An electron and positron orbiting around their common centre of mass. (An s state has zero angular momentum; so orbiting around each other would mean going straight at each other until scattered or until annihilation, whichever comes first). This is a bound quantum state known as positronium.
The mass of positronium is 1.022 MeV, which is twice the electron mass minus the binding energy of a few eV. The ground state of positronium, like that of hydrogen, has two possible configurations depending on the relative orientations of the spins of the electron and the positron.
The singlet state, 1 S 0, with antiparallelspins (S = 0, Ms = 0) is known as para-positronium (p-Ps). It has a mean lifetime of 0.125 ns and decays preferentially into two gamma rays with energy of 511 keV each (in the center-of-mass frame). By detecting these photons the position at which the decay occurred can be determined. This process is used in positron-emission tomography. Para-positronium can decay into any even number of photons (2, 4, 6, ...), but the probability quickly decreases with the number: the branching ratio for decay into 4 photons is 1.439(2)×10−6.
Para-positronium lifetime in vacuum is approximately
The triplet state, 3S1, with parallel spins (S = 1, Ms = −1, 0, 1) is known as ortho-positronium (o-Ps). It has a mean lifetime of 142.05±0.02 ns, and the leading decay is three gammas. Other modes of decay are negligible; for instance, the five-photons mode has branching ratio of ≈10−6.
Ortho-positronium lifetime in vacuum can be calculated approximately as:
However more accurate calculations with corrections to O(α²) yield a value of 7.040 μs−1 for the decay rate, corresponding to a lifetime of 142 ns.
Positronium in the 2S state is metastable having a lifetime of 1100 ns against annihilation. The positronium created in such an excited state will quickly cascade down to the ground state, where annihilation will occur more quickly.
Annihilation can proceed via a number of channels, each producing gamma rays with total energy of 1022 keV (sum of the electron and positron mass-energy), usually 2 or 3, with up to 5 recorded.[clarification needed]
The annihilation into a neutrino–antineutrino pair is also possible, but the probability is predicted to be negligible. The branching ratio for o-Ps decay for this channel is 6.2×10−18 (electron neutrino–antineutrino pair) and 9.5×10−21 (for other flavour) in predictions based on the Standard Model, but it can be increased by non-standard neutrino properties, like relatively high magnetic moment. The experimental upper limits on branching ratio for this decay (as well as for a decay into any "invisible" particles) are <4.3×10−7 for p-Ps and <4.2×10−7 for o-Ps.
While precise calculation of positronium energy levels uses the Bethe–Salpeter equation or the Breit equation, the similarity between positronium and hydrogen allows a rough estimate. In this approximation, the energy levels are different because of a different effective mass, m*, in the energy equation (see electron energy levels for a derivation):
where me and mp are, respectively, the mass of the electron and the positron (which are the same by definition as antiparticles).
Thus, for positronium, its reduced mass only differs from the electron by a factor of 2. This causes the energy levels to also roughly be half of what they are for the hydrogen atom.
So finally, the energy levels of positronium are given by
The lowest energy level of positronium (n = 1) is −6.8 electronvolts (eV). The next level is −1.7 eV. The negative sign is a convention that implies a bound state. Positronium can also be considered by a particular form of the two-body Dirac equation; Two particles with a Coulomb interaction can be exactly separated in the (relativistic) center-of-momentum frame and the resulting ground-state energy has been obtained very accurately using finite element methods of J. Shertzer. The Dirac equation whose Hamiltonian comprises two Dirac particles and a static Coulomb potential is not relativistically invariant. But if one adds the 1/c2n (or α2n, where α is the fine-structure constant) terms, where n = 1,2…, then the result is relativistically invariant. Only the leading term is included. The α2 contribution is the Breit term; workers rarely go to α4 because at α3 one has the Lamb shift, which requires quantum electrodynamics.
Stjepan Mohorovičić predicted the existence of positronium in a 1934 article published in Astronomische Nachrichten, in which he called it the "electrum". Other sources credit Carl Anderson as having predicted its existence in 1932 while at Caltech. It was experimentally discovered by Martin Deutsch at MIT in 1951 and became known as positronium. Many subsequent experiments have precisely measured its properties and verified predictions of quantum electrodynamics. There was a discrepancy known as the ortho-positronium lifetime puzzle that persisted for some time, but was eventually resolved with further calculations and measurements. Measurements were in error because of the lifetime measurement of unthermalised positronium, which was only produced at a small rate. This had yielded lifetimes that were too long. Also calculations using relativistic quantum electrodynamics are difficult to perform, so they had been done to only the first order. Corrections that involved higher orders were then calculated in a non-relativistic quantum electrodynamics.
Molecular bonding was predicted for positronium. Molecules of positronium hydride (PsH) can be made. Positronium can also form a cyanide and can form bonds with halogens or lithium.
Czarnecki, Andrzej; Karshenboim, Savely G. (2000). "Decays of Positronium". In Levchenko, B. B.; Savrin, V. I. (eds.). Proceedings of the International Workshop on High Energy Physics and Quantum Field Theory (QFTHEP). 14. pp. 538–544. arXiv:hep-ph/9911410. Bibcode:1999hep.ph...11410C.
Cooke, D. A.; Crivelli, P.; Alnis, J.; Antognini, A.; Brown, B.; Friedreich, S.; Gabard, A.; Haensch, T. W.; Kirch, K.; Rubbia, A.; Vrankovic, V. (2015). "Observation of positronium annihilation in the 2S state: towards a new measurement of the 1S-2S transition frequency". Hyperfine Interact. 233 (1–3): 67–73. arXiv:1503.05755. Bibcode:2015HyInt.233...67C. doi:10.1007/s10751-015-1158-4.