Formation mechanism of chemically precompressed hydrogen clathrates in metal superhydrides
Abstract
Recently, the experimental discovery of high- superconductivity in compressed hydrides H3S and LaH10 at megabar pressures has triggered searches for various superconducting superhydrides. It was experimentally observed that thorium hydrides, ThH10 and ThH9, are stabilized at much lower pressures compared to LaH10. Based on first-principles density-functional theory calculations, we reveal that the isolated Th frameworks of ThH10 and ThH9 have relatively more excess electrons in interstitial regions than the La framework of LaH10. Such interstitial excess electrons easily participate in the formation of anionic H cage surrounding metal atom. The resulting Coulomb attraction between cationic Th atoms and anionic H cages is estimated to be stronger than the corresponding one of LaH10, thereby giving rise to larger chemical precompressions in ThH10 and ThH9. Such a formation mechanism of H clathrates can also be applied to another experimentally synthesized superhydride CeH9, confirming the experimental evidence that the chemical precompression in CeH9 is larger than that in LaH10. Our findings demonstrate that interstitial excess electrons in the isolated metal frameworks of high-pressure superhydrides play an important role in generating the chemical precompression of H clathrates.
I INTRODUCTION
In recent years, hydrides have attracted much attention theoretically and experimentally because of their promising possibility for the realization of room-temperature superconductivity (SC) [1, 2]. Motivated by the pioneering idea of Neil Ashcroft on high-temperature SC in metallic hydrogen [3] and the incessant theoretical predictions of high superconducting transition temperature in a number of hydrides [4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20], experiments have confirmed that sulfur hydride H3S and lanthanum hydride LaH10 exhibit around 203 K at 155 GPa [21] and 250260 K at 170 GPa [22, 23], respectively. Subsequently, such a conventional Bardeen-Cooper-Schrieffer type SC has also been experimentally observed in various compressed hydrides at high pressures [25, 26, 27, 24, 28, 30, 29]. For examples, ThH10 (ThH9) exhibits = 159161 (146) K between 170 and 175 GPa [24], while CeH9 exhibits a of 100 K at 130 GPa [27]. More recently, carbonaceous sulfur hydride was observed to exhibit a room-temperature SC with a of 288 K at 267 GPa [30]. Therefore, the experimental observations of high-temperature SC in either surfur-containing hydrides [30, 21] or superhydrides containing an abnormally large amount of hydrogen [22, 23, 25, 26, 27, 28, 24] has launched a new era of high- superconductors.
Compared to the existence of metallic hydrogen at high pressures over 400 GPa [31, 32], the syntheses of superhydrides with H-rich clathrate structures have been achieved at relatively much lower pressures, because H atoms can be “chemically precompressed” by chemical forces between metal atoms and H cages [33]. Using density-functional theory (DFT) calculations, the high- superconducting phases of various superhydrides have been predicted to be metastable at higher pressures than a critical pressure [34, 35, 36, 37]. It is noticeable that the magnitude of reflects the strength of chemical precompression in superhydrides. Experimentally, the value of ThH10 having a fcc-Th framework [see Fig. 1(a)] was measured to be 85 GPa [24], much lower than 170 of an isostructural superhydride LaH10 [22, 23]. Furthermore, ThH9 (CeH9) having a hcp metal framework [see Fig. 1(b)] was observed to exhibit a of 86 (80) GPa [24, 25, 26, 27]. Based on these existing experimental data [22, 23, 24], it is most likely that changes with respect to metal species: i.e., Group-4 metal hydrides ThH10, ThH9, and CeH9 with occupied -subshell electrons have lower values or larger chemical precompressions compared to a Group-3 metal hydride LaH10. Our recent DFT calculations for LaH10 [38] revealed that the isolated La framework without H atoms behaves as an electride at high pressures, where some electrons detached from La atoms are well localized in interstitial regions. These interstitial excess electrons are easily captured to H atoms, forming a H clathrate structure in LaH10. In the present study, such an electride feature in the La framework of LaH10 is compared with other metal frameworks of the above-mentioned superhydrides ThH10, ThH9, and CeH9. By the estimation of Coulomb attractions between metal atoms and H cages, we provide an explanation for the different chemical precompressions observed in such superhydrides [22, 23, 24, 25, 26, 27], as will be discussed below.
In this paper, using first-principles DFT calculations, we perform a comparative study of chemical precompressions in ThH10, ThH9, CeH9, and LaH10. We find that the isolated metal frameworks of ThH10, ThH9, and CeH9 possess more interstitial excess electrons than that of LaH10 at an equal pressure of 300 GPa. Such loosely bound electrons can be easily captured to form H clathrate structures with attractive Coulomb interactions between cationic metal atoms and anionic H cages. Using the calculated Bader charges [39] and positions of metal and H atoms in each superhydride, we estimate a chemical pressure acting on H cage around a metal atom. As a result, ThH10, ThH9, and CeH9 are found to have larger chemical precompressions than LaH10, consistent with the experimentally observed values in these superhydrides [22, 23, 24, 25, 26, 27]. It is thus demonstrated that Group-4 metal hydrides occupying electrons can be more chemically precompressed compared to Group-3 metal hydride, thereby contributing to lower . The present findings illuminate that interstitial excess electrons in the metal frameworks of superhydrides are of importance to generate the chemical precompression of H cages around metal atoms.

II Calculational Methods
Our first-principles DFT calculations were performed using the Vienna ab initio simulation package with the projector-augmented wave method [40, 41, 42]. Here, we treated Th 66567, Ce 55456, La 5556 and H 1 as valence electrons, including 66, 55, and 55 semicore electrons for Th, Ce, and La, respectively. For the exchange-correlation energy, we employed the generalized-gradient approximation functional of Perdew-Burke-Ernzerhof [43, 44]. A plane-wave basis was used with a kinetic energy cutoff of 500 eV for ThH10 and ThH9. The -space integration was done with 242424 and 181812 points for ThH10 and ThH9, respectively. All atoms were allowed to relax along the calculated forces until all the residual force components were less than 0.005 eV/Å. We calculated phonon frequencies with the 666 (553) points for ThH10 (ThH9) using the QUANTUM ESPRESSO package [45]. For CeH9 and LaH10, we chose the calculation parameters used in our previous works [38, 46].
III Results
We first optimize the structures of experimentally synthesized superhydrides ThH10, ThH9, CeH9, and LaH10 as a function of pressure using DFT calculations. These superhydrides have hydrogen sodalitelike clathrate structures with high-symmetry space groups: i.e., (No. 225) for ThH10 and LaH10, while P63/ (No. 194) for ThH9 and CeH9. As shown in Fig. 1(a), ThH10 (LaH10) is constituted by the fcc metal framework, where each Th (La) atom is surrounded by the H32 cage consisting of 32 H atoms. Meanwhile, ThH9 (CeH9) have the hcp metal framework with the H29 cage surrounding a Th (Ce) atom [see Fig. 1(b)]. Note that there are two (three) species of H atoms composing the H32 (H29) cages in ThH10 and LaH10 (ThH9 and CeH9). The optimized structures of these superhydrides show that the lattice constants decrease monotonously with increasing pressure (see Fig. S1 in the Supplemental Material). Accordingly, the averaged bond lengths between metal and H atoms decrease with increasing pressure [see Fig. 2(a)]. We find that for ThH10 having the H32 cage is longer than ThH9 having the H29 cage at a given pressure. Meanwhile, for CeH9 is shorter than that for ThH9, possibly because of the smaller size of Ce atom with the atomic number of = 58 compared to Th atom with = 90. Interestingly, despite the larger atomic number of Th than La ( = 57), the values for ThH10 and ThH9 are close to that for LaH10 at a given pressure, implying that the former superhydrides have larger chemical precompressions than the latter one. We note that the charges of cationic metal and anionic H atoms are also essential ingredients for determining chemical precompression, as discussed below. Figure 2(b) displays the averaged HH bond lengths for ThH10, ThH9, CeH9, and LaH10, which also decrease monotonously with increasing pressure. It is seen that the values for ThH10 and LaH10 are shorter than those for ThH9 and CeH9, indicating that the H32 cages composed of larger number of H atoms give rise to shorter compared to the H29 cages.

In order to examine the dynamical stability of ThH10, ThH9, CeH9, and LaH10, we calculate their phonon spectra as a function of pressure. As shown in Fig. 3(a), the phonon spectrum of ThH10, calculated at 130 GPa, exhibits the softening of low-energy phonon modes (marked by arrows) along the and lines. Such H-derived phonon modes finally have imaginary frequencies at 120 GPa [see Fig. 3(b)]. This indicates that the fcc-ThH10 phase becomes unstable as pressure decreases. Therefore, the phonon spectra as a function of pressure allow us to predict the values of about 130, 110, 100, and 220 GPa for ThH10, ThH9, CeH9 [46], and LaH10 [36], respectively (see Fig. S2 in the Supplemental Material). We find that ThH10 and ThH9 have much lower than LaH10, while their values are close to that of CeH9. The overall trend of these predicted values in four superhydrides are well consistent with the experimentally measured ones of about 85, 86, 80, and 170 GPa for ThH10 [24], ThH9 [24], CeH9 [25, 26, 27], and LaH10 [22, 23], respectively. It was previously pointed out that for LaH10, the anharmonic effects on phonons and the quantum ionic zero-point energy are of importance for a proper prediction of [47]. Therefore, the above overestimation of predicted values is likely due to the ignorance of anharmonic and quantum effects [48, 49, 50] in the present theory. Nevertheless, we can say that ThH10, ThH9, and CeH9 have significantly larger chemical precompressions compared to LaH10.

We next explore the electride-like characteristics of the isolated metal frameworks of ThH10, ThH9, CeH9, and LaH10. Here, the structure of each metal framework is taken from the optimized structure of the corresponding superhydride. The valence charge densities (without including of semicore electrons) of the metal frameworks of ThH10, ThH9, CeH9, and LaH10, calculated at an equal pressure of 300 GPa, are displayed in Figs. 4(a), 4(c), 4(e), and 4(g), respectively. It is seen that some electrons detached from metal atoms are localized in the interstitial regions around the and sites. These interstitial excess electrons of the so-called and anions are well confirmed by the corresponding electron localization function (ELF) [51]. Figures. 4(b), 4(d), 4(f), and 4(h) represent the calculated ELFs of the metal frameworks of ThH10, ThH9, CeH9, and LaH10, respectively. For the Th framework of ThH10, the number of electrons () within the muffin-tin sphere of the () anion is 0.281 (0.234), larger in magnitude than 0.204 (0.194) for the La framework of LaH10 [see Figs. 4(a) and 4(g)]. Similarly, as shown in Figs. 4(c) and 4(e), () in ThH9 is 0.124 (0.080), larger in magnitude than the corresponding value of 0.118 (0.072) in CeH9. Therefore, the metal framework of ThH10 (ThH9) exhibits a more electride feature than that of LaH10 (CeH9). It is noted that the magnitudes of and change as a function of pressure (see Figs. S3 and S4 in the Supplemental Material), showing that the electride-like characteristics of metal frameworks are enhanced with increasing pressure. Indeed, the localization of interstitial excess electrons also emerges in compressed alkali metals at high pressures [52, 53, 54], in order to reduce Coulomb repulsions arising from the overlap of atomic valence electrons. Such loosely bound anionic electrons residing in the metal frameworks of compressed superhydrides can be easily captured to H atoms, forming H cages with attractive Coulomb interactions between cationic metal and anionic H atoms. It is remarkable that the anionic electrons in H cages are mostly supplied because of the electride nature of metal frameworks at high pressures, rather than due to the different electronegativities of metal and H atoms [11].

Figures 5(a) and 5(b) show the calculated total charge densities of ThH10 and ThH9 at 300 GPa, respectively. It is seen that the H atoms in each H cage are bonded to each other with covalent bonds, where each HH bond has a saddle point of charge density at its midpoint, similar to the CC covalent bond in diamond [55]. For ThH10 (TH9), the charge densities at the midpoints of the HH bonds are 0.911 and 0.729 (0.992, 0.769, and 0.601) /Å3: see the arrows in Figs. 5(a) and 5(b). In order to confirm that the interstitial excess electrons of the Th framework of ThH10 (ThH9) are captured to form the H32 (H29) cages, we calculate the charge densities of the isolated H32 (H29) cages without Th atoms. Here, the structure of each isolated H cage is taken from the optimized structure of the corresponding superhydride. As shown in Fig. 5(c) [5(d)], we find that decreases as 0.742 and 0.621 (0.843, 0.580, and 0.479) /Å3, smaller than those in ThH10 (TH9). This indicates that the HH covalent bonds in ThH10 and TH9 are strengthened by capturing the interstitial excess electrons of isolated Th frameworks.

To provide an explanation for the variation of in ThH10, ThH9, CeH9, and LaH10, we estimate chemical precompression by calculating the attractive Coulomb forces between a metal atom and its surrounding H atoms [see the lower panel in Figs. 1(a) and 1(b)]. This simple estimation is based on the complete screening of the electric field arising from metal atoms within H cages. Using the Bader [39] analysis, we calculate the cationic charge inside the Bader basin [see Figs. 5(a) and 5(b)] of metal atom in each superhydride. For ThH10, ThH9, CeH9, and LaH10, we obtain values of 1.486, 1.464, 1.199, and 1.036, respectively (see Fig. 6). Assuming that is the point charge at the position of corresponding metal atom and the anionic charge () of H atoms is uniformly distributed on the spherical shell with a radius of , we evaluate the magnitudes of Coulomb forces acting on the H atoms composing the H32 or H29 cage, and divide it by the surface area of the spherical shell. Figure 6 shows such estimated chemical pressures of ThH10, ThH9, CeH9, and LaH10 at 300 GPa, with ratios relative to the value of LaH10. We find that the chemical pressures of ThH10 and ThH9 (CeH9) are about two (one and half) times higher than that of LaH10, indicating that the former Group-4 metal hydrides have larger chemical precompressions to attain lower values than the latter Group-3 metal hydride. Considering that the values of ThH10 and ThH9 are close to that of LaH10 [see Fig. 2(a)], the higher chemical pressures in Th superhydrides are likely attributed to more cationic and anionic charges compared to LaH10. As shown in Fig. 6, the chemical pressures of four superhydrides are well consistent with their variations of . It is noted that the estimated chemical pressure of CeH9 is lower than that of isostructural ThH9 at 300 GPa (see Fig. 6), while the predicted value of = 100 GPa for the former is lower than that (110 GPa) for the latter. This inconsistency of chemical precompression and between CeH9 and ThH9 may be due to the delocalized nature of Ce 4 electrons [46], which could lower via a more hybridization with the H 1 state. Nevertheless, despite their crude simulations, the estimated relative chemical pressures of ThH10, ThH9, CeH9, and LaH10 are in reasonable agreement with the variation of experimentally measured values [22, 23, 24, 25, 26, 27].

IV Summary
Using first-principles DFT calculations, we have conducted a comparative study of chemical precompressions in experimentally synthesized superhydrides including ThH10, ThH9, CeH9, and LaH10. We found that these superhydrides form H clathrates by capturing excess electrons in interstitial regions of their isolated fcc- and hcp-metal frameworks. By taking into account the attractive Coulomb interactions between cationic metal atom and its surrounding H atoms, we estimated chemical precompressions in ThH10, ThH9, CeH9, and LaH10. It was found that ThH10, ThH9, and CeH9 have larger chemical precompressions than LaH10, consistent with the variation of experimentally measured values [22, 23, 24, 25, 26, 27]. Our findings not only demonstrated that interstitial excess electrons in the metal frameworks of superhydrides play an importnt role in generating the chemical precompression of H cages around metal atoms, but also have important implications for the exploration of new superhydrides which can be synthesized at moderate pressures below 100 GPa.
V Supplementary Material
See Supplemental Material for the lattice constants of ThH10, ThH9, CeH9, and LaH10, the phonon spectrum of ThH9, and the valence charge densities of the Th frameworks of ThH10 and ThH9.
VI Author’s contributions
S. Y., C. W., and S. L. contributed equally to this work.
VII Acknowledgement
This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korean Government (Grants No. 2019R1A2C1002975, No. 2016K1A4A3914691, and No. 2015M3D1A1070609). The calculations were performed by the KISTI Supercomputing Center through the Strategic Support Program (Program No. KSC-2020-CRE-0163) for the supercomputing application research.
VIII DATA AVAILABILITY
The data that support the findings of this study are available from the corresponding author upon reasonable request.
∗ Corresponding author: [email protected]
References
- [1] T. Bi, N. Zarifi, T. Terpstra, and E. Zurek, Reference Module in Chemistry, Molecular Science and Chemical Engineering (Elsevier, New York, 2019).
- [2] J. A. Flores-Livas, L. Boeri, A. Sanna, G. Profeta, R. Arita, and M. I. Eremets, Phys. Rep. 856, 1 (2020) and references therein.
- [3] N. W. Ashcroft, Phys. Rev. Lett. 21, 1748 (1968).
- [4] E. Zurek, R. Hoffmann, N. W. Ashcroft, A. R. Oganov, and A. O. Lyakhov, Proc. Natl. Acad. Sci. USA 106, 42 (2009).
- [5] Y. Xie, Q. Li, A. R. Oganov, and H. Wang, Acta Cryst. C 70, 104 (2014).
- [6] J. Hooper and E. Zurek, J. Phys. Chem. C 116, 13322 (2012).
- [7] H. Wang, J. S. Tse, K. Tanaka, T. Litaka, and Y. Ma, Proc. Natl. Acad. Sci. USA 109, 6463 (2012).
- [8] D. Duan, Y. Liu, F. Tian, D. Li, X. Huang, Z. Zhao, u. Yu, B. Liu, W. Tian, and T. Cui, Sci. Rep. 4, 6968 (2014).
- [9] Y. Li, J. Hao, H. Liu, Y. Li, and Y. Ma, J. Chem. Phys. 140, 174712 (2014).
- [10] X. Feng, J. Zhang, G. Gao, H. Liu, and H. Wang, RSC Adv. 5, 59292 (2015).
- [11] F. Peng, Y. Sun, C. J. Pickard, R. J. Needs, Q. Wu, and Y. Ma, Phys. Rev. Lett. 119, 107001 (2017).
- [12] H. Liu, I. I. Naumov, R. Hoffmann, N. W. Ashcroft, and R. J. Hemley, Proc. Natl. Acad. Sci. USA 114, 6990 (2017).
- [13] Y. Sun, J. Lv, Y. Xie, H. Liu, and Y. Ma, Phys. Rev. Lett. 123, 097001 (2019).
- [14] H. Xie, Y. Yao, X. Feng, D. Duan, H. Song, Z. Zhang, S. Jiang, S. A. T. Redfern, V. Z. Kresin, C. J. Pickard, and T. Cui, Phys. Rev. Lett. 125, 217001 (2020).
- [15] L. Liu, C. Wang, S. Yi, K. W. Kim, J. Kim, and J.-H. Cho, Phys. Rev. B 99, 140501(R) (2019).
- [16] C. Wang, S. Yi and J.-H. Cho, Phys. Rev. B 101, 104506 (2020).
- [17] A. P. Durajski, R. Szczȩśniak, Y. Li, C. Wang, and J.-H. Cho, Phys. Rev. B 101, 214501 (2020).
- [18] C. Heil, S. diCataldo, G. B. Bachelet, and L. Boeri, Phys. Rev. B 99, 220502(R) (2019).
- [19] Y. Quan, S. S. Ghosh, and W. E. Pickett, Phys. Rev. B 100, 184505 (2019).
- [20] D. A. Papaconstantopoulos, M. J. Mehl, and P.-H. Chang, Phys. Rev. B 101, 060506(R) (2020).
- [21] A. P. Drozdov, M. I. Eremets, I. A. Troyan, V. Ksenofontov, and S. I. Shylin, Nature (London) 525, 73 (2015).
- [22] M. Somayazulu, M. Ahart, A. K. Mishra, Z. M. Geballe, M. Baldini, Y. Meng, V. V. Struzhkin, and R. J. Hemley, Phys. Rev. Lett. 122, 027001 (2019).
- [23] A. P. Drozdov, P. P. Kong, V. S. Minkov, S. P. Besedin, M. A. Kuzovnikov, S. Mozaffari, L. Balicas, F. F. Balakirev, D. E. Graf, V. B. Prakapenka, E. Greenberg, D. A. Knyazev, M. Tkacz, and M. I. Eremets, Nature (London) 569, 528 (2019).
- [24] D. V. Semenok, A. G. Kvashnin, A. G. Ivanova, V. Svitlyk, V. Y. Fominski, A. V. Sadakov, O. A. Sobolevskiy, V. M. Pudalov, I. A. Troyan, and A. R. Oganov, Mater. Today 33, 36 (2020).
- [25] X. Li, X. Huang, D. Duan, C. J. Pickard, D. Zhou, H. Xie, Q. Zhuang, Y. Huang, Q. Zhou, B. Liu, and T. Cui, Nat. Commun. 10, 3461 (2019).
- [26] N. P. Salke, M. M. Davari Esfahani, Y. Zhang, I. A. Kruglov, J. Zhou, Y. Wang, E. Greenberg, V. B. Prakapenka, J. Liu, A. R. Oganov, and J.-F. Lin, Nat. Commun. 10, 4453 (2019).
- [27] W. Chen, D. V. Semenok, X. Huang, H. Shu, X. Li, D. Duan, T. Cui, and A. R. Oganov, arXiv:2101.01315 (2021).
- [28] P. P. Kong, V. S. Minkov, M. A. Kuzovnikov, S. P. Besedin, A. P. Drozdov, S. Mozaffari, L. Balicas, F. F. Balakirev, V. B. Prakapenka, E. Greenberg, D. A. Knyazev, and M. I. Eremets, arXiv:1909.10482 (2019).
- [29] D. Zhou, D. V. Semenok, D. Duan, H. Xie, W. Chen, X. Huang, X. Li, B. Liu, A. R. Oganov, and T. Cui, Sci. Adv. 6, eaax6849 (2020).
- [30] E. Snider, N. Dasenbrock-Gammon, R. McBride, M. Debessai, H. Vindana, K. Vencatasamy, K. V. Lawler, A. Salamat, and R. P. Dias, Nature (London) 586, 373 (2020).
- [31] R. P. Dias and I. F. Silvera, Science 355, 715 (2017).
- [32] P. Loubeyre, F. Occelli, and P. Dumas, Nature (London) 577, 631 (2020).
- [33] N. W. Ashcroft, Phys. Rev. Lett 92, 187002 2004.
- [34] Z. M. Geballe, H. Liu, A. K.Mishra, M. Ahart, M. Somayazulu, Y. Meng, M. Baldini, and R. J. Hemley, Angew. Chem., Int. Ed. 57, 688 (2018).
- [35] H. Liu, I. I. Naumov, Z. M. Geballe, M. Somayazulu, J. S. Tse, and R. J. Hemley, Phys. Rev. B 98, 100102(R) (2018).
- [36] C. Wang, S. Yi, and J.-H. Cho, Phys. Rev. B 100, 060502(R) (2019).
- [37] A. M. Shipley, M. J. Hutcheon, M. S. Johnson, R. J. Needs, and C. J. Pickard, Phys. Rev. B 101, 224511 (2020).
- [38] S. Yi, C. Wang, H. Jeon, and J.-H. Cho, Phys. Rev. Materials 5, 024801 (2021).
- [39] R. F. W. Bader, Acc. Chem. Res. 18, 9 (1985).
- [40] G. Kresse and J. Hafner, Phys. Rev. B 48, 13115 (1993).
- [41] G. Kresse and J. Furthmüller, Comput. Mater. Sci. 6, 15 (1996).
- [42] P. E. Blöchl, Phys. Rev. B 50, 17953 (1994).
- [43] J. P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 77, 3865 (1996); 78, 1396(E) (1997).
- [44] L. Kývala and D. Legut, Phys. Rev. B 101, 075117 (2020).
- [45] P. Giannozzi, S. Baroni, N. Bonini, M. Calandra, R. Car, C. Cavazzoni, D. Ceresoli, G. L. Chiarotti, M. Cococcioni, I. Dabo, et al. J. Phys.: Condens. Matter 21, 395502 (2009).
- [46] H. Jeon, C. Wang, S. Yi, and J.-H. Cho, Sci. Rep. 10, 16878 (2020).
- [47] I. Errea, F. Belli, L. Monacelli, A. Sanna, T. Koretsune, T. Tadano, R. Bianco, M. Calandra, R. Arita, F. Mauri, and J. A. Flores-Livas, Nature (London) 578, 66 (2020).
- [48] I. Errea, M. Calandra, C. J. Pickard, J. Nelson, R. J. Needs, Y. Li, H. Liu, Y. Zhang, Y. Ma, and F. Mauri, Phys. Rev. Lett. 114, 157004 (2015).
- [49] I. Errea, M. Calandra, C. J. Pickard, J. R. Nelson, R. J. Needs, Y. Li, H. Liu, Y. Zhang, Y. Ma, and F. Mauri, Nature (London) 532, 81 (2016).
- [50] A. P. Durajski, Sci. Rep. 6, 38570 (2016).
- [51] B. Silvi and A. Savin, Nature (London) 371, 683 (1994).
- [52] M.-S. Miao and R. Hoffmann, J. Am. Chem. Soc. 137, 3631 (2015).
- [53] J. Wang, Q. Zhu, Z. Wang, and H. Hosono, Phys. Rev. B 99, 064104 (2019) and referenes therein.
- [54] Z. Zhao, S. Zhang, T. Yu, H. Xu, A. Bergara, and G. Yang, Phys. Rev. Lett. 122, 097002 (2019).
- [55] E. Kaxiras, Atomic and Electronic Structure of Solids. (Cambridge University Press, New York, 2003) p. 152.