Topological quantum computing has a measurement problem before it has a computing problem. Its proposed hardware — anyons, quasiparticles that carry a fraction of an electron’s charge and remember the order in which they are exchanged — cannot be used until they can be found. Anyons are not particles sitting in a trap waiting to be addressed. They are collective rearrangements of many electrons, spread over a region of material, and in the fractional Chern insulators now realized in moiré semiconductors they are typically pinned by disorder at unknown locations. A new paper from Emory University, published in Physical Review Letters, proposes a way to locate them with light: attach an exciton, and read the shift in its spectrum.
The paper, “Exciton-Anyon Binding in Fractional Chern Insulators: Spectral Fingerprints” by Tianhong Lu and Luiz H. Santos, both of Emory’s physics department, first circulated as a preprint in January 2026 and appears in Physical Review Letters this month. Its claim is deliberately narrower than the headlines suggest, and more interesting for it. It does not report the detection of an anyon. It calculates, by exact diagonalization of a model system, that an exciton bound to a quasihole should produce a spectral signature — a redshift of the exciton’s energy, separated from the scattering continuum — with a magnitude that photoluminescence spectroscopy can resolve. If it holds up in real devices, optics gains not just a way to see that anyons exist, but a way to say where they are.
This is a theory result, on small lattices, at zero temperature, in a model that is not a specific device. What follows is a prediction with a number attached, not a measurement.
Why anyons need a locator
The case for anyons as quantum hardware rests on a property no ordinary qubit has: topological protection. An anyon-based qubit would encode information in the braiding history of quasiparticles — in how they wind around each other — which local disturbances cannot easily erase. As Santos puts it in Emory’s account of the work, “anyons can hide themselves and avoid being perturbed by many environmental disturbances,” which makes them “more stable than typical quantum materials.” Lu, the paper’s first author and an Emory Ph.D. candidate, states the long-term goal plainly: “The dream goal is the ability to manipulate anyons to achieve topological quantum computation.”
Manipulation presupposes location. The fractional Chern insulators at the center of this work — lattice analogues of the fractional quantum Hall states, realized without any external magnetic field in twisted bilayer MoTe2 and rhombohedral multilayer graphene — host quasiholes carrying charge e/3 at filling fraction ν = 1/3. But transport measurements establish fractional charge in aggregate; they do not point at an individual excitation. And the quasiholes themselves are localized by disorder at positions no one chose and no one knows. Santos’s summary of the new result is worth quoting exactly, because it names the capability being claimed: “Using a particular frequency of light, we’ve developed a sensor that can pinpoint the location of an anyon in a material.”
The model: a kagome lattice, a mobile exciton, a repulsive interaction
The calculation is built on a kagome-lattice fermion-boson model with three ingredients. First, an electronic fractional Chern insulator: spinless fermions in an isolated Chern band with Chern number 1, filled to ν = 1/3, stabilized by repulsive interactions. Second, a mobile exciton — treated as a single boson — whose dispersion can be tuned continuously from a parabolic band to a flat band by changing the sign of its hopping parameter. Third, a local repulsive coupling between the exciton and the electrons, modeling an interlayer exciton in a neighboring TMD heterobilayer whose electric dipole points away from the fractional Chern insulator layer.
The mechanism is the same counterintuitive one this desk described in its coverage of the authors’ companion MoTe2 study earlier this month: the exciton-electron interaction is repulsive, but a quasihole is a local deficit of electrons, and repulsion from electrons is equivalently attraction to an electron deficit. What the new paper adds is the spectral analysis — where the bound states sit in energy, and what an optical experiment would actually see.
The method is exact diagonalization on periodic lattices, principally 4 × 4 with five fermions, with band-projected calculations extended to 5 × 5 lattices reaching roughly a million states. The authors identify bound states by comparing spectra with and without an exciton as a quasihole is created — in practice, by adding one unit cell at fixed fermion number, moving the filling from ν = 1/3 to ν = 1/3 − δ. The diagnostic is a threshold energy: the sum of the exciton’s ground-state energy at ν = 1/3 and the quasihole energy, denoted ε_X^0 + Δ_qh. States below that line are bound; states above it belong to the scattering continuum. The bound states are there, and they are not marginal: mid-gap states appear between the spectral minimum at E/V_FF ≈ 0.02 and the bottom of the continuum at E/V_FF ≈ 0.08, and at moderate coupling, V_FB/V_FF = 0.4, the binding energy amounts to roughly 0.73 times the fractional Chern insulator gap. The ground state satisfies ε_X < ε_X^0 + Δ_qh: the bound configuration is energetically favored over the separated one.
The control knob: how heavy the exciton is
The paper’s most useful result for experimentalists is not a single number but a dependence — how binding responds to the exciton’s mobility. The boson hopping parameter t_B tunes the exciton’s kinetic energy, and its sign selects two qualitatively different regimes.
When t_B is positive, the exciton’s lowest band is exactly flat: zero group velocity, effectively infinite mass. A flat-band exciton pays almost no kinetic penalty for localizing onto a quasihole, so the binding energy is governed almost entirely by the interaction strength and depends only weakly on the hopping.
When t_B is negative, the lowest band is parabolic, and the hopping tunes the exciton’s effective mass directly. Here the competition is explicit: a light, dispersive exciton resists binding because localization costs kinetic energy; a heavy, weakly dispersive one binds more readily. The binding energy falls systematically with system size for sufficiently negative t_B, consistent with the exciton and quasihole decoupling in the thermodynamic limit. The ground-state exciton-anyon entanglement entropy tells the same story: it saturates at about 0.67 of its maximum for the flat-band case, but tends to zero for large negative t_B, where the ground state factorizes into an independent quasihole and a free exciton.
The crossover between the regimes is sharp enough to name. The paper identifies a scale t* ≈ −0.1 V_FF, roughly −2.1 meV, at which both the binding energy and the entanglement drop by half. That is a falsifiable landmark: an experiment that can tune exciton mass should see binding strengthen as the exciton is made heavy, and weaken as it is made light.
The predicted numbers: a millielectronvolt within reach of photoluminescence
The question that matters for the laboratory is the size of the shift in a real material. The paper translates its model units into estimates for the most concrete platform: twisted bilayer MoTe2 hosting a fractional Chern insulator, proximitized by an excitonic TMD heterobilayer such as MoSe2/WSe2 or WSe2/WS2.
The translation uses representative values: a moiré length a_M ≈ 6.8 nanometers, corresponding to a 3-degree twist angle; an effective dielectric constant ε/ε_0 ≈ 10; an interlayer exciton charge displacement ℓ ≈ 0.6 nanometers. From these, the fermion-fermion interaction scale comes out at V_FF ≈ 21.2 meV, and the exciton-electron interaction V_FB spans roughly 0.9 to 21.6 meV for layer separations d between 2 and 10 nanometers. At strong coupling, V_FB/V_FF ≈ 1, the binding energy reaches 0.55 to 0.85 times the fractional Chern insulator gap — which for the MoTe2 parameters corresponds to 0.8 to 1.2 meV.
That range is the paper’s headline number, and it is chosen for a reason: a shift of order a millielectronvolt sits comfortably within the resolution of photoluminescence spectroscopy, the workhorse technique for excitons in TMDs. The authors state the point directly — these effects are “within reach of photoluminescence spectroscopy” — and add a comparison that sharpens it: comparable binding scales have been reported in graphene fractional quantum Hall systems, but those required a magnetic field of 16 tesla. The fractional Chern insulator version, if realized, needs no field at all.
Two qualifications belong in the same paragraph as the number. First, the quasihole energy Δ_qh itself comes out nearly zero in the calculation, which the authors attribute to finite-size effects and to the well-known smallness of the quasihole gap relative to the charge-neutral gap; treating it as zero makes their binding estimate conservative, a lower bound on the magnitude. Second, the 0.8–1.2 meV range is a model estimate for a specific parameter window — separations of a few nanometers, strong coupling, heavy excitons — not a prediction that any MoTe2 device will show a 1 meV shift.
What the fingerprint would and would not prove
A spectral fingerprint is only useful if its meaning is unambiguous, and here the honest accounting cuts both ways. The predicted signature is a redshift of the exciton energy when it binds to a quasihole, with the bound state well separated from the scattering continuum. In a photoluminescence experiment, that would appear as a new spectral line, shifted by roughly a millielectronvolt from the unbound exciton line, appearing at fractional fillings where quasiholes exist.
But a redshift, on its own, proves very little. Excitons shift for many mundane reasons — screening by nearby charges, strain, local charge density, disorder — and the paper’s own mechanism begins from ordinary Coulomb repulsion. The fingerprint earns its name only if it passes a checklist. It should appear at fractional fillings and vanish at integer fillings, where no quasiholes exist. Its magnitude should grow as the exciton is made heavier and as the layer separation shrinks. It should sit at the predicted energy scale, of order a millielectronvolt, rather than drifting with unrelated sample parameters. And it should be distinguished from the related but distinct phenomenon reported earlier this year in twisted MoTe2 — anyon-trion binding, observed experimentally, in which a charged trion rather than a neutral exciton binds to a fractional excitation.
There is also a limit to what optics can establish. A photoluminescence redshift can show that something binds to the exciton at fractional filling with the right energy scale and parameter dependence. It cannot directly measure the fractional charge of the object bound, nor its exchange statistics. Those require local charge sensing or interferometry — the same tools that have always been the bottleneck. The exciton is a locator and a probe, not a charge meter or a braiding apparatus. The paper does not propose a qubit encoding, does not address braiding, and does not claim any gate operation. Its contribution is one step earlier in the chain: find the anyon first.
Where this sits in the authors’ program — and what comes next
Readers of this desk will recognize the research program. Earlier this month we covered Lu and Santos’s companion study, which modeled the same physics in a realistic twisted MoTe2 continuum model and predicted the bound anyon’s internal structure — a charge core of roughly e/3 with a threefold, moiré-imposed shape. The new Physical Review Letters paper is the earlier, more general work: submitted in January 2026, it establishes binding and its spectral consequences in a minimal kagome model. The two papers are complementary, not redundant: one asks what the bound state looks like in a specific material, the other asks what it looks like in a spectrum.
The authors sketch three directions, and they are worth listing because they mark the distance between this result and a topological qubit. First, extending the framework to the full set of fractional states in twisted bilayer MoTe2, and assessing how the quantum geometry of real Chern bands shapes the binding scales — the step that would turn a model estimate into a device design rule. Second, coupling excitons to non-Abelian fractional Chern insulator anyons, the class that could actually support topological qubits, as opposed to the Abelian quasiholes studied here. Nothing in this paper demonstrates non-Abelian behavior, and the optical interface to a non-Abelian anyon is a proposal, not a result. Third, exploring finite exciton densities, where exciton-exciton interactions meet the collective excitations of the fractional Chern insulator.
For now, the right summary is modest and precise. Two Emory theorists have calculated that a heavy, weakly dispersive exciton should bind to an e/3 quasihole in a fractional Chern insulator, producing a spectral line shifted by an estimated 0.8 to 1.2 meV in twisted MoTe2 heterostructures — a shift large enough for photoluminescence to resolve, at zero magnetic field, with a dependence on exciton mass and layer separation that gives experimenters a checklist rather than a single number to chase. If an experiment confirms the redshift and its parameter dependence, optics gains a locator for individual anyons. If it does not, the failure will itself be informative, because the prediction is specific enough to be wrong in identifiable ways. The sensor comes first; the qubit comes later, if at all.
Sources and further reading
- T. Lu and L. H. Santos, “Exciton-Anyon Binding in Fractional Chern Insulators: Spectral Fingerprints,” Physical Review Letters (2026). DOI: 10.1103/8bp8-g8g4. Preprint: arXiv:2601.14365 (submitted January 20, 2026; revised June 26, 2026). All binding energies, crossover scales, lattice parameters, and material estimates quoted above are taken from this paper.
- Emory University / phys.org, “Theoretical method uses light to locate anyons, potential quantum computing building blocks,” October 29, 2026 — source of the Santos and Lu quotations.
- T. Lu and L. H. Santos, “Anyon bound states in MoTe2 Fractional Chern Insulators,” arXiv:2610.22473 (October 15, 2026) — the companion continuum-model study covered by this desk on October 24, 2026.
- W. Li et al., Nature 651, 48 (2026): experimental observation of anyon-trion signatures in twisted MoTe2, the closest experimental analogue cited in the paper.
- Experimental realizations of fractional Chern insulators in twisted bilayer MoTe2: J. Cai et al., Nature 622, 63 (2023); Y. Zeng et al., Nature 622, 69 (2023); H. Park et al., Nature 622, 74 (2023); F. Xu et al., Physical Review X 13, 031037 (2023).
- Prior predictions of exciton-anyon binding in fractional quantum Hall systems: N. Mostaan et al., arXiv:2507.08933 (2025); G. Wagner and T. Neupert, arXiv:2507.08928 (2025).
- Note on verification: this article reports a theoretical prediction. No exciton-anyon bound state of the kind described has been experimentally demonstrated as of this writing, and all figures attributed to the Emory paper are results of model calculations on finite lattices, not measurements. The 0.8–1.2 meV binding estimate is the paper’s latest-version figure for twisted bilayer MoTe2 parameters; earlier preprint versions quoted 0.6–1.1 meV.