was described using the local-density approximation
in the parameterization of Perdew and Zunger [38].
The Brillouin zone was sampled by a 15 × 15 × 1
k-point mesh. The energy convergence for the relaxa-
tion was selected to be less than 10–5 eV/A. We added
a vacuum layer of 20 A along the z-axis to avoid
mirror interactions between neighboring images.
The phonon spectrum and Raman intensities were
calculated within density-functional perturbation
theory (DFPT) as introduced by Lazzeri and Mauri
[39]. To perform the spin projection calculations,
we computed the Wannier functions using the
maximally localized algorithm [40] implemented in
the package WANNIER90 [41] using Mo d orbitals
and S p orbitals.
Acknowledgements
This research was supported by MOE under AcRF
Tier 2 (No. MOE2012-T2-2-124) and AcRF Tier 3 (No.
MOE2011-T3-1-005) in Singapore. X. L. W. and B. K. T.
would like to acknowledge the funding support from
NTU-A*STAR Silicon Technologies Centre of Excellence
under the program grant No. 112 3510 0003. L. Z.
would like to acknowledge the funding support from
the Singapore National Research Foundation under
NRF RF Award No. NRF-RF2013-08. J. X. Y. and J. X.
acknowledge the technical support from H. L. H. at
WITec. We thank Dr. Jer-Lai Kuo for helpful discussions.
Electronic Supplementary Material: Supplementary
material (systematic Raman and PL characterization
of the sample, Raman selection rules of HS and LS
mode in 3R and AAB trilayer MoS2, circularly
polarized PL set-up details, calculated band structures
and charge density distribution) is available in the
online version of this article at http://dx.doi.org/
10.1007/s12274-016-1329-x.
References
[1] Žutić, I.; Fabian, J.; Das Sarma, S. Spintronics: Fundamentals
and applications. Rev. Mod. Phys. 2004, 76, 323–410.
[2] Pesin, D.; MacDonald, A. H. Spintronics and pseudospintronics
in graphene and topological insulators. Nat. Mater. 2012,
11, 409–416.
[3] Xiao, J.; Ye, Z. L.; Wang, Y.; Zhu, H. Y.; Wang, Y.; Zhang,
X. Nonlinear optical selection rule based on valley-exciton
locking in monolayer WS2. Light: Sci. Appl. 2015, 4, e366.
[4] Rycerz, A.; Tworzydlo, J.; Beenakker, C. W. J. Valley filter
and valley valve in graphene. Nat. Phys. 2007, 3, 172–175.
[5] Isberg, J.; Gabrysch, M.; Hammersberg, J.; Majdi, S.; Kovi,
K. K.; Twitchen, D. J. Generation, transport and detection
of valley-polarized electrons in diamond. Nat. Mater. 2013,
12, 760–764.
[6] Takashina, K.; Ono, Y.; Fujiwara, A.; Takahashi, Y.;
Hirayama, Y. Valley polarization in Si(100) at zero magnetic
field. Phys. Rev. Lett. 2006, 96, 236801.
[7] Shkolnikov, Y. P.; De Poortere, E. P.; Tutuc, E.; Shayegan,
M. Valley splitting of AlAs two-dimensional electrons in a
perpendicular magnetic field. Phys. Rev. Lett. 2002, 89,
226805.
[8] Zhu, Z. W.; Collaudin, A.; Fauqué, B.; Kang, W.; Behnia, K.
Field-induced polarization of Dirac valleys in bismuth. Nat.
Phys. 2012, 8, 89–94.
[9] Jones, A. M.; Yu, H. Y.; Ghimire, N. J.; Wu, S. F.; Aivazian,
G.; Ross, J. S.; Zhao, B.; Yan, J. Q.; Mandrus, D. G.; Xiao, D.
et al. Optical generation of excitonic valley coherence in
monolayer WSe2. Nat. Nanotechnol. 2013, 8, 634–638.
[10] Cao, T.; Wang, G.; Han, W. P.; Ye, H. Q.; Zhu, C. R.; Shi,
J. R.; Niu, Q.; Tan, P. H.; Wang, E. G.; Liu, B. L. et al.
Valley-selective circular dichroism of monolayer molybdenum
disulphide. Nat. Commun. 2012, 3, 887.
[11] Mak, K. F.; He, K. L.; Shan, J.; Heinz, T. F. Control of valley
polarization in monolayer MoS2 by optical helicity. Nat.
Nanotechnol. 2012, 7, 494–498.
[12] Sie, E. J.; McIver, J. W.; Lee, Y.-H.; Fu, L.; Kong, J.;
Gedik, N. Valley-selective optical Stark effect in monolayer
WS2. Nat. Mater. 2015, 14, 290–294.
[13] Zeng, H. L.; Dai, J. F.; Yao, W.; Xiao, D.; Cui, X. D. Valley
polarization in MoS2 monolayers by optical pumping. Nat.
Nanotechnol. 2012, 7, 490–493.
[14] Xiao, D.; Liu, G.-B.; Feng, W. X.; Xu, X. D.; Yao, W.
Coupled spin and valley physics in monolayers of MoS2
and other group-VI dichalcogenides. Phys. Rev. Lett. 2012,
108, 196802.
[15] Mak, K. F.; Lee, C.; Hone, J.; Shan, J.; Heinz, T. F.
Atomically thin MoS2: A new direct-gap semiconductor.
Phys. Rev. Lett. 2010, 105, 136805.
www.theNanoResearch.com∣www.Springer.com/journal/12274 | Nano Research
9 Nano Res.
[16] Splendiani, A.; Sun, L.; Zhang, Y. B.; Li, T. S.; Kim, J.;
Chim, C.-Y.; Galli, G.; Wang, F. Emerging photoluminescence
in monolayer MoS2. Nano Lett. 2010, 10, 1271–1275.
[17] Ross, J. S.; Wu, S. F.; Yu, H. Y.; Ghimire, N. J.; Jones, A. M.;
Aivazian, G.; Yan, J. Q.; Mandrus, D. G.; Xiao, D.; Yao, W.
et al. Electrical control of neutral and charged excitons in a
monolayer semiconductor. Nat. Commun. 2013, 4, 1474.
[18] Gong, Z. R.; Liu, G.-B.; Yu, H. Y.; Xiao, D.; Cui, X. D.;
Xu, X. D.; Yao, W. Magnetoelectric effects and valley-
controlled spin quantum gates in transition metal dichalcogenide
bilayers. Nat. Commun. 2013, 4, 2053.
[19] Jones, A. M.; Yu, H. Y.; Ross, J. S.; Klement, P.; Ghimire,
N. J.; Yan, J. Q.; Mandrus, D. G.; Yao, W.; Xu, X. D. Spin-
layer locking effects in optical orientation of exciton spin in
bilayer WSe2. Nat. Phys. 2014, 10, 130–134.
[20] Zhu, B. R.; Zeng, H. L.; Dai, J. F.; Gong, Z. R.; Cui, X. D.
Anomalously robust valley polarization and valley coherence
in bilayer WS2. Proc. Natl. Acad. Sci. USA 2014, 111,
11606–11611.
[21] Xu, X. D.; Yao, W.; Xiao, D.; Heinz, T. F. Spin and
pseudospins in layered transition metal dichalcogenides.
Nat. Phys. 2014, 10, 343–350.
[22] Wu, S. F.; Ross, J. S.; Liu, G.-B.; Aivazian, G.; Jones, A.;
Fei, Z. Y.; Zhu, W. G.; Xiao, D.; Yao, W.; Cobden, D. et al.
Electrical tuning of valley magnetic moment through symmetry
control in bilayer MoS2. Nat. Phys. 2013, 9, 149–153.
[23] Yao, W.; Xiao, D.; Niu, Q. Valley-dependent optoelectronics
from inversion symmetry breaking. Phys. Rev. B 2008, 77,
235406.
[24] Liu, Q. H.; Zhang, X. W.; Zunger, A. Intrinsic circular
polarization in centrosymmetric stacks of transition-metal
dichalcogenide compounds. Phys. Rev. Lett. 2015, 114,
087402.
[25] Zhang, X. W.; Liu, Q. H.; Luo, J.-W.; Freeman, A. J.;
Zunger, A. Hidden spin polarization in inversion-symmetric
bulk crystals. Nat. Phys. 2014, 10, 387–393.
[26] Aivazian, G.; Gong, Z. R.; Jones, A. M.; Chu, R.-L.; Yan, J.;
Mandrus, D. G.; Zhang, C. W.; Cobden, D.; Yao, W.; Xu, X.
Magnetic control of valley pseudospin in monolayer WSe2.
Nat. Phys. 2015, 11, 148–152.
[27] Li, X.; Zhang, F.; Niu, Q. Unconventional quantum hall
effect and tunable spin hall effect in dirac materials:
Application to an isolated MoS2 trilayer. Phys. Rev. Lett.
2013, 110, 066803.
[28] MacNeill, D.; Heikes, C.; Mak, K. F.; Anderson, Z.;
Kormányos, A.; Zólyomi, V.; Park, J.; Ralph, D. C. Breaking
of valley degeneracy by magnetic field in monolayer MoSe2.
Phys. Rev. Lett. 2015, 114, 037401.
[29] Scrace, T.; Tsai, Y.; Barman, B.; Schweidenback, L.;
Petrou, A.; Kioseoglou, G.; Ozfidan, I.; Korkusinski, M.;
Hawrylak, P. Magnetoluminescence and valley polarized
state of a two-dimensional electron gas in WS2 monolayers.
Nat. Nanotechnol. 2015, 10, 603–607.
[30] Suzuki, R.; Sakano, M.; Zhang, Y. J.; Akashi, R.; Morikawa,
D.; Harasawa, A.; Yaji, K.; Kuroda, K.; Miyamoto, K.;
Okuda, T. et al. Valley-dependent spin polarization in bulk
MoS2 with broken inversion symmetry. Nat. Nanotechnol.
2014, 9, 611–617.
[31] Jiang, T.; Liu, H. R.; Huang, D.; Zhang, S.; Li, Y. G.; Gong,
X. G.; Shen, Y.-R.; Liu, W.-T.; Wu, S. W. Valley and
band structure engineering of folded MoS2 bilayers. Nat.
Nanotechnol. 2014, 9, 825–829.
[32] Akashi, R.; Ochi, M.; Bordács, S.; Suzuki, R.; Tokura, Y.;
Iwasa, Y.; Arita, R. Two-dimensional valley electrons and
excitons in noncentrosymmetric 3R-MoS2. Phys. Rev. Appl.
2015, 4, 014002.
[33] Lui, C. H.; Ye, Z. P.; Keiser, C.; Barros, E. B.; He, R.
Stacking-dependent shear modes in trilayer graphene. Appl.
Phys. Lett. 2015, 106, 041904.
[34] Tan, P. H.; Han, W. P.; Zhao, W. J.; Wu, Z. H.; Chang, K.;
Wang, H.; Wang, Y. F.; Bonini, N.; Marzari, N.; Pugno, N.
et al. The shear mode of multilayer graphene. Nat. Mater.
2012, 11, 294–300.
[35] Tongay, S.; Suh, J.; Ataca, C.; Fan, W.; Luce, A.; Kang,
J. S.; Liu, J.; Ko, C.; Raghunathanan, R.; Zhou, J. et al.
Defects activated photoluminescence in two-dimensional
semiconductors: Interplay between bound, charged, and
free excitons. Sci. Rep. 2013, 3, 2657.
[36] Giannozzi, P.; Baroni, S.; Bonini, N.; Calandra, M.; Car, R.;
Cavazzoni, C.; Ceresoli, D.; Chiarotti, G. L.; Cococcioni,
M.; Dabo, I. et al. QUANTUM ESPRESSO: A modular and
open-source software project for quantum simulations of
materials. J. Phys.: Condens. Matter 2009, 21, 395502.
[37] Hamann, D. R. Generalized norm-conserving pseudopotentials.
Phys. Rev. B 1989, 40, 2980–2987.
[38] Perdew, J. P.; Zunger, A. Self-interaction correction to
density-functional approximations for many-electron systems.
Phys. Rev. B 1981, 23, 5048–5079.
[39] Lazzeri, M.; Mauri, F. First-principles calculation of vibrational
Raman spectra in large systems: Signature of small rings in
crystalline SiO2. Phys. Rev. Lett. 2003, 90, 036401.
[40] Marzari, N.; Vanderbilt, D. Maximally localized generalized
Wannier functions for composite energy bands. Phys. Rev.
B 1997, 56, 12847–12865.
[41] Mostofi, A. A.; Yates, J. R.; Lee, Y.-S.; Souza, I.;
Vanderbilt, D.; Marzari, N. Wannier90: A tool for obtaining
maximally-localised Wannier functions. Comput. Phys.
Commun. 2008, 178, 685–699.
The author has requested enhancement of the downloaded file. All in-text references underlined in blue are linked to publications on ResearchGate.The author has requested enhancement of the downloaded file. All in-text references underlined in blue are linked to publications on ResearchGate.