Can We Master the Exotic and Topological Phases of Quantum Matter?

In a basement laboratory cooled to a hundredth of a degree above absolute zero, electrons confined to a sheet of gallium arsenide two atoms thick are doing something that no reasonable person would predict from first principles. Squeezed by a strong magnetic field into a flat, frozen sea, they conspire — through nothing but their mutual repulsion — to produce particles that carry exactly one-third of an electron's charge. Not approximately a third. Not a third on average. Exactly a third, protected to many decimal places by nothing more than the shape of a wavefunction. If you tried to pull one of these fractional particles apart, you would not get two smaller pieces. You would get nothing, because the "particle" was never really there in the first place — it is a pattern in the collective motion of trillions of electrons, a knot tied in the fabric of a many-body quantum state.

This is the signature strangeness of topological matter: phases of matter defined not by what breaks — the melting of ice, the alignment of spins in a magnet — but by what cannot be undone without tearing the fabric itself. It is a shift as deep as the one from Newton to Einstein, except this time the terrain being remapped is the internal logic of solids, liquids, and quantum fluids. And buried inside this remapping is a prize with obvious technological weight: particles called non-abelian anyons, and in particular the ghostly Majorana zero mode, which physicists believe could encode quantum information in a form so robust that the errors plaguing every quantum computer built so far would largely cease to matter.

Whether we can actually manufacture, control, and exploit this kind of matter — reliably, at scale, on demand — is one of the great open questions of condensed matter physics. It sits at a strange crossroads: part of it is settled Nobel-winning science, part of it is a controversial frontier where claimed discoveries have been retracted in full public view, and part of it reaches outward to touch questions about high-temperature superconductivity and even the search for new particles beyond the Standard Model.

svg[Two routes to non-abelian matter: a fractional quantum Hall liquid whose bulk excitations are anyons, and a semiconductor–superconductor nanowire whose ends host Majorana zero modes.]

Both routes share the same deep logic: a bulk energy gap protects a set of special, zero-energy or fractional excitations whose properties are pinned by topology rather than by fine-tuning. Both connect to genuinely settled physics — the integer and fractional quantum Hall effects are among the most precisely measured phenomena in all of physics, so precise that the international resistance standard is defined via the von Klitzing constant. The theoretical existence of Majorana zero modes in idealized model Hamiltonians is mathematically airtight. What remains unsettled is whether real, messy, disordered materials realize these idealized models cleanly enough — and whether we can do things with the resulting excitations.

The trouble with hunting for Majorana zero modes is that nature is a superb forger. The most-used experimental signature — a peak in the differential conductance pinned at exactly zero bias voltage when you tunnel electrons into the end of a nanowire — is a real, predicted consequence of a Majorana mode. But it is also, maddeningly, a generic consequence of several mundane phenomena that have nothing to do with topology: disorder-induced Andreev bound states, smooth confinement potentials, and quasi-Majorana states formed by ordinary physics near the wire's ends that mimic quantization without any actual non-abelian character.

This is not a hypothetical worry; it played out publicly and, for the field, painfully. In 2018, a Delft University group led by Leo Kouwenhoven published a paper in Nature reporting a nearly perfectly quantized zero-bias conductance peak, of height $2e^2/h$, in an InSb nanowire device — a result trumpeted as strong evidence for Majorana modes because that specific quantization value is a nontrivial theoretical prediction, harder to fake than a peak alone. In 2021, after a routine internal data-integrity review flagged selective presentation of the data (only some traces, not the full dataset, were shown, and background subtraction that shaped the quantization was not disclosed), the paper was retracted by the authors themselves. This was, notably, science working as it should — the retraction came from within the field's own scrutiny, not from an outside debunking — but it left the community rattled and considerably more skeptical of zero-bias-peak evidence generally.

Since then, the field has tried to armor itself against being fooled again. Researchers now demand a much wider battery of tests: peak height quantized at $2e^2/h$ over an extended, "stable" range of magnetic field and gate voltage (not just a single lucky point), consistency across both ends of the wire simultaneously, and correlation between conductance and independent measurements like the Coulomb-blockade-based "topological gap protocol" championed by Microsoft's Station Q effort. In 2023, Microsoft researchers reported evidence for a topological phase in an improved nanowire device using exactly this multi-signature protocol, and in early 2025 the company announced a small semiconductor-based qubit device ("Majorana 1") they claimed showed a topological phase transition — a claim that drew immediate, sharp pushback from parts of the community, with several prominent theorists arguing the published evidence fell well short of demonstrating actual non-abelian statistics or the presence of Majorana modes at all, as opposed to a topologically trivial phase with similar transport signatures. As of this writing, that dispute is unresolved in the peer-reviewed literature; braiding — the operation that would most directly demonstrate non-abelian statistics — has not been convincingly shown in any solid-state platform.

There is a second, quieter place where the picture breaks down: even setting aside whether Majorana modes exist in a given device, nobody has performed a clean, statistically overwhelming demonstration of non-abelian braiding statistics itself — the actual signature that makes these particles interesting rather than merely exotic. Proposals exist (T-junction wire networks, measurement-based braiding using Coulomb-blockaded islands), and small-scale experiments have reported braiding-consistent phase measurements, but the community consensus is that a decisive, repeatable, loophole-free braiding experiment is still ahead of us, not behind us.

Finally, there's a structural difficulty that goes beyond any single experiment: quasiparticle poisoning. A topological qubit stores information in the fermion parity of a pair of Majoranas — even or odd occupation of the shared nonlocal mode. But stray quasiparticles from imperfect superconductors, or thermally excited quasiparticles hopping onto the wire, can flip that parity invisibly, corrupting the qubit without any local "measurement." Suppressing this below useful thresholds has proven to be a serious materials-science battle, tied up with disorder in epitaxial superconductor-semiconductor interfaces, subgap states, and quasiparticle trapping — problems that are unglamorous, incremental, and absolutely central.

There is no single leading theory here in the sense of a rival explanation for the same data; there is instead a landscape of competing platforms, each betting on a different physical system to host non-abelian order, plus a genuine scientific disagreement about how close any of them actually are.

Fractional quantum Hall states, especially $\nu = 5/2$. The $\nu = 5/2$ plateau, discovered in 1987 by Willett, Eisenstein, Störmer, Tsui, Gossard, and English, sits at an even denominator — forbidden for the simplest Laughlin-type states, which require odd denominators from Fermi statistics of the underlying electrons. The leading theoretical candidate, the Moore-Read Pfaffian state (and its particle-hole conjugate, the "anti-Pfaffian"), is built from a p-wave pairing of composite fermions and is predicted to host genuine non-abelian anyons with quantum dimension $\sqrt{2}$. Interferometry experiments — most notably a 2020 Fibonacci-anyon-adjacent series of measurements by the Weizmann Institute group of Mordechai Heiblum, using quantum-Hall interferometers and shot-noise thermometry — have produced results consistent with non-abelian statistics at $\nu=5/2$, reporting a fractional thermal Hall conductance quantized at a half-integer value in units expected for a non-abelian state. This is genuinely exciting evidence, arguably the closest thing the field has to a clean non-abelian signature. Its drawback is platform difficulty: fractional quantum Hall physics demands extremely high-mobility two-dimensional electron gases, millikelvin temperatures, and large magnetic fields, and manipulating individual anyons for computation (rather than just detecting their existence via bulk transport) is an enormous additional engineering step that nobody has achieved.

Semiconductor-superconductor nanowires (the Majorana wire program). As described above, this is the most industrially invested approach, championed heavily by Microsoft's quantum computing division alongside academic groups at Delft, Copenhagen, and elsewhere. Its appeal is that it maps onto a clean theoretical model (the Kitaev chain) with a small number of tunable parameters, and it uses relatively conventional semiconductor fabrication, meaning it could in principle scale using adapted chip-manufacturing techniques. Its trouble is exactly the forgery problem described above: the devices are exquisitely sensitive to disorder, and distinguishing genuine topological zero modes from trivial impostors has proven far harder than early theorists expected, contributing to the 2021 retraction and the ongoing dispute over recent Microsoft claims.

Iron-based and kagome superconductors, and vortex Majorana modes. Rather than nanowires, another approach looks for Majorana zero modes bound to the cores of magnetic vortices in intrinsically topological superconductors, such as the iron-based superconductor FeTe$_{0.55}$Se$_{0.45}$. Scanning tunneling microscopy studies (notably by groups including Hong Ding's and Jin-Feng Jia's in China, and Kobayashi and Hasan-affiliated groups) have reported zero-bias conductance peaks localized precisely at vortex cores, with some spatial and magnetic-field dependence favoring a genuine bound-state interpretation. This route sidesteps some of the fabrication-disorder issues of nanowires because the material is a single, cleaner crystalline compound, but it introduces its own difficulty: vortices are hard to move and braid in a controlled way, since manipulating individual magnetic flux vortices with the necessary precision is its own unsolved engineering problem.

Kitaev's honeycomb model and candidate quantum spin liquids. A conceptually distinct proposal, due again to Kitaev, shows that certain frustrated magnets on a honeycomb lattice with strongly anisotropic, bond-dependent exchange interactions can realize a quantum spin liquid whose emergent excitations, under an applied magnetic field, include non-abelian anyons described by the same Ising-type theory as the Moore-Read state. The candidate material $\alpha$-RuCl$_3$ has been a major experimental focus; a landmark 2018 paper from a UC Berkeley/Lawrence Berkeley/Seoul National University collaboration reported a thermal Hall conductance plateau, $\kappa_{xy}/T$, quantized at half the expected value for a chiral Majorana edge mode — exactly the signature predicted for the non-abelian Kitaev spin liquid phase. This result generated enormous excitement, but subsequent experiments by other groups have had difficulty reproducing the exact half-quantization in the same material, and the interpretation remains genuinely contested; some groups see the plateau, others see a similar-looking but non-quantized signal, and the community has not converged. This is a case where "where we know" and "where it breaks down" essentially collapse into the same still-open experiment.

Topological superconductivity from proximitized topological insulators. A fifth route, pursued by groups including those at Stanford and Tsinghua, uses the surface states of a topological insulator (itself a genuine, well-established topological phase of noninteracting electrons, distinct from but related in spirit to the quantum Hall effect) placed in contact with a conventional superconductor, aiming to induce a 2D analog of a p+ip superconductor whose vortices trap Majorana zero modes. This connects the Majorana search directly to the enormous, independently important field of topological band insulators, but the platform faces similar disorder and interface-quality challenges as the nanowires.

It is worth being honest about the sociology here, because it matters for how to read the science: this is a field where enormous industrial money (Microsoft's decade-plus, multi-hundred-million-dollar bet on topological qubits, alongside smaller but serious efforts elsewhere) creates real pressure toward optimistic announcements, and where the 2021 retraction has made peer reviewers and rival groups considerably more demanding of extraordinary evidence. That skepticism is healthy. It does not mean the underlying theory is wrong — Kitaev's mathematics, the Moore-Read wavefunction, and the topological classification of superconductors all rest on solid theoretical ground, largely independent of any specific experimental claim. What is genuinely open is whether any real material realizes these idealized possibilities cleanly enough to use.

Several distinct classes of experiment, in principle, could turn this from an open question into a closed one, and some are plausibly within reach on a timescale of years rather than decades.

The most direct test is a braiding experiment with a quantitative, model-independent signature. The theoretical prediction is precise: exchanging two Majorana-carrying vortices or wire-junctions should produce a specific, calculable modification to a subsequent interference or parity measurement — not just "some noise changed," but a predicted phase or outcome probability tied to the non-abelian character of the exchange. Proposed "T-junction" nanowire networks, where wires are arranged so that Majorana modes can be physically shuttled around each other by gate-voltage control, are designed for exactly this. A convincing demonstration would need repeated exchanges in different orders, showing non-commutativity explicitly (braid $A$ then $B$ gives a different outcome than $B$ then $A$), reproduced across multiple devices and multiple labs. Nothing at this level of rigor has yet been published; it is a reasonable bet for the next five to ten years if nanowire disorder problems continue improving, though it could also stall if material quality plateaus below the needed threshold.

A second decisive test is improved, artifact-resistant transport spectroscopy combined with independent thermodynamic probes. The topological gap protocol pursued by Microsoft, which requires quantized zero-bias conductance to persist as a stable plateau across a two-dimensional parameter space of magnetic field and gate voltage (rather than a lucky single point), raises the bar considerably against disorder-mimicry. Combining this with completely independent measurements — for instance Coulomb-blockade parity readout, or microwave-frequency measurements of the topological gap itself — that all must agree, shrinks the space for coincidental false positives. The current scientific controversy over Microsoft's 2025 device claims is precisely a disagreement about whether existing data meets this bar; resolution will likely come through independent replication by outside academic groups with access to similar materials, something that typically takes a few years given how specialized the fabrication is.

A third avenue, less flashy but potentially more clean in principle, is quantized thermal Hall conductance measurements in fractional quantum Hall and Kitaev spin-liquid candidates, of the kind reported at $\nu=5/2$ and in $\alpha$-RuCl$_3$. Thermal Hall quantization is a bulk topological measurement, less susceptible to some of the local-disorder pitfalls of tunneling spectroscopy, but it demands exquisite thermal isolation and calibration, and as the RuCl$_3$ saga shows, different groups using nominally similar samples have obtained conflicting results — suggesting either extreme sample sensitivity or subtle systematic effects that the field has not yet fully identified. Convergence of independent groups, using different crystal growth methods, on the same quantized value would be strong evidence; continued disagreement would suggest either that the phase is present only in a narrow, hard-to-control regime, or that non-topological contributions are being misattributed.

Finally, the most demanding but ultimately most convincing benchmark is operational: a small logical qubit built from topological hardware whose error rate, measured directly, is shown to decrease as the topological protection is engineered to be stronger (longer wire separation, cleaner superconducting gap), in the specific functional way theory predicts — rather than an error rate that merely happens to be low. This would be the condensed-matter analogue of demonstrating not just that a superconducting magnet exists, but that a working accelerator ring built from it delivers beams reliably. No group has come close to this milestone yet; realistic estimates from people inside the field range from "several years, if current disorder problems are solved" to "possibly never, if some overlooked materials obstruction exists" — and reasonable, well-informed physicists disagree sharply on which estimate is right.

The most quoted stake is quantum computing, and it is real: today's leading quantum computers, built from superconducting transmon qubits or trapped ions, are extraordinarily impressive but still require heavy software-level quantum error correction, encoding one robust "logical" qubit in dozens or hundreds of noisy physical qubits, because any stray photon or vibration can flip a bit. A genuine topological qubit would push much of that protection into the hardware itself: the information would be nonlocal by construction, so local noise literally cannot access it. Whether that promise is realizable, or whether real materials are too imperfect to ever reach the idealized protection of the theory, is exactly the unresolved question at the heart of this essay. If it works, it could be the difference between quantum computers that need millions of physical qubits for useful algorithms and machines needing orders of magnitude fewer.

But the stakes reach well past engineering. Topological order forced physicists to expand their entire ontology of what a "phase of matter" is. Before the 1980s, essentially every phase transition fit inside the Landau symmetry-breaking framework, part of the bedrock of condensed matter theory. The fractional quantum Hall effect showed that two phases can be locally indistinguishable — identical symmetry, identical local order parameters — and yet be fundamentally, provably different, distinguished only by global, topological data: ground-state degeneracy on a torus, edge-mode content, braiding statistics of excitations. This is now understood to be a genuinely new organizing principle sitting alongside Landau's, not replacing it, and it has reorganized how we think about insulators (topological insulators), superconductors (topological superconductors), and even certain magnets (quantum spin liquids) across all of many-body physics.

There are quieter but equally important connections outward. The search for unconventional pairing mechanisms that could produce topological superconductivity shares deep methodological terrain with the room-temperature-superconductivity frontier: both demand understanding strongly correlated electron pairing beyond the conventional phonon-mediated BCS picture, and insights into one sometimes illuminate the other — iron-based superconductors, for instance, are of interest both as candidate high-temperature superconductors and as hosts for vortex Majorana modes. And more speculatively, the anyons of two-dimensional topological order are a rare instance where physicists have built and measured, in an actual laboratory, particles obeying exotic statistics predicted decades earlier by pure mathematics — a small but genuine echo of the larger project, alive in beyond-the-standard-model physics, of testing whether nature's fundamental building blocks include structures beyond the familiar fermions and bosons. Majorana fermions themselves were first proposed by Ettore Majorana in 1937 as a possible description of a fundamental neutral fermion — potentially the neutrino — that is its own antiparticle; condensed matter's Majorana zero modes are not that particle, but emergent quasiparticle excitations that share its defining mathematical property, and the search for both is, in a real sense, a search for the same peculiar structure in different arenas of nature.

There is also a philosophical undertone worth naming honestly. Topological order suggests that some of the most robust, information-theoretically protected structures nature builds are not "things" in the ordinary sense at all — not local excitations of a field, not particles you could isolate and hold — but patterns in the correlations of many particles at once, existing only collectively, only nonlocally, only because of the topology of the whole system. That a society's most reliable form of quantum memory might have to be built this way, out of something that is nowhere in particular, is a strange and genuinely humbling lesson about what "solidity" and "robustness" mean at the deepest level.

The theoretical foundations were laid by a small set of physicists whose insights connected condensed matter to topology and to computer science.

physicist[Michael Freedman|born 1951|Fields Medalist mathematician who, working with Kitaev and others at Microsoft Station Q, helped formalize the mathematics of anyonic quantum computation and has led Microsoft's long-running experimental push toward topological qubits]

physicist[Leo Kouwenhoven|born 1963|Dutch experimental physicist at Delft University of Technology whose group reported early nanowire Majorana signatures, including the 2018 result later retracted in 2021 — an episode that reshaped evidentiary standards across the field]

Other essential contributors include Xiao-Gang Wen, who developed much of the mathematical language of topological order and its connection to long-range quantum entanglement; Duncan Haldane and David Thouless, who (with Kosterlitz) shared the 2016 Nobel Prize for foundational work on topological phases and topological phase transitions, including the Haldane model that first showed a Chern insulator could exist without any net magnetic field; Charles Kane and Eugene Mele, who predicted the quantum spin Hall effect and topological insulators; and Mordechai Heiblum's group at the Weizmann Institute, whose interferometry and noise measurements have produced some of the most direct evidence for anyonic and possibly non-abelian statistics in the fractional quantum Hall regime.

Stand back from the specific devices and disputes, and a clear shape emerges. We are certain that nature permits topological order: the fractional quantum Hall effect proved that beyond argument more than forty years ago, and its anyons are as real and as measured as any particle in condensed matter physics. We are reasonably confident, on strong theoretical grounds, that non-abelian anyons — and Majorana zero modes in particular — are consistent with quantum mechanics and should exist in the right material or engineered structure. What we do not have, despite a decade and more of an extraordinarily well-funded, well-motivated, technically brilliant search, is an experiment that everyone in the field agrees demonstrates non-abelian braiding beyond reasonable doubt, nor a working device that uses that braiding to store and manipulate quantum information reliably.

That gap between theoretical near-certainty and experimental contestation is the real frontier — not a question of whether the physics is possible, but of whether this universe's actual materials, with their disorder, their impurities, their stubborn refusal to be perfectly clean, will let us get close enough to the idealized theory to use it. It is a question as much about materials science, nanofabrication, and metrology as about fundamental physics, which is itself an interesting lesson: sometimes the deepest open questions in physics are not "what are the laws" but "can the laws, as we already understand them, actually be engineered."

Several sub-questions remain genuinely, specifically open:

• Does any real material or device unambiguously host non-abelian anyons, verified by a direct, loophole-free braiding measurement rather than indirect spectroscopic signatures?
• Can quasiparticle poisoning and disorder-induced trivial states be suppressed far enough, in scalable fabrication, to make topological qubits practically useful rather than merely theoretically possible?
• Is the $\nu=5/2$ fractional quantum Hall state truly the non-abelian Moore-Read Pfaffian, or one of several competing, topologically distinct candidate states consistent with current data?
• Why do thermal Hall measurements in candidate Kitaev spin liquids like $\alpha$-RuCl$_3$ disagree so sharply between research groups, and what does that disagreement reveal about the true nature of the phase?
• Can topological quantum computation ever be scaled to the thousands of coupled non-abelian anyons a useful algorithm would require, or does it hit a materials ceiling no amount of engineering can lift?

Ask the tutor: If a device shows a beautifully quantized zero-bias conductance peak but braiding is never demonstrated, how much should that actually shift our confidence that it hosts a genuine Majorana zero mode — and what would it take to convince you either way?