Quantum Matter Frontiers and Open Problems

A frontier claim is not made trustworthy by being recent, surprising, or widely discussed. It becomes assessable when the physical system, prepared state, measured record, inference chain, alternatives, uncertainties, and stopping point are all explicit. This gateway supplies that discipline for quantum matter. It routes durable physics to its canonical teaching page and reserves dated status, change tracking, and living reviews for the site’s frontier volume.

The page is deliberately not a news feed or a second catalog of frontier topics. Its object is a transferable skill: turn a fast-moving material claim into a bounded record that another reader can inspect, reproduce, challenge, and update without silently changing the question.

Required background. Quantum Matter Map supplies the system–state–observable branches and the ability to locate a stable canonical owner.

Helpful background. How to Use This Volume develops curriculum and task routing. How Quantum Matter Is Measured separates detector records from observables, while Data Interpretation and Pitfalls develops cross-probe alternatives and artifact checks. No single probe, computational method, or specialist phase page is required for every branch.

Enter Quantum Matter Frontiers

A useful frontier question has a stable physical core and a genuinely moving edge. The stable core may be a response function, phase criterion, excitation, Hamiltonian, symmetry class, or probe forward model. The moving edge may be a candidate material, disputed mechanism, unreplicated signal, numerical extrapolation, engineering milestone, or proposed decisive experiment.

Begin by asking what kind of statement is actually under review:

• an observation about a calibrated record;
• an inference of an observable, parameter, excitation, or phase through a declared forward model;
• a mechanism claim explaining why the object occurs;
• a numerical claim about a finite model, method, or representation; or
• an engineering claim about control, yield, integration, or task value.

Universality describes the reach claimed for one of these statements. An open problem describes a precise unresolved question. Neither is a substitute for the claim class.

These claims do not inherit one another’s evidence automatically. A robust spectral feature need not identify a phase. A phase need not identify its mechanism. A record material property need not establish a scalable device. The central habit of this chapter is therefore to stop at the strongest layer that the evidence actually reaches.

Research-status words also answer a different question from publication or page-lifecycle words. Established denotes evidence that has survived the relevant checks. Active means that researchers are currently working on a question; it does not mean that one proposed answer is probably true. Conjectural, experimentally unresolved, benchmark dependent, and controversial each identify a different limitation. Attach a date to every such label because the label is designed to change.

By the end of the page, a reader should be able to accept a new claim, freeze its meaning, find the stable physics owner, audit the inference, identify a decisive next test, and name either a live frontier owner or an explicit coverage gap.

Separate Canonical Physics from Living Status

Three owners may be involved in one frontier question, and they should not be collapsed.

The stable concept owner defines the mathematical or physical object. It owns derivations, conventions, limiting cases, and durable diagnostic standards. For example, Strange Metals owns the meaning and limitations of a Planckian-rate extraction, while Topological Superconductors owns the BdG and boundary-evidence baseline for a topological-superconductivity claim.

The method or probe owner explains how a raw record becomes the quantity being compared. It owns calibration, response kernels, matrix elements, resolution, contacts, backgrounds, inversion, numerical convergence, and software-level reproducibility. A detector feature cannot skip this layer on its way to a phase claim.

The living frontier owner records which claims strengthened, weakened, or changed status; which open questions remain; and what changed during a dated review interval. Those pages belong in Research. The relevant Research entries are currently marked Planned: their dated assessments are not yet written. They map intended coverage; the stable treatments below supply the existing physics explanations.

This gateway owns the seam between the three. It does not reproduce the topical derivation, maintain a current material leaderboard, or keep a second annual archive. If no living frontier page is yet substantive, record that coverage gap and use the stable owner plus this claim-audit method.

Dates require the same care as owners. Distinguish the date of an event, the date a record was acquired, the date a paper or correction appeared, the evidence cutoff used in an assessment, and the date this page was reviewed. Changing one date without the others can make a stale status statement appear current.

Write the Ten-Field Frontier Claim Ledger

Every worked frontier record uses the same ten fields. A field may be marked not applicable, but only with a reason. It may not disappear silently.

1. Frontier question and canonical owner. State one enduring question, the stable concept or model it requires, and the exact live canonical route.

2. System, specimen, and prepared state. Record composition or platform, batch or device, geometry, dimension and boundaries, temperature, pressure, field, filling or doping, strain, disorder, drive, and preparation history.

3. Claimed object, claim class, and dated status. Name the phase, excitation, response, mechanism, or platform; identify whether the statement is an observation, inference, mechanism, numerical, or engineering claim; and attach the applicable frontier label and as-of date.

4. Scales, regime, and limits. Give the energy, temperature, frequency, time, and length windows; equilibrium or driven status; finite or thermodynamic target; and any noncommuting order of limits.

5. Observable, probe, and forward model. Identify the operator, acquisition, calibration, matrix elements, resolution, estimator, and map from the raw record to the claimed object.

6. Evidence provenance and reproduction. State the specimens and replicas, raw and processed artifacts, code and version, independent replication, and publication, correction, or retraction status.

7. Nuisance controls and alternatives. List systematics, null tests, false positives, competing phases or mechanisms, and relevant negative evidence.

8. Theory/computation validity and uncertainty. State assumptions and control, parameter provenance, convergence, fit covariance, model sensitivity, and the complete uncertainty budget.

9. Cross-evidence, progress marker, and falsifier. Name the independent evidence that must cohere, a decisive next test or calculation, and an explicit result that would weaken the claim.

10. Licensed conclusion, stopping point, ownership, and review. Write the strongest justified statement, what is not licensed, the live canonical route, the live frontier route or explicit coverage gap, and the review date, interval, and change note.

The ledger is intentionally more demanding than a bibliography or a list of pros and cons. It makes comparison possible without pretending that every uncertainty is statistical or that every disagreement concerns the same physical object.

Use a compact local vocabulary so that the ledger remains usable while the living frontier assessments are being developed. A claim class describes what the sentence is doing:

• observation records a calibrated feature without promoting it through a model;
• inference obtains a phase, excitation, or parameter through a declared forward model;
• mechanism proposes a causal microscopic explanation;
• numerical reports the result of a finite model, method, or representation; and
• engineering judges a platform or milestone against stated acceptance criteria.

Give a claim one primary class. Add a secondary class only when the sentence genuinely contains both. Universality and open-problem statements describe the content or reach of a claim; they do not replace this classification.

The allowed frontier-status labels are likewise precise:

• established means mature evidence or theory supports the scoped claim;
• standard means the object or method belongs to the accepted research toolkit;
• active means the topic is currently researched, not that the claim is probably true;
• conjectural means plausible but unsettled;
• speculative means weak, indirect, or strongly model-dependent evidence;
• controversial means credible experts disagree about interpretation or significance;
• engineering-limited means the principle is accepted but implementation is the present bottleneck;
• experimentally-unresolved means a decisive experiment is absent;
• theoretically-unresolved means competing or incomplete theoretical accounts remain;
• mathematically-open means a precise formal question remains unsolved;
• benchmark-dependent means the conclusion changes materially with the baseline, metric, or comparison set;
• claim-under-review means a challenge, correction, or consequential new result is being assessed;
• historical-frontier retains a claim for historical context rather than current status; and
• retired-frontier marks a former frontier as resolved, superseded, or abandoned.

These labels may be combined when their meanings are independently true. They classify a dated scientific claim and must never replace this page’s scalar lifecycle status.

Use evidence labels as a separate axis. definition, postulate, theorem, derivation, and standard-model-result identify theoretical provenance. accepted-experiment, reproduced-experiment, and single-experiment-claim distinguish empirical maturity. benchmark-claim, numerical-evidence, simulation-evidence, and engineering-demonstration identify computational or platform evidence. peer-reviewed-result, preprint-result, and review-article-summary identify publication form, not automatic evidential strength. expert-consensus, minority-view, and speculation identify interpretive status. commercial-claim and press-claim identify source provenance, not independent validation. Apply every relevant evidence label to a major claim; do not infer scientific status from venue or publicity.

Pass the Readiness and Evidence Gate

Before comparing claims, pass five gates.

Object gate. Can the claim be written as one physical statement with units, normalization, basis, sign, frame, geometry, window, and limiting procedure? If two groups mean different objects by the same word, a numerical comparison is premature.

State gate. Are composition, batch, boundaries, disorder, tuning, field, temperature, drive, and history aligned? Reprocessing one specimen several ways is not material replication, and samples from different phases cannot be averaged into one claim without a model of that variation.

Record gate. Is the reported quantity observed directly, or obtained by a forward model, inversion, fit, continuation, extrapolation, or finite-size limit? State the operator and the map before discussing agreement.

Evidence gate. Are sensitivity, resolution, nuisance controls, alternatives, covariance, convergence, and independent reproduction adequate for the claimed layer? A null result is bounded by its acceptance and power; it is not automatically evidence that an object is absent.

Ownership gate. Is the stable physics linked to its canonical page, and is changing status assigned to a live frontier page or marked as a coverage gap? A fashionable label is not a reason to create another permanent route.

When two aligned estimates $x_1$ and $x_2$ share calibration or analysis uncertainty, a standardized difference must retain their covariance:

$$z_\Delta := \frac{|x_1-x_2|} {\sqrt{\sigma_1^2+\sigma_2^2 -2\operatorname{Cov}(x_1,x_2)}}.$$

This number is not the probability that either scientific claim is true. It only standardizes a declared difference under the stated covariance model.

For data $d$, model $f(\theta)$, residual $r=d-f(\theta)$, and a validated covariance matrix $C$, one may write

$$\chi^2(\theta) := r^{\mathsf T}C^+r,$$

where $C^+$ is the inverse on the retained, numerically conditioned covariance subspace. Report that retained rank, nuisance parameters, constraints or priors, and residual structure. An acceptable residual does not establish identifiability or a unique mechanism.

If the same data vector and response object are fitted with likelihoods on the same scale, and a declared candidate set justifies an Akaike comparison, define

$$\begin{aligned} \operatorname{AIC}_i &:= 2k_i-2\ln L_{i,\max}, \\ \Delta_i &:= \operatorname{AIC}_i-\min_j\operatorname{AIC}_j, \\ w_i &:= \frac{e^{-\Delta_i/2}} {\sum_j e^{-\Delta_j/2}}. \end{aligned}$$

Akaike weights compare only those candidate models. They are not posterior model probabilities, and they cannot protect against omitted physics, a wrong likelihood, or an invalid noise model.

For the transport audit below, let $\rho(T)=\rho_{\rm fit}+AT$, where $\rho_{\rm fit}$ is the intercept of the declared finite-temperature fit and is not automatically a measured $T\to0$ residual resistivity. Under a one-band Drude map with temperature- independent $n$ and $m^\star$, the slope defines

$$\begin{aligned} \hbar\frac{d\tau_{\rm tr}^{-1}}{dT} &:= \alpha_{\rm tr}k_{\rm B}, \\ \alpha_{\rm tr} &:= \frac{\hbar n e^2 A} {m^\star k_{\rm B}}. \end{aligned}$$

Only with an additive residual-rate or Matthiessen assumption may this be integrated as

$$\hbar\!\left[ \tau_{\rm tr}^{-1}(T)-\tau_{\rm fit}^{-1} \right] = \alpha_{\rm tr}k_{\rm B}T, \qquad \tau_{\rm fit}^{-1} := \frac{ne^2\rho_{\rm fit}}{m^\star}.$$

The dimensionless $\alpha_{\rm tr}$ is therefore a rate-slope parameter. It inherits the carrier, mass, vertex, multiband, additivity, and geometry assumptions used in that conversion. An order-one value is not, by itself, a universal bound or microscopic mechanism.

For the optical audit, use the $e^{-i\omega t}$ convention. Quoted THz values are ordinary frequencies $f=\omega/(2\pi)$; $\Gamma$ below is an angular relaxation rate, so its value in THz is $\Gamma/(2\pi)$. For a stationary, time-translation-invariant response, an ideal stiffness term is distributionally

$$\sigma(\omega) := \pi D_s\delta(\omega) +iD_s\,\mathcal P\!\left(\frac{1}{\omega}\right) +\sigma_{\rm reg}(\omega),$$

where $\mathcal P$ denotes principal value. A narrow Drude term,

$$\sigma_D(\omega) := \frac{D_D}{\Gamma-i\omega},$$

also has $\sigma_{2,D}\simeq D_D/\omega$ over a finite window when $\Gamma\ll\omega$. A measured $1/\omega$ trend over a finite bandwidth is therefore not a standalone proof of phase stiffness or superconductivity.

A pump–probe experiment generally measures a two-time, delay-dependent response $\Sigma(\omega;t_d)$ rather than an instantaneous equilibrium-like $\sigma(\omega;t_d)$. Candidate models must be propagated through the same nonstationary response and detector kernel at the reflected- or transmitted- field level. Calling the result an instantaneous conductivity requires a validated quasistationary approximation.

Follow the Nonlinear Frontier Graph

Frontier reasoning is a graph, not a reading list with one compulsory order. Use the shortest branch that reaches the claimed object, then add method and evidence owners as required.

1. Freeze the claim. Fill fields 1–4 before searching for a favored explanation. Record what would count as changing the system or question.

2. Locate the stable owner. Use the Quantum Matter Map and the live routes below. A canonical page fixes the definition and diagnostic standard; it does not automatically settle a candidate material.

3. Attach the method owner. Add the probe, response, or computational page that maps the raw record to the reported observable. Preserve its conventions, calibration, and limits.

4. Audit alternatives and uncertainty. Separate measurement precision, sample variation, fit covariance, numerical convergence, and model spread. Do not combine unlike uncertainties into one reassuring error bar.

5. Write the bounded conclusion. State what the record licenses, what it does not license, the next discriminator, and the review trigger.

6. Hand off living status. Identify the relevant Research entry and check whether it contains a substantive dated assessment. An entry marked Planned identifies a coverage gap; it supplies no current evidence or scientific conclusion.

Several branches can meet. A proposed light-induced topological superconductor, for example, needs a driven-state owner, a pump–probe forward model, superconducting evidence, a topological invariant and boundary test, and finally a dated status assessment. No one page should absorb all five.

Route Correlations and Superconductivity Claims

High-temperature and unconventional superconductivity. Begin with Unconventional Superconductivity for fermionic antisymmetry, crystal and pseudospin pairing structure, nodes, multicomponent order, and the stable multi-probe evidence ladder. Use Competing Orders, Mott Insulators, and the relevant material pages for neighboring phases. The planned Research Unconventional Superconductivity Frontiers page, not a duplicate local leaf, will own dated material-family and microscopic-mechanism status.

Strange metals and Planckian language. Strange Metals owns the evidence ladder from low-temperature linear resistivity through model-dependent rate extraction, optics, thermodynamics, and competing mechanisms. Transport, Response, and Optics routes the actual conductivity calculation and measurement limits. Non-Fermi Liquids and Quantum Criticality own the corresponding stable mechanism classes. Dated comparative status will belong to those planned Research topics when live.

Quantum spin liquids. Quantum Spin Liquids owns the material-phase taxonomy and the evidence needed beyond absence of order. Fractionalization and Emergent Gauge Fields own the excitation and gauge descriptions. A continuum, missing Bragg peak, or low-temperature heat-capacity term remains compatible with disorder, weak order, freezing, phonons, or unresolved conventional modes until independent tests separate them.

For every correlation branch, distinguish a model realizing a phenomenon from a material establishing it. A Hubbard, Kondo, Kitaev, or SYK-like calculation can sharpen a mechanism and predict discriminators; agreement with one observable does not make the material identical to that model.

Route Topology, Fractionalization, and Moiré Claims

Anyons and non-Abelian evidence. Anyons and Braiding owns fusion spaces, braid operations, and interferometric logic. Topological Quantum Computation Bridge connects those data to computational operations. A localized zero-energy-like state does not establish a non-Abelian excitation, fusion rule, braid, or fault-tolerant qubit.

Topological-superconductivity platforms. Topological Superconductors owns the BdG invariant, bulk gap, boundary and vortex modes, and its evidence ladder. Proximity and Andreev Physics owns interface-induced pairing and ordinary Andreev alternatives. Platform status will ultimately split between the Research anyon and unconventional- superconductivity records rather than live in a second local platform page.

Fractional Chern claims. Moiré Topology owns the platform-specific route from Chern minibands to integer and fractional anomalous Hall evidence. Fractional Quantum Hall Effect and Topological Order own the stable many-body baseline. Fractional filling or a Hall-like plateau is not enough: require calibrated filling, low dissipation, thermodynamic incompressibility or a gap, Středa consistency where accessible, charge-order controls, and phase-specific evidence before approaching an anyon claim.

Moiré and flat-band claims. Moiré Superlattices owns the emergent cell, reconstruction, mini Brillouin zone, filling, and scale hierarchy. Flat Bands separates energy flatness from projector geometry and interaction dominance. A narrow band does not select one correlated phase; retain isolation, geometry, filling, interactions, disorder, and competing symmetry breaking.

The Topological Quantum Matter gateway routes stable topological objects. Future Research pages will track changing candidate and platform status. This page supplies the ownership and stopping logic between them.

Route Driven Matter, Design, and Technology Claims

Periodic and pump-prepared matter. Floquet Quantum Matter is the material-and-experiment bridge for periodic driving, quasienergy, occupation, heating, prethermal windows, and observable signatures. Driven-Dissipative Matter owns open steady states and attractors. A pump-prepared state must retain its envelope, occupation, dephasing, heating, spatial profile, and useful lifetime; it is not an equilibrium phase by relabeling.

Light-controlled claims. Pump–Probe Spectroscopy owns the time-domain acquisition and inversion workflow. Terahertz and Infrared Probes owns low-frequency electrodynamic calibration, while Raman and Optical Spectroscopy owns equilibrium optical channels. A transient gap-like edge, inductive trend, or coherence feature must survive finite bandwidth, depth inhomogeneity, heating, narrow-Drude, and spectral-redistribution alternatives.

Materials discovery. Quantum Materials by Design owns the closed loop from target property and representation through candidate generation, synthesis, characterization, and model revision. Computational Quantum Matter owns method selection, numerical tolerances, validation, and provenance. Machine-learning splits must exclude structure and composition duplicates, domain shift must be visible, and a ranked candidate must not be described as synthesized or property verified.

Technology claims. Nanostructures for Quantum Technology owns the material-to-device ledger. Materials and Fabrication Interface owns process, variability, yield, contamination, packaging, and hardware evidence. A material metric or device spectrum does not establish control, readout, error performance, integration, or task value.

Across these branches, a calculation, database entry, press release, and peer-reviewed experimental record are different evidence classes. They may inform one another, but they should never be merged into one undifferentiated claim count.

Worked Audit: A Planckian-Metal Claim

This synthetic record asks how far a linear-$T$ resistivity repeated across preparations in one study can support Planckian language. The arithmetic is real; the material label and data are pedagogical so that the stopping rule, rather than a current-material verdict, remains the object of study.

1. Frontier question and canonical owner. Does linear-$T$ resistivity in synthetic tetragonal metal M at doping $x=0.24$ license a universal Planckian rate? Strange Metals is the live stable owner. Non-Fermi Liquids and Quantum Criticality in Research are the planned dated owners and are not yet learner links.

2. System, specimen, and prepared state. The record contains four lithographically defined four-probe bars from two batches, patterned along the tetragonal $a$ axis on an $ab$ face. The longitudinal $\rho_{aa}$ record is taken at ambient pressure, zero magnetic field, weak linear-response current, and $5\le T\le60\ \text K$. Contact placement, bar dimensions, current-reversal protocol, and doping record are frozen before comparison.

3. Claimed object, claim class, and dated status. The primary observation claim is a linear-$T$ dc-resistivity window. The secondary inference is an order-one one-band transport-rate slope. A universal bound and a microscopic mechanism are not claimed. The inference is active and benchmark-dependent as of 2026-08-21. The study remains a single-experiment-claim; repeat preparations across two batches are recorded separately and do not constitute independent reproduction.

4. Scales, regime, and limits. The fit is restricted to the measured $5$–$60\ \text K$ normal-metal window at $B=0$. For each finite specimen, the weak-current dc limit is taken within the experimental resolution. The fit is not extrapolated to $T=0$, and no thermodynamic critical regime is inferred.

5. Observable, probe, and forward model. Four-terminal transport gives the finite-window fit intercept $\rho_{\rm fit}=12.0\pm0.4\ \mu\Omega\,\text{cm}$ and $A=0.300\pm0.010\ \mu\Omega\,\text{cm K}^{-1}$. A separate antisymmetrized Hall sweep on the same composition uses current along $a$, $B\parallel c$, and $-9\le B\le9\ \text T$. With a positive linear Hall slope and the declared one-band Hall factor $r_H=1$, it gives $n=(8.0\pm1.5)\times10^{27}\ \text m^{-3}$. Cyclotron resonance for $B\parallel c$ gives $m_c^\star=(5.0\pm1.0)m_e$; using this cyclotron mass as the in-plane Drude current mass is an explicit model assumption, not an identity. The Drude map converts the fitted slope to $\alpha_{\rm tr}$; it neither measures the total $\tau_{\rm tr}$ nor proves an additive residual and inelastic rate.

6. Evidence provenance and reproduction. All four bars across both batches show the linear window. The retained package contains raw voltage, current, field, thermometer, geometry, and reversal records; specimen-level fits and covariance; analysis code and version; and the publication and correction state. This is repeat preparation and batch replication, not an independent-laboratory reproduction.

7. Nuisance controls and alternatives. Changing the fit window moves $A$ by less than $6\%$, while geometry and contact variants move it by less than $2\%$. Ordinary phonon scattering, multiband transport, anisotropic current vertices, and doping inhomogeneity remain credible alternatives to a universal microscopic mechanism.

8. Theory/computation validity and uncertainty. First convert $A=3.00\times10^{-9}\ \Omega\,\text m\,\text K^{-1}$. With the quoted one-band inputs,

   $$\alpha_{\rm tr} = \frac{\hbar n e^2 A}{m^\star k_{\rm B}} = 1.033.$$

   Treating $n$, $m^\star$, and $A$ as independent gives

   $$\frac{\sigma_\alpha}{\alpha} = \sqrt{ \left(\frac{1.5}{8.0}\right)^2 +\left(\frac{1.0}{5.0}\right)^2 +\left(\frac{0.010}{0.300}\right)^2} = 0.276,$$

   so $\sigma_\alpha=0.285$ and $\alpha_{\rm tr}=1.03\pm0.29$. A real analysis must restore covariance. Admissible two-band spectral-weight allocations instead span $0.55$–$1.75$; that is model spread, not another statistical error bar.

9. Cross-evidence, progress marker, and falsifier. An aligned optical, thermodynamic, or quantum-oscillation carrier audit and a controlled tuning series are required to distinguish phonon, multiband, and critical accounts. Reproducible curvature, reassignment of carrier weight, or loss of slope collapse under controlled tuning would weaken universality.

10. Licensed conclusion, stopping point, ownership, and review. The record licenses a repeat-preparation linear-$T$ window and a model-dependent Planckian-scale rate-slope estimate. It does not license a universal bound or unique mechanism. Stop at Strange Metals and the transport owners; record the absent live Research handoff as a coverage gap. Review in six months, or sooner after an independent reproduction, carrier reassignment, correction, or decisive tuning result.

The lesson is not that $\alpha_{\rm tr}\simeq1$ is uninteresting. It is that this rate-slope parameter’s uncertainty hierarchy is dominated by what the chosen representation means, not merely by the last decimal place in the resistivity fit.

Worked Audit: A Light-Induced Superconducting-Like Response

This second synthetic record tests a stronger inference: whether transient terahertz electrodynamics above an equilibrium $T_c$ establishes a light-induced superconducting phase.

1. Frontier question and canonical owner. Does a transient THz response in synthetic layered superconductor Q above its equilibrium $T_c$ establish light-induced superconductivity? Pump–Probe Spectroscopy, Terahertz and Infrared Probes, and Unconventional Superconductivity are the live owners. Floquet and Driven Systems and Unconventional Superconductivity Frontiers are planned Research owners.

2. System, specimen, and prepared state. The study uses three free-standing $100\ \mu\text m$-thick platelets and six spots, with the $ab$ face exposed. Their equilibrium $T_c$ is $45\ \text K$; the initial state is $70\ \text K$, ambient pressure, and $B=0$. Pump and THz probe are normally incident on the $ab$ face and co-polarized along $a$; a $1.2\ \text{mm}$ pump spot overfills the $0.40\ \text{mm}$ THz spot. A $17\ \text{THz}$, $250\ \text{fs}$ pump has fluence $1.5\ \text{mJ cm}^{-2}$ and calibrated baseline excitation profile $e^{-z/(0.8\,\mu\mathrm m)}$.

3. Claimed object, claim class, and dated status. The primary observation is a transient complex-optical response. Superconducting stiffness or a phase is a model-mediated inference. Mark the response active and the phase interpretation experimentally-unresolved as of 2026-08-21. The study is a single-experiment-claim, even though several specimens and spots were measured.

4. Scales, regime, and limits. The probe spans ordinary frequency $0.4\le f\le2.5\ \text{THz}$ and delays $0.2\le t_d\le3.0\ \text{ps}$. The state is driven, depth-inhomogeneous, and finite-lived. No equilibrium, $\omega\to0$, dc, infinite-time, or thermodynamic-limit conclusion is available from this record.

5. Observable, probe, and forward model. At $t_d=0.6\ \text{ps}$, fit the raw co-polarized reflected fields with a nonstationary layered kernel for a pumped surface profile above the finite unpumped crystal and its two vacuum boundaries. The direct output is the delay-dependent in-plane response $\Sigma_{aa}(2\pi f;t_d)$, whose imaginary part scales approximately as $1/(2\pi f)$ over the measured band and whose loss spectrum has an edge at $f=1.10\pm0.08\ \text{THz}$. The retained forward model declares the $e^{-i\omega t}$ convention, pump and probe profiles, equilibrium optical constants, detector response, temporal convolution, and inversion covariance. It reports an instantaneous $\sigma_{aa}(\omega;t_d)$ only if a quasistationary approximation passes an explicit timescale test.

6. Evidence provenance and reproduction. The response occurs on all three crystals and six spots with $20\%$ amplitude spread. Raw reflected fields, equilibrium references, depth calibration, the processed delay-dependent response, any model-inferred conductivity with its quasistationary test, inversion code, versions, and correction status are archived. No independent-laboratory replication is claimed.

7. Nuisance controls and alternatives. Calibrated heating is $\Delta T=8\pm3\ \text K$. An inductive term and a sub-resolution narrow Drude term remain viable alternatives, as do excitation-depth inhomogeneity, fluence-dependent optical saturation and departures from the baseline exponential profile, photocarriers, phonons, and redistribution of normal spectral weight. There is no dc, Meissner, or phase-sensitive evidence.

8. Theory/computation validity and uncertainty. The transient lifetime is $1.40\pm0.20\ \text{ps}$, comparable to periods within the probe band, so quasistationarity is not assumed. Propagating inductive and narrow-Drude candidates through the same two-time kernel, raw-field data, covariance, and likelihood gives $\Delta\operatorname{AIC}=2.1$. Even if the inductive fit is assigned the lower AIC in this two-candidate illustration, its normalized Akaike weight is

   $$\frac{1}{1+e^{-2.1/2}} \simeq 0.74,$$

   which is not a posterior phase probability and says nothing about omitted models. Admitted excitation-depth and forward-model variants move the inferred stiffness by $30\%$. Fit covariance and model spread therefore remain separate entries in the uncertainty budget.

9. Cross-evidence, progress marker, and falsifier. Require a temporally matched phase-sensitive or magnetic-stiffness observable, lower-$f$ coverage that resolves or excludes $\Gamma/(2\pi)$, and a spectral-weight audit. Resolution of a finite normal-state $\Gamma$, loss of the signal after calibrated depth correction, or absence of the correlated phase marker weakens the superconducting interpretation.

10. Licensed conclusion, stopping point, ownership, and review. License a transient superconducting-like or inductive optical response. Do not license a superconducting phase, equilibrium transition, or microscopic mechanism. Stop at the pump–probe and THz owners and hand pairing semantics to Unconventional Superconductivity. The Research owners remain an explicit coverage gap. Review in six months, or sooner after independent replication, lower-frequency data, a phase-sensitive test, correction, or retraction.

The finite-bandwidth $1/\omega$ trend is valuable evidence, but the narrow Drude construction proves that it is not a phase discriminator by itself. The proper next step is a measurement that makes the alternatives predict different records in the same prepared state.

Exit Checkpoint

You are ready to use this gateway when you can take an unfamiliar claim and do all of the following without choosing a preferred mechanism first:

• freeze the system, prepared state, object, regime, and evidence cutoff;
• classify each sentence as observation, inference, mechanism, numerical, or engineering, then apply status and evidence labels on separate axes;
• identify a live stable owner and the probe or computation owner;
• reconstruct the forward model from raw record to claimed object;
• separate precision, shared covariance, specimen variation, convergence, representation dependence, and model spread;
• name at least one credible alternative, one independent discriminator, and one result that would weaken the claim;
• write a conclusion that states both what is licensed and where it stops; and
• assign changing status to a live frontier route or record an explicit coverage gap with a review trigger.

A record is not ready merely because all ten headings contain text. Each field must be specific enough that another reader can discover a changed specimen, hidden limit, missing alternative, or shifted conclusion. If that cannot be done, the next task is record repair, not confidence scoring.

Review a frontier record at its declared interval and immediately after an independent reproduction, failed reproduction, correction, retraction, material state change, new decisive probe, altered forward model, or newly substantive living owner. Preserve the previous dated conclusion in the living status system rather than silently overwriting history here.

Canonical Boundaries and Coverage Gaps

This gateway owns one thing: the stable material-facing method for decomposing, routing, and stopping a frontier claim. Its synthetic audits teach that method. The following boundaries keep the rest of the site coherent.

• Quantum Matter concept, phase, model, material, response, and probe pages own stable definitions, derivations, diagnostic criteria, forward models, and material ledgers. This gateway consumes those results; it does not reproduce their technical development.
• How Quantum Matter Is Measured owns the detector-record-to-observable map. Individual probe pages own acquisition, calibration, matrix elements, resolution, and inversion. Data Interpretation and Pitfalls owns cross-probe artifacts and alternative explanations.
• Computational Quantum Matter and the computational-methods volume own algorithm selection, convergence, benchmarks, software, and reproducibility. This gateway audits what a frontier calculation licenses rather than reteaching its implementation.
• Quantum Information hardware pages own fabrication, control, readout, yield, drift, correlated error, integration, and architecture claims. A material metric alone cannot establish a useful technology.
• Research owns dated claim trackers, topical status, annual change logs, living reviews, and the history of upgrades, downgrades, corrections, and retired claims. This page keeps no competing current-status table.

The Research entries below are marked Planned and do not yet contain dated assessments: Unconventional Superconductivity Frontiers, Non-Fermi Liquids, Quantum Criticality, Quantum Spin Liquids, Anyons and Non-Abelian Statistics, Fractional Chern Insulators, Moiré Quantum Systems, Flat-Band and Correlated Topological Matter, Floquet and Driven Systems, and Quantum Materials Discovery. Research has no declared dedicated material-to- technology status route yet; record that as an explicit coverage gap rather than inventing a title. Use the live stable owner and this gateway in the meantime.

Do not fill those gaps with one page per fashionable material, mechanism, or press cycle. A durable new route needs a distinct learner exit capability, a canonical ownership boundary, and a maintenance plan. Otherwise, develop a dated record within the appropriate Research topic.

Exercises

1. Classify and repair eight frontier statements

Classify each statement below by primary claim class, frontier-status label, and evidence label. More than one status or evidence label may apply, but do not give a second claim class unless the sentence genuinely contains two claims.

1. A calibrated THz loss edge was observed on six spots from one laboratory and reported in a peer-reviewed paper.
2. A transfer-matrix inversion assigns that edge to a nonzero transient stiffness, but narrow-Drude and depth-profile fits remain viable.
3. Exact diagonalization of a finite model finds a many-body Chern number equal to one after size and twist-angle convergence checks.
4. A phonon mechanism is proposed from an isotope shift and a controlled microscopic calculation, while a competing electronic account still fits the spectra.
5. A fabrication line reaches its preregistered yield target on three lots.
6. Two independent laboratories reproduce a calibrated response, but disagree about its microscopic origin.
7. “Artificial intelligence discovered a room-temperature quantum material” when only a held-out computational ranking exists.
8. “A zero-bias peak proves non-Abelian anyons” when it comes from one local tunneling experiment on one device family.

Repair the last two statements without erasing what was actually achieved.

Solution

1. This is an observation, plausibly active and experimentally-unresolved, with single-experiment-claim and peer-reviewed-result evidence. Several spots are not independent labs.
2. This is an inference, active, experimentally-unresolved, and benchmark-dependent. It inherits the same empirical provenance and adds a model-dependent inference; the listed alternatives prevent promotion to established.
3. This is numerical. The result may be established within the declared finite model once Hilbert space, boundary, size, twist-mesh, and convergence controls pass, while material relevance remains theoretically-unresolved and/or benchmark-dependent. Active would indicate ongoing research, not the result’s maturity. Applicable evidence includes numerical-evidence and peer-reviewed-result if published; neither turns it into an accepted experiment.
4. This is a mechanism claim, theoretically-unresolved, with evidence labels appropriate to the actual records—for example accepted-experiment for a mature isotope result and numerical-evidence for the calculation. The fit degeneracy blocks a unique-mechanism conclusion.
5. This is engineering. If the target and sampling plan were frozen in advance, engineering-demonstration applies; engineering-limited may describe the broader platform if scaling remains the bottleneck.
6. The response is an observation with reproduced-experiment. Its scoped existence may be established, while the mechanism remains controversial or theoretically-unresolved. Those labels belong to different claims and may coexist.
7. The available claim is numerical, benchmark-dependent, and supported by a benchmark-claim or numerical-evidence, not an experimental discovery. A repair is: “A held-out computational benchmark ranks the candidate for synthesis and measurement; its predicted property has not yet been experimentally verified.”
8. The peak is an observation, active, and experimentally-unresolved, with single-experiment-claim evidence. A repair is: “Local tunneling resolves a zero-bias peak compatible with the topological-device model, while ordinary Andreev, disorder, and confinement alternatives remain to be excluded; no non-Abelian fusion or braiding claim is licensed.”

Publication form may be recorded separately as peer-reviewed-result or preprint-result; neither label determines truth. Likewise, active reports research activity rather than confidence.

2. Distinguish stable treatments from planned assessments

For each topic, name one live stable owner and one planned living-status owner: high-$T_c$ mechanism, Planckian metal, quantum spin liquid, Majorana platform, fractional Chern insulator, moiré correlated phase, light-controlled state, and AI-ranked material. Distinguish the substantive physics treatment from the unwritten Research assessment.

Solution

• High-$T_c$ mechanism: use Unconventional Superconductivity; the planned living owner is Unconventional Superconductivity Frontiers.
• Planckian metal: use Strange Metals; the planned owner is Non-Fermi Liquids; use Quantum Criticality instead only when the frozen claim is specifically about a critical mechanism.
• Quantum spin liquid: use Quantum Spin Liquids; the planned owner is Quantum Spin Liquids in Research.
• Majorana platform: use Topological Superconductors; the planned owner is Anyons and Non-Abelian Statistics.
• Fractional Chern insulator: use Moiré Topology; the planned owner is Fractional Chern Insulators.
• Moiré correlated phase: use Moiré Superlattices; the planned owner is Moiré Quantum Systems.
• Light-controlled state: use Floquet Quantum Matter for periodic-drive semantics or Pump–Probe Spectroscopy for a pulsed experiment; the planned owner is Floquet and Driven Systems.
• AI-ranked material: use Quantum Materials by Design; the planned owner is Quantum Materials Discovery.

The Research links in this solution lead to entries marked Planned. They identify the intended assessment topics; use this gateway and the substantive stable treatments to evaluate a claim meanwhile.

3. Recompute and bound the Planckian rate-slope estimate

Use $A=0.300\pm0.010\ \mu\Omega\,\text{cm K}^{-1}$, $n=(8.0\pm1.5)\times10^{27}\ \text m^{-3}$, and $m^\star=(5.0\pm1.0)m_e$. Treat the three quoted errors as independent. Compute $\alpha_{\rm tr}$ and its standard uncertainty. Then incorporate the separate two-band range $0.55$–$1.75$ and write the strongest licensed claim.

Solution

The slope is $A=0.300\times10^{-8}=3.00\times10^{-9}\ \Omega\,\text m\,\text K^{-1}$. Substitution into

$$\alpha_{\rm tr} = \frac{\hbar n e^2A}{m^\star k_{\rm B}}$$

gives $\alpha_{\rm tr}=1.033$. Independent relative propagation yields

$$\begin{aligned} \left(\frac{\sigma_\alpha}{\alpha}\right)^2 &= \left(\frac{1.5}{8.0}\right)^2 +\left(\frac{1.0}{5.0}\right)^2 +\left(\frac{0.010}{0.300}\right)^2, \\ \frac{\sigma_\alpha}{\alpha} &=0.276, \qquad \sigma_\alpha=0.285. \end{aligned}$$

Thus the one-band branch gives $\alpha_{\rm tr}=1.03\pm0.29$. The $0.55$–$1.75$ interval comes from admissible carrier-weight representations and must remain a separate model-sensitivity range. The licensed statement is: “The samples show a repeat-preparation linear-$T$ resistivity window whose one-band Drude conversion gives a model-dependent, order-one Planckian-scale rate-slope parameter.” Neither a universal bound nor a unique mechanism follows.

4. Audit an absence-of-order plus continuum claim

A candidate frustrated magnet has no resolved elastic magnetic Bragg peak down to $80\ \text{mK}$ and shows a broad inelastic-neutron continuum. The elastic sensitivity is $0.03\mu_B$ per magnetic ion, two samples come from one growth, and no local magnetic probe or isotope control has been reported. Identify the licensed conclusion, three important alternatives, and three genuinely independent discriminators.

Solution

The null result licenses only “no static order above the stated moment, wave-vector, acceptance, and time-window sensitivity in the measured specimens.” The continuum is an observed neutron response, not yet a direct spinon or topological-order measurement.

Three important alternatives are disorder-induced random singlets or a broad distribution of local scales; slow or glassy freezing outside the neutron time window; and unresolved conventional modes, multiparticle scattering, or phonon contamination. Valence-bond order is another phase alternative.

Three independent discriminators are:

1. a local dynamic probe such as $\mu$SR or NMR, with temperature and frequency dependence, to test for slow freezing invisible to the neutron window;
2. polarized-neutron and isotope- or temperature-dependent measurements with absolute normalization and full momentum mapping to separate magnetic weight from phonons and test continuum sum rules; and
3. independently grown samples with quantified defects, paired with a disorder-tuning and thermodynamic or thermal-transport study, to test whether the continuum tracks intrinsic correlations rather than defect density.

These probes differ in coupling, time window, and specimen history. Repeating the same neutron inversion three ways would not create the same independence. The claim stops at Quantum Spin Liquids with status still experimentally unresolved.

5. Stop a proximitized-wire zero-bias claim correctly

A proximitized nanowire device displays a zero-bias conductance peak over a finite gate and field window. The peak is local to one tunnel contact; the gap softens, and no opposite-end, island-parity, fusion, or braiding measurement is available. Build an evidence ladder from the stable baseline to a decisive progress marker, and state what the present record licenses.

Solution

The stable baseline has three owners. Proximity and Andreev Physics owns induced pairing and ordinary Andreev structure. Topological Superconductors owns the BdG invariant, bulk-gap closing and reopening, and boundary-mode criteria. Anyons and Braiding owns fusion and exchange claims.

The current observation is a local zero-bias peak. Ordinary Andreev bound states, smooth confinement, disorder, quantum-dot or Kondo structure, heating, soft-gap spectral weight, and contact-dependent broadening remain alternatives. A stronger platform test would correlate both wire ends across the same topological-transition window while tracking the bulk gap and device parameters. Island-parity and nonlocal fusion protocols would be later progress markers. A controlled fusion outcome is still not a braid.

The present record licenses a zero-bias feature compatible with the candidate device model. It does not license a topological phase, a Majorana zero mode, a non-Abelian anyon, a fusion rule, or braiding. A disappearing opposite-end correlation, lack of a bulk transition, or reproduction by a calibrated ordinary-Andreev model would weaken the topological interpretation.

6. Bound a fractional-filling Hall claim

A moiré device near fractional filling shows a Hall plateau and small but nonzero longitudinal resistance over a narrow field interval. Filling has a $6\%$ electrostatic-model uncertainty; compressibility and real-space order have not been measured. State the minimum evidence needed to promote the record toward a fractional Chern-insulator inference, and identify the stopping point even if those tests pass.

Solution

First align the filling using an independently calibrated density and report contacts, geometry, tensor inversion, and the residual dissipation sensitivity. Track the same feature through a density–field fan and test its Středa slope where that continuation is experimentally accessible. A plateau with appreciable $\rho_{xx}$ may be a contact, inhomogeneity, or crossover feature rather than a quantized bulk response.

Next require an incompressibility or thermodynamic-gap signature consistent with the same state. Exclude charge-density order, translation-symmetry breaking, ordinary integer reconstruction, percolation, and nearby competing states using real-space, diffraction, or symmetry-sensitive information. Check robustness across devices, controlled disorder, temperature, and the relevant finite-size or finite-field window. The stable owners are Moiré Topology and Fractional Quantum Hall Effect.

If calibrated filling, low dissipation, incompressibility or a thermodynamic gap, Středa consistency where accessible, and exclusion of charge order cohere, the record may license an FCI candidate or phase inference at the stated conditions. It still does not measure fractional charge, exchange statistics, non-Abelian content, fusion, or braiding. Those require their own interferometric, noise, charge-sensing, or operational evidence.

7. Compare the two transient-optics alternatives

For the light-induced record, propagate the inductive and narrow-Drude candidates through the same two-time response and detector kernel, then fit the same raw-field vector with the same covariance and likelihood scale. Their fits differ by $\Delta\operatorname{AIC}=2.1$, depth and forward-model variants shift the inferred stiffness by $30\%$, and no dc, Meissner, or phase-sensitive evidence exists. If the inductive model has the smaller AIC, compute its normalized two-model Akaike weight. Then give the licensed conclusion and one temporally matched decisive test.

Solution

Within the declared two-model candidate set,

$$w_{\rm ind} = \frac{1}{1+e^{-2.1/2}} = 0.74, \qquad w_{\rm Drude}=0.26.$$

These weights are not posterior probabilities that either physical state is true. They omit every untested model and inherit the likelihood, covariance, depth profile, and candidate definitions. The $30\%$ model shift is also much larger than a precision-only reading would suggest.

The licensed conclusion is a finite-lived transient inductive optical response over the measured bandwidth and delay window. A temporally matched phase-sensitive Josephson response would be a discriminating next test; alternatively, a magnetic-stiffness measurement in the same prepared state would test a different consequence of phase coherence. Either must be aligned to the pump envelope and lifetime. Without such evidence, do not claim a superconducting phase, equilibrium transition, or mechanism.

8. Complete a frontier record for an ML-ranked moiré material

A model trained on $2400$ relaxed two-dimensional heterostructures ranks synthetic MX2/MX2 at target twist $1.30^\circ$ for a narrow valence band. An additional $240$-structure, composition-family-disjoint test set gives an unweighted per-structure bandwidth RMSE of $6\ \text{meV}$. The candidate prediction is $7\pm3\ \text{meV}$, where the quoted spread is the mean plus or minus one sample standard deviation across ten independently initialized surrogate models, not a candidate-specific confidence interval. The preregistered numerical screen requires bandwidth $W\le12\ \text{meV}$ and minimum direct isolation from excluded bands $\Delta_{\rm sep}\ge20\ \text{meV}$. One batch yields two devices with measured twist $1.33^\circ\pm0.03^\circ$; at $1.6\ \text K$, $B=0$, and filling $\nu=-1$—defined as one hole per moiré cell relative to charge neutrality, summed over the active spin and valley flavors—both show a resistance maximum but no activation or compressibility record. A fresh relaxed calculation gives $W=9\ \text{meV}$, while an admissible relaxation model gives $W=16\ \text{meV}$; neither branch reports $\Delta_{\rm sep}$. Fill all ten ledger fields, including the stable and planned owners, progress marker, falsifier, and review cadence.

Solution

1. Frontier question and canonical owner. Ask whether the ranked and fabricated heterostructure realizes the isolated narrow-band conditions needed for a correlated moiré candidate. Quantum Materials by Design owns the discovery loop, Moiré Superlattices owns the platform variables, and Flat Bands owns the stable band criteria. Quantum Materials Discovery and Moiré Quantum Systems are planned living owners.

2. System, specimen, and prepared state. Record the MX2/MX2 composition, stacking convention, measured twist $1.33^\circ\pm0.03^\circ$, relaxation and encapsulation, one fabrication batch and two devices, active-area and boundary geometry, disorder and strain maps, $T=1.6\ \text K$, $B=0$, $\nu=-1$ with the stated one-hole-per-moiré-cell convention, gate history, and measurement current. The nominal $1.30^\circ$ design and the measured twist are not interchangeable.

3. Claimed object, claim class, and dated status. The ranking and bandwidth are numerical claims; fabrication to the measured geometry is an engineering claim; the resistance maximum is an observation. A narrow correlated band or phase is an unlicensed inference. Mark the ranking active and benchmark-dependent, and the phase experimentally-unresolved, as of 2026-08-21. Applicable evidence includes benchmark-claim, numerical-evidence, and engineering-demonstration, but not reproduced-experiment.

4. Scales, regime, and limits. The calculation targets valence-band energies near $\nu=-1$ at zero physical temperature, while the devices are finite, disordered, and measured at $1.6\ \text K$. State the momentum mesh, real-space relaxation scale, gate and filling window, sample dimensions, and the order in which size, mesh, broadening, and temperature limits are taken. No thermodynamic-limit phase is available.

5. Observable, probe, and forward model. The ML output is a predicted bandwidth, mapped from relaxed structural descriptors through a frozen model. The fresh electronic-structure calculation supplies a surrogate- independent numerical branch, but it is not independent physical validation when it shares the electronic-structure approximation used to label the training set. Four-terminal resistance supplies the experimental observation; it does not directly measure bandwidth, compressibility, or an order parameter. Archive the density calibration, contact geometry, tensor convention, and raw transport record.

6. Evidence provenance and reproduction. Retain the $2400$ training and $240$ family-disjoint test structures, composition-family group labels, split hashes, ten model seeds and weights, code and environment, hyperparameter search, held-out predictions, relaxed input files, and both fresh calculations. Retain microscopy, twist extraction, raw transport, and device histories. Two devices from one batch establish neither independent synthesis nor independent-laboratory reproduction.

7. Nuisance controls and alternatives. Audit composition and structure duplicates across splits, candidate distance from the training domain, relaxation choice, exchange-correlation and screening models, twist and strain inhomogeneity, contacts, density calibration, disorder, and an ordinary van Hove or percolative resistance maximum. Charge order and other symmetry-broken competitors remain open.

8. Theory/computation validity and uncertainty. The family-disjoint, unweighted $6\ \text{meV}$ test RMSE is a benchmark statistic, not a candidate-specific posterior error bar. The candidate’s $3\ \text{meV}$ ten-seed sample standard deviation measures surrogate variation only. The $9$–$16\ \text{meV}$ relaxation branch is a separate material-model spread: one value passes and one fails the $12\ \text{meV}$ bandwidth threshold. Neither establishes the required $20\ \text{meV}$ isolation because $\Delta_{\rm sep}$ is missing. Record mesh, basis, structural-force, and self-consistency convergence and report split-to-split variation separately. The current data show neither robust isolation nor interaction dominance.

9. Cross-evidence, progress marker, and falsifier. A decisive next step is an independently calibrated spectroscopy or compressibility measurement on the same prepared state, paired with spatial twist and strain maps and a second-batch reproduction. A resolved $W>12\ \text{meV}$, $\Delta_{\rm sep}<20\ \text{meV}$, loss of the feature under corrected filling, or correlation of the resistance maximum with inhomogeneity would weaken the narrow-band phase claim.

10. Licensed conclusion, stopping point, ownership, and review. License an ML-ranked candidate that was fabricated near its target twist, plus a resistance maximum at the stated filling. Do not license a verified narrow isolated band, correlated phase, mechanism, or useful device. Stop at the three stable moiré/design owners and record Quantum Materials Discovery and Moiré Quantum Systems as planned Research coverage gaps. Review in six months or immediately after spectroscopy, compressibility, a second batch, model revision, correction, or failed reproduction.

References

• H. Akaike, “A New Look at the Statistical Model Identification,” IEEE Transactions on Automatic Control 19, 716–723 (1974), doi:10.1109/TAC.1974.1100705.
• L. Balents, C. R. Dean, D. K. Efetov, and A. F. Young, “Superconductivity and Strong Correlations in Moiré Flat Bands,” Nature Physics 16, 725–733 (2020), doi:10.1038/s41567-020-0906-9.
• D. N. Basov, R. D. Averitt, and D. Hsieh, “Towards Properties on Demand in Quantum Materials,” Nature Materials 16, 1077–1088 (2017), doi:10.1038/nmat5017.
• J. A. N. Bruin, H. Sakai, R. S. Perry, and A. P. Mackenzie, “Similarity of Scattering Rates in Metals Showing $T$-Linear Resistivity,” Science 339, 804–807 (2013), doi:10.1126/science.1227612.
• K. T. Butler, D. W. Davies, H. Cartwright, O. Isayev, and A. Walsh, “Machine Learning for Molecular and Materials Science,” Nature 559, 547–555 (2018), doi:10.1038/s41586-018-0337-2.
• S. Curtarolo et al., “AFLOWLIB.ORG: A Distributed Materials Properties Repository from High-Throughput Ab Initio Calculations,” Computational Materials Science 58, 227–235 (2012), doi:10.1016/j.commatsci.2012.02.002.
• A. de la Torre, D. M. Kennes, M. Claassen, S. Gerber, J. W. McIver, and M. A. Sentef, “Colloquium: Nonthermal Pathways to Ultrafast Control in Quantum Materials,” Reviews of Modern Physics 93, 041002 (2021), doi:10.1103/RevModPhys.93.041002.
• J. S. Dodge, L. Lopez, and D. G. Sahota, “Optical Saturation Produces Spurious Evidence for Photoinduced Superconductivity in $\mathrm{K}_3\mathrm{C}_{60}$,” Physical Review Letters 130, 146002 (2023), doi:10.1103/PhysRevLett.130.146002.
• S. A. Hartnoll and A. P. Mackenzie, “Colloquium: Planckian Dissipation in Metals,” Reviews of Modern Physics 94, 041002 (2022), doi:10.1103/RevModPhys.94.041002.
• Joint Committee for Guides in Metrology, Evaluation of Measurement Data—Guide to the Expression of Uncertainty in Measurement, JCGM 100:2008, corrected 2010, doi:10.59161/JCGM100-2008E.
• B. Keimer et al., “From Quantum Matter to High-Temperature Superconductivity in Copper Oxides,” Nature 518, 179–186 (2015), doi:10.1038/nature14165.
• B. Keimer and J. E. Moore, “The Physics of Quantum Materials,” Nature Physics 13, 1045–1055 (2017), doi:10.1038/nphys4302.
• M. Mitrano et al., “Possible Light-Induced Superconductivity in $\mathrm{K}_3\mathrm{C}_{60}$ at High Temperature,” Nature 530, 461–464 (2016), doi:10.1038/nature16522.
• M. R. Munafò et al., “A Manifesto for Reproducible Science,” Nature Human Behaviour 1, 0021 (2017), doi:10.1038/s41562-016-0021.
• National Academies of Sciences, Engineering, and Medicine, Reproducibility and Replicability in Science, National Academies Press (2019), doi:10.17226/25303.
• C. Nayak et al., “Non-Abelian Anyons and Topological Quantum Computation,” Reviews of Modern Physics 80, 1083–1159 (2008), doi:10.1103/RevModPhys.80.1083.
• D. Nicoletti and A. Cavalleri, “Nonlinear Light–Matter Interaction at Terahertz Frequencies,” Advances in Optics and Photonics 8, 401–464 (2016), doi:10.1364/AOP.8.000401.
• J. Orenstein and J. S. Dodge, “Terahertz Time-Domain Spectroscopy of Transient Metallic and Superconducting States,” Physical Review B 92, 134507 (2015), doi:10.1103/PhysRevB.92.134507.
• N. Regnault and B. A. Bernevig, “Fractional Chern Insulator,” Physical Review X 1, 021014 (2011), doi:10.1103/PhysRevX.1.021014.
• M. Sato and Y. Ando, “Topological Superconductors: A Review,” Reports on Progress in Physics 80, 076501 (2017), doi:10.1088/1361-6633/aa6ac7.
• L. Savary and L. Balents, “Quantum Spin Liquids: A Review,” Reports on Progress in Physics 80, 016502 (2017), doi:10.1088/0034-4885/80/1/016502.
• R. L. Wasserstein and N. A. Lazar, “The ASA Statement on $p$-Values: Context, Process, and Purpose,” The American Statistician 70, 129–133 (2016), doi:10.1080/00031305.2016.1154108.