Magnetic Monopole

Every magnet is a dipole

Cut a bar magnet in half. You do not get an isolated north pole on one side and an isolated south pole on the other — you get two smaller bar magnets, each with its own north and south. Cut again and the pattern persists. Down to atomic level, every magnetic source we have ever seen consists of dipoles: closed loops of magnetic field where every line that leaves a north pole eventually re-enters a south. This is what Maxwell's equation ∇·B = 0 says formally: the magnetic field has no sources or sinks, no places where field lines simply begin or end.

Contrast this with electricity. An electron sits at one end of an electric-field line; the field radiates outward and terminates only on a positive charge elsewhere. Take away the positive charge and the electric field still emerges from the electron and runs off to infinity. That kind of isolated source has never been seen for magnetism. A magnetic monopole would be the magnetic version: a particle whose magnetic-field lines emerge radially from a point and stretch to infinity without ever coming back.

The two fields look mathematically similar but obey fundamentally different constraints. The asymmetry has bothered physicists for 200 years: why does electricity have monopoles but magnetism doesn't? Maxwell's equations as written enforce this asymmetry by assuming ∇·B = 0; the moment a real monopole appears, the equations must be modified.

Dirac's 1931 argument: monopoles force charge quantization

In 1931 Paul Dirac was wrestling with quantum mechanics in the presence of electromagnetic potentials. The vector potential A (where B = ∇×A) is needed to write Schrödinger's equation in a magnetic field, but A is gauge-dependent: a wave function with potential A + ∇χ is physically equivalent to the original via a phase shift ψ → e^(ieχ/ℏ) ψ. The phase shift must be single-valued — go around any closed loop and end at the same value.

Dirac asked: what happens if we try to write A for a magnetic monopole? The radial B field of a monopole is the curl of no globally smooth A. To even write down A you need a singular line — the "Dirac string" — running from the monopole to infinity, along which A has a coordinate singularity. The Dirac string is mathematical fiction; physically it must be undetectable.

Undetectability of the string requires that a quantum wave function picking up a phase as it loops around the string return to its original value. That phase shift is e·Φ/ℏ where Φ = 4πg/μ₀ is the magnetic flux exiting the monopole. Setting this equal to 2πn (so the wave function is single-valued) gives

e · g = n · h / 2,  n = 1, 2, 3, ...

This is the Dirac quantization condition. With even one monopole anywhere in the universe — Dirac emphasized that "one" is enough — every electric charge in existence must be an integer multiple of the smallest allowed unit e₀ = h/(2g). The argument runs both ways: if charge is quantized (which it observably is), there exists a self-consistent quantum mechanics that includes monopoles. Dirac considered it "surprising that nature has not made use of this possibility."

Maxwell's equations: symmetric vs not

Equation	Standard Maxwell	With monopoles	What changed
Gauss for E	∇·E = ρ_e / ε₀	∇·E = ρ_e / ε₀	Same — electric monopoles are real
Gauss for B	∇·B = 0	∇·B = μ₀ ρ_m	Magnetic charge ρ_m sources B
Faraday	∇×E = −∂B/∂t	∇×E = −μ₀ J_m − ∂B/∂t	Magnetic current J_m sources circulating E
Ampère-Maxwell	∇×B = μ₀J_e + μ₀ε₀ ∂E/∂t	∇×B = μ₀J_e + μ₀ε₀ ∂E/∂t	Same — already symmetric
Duality	Broken (E ↔ B asymmetric)	Restored under (E,B) → (cB, −E/c) and (e,g) → (g, −e)	Perfect symmetry
Field around point source	B is a dipole; E is monopole	Both have monopole solutions	Magnetic radial field B = g/(4πr²)

The corrected equations are mathematically beautiful — perfect E ↔ B symmetry — and quantum-mechanically allowed by Dirac's argument. The reason we use the asymmetric version of Maxwell's equations is purely empirical: no monopoles have ever been observed, so we set ρ_m = J_m = 0.

't Hooft-Polyakov: the topological monopole

Dirac's monopole is a fundamental point particle with a singular Dirac string. In 1974 Gerard 't Hooft and Alexander Polyakov independently found a fundamentally different kind: a smooth, non-singular soliton solution of a non-abelian gauge theory. If a simple gauge group like SU(2) breaks down to U(1) via the Higgs mechanism, the broken-symmetry vacuum has non-trivial topology — the second homotopy group π₂(G/H) is non-trivial — and finite-energy field configurations include monopoles.

The 't Hooft-Polyakov monopole has mass M ~ M_W / α, where M_W is the gauge-boson mass and α the coupling constant. For an SU(5) grand unified theory with symmetry breaking at the GUT scale ~10¹⁶ GeV, the predicted monopole mass is ~10¹⁷ GeV (~1.8 × 10⁻⁸ g — comparable to a flake of dust). They carry magnetic charge g = 2π/e (twice the Dirac value), and their formation should have been prolific in the early hot universe.

This is the "monopole problem" in cosmology: hot Big Bang models predict roughly one monopole per Hubble volume at the GUT epoch, which redshifts to a present-day density that would dwarf all observed matter. Alan Guth's 1980 inflation proposal was originally motivated by the need to dilute these unobserved monopoles below the matter density of the universe. Inflation's prediction of a flat, large-scale-homogeneous universe was, in the original telling, a by-product of solving the monopole problem.

Experimental searches: a century of nothing

• The 1982 Cabrera event. Blas Cabrera's SQUID coil at Stanford, on the night of 14 February 1982, recorded a single magnetic-charge-like flux transition consistent with one monopole passing through. The signal was at the Dirac quantization (Φ₀ = 2g_D). Despite years of additional running with the same and larger detectors, the event was never reproduced. Cabrera's setup limit became roughly 10⁻¹⁰ monopoles/cm²/s.
• MoEDAL at the LHC. The Monopole and Exotics Detector at LHCb uses nuclear-track plastic and trapping arrays to look for monopoles in the 10⁻¹⁴ kg mass range produced by proton-proton collisions at 13 TeV. As of the 2024 results, no monopole signal has been seen, with limits below ~10⁻⁴ pb for masses up to 6 TeV.
• IceCube at the South Pole. A km³ of instrumented Antarctic ice searches for ultra-relativistic monopoles via their predicted Cherenkov-like radiation pattern (a monopole moving at v > c/n radiates with intensity 8200× a same-velocity electron, since its coupling is g²/ℏc instead of e²/ℏc). IceCube limits cosmic monopole flux below 10⁻¹⁹ /cm²/s/sr at energies above 10⁸ GeV.
• Lunar samples and meteorites. Two grams of lunar regolith examined by SQUID, terrestrial deep-mine ferromagnetic samples, and chunks of iron meteorites have all been screened for trapped monopoles. The cumulative null result limits monopole density in ordinary matter below ~10⁻²⁹ per nucleon.
• The Parker bound. If too many monopoles existed in the galaxy, they would drain energy from the galactic magnetic field faster than the dynamo can replenish it. Eugene Parker's 1970 argument gives an upper limit on cosmic monopole flux of ~10⁻¹⁵ /cm²/s/sr — a model-independent constraint from astrophysics.

Emergent monopoles in condensed matter

The Standard Model has no fundamental magnetic monopole. But certain materials — specifically "spin ice" pyrochlores like Dy₂Ti₂O₇ and Ho₂Ti₂O₇ — host quasiparticle excitations that behave mathematically like monopoles within the material. The spins on each tetrahedron of the pyrochlore lattice obey "two-in/two-out" ice rules; a violation (three-in/one-out or vice versa) creates a localized source of magnetization divergence — formally identical to a magnetic monopole obeying ∇·M ≠ 0.

Castelnovo, Moessner, and Sondhi showed in 2008 that these emergent monopoles interact via a Coulomb law with effective magnetic charge g* ≈ 5 × 10⁻¹³ J/T·m. They are real in the sense that they can be observed by neutron scattering and respond to applied magnetic fields, but they are not fundamental: they exist only inside the spin-ice material and have no relevance to particle physics or charge quantization. The distinction is essential: emergent monopoles do not satisfy Dirac's quantization for free-space electric charge.

Why physicists keep looking

• It is the only known explanation for charge quantization. Every electron carries exactly −e, every proton exactly +e, every quark ±e/3 or ±2e/3. There is no other reason — gauge symmetry alone permits any real number for the charge. Dirac's argument is the only fundamental mechanism that makes integer charges inevitable.
• It is generic in grand unified theories. SU(5), SO(10), E₆, and every other GUT candidate predicts monopoles automatically as topological defects of symmetry breaking. Finding one would be direct evidence of physics far above the electroweak scale.
• It would change electromagnetism's fundamental status. Maxwell's equations become fully E ↔ B dual the moment a monopole appears. This is exactly the kind of duality (analogous to S-duality in string theory) that mathematical physics has spent decades pursuing.
• Inflation needs them gone. Cosmic inflation's original motivation was the monopole problem. If we ever found monopoles, it would tell us either inflation didn't work as advertised or there is a post-inflation production mechanism we don't understand.
• The Cabrera event remains unexplained. A single, well-instrumented SQUID transition in 1982 is consistent with a monopole and inconsistent with known noise sources. It has never been ruled out, only never reproduced.

Variants and extensions

• Dyon. A particle carrying both electric and magnetic charge. Julian Schwinger and Edward Witten studied dyons; they obey a generalized quantization condition e₁g₂ − e₂g₁ = nh/2 (Witten effect: a CP-violating angle in the gauge theory shifts the electric charge of magnetic monopoles).
• BPS monopole. A 't Hooft-Polyakov monopole at the Bogomol'nyi-Prasad-Sommerfield limit, where the Higgs self-coupling is zero. It is exactly stable, saturates the BPS mass bound, and is a key object in studies of N = 2 supersymmetric gauge theories.
• Cosmic strings and texture. Other topological defects from symmetry breaking. Where π₂(G/H) gives monopoles, π₁ gives cosmic strings and π₃ gives textures. None has been observed, but all are predicted by various GUT scenarios.
• Monopole catalysis of proton decay. Rubakov and Callan showed that a passing GUT monopole could catalyze the decay p → e⁺ π⁰ with a cross-section of geometrical size. Failure to observe such induced proton decay places further limits on terrestrial monopole flux.
• Synthetic monopoles in ultracold atoms. Spinor Bose-Einstein condensates with engineered synthetic gauge fields can host Dirac-monopole field configurations in the synthetic vector potential. Saari and others demonstrated synthetic Dirac monopoles in ⁸⁷Rb BECs in 2014 — proof of principle that the mathematical structure is realizable even if the fundamental particle isn't.

Common pitfalls

• Confusing bar-magnet poles with monopoles. The north end of a bar magnet looks like a monopole at long range but is half of a dipole at short range. Cutting it doesn't isolate the pole; it makes two new dipoles. True monopoles are point particles, not regions of polarized matter.
• Treating ∇·B = 0 as a fundamental law. It is an empirical statement, not a derived one. Maxwell wrote it as an axiom because no magnetic charge has been observed. The moment a monopole turns up, the law becomes ∇·B = μ₀ ρ_m.
• Forgetting that one monopole is enough. Dirac's quantization condition kicks in if any monopole exists anywhere in the universe at any time. We don't need to find one in our lab — we need to know they exist somewhere. The fact that all observed electric charges are integer multiples of e is consistent with there being monopoles, but doesn't prove it.
• Confusing emergent and fundamental monopoles. Spin-ice quasiparticles obey ∇·M ≠ 0 inside the material — they behave like monopoles for the magnetization field. They are not fundamental magnetic charges. They cannot leave the spin ice, and their existence has no bearing on Dirac quantization.
• Mistaking the Dirac string for a real object. The Dirac string is a coordinate artifact, not a physical line. It must carry zero observable effect — and the condition that it be invisible is exactly what gives charge quantization. Modern fiber-bundle formulations avoid the string entirely.