A numerical method for designing topological superconductivity induced by s-wave pairing

Abstract

Topological superconductors, as one of the most important research directions at present, have attracted much attention because of their potential to realize topological quantum computation. However, a universal computational tool based on first-principle calculations for topological superconductivity was not yet fully developed, and eventually significant challenges in predicting topological superconducting materials. In this paper, we develop a numerical method to characterize the superconducting band spectrum and superconducting topological invariants of two-dimensional (2D) slab system from first-principles calculations and implemented in an open-source software WannierTools. The program is performed on many typical models and materials, such as Fu-Kane model, Dirac cone model of topological insulator, Bi₂Se₃ slab and SnTe slab, to identify its validity. We hope that it would accelerate the discovery of the topological superconductor candidates.

I Introduction

The concept of topological superconductors was formalized in 2000 in a 2D model by Read and Green [1] and in a one-dimensional (1D) model by Kitaev [2]. Both models consider spinless, time-reversal-breaking p-wave pairing for the superconducting state, which gives rise to a non-Abelian Majorana zero mode in the vortex core in the 2D case or at the edge in the 1D case.

As one can see in this history, the interest in topological superconductors are strongly tied to the Majorana fermions that are exotic in that particles are their own antiparticles [3, 4, 5, 6]. It has long been known that Majorana zero mode can be used as qubit in topological quantum computation [7, 8]. More recently, chiral Majorana states (MSs) have also been shown to be usable in topological quantum computation [9]. Therefore, topological superconductors have attracted increasing attention from the academic community [10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22]. The initial system to realize TSC involves the p-wave pairing superconductor (SC) [23, 24, 25, 26, 27]. However, intrinsic p-wave SCs are very rare and MSs have not been clearly distinguished in these systems. In the past decade, many proposals are raised to induce TSC at the surface/interface of the topological materials [28, 29, 30, 31, 32, 33], semiconductors with Rashba spin-orbit coupling (Rashba-SOC) [34, 35, 36, 37, 38, 39], magnetic atom chains [40, 41, 42, 43, 44], or obstructed atomic insulator [45] by the proximity effect with the s-wave SC. The other route is to achieve TSCs in superconducting topological metals that host both topological electronic structures at the Fermi level and superconductivity in one compound, typical of such materials are iron-based superconductors [46, 47, 48, 49, 50, 51, 52].

Fig. 1: A flow chart of our numerical method to find topological superconductor candidates.

However, experimental confirmation of topological superconductors remains scarce compared to theoretical predictions. A major reason for this is that theoretical predictions of topological superconductors are almost always made based on effective models. This makes it difficult to take into account the complex electronic structures and interactions in real systems in theoretical analyses. In this paper, we develop a numerical method to characterize the superconducting band spectrum and superconducting topological invariants of two-dimensional (2D) slab system from first-principles calculations as implemented in an open-source software WannierTools [53]. This program will greatly increase the degree of conformity between theoretical predictions and experimental confirmation.

The flow of the program is shown in Fig. 1. First, the Wannier function, which accurately describes the electronic structure of the bulk of the topological material, is constructed based on the first-nature principle calculations. Further, the superconducting pair and the applied magnetic field are introduced at the interface of the 2D film to simulate the superconducting proximity effect and magnetic proximity effect, and the corresponding 2D Bogoliubov-de Gennes (BdG) Hamiltonian is constructed. By using this program, a physical platform for realizing 2D chiral topological superconducting phases, i.e., superconductor/topological insulator thin film/ferromagnetic insulator heterojunction (hereafter referred to as SC/TI/FMI), is investigated in this paper.

II Methods

II.1 BdG Hamiltonian in thin film system

A common and convenient method to construct tight-binding (TB) Hamiltonians for real materials is the maximum localized Wannier functions (MLWF)  [54] as implemented in the Wannier90 software package [55], and interfaces are available to a number of first-principles calculation packages, such as VASP, Wien2k, and QE. Wien2k and QE, among others. Therefore, the MLWF Hamiltonian for real materials can be obtained automatically from first-principles calculations. The real-space hop** parameters of electrons in real materials can be obtained from first-principles calculations [54]:

$$H_{mn}(\bm{R})=\langle\mathcal{W}_{m\bm{0}}|\hat{H}_{KS}|\mathcal{W}_{n\bm{R}}\rangle$$ (1)

where $\hat{H}_{KS}$ is the Kohn-Sham (KS) Hamiltonian from first principles calculations, $H_{mn}(R)$ is the matrix element of the projection of the Kohn-Sham (KS) Hamiltonian onto the MLWF Wannier basis, and $R$ is the lattice vector in the crystal. In the atomic position norm, the Fourier transform gives the bulk Hamiltonian:

$$H_{mn}(\bm{k})=\sum_{\bm{R}}H_{mn}(\bm{R})e^{-i\bm{kR}}$$ (2)

In order to calculate the superconducting near-neighbor effect or magnetic near-neighbor effect of SC/TI/FMI, it is necessary to further adopt open boundary conditions to simulate the surface or interface of the material. For convenience, the three-dimensional vector $\bm{R}$ is divided into the longitudinal vector $\bm{R}_{\perp}$ and the transverse vector $\bm{R}_{\parallel}$. where $\bm{R}_{\perp}$ is perpendicular to the material surface and $\bm{R}_{\parallel}$ is parallel to the material surface. After opening the boundary in the $\bm{R}_{\perp}$ direction, a two-dimensional film system is formed, where only the in-plane $\bm{R}_{\parallel}$ direction is periodic. In general, the experimentally easy dissociation plane conflicts with the basis vectors of the protocells calculated by first principles, and therefore the basis vectors $\bm{R}^{{}^{\prime}}_{1,2,3}$ need to be re-selected:

$$\begin{split}\bm{R}^{{}^{\prime}}_{1}=U_{11}\bm{R}_{1}+U_{12}\bm{R}_{2}+U_{13}\bm{R}_{3}\\ \bm{R}^{{}^{\prime}}_{2}=U_{21}\bm{R}_{1}+U_{22}\bm{R}_{2}+U_{23}\bm{R}_{3}\\ \bm{R}^{{}^{\prime}}_{3}=U_{31}\bm{R}_{1}+U_{32}\bm{R}_{2}+U_{33}\bm{R}_{3}\\ \end{split}$$ (3)

where $\bm{R}^{{}^{\prime}}_{1}$ and $\bm{R}^{{}^{\prime}}_{2}$ lattice vectors form $\bm{R}_{\parallel}$ and $\bm{R}^{{}^{\prime}}_{3}$ corresponds to the $\bm{R}_{\perp}$ direction. With the basis-vector transformation matrix $U$, it is possible to generalize the physical properties of any dissociation surface.

The Hamiltonian of the 2D thin film system can be obtained by taking the open boundary condition in the $\bm{R}^{{}^{\prime}}_{3}$ direction and Fourier transforming it:

$$H_{mn}^{slab}(\bm{k}_{\parallel})=\sum_{\parallel\bm{R}^{{}^{\prime}}\parallel}H_{mn}(\bm{R}^{{}^{\prime}})e^{-i\bm{k}_{\parallel}\bm{R}^{{}^{\prime}}}$$ (4)

where the $\parallel\bm{R}^{{}^{\prime}}\parallel$ restriction sums only in the direction of the basis vectors $\bm{R}^{{}^{\prime}}_{1}$ and $\bm{R}^{{}^{\prime}}_{2}$, and different $\bm{R}^{{}^{\prime}}_{3}$ corresponds to different film layers, constituting the longitudinal dimension of the Hamiltonian of the 2D film. We label the layer index along $\bm{R}^{{}^{\prime}}_{3}$ as i, j. When the system has $N_{s}$ layers, the matrix form of the Hamiltonian of the 2D film can be written as:

$$H_{mn}^{slab}(\bm{k}_{\parallel})=\left[\begin{array}[]{cccc}H_{mn}^{11}(\bm{k}_{\parallel})&H_{mn}^{12}(\bm{k}_{\parallel})&\cdots&H_{mn}^{1N_{s}}(\bm{k}_{\parallel})\\ H_{mn}^{21}(\bm{k}_{\parallel})&H_{mn}^{22}(\bm{k}_{\parallel})&\cdots&H_{mn}^{2N_{s}}(\bm{k}_{\parallel})\\ \vdots&\vdots&\ddots&\vdots\\ H_{mn}^{N_{s}1}(\bm{k}_{\parallel})&H_{mn}^{N_{s}2}(\bm{k}_{\parallel})&\cdots&H_{mn}^{N_{s}N_{s}}(\bm{k}_{\parallel})\end{array}\right]$$ (5)

where diagonal matrix elements represent intra-layer (intra-layer) Hamiltonians and non-diagonal matrix elements represent inter-layer (inter-layer) Hamiltonians. The specific form of the matrix element can be written as:

$$H_{mn}^{ij}(\bm{k}_{\parallel})=\sum_{\bm{R}^{{}^{\prime}}=\{\bm{R}^{{}^{\prime}}_{1},\bm{R}^{{}^{\prime}}_{2},(i-j)\bm{R}^{{}^{\prime}}_{3}\}}H_{mn}(\bm{R}^{{}^{\prime}})e^{-i\bm{k}_{\parallel}\bm{R}^{{}^{\prime}}}$$ (6)

When atomic-scale exchange coupling due to magnetic proximity effects is present, the Hamiltonian of the surface layer (i = j = Ns) becomes:

$$H_{mn}^{N_{s}N_{s}}(\bm{k}_{\parallel},\bm{B})=H_{mn}^{N_{s}N_{s}}(\bm{k}_{\parallel})+\bm{B}\cdot\bm{\sigma}I_{orbital}$$ (7)

where $\bm{B}$ is the Zeeman field due to exchange coupling and $\bm{\sigma}$ is the Pauli matrix in spin space, and I is the unitary matrix in Wannier orbital space.

Finally, the BdG Hamiltonian of a superconductor-topological insulator film-ferromagnetic insulator heterojunction can be obtained by inducing the superconducting pairing on the top surface into the whole system, taking into account the exchange coupling. In the Numbu representation, this BdG Hamiltonian can be written as:

$$H_{BdG}^{slab}(k,\bm{B})=\left[\begin{array}[]{cc}H_{mn}^{slab}(\bm{k}_{\parallel},\bm{B})-\mu&\Delta\\ \Delta^{\dagger}&-H_{mn}^{slab}(\bm{k}_{\parallel},\bm{B})+\mu\end{array}\right]$$ (8)

where $\mu$ is the chemical potential of the system; since this paper focuses on the proximity effects of s-wave superconductors, $\Delta$ here denotes the intra-orbital s-wave pairing strength.

II.2 Superconducting topological invariant

Theoretically, the topological classification of superconductors has been perfected [56]. As shown in Table 1, superconductors can be categorized into class D and class D$\rm III$ based on the presence or absence of time-reversal symmetry. As can be seen from the table, for 2D systems, when the time-reversal symmetry is broken, the system belongs to the Z classification, and the corresponding topological invariant is the Chern number [57]; when the time-reversal symmetry is exists, the system belongs to the $Z_{2}$ classification, and the corresponding topological invariant is the $Z_{2}$ topological number defined by Kane-Mele [58]. In addition, Qi et al. demonstrated that in a superconducting system with both time-reversal and inversion symmetry, s-wave superconducting pairing leads the system into a topologically trivial phase [59].

Table 1: Time-reversal-breaking (TRB) and time-reversal-invariant (TRI) Topological periodic table of superconductors. In the Altland-Zirnbauer classification [56], they belong to categories D and D$\rm III$, respectively.

AZ class	TRS	PHS	1D	2D	3D
TRB SCs (class D)	—	+1	$Z_{2}$	$Z$	0
TRI SCs (class D$\rm III$)	-1	+1	$Z_{2}$	$Z_{2}$	$Z$

There are many methods for calculating topological invariants [60, 61, 62, 63], among which the Wannier charge centers (Wilson loop) method [60, 61], which can calculate the $Z_{2}$ topological number of a system as well as the Chern number. In the following, the topological invariants of BdG Hamiltonian are calculated using the Wannier charge centers (Wilson loop) method in the 2D slab system.

In WannierTools, we take the algorithm presented in Refs [60, 64]. The hybrid Wannier functions [65] are defined as:

$$|nk_{x}l_{y}\rangle=\frac{1}{2\pi}\int^{2\pi}_{0}dk_{y}e^{-ik_{y}l_{y}}|\psi_{n\bm{k}}^{BdG}\rangle$$ (9)

where $|\psi_{n\bm{k}}^{BdG}\rangle$ is the Bloch wave function of BdG Hamiltonian. The hybrid Wannier centers are defined as:

$$\begin{split}\bar{y}_{n}(k_{x})&=\langle nk_{x}0|y|nk_{x}0\rangle\\ &=\frac{i}{2\pi}\int^{\pi}_{\pi}dk_{y}\langle u_{n,k_{x},k_{y}}^{BdG}|\partial_{k_{y}}|u_{n,k_{x},k_{y}}^{BdG}\rangle\end{split}$$ (10)

where $|u_{n,k_{x},k_{y}}^{BdG}\rangle$ is the periodic part of Bloch function $|\psi_{n\bm{k}}^{BdG}\rangle$. In practice, the integration over $k_{y}$ is transformed by a summation over the discretized $k_{y}$. Eq. 10 can be reformulated using the discretized Berry phase formula [66]:

$$\begin{split}\bar{y}_{n}(k_{x})=\frac{1}{2\pi}\mathrm{Im}\ln\prod_{j}M_{mn}^{(j)}\end{split}$$ (11)

where the gauge-dependent overlap matrix $M_{mn}^{(j)}=\langle u_{m,k_{x},k_{y_{i}}}^{BdG}|\partial_{k_{y}}|u_{n,k_{x},k_{y_{i+1}}}^{BdG}\rangle$ is introduced. However, the summation of the hybrid Wannier centers $\bar{y}_{n}(k_{x})$ over $k_{x}$ is gauge invariant [67]. As shown in Ref. [66, 68], there is another way to obtain $\bar{y}_{n}(k_{x})$. Firstly, we get the ”unitary part” $\tilde{M^{(j)}}$ of each overlap matrix $M_{mn}^{(j)}$ by carrying out the single value decomposition $M=V\Sigma W^{\dagger}$ , where $V$ and $W$ are unitary and $\Sigma$ is real-positive and diagonal. Then we set $\tilde{M^{(j)}}=VW^{\dagger}$. The eigenvalues $\lambda_{n}$ of matrix $\Lambda=\prod_{j}\tilde{M^{(j)}}$ are all of unit modulus. The hybrid Wannier centers are defined with the phases of $\lambda_{n}$ eventually

$$\begin{split}\bar{y}_{n}(k_{x})=\frac{1}{2\pi}\mathrm{Im}\ln\lambda_{n}\end{split}$$ (12)

We can get the topological properties of 2D system from the evolution of $\bar{y}_{n}(k_{x})$ along a $k_{x}$ string. The details of such classification of Wannier charge centers or Wilson loop are discussed in Refs. [60, 61, 68]. More information can be found in Ref. [64].

In our program, the positions of the atoms in the slab system are in Cartesian coordinates, and the positions of the particles and holes are assumed to correspond one-to-one. In addition, the number of occupied BdG bands is always half the total number of BdG bands, so we always compute the wilson loop for half the total number of BdG bands.

II.3 Proposals to induce TSC by s-wave SC

II.3.1 Fu-kane’s proposal

Due to the superconducting proximity effect, the Bogoliubov-de Gennes (BdG) Hamiltonian at the interface of a 3D strong topological insulator and an s-wave superconductor can be written as [28]:

$$H_{interface}(k)=k_{x}\tau_{0}\sigma_{x}+k_{y}\tau_{z}\sigma_{y}-\mu\tau_{z}\sigma_{0}+\Delta\tau_{y}\sigma_{y}$$ (13)

This Hamiltonian is written in the Nambu basis $(c_{\uparrow,k}^{\dagger},c_{\downarrow,k}^{\dagger},c_{\uparrow,-k},c_{\downarrow,-k})$, where $\tau$ and $\sigma$ represent the Pauli matrix of particle-hole space and spin space respectively, $\mu$ represents the chemical potential and $\Delta$ represents the superconducting pairing strength. Take the gauge transformation $a_{k}=c_{\uparrow,k}+e^{i\theta_{k}}c_{\downarrow,k}$ of this Hamiltonian. Where $\theta_{k}$ is defined by $|k|(cos\theta_{k},sin\theta_{k})=(k_{x},k_{y})$, a effective p+ip-wave pairing can be obtained:

$$H_{eff}^{p_{x}+ipy}(k)=\left[\begin{array}[]{cccc}-|k|-\mu&0&0&-\Delta e^{i\theta_{k}}\\ 0&|k|-\mu&\Delta e^{i\theta_{k}}&0\\ 0&\Delta e^{-i\theta_{k}}&-|k|+\mu&0\\ -\Delta e^{-i\theta_{k}}&0&0&|k|+\mu\end{array}\right]$$ (14)

The physical meaning of this process is to treat $\Delta$ as a perturbation and project the Hamiltonian onto an eigenstate without the perturbation. However, an important difference between this Hamiltonian and p+ip-wave superconductors is that the Hamiltonian still has time-inversion symmetry, so the chiral boundary state of p+ip-wave superconductivity is forbidden in this system. This can also be seen from the dispersion of this Hamiltonian, which is: $E=\pm\sqrt{(\pm|k|-\mu)^{2}+\Delta^{2}}$. It always has a gap, so there is no topology in the system. When the Zeman field $M_{z}\tau_{z}\sigma_{z}$ is added to the system to break time-inversion symmetry (for simplicity, let $\mu=0$), the dispersion becomes $E=\pm\sqrt{|k|^{2}+(M_{z}\pm\Delta)^{2}}$. It can be seen that topological phases can exist in this systems that belong to class D, and this topological phase can be characterized by the Chern number [69]. The condition for the phase transition is that $|B_{z}|>\sqrt{\mu^{2}+\Delta^{2}}$.

In the Fu-kane’s proposal, we assume that the superconducting proximity effect can introduce equal superconducting pairings throughout the film system. In addition, the Fu-kane’s proposal can also be extended to 2D systems with Rashba Spin-orbit coupling (SOC) [70].

II.3.2 Nagaosa’s proposal

Consider a superconductor-TI film-ferromagnet heterostructure as shown in the Fig.2 [32]:

Fig. 2: Nagaosa’s proposal

Since the low-energy physics of this system is determined by the two Dirac surface states of TI, the BdG Hamiltonian of the system under the Nambu basis $(c_{t\uparrow,k}^{\dagger},c_{t\downarrow,k}^{\dagger},c_{b\uparrow,k}^{\dagger},c_{b\downarrow,k}^{\dagger},c_{t\uparrow,-k},c_{t\downarrow,-k},c_{t\uparrow,-k},c_{t\downarrow,-k})$ is:

$$\begin{split}H_{twosurface}(k)=&v(k_{x}\sigma_{y}\tau_{z}-k_{y}\sigma_{x}\tau_{0})s_{z}+t_{c}\sigma_{0}s_{x}\tau_{z}\\ &+M_{z}\sigma_{z}\tau_{z}\frac{s_{z}-s_{0}}{2}+\Delta\sigma_{y}\tau_{y}\frac{s_{z}+s_{0}}{2}\\ &-\sigma_{0}(\mu s_{0}+\delta Es_{z})\tau_{z}\end{split}$$ (15)

$\tau$, $\sigma$, and $s$ represent the Pauli matrix of particle-hole space, spin space, top and bottom space, respectively. $M_{z}$ represents the Zeeman field, $\Delta$ represents the s-wave superconducting pairing strength, $t_{c}$ represents the coupling between the top and bottom surfaces, $\mu$ represents the chemical potential, and $\delta E$ represents the energy difference between the upper and lower surfaces. It can be seen that this Hamiltonian describes the interface between two Dirac cones with different mass terms ($\Delta$ on top and $M_{z}$ on bottom), which is a chiral topological superconducting state according to the criteria in [71].

Let’s briefly analyze the topological properties of this system. After ignoring $t_{c}$and $\delta E$, the Hamiltonian $H_{twosurface}(k)$ becomes block diagonalized, and it is pointed out in the Fu-kane’s proposal that the top surface without Zeeman field is Topological trivial. In this case, the topological properties of the system are completely determined by the bottom surface, which is:

$$H_{bottom}(k)=k_{x}\tau_{0}\sigma_{x}+k_{y}\tau_{z}\sigma_{y}-\mu\tau_{z}\sigma_{0}+M_{z}\tau_{z}\sigma_{z}$$ (16)

This is also a block diagonal Hamiltonian corresponding to two Dirac Hamiltonians with opposite helicity and opposite masses. That is, if $M_{z}$ is not 0, the system must be topological (one of the two Dirac Hamiltonians must be topological), and the topological property is determined by the mass term (that is, $M_{z}$). In addition, $t_{c}$ and $\delta E$ can be regarded as perturbations to the Hamiltonian $H_{bottom}(k)$, which are small in general topological insulators. As long as the perturbation is insignificant relative to the mass term $M_{z}$, it does not change the topological properties of the system.

In the Nagaosa’s proposal, we assume that the superconducting proximity effect introduces superconducting pairings only in the TI surface.

III Examples

III.1 TI model

Fig. 3: a,b,c: Slab band structures, BdG band spectrum, and the wilson spectrum of the BdG Hamiltonian in Fu-kane’s proposal of TI model. d,e,f: Slab band structures, BdG band spectrum, and the wilson loop spectrum of the BdG Hamiltonian in Nagaosa’s proposal of TI model.

Fig. 4: a: Bulk band structures of TI model. b: Band structures of 10 layers slab of TI model.

Before calculating the real material, we numerically calculate and analyze the BdG Hamiltonian quantities for the TI tight-binding model. These numerical results are useful to establish a clear physical picture and also to verify the correctness of the program.

The model is built on a simple cubic lattice with two orbitals per primitive cell:

$$H_{\mathrm{TI}}(k)=m(k)\tau_{z}\sigma_{0}+\tau_{x}\bm{d(k)}\cdot\bm{\sigma}$$ (17)

where $\tau$ ($\sigma$) are Pauli matrices in the orbital (spin) space, $m(k)=M+m_{0}\sum_{i}\cos(k_{i}),d(k)^{i}=2t\sin{k_{i}},i=x,y,z$, and $t$, $m_{0}$, and $M$ are parameters of the model. The model is in the strong TI phase if $-3<\frac{M}{m_{0}}<-1$. Throughout the calculations in this paper, we have used the parameter $t=0.5$, $M=2.5$, $m_{0}=-1$. Its bulk band spectrum is shown in Fig. 4 a. A 2D film system with 10 layers was constructed by opening the boundary in the (001) direction, and the calculated band spectrum is shown in Fig. 4 b.

This section investigates the topological superconducting properties of this slab Hamiltonian by bringing it into Eq. , considering the magnetic proximity effect and the superconducting pairing forms corresponding to the two schemes in Section II.3.

In the Fu-kane’s proposal, we take the superconducting pairing $\Delta=0.1eV$ and the Zeeman field $B_{z}=0.2eV$. And in the Nagaosa’s proposal, we take $\Delta=0.1eV$ and $B_{z}=0.1eV$.

In Fig. 3 a,b,c, we can see that Fu-kane’s proposal does realize a topological superconducting state with Chern number 1 when $B_{z}>\Delta$. Which is in accordance with the expected. And in Fig. 3 d,e,f, we can see that Nagaosa’s proposal also realize a topological superconducting state with Chern number 1 even $B_{z}=\Delta$. As mentioned in Section II.3, as long as the coupling between the top and bottom surfaces is negligible with respect to $B_{z}$, the topological superconducting can be realized.

III.2 Bi₂Se₃

Fig. 5: a,b,c: Slab band structures, BdG band spectrum, and the wilson spectrum of the BdG Hamiltonian in Fu-kane’s proposal of Bi₂Se₃. d,e,f: Slab band structures, BdG band spectrum, and the wilson loop spectrum of the BdG Hamiltonian in Nagaosa’s proposal of Bi₂Se₃.

Fig. 6: a: Bulk band structures of Bi₂Se₃. b: Band structures of 10 layers slab of Bi₂Se₃.

In this section, we will demonstrate the ability of the program to analyze topological superconductivity in real material systems, using Bi₂Se₃ as an example. Bi2Se3 is a topological insulator. We have obtained the TB model for the p-orbitals of Bi and Se atoms using Wannier90, based on first-principles calculations. its bulk band spectrum and the 10-layer slab band spectrum in the [001] direction obtained by this TB model are shown in Fig. 6 a, b, with the Dirac cone appearing in the middle of the bulk energy gap [72].

Bringing the TB Hamiltonian of Bi₂Se₃ into Eq. 4 by taking Ns=10 and considering the effects of superconducting proximity effects and magnetic proximity effects, the BdG Hamiltonian of the 10-layer Bi2Se3 is obtained.

In the Fu-kane’s proposal, we take the superconducting pairing $\Delta=0.1eV$ and the Zeeman field $B_{z}=0.4eV$. And in the Nagaosa’s proposal, we take $\Delta=0.1eV$ and $B_{z}=0.1eV$.

In Fig. 5 a,b,c, we can see that Fu-kane’s proposal does realize a topological superconducting state with Chern number -1 when $B_{z}>\Delta$. The $B_{z}$ required to realize the topological phase here is larger than in the TI model, this is because the topological surface states in real materials are less localized than in the model, so the effective Zeeman field on the surface states is less than $B_{z}$. And in Fig. 5 d,e,f, we can see that Nagaosa’s proposal also realize a topological superconducting state with Chern number -1 even $B_{z}=\Delta$. This indicates that the coupling between the top and bottom surfaces is still small.

III.3 SnTe

Fig. 7: a,b,c: Slab band structures, BdG band spectrum, and the wilson spectrum of the BdG Hamiltonian in Fu-kane’s proposal of SnTe. d,e,f: Slab band structures, BdG band spectrum, and the wilson loop spectrum of the BdG Hamiltonian in Nagaosa’s proposal of SnTe.

Fig. 8: a: Bulk band structures of SnTe. b: Band structures of 10 layers slab of SnTe.

Fig. 9: a: The wilson spectrum of the BdG Hamiltonian in Fu-kane’s proposal of SnTe. b: The wilson spectrum of the BdG Hamiltonian in Nagaosa’s proposal of SnTe.

From the calculations in the previous two subsections, we have determined that Dirac cone surface states can be used to construct 2D chiral topological superconductors. Next, we will use WannierTools to explore whether systems with multiple surface Dirac cones can realize 2D topological superconductors with high Chern numbers, using SnTe as an example.

SnTe is a topological crystal insulator protected by mirror symmetry with four Dirac cone surface states protected by crystal symmetry on its [001] surface [73]. We have obtained the TB model for the p-orbitals of Sn and Te atoms using Wannier90, based on first-principles calculations. The bulk band spectrum and the 40-layer slab band spectrum in the [001] direction obtained by this TB model are shown in Fig. 8.

In the Fu-kane’s proposal, we take the superconducting pairing $\Delta=0.01eV$ and the Zeeman field $B_{z}=0.2eV$. The reason for setting $\Delta$ small here is that the bulk energy gap of SnTe is small and the surface states are less localized, so the energy scale needs to be carefully chosen to avoid interference from the bulk states. In the Nagaosa’s proposal, we take $\Delta=0.1eV$ and $B_{z}=0.1eV$.

In Fig. 7 a,b,c, we can see that Fu-kane’s proposal does realize a topological superconducting state with Chern number -4. And in Fig. 7 d,e,f, we can see that Nagaosa’s proposal also realize a topological superconducting state with Chern number -4 even. This indicates that the coupling between the top and bottom surfaces is still small.

Since it is inevitable to set both Dirac cones in the same integration path when calculating the wilson loop spectrum, the wilson loop spectrum of SnTe is very steep near this overlap** path. We can zoom in on the wilson loop spectrum near the $\Gamma$ point as shown in Fig. 9 a, b. This will clearly distinguish that the Chern number of the system is indeed -4.

This is an expected result, as the system has no time-reversal symmetry, so the four surface Dirac cones are related only by a fourfold rotational symmetry. Thus, these four Dirac cones should provide equal Berry curvature, resulting in a fourfold Chern number.

IV Conclusions

In conclusion, we developed a program [74] to compute the superconducting BdG spectrum and the superconducting topological invariants of two-dimensional systems. We show how it can be used to analyze topological superconducting properties in the SC/TI/FMI heterostructures. As one of the examples, we demonstrate that SnTe can be used to construct chiral topological superconductors with high Chern number. These results verify the validity and rationality of the program, which will play the role of first-principles calculations in the design and prediction of topological superconductors, and provide theoretical support for the enrichment of candidate materials for topological superconductors. On the other hand, the computational procedure can be extended to study other topological superconducting systems, such as magnetic topological insulator/superconductor heterojunctions, superconductor/ferromagnetic insulator heterostructures, and superconductor/topological insulator/superconductor heterojunctions.

V Acknowledgments

This work was supported by the National Key R&D Program of China (Grant No. 2023YFA1607400), the National Natural Science Foundation of China (Grant No. 12274154, 12274436), the Science Center of the National Natural Science Foundation of China (Grant No. 12188101) and the Center for Materials Genome, China.