Abstract
Quantum spin Hall (QSH) insulators are a peculiar phase of matter exhibiting excellent quantum transport properties with potential applications in lowerpowerconsuming electronic devices. Currently, among all predicted or synthesized QSH insulators, square and hexagonal atomic rings are the dominant structural motifs, and QSH insulators composed of pentagonal rings have not yet been reported. Here, based on firstprinciples calculations, we predict a family of largegap QSH insulators in SnX_{2} (X=S, Se, or Te) twodimensional (2D) crystals by the direct calculation of Z_{2} topological invariants and edge states. Remarkably, in contrast to all known QSH insulators, the QSH insulators predicted here are composed entirely of pentagonal rings. Moreover, these systems can produce sizeable nontrivial gaps ranging from 121 to 224 meV, which is sufficiently large for practical applications at room temperature. Additionally, we propose a quantum well by sandwiching an SnTe_{2} 2D crystal between two BiOBiS_{2} sheets and reveal that the considered 2D crystal remains topologically nontrivial with a sizeable gap. This finding demonstrates the robustness of its band topology against the effect of the substrate and provides a viable method for further experimental studies.
Introduction
Topological insulators (TIs) have sparked extensive research activities in recent years because of their rich physics and promising applications in quantum devices and spintronics.^{1, 2} The resultant Dirac surface states in threedimensional (3D) TIs as well as the helical edge states in twodimensional (2D) TIs are spin locked because of protection by timereversal symmetry; thus, they are robust against perturbations. Particularly in 2D TIs, also known as quantum spin Hall (QSH) insulators,^{3, 4} all the lowenergy scatterings of the edge states caused by the nonmagnetic defects are completely forbidden because the edge electrons can only propagate along two directions with opposite spins, which makes 2D TIs more suitable for lowpowerconsuming applications than 3D TIs. Unfortunately, the material realization of 2D TIs is challenging: while 3D TIs have been discovered in many materials, such as the Bi_{2}Se_{3} family,^{5, 6} the experimental realization of 2D TIs is thus far limited to the HgTe/CdTe^{7} and InAs/GaSb^{8} quantum wells. Moreover, the QSH effect in these two quantum wells can occur only at ultralow temperature (<10 K) because of their extremely small bulk band gaps, and this limitation greatly obstructs their possible applications. The search for new 2D TIs with large band gaps that could support room temperature applications has thus become critically important.
Considerable effort has been devoted thus far to search for new 2D materials or new schemes to realize 2D TIs, and a variety of largegap 2D TIs have been proposed theoretically. These largegap 2D TIs include silicene,^{9} Bi(1 1 1) bilayers,^{10} IIIBi bilayers,^{11} BiF 2D crystals,^{12} Bi_{4}Br_{4} single layers,^{13} chemically modified Ge/Sn^{14, 15, 16} and Bi/Sb^{17, 18} honeycomb lattices, ZrTe_{5}/HfTe_{5}^{19} and 2D transition metal dichalcogenides.^{20, 21, 22, 23} In terms of geometrical motifs, square, hexagonal and pentagonal rings are considered to be the three basic building blocks of 2D materials. However, among all predicted or synthesized 2D TIs,^{9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23} square and hexagonal rings are almost the only building blocks that have been found. Such a limitation mostly arises from the fact that the presence of a gapless band structure in these 2D materials makes them promising for harboring QSH states. Therefore, the common expectation is that the QSH effect cannot survive in pentagonal ringbased 2D crystals. To the best of our knowledge, extended 2D TIs composed exclusively of pentagonal rings have never been reported at all, although recent efforts have been conducted to search for such pentagonal ringbased 2D materials.^{24, 25}
The purpose of this work is to explore novel 2D topological phases, especially at room temperature, in the supposedly ordinary pentagonal ringbased 2D crystals. The practical means of doing so will not only significantly enrich our fundamental understanding of the topological phenomena but also greatly broaden the possibilities for realistic applications of the QSH effect. After extensive research, we show that 2D TIs, consisting entirely of pentagonal rings, can indeed exist, namely in the family of SnX_{2} 2D crystals, in which X=S, Se or Te. These SnX_{2} systems are identified to be roomtemperature 2D TIs with sizeable nontrivial band gaps ranging from 121 to 224 meV, and they form robust QSH systems. These new QSH insulators present single Dirac cone edge states crossing the bulk band gap, which is ideal for dissipationless transport. More notably, by sandwiching the SnTe_{2} 2D crystal between two BiOBiS_{2} sheets, we propose a promising quantum well configuration in which the considered 2D crystal remains topologically nontrivial with a sizeable band gap, thus inducting the robustness of its band topology against the effect of the substrate.
Methods
Density functional theory (DFT) calculations are performed using the plane wave basis Vienna ab initio simulation package.^{26, 27} The exchangecorrelation potential is described by the Perdew–Burke–Ernzerhof^{28} version of the generalized gradient approximation.^{29} The planewave cutoff is set to 500 eV, and the convergence threshold for energy is set to 10^{−6} eV. Monkhorst–Pack kpoint meshes^{30} of 9 × 9 × 1 and 13 × 13 × 1 are used for geometry optimization and selfconsistent electronic structure calculations, respectively. The 2D crystals are modeled by the periodic slab approach with a vacuum layer of more than 18 Å to ensure decoupling between periodic images. All structures, including the lattice parameters and the internal ion coordinates, are fully optimized with a residual force tolerance of 0.01 eV Å^{−1}. The spinorbit coupling (SOC) is included in the selfconsistent electronic structure calculations. For the quantum well structure, the van der Waals interaction is included by using the density functional theoryD2 approach.^{31}
Results
The schematic crystal structures of SnX_{2} 2D crystals are shown in Figures 1a and b. We refer to these two configurations as α and βSnX_{2}, respectively. βSnX_{2} is more stable than αSnX_{2} by 0.7, 22.8 and 25.8 meV per atom in energy for X=S, Se and Te, respectively. Supplementary Table S1 gives the corresponding structural details. From the top view, we can see that both α and βSnX_{2} 2D crystals resemble the structure of experimentally identified layered silver azide^{32} and are composed entirely of the pentagonal rings; they present an amazing pattern that is well known as Cairo pentagonal tiling.^{33} The crystal structure of βSnX_{2} displays P42_{1}m symmetry (point group D_{2d}) and a square lattice that contains two Sn and four X atoms in one unit cell. Its unit cell can be considered a triplelayer structure with an Sn layer sandwiched between two X layers, as shown in Figure 1b. αSnX_{2} 2D crystals also present a square lattice, but the symmetry is P4/mbm (point group D_{4H}). This indicates that αSnX_{2} 2D crystals have inversion symmetry, whereas the inversion symmetry does not hold for the βSnX_{2} 2D crystals. Figure 1c shows the first Brillouin zone of SnX_{2} 2D crystals, which is also square with four X points on the side centers and four M points on the corners. These systems predicted here are different from the known SnX_{2} 2D crystals (labeled as TSnX_{2}). The TSnX_{2} 2D crystals display the CdI_{2}type structure with a hexagonal lattice instead of a square lattice. The symmetry space group of TSnX_{2} 2D crystals is P3m1 (point group D_{3D}), and it contains one Sn and two X atoms in the unit cell. For more details, please see Supplementary Figure S1 and the corresponding discussion.
The electronic band structures of SnX_{2} 2D crystals without and with the inclusion of SOC are shown in Figure 2. Evidently, the valence and conduction bands away from the Γ point are well separated. Thus, we only need to focus on the bands around the Γ point. By projecting the bands onto different atomic orbitals, we find that the bands around the Fermi level are mainly contributed by p_{z}^{−}, degenerate p_{z}^{+}, p_{x, y}^{+} and degenerate p_{x, y}^{−} orbitals of the Sn/X atoms. Especially for αSnSe_{2}/SnTe_{2} 2D crystals at the Γ point, when excluding SOC, the Fermi level separates the (p_{x, y}^{+}, degenerate p_{x, y}^{−}) orbitals from the (p_{z}^{−}, degenerate p_{z}^{+}) orbitals, with the (p_{x, y}^{+}, degenerate p_{x, y}^{−}) orbitals located above the (p_{z}^{−}, degenerate p_{z}^{+}) orbitals. Furthermore, for these two 2D crystals without including SOC, one energy band with the p_{x, y}^{−} orbital character overlaps with another energy band with the p_{z}^{−} orbital character around the Fermi level, as shown in Figures 2b and c. This band overlap leads to the appearance of two band crossings located in very close proximity to the Fermi level near the Γ point for αSnSe_{2}/SnTe_{2} 2D crystals. Such band crossings can be understood as a result of the symmetry incompatibility of these two bands. Alternately, compared with αSnSe_{2}/SnTe_{2} 2D crystals, the bands around the Fermi level are markedly modified in βSnSe_{2}/SnTe_{2} 2D crystals. For βSnSe_{2}/SnTe_{2} 2D crystals, the energy band with the p_{z}^{−} orbital character is located far above the Fermi level, which leaves the degenerate p_{x, y}^{−} orbitals on the Fermi level. As a result, without including SOC, βSnSe_{2}/SnTe_{2} 2D crystals are gapless, and the valence band maximum and conduction band minimum degenerate at the Fermi level at the Γ point.
After including SOC, for αSnX_{2} 2D crystals, the band gap opens at the band crossing; see Figures 2a–c. The systems have indirect band gaps of 130, 224 and 121 meV for X=S, Se and Te, respectively (see Figure 1d). To understand the role of SOC on the band structures, we show the schematic diagram of the band evolution at the Γ point of αSnSe_{2} and αSnTe_{2} 2D crystals under SOC in Figure 1e. λ, ξ and δ in Figure 1e denote the band gap without SOC, the band splitting between degenerate p_{x, y}^{−} orbitals, and the band gap with SOC, respectively. We focus on the p_{z}^{−}, degenerate p_{z}^{+}, p_{x, y}^{+} and degenerate p_{x, y}^{−} orbitals around the Fermi level and neglect other atomic orbitals; see Figure 1e. Here, the superscripts ‘+’ and ‘−’ represents the bonding and antibonding states, respectively. For αSnSe_{2}/SnTe_{2} 2D crystals, without including SOC, one p_{z}^{−} orbital lies below the degenerate p_{x, y}^{−} orbitals, and the Fermi level stays between them; see Figure 1e. When switching on SOC, the degenerate p_{x, y}^{−} orbitals split, and the order of one p_{x, y}^{−} orbital and the p_{z}^{−} orbital is exchanged. According to the band overlap features shown in Figures 2b and c, it seems that such nontrivial band inversion before including SOC is induced by the inversion between the p_{z}^{−} and p_{x, y}^{−} orbitals near the Fermi level at the Γ point; however, this is not that case. We know that a band inversion that can affect the band topology, namely a nontrivial band inversion, refers to the inversion between bands with different parities. Otherwise, if a band inversion occurs between bands with the same parity, then it is a trivial band inversion; therefore, it could not affect the band topology. Our parity analysis shows that the p_{z}^{−} and p_{x, y}^{−} orbitals all display ‘−’ parity, which indicates that a nontrivial band inversion can never be induced by the inversion between these orbitals. In other words, here, the band overlap features near the Fermi level, as shown in Figures 2b and c, have nothing to do with the intrinsic nontrivial band inversion. Actually, the nontrivial band order before including SOC in these systems stems from the band inversion between the (p_{z}^{−}, degenerate p_{z}^{+}) and (p_{x, y}^{+}, degenerate p_{x, y}^{−}) orbitals, which is caused by the crystal field. To elucidate the underlying band inversion mechanism explicitly, we display the orbital evolution at the Γ point under the crystal field effect in Figure 1e (namely, from (I) to (II)). The crystal field effect is estimated by comparing the orbital orders with those of the same crystal that is strained to 170%, as suggested by Zhou et al.^{34} and Huang et al.^{35} Without including the crystal field effect (see (I) of Figure 1e), the (p_{z}^{−}, degenerate p_{z}^{+}) orbitals are located above the Fermi level, whereas the (p_{x, y}^{+}, degenerate p_{x, y}^{−}) orbitals are located below the Fermi level. At this moment, the band topologies of αSnSe_{2}/SnTe_{2} are trivial. Under the crystal field effect (see (II) of Figure 1e), a band inversion occurs between the (p_{z}^{−} orbital with ‘−’ parity, degenerate p_{z}^{+} orbitals with ‘+’ parity) and (p_{x, y}^{+} orbital with ‘+’ parity, degenerate p_{x, y}^{−} orbitals with ‘−’ parity), indicating a topological phase transition from a trivial insulator to a nontrivial insulator. Obviously, the subsequent SOCinduced band exchange between the p_{z}^{−} and p_{x, y}^{−} orbitals does not contribute to the band inversion because they both display ‘−’ parity. This is different from the situation in the previously reported systems, such as Bi_{2}Te_{3},^{36} where the SOCinduced band exchange usually changes the band topology.
To understand the difference in band structures between αSnS_{2} and αSnSe_{2}/SnTe_{2} 2D crystals, we further display the orbital evolution at the Γ point for αSnS_{2} 2D crystal in Supplementary Figure S2. As shown in Supplementary Figure S2, similar to αSnSe_{2}/SnTe_{2} 2D crystals, the (p_{z}^{−}, degenerate p_{z}^{+}) orbitals also lie above the (p_{x, y}^{+}, degenerate p_{x, y}^{−}) orbitals before including the crystal field effect, and then a band inversion occurs between the (p_{z}^{−}, degenerate p_{z}^{+}) and (p_{x, y}^{+}, degenerate p_{x, y}^{−}) orbitals after including the crystal field effect, thus driving the αSnS_{2} 2D crystal into a topological phase. Here, the SOC also does not contribute to the band inversion. It should be noted that the crystal field effect shifts the (p_{z}^{−}, degenerate p_{z}^{+}) orbitals of the αSnS_{2} 2D crystal far above the Fermi level with respect to the case of αSnSe_{2}/SnTe_{2}, as shown in Supplementary Figure S2. Therefore, the subsequent SOCinduced band exchange between the p_{z}^{−} and p_{x, y}^{−} orbitals observed in αSnSe_{2}/SnTe_{2} does not occur in the αSnS_{2} 2D crystal (see Figure 1e and Supplementary Figure S2).
For βSnSe_{2} and βSnTe_{2} 2D crystals, SOC also lifts the degeneracy of the bands at the Fermi level, and the conduction band shifts upwards while the valence band shifts downwards, which forms a band gap of 146 and 160 meV, respectively. The corresponding schematic diagram of the band evolution at the Γ point under the crystal field effect and SOC in sequence is plotted in Figure 1f. Before including the crystal field effect (see (I) of Figure 1f), the (p_{z}^{−}, degenerate p_{z}^{+}) and (p_{x, y}^{+}, degenerate p_{x, y}^{−}) orbitals are located above and below the Fermi level, respectively. After including the crystal field effect (see (II) of Figure 1f), a band inversion occurs between the degenerate p_{z}^{+} and (p_{x, y}^{+}, degenerate p_{x, y}^{−}) orbitals, indicting a topological phase transition from a normal insulator to a TI. This leaves the Fermi level right on the degenerate p_{x, y}^{−} orbitals; subsequently, after turning on SOC, the degeneracy of the two p_{x, y}^{−} orbitals is lifted; see the SOC part of Figure 1f. Therefore, βSnSe_{2}/SnTe_{2} 2D crystals also display a nontrivial band order before including SOC, which also arises from the crystal field effect. In these systems, the SOCinduced band exchange also does not contribute to the band inversion, and the only effect of SOC is to create energy gaps at the touching points. By comparing Figures 1e and f, we can see that, without including the crystal field, the orbital order of the βSnSe_{2}/SnTe_{2} 2D crystals is similar to that of αSnSe_{2}/SnTe_{2}, that is, the (p_{z}^{−}, degenerate p_{z}^{+}) orbitals are located above the (p_{x, y}^{+}, degenerate p_{x, y}^{−}) orbitals. This is due to the similar elemental composition of α and βSnSe_{2}/SnTe_{2} 2D crystals. However, as shown in Figures 1a and b, the geometric structures of the α and β phases are very different, which indicates that the crystal field effects in the α and βSnSe_{2}/SnTe_{2} 2D crystals are very different. Thus, after including the crystal field effect, a band inversion occurs between the degenerate p_{z}^{+} and (p_{x, y}^{+}, degenerate p_{x, y}^{−}) orbitals for βSnSe_{2}/SnTe_{2} 2D crystals, whereas the (p_{z}^{−}, degenerate p_{z}^{+}) and (p_{x, y}^{+}, degenerate p_{x, y}^{−}) orbitals are inverted for the α phase (see Figures 1e and f). As a result, in contrast to the α phase in which the p_{z}^{−} orbital shifts below the degenerate p_{x, y}^{−} orbitals, the p_{z}^{−} orbital of βSnSe_{2}/SnTe_{2} 2D crystals remains above the degenerate p_{x, y}^{−} orbitals after including the crystal field effect. Moreover, by comparing Figures 2d–f, we notice that SOC is responsible for a significant spin splitting of both the valence and conduction bands of βSnX_{2} 2D crystals around the Fermi level. The origin lies in the break of inversion symmetry in βSnX_{2} 2D crystals, which induces the local dipole field and thus leads to the spin splitting. Alternately, because αSnX_{2} 2D crystals possess inversion symmetry, bands in Figures 2a–c are thus spin degenerate in the presence of SOC. For the βSnS_{2} 2D crystal, the scenario of the band evolution under SOC is very special. This scenario is different from the other five systems because after including SOC, the band gap does not open, and the entire system is metallic. Therefore, for the remainder of our discussion, we will focus on the αSnX_{2}, βSnSe_{2} and βSnTe_{2} 2D crystals.
To identify the topological nature of the insulating phases in these systems, we investigate the Z_{2} topological invariants within the density functional theory frame. Z_{2}=1 characterizes a topologically nontrivial phase, whereas Z_{2}=0 corresponds to a topologically trivial phase. For αSnX_{2} 2D crystals, the existence of inversion symmetry simplifies the calculations of Z_{2} topological invariants. According to the method developed by Fu and Kane,^{37} topological invariants can be easily obtained by evaluating the parity eigenvalues of each pair of Kramer’s degenerate occupied energy bands at the four timereversalinvariant momenta points (one Γ, one M and two X) of the Brillouin zone. Owing to the square symmetry, the parities at the two X points are identical, δ(X)^{2}≡1, which has no effect on the band topology; thus, the topological invariants can be reduced to (−1)^{ν}=δ(Γ) δ(M). In αSnX_{2} 2D crystals, the product of the parity eigenvalues at the Γ point is −1, whereas it is +1 at the M point; thus, we conclude that αSnX_{2} 2D crystals are nontrivial 2D TIs with Z_{2}=1. For more details, see Supplementary Table S2. For βSnSe_{2} and βSnTe_{2} 2D crystals, the inversion symmetry is broken, and the Z_{2} topological invariants cannot be determined by the method mentioned above. Instead, we use the method in Fukui and Hatsugai^{38} to directly perform the lattice computation of the Z_{2} topological invariants in terms of the nfield configuration. The resultant Z_{2}=1 verifies that βSnSe_{2} and βSnTe_{2} 2D crystals are also 2D TIs. The roomtemperature QSH effect can be readily realized in all these systems if one considers their nontrivial band gaps. We emphasize that these results go beyond all predicted or synthesized 2D TIs^{9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23} in which square and hexagonal atomic rings are almost the only building blocks. We show, for the first time, that the QSH effect can also occur in the supposedly ordinary 2D crystals composed entirely of pentagonal rings. Our work opens new avenues for the discovery of additional largegap 2D TIs in different and unexplored classes of systems.
To further reveal the nontrivial topological nature of these systems, we perform calculations to examine the topological edge states on the boundaries of αSnX_{2}, βSnSe_{2} and βSnTe_{2} 2D crystals. They are calculated using a nanoribbon structure with symmetric edges, which leads to two energetically degenerate edge states located at opposite sides. The corresponding schematic ribbon model is shown in Figure 3a. All the dangling bonds of the edge atoms are saturated by hydrogen atoms. The widths of the nanoribbons are set large enough to avoid interactions between the two edges. Figures 3b–f presents the calculated band structures of αSnX_{2}, βSnSe_{2} and βSnTe_{2} nanoribbons. One can explicitly see that each edge has a single pair of helical edge states (yellow lines) for all five of these systems, and the helical edge states disperse in the bulk band gap and cross linearly at the point. These features further prove the nontrivial nature of these systems, which is consistent with the Z_{2} calculations. Remarkably, in the cases of αSnS_{2}, αSnSe_{2}, αSnTe_{2} and βSnTe_{2}, the Dirac points formed by the helical edge states lie exactly at the Fermi level. Such placement of Dirac points is important for actual applications. Moreover, their sizeable bulk band gaps can stabilize the edge states against the interference of the thermally activated carriers, which are beneficial for observing the roomtemperature QSH effect in these 2D crystals.
Discussion
In the following section, we take αSnTe_{2} 2D crystal as an example and discuss the realistic possibility of experimentally realizing these 2D crystals. We find that the widely used square BiOBiS_{2} sheet^{39} can be used as a suitable substrate for an αSnTe_{2} 2D crystal in experiments. The inplane lattice constant of an αSnTe_{2} 2D crystal is very close to that of the insulating 2 × 2 BiOBiS_{2} sheet. The lattice mismatch between them is only 2.3%. To this end, we propose a quantum well structure between the αSnTe_{2} 2D crystal and BiOBiS_{2} sheet. As shown in Figure 4a, in the BiOBiS_{2}/αSnTe_{2} quantum well, one αSnTe_{2} layer is sandwiched between two BiOBiS_{2} sheets. Our calculated band structure displayed in Figure 4b shows that the states around the Fermi level mainly come from the p_{z}^{−} and p_{x, y}^{−} orbitals of αSnTe_{2}. In the absence of SOC for the BiOBiS_{2}/αSnTe_{2} quantum well, which is similar to the case of a freestanding αSnTe_{2} 2D crystal, there is still one energy band with the p_{x, y}^{−} orbital character that overlaps with another energy band with the p_{x, y}^{−} orbital character around the Fermi level. However, such band overlapping only forms one Dirac point located at the Fermi level along the Γ–X direction. After including SOC, the Dirac point is deformed, and a band gap appears, which is a strong indication for the existence of a topologically nontrivial phase. Therefore, the BiOBiS_{2}/αSnTe_{2} quantum well may still be a 2D TI. To firmly confirm its band topology, we investigated the Z_{2} topological invariant. The corresponding results are listed in Supplementary Table S3. The result of Z_{2}=1 indicates that a BiOBiS_{2}/αSnTe_{2} quantum well is indeed a QSH insulator. More remarkably, the nontrivial bulk band gap is 44 meV, which is still well above the energy scale of room temperature.
To deeply understand the effect of the insulating substrate BiOBiS_{2} sheet on the topological nature of an αSnTe_{2} 2D crystal, we investigated the related properties of an isolated αSnTe_{2} layer that is taken directly from the BiOBiS_{2}/αSnTe_{2} quantum well without any further structure relaxation. The corresponding band structures are plotted in Figure 4c, and we can observe that the band structure near the Fermi level retains linear band crossing in the absence of SOC, which resembles that of the BiOBiS_{2}/αSnTe_{2} quantum well. When including SOC, a band gap of 214 meV is produced in the isolated αSnTe_{2} layer. Our band topology analysis estimates that the topologically nontrivial state remains intact in the isolated αSnTe_{2} layer. However, by comparing Figures 4b and c, it is important to notice that the nontrivial band gap is significantly reduced (from 214 to 44 meV) when the isolated αSnTe_{2} layer is sandwiched between the BiOBiS_{2} sheets. This reduction reflects the band realignment effect when placing thin films on the substrate. The conduction band maximum and valence band maximum of the BiOBiS_{2}/αSnTe_{2} quantum well are contributed by the BiOBiS_{2} sheets and isolated αSnTe_{2} layer, respectively. Obviously, these are the unique features of the typeII quantum well: the electron subband and hole subband are separated in two different layers. Considering these results, one can easily understand this reduction. Furthermore, these results also reveal the importance of the gap size of the substrate for maintaining the nontrivial topological states.
Finally, we wish to note that topological states have also been reported recently in IV–VI 2D crystals.^{40} Unlike the SnX_{2}based 2D TIs studied here, the topological states of IV–VI 2D crystals are protected by crystal symmetry instead of by timereversal symmetry. Namely, these IV–VI 2D crystals are 2D topological crystalline insulators.^{40} Therefore, these two classes of 2D crystals with the same elements but different compositions/structures can lead to very distinct topological phases: 2D TI and 2D topological crystalline insulator. This finding is certainly interesting and suggests greater room in the design of topological materials beyond conventional structures.
In summary, we show by firstprinciples calculations that αSnX_{2}, βSnSe_{2} and βSnTe_{2} 2D crystals, which are composed exclusively of pentagonal rings, are 2D TIs. This result challenges the current knowledge 2D TIs, that is, that square or hexagonal rings are the commonly building blocks. Their topologically nontrivial characteristics are confirmed by the nontrivial Z_{2}type topological invariants and the existence of topologically protected helical edge states. Sizeable nontrivial band gaps of 121–224 meV are obtained, suggesting that the QSH effect can be realized in these systems above room temperature. For βSnSe_{2} and βSnTe_{2} 2D crystals, their bands exhibit the SOCinduced spinsplitting effect due to their broken inversion symmetry. αSnX_{2} 2D crystals are inversion symmetric; thus, no spin splitting can be observed. To probe the possibility of experimentally realizing these 2D crystals, we further propose the BiOBiS_{2}/αSnTe_{2} quantum well. Additionally, we find that the αSnTe_{2} 2D crystal remains topologically nontrivial with a sizeable gap (44 meV) when sandwiched between the BiOBiS_{2} sheets, thus suggesting that the 2D TIs predicted here can be sandwiched between appropriate nanostructures for realistic applications. Our work predicts that more pentagonal ringbased 2D TIs will be discovered in the future, which will greatly broaden the scientific and technological impact of the QSH effect.
References
 1.
Hasan, M. Z. & Kane, C. L. Colloquium: topological insulators. Rev. Mod. Phys. 82, 3045–3067 (2010).
 2.
Qi, X.L. & Zhang, S.C. Topological insulators and superconductors. Rev. Mod. Phys. 83, 1057 (2011).
 3.
Kane, C. L. & Mele, E. J. Quantum spin Hall effect in graphene. Phys. Rev. Lett. 95, 226801 (2005); Phys. Rev. Lett. Z_{2} topological order and the quantum spin Hall effect. 95, 146802 (2005).
 4.
Bernevig, B. A. & Zhang, S. C. Quantum spin Hall effect. Phys. Rev. Lett. 96, 106802 (2006).
 5.
Zhang, H., Liu, C.X., Qi, X.L., Dai, X., Fang, Z. & Zhang, S.C. Topological insulators in Bi_{2}Se_{3}, Bi_{2}Te_{3} and Sb_{2}Te_{3} with a single Dirac cone on the surface. Nat. Phys. 5, 438–442 (2009).
 6.
Xia, Y., Qian, D., Hsieh, D., Wray, L., Pal, A., Lin, H., Bansil, A., Grauer, D., Hor, Y. S., Cava, R. J. & Hasan, M. Z. Observation of a largegap topologicalinsulator class with a single Dirac cone on the surface. Nat. Phys. 5, 398–402 (2009).
 7.
Konig, M., Wiedmann, S., Brune, C., Roth, A., Buhmann, H., Molenkamp, L. W., Qi, X. L. & Zhang, S. C. Quantum spin Hall insulator state in HgTe quantum wells. Science 318, 766–770 (2007).
 8.
Knez, I., Du, R.R. & Sullivan, G. Evidence for helical edge modes in inverted InAs/GaSb quantum wells. Phys. Rev. Lett. 107, 136603 (2011).
 9.
Liu, C.C., Feng, W. & Yao, Y. Quantum spin Hall effect in silicene and twodimensional germanium. Phys. Rev. Lett. 107, 076802 (2011).
 10.
Murakami, S. Quantum spin Hall effect and enhanced magnetic response by spinorbit coupling. Phys. Rev. Lett. 97, 236805 (2006).
 11.
Chuang, F.C., Yao, L.Z., Huang, Z.Q., Liu, Y.T., Hsu, C.H., Das, T., Lin, H. & Bansil, A. Prediction of largegap twodimensional topological insulators consisting of bilayers of group III elements with Bi. Nano Lett 14, 2505 (2014).
 12.
Luo, W. & Xiang, H. J. Room temperature quantum spin Hall insulators with a buckled square lattice. Nano Lett. 15, 3230–3235 (2015).
 13.
Zhou, J. J., Feng, W. X., Liu, C.C., Guan, S. & Yao, Y. G. Largegap quantum spin Hall insulator in single layer bismuth monobromide Bi_{4}Br_{4}. Nano Lett. 14, 4767–4771 (2014).
 14.
Ma, Y. D., Kou, L. Z., Du, A. J. & Heine, T. Group 14 elementbased noncentrosymmetric quantum spin Hall insulators with large bulk gap. Nano Res. 8, 3412–3420 (2015).
 15.
Si, C., Liu, J. W., Xu, Y., Wu, J., Gu, B.L. & Duan, W. H. Functionalized germanene as a prototype of largegap twodimensional topological insulators. Phys. Rev. B 89, 115429 (2014).
 16.
Xu, Y., Yan, B. H., Zhang, H.J., Wang, J., Xu, G., Tang, P., Duan, W. & Zhang, S. C. Largegap quantum spin Hall insulators in tin films. Phys. Rev. Lett. 111, 136804 (2013).
 17.
Ma, Y. D., Dai, Y., Kou, L. Z., Frauenheim, T. & Heine, T. Robust twodimensional topological insulators in methylfunctionalized bismuth, antimony, and lead bilayer films. Nano Lett. 15, 1083–1089 (2015).
 18.
Song, Z. G., Liu, C. C., Yang, J. B., Han, J. Z., Fu, B. T., Ye, M., Fu, B. T., Yang, Y. C., Niu, Q., Lu, J. & Yao, Y. G. Quantum spin Hall insulators and quantum valley Hall insulators of BiX/SbX (X=H, F, Cl and Br) monolayers with a record bulk band gap. NPG Asia Mater. 6, 147 (2014).
 19.
Weng, H., Dai, X. & Fang, Z. Transitionmetal pentatelluride ZrTe_{5} and HfTe_{5}: a paradigm for largegap quantum spin Hall insulators. Phys. Rev. X 4, 011002 (2014).
 20.
Ma, Y. D., Kou, L. Z., Li, X., Dai, Y., Smith, S. C. & Heine, T. Quantum spin Hall effect and topological phase transition in twodimensional square transitionmetal dichalcogenides. Phys. Rev. B 92, 085427 (2015).
 21.
Ma, Y. D., Kou, L. Z., Li, X., Dai, Y. & Heine, T. Twodimensional transition metal dichalcogenides with a hexagonal lattice: roomtemperature quantum spin Hall insulators. Phys. Rev. B 93, 035442 (2016).
 22.
Nie, S. M., Song, Z., Weng, H. M. & Fang, Z. Quantum spin Hall effect in twodimensional transitionmetal dichalcogenide haeckelites. Phys. Rev. B 91, 235434 (2015).
 23.
Qian, X. F., Liu, J. W., Fu, L. & Li, J. Quantum spin Hall effect in twodimensional transition metal dichalcogenides. Science 346, 1344–1347 (2014).
 24.
Zhang, S. H., Zhou, J., Wang, Q., Chen, X. S., Kawazoe, Y. & Jena, P. Pentagraphene: a new carbon allotrope. Proc Natl Acad Sci USA 112, 2372–2377 (2015).
 25.
Tang, C. P., Xiong, S. J., Shi, W. J. & Cao, J. Twodimensional pentagonal crystals and possible spinpolarized Dirac dispersion relations. J. Appl. Phys. 115, 113702 (2014).
 26.
Kresse, G. & Furthmüller, J. Efficiency of abinitio total energy calculations for metals and semiconductors using a planewave basis set. Comput. Mater. Sci. 6, 15–50 (1996).
 27.
Kresse, G. & Furthmller, J. Efficient iterative schemes for ab initio totalenergy calculations using a planewave basis set. Phys. Rev. B 54, 11169 (1996).
 28.
Perdew, J. P., Burke, K. & Ernzerhof, M. Generalized gradient approximation made simple. Phys. Rev. Lett. 77, 3865 (1996).
 29.
Perdew, J. P. & Wang, Y. Accurate and simple analytic representation of the electrongas correlation energy. Phys. Rev. B 45, 13244 (1992).
 30.
Monkhorst, H. J. & Pack, J. D. Special points for Brillouinzone integrations. Phys. Rev. B 13, 5188 (1976).
 31.
Grimme, S. Semiempirical GGAtype density functional constructed with a longrange dispersion correction. J. Comput. Chem. 27, 1787–1799 (2006).
 32.
Schmidt, C. L., Dinnebier, R., Wedig, U. & Jansen, M. Crystal structure and chemical bonding of the hightemperature phase of AgN_{3}. Inorg. Chem. 46, 907–916 (2007).
 33.
Ressouche, E., Simonet, V., Canals, B., Gospodinov, M. & Skumryev, V. Magnetic frustration in an ironbased Cairo pentagonal lattice. Phys. Rev. Lett. 103, 267204 (2009).
 34.
Zhou, L. J., Kou, L. Z., Sun., Y., Felser, C., Hu, F. M., Shan, G. C., Smith, S. C., Yan, B. & Frauenheim, T. New family of quantum spin Hall insulators in twodimensional transitionmetal halide with large nontrivial band gaps. Nano Lett. 15, 7867–7872 (2015).
 35.
Huang, H. Q., Liu, J. P. & Duan, W. H. Nontrivial Z_{2} topology in bismuthbased III–V compounds. Phys. Rev. B 90, 195105 (2014).
 36.
Zhang, Y., He, K., Chang, C.Z., Song, C.L., Wang, L.L., Chen, X., Jia, J.F., Fang, Z., Dai, X., Shan, W.Y., Shen, S.Q., Niu, Q., Qi, X.L., Zhang, S. C., Ma, X.C. & Xue, Q.K. Crossover of the threedimensional topological insulator Bi_{2}Se_{3} to the twodimensional limit. Nat. Phys. 6, 584–588 (2010).
 37.
Fu, L., Kane, C. L. & Mele, E. J. Topological insulators in three dimensions. Phys. Rev. Lett. 98, 106803 (2007).
 38.
Fukui, T. & Hatsugai, Y. Quantum spin Hall effect in three dimensional materials: Lattice computation of Z_{2} topological invariants and its application to Bi and Sb. J. Phys. Soc. Jpn. 76, 053702 (2007).
 39.
Phelan, W. A., Wallace, D. C., Arpino, K. E., Neilson, J. R., Livi, K. J., Seabourne, C. R., Scott, A. J. & McQueen, T. M. Stacking variants and superconductivity in the Bi–O–S system. J. Am. Chem. Soc. 135, 5372–5374 (2013).
 40.
Liu, J. W., Qian, X. F. & Fu, L. Crystal field effect induced topological crystalline insulators in monolayer IV−VI semiconductors. Nano Lett. 15, 2657–2661 (2015).
Acknowledgements
Financial support from the European Research Council (ERC, StG 256962) and the Taishan Scholar Program of Shandong is gratefully acknowledged.
Author contributions
YM and TH conceived and designed this study, YM performed the calculations, and all authors discussed the results and contributed to the writing of the manuscript.
Author information
Affiliations
Department of Physics and Earth Sciences, Jacobs University Bremen, Bremen, Germany
 Yandong Ma
 & Thomas Heine
School of Chemistry, Physics and Mechanical Engineering Faculty, Queensland University of Technology, Brisbane, QLD, Australia
 Liangzhi Kou
Department of Physics, University of Texas at Austin, Austin, TX, USA
 Xiao Li
School of Physics, Shandong University, Jinan, People’s Republic of China
 Ying Dai
WilhelmOstwaldInstitut für Physikalische und Theoretische Chemie, Universität Leipzig, Leipzig, Germany
 Thomas Heine
Authors
Search for Yandong Ma in:
Search for Liangzhi Kou in:
Search for Xiao Li in:
Search for Ying Dai in:
Search for Thomas Heine in:
Competing interests
The authors declare no conflict of interest.
Corresponding authors
Correspondence to Yandong Ma or Thomas Heine.
Supplementary information
Word documents
Rights and permissions
This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/
About this article
Supplementary Information accompanies the paper on the NPG Asia Materials website (http://www.nature.com/am)
Further reading

1.
Construction of bilayer PdSe2 on epitaxial graphene
Nano Research (2018)

2.
npj 2D Materials and Applications (2017)

3.
Nature Communications (2016)