Positronium: The Atom Built From Matter and Its Own Antimatter

A hydrogen atom with the mass symmetry cranked to the maximum

Positronium is the bound state of an electron (e⁻) and a positron (e⁺), held together by the same Coulomb attraction that binds hydrogen: F = e²/(4πε₀ r²). Structurally it looks like hydrogen — an S/P/D level scheme, a Rydberg-like spectrum — but there is one decisive difference. In hydrogen the proton is 1836× heavier than the electron, so it sits essentially still while the electron orbits. In positronium both partners have identical mass mₑ = 9.109 × 10⁻³¹ kg (rest energy 511 keV), so both orbit their common center of mass.

• The correct mass to use is the reduced mass μ = mₑmₑ/(mₑ + mₑ) = mₑ/2. Every energy scales with μ, so positronium's binding is exactly half hydrogen's.
• Ground-state binding energy: E₁ = −μe⁴/(8ε₀²h²) = −6.80 eV, precisely one-half the 13.6 eV Rydberg of hydrogen.
• The Bohr radius, ∝ 1/μ, doubles: a_Ps = 2a₀ = 1.06 × 10⁻¹⁰ m. Positronium is a physically bigger, more loosely bound atom.

The full spectrum follows E_n = −6.80 eV/n². The Lyman-α analog (2P → 1S) sits at 5.10 eV, and the whole optical fingerprint is red-shifted relative to hydrogen by the factor of two in the Rydberg.

Spin decides everything: para versus ortho

Because e⁻ and e⁺ are both spin-½ fermions, their spins couple into a total spin S = 0 (singlet, antiparallel) or S = 1 (triplet, parallel). These are the two flavors of positronium, and their fates could not be more different.

• para-positronium (p-Ps, ¹S₀): S = 0, one spin substate. It can annihilate into an even number of photons, dominantly two.
• ortho-positronium (o-Ps, ³S₁): S = 1, three spin substates. Charge-conjugation and angular-momentum rules forbid two-photon decay; it must emit an odd number, dominantly three.

When positronium forms from unpolarized positrons and electrons, spin statistics give a 3:1 ratio — 75% ortho, 25% para — set purely by the 3 versus 1 substate multiplicity. The two states are split in energy by the hyperfine (spin-spin) interaction. Uniquely for positronium this splitting is huge — Δν_HFS = 203.4 GHz, corresponding to 0.84 meV — because on top of the usual magnetic spin-spin term there is a genuinely relativistic virtual annihilation contribution (e⁺e⁻ → single virtual photon → e⁺e⁻) that has no analog in hydrogen. Measuring this 203 GHz splitting to parts-per-million is a headline QED test.

Why it dies: annihilation and the α-scaling of lifetimes

Positronium is metastable because matter and antimatter, brought together, mutually annihilate. The two-body annihilation rate is governed by the probability that the two point particles overlap — the wavefunction density at the origin, |ψ(0)|² — times a cross-section built from the electromagnetic coupling α = e²/(4πε₀ℏc) ≈ 1/137.

• para-Ps → 2γ. The rate is Γ_p = ½ mₑc²α⁵/ℏ, giving a lifetime τ_p ≈ 125 ps and a decay rate of about 8.0 × 10⁹ s⁻¹. The two 511 keV photons fly out back-to-back (momentum conservation in the rest frame) — this is exactly the annihilation signal a PET scanner detects.
• ortho-Ps → 3γ. Forbidden from two-photon decay, it needs an extra vertex, so its rate carries an extra factor of α: Γ_o ∝ α⁶. The result is a lifetime of τ_o ≈ 142 ns — about 1140× longer than para-Ps. The three photons share the 1.022 MeV total energy with a continuous spectrum, each ≤ 511 keV.

The controlling numbers are stark: each additional photon costs a factor of α ≈ 1/137 in rate, and the whole scale is set by the natural QED clock ℏ/(mₑc²α²) ≈ 2.4 × 10⁻¹⁷ s multiplied up by powers of 1/α. The measured o-Ps rate, 7.040 μs⁻¹, matches QED including higher-order corrections to better than a part in 10⁴ — a famous confirmation once clouded by the 1980s–90s "orthopositronium lifetime puzzle," now resolved in QED's favor.

A worked feel: sizes, speeds, and how long an orbit lasts

Numbers make positronium concrete. Take the ground state and use the Bohr picture with reduced mass μ = mₑ/2.

• Orbital speed: v ≈ αc ≈ (1/137)(3.0 × 10⁸ m/s) ≈ 2.2 × 10⁶ m/s — the same αc as hydrogen, because v ∝ Z but is mass-independent. So positronium is non-relativistic to a part in ~10⁴, and relativistic (v²/c² ~ α²) corrections enter at the 10⁻⁴ level, right where the fine and hyperfine structure live.
• Orbital period: with radius a_Ps = 1.06 × 10⁻¹⁰ m and the two particles circling the midpoint, T = 2π(a_Ps/2)/v ≈ 1.5 × 10⁻¹⁶ s. That means para-Ps completes roughly τ_p/T ≈ 125 ps / 0.15 fs ≈ 8 × 10⁵ orbits before annihilating, and ortho-Ps nearly a billion. It is genuinely an atom, not a fleeting collision.
• Wavefunction at the origin: for the 1S state |ψ(0)|² = 1/(π a_Ps³) ≈ 1/[π(1.06 × 10⁻¹⁰ m)³] ≈ 2.7 × 10²⁹ m⁻³. This finite overlap is precisely what makes annihilation possible — and why S-states (which have nonzero density at r = 0) annihilate promptly while P-states, with ψ(0) = 0, must first radiatively cascade down.

The controlling variables: what you can tune and what you can't

Positronium's behavior is dictated by a short list of physical inputs, and knowing which knobs exist explains both the laboratory subtleties and the technology.

• Spin state (fixed at formation): ortho vs para sets the entire decay channel and the 1140× lifetime gap. External magnetic fields mix the m = 0 substates of ortho and para (Zeeman effect), so a field can quench the long-lived o-Ps by giving it a para-like 2γ channel — the basis of magnetic-quenching lifetime measurements.
• Environment / density: in matter, o-Ps rarely lives its full 142 ns. Pick-off annihilation — the positron annihilating with a nearby electron of opposite spin from the surrounding medium — shortens its life. The residual o-Ps lifetime in a material maps directly onto the size of the free-volume voids it samples (typical pore diameters 0.3–1 nm), which is the entire basis of positron annihilation lifetime spectroscopy (PALS).
• Kinetic energy: positronium can be formed hot and then laser-cooled; recent experiments have chilled Ps clouds toward the sub-kelvin regime, a prerequisite for making a Bose-Einstein condensate of positronium or a gamma-ray annihilation laser.
• Nuclear charge: unlike hydrogen there is none — the "nucleus" is a positron of charge +e. This is why Z is locked at 1 and why there is no isotope structure, no nuclear spin, and no volume shift.

Where it shows up: PET scanners, materials science, and precision QED

Positronium is not a curiosity — it is a working tool and a precision instrument.

• Medical PET imaging. A tracer like ¹⁸F emits a positron that thermalizes and annihilates, roughly 40% of the time via positronium, producing the pair of 511 keV back-to-back photons that PET detectors register in coincidence. Emerging positronium-lifetime PET reads the o-Ps lifetime inside tissue to probe oxygenation and microstructure, adding biochemical contrast beyond anatomy.
• Materials & polymer science (PALS). By timing o-Ps decays, engineers measure sub-nanometer free volume in plastics, membranes, aerogels, and low-κ dielectrics. A shift of a fraction of a nanosecond in the o-Ps lifetime signals a change in pore size, aging, or crystallinity that no optical probe can see.
• Precision QED. Because positronium is a pure leptonic bound state — no messy nuclear structure — its energies are calculable in QED from first principles. The 1S–2S interval (1 233 607 216.4 MHz) and the 203.4 GHz hyperfine splitting are measured and compared with theory to test QED, search for exotic decays, and hunt for tiny discrepancies that could signal new physics. Positronium is also a benchmark for the gravitational behavior of antimatter and for putative fifth forces.

Subtleties and misconceptions

Several intuitive-sounding statements about positronium are wrong, and untangling them sharpens the physics.

• "Matter and antimatter annihilate instantly on contact." No — they orbit for up to 10⁹ revolutions. Annihilation is a probabilistic QED process whose rate depends on |ψ(0)|²; it is fast but not instantaneous, and a P-state with ψ(0) = 0 barely annihilates at all until it cascades to an S-state.
• "All the mass turns into two 511 keV gammas." Only for para-Ps. Ortho-Ps forbids the 2γ channel and instead makes three photons with a continuous energy spectrum. The clean back-to-back pair belongs specifically to the singlet.
• "The positron is the nucleus." There is no heavy nucleus — the reduced mass is mₑ/2, so both particles share the orbital motion equally. This is why every energy and length differs from hydrogen by factors of two, not by a small correction.
• "Hyperfine structure is a minor detail." In positronium the ortho–para splitting is 0.84 meV (203 GHz) — a substantial fraction of the fine structure — because the virtual annihilation term, absent in ordinary atoms, adds a genuinely relativistic contribution comparable to the magnetic spin-spin term.

Para-positronium (singlet) versus ortho-positronium (triplet): the two spin states differ dramatically in how they die.

Property	para-Ps (¹S₀)	ortho-Ps (³S₁)
Total spin S	0 (antiparallel)	1 (parallel)
Statistical weight	1 (25% formed)	3 (75% formed)
Dominant decay	2 photons	3 photons
Vacuum lifetime	125 ps	142 ns (~1140× longer)
Photon energy each	511 keV (back-to-back)	≤ 511 keV (continuous)
Decay rate ∝	α⁵ (m c²/ℏ)	α⁶ (m c²/ℏ)