Experiment: The Positron
On 2 August 1932 a cosmic-ray track crossed Carl Anderson's cloud
chamber at Caltech, curved the wrong way for an electron, and passed
through a \(6\,\mathrm{mm}\) lead plate emerging with less
momentum than it entered — which fixed its direction of travel, and
so its sign. Anderson announced the particle in Science within
weeks [Anderson:1932] and published the full analysis the
following March under the title “The Positive Electron”
[Anderson:1933]. It was the first antiparticle ever seen,
and it had been predicted: Dirac's equation
(The Dirac Equation) required a particle of the electron's
mass and opposite charge [Dirac:1930a] [Dirac:1931], and no other
theory then in existence asked for one. This chapter records the
measurement in full phenomenological detail — apparatus, procedure,
numbers with their uncertainties in SI units, and the inferences the
numbers do and do not support — using the structured
experiment environment of this treatise when the content is
written.
Two things make the case worth the space. First, the inference chain is short and checkable: a track's curvature in a known field gives \(B\rho\) and hence momentum, its ionization density gives a bound on the charge and mass, and the lead plate gives the direction; nothing else is needed to establish that the particle is positive, light, and not a proton. Second, the confirmation was immediate and independent — Blackett and Occhialini, using a counter-triggered chamber, published photographs of electron–positron showers within months [Blackett:1933], identifying pair production as the mechanism [Oppenheimer:1933] [Bethe:1934] — and the sequel has run for ninety years, through annihilation radiation [Klemperer:1934], positronium [Deutsch:1951], the antiproton [Chamberlain:1955] and trapped antihydrogen [Ahmadi:2017], to a hospital imaging technique that works by annihilating antimatter [TerPogossian:1975].
Experiment: The Positron: all derivations of this chapter are pending.
Historical context and the prediction under test
Dirac's hole theory as a falsifiable prediction
[Reserved: the negative-energy solutions of the Dirac equation, Dirac's filled sea and his first, wrong identification of a hole with the proton [Dirac:1930a]; the retraction and the 1931 prediction of an “anti-electron” of equal mass and opposite charge [Dirac:1931]; what exactly was predicted — mass equal to \(m_{e}\), charge \(+e\), and production in pairs — and therefore what the experiment had to measure to confirm or refute it; the state of the theory as of The Dirac Equation, and the honest note that Anderson was not looking for Dirac's particle and did not cite him.]
Cosmic rays as the available source
[Reserved: the penetrating radiation established by Hess's balloon ascents [Hess:1912], the only source of particles above \(100\,\mathrm{MeV}\) available in 1932; the Caltech programme of measuring cosmic-ray energies in a magnetic cloud chamber; why the question “are the primaries positive or negative?” motivated the apparatus, and how that question made the positron an unintended discovery; the cosmic-ray physics itself is Cosmic Rays and Astroparticle Physics.]
What else the track could have been
[Reserved: the exhaustive list of alternatives Anderson had to exclude [Anderson:1933] — a negative electron travelling upwards, excluded by the lead plate; a proton of the observed curvature, excluded by ionization density and by range, since a proton of \(0.21\,\mathrm{T}\,\mathrm{m}\) would have stopped within about \(5\,\mathrm{mm}\) while the track ran over \(5\,\mathrm{cm}\); a light nucleus; a chance coincidence of two unrelated tracks. The subsection should state each alternative with the observable that kills it, because this is the anatomy of an experimental claim.]
Apparatus
The cloud chamber
[Reserved: Wilson's expansion chamber, in which a sudden adiabatic expansion supersaturates a vapour and ions left along a charged particle's path seed visible droplets [Wilson:1912]; Anderson's vertical chamber of about \(17\,\mathrm{cm}\) diameter and \(3\,\mathrm{cm}\) depth, its gas filling and expansion ratio, and the stereoscopic camera pair [Anderson:1933]; the physical mechanism — ionization energy loss — linking droplet density to charge and velocity, and its treatment in Scattering Theory.]
The magnet
[Reserved: the water-cooled electromagnet producing a field of about \(1.5\,\mathrm{T}\) across the chamber, then among the strongest available; its calibration and uniformity, which set the systematic floor on every momentum quoted; the relation \(p=qB\rho\) from Magnetostatics and its relativistic validity (Relativistic Dynamics), so that a measured sagitta gives a momentum without any assumption about the particle's mass.]
The lead plate
[Reserved: the \(6\,\mathrm{mm}\) lead plate mounted across the middle of the chamber [Anderson:1933] — the single component that turns an ambiguous photograph into a measurement, since a particle loses energy in crossing it and therefore curves more tightly on the far side, which fixes the direction of travel and hence the sign of the charge; the design as the model for every later use of a degrader to break the sign–direction degeneracy.]
Procedure
Expansion, triggering and photography
[Reserved: random expansion at fixed intervals in Anderson's setup, with roughly one usable cosmic-ray track in fifty photographs [Anderson:1933]; the counter-controlled trigger introduced by Blackett and Occhialini, in which coincident Geiger counters above and below fire the expansion so that most photographs contain a track [Blackett:1933] — an efficiency gain of well over an order of magnitude and the reason their shower photographs came so quickly; scanning and selection criteria, and the selection bias each introduces.]
Track reconstruction and momentum from curvature
[Reserved: reprojection of the stereoscopic pair, fitting a circle to the droplet trail, and the propagation of the sagitta uncertainty into \(B\rho\) and thence into momentum, following the error-analysis conventions of Measurement, SI Units, and the Theory of Errors; the dominant systematics — multiple scattering in the chamber gas, turbulence and droplet diffusion, distortion of the optics — and why they matter more for high momenta, where the track is nearly straight.]
Observations
The quantities below are what the chapter will tabulate, each with its uncertainty, in SI units.
The photograph of 2 August 1932
[Reserved: the single track that carries the claim [Anderson:1933]: \(B\rho=0.21\,\mathrm{T}\,\mathrm{m}\) above the plate and \(0.075\,\mathrm{T}\,\mathrm{m}\) below, corresponding to momenta of \(63\,\mathrm{MeV}/c\) and \(23\,\mathrm{MeV}/c\); curvature sense opposite to that of an electron in the same field; visible track length over \(5\,\mathrm{cm}\) beyond the plate; droplet density indistinguishable from that of a fast electron. The figure itself is to be re-authored as a labelled TikZ diagram of the track geometry, not reproduced as an image.]
Cosmic radiation contains singly charged particles whose tracks curve in a magnetic field in the sense opposite to an electron's, whose droplet density along the path is that of a fast electron rather than that of a proton of the same curvature, and whose range exceeds by an order of magnitude the few millimetres such a proton could traverse. Their charge is therefore \(+e\) and their mass is the electron's, not the proton's. The particle is stable in vacuum, and its mass and the magnitude of its charge equal the electron's to within the precision of present comparisons [Navas:2024].
Derivation. One relation turns the photograph into a measurement, and it is exact relativistically. A particle of charge \(q\) in a uniform field \(\vect{B}\) obeys \(\dd\vect{p}/\dd t=q\,\vect{v}\times\vect{B}\). The force is perpendicular to \(\vect{v}\), so it does no work; the speed is constant, hence so is \(\gamma\), hence so is the magnitude of \(\vect{p}=\gamma m\vect{v}\). Motion in the plane normal to \(\vect{B}\) is then uniform circular motion of some radius \(\rho\), for which \(\abs{\dd\vect{p}/\dd t}=\abs{\vect{p}}v/\rho\), while the right-hand side has magnitude \(\abs{q}vB\). Equating and cancelling \(v\),
Equation (95.1) contains no reference to the mass, so a measured sagitta in a calibrated field yields a momentum whatever the particle turns out to be. Three further inferences follow, and between them they exhaust the alternatives.
First, the sense of the curvature gives the sign of \(q\) once the direction of travel is known, and the lead plate supplies that: a particle can only lose energy in crossing it, so by Equation (95.1) it must emerge with the smaller \(B\rho\), and the side of the plate carrying the tighter curve is the far side. Second, the mass enters the other observables: at fixed momentum the velocity is \(v/c=\abs{\vect{p}}c/\sqrt{\abs{\vect{p}}^{2}c^{2}+m^{2}c^{4}}\), and ionization per unit path rises steeply as \(v\) falls. A proton carrying the measured momentum would be slow, would ionize far more densely than the track shows, and would come to rest within a few millimetres of chamber gas; a particle of electronic mass at the same momentum is ultrarelativistic, near minimum ionizing, and crosses the chamber. Third, an ordinary electron travelling upward would reproduce the curvature but not the energy loss across the plate, and is excluded by the same argument that fixed the direction. Charge \(+e\) and mass \(m_{e}\) are what remains.
∎The tabulated quantities
[Reserved: a booktabs table, set with captionof inside
the experiment box, listing for each of the fifteen positive tracks
reported in [Anderson:1933]: momentum above and below the plate in
\(\mathrm{MeV}/c\), inferred energy loss in the lead, specific
ionization relative to a minimum-ionizing electron, and the resulting
upper bound on the mass, quoted by Anderson as under twenty electron
masses and tightened by later work to consistency with \(m_{e}\)
[Navas:2024]; the modern reference values of \(m_{e}\) and \(e\) from
CODATA [Mohr:2025].]
Uncertainties and systematics
[Reserved: the uncertainty budget — field calibration, sagitta measurement, energy loss in lead computed from the stopping-power formula, and the statistical weight of a single event; the reason the claim survived on one photograph, namely that the alternatives of Section 95.1.3 are excluded by qualitative observables that do not depend on the momentum precision; the contrast with modern discovery thresholds (Experiment: The Higgs Boson Discovery), which this chapter should draw explicitly rather than leave implicit.]
Interpretation
Identification with Dirac's anti-electron
[Reserved: the inference from charge sign, mass bound and ionization to a positive particle of electronic mass [Anderson:1932] [Anderson:1933]; the identification with the hole of The Dirac Equation [Dirac:1931], made by Blackett and Occhialini rather than by Anderson [Blackett:1933]; the methodological point that a confirmed prediction and an unmotivated discovery here coincide, and that the treatise records who claimed what.]
Independent confirmation and pair production
[Reserved: the counter-triggered showers of Blackett and Occhialini, showing positive and negative tracks in equal numbers and emerging from a common origin [Blackett:1933]; the interpretation as materialization of a photon into an electron–positron pair above the threshold \(2m_{e}c^{2}=1.022\,\mathrm{MeV}\) [Oppenheimer:1933], with the cross section computed by Bethe and Heitler [Bethe:1934]; positrons from artificial radioactivity observed independently by Curie and Joliot [Curie:1934], which made the particle available in the laboratory without cosmic rays and connects to the \(\beta^{+}\) decays of Nuclear Forces and Nuclear Structure.]
A photon of sufficient energy passing through matter disappears, and an electron and a positron appear together at the same point; in showers the positive and negative tracks emerge from a common vertex in equal numbers. The process has a threshold — no pair appears below a photon energy of \(2m_{e}c^{2}=1.022\,\mathrm{MeV}\) [Mohr:2025] — and it requires the presence of a nucleus: a photon travelling in empty space never converts, however energetic.
Derivation. Both statements are kinematic and need no cross section. A single photon has \(p_{\gamma}\cdot p_{\gamma}=0\), whereas any electron–positron pair has \(\left(p_{+}+p_{-}\right)\cdot\left(p_{+}+p_{-}\right) \geq\left(2m_{e}c\right)^{2}\), the minimum being reached when the two are at relative rest. The total four-momentum is conserved and its square is invariant, so \(\gamma\to e^{+}e^{-}\) in vacuum would demand \(0\geq\left(2m_{e}c\right)^{2}\), which is false: a third body must be present to absorb momentum. Take that body to be a nucleus of mass \(M\) initially at rest. The invariant squared energy is
and the lightest final state allowed by charge and lepton-number conservation is the nucleus together with the pair, all at relative rest, so that at threshold \(\sqrt{s}=\left(M+2m_{e}\right)c^{2}\). Expanding and cancelling \(M^{2}c^{4}\),
For any nucleus \(m_{e}/M\) is below \(10^{-4}\), so the threshold is \(2m_{e}c^{2}\) to far better than a cloud chamber can measure; the correction is the recoil the nucleus must take, and it is the same term that vanishes only in the unphysical limit \(M\to\infty\). Finally, the equality of the numbers of positive and negative tracks is charge conservation applied to a neutral initial state.
∎Annihilation radiation
[Reserved: Klemperer's measurement of the radiation emitted when positrons stop in matter [Klemperer:1934] — two photons of \(511\,\mathrm{keV}\) emitted back to back, as energy and momentum conservation require for a pair annihilating at rest; the process as the exact inverse of pair production and as the first direct evidence that matter and antimatter destroy each other; the astrophysical counterpart, the \(511\,\mathrm{keV}\) line from the Galactic centre mapped by INTEGRAL [Knoedlseder:2005]; the cross section belongs to Quantum Electrodynamics and Renormalization.]
Positrons brought to rest in matter vanish, and in their place two photons appear, each of energy \(m_{e}c^{2}=511.0\,\mathrm{keV}\) [Mohr:2025] and travelling in opposite directions. Matter and antimatter are not merely oppositely charged: brought together they are destroyed entirely, their whole rest energy converted into radiation. No single photon is ever observed from such an event.
Derivation. Take the pair at rest, so that the initial four-momentum is \(\left(2m_{e}c,\vect{0}\right)\). If two photons are produced, conservation of momentum makes their three-momenta sum to zero, so \(\vect{k}_{2}=-\vect{k}_{1}\); both being massless their energies are then equal, and conservation of energy gives \(E_{1}=E_{2}=m_{e}c^{2}\), the two emerging back to back. Exactly one photon is impossible in any frame: a single photon has \(p\cdot p=0\) while the initial state has \(p\cdot p=\left(2m_{e}c\right)^{2}\neq0\), and no change of frame alters an invariant — the same obstruction, read backwards, that forbids a free photon from converting in Phenomenon 95.2. Three or more photons are kinematically allowed and do occur, at a rate suppressed by a further power of the fine-structure constant; which channel is permitted at all is decided not by kinematics but by the charge-conjugation quantum number of the annihilating state, a selection rule belonging to Discrete Symmetries and CPT.
∎The muon: the same apparatus, a different particle
[Reserved: the honest sequel, in which Anderson's own chamber produced tracks too penetrating for electrons and too light for protons, announced as a particle of intermediate mass [Neddermeyer:1937] and long mistaken for Yukawa's nuclear-force meson; the muon's actual identity as a heavy lepton (Cosmic Rays and Astroparticle Physics); included here because it shows the same apparatus and the same style of argument yielding one correct identification and one wrong one, and because it disciplines the reading of Section 95.5.1.]
The chambers that showed the positron also showed singly charged particles that penetrate far more matter than an electron of the same momentum and ionize far less than a proton of the same momentum: particles of intermediate mass. They are the muons — charge \(\pm e\), spin \(\tfrac{1}{2}\), mass \(105.658\,\mathrm{MeV}/c^{2}\), some 207 times the electron's, and mean life about \(2.2\,\mu\mathrm{s}\) [Navas:2024]. They are not the quantum of the nuclear force, which had been predicted with roughly that mass; they behave in every respect as a heavy copy of the electron, and nothing in this part of the treatise explains why a second copy should exist.
The identification of the penetrating cosmic-ray component as a heavy lepton rather than as the quantum of the nuclear force: its anomalously weak absorption in nuclei against the strong coupling a nuclear-force carrier must have, and the three-body shape of its decay electron spectrum. The muon's own phenomenology is taken up in the cosmic-ray chapter of Part XII.
Modern repetitions and precision
Positronium
[Reserved: Deutsch's observation of the bound electron–positron system in gases [Deutsch:1951]; the two ground states — the singlet decaying to two photons in about \(125\,\mathrm{ps}\) and the triplet to three photons in about \(142\,\mathrm{ns}\) [Navas:2024] — as a direct test of the annihilation selection rules and of charge-conjugation symmetry (Discrete Symmetries and CPT); positronium as a purely leptonic “atom” whose spectrum tests the bound-state theory of The Hydrogen Atom with no nuclear structure at all.]
The antiproton and antihydrogen
[Reserved: the antiproton produced at the Bevatron and identified by simultaneous momentum and velocity measurement [Chamberlain:1955], closing the argument that antimatter is general rather than a peculiarity of the electron; antihydrogen trapped at CERN and its \(1S\)–\(2S\) transition measured and found consistent with hydrogen [Ahmadi:2017], the sharpest laboratory test of \(CPT\) available (Discrete Symmetries and CPT); the direct observation that antihydrogen falls downwards [Anderson:2023], a test of the equivalence principle of The Equivalence Principle and Classical Tests on antimatter.]
Precision comparisons of positron and electron
[Reserved: the Penning-trap comparison of the \(g\) factors of the electron and the positron, agreeing to about two parts in \(10^{12}\) [VanDyck:1987], which is simultaneously the tightest test of \(CPT\) for leptons and a measurement of the anomaly discussed in Experiment: The Electron Anomalous Magnetic Moment; the charge-to-mass ratio comparisons; how each such null result is quoted as a bound rather than as a confirmation, per the conventions of Measurement, SI Units, and the Theory of Errors.]
Every comparison so far made between a particle and its antiparticle has found their masses equal, their lifetimes equal, their magnetic moments equal in magnitude and opposite in sign, and their charges equal in magnitude and opposite in sign. For the electron and the positron the agreement of the \(g\) factors and of the charge-to-mass ratios is among the sharpest such tests available, and no significant difference has been reported in any of them [Navas:2024].
The CPT theorem — that any local, Lorentz-invariant quantum field theory with a Hermitian Hamiltonian and the normal spin–statistics connection is invariant under the combined operation of charge conjugation, parity and time reversal — together with the equalities of mass, lifetime and magnetic moment that follow from it, and the reasons a null result here bounds rather than confirms.
Applied annihilation: positron emission tomography
[Reserved: the imaging technique built directly on Section 95.5.3 — a \(\beta^{+}\) emitter administered to a patient, the two \(511\,\mathrm{keV}\) photons detected in coincidence, and the line of response reconstructed tomographically, as realized in the first transaxial positron tomographs [TerPogossian:1975] [Phelps:1975]; the back-to-back emission geometry, whose small departure from \(\pi\) from residual momentum sets a floor on spatial resolution; included because it is the plainest available demonstration that a 1932 cosmic-ray track is now routine engineering.]