WorksPhysics and Astronomy
Mapping out the emergence of topological features in the highly alloyed topological Kondo insulators SmB (=Eu, Ce)
Abstract
SmB6 is a strongly correlated material that has been attributed as a topological insulator and a Kondo insulator. Recent studies have found the topological surface states and low temperature insulating character to be profoundly robust against magnetic and non-magnetic impurities. Here, we use angle resolved photoemission spectroscopy to chart the evolution of topologically-linked electronic structure features versus magnetic doping and temperature in SmMB (M=Eu, Ce). Topological coherence phenomena are observed out to unprecedented ~30% Eu and 50% Ce concentrations that represent extreme nominal hole and electron doping, respectively. Theoretical analysis reveals that a recent re-designation of the topologically inverted band symmetries provides a natural route to reconciling the persistence of topological surface state emergence even as the insulating gap is lost through decoherence.
The compound SmB has been of great ongoing research interest as a Kondo insulator [1, 2], a mixed-valent system [3, 4, 5] and as the prototypical example of a strongly correlated topological insulator [6, 7, 8, 9, 10]. A topologically nontrivial interpretation of the band structure is supported by angle resolved photoemission spectroscopy (ARPES) investigations [10, 11, 12, 13, 14, 15], and is consistent with measurements of surface state spin texture [16]. However, the material remains enigmatic in many ways. Understanding of the low energy electronic band structure is fragmentary, and the symmetry character of the topological gap has recently been reattributed [17, 18, 19, 20, 21]. Moreover, studies have shown that the bulk insulating character and surface states are phenomenally robust and can persist in alloys doped at the 10s of percent level, even with impurities that are nominally charge-donating and magnetic [22, 23, 24, 25]. Together, these factors raise the fundamental question of how the electronic structure should be conceptualized at the boundary to the topological regime. Here we report ARPES measurements of band coherence phenomena in magnetically doped SmMB (=Eu, Ce), tracking the band evolution and emergence of surface states as a function of both temperature and sample composition. Theoretical analysis reveals that the recent symmetry attribution of the topological gap enables a framework for understanding the large doping, temperature, and energy scales of coherence effects in the band structure, which greatly exceed expectations based on the size and temperature dependence of the insulating gap.
For the easy-cleave [1] surface, the topological band inversion of SmB projects to the two dimensional (2D) -point and -point, meaning that topological surface states must lie between these momentum coordinates and the non-inverted -point [see Fig. 1(a)]. Band structure is relatively easy to observe along the momentum axis [13, 25], and can be understood from a tight binding model that includes one highly dispersive band derived from the Sm orbital, crossing the three bands (Figure 1(b)). The positive parity eigenvalue of the band introduces a symmetry inversion at the 3D -point, leading to a gapless surface state (green line) that appears inside the hybridization gap (see Fig. 1(c) slab calculation). Within the surface-sensitive ARPES spectral function, the surface state dispersion is only highly visible very close to the Fermi level (Figure 1(f)), and results in a zig-zag intensity contour that can be traced by applying a single-band fit to low temperature ARPES data (Figure 1(d)). Raising the temperature greatly increases the energy width of the -bands [14], causing them to vanish from the apparent band dispersions (Figure 1(d-Figure 1h)). The resulting spectrum shows only the band, which has a larger Fermi momentum than the low temperature surface states, and an enhanced group velocity due to the loss of hybridization.
Figure 1. Topological band structure of SmB. (a) A constant energy ARPES map of CeSmB in the plane ( eV photon energy) is shown at binding energy meV, just beneath the bands. Symmetry inversion regions surrounding the and point are delineated by the outer contour of the band, and have been highlighted in white. The point Fermi surface is not visible due to the choice of photon energy, and the point symmetry inversion region is shown with reduced size for visual clarity. (b) The bulk band structure from a tight binding model of SmB is color coded to show orbital symmetry. (c) Simulated two-dimensional band structure of a slab configuration with unit cells stacked along the -axis. In-gap topological surface states are indicted by the green solid line. (d-e) Experimental ARPES spectrum along the cut of CeSmB at and K. The intensity maximum as a function of binding energy is traced with the dashed line. (f-h) Simulated spectral functions at , , and K, projected to the top 3 unit cells of the crystal.
Single crystals of SmMB ( Eu, Ce) were prepared by the alumina flux method, and details of the growth method are described in Ref. [25]. High-resolution ARPES measurements were performed at beamline 4.0.3 at the Advanced Light Source and ESM beamline at NSLS-II, with a base pressure better than Torr. Both measurements were performed at eV corresponding to a bulk -plane () of the cubic Brillouin zone. X-ray diffraction characterization confirms single phase growth with lattice parameters that evolve monotonically with doping [see supplemental material (SM)] [26]. Extensive sample characterization can also be found in Ref. [25], including measurements of dopant homogeneity, crystallographic structure, impurity multiplet states, surface stability, and the 2D nature of surface states. All Eu alloyed samples retain a Kondo insulating resistivity trend [27], despite nominal hole doping by Eu as Eu[25]. Ce alloys are nominally electron doped, and have relatively flat low temperature resistivity by 30% doping that resembles the Kondo metal end point of CeB. Detailed doping- and temperature-resolved measurements were performed with an in-house ARPES instrument equipped with a He lamp at photon energy eV, for which the 2D -point is observed at Å[13]. Momentum space alignment was performed in a fixed sample geometry, using a DA30 electron deflector analyzer. The base pressure was better than Torr for the He lamp ARPES system. All samples were cleaved in situ K, and all ARPES data were obtained within 12 hours after cleavage. The tight binding model is based on Ref. [28], with details described in the SM [26]. We note that although the binding energy of the Dirac point can vary depending on surface modeling parameters, the ARPES spectral function is relatively insensitive to this factor [26].
An example of low temperature He lamp ARPES data from a highly alloyed SmB sample (SmCeB, referenced as Ce 0.5) along the direction is presented in Figure 2(a), and requires some contextual analysis to correctly interpret. The image shows the band dispersing through less visible non-dispersive -derived features. The outer contour of the band is clearly visible at Å, closely resembling the plane synchrotron data in Figure 1(e). This is surprisingly incompatible with the nominal surface-normal momentum of Å, where one expect to see the very edge of the electron pocket. Moreover, strong background intensity is present at smaller in-plane momenta of . Both of these effects can be understood by considering the limited resolution of ARPES. In the absence of -axis momentum resolution, the spectral function is expected to reflect the -projected density of states (DOS) profile of the band (blue curve in Figure 2(c)), which has pronounced peaks at the outer band contour (Å), corresponding to the band location in the plane.
Figure 2. Fits of He lamp ARPES data. (a) The symmetrized ARPES spectrum along obtained by He lamp for CeSmB. (b) Momentum distribution curves at energy , and meV are overlaid with a two-component fit decomposition. The fit includes (green shaded area) two identical Lorentz peaks and (gray shaded area) a negative-curvature parabolic background, convoluted by experimental resolution. (c) A decomposition of the momentum distribution curve at meV, showing (red) the fully modeled lineshape obtained from factoring in resolution, (blue) the underlying -projected DOS, (orange) an 2-Lorentzian fit of the outer intensity contour, which resembles a spectrum, and (black dashed) the difference between the fully modeled lineshape and the 2-Lorentzian approximation. This difference resembles a negative-curvature parabola, and motivates the two-component fitting approach in panel (b).
Modeling the resolution function as a Lorentzian (half-width of Å) causes these intense DOS peaks to show up as the dominant spectral feature, while photoemission from values near the nominal Å plane appears only as background. The resulting intensity profile closely matches experimental data (Figure 2(d)). Further contextual discussion of the ARPES resolution function can be found in the SM [26]. Note that the data are left-right symmetrized to restore the reflection symmetry of the spectral function, which is otherwise lost due to matrix elements associated with the geometry of the measurement. For continuum-like spectral features and highly disordered systems, matrix element effects can cause band dispersions to appear to differ slightly on the positive and negative momentum axes [29]. For fitting purposes, this resolution limited spectral function can be approximated by treating the outer edge of the spectral feature as a single band, and filling in the small-momentum background with a downward-opening parabola (see comparison in Figure 2(c)). Single band fits employing this simplified method are shown at multiple binding energies in Figure 2(b).
Measurements performed with the He lamp ARPES system as a function of temperature across the Eu and Ce doping series are shown in Figure 3. For all samples, the dispersions at high temperature (K) show a mostly smooth linear dispersion, when fitted as described above [see fit summaries in Figure 3(b)]. As temperature is lowered, a clear deviation from linearity can be observed in all samples at K, and the resulting intensity profile conforms to the zig-zag pattern that is expected upon the emergence of surface states. The linear band dispersion remains visible at energies meV beneath the Fermi level, however the group velocity (slope) of the band is reduced. At the K base temperature, the more highly doped samples (SmCeB, and SmEuB referenced as Ce 0.5 and Eu 0.3) show smaller deviations from the extrapolated band dispersion. The deviation of band momentum from the linear trajectory is tracked as a metric for identifying the emergence of surface states in Figure 4(b). The maximum deviation of band momentum () is shown as a function of temperature and doping. At low temperature, the amplitude of is relatively flat across the doping series from Ce 0.3 to Eu 0.2 with an average value of Å. Outside of this range, we observe a sudden change to Å and Å for Ce 0.5 and Eu 0.3 respectively. The thermal onset of behaves as a roughly linear function within each doping series, with an origin between and 180K for most samples, in keeping with the onset of surface states documented in ARPES measurements on the undoped compound.
Figure 3. Raw ARPES data and fitted dispersions. (a) Momentum distribution curves within 150 meV of the Fermi level are plotted with a 10 meV spacing for each doping series, at temperatures T=6, 80, and 180K. The Fermi level is indicated with a red curve, and a black dashed line traces the fitted peak positions. (b) Peak positions are plotted as a function of binding energy for each the doping series at T=6, 80 and 180K. Linear fits for the 5d dispersion are obtained from the -150 to -50 meV energy window, and shown as black dashed lines.
Figure 4. Identifying the topological Kondo cross-over. (a) The Fermi momentum observed at T=6K (blue) for each sample is compared with (black) the extrapolated T=6K Fermi momentum of the bulk band from Figure 3(b). (b) A map of surface state emergence, showing the maximum deviation between the fitted band contour and the extrapolated linear dispersion of the band near the Fermi level. Red and black dashed lines show expectations for surface state emergence based on treating the impurity scattering as a contribution to (black) imaginary and (red) real self energy, as discussed in the text. (c) The fitted band velocity in the -150 to -50 meV energy range, as a function of temperature. To reduce noise, 2-point averaging has been applied along the temperature axis.
The merging of the low temperature Fermi momentum with the extrapolated band dispersion at high doping does not necessarily indicate the loss of the topological surface state. Separately tracking the extrapolated Fermi momentum and the low temperature Fermi momentum in Figure 4(a) reveals that the near-intersection of these momenta is associated with different kinds of anomalies at the two ends of the doping series. In the case of Ce 0.5, the Fermi momentum appears to converge towards the band, consistent with the loss of the topological band coherence, and with nominal electron doping from Ce promoting a larger electron pocket. However, in the case of Eu doping, the Fermi momentum of the Eu 0.3 sample is sharply reduced, suggesting that the surface state may still be very robust but is simply not greatly offset from the band trajectory.
An interplay is also observed between the appearance of the surface state and renormalization of the band velocity. Fits of the band velocity at energies beneath the bands (from to meV) are shown in Figure 4(c). The bare band velocity obtained from overlaying theory with data on a large ( eV) energy scale is roughly eV-Å [26], and is reduced to eV-Å at low temperature. A factor of recovery of band velocity is observed upon heating to K for samples with lower doping (hollow circles), compared to an average factor of just for both of the highly doped Ce 0.5 and Eu 0.3 samples. The trend for lightly doped samples can be understood by noting that as the bands gain coherence at low temperature and a topological gap opens, the level repulsion effect of hybridization directly reduces the energy of states that draw close to the band beneath the Fermi level. The effect of this on the fitted band velocity within the modeled spectral function [from Fig. 1(f-h)] is plotted with a dashed curve, and accurately reproduces the convergence towards eV-Å in the K regime. The divergence of the prediction from the fit at K corresponds to a region in which the state width is much more difficult to characterize, due to the states being incompletely visible beneath the Fermi level and having poor contrast on top of the incoherent background intensity.
The loss of surface states in the undoped sample occurs near K [12, 14, 25], and has been associated with the increase of the -electron imaginary self energy beyond meV. The associated K crossover behavior and surface state emergence are theoretically reproduced in the Figure 1(g-Figure 1h) model, which is parameterized with self energies from Ref. [14]. This value is estimated as full width at half maximum of the fitted ARPES band feature, which shrinks to meV at low temperature. The low temperature f-electron energy width of Ce 0.3 and Eu 0.2 samples has been measured under similar conditions [25], and has values of and meV, respectively. Extrapolating linearly with doping, we naively expect Ce 0.5 to be well beyond the critical value ( meV), and Eu 0.3 to be just inside it ( meV), consistent with the apparent decline of topological surface state visibility in Ce 0.5.
The amplitude of this energy and temperature scale for losing topological surface states is perplexingly large if we associate surface topology with the insulating gap, which is filled in due to thermalization at a much lower temperature K [30] that marks the end of the exponential Kondo resistivity regime [1]. The insulating gap is consistently assigned relatively small values of meV [10] ( meV in Ref. [30]). These factors make it important to distinguish between the hybridization gap that introduces topology and the dispersion extrema that define conductivity.
The highest energy bulk valence band features identified by ARPES are local dispersion maxima projecting to the 2D point, which are expected to lie within a few millielectron volts of the lower edge of the insulating gap [13, 14, 21]. These features derive directly from the kinetic Hamiltonian of the orbital [see Figure 5(c)], and place a constraint on the $\Gamma_7$ orbital energy. However, recent spin texture and resonant spectroscopy analyses have revealed that the vacant $4f$ states above the Fermi level do not arise from the $\Gamma_7$, but rather from the $4f$ $\Gamma_8$ orbitals [17, 18, 19, 20, 21], and it has consequently been proposed that the topological physics should be interpreted with respect to a Hilbert space basis that neglects $\Gamma_7$. These two experimentally constraints are mutually satisfied when the $\Gamma_7$ and $\Gamma_8$ orbital energies are approximately equal, leading to a scenario in which all local maxima of the valence band emerge from the $\Gamma_7$ orbital [see {xref}Fig. 5(a) symmetry decomposition], but the $4f/5d$ hybridization that gives rise to the surface state is defined by the $\Gamma_8$ bands [see {xref}Fig. 5(b)]. A similar band structure is achieved by recent DFT+DMFT modeling in {citep}Ref. [21]`.
Figure 5. Distinct gaps for topology and conductivity. (a) Tight binding model band structure. Occupancy of orbital is shown by the thickness of a black overlay. Arrows indicate local maxima that roughly coincide with the 2D point. The 2D point is found at [0.25 0.25] in reciprocal lattice units, and is projected to from the midpoints of the - and - momentum segments. (b) The same model is shown with the states removed. (c) The dispersion is shown with all other orbitals removed. (d) The bulk band structure is shown from a 10-layer slab calculation. (e) The slab calculation is shown with the orbitals removed.
In spite of the vital role played by the electrons in defining the insulating gap, they are effectively a spectator to the surface topology, and can be removed without qualitatively modifying the surface state dispersion [see Fig. Fig. 5(d-e)]. The resulting picture includes a topological bulk gap of meV that is commensurate to the robustness of surface states against doping and temperature, and a much smaller insulating gap. A more detailed overview of the spectator nature of with respect to modeling parameters, surface spin texture, symmetry mixing, and bulk Berry curvature is found in the SM [26]. A remarkable feature of this electronic structure is that the insulating gap is not primarily a hybridization gap, contrary to the conventional expectation for Kondo insulators, but can be tuned over a broad range by changing the energy difference between the and orbitals without modifying the strength of hybridization.
In spite of heat and doping playing similar roles in broadening the spectral features, the relative lack of doping dependence in the coherence onset temperature [Fig. 4(b)] suggests that impurity scattering does not have a simple additive relationship with thermal decoherence. First principles-based DMFT simulations and ARPES investigations on undoped SmB have attributed f-electron imaginary self energy as growing linearly with temperature [14, 26]. Combining this with a scattering rate proportional to impurity density would yield a downward-opening ‘V’-shaped phase boundary [black dashed curve in Figure 4(b)] that is relatively inconsistent with the experimental results.
Moreover, contrary to expectations for a scenario with large imaginary self energy, the highly doped Eu 0.3 and Ce 0.5 samples have low temperature band velocities that are similar to the heavily reduced values seen in the K Kondo coherent regime for lightly doped samples [ eV-Å, see Figure 4(c)]. A likely explanation is that rather than generating incoherent scattering, impurity doping results in strong coherent local screening as suggested in recent theoretical work [5, 31], and should therefore be thought of as contributing to real self energy rather than imaginary [32]. As such, the self energy components would not be linearly additive. An example of how this may play out for the phase boundary is plotted as a red curve in Figure 4(b). For this curve, the energy width contribution from scattering is treated as a continuum of real-valued Gaussian-distributed energies of local domains, and imaginary self energy is temperature-linear. The phase boundary is associated with peak width at half maximum reaching a critical threshold that corresponds to the T=120K thermal self energy. Comparing with the imaginary impurity scattering self energy scenario (black curve), the real self energy scenario has a flat-topped phase dome, consistent with the experimental data.
Characterization by ARPES in Ref. [25] found the coherence length of Eu 0.2 surface state electrons and the size scale of spatial fluctuations seen in STM to both be just crystallographic unit cells, supporting a highly localized picture. The resistance of the topological surface state spectral features to combinations of strong disorder (as in Eu 0.3) and high temperature (K) may not be surprising if the surface state spectral feature can be thought of as emerging from semi-localized single-nanometer scale domains.
In summary, we have used ARPES to map the emergence of electronic coherence phenomena as a function of doping and temperature in (=Eu, Ce). Evidence of surface states is observed extending to 50% Ce and 30% Eu doping, raising the question of how a topological surface band structure at such high impurity density should be conceptualized, and positioning as a parent system for broad chemical exploration of strongly correlated topological physics. A crossover is found to occur beyond 30% Ce and 20% Eu doping, inducing the partial loss of surface state visibility and a significant change in the temperature dependence of hybridization. Analysis of the low energy band symmetries reveals that the topological robustness can be understood by attributing the insulating and topological gaps to band pairs originating from different sets of orbitals.
This has the further implication that the insulating gap of is not a hybridization gap as would be expected for most Kondo systems, and can be tuned by chemical changes that manipulate the relative energy of and orbitals. Our results shed light on the broader question of how topological insulator surface states are lost upon the incremental loss of coherence, and how the unique topological Kondo electron system realized by can be chemically tuned to explore phenomena such as the much-sought antiferromagnetic TI state that is thought to occur within Eu doped ( [25, 27]; is maximized at ).
This research used resources of the Advanced Light Source, a U.S. DOE Office of Science User Facility under Contract No. DE-AC02-05CH11231. Operation of the ESM beam line at the National Synchrotron Light Source is supported by DOE Office of Science User Facility Program operated for the DOE Office of Science by Brookhaven National Laboratory under Contract No. DEAC02-98CH10886. ARPES measurements at N. Y. U. were supported by the National Science Foundation under Grant No. DMR-2105081. M. S. S., B. Y. K., J. W. L., and B. K. C. were supported by National Research Foundation of Korea (NRF), funded by the Ministry of Science, ICT and Future Planning (No. NRF2017R1A2B2008538). L. M. is supported by the National Natural Science Foundation of China (Grants No. U2032156 and No. 12004071) and Natural Science Foundation of Jiangsu Province, China (Grant No. BK20200348).
References
References
- [1]A. Menth, E. Buehler, and T. H. Geballe, “Magnetic and semiconducting properties of smb₆,” Phys. Rev. Lett. 22, 295–297 (1969).DOI
- [2]R. L. Cohen, M. Eibschütz, and K. W. West, “Electronic and magnetic structure of smb₆,” Phys. Rev. Lett. 24, 383–386 (1970).
- [3]T Kasuya, “Mixed-valence state in smb₆,” Europhysics Letters (EPL) 26, 283–287 (1994).
- [4]Nicholas P. Butch, Johnprierre Paglione, Paul Chow, Yuming Xiao, Chris A. Marianetti, Corwin H. Booth, and Jason R. Jeffries, “Emergent mystery in the kondo insulator samarium hexaboride,” Phys. Rev. Lett. 116, 156401 (2016).
- [5]Brian Skinner, “Properties of the donor impurity band in mixed valence insulators,” Phys. Rev. Mater. 3, 104601 (2019).DOI
- [6]Maxim Dzero, Kai Sun, Victor Galitski, and Piers Coleman, “Topological kondo insulators,” Phys. Rev. Lett. 104, 106408 (2010).
- [7]Maxim Dzero, Kai Sun, Piers Coleman, and Victor Galitski, “Theory of topological kondo insulators,” Phys. Rev. B 85, 045130 (2012).DOI
- [8]F. Lu, J. Z. Zhao, H. Weng, Z. Fang, and X. Dai, “Correlated topological insulators with mixed valence,” Phys. Rev. Lett. 110, 096401 (2013).pubmed.ncbi.nlm.nih.gov/23496729
- [9]Maxim Dzero, Jing Xia, Victor Galitski, and Piers Coleman, “Topological kondo insulators,” Annu. Rev. Condens. Matter Phys. 7, 249–280 (2016), https://doi.org/10.1146/annurev-conmatphys-031214-014749.
- [10]Lu Li, Kai Sun, Caglayan Kurdak, and J. W. Allen, “Emergent mystery in the kondo insulator samarium hexaboride,” Nature Reviews 2, 463–479 (2020).DOI
- [11]M. Neupane, N. Alidoust, S. Y. Xu, T. Kondo, Y. Ishida, D. J. Kim, Chang Liu, I. Belopolski, Y. J. Jo, T. R. Chang, H. T. Jeng, T. Durakiewicz, L. Balicas, H. Lin, A. Bansil, S. Shin, Z. Fisk, and M. Z. Hasan, “Surface electronic structure of the topological kondo-insulator candidate correlated electron system smb₆,” Nat. Commun. 4, 2991 (2013).DOI
- [12]J. Jiang, S. Li, T. Zhang, Z. Sun, F. Chen, Z. R. Ye, M. Xu, Q. Ge, S. Y. Tan, X. Niu, M. Xia, B. P. Xie, Y. F. Li, X. H. Chen, H. H. Wen, and D. L. Feng, “Observation of possible topological in-gap surface states in the kondo insulator smb₆ by photoemission,” Nat. Commun. 4, 3010 (2013).DOI
- [13]Jonathan D. Denlinger, James W. Allen, Jeong-Soo Kang, Kai Sun, Byung-II Min, Dae-Jeong Kim, and Zachary Fisk, “SmB₆ photoemission: Past and present,” in Proceedings of the International Conference on Strongly Correlated Electron Systems (SCES2013) (2014) journals.jps.jp/doi/pdf/10.7566/JPSCP.3.017038.DOI
- [14]J. D. Denlinger, J. W. Allen, J. S. Kang, K. Sun, J. W. Kim, J. H. Shim, B. I. Min, Dae-Jeong Kim, and Z. Fisk, “Temperature dependence of linked gap and surface state evolution in the mixed valent topological insulator SmB₆,” (2014), arXiv:1312.6637 [cond-mat.str-el].arxiv.org/abs/1312.6637
- [15]Yoshiyuki Ohtsubo, Yuki Yamashita, Kenta Hagiwara, Shin-ichiro Ideta, Kiyohisa Tanaka, Ryu Yukawa, Koji Horiba, Hiroshi Kumigashira, Koji Miyamoto, Taichi Okuda, Wataru Hirano, Fumitoshi Iga, and Shin-ichi Kimura, “Non-trivial surface states of samarium hexaboride at the (111) surface,” Nat. Commun. 10, 2298 (2019).
- [16]N. Xu, P. K. Biswas, J. H. Dil, R. S. Dhaka, G. Landolt, S. Muff, C. E. Matt, X. Shi, N. C. Plumb, M. Radović, E. Pomjakushina, K. Conder, A. Amato, S. V. Borisenko, R. Yu, H. M. Weng, Z. Fang, X. Dai, J. Mesot, H. Ding, and M. Shi, “Direct observation of the spin texture in smb₆ as evidence of the topological kondo insulator,” Nat. Commun. 5, 4566 (2014).DOI
- [17]Pier Paolo Baruselli and Matthias Vojta, “Distinct topological crystalline phases in models for the strongly correlated topological insulator SmB₆,” Phys. Rev. Lett. 115, 156404 (2015).DOI
- [18]Markus Legner, Andreas Rüegg, and Manfred Sigrist, “Surface-state spin textures and mirror chern numbers in topological kondo insulators,” Phys. Rev. Lett. 115, 156405 (2015).DOI
- [19]Pier Paolo Baruselli and Matthias Vojta, “Spin textures on general surfaces of the correlated topological insulator SmB₆,” Phys. Rev. B 93, 195117 (2016).DOI
- [20]M. Sundermann, H. Yavaş, K. Chen, D. J. Kim, Z. Fisk, D. Kasatkin, M. W. Haverkort, P. Thalmeier, A. Seyering, and L. H. Tjeng, “4f crystal field ground state of the strongly correlated topological insulator SmB₆,” Phys. Rev. Lett. 120, 016402 (2018).DOI
- [21]Patrik Thunström and Karsten Held, “Topology of SmB₆ determined by dynamical mean field theory,” Phys. Rev. B 104, 075131 (2021).
- [22]D.-J. Kim, J. Xia, and Z. Fisk, “Topological surface state in the kondo insulator samarium hexaboride,” Nat. Mater. 13, 466 (2014).DOI
- [23]B. Y. Kang, C.-H. Min, S. S. Lee, M. S. Song, K. K. Cho, and B. K. Cho, “Magnetic and nonmagnetic doping dependence of the conducting surface states in smb₆,” Phys. Rev. B 94, 165102 (2016).
- [24]Y. S. Eo, A. Rakoski, J. Lucien, D. Mihaliov, C. Kurdak, P. F. S. Rosa, and Z. Fisk, “Transport gap in smb₆ protected against disorder,” Proc. Natl. Acad. Sci. U. S. A. 116, 12638 (2019).DOI
- [25]Lin Miao, Chul-Hee Min, Yishuai Xu, Zengle Huang, Erica C. Kotta, Rourav Basak, M. S. Song, B. Y. Kang, B. K. Cho, K. Kißner, Friedrich Reinert, Turgut Yilmaz, Elio Vescovo, Yi-De Chuang, Weida Wu, Jonathan D. Denlinger, and L. Andrew Wray, “Robust surface states and coherence phenomena in magnetically alloyed SmB₆,” Phys. Rev. Lett. 126, 136401 (2021).
- [26]See Supplemental Material [url] for additional modeling details and sample characterization, which includes Refs. [33–39].
- [27]Sunmog Yeo, Kimyung Song, Namjung Hur, Z. Fisk, and Pedro Schlottmann, “Effects of eu doping on SmB₆ single crystals,” Phys. Rev. B 85, 115125 (2012).DOI
- [28]Pier Paolo Baruselli and Matthias Vojta, “Scanning tunneling spectroscopy and surface quasiparticle interference in models for the strongly correlated topological insulators SmB₆ and PuB₆,” Phys. Rev. B 90, 201106(R) (2014).DOI
- [29]Lin Miao, Chul-Hee Min, Yishuai Xu, Wenhan Zhang, Daniel Older, S. Alexander Breitweiser, Erica Kotta, Takehito Suzuki Haowei He, Jonathan D. Denlinger, Rudro R. Biswas, Joseph Checkelsky, Weida Wu, and L. Andrew Wray, “Observation of a topological insulator dirac cone reshaped by non-magnetic impurity resonance,” npj Quantum Mater. 3, 29 (2018).arxiv.org/abs/1805.00084
- [30]Harris Pirie, Yu Liu, Anjan Soumyanarayanan, Pengcheng Chen, Yang He, M. M. Yee, P. F. S. Rosa, J. D. Thompson, Dae-Jeong Kim, Z. Fisk, Xi-angfeng Wang, Johnpierre Paglione, Dirk K. Morr, M. H. Hamidian, and Jennifer E. Hoffman, “Imaging emergent heavy dirac fermions of a topological kondo insulator,” Nat. Phys. 16, 52–56 (2020).DOI
- [31]Pier Paolo Baruselli and Matthias Vojta, “Kondo holes in topological kondo insulators: Spectral properties and surface quasiparticle interference,” Phys. Rev. B 89, 205105 (2014).DOI
- [32]R. Abou-Chacra, D. J. Thouless, and P. W. Anderson, “A selfconsistent theory of localization,” J. Phys. C: Solid State Phys. 6, 1734 (1973).