Article | Open

Phase Diagram and High-Temperature Superconductivity of Compressed Selenium Hydrides

  • Scientific Reports 5, Article number: 15433 (2015)
  • doi:10.1038/srep15433
  • Download Citation
Published online:


Recent discovery of high-temperature superconductivity (Tc = 190 K) in sulfur hydrides at megabar pressures breaks the traditional belief on the Tc limit of 40 K for conventional superconductors, and opens up the doors in searching new high-temperature superconductors in compounds made up of light elements. Selenium is a sister and isoelectronic element of sulfur, with a larger atomic core and a weaker electronegativity. Whether selenium hydrides share similar high-temperature superconductivity remains elusive, but it is a subject of considerable interest. First-principles swarm structure predictions are performed in an effort to seek for energetically stable and metallic selenium hydrides at high pressures. We find the phase diagram of selenium hydrides is rather different from its sulfur analogy, which is indicated by the emergence of new phases and the change of relative stabilities. Three stable and metallic species with stoichiometries of HSe2, HSe and H3Se are identified above ~120 GPa and they all exhibit superconductive behaviors, of which the hydrogen-rich HSe and H3Se phases show high Tc in the range of 40–110 K. Our simulations established the high-temperature superconductive nature of selenium hydrides and provided useful route for experimental verification.


Since Onnes discovered superconductivity of mercury in 19111, intensive research activities were stimulated to search for new superconductors with high critical temperatures (Tc). With this thrust, unconventional superconductors such as cuprates2,3,4 and Fe-pnictides5,6,7, in which superconducting mechanism cannot be described by Bardeen-Cooper-Schrieffer (BCS) theory8 have attracted major attention since they often exhibit high Tc values, reaching as high as 164 K (at high pressures)9. In spite of extensive research for several decades, the superconducting mechanism in these unconventional superconductors is still controversial, preventing them from being an optimal platform for designing the superconductors with the higher Tc.

On the side of conventional superconductors, the situation is rather disappointing since they all have low Tc values. The best-known conventional superconductor of MgB2 has the Tc of 39 K10. As such, there is a traditional belief on the Tc limit of 40 K for conventional superconductors in the field.

BCS theory gives a clear count on superconducting mechanism of conventional superconductors, making the design of high-Tc superconductors possible. According to BCS theory, a necessary condition for a high-Tc superconductor is that the metallic compounds shall have large electron density of states at the Fermi level, high phonon frequencies and strong electron-phonon coupling. Hydrogen is the lightest element, and therefore naturally gives rise to high phonon frequencies, and also owes the unique strong bare electron-ion interaction. Ashcroft firstly proposed that solid hydrogen once being metallic under pressure has the potential to be a high-temperature superconductor11. Later on, the idea on metallic superconducting hydrogen was extended into hydrogen-rich compounds12, where metallization pressure can be significantly lowered than that in pure hydrogen. A number of hydrogen-rich compounds were subsequently predicted to be good superconductors13 with estimated Tc reaching remarkably high values (e.g. 64 K for GeH4 at 220 GPa14, 107 K for SiH4(H2)2 at 250 GPa15, 235 K for CaH6 at 150 GPa16, and 204 K for (H2S)2H2 at 200 GPa17). Experimental syntheses of these potential high-temperature superconductors are excitingly ongoing, but challenging.

Sulfur dihydride (H2S) has not been considered as the candidate for superconducting hydrides since it was proposed to dissociation into elemental sulfur and hydrogen before the metallization18,19. Only recently, first-principles swarm structure searches on high-pressure structures of H2S was conducted and H2S was excluded from the elemental dissociation, making the prediction of its superconductivity with a high Tc ~80 K at 160 GPa possible20. Shortly after this report, breakthrough electrical measurement observed high-temperature superconductivity in compressed H2S with an unprecedentedly high Tc up to 190 K at megabar pressures21. The decrease of Tc with magnetic field, and the strong isotope shift of Tc suggest that H2S is a conventional superconductor. There exist two observed superconducting states: the sample prepared at low temperature of 100–150 K has a maximal Tc of 150 K at 200 GPa, while the 190 K superconductivity comes from the sample prepared at high temperature of 220–300 K, which likely associates with the dissociation of H2S into H3S21,22,23. This discovery has stimulated significant interest in studying the underlying superconducting mechanism23,24,25,26,27 and searching for new high-Tc superconductors in other dense hydride systems.

Selenium is a sister and isoelectronic element of sulfur with a larger atomic radius and a weaker electronegativity. By witness of high-Tc superconductivity of sulfur hydrides (i.e. H-S system), a natural and immediate thought is to examine whether selenium hydrides (H-Se system) are also high-temperature superconductors at high pressures. To our best knowledge, there is less report on solid phases of H-Se system, except for the existence of three temperature-dependent H2Se phases at ambient pressure28.

We herein explored the hitherto unknown high-pressure phase diagram of the H-Se system via the first-principles swarm-intelligence based structure search. In the moderate pressure region, different from the H-S system all the H-Se compounds are energetically unstable against the elemental decomposition including the normally expectant H2Se stoichiometry. At megabar pressures above 120 GPa, three stable species with stoichiometries of HSe2, HSe and H3Se were identified. Among them HSe with the high-symmetry P4/nmm structure is the most stable phase and H3Se becomes marginally stable at high pressures. They are all metallic and exhibit superconductive behaviors. The latter two H-rich phases, benefiting from fairly strong electron-phonon coupling, are predicted to owe high Tc values up to 110 K. Our simulations provide a useful roadmap for discovering high-temperature superconductors in selenium hydrides.


The variable-composition structure searches are performed at a variety of H-Se stoichiometries containing up to 4 formula units per simulation cell at 0, 50, 100, 200 and 300 GPa. At 0 GPa, we find only one stable stoichiometry with respect to the elemental decomposition, H2Se (see Supplemental Fig. S1a), which is consistent with available experimental reports28. It is in the P3121 symmetry, where the arrangement of Se atoms is nearly the same as that of the Se-I phase29 and H atoms are accommodated on the line of two adjacent Se to form covalent H-Se bonds and hydrogen bonds simultaneously (Fig. S1c). However, with increasing pressure this phase becomes dramatically unstable (Fig. S1b). The structure search results are summarized in the convex hulls constructed with solid H2 and Se as the binary variables in Fig. 1 (and Fig. S1a). In the low-pressure region up to 100 GPa, all the stoichiometries are energetically unstable against the elemental decomposition. This is in sharp contrast to the H-S system, where all the investigated stoichiometries are stable with respect to the elemental decomposition22. With increasing pressure to 200 GPa, while most of phases still lie above the elemental decomposition line, three stoichiometries (HSe2, HSe and H3Se) show the tendency of being stabilized and move downward below the line. Eventually HSe2 and H3Se become stable stoichiometries on the hull against any way of decomposition. At 300 GPa, the HSe2, HSe and H3Se stoichiometries are all clearly located on the hull, and HSe emerges as the most stable phase over other species. The energetic instability of these phases in the low-pressure region and their being restabilized at high pressures may be attributed to the weaker covalent bond of H-Se than that of H-S (resulted from the larger size of Se), which is cumulatively strengthened by volume contraction under compression.

Figure 1: Calculated formation enthalpies (H in eV/atom) of various selenium hydrides with respect to the elemental decomposition into solidified H2 and Se at 100 (violet), 200 (red) and 300 (blue) GPa, respectively.
Figure 1

At each stoichiometry, only H of the lowest-energy structure is shown. The phase IV (C2/m) and phase VI (Im-3m) of Se30, the P63m and C2/c structure of solid H2 in respective stable pressure regions are chosen for calculating H. The inset plot shows the pressure range in which each stable stoichiometry is stabilized.

The Se-rich HSe2 stoichiometry is the most stable phase at 200 GPa. Its lowest-energy structure (Fig. 2a) has a C2/m symmetry, in which the sublattice of Se atoms is isostructural to the Se-IV phase30 and H atoms passivate alternatively from both sides of the infinite zigzag Se chains. It is stabilized above 124 GPa (as in the inset of Fig. 1). The HSe stoichiometry is stabilized above 249 GPa as a highly symmetric PbO-type structure (space group P4/nmm, Fig. 2b). This compound, which is isostructural to superconductive Fe-based chalchogenides31, consists of the stack of two-dimensional layered edge-sharing SeH4-tetrahedra networks. Within the layer both Se and H are four-fold coordinated. It is worth mentioning that in addition to the P4/nmm HSe phase, our structure search finds an energetically competitive P21/c structure (though metastable, Fig. 2d). It consists of layered three-fold coordinated Se/H networks. As to the H3Se stoichiometry (stable above 166 GPa), the lowest-energy structure has a high symmetry of Im-3 m, isostructural to that of H3S17, where Se atoms occupy a body-centered cubic sublattice, each Se being six-fold coordinated by H. This structure is transformed from a molecular R3m phase as the result of pressure-induced hydrogen-bond symmetrization.

Figure 2
Figure 2

The energetically stable H-Se compounds identified by the structure search: (a) HSe2 in the C2/m structure, (b) HSe in the P4/nmm structure and (c) H3Se in the Im-3m structure. For the HSe stoichiometry, the metastable P21/c structure with competitive enthalpy is shown in (d). See Supplementary Table S1 for their detailed structural information and Fig. S2 for more metastable structures and Figs S5–S7 for specific enthalpy-pressure relationship of each stoichiometry.

We then studied the electronic, phonon and electron-phonon coupling (EPC) properties for the lowest-energy structures of three stable stoichiometries. The band structures of them all exhibit metallic features in their stable pressure regions (see Fig. S3). Phonon calculations indicate their lattice dynamical stabilities evidenced by the absence of any imaginary phonon mode in the whole Brillouin zone.

For the most H-rich H3Se stoichiometry, most of conducting states across the Fermi level (Ef) have significant contribution from the H-s orbital, hybridizing strongly with the Se-p states, and there exists the combination of flat bands and steep bands32 in proximity to the Ef (Fig. 3a). These features, resembling those of isostructural H3S17, are potentially favorable for strong EPC. By consideration of the heavier atomic weight of Se than that of S, one expects H3Se will have the generally lower phonon frequencies than those of H3S at the same pressure. This is indeed the case for the low-frequency Se-derived vibrations (below 15 THz) and mid-lying H-derived wagging and bending modes (between 15 and 45 THz) as shown in Fig. 3b. However, for the high-lying H-stretching vibrations (above 50 THz), one observes an opposite behavior, i.e. they show the higher frequencies and separate clearly from the mid-lying regime, unlike the situation of H3S in which the H-stretching vibrations mix together with the mid-lying phonons. This may be attributed to the stronger H-Se covalent bond as the result of the larger chemical precompression effect12 induced by Se with the larger atomic radius. The calculated phonon linewidths (Fig. 3b) and EPC spectral function (Fig. 3c) indicate a similar mechanism of EPC to that of H3S, where the high-frequency H-stretching modes give the notable contribution (31%) to the integral EPC parameter λ. This mechanism is different from the cases of superconducting CaH616 and SnH433 containing quasi-molecular H-units, where the mid-lying H-derived vibrations contribute most significantly to the EPC. For both of compounds, the EPC shows strong anisotropy along different phonon momentum vectors. Because of the hardening of H-stretching phonons, H3Se shows a slightly reduced λ (1.04 compared with 1.33 of H3S), but still falling in the range of fairly strong EPC. Note that owing to the choice of relatively large broadening parameters in our calculations (see Supplementary Methods), our calculated λ of H3S is a rather conservative evaluation, smaller than the previously reported values, ~2.022,27,34,35. It is thus highly possible that the calculated λ of H3Se here is underestimated in a similar way. Different from the linearly decreased λ with pressure in H3S, the λ of H3Se shows negligible pressure dependence (Fig. 3d). By substituting λ into the Allen-Dynes modified McMillan equation36, we get a weakly pressure-dependent superconducting Tc around 110 K for H3Se, mildly lower than the values of 160–170 K for H3S (see also Table S3).

Figure 3
Figure 3

(a) Electronic band structure of H3Se in the Im-3 m structure at 250 GPa. The projection onto the H-s orbital is depicted by the sizes of green circles. (b) Comparison of phonon spectra of H3Se (red) and H3S (gray) at 250 GPa. The phonon linewidth γq,j(ω) of each mode (q, j) caused by EPC is illustrated by the size of circle. (c) Eliashberg EPC spectral function α2F(ω) and EPC integration λ(ω) of H3Se and H3S. (d) Pressure dependence of the EPC parameter λ (right axis) and Tc (left axis) for H3Se and H3S. The typical value of the Coulomb pseudopotential μ* = 0.1 is used for calculating Tc.

Turning to the HSe stoichiometry, there is moderate contribution of the H-s state to the conducting states, i.e. only several bands across the Ef (e.g. along the A-M, M-Γ and X-Γ lines) are derived from the H-s orbital (Fig. 4a). The Fermi surface (Fig. 4b and Fig. S8) consists of the relatively small pockets around the M and Z points, and two expanding sheets at the larger wave-vectors. No notable Fermi nesting can be observed. Analysis of bonding feature via the electron localization function (ELF, Fig. 4c) indicates a much weaker H-Se covalent bond (with the maximum ELF magnitude of ~0.5) by comparison with that of H3Se (with the maximum ELF of ~0.9). This is further supported by the elongated bond length of HSe (1.69 Å) than that of H3Se (1.51 Å) at the same pressure (300 GPa). The weakening of the covalent H-Se bonding results in a remarkably softening of phonon spectrum (solid blue line in the phonon density of states plot of Fig. 4d) compared with the case of H3Se (black dash line). The EPC calculations (the upper panel of Fig. 4d) gives a moderately strong EPC parameter λ of ~0.8 (see also Table S3), where the H-derived vibrations make a ~48% contribution. The smaller magnitude of λ in HSe than that of H3Se can be rationalized by the weakened bonding strength in the covalent-bond system37. Meanwhile, due to the relatively low logarithmic averaged phonon frequency ωlog (800–900 K) as the result of phonon softening, HSe exhibits a moderate Tc around 40 K.

Figure 4
Figure 4

(a) Electronic band structure of HSe in the P4/nmm structure at 250 GPa. Similar to Fig. 3a, the projection onto the H-s orbital is indicated by green circles. (b) Fermi surface of HSe at 250 GPa. (c) Comparison of the electron localization function within the (010) plane of HSe (left) and H3Se (right) at the same pressure. (d) Phonon density of states (lower panels), Eliashberg EPC spectral function α2F(ω) and EPC integration λ(ω) (upper panels) of HSe and H3Se.

For the Se-rich HSe2 stoichiometry, as expected its electronic structure is predominated by Se rather than H. The calculated rather small λ of 0.45 and low Tc of ~5 K (at 300 GPa) is reminiscent of superconducting solid Se at high pressures38.


After the completion of our work, we were aware of the work by Flores-Livas et al.35 predicting high-temperature superconductivity (with Tc up to 131 K) in selenium hydrides. Our work is different from theirs in several aspects: (i) they only focus on the H3Se stoichiometry in analogy to H3S, whereas we investigated the entire energy landscape of H-Se system by a more comprehensive global structure search. In addition to H3Se, we identified two new energetically stable stoichiometries i.e., HSe2 and HSe; (ii) we got a qualitatively different picture of energetic stability for selenium hydrides with respect to the elemental decomposition from theirs. Most of their structures are highly stable relative to the elemental decomposition, but in our work the stable stoichiometries against the elemental decomposition can only emerge at quite high pressures (above 100 GPa); (iii) in contrast to their work assuming H3Se as the most stable stoichiometry, we find in fact H3Se is marginally stable, and HSe2 and HSe are the most stable species at medium and high pressure region, respectively. The implication of our results is that the synthesis of H3Se in experiments may require particular kinetic control process.

The HSe phase in the P4/nmm symmetry represents another interesting high-symmetry structure in addition to the intriguing Im-3m structure discovered originally in H3S17. It should be pointed out that this structure is completely different from the one of HS (in the low symmetry of C2/m) predicted by Errea et al.34. For such a highly symmetric structure, the transition barrier between it and other isomers is usually high, which points to a strong kinetic stability with respect to variations of external conditions, thus favorable to experimental synthesis.

It should be mentioned that since the screened Coulomb repulsion parameter μ* (in the Allen-Dynes modified McMillan equation) cannot be evaluated by any first principle method, our theoretically predicted Tc has an intrinsic uncertainty originating from the empirical choice of μ*. Depending on particular materials, μ* usually falls in the range between 0.1 and 0.239. In current study the typical value μ* = 0.1 is used. For completeness, we have calculated the Tc values of H3Se and HSe (at 250 GPa) by taking a series of μ* = 0.05, 0.10, 0.15 and 0.20, as shown in Fig. S10. As seen, as μ* increases from 0.1 to 0.13, a somehow ubiquitous value for metallic hydrides12, Tc of H3Se and HSe change slightly from ~110 K and ~39 K to ~96 K and ~32 K, respectively.


In summary, with the aim of finding stable and metallic selenium hydrides for potential high-Tc superconductors, we explore via a global-minimum structure search method the hitherto unknown energy landscape of H-Se system at high pressures. Despite of similar electronic properties of Se and S, the high-pressure phase diagram of H-Se system is distinct from its H-S analogy. The H2Se stoichiometry expected from the normal valence states of H (+1) and Se (−2) is surprisingly unstable against the elemental decomposition. Three energetically stable phases, i.e. HSe2, HSe and H3Se, are identified above 120 GPa. While the H3S stoichiometry dominates as the most stable phase in the H-S system, H3Se turns out to be marginally stable and the most stable stoichiometry is the highly symmetric HSe. All of stable phases exhibit metallic features and superconducting activities. The latter two are predicted to have a high Tc of 40 K (HSe) and 110 K (H3Se). Experimental attempt to synthesize these new phases and verification of their superconductivity are called for.


The energetic stability of H-Se system is investigated by globally minimizing the potential energy surface at varied stoichiometries via an in-house developed swarm-intelligence based CALYPSO method40,41 in combination with ab initio density functional theory (DFT) total-energy calculations. Its validity in rapidly finding the stable ground-state structures has been demonstrated by its applications in various material systems ranging from elements to binary and ternary compounds40,42,43,44. The energetic calculations are performed using the plane-wave pseudopotential method within the generalized gradient approximation through the Perdew–Burke–Ernzerhof (PBE) exchange-correlation functional45, as implemented in the VASP code46. The electron-ion interaction was described by the projected-augmented-wave potentials with 1s1 and 4s24p4 as valence electrons for H and Se, respectively. During the structure search, an economy set of parameters are used to calculate the relative energetics of sampled structures, following which the cutoff energy of 600 eV for the expansion of wave-function and Monkhorst−Pack k-point sampling with grid spacing of 2π × 0.03 Å−1 were chosen to ensure the enthalpy converged to better than 1 meV/atom. The validity of pseudopotentials used at high pressures is carefully examined with the full-potential linearized augmented plane-wave method through the WIEN2k package47. The phonon spectrum for evaluating the lattice dynamic stability and electron-phonon coupling for superconducting properties of stable phases are performed within the framework of the linear-response theory via Quantum-ESPRESSO package48. The spin-orbit coupling is found to have a negligible effect on the electronic and superconductive properties of the current system (Fig. S11 and Table S4), and is thus reasonably neglected. See Supplementary information for more details.

Additional Information

How to cite this article: Zhang, S. et al. Phase Diagram and High-Temperature Superconductivity of Compressed Selenium Hydrides. Sci. Rep. 5, 15433; doi: 10.1038/srep15433 (2015).


  1. 1.

    & The discovery of superconductivity. Phys. Today. 63, 38–43 (2010).

  2. 2.

    & Possible high Tc superconductivity in the Ba-La-Cu-O system. Zeitschrift für Physik B, 64, 189–193 (1986).

  3. 3.

    , , , & Tc enhancement of HgBa2Ca2Cu3O8+δ by Tl substitution. Phys. Lett. A. 192, 122–124 (1994).

  4. 4.

    A Possible Path to RTS. AAPPS Bulletin 18, 9–21 (2008).

  5. 5.

    et al. Iron-Based Layered Superconductor:  LaOFeP. J. Am. Chem. Soc. 128, 10012–10013 (2006).

  6. 6.

    , , & Iron-Based Layered Superconductor La[O1-xFx]FeAs (x = 0.05−0.12) with Tc = 26 K. J. Am. Chem. Soc. 130, 3296–3297 (2008).

  7. 7.

    Superconductivity in iron compounds. Rev. Mod. Phys. 83, 1589–1652 (2011).

  8. 8.

    , & Theory of Superconductivity. Phys. Rev. 108, 1175 (1957).

  9. 9.

    et al. Superconductivity up to 164 K in HgBa2 Cam−1 CumO2m+2+δ (m = 1, 2, and 3) under quasihydrostatic. Phys. Rev. B 50, 4260–4263 (1994).

  10. 10.

    et al. Structure of the superconducting gap in MgB2 from point-contact spectroscopy. Supercond. Sci. Tech. 15, 526–532 (2002).

  11. 11.

    Metallic Hydrogen: A high-temperature superconductor? Phys. Rev. Lett. 21, 1748–1749 (1968).

  12. 12.

    Hydrogen dominant metallic alloys: high temperature superconductors? Phys. Rev. Lett. 92, 187002 (2004).

  13. 13.

    Superconductivity in compressed hydrogen-rich materials: Pressing on hydrogen. Physica. C, 514, 77–85 (2015).

  14. 14.

    et al. Superconducting High Pressure Phase of Germane. Phys. Rev. Lett. 101, 107002 (2008).

  15. 15.

    et al. Superconductivity at ~100 K in dense SiH4(H2)2 predicted by first principles. Proc. Natl. Acad. Sci. USA 107, 15708–15711 (2010).

  16. 16.

    , , , & Superconductive sodalite-like clathrate calcium hydride at high pressures. Proc. Natl. Acad. Sci. USA 109, 6463–6466 (2011).

  17. 17.

    et al. Pressure-induced metallization of dense (H2S)2H2 with high-Tc superconductivity. Sci. Rep. 4, 6968 (2014).

  18. 18.

    , , , & Ab initio simulation of phase transitions and dissociation of H2S at high pressure. Phys. Rev. Lett. 85, 1254–1257 (2000).

  19. 19.

    et al. Pressure-induced molecular dissociation and metallization in hydrogen-bonded H2S solid. Phys. Rev. Lett. 79, 1082–1085 (1997).

  20. 20.

    , , , & The metallization and superconductivity of dense hydrogen sulfide. J. Chem. Phys. 140, 174712 (2014).

  21. 21.

    , & Conventional superconductivity at 190 K at high pressures. arXiv. 1412, 0460 (2014).

  22. 22.

    et al. Pressure-induced decomposition of solid hydrogen sulfide. arXiv, 1501, 01784 (2015).

  23. 23.

    , , , & What superconducts in sulfur hydrides under pressure, and why. arXiv. 1501, 00196 (2014).

  24. 24.

    , , & Cubic H3S around 200 GPa: an atomic hydrogen superconductor stabilized by sulfur. arXiv, 1501, 03950 (2015).

  25. 25.

    & Hole superconductivity in H2S and other sulfides under high pressure. arXiv, 1412, 6251 (2014).

  26. 26.

    , & Non-BCS thermodynamic properties of H2S superconductor. arXiv. 1412, 8640 (2015).

  27. 27.

    , & First-principles study of the pressure and crystal-structure dependences of the superconducting transition temperature in compressed sulfur hydrides. arXiv. 1502, 00936 (2015).

  28. 28.

    , & Proton Resonance Study of Molecular Motion and Phase Behavior of Solid H2S and H2Se. J. Chem. Phys. 44, 3912 (1966).

  29. 29.

    & The crystal structure of trigonal selenium. Inorg. Chem. 6, 1589–1591 (1967).

  30. 30.

    , , , & Confirmation of the incommensurate nature of Se-IV at pressures below 70 GPa. Phys. Rev. B 70, 054101 (2004).

  31. 31.

    et al. Superconductivity in the PbO-type structure α-FeSe. Proc. Natl. Acad. Sci. USA 105, 14262 (2008).

  32. 32.

    Superconductivity and Chemistry. Angew. Chem. Int. Edit. 36, 1788–1806 (1997).

  33. 33.

    et al. High-pressure crystal structures and superconductivity of Stannane (SnH4). Proc. Natl. Acad. Sci. USA 107, 1317–1320 (2010).

  34. 34.

    et al. High-pressure hydrogen sulfide from first principles: a strongly anharmonic phonon-mediated superconductor. Phys. Rev. Lett. 114, 157004 (2015).

  35. 35.

    , & High temperature superconductivity in sulfur and selenium hydrides at high pressure. arXiv. 1501, 06336 (2015).

  36. 36.

    & Transition temperature of strong-coupled superconductors reanalyzed. Phys. Rev. B 12, 905–922 (1975).

  37. 37.

    & Superconductivity of MgB2: Covalent bonds driven metallic. Phys. Rev. Lett. 86, 4366–4369 (2001).

  38. 38.

    , , , & Crystal structure of the superconducting phases of S and Se. Phys. Rev. B 71, 214104 (2005).

  39. 39.

    Transition temperature of strong-coupled superconductors. Phys. Rev. 167, 331 (1968).

  40. 40.

    , , & Crystal structure prediction via particle-swarm optimization. Phys. Rev. B 82, 094116 (2010).

  41. 41.

    , , & CALYPSO: a method for crystal structure prediction. Comput. Phys. Commun. 183, 2063–2070 (2012).

  42. 42.

    , , & Predicted novel high-pressure phases of lithium. Phys. Rev. Lett. 106, 015503 (2011).

  43. 43.

    et al. High pressure partially ionic phase of water ice. Nat. Commun. 2, 563 (2011).

  44. 44.

    , , , & Reactions of xenon with iron and nickel are predicted in the Earth’s inner core. Nat. Chem. 6, 645–649 (2014).

  45. 45.

    et al. Atoms, molecules, solids, and surfaces: Applications of the generalized gradient approximation for exchange and correlation. Phys. Rev. B 46, 6671–6687 (1992).

  46. 46.

    & Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set. Phys. Rev. B 54, 11169–11186. (1996).

  47. 47.

    , , & Full-potential, linearized augmented plane wave programs for crystalline systems. Comput. Phys. Commun 59, 399–415 (1990).

  48. 48.

    et al. QUANTUM ESPRESSO: a modular and open-source software project for quantum simulations of materials. J. Phys. Condens. Mat. 21, 395502 (2009).

Download references


This research was supported by the China 973 Program (2011CB808200), Natural Science Foundation of China under No. 11274136, the 2012 Changjiang Scholars Program of China, the Natural Science Foundation of Jilin Province (20150101042JC), the fund of CAEP-CSCNS (R2015-03) and the Postdoctoral Science Foundation of China under grant 2013M541283. L.Z. acknowledges funding support from the Recruitment Program of Global Experts (the Thousand Young Talents Plan).

Author information


  1. State Key Laboratory of Superhard Materials, Jilin University, Changchun 130012, China

    • Shoutao Zhang
    • , Yanchao Wang
    • , Jurong Zhang
    • , Hanyu Liu
    • , Xin Zhong
    • , Guochun Yang
    • , Lijun Zhang
    •  & Yanming Ma
  2. Faculty of Chemistry, Northeast Normal University, Changchun 130024, China.

    • Guochun Yang
  3. College of Materials Science and Engineering and Key Laboratory of Automobile Materials of MOE, Jilin University, Changchun 130012, China

    • Lijun Zhang
  4. LCP, Institute of Applied Physics and Computational Mathematics, Beijing 100088, China

    • Hai-Feng Song
  5. Software Center for High Performance Numerical Simulation, China Academy of Engineering Physics, Beijing 100088, China

    • Hai-Feng Song


  1. Search for Shoutao Zhang in:

  2. Search for Yanchao Wang in:

  3. Search for Jurong Zhang in:

  4. Search for Hanyu Liu in:

  5. Search for Xin Zhong in:

  6. Search for Hai-Feng Song in:

  7. Search for Guochun Yang in:

  8. Search for Lijun Zhang in:

  9. Search for Yanming Ma in:


Y.M., L.Z. and G.Y. conceived the idea and supervised the project. S.Z., Y.W., J.Z., H.L. and X.Z. performed the calculations. All the authors contributed to analyzing the results. Y.M., L.Z., G.Y., S.Z. and H.S. wrote the paper.

Competing interests

The authors declare no competing financial interests.

Corresponding authors

Correspondence to Guochun Yang or Lijun Zhang or Yanming Ma.

Supplementary information


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Creative Commons BYThis work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit