Abstract
Magnetic topological insulators (TI) provide an important material platform to explore quantum phenomena such as quantized anomalous Hall effect and Majorana modes, etc. Their successful material realization is thus essential for our fundamental understanding and potential technical revolutions. By realizing a bulk van der Waals material MnBi_{4}Te_{7} with alternating septuple [MnBi_{2}Te_{4}] and quintuple [Bi_{2}Te_{3}] layers, we show that it is ferromagnetic in plane but antiferromagnetic along the c axis with an outofplane saturation field of ~0.22 T at 2 K. Our angleresolved photoemission spectroscopy measurements and firstprinciples calculations further demonstrate that MnBi_{4}Te_{7} is a Z_{2} antiferromagnetic TI with two types of surface states associated with the [MnBi_{2}Te_{4}] or [Bi_{2}Te_{3}] termination, respectively. Additionally, its superlattice nature may make various heterostructures of [MnBi_{2}Te_{4}] and [Bi_{2}Te_{3}] layers possible by exfoliation. Therefore, the low saturation field and the superlattice nature of MnBi_{4}Te_{7} make it an ideal system to investigate rich emergent phenomena.
Introduction
Magnetic topological insulators (MTIs), including Chern insulators with a Zinvariant and antiferromagnetic (AFM) topological insulators (TIs) with a Z_{2}invariant, provide fertile ground for the exploration of emergent quantum phenomena such as the quantum anomalous Hall (QAH) effect, Majorana modes, the topological magnetoelectric effect, the proximity effect, etc^{1,2}. In the twodimensional (2D) limit of ferromagnetic (FM) TIs, the QAH effect arising from chiral edge states exists under zero external magnetic fields, which has been experimentally observed in doped FM TI Cr_{0.15}(Bi_{0.1}Sb_{0.9})_{1.85}Te_{3} thin films^{3}. However, the unavoidable sample inhomogeneity in doped materials restrains the investigation of associated emergent phenomena below temperatures of hundreds of mK^{2}. Stoichiometric MTIs are expected to have homogeneous electronic and magnetic properties, which may provide new opportunities to study the QAH effect. Recently, MnBi_{2}Te_{4} was discovered to be an intrinsic AFM TI^{4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24}. In its 2D limit, quantized Hall conductance originating from the topological protected dissipationless chiral edge states was realized in fewlayer slabs^{15,16}. However, probably because the uncompensated AFM spin configuration cannot provide enough Zeeman field to realize the band inversion in only one spin channel, to observe such a QAH effect, a high magnetic field of 12 T at 4.5 K or 6 T at 1.5 K is required to fully polarize the AFM spins into a forced FM state^{15}.
A FM state is crucial to realize the QAH effect experimentally^{15}; however, as we await an ideal candidate that has both TI and FM properties, an intrinsic AFM TI with low saturation fields and clean band structure where only nontrivial bands cross the Fermi level can also provide a good material platform. By this, the QAH effect may be realized with higher temperatures and reasonably low magnetic fields, which allows us to study their associated emergent phenomena at more accessible conditions. How can we realize such intrinsic AFM TIs? Recall that MnBi_{2}Te_{4} crystalizing in the GeBi_{2}Te_{4} structure with septuple layers (SL) of [MnBi_{2}Te_{4}] is an AFM material with inplane FM and outofplane AFM exchange interaction. Hence, based on the SL building block, one strategy to achieve AFM with small saturation fields or even FM is to reduce the interlayer MnMn exchange interaction by increasing the interlayer distance with extra spacer layers added. Structurally, SL blocks have great compatibility with [Bi_{2}Te_{3}] quintuple layers (QL), whose bulk form is a TI with a preserved timereversal symmetry. As an example, GeBi_{4}Te_{7} with alternating [GeBi_{2}Te_{4}] and [Bi_{2}Te_{3}] building blocks has been synthesized^{25}. This superior compatibility provides us with flexible structural control to achieve our goal. Furthermore, not only can such superlattices manifest weak interlayer magnetic coupling, but they can also serve as natural heterostructures by exfoliation, which may enable the realization of various topological states.
The exploration of the MnTe–Bi_{2}Te_{3} ternary system^{26} has shown that MnBi_{2n}Te_{3n+1} (n = 1, 2, and 3) series exist with alternating [MnBi_{2}Te_{4}] and (n−1)[Bi_{2}Te_{3}] layers. In this work, we focus on MnBi_{4}Te_{7} (n = 2) with a hexagonal superlattice crystal structure of alternate stackings of one [MnBi_{2}Te_{4}] SL and one [Bi_{2}Te_{3}] QL. Through our transport, thermodynamic, angleresolved photoemission spectroscopy (ARPES) and density functional theory (DFT) calculations, we discovered that MnBi_{4}Te_{7} is a Z_{2} AFM TI with an outofplane saturation field as low as 0.22 T at 2 K, 40 times lower than that of MnBi_{2}Te_{4}. Furthermore, the naturalheterostructurelike construction of MnBi_{4}Te_{7} can host two distinct (001) surface states. For the [Bi_{2}Te_{3}] termination, clean gapped surface states are observed as has long been desired; while for the [MnBi_{2}Te_{4}] termination, nearly gapless surface Dirac cone is observed, similar to the case of the MnBi_{2}Te_{4} compound^{20,21,22,23}. Our finding provides a superior new material realization to explore the QAH effect, quantum spin Hall (QSH) effect and associated phenomena^{27}.
Results
Atype antiferromagnetism in MnBi_{4}Te_{7} with strong FM fluctuations and weak interlayer exchange interaction
Figure 1b shows the (00l) Xray diffraction peaks of a piece of representative single crystal, which can be well indexed by the MnBi_{4}Te_{7} crystal structure^{26}. The Rietveld refinement of the powder Xray diffraction pattern agrees well the MnBi_{4}Te_{7} structure model^{26} and suggests Bi_{2}Te_{3} is the only impurity inside with a molar ratio of 14% (Supplementary Fig. 1). The refined lattice parameters are a = 4.3454(5) Å, and c = 23.706(4) Å, indicating the distance between two adjacent Mn layers in MnBi_{4}Te_{7} is 23.706(4) Å, much longer than the 13.8 Å of MnBi_{2}Te_{4}. The inset of Fig. 1b shows a picture of a MnBi_{4}Te_{7} single crystal against a 1mm scale, where the shiny cleaved ab surface can be seen.
The magnetic properties are depicted in Fig. 1c–e. Figure 1c presents the fieldcooled (FC) magnetic susceptibility data of x^{ab} (H  ab) and x^{c} (H  c) measured at 0.1 T. The abrupt halt in the rise of x^{c} on cooling suggests the onset of AFM ordering, similar to that seen in other vdW antiferromagnets MnBi_{2}Te_{4} and CrCl_{3},^{9,28} but different from the FM one^{29}, suggesting that long range AFM ordering takes place at 13 K. This is consistent with the specific heat measurement in Supplementary Fig. 2, where a specific heat anomaly associated with the AFM transition emerges at 13 K. As seen from Fig. 1c, fitting the inverse susceptibilities up to 80 K to the CurieWeiss law results in Weiss temperatures of \(\theta _{\mathrm{w}}^{{\mathrm{ab}}}\) = 11.5 K, \(\theta _{\mathrm{w}}^{\mathrm{c}}\) = 12.2 K, \(\theta _{\mathrm{w}}^{{\mathrm{ave}}}\) = 11.7 K, and effective moments of \(\mu _{{\mathrm{eff}}}^{{\mathrm{ab}}}\) = 5.4μ_{B}/Mn, \(\mu _{{\mathrm{eff}}}^{\mathrm{c}}\) = 5.1μ_{B}/Mn and \(\mu _{{\mathrm{eff}}}^{{\mathrm{ave}}}\) = 5.3μ_{B}/Mn. These values indicate magnetic isotropy above T_{N} and thus negligible single ion anisotropy in the material. Despite the fact that MnBi_{4}Te_{7} is AFM below 13 K, the positive \(\theta _{\mathrm{w}}^{{\mathrm{ave}}}\) of 11.7 K suggests strong ferromagnetic (FM) exchange interactions. Recall that MnBi_{2}Te_{4} has a much higher T_{N} of 25 K and a much lower θ_{w} of 3–6 K^{9,18}, this may indicate that the energy scales of the FM and AFM exchange interaction are much closer in MnBi_{4}Te_{7}. This is consistent with the fact that the extra insulating [Bi_{2}Te_{3}] layer reduces the interlayer exchange interaction between adjacent Mn layers as we initially designed. The AFM orders of both MnBi_{2}Te_{4} and MnBi_{4}Te_{7} are formed under the superexchange scenario, where the magnetic interaction between the adjacent Mn layers is mediated by the electrons of the common neighbors. Despite the long distance between the adjacent Mn layers (23.7 Å), our DFT calculation reveals an Atype AFM configuration in MnBi_{4}Te_{7} with the interlayer exchange coupling about −0.15 meV/Mn, which is about one order of magnitude smaller than the counterpart of MnBi_{2}Te_{4}. More details are given in Supplementary Note 2.
Figure 1d, e present the hysteresis loops of isothermal magnetization data for M^{c}(H) (H  c) and M^{ab}(H) (H  ab), respectively. As shown in Fig. 1d, in sharp contrast to MnBi_{2}Te_{4} where a spinflop transition takes place at 3.5 T and saturates at 8 T in M^{c}(H)^{9,17,18}, MnBi_{4}Te_{7} undergoes a firstorder spinflip transition with hysteresis starting at a much lower field of H_{f} = 0.15 T. It quickly enters the forced FM state and saturates at H_{c} = 0.22 T. The small saturation field again indicates weaker interlayer AFM exchange interactions than in MnBi_{2}Te_{4.} Upon warming up to 10 K, the hysteresis area is gradually reduced to zero, but H_{c }remains little changed, indicating a sharp triggering of the spinflipping between 10 K and T_{N}. With H  ab, the saturation field is 1.0 T, indicating the c axis as the magnetic easy axis and likely Ising form. As shown in Fig. 1e, the saturation moment is 3.5μ_{B}/Mn at 7 T, which is very similar to the value of 3.6μ_{B}/Mn^{18} in MnBi_{2}Te_{4} but smaller than the DFT calculated value of 4.6μ_{B}/Mn. The reduced Mn saturation moments in this family may arise from Mn disorders, which were observed in MnBi_{2}Te_{4}^{10}.
Figure 1f shows the temperature dependent inplane (ρ_{xx}) and outofplane resistivity (ρ_{zz}). Above 20 K, both ρ_{xx} and ρ_{zz} decrease nearly linearly upon cooling with ρ_{zz}/ρ_{xx}~53 at 300 K (Supplementary Fig. 3), suggesting a large transport anisotropy that is consistent with its vdW nature. With further cooling, ρ_{xx} and ρ_{zz} increase slightly, which is likely caused by the enhanced scattering from spin fluctuations, a phenomenon frequently observed in low dimensional magnetic materials^{30,31}. Then at 13 K, a sudden drop of ρ_{xx} and a sharp increase of ρ_{zz} are observed. This is in agreement with the Atype magnetic structure shown in Fig. 1a since the antiparallel alignment of Mn moments can reduce the conductivity via spinslip scattering, while parallel alignment of the Mn moments will eliminate such scattering and thus enhance the conductivity^{30}.
Figure 1g shows the transverse magnetoresistance (TMR), defined as MR = (ρ_{xx}(H)ρ_{xx}(0))/ρ_{xx}(0). The main feature of the figure is the overall W shape of the TMR. The W shape becomes deeper upon warming, with the largest negative TMR of 8% appearing at 12 K, which is close to T_{N}. Above T_{N}, it starts to become shallower and finally transforms into an ordinary parabolic shape at 50 K. The overall W shape can be understood in the framework of FM fluctuations. Above 50 K, the lack of magnetic fluctuations leads to the parabolic TMR. Upon cooling, FM fluctuations begin to appear and become increasingly stronger with maxima around T_{N}. As a result, the summation of the positive parabolic TMR and the negative TMR arising from the FM fluctuations under fields leads to a progressively deeper W shape of TMR upon cooling. Below T_{N}, the FM fluctuations are reduced, but still with a strong presence, leading to the shallower W shape under field.
The spinflip transition strongly affects the transport properties, as shown in Fig. 2. ρ_{xx}(H), ρ_{zz}(H) and ρ_{xy}(H) follow the same hysteresis as that in M(H). With H  c, the transverse magnetoresistivity of ρ_{xx} with I  ab (Fig. 2a) and the longitudinal magnetoresistivity of ρ_{zz} with I  c (Fig. 2b) slightly change between 0 T to H_{f}. Then up to H_{c}, since the system enters the forced FM state and the loss of spin scattering occurs, ρ_{xx }drops by 3.8% whereas ρ_{zz} decreases by 34%. With H  ab, up to the saturation field of 1.0 T, ρ_{zz} (Fig. 2e) decreases by 39% whereas ρ_{xx} (Fig. 2f) drops by 2.6%. Our data show that the transition from AFM to FM spin alignment along the c axis has much stronger effect on ρ_{zz} than ρ_{xx}. MnBi_{4}Te_{7} displays evident anomalous Hall effect (AHE) as seen in the bottom panel of Fig. 2a. Our ρ_{xy}(H) is linear up to 9 T above 50 K (Supplementary Fig. 3), suggesting single band transport here. Using n = H/eρ_{xy}, our 50 K data corresponds to an electron carrier density of 2.84 × 10^{20} cm^{−3}, similar to that of MnBi_{2}Te_{4}^{17,18,32}. Our Hall resistivity below 13 K can be described by \(\rho _{{\mathrm{xy}}} = R_0H + \rho _{{\mathrm{xy}}}^{\mathrm{A}}\), where the R_{0}H is the trivial linear contribution and \(\rho _{{\mathrm{xy}}}^{\mathrm{A}}\) represents the anomalous Hall resistivity. At 2 K, \(\rho _{{\mathrm{xy}}}^{\mathrm{A}}\) is extracted to be 3.3 μΩ cm, which is half of the one in MnBi_{2}Te_{4}^{17}. Consequently, the anomalous Hall conductivity \(\sigma _{{\mathrm{xy}}}^{\mathrm{A}}\left( { = \rho _{{\mathrm{xy}}}^{\mathrm{A}}/\rho _{{\mathrm{xx}}}^2} \right)\) is 25.5 Ω^{−1} cm^{−1} and the anomalous Hall angle (AHA ~ \(\rho _{{\mathrm{xy}}}^{\mathrm{A}}/\rho _{{\mathrm{xx}}}\)) is ~1%.
Z_{2} AFM TI predicted by theoretical calculation
MnBi_{4}Te_{7} crystalizes in the space group (G) P3m1 (No. 164). By taking into account the Atype AFM, the primitive cell doubles along the c axis, rendering a magnetic space group P_{c}3c1 (No. 165.96) under the Belov–Neronova–Smirnova notation^{33}, as shown in Fig. 1a. This magnetic space group is derived from its nonmagnetic space group by adding an extra sublattice generated by an operation that combines timereversal T with a fractional translation τ_{1/2}. Then the full magnetic group is built as G_{M} = G + GS, where S is a combinatory symmetry S = Tτ_{1/2} with τ_{1/2} the half translation along the c axis of the AFM primitive cell. Although the explicit Tsymmetry is broken, the Ssymmetry (also referred to nonsymmorphic timereversal^{34}) still exists in bulk MnBi_{4}Te_{7}. In addition, MnBi_{4}Te_{7} has inversion symmetry P, while the square of the symmetry operator PS equals −1 at an arbitrary k in momentum space. Therefore, analogous to TI with Tsymmetry where Kramer’s degeneracy is induced by T^{2} = −1, in MnBi_{4}Te_{7} the existence of the PS symmetry ensures an equivalent Kramer’s degeneracy in the whole Brillion zone, and thus a Z_{2} topological classification.
Figure 3a shows the calculated band structure of bulk AFM MnBi_{4}Te_{7} with the presence of spin–orbit coupling (SOC). The conduction band minimum is located at the Γ point, while the valence band maximum in the vicinity of Γ shows a slightly curved feature. The calculated bulk band gap is about 160 meV. The projection of band eigenstates onto the porbitals of Bi and Te (as indicated by the blue and red coloring) clearly indicates an inverted order between several conduction and valence bands around the Γ point, which is strong evidence of the possible nontrivial topological nature. On the other hand, the Mn3d^{5} states form nearly flat bands far away from the Fermi level (Supplementary Fig. 6), indicating that the main effect of Mn is to break Tsymmetry by introducing staggered Zeeman field into the lowenergy Hamiltonian.
To determine the topological properties of AFM MnBi_{4}Te_{7}, we first apply the FuKane formula^{35} to calculate the Z_{2} invariant. The topological insulator phase of AFM materials is protected by Ssymmetry, under which there are only four invariant kpoints forming a 2D plane in the momentum space. Thus, analogous to weak Z_{2} indices in nonmagnetic materials, the Ssymmetry indeed protects weak Z_{2} topological phases in AFM materials. In AFM MnBi_{4}Te_{7}, four TRIM points, including Γ(0, 0, 0) and three equivalent M(π, 0, 0), need to be considered here with k τ_{1/2} = nπ. Due to the abovementioned band inversion at the Γ point, we find that the parities for the occupied bands at Γ are opposite to that of the other three M points, indicating a nontrivial Z_{2} = 1. To verify our results, we also calculate the evolution of Wannier charge centers (WCCs) using the Wilson loop approach^{36}. As show in Fig. 3b, the largest gap function and the WCCs line cross each other an odd number of times through the evolution, confirming that MnBi_{4}Te_{7} is indeed a Z_{2} AFM topological insulator. Compared with TIs with Tsymmetry, the protection of gapless surface states in AFM TIs requires that the cleaved surface respects Ssymmetry that contains translation along the c axis. Figure 3c clearly shows the gapless surface Dirac cone at the Γ point for the (010) surface, partially validating the bulksurface correspondence of MnBi_{4}Te_{7} as an AFM TI. The easycleaved (001) plane, where the Ssymmetry is broken, are measured by ARPES and compared with our theoretical calculations, as discussed in the following.
Surface and bulk states measured by ARPES
In contrast to the recently discovered AFM TI MnBi_{2}Te_{4} where only one type of surface termination exists, MnBi_{4}Te_{7} can terminate on two different sublattice surfaces on the (001) plane, i.e., the [Bi_{2}Te_{3}] QL termination and the [MnBi_{2}Te_{4}] SL termination, resulting in different surface states. ARPES with 47 eV, linear horizontal polarized light and a small beam spot reveals two different types of Ek maps by scanning across different parts of the sample in real space, as plotted in Fig. 4d, e and Fig. 4h, i. There are several distinguishing features between the two types of surface spectra: Fig. 4h, i appear to show a gap with massive quasiparticles while Fig. 4d, e show a sharp Diraclike crossing, possibly with a small gap. The spectra of Fig. 4d, e are reminiscent of recent high resolution ARPES spectra of the MnBi_{2}Te_{4} compound^{20,21,22,23} that show Diraclike spectra, and we assign these states to the [MnBi_{2}Te_{4}] SL termination, while we assign the other set of surface states to the [Bi_{2}Te_{3}] QL termination.
On these two terminations, symmetry operations combined with τ_{1/2} are not preserved. In the ideal case that the surface magnetic structure perfectly inherits the bulk property, due to the Atype outofplane magnetization of the Mn sublayers, the gapped surface states are described by adding an exchange term to the ordinary Rashbatype surface Hamiltonian for TI with Tsymmetry, i.e., \(H_{{\mathrm{surf}}}({\mathbf{k}}) = ( {\sigma _{\mathrm{x}}k_{\mathrm{y}}  \sigma _{\mathrm{y}}k_{\mathrm{x}}} ) + m_{{\mathrm{S}}/{\mathrm{Q}}}\sigma _{\mathrm{z}}\), where σ is the Pauli matrix for spin, and m_{S/Q} the surface exchange field that distinguishes the [MnBi_{2}Te_{4}] SL and [Bi_{2}Te_{3}] QL surfaces. Our calculation shows that the surface state terminated at the [Bi_{2}Te_{3}] QL has a massive Dirac cone with a surface gap around 60 meV (Fig. 4f, g), and an overall structure that agrees very well with the experimental data of Fig. 4h, i, confirming the assignment of the experimental data as arising from the [Bi_{2}Te_{3}] QL termination. When comparing Fig. 4i with the bulk states calculated by DFT (Fig. 4b), we can easily distinguish the surface states from the bulk states. To measure gap sizes in Fig. 4i, we extract an energy distribution curve (EDC) at the Γ point and fit it to several Voigt profiles, as shown in Fig. 4j. We find that despite the appearance of some spectral weight in the gapped region in Fig. 4i, the EDC does not show any signature of a peak in the gapped region, indicating that the surface state is gapped by ~100 meV while the bulk gap is nearly 225 meV.
The equivalent calculation on the [MnBi_{2}Te_{4}] SL termination is shown in Fig. 4a, c and does not agree well with the experimental data of Fig. 4d, e. While the theory shows that surface states merge with the bulk valence bands, the experiment suggests a Diraclike structure inside the gap. By taking full account of experimental resolution functions in both momentum directions and in energy, the ARPES data are consistent with either no gap or a maximum gap size of 10 meV. More details are given in Supplementary Note 3. A similar feature, i.e., nearly gapless surface Dirac cone at the SL termination, was observed recently in MnBi_{2}Te_{4} single crystals^{20,21,22,23}, where the deviation between ARPES and DFT calculation is suggested to be due to the surfacemediated spin reconstruction at the top layers of the [MnBi_{2}Te_{4}] SL termination.
Figure 5a, b shows stacks of measured isoenergy surfaces for the [MnBi_{2}Te_{4}] SL and [Bi_{2}Te_{3}] QL terminations over a wide range of energies both above and below the Dirac point, while Fig. 5c shows equivalent DFT calculations for the [Bi_{2}Te_{3}] QL termination. The sixfold symmetric isoenergy surfaces are seen in all cases, including the hexagonal warping or snowflake effect^{37}. We comment that while both terminations collapse to a single resolutionlimited point in kspace in the middle panels in Fig. 5a, b, this is expected whether or not there is a gapless or gapped Dirac point, due to the broad energy band width of the nearby valence and conduction bands (Fig. 4j).
Discussion
The vdW AFM TI MnBi_{4}Te_{7} single crystal reported here is in fact a 1:1 superlattice composing the building blocks of AFM TI [MnBi_{2}Te_{4}] and Tinvariant TI [Bi_{2}Te_{3}]. Our realization of the superlattice design has three advantages. First, as discussed above, it serves as a “buffer layer” that separates and thus effectively decreases the AFM coupling between the two neighboring [MnBi_{2}Te_{4}] SLs, leading to a weaker magnetic field to trigger the QAH. Second, by interlayer coupling between [Bi_{2}Te_{3}] QL and the adjacent [MnBi_{2}Te_{4}] SLs, the SOCinduced nontrivial topology of [Bi_{2}Te_{3}] ensures the band inversion in the 2D limit. As a result, QAH is well expected in fewlayer MnBi_{4}Te_{7}. Third, when MnBi_{4}Te_{7} is exfoliated into the 2D limit, natural heterostructures are made, which provides more 2D configurations than MnBi_{2}Te_{4} or Bi_{2}Te_{3} single crystal since the latter ones are only stacked by one type of building block. One can exfoliate MnBi_{4}Te_{7} with designed termination and different film thickness. For example, two types of threelayer systems with distinct topological properties, [MnBi_{2}Te_{4}]/[Bi_{2}Te_{3}]/[MnBi_{2}Te_{4}] and [Bi_{2}Te_{3}]/[MnBi_{2}Te_{4}]/[Bi_{2}Te_{3}], should be easily obtained by exfoliation. Recent calculations^{27} show that [MnBi_{2}Te_{4}]/[Bi_{2}Te_{3}]/[MnBi_{2}Te_{4}] is a QAH insulator if a small magnetic field around 0.2 T is applied to stabilize the forced FM phase. On the other hand, [Bi_{2}Te_{3}]/[MnBi_{2}Te_{4}]/[Bi_{2}Te_{3}] is suggested to be a QSH insulator with timereversalsymmetry breaking^{27} which cannot be achieved from the thin films of either MnBi_{2}Te_{4} or Bi_{2}Te_{3}. Therefore, the 2D version exfoliated from bulk vdW TI MnBi_{4}Te_{7} paves an avenue to chase the longsought emergent properties such as QAH effect and QSH effect. As the foundation of engineering 2D heterostructures, such topological vdW materials could open up unprecedented opportunities in discovering novel fundamental physics as well as making new quantum devices^{38}.
Methods
Sample growth and characterization
Single crystals of MnBi_{4}Te_{7} were grown using selfflux^{11}. Mn, Bi and Te elements are mixed so the molar ratio of MnTe: Bi_{2}Te_{3} is 15:85. The mixtures are loaded in a 2 mL crucibles, sealed in quartz tube, heated to and held at 900 °C for 5 h. After a quick cooling to 595 °C, the mixtures are slowly cooled down to 582 °C over one to three days, where sizable single crystals are obtained after centrifuging. Although Bi_{2}Te_{3} is the inevitable side product, we can differentiate MnBi_{4}Te_{7} pieces by measuring their (00l) diffraction peaks. In each growth, a few sizable platelike MnBi_{4}Te_{7} single crystals with typical dimensions of 3 × 3 × 0.5 mm^{3} were obtained.
To confirm the phase, Xray diffraction data were collected using a PANalytical Empyrean diffractometer (Cu Kα radiation). Samples used for powder Xray diffraction were ground into powder inside acetone to reduce the preferred orientation. Electric resistivity and heat capacity data were measured in a Quantum Design (QD) DynaCool Physical Properties Measurement System (DynaCool PPMS). The magnetization data were measured in a QD Magnetic Properties Measurement System (QD MPMS). All magnetic data were calculated assuming the molar ratio between MnBi_{4}Te_{7} and Bi_{2}Te_{3} impurity is 86:14 in the sample suggested by powder Xray refinement (Supplementary Fig. 1). Magnetic data measured for H  c were corrected with a demagnetization factor.
ARPES measurements
ARPES measurements on single crystals of MnBi_{4}Te_{7} were carried out at the Advanced Light Source beamline 7.0.2 with photon energies between 40 and 55 eV with linear horizontal polarization. Single crystal samples were topposted on the (001) surface, and cleaved insitu in an ultrahigh vacuum better than 4 × 10^{−11} Torr and a temperature of 15 K. ARPES spectra were taken at 12 K, slightly smaller than T_{N}. As the cleaved terrain is expected to consist of patches of exposed [Bi_{2}Te_{3}] QL and [MnBi_{2}Te_{4}] SL, to eliminate the effect of possible QL and SL mixing on the ARPES data, we scanned a 1 mm square surface of the sample in 50 μm steps with a 50 μm beam spot and collected spectra from over 200 different spots on the sample. We looked at each spectrum, finding many regions with clear, sharp features. We also narrowed the beam spot down to 20 μm × 20 μm and scanned more finely, in 15 µm steps, in smaller regions of interest. We found that there were regions on the order of 50 × 50 μm that were spectroscopically stable, meaning the ARPES spectra were not changing from spot to spot. We then took our data with a 20 μm × 20 μm beam spot and studied the centroid of the spectroscopically stable regions, which we believe will minimize any contamination due to another surface.
Firstprinciples calculations
We apply density functional theory (DFT) by using the projectoraugmented wave (PAW) pseudopotentials^{39} with the exchangecorrelation of Perdew–Burke–Ernzerhof (PBE) form^{40} and GGA + U^{41} approach within the Dudarev scheme^{42} as implemented in the Vienna abinitio Simulation Package (VASP)^{43}. The energy cutoff is chosen 1.5 times as large as the values recommended in relevant pseudopotentials. The U value is set to be 5 eV^{6}. The kpointsresolved value of BZ sampling is 0.02 × 2π Å^{−1}. The total energy minimization is performed with a tolerance of 10^{−6} eV. The crystal structure and atomic position are fully relaxed until the atomic force on each atom is <10^{−2} eV Å. SOC is included selfconsistently throughout the calculations. We constructed Wannier representations^{44,45} by projecting the Bloch states from the DFT calculations of bulk materials onto the Mn3d, Bi6p, and Te5p orbitals. The band spectra of the surface states are calculated in the tightbinding models constructed by these Wannier representations and by the iterative Green’s function technique as implemented in WannierTools package^{46}.
Data availability
Data supporting the findings in this study are available from the corresponding authors on reasonable request.
References
 1.
Mong, R. S. K., Essin, A. M. & Moore, J. E. Antiferromagnetic topological insulators. Phys. Rev. B 81, 245209 (2010).
 2.
Tokura, Y., Yasuda, K. & Tsukazaki, A. Magnetic topological insulators. Nat. Rev. Phys. 1, 126–143 (2019).
 3.
Chang, C.Z. et al. Experimental observation of the quantum anomalous Hall effect in a magnetic topological insulator. Science 340, 167–170 (2013).
 4.
Hirahara, T. et al. Largegap magnetic topological heterostructure formed by subsurface incorporation of a ferromagnetic layer. Nano Lett. 17, 3493–3500 (2017).
 5.
Hagmann, J. A. et al. Molecular beam epitaxy growth and structure of selfassembled Bi_{2}Se_{3}/Bi_{2}MnSe_{4} multilayer heterostructures. N. J. Phys. 19, 085002 (2017).
 6.
Otrokov, M. M. et al. Highlyordered wide bandgap materials for quantized anomalous Hall and magnetoelectric effects. 2D Mater. 4, 025082 (2017).
 7.
Gong, Y. et al. Experimental realization of an intrinsic magnetic topological insulator. Chin. Phys. Lett. 36, 076801 (2019).
 8.
Lee, D. S. et al. Crystal structure, properties and nanostructuring of a new layered chalcogenide semiconductor, Bi_{2}MnTe_{4}. CrystEngComm 15, 5532–5538 (2013).
 9.
Otrokov, M. M. et al. Prediction and observation of the first antiferromagnetic topological insulator. Preprint at https://arxiv.org/abs/1809.07389 (2018).
 10.
Zeugner, A. et al. Chemical aspects of the candidate antiferromagnetic topological insulator MnBi_{2}Te_{4}. Chem. Mater. 31, 2795–2806 (2019).
 11.
Yan, J.Q. et al. Crystal growth and magnetic structure of MnBi_{2}Te_{4}. Phys. Rev. Mater. 3, 064202 (2019).
 12.
Otrokov, M. et al. Unique thicknessdependent properties of the van der Waals interlayer antiferromagnet MnBi_{2}Te_{4} films. Phys. Rev. Lett. 122, 107202 (2019).
 13.
Zhang, D. et al. Topological axion states in the magnetic insulator MnBi_{2}Te_{4} with the quantized magnetoelectric effect. Phys. Rev. Lett. 122, 206401 (2019).
 14.
Li, J. et al. Intrinsic magnetic topological insulators in van der Waals layered MnBi_{2}Te_{4} family materials. Sci. Adv. 5, eaaw5685 (2019).
 15.
Deng, Y. et al. Magneticfieldinduced quantized anomalous Hall effect in intrinsic magnetic topological insulator MnBi_{2}Te_{4}. Preprint at https://arxiv.org/abs/1904.11468 (2019).
 16.
Liu, C. et al. Quantum phase transition from axion insulator to Chern insulator in MnBi_{2}Te_{4}. Preprint at https://arxiv.org/abs/1905.00715 (2019).
 17.
Lee, S. H. et al. Spin scattering and noncollinear spin structureinduced intrinsic anomalous Hall effect in antiferromagnetic topological insulator MnBi_{2}Te_{4}. Phys. Rev. Res. 1, 012011 (2019).
 18.
Yan, J. Q. et al. Evolution of structural, magnetic and transport properties in MnBi_{2x}Sb_{x}Te_{4}. Phys. Rev. B 100, 104409 (2019).
 19.
Chen, B. et al. Intrinsic magnetic topological insulator phases in the Sb doped MnBi_{2}Te_{4} bulks and thin flakes. Nat. Commun. 10, 1–8 (2019).
 20.
Hao, Y.J. et al. Gapless surface Dirac cone in antiferromagnetic topological insulator MnBi_{2}Te_{4}. Phys. Rev. X 9, 041038 (2019).
 21.
Chen, Y. et al. Topological electronic structure and its temperature evolution in antiferromagnetic topological insulator MnBi_{2}Te_{4}. Phys. Rev. X 9, 041040 (2019).
 22.
Swatek, P. et al. Gapless Dirac surface states in the antiferromagnetic topological insulator MnBi_{2}Te_{4}. Preprint at https://arxiv.org/abs/1907.09596 (2019).
 23.
Li, H. et al. Dirac surface states in intrinsic magnetic topological insulators EuSn_{2}As_{2} and MnBi_{2n}Te_{3n+1}. Phys. Rev. X 9, 041039 (2019).
 24.
Vidal, R. et al. Surface states and Rashbatype spin polarization in antiferromagnetic MnBi_{2}Te_{4} (0001). Phys. Rev. B 100, 121104 (2019).
 25.
Muff, S. et al. Separating the bulk and surface nto ptype transition in the topological insulator GeBi_{4x}Sb_{x}Te_{7}. Phys. Rev. B 88, 035407 (2013).
 26.
Aliev, Z. S. et al. Novel ternary layered manganese bismuth tellurides of the MnTeBi_{2}Te_{3} system: synthesis and crystal structure. J. Alloy. Compd. 789, 443–450 (2019).
 27.
Sun, H. et al. Rational design principles of the quantum anomalous hall effect in superlatticelike magnetic topological insulators. Phys. Rev. Lett. 123, 096401 (2019).
 28.
McGuire, M. A. et al. Magnetic behavior and spinlattice coupling in cleavable van der Waals layered CrCl_{3} crystals. Phys. Rev. Materials 1, 014001 (2017).
 29.
Liu, Y. et al. Anomalous Hall effect in the van der Waals bonded ferromagnet Fe_{3x}GeTe_{2}. Phys. Rev. B 97, 165415 (2018).
 30.
Masuda, H. et al. Quantum Hall effect in a bulk antiferromagnet EuMnBi_{2} with magnetically confined twodimensional Dirac fermions. Sci. Adv. 2, e1501117 (2016).
 31.
Shen, B. et al. Structural distortion and incommensurate noncollinear magnetism in EuAg_{4}As_{2}. Preprint at https://arxiv.org/abs/1809.07317 (2018).
 32.
Cui, J. et al. Transport properties of thin flakes of the antiferromagnetic topological insulator MnBi_{2}Te_{4}, Phys. Rev. B 99, 155125 (2019).
 33.
Belov, N., Neronova, N. & Smirnova, T. Shubnikov groups. Kristallografiya 2, 315–325 (1957).
 34.
Hua, G. et al. Dirac semimetal in typeIV magnetic space groups. Phys. Rev. B 98, 201116 (2018).
 35.
Fu, L. & Kane, C. L. Topological insulators with inversion symmetry. Phys. Rev. B 76, 045302 (2007).
 36.
Soluyanov, A. A. & Vanderbilt, D. Computing topological invariants without inversion symmetry. Phys. Rev. B 83, 235401 (2011).
 37.
Fu, L. Hexagonal warping effects in the surface states of the topological insulator Bi_{2}Te_{3}. Phys. Rev. Lett. 103, 266801 (2009).
 38.
Geim, A. K. & Grigorieva, I. V. Van der Waals heterostructures. Nature 499, 419 (2013).
 39.
Kresse, G. & Joubert, D. From ultrasoft pseudopotentials to the projector augmentedwave method. Phys. Rev. B 59, 1758 (1999).
 40.
Perdew, J. P., Burke, K. & Ernzerhof, M. Generalized gradient approximation made simple. Phys. Rev. Lett. 77, 3865–3868 (1996).
 41.
Anisimov, V. I., Zaanen, J. & Andersen, O. K. Band theory and Mott insulators: Hubbard U instead of Stoner I. Phys. Rev. B 44, 943 (1991).
 42.
Dudarev, S., Botton, G., Savrasov, S., Humphreys, C. & Sutton, A. Electronenergyloss spectra and the structural stability of nickel oxide: An LSDA+ U study. Phys. Rev. B 57, 1505 (1998).
 43.
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).
 44.
Mostofi, A. A. et al. An updated version ofwannier90: a tool for obtaining maximallylocalised Wannier functions. Comput. Phys. Commun. 185, 2309–2310 (2014).
 45.
Marzari, N. & Vanderbilt, D. Maximally localized generalized Wannier functions for composite energy bands. Phys. Rev. B 56, 12847 (1997).
 46.
Wu, Q., Zhang, S., Song, H.F., Troyer, M. & Soluyanov, A. A. WannierTools: an opensource software package for novel topological materials. Comput. Phys. Commun. 224, 405–416 (2018).
Acknowledgements
We thank Paul C. Canfield, Quansheng Wu, Suyang Xu, Filip Ronning and Chris Regan for helpful discussions, and Chris Jozwiak and Roland Koch at the Advanced Light Source for experimental help. Work at UCLA and UCSC was supported by the U.S. Department of Energy (DOE), Office of Science, Office of Basic Energy Sciences (BES) under award number DESC0011978 and DESC0017862, respectively. Work at CU Boulder was supported by the U.S. National Science FoundationDivision of Material Research under NSFDMR1534734. Work at SUSTech was supported by the NSFC under Grant No. 11874195, the Guangdong Provincial Key Laboratory of Computational Science and Material Design under Grant No. 2019B030301001, “Climbing Program” Special Funds under Grant No. pdjhb0448 and Center for Computational Science and Engineering of SUSTech. H.C. acknowledges the support from U.S. DOE BES Early Career Award KC0402010 under contract no. DEAC05 00OR22725. This research used resources of the Advanced Light Source, which is a DOE Office of Science User Facility under contract no. DEAC0205CH11231.
Author information
Affiliations
Contributions
N.N. conceived the idea and organized the research. N.N., Q.L., and D.D. supervised the research. C.H., J.L., E.E., H.B., and N.N. grew the bulk single crystal and carried out Xray and transport measurements. A.R. and C.H. performed magnetic measurements. K.G., X.Z., P.H., D.N., and D.D. carried out the ARPES measurements and data analysis. Q.L., P.L., H.S., and Y.L. performed the firstprinciples calculations. H.C., L.D. and C.H. carried out structure determination. N.N., Q.L., D.D., and K.G. prepared the manuscript with contributions from all authors.
Corresponding authors
Correspondence to Qihang Liu or Dan Dessau or Ni Ni.
Ethics declarations
Competing interests
The authors declare no competing interests.
Additional information
Peer review information Nature Communications thanks Guang Bian and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Hu, C., Gordon, K.N., Liu, P. et al. A van der Waals antiferromagnetic topological insulator with weak interlayer magnetic coupling. Nat Commun 11, 97 (2020). https://doi.org/10.1038/s4146701913814x
Received:
Accepted:
Published:
Further reading

Dynamical axion state with hidden pseudospin Chern numbers in MnBi2Te4 based heterostructures
Physical Review B (2020)

Crystal Chemistry and Bonding Patterns of BismuthBased Topological Insulators
Inorganic Chemistry (2020)
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.