CM01 — Spin Ice Magnetic Monopoles
| Field | Value |
|---|---|
| Domain | Condensed Matter / Frustrated Magnetism |
| System | Pyrochlore spin ice Dy₂Ti₂O₇ or Ho₂Ti₂O₇ |
| Group | U(1) (emergent gauge field) |
| H^k tier | H⁰ (ice rules) → H² (monopole pair creation via BIND) |
| ISA | Forge (ice rules) → Origami Valence (monopole dynamics) |
| Status | Validated (Castelnovo et al. 2008; Morris et al. 2009) |
| Opcodes | ORBIT · LINK · BIND · MERGE · TWIST · FLIP |
| Papers | Paper 672 (notes); OC01 (related) |
Physical system
In pyrochlore spin ice (Dy₂Ti₂O₇, Ho₂Ti₂O₇), rare-earth magnetic moments (“spins”) sit on a network of corner-sharing tetrahedra. The ice rules — two spins pointing in, two pointing out per tetrahedron — define the ground state manifold. Castelnovo, Moessner and Sondhi (2008) showed that topological defects in the ice rule are emergent magnetic monopoles: a tetrahedron with 3-in/1-out carries effective magnetic charge +q; one with 1-in/3-out carries −q. These charges obey the Dirac quantisation condition 2eg = nℏc, just as Dirac monopoles do.
The monopoles interact via a magnetic Coulomb law V(r) = μ₀q²/4πr, are connected by Dirac strings (chains of reversed spins carrying no energy in spin ice — the string tension is zero), and have been directly observed via neutron scattering (Morris et al. 2009, Fennell et al. 2009).
Why this entry requires BIND
This is the first ISA zoo entry in condensed matter where BIND is irreducible — the programme cannot be written using only H⁰/H¹ opcodes.
The reason: monopole pair creation from the vacuum (the ice-rule ground state) is the Frobenius comultiplication
\[\delta: I \;\longrightarrow\; |\mathrm{mono}^+\rangle \otimes |\mathrm{mono}^-\rangle\]This maps the vacuum (unit object I) to a two-monopole state. In the ISA, this is exactly BIND: a single state splitting into two entangled objects with opposite topological charge. No sequence of ORBIT + TWIST + MERGE + LINK can produce two objects from one — only BIND (δ: A → A⊗A) can, and its degenerate form (δ: I → A⊗A, creation from vacuum) is the monopole pair creation event.
Contrast with G01 (Yang-Mills instantons): there, BIND computes the second Chern class ∫tr(F∧F) — a topological invariant of an existing configuration. Here, BIND creates the topological objects themselves.
Target category
IceGauge — the emergent U(1) gauge theory of spin ice, whose objects are monopole charge sectors (charge q ∈ ℤ × q₀) and whose morphisms are gauge-equivariant spin-flip sequences. The Dirac string is the morphism connecting +q and −q objects; its length is gauge-dependent (unobservable) while the endpoint charges are gauge-invariant.
Interpretation functor
F: Origami ISA → IceGauge defined by:
| Opcode | F(opcode) |
|---|---|
| ORBIT | Ice-rule ground state: 2-in/2-out tetrahedron. Eigenvalue = zero magnetic charge. The “orbital” is the tetrahedron’s charge sector. |
| BIND | Monopole pair creation: flip a spin to create a 3-in/1-out (+q) and 1-in/3-out (−q) pair from a 2-in/2-out ground state. The two monopoles are entangled — they must be created together (total charge conservation). |
| LINK(+q, −q, ν=1) | Dirac string connecting the monopole pair. Linking number ν = number of reversed spins in the connecting chain. String tension = 0 in spin ice (unlike in a real magnet). |
| TWIST | Extend or reroute the Dirac string — a Reidemeister move on the spin chain. Physical realisation: flipping a spin along the string moves one monopole by one lattice step. |
| MERGE | Monopole annihilation: +q and −q meet and annihilate back to the ice-rule ground state. Frobenius multiplication μ: A⊗A → A (or → I at the vacuum). |
| FLIP | Neutron scattering measurement: project onto a specific monopole charge sector. Readout of the ±q eigenvalue. |
| SNAP↑ | Freeze-out transition: below T* ≈ 1K, monopole pair creation becomes thermally suppressed; the system freezes into the ice-rule manifold. SNAP↑ marks the entry into the monopole-active (H²) regime as T increases above T*. |
ISA programme
-- Monopole pair creation and separation in spin ice
GROUND: ORBIT[|ice⟩ | charge=0] -- ice-rule ground state
CREATE: BIND[|mono+⟩ ⊗ |mono−⟩] -- pair creation (one spin flip)
STRING: LINK[mono+, mono−, ν=1] -- Dirac string (1 reversed spin)
SEPARATE: TWIST × N -- N spin flips extend string by N steps
COULOMB: LABEL[V = μ₀q²/4πr] -- magnetic Coulomb potential (r = Nа)
OBSERVE: FLIP[neutron scattering | ±q sector] -- experimental readout
ANNIHILATE: MERGE[mono+, mono−] -- pair annihilation → ground state
The full programme is: ORBIT → BIND → LINK → TWIST^N → FLIP → MERGE
The minimum programme that cannot be shortened: BIND is irreducible. No rewriting in {ORBIT, TWIST, MERGE, LINK, FLIP} produces |mono+⟩⊗|mono−⟩ from |ice⟩.
Computable output
- Monopole charge: q = ±2|J|/a (J = nearest-neighbour exchange, a = pyrochlore lattice constant). For Dy₂Ti₂O₇: q ≈ 4.6 μ_B/Å.
- Dirac string tension: exactly zero in nearest-neighbour spin ice (the “Coulomb phase”). Non-zero corrections appear from next-nearest-neighbour interactions — a Forge soft-threshold effect at finite β.
- Monopole density: ρ ∝ exp(−Δ/k_BT) where Δ = energy cost of one BIND event (one spin flip out of ice rule). For Dy₂Ti₂O₇: Δ ≈ 4K → ρ peaks around T ≈ 2K.
- Magnetic Coulomb law: V(r) = μ₀q²/4πr confirmed by neutron diffuse scattering (Fennell et al. 2009 Science 326, 415) — direct validation that the emergent monopoles are genuine magnetic charges.
Connection to the ISA framework
Spin ice monopoles are the only experimental H² condensed-matter system where BIND creates topological objects (rather than computing a topological invariant of an existing field configuration):
| Entry | BIND role | Status | ||
|---|---|---|---|---|
| G01 (YM instanton) | Computes c₂ = ∫tr(F∧F) | Classical solution | ||
| G03 (Higgs) | SNAP↑ removes LINK (gauge breaking) | QFT | ||
| CM01 (spin ice) | **Creates | mono+⟩⊗ | mono−⟩ from vacuum** | Experimental ✓ |
| OC01 (Furey) | TWIST within octonionic H² fibre | Algebraic |
Is this BIND associative or non-associative? The spin ice monopole BIND is associative: the underlying algebra is ℝ³ (the crystal lattice), and magnetic charge addition is commutative and associative. The non-associative octonionic BIND (G₂ 3-form) would require a material with G₂ lattice symmetry and monopole-monopole-monopole three-body interactions of the form φ_{ijk}. Such a material is not yet synthesised but is predicted by Paper 672 (§7, G₂ 3-form test) to show a qualitatively different phase diagram.
Relation to Dirac monopoles (OC01 §6): spin ice monopoles are emergent analogues of Dirac monopoles. The Dirac quantisation condition (LINK linking number quantisation) holds for both — this is not a coincidence but the same categorical statement (c₁ of the U(1) bundle ∈ ℤ) realised on different substrates: fundamental (Dirac) vs emergent (spin ice).
MGE connection: the Forge ISA (finite β) describes the thermal fluctuations of the ice-rule ground state — the partition function at temperature T = 1/β. The spin ice phase transition at T* is a Forge β* snap: below T* the system is frozen in H⁰ (ice rules dominate), above T* it is active in H² (monopole pair creation BIND events). The hysteresis of the T* transition maps to the MGE snap hysteresis of Paper 596.
Validation
- Pair creation: Castelnovo, Moessner & Sondhi (2008) Nature 451, 42. Showed ice-rule defects carry magnetic charge ±q with Coulomb interactions. Exact mapping to Dirac monopoles via emergent U(1) gauge field.
- Neutron scattering: Morris et al. (2009) Science 326, 411; Fennell et al. (2009) Science 326, 415. Direct observation of monopole diffuse scattering and magnetic Coulomb law V(r) ∝ 1/r confirmed.
- Dirac string: Giblin et al. (2011) Nature Physics 7, 252. Direct visualisation of Dirac strings as chains of reversed spins.
- Monopole current: Bramwell et al. (2009) Nature 461, 956. Measurement of monopole drift in applied field — confirms charge ±q and mobility.
Part of the ISA Zoo. See also: G01 — Yang-Mills Instantons, OC01 — Furey Fermion Ladder. Background theory: Paper 672 (where H² begins).