[Paper Review] A consistent view of the samarium hexaboride terminations to resolve the nature of its surface states

Hartmut Herrmann, P. Hlawenka+12 more|arXiv (Cornell University)|Oct 31, 2018

Rare-earth and actinide compounds4 citations

TL;DR

This study reconciles conflicting STM and ARPES observations of SmB6 (100) surfaces by linking specific surface topographies to distinct crystal terminations. Using bias-dependent STM contrast and element-specific density of states, it identifies the unreconstructed (1×1) surface as boron-terminated and the (2×1) structure as samarium-terminated, resolving long-standing discrepancies and supporting a trivial origin for surface conductivity via Sm 4f-level shifts.

ABSTRACT

The research effort prompted by the prediction that SmB$_6$ could be the first topological Kondo insulator has produced a wealth of new results, though not all of these seem compatible. A major discrepancy exists between scanning tunneling microscopy / spectroscopy (STM/S) and angle-resolved photoemission spectroscopy (ARPES), because the two experimental methods suggest a very different number of terminations of the (100) surface with different properties. Here we tackle this issue in a combined STM/S and ARPES study. We find that two of the well-ordered topographies reported in earlier STM studies can be associated with the crystal terminations identified using photoemission. We further observe a reversal of the STM contrast with bias voltage for one of the topographies. We ascribe this result to a different energy dependence of Sm and B-derived states, and show that it can be used to obtain element specific images of SmB$_6$ and identify which topography belongs to which termination. We finally find STS results to support a modification of the low-energy electronic structure at the surface that has been proposed as the trivial origin of surface metallicity in this material.

Motivation & Objective

• To resolve the long-standing discrepancy between STM and ARPES results on SmB6 (100) surface terminations.
• To establish a direct link between well-ordered STM topographies and chemically distinct crystal terminations identified by photoemission.
• To clarify the origin of surface conductivity in SmB6 by testing the hypothesis of Sm 4f-level shifts as a trivial mechanism.
• To determine whether surface reconstructions like (2×1) are consistent with pure terminations or require complex structural models.

Proposed method

• Combined STM and ARPES experiments on cleaved surfaces from the same single-crystalline SmB6 sample.
• Bias-dependent STM measurements to probe energy-dependent tunneling currents from Sm and B-derived states.
• Analysis of angle-resolved photoemission spectra to identify Sm 4f and B 2p core levels and their termination-specific intensity shifts.
• Use of differential conductance (dI/dV) to compare local electronic structures of different topographies.
• Comparison of umklapp intensity in ARPES with STM topography to assess superstructure origins.
• Application of Fermi-Dirac correction to photoemission data to reveal state evolution near the Fermi level.

Experimental results

Research questions

• RQ1 Which STM topography corresponds to the B-terminated surface, and which to the Sm-terminated surface?
• RQ2 How does the bias-dependent STM contrast enable element-specific imaging of SmB6 surfaces?
• RQ3 Can the apparent contradiction between STM’s multiple topographies and ARPES’s two distinct terminations be reconciled?
• RQ4 What is the origin of the umklapp intensity observed in ARPES for the B-terminated surface?
• RQ5 Does the observed surface conductivity in SmB6 arise from a trivial shift of Sm 4f-like states rather than topological surface states?

Key findings

• The unreconstructed (1×1) STM topography is assigned to the B-terminated surface, based on a reversal of STM contrast with increasing bias voltage due to energy-dependent B 2p state contribution.
• The (2×1) STM topography is assigned to the Sm-terminated surface, consistent with the absence of B-state contribution and dominance of Sm 4f-derived states in tunneling.
• ARPES data confirm two distinct terminations with chemically pure Sm and B surface layers, supporting the assignment of STM topographies to specific terminations.
• The differential conductance of the (1×1) and (2×1) topographies shows distinct electronic structures, supporting a termination-dependent shift of Sm 4f-like intensity near the Fermi level.
• Umklapp intensity in ARPES for the B-terminated surface is observed but not fully explained, possibly due to subtle structural distortions like B-octahedral tilts not visible in STM.
• The study provides strong evidence that surface metallicity in SmB6 may arise from a trivial mechanism involving Sm 4f-level shifts, rather than topological surface states.