betaFeedback

WorksPhysics and Astronomy

Hydrogen-impurity induced unconventional magnetism in semiconducting molybdenum ditelluride

Abstract

Layered transition-metal dichalcogenides are proposed as building blocks for van der Waals (vdW) heterostructures due to their graphene-like two dimensional structure. For this purpose, a magnetic semiconductor could represent an invaluable component for various spintronics and topotronics devices. Here, we combine different local magnetic probe spectroscopies with angle-resolved photoemission and density-functional theory calculations to show that 2H-MoTe2 is on the verge of becoming magnetic. Our results present clear evidence that the magnetism can be "switched on" by a hydrogen-like impurity. We also show that this magnetic state survives up to the free surface region, demonstrating the material's potential applicability as a magnetic component for thin-film heterostructures.

Layered transition-metal dichalcogenides (TMDs) are a promising class of materials featuring interesting optoelectronic, superconducting, catalytic, and topological properties. They can also be exfoliated down to monolayers, making them suitable for potential applications in complex van der Waals (vdW) heterostructures [1, 2]. Among the TMDs, polymorphic MoTe2\mathrm{MoTe}_2 exhibits a variety of electronic and magnetic properties: Its metallic 1T′1T' phase turns into a TdT_d-phase at ≈250\approx250 K [3]. The latter is a topological Weyl semimetal that features unconventional superconductivity at low temperature [4, 5, 6]. The 2H2H-phase, the most stable one at ambient conditions, is an indirect band gap semiconductor and has been successfully integrated in transistors and photodetectors [7, 8, 9]. Moreover, it has been demonstrated that the 2H2H-phase can be distorted into the 1T′1T'-phase as a function of strain or gate voltage [10, 11]. In addition, room temperature (RT) magnetic hysteresis has been observed for small amounts of vanadium deposited on 2H2H-MoTe2\mathrm{MoTe}_2 [12]. However, most interestingly, 2H2H-MoTe2\mathrm{MoTe}_2 was shown by muon spin spectroscopy (μ\muSR) to be a magnetic semiconductor, exhibiting a long-range magnetic order in the bulk at low temperatures, which was tentatively attributed to Mo anti-site defects [13]. However, no other technique has reported such long range magnetic order and the origin of this magnetism remains still unclear. In fact, theoretical calculations also predict that, apart from anti-site defects, hydrogen and transition-metal dopants are able to induce spin polarization in MoTe2\mathrm{MoTe}_2 in the monolayer limit [14, 15].

Since a muon can be considered to be a light hydrogen isotope, one possible explanation is that the magnetism observed in μ\muSR is not intrinsic, but rather induced by the muon itself. Another important consideration is that in order to exploit these magnetic properties, e.g. in thin-film heterostructures, one should first elucidate their origin and evolution as a function of depth in the vicinity of an interface (with vacuum or other materials).

Here, we address these questions by utilizing complementary local spin probe measurements, with implanted muons, as well as with the host’s 125^{125}Te-NMR to reveal the origin of the reported magnetism. We find that while some of the implanted muons, which form a hydrogen-like state in 2H2H-MoTe2\mathrm{MoTe}_2, detect magnetism in this material, 125^{125}Te-NMR shows no evidence of any intrinsic magnetic order. These findings were then confirmed by detailed density-functional theory (DFT) calculations, reproducing the experimental 3D electronic band structure. When including isolated hydrogen (i.e., muon like) impurities, the DFT unveils an induced magnetic moment at the impurity. Furthermore, we show that this hydrogen-induced magnetic state remains unchanged towards the crystal surface, establishing 2H2H-MoTe2\mathrm{MoTe}_2 as a switchable magnetic semiconductor by hydrogen doping, thus paving the path towards new possibilities for integrating this intriguing TMD into future applications in thin-film heterostructures.

In order to gain a better understanding of the nature of the previously observed magnetic signal in 2H2H-MoTe2_2, we combine different local magnetic spectroscopies, (i) μ+\mu^+SR which uses individually implanted, spin-polarized muons, stopping at interstitial sites in the lattice, and (ii) 125^{125}Te-NMR on the tellurium atoms, intrinsic to the host compound. This is illustrated in Figure 1(a). Both the muon and the 125^{125}Te nuclei have a spin of 1/2 and are, therefore, unaffected by quadrupolar interactions and thus act as purely magnetic probes of the system. The 2H2H-MoTe2_2 single crystals used in this study were grown by the Chemical Vapour Transport technique, using TeCl4_4 as the transporting agent. In contrast, the previously studied samples from Ref. [13] were prepared using Te flux. Therefore, we expect the density and nature of native defects in our samples to be different. This subtle difference should affect their magnetic properties, if long-range magnetism were induced by structural defects.

Image not extracted

Figure 1. (a) Schematic illustration of the different probing sites. The muon’s stopping sites are interstitial, inside the layer and in the vdW gap, whereas 125^{125}Te-NMR is sensitive to the local magnetism at the Te site. The μ+\mu^{+}SR asymmetry spectra measured in zero field at T=5.5T=5.5 K with the initial muon spin polarization (b) perpendicular (⊥\perp) and (c) parallel (∥\parallel) to the cc-axis of the crystal. Coherent zero-field precession is observed only in (b), indicating that the muons probe a magnetic field along the cc-axis. (d) 125^{125}Te-NMR spectral line shapes of a 2H2H-MoTe2_2 single crystal in a field of 5.01 T perpendicular to the cc-axis at different temperatures. (e) 125^{125}Te-NMR powder line shapes at 15 K at different applied fields. The lack of a pronounced temperature or field dependence in the 125^{125}Te-NMR lines implies the absence of an intrinsic magnetic order.

Indeed we observe a coherent precession of the muon spin in zero magnetic field and at low temperature, see Figure 1(b). However, our μ+\mu^+SR measurements show identical results to those published earlier, namely yielding the same transition temperature and the same internal magnetic field magnitude. The precession frequency at 5.5 K in Figure 1(b) corresponds to an internal magnetic field of 199.1(3) mT, consistent with the previously reported value of μ0Hint=200\mu_0 H_{\mathrm{int}} = 200 mT [13]. Additional measurements as a function of temperature and using low-energy muons stopping within 10 nm to 100 nm from the surface are presented in Supplementary Fig. S8 and show a similar behavior [16]. Therefore, the local magnetic properties observed with μ+\mu^+SR in 2H2H-MoTe2_2 are robust and do not change in the near-surface region of the crystals, nor do they depend on the details of the sample synthesis. This is an indication that the previously reported structural defects [13] are unlikely to be the origin of long-range magnetic order.

Furthermore, measurements as a function of the angle between the cc-axis and initial muon spin polarization, as shown in Figure 1(b,Figure 1c), reveal that the internal magnetic field sensed by the muon is oriented along the cc-axis. A maximal Larmor precession amplitude is detected when the initial muon spin S ⊥\perp cc-axis (Figure 1(b)) while it almost vanishes for S ∥\parallel cc-axis (Figure 1(c)). This is a clear indication that the local magnetic field sensed by the implanted muons points along the crystallographic cc-axis. A detailed analysis of the full angular dependence between the initial muon spin and the cc-axis is presented in Supplementary Fig. S6 and supports this conclusion [16].

Contrary to the μ+\mu^+SR results, we find no signatures of intrinsic magnetism when we investigate 2H2H-MoTe2_2 with 125^{125}Te-NMR. Example spectra on single crystalline 2H2H-MoTe2_2 in a field of ≈5.01\approx5.01 T are shown in Figure 1(d). The spectra neither exhibit a significant broadening nor a resonance shift as a function of temperature, which is strong evidence that the 125^{125}Te nuclear spin experiences no significant changes in the local spin susceptibility.

This is in stark contrast to the typical NMR hallmarks of magnetic ordering, which consist of a temperature-dependent NMR line shift and broadening, caused by the emergence of an internal magnetic field near the magnetic transition, as observed for example in unr­enriched Fe [17] or in the RAlGe material family [18]. Similar distinctive NMR magnetic signatures are also expected for defect-induced magnetism, as has been observed e.g. in the 2D system SiC [19].

Furthermore, we exclude the presence of a static magnetic field at the 125^{125}Te site, by measuring powder line shapes at 15 K in different applied fields, see Figure 1(e). These spectra are significantly broader than those of single crystals, a broadening which we attribute to an anisotropic chemical shift. However, apart from a small change in the line shape, the line width scales perfectly with the applied field. This lack of field dependence is indicative of the non-magnetic origin of the line broadening, as it excludes the presence of a static internal magnetic field at the 125^{125}Te site. Therefore, these results clearly show that an intrinsic spin probe exhibits no evidence of any long-range magnetic order in 2H2H-MoTe2_2. This is in clear contrast to our observation with μ+\mu^+SR and a strong indication that the local static fields observed with muons originate from a muon-induced effect.

In order to reveal the origin of this effect, we use DFT to calculate the ground state of 2H2H-MoTe2_2 and how it is affected by an individual implanted muon. We start by refining our DFT calculation to reproduce the electronic structure of 2H2H-MoTe2_2. For this purpose, we determine the three dimensional electronic band dispersion with soft X-ray angle-resolved photoemission spectroscopy (ARPES). Spectra along high symmetry directions at photon energies corresponding to the Γ\Gamma and A planes are shown in Figure 2(a-Figure 2e). These reveal sharp bands over the whole range of binding energies. We find that the valence band maximum is formed at the Γ\Gamma point by a band derived from states that are symmetric along Γ\Gamma–M. In addition, we measure the band dispersions in kzk_z along Γ\Gamma–A as a function of photon energy (Figure 2(f)). Along this direction we expect that a two dimensional state forms a flat band with no clear dispersion. In contrast, the dispersive spectral weight in Figure 2(f) reveals that most bands hybridize across the vdW gap and gain a strong 3D character, despite the 2D layered structure of 2H2H-MoTe2_2. Interestingly, the spectral weight exhibits modulations with longer periodicity in kzk_z than the period of the reciprocal lattice. Such an effect has been previously observed in other 2H2H-TMDs and is connected with the non-symmorphic symmetry group of the 2H2H structure of MoTe2_2 [20].

Image not extracted

Figure 2. Comparison of soft X-ray ARPES spectra at ∼12\sim12 K with the calculated band structure (solid lines). (a,b) Dispersion along Γ−M\Gamma-M at Γ25\Gamma_{25}, (c) along Γ−M\Gamma-M at Γ27\Gamma_{27}, (d) along K-M at Γ25\Gamma_{25}, (e) along A-L between Γ26\Gamma_{26} and Γ27\Gamma_{27} and (f) along Γ\Gamma-A. (g) Constant energy cuts at Γ25\Gamma_{25}. In (a) and (b) symmetric and antisymmetric bands were probed by using pp- and ss- polarized light, respectively. There, the width of the calculated lines represents the contribution of symmetric or anti-symmetric orbitals to the band structure. All other spectra were probed with circularly polarized light, which does not distinguish between symmetric and anti-symmetric states. (h) Brillouin zone and high symmetry points of 2H2H-MoTe2\mathrm{MoTe}_2.

We compare the spectral weight in Figure 2 to the calculated band structure. Note that in general, ab-initio DFT of vdW materials with strong spin-orbit coupling can be misleading, as exemplified by TdT_d-MoTe2_2, where an on-site correlation term (U) had to be introduced, to correctly reproduce the experimental results [6, 21, 22]. However, our DFT calculations, performed without a UU term, are in excellent agreement with the experimental bands. This indicates that, unlike TdT_d-MoTe2_2, in 2H2H-MoTe2_2 there are only weak correlations. We further note that there are no signatures of intrinsic magnetism in the calculations of bulk 2H2H-MoTe2_2.

By correctly reproducing the 3D band dispersions, we have identified a reliable representation of 2H2H-MoTe2_2’s electronic structure and charge density within DFT. We can now employ it to predict the behavior of implanted μ+\mu^{+} spin probes, introduced as isolated impurities at interstitial positions within a supercell. We consider two charge states of the muon by adding a charge on the supercell (+1 for a diamagnetic μ+\mu^{+} and +0 for a neutral muon state) [23, 24]. The electrically neutral state of a muon corresponds to a muonium (Mu), a hydrogen-like bound state between a muon and an electron. Figure 3 shows the resulting stable stopping sites and the diffusion energy barrier along the most energetically favorable paths between them. In particular, we find that the diamagnetic μ+\mu^{+} has a stable site b (denoted by its Wyckoff letter) inside the Mo-layer and a meta-stable site g in the vdW gap. On the other hand, the neutral muonium state can be stable in b or meta-stable in a.

Image not extracted

Figure 3. (a) Stopping sites and diffusion paths indicated in the 3×3×13 \times3 \times1 supercell. The muon (μ+\mu^{+}) at sites b and g and muonium (Mu) at sites b and a. Energy barrier for (b) μ+\mu^{+}, and (c) Mu diffusion along the different paths. The black lines show the ground-state energy level of the particles at the local minima. (d) Calculated local spin density of states for a muon stopping at site a, assuming a nonzero Hubbard UU.

Finally, we consider Mu- and muon-induced magnetism in these sites. Indeed, for the Mu at the site a, and when employing a finite Hubbard term UU on the hydrogen (i.e. Mu) potential, a significant spin density is induced in the material, Figure 3(d), while all other sites remain nonmagnetic (Supplementary Fig. S5) [16].

Such muon-induced magnetism is very similar to what was found for hydrogen adsorbed on a monolayer of 2H2H-MoTe2_2 or on graphene [14, 25]. As we discuss in more detail in the Supplementary [16], such a magnetic muonium state is consistent with both the zero-field as well as the transverse field behavior (Supplementary Fig. S7) of the observed μ+\mu^{+}SR signals. Therefore, we conclude that the 2H2H-MoTe2_2 does not exhibit an intrinsic magnetic order and that the magnetic signal observed with μ+\mu^{+}SR originates from a magnetic muonium state in site a. As the muon can be considered a light hydrogen isotope, this induced magnetic state should be the same for a hydrogen atom located at site a.

To conclude, using local magnetic probe measurements, ARPES and theoretical DFT calculations, we are able to reveal that 2H2H-MoTe2_2 is a semiconductor on the verge of becoming magnetic. The comparison of the intrinsic 125^{125}Te-NMR probe with the implanted muon spin probe confirms that the observed local field in μ+\mu^{+}SR does not reflect an intrinsic long-range magnetic order. Instead, we find that the muonium, being a hydrogen-like isolated impurity, can be used to “switch-on” a local magnetic moment in this system.

Importantly, we find that the effect of muonium on the local magnetic properties remains unchanged up to the surface of the crystals. This paves the path towards the use of hydrogen impurities to transform 2H2H-MoTe2_2 into a magnetic semiconductor building block for vdW heterostructures. We expect that the use of such a material is not limited to applications in spintronics devices [26] but may also include, e.g., Majorana heterostructures [27] or introduction of magnetic proximity at the surface of topological materials [28].

Acknowledgements

This work is partially based on experiments performed at the Swiss Muon Source (Sμ\muS) and Swiss Light Source (SLS), Paul Scherrer Institute, Villigen, Switzerland. J.A.K. acknowledges support by the Swiss National Science Foundation (SNF-Grant No. 200021_165910). I.P.R. acknowledges support from the Ministry of Education and Science of the Russian Federation within State Task Nr. 0721-2020-0033. The work at the University of Warwick was supported by EPSRC, UK, through through Grants EP/M028771/1 and EP/T005963/1. The magnetization measurements were supported by the Swiss National Science Foundation (SNF-Grant No. 206021_139082). We are grateful to C. W. Schneider for help with single crystal XRD measurements. We thank Z. Guguchia, and S. Holenstein from PSI, CH, and A. Chatzichristos, D. Fujimoto, V. L. Karner, R. M. L. McFadden, J. O. Ticknor, W. A. MacFarlane, and R. F. Kiefl from the University of British Columbia, CA, for helpful discussions. Further, we note that a preliminary version of these results was published as part of a PhD thesis, Ref. [29].

Sample growth and characterization

High-quality single crystals of 2H2H-MoTe2_2 were grown by chemical vapor transport method. Stoichiometric amounts of the starting materials, Mo and Te, along with the transport agent TeCl4_4 (3 mg/cm3^3), were mixed together in a quartz ampule and sealed in vacuum. The quartz tube was then placed for 3 weeks in a two-zone horizontal tube furnace with the charge end at 800 °C and the end where the crystals form at 750 °C. The high quality of the single crystals and the crystal structure was confirmed by x-ray diffraction (XRD) and x-ray photoelectron spectroscopy (XPS), as shown in Figure S1. The additional XPS peaks around −460 eV are the Te M4,5_{4,5}-N4,5_{4,5}N4,5_{4,5} Auger lines.

Image not extracted

Figure S1. (a) The Laue back reflection X-ray diffraction pattern for one of the single crystals. (b) X-ray (0,0,nn) diffraction peaks. The inset shows a photograph of one of the single crystals. (c) XPS core level spectrum measured at hν=950h\nu= 950 eV. The blue (red) lines show the expected positions of the Mo (Te) peaks.

Magnetization measurements

Magnetometry measurements have been performed with a Quantum Design superconducting quantum interference device (SQUID) magnetic properties measurement system (MPMS) and are shown in Figure S2. For an applied in-plane field of 100 mT, the sample shows a Curie-Weiss like behavior. In contrast, if the field is applied out-of-plane there is a signature of a transition around 5 K, see Figure S2. Moreover, for an out-of-plane field of 10 mT this transition features a small hysteresis in the magnetization signal, Figure S2. Its magnitude is too small for an intrinsically ferromagnetic sample and it cannot account for the observed magnetic transition in μ+\mu^+SR, which appears at much higher temperatures (∼40\sim40 K). However, it could be attributed to various origins, including magnetic defects [13] or magnetic grain-boundaries [S2].

Image not extracted

Figure S2. (a) Zero field cooled (ZFC) magnetization in an applied field of 100 mT. (b) Zoom into the low temperature region of (a). The solid line shows the fit of a Curie-Weiss law (Tc=−0.9(5)T_{\mathrm{c}} = -0.9(5) K). The arrow shows the onset of the magnetic zero field oscillations observed in μ+\mu^+SR. (c) Field cooled (below 50 K) and zero field cooled magnetization for a 10 mT out-of-plane field.

ARPES

The soft X-ray ARPES experiments were performed on the X03MA beamline at the Swiss Light Source of the Paul Scherrer Institute, Villigen, Switzerland [S3, S4]. The analyzer slit was oriented along the incoming beam direction and the photon energy hνh\nu was tuned in the range of 320 eV to 900 eV. The high photon energies allow a better kzk_z resolution [38] than in previous investigations of 2H2H-MoTe2_2 [39]. The combined beamline and analyzer resolution was better than 0.15 eV at hν=900h\nu= 900 eV and better than 63 meV around hν∼500h\nu\sim500 eV. The samples were cleaved in-situ and kept at ∼12\sim12 K in a vacuum lower than 10−1010^{-10} mbar. We only present the binding energy with respect to the valence band maximum, because 2H2H-MoTe2\mathrm{MoTe}_2 tends to statically charge upon x-ray illumination, which causes a rigid shift of the spectrum, see below.

Out-of plane dispersion

In the soft X-rays energy range the photoelectrons have a larger escape depth compared to standard ultraviolet ARPES. This reduces the kzk_z broadening and allows to determine the out-of plane dispersion. For a measured photoelectron kinetic energy EkinE_{\mathrm{kin}}, the perpendicular momentum is given by:

kz=2me(Ekin−V000)ℏ2−k∥+pγ,z,(1)k_z = \sqrt{\frac{2m_{\mathrm{e}}\left(E_{\mathrm{kin}}-V_{000}\right)}{\hbar^2}}-\mathbf{k}_{\parallel}+\mathbf{p}_{\gamma,z}, \tag*{(1)}

where mem_{\mathrm{e}} is the electron mass, k∥k_{\parallel} the in-plane momentum, V000V_{000} the inner potential and pγ,zp_{\gamma,z} the component of the photon momentum (hν/ch\nu/c) perpendicular to the sample surface [S28]. The inner potential is a material and energy dependent parameter. We have estimated it by comparing the measured band structure with the expected periodicity of the Brillouin zone (cf. Fig. 1(f) in the main paper). We used a value of V000≈10 eVV_{000} \approx10\,\mathrm{eV}, which gives a fair agreement between theory and experiment, but slightly differs from the previously used V000=16(1) eVV_{000} = 16(1)\,\mathrm{eV} [39].

Static charging

Because 2H2H-MoTe2\mathrm{MoTe}_2 is a semiconductor, it tends to statically charge upon x-ray irradiation at low temperature. This is shown in Figure S3, which displays the binding energy of several features in the band structure as a function of photon flux. The binding energies were calibrated by measuring the Fermi level of Au. For some measurements the

Image not extracted

Figure S3. (a) Binding energy of several features indicated in (b) as a function of x-ray flux. (c) Angle integrated spectra corresponding to (a), after rigidly aligning the valence band maximum.

sample was additionally irradiated with a near UV laser. This excites some charge carriers and helps to significantly reduce the charging (Figure S3). In both cases, the charging can be assumed to be a rigid shift of the spectrum. This is evident from the linear dependence on the flux in Figure S3 and from the collapse of angle integrated spectra after alignment of the valence band maximum in Figure S3. Therefore, in the main paper we only show the binding energy with respect to the valence band maximum.

n-type samples

All measurements shown in the main text were performed on samples that were freshly cleaved in-situ at ∼12 K\sim12\,\mathrm{K} directly before the measurement. However, after a couple of hours of ARPES measurements the 2H2H-MoTe2\mathrm{MoTe}_2 samples usually degrade and the material becomes slightly nn-type. This is exemplified by the spectra in Figure S4, where a small spectral weight above the valence band was observed. These measurements clearly confirm that the band gap in 2H2H-MoTe2\mathrm{MoTe}_2 is indirect.

Image not extracted

Figure S4. (a,c) Spectra measured on a sample that became n-type over time. (b,d) Zooms into the region above the valence band maximum. The dashed line indicates the estimated, effective position of the Fermi level.

Density-functional theory calculations

The electronic structure was calculated with Density-Functional Theory (DFT) and the Projected Augmented Wave (PAW) method from the VASP code [40, 41]. Spin-orbit coupling was treated within the scalar-relativistic approach by the second variation method. The structure was relaxed and the exchange-correlation functional was chosen to be Perdew-Burke-Ernzerhof (PBE) [42]. The parameter for the relaxed cell used in the ARPES calculation were given by a=3.491a = 3.491 Å, c=13.647c = 13.647 Å, where the Mo and Te sit at the 2c2c and 4f4f Wyckoff positions of space group 194, respectively, with relative zz-coordinates 0.116 and 0.383 for Te. Along Γ\Gamma-M an orbital decomposition into symmetric (pz,py,dz2,dx2−y2,dzyp_z,p_y,d_{z^2},d_{x^2-y^2},d_{zy}) and antisymmetric (px,dxy,dxzp_x,d_{xy},d_{xz}) orbitals was preformed. The orbital character does not fully separate into symmetric and antisymmetric bands because of spin-orbit coupling. Nevertheless, when using p−p- (s−s-) polarized light [Figure 2(a) (2(b)) in the main text], only the symmetric (antisymmetric) part of the electronic wave function, which primarily consists of these orbitals, will be probed.

Muon site determination

In order to simulate a single muon impurity, we used a hydrogen pseudopotential in the relaxed 3×3×13 \times3 \times1 supercell of 2H2H-MoTe2_2. The assumed lattice constants in the 3×3×13 \times3 \times1 supercell, c=13.647c = 13.647 Å and a corresponding a=10.4745a = 10.4745 Å, were the same as for the band structure calculations. The charge state of the muon is accounted for by assuming a neutral/charged supercell [23, 24]. Therein, we preformed nudged elastic band (NEB) [43] as well as density of states calculations on the basis of the Linear Combination of Pseudo Atomic Orbitals (LCPAO) approach as implemented in the OPENMX code [S13]. The muon ground state energies were estimated by approximating the NEB diffusion energy barriers with an anisotropic harmonic oscillator potential. The crystal structures were visualized with VESTA [45]. Some further details regarding the stable muon positions are listed in Table I.

SpeciesWyckoff positionfrac. coordinatesΔE\Delta E (eV)ZPE (eV)
μ+\mu^{+}2b2b0 0 0.2500.57
μ+\mu^{+}6g6g0.5 0 0.51.50.22
Mu2b2b0 0 0.2500.54
Mu2a2a0 0 0.51.240.26

Table I. Details of the identified stable and meta-stable site for several implanted probe species. ΔE\Delta E denotes the difference in the total energy of the relaxed supercell compared to the most stable site and ZPE is the approximate ground state energy of the muon.

We further show the local spin density of states at the different stopping sites in Figure S5. It is evident from this figure, that there is no induced magnetism at the stable muon stopping site b and g, see Figure S5,Figure S5). For site b, the absence of an induced spin polarization can be understood because of the strong pairwise interaction of the muon with the neighboring Mo atoms. However, in the case of a muonium (and also a muon) sitting at the hexagonal site inside the vdW gap (site a) and when employing a false Hubbard term UU, a significant spin density is induced, see Figure S5,Figure S5). This is very similar to what was found for hydrogen adsorbed on a monolayer 2H2H-MoTe2_2 or on graphene [14 S16]. Note that site a is only a stable stopping site of muonium but not of a muon. As argued in more detail in the main text and below, such a muonium induced magnetic state can explain the observed coherent muon spin precession, while being consistent with the absence of a magnetic signal in magnetization and 125^{125}Te-NMR.

Image not extracted

Figure S5. Total density of states and local spin density of states calculated for a muon (i.e. a hydrogen potential in a charged supercell) at site b (a,b), site a (c,d), and site g (e,f), respectively. Note that site a is not a stable stopping site for a diamagnetic muon.

Muon spin spectroscopy

The μ+\mu^{+}SR experiments were performed on the general purpose spectrometer (GPS) at the Swiss Muon Source, Paul Scherrer Institute, Switzerland [30]. The temperature was controlled using a 4He{}^{4}\mathrm{He} flow cryostat. In order to vary the orientation of the sample with respect to the initial muon spin direction, the sample was mounted on a mechanical rotation-stage. Nearly 100 % spin polarized muons were implanted. The asymmetric emission of the muon’s decay positrons (lifetime τμ=2.2 μs\tau_{\mu}=2.2\ \mu\mathrm{s}) enables us to use them as a local magnetic probe [S18]. Note, that at any given moment, there is only one muon in the sample. Therefore, we can exclude that a muon probes magnetism induced by a nearby Mu or μ+\mu^{+}. The data were analyzed with musrfit software suite [32].

Zero field angle dependence

In order to determine the variation in the zero field (ZF) μ+\mu^{+}SR spectra as a function of sample orientation, the asymmetry spectra (such as the ones shown in Figure 1(b,Figure 1c) of the main paper) were fitted to a sum of a Gaussian damped oscillating part and two exponentially damped components:

A(t)=A1e−12(σt)2cos⁡(γμBt+ϕ)+A2e−λslowt+A3e−λfastt.(2)A(t)=A_{1}e^{-\frac{1}{2}(\sigma t)^{2}}\cos(\gamma_{\mu}Bt+\phi)+A_{2}e^{-\lambda_{\mathrm{slow}}t}+A_{3}e^{-\lambda_{\mathrm{fast}}t}. \tag*{(2)}

We further assumed that the internal field BB and the depolarization rate σ\sigma of the oscillating part to be independent of the measured projection of the polarization. Figure S6 shows the resulting oscillating ZF asymmetry A1A_{1} in different sets of detectors (corresponding to different projections of the muon spin polarization) as a function of the sample orientation. In order to understand this rotation dependence, one has to consider the Larmor precession of the muon polarization P(t)\mathbf{P}(t), defined as

Image not extracted

Figure S6. Oscillating asymmetry in different sets of detectors, measured for different orientations of the sample. The inset depicts the measurement geometry.

dP(t)dt=γμP(t)∧B.(3)\frac{\mathrm{d}\mathbf{P}(t)}{\mathrm{d}t}=\gamma_{\mu}\mathbf{P}(t)\wedge\mathbf{B}. \tag*{(3)}

Here, B\mathbf{B} denotes the local magnetic field vector and γμ=2π×135.5 MHz/T\gamma_{\mu}=2\pi\times135.5\ \mathrm{MHz/T} is the gyromagnetic ratio of the muon. The resulting precession can be written as [S18, S20]:

P(t)=p∥+p1⊥cos⁡(γμBt)+p2⊥sin⁡(γμBt),(4)\mathbf{P}(t)=\mathbf{p}^{\parallel}+\mathbf{p}_{1}^{\perp}\cos(\gamma_{\mu}Bt)+\mathbf{p}_{2}^{\perp}\sin(\gamma_{\mu}Bt), \tag*{(4)}
p∥=(P0⋅B)BB2,p1⊥=P0−p∥,p2⊥=p1⊥∧BB,(5)\mathbf{p}^{\parallel}=\frac{(\mathbf{P}_{0}\cdot\mathbf{B})\mathbf{B}}{B^{2}},\qquad\mathbf{p}_{1}^{\perp}=\mathbf{P}_{0}-\mathbf{p}^{\parallel},\qquad\mathbf{p}_{2}^{\perp}=\mathbf{p}_{1}^{\perp}\wedge\frac{\mathbf{B}}{B}, \tag*{(5)}

where P0\mathbf{P}_{0} is initial polarization vector and B=∥B∥B=\lVert\mathbf{B}\rVert denotes the magnitude of the field.

We choose the coordinate system (x,y,z)(x,y,z) depicted in the inset of Figure S6, with P0∝(0,1,1)\mathbf{P}_0 \propto(0,1,1), and the local field B=B(0,sin⁡(θ+α),cos⁡(θ+α))\mathbf{B}=B(0,\sin(\theta+\alpha),\cos(\theta+\alpha)). The angle θ\theta denotes the rotation of the sample’s cc-axis with respect to the detectors and α\alpha is a hypothetical offset between the local magnetic field seen by the muon and the cc-axis of the crystal.

The asymmetry probed by a set of opposite detectors is proportional to the projection of the polarization along that direction. Therefore, in the inset of Figure S6 the asymmetry in the up-down detectors is Aup-do∝P(t)⋅e^yA_{\mathrm{up\text{-}do}}\propto\mathbf{P}(t)\cdot\hat{\mathbf{e}}_y, and the one in the forward backward detectors Afw-ba∝P(t)⋅e^zA_{\mathrm{fw\text{-}ba}}\propto\mathbf{P}(t)\cdot\hat{\mathbf{e}}_z, where e^y\hat{\mathbf{e}}_y and e^z\hat{\mathbf{e}}_z are the unit vectors in the yy and zz directions.

Inserting these conventions into Eq. (4), we find that the oscillating part of the asymmetry is given by p1⊥\mathbf{p}^{\perp}_1 and that accordingly

Aup-do∝1−2sin⁡(α+θ+π/4)sin⁡(θ+α).(6)A_{\mathrm{up\text{-}do}}\propto1-\sqrt{2}\sin(\alpha+\theta+\pi/4)\sin(\theta+\alpha). \tag*{(6)}
Afw-ba∝1−2sin⁡(α+θ+π/4)cos⁡(θ+α).(7)A_{\mathrm{fw\text{-}ba}}\propto1-\sqrt{2}\sin(\alpha+\theta+\pi/4)\cos(\theta+\alpha). \tag*{(7)}

These curves are depicted as solid lines in Figure S6, where α=−4(1)∘\alpha=-4(1)^\circ was determined using nonlinear least-squares optimization. As the sample was aligned with respect to the detector frame by hand, this small value of α\alpha could easily correspond to a small misalignment of the sample. Therefore, we conclude that within our measurement precision, the local field points along the cc-axis of the crystal.

Field dependence in transverse field {#sec:Field dependence in transverse field}

We have also determined the evolution of the muon oscillation frequencies as a function of a transverse field applied along the crystallographic cc-axis. The fast Fourier transforms of representative spectra are shown in Figure S7. We find that there is one component in the spectra, which oscillates at a frequency corresponding to the applied field, and which allows an accurate calibration of applied field (BextB_{\mathrm{ext}}). This component is partially attributed to a background contribution from a silver plate, onto which the sample was mounted. Other contributions to this component could also include some diamagnetic muons, which experience no strong magnetic field in 2H{}^{2}\mathrm{H}-MoTe2{\mathrm{MoTe}}_2. In addition, we observe other oscillating components which split up into peaks at γμ∣Bloc±Bext∣\gamma_\mu\lvert B_{\mathrm{loc}}\pm B_{\mathrm{ext}}\rvert, where Bloc∼200 mTB_{\mathrm{loc}}\sim200\,\mathrm{mT} is the field corresponding to the ZF oscillation. The accurate values of these frequencies were extracted by fitting the μ+SR\mu^{+}\mathrm{SR} time spectra to a sum of Gaussian damped oscillating components plus a non-oscillating Gaussian damped contribution:

Image not extracted

Figure S7. (a) Fourier transform of the μ+\mu^{+}SR asymmetry spectra measured in different applied fields. The curves have been offset for clarity. (b) Evolution of the oscillation frequencies as a function of the applied field.

A(t)=A0e−12(σt)2+∑nAne−12(σnt)2cos⁡(γμBnt+φ).(8)A(t)=A_0e^{-\frac{1}{2}(\sigma t)^2}+\sum_n A_ne^{-\frac{1}{2}(\sigma_nt)^2}\cos(\gamma_\mu B_nt+\varphi). \tag*{(8)}

The resulting field values BnB_n are shown in Figure S7. We note that the internal field BlocB_{\mathrm{loc}} seems to point equally likely along the positive and negative cc-axis direction. At a first glance, this seems incompatible with a muonium state, to which we attribute this signal in the main part of the paper. Usually the behavior of muonium in a field is dominated by the hyperfine coupling and described by the Hamiltonian

H=ℏIμASe+ℏγeSe⋅Bext+ℏγμIμ⋅Bext,(9)\mathcal{H} = \hbar I_{\mu} A S_{e} + \hbar\gamma_{e} S_{e} \cdot\mathbf{B}_{\mathrm{ext}} + \hbar\gamma_{\mu} I_{\mu} \cdot\mathbf{B}_{\mathrm{ext}}, \tag*{(9)}

where AA is the hyperfine coupling tensor, IμI_{\mu}, and SeS_{e} are the spins of the muon and the bound electron and γe=2π×28 GHz/T\gamma_{e} = 2\pi\times28\ \mathrm{GHz/T} is the electron’s gyromagnetic ratio. In a system without locally induced magnetism, this results in the presence of four different oscillation frequencies, which scale non-linearly as a function of field, see e.g. Refs. [S18, S21, S22].

However, in the case of muonium induced magnetism, we can assume that the electron spin SeS_{e} is locked along its local easy axis, and remains unaffected by the (relatively small) applied fields. Then Eq. (9) can be simplified to,

H=ℏIμASe+ℏγμIμ⋅Bext=ℏγμIμ⋅(Bloc+Bext),(10)\mathcal{H} = \hbar I_{\mu} A S_{e} + \hbar\gamma_{\mu} I_{\mu} \cdot\mathbf{B}_{\mathrm{ext}} = \hbar\gamma_{\mu} I_{\mu} \cdot\left(\mathbf{B}_{\mathrm{loc}} + \mathbf{B}_{\mathrm{ext}}\right), \tag*{(10)}

where ASe/γμ=:BlocA S_{e}/\gamma_{\mu} =: \mathbf{B}_{\mathrm{loc}} is an effective local magnetic field characterizing the ZF oscillation. The field dependence of the oscillation frequencies resulting from Eq. (10) is fully consistent with our observation presented in Figure S6, Figure S7, with Bloc=(±200 mT)c^\mathbf{B}_{\mathrm{loc}} = (\pm200\ \mathrm{mT})\hat{\mathbf{c}}, where c^\hat{\mathbf{c}} is a unit vector along the cc-axis.

Low-energy μ\muSR

We also investigate the evolution of the observed magnetism within ∼100 nm\sim100\ \mathrm{nm} from the surface using low energy μ+\mu^{+}SR at the μ\muE4 beamline of the Swiss Muon Source, Paul Scherrer Institute, Switzerland [33]. For these measurements, multiple single crystals were glued on a Ni coated sample plate with the cc-axis parallel to the applied field, but perpendicular to the initial muon spin. The implantation depth was controlled by varying the kinetic energy, EE, of the incident muon beam. The resulting stopping profiles were calculated with the Trim.SP software [34] and are shown in Figure S8. At 10 K and E=22.5E = 22.5 keV (corresponding to an average implantation depth of 99 nm), the muon spins exhibit a coherent precession in zero field (ZF) similar to what we observed in the bulk, see Figure S8. This spectrum was fitted to the following function:

Image not extracted

Figure S8. (a) μ+\mu^{+}SR time spectrum measured in zero field at 10 K with an implantation energy of 22.5 keV. (b) Simulated implantation profiles for different implantation energies. The average stopping depths beneath the surface for 3 keV or 22.5 keV are 18 nm and 99 nm, respectively. (c) Representative weak transverse field asymmetry spectra at 22.5 keV corresponding to (d). (d) Fraction of diamagnetic muons as a function of temperature, normalized at RT. The dashed line shows the bulk data from Ref. [13]. The solid lines in (a,c) show fits to the data.

A(t)=A1e−12(σosct)2cos⁡(γμBintt)+A2e−λslowt+A3e−12(σfastt)2.(11)A(t) = A_{1}e^{-\frac{1}{2}(\sigma_{\mathrm{osc}}t)^{2}}\cos\left(\gamma_{\mu}B_{\mathrm{int}}t\right) + A_{2}e^{-\lambda_{\mathrm{slow}}t} + A_{3}e^{-\frac{1}{2}(\sigma_{\mathrm{fast}}t)^{2}}. \tag*{(11)}

The first term describes the muons that oscillate in a local magnetic field BintB_{\mathrm{int}}. The second term represents the slowly depolarized (nonmagnetic) signal and the third accounts for the strongly damped part due to background contribution from muons stopping in the Ni coated sample plate. This analysis is similar to that used in Ref. [13], except for the background contribution. We find that the oscillation frequency in Figure S8 corresponds to an internal static field Bint=200(2)B_{\mathrm{int}} = 200(2) mT. Within the statistical error of the measurement, this is consistent with what is found in the bulk [13].

We then use a small magnetic field (5 mT, transverse to the initial muon spin) to identify the fraction of diamagnetic muons, which is proportional to the non-magnetic fraction of the sample [Fig. S8(c,d)]. These spectra were fitted with a single, Gaussian damped oscillating polarization.

A(t)=A0e−12(σt)2cos⁡(γμBt).(12)A(t) = A_0 e^{-\frac{1}{2}(\sigma t)^2} \cos(\gamma_{\mu} B t). \tag*{(12)}

This function only describes the muons in a nonmagnetic environment oscillating at a frequency close to the Larmor frequency in the applied field. Any magnetic or strongly damped signal (cf. Eq. (11)) leads to an effective loss of A0A_0. The background contribution to A0A_0 was determined by measuring an empty Ni plate with the same beamline settings. This temperature independent value was subtracted from the measured A0A_0 and the resulting curve is shown in Figure S8. Independent of depth and within the experimental uncertainties, all points in Figure S8 closely follow the bulk behavior reported in Ref. [13]. Therefore, we conclude that the bulk magnetic properties observed with μ+\mu^{+}SR in 2H{}^{2}\mathrm{H}-MoTe2_2 do not change in the near-surface region of the crystals.

125Te-NMR

The 125Te{}^{125}\mathrm{Te}-NMR signals were collected by means of spin-echo detection in a π/2\pi/2–π\pi sequence, using a 62.5 MHz to 125 MHz duplexer, with a 63 dB preamplifier. The line shapes were then extracted via a fast Fourier transformation of those signals. Typical pulse widths of 5 μ\mus and repetition times of 20 s were used, with each line shape corresponding to more than 20’000 records. Measurements on a single crystal were performed in a field of 5.007 T perpendicular to the cc-axis, while those on powders were performed in fields of 7.067 T and 4.011 T. The 125Te{}^{125}\mathrm{Te} reference frequencies were calibrated using 27Al{}^{27}\mathrm{Al}-NMR on aluminum foil, which is known to have a very small shift as a function of temperature [35].

Powder line shapes

Figure S9 shows the detailed temperature dependence of the powder line shape of 2H{}^{2}\mathrm{H}-MoTe2_2 in 7.067 T. Similar to the spectra of the single crystal measurements shown in the main part of this paper, here, too, we observe no significant peak-shift or broadening, that would be the two common fingerprints of a magnetic transition. We note that there are some small changes in the line shape as a function of temperature. However, these are most likely due to a change in spectral weight between different portions of the resonance line, owing to a difference in their spin-lattice relaxation (SLR) temperature dependence.

${}^{125}\mathrm{Te}$-NMR powder line shapes of ${}^{2}\mathrm{H}$-MoTe$_2$

Figure S9. 125Te{}^{125}\mathrm{Te}-NMR powder line shapes of 2H2H-MoTe2_2 in a field of 7.067 T at different temperatures. The spectra have been offset for clarity. The black dashed line indicates the expected position of the reference line.

Spin-lattice relaxation

The spin-lattice relaxation rate T1−1T_1^{-1} of the 125Te{}^{125}\mathrm{Te} moments is shown in Figure S10. At high temperature, the SLR shows an exponential dependence on temperature (1/T1∝eΔ/T1/T_1 \propto e^{\Delta/T}). This is a typical behavior observed in semiconductors. However, we find that here the gap Δ=27.7(3)\Delta= 27.7(3) meV is much smaller than the spectroscopic band gap, which is on the order of ∼1\sim1 eV, cf. Figure S4. This difference is likely due to, e.g., the presence of a small amount of impurity-induced in-gap states, which have too little spectral weight to be detected with ARPES, but still can depolarize the 125Te{}^{125}\mathrm{Te} nuclear spins. At present, we are unable to determine the mechanism for the increase in relaxation rate at low temperatures in 2H2H-MoTe2_2. However, we note that a similar increase in relaxation was observed in the analogue compounds ZrTe2_2 and ZrSiTe [50, 51]. In both cases, the increase in relaxation rate is suggested to arise from a contribution from Dirac electrons. However, 2H2H-MoTe2_2 is not a Dirac system, so the fact that there is a similar increase in relaxation rate suggests that the true origin of this effect might be due to some new mechanism, yet to be understood. Finally, we note that the SLR has a minimum around 40 K, i.e., the onset temperature of the magnetic oscillations observed by μ+\mu^{+}SR. However, in the presence of an intrinsic magnetic transition, the SLR should instead show a peak at this temperature due to critical magnetic fluctuations. This finding confirms once more that the observed local magnetic field in μ+\mu^{+}SR is not intrinsic to 2H2H-MoTe2_2.

Image not extracted

Figure S10. 125^{125}Te-NMR spin lattice relaxation rate in 2^{2}H-MoTe2_2 in a field of 7.067 T as a function of temperature. The dashed line indicates the activated Arrhenius-like behavior at high temperature.

References

References

  1. [1]A. K. Geim and I. V. Grigorieva, Nature 499, 419 (2013).
  2. [2]C. R. Rajamathi, U. Gupta, N. Kumar, H. Yang, Y. Sun, V. Süß, C. Shekhar, M. Schmidt, H. Blumtritt, P. Werner, et al., Advanced Materials 29, 1606022 (2017).
  3. [3]R. Clarke, E. Marsiglio, and H. P. Hughes, Philosophical Magazine B 38, 121 (1978).
  4. [4]Y. Sun, S.-C. Wu, M. N. Ali, C. Felser, and B. Yan, Physical Review B 92, 161107(R) (2015).
  5. [5]Y. Qi, P. G. Naumov, M. N. Ali, C. R. Rajamathi, W. Schnelle, O. Barkalov, M. Hanfland, S.-C. Wu, C. Shekhar, Y. Sun, et al., Nature Communications 7, 11038 (2016).
  6. [6]N. Xu, Z. W. Wang, A. Magrez, P. Bugnon, H. Berger, C. E. Matt, V. N. Strocov, N. C. Plumb, M. Radovic, E. Pomjakushina, K. Conder, J. H. Dil, J. Mesot, R. Yu, H. Ding, and M. Shi, Physical Review Letters 121, 136401 (2018).
  7. [7]I. G. Lezama, A. Ubaldini, M. Longobardi, E. Giannini, C. Renner, A. B. Kuzmenko, and A. F. Morpurgo, 2D Materials 1, 021002 (2014).
  8. [8]H. Xu, S. Fathipour, E. W. Kinder, A. C. Seabaugh, and S. K. Fullerton-Shirey, ACS Nano 9, 4900 (2015).
  9. [9]Y. Ding, N. Zhou, L. Gan, X. Yan, R. Wu, I. H. Abidi, A. Waleed, J. Pan, X. Ou, Q. Zhang, et al., Nano Energy 49, 200 (2018).
  10. [10]S. Song, D. H. Keum, S. Cho, D. Perello, Y. Kim, and Y. H. Lee, Nano Letters 16, 188 (2016).
  11. [11]Y. Wang, J. Xiao, H. Zhu, Y. Li, Y. Alsaid, K. Y. Fong, Y. Zhou, S. Wang, W. Shi, Y. Wang, et al., Nature 550, 487 (2017).
  12. [12]P. M. Coelho, H.-P. Komsa, K. Lasek, V. Kalappattil, J. Karthikeyan, M.-H. Phan, A. V. Krasheninnikov, and M. Batzill, Advanced Electronic Materials 0, 1900044 (2019).
  13. [13]Z. Guguchia, A. Kerelsky, D. Edelberg, S. Banerjee, F. v. Rohr, D. Scullion, M. Augustin, M. Scully, D. A. Rhodes, Z. Shermadini, et al., Science Advances 4, eaat3672 (2018).
  14. [14]Y. Ma, Y. Dai, M. Guo, C. Niu, J. Lu, and B. Huang, Physical Chemistry Chemical Physics 13, 15546 (2011).
  15. [15]M. B. Kanoun, Journal of Alloys and Compounds 748, 938 (2018).
  16. [16]Additional details are available in the Supplementary Information, containing Refs. [13, 14, 23–25, 30–51].
  17. [17]J. I. Budnick, L. J. Bruner, R. J. Blume, and E. L. Boyd, Journal of Applied Physics 32, S120 (1961).
  18. [18]D. Tay, T. Shang, P. Puphal, E. Pomjakushina, H.-R. Ott, and T. Shiroka, Physical Review B 102, 241109(R) (2020).
  19. [19]Z. T. Zhang, D. Dmytrieva, S. Molatta, J. Wosnitza, Y. Wang, M. Helm, S. Zhou, and H. Kühne, Physical Review B 95, 085203 (2017).
  20. [20]F. Weber, R. Hott, R. Heid, L. L. Lev, M. Caputo, T. Schmitt, and V. N. Strocov, Physical Review B 97, 235122 (2018).
  21. [21]D. Rhodes, R. Schönemann, N. Aryal, Q. Zhou, Q. R. Zhang, E. Kampert, Y.-C. Chiu, Y. Lai, Y. Shimura, G. T. McCandless, J. Y. Chan, D. W. Paley, J. Lee, A. D. Finke, J. P. C. Ruff, S. Das, E. Manousakis, and L. Balicas, Physical Review B 96, 165134 (2017).
  22. [22]N. Aryal and E. Manousakis, Physical Review B 99, 035123 (2019).
  23. [23]J. S. Möller, D. Ceresoli, T. Lancaster, N. Marzari, and S. J. Blundell, Physical Review B 87, 121108(R) (2013).
  24. [24]F. Bernardini, P. Bonfà, S. Massidda, and R. De Renzi, Physical Review B 87, 115148 (2013).
  25. [25]H. González-Herrero, J. M. Gómez-Rodríguez, P. Mallet, M. Moaied, J. J. Palacios, C. Salgado, M. M. Ugeda, J.-Y. Veuillen, F. Yndurain, and I. Brihuega, Science 352, 437 (2016).
  26. [26]V. Baltz, A. Manchon, M. Tsoi, T. Moriyama, T. Ono, and Y. Tserkovnyak, Reviews of Modern Physics 90, 015005 (2018).
  27. [27]J. D. Sau, S. Tewari, R. M. Lutchyn, T. D. Stanescu, and S. Das Sarma, Physical Review B 82, 214509 (2010).
  28. [28]T. Hesjedal and Y. Chen, Nature Materials 16, 3 (2017).
  29. [29]J. A. Krieger, Doctoral Thesis, ETH Zurich (2020), URL https://doi.org/10.3929/ethz-b-000460760.
  30. [30]A. Amato, H. Luetkens, K. Sedlak, A. Stoykov, R. Scheuermann, M. Elender, A. Raselli, and D. Graf, Review of Scientific Instruments 88, 093301 (2017).
  31. [32]A. Suter and B. M. Wojek, Physics Procedia 30, 69 (2012).
  32. [33]T. Prokscha, E. Morenzoni, K. Deiters, F. Foroughi, D. George, R. Kobler, A. Suter, and V. Vrankovic, Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment 595, 317 (2008).
  33. [34]W. Eckstein, Computer Simulation of Ion-Solid Interactions, Springer Series in Materials Science (Springer-Verlag, Berlin Heidelberg, 1991), ISBN 978-3-642-73515-8.DOI
  34. [35]R. K. Harris, E. D. Becker, S. M. C. d. Menezes, R. Goodfellow, and P. Granger, Magnetic Resonance in Chemistry 40, 489 (2002).
  35. [38]V. N. Strocov, M. Shi, M. Kobayashi, C. Monney, X. Wang, J. Krempasky, T. Schmitt, L. Patthey, H. Berger, and P. Blaha, Physical Review Letters 109, 086401 (2012).
  36. [39]T. Böker, R. Severin, A. Müller, C. Janowitz, R. Manzke, D. Voß, P. Krüger, A. Mazur, and J. Pollmann, Physical Review B 64, 235305 (2001).
  37. [40]G. Kresse and J. Furthmüller, Physical Review B 54, 11169 (1996).
  38. [41]G. Kresse and D. Joubert, Physical Review B 59, 1758 (1999).
  39. [42]J. P. Perdew, K. Burke, and M. Ernzerhof, Physical Review Letters 77, 3865 (1996).
  40. [43]G. Henkelman and H. Jónsson, The Journal of Chemical Physics 113, 9978 (2000).
  41. [45]K. Momma and F. Izumi, Journal of Applied Crystallography 44, 1272 (2011).
  42. [50]Y. Tian, N. Ghassemi, and J. H. Ross, Physical Review B 102, 165149 (2020).
  43. [51]Y. Tian, Y. Zhu, R. Li, Z. Mao, and J. H. Ross, Physical Review B 104, L041105 (2021).

Paper details

vlnc:W_2wbnq3yh
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents