Search

MoS2 has attracted intense interest for many applications. Natural MoS2 and field-effect transistors made of it generally exhibit n-type characteristics, but its origin is unknown. Herein, we show that C is the origin of the universal n-type doping of natural MoS2. Photoemission spectroscopies reveal that while many MoS2 samples with C detected are n-type, some without C exhibit p-type characteristics. The C-free, p-type MoS2 changes to n-type over time with the concurrent appearance of C that is out-diffused from bulk, indicating that C induces the n-type doping. The C-origin is verified by C-deposition and supported by theoretical calculations. This carbon appears as nanometer-scale defects frequently observed in scanning tunneling microscopy. In addition, we propose based on the calculations that S vacancies are responsible for the p-type characteristics, which contrasts the widespread belief. This work provides new perspectives on MoS2 doping and presents a new direction for fabricating reliable MoS2 devices.

Vanadium ditelluride (VTe2) has been intensively explored to understand the charge density wave (CDW) phase and its connection to magnetic properties. Here, we conduct a systematic study to understand the fine structure of CDW via scanning tunneling microscopy (STM) combined with density functional theory (DFT) calculations. STM topograph at 79 K shows that a CDW phase in VTe2 has a stripe modulation with 3 × 1 periodicity, following the double zigzag chain of distorted Te lattices. Interestingly, the 3 × 1 CDW modulation undergoes contrast inversion between filled and empty state topographs. Atomistic features and contrast changes of CDW observed in STM are clearly reproduced in our DFT simulation images. Charge distribution calculation indicates that the spatial extension and density of Te 5 p orbitals have strong variations with filled and empty states, explaining the fine structure of 3 × 1 CDW in VTe2. Our finding provides an inspiring insight to further research on the less explored electronic structure of VTe2.

The identification of the surface atomic structure having a large unit cell remains a major challenge, while this surface structure is the starting point for templated growth on it. The atomic structure of the carbon-induced R(15 × 3)-W(110) is investigated by scanning tunneling microscopy (STM) operated around 80 K in near-contact mode, combined with the first-principles calculations. It is shown that the R(15 × 3) reconstruction results from the moire pattern, which a W-C bilayer of α-W2C(0001) forms with the W(110) substrate, confirming an earlier prediction by Bauer et al. Besides the moire pattern, the present structure contains the surface lattice deformation, the main ingredient of the other model, and provides a comprehensive picture, well reconciling the two previous models.

Transition-metal containing metal phthalocyanine (MPc) molecules are representative molecular magnets and their adsorption on two-dimensional metallic surfaces has been extensively studied. In/Si(111)-4×1 is a prototypical quasi-one-dimensional metallic surface made up of alternating In nanowires and Si Seiwatz chains. Here, we investigated the adsorption of MPc (M = Mn, Fe, Co) on In/Si(111)-4×1 using density functional theory calculations. We found that there are three locally stable adsorption sites of MPc, two on In nanowires and one on Seiwatz chains. Also, the stable orientation of MPc is different depending on adsorption sites. Among three MPc’s, MnPc are found to carry a finite magnetic moment of 1.6 ~ 2.1 μB in its three almost-degenerate adsorption structures. The interaction between MPc molecules and the In/Si(111)-4×1 surface turns out to occur via the charge transfer, and the charge transfer channel varies for different adsorption sites: Transition-metal 3dz2 and two N 2pz orbitals on In nanowires vs. four N 2pz orbitals on Seiwatz chains.

The adsorption of acetonitrile (CH3CN) on Si(111)-(7 × 7) at a room temperature has been investigated using scanning tunneling microscopy (STM) and first-principles calculations. The site-specific ...

One-dimensional indium chains on Si(111) exhibit both temperature-and defect-induced perioddoubling (×2) structural transitions, and their natures have been focus of recent investigations. Using density functional perturbation theory calculations, we examined the vibrational properties of the room-temperature Si(111)In-4×1 structure. The phonon band structure revealed two unstable modes at the zone-boundary in the chain direction, which lead to a lattice instability towards a structure with parallel trimers. This lattice-instability-driven 4 × 2 structure is different from the low-temperature 4×2 (LT-4×2) structure (hexagon structure). The result suggests that the roomtemperature phase is neither a static 4×1 phase nor the dynamical fluctuation of the LT-4×2 phase. We demonstrate that the room-temperature phase is a dynamically fluctuating parallel-trimer 4×2 phase, and the fluctuations become suppressed near the defects to result in the defect-induced ×2 structure.

Macrocyclic crown ethers (CEs) have tunable cavity sizes and site-selective binding with metal ions, making the CE-ion complex a promising candidate as a two-dimensional (2D) electrolyte. In this work, density functional theory method is used to determine the energetically stable structures of 12-crown-4 ether (CE4) and 15-crown-5 ether (CE5) complexed with four cations: Li + , Na + , Mg 2 + , Ca 2 + . In addition to the CE-ion binding energies, the diffusion barriers for ion transport through the CE cavities are calculated. Among the complexes investigated, CE5 presents the lowest energy barrier for ion diffusion. The barriers for Li + travelling through a single CE5 and moving between two CE5s are 0.29 eV and 0.16 eV, respectively. Field-controlled modulation of the diffusion barrier is also demonstrated. By applying a 0.15 V/A electric field perpendicular to the plane of the CE, the diffusion barrier of Li + through one CE5 can be reduced from 0.29 to 0.20 eV to facilitate the ion transport.

Adsorption of CO molecules on Si(111)-(7×7) is investigated by using scanning tunneling microscopy (STM) and density-functional theory calculations. The most reactive site on the Si(111)-(7×7) surface is the corner adatom in the faulted half unit (FHU), followed by the center adatom in the FHU. The initial sticking probability of CO molecules on Si(111)-(7×7) at room temperature (RT) is estimated to be ∼ 4 × 10 10 molecules/( cm 2 ·Langmuir), which is comparable with that on Si(001)-(2×1). From the experiments and theoretical calculations, the adsorption of CO molecules are found to occur on Si adatoms either in the upright on-top configuration or in the back-bond inserted configuration, while the adsorption on the rest and the corner hole atoms (which are theoretically probable) are not observable using STM due to their low-lying geometries. Though the sticking probability is very low, every surface dangling bond on the surface can bind strongly with the C atom of a CO molecule even at temperatures higher than RT.

A triangular domain structure within hexagonal monolayer ${\mathrm{WS}}_{2}$ is found to be induced by edge polarization. Based on density-functional theory calculations, the local electrostatic potential over ${\mathrm{WS}}_{2}$ is investigated, and with the S-adatom and ${\mathrm{S}}_{2}$-addimer reconstructions at the W edges, the domains facing the W edges are found to be lower by up to 1.8 eV in the electrostatic potential than those facing the S edges. The edge-induced band bending in the terminated two-dimensional system is a two-dimensional analogy of the surface band bending effect in the terminated three-dimensional system. The largest potential difference is found to be comparable to the electronic gap of ${\mathrm{WS}}_{2}$, and near the edges, the W and S domains can be $n$- and $p$-type doped, respectively, by which the hexagonal monolayer ${\mathrm{WS}}_{2}$ can be applied to a bipolar electronic junction device.

Recent studies on a quasi-one-dimensional (quasi-1D) charge-density-wave (CDW) system, In atomic wires on Si(111), have brought intriguing issues on topological solitons in quasi-1D systems: the existence of few-atom-sized solitons, chirality of solitons, and the realization of logic exploiting the chirality switching. Using scanning tunneling microscopy and first-principles calculations, we show that the previously reported ``short'' phase-flip defects are In adatoms and thus have nonsolitonic nature, resolving the controversy over the existence of highly localized solitons. The observed ``long'' phase-flip and phase-slip defects are genuine solitons with and without chirality, respectively. While achiral solitons (phase-slip defects) can exist on the pristine CDW ($8\ifmmode\times\else\texttimes\fi{}2$) surface, chiral solitons (phase-flip defects) cannot due to their breakage of $8\ifmmode\times\else\texttimes\fi{}2$ ordering. The chiral solitons can exist only when they are trapped by In adatoms and constitute a part of a closed-loop domain wall. The intertwinement of chiral solitons and In adatoms implies the limitations of the previously proposed logic utilizing soliton chirality, but it provides an opportunity to realize this by controlling the In-adatom defect.

We report density functional theory calculations on the vacancy-induced magnetic properties of the wurzite GaN $$(10\bar 10)$$ surface. Among various intrinsic vacancies on the surface, we focus on the monatomic Ga and N vacancies. We have found two stable geometries of a surface Ga vacancy and one stable geometry of a surface N vacancy. Two stable surface Ga vacancy structures are shown to have magnetic moments of 3 μ B , which is the same as for the bulk Ga vacancy. Differently from the bulk N vacancy carrying null magnetic moment, the monatomic surface N vacancy has a finite magnetic moment of 1 μ B . These results support the experimental claim that the roomtemperature ferromagnetism observed in inorganic nanoparticles is due to surface vacancies and their complexes. We also present simulated scanning tunneling microscope images for comparison with those obtained in future experimental studies.

Incorporation of carbon (C) atoms in the Si(111)-(7 × 7) surface is investigated by using scanning tunneling microscopy and density-functional theory calculations. C atoms are supplied through thermal dissociation of CO molecules adsorbed on the Si(111)-(7 × 7) surface. One C atom can be incorporated substitutionally below the Si adatoms of the (7 × 7) structure. However, the stable (7 × 7) surface can accommodate only a very small amount of C atoms and the surplus C atoms are swept away to the step edges in the form of silicon-carbon clusters, restoring the clean Si(111)-(7 × 7) structure on the flat terrace. In contrast to the B-induced 3 × 3 reconstruction, high density C atoms do not induce a single crystalline 3 × 3 reconstruction over a wide area of Si(111).

Recent studies on a quasi-one-dimensional (quasi-1D) charge-density-wave (CDW) system, In atomic wires on Si(111), have brought intriguing issues on topological solitons in quasi-1D systems: the existence of few-atom-sized solitons, chirality of solitons, and the realization of logic exploiting the chirality switching. Using scanning tunneling microscopy and first-principles calculations, we show that the previously reported "short" phase-flip defects are In adatoms and thus have nonsolitonic nature, resolving the controversy over the existence of highly localized solitons. The observed "long" phase-flip and phase-slip defects are genuine solitons with and without chirality, respectively. While achiral solitons (phase-slip defects) can exist on the pristine CDW (8×2) surface, chiral solitons (phase-flip defects) cannot due to their breakage of 8×2 ordering. The chiral solitons can exist only when they are trapped by In adatoms and constitute a part of a closed-loop domain wall. The intertwinement of chiral solitons and In adatoms implies the limitations of the previously proposed logic utilizing soliton chirality, but it provides an opportunity to realize this by controlling the In-adatom defect.

We report density-functional-theory calculations on vacancies on the ZnO\((10\bar 10)\) surface with the on-site Coulomb interaction correction. Zn and O vacancies and three different configurations of ZnO divacancies were examined considering all the possible charge states. Using the calculated vacancy formation energies, we constructed the phase diagram by spanning the allowed range of electronic and ionic chemical potential. Differently from the bulk where the divacancy is not stabilized at all, one configuration of surface divacancies can be stable in a certain region of the chemical potentials. Among all the stable vacancies, only the neutral and the singly negatively-charged monatomic Zn vacancies exhibit a finite magnetic moment of 1 μB, which implies that the surface Zn vacancies may be responsible for the experimentally-observed room-temperature ferromagnetism in ZnO nanocrystals.

We have investigated the vibrational properties of orthorhombic PtSi by performing first-principles calculations. The calculated phonon band structure shows unstable phonon modes around the Γ-point. The ionic displacement vectors associated with the unstable Γ-point phonon mode suggest a structural modulation to a lower-symmetry phase, which was confirmed by comparing the theoretical equations of states of the modulated (P212121) and the unmodulated (Pnma) PtSi. The symmetry-reduced modulated orthorhombic PtSi shows a phonon band gap. We attribute the origin of the phonon band gap to the separation of the low-frequency Pt phonon modes and the high-frequency Si phonon modes due to the large atomic mass difference between Pt and Si.

We report density functional theory calculations for the bulk ZnO divacancy in hexagonal ZnO and compare the results with the monatomic bulk vacancies. We find that the ZnO divacancy could be viewed as a tightly-bound state of a doubly-negative Zn vacancy and the doubly-positive O vacancy from the structural and the electronic points of view. The singly-positive ZnO divacancy is found to carry a magnetic moment of 1µB, but it can only be stabilized in very high p-type samples. Otherwise, the non-magnetic neutral divacancy is stable. These results suggest that the ZnO divacancies in the bulk region cannot play a significant role in stabilizing the recently-observed ferromagnetism in ZnO nanoparticles.

We investigated the structural and the binding properties of black phosphorus (black-P) by employing density functional theory (DFT) calculations in combination with various implementations of the van der Waals (vdW) interaction. Both the binding energy curve of the two isolated puckered layers and the equation of states of the bulk black-P suggest that the conventional generalized gradient approximation (GGA) functional is incapable of describing the interlayer vdW interaction. From the comparison of the seven different vdW implementations, the appropriate vdW schemes in describing layer-structured black-P are found to be either the Grimme’s dispersion correction (DFT-D2) or the optB86b vdW denisty-functional approach.

Industrial demands for sustainable and renewable energy resources have inspired studies on photonic and electronic properties of graphitic-carbon nitride (g-C3N4) as a promising photocatalyst without precious metal. The absorption and the yield by metal-free pristine g-C3N4 are, however, still limited with hydrogen/oxygen evolution reaction (HER/OER) mostly around ultraviolet-light. Here, we propose the graphene-decorated g-C3N4 as a metal-free photocatalyst under visible-light, based on our atomic-scale measurements and calculations. The g-C3N4 nanosheets on highly-oriented pyrolytic graphite (HOPG) exhibit band-gaps appropriate for visible-light absorption and work-functions tuned for band alignments to supply electrons and holes for HER/OER. Scanning probe microscopy (SPM) measurements for local density of states (LDOS) in atomic scale and work-functions in nanometer scale with ab initio calculations confirmed the various electronic transitions for each nitrogen and carbon atom in different atomic registries. The graphene-decorated g-C3N4, therefore, could provide a breakthrough enabling the efficient water-splitting reactions under visible-light without precious metal.

We have performed spin-polarized density functional theory calculations on κ-carbides, (Fe, Mn)3AlC, to investigate both the effect of different exchange-correlation functionals and the effect of Mn substitution. Differently from our previous calculations using the Perdew-Wang functional, Fe2MnAlC is found to be the most stable crystal among crystalline κ-carbides by using the Perdew, Burke, and Ernzerhof functional. Supercell calculations to simulate low Mn concentration show that substitutional Mn atoms hardly interact with each other and suggest that a random alloy model can be applied. The stabilization of Fe2MnAlC and its enhanced magnetization are attributed to the formation of a -Mn-C- linear chain structure.

We perform spin-polarized density functional theory calculations for a Zn vacancy on the ZnO(10 $\bar 1$ 0) surface. Two stable configurations of the surface Zn vacancy are found, and the activation energy barrier is estimated to be ∼0.01 eV. The lower energy configuration has a newly formed surface Zn-O bond to restore the bulk-like structure on the surface. Due to the newly formed bond, the vacancy state in the band gap is characterized by a complicated hybridization of neighboring surface and subsurface atoms and by a more extended electron density. Despite such a hybridization, the surface Zn vacancy is found to have a robust magnetic moment of 1 μ B , implying that surface Zn vacancies may be responsible for the ferromagnetism observed in ZnO thin films and nanoparticles. Simulated scanning tunneling microscope images show that the two structures of the surface Zn vacancy can be distinguished in the filled-state images.

Understanding and effectively predicting the thermal stability of ternary transition metal oxides with heavy elements using first principle simulations are vital for understanding performance of advanced materials. In this work, we have investigated the thermal stability of mullite RMn2O5 (R = Bi, Pr, Sm, or Gd) structures by constructing temperature phase diagrams using an efficient mixed generalized gradient approximation (GGA) and the GGA + U method. Simulation predicted stability regions without corrections on heavy elements show a 4-200 K underestimation compared to our experimental results. We have found the number of d/f electrons in the heavy elements shows a linear relationship with the prediction deviation. Further correction on the strongly correlated electrons in heavy elements could significantly reduce the prediction deviations. Our corrected simulation results demonstrate that further correction of R-site elements in RMn2O5 could effectively reduce the underestimation of the density functional theory-predicted decomposition temperature to within 30 K. Therefore, it could produce an accurate thermal stability prediction for complex ternary transition metal oxide compounds with heavy elements.

The initial adsorption of oxygen molecules on Si(001) is investigated at room temperature. The scanning tunneling microscopy images reveal a unique bright O2-induced feature. The very initial sticking coefficient of O2 below 0.04 Langmuir is measured to be ∼0.16. Upon thermal annealing at 250–600 °C, the bright O2-induced feature is destroyed, and the Si(001) surface is covered with dark depressions that seem to be oxidized structures with -Si-O-Si- bonds. This suggests that the observed bright O2-induced feature is an intermediate precursor state that may be either a silanone species or a molecular adsorption structure.

Silicon nanoclusters have become significant research interest due to their potential application to optoelectronic devices in visible range. We investigate the electronic and optical properties of hydrogenated and nitrogen-passivated silicon nanoclusters using density functional theory calculations. The energy gap between the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) of nanoclusters have varying sizes. They are systematically studied using the conventional local density approximation, the generalized gradient approximation, and the time-dependent density functional theory calculations with hybrid functional. The HOMO-LUMO gap is found to decrease monotonically as the size of nanocluster increases. Introducing one and two nitrogen passivants to a Si29H36 nanocluster, we find that the HOMO-LUMO gap decreases as the number of nitrogen passivants increases. It suggests that multi-nitrogen passivants may enable light emission in visible range from smaller clusters.

The Cl-saturated Si(001) surface is investigated by employing density functional theory calculations. From the electronic density of states, the low- and the high-bias filled-state images and the empty-state images are found to originate from Cl(3p) states directed in-plane, in parallel with the Si dimer row, and perpendicular to the surface (z), respectively. The bright triplet, which is a characteristic scanning tunneling microscopy (STM) feature of an unchlorinated bare Si dimer, is attributed to an occupied dangling bond state of the bare Si dimer that is hybridized with the neighboring Cl(3pz) orbital. The previously observed split-dimer feature is confirmed to be the dimer-bond oxidized structure. Finally, we suggest that the initial back-bond oxidation does not occur on the chlorinated Si(001) surface because it is energetically not favored compared with dimer-bond oxidation and because its STM simulation does not match experimental images.

We investigated the structural and the binding properties of tetrahedral P4 molecular pairs by employing density functional theory (DFT) calculations using both the local density approximation (LDA) and the generalized gradient approximation (GGA). To consider the van der Waals interaction, we employed the GGA with dispersion correction, specifically Grimme’s approach (DFT-D2). This DFT-D2 in conjunction with the GGA (GGA-D2) well fixes the failure of the GGA in reproducing a reasonable binding energy curve. The binding energy and the equilibrium intermolecular distance are found to be in close correlation with the number of adjacent atomic pairs between two molecules. Application of the LDA and the GGA-D2 to a molecular crystal β-P4 and comparison with experiments confirms that the GGA-D2 should be used for a reliable description of P4-based molecular systems.

The dissociative adsorption of water molecules on Si(001) was studied up to 850 K by scanning tunneling microscopy. Water molecules dissociate into OH and H fragments to form on-dimer (OD) and interdimer (ID) configurations, and the population ratio nID/nOD is measured to change from ∼5 at room temperature to the saturation value of ∼0.6 above 500 K, contradicting to the prediction from the static transition-state theory. The sticking coefficient also shows an abnormal temperature dependence. The temperature dependence in the H dissociation and in the molecular adsorption can be elucidated only by taking into account the flip-flop motion of Si dimers.

The diffusion properties of He and Xe in UO2 have been investigated, using density-functional calculations employing the projector–augmented–wave (PAW) method and the generalized gradient approximation (GGA). The migration energies corresponding to both interstitial and vacancy-assisted mechanisms have been calculated and the results for the two noble gas atoms are compared with each other. We suggest that He likely diffuses by hopping through a single vacancy with computed low migration energies smaller than 0.79 eV and its diffusivity is much higher than that of Xe. Xe has a quite large migration energy compared to He; the strain energy plays a key role in Xe diffusion in UO2.

We have performed ab initio total energy calculations to investigate the defect energetics and diffusion behavior of Xe in UO 2 and ThO 2 matrices. All calculations have been carried out using density functional theory within the generalized gradient approximation and applying the projector-augmented-wave method. Our results show that the formation and migration energies of vacancy defects are more than twice as high in ThO 2 compared with UO 2 . Another notable difference between the two oxides is the role played by an oxygen vacancy in the movement of a cation vacancy. An vacancy enhances the movement of a uranium vacancy by lowering its migration energy by about 1 eV, but a similar effect is not observed in ThO 2 . The different behavior of cation vacancies in the two oxides strongly affects the mobility of fission gases and leads to differences in their respective diffusion behavior. We suggest that the strong resistance against oxidation of ThO 2 prevents the creation and migration of defects, and results in a lower mobility of fission gases in ThO 2 as compared to UO 2 .

We have investigated the optimal growth conditions of carbon nanotubes (CNTs) using the chemical vapor deposition and the Ni nanoparticle arrays. The diameter of the CNT is shown to be controlled down to below 20 ㎚ by changing the size of Ni particle. The position and size of Ni particles are controlled continuously by using wafer-scale compatible methods such as lithography, ion-milling, and chemical etching. Using optimal growth conditions of temperature, carbon feedstock, and carrier gases, we have demonstrated that an individual CNT can be grown from each Ni nanoparticle with almost 100% probability over wide area of SiO₂/Si wafer. The position, diameter, and wall thickness of the CNT are shown to be controlled by adjusting the growth conditions.

We have investigated vacancy-assisted diffusion of Xe in uranium dioxide (UO 2 ) calculating incorporation, binding, and migration energies. All the energy values have been obtained using the density functional theory (DFT) within the generalized gradient approximation (GGA) and the projector-augmented-wave (PAW) method. Considering spin-polarization effect, we find that the computed migration energy is reduced by and agrees well with experimental data compared to those obtained from non-magnetic calculations. We also find that an oxygen vacancy lowers the migration energy of a uranium vacancy by about 1 eV, enhancing an effective movement of vacancy clusters consisting of both uranium and oxygen vacancies. Furthermore, the strain energy of Xe is large enough to contribute to the clustering of vacancies making it the driving force for the vacancy-assisted diffusion of Xe in UO 2 . In summary all the calculated results suggest that the trivacancy is a major diffusion pathway of Xe in UO 2 .

Scanning tunneling microscopy was used to study the adsorption behaviors of NH 3 molecules across Si dimer rows on a Si(0 0 1)-2 × 1 surface, taking into account the existence of both the on-dimer (OD) and inter-dimer (ID) adsorption configurations. In the OD configuration, the dissociated NH 2 and H form bonds with two Si atoms of a dimer. On the other hand, in the ID configuration, they adsorb at two Si atoms on the same side of two adjacent dimers along the dimer row. The OD configurations on the neighboring dimer rows were preferentially aligned perpendicular to the dimer rows, indicating a strong attractive interaction across the dimer rows, whereas the NH 3 molecules adsorbed in the ID configuration rarely showed ordering behavior.

We investigate the effect of Ge on the retardation of B diffusion in SiGe alloys through first-principle calculations, and find that the Ge bonding effect is most significant in the nearest-neighborhood of B. The B dopant diffuses from a self-interstitial–B pair via an interstitialcy mechanism for neutral charge state, while a kick-out mechanism is also possible for 1+ charge state. The migration and activation energies depend on the number and positions of the Ge atoms and are generally enhanced by the presence of Ge, reducing the B diffusivity.

We investigate the effects of dielectric environments on the transition levels of Re and Nb dopants in monolayer ${\mathrm{MoS}}_{2}$ through density functional theory calculations. The inherently weakly screened Coulomb interaction in free-standing monolayer ${\mathrm{MoS}}_{2}$ makes the dopant electrons and holes strongly bound, and the Re and Nb impurities are found to produce deep levels. It is shown that when the monolayer ${\mathrm{MoS}}_{2}$ is placed near high permittivity dielectrics the screened Coulomb interaction induces carrier delocalization with generating shallow levels.

There are many interests to achieve long-range magnetic order in topological insulators of Bi2Se3 or Bi2Te3 by doping magnetic transition metals such as Fe and Mn. The transition metals act as not only magnetic dopants but also electric dopants because they are usually divalent. However, if the doping elements are rare-earth metals such as Gd, which are trivalent, only magnetic moments can be introduced. We fabricated single crystals of Bi2-xGdxTe3 (0 ≤ × ≤ 0.2), in which we observed magnetic phase change from paramagnetic (PM) to antiferromagnetic (AFM) phase by increasing x. This PM-to-AFM phase transition agrees with the density functional theory calculations showing a weak and short-ranged Gd-Gd AFM coupling via the intervening Te ions. The critical point corresponding to the magnetic phase transition is x = 0.09, where large linear magnetoresistance and highly anisotropic Shubnikov-de Haas oscillations are observed. These results are discussed with two-dimensional properties of topological surface state electrons.

We have performed total-energy calculations to study theoretical scanning tunneling microscopy (STM) images of the Si(1 1 1)3 × 2 surfaces induced by the adsorption of alkaline-earth metals (AEMs). Previously, in a series of works on Ba/Si(1 1 1) system, we have found that the observed Si(1 1 1)3 × 1–Ba LEED phase indeed has a 3 × 2 periodicity with a Ba coverage of 1/6 ML and the HCC substrate structure. Based on results of the Ba case, we proposed that the HCC structure is also adopted for other AEM atoms, which was confirmed by our recent work. In this paper, we mainly report the STM simulations for different AEM systems to compare with existing experimental data. We discuss the difference in the detailed STM images for different AEM adsorbates. Especially, the difference in filled-state images between Mg and other AEM atoms is attributed to the strong Mg–Si interaction.