One Positron in a Quantum Cyclotron: Antimatter's CPT Test

Somewhere inside a sealed silver cavity, cooled to a fraction of a degree above absolute zero, a single particle of antimatter hangs in empty space. Nothing touches it. The vacuum around it is so good that no stray atom will ever bump into it and destroy it, and physicists at Northwestern University report that one such particle has stayed there for months. Every so often they nudge it with microwaves, and it answers by climbing exactly one rung of an energy ladder, or by flipping its spin, without being disturbed by the act of looking.

The particle is a positron, the antimatter twin of the electron. In a paper published on October 21 in Physical Review Letters, Gerald Gabrielse and colleagues T. G. Myers, L. Soucy, B. A. D. Sukra and B. Sinha describe what they call a one-positron quantum cyclotron: a single positron whose lowest motion and spin states can be told apart one by one. It is a first for the positron. It is also not yet the measurement the team wants. It is the machine built to make that measurement, switched on and shown to work. This article explains what the machine is, why it is so hard to build, and what question it is meant to answer. We covered the announcement itself in a short news report; here is the longer story.

The twin that could break the rulebook

The Standard Model of particle physics, the mathematical description of every known elementary particle and three of the four forces, contains a quiet but rigid rule. Physicists describe it with three letters. C, charge conjugation, swaps every particle for its antiparticle. P, parity, flips space into its mirror image. T, time reversal, runs all motion backward. According to the Standard Model, the weak force breaks each of these symmetries separately, so a world with antiparticles, or a mirror-image world, or a backward-running world, would not behave exactly like ours. But apply all three at once, and the Standard Model says reality comes out unchanged. This is called CPT invariance, and in the theory it is unavoidable rather than optional.

One consequence is simple to state. A particle and its antiparticle must have identical masses and lifetimes, and magnetic properties that are identical in size and opposite in sign. If an electron and a positron ever turned out to differ in the strength of their tiny internal magnets, CPT invariance would fail, and the Standard Model would need to be rebuilt from its foundations. The Standard Model gives no reason to expect it. The point of the test is that a failure would be a discovery that could not be explained away.

The Standard Model is, in the words of the group's own review of the field, both a great triumph and a great frustration of modern physics. It makes astonishingly precise predictions that experiments confirm, yet it cannot explain how basic features of the universe, such as the dominance of matter over antimatter, could emerge from a big bang. Testing its deepest symmetries ever more sharply is one of the few ways to look for cracks.

Why the magnet inside a particle is the thing to compare

An electron is not just a point of charge. It also behaves like a minuscule bar magnet, and the strength of that magnet is called its magnetic moment. Physicists usually express it as a pure number, g/2, which says how many natural units of magnetism the particle carries. The natural unit is the Bohr magneton. For a particle with no inner structure and no interactions beyond the simplest, the number would be exactly one. In reality it is larger than one by only about one part in a thousand, and that tiny excess is where the physics lives. It comes from the electron constantly interacting with the vacuum around it, emitting and reabsorbing virtual particles.

Because the excess is so sensitive to everything the electron does, it is both the most precisely measured property of any elementary particle and the subject of the most precise prediction of the Standard Model. The best electron value, measured in 2023 by the same group, is g/2 = 1.001 159 652 180 59, with an uncertainty of 13 in the last two digits, a fractional precision of 0.13 parts per trillion. Theory and measurement agree to roughly one part in a trillion, limited at present by a disagreement between two independent measurements of the fine-structure constant, a number the theoretical prediction needs as input.

The positron is a different story. The best positron value dates from 1987, from experiments that tracked the particle's classical orbit, and it is less precise than the electron's by a factor of about 30. The group's review puts the two values side by side: the electron and positron figures agree within two standard deviations, but the positron's much larger uncertainty, about 4.3 parts per trillion, dominates the comparison. The difference between them is the most sensitive test of CPT invariance ever made with leptons, the family that includes the electron, and it is limited almost entirely by the positron half.

That is the gap this experiment aims to close. If a quantum-cyclotron measurement of the positron reached the precision already achieved for the electron, the paper notes, it would test CPT invariance in the lepton sector more than 30 times more precisely than is possible today.

What a quantum cyclotron is

A cyclotron, in the everyday sense, is an accelerator in which charged particles circle in a magnetic field. A single charged particle in a magnetic field does the same on a tiny scale: it loops around the field lines at a fixed rate, the cyclotron frequency, which depends only on its charge, its mass and the field strength. In this experiment the field is 5.6 tesla, produced by a superconducting solenoid, and the positron circles at about 156 billion times per second, a frequency in the microwave range.

At ordinary temperatures such motion looks smooth and can have any energy. But a particle this light, cooled enough, obeys quantum rules. Its circling energy can only take values on a ladder of equally spaced rungs, numbered 0, 1, 2 and so on, and the spacing between rungs equals Planck's constant times the cyclotron frequency. Add the particle's spin, which can point either along the field or against it, and there are two such ladders, one for each spin direction. The spin has its own transition frequency, and here is the key fact: for an electron or positron the spin frequency and the cyclotron frequency are almost the same, differing by only about one part in a thousand. That small difference, called the anomaly frequency, is exactly the excess of g/2 over one.

So the recipe for measuring the magnetic moment is to measure two frequencies on the same particle in the same field, the cyclotron frequency and the anomaly frequency, and take their ratio. The magnetic field cancels out of the ratio, provided it does not drift between the two measurements. The particle acts as its own magnetometer. The solenoid used here drifts by only about two parts per billion per day.

The word "quantum" in quantum cyclotron means something strict: the particle must sit on the lowest rungs, and the experimenters must be able to tell rung 0 from rung 1 for a single particle. That is hard, because the rungs of an ideal ladder are evenly spaced, so a microwave pulse that lifts the particle from rung 0 would also lift it from rung 1, and 2, and so on, without telling them apart. The way out comes from Einstein. A fast-moving particle is slightly heavier, and the cyclotron frequency shifts accordingly. In the energy levels this shows up as a tiny shift of about one part in a billion, enough to make the lowest rungs individually addressable. The paper treats this shift as small enough to ignore for the present demonstration, but the group's review explains that it is exactly what lets the bottom rungs behave like a pair of cleanly separable states.

Step one: catching a positron

Before any of this can be measured, a positron has to be caught, and positrons do not sit around waiting. The experiment gets them from a small piece of sodium-22, a radioactive isotope whose nuclei decay by emitting positrons. The emitted positrons are fast: up to 546 thousand electron-volts of kinetic energy. The trapped positron, by contrast, must end up with a thermal cyclotron energy of only 26 millionths of an electron-volt. That is a drop of roughly ten orders of magnitude, and the paper calls this the significant challenge of the whole setup.

The source is weak: 18 microcuries. The authors point out that this is a practical choice. Modern safety practice in a university laboratory would make a much stronger source extremely difficult to install, such as the 500-microcurie source used in the best earlier, classical-orbit measurement of the positron's magnetic moment. A weak source means few positrons, so the slowing and capture process has to be efficient.

The path from source to trap runs through several stages, all inside a sealed vacuum container kept near 300 millikelvin. First the positrons pass through a titanium window ten micrometres thick. They then enter a single crystal of tungsten, which acts as a moderator: collisions inside the crystal rob them of energy, and for the few that reach the exit surface of the tungsten, the metal's work function has the right sign to push them out gently at very low speed.

What happens next is the cleverest part. A slow positron leaving the tungsten along a vertical magnetic field line can pick up an electron from the tungsten surface. The two particles travel together along parallel field lines as a loosely bound pair, positronium, in an extremely stretched-out state that physicists call a Rydberg state. Because positronium is electrically neutral, it passes unaffected through the electric fields of the trap, going through a hole just 625 micrometres wide in a silver electrode. Only when it reaches the interior of the trap does the electric field grow strong enough to tear the pair apart. The negative electron is lost, and the positive positron, with nowhere to go, falls into the electrostatic well of the trap.

The hole must be large enough to let positronium through before it breaks up, and small enough not to ruin the cavity properties described below. This trade-off is one of the design challenges the authors highlight. The price of the simple, robust design is time: loading one positron takes 8 to 11 hours. That is slow, the paper acknowledges, compared with more elaborate positron-accumulation schemes. But once the trap potentials are tuned so that a single positron's motion is easy to detect, loading is rarely needed again. The authors report that a single positron has been trapped for months.

After loading, the source is raised on a thin nylon cable to a storage position away from the trap, so that no further particles can enter. The same apparatus can load an electron when the experiment needs one.

Step two: a trap that is also a bell

A positron cannot be held by a magnetic field alone, which only guides it along the field lines. To confine it along the field direction too, the team adds an electric field shaped like a saddle, produced by a stack of cylindrical electrodes. This arrangement is called a Penning trap, and it is the same kind of device that let physicists keep single charged particles for long periods and measure their properties with extreme precision. In the trap the positron oscillates back and forth along the field at about 206 million cycles per second, drifts in a slow circle at about 135 thousand cycles per second, and loops at its 156-gigahertz cyclotron rate. The slow circle, called magnetron motion, is cooled by a technique known as axial sideband cooling so that it does not matter for this demonstration.

The trap does a second job that is less obvious. A charged particle moving in a circle radiates energy, called synchrotron radiation, and a positron circling in a field of 5.6 tesla would lose its first quantum of energy in only 83 milliseconds, jumping from rung 1 down to rung 0. That is far too fast to see. To catch a state before it disappears, the signal from the single positron has to be averaged for longer than its lifetime.

The solution is to make the trap electrodes into a microwave cavity, a closed metal chamber with resonances, like a bell. In the group's earlier electron experiments the cavity measured roughly 14 millimetres across. The positron sits at the centre. A radiating particle can only emit light that the surroundings allow, and in a cavity the allowed frequencies are restricted to the cavity's resonances. If the cyclotron frequency falls between two resonances, the emission is suppressed. This is a well-established effect known as inhibited spontaneous emission. The paper says the rate can be cut by a factor of 200 or more, depending on how far the cyclotron frequency is from the cavity's modes. For this positron the first excited state lives about 4 seconds instead of 83 milliseconds.

That is the reason for the small hole. If it were larger, positrons would be easier to load, but the cavity's quality would fall and the inhibition would be spoiled. If it were smaller, the positronium would not get in. Getting both right is why the authors write that the positron's path into the trap must not compromise the cavity's ability to inhibit spontaneous emission.

Step three: a deep freeze

To keep the positron on the lowest rung, the surroundings must not hand it energy. The enemy is blackbody radiation, the faint glow every warm object emits. The gap between rung 0 and rung 1 corresponds to a temperature of about 7.5 kelvin, so a cavity at 4 kelvin would still pump the positron up now and then. At 300 millikelvin, the paper says, the chance that a blackbody photon has put the positron in its first excited state is only slightly larger than one part in ten to the eleventh power. The cold cavity also acts as a cryopump, freezing out leftover gas and producing a nearly perfect vacuum, good enough to keep a positron for as long as desired without it annihilating against a gas atom.

The cooling comes from a dilution refrigerator, a machine that uses a mixture of helium isotopes to reach temperatures far below that of liquid helium. In this design the refrigerator hangs down into the bore of the superconducting solenoid, which itself sits at 4.2 kelvin. The cold trap and its vacuum enclosure hang beneath the refrigerator, but, as the authors note, they rest mechanically on the 4.2-kelvin aluminium tube on which the solenoid is wound. That mechanical link matters: it keeps the trap from moving through small residual variations of the magnetic field, which is how the field stays stable to two parts per billion per day. According to the group's review, this arrangement was developed for their electron work and it is the same philosophy here.

Reading the state without touching it

The hardest requirement is how to find out which rung the positron is on without kicking it to another one. Ordinary measurements of a quantum system disturb it. What is needed is a quantum nondemolition measurement, in which repeating the observation keeps giving the same answer, because the measurement does not change the quantity being measured.

The trick uses the third motion, the back-and-forth bounce along the field. Around the trap the team places rings of nickel, which become magnetised in the strong field and create a magnetic "bottle": a field that varies slightly with position. In such a field the bounce frequency depends a little on the particle's cyclotron and spin quantum numbers. Each cyclotron rung, and each spin state, shifts the bounce frequency by a fixed small amount. In this experiment the shift per quantum is about 3.8 hertz out of 206 million, a change of roughly twenty parts per billion. Watching the bounce frequency therefore reveals the positron's cyclotron and spin state, while not changing either.

To observe the bounce, the team drives it with a weak oscillating voltage and monitors the electrical current that a moving charge induces in the electrodes. The electrodes are wired into a resonant circuit, and the signal is amplified by a cryogenic transistor amplifier, known as a HEMT, before being digitised. The paper describes a coaxial resonator bent into a circle around the electrodes, a geometry the authors say has not been used before for this kind of particle detection. The resonant circuit also damps the bounce, at a rate of about 2.7 hertz, and the damping sets the width of the signal.

That damping offers a neat check on how many positrons are trapped. For several particles bouncing together, the damping grows in proportion to their number. A second check comes from the size of the quantum jumps themselves. The authors point out that with two positrons the bounce shift per quantum would be half as large, 1.9 hertz rather than 3.8, because the shift scales inversely with the total mass being observed. The jumps they see are all 3.8 hertz. And because a jump of one spin among two particles would sometimes show half the shift, or none at all if one spin went up while the other went down, the clean, repeated 3.8-hertz steps confirm that only one positron is in the trap.

What the team actually saw

With the apparatus working, the group demonstrated the three things a quantum cyclotron has to do. First, the cyclotron motion starts in its ground state, since the cold surroundings leave it nowhere else to be. Second, it can be excited one quantum at a time with a microwave drive near 156 gigahertz, and the bounce frequency steps up by the expected amount. Third, after a while the excited state decays by emitting a single quantum of synchrotron radiation, and the frequency steps back down.

The paper shows several such traces. In some, a single excitation is followed by a single spontaneous decay. In a trace taken with a stronger drive, the cyclotron energy rises by two quanta in quick succession, then falls by two spontaneous emissions. The authors call the clean steps the realisation of a one-positron quantum cyclotron.

The spin was also controlled. Flipping a spin from down to up while both cyclotron and spin states are monitored takes two steps. First, a drive near the cyclotron frequency lifts the positron from rung 0 to rung 1 with spin down. Then a second drive at the anomaly frequency takes it to rung 0 with spin up. The reverse flip, from up to down, uses the anomaly drive to move the positron from spin up on rung 0 to spin down on rung 1, after which the excited cyclotron state decays spontaneously, leaving the positron on rung 0 with spin down. In both cases the net effect is a flipped spin, and the experiment tracked it jump by jump.

The spin states are far more stable than the cyclotron ones. The ground state with spin down cannot decay at all, and the one with spin up decays by a magnetic transition about nine orders of magnitude more slowly than the cyclotron excitation does, with a calculated lifetime of 2.1 years. The paper notes that the cavity probably suppresses even this. Both lowest states are therefore effectively stable, and no inhibited emission is needed to read the spin.

Three decades of electrons behind it

The positron result stands on a long chain of electron work. According to the group's review of lepton magnetic moments, the era of quantum measurements began in 1987, when Gabrielse's group at Harvard University set out to realise a one-electron quantum cyclotron. They first did so in 1999. The first quantum-cyclotron measurement of the electron's magnetic moment came in 2006, a second in 2008, and a third in 2023, after the group had moved to Northwestern. The review says the methods were developed one PhD thesis at a time, in a sequence of 14 theses over 35 years.

Before that, the best electron measurements used large classical orbits. A single trapped electron was first seen directly at the University of Washington in 1973, and the last classical-orbit measurement, in 1987, reached a fractional precision of 4.3 parts per trillion, an improvement by a factor of about 800 over earlier free-electron experiments. It stood for almost two decades. The quantum measurements are far more precise, and the review notes that they disagree with the classical result by about 1.7 standard deviations.

The electron value now serves as a test of quantum electrodynamics, the theory of light and charged matter, through its tenth order of calculation. In the review's account, Freeman Dyson, one of the architects of that theory, expressed profound surprise, in a congratulatory letter to Gabrielse quoted in Physics Today, that the framework could make such precise predictions. The current weak point is not the theory but the input: two independent determinations of the fine-structure constant disagree by more than five standard deviations, so the prediction cannot yet be pushed past about one part in a trillion.

The group is also developing a successor electron apparatus at Northwestern. The review describes ideas to measure the electron moment ten times more precisely: using special relativity itself as the coupling that reads out the state, instead of the magnetic bottle; adding a quantum-limited superconducting detector; and circumventing detector backaction. It says the new apparatus can handle both positrons and electrons. The positron demonstration reported now uses the older magnetic-bottle method, which the review identifies as a source of broadened lines and systematic error.

What it does not yet show

It is worth being clear about what the paper does not claim. It does not report a new value for the positron's magnetic moment. It reports that a single positron can be loaded, cooled to the cyclotron ground state, driven through quantum jumps and spin flips, and read out without being disturbed. The authors say the system "opens the possibility" of far more precise measurements. In their summary they write that, with further optimisation and the considerable time it takes to investigate uncertainties, it now seems feasible to measure the positron moment at the precision of the most recent electron measurement. They give no schedule.

The experiment also has known limits. The magnetic bottle that makes the readout possible also broadens the resonance lines, because the positron's frequency then depends slightly on where it sits in the trap. The cavity that makes the quantum states observable also pulls the cyclotron frequency by a small amount, a cavity shift that has to be calculated and corrected, and the correction adds uncertainty. In the electron work, the largest uncertainty has come from a mismatch between predicted and observed cyclotron resonance shapes, which the group attributes tentatively to tiny vibrations of the solenoid.

Nor is this the only way of testing CPT. The paper notes that baryons are tested with protons and antiprotons in classical cyclotron motion, most recently to one part in a billion, much less precisely than the electron-positron comparison could be. Mesons are tested much more precisely by comparing the decay patterns of long- and short-lived kaons. Each test probes a different corner of the theory, and a violation could appear in one and not the others, which is why physicists want all of them.

What comes next

The next step for the Northwestern group is to turn the demonstration into a measurement: find the cyclotron and anomaly frequencies of the positron by sweeping drive frequencies and counting quantum jumps, correct for the trap's small shifts, and compare the resulting g/2 with the electron's. If the two numbers match to the precision available, the result will be the most stringent check of CPT invariance in the lepton sector, and a further confirmation of the Standard Model, which has so far passed each such test. If they differ, the Standard Model would need to be rebuilt from its foundations.

The Standard Model predicts that the two numbers will match, and every test so far agrees. The reason to keep testing is to push the comparison to the sharpest edge available. A single positron, held for months in a sealed cavity at 300 millikelvin, is a new instrument for doing exactly that. The experiment shows that antimatter can be handled with the same quantum finesse as the electron; what that finesse will reveal is the question the next measurement is built to answer.

More of our coverage of this area is in the science section.

Sources

• T. G. Myers, L. Soucy, B. A. D. Sukra, B. Sinha, G. Gabrielse, "One-Positron Quantum Cyclotron", Physical Review Letters, October 21, 2026: doi.org/10.1103/gbkc-6nlj
• Preprint of the same paper (v3, August 2026): arXiv:2605.08147
• G. Gabrielse, "Measured Lepton Magnetic Moments" (review): arXiv:2507.11268
• X. Fan, T. G. Myers, B. A. D. Sukra, G. Gabrielse, "Measurement of the Electron Magnetic Moment", Physical Review Letters 130, 071801 (2023): arXiv:2209.13084