communications
Angle-Resolved Raman Spectra Measurements:
A
Witec maximally localized algorithm[39] and were implemented in the
CRM200 backscattering Raman system was used equipped with a package WANNIER90[40] using W s and d orbitals and Te p orbitals.
solid-state (yttrium aluminum garnet) YAG laser at 2.33 eV (532 nm) The Weyl chirality was calculated by the software package Wanni-
with a proper power avoiding sample heating (below 1 mW). The erTools, which was based on the iterative Green’s function.[41,42]
size of the laser beam is around 250 nm which was focused on
the sample using an objective (Olympus, 100×/NA0.95). To obtain
stronger Raman signal with high (signal to noise) SN ratio, for the
excitation path, the laser polarization was fixed and the sample
was rotated from 0o to 360o with respected to the laser polariza-
tion direction. For the collection path, both the Raman scattering
Supporting Information
signals with z(xx)z and z(xy)z configurations can be detected simul-
taneously without any analyzer before output fiber. As the few-
layer WTe2 samples are easily degraded, all the related Raman
experiments for each sample were completed within 2 h.[31]
High Pressure Measurements: The high-pressure cell used
in this experiment is based on a symmetric diamond anvil cell
(DAC) having two diamonds with 400 µm culet size. The sample
and pressure transmitting medium (methanol/ethanol, 4:l) with a
small ruby chip (10 µm) are loaded in a 150 µm hole drilled in a
250 µm thick T301 gasket preindented to 60 µm thickness. The
high-pressure Raman spectra were collected using a Renishaw
InVia micro Raman system under backscattering configuration,
with excitation laser wavelength of 532 nm and 20× objective
lens. The laser power is controlled below 0.4 mW to avoid possible
laser-induced heating. The signal was dispersed by a 2400 g mm−1
grating under the triple subtractive mode with a spectra resolu-
tion of 1 cm−1. The acquisition time of each spectrum was 200 s.
Frequency calibration of the Raman spectrum is realized using the
characteristic 520 cm−1 line of silicon. For WTe2 sample loading in
DAC, a very fine tip was used to cut the WTe2 crystal and adjust its
orientation within the punched gasket to obtain parallel and ver-
tical alignments.
Supporting Information is available from the Wiley Online Library
or from the author.
Acknowledgements
This research was supported by Ministry of Education (MOE) under
Academic Research Fund (AcRF) Tier 1 (Reference No. RG103/16),
AcRF Tier 2 (MOE2015-T2-1-148), and AcRF Tier 3 (MOE2011-
T3-1-005) in Singapore. Z.L. would like to acknowledge the
funding support from the Singapore National Research Foundation
(NRF) under NRF RF Award No. NRF-RF2013-08. J.-H.L. acknowl-
edge (Japan Society for the Promotion of Science) JSPS KAKENHI
(P16823) for financial support. J.-X.Y. and J.X. acknowledge the
technical support from H.-L.H at WITec. J.X. and D.-F.L. performed
the experiments. J.-X.Y. performed simulation studies. J.-D.Z, P.Y.,
and Z.L. helped to prepare the samples. J.-H.L provided the STEM
measurements. All authors discussed the results. J.X., J.-X.Y., and
Z.-X.S. conceived the study. J.X. and J.-X.Y. wrote the paper.
Conflict of Interest
Ab Initio Calculations: The first-principles calculations
including geometry optimization, vibrational properties, and mole-
cule dynamics were all carried out using the projector-augmented
wave[32] method as implemented in the Vienna ab initio simula-
tion package (VASP) code[33] including the spin–orbit coupling
effects. The exchange correlation potential was described using
the Perdew, Burke and Ernzerhof functional within the general-
ized gradient approximation.[34] The kinetic energy cutoff was set
to 500 eV. The BZ was sampled by a 16 × 8 × 1 k-point mesh using
Monkhorst–Pack (MP) method. The energy convergence for the
relaxation was chosen to be less than 10−5 eV Å−1. To account for
the interlayer van der Waals interactions, all the simulations were
performed using the nonlocal vdW-DF-optB86b functional.[35] The
phonon frequencies at the Γ point and Raman intensities were
calculated within DFPT as introduced in Phonopy.[36] For the MD
simulation study, a 96-atoms 2 × 2 × 2 supercell and 2 × 2 × 2 MP
k-point sampling grid. To adopt Normal Pressure and Temperature
(NPT) ensemble, the simulation system was coupled to an Par-
rinello–Rahman barostat[37] running with Langevin stochastic ther-
mostat, as recently implemented in VASP code. The time step in
the molecular-dynamics simulations was 1 fs and the external pres-
sure maintains at 20 GPa. The Raman peak intensity of a particular
The authors declare no conflict of interest.
[1] X. Dou, K. Ding, D. Jiang, B. Sun, ACS Nano 2014, 8, 7458.
[2] A. P. Nayak, T. Pandey, D. Voiry, J. Liu, S. T. Moran, A. Sharma,
C. Tan, C.-H. Chen, L.-J. Li, M. Chhowalla, J.-F. Lin, A. K. Singh,
D. Akinwande, Nano Lett. 2015, 15, 346.
[3] M. Peña-Álvarez, E. del Corro, Á. Morales-García, L. Kavan,
M. Kalbac, O. Frank, Nano Lett. 2015, 15, 3139.
[4] A. P. Nayak, S. Bhattacharyya, J. Zhu, J. Liu, X. Wu, T. Pandey,
C. Jin, A. K. Singh, D. Akinwande, J.-F. Lin, Nat. Commun. 2014, 5,
3731.
[5] Z.-H. Chi, X.-M. Zhao, H. Zhang, A. F. Goncharov, S. S. Lobanov,
T. Kagayama, M. Sakata, X.-J. Chen, Phys. Rev. Lett. 2014, 113,
036802.
[6] Z. Zhao, H. Zhang, H. Yuan, S. Wang, Y. Lin, Q. Zeng, G. Xu, Z. Liu,
G. K. Solanki, K. D. Patel, Y. Cui, H. Y. Hwang, W. L. Mao, Nat.
Commun. 2015, 6, 7312.
[7] Y. Qi, P. G. Naumov, M. N. Ali, C. R. Rajamathi, W. Schnelle,
O. Barkalov, M. Hanfland, S.-C. Wu, C. Shekhar, Y. Sun, V. Susz,
M. Schmidt, U. Schwarz, E. Pippel, P. Werner, R. Hillebrand,
T. Forster, E. Kampert, S. Parkin, R. J. Cava, C. Felser, B. Yan,
S. A. Medvedev, Nat. Commun. 2016, 7, 11038.
[8] M. N. Ali, J. Xiong, S. Flynn, J. Tao, Q. D. Gibson, L. M. Schoop,
T. Liang, N. Haldolaarachchige, M. Hirschberger, N. P. Ong,
R. J. Cava, Nature 2014, 514, 205.
[9] I. Pletikosic´, M. N. Ali, A. V. Fedorov, R. J. Cava, T. Valla, Phys. Rev.
Lett. 2014, 113, 216601.
normal mode v was calculated as[38] ν ∝|ei ⋅ Aν ⋅es |(1/ωv )(nv +1) ,
I
where ei(es ) is the incident (scattered) light polarization, nv is
the Bose–Einstein distribution, and Aν is the Raman suscep-
tibility associated with the vth normal mode. To do the spin pro-
jection calculations, the Wannier functions were computed by the
1701887 (6 of 7)
© 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2017, 13, 1701887