Magnetism and Spin Systems

Magnetism in a material is not selected by seeing a hysteresis loop, a Curie tail, a split band, or a neutron peak and then attaching the nearest familiar model. The same experimental feature can arise from different magnetic operators, exchange mechanisms, dimensionalities, domain histories, and probe weights. Conversely, the same microscopic Hamiltonian can support distinct ordered, fluctuating, screened, or textured regimes as temperature, field, pressure, doping, and disorder change.

This gateway turns a declared magnetic question into the shortest defensible route. It asks what the moment means, which degrees of freedom remain active, which symmetry and scale assumptions license a model, what the experiment actually records, and what observation would force the interpretation to stop. The linked pages own the physics and derivations; this page owns route selection, evidence discipline, and the boundaries between them.

The Quantum Matter Map owns the system–state–observable problem tuple, while Choosing a Model for Quantum Matter owns cross-family candidate comparison and smallest-adequate-model tests. This gateway begins after the question has become magnetic and specializes those tools to magnetic material, model, and probe owners.

Helpful background. Use Magnetic Moments and g-Factors for the operator and sign conventions of bare spin and orbital moments. Use Lattice Models and Spin Systems for model selection and Phases, Order, and Criticality for order parameters, symmetry breaking, and finite-size or thermodynamic-limit language. These are branch-specific capabilities, not universal prerequisites.

Enter This Chapter

Choose the question before choosing the magnetic noun.

• What magnetic object is being quoted? Begin with Magnetic Moments in Matter to identify the physical operator, retained subspace, moment notion, scale and probe window, and whether a fixed-length spin or pseudospin is licensed. Use the bare moment, atomic, spin–orbit, itinerant, and susceptibility owners linked there for their specialist derivations and measurements.
• Where does the coupling come from? Use Exchange Interactions, then specialize to RKKY Interaction, Spin–Orbit Coupling in Solids, or Magnetic Anisotropy.
• Which ordered phase is compatible with the evidence? Compare Ferromagnetism, Antiferromagnetism, and Ferrimagnetism.
• What is the magnetic excitation? Read canonical Magnons, then Spin Waves and Magnons for the material Hamiltonian, spectral weights, linewidths, and probe fit.
• Are mobile electrons essential? Use Itinerant Magnetism and the Stoner Criterion for the band-electron branch. Use the Kondo Effect and RKKY Interaction for a local moment embedded in a metal.
• Is the goal control or texture rather than phase identification? These are mechanism-dependent branches, not one universal chain. Use Spin–Orbit Coupling in Solids when relativistic projection or charge–spin conversion is relevant; Magnetic Anisotropy for orientation energies; Spintronics for transport, torque, and device control; or Skyrmions and Magnetic Textures for a declared order manifold, texture, and stability question.

Consolidated Model Boundaries

Two empty prelaunch routes were consolidated. The generic Heisenberg model, sign convention, graph, and exact limits remain with the Heisenberg Model and Common Spin Hamiltonians. Material exchange provenance belongs to Exchange Interactions, phases belong to the three order pages, and spectra belong to Spin Waves and Magnons. Ising, XY, and XXZ Hamiltonian forms likewise remain with the model and reference owners; material symmetry reduction belongs to Magnetic Anisotropy, while walls and vortices belong to the texture page. A duplicate material-model bridge would obscure those canonical boundaries.

Readiness Check

Before accepting a magnetic interpretation, be able to answer:

1. What sample, composition, lattice, dimension, boundary, and magnetic cell are being discussed?
2. Is the claimed object a microscopic moment, magnetization, susceptibility, correlation, order parameter, excitation, spin current, or texture?
3. Are the active variables ionic multiplets, projected pseudospins, itinerant quasiparticles, impurities, orbitals, or a mixed set?
4. Which exact or approximate symmetries remain, and which are spontaneously or explicitly broken?
5. Which temperature, field, pressure, doping, disorder, history, frequency, momentum, and size window defines the claim?
6. What instrument record and forward model connect the sample to the reported quantity?

If one item is missing, repair it before selecting a spin Hamiltonian or phase label. A material name is not a Hamiltonian, and a fitted coupling is not a microscopic mechanism without a provenance and validity test.

Write the Magnetic-Claim Ledger

Record ten fields before comparing magnetic theories.

1. Material and state. Give composition, lattice, effective dimension, magnetic cell if known, filling or valence, boundaries, defects, disorder, temperature, field, pressure, doping, preparation, and history.
2. Moment object and normalization. Name the microscopic dipole, spin, orbital or total moment, projected pseudospin, spin density, sample moment, magnetization, or susceptibility. State per site, formula unit, cell, mole, area, or volume; declare SI or other units, axes, sign, and internal versus applied field.
3. Active degrees of freedom. Declare local, itinerant, ionic, covalent, orbital, multipolar, impurity, carrier, or mixed variables and the retained Hilbert space or eliminated states.
4. Symmetry. Separate exact from approximate spin, lattice, inversion, translation, time-reversal, and magnetic-space-group symmetries; state which are proposed to be spontaneously or explicitly broken.
5. Interactions. Record exchange tensors, anisotropy, dipolar, Zeeman, spin–orbit, carrier, Kondo, or other couplings with sign and pair-counting conventions.
6. Scale hierarchy and regime. Compare charge, crystal-field, spin–orbit, exchange, anisotropy, dipolar, field, thermal, disorder, damping, finite-size, correlation-length, drive, and probe-resolution scales. Declare equilibrium or driven and local-moment or itinerant limits only over a stated window.
7. Order or correlation object. Specify uniform, finite-$\mathbf Q$, multi-sublattice, multi-$\mathbf Q$, multipolar, glassy, domain, or correlation-only structure, its magnetic cell, and the order of limits—or state explicitly that long-range order has not been established.
8. Requested observable. Name the thermodynamic derivative, static or dynamic structure factor, spectral pole and weight, band reconstruction, transport tensor, torque, texture, switching metric, or device output.
9. Framework and provenance. State the spin, itinerant-electron, impurity, micromagnetic, response, or transport framework; mark exact representation changes versus approximations; record how parameters were derived, fitted, or measured and where the reduction is controlled.
10. Forward model and audit. Connect the intrinsic object to the specified probe, geometry, domains, resolution, contacts, and backgrounds. Attach an uncertainty, competing explanation, convergence or scaling check, and a falsification, stopping, or escalation test.

A defensible conclusion has the form: “within this state, scale window, operator convention, model, and forward model, the observations support this bounded magnetic claim over the stated alternatives.” It does not turn one loop, peak, exponent, or image into a unique Hamiltonian.

Keep the Magnetic Objects Separate

A local moment is an operator or correlation property of a retained degree of freedom. Magnetization is a signed moment density. Susceptibility is a response coefficient. An ordered moment is a symmetry-selected Fourier component. A magnon is a collective excitation about a stationary ordered reference. None of these definitions implies the others.

A candidate spin-only reduction may be organized as follows. When $\mathbf s_i$ represents genuine spin, it is dimensionless and the corresponding angular momentum is $\hbar\mathbf s_i$. A projected pseudospin is instead a basis operator whose physical angular momentum and magnetic moment require declared projection tensors. A classical reduction must declare its vector length and normalization. The pair sum runs over ordered distinct sites $i\ne j$, and the factor $1/2$ removes double counting.

$$\begin{aligned} H_{\mathrm{eff}} &= \frac12\sum_{i\ne j} \mathbf s_i^{\mathsf T}\mathsf J_{ij}\mathbf s_j +\sum_i\mathbf s_i^{\mathsf T}\mathsf A_i\mathbf s_i \\ &\quad- \sum_i\mathbf h_i\cdot\mathbf s_i+⋯ . \end{aligned}$$

This is not a universal magnetic Hamiltonian. It presumes a stable retained moment manifold and scale separation from charge, orbital, crystal-field, and longitudinal fluctuations. For the displayed Hermitian convention, $\mathsf J_{ij}$ and $\mathsf A_i$ have energy units, $\mathsf J_{ij}$ is real with $\mathsf J_{ji}=\mathsf J_{ij}^{\mathsf T}$, each $\mathsf A_i$ is real and symmetric, and $\mathbf h_i$ has energy units. The sign of an entry of $\mathsf J_{ij}$ is meaningless until the Hamiltonian sign, bond counting, spin normalization, and orbital basis have been declared.

For a dipolar magnetic-moment order parameter at wavevector $\mathbf Q$, define $\hat{\boldsymbol\mu}_{\mathbf Q}=\sum_j e^{-i\mathbf Q\cdot\mathbf R_j}\hat{\boldsymbol\mu}_j$. A physical selecting source at a generic, possibly incommensurate wavevector must include both Fourier components:

$$\begin{aligned} H_h &=H- \mathbf h_{\mathbf Q}\cdot\hat{\boldsymbol\mu}_{-\mathbf Q} -\mathbf h_{\mathbf Q}^{*}\cdot\hat{\boldsymbol\mu}_{\mathbf Q},\\ \mathbf h_{\mathbf Q}&=\lambda\mathbf e, \qquad \lambda\to0^+, \end{aligned}$$

where the complex polarization and phase $\mathbf e$ are held fixed. The thermodynamic limit must precede removal of that source:

$$\mathbf m_{\mathbf Q}(\mathbf e) = \lim_{\lambda\to0^+} \lim_{N\to\infty} \frac1N \left\langle\hat{\boldsymbol\mu}_{\mathbf Q}\right\rangle_{\lambda,\mathbf e} .$$

A symmetric finite-system state can have zero one-point order while its correlations, tower states, and finite-size scaling reveal incipient broken symmetry. Conversely, a field-polarized finite sample need not possess spontaneous order.

With the angular-frequency convention and the normalized Fourier operator

$$\hat S^\alpha_{\mathbf Q} = \frac{1}{\sqrt N} \sum_j e^{-i\mathbf Q\cdot\mathbf R_j}\hat S_j^\alpha,$$

the dynamical correlation tensor is

$$S^{\alpha\beta}(\mathbf Q,\omega) = \frac{1}{2\pi} \int dt\,e^{i\omega t} \left\langle \hat S^\alpha_{\mathbf Q}(t) \hat S^\beta_{-\mathbf Q}(0) \right\rangle$$

Other Fourier and extensive normalizations are valid when stated consistently. This one-frequency form assumes a stationary state. A driven nonstationary state generally requires a two-time correlator or a declared Wigner transform. The tensor is not itself a detector count. Magnetic form factors, polarization factors, domains, kinematics, backgrounds, resolution, and instrumental normalization belong to the probe forward model.

Read the Dependency Graph

The sidebar is a catalog, not a linear prerequisite chain.

1. Moment algebra and material projection feed Exchange Interactions, but moment formation does not establish order.
2. Exchange Interactions plus order-parameter language branches independently to ferromagnetism and antiferromagnetism. Neither ordered phase is a universal prerequisite for the other.
3. Ferro- and antiferromagnetic sublattice language converge for ferrimagnetism; compensation does not mean the sublattice order vanished.
4. A selected ordered reference plus Exchange Interactions and canonical Magnons leads to the material spin-wave route.
5. Band Theory plus susceptibilities leads to Itinerant Magnetism. The Stoner Criterion is a scalar uniform mean-field test inside that larger branch, not a verdict on every magnetic instability.
6. A controlled impurity model leads to the Kondo Effect. The host’s full static spin-susceptibility tensor leads separately to RKKY Interaction; Fermi-surface geometry controls common metallic asymptotics, but interband and Fermi-sea contributions can matter. Their competition continues to Kondo-lattice and heavy-fermion owners.
7. General spin–orbit coupling and Bloch-state symmetry lead to Spin–Orbit Coupling in Solids. Relativistic mechanisms can then feed anisotropy, charge–spin conversion, torque, or chiral-texture branches, but this is not a universal chain: GMR and spin-transfer physics can exist with negligible spin–orbit coupling.
8. Textures require the selected order manifold and stabilizing interactions. Depending on the material, those may be anisotropy, bond-symmetry-constrained antisymmetric exchange, frustrated symmetric exchange, dipolar coupling, or field. Spintronics is a control and device branch, not a hard prerequisite for texture existence.

Choose the Shortest Live Route

Moment formation and projection

Use Magnetic Moments and g-Factors for the operator dictionary, Orbital Magnetic Moments for the bare orbital contribution, and Hund’s Rules for a free-ion starting point. Then use Spin–Orbit Coupling in Solids for crystal-field and pseudospin structure, Itinerant Magnetism when the moment is a collective band-electron property, and Magnetic Susceptibility for the measured Curie, Pauli, Van Vleck, orbital, ordered, or impurity response. Do not equate a Curie effective moment with a saturation or ordered moment.

Exchange and spin-model choice

Use Exchange Interactions to compare direct exchange, superexchange, double exchange, carrier-mediated exchange, and itinerant response. Use the Heisenberg Model and Common Spin Hamiltonians for model definitions and signs. The simplest Hubbard-to-Heisenberg reduction $J=4t^2/U$ requires a repulsive single-band model with $U\gg|t|$, one particle per site, a retained singly occupied sector, and targets below the charge gap. At appreciable doping, retain charge motion and route toward the t–J Model or a more complete itinerant description; the pure Heisenberg model is not a universal material identity.

Ordered phase and thermodynamics

Choose Ferromagnetism when uniform magnetization is the primary equilibrium order parameter, while remembering that domains can cancel the macroscopic sample moment. A secondary uniform component can instead occur in a canted antiferromagnet or ferrimagnet. Choose Antiferromagnetism from the propagation vector, magnetic basis or space group, and staggered order; an antiferromagnet can have $\mathbf k=0$. Choose Ferrimagnetism for oppositely aligned inequivalent sublattices; it remains ferrimagnetic when its uniform component vanishes at a compensation point. Use Magnetic Susceptibility and Neutron Scattering to connect phase language to field history, moment normalization, magnetic Bragg structure, and finite-temperature evidence.

Spin waves, magnons, and spectra

Use canonical Magnons for reusable quantization, stability, interactions, and exactness language. Then use Spin Waves and Magnons for the magnetic cell, local frames, material Hamiltonian, eigenvectors, spectral weights, linewidths, continua, resolution, and parameter validation. A fitted dispersion without intensity or stability checks does not establish the Hamiltonian. A broad continuum or linewidth does not by itself establish fractionalization: test decay, particle–hole, multiparticle, disorder, domain, hybridization, and instrumental-resolution alternatives.

Itinerant and Stoner branches

Use Itinerant Magnetism for exchange-split bands, finite-$\mathbf Q$ instabilities, continua, collective modes, and local–itinerant crossover. Use the Stoner Criterion only for its declared scalar static uniform mean-field test. In a multiband response, a continuous local loss of stability of the declared reference state, within the chosen approximation, is instead associated with a vanishing eigenvalue of the appropriate static kernel, such as

$$\det\!\left[ \mathbf 1- \boldsymbol\Gamma(\mathbf q,0) \boldsymbol\chi_0(\mathbf q,0) \right]=0 .$$

The required static limit and order of limits must be declared. A first-order transition can occur before this local criterion is reached. At nonzero real or complex frequency, zeros of the corresponding retarded kernel instead mark collective poles, damped resonances, or—only with an explicit analytic convention—dynamical instabilities.

The scalar $IN(E_F)>1$ limit requires mutually matched conventions for $I$ and for a per-spin or spin-summed density of states. It does not predict the ordered moment, Curie temperature, finite-$\mathbf Q$ competitor, or correlation corrections.

Kondo and RKKY branches

Use the Kondo Effect for dilute-alloy and quantum-dot screening phenomenology, scale definitions, transport alternatives, and screening-cloud evidence. Use RKKY Interaction for the sign, range, tensor structure, disorder dependence, and full host spin-susceptibility dependence of mediated exchange. Continue to Kondo Lattices or Heavy Fermions when the moments are dense and coherence or Fermi-volume questions become central. A resistance minimum or logarithm is not uniquely Kondo physics, and a cartoon crossing of Kondo and RKKY scales is not a phase theorem. The Kondo screening cloud is a many-body correlation pattern, not one spatially bound screening electron.

Spin–orbit coupling, anisotropy, and control

Use Spin–Orbit Coupling in Solids for crystal-field projection, Kramers and $\mathcal P\mathcal T$ structure, Rashba and Dresselhaus regimes, and the spin-Hall bridge. Use Magnetic Anisotropy to separate magnetocrystalline, shape, exchange, dipolar, surface, interface, and strain contributions. Use Spintronics only after defining spin accumulation, the chosen spin-current and torque convention, interfaces, relaxation, and the charge–spin conversion forward model. Spin is generally not conserved in the presence of spin–orbit coupling. A bulk spin-Hall conductivity, long spin-diffusion length, or quoted torque efficiency does not by itself demonstrate injection, transmission, detection, or switching in a specified device.

Textures and topological claims

Use Skyrmions and Magnetic Textures for walls, vortices, skyrmion topology, energetic stability, dynamics, emergent electrodynamics, and evidence. For a smooth unit vector $\mathbf n$ with a declared compactifying boundary,

$$Q_{\mathrm{sk}} = \frac{1}{4\pi} \int d^2r\, \mathbf n\cdot \left(\partial_x\mathbf n\times\partial_y\mathbf n\right) .$$

Topology does not guarantee stability, accessibility, or lifetime. A residual Hall hump is not proof of skyrmions; require imaging, scattering, or an equivalent orthogonal texture test together with phase-boundary and artifact controls.

When the answer is not an ordered magnet

Use Quantum Spin Liquids for long-range-entangled and fractionalized candidates, Glasses and Spin Glasses for disorder and slow collective freezing, and Kondo Lattices for dense screening and coherence. Absence of a bulk uniform moment is not enough to select any of them.

Worked Audit: Why Can the Net Magnetization Vanish?

Suppose a sample’s low-field magnetization is indistinguishable from zero below a sharp thermal anomaly. Four explanations remain live: antiferromagnetic order, a compensated ferrimagnet, a multidomain ferromagnet, or a paramagnet with a nonmagnetic transition.

Fill all ten ledger fields. Do not silently infer the missing entries from the phrase “zero magnetization.”

1. Material and state: record composition, structure, effective dimension, filling or ionic valence, sample boundaries and disorder, temperature relative to the anomaly, applied field, and zero-field-cooled or field-cooled history.
2. Moment object and normalization: state whether zero refers to total sample moment, magnetization in A/m, moment in $\mu_B$ per ion or formula unit, remanence, a high-field extrapolation, or a susceptibility in declared SI or cgs units.
3. Active degrees of freedom: keep ionic or projected local moments, itinerant electrons, orbital moments, and mixed descriptions live until the data eliminate them; list any states removed by a projection.
4. Symmetry: record the chemical and candidate magnetic space groups, exact and approximate spin symmetries, time reversal, and the symmetry each candidate would break.
5. Interactions: list candidate exchange tensors, anisotropy, dipolar, Zeeman, spin–orbit, carrier, and domain energies without assigning their values from the bulk trace alone.
6. Scale hierarchy and regime: compare temperature, measuring field, demagnetizing field, coercive or anisotropy fields, interlayer coupling, domain and sample sizes, correlation length, sweep time, and probe sensitivity.
7. Order or correlation object: distinguish uniform order, a propagation vector and magnetic basis, unequal sublattices, a domain population, diffuse correlations, and the explicit possibility that no magnetic long-range order has been established.
8. Requested observable: treat bulk magnetometry as the starting record and specify which susceptibility, diffraction, local-probe, imaging, or inelastic observable would discriminate the candidates.
9. Framework and provenance: compare a symmetry-refined magnetic structure, a sublattice model, a domain model, and a paramagnetic explanation; mark each representation change and approximation and record every fitted or measured input.
10. Forward model and audit: retain sample geometry, field orientation, domains, demagnetizing correction, backgrounds, resolution, and detection threshold. Contacts are not applicable unless a transport record is added. Require an orthogonal probe or a stated stopping test for each candidate.

Discriminate the candidates. Refine the magnetic propagation vector $\mathbf k$ and magnetic basis or magnetic space group from peaks at $\mathbf Q=\boldsymbol\tau\pm\mathbf k$; nonzero absolute momentum transfer by itself does not distinguish ferro-, antiferro-, or ferrimagnetism. A consistent magnetic structure can support antiferromagnetic or ferrimagnetic order. Element- or site-resolved moments and their temperature dependence can distinguish unequal opposing sublattices and a compensation temperature. Hysteresis, imaging, field training, and geometry dependence can expose ferromagnetic domains whose net moment cancels. Local probes and diffuse or inelastic scattering test whether moments and correlations exist when no Bragg order is resolved.

Stop at the licensed conclusion. Zero magnetization alone establishes none of the four explanations. Report the ordering wavevector, magnetic symmetry, sublattice moments, domain evidence, field and history window, and unresolved alternatives. A compensation point is not the Curie transition, and absence of one probe signature is not absence of magnetism.

Worked Audit: A Layered Magnet with a Magnon Gap

Suppose a layered material orders at nonzero temperature and inelastic neutron scattering shows a low-energy magnetic peak with a gap. A nearest-neighbor two-dimensional isotropic Heisenberg fit is not yet an explanation.

Fill all ten ledger fields.

1. Material and state: give composition, layer registry, magnetic cell, effective thickness, filling or ionic valence, defects, temperature, field, pressure, preparation, and domain history.
2. Moment object and normalization: define the spin, orbital, total, or projected moment, its units and axes, the ordered moment, and the precise normalization of $S^{\alpha\beta}(\mathbf Q,\omega)$.
3. Active degrees of freedom: state whether the retained variables are rigid local moments, crystal-field pseudospins, itinerant quasiparticles, or a mixed manifold, and identify the charge or orbital states eliminated.
4. Symmetry: distinguish exact from approximate lattice, spin, time-reversal, and magnetic-space-group symmetries; state which symmetry the ordered state and each proposed gap term break.
5. Interactions: declare exchange range and tensors, interlayer exchange, single-ion and exchange anisotropy, dipolar, Zeeman, spin–orbit, and possible magnon–phonon terms with sign and bond-count conventions.
6. Scale hierarchy and regime: compare charge and crystal-field gaps, exchange bandwidth, anisotropy, interlayer, dipolar, Zeeman, thermal, damping, finite-thickness, correlation-length, and energy–momentum resolution scales.
7. Order or correlation object: give the propagation vector, magnetic basis, ordered direction, domains, order of limits, and whether the claimed branch is the lowest acoustic mode or an optical or folded mode.
8. Requested observable: require energies, eigenvectors, polarization and intensity, linewidths, continua, temperature and field dependence, and the elastic evidence for the reference state.
9. Framework and provenance: identify the spin-wave or itinerant-response approximation, the parameter source and covariance, the stationarity and stability tests, and the window over which the reduction is controlled.
10. Forward model and audit: include magnetic form and polarization factors, domain populations, sample orientation, kinematics, backgrounds, instrumental resolution, and an alternative-model fit. Contacts are not applicable here. Define the test that would reopen the retained degrees of freedom or reject the claimed gap mechanism.

Test the order mechanism. Declare the magnetic cell, ordering wavevector, moment direction, interlayer coupling, exchange range, single-ion or exchange anisotropy, dipolar scale, spin–orbit scale, defects, domains, and finite sample thickness. The Mermin–Wagner constraint applies to the ideal short-range continuous-symmetry two-dimensional limit; discrete anisotropy, dipolar or interlayer coupling, finite size, and Berezinskii–Kosterlitz–Thouless physics change the conclusion in different ways.

Test the excitation. Verify that the candidate ordered reference is stable, that the bosonic eigenproblem has physical modes, and that the calculated eigenvectors reproduce momentum, polarization, field, temperature, and intensity dependence after magnetic form factors and instrumental resolution are included. A genuine thermodynamic gap in the lowest acoustic branch at the ordering wavevector requires explicit breaking of the relevant continuous spin symmetry—for example by anisotropy, an applied field, or dipolar coupling—or the absence of that continuous symmetry in the first place. Symmetry-preserving interlayer exchange creates three-dimensional dispersion and can produce gapped optical or folded branches, but it does not by itself gap a Goldstone acoustic mode. Likewise, symmetry-preserving hybridization can create avoided crossings away from the ordering wavevector without removing a protected Goldstone mode. Distinguish a true gap from an optical, folded, spectrally dark, or resolution-limited apparent gap; a finite-size gap must close in the declared size limit. Its magnitude alone does not identify the responsible term.

Bound the result. Quote the Hamiltonian convention, parameter provenance, fit covariance and alternatives, continuum and linewidth treatment, and the temperature and field window. The strongest conclusion is that one declared model jointly accounts for the order and probe-weighted spectrum within that window—not that the fitted model is the unique microscopic Hamiltonian.

Canonical Boundaries

• This gateway owns question-first routing, the magnetic-claim ledger, chapter status, the live-route map, and stopping rules. It owns no leaf derivation.
• The Quantum Matter Map owns the general system–state–observable tuple, and Choosing a Model for Quantum Matter owns cross-family adequacy comparison; this gateway applies those contracts only after the problem is magnetic.
• General spin operators, angular momentum, moment algebra, symmetry, and topology remain in their respective formal volumes.
• Lattice Models and Spin Systems owns generic Hamiltonians, exact limits, and model solution status. Phases, Order, and Criticality owns general symmetry breaking, universality, criticality, and order-of-limits theory.
• Strong Correlations owns spin liquids, Kondo lattices, heavy fermions, fractionalization, and emergent gauge fields. Disorder owns spin glasses.
• Probe pages own magnetometer, neutron, X-ray, Raman, Hall, and imaging records, reductions, calibration, backgrounds, resolution, and instrument-specific forward models.
• Computational owners retain production exchange extraction, Monte Carlo, tensor-network and exact-diagonalization calculations, spin dynamics, convergence, uncertainty, and reproducibility.

Exit Checkpoint

Before leaving the gateway, you should be able to:

• name the magnetic operator, sign, normalization, active degrees of freedom, state, symmetry, and scale window;
• distinguish moment formation, magnetization, susceptibility, spontaneous order, correlation, excitation, transport, and texture;
• choose the narrowest live material and formal owners without treating the sidebar as a prerequisite chain;
• state the Hamiltonian convention, approximation, parameter provenance, and retained versus eliminated variables;
• attach the correct probe forward model and identify the recorded quantity;
• give at least one competing mechanism and one falsification, scaling, reciprocity, stability, or convergence test; and
• stop before converting a fitted feature into a unique phase, topology, or microscopic interaction.

Common Routing Errors

• A local moment is not automatically long-range magnetic order.
• Exchange is not an additional force, and its sign is undefined without the Hamiltonian and pair-counting convention.
• Ferromagnetism is equilibrium uniform spontaneous order; hysteresis and coercivity are protocol-dependent domain phenomena.
• Antiferromagnetism is not defined only by neighboring antiparallel arrows; state the ordering wavevector, magnetic cell, and symmetry.
• Magnetization compensation, angular-momentum compensation, and the Curie transition of a ferrimagnet are different phenomena.
• A two-dimensional label alone does not forbid finite-temperature order or license it; the symmetry, range, anisotropy, coupling, and finite-size limits matter.
• Local and itinerant descriptions can coexist over different scales.
• A Curie–Weiss temperature does not uniquely determine the ordered phase or exchange sign pattern.
• A Stoner criterion is not a Curie-temperature formula or a proof of a ferromagnetic ground state.
• A magnon dispersion without spectral weight, linewidth, continuum, stability, and resolution checks is incomplete.
• A broad magnetic continuum or linewidth does not by itself prove fractionalization; decay, particle–hole, multiparticle, disorder, domain, hybridization, and resolution alternatives remain live.
• Spin–orbit coupling alone does not break time reversal or prove topology.
• Dzyaloshinskii–Moriya coupling is bond-symmetry-constrained antisymmetric exchange, not generic single-ion anisotropy.
• Spin current is convention dependent and generally not conserved with spin–orbit coupling.
• Bulk spin-Hall conductivity, spin-diffusion length, or torque efficiency does not by itself demonstrate device injection, transmission, detection, or switching.
• Texture topology does not establish energetic stability, lifetime, or an experimental identification.

Routing Exercises

1. Classify five moment notions

An equal-time sum rule gives $\langle\hat{\boldsymbol\mu}_i^2\rangle=(4.2\mu_B)^2$, a Curie–Weiss fit gives $\mu_{\mathrm{eff}}=3.9\mu_B$, high-field data approach $2.8\mu_B$ per ion, diffraction gives $2.1\mu_B$, and an inelastic integral gives a larger spectral weight over a stated energy window. Are these inconsistent measurements of one number?

Solution

No. The equal-time local quantity is the square root of a local moment-squared expectation under a declared elastic-plus-inelastic normalization. The Curie effective moment is a thermal fluctuation scale, the high-field value is a protocol- and state-dependent approach to saturation, and diffraction measures an ordered Fourier component within its time and momentum window. The inelastic integral has moment-squared units, omits elastic Bragg weight, and depends on its momentum and energy window, form factor, and normalization; only a declared square root may be called an effective fluctuating moment. Crystal fields, covalency, spin–orbit coupling, itinerancy, incomplete saturation, quantum fluctuations, and missing experimental windows affect these notions differently. Preserve each definition, units, and elastic–inelastic split rather than averaging five values.

2. Audit an exchange sign

One paper writes $H=J\sum_{\langle ij\rangle}\mathbf S_i\cdot\mathbf S_j$ and another writes $H=-2J'\sum_{i<j}\mathbf S_i\cdot\mathbf S_j$. Can their quoted positive couplings be compared directly?

Solution

No. State the overall sign, whether each bond is counted once or twice, the spin normalization, and the mapped orbital basis. In the first convention positive $J$ favors antiparallel alignment on a single bond; in the second, positive $J'$ favors parallel alignment. The factor of two and bond set must be matched before comparing magnitudes. A fitted sign still does not by itself identify direct exchange, superexchange, double exchange, or carrier-mediated exchange.

3. Audit six explanations of bulk magnetometry

A bulk trace is small or zero below a thermal anomaly. What additional evidence would distinguish a ferromagnet, an antiferromagnet, a ferrimagnet, a ferrimagnet at compensation, a multidomain ferro- or ferrimagnet, and a paramagnet with a nonmagnetic anomaly?

Solution

Bulk magnetometry alone does not decide. A selected ferromagnetic state has a uniform spontaneous component, whereas a multidomain sample can cancel it; field training, hysteresis, demagnetizing analysis, and imaging test that distinction. Antiferromagnetism requires a propagation vector, magnetic basis or space group, and staggered order; it can have $\mathbf k=0$. Ferrimagnetism requires opposing inequivalent sublattices with unequal moments away from a possible compensation point. Their temperature dependences can cancel the net magnetization at $T_{\mathrm{comp}}$ while both remain ordered, so compensation is not the Curie transition; angular-momentum compensation is separate when gyromagnetic factors differ. Diffraction, site-resolved moments, local probes, and diffuse correlations test these ordered candidates. If no static order or magnetic correlations survive those tests, retain a paramagnetic state and test whether the anomaly is structural, electronic, or otherwise nonmagnetic.

4. Audit finite-temperature order in two dimensions

A monolayer magnet has a finite ordering temperature. Does that contradict the Mermin–Wagner theorem?

Solution

Not without further information. The theorem concerns an ideal infinite short-range system with continuous symmetry under its stated hypotheses. Magnetocrystalline or exchange anisotropy can reduce the symmetry, dipolar interactions are long ranged, substrate or interlayer coupling can make the system effectively three dimensional, and finite size can produce a crossover. An XY regime can instead support BKT physics. Measure or bound the relevant anisotropy, coupling, size, and correlation lengths before claiming either a violation or the mechanism of order.

5. Diagnose local versus itinerant magnetism

A metal shows a Curie–Weiss regime at high temperature, coherent quasiparticles at low temperature, and broad magnetic spectral weight. Must it be classified as purely local or purely itinerant?

Solution

No. Local and itinerant are scale-dependent limits. Compare instantaneous and ordered moments, spectral-weight redistribution, band reconstruction, susceptibility in momentum and frequency, charge and spin coherence scales, and material-specific calculations. A system can retain sizable fluctuating moments while supporting coherent low-energy quasiparticles. The bounded claim should state the energy, temperature, momentum, and time window in which each description works.

6. Fit a magnon responsibly

Two spin Hamiltonians give nearly identical magnon energies along one high-symmetry path. What additional evidence can distinguish them?

Solution

Compare eigenvectors and probe-weighted intensities across the full accessible Brillouin zone, polarization selection, branch degeneracies, field and temperature dependence, linewidths and continua, sum rules, and other magnetic domains. Convolve both predictions with the same resolution and fit shared structural and form-factor inputs. Stable energies alone do not certify the reference state, exchange tensor, anisotropy, or omitted interactions.

7. Limit a Stoner conclusion

A paramagnetic calculation reports a spin-summed density of states $N_{\mathrm{tot}}(E_F)=2.4$ states/(eV cell) and $I=0.50$ eV, then claims $IN_{\mathrm{tot}}(E_F)=1.2>1$. In the same interaction convention, $\lambda_{\max}[\boldsymbol\Gamma\boldsymbol\chi_0]$ is $0.60$ at $\mathbf q=0$ but $1.08$ at a concrete nesting vector $\mathbf Q$. Audit the claim.

Solution

For the common paramagnetic convention in which the Stoner criterion uses the per-spin density of states, spin degeneracy gives $N_\uparrow(E_F)=N_{\mathrm{tot}}(E_F)/2=1.2$ states/(eV cell), so the matched product is $IN_\uparrow(E_F)=0.60$, not $1.2$. A different definition of $I$ could absorb that factor, which is why both conventions must be stated together. The uniform channel is locally stable in the declared calculation, while the $1.08$ eigenvalue of $\boldsymbol\Gamma\boldsymbol\chi_0$ makes an eigenvalue of $\mathbf1-\boldsymbol\Gamma\boldsymbol\chi_0$ pass through zero and therefore identifies a finite-$\mathbf Q$ local instability of the chosen reference state within the chosen static approximation. It does not by itself determine the ordered structure, moment, transition temperature or order, fluctuation corrections, or validity of the band input. Resolve the eigenvector and magnetic basis, compare other wavevectors and first-order competitors, and attach experimental evidence before naming the phase.

8. Compare Kondo and RKKY claims

A dense magnetic metal has a resistivity minimum and an ordering anomaly. Does that locate a universal crossing of Kondo and RKKY scales?

Solution

No. First exclude weak localization, interaction corrections, structural or magnetic scattering, multiband transport, and phonon backgrounds. Define the chosen Kondo scale operationally and test impurity or lattice coherence using thermodynamics, spectroscopy, field response, and dilution where possible. RKKY depends on the full susceptibility, dimension, Fermi surface, spin–orbit coupling, temperature, and disorder. A Doniach-style competition is a useful organizer, not a universal phase-boundary equation.

9. Audit a topological Hall and torque claim

A spin–orbit-coupled film shows a residual Hall hump and current-driven domain motion. What is still required before claiming skyrmions and a spin-transfer torque mechanism?

Solution

Decompose ordinary and anomalous Hall contributions with reversal, history, thickness, contact, inhomogeneity, and multiband checks. Establish the magnetic phase boundaries and use imaging, scattering, or another orthogonal probe to identify the texture and its density. Define the spin-current and torque convention, account for spin–orbit torque, Oersted fields, heating, pinning, and interface conversion, and verify the predicted current, field, symmetry, and polarity dependence. A Hall residual and motion alone prove neither texture topology nor one torque mechanism.

References

• A. Auerbach, Interacting Electrons and Quantum Magnetism (Springer, 1994).
• S. Blundell, Magnetism in Condensed Matter (Oxford University Press, 2001).
• J. M. D. Coey, Magnetism and Magnetic Materials (Cambridge University Press, 2010).
• P. Coleman, Introduction to Many-Body Physics (Cambridge University Press, 2015).
• I. Dzyaloshinsky, “A thermodynamic theory of ‘weak’ ferromagnetism of antiferromagnetics,” Journal of Physics and Chemistry of Solids 4, 241–255 (1958), doi:10.1016/0022-3697(58)90076-3.
• A. C. Hewson, The Kondo Problem to Heavy Fermions (Cambridge University Press, 1993).
• A. Hubert and R. Schäfer, Magnetic Domains (Springer, 1998).
• T. Moriya, Spin Fluctuations in Itinerant Electron Magnetism (Springer, 1985).
• T. Moriya, “Anisotropic Superexchange Interaction and Weak Ferromagnetism,” Physical Review 120, 91–98 (1960), doi:10.1103/PhysRev.120.91.
• N. Nagaosa and Y. Tokura, “Topological properties and dynamics of magnetic skyrmions,” Nature Nanotechnology 8, 899–911 (2013), doi:10.1038/nnano.2013.243.
• R. M. White, Quantum Theory of Magnetism, 3rd ed. (Springer, 2007).
• I. Žutić, J. Fabian, and S. Das Sarma, “Spintronics: Fundamentals and applications,” Reviews of Modern Physics 76, 323–410 (2004), doi:10.1103/RevModPhys.76.323.