InSe 2
InSe 2
InSe 2
‡
Société Civile Synchrotron SOLEIL, L’Orme des Merisiers, Saint-Aubin, BP 48, Gif-sur-Yvette 91192, France
§
Laboratoire des Solides Irradiés, Ecole Polytechnique, CNRS, CEA/DRF/IRAMIS, Institut Polytechnique de Paris, Palaiseau
91128, France
∥
European Theoretical Spectroscopy Facility, Palaiseau 91128, France
⊥
Institut de Minéralogie, de Physique des Matériaux et de Cosmochimie, CNRS-UMR7590, Sorbonne Université, 4 Place Jussieu,
Paris 75252, France
#
Engineering Technology Research Center for 2D Material Information Function Devices and Systems of Guangdong Province,
Shenzhen University, Shenzhen 518060, China
*
S Supporting Information
Figure 1. (a) Crystal structure and (b) the first Brillouin zone of ε polytype InSe. (c) Calculated band structure of InSe along out-of-plane
Γ−A direction. (d) ARPES intensity map of InSe under hν = 120 eV along K−M−Γ−K direction. (e) ARPES intensity map of InSe under hν
= 111 eV along H−A−H direction. The calculated band structures (solid curves) under kz = 0 and kz = π/c are superimposed on (d) and (e),
respectively.
Being able to change a direct band gap by a few hundred meV which are shown in Figure 1a,b. The band structure of ε-InSe
changes the optical response in a significant way. Band gap was calculated by density functional theory (DFT) with the
renormalization is a powerful technique to overcome intrinsic local density approximation (LDA) using the ABINIT code.
limits and to design novel materials properties and Due to the well-known underestimation of the band gap in
functionalities. Calculations predict that the band gap of DFT-LDA, we have added an additional rigid shift to the
bilayer InSe decreases dramatically and semiconductor to valence bands to reach the measured band gap (see the
metal phase transition is achieved by tuning carrier density, Supporting Information and Figure S1 for the details of
which can promote InSe as data storage devices.20 However, calculation). From band dispersion along Γ−A direction
the experimental engineering of InSe electronic structure with shown in Figure 1c, both the first valence band and conduction
carrier concentration has not been achieved systematically as band indicate significant kz dependence, which arises from the
yet. pz-like orbital (out-of-plane) contribution. Therefore, in order
Here, we report the direct observation of band gap to investigate the band gap modification experimentally, it is
renormalization in InSe by combining high-resolution angle- essential to project electrons from Γ−M−K plane as both
resolved photoemission spectroscopy (ARPES) with in situ CBM and VBM are located at the Γ point. ARPES experiments
surface sodium (Na) atom deposition under low temperature. can determine the in-plane (k//) of electrons in solids by the
Akin to applying a gate voltage in few-layer FETs, depositing momentum conservation of photoelectrons directly, while
alkali atom on the surface of layered semiconductors is an determining the kz requires a set of photon energy (hν)
efficient way to modulate the carrier density and the dependent ARPES measurements. Based on the nearly free
corresponding electronic structure, already achieved in several electron final state approximation with a potential parameter
2D semiconductors. With Na atom deposition, we confirm that V0, kz can be written as kz = 2m(Ek cos θ 2 + V0)/ℏ2 where θ
the band gap of InSe decreases ∼120 meV at the induced
carrier concentration of 8.1 × 1012 cm−2 as deduced from the is the emission angle and Ek is the kinetic energy of the emitted
shifts of both conduction band minimum (CBM) and valence free electrons satisfied Ek = hν − w − EB, where w is the work
band maximum (VBM). By increasing the deposition time, we function of the sample and EB is the electron binding energy.
observed a band-tail inside the gap, which might be a sign of a We performed hν dependent ARPES experiments and found
semiconductor-metal phase transition or caused by the Na V0 = 11 eV by fitting the experimental data with the kz
intercalation into the sheet of InSe. The application of electron expression (hν = 95−154 eV, see Figure S2). We further
doping reduces, in a controlled manner, the band gap of InSe, confirmed that the Γ−M−K plane can be probed by hν = 120
and therefore makes it optically active for a broad range of eV and A−H−L plane can be probed by hν = 111 eV, as
wavelengths, which can be used for near-infrared optoelec- shown in Figure 1d,e. From the ARPES result along K−M−Γ−
tronic devices and optical modulators or switches subse- K direction with hν = 120 eV shown in Figure 1d, one can
quently. clearly visualize the dispersion of valence bands, indicating the
high quality of our sample. Moreover, the experimental data
are in excellent agreement with the calculation kz = 0. From
RESULTS AND DISCUSSION our data, the energy difference from the VBM to the Fermi
van der Waals layered InSe used here belongs to ε polytype as energy is 1.27 eV. Since the band gap of pristine InSe is ∼1.28
determined by both Raman and transmission electron eV from our photoluminescence spectrum (Figure S3), we can
microscopy.16 The unit cell of ε-InSe contains two parallel deduce that our pristine sample is electron doped with the
layers obtained by a translation of one layer onto another, and chemical potential just below the CBM. The fact that our
the corresponding Brillouin zone (BZ) has hexagonal shape, pristine sample is electron doped with the chemical potential
13487 DOI: 10.1021/acsnano.9b07144
ACS Nano 2019, 13, 13486−13491
ACS Nano Article
Figure 2. Band structure evolution of InSe as a function of carrier concentration. (a) ARPES intensity map of InSe along Γ−Κ direction
under hν = 120 eV. (b−g) The zoom-in ARPES of InSe without Na deposition (b) and with Na deposition time of 3 (c), 5 (d), 11 (e), 14 (f),
and 17 min (e), respectively. (h) Comparison of EDCs at the momentum of k// = 0 ± 0.01 Å−1 in ARPES of pristine InSe and InSe with 5
min Na deposition. The dotted red curves are the calculated conduction bands of InSe and the dotted white lines indicate the Fermi levels.
Figure 3. InSe band gap renormalization as a function of electron doping. (a) EDCs of conduction band as a function of carrier
concentration. The CBM is determined as two times of peak energy related to EF of Gaussian curve fitting (dashed line). (b) EDCs of
valence band as a function of carrier concentration. The VBM is obtained by extrapolating a linear fit (dashed line) to the leading edge of the
spectra to the baseline. (c) Energy shift of CBM, VBM, core levels of In and Se, and band gap as a function of carrier concentration (or Na
deposition time).
just below the CBM is a boon for our experiment. It means the VBM and CBM are located at the Γ point, in the following,
that doped electrons will be immediately visible in the ARPES we examine the band gap of InSe by using ARPES at the
spectrum and that one can extract the band gap and further photon energy of 120 eV so as to probe the Γ−M−K plane.
study band gap modification as a function of electron doping. Electron doping by depositing alkali atom on the surface of
We now consider the appropriate energy and BZ region for sample is a stable and systematic way to precisely control
such studies. From the ARPES result along the A−H direction amount of carriers and monitor the variation of the surface
with hν = 111 eV shown in Figure 1e, the VBM is not at the A electronic structure.21−25 After the alkali atom deposition,
point but along the A−H direction, in good agreement with electrons are transferred from the alkali atom to the
the band dispersion at kz = π/c from the calculation. Since both semiconductor, which induces 2D electric dipole layers and
13488 DOI: 10.1021/acsnano.9b07144
ACS Nano 2019, 13, 13486−13491
ACS Nano Article
Figure 5. (a−f) Series of laser-based ARPES maps of InSe as a function of carrier concentration along Γ−K direction. The red dotted curves
are the first conduction band from theory calculation. (g) EDCs of conduction band as a function of carrier concentration. A sign of weak
conduction band and band-tail formation are observed under heavily electron doped InSe (deposition time t = 18 min). The dotted red
curves are the calculated conduction bands of InSe and the dotted white lines indicate the Fermi levels.
(Na deposition time of 18 min), the dispersion of conduction experiment of InSe in Figure 1 and Figure S2 was carried out at the
band gets blurry and the intensity deceases dramatically, with ANTARES beamline of synchrotron SOLEIL. The ARPES, XPS
the formation of a band-tail in the band gap shown in Figure combining with in situ Na atom deposition experiments in Figures
5f,g. The observations of conduction band shift, the carrier 2−4 were performed with hν = 120 eV at the TEMPO beamline of
synchrotron SOLEIL. In both beamlines, the photoelectrons were
concentration estimation under moderate doping, and band- collected with new generation MBS A-1 analyzers equipped with
tail formation under heavy electron doping from laser-based electrostatic lenses allowing Fermi surface mapping in order to
ARPES are consistent with the synchrotron ARPES data determine the sample orientation. The iso-energy coutours data are
shown in Figures 2−4, supporting our findings of band gap shown in the Figure S4. The laser-based ARPES and in situ Na atom
renormalization ∼120 meV and possible semiconductor to deposition experiments in Figure 5 were performed under hν = 6.3 eV
metal phase transition in the layered InSe semiconductor. at the femtoARPES platform located in synchrotron SOLEIL.32 The
sample orientation was checked by low energy electron diffraction
(LEED) before performing the laser-based ARPES experiment. The
CONCLUSIONS overall energy and momentum resolutions for all the ARPES setups
In summary, we study the electronic band structure of were better than 35 meV and 0.01 Å−1.
electron-doped InSe by means of ARPES and in situ surface Na Atom Deposition. The Na atom deposition experiments at
Na deposition. Deposition of Na atom on the surface of InSe TEMPO beamline and femtoARPES platform were performed using
donates electrons to adjacent InSe layers, and forms a dipole commercial getter sources. For the entire duration of the experiment,
layer. After moderate Na deposition, the conduction band and the sample was kept at the base temperature of 125 K and has never
been decoupled from the cryogenic manipulator.
its dispersion are clearly observed and are consistent with
calculations. We observe a band gap renormalization of ∼120
ASSOCIATED CONTENT
meV in InSe due to the strong vertical electrical field generated
by ionized Na atoms, which corresponds to a carrier *
S Supporting Information
concentration of 8.1 × 1012 cm−2 in the surface of InSe due The Supporting Information is available free of charge on the
to the Stark effect. We also find that a band-tail appears under ACS Publications website at DOI: 10.1021/acsnano.9b07144.
strong electron doping which might be a sign of semiconductor Theory calculation of InSe band structure; photon
to metal phase transition. This basic research project mimics energy dependent ARPES measurement; photolumines-
the effects of field-effect doping in InSe, which will offer cence spectrum of pristine InSe; determination of InSe
insights into the optoelectronic properties of InSe and will orientation (PDF)
potentially boost the performance of multifunctional devices.
AUTHOR INFORMATION
METHODS
Corresponding Authors
Materials and Photoluminescence Spectrum. Single crystals
of InSe in our work have been grown using the Bridgmann method.31
*E-mail: zheshengchen@gmail.com.
Photoluminescence spectrum of pristine InSe shown in Figure S3 was *E-mail: queenly@vip.sina.com.
performed on a Horiba Jobin-Yvon Xplora spectrometer in a back *E-mail: azzedine.bendounan@synchrotron-soleil.fr.
scattering geometry under ambient conduction with incident ORCID
wavelength of 638 nm. Zhesheng Chen: 0000-0001-7978-5637
X-ray-Based ARPES and Laser-Based ARPES. High-quality
single crystals of InSe were cleaved at the base pressure of 1 × 10−10 Bingbing Tian: 0000-0003-1508-6217
mbar for all ARPES experiment. The sample temperature was ∼80 K Luca Perfetti: 0000-0002-8426-557X
during experiment. The photon energy (hν) dependent ARPES Azzedine Bendounan: 0000-0001-7557-4322
13490 DOI: 10.1021/acsnano.9b07144
ACS Nano 2019, 13, 13486−13491
ACS Nano Article