This paper was converted on www.awesomepapers.org from LaTeX by an anonymous user.
Want to know more? Visit the Converter page.

Formation mechanism of chemically precompressed hydrogen clathrates in metal superhydrides

Shichang Yao, Chongze Wang, Shuyuan Liu, Hyunsoo Jeon, and Jun-Hyung Cho Department of Physics, Research Institute for Natural Science, and Institute for High Pressure at Hanyang University, Hanyang University, 222 Wangsimni-ro, Seongdong-Ku, Seoul 04763, Republic of Korea
Abstract

Recently, the experimental discovery of high-TcT_{c} 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 TcT_{c} 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 TcT_{\rm c} around 203 K at {\sim}155 GPa [21] and 250-260 K at {\sim}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 TcT_{c} = 159-161 (146) K between 170 and 175 GPa [24], while CeH9 exhibits a TcT_{c} of {\sim}100 K at 130 GPa [27]. More recently, carbonaceous sulfur hydride was observed to exhibit a room-temperature SC with a TcT_{\rm c} of 288 K at {\sim}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-TcT_{\rm c} superconductors.

Compared to the existence of metallic hydrogen at high pressures over {\sim}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-TcT_{c} superconducting phases of various superhydrides have been predicted to be metastable at higher pressures than a critical pressure PcP_{c} [34, 35, 36, 37]. It is noticeable that the magnitude of PcP_{c} reflects the strength of chemical precompression in superhydrides. Experimentally, the PcP_{c} value of ThH10 having a fcc-Th framework [see Fig. 1(a)] was measured to be {\sim}85 GPa [24], much lower than PcP_{c} {\approx} 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 PcP_{c} of {\sim}86 (80) GPa [24, 25, 26, 27]. Based on these existing experimental data [22, 23, 24], it is most likely that PcP_{c} changes with respect to metal species: i.e., Group-4 metal hydrides ThH10, ThH9, and CeH9 with occupied ff-subshell electrons have lower PcP_{c} 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 PcP_{c} values in these superhydrides [22, 23, 24, 25, 26, 27]. It is thus demonstrated that Group-4 metal hydrides occupying ff electrons can be more chemically precompressed compared to Group-3 metal hydride, thereby contributing to lower PcP_{c}. 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.

Refer to caption
Figure 1: Optimized structures of (a) ThH10 and (b) ThH9 at 300 GPa. ThH10 (ThH9) has the fcc (hcp) Th framework with the H32 (H29) cage surrounding each Th atom. There are two (three) different types of H atoms: i.e., H1 and H2 for ThH10 (H1, H2, and H3 for ThH9). The (110) planes are drawn in the fcc and hcp lattices.

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 6s2s^{2}6p6p^{6}5f1f^{1}6d1d^{1}7s2s^{2}, Ce 5s2s^{2}5p6p^{6}4f1f^{1}5d1d^{1}6s2s^{2}, La 5s2s^{2}5p6p^{6}5d1d^{1}6s2s^{2} and H 1s1s^{1} as valence electrons, including 6s2s^{2}6p6p^{6}, 5s2s^{2}5p6p^{6}, and 5s2s^{2}5p6p^{6} 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 𝐤{\bf k}-space integration was done with 24×{\times}24×{\times}24 and 18×{\times}18×{\times}12 kk 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 6×{\times}6×{\times}6 (5×{\times}5×{\times}3) qq 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., FmFm3¯m\overline{3}m (No. 225) for ThH10 and LaH10, while P63/mmcmmc (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 dMHd_{\rm M-H} between metal and H atoms decrease with increasing pressure [see Fig. 2(a)]. We find that dMHd_{\rm M-H} for ThH10 having the H32 cage is longer than ThH9 having the H29 cage at a given pressure. Meanwhile, dMHd_{\rm M-H} for CeH9 is shorter than that for ThH9, possibly because of the smaller size of Ce atom with the atomic number of ZZ = 58 compared to Th atom with ZZ = 90. Interestingly, despite the larger atomic number of Th than La (ZZ = 57), the dMHd_{\rm M-H} 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 H-H bond lengths dHHd_{\rm H-H} for ThH10, ThH9, CeH9, and LaH10, which also decrease monotonously with increasing pressure. It is seen that the dHHd_{\rm H-H} 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 dHHd_{\rm H-H} compared to the H29 cages.

Refer to caption
Figure 2: Calculated averaged bond lengths (a) dMHd_{\rm M-H} between metal and H atoms and (b) dHHd_{\rm H-H} between H atoms for ThH10, ThH9, CeH9, and LaH10 as a function of pressure.

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 ΓL\Gamma-L and ΓK\Gamma-K 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 PcP_{c} 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 PcP_{c} than LaH10, while their PcP_{c} values are close to that of CeH9. The overall trend of these predicted PcP_{c} 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 PcP_{c} [47]. Therefore, the above overestimation of predicted PcP_{c} 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.

Refer to caption
Figure 3: Calculated phonon spectra of ThH10 at (a) 130 and (b) 120 GPa. The arrows in (a) indicate the softened phonon modes. The imaginary phonon frequencies appear at 120Gpa along the ΓL\Gamma-L line.

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 ρM{\rho_{\rm M}} (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 A1A_{1} and A2A_{2} sites. These interstitial excess electrons of the so-called A1A_{1} and A2A_{2} 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 QA1Q_{A_{1}} (QA2Q_{A_{2}}) within the muffin-tin sphere of the A1A_{1} (A2A_{2}) anion is -0.281 (-0.234)ee, larger in magnitude than -0.204 (-0.194)ee for the La framework of LaH10 [see Figs. 4(a) and 4(g)]. Similarly, as shown in Figs. 4(c) and 4(e), QA1Q_{A_{1}} (QA2Q_{A_{2}}) in ThH9 is -0.124 (-0.080)ee, larger in magnitude than the corresponding value of -0.118 (-0.072)ee 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 QA1Q_{A_{1}} and QA2Q_{A_{2}} 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].

Refer to caption
Figure 4: Calculated valence charge density plot of the metal frameworks of (a) ThH10, (c) ThH9, (e) CeH9, and (g) LaH10 at 300 GPa, together with the ELFs of the metal frameworks of (b) ThH10, (d) ThH9, (f) CeH9, and (h) LaH10. The charge densities in (a), (c), (e), and (g) are drawn on the (110) plane with a contour spacing of 0.005 ee3. The ELF in (b), (d), (f), and (h) are drawn with a contour spacing of 0.05. A1 and A2 indicate the two anions in interstitial regions, and the dashed circles represent the muffin-tin spheres around A1 and A2 with the radii of 0.75 (0.60) and 0.75 (0.60) Å in ThH10 and LaH10 (ThH9 and CeH9), respectively.

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 H-H bond has a saddle point of charge density at its midpoint, similar to the C-C covalent bond in diamond [55]. For ThH10 (TH9), the charge densities ρHH{\rho}_{\rm H-H} at the midpoints of the H-H bonds are 0.911 and 0.729 (0.992, 0.769, and 0.601) ee3: 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 ρHH{\rho}_{\rm H-H} decreases as 0.742 and 0.621 (0.843, 0.580, and 0.479) ee3, smaller than those in ThH10 (TH9). This indicates that the H-H covalent bonds in ThH10 and TH9 are strengthened by capturing the interstitial excess electrons of isolated Th frameworks.

Refer to caption
Figure 5: Calculated total charge densities of (a) ThH10 and (b) ThH9 at 300 GPa, with a contour spacing of 0.1 ee3. The Bader basins of Th atoms are also drawn in (a) and (b). The calculated charge densities of the isolated H32 and H29 cages of ThH10 and ThH9 without Th atoms are displayed in (c) and (d), respectively. The numbers represent the values of ρHH{\rho}_{\rm H-H} at the midpoints (marked by the arrows) of the H-H bonds.

To provide an explanation for the variation of PcP_{c} 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 QMQ_{\rm M} 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 QMQ_{\rm M} values of 1.486, 1.464, 1.199, and 1.036ee, respectively (see Fig. 6). Assuming that QMQ_{\rm M} is the point charge at the position of corresponding metal atom and the anionic charge (-QMQ_{\rm M}) of H atoms is uniformly distributed on the spherical shell with a radius of dMHd_{\rm M-H}, 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 PcP_{c} values than the latter Group-3 metal hydride. Considering that the dMHd_{\rm M-H} 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 QMQ_{\rm M}. 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 PcP_{c} = 100 GPa for the former is lower than that (110 GPa) for the latter. This inconsistency of chemical precompression and PcP_{c} between CeH9 and ThH9 may be due to the delocalized nature of Ce 4ff electrons [46], which could lower PcP_{c} via a more hybridization with the H 1ss 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 PcP_{c} values [22, 23, 24, 25, 26, 27].

Refer to caption
Figure 6: Calculated chemical pressures of ThH10, ThH9, CeH9, and LaH10 at 300 GPa. The cationic charges QMQ_{\rm M} of metal atoms obtained from Bader charge analysis are also given.

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 PcP_{c} 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 {\sim}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.