VASP inputs and outputs for the:\n\n\n\tAr2 binding energy curve (Ar2_bpara.tar.gz)\n\tL28 set of layered-solid geometries and interlayer binding energies (L28.tar.gz)\n\tS22 set of interaction energies of weakly-bound complexes (S22.tar.gz)\n\n\nAll POTCAR files have been replaced by "potcar.txt" files containing the title(s) of the POTCAR(s) needed to reproduce the calculations. A preprint is available from the arXiv:2204.11717 (link).
First Page
In addition to its technological relevance, silicon poses a challenge for first principles simulations because it undergoes a semiconductor-to-metal transition upon melting. Moreover, the resulting metallic liquid contains a mixture of metallic and covalent bonding. This coexistence of fundamentally different interactions is difficult to describe within approximate density functional methods, which oftentimes cannot accurately describe these two extremes simultaneously. We report an investigation of the structure, dynamics, and thermodynamics of liquid silicon using ab initio molecular dynamics simulations with three density functional approximations: the local density approximation, the Perdew-Burke-Ernzerhof generalized gradient approximation, and the strongly constrained and appropriately normed (SCAN) meta-generalized gradient approximation. We demonstrate that SCAN describes this liquid with better accuracy than the other often-used functionals because it can simultaneously capture covalent and metallic bonding with similar high accuracy.
End-to-end models have gradually become the preferred option for automatic speech recognition (ASR) applications. During the training of end-to-end ASR, data augmentation is a quite effective technique for regularizing the neural networks. This paper proposes a novel data augmentation technique based on semantic transposition of the transcriptions via syntax rules for end-to-end Mandarin ASR. Specifically, we first segment the transcriptions based on part-of-speech tags. Then transposition strategies, such as placing the object in front of the subject or swapping the subject and the object, are applied on the segmented sentences. Finally, the acoustic features corresponding to the transposed transcription are reassembled based on the audio-to-text forced-alignment produced by a pre-trained ASR system. The combination of original data and augmented one is used for training a new ASR system. The experiments are conducted on the Transformer[2] and Conformer[3] based ASR. The results show that the proposed method can give consistent performance gain to the system. Augmentation related issues, such as comparison of different strategies and ratios for data combination are also investigated.
Most of the group IV, III-V, and II-VI compounds crystallize in semiconductor structures under ambient conditions. Upon application of pressure, they undergo structural phase transitions to more closely packed structures, sometimes metallic phases. We have performed density functional calculations using projector augmented wave (PAW) pseudopotentials to determine the transition pressures for these transitions within the local density approximation (LDA), the Perdew-Burke-Ernzerhof (PBE) generalized gradient approximation (GGA), and the strongly constrained and appropriately normed (SCAN) meta-GGA. LDA underestimates the transition pressure for most of the studied materials. PBE under- or overestimates in many cases. SCAN typically corrects the errors of LDA and PBE for the transition pressure. The accuracy of SCAN is comparable to that of computationally expensive methods like the hybrid functional HSE06, the random phase approximation (RPA), and quantum Monte Carlo (QMC), in cases where calculations with these methods have been reported, but at a more modest computational cost. The improvement from LDA to PBE to SCAN is especially clearcut and dramatic for covalent semiconductor-metal transitions, as for Si and Ge, where it reflects the increasing relative stabilization of the covalent semiconducting phases under increasing functional sophistication.
Jellium, a simple model of metals, is a standard testing ground for density functionals both for bulk and for surface properties. Earlier tests show that the Tao--Perdew--Staroverov--Scuseria (TPSS) nonempirical metageneralized gradient approximation (meta-GGA) for the exchange-correlation energy yields more accurate surface energies than the local spin density (LSD) approximation for spin-unpolarized jellium. In this study, work functions and surface energies of a jellium metal in the presence of ``internal'' and external magnetic fields are calculated with LSD, Perdew--Burke--Ernzerhof (PBE) GGA, and TPSS meta-GGA and its predecessor, the nearly nonempirical Perdew--Kurth--Zupan--Blaha meta-GGA, using self-consistent LSD orbitals and densities. The results show that (i) For normal bulk densities, the surface correlation energy is the same in TPSS as in PBE, as it should be since TPSS strives to represent a self-correlation correction to PBE; (ii) Normal surface density profiles can be scaled uniformly to the low-density or strong-interaction limit, and TPSS provides an estimate for that limit that is consistent with (but probably more accurate than) other estimates; (iii) For both normal and low densities, TPSS provides the same description of surface magnetism as PBE, suggesting that these approximations may be generally equivalent for magnetism. The energies of jellium spheres with up to 106 electrons are calculated using density functionals and compared to those obtained with diffusion quantum Monte Carlo data, including our estimate for the fixed-node correction. Typically, while PBE energies are too low for spheres with more than about two electrons, LSD and TPSS are accurate there. We confirm that curvature energies are lower in PBE and TPSS than in LSD. Finally, we calculate the linear response of bulk jellium using these density functionals and find that not only LSD but also PBE GGA and TPSS meta-GGA yield a linear response in good agreement with that of the quantum Monte Carlo method, for wave vectors of the perturbing external potential up to twice the Fermi wave vector.
Among the computationally efficient semilocal density functionals for the exchange-correlation energy, meta-generalized-gradient approximations (meta-GGAs) are potentially the most accurate. Here, we assess the performance of three new meta-GGAs (revised Tao–Perdew–Staroverov–Scuseria or revTPSS, regularized revTPSS or regTPSS, and meta-GGA made simple or MGGA_MS), within and beyond their “comfort zones,” on Grimme’s big test set of main-group molecular energetics (thermochemistry, kinetics, and noncovalent interactions). We compare them against the standard Perdew–Burke–Ernzerhof (PBE) GGA, TPSS, and Minnesota M06L meta-GGAs, and Becke-3-Lee–Yang–Parr (B3LYP) hybrid of GGA with exact exchange. The overall performance of these three new meta-GGA functionals is similar. However, dramatic differences occur for different test sets. For example, M06L and MGGA_MS perform best for the test sets that contain noncovalent interactions. For the 14 Diels–Alder reaction energies in the “difficult” DARC subset, the mean absolute error ranges from 3 kcal mol–1 (MGGA_MS) to 15 kcal mol–1 (B3LYP), while for some other reaction subsets the order of accuracy is reversed; more generally, the tested new semilocal functionals outperform the standard B3LYP for ring reactions. Some overall improvement is found from long-range dispersion corrections for revTPSS and regTPSS but not for MGGA_MS. Formal and universality criteria for the functionals are also discussed.
The wave-vector analysis $\ensuremath{\gamma}(K)$, whose integral gives the nonuniform part of the exchange-correlation energy, is studied for a planar metallic surface. It is shown (1) that for "exchange only" $\ensuremath{\gamma}(K)\ensuremath{\rightarrow}\ensuremath{-}\frac{0.009662}{{{r}_{s}}^{3}}$ a.u. as $K\ensuremath{\rightarrow}0$ for the infinite barrier surface; (2) that this limit is universal---for all planar surfaces in an exchange-only approximation, $\ensuremath{\gamma}$ approaches this same universal number; (3) how the limit above reverts to that previously derived by the authors [$\ensuremath{\gamma}(K)\ensuremath{\propto}|K|$ as $K\ensuremath{\rightarrow}0$] when electron correlation is included. The method of wave-vector interpolation, which relies on the latter limit for the Coulomb-correlated surface, thus withstands recent questioning based on the "exchange-only" approximation. Finally, the probable small-$K$ behavior of $\ensuremath{\gamma}(K)$ for a system of neutral fermions is discussed.
Perfekt verdrahtet: Hoch kristalline kolloidale PbS- und CdS-Nanodrähte wurden über den SLS-Mechanismus (solution–liquid–solid) an Bi-Nanopartikeln gezüchtet. Die Verwendung von Einkomponentenvorstufen könnte ein allgemeiner Ansatz für die Synthese von kolloidalen Halbleiternanodrähten sein.
Triangulene and its analogue metal-free magnetic systems have garnered increasing attention since their discovery. Predicting the magnetic couplings and spin polarization energy is beyond the predictive power of todays density-functional theory (DFT) due to their intrinsic multi-reference character. Herein, we create a benchmark dataset of 25 magnetic systems with non-local spin densities, including the triangulene monomer, dimer, and their analogues. We calculate the magnetic coupling (J) and spin-polarization energy (Espin) of these systems using complete active space self-consistent field (CASSCF) and coupled cluster methods as high-quality reference values. This reference data is then used to benchmark 22 DFT functionals commonly used in material science. Our results show that, while some functionals consistently correctly predict the qualitative character of the ground state, achieving quantitative accuracy with small relative errors is currently not feasible. PBE0, M06-2X, and MN15 are predicting the correct electronic ground state for all systems investigated here, and also have the lowest mean absolute error for predicting both Espin (0.34 eV, 0.32 eV and 0.31 eV) and J (13.36 meV, 13.39 meV and 11.76 meV). They may therefore also serve as starting points for higher-level methods such as the GW or the random phase approximation. As other functionals fail for the prediction of the ground state, they cannot be recommended for metal-free magnetic systems.
Surface energies and work functions are calculated for the (110), (100), and (111) faces of the bcc metals Li, Ba, Na, K, Rb, and Cs and the fcc metals Al, Pb, Ca, Sr, and for the (0001) faces of the hcp metals Zn and Mg. In the Kohn-Sham energy functional employed, the crystal lattice of ions is represented by the Ashcroft pseudopotential, and nonlocal exchange-correlation energy corrections included via the wave-vector analysis method. The surface energies are determined by application of the Rayleigh-Ritz variational principle. The work functions are obtained for these energy-minimized densities by the variationally accurate, "displaced-profile change-in-self-consistent-field" (DP▵SCF) expression, which is tested for real metals here for the first time. The variational electronic densities employed are those generated by the linear-potential model, which permits the calculations to be primarily analytical. It is observed that the surface energies for each metal with the exception of Rb and Cs increase with decreasing packing density of the exposed crystal face, and that the Smoluchowski rule of decreasing work functions with decreasing packing density is obeyed by each metal except Al and Pb. An analysis of these numerical results indicates them to be superior to those of perturbation theory, and to be equivalent or generally superior to the results of other variational calculations for metals with ${r}_{s}\ensuremath{\gtrsim}3$. The Mahan-Schaich derivation for the work function of jellium metal is extended to include local ionic pseudopotentials, and the result shown to be equivalent to the DP▵SCF expression. A general expression for polycrystalline work functions is also derived, and an empirical formula given for the work function of alkali metals. It is further argued that it is meaningful, at least for the alkali metals, to compare the polycrystalline work function to the minimum work function, and this equivalence is demonstrated by comparison with experiment.
Density functional approximations for the exchange-correlation energy EDFAxc of an electronic system are often improved by admixing some exact exchange Ex: Exc≊EDFAxc+(1/n)(Ex−EDFAx). This procedure is justified when the error in EDFAxc arises from the λ=0 or exchange end of the coupling-constant integral ∫10 dλ EDFAxc,λ. We argue that the optimum integer n is approximately the lowest order of Görling–Levy perturbation theory which provides a realistic description of the coupling-constant dependence Exc,λ in the range 0≤λ≤1, whence n≊4 for atomization energies of typical molecules. We also propose a continuous generalization of n as an index of correlation strength, and a possible mixing of second-order perturbation theory with the generalized gradient approximation.
