From Kondo Lattice to Spin-Textured Nematic Heavy Fermions in USbTe

What the Kondo lattice already explained, and where it stopped

The Kondo lattice model describes local moments hybridizing with itinerant electrons to form heavy quasiparticles, and it has been used for f-electron heavy fermions, transition metal dichalcogenides, and moiré systems [1]. Foundational work on magic-angle twisted bilayer graphene recast the system as a topological heavy-fermion problem with localized f electrons and conduction c electrons, and showed that ordered states compete with a symmetric Kondo state at integer fillings [7][8]. In parallel, a 2D Kondo lattice was realized in 1T/1H-TaSe2 heterobilayers, where sub-kelvin STM/STS found a split Kondo peak consistent with a coherent, magnetically ordered ground state rather than a simple Kondo paramagnet [9]. These studies established that Kondo coherence can coexist with magnetic order, but they did not resolve whether rotational symmetry breaking can emerge from the hybridization itself.

The prior frontier for nematic order came largely from iron-based superconductors, where an electronic mechanism of nematicity was debated and compared with other correlated systems including heavy-fermion materials [4]. That context made the USbTe result a direct test: can a Kondo lattice host a nematic state whose energy and temperature scales track heavy-quasiparticle formation rather than a separate structural or magnetic transition?

USbTe: nematicity appears where heavy quasiparticles form

In USbTe, a heavy-fermion ferromagnet with Curie temperature Tc = 125 K, STM/STS on the U termination visualized a stripe-like dI/dV pattern at the Fermi level oriented 45 degrees to the a and b axes, with domains rotated by 90 degrees [1]. The nematic signal is confined to about ±2 meV around the Fermi level and disappears by 15 K, matching the temperature and energy scale of a peak-dip feature in the tunneling spectra that is associated with heavy f-electron spectral weight and a partial hybridization gap [1]. A symmetry factor alpha, which quantifies fourfold rotational symmetry breaking in Fourier-transformed dI/dV maps, peaks near the Fermi level at 4.2 K and becomes featureless at 15 K [1].

Quasiparticle interference maps evolve from C4 to C2 symmetry toward the Fermi level, and the QPI dispersion along one diagonal becomes flat around 5 meV while the orthogonal cut remains light-band-like [1]. ARPES at 6 K shows a flat band at the Fermi level hybridizing with a conduction band, consistent with heavy quasiparticle formation [1]. The authors interpret these combined data as a heavy electronic liquid-crystal phase: the rotational symmetry breaking is tied to Kondo hybridization rather than to a pre-existing structural distortion [1].

Spin-polarized STM turns the nematic state magnetic

Using a nickel tip whose polarization can be flipped with ±0.5 T while the sample magnetization remains fixed, the authors observed a contrast reversal of dI/dV at the Fermi level on the U sites: tunneling is enhanced when tip and sample spins align and suppressed when they anti-align [1]. The reversal is robust across repeated field sequences and disappears with a nonmagnetic Pt-Ir tip, supporting a spin-channel origin [1]. Zoomed-in dI/dV maps show that the nematic stripe pattern is visible under one tip polarization but not the other, and the symmetry factor alpha peaks near zero bias only for the aligned polarization [1].

The paper's phenomenological model requires two hybridization components to break fourfold symmetry: a symmetry-preserving term that turns on at the Kondo coherence temperature and a B2g nematic order parameter that must condense through a phase transition [1]. Group theory identifies three pure B2g hybridization channels, all involving f-p orbital combinations, while f-d combinations cannot produce a pure nematic order parameter [1]. This is a mechanism proposal, not a settled identification: the authors note that further experiments are needed to determine whether the nematic phase preserves inversion and vertical mirrors, which would distinguish B2g, Eg, and Eu scenarios [1].

How the USbTe claim compares with other Kondo platforms

The closest conceptual relatives are artificial Kondo lattices. In 1T/1H-TaSe2, a split Kondo peak at 340 mK was interpreted as a coherent Kondo lattice with in-plane magnetic order on the magnetic side of the Doniach phase diagram [9]. In monolayer VSe2 on NbSe2, a uniform Kondo resonance with a Kondo temperature of about 44 K coexists with a charge density wave and a superconducting proximity gap [10]. In CsCr6Sb6, a kagome Kondo lattice shows heavy fermions with effective mass over 100 times the vanadium counterpart, Kondo insulating behavior at ultralow carrier density, and dimensionality-induced Kondo breakdown [12]. These systems establish that Kondo coherence, magnetism, and nontrivial band topology can coexist, but none reports a spin-polarized nematic state tied to hybridization.

Kagome and topological heavy-fermion work provides a different comparison. YbCr6Ge6 hosts coexisting kagome flat bands and Yb 4f Kondo resonance states, with filling-tunable weak and strong topological Kondo insulating regimes and a Dirac-Kondo semimetal [2]. That system shows how flat-band geometry and f-electron correlations can produce topological heavy fermions, but its symmetry-breaking physics is not nematic [2]. The USbTe result is therefore not a generic Kondo-lattice phenomenon: it is a specific observation in one layered Kondo magnet, and the paper's own evidence boundary is that the mechanism cannot be directly generalized to all Kondo lattice materials [1].

The broader correlated-electron context supports the significance without settling the mechanism. A community review identifies quantum criticality beyond the Landau-Ginzburg-Wilson paradigm and correlated topological matter as major open problems, and notes that nematic order can be continuous in some correlated systems [3]. Quantum criticality in multipolar Kondo lattices has been argued to involve intertwined spin and orbital channels with two distinct Kondo-entanglement destruction points [11]. USbTe adds a candidate platform where nematic symmetry breaking, spin polarization, and Kondo hybridization meet, but whether it hosts a quantum critical point and how its order parameter couples to the lattice remain open [1].

What validates the approach and what remains uncertain

The spin-polarized STM methodology is validated by single-atom Kondo studies. Individual Sm adatoms on graphene/Ir(111) show a Kondo resonance whose Zeeman splitting reveals a large g-factor, and cotunneling analysis identifies the charge state and crystal-field levels [5]. That work demonstrates that spin-resolved tunneling can extract magnetic information from Kondo features, supporting the logic of using a magnetic tip to test spin polarization in USbTe [1][5]. However, the USbTe measurement uses a soft magnetic tip and a ferromagnetic sample with a coercive field of about 1.3 T, so the spin contrast is inferred from field-dependent tip polarization rather than from a direct single-spin measurement [1].

A key limitation comes from related itinerant magnets. In GdRu2Ge2, intra-unit-cell STM shows that multi-Q magnetic phases produce multi-Q LDOS patterns, and the modeling requires four-spin interactions beyond RKKY to explain the double-Q ground state [6]. This is a reminder that spin and charge textures in Kondo-lattice-like magnets can involve mechanisms beyond simple hybridization, and that LDOS patterns may reflect complex magnetic reconstructions rather than a single order parameter [6]. For USbTe, the nematic state is observed in five samples and disappears above 15 K, but the paper does not establish whether time-reversal symmetry is broken by the nematic order itself or whether the spin polarization is a by-product of the underlying ferromagnetic local moments [1]. The theoretical model also allows several orbital channels, so the microscopic identity of the nematic order parameter remains to be pinned down [1].

About These Sources

This research page is built on 12 peer-reviewed studies — published from 2014 to 2026, 7 from 2024 or later, collectively cited 1,543 times — selected as the most relevant from 13 studies that passed quality screening, drawn from 72 papers retrieved from a database of over 500 million.

Sources used in this answer