Landau Level Spectroscopy

• Landau level spectroscopy is a set of experimental and theoretical methods that quantize cyclotron motion into discrete energy levels under strong magnetic fields.
• It employs optical, tunneling, microwave, and Raman techniques to extract parameters like effective mass, carrier density, and Fermi velocity in materials such as graphene and topological insulators.
• The method also uncovers quantum signatures like Berry phase shifts and wave-function nodes, thereby diagnosing disorder, band inversion, and many-body interactions in complex systems.

Landau level spectroscopy is the set of experimental and theoretical methods that interrogate the discrete electronic states produced when carriers in a solid are subjected to a quantizing magnetic field. In this regime, cyclotron motion is quantized into Landau levels (LLs), and optical, tunneling, microwave, transport-adjacent, or polarization-resolved probes convert the underlying band structure into a measurable ladder of resonances, peaks, and fan diagrams. In modern condensed-matter physics, the technique is used not only to extract effective mass, carrier density, mobility, Fermi velocity, and band gap, but also to diagnose Berry-phase-shifted quantization, band inversion, wave-function structure, disorder, screening, and interaction effects in semiconductors, semimetals, graphene-based systems, topological materials, and other narrow- or zero-gap systems (Akrap et al., 26 Feb 2026).

1. Conceptual foundations

Landau quantization begins with cyclotron motion in a magnetic field. In the quasi-classical description, the cyclotron frequency is

$\omega_c=\frac{|q|B}{m},$

while in a crystal the appropriate quantity is the cyclotron mass

$m_c=\frac{\hbar^2}{2\pi}\frac{dA}{dE},$

with $A(E)$ the cyclotron-orbit area in reciprocal space (Akrap et al., 26 Feb 2026). The corresponding Landau-level degeneracy per unit area is

$\zeta=\frac{eB}{h},$

which explains why the spectral weight and filling of LLs evolve strongly with field (Akrap et al., 26 Feb 2026).

For ordinary Schrödinger-like carriers with parabolic dispersion, the Landau ladder is equally spaced: $$E_n=\hbar\omega_c\left(n+\frac12\right), \qquad \omega_c=\frac{eB}{m^\ast},$$ and spectroscopic transitions are therefore linear in $B$ to leading order (Luican et al., 2011). By contrast, for massless Dirac fermions in graphene and related systems,

$$E_N=E_D+\mathrm{sgn}(N)\,v_F\sqrt{2e\hbar B|N|}, \qquad N=0,\pm1,\pm2,\ldots,$$

so LL spacing is nonuniform and largest near the Dirac point (Luican et al., 2011). The field-independent $N=0$ level at $E_D$ is the defining spectroscopic signature of Dirac quantization in graphene and topological surface states (Luican et al., 2011, Hanaguri et al., 2010). In the massive-Dirac formulation emphasized by the recent review,

$E_c^n=-E_v^n=\sqrt{\Delta^2+v^2 2e\hbar nB},$

which interpolates between narrow-gap semiconductors and gapless Dirac systems (Akrap et al., 26 Feb 2026).

A central practical distinction of Landau level spectroscopy is that it is not limited to states at the Fermi energy. Because optical and tunneling probes directly measure transitions or local density of states, they can access occupied and unoccupied states and thereby reconstruct portions of the band structure away from $m_c=\frac{\hbar^2}{2\pi}\frac{dA}{dE},$0 (Akrap et al., 26 Feb 2026). This is one reason the method remains useful in semiconductors, semimetals, topological insulators, and graphene heterostructures alike.

2. Spectroscopic observables and inference from Landau fans

The basic observables of Landau level spectroscopy are resonance energies, field-dependent peak positions, linewidths, oscillator strengths, and their polarization dependence. In optical conductivity language, the real part of the response can be organized as a sum over inter-LL transitions weighted by LL degeneracy, spin and valley multiplicities, occupation factors, and optical matrix elements, with disorder replacing ideal delta functions by broadened line shapes (Akrap et al., 26 Feb 2026). In STM/STS, LLs appear as peaks in $m_c=\frac{\hbar^2}{2\pi}\frac{dA}{dE},$1, i.e. in the local density of states, and their field evolution can be read directly in bias–field maps (Luican et al., 2011, Pauly et al., 2015).

Landau-fan analysis typically proceeds by assigning LL indices and comparing measured peak energies against the scaling variable appropriate to the candidate dispersion. For graphene, plotting $m_c=\frac{\hbar^2}{2\pi}\frac{dA}{dE},$2 versus $m_c=\frac{\hbar^2}{2\pi}\frac{dA}{dE},$3 produces a straight line whose slope yields $m_c=\frac{\hbar^2}{2\pi}\frac{dA}{dE},$4 (Luican et al., 2011). For the Bi$m_c=\frac{\hbar^2}{2\pi}\frac{dA}{dE},$5Se$m_c=\frac{\hbar^2}{2\pi}\frac{dA}{dE},$6 surface state, the semiclassical relation

$m_c=\frac{\hbar^2}{2\pi}\frac{dA}{dE},$7

lets one convert LL indices into momenta and reconstruct the dispersion $m_c=\frac{\hbar^2}{2\pi}\frac{dA}{dE},$8 even when quasiparticle interference is ineffective (Hanaguri et al., 2010). In TaAs, the same Lifshitz–Onsager logic is used operationally to distinguish a conical $m_c=\frac{\hbar^2}{2\pi}\frac{dA}{dE},$9 response from narrow-gap parabolic bands associated with the $A(E)$0 region (Santos-Cottin et al., 2021).

The field dependence itself is often the first discriminator between band topologies. A $A(E)$1 law characterizes massless Dirac-like carriers in graphene, topological surface states, H-point graphite, and low-index Rashba-split BiTeI LLs (Faugeras et al., 2014, Hanaguri et al., 2010, Kühne et al., 2017, Bordács et al., 2013). Approximately linear-in-$A(E)$2 transitions indicate conventional parabolic bands or the Schrödinger limit of a massive-Dirac system (Santos-Cottin et al., 2021, Bovkun et al., 2017, Akrap et al., 26 Feb 2026). More exotic scaling can also occur. In monolayer FeSe, the shallow electron band near X/Y was reported to follow

$A(E)$3

and the authors connected that to a nonparabolic band dominated by a $A(E)$4 term (Huang et al., 2024). This suggests that Landau spectroscopy can resolve not only whether a band is parabolic or linear, but whether higher-order terms dominate the low-energy orbital quantization.

Linewidths and level visibility are themselves spectroscopic observables. In graphene on chlorinated SiO$A(E)$5, LLs were resolved only above about $A(E)$6, consistent with the magnetic length becoming smaller than the charge-puddle scale (Luican et al., 2011). In graphene disorder metrology, the onset field

$A(E)$7

was used to infer local disorder broadening and, through $A(E)$8 and $A(E)$9, quasiparticle lifetime and mean free path (Lu et al., 2015). In Sb$\zeta=\frac{eB}{h},$0Te$\zeta=\frac{eB}{h},$1, LL widths increased approximately linearly with $\zeta=\frac{eB}{h},$2, supporting dominant hot-quasiparticle decay by electron-electron scattering rather than disorder or phonons (Pauly et al., 2015).

3. Experimental modalities

Landau level spectroscopy is not a single technique but a family of spectroscopies adapted to different materials classes and energy scales. Broadband magneto-optical spectroscopy in transmission or reflection remains the canonical implementation in semiconductors, semimetals, graphene, and topological materials (Akrap et al., 26 Feb 2026). Reflection-type optical Hall effect spectroscopy extends this by measuring the Mueller matrix and fitting the full magneto-optical dielectric tensor, thereby distinguishing polarization-mode-preserving from polarization-mode-mixing LL transitions and extracting circular-polarization-averaged transition energies and left–right splittings (Kühne et al., 2017).

Scanning tunneling microscopy and spectroscopy provide a local alternative. In exfoliated graphene on chlorinated SiO$\zeta=\frac{eB}{h},$3, STM/STS in magnetic field revealed the single-layer graphene LL sequence on an insulating, gateable substrate, enabling simultaneous local spectroscopy and back-gate control of filling (Luican et al., 2011). The same local approach on Sb$\zeta=\frac{eB}{h},$4Te$\zeta=\frac{eB}{h},$5(0001) resolved topological surface-state LLs above $\zeta=\frac{eB}{h},$6, mapped spatial shifts of the Dirac point by the field-independent zeroth LL, and extracted distinct upper- and lower-cone velocities (Pauly et al., 2015). In Bi$\zeta=\frac{eB}{h},$7Se$\zeta=\frac{eB}{h},$8, low-temperature STS in field produced more than 20 LLs on the positive-energy side and enabled momentum-resolved reconstruction of the surface-state dispersion without relying on quasiparticle interference (Hanaguri et al., 2010).

Fourier-transform STS adds a complementary dimension by resolving the real-space internal structure of LL wave functions. For a disordered 2DES on Cs/InSb(110), the Fourier transform of the LDOS map obeyed

$\zeta=\frac{eB}{h},$9

with

$$E_n=\hbar\omega_c\left(n+\frac12\right), \qquad \omega_c=\frac{eB}{m^\ast},$$0

so the $$E_n=\hbar\omega_c\left(n+\frac12\right), \qquad \omega_c=\frac{eB}{m^\ast},$$1 radial minima in momentum space directly counted the $$E_n=\hbar\omega_c\left(n+\frac12\right), \qquad \omega_c=\frac{eB}{m^\ast},$$2 nodes of the cyclotron orbit (Hashimoto et al., 2012). This established FT-STS as a wave-function-sensitive form of LL spectroscopy rather than merely an energy probe.

Raman scattering accesses yet another sector of the LL spectrum. In quasi-neutral graphene, micro-magneto-Raman spectroscopy resolved zero-momentum electronic inter-LL excitations $$E_n=\hbar\omega_c\left(n+\frac12\right), \qquad \omega_c=\frac{eB}{m^\ast},$$3, and the associated velocity

$$E_n=\hbar\omega_c\left(n+\frac12\right), \qquad \omega_c=\frac{eB}{m^\ast},$$4

was found to depend on dielectric environment, LL index, and magnetic field, directly exposing electron-electron interaction effects and the failure of Kohn’s theorem for Dirac carriers (Faugeras et al., 2014).

Microwave cyclotron resonance can probe very low-energy LL transitions. In air-exposed Bi$$E_n=\hbar\omega_c\left(n+\frac12\right), \qquad \omega_c=\frac{eB}{m^\ast},$$5Te$$E_n=\hbar\omega_c\left(n+\frac12\right), \qquad \omega_c=\frac{eB}{m^\ast},$$6, X-band microwave spectroscopy at $$E_n=\hbar\omega_c\left(n+\frac12\right), \qquad \omega_c=\frac{eB}{m^\ast},$$7 resolved three resonances with $$E_n=\hbar\omega_c\left(n+\frac12\right), \qquad \omega_c=\frac{eB}{m^\ast},$$8 angular scaling, interpreted as adjacent transitions between LLs of relativistic surface fermions (Wolos et al., 2012). Contactless Shubnikov–de Haas oscillations measured in the same cavity provided the bulk Fermi-level position, allowing the surface-state occupancy to be interpreted in terms of surface band bending (Wolos et al., 2012).

A more recent tunneling implementation uses atomic defect quantum dots inside an hBN tunnel barrier as local spectrometers. In graphene, the dot chemical potential

$$E_n=\hbar\omega_c\left(n+\frac12\right), \qquad \omega_c=\frac{eB}{m^\ast},$$9

made resonant tunneling sensitive to both the local compressibility and the excitation spectrum, while fixed-density bias scans accessed broken-symmetry LL structure with minimal local screening (Keren et al., 2020). This suggests a new spectroscopic niche between planar tunneling, local SET electrometry, and STM.

4. Material platforms and characteristic signatures

Graphene remains the paradigmatic LL-spectroscopy platform. Its nonuniform Dirac LL spacing, gate tunability, and sensitivity to substrate disorder made it possible to study disorder thresholds, pinning of the chemical potential to LLs, and interaction-induced velocity renormalization near charge neutrality (Luican et al., 2011). Gate-dependent LL spectroscopy further evolved into a disorder metrology: local fluctuations were inferred from the LL onset field and linewidth, while global potential fluctuations were extracted from the tilt of gate-dependent LL staircases in graphene on SiO$B$0, in graphene with an intermediate graphene screening layer, and near hBN (Lu et al., 2015). In bilayer graphene, ultralow-temperature STM/STS resolved valley, spin, and orbital splittings of the lowest $B$1 manifold and directly visualized the orbital-polarized intermediate $B$2 state at atomic scale (Yin et al., 2019).

Topological insulators and related topological materials form a second major class. In Bi$B$3Se$B$4, LL spectroscopy established the field-independent $B$5 level at the Dirac point and reconstructed a convex surface-state $B$6 consistent with ARPES (Hanaguri et al., 2010). In Sb$B$7Te$B$8, the same method showed different upper- and lower-cone velocities and correlated local Dirac-point shifts with defect-induced potential fluctuations (Pauly et al., 2015). Magneto-optics on Bi$B$9Sb$$E_N=E_D+\mathrm{sgn}(N)\,v_F\sqrt{2e\hbar B|N|}, \qquad N=0,\pm1,\pm2,\ldots,$$0 detected multiple LL transitions all following $$E_N=E_D+\mathrm{sgn}(N)\,v_F\sqrt{2e\hbar B|N|}, \qquad N=0,\pm1,\pm2,\ldots,$$1, which the authors interpreted as evidence for optically active Dirac-like surface states in the bulk gap (Schafgans et al., 2012). In Bi$$E_N=E_D+\mathrm{sgn}(N)\,v_F\sqrt{2e\hbar B|N|}, \qquad N=0,\pm1,\pm2,\ldots,$$2Te$$E_N=E_D+\mathrm{sgn}(N)\,v_F\sqrt{2e\hbar B|N|}, \qquad N=0,\pm1,\pm2,\ldots,$$3, bulk infrared LL spectroscopy instead supported a direct-gap massive-Dirac picture with

$$E_N=E_D+\mathrm{sgn}(N)\,v_F\sqrt{2e\hbar B|N|}, \qquad N=0,\pm1,\pm2,\ldots,$$4

a gap $$E_N=E_D+\mathrm{sgn}(N)\,v_F\sqrt{2e\hbar B|N|}, \qquad N=0,\pm1,\pm2,\ldots,$$5, and no spectroscopic indication that the fundamental gap itself is inverted (Mohelsky et al., 2020). In the topological crystalline insulator Pb$$E_N=E_D+\mathrm{sgn}(N)\,v_F\sqrt{2e\hbar B|N|}, \qquad N=0,\pm1,\pm2,\ldots,$$6Sn$$E_N=E_D+\mathrm{sgn}(N)\,v_F\sqrt{2e\hbar B|N|}, \qquad N=0,\pm1,\pm2,\ldots,$$7Se, the spectra were described by a 3D massive-Dirac Hamiltonian with a negative quadratic coefficient $$E_N=E_D+\mathrm{sgn}(N)\,v_F\sqrt{2e\hbar B|N|}, \qquad N=0,\pm1,\pm2,\ldots,$$8, establishing band inversion in the bulk, while no unambiguous surface-state LL signal was identified (Tikuišis et al., 2021).

Weyl and Dirac semimetals add a third layer of complexity because multiple cones and trivial bands often coexist. In TaAs, magneto-optical LL spectroscopy identified a nearly isotropic conical response extending over more than $$E_N=E_D+\mathrm{sgn}(N)\,v_F\sqrt{2e\hbar B|N|}, \qquad N=0,\pm1,\pm2,\ldots,$$9, assigned to the $N=0$0 Weyl cones with velocity

$N=0$1

whereas the $N=0$2 region yielded narrow-gap, nearly parabolic optical transitions rather than a clean Weyl fan (Santos-Cottin et al., 2021). This shows that in multiband Weyl materials, LL spectroscopy can be selective: accessible cones may be only a subset of the band structure.

Strong spin-orbit systems without topological surface-state emphasis also exhibit distinct LL phenomenology. In BiTeI, infrared LL spectroscopy revealed $N=0$3-scaling transitions characteristic of Dirac-like electrons emerging from the inner Rashba-split conduction band, with the strongest $N=0$4 mode still observable at room temperature (Bordács et al., 2013). In proximitized graphene, analytic LL formulas for the continuum model with staggered potential, Rashba SOC, and sublattice-resolved intrinsic SOC showed that the symmetry and crossings of the low-energy fan can diagnose whether the intrinsic SOC is Kane–Mele-like or valley-Zeeman-like, and in the inverted regime the critical crossing field

$N=0$5

provides a route to extracting the Rashba term (Frank et al., 2020). In monolayer FeSe, STS-based LL spectroscopy was used to argue for Rashba-split hole bands and a shallow electron band with anomalous $N=0$6 scaling (Huang et al., 2024).

Bulk narrow-gap and quantum-well systems have long been central to the field and remain conceptually important. In ZrTe$N=0$7, transmission magneto-spectroscopy on thin flakes supported a 2D massive-Dirac interpretation with

$N=0$8

a gap $N=0$9, and circular-polarization-resolved splitting governed by electron-hole asymmetry and Zeeman terms in a BHZ-like model (Jiang et al., 2017). In p-type HgTe quantum wells, LL spectroscopy of valence bands showed that the standard axial-symmetric 4-band Kane model fails to explain forbidden transitions and avoided crossings, whereas including bulk and interface inversion asymmetry activates those lines and captures the valence-band LL mixing (Bovkun et al., 2017).

5. Disorder, interactions, and symmetry breaking

A major evolution of Landau level spectroscopy is its transformation from a band-parameter tool into a probe of disorder, screening, and many-body physics. In graphene on SiO$E_D$0, electron-hole puddles of size $E_D$1 set the threshold field for resolving LLs, and the line broadening corresponded to carrier lifetimes of order $E_D$2–$E_D$3 near $E_D$4 (Luican et al., 2011). In the screening study of graphene on SiO$E_D$5, a graphene buffer layer reduced the LL onset field from about $E_D$6 to about $E_D$7, narrowed the linewidth from about $E_D$8–$E_D$9 to about $E_c^n=-E_v^n=\sqrt{\Delta^2+v^2 2e\hbar nB},$0, increased the mean free path from about $E_c^n=-E_v^n=\sqrt{\Delta^2+v^2 2e\hbar nB},$1 to about $E_c^n=-E_v^n=\sqrt{\Delta^2+v^2 2e\hbar nB},$2, and reduced global potential fluctuations from $E_c^n=-E_v^n=\sqrt{\Delta^2+v^2 2e\hbar nB},$3 to $E_c^n=-E_v^n=\sqrt{\Delta^2+v^2 2e\hbar nB},$4; near hBN, the global fluctuation scale fell further to $E_c^n=-E_v^n=\sqrt{\Delta^2+v^2 2e\hbar nB},$5 while local broadening remained similar (Lu et al., 2015).

Interactions can alter not only widths and onset fields but the LL energies themselves. In graphene Raman spectroscopy, the velocity inferred from symmetric inter-LL excitations depended on dielectric environment, field, and LL index, contradicting a single-particle Dirac model with one fixed $E_c^n=-E_v^n=\sqrt{\Delta^2+v^2 2e\hbar nB},$6. The data were instead interpreted using self-energy and vertex corrections, with the crucial conceptual point that Kohn’s theorem does not protect these zero-momentum transitions in graphene’s nonparabolic Dirac spectrum (Faugeras et al., 2014). Gate-dependent STM/STS on graphene found that the LL spacing increased markedly as charge neutrality was approached, with the extracted $E_c^n=-E_v^n=\sqrt{\Delta^2+v^2 2e\hbar nB},$7 rising from roughly $E_c^n=-E_v^n=\sqrt{\Delta^2+v^2 2e\hbar nB},$8 at higher doping to about $E_c^n=-E_v^n=\sqrt{\Delta^2+v^2 2e\hbar nB},$9 near the Dirac point, consistent with interaction-induced renormalization of the Dirac cone (Luican et al., 2011).

Broken-symmetry LL structure provides a more direct spectroscopic window into correlated quantum Hall states. In bilayer graphene, both spin and orbital splittings of the lowest $m_c=\frac{\hbar^2}{2\pi}\frac{dA}{dE},$00 LL were enhanced at partial filling, and even the splitting of a fully empty valley-polarized LL was enhanced by partial filling of the opposite valley, indicating strong many-body coupling across internal quantum numbers (Yin et al., 2019). In quantum-dot-assisted tunneling spectroscopy of graphene, the onset of SU(4) degeneracy lifting was directly resolved in both occupied and unoccupied LL manifolds. Effective gaps in the $m_c=\frac{\hbar^2}{2\pi}\frac{dA}{dE},$01 and $m_c=\frac{\hbar^2}{2\pi}\frac{dA}{dE},$02 excited states evolved linearly with field and corresponded to phenomenological $m_c=\frac{\hbar^2}{2\pi}\frac{dA}{dE},$03 factors as high as about 160, while high-field measurements resolved a primary spin splitting and a secondary spin-dependent valley splitting of the $m_c=\frac{\hbar^2}{2\pi}\frac{dA}{dE},$04, $m_c=\frac{\hbar^2}{2\pi}\frac{dA}{dE},$05 manifold (Keren et al., 2020).

Disorder and interactions can also be disentangled through Fourier-space wave-function analysis. In the InSb 2DES study, disorder randomizes guiding-center motion in real space, but the LL form factor remains visible in momentum space as robust radial minima governed by the zeros of Laguerre polynomials (Hashimoto et al., 2012). This suggests a general methodological principle: when disorder entangles internal quantum structure with spatial meandering, Fourier analysis can separate them if they occupy distinct momentum scales.

6. Parameter extraction, interpretation, and limits

The practical value of Landau level spectroscopy lies in parameter extraction from highly structured spectra. Effective masses are obtained from linear-in-$m_c=\frac{\hbar^2}{2\pi}\frac{dA}{dE},$06 cyclotron resonance or from parabolic-sector LL fans (Akrap et al., 26 Feb 2026, Santos-Cottin et al., 2021). Fermi velocities are obtained from the slope of Dirac-fan plots, as in graphene, Bi$m_c=\frac{\hbar^2}{2\pi}\frac{dA}{dE},$07Se$m_c=\frac{\hbar^2}{2\pi}\frac{dA}{dE},$08, Sb$m_c=\frac{\hbar^2}{2\pi}\frac{dA}{dE},$09Te$m_c=\frac{\hbar^2}{2\pi}\frac{dA}{dE},$10, BiTeI, and multilayer graphene optical Hall spectra (Luican et al., 2011, Hanaguri et al., 2010, Pauly et al., 2015, Bordács et al., 2013, Kühne et al., 2017). Band gaps follow from zero-field extrapolations of inter-LL transitions in massive-Dirac systems such as Bi$m_c=\frac{\hbar^2}{2\pi}\frac{dA}{dE},$11Te$m_c=\frac{\hbar^2}{2\pi}\frac{dA}{dE},$12 and PbSnSe (Mohelsky et al., 2020, Tikuišis et al., 2021). In graphene and related 2D systems, carrier density can be inferred from filling-factor-controlled appearance of particular transitions and from gate-dependent LL plateaus (Akrap et al., 26 Feb 2026, Lu et al., 2015).

Yet the method has systematic limits. Multiband materials can exhibit overlapping LL fans and nearby trivial bands, as in TaAs and HgTe quantum wells (Santos-Cottin et al., 2021, Bovkun et al., 2017). Surface-state spectroscopy can be masked by bulk conductivity or band overlap, as in Bi$m_c=\frac{\hbar^2}{2\pi}\frac{dA}{dE},$13Se$m_c=\frac{\hbar^2}{2\pi}\frac{dA}{dE},$14, Bi$m_c=\frac{\hbar^2}{2\pi}\frac{dA}{dE},$15Te$m_c=\frac{\hbar^2}{2\pi}\frac{dA}{dE},$16, and PbSnSe (Hanaguri et al., 2010, Mohelsky et al., 2020, Tikuišis et al., 2021). Disorder broadening can obscure LLs entirely below a threshold field or smear nominally sharp chemical-potential jumps (Luican et al., 2011, Lu et al., 2015). Symmetry lowering can activate forbidden transitions and complicate simple selection-rule assignments, as shown explicitly for HgTe valence bands (Bovkun et al., 2017). And in several cases the evidence for interaction-induced renormalization remains indirect, inferred from effective parameters rather than full many-body line-shape analysis (Luican et al., 2011).

A broader implication is that Landau level spectroscopy now spans real-space, momentum-space, and polarization-resolved implementations. It can reveal universal LL energetics, internal wave-function nodes, and subtle symmetry or interaction effects within the same material class, depending on the probe. This suggests that the subject is no longer reducible to cyclotron resonance alone. A plausible synthesis is that contemporary Landau level spectroscopy functions as a unifying language for quantized-band diagnostics across conventional semiconductors, Dirac materials, Weyl systems, proximitized heterostructures, and low-density superconductors (Akrap et al., 26 Feb 2026). The technique’s continuing expansion into local tunneling, optical Hall effect, Raman, and quantum-dot-assisted tunneling further suggests that its future importance will lie as much in resolving internal quantum structure and interaction effects as in measuring masses and gaps.