WorksPhysics and Astronomy
Superconducting topological Dirac semimetals: -Si and -NaSi
Abstract
We theoretically propose that hexagonal silicon-based crystals, -Si and -NaSi, are topological Dirac semimetals with superconducting critical temperatures of 12 K and 13 K, respectively, at ambient pressure. Band inversion occurs with the Fu-Kane topological invariant , even in the absence of spin-orbit coupling. The Dirac nodes protected by crystal rotational symmetry remain gapless with spin-orbit coupling. Using first-principles calculations, we find pressure-induced topological phase transitions for -Si and -NaSi with critical external pressures of 11.5 GPa and 14.9 GPa, respectively. Above the critical pressures, the Dirac bands are gapped with , while the superconducting states and the crystal symmetries are retained.Our results may shed light into a search for silicon-based topological materials with superconductivity.
Introduction
Semiconducting silicon becomes indispensable in electronics due to its versatile features such as the ease of electron or hole doping in a wide range, high-temperature stability, non toxicity, and natural abundance. It is not uncommon to modify phases of solids by varying crystal structures and/or applying external stimuli. There has been a great effort in fabricating silicon in different condensed matter phases, especially a superconducting phase. Some metallic silicon phases were proposed to be superconducting under high pressures [1, 2, 3, 4, 5]. For doped silicon clathrates and boron-doped cubic silicon, superconductivity was observed at ambient pressure [6, 7, 8]. Recently, hexagonal silicon-based crystals -Si and -NaSi, which can be synthesized similarly to Ref. 9, were predicted to show superconductivity (transition temperature of 12 K and 13 K, respectively) at ambient pressure [10].
Topology in the reciprocal space plays an important role in quantum materials properties due to topological protection. Based on symmetries, gapped and gapless quantum phases including a superconducting phase can be topologically classified [11]. For example, three-dimensional Dirac semimetals which can exist in the presence of time-reversal and inversion symmetries, are categorized into two classes [12]. In the first class, band inversion occurs with a nontrivial Fu-Kane topological invariant [13] and associated Fermi arc surface states, and Dirac nodes are protected by crystal rotational symmetry [15, 44]. This class is henceforth referred to as a topological Dirac semimetal phase, following Ref. [12]. In the second class, Dirac nodes are protected by nonsymmorphic space group [16] without band inversion.
Most topological materials are not based on silicon. A topological Dirac semimetal phase was discovered in a silicon-containing compound, CaAlSi (space group No. 164, point group ) [17, 18] in the presence of spin-orbit coupling (SOC). Silicon-based crystals -AHT-Si and -VFI-Si have been proposed to be topological nodal line semimetals only when SOC is excluded [19]. With SOC, the nodal lines are gapped with a small energy gap of 1 meV [19].
Recently, superconductivity was observed in three-dimensional topological Dirac semimetals such as CdAs[20, 21, 22], AuPb [23, 24], KZnBi [25], PdTe [26, 27], and BaSn[28]. To the best of our knowledge, silicon-based topological Dirac semimetals with superconductivity have not been proposed or synthesized yet. Superconductivity in topological Dirac semimetals is appealing considering the possibility to create odd-parity or spin-triplet Cooper pair interaction by doping and/or breaking time-reversal symmetry [29, 30, 31]. Furthermore, superconductors with a nontrivial Fu-Kane topological invariant were proposed to be platforms for realization of Majorana zero modes [32] whose braiding is used for topological quantum computation [33, 34]. Families of superconductors with a nontrivial invariant showed promising experimental signatures of Majorana zero modes at the ends of magnetic vortices [35, 36].
In this work, we propose that -Si and -NaSi are topological Dirac semimetals with superconductivity at ambient pressure even when SOC is included. We find that the silicon-based crystals undergo a topological phase transition driven by hydrostatic pressure without any structural changes. Above slightly different critical pressures, both silicon-based crystals become topologically trivial in the superconducting state.
Methods
We employ density functional theory (DFT) with projector augmented wave (PAW) [37] pseudopotentials and the Perdew-Burke-Ernzerhof generalized gradient approximation (PBE-GGA) [38] for the exchange-correlation functional, as implemented in the VASP software [39]. A plane-wave basis set with a kinetic energy cutoff of 500 eV is used. We use -centered -point meshes of for bulk -Si and -NaSi and for finite slabs. Starting with the geometries from Ref. 10, we further optimize the atomic coordinates and lattice constants until the residual forces are less than 0.01 eV/Å with and without pressure. The optimized bulk hexagonal lattice parameters for -Si (-NaSi) are (6.76) Å and (2.44) Å without pressure, which are close to those in Ref. 10. The relaxed atomic coordinates can be found in Appendix § 5.
We compute the Fu-Kane topological invariant [13] from the parity of all valence or occupied bands at the time-reversal invariant momentum (TRIM) points as well as by using the Wilson loop method [40, 41]. To apply the Wilson loop method and to investigate the surface states, we first construct a tight-binding Hamiltonian for the Si bands by generating 24 (33) maximally localized Wannier functions for -Si (-NaSi), using WANNIER90 code [42, 56] (see Appendix § 5.1). Then, we calculate the surface Green’s function of the semi-infinite system based on the Wannier-function tight-binding model, using WannierTools code[41]. Irreducible representations of bands at the high symmetry points and along the - axis are obtained using Quantum Espresso [43].
Results and discussion
The atomic structure of -NaSi (space group no. 175, point group ) is shown in Figure 1(a), where the stacking along the axis is identical for all atomic layers. The structure of -Si is obtained by simply removing Na atoms from that of -NaSi. (see Appendix § 5.) The Na atoms can be easily removed by the degassing process, since the migration barrier of Na atoms along the cylindrical holes is only 0.48 eV [10]. This barrier is 0.17 eV lower than that for the phase [9].
Figure 1. (a) Top view of the atomic structure of -NaSi with a unit cell. Isostructure -Si can be obtained by removing Na atoms from -NaSi. (b) First Brillouin zone for -Si and -NaSi. SOC-included DFT band structures of -Si at (c) 0 GPa and (d) 15 GPa, where the Fermi level is set to zero. All bands are at least doubly degenerate. Relative proportions of Si and -orbital characters are encoded in colors. The Dirac points are indicated as arrows and the parity of a few occupied bands are shown.
Let us first discuss electronic and topological properties without pressure and then with pressure. Figure 1(c) shows the electronic structure of -Si at symmetric points with SOC in the absence of pressure. Mostly -orbital character is dominant near the Fermi level. However, -orbital character is locally dominant in the plane, i.e., along the A-L-H-A lines. At the L point, bands at eV and eV (relative to the Fermi level) have strong -orbital and -orbital characters, respectively, while bands at eV have characters of mixed and orbitals. (see Appendix § 5.2.) Strong Si -orbital characteristics also appear in most of the space except for the plane for -NaSi, as presented in Figure 2.

Figure 2. Electronic band structure obtained from first-principles calculations using DFT. First-principles band structures of -NaSi at pressure (a) 0 GPa and (b) 15 GPa, including spin-orbit coupling. According to the relative weight of the Si character and the Si character, the band color was encoded as shown in the legend. The band structure calculation at pressure 15 GPa shows the band structure just before band inversion occurs. The situation of band inversion that occurs with changes in pressure can be seen, especially in the L-H section.
For -Si at 0 GPa, the parity analysis at the TRIM points gives (see Appendix § 5.1), which results from the band inversion at the L points, as shown in Figure 1(c). (Note that a unit cell of -Si has 24 valence electrons.) At the L points, the 23th and 24th (25th and 26th) bands at eV ( eV) have parity, as depicted in Figure 1(c). The band inversion is not induced by SOC because without SOC.
The analysis of irreducible representations at the L point indicates that the opposite parity for the two band pairs originates from the horizontal mirror symmetry . The irreducible representation and eigenvalues of of the bands at the L point are listed in Table I. The 23th/24th bands belong to irreducible representation (where an eigenvalue of horizontal mirror reflection is ), while the 25th/26th bands belong to irreducible representation (where an eigenvalue of is ). We also confirm that using the Wilson loop method. (see Appendix § 5.1.) Even when the parity of all occupied bands below the Fermi level is counted, we find that persists.
For -Si with SOC in the absence of pressure, along the -A direction, we find several possible band crossing points in the vicinity of the Fermi level. To identify gapless Dirac nodes, we zoom-in the bands near the crossing points with numerical accuracy of 10 eV, as shown in Figure 3. The bands at , , and are gapped

Figure 3. SOC-included DFT bands for (a) -Si at 0 GPa, between and A. (b)-(e) enlarged bands near 5 possible band crossing points (-) for (a). represent irreducible representations of double group along the -A direction. Points and are gapless Dirac nodes (when bands and bands meet each other).
| band # | irrep | |||
| 1, 2 | ||||
| 3, 4 | ||||
| 5, 6 | ||||
| 7, 8 | ||||
| 9, 10 | ||||
| 11, 12 | ||||
| 13, 14 | ||||
| 15, 16 | ||||
| 17, 18 | ||||
| 19, 20 | ||||
| 21, 22 | ||||
| 23, 24 | ||||
| 25, 26 |
Table 1. Irreducible representation (irrep), point group symmetries, and their eigenvalues of bands at L point, for -Si at 0 GPa. is two-fold rotation operator with respect to the -axis. and are two-fold inversion and horizontal mirror plane operators, respectively. The highest occupied band is 24th band.
with 0.07-28.9 meV, while the crossing points and are gapless within numerical accuracy. We also analyze irreducible representations of the bands near the crossing points, considering that double group is applied to the bands along the -A direction. This symmetry analysis [Figure 3(b)-(Figure 3e)] agrees with the numerical result. The Dirac nodes along the axis are protected by the crystal sixfold rotational symmetry . The gapless protected Dirac nodes in conjunction with our analysis of the invariant, suggest that -Si is a topological Dirac semimetal.
For -NaSi, it is tricky to compute the invariant by counting the parity of bands at the TRIM points, where is the number of valence electrons, since is now an odd number due to the Na atom. (In Ref. 13, each degenerate pair was counted only once for centrosymmetric metals.) In order to circumvent this, we take into account all occupied bands below the Fermi level at the TRIM points in our calculation of the invariant. Note that there are different numbers of occupied bands at different TRIM points. For -NaSi at 0 GPa, the product of the parity values of all occupied bands at each L point is positive, while the corresponding products at the other 5 TRIM points are negative. (See Appendix § 5.1.) This gives rise to and the band inversion also occurs at the L points, as shown in Figure 2(a). Therefore, similarly to -Si, -NaSi is also a topological Dirac semimetal. The gapless Dirac nodes [crossing points iii and v in Figure 4(d)] are found below the Fermi level.

Figure 4. (a) DFT bands in -NaSi at 0 GPa, between and A. (b)-(d) enlarged bands near 5 possible band crossing points (i-v) for (a). Points iii and v are gapless Dirac nodes (when bands and bands meet each other).
The nontrivial topology of three-dimensional Dirac bands in -Si and -NaSi indicates that there are nontrivial surface states. We calculate surface states using the surface Green’s function of the semi-infinite system described earlier, considering two surface types: surfaces parallel and perpendicular to the -A direction, i.e., (100) and (001) surfaces, respectively. Figure 5 shows the calculated local density of states which is an imaginary part of the surface Green’s function, for the (100) and (001) surfaces. At the chemical potential, corresponding to one of the Dirac node energies, for the (100) surface, two arc-shaped (left and right sides of hourglass) surface states connect the two Dirac node projection points (with bulk characteristics) indicated as solid dots along the axis in Figure 5(b). This is expected because each Dirac node consists of degenerate Weyl nodes with opposite chirality in topological Dirac semimetals [44, 45].

Figure 5. Electronic structure or local density of states of surface states along the two-dimensional symmetric points (a,c), in the - plane (b) and in the - plane (d) for the semi-infinite (100) and (001) surfaces of -Si, respectively. Horizontal dashed lines in (a) and (c) represent the Dirac point energies, respectively. Filled dots in (b) represent the projection points of the Dirac nodes onto the - plane. In (a)-(d), bright color indicates higher density of states at the surface.
For the (001) surface, the surface states appear as a point [Figure 5(d)]. In this case, the two Dirac nodes along the axis are projected onto the same point, i.e., the origin, in the - plane. Thus, point-like surface states are expected rather than arc-shaped surface states. The features of the surface states for the (100) and (001) surfaces agree with those of prototype topological Dirac semimetals NaBi[44] and CdAs[15].
Now with pressure up to 15 GPa, the crystal symmetry is still maintained for -Si and -NaSi. Since the effect of pressure is similar for both crystals, as shown in Figure 1 and Figure 2, we mainly discuss -Si. At 15 GPa, the amount of changes of the band structure depends on orbital characters. Specifically, the bands of -orbital character are sensitive to pressure, while the bands of -orbital character are somewhat rigid [Figure 1(d)]. The band structure near the Fermi level changes significantly only in the plane [Figure 1(c) vs (d)]. At the L points, the -orbital band is shifted by 0.51 eV, whereas the highest-energy occupied -orbital band is shifted by 0.16 eV.
While the topological invariant is insensitive to small local perturbations such as disorder or impurities, pressure-induced deformation can be used to control the electronic structure [46, 47, 48, 49]. For -Si, under high pressure, the parity of the 23th/24th bands is reversed to that of the 25th/26th bands at the L point [See the bands in the energy window between and eV in Figure 1(d)], and so the bands with − parity become higher in energy than the bands with + parity, which removes the band inversion. This leads to a topological phase transition from nontrivial to trivial . The critical external pressure for the topological phase transition in -Si is 11.5 GPa. [See Figure 6(a).] The gapless Dirac nodes which exist below the critical pressure now open up a gap above the critical pressure, as shown in Fig. 1(d)`, due to the topological phase transition. In the case of -NaSi, a similar topological phase transition occurs at 14.9 GPa, which is somewhat higher than the critical pressure of -Si.
Figure 6. (a) topological invariants for -Si and -NaSi are shown as a function of external hydrostatic pressure. Critical external pressures of the topological phase transitions for -Si and -NaSi are 11.5 GPa and 14.9 GPa, respectively. (b) Superconducting critical temperatures of -Si and -NaSi, obtained from Ref. 10 are shown as a function of external pressure. In (a) and (b), the filled (empty) symbols indicate the () phase.
Theoretically, -Si and -NaSi were proposed to be superconductors at ambient pressure[10]. Figure 6(b) shows the superconducting critical temperature as a function of hydrostatic pressure. While decreases monotonically as pressure () increases, both -Si and -NaSi are superconducting within the pressure range we consider, GPa. Therefore, we propose that -Si (-NaSi) is a superconducting topological Dirac semimetal below 11.5 (14.9) GPa. We emphasize that -Si and -NaSi consist of light elements with atomic number 14 or less, based on silicon, in contrast to reported superconducting topological Dirac semimetals such as CdAs[20, 21, 22], AuPb[23], KZnBi[25], PdTe[26, 27], and BaSn[28] which consist of heavy elements.
Due to the nesting vectors along , it turns out that there is a modulation in Si-Si bond lengths (out-of-hexagonal plane) formed in the direction. Indeed, we found a small oscillation of the out-of-plane bond length in the slab geometry. (see Appendix § 6.) The modulation is strongest on the surface and decreases near the center of the slab.
We calculated the Fermi surfaces from the two crystal structures, -Si and -NaSi, as presented in Figure 7. -Si and -NaSi can provide Dirac cones near or at the Fermi energy and show an anisotropic conducting channel due to their anisotropic bonding nature. Evidence of our proposed pressure-induced topological phase transition in -Si and -NaSi may be explored by transport experiments. The three-dimensional Dirac nodes and resultant arc-shaped surface states are supposed to induce a unique field-direction dependence in Shubnikov-de Haas oscillations[50]. The chiral anomaly associated with Weyl nodes suggests the following interesting properties: (i) A large negative magnetoresistance is found when an external magnetic field is parallel to a current direction[51, 52]; (ii) Thermoelectric properties depends on the relative direction between an external magnetic field and a temperature gradient[53]; (iii) The giant planar Hall effect is expected when a current direction is not parallel to a magnetic field direction in plane[54, 55]. By measuring variations of the above properties upon external pressure, the proposed topological phase transition can be experimentally probed, similarly to Ref. 49.

Figure 7. Fermi surfaces of (a) -Si at 0 GPa, (b) -Si at 15 GPa, (c) -NaSi at 0 GPa, and (d) -NaSi at 15 GPa are shown, respectively. The Fermi surface is the surface of constant energy in the first BZ which separates occupied from unoccupied electron states at zero temperature. Electronic band structure obtained from first-principles calculations using DFT.
Conclusions
In summary, we have proposed that -Si and -NaSi crystals are superconducting topological Dirac semimetals at ambient pressure with . The two gapless bulk Dirac nodes appear at eV and eV for the former and at eV and eV for the latter below the Fermi level. With hole doping, the Fermi level may be lowered to one of the two Dirac node energies for experimental signatures of the topological phase. The gapless Dirac nodes are protected by the crystal rotational symmetry in the presence of time-reversal symmetry. The topological Dirac semimetal phase for the crystals becomes topologically trivial beyond a critical external pressure. This pressure-induced topological phase transition retains superconductivity and the original crystal symmetry group. The coexistence of a topological Dirac semimetal state and a superconducting state in the present crystals, along with a pressure-induced topological phase transition, will provide an interesting platform to study the interplay between topology in electronic structure and superconductivity.
Acknowledgments
We thank Sohrab Ismail-Beigi for helpful discussions. This work also used the Extreme Science and Engineering Discovery Environment (XSEDE), which is supported by the National Science Foundation grant number ACI-1548562, by using computer time on the Comet supercomputer as enabled by XSEDE allocation MCA08X007.
I.H.L was supported by the National Center for Materials Research Data (NCMRD) through the National Research Foundation of Korea (NRF) funded by the Ministry of Science and ICT (NRF-2021M3A7C2089748).
Atomic structure of -
Figure 8 shows the atomic structure of -, which can be obtained by removing Na atoms from -. In Table II, we provide the information of the atomic structure including lattice parameters and atomic coordinate. Note that all the structures have space group symmetry, at both 0 GPa and 15 GPa.
Figure 8. (a) Top view and (b) side view of the atomic structure of -. (c) First Brillouin zone (BZ) of -. , , , and points are time reversal invariant momentum points, which are related to topological invariant calculation especially for crystals with inversion symmetry. (space group no. 175) is belong to the hexagonal Bravais lattice crystal system. The space group no. 175 has inversion symmetry.
| Material | (GPa) | (Å) | (Å) | Atom | |||
| -Si | 0 | 6.813 | 2.501 | Si (6j) | 0.154 | 0.733 | 0 |
| -Si | 15 | 6.542 | 2.402 | Si (6j) | 0.145 | 0.725 | 0 |
| -NaSi | 0 | 6.755 | 2.444 | Na (1b) | 0 | 0 | 0.5 |
| -NaSi | 0 | 6.755 | 2.444 | Si (6j) | 0.145 | 0.710 | 0 |
| -NaSi | 15 | 6.536 | 2.377 | Na (1b) | 0 | 0 | 0.5 |
| -NaSi | 15 | 6.536 | 2.377 | Si (6j) | 0.144 | 0.713 | 0 |
Table II. Optimized lattice parameters and atomic coordinates of -Si and -NaSi phases at external pressure () of 0 GPa and 15 GPa, obtained from first-principles calculations using density-functional theory (DFT). Both -NaSi and -Si phases are hexagonal with , and .
topological invariant calculations
The topological phase of band structure can be classified by calculating topological invariant. We check the topology by calculating the parities at time-reversal invariant momenta, as shown in Table III, and using the hybrid Wannier charge center flow technique, which is equivalent to the Wilson loop method. To investigate the topological nature of the electronic states, we obtained maximally localized Wannier functions using the WANNIER90 code [42, 56] (see Figure 9) and performed hybrid Wannier charge center flow calculations. With this Wannier model, the symmetry properties and topological invariant could be directly computed. topological invariants were calculated by observing hybrid Wannier charge center flows. The topological surface states were calculated using the Green’s function approach, as implemented in the WannierTools code [41].

Figure 9. First-principles bands (solid line) and Wannier bands (red filled circles) of -Si at (a) 0 GPa and (b) 15 GPa. In the eV energy scale, there are no visible differences between two different bands. Spin-orbit coupling is included.
| k-point | -Si 0 GPa | -Si 15 GPa | -NaSi 0 GPa | -NaSi 15 GPa |
| (0,0,0) | + | + | − | − |
| (0.5, 0, 0) | + | + | − | − |
| (0,, 0.5, 0) | + | + | − | − |
| (0, 0, 0.5) | + | + | − | − |
| (0.5, 0.5, 0) | + | + | − | − |
| (0.5, 0, 0.5) | − | + | + | − |
| (0, 0.5, 0.5) | − | + | + | − |
| (0.5, 0.5, 0.5) | − | + | + | − |
Table III. Parities at time-reversal invariant momentum points of -Si and -NaSi at 0 GPa and 15 GPa.
Space group no. 175 (, hexagonal) has inversion symmetry. Because of the existence of inversion symmetry in both -Si and -NaSi, the Fu-Kane parity criterion can be used to easily calculate the topological invariants. We were able to obtain the same topological invariants for both -Si and -NaSi by using two different methods, observing hybrid Wannier charge center flows and investigating Fu-Kane parity criterion. is the strong topological index. () are weak topological indices. Table IV show the example for -Si at 0 GPa. From the hybrid
Wannier charge center flow calculations, we calculated (). is obtained by
and thus we have . Therefore, -Si at 0 GPa has a strong topological nature. Similarly, we find that -NaSi at 0 GPa has a strong topological nature.
| -plane | ||
| - plane | ||
| - plane | ||
| - plane | ||
| - plane | ||
| - plane | ||
| - plane |
caption: TABLE IV. and () for P6/-Si obtained from hybrid Wannier charge center flow calculations, and correspond points and -planes.
:::
Band structures of P6/-NaSi and P6/-Si
In Figure 10, we present the the weight of in-plane Si character () and out-of-plane Si character () on the band structure of P6/-Si. At 0 GPa, -dominant band is higher in energy than the band along the L-H-A interval, whereas the band has lower energy than the band at 15 GPa.

Figure 10. Electronic band structure obtained from first-principles calculations using DFT. In the presence of SOC, band structures of P6/-Si at (a) 0 GPa and (b) 15 GPa and their projections on the the Si character are shown. According to the relative weight of in-plane Si character () and out-of-plane Si character () the band colors were encoded as shown in the legend. The situation of band inversion that occurs with changes in external pressure ( GPa) can be seen, especially in the L-H-A interval.
Si-Si bond lengths in the slab geometry
In Figure 11, we plot the Si-Si bond length along the direction (out of plane from the hexagonal surface). Note that the bond length distribution has an oscillation, while the oscillation of the bond length is largest and decreases near the bulk-like region.
Figure 11. Si-Si bond lengths, formed in the direction also denoted by out-of-plane (oop bond lengths) from the hexagonal plane, in the relaxed 40-layer -Si slab are shown. Position of the surface along the out-of-plane, is set to be zero. Dashed line represents the Si-Si bond length in the unrelaxed slab.

(a)

(b)

(c)

(d)
Figure 12 (PDF p. 11).
Figure 13 (PDF p. 12).

Figure 14 (PDF p. 13).
(a)
(b)
Figure 15 (PDF p. 14).
Figure 16 (PDF p. 15).
Figure 17 (PDF p. 16).

(a)

(b)
Figure 18 (PDF p. 17).

Figure 19 (PDF p. 18).
Superconducting Dirac semimetals: - and -
Alex Taekyung Lee; Kyungwha Park; In-Ho Lee
Supplemental Material Superconducting Dirac semimetals: - and -
Atomic structure of -
Figure S1. (a) Top view and (b) side view of the atomic structure of -. - can be obtained by removing Na atoms from -NaSi. (c) First Brillouin zone (BZ) of -. , M, A, and L points are time reversal invariant momentum points, which are related to topological invariant calculation especially for crystals with inversion symmetry. (space group no. 175) is belong to the hexagonal Bravais lattice crystal system. The space group no. 175 has inversion symmetry.
| Material | pressure (GPa) | (Å) | (Å) | Atom | |||
| - | 0 | 6.813 | 2.501 | Si (6j) | 0.154 | 0.733 | 0 |
| - | 15 | 6.542 | 2.402 | Si (6j) | 0.145 | 0.725 | 0 |
| -NaSi | 0 | 6.755 | 2.444 | Na (1b) | 0 | 0 | 0.5 |
| Si (6j) | 0.145 | 0.710 | 0 | ||||
| -NaSi | 15 | 6.536 | 2.377 | Na (1b) | 0 | 0 | 0.5 |
| Si (6j) | 0.144 | 0.713 | 0 |
Table S1. Optimized lattice parameters and atomic coordinates of - and -NaSi phases at external pressure of 0 GPa and 15 GPa, obtained from first-principles calculations using density-functional theory (DFT). Both -NaSi and - phases are hexagonal with , and .
Band structures of -NaSi and -Si

Figure S2. Electronic band structure obtained from first-principles calculations using DFT. First-principles band structures of -NaSi at pressure (a) 0 GPa and (b) 15 GPa, including spin-orbit coupling. According to the relative weight of the Si character and the Si character, the band color was encoded as shown in the legend. The band structure calculation at pressure 15 GPa shows the band structure just before band inversion occurs. The situation of band inversion that occurs with changes in pressure can be seen, especially in the L-H section.
Figure S3. Electronic band structure obtained from first-principles calculations using DFT. In the presence of SOC, band structures of -Si at (a) 0 GPa and (b) 15 GPa and their projections on the Si character are shown. According to the relative weight of in-plane Si character () and out-of-plane Si character () the band colors were encoded as shown in the legend. The situation of band inversion that occurs with changes in external pressure ( GPa) can be seen, especially in the L-H-A interval.
Z2 topological invariant calculations
The topological phase of band structure can be classified by calculating topological invariant. The hybrid Wannier charge center flow technique, which is equivalent to the Wilson loop method. To investigate the topological nature of the electronic states, we obtained maximally localized Wannier functions using the WANNIER90 code [? ] and performed hybrid Wannier charge center flow calculations. With this Wannier model, the symmetry properties and topological invariant could be directly computed. topological invariants were calculated by observing hybrid Wannier charge center flows. The topological surface states were calculated using the Green’s function approach, as implemented in the WannierTools code [? ].
Space group no. 175 (, hexagonal) has inversion symmetry. Because of the existence of inversion symmetry in both - and -, the Fu-Kane parity criterion can be used to easily calculate the topological invariants. We were able to obtain the same topological invariants for both - and - by using two different methods, observing hybrid Wannier charge center flows and investigating Fu-Kane parity criterion. is the strong topological index. () are weak topological indices. Table S3 show the example for - at 0 GPa. From the hybrid Wannier charge center flow calculations, we calculated (). is obtained by
| -plane | ||
| - plane | ||
| - plane | ||
| - plane | ||
| - plane | ||
| - plane | ||
| - plane |
Table S3. and () for - obtained from hybrid Wannier charge center flow calculations, and corresponding points and -planes.
| k-point | - 0 GPa | - 15 GPa | - 0 GPa | - 15 GPa | |
| A | |||||
Table S2. Parities at time-reversal invariant momentum points of - and - at 0 GPa and 15 GPa.
and thus we have . Therefore, - at 0 GPa has a strong topological nature. Similarly, we find that - at 0 GPa has a strong topological nature.
| band # | irrep | |||
| 1 | ||||
| 2 | ||||
| 3 | ||||
| 4 | ||||
| 5 | ||||
| 6 | ||||
| 7 | ||||
| 8 | ||||
| 9 | ||||
| 10 | ||||
| 11 | ||||
| 12 | ||||
| 13 |
Table S4. Irreducible representation (irrep), point group symmetries, and their eigenvalues of bands at L point. is two-fold rotation operator with respect to the -axis. and are two-fold inversion and horizontal mirror plane operators, respectively. The highest occupied band is 12th band.
Figure S4. First-principles bands (solid line) and Wannier bands (red filled circles) of - at (a) 0 GPa and (b) 15 GPa. In the eV energy scale, there are no visible differences between two different bands. Spin-orbit coupling is included.
Si-Si bond lengths in the slab geometry

Figure S5. Si-Si bond lengths, formed in the direction also denoted by out-of-plane (oop bond lengths) from the hexagonal plane, in the relaxed 40-layer -Si slab are shown. Position of the surface along the out-of-plane, is set to be zero. Dashed line represents the Si-Si bond length in the unrelaxed slab.
Fermi surfaces for P6/m-Si6 and P6/m-NaSi6

Figure S6. Electronic band structure obtained from first-principles calculations using DFT. Fermi surfaces of (a) -Si at 0 GPa, (b) -Si at 15 GPa, (c) -NaSi at 0 GPa, and (d) -NaSi at 150 GPa are shown, respectively. The Fermi surface is the surface of constant energy in the first BZ which separates occupied from unoccupied electron states at zero temperature.
References
- [1]W. Buckel and J. Wittig, Physics Letters 17, 187 (1965).
- [2]G. Stepanov, T. Valyanskaya, and E. Yakovlev, Sov. Phys.-Solid State (Engl. Transl.);(United States) 22 (1980).
- [3]G. Martinez, J. M. Mignot, G. Chouteau, K. J. Chang, M. M. Dacorogna, and M. L. Cohen, Physica Scripta T13, 226 (1986).
- [4]D. Erskine, P. Y. Yu, K. J. Chang, and M. L. Cohen, Phys. Rev. Lett. 57, 2741 (1986).
- [5]K. J. Chang, M. M. Dacorogna, M. L. Cohen, J. M. Mignot, G. Chouteau, and G. Martinez, Phys. Rev. Lett. 54, 2375 (1985).
- [6]H. Kawaiji, H.-O. Horie, S. Yamanaka, and M. Ishikawa, Phys. Rev. Lett. 74, 1427 (1995).
- [7]D. Connétable, V. Timoshevskii, B. Masenelli, J. Beille, J. Marcus, B. Barbara, A. M. Saitta, G.-M. Rignanese, P. Mélinon, S. Yamanaka, and X. Blase, Phys. Rev. Lett. 91, 247001 (2003).
- [8]E. Bustarret, C. Marcenat, P. Achatz, J. Kačmarčík, F. Lévy, A. Huxley, L. Orégue, E. Bourgeois, X. Blase, D. Débarre, and J. Boulmer, Nature 444, 465 (2006).
- [9]D. Y. Kim, S. Stefanoski, O. O. Kurakevych, and T. A. Strobel, Nature Materials 14, 169 (2015).
- [10]H.-J. Sung, W. H. Han, I.-H. Lee, and K. J. Chang, Phys. Rev. Lett. 120, 157001 (2018).
- [11]C.-K. Chiu, J. C. Y. Teo, A. P. Schnyder, and S. Ryu, Rev. Mod. Phys. 88, 035005 (2016).
- [12]B.-J. Yang and N. Nagaosa, Nature Communications 5, 4898 (2014).
- [13]L. Fu and C. L. Kane, Phys. Rev. B 76, 045302 (2007).
- [15]Z. Wang, H. Weng, Q. Wu, X. Dai, and Z. Fang, Phys. Rev. B 88, 125427 (2013).
- [16]S. M. Young, S. Zaheer, J. C. Y. Teo, C. L. Kane, E. J. Mele, and A. M. Rappe, Phys. Rev. Lett. 108, 140405 (2012).
- [17]H. Su, X. Shi, W. Xia, H. Wang, X. Hanli, Z. Yu, X. Wang, Z. Zou, N. Yu, W. Zhao, G. Xu, and Y. Guo, Phys. Rev. B 101, 205138 (2020).
- [18]T. Deng, C. Chen, H. Su, J. He, A. Liang, S. Cui, H. Yang, C. Wang, K. Huang, C. Jozwiak, A. Bostwick, E. Rotenberg, D. Lu, M. Hashimoto, L. Yang, Z. Liu, Y. Guo, G. Xu, Z. Liu, and Y. Chen, Phys. Rev. B 102, 045106 (2020).
- [19]Z. Liu, H. Xin, L. Fu, Y. Liu, T. Song, X. Cui, G. Zhao, and J. Zhao, The Journal of Physical Chemistry Letters 10, 244 (2019), https://doi.org/10.1021/acs.jpclett.8b03345.
- [20]L. Aggarwal, A. Gaurav, G. S. Thakur, Z. Haque, A. K. Ganguli, and G. Sheet, Nature Materials 15, 32 (2016).
- [21]H. Wang, H. Wang, H. Liu, H. Lu, W. Yang, S. Jia, X.-J. Liu, X. C. Xie, J. Wei, and J. Wang, Nature Materials 15, 38 (2016).
- [22]C. Huang, B. T. Zhou, H. Zhang, B. Yang, R. Liu, H. Wang, Y. Wan, K. Huang, Z. Liao, E. Zhang, S. Liu, Q. Deng, Y. Chen, X. Han, J. Zou, X. Lin, Z. Han, Y. Wang, K. T. Law, and F. Xiu, Nature Communications 10, 2217 (2019).
- [23]L. M. Schoop, S. L. Xie, R. Chen, Q. D. Gibson, S. H. Lapidus, I. Kimchi, M. Hirschberger, N. Haldolaarachchige, M. N. Ali, C. A. Belvin, T. Liang, J. B. Neaton, N. P. Ong, A. Vishwanath, and R. J. Cava, Phys. Rev. B 91, 214517 (2015).
- [24]Y. Xing, H. Wang, C.-K. Li, X. Zhang, J. Liu, Y. Zhang, J. Luo, Z. Wang, Y. Wang, L. Ling, M. Tian, S. Jia, J. Feng, X.-J. Liu, J. Wei, and J. Wang, npj Quantum Materials 1, 16005 (2016).
- [25]J. Song, S. Kim, Y. Kim, H. Fu, J. Koo, Z. Wang, G. Lee, J. Lee, S. H. Oh, J. Bang, T. Matsushita, N. Wada, H. Ikegami, J. D. Denlinger, Y. H. Lee, B. Yan, Y. Kim, and S. W. Kim, Phys. Rev. X 11, 021065 (2021).
- [26]S. Teknowijoyo, N. H. Jo, M. S. Scheurer, M. A. Tanatar, K. Cho, S. L. Bud’ko, P. P. Orth, P. C. Canfield, and R. Prozorov, Phys. Rev. B 98, 024508 (2018).
- [27]Amit and Y. Singh, Phys. Rev. B 97, 054515 (2018).
- [28]K. Huang, A. Y. Luo, C. Chen, G. N. Zhang, X. L. Liu, Y. W. Li, F. Wu, S. T. Cui, Z. Sun, C. Jozwiak, A. Bostwick, E. Rotenberg, H. F. Yang, L. X. Yang, G. Xu, Y. F. Guo, Z. K. Liu, and Y. L. Chen, Phys. Rev. B 103, 155148 (2021).
- [29]S. Kobayashi and M. Sato, Caya, and N. P. Ong, Science 350, 413 (2015), https://www.science.org/doi/pdf/10.1126/science.aac6089.
- [30]T. Hashimoto, S. Kobayashi, Y. Tanaka, and M. Sato, Phys. Rev. B 94, 014510 (2016).
- [31]M. Sato and Y. Ando, Reports on Progress in Physics 80, 076501 (2017).
- [32]J. Alicea, Rep. Prog. Phys. 75, 076501 (2012).
- [33]J. Kitaev, Annals of Physics 303, 2 (2003).
- [34]S. D. Sarma, M. Freedman, and C. Nayak, npj Quantum Information 1, 15001 (2015).
- [35]A. Kreisel, P. J. Hirschfeld, and B. M. Andersen, Symmetry 12 (2020), 10.3390/sym12091402.
- [36]L. Kong, L. Cao, S. Zhu, M. Papaj, G. Dai, G. Li, P. Fan, W. Liu, F. Yang, X. Wang, S. Du, C. Jin, L. Fu, H.-J. Gao, and H. Ding, Nature Communications 12, 4146 (2021).
- [37]P. E. Blöchl, Phys. Rev. B 50, 17953 (1994).
- [38]J. P. Perdew, K. Burke, and M. Ernzerhof, Phys. Rev. Lett. 77, 3865 (1996).
- [39]G. Kresse and D. Joubert, Phys. Rev. B 59, 1758 (1999).
- [40]A. Soluyanov and D. Vanderbilt, Phys. Rev. B 83, 035108 (2011).
- [41]Q. Wu, S. Zhang, H.-F. Song, M. Troyer, and A. A. Soluyanov, Computer Physics Communications 224, 405 (2018).
- [42]N. Marzari, A. A. Mostofi, J. R. Yates, I. Souza, and D. Vanderbilt, Rev. Mod. Phys. 84, 1419 (2012).
- [43]P. Giannozzi, O. Andreussi, T. Brumme, O. Bunau, M. B. Nardelli, M. Calandra, R. Car, C. Cavazzoni, D. Ceresoli, M. Cococcioni, N. Colonna, I. Carnimeo, A. D. Corso, S. de Gironcoli, P. Delugas, R. A. D. Jr, A. Ferretti, A. Floris, G. Fratesi, G. Fugallo, R. Gebauer, U. Gerstmann, F. Giustino, T. Gorni, J. Jia, M. Kawamura, H.-Y. Ko, A. Kokalj, E. Küçükbenli, M. Lazzeri, M. Marsili, N. Marzari, F. Mauri, N. L. Nguyen, H.-V. Nguyen, A. O. de-la Roza, L. Paulatto, S. Poncé, D. Rocca, R. Sabatini, B. Santa, M. Schlipf, A. P. Seitsonen, A. Smogunov, I. Timrov, T. Thonhauser, P. Umari, N. Vast, X. Wu, and S. Baroni, Journal of Physics: Condensed Matter 29, 465901 (2017).
- [44]Z. Wang, Y. Sun, X.-Q. Chen, C. Franchini, G. Xu, H. Weng, X. Dai, and Z. Fang, Phys. Rev. B 85, 195320 (2012).
- [45]J. W. Villanova, E. Barnes, and K. Park, Nano Letters 17, 963 (2017).
- [46]S. M. Young, S. Chowdhury, E. J. Walter, E. J. Mele, C. L. Kane, and A. M. Rappe, Phys. Rev. B 84, 085106 (2011).
- [47]M. S. Rahamy, B.-J. Yang, R. Arita, and N. Nagaosa, Nature Communications 3, 679 (2012).
- [48]A. Bansil, H. Lin, and T. Das, Rev. Mod. Phys. 88, 021004 (2016).
- [49]T. Ideue, M. Hu, H. Taiko, T. Takahashi, M. Murase, T. Miyake, S. Murakami, T. Sasagawa, and Y. Iwasa, Proceedings of the National Academy of Sciences 116, 25530 (2019), https://www.pnas.org/content/116/51/25530.full.pdf.
- [50]P. J. W. Moll, N. L. Nair, T. Helm, A. C. Potter, I. Kimchi, A. Vishwanath, and J. G. Analytis, Nature 535, 266 (2016).
- [51]A. A. Burkov, M. D. Hook, and L. Balents, Phys. Rev. B 84, 235126 (2011).
- [52]J. Xiong, S. K. Kushwaha, T. Liang, J. W. Krizan, M. Hirschberger, W. Wang, R. J. Cava, and N. P. Ong, Science 350, 413 (2015), https://www.science.org/doi/pdf/10.1126/science.aac6089.
- [53]Z. Jia, C. Li, X. Li, J. Shi, Z. Liao, D. Yu, and X. Wu, Nature Communications 7, 13013 (2016).
- [54]A. A. Burkov, Phys. Rev. B 96, 041110 (2017).
- [55]M. Wu, G. Zheng, W. Chu, Y. Liu, W. Gao, H. Zhang, J. Lu, Y. Han, J. Zhou, W. Ning, and M. Tian, Phys. Rev. B 98, 161110 (2018).
- [56]G. Pizzi, V. Vitale, R. Arita, S. Blügel, F. Freimuth, G. Géranton, M. Gibertini, D. Gresch, C. Johnson, T. Koretsune, J. Ibañez-Azpiroz, H. Lee, J.-M. Lihm, D. Marchand, A. Marrazzo, Y. Mokrousov, J. I. Mustafa, Y. Nohara, Y. Nomura, L. Paulatto, S. Poncé, T. Ponweiser, J. Qiao, F. Thöle, S. S. Tsirkin, M. Wierzbowska, N. Marzari, D. Vanderbilt, I. Souza, A. A. Mostofi, and J. R. Yates, Journal of Physics: Condensed Matter 32, 165902 (2020).













