• Matter and Radiation at Extremes
  • Vol. 10, Issue 2, 027802 (2025)
Yan Liu1,*, Tian Cui2, and Da Li1
Author Affiliations
  • 1State Key Laboratory of High Pressure and Superhard Materials and Key Laboratory of Material Simulation Methods and Software of Ministry of Education, College of Physics, Jilin University, Changchun 130012, Jilin, China
  • 2School of Physical Science and Technology, Ningbo University, Ningbo 315211, Zhejiang, China
  • show less
    DOI: 10.1063/5.0252519 Cite this Article
    Yan Liu, Tian Cui, Da Li. Leading role of satellite interstitial electrons in superconductivity in ternary superlithide Li14CP[J]. Matter and Radiation at Extremes, 2025, 10(2): 027802 Copy Citation Text show less

    Abstract

    The discovery of pressure-induced superconducting electrides has sparked a intense wave of interest in novel superconductors. However, opinions vary regarding the relationship between non-nuclear attractors (NNAs) and superconductivity, with two opposing views currently represented by the materials Li6P and Li6C. Here, we choose the ternary Li–C–P as a model system and reveal the underlying mechanism by which NNAs contribute to superconductivity. The loosely bound NNAs in the superlithide Li14CP covalently bond with Li and form unique satellite interstitial electrons (SIEs) around Li near the Fermi level, dominating the superconductivity. First-principles calculations show that the SIEs progressively increase in number and couple strongly with phonons at high pressure. Moreover, the Fermi surface nesting associated with SIEs induces phonon softening, further enhancing the electron–phonon coupling and giving the superlithide Li14CP a Tc of 10.6 K at 300 GPa. The leading role of SIEs in superconductivity is a general one and is also relevant to the recently predicted Li6P and Li6C. Our work presented here reshapes the understanding of NNA-dominated superconductivity and holds promise for guiding future discoveries and designs of novel high-temperature superconductors.

    I. INTRODUCTION

    Under compression, the increase in orbital energy at interstitial cavities is less pronounced than that of atomic orbitals,1 causing electrons to detach from atoms and localize in the cavities to form off-center lobes of electrons, also known as non-nuclear attractors (NNAs).2,3 These NNAs are the hallmark feature of high-pressure electrides and can be identified in electron-rich, -neutral, and -deficient systems.4–6 The alkali metal lithium (Li), with its strong electron-donating capacity, serves as an ideal candidate as a cation.7–10 To date, a multitude of high-pressure electrides involving the alkali metal Li has been identified.11–15 The magnitude and topology of NNAs depend on the stoichiometric ratios and chemical properties of the electron donors and acceptors, as well as external conditions, particularly pressure. These variabilities significantly expand the potential chemical space and the fascinating variety of physical properties of electrides.16,17 Even more intriguingly, unexpected superconductivity has been induced in pressure-stabilized Li-rich electrides such as CaLi2, Li3Ar, Li5He2, Li5C, Li5N, Li6Al, Li8Cs, Li9Te, Li10Se, and Li11Sb2.18–25 Not only that, but the critical temperatures Tc of these electrides have continuously rewritten the superconducting records, ranging from 10 K for Li6C to 39.3 K for Li6P,26,27 closing in on the McMillan limit (40 K), and reaching 73.1 K for the designed Li8Au,28 and progressively approaching the temperature of liquid nitrogen (77.4 K).

    The discovery of these pressure-induced superconducting electrides has sparked an intense wave of interest in novel Li-based superconductors. Unfortunately, however, the detailed influence of NNAs on superconductivity remains to be elucidated, and a consensus concerning their role in superconductivity has yet to be reached. Carbon (C) and phosphorus (P), as main group nonmetallic elements, follow the diagonal rule and exhibit similar chemical properties.29,30 When combined with Li, they have demonstrated the ability to form superconducting Li-based electrides. Although the NNAs in both Li6C and Li6P act as bridges linking anions to form electron-conducting channels, their contributions to superconductivity are strikingly distinct, resulting in a considerable disparity in Tc. Specifically, the dumbbell-like interconnected NNAs in Li6P play a dominant role in the superconducting transition, with the Fermi surface nesting associated with bands of these NNAs inducing phonon softening and thereby enhancing electron–phonon coupling (EPC) and favoring a high Tc close to the McMillan limit.27 By contrast, the superconductivity of Li6C is attributed to synergy of the interatomic coupling effect and phonon-coupled bands, which are predominantly governed by host sp-hybridized electrons rather than the cage-like NNAs.26 In addition to these two typical instances, other theoretical studies have also suggested that NNAs make a remarkable contribution to the superconductivity of electrides such as Li6Al,31 Li10Te,23 Li8Au,28 and Li8H4.32 However, an alternative point of view holds that the contribution of NNAs to superconductivity is negligible, as observed in electrides like Li5C20 and Li5Si.33

    Therefore, it is imperative to elucidate the underlying mechanism by which NNAs contribute to superconductivity, which is vital for understanding their properties as well as providing potentially valuable insights to aid in the further development of superconducting electrides. Given the current opposing views regarding the relationship between NNAs and superconductivity in Li6C and Li6P, further exploration of ternary compounds that combine them is warranted. Ternary compounds, compared with binary ones, can integrate the merits of different elements and provide a more abundant range of chemical compositions and stronger synergistic charge transfer.34,35 To date, a variety of ternary electrides have been successfully predicted.6,36 Consequently, the reaction between Li, C, and P under high pressure is expected to produce diverse NNA morphology, allowing for a more thorough study of the contribution of NNAs to superconductivity.

    Here, we choose ternary Li–C–P as an ideal system, focusing on the influence of NNAs on superconductivity at high pressure and attempting to reconcile the conflicting views on the contribution of NNAs to superconductivity in Li6P and Li6C. A novel high-pressure electride Li14CP, with an R3 space group, is predicted using a crystal structure search technique combined with first-principles calculations. Excitingly, R3-Li14CP is found to be a potential superconductor, with a Tc of 10.6 K at 300 GPa, positioned between the Tc values of Li6C and Li6P. The cage-like NNAs bond covalently with Li, leading to the formation of satellite interstitial electrons (SIEs) near the Fermi level EF. Our first-principles studies reveal that the SIEs play a leading role in determining Tc by softening the low-frequency phonon vibrations. Most importantly, the mechanism by which SIEs contribute to superconductivity revealed here is general and can be applied to the previously predicted Li6P and Li6C, effectively resolving the controversy regarding whether NNAs contribute to superconductivity.

    II. COMPUTATIONAL METHOD

    High-pressure structural predictions of the ternary Li–C–P system were performed from first principles with the implementation of the Ab Initio Random Structure Searching (AIRSS) and Crystal structure AnaLYsis by Particle Swarm Optimization (CALYPSO) codes,37,38 whose effectiveness has been confirmed by successful applications to the discovery of numerous structures of solids, surfaces, and clusters.39–41 The energy stability of different ternary stoichiometries was evaluated by their enthalpy of dissociation into the most competing element and binary compounds.20,27,42–44 Structural relaxations during the predictions used ultrasoft pseudopotentials in the Cambridge Sequential Total Energy Package (CASTEP) code.45 High-accuracy structural relaxations and enthalpy calculations employed the all-electron projector augmented wave (PAW) method pseudopotentials in the Vienna Ab initio Simulation Package (VASP) code with a kinetic cutoff energy of 500 eV.46 The valence electrons of the pseudopotentials were 1s22s1 for Li, 2s22p2 for C, and 3s23p3 for P. The exchange–correlation functional was given by the Perdew–Burke–Ernzerhof (PBE) parametrization within the generalized gradient approximation (GGA).47 A Monkhorst–Pack k-point grid was generated with a reciprocal space resolution of 2π × 0.03 Å−1 to ensure that the enthalpy calculations converged to <1 meV/atom. Bader charge analysis in the quantum theory of atoms in molecules (QTAIM) theory was used to reveal the extent of charge transfer.48 Chemical bonding analysis was investigated by the crystal orbital Hamilton population (COHP) using the Local Orbital Basis Suite Toward Electronic-Structure Reconstruction (LOBSTER) package.49 Density functional perturbation theory was adopted for calculation of EPC through the Quantum ESPRESSO (QE) code.50 Ultrasoft pseudopotentials were used with kinetic energy cutoffs of 80 Ry and 800 Ry for wave functions and charge density, respectively.51 A 24 × 24 × 24 k-point grid was used for Brillouin zone (BZ) integration in the electronic calculations and a 4 × 4 × 4 q-point grid was used to compute the EPC matrix elements. The superconducting critical temperatures Tc of all stable structures were estimated from the McMillan–Allen–Dynes equation.52

    III. RESULTS AND DISCUSSION

    A. Stability of Li–C–P compounds

    Comprehensive structure predictions were performed for ternary LixCyPz (x = 1–16, y = 1–3, and z = 1–3) compounds utilizing the AIRSS and CALYPSO structure search methods under pressures ranging from 150 to 350 GPa, with the main focus on lithium-rich species. A detailed description of the structure prediction is provided in the “Calculational details” section of the supplementary material, and the high-pressure phase diagrams of the ternary Li–C–P system are constructed in Figs. 1(a) and S1 (supplementary material). Thermodynamic stability of a compound is quantified by its energy above the convex hull Ehull.53 In Fig. 1(a), the stars represent stable ternary compounds with Ehull = 0 that are stable against decomposition into pure elements, together with other energetically most favorable binary and ternary phases, while the gray squares represent other metastable and unstable compounds. Four stable compounds Li3CP, Li8C2P, Li16C2P, and Li14CP were predicted at 300 GPa (detailed structural information is presented in Table S1, supplementary material). There is a near-linear increase in Li content, resulting in an enhanced number of electron donors and a greater abundance of electron-confined spaces available for accommodating NNAs, thereby providing the conditions necessary for the formation of diverse NNAs. All predicted stable Li–C–P ternaries are metallic and dynamically stable on the basis of electronic energy band structures and phonon dispersion curves (Figs. S2 and S3, supplementary material), but only Li14CP and Li16C2P exhibit NNA characteristics (Fig. S4, supplementary material). Subsequently, we concentrate on analyses of Li14CP, owing to its high crystal symmetry and lithium concentration.

    Phase stability and crystal structures of Li–C–P ternary compounds. (a) Ternary enthalpy phase diagram of Li–C–P system at 300 GPa. Gray squares represent metastable or unstable phases (above the convex hull) with different Ehull. The stars within the triangle represent thermally stable phases (located on the convex hull), while pink stars positioned on the vertices and edges of the triangle represent the corresponding stable elements and binary compounds. (b) Crystal structure of R3-Li14CP with rhombohedral primitive cell at 300 GPa. The excess electrons (yellow spheres) are located at the body-center position of the double hexagonal pyramid (NNA1) and dodecahedron (NNA2). The isosurface value is 0.75. Pink, blue, and brown spheres represent Li, P, and C atoms, respectively.

    Figure 1.Phase stability and crystal structures of Li–C–P ternary compounds. (a) Ternary enthalpy phase diagram of Li–C–P system at 300 GPa. Gray squares represent metastable or unstable phases (above the convex hull) with different Ehull. The stars within the triangle represent thermally stable phases (located on the convex hull), while pink stars positioned on the vertices and edges of the triangle represent the corresponding stable elements and binary compounds. (b) Crystal structure of R3-Li14CP with rhombohedral primitive cell at 300 GPa. The excess electrons (yellow spheres) are located at the body-center position of the double hexagonal pyramid (NNA1) and dodecahedron (NNA2). The isosurface value is 0.75. Pink, blue, and brown spheres represent Li, P, and C atoms, respectively.

    B. Crystal structure and electronic properties of Li14CP

    Li14CP crystallizes in a rhombohedral primitive cell with an R3 space group at 300 GPa [Fig. 1(b)]. The thermodynamically stable pressure of Li14CP can be ascertained from enthalpy difference calculations (Fig. S5, supplementary material), and it is found that the rhombohedral phase is the most stable phase against decomposition into Li6P, Li2C, and Li above 145 GPa. Within the primitive cell, the Li atoms occupy two different crystallographic Wyckoff sites 3b and 1a (Fig. S6, supplementary material). The 12 Li atoms at the 3b site (Li1, Li2, Li3, and Li4) are arranged in four equilateral triangles that are parallel to each other, with the body diagonal of the primitive cell as the central axis. Additional Li5 and Li6 atoms (1a site) are distributed along this central axis in conjunction with C and P atoms. Each C and P atom has 16-fold coordination, assembling the C–Li and P–Li polyhedrons that serve as the building block in the construction of the three-dimensional crystal structure [Fig. 1(b)]. In Li14CP, the Li atom serves as the electron donor, with a portion of the electrons occupying the electron shells of C and P atoms, and the excess electrons are accommodated in the interstitial cavities formed by coplanar connected polyhedrons, resulting in the formation of the NNAs. The calculated electron localization function (ELF) reveals the localized distribution of electrons within Li14CP [Fig. 1(b)], where the electron densities at the centers of both the double hexagonal pyramid (NNA1) and dodecahedron (NNA2) exceed the superposition of unbound atoms’ free electrons, thus providing evidence for the existence of two cage-like NNAs. It is worth mentioning that the NNAs in Li14CP exhibit similar characteristics to those in Li6C and Li6P, with the cage-like NNAs interconnecting with each other across intermediate extranuclear electrons of C and P atoms, forming the electron-conducting channels. However, the distinction lies in the presence of the electron conduction channel between the two cage-like NNAs in Li14CP.

    The electronic band structure in Fig. 2(a) shows that three bands, labeled I, II, and III, cross EF and form a broad density of states (DOS) peak around EF, which is predominantly attributed to the Li 2p orbitals and the non-nuclear centered orbital of the NNAs, with minor contributions from the atomic orbitals of C and P [Fig. 2(b)]. Furthermore, a pronounced Van Hove singularity (VHS) is observed near EF, corresponding to the flat band structure around the high-symmetry point T. Meanwhile, bands I, II, and III passing through EF along the Γ–T, H2–L, and Γ–S0 directions exhibit steep slopes. These not only characterize the metallicity but also imply potential superconductivity in Li14CP. Subsequent EPC calculations reveal that Li14CP is a superconductor with a Tc of 10.6 K at 300 GPa. Given that the interaction between NNAs and lattice originates primarily from the interaction of NNAs with the valence electrons of atoms, the electron distribution near the Fermi surface is crucial in elucidating the role of NNAs in superconducting electrides. The band-decomposed electron density maps for bands I, II, and III crossing EF on the (11̄0) plane in the rhombohedral primitive cell are displayed in Fig. 2(c). The conduction electrons occupying electronic states within band I are distributed in the interstices of the double hexagonal pyramid and diffuse toward adjacent Li atoms, emphasizing a more pronounced influence of NNA1 than NNA2 on the electrical conductivity of Li14CP. Specifically, the diffused NNA1 surrounds the Li5 atom like a satellite, forming the striking satellite interstitial electrons (SIEs). Furthermore, the band-decomposition electron density in the energy window −1 eV < EEF < 0 eV is also calculated (Fig. S7, supplementary material), and the electron distribution thus obtained is consistent with the above. The presence of SIEs is attributed to the covalent interaction between NNA1 and Li5, as evidenced by the two-dimensional ELF of the total electron density on the (11̄0) plane [Fig. 2(c)]. Valuable bonding information can be gleaned from the crystal orbital Hamilton population (COHP), and the integrated COHP (ICOHP) can be used as a qualitative measure of covalent bonding.41 The calculated ICOHP value of the NNA1–Li5 pair is −0.6, and its absolute value exceeds that of the typical NNA–Mg pair in Mg2Xe (−0.4), indicating strong chemical interactions with a large covalent component. Consequently, the striking SIEs occupying the electronic state of band I contribute significantly to the electrical conductivity, even superconductivity, of Li14CP.

    Electronic structure of Li14CP at 300 GPa. (a) Band structure. The Fermi level EF is defined as the origin (dashed line), and the three bold dispersion curves represent the bands across EF, primarily composed of electronic states located around Li and NNAs. (b) Projected density of states (PDOS) on the spheres located at Li, C, P, and NNA positions. (c) Two-dimension ELF (top left) and band-decomposed electron density maps of the three bands crossing EF on the (11̄0) plane (top right, bottom left, and bottom right), bisecting the rhombohedral primitive cell. The conduction electrons occupying electronic states within band I are distributed around NNA1 and diffuse toward adjacent Li atoms to form satellite interstitial electrons (SIEs).

    Figure 2.Electronic structure of Li14CP at 300 GPa. (a) Band structure. The Fermi level EF is defined as the origin (dashed line), and the three bold dispersion curves represent the bands across EF, primarily composed of electronic states located around Li and NNAs. (b) Projected density of states (PDOS) on the spheres located at Li, C, P, and NNA positions. (c) Two-dimension ELF (top left) and band-decomposed electron density maps of the three bands crossing EF on the (11̄0) plane (top right, bottom left, and bottom right), bisecting the rhombohedral primitive cell. The conduction electrons occupying electronic states within band I are distributed around NNA1 and diffuse toward adjacent Li atoms to form satellite interstitial electrons (SIEs).

    C. Superconductivity of Li14CP

    To investigate the influence of SIEs on the superconductivity in Li14CP, we analyzed the electron density alongside the most significant variations in SIEs (NSIEs) and the related EPC parameters at pressures of 200, 250, and 300 GPa (Table S2, supplementary material). As shown in Figs. 3(a) and 3(b), NSIEs is proportional to Tc and the EPC parameter λ, both of which progressively increase under pressure (5.4 K and 0.5 at 200 GPa, 10.6 K and 0.68 at 300 GPa). Strong compression reduces the volume of the interstitial cavity within the double hexagonal pyramid, hindering the entry of electrons released from the Li atoms. Consequently, a portion of the electrons are localized between Li and NNA1, thereby enhancing the covalent interaction between them. The pink histograms in Fig. 3(b) illustrate the variation of the ICOHP value of the NNA1–Li5 pair. It is evident that all ICOHPs are negative, and their absolute values gradually increase with increasing pressure, which provides compelling evidence for the enhancement of the covalent interaction between NNA1 and Li5; that is more SIEs are formed under strong compression. As expected, NSIEs grows by 28.6%, from 6.9 × 10−3 at 200 to 8.7 × 10−3 Å−3 at 300 GPa. Meanwhile, the electronic energy band I associated with the SIEs gradually flattens out near the high-symmetry point T (Fig. S8, supplementary material), promoting the formation of Cooper pairs and enhancing electron–phonon interactions.

    Electron–phonon coupling and mechanism of superconductivity in Li14CP. (a) Pressure-modulated electron density maps of band I crossing EF on the (11̄0) plane, in which the electron densities of the SIEs increase with increasing pressure. (b) Pressure-modulated EPC parameter λ, superconducting critical temperature Tc, number of electrons at the centers of SIEs NSIEs, and negative integrated COHP (−ICOHP) of the NNA1–Li5 pair. (c) Phonon dispersion curve with q-resolved EPC parameter λq at 300 GPa, where the radii of the brown circles are proportional to the strength of λq. (d) Eliashberg spectral function α2F(ω) and integral EPC parameter λ as a function of frequency. (e) Fermi surface nesting function ξ(q) along special q trajectories.

    Figure 3.Electron–phonon coupling and mechanism of superconductivity in Li14CP. (a) Pressure-modulated electron density maps of band I crossing EF on the (11̄0) plane, in which the electron densities of the SIEs increase with increasing pressure. (b) Pressure-modulated EPC parameter λ, superconducting critical temperature Tc, number of electrons at the centers of SIEs NSIEs, and negative integrated COHP (−ICOHP) of the NNA1–Li5 pair. (c) Phonon dispersion curve with q-resolved EPC parameter λq at 300 GPa, where the radii of the brown circles are proportional to the strength of λq. (d) Eliashberg spectral function α2F(ω) and integral EPC parameter λ as a function of frequency. (e) Fermi surface nesting function ξ(q) along special q trajectories.

    There is obvious phonon softening in the low-frequency acoustic branches along the F–Γ, Γ–T, H2–L, and Γ–S0 directions within the phonon dispersion curves of Li14CP at 300 GPa [Fig. 3(c)], and the q-resolved λq is especially large in these softened regions. At the same time, the Eliashberg spectral function α2F(ω) illustrated in Fig. 3(d) reveals that the low-frequency acoustic branches dominate the superconductivity so that modes below 8 THz make a contribution of 64.7% to the EPC parameter λ at 300 GPa. It is noteworthy that as pressure increases, the softening along the F–Γ, Γ–T, H2–L, and Γ–S0 directions progressively intensifies (Fig. S9, supplementary material), and the low-frequency vibrations contribute more significantly to the EPC, which is usually related to Fermi surface nesting (FSN). To elucidate this phenomenon, the nesting function ξ(q) of Li14CP at 300 GPa is calculated. This is a geometrical property of the Fermi surface and is particularly large for wave vectors that connect parallel portions of the surface.54 As can be seen in Fig. 3(e), ξ(q) exhibits a notable feature of sharp peaks along the Γ–T, H2–L, and Γ–S0 directions, which, except for that at Γ (i.e., q = 0, where the peak represents the entire Fermi surface nesting into itself and has no actual physical meaning),55 signify the occurrence of FSN. Coincidently, the nesting direction corresponds not only to the phonon softening direction, but also to the region where the electronic energy band I forms the SIEs crossing EF in Li14CP [Fig. 2(a)]. These results indicate that the highly localized SIEs play a major role in the coupling with phonons, thus contributing to enhancing the superconductivity of Li14CP.

    NNAs, including the SIEs proposed in this study, are essentially anions, and so the Li14CP framework is positively charged at high pressure. To further substantiate the leading role of the SIEs in the superconductivity of Li14CP, a hypothetical charged system is constructed by removing the valence electrons, in a process commonly referred to as hole doping.56–58 The electrons around EF are gradually neutralized by this operation (Table S3, supplementary material). The hole-doping-dependent EPC parameters at 300 GPa are calculated and presented in Fig. 4(a), from which it can be seen that both Tc and λ gradually decrease with increasing doping concentration and ultimately almost disappear as the doping level approaches 1e. By comparing the phonon dispersion curves, it is found that there is a clear disappearance of the softening in the low-frequency acoustic branches [Figs. 4(b) and S10, supplementary material], leading to a significant reduction in the percentage contribution of acoustic branches to λ [pink columns in Fig. 4(a)]. The disappearance of acoustic branches induced by hole doping is attributable to the vanishing of FSN, which can be corroborated by the changes in the nesting function ξ(q) in Fig. 4(c). Most importantly, a topological analysis of Li14CP reveals a bond critical point (BCP) between NNA1 and Li5 [Fig. 4(d) insert]. The charge density distribution at the BCP provides valuable insights into the bonding types and properties, while the sign of the Laplacian of the electron density [∇2ρ(r)] indicates whether the density is locally concentrated (negative) or depleted (positive). The calculated ∇2ρ(r) has a negative value in the primitive structure, confirming the covalent interaction between NNA2 and Li5, which manifests as SIEs near EF. Furthermore, as the doping level increases, ∇2ρ(r) gradually transforms from negative to positive, with its absolute value decreasing first and then increasing [Fig. 4(d)]. These results directly illustrate how the electron density of SIEs and the soft phonon modes caused by FSN lead to the strong EPC, thereby inducing the superconductivity of Li14CP at high pressure.

    Hole-doping-modulated electronic properties and superconductivity at 300 GPa. (a) EPC parameter λ, superconducting critical temperature Tc, and percentage contributions to λ of low-frequency acoustic branches below 8 THz. (b) Low-frequency acoustic branches with the q-resolved EPC parameter λq; the doping concentrations from top to bottom are 1, 0.4, and 0e, respectively. (c) Fermi surface nesting function ξ(q) along special q trajectories. (d) Sign of the Laplacian of the electron density [∇2ρ(r)] between NNA1 and Li5; the small black sphere in the inset corresponds to the bond critical point (BCP).

    Figure 4.Hole-doping-modulated electronic properties and superconductivity at 300 GPa. (a) EPC parameter λ, superconducting critical temperature Tc, and percentage contributions to λ of low-frequency acoustic branches below 8 THz. (b) Low-frequency acoustic branches with the q-resolved EPC parameter λq; the doping concentrations from top to bottom are 1, 0.4, and 0e, respectively. (c) Fermi surface nesting function ξ(q) along special q trajectories. (d) Sign of the Laplacian of the electron density [∇2ρ(r)] between NNA1 and Li5; the small black sphere in the inset corresponds to the bond critical point (BCP).

    D. Leading role of SIEs in superconductivity

    A leading role of SIEs in superconductivity has been proposed in the present study. Here, based on our point of view, the superconducting mechanisms of the previously predicted Li6P and Li6C are reanalyzed. From the available information,26,27 it can be concluded that both the connected dumbbell-like NNAs in Li6P and the cage-state NNAs in Li6C contribute to their metallic state [Fig. 5(a)], with these contributions ranking just slightly below that of Li. Calculations of the decomposed electron densities of the energy bands crossing EF in these two electrides provide further understanding of the mechanisms of superconductivity. As depicted in Fig. 5(b), there are obvious SIEs between two types of NNAs and their surrounding Li atoms, and the electron density of SIEs NSIEs exhibits an opposite trend with increasing pressure. Nevertheless, it is exciting to note that the variations of NSIEs in Li6P and Li6C are proportional to their respective critical temperatures Tc [Fig. 5(c)]. The larger value of NSIEs enhances the electron–phonon interaction, directly facilitating the improvement in Tc. Overall, in the binary Li–C and Li–P systems, as well as the ternary Li–C–P system, the existence of SIEs is conducive to superconductivity, and regulating the electron density of the SIEs is expected to uncover novel superconducting electrides with high Tc. We have thus unified the mechanisms of superconductivity in these three electrides and have successfully revealed the relationship between NNAs and superconductivity.

    Mechanisms of superconductivity in Li6P and Li6C. (a) Two-dimensional ELF. There are connected dumbbell-like and cage-state NNAs in Li6P and Li6C. (b) Electron density maps of the bands crossing EF, in which there are obvious SIEs between NNAs and the adjacent Li atoms. (c) Pressure-modulated superconducting critical temperature Tc and electron density of SIEs NSIEs.

    Figure 5.Mechanisms of superconductivity in Li6P and Li6C. (a) Two-dimensional ELF. There are connected dumbbell-like and cage-state NNAs in Li6P and Li6C. (b) Electron density maps of the bands crossing EF, in which there are obvious SIEs between NNAs and the adjacent Li atoms. (c) Pressure-modulated superconducting critical temperature Tc and electron density of SIEs NSIEs.

    IV. CONCLUSIONS

    We have presented a new mechanism by which NNAs contribute to superconductivity in the ternary superlithide Li14CP. Under compression, the excess electrons accumulate in double hexagonal pyramid interstitial cavities and form cage-like NNAs. These loosely bound NNAs have a covalent interaction with surrounding Li, leading to the formation of unique SIEs near EF. First-principles calculations show that the SIEs progressively increase in number and couple more strongly with phonons at higher pressure. Moreover, the Fermi surface nesting associated with SIEs induces phonon softening, further enhancing the electron–phonon coupling and enabling the superlithide Li14CP to achieve a Tc of up to 10.6 K at 300 GPa. This mechanism by which SIEs contribute to superconductivity also applies to the predicted compounds Li6P and Li6C, effectively reconciling previously conflicting views. Our work presented here thus provides deep insights into the mechanism of superconductivity in high-pressure electrides.

    SUPPLEMENTARY MATERIAL

    ACKNOWLEDGMENTS

    Acknowledgment. This work was supported by the National Key R&D Program of China (Grant No. 2023YFA1406200), the National Natural Science Foundation of China (Grant Nos. 12374004 and 12174141), and the High Performance Computing Center of Jilin University, China.

    References

    [1] R.Hoffmann, M. S.Miao. High pressure electrides: A predictive chemical and physical theory. Acc. Chem. Res., 47, 1311-1317(2014).

    [2] X.Dong, A. F.Goncharov, S.Lobanov, A. R.Oganov, E.Stavrouet?al.. A stable compound of helium and sodium at high pressure. Nat. Chem., 9, 440-445(2017).

    [3] J.Cioslowski. Nonnuclear attractors in the lithium dimeric molecule. J. Phys. Chem., 94, 5496-5498(1990).

    [4] Y. T.Gong, H.Hosono, K.Li, J. J.Wang. Electron-deficient-type electride Ca5Pb3: Extension of electride chemical space. J. Am. Chem. Soc., 143, 8821-8828(2021).

    [5] T.Katase, J.Kim, H.Mizoguchi, S.-W.Park, G. V.Vazheninet?al.. Origin of metallic nature of Na3N. J. Am. Chem. Soc., 143, 69-72(2021).

    [6] M. L.Evans, Y. T.Gong, S. Y.Wang, Z. Q.Wang, Y. J.Yanet?al.. Machine learning-accelerated discovery of A2BC2 ternary electrides with diverse anionic electron densities. J. Am. Chem. Soc., 145, 26412-26424(2023).

    [7] L. A.Burton, C.Liu, S. A.Nikolaev, W.Ren. Electrides: A review. J. Mater. Chem. C, 8, 10551-10567(2020).

    [8] R. J.Needs, C. J.Pickard. Dense low-coordination phases of lithium. Phys. Rev. Lett., 102, 146401(2009).

    [9] H.Hosono, M.Kitano. Advances in materials and applications of inorganic electrides. Chem. Rev., 121, 3121-3185(2021).

    [10] X.Chen, P.Jiang, Y. G.Li, D.Zhang, X. H.Zhenget?al.. Pressure-tuned one- to quasi-two-dimensional structural phase transition and superconductivity in LiP15. Phys. Rev. B, 105, 094109(2022).

    [11] J.Cho, S. Y.Liu, C. Z.Wang, S. H.Yi. Underlying mechanism of charge transfer in Li-doped MgH16 at high pressure. Phys. Rev. B, 102, 184509(2020).

    [12] L.Boeri, C.Heil, C.Kokail. Search for high-Tc conventional superconductivity at megabar pressures in the lithium-sulfur system. Phys. Rev. B, 94, 060502(2016).

    [13] R.Hoffmann, X. D.Wen, E.Zurek. (Barely) Solid Li(NH3)4: The electronics of an expanded metal. J. Am. Chem. Soc., 133, 3535-3547(2011).

    [14] N. W.Ashcroft, A.Hermann, R.Hoffmann, A.McSorley. From Wade–Mingos to Zintl–Klemm at 100 GPa: Binary compounds of boron and lithium. J. Am. Chem. Soc., 134, 18606-18618(2012).

    [15] J.Botana, J.Brgoch, C. J.Hou, M. S.Miao. Iodine anions beyond −1: Formation of LinI (n = 2–5) and its interaction with quasiatoms. Inorg. Chem., 55, 9377-9382(2016).

    [16] M.Eremets, Y. M.Ma, A. R.Oganov, I.Trojan, Y.Xieet?al.. Transparent dense sodium. Nature, 458, 182-185(2009).

    [17] B.Wan, Y.Xu, Y. F.Yuan, S. J.Zhao, L.Zhenget?al.. BaCu, a two-dimensional electride with Cu anions. J. Am. Chem. Soc., 146, 17508-17516(2024).

    [18] Y. M.Ma, A. R.Oganov, Y.Xie. Novel high pressure structures and superconductivity of CaLi2. Phys. Rev. Lett., 104, 177005(2010).

    [19] X.Dong, C. M.Hao, J. Y.Hou, A. R.Oganov, X. J.Wenget?al.. Helium-bearing superconductor at high pressure. Phys. Rev. B, 106, L220501(2022).

    [20] J.Chen, W. W.Cui, K.Gao, T. T.Gu, Q. F.Wanget?al.. Pressure-stabilized superconducting electride Li5C. Phys. Rev. B, 106, 054519(2022).

    [21] Z. Y.Wan, W. J.Xu, T. Y.Yang, C.Zhang, R. Q.Zhang. Predicted superconductivity and superionic state in the electride Li5N under high pressure. New J. Phys., 24, 113012(2022).

    [22] V. A.Blatov, H. M.Huang, A. R.Oganov, X. T.Wei, Q.Zhuet?al.. Novel topological motifs and superconductivity in Li-Cs system. Nano Lett., 23, 5012-5018(2023).

    [23] A.Bergara, F.Li, G. C.Yang, X. H.Zhang. Pressure-induced superconductivity in Li-Te electrides. Phys. Rev. B, 104, 134505(2021).

    [24] A.Bergara, G. C.Yang, X. H.Zhang, Y. P.Zhao. Superconducting Li10Se electride under pressure. J. Chem. Phys., 156, 194112(2022).

    [25] S. C.Ding, J. Y.Gao, Y.Liu, X. H.Zhang, Y. P.Zhaoet?al.. Superconducting Li11Sb2 electride at ambient pressure. J. Mater. Chem. C, 11, 17087-17092(2023).

    [26] D. F.Duan, Z.Liu, H.Song, F.Tian, Q.Zhuanget?al.. Proposed superconducting electride Li6C by sp-hybridized cage states at moderate pressures. Phys. Rev. Lett., 127, 157002(2021).

    [27] A.Bergara, H. Y.Xu, T.Yu, S. T.Zhang, Z. Y.Zhaoet?al.. Predicted pressure-induced superconducting transition in electride Li6P. Phys. Rev. Lett., 122, 097002(2019).

    [28] A.Bergara, S. C.Ding, F.Li, Y. S.Yao, X. H.Zhanget?al.. Superconductivity in Li8Au electride. Phys. Rev. B, 107, L100501(2023).

    [29] C. S.Cao, H. S.Hu, J.Li, W. H. E.Schwarz. Physical origin of chemical periodicities in the system of elements. Pure Appl. Chem., 91, 1969-1999(2019).

    [30] P.Pyykkö, P.Schwerdtfeger, O. R.Smits. The periodic table and the physics that drives it. Nat. Rev. Chem., 4, 359-380(2020).

    [31] Q.Lu, S. N.Pan, J. J.Wang, X. M.Wang, Y.Wanget?al.. Pressure stabilized lithium–aluminum compounds with both superconducting and superionic behaviors. Phys. Rev. Lett., 129, 246403(2022).

    [32] A.Bergara, S. C.Ding, Z. X.Guo, X.Li, X. H.Zhanget?al.. Superconductivity in Li8Hn electrides: The effect of interstitial anionic electrons on electron-phonon coupling. Phys. Rev. B, 109, 134505(2024).

    [33] Y. P.Feng, B.Gu, G.Su, J. Y.You. Emergent Kagome electrides. J. Am. Chem. Soc., 144, 5527-5534(2022).

    [34] X.Du, X.Li, Y. D.Wei, Z.Yang, X. H.Zhanget?al.. Au with sp3 hybridization in Li5AuP2. J. Phys. Chem. Lett., 13, 236-242(2022).

    [35] G.Yang, X.Zhang, Y.Zhao. Superconducting ternary hydrides under high pressure. Comput. Mol. Sci., 12, e1582(2022).

    [36] H.Hosono, J. J.Wang, Z. H.Wang, Q.Zhu. Ternary inorganic electrides with mixed bonding. Phys. Rev. B, 99, 064104(2019).

    [37] R. J.Needs, C. J.Pickard. Ab initio random structure searching. J. Phys. Condens. Matter, 23, 053201(2011).

    [38] J.Lv, Y. M.Ma, Y. C.Wang, L.Zhu. Crystal structure prediction via particle-swarm optimization. Phys. Rev. B, 82, 094116(2010).

    [39] T.Cui, M. J.Hutcheon, A. M.Shipley, H.Song, Z. H.Zhanget?al.. Design principles for high-temperature superconductors with a hydrogen-based alloy backbone at moderate pressure. Phys. Rev. Lett., 128, 047001(2022).

    [40] Y. R.Gao, J.Jiang, L.Li, Y.Liu, W. D.Zhuet?al.. Rich proton dynamics and phase behaviours of nanoconfined ices. Nat. Phys., 20, 456-464(2024).

    [41] T.Cui, D.Li, Y.Liu, R.Wang, Z. G.Wang. Formation of twelve-fold iodine coordination at high pressure. Nat. Commun., 13, 412(2022).

    [42] J.Lv, Y. M.Ma, Y. C.Wang, L.Zhu. Predicted novel high-pressure phases of lithium. Phys. Rev. Lett., 106, 015503(2011).

    [43] M.Amsler, J. A.Flores-Livas, S.Goedecker, L.Lehtovaara, S.Pailhèset?al.. Raman activity of sp3 carbon allotropes under pressure: A density functional theory study. Phys. Rev. B, 85, 155428(2012).

    [44] L.Boeri, A. P.Drozdov, J. A.Flores-Livas, G.Profeta, A.Sannaet?al.. Interplay between structure and superconductivity: Metastable phases of phosphorus under pressure. Phys. Rev. Materials, 1, 024802(2017).

    [45] P. J.Hasnip, J. D. L.Philip, C. J.Pickard, M. J.Probert, M. D.Segallet?al.. First-principles simulation: Ideas, illustrations and the CASTEP code. J. Phys. Condens. Matter, 14, 2717(2002).

    [46] J.Furthmüller, G.Kresse. Efficiency of ab-initio total energy calculations for metals and semiconductors using a plane-wave basis set. Comput. Mater. Sci., 6, 15-50(1996).

    [47] K.Burke, M.Ernzerhof, J. P.Perdew. Generalized gradient approximation made simple. Phys. Rev. Lett., 77, 3865-3868(1996).

    [48] G.Henkelman, E.Sanville, W.Tang. A grid-based Bader analysis algorithm without lattice bias. J. Phys. Condens. Matter, 21, 084204(2009).

    [49] V. L.Deringer, R.Dronskowski, A. L.Tchougréeff. Crystal Orbital Hamilton Population (COHP) analysis as projected from plane-wave basis sets. J. Phys. Chem. A, 115, 5461-5466(2011).

    [50] S.Baroni, N.Bonini, M.Calandra, R.Car, P.Giannozziet?al.. Quantum ESPRESSO: A modular and open-source software project for quantum simulations of materials. J. Phys. Condens. Matter, 21, 395502(2009).

    [51] D.Joubert, G.Kresse. From ultrasoft pseudopotentials to the projector augmented-wave method. Phys. Rev. B, 59, 1758-1775(1999).

    [52] P. B.Allen, R. C.Dynes. Transition temperature of strong-coupled superconductors reanalyzed. Phys. Rev. B, 12, 905-922(1975).

    [53] X.Chen, Y.Dan, B.Suo, H.Xiao. Comment on ‘Accelerated discovery of new 8-electron half-Heusler compounds as promising energy and topological quantum materials. J. Phys. Chem. C, 124, 2247-2249(2020).

    [54] T.Bazhirov, M. L.Cohen, J.Noffsinger. Superconductivity and electron-phonon coupling in lithium at high pressures. Phys. Rev. B, 82, 184509(2010).

    [55] D. F.Duan, H. Y.Liu, Y. B.Ma, Z. J.Shao, H. Y.Yuet?al.. Divergent synthesis routes and superconductivity of ternary hydride MgSiH6 at high pressure. Phys. Rev. B, 96, 144518(2017).

    [56] Y.Bang, H.Jeon, S. Y.Liu, C. Z.Wang, S.Yiet?al.. Effect of hole doping on superconductivity in compressed CeH9 at high pressures. Phys. Rev. B, 104, L020504(2021).

    [57] W. W.Cui, J.Hao, Y. W.Li, J. M.Shi, K.Yang. Superconductivity of graphenelike hydrogen in H2He at high pressure. Phys. Rev. B, 107, 024501(2023).

    [58] T.Cui, D. F.Duan, Z.Liu, Q.Zhuang. High-temperature superconductivity in electrides dominated by hybridized p-orbital-like electride states. Phys. Rev. B, 108, L100507(2023).

    Yan Liu, Tian Cui, Da Li. Leading role of satellite interstitial electrons in superconductivity in ternary superlithide Li14CP[J]. Matter and Radiation at Extremes, 2025, 10(2): 027802
    Download Citation