Text from the interactive explanation

Production and magnetic self-confinement of e−e+ plasma by an extremely intense laser pulse incident on a structured solid target

Alexander Samsonov, Alexander Pukhov

Could a single laser flash make a speck of pulsar-like matter-and-antimatter plasma on Earth — and hold on to it long enough to study?

laser pulse

electrons

positrons (antimatter)

gamma rays

magnetic field lines

solid target, cut open (block ≈ 40 µm long)

Illustrative re-enactment of the authors’ 3D simulation (watch their own renderings of it ↓); readouts traced from Fig. 4 (approximate). tfocus = the moment the pulse peak reaches the cone tip; 1 fs (femtosecond) = 10⁻¹⁵ s; 1 GG (gigagauss) = 10⁹ gauss.

In supercomputer simulations by Alexander Samsonov and Alexander Pukhov, one extremely intense laser pulse fired into a cone-shaped hole in a solid block creates almost 2 × 10¹³ matter–antimatter pairs — and wraps them in a magnetic field of its own making.

Scroll to take it apart, one step at a time.

The real run

The same shot, straight from the supercomputer

The animation above is our sketch. These two clips are the authors’ own 3D renderings of their main simulated shot, made from the simulation data and published alongside the paper. They share one clock and one camera, so you can line them up moment by moment.

Left: the gold surface marks where the solid target is (a surface of constant ion density). The violet-blue cloud is the positron density. Right: the target is drawn see-through so the magnetic field lines show. Red/blue lines mostly wrap around the axis; green/magenta lines mostly run along it. Where the lines are denser, the field is stronger.

Try this: drag the clock to about +260 fs, then to the end (about +625 fs, more than twice as long as our sketch runs). Does the cloud on the left leave the region the field lines on the right are wrapped around?

Simulation output (3D QED particle-in-cell, reference run), not an experiment. The time stamp in each frame is t − tfoc, measured from the moment the pulse peak reaches the cone tip. Videos by A. Samsonov and A. Pukhov, supplementary material to the paper (doi.org/10.60893/figshare.mre.c.7933274), CC BY-NC 4.0; resized and re-encoded, no sound.

1 · Why bother?

Matter and antimatter, mixed into a plasma

Every electron has an antimatter twin, the positron : same mass, opposite electric charge. Mix equal numbers of both, hot enough that they fly around freely, and you get a pair plasma — an electrically neutral gas of matter and antimatter.

Pair plasmas exist in some of the most extreme places in the Universe. But those places are far too distant to probe directly. A lab version would never match them in absolute size — what matters is matching the ratios between the plasma’s natural length scales. That would let theories of these objects be tested by experiment.

Earlier approaches — firing fast electrons from accelerators, or from lasers, through heavy metal — make plenty of positrons, but usually as a sparse beam rather than a true plasma. Only recently has a lab beam been dense enough to show plasma-like collective behaviour, and holding it together has not yet been achieved.

Pulsars and neutron stars Spinning, super-magnetised remnants of dead stars.

Black holes Matter swirling in before it disappears.

Quasar jets Beams shot out from giant black holes in distant galaxies.

Three hurdles for a lab pair plasma

Make fast “seed” particles to start the process.not yet cleared

Turn them into a pair cloud dense and big enough to act collectively — as a plasma, not a spray.not yet cleared

Hold it together long enough to watch it.not yet cleared

The paper’s scheme tackles all three with a single laser pulse. We’ll tick them off as we go.

2 · Light into matter

How a laser makes antimatter

Energy can become matter (E = mc²), but it needs the right conditions. The paper relies on a two-step chain:

A fast electron, shaken by intense laser light, flings out a gamma ray — a particle of very high-energy light.

If that gamma ray is energetic enough compared with the field around it, it splits into an electron and a positron .

Each new particle can shed more gamma rays, which make more pairs: a cascade .

Physicists track one number, χ (“chi”): how strong the field looks from the particle’s point of view, compared with the field at which empty space itself would start making pairs. Around χ ≈ 1 and above, the gamma rays can make pairs.

Predict first: which way should the electron move to make the brightest gamma rays?

Make a guess, then check it with the sliders.

Riding along with the light, its electric and magnetic pushes almost cancel. Head-on, they add up. So you need fast electrons and light coming back at them — exactly what a cone with a reflecting tip provides.

χ ≈ γ · (E / ES) · (1 − β cos α)

γ: the electron’s energy in units of its rest energy (0.511 MeV); β: its speed as a fraction of light speed.

E: the laser’s electric field, ≈ 3.5 × 10¹⁵ V/m at the paper’s peak strength (a₀ = 1000).

ES ≈ 1.32 × 10¹⁸ V/m: the Sauter–Schwinger field, where empty space itself would break down into pairs.

α: angle between the electron’s motion and the light’s direction. Riding along (α = 0), 1 − β cos α is almost zero; head-on (α = 180°) it is 2.

A simplified plane-wave estimate of the paper’s Eq. (2); real fields in the cone are more complex.

3 · Why drill a cone?

A hole with two jobs

The target isn’t flat. A cone-shaped hole, 30 µm deep and 30 µm wide at the mouth (a third of a hair’s width), is cut along the laser’s path.

Walls The sloping walls let the pulse graze the surface, ripping electrons out and accelerating them along it. hurdle ①

Tip The tip reflects the pulse straight back into those electrons — head-on, the recipe for pair-making gamma rays. hurdle ②

Predict first: which shape keeps the largest share of its antimatter?

Pick one to see what the simulations found.

“Core”: an ellipsoid around the cloud’s centre, two standard deviations wide in each direction — the authors’ measure of how much of the antimatter stays confined.

Numbers: simulated (Table I, same laser, only the cone angle changes; core share at 100 fs after focus). Animation: illustrative.

Fast seed particles — from the grazing wallsnot yet cleared

Dense pair cloud — from the head-on collision at the tipnot yet cleared

Hold it togethernot yet cleared

4 · A bottle that builds itself

The laser makes its own magnet

Nobody brings a magnet. The pulse creates the magnetic field itself, in two ways:

Current The electrons racing along the cone are an electric current. Any current wraps a magnetic field around itself, like the field around a wire. (red/blue rings)

Twist The laser is circularly polarised: its electric field rotates like a corkscrew, so the light carries spin. It sets electrons circling around the axis, and a circling current makes a field along the axis, like a coil. This is the inverse Faraday effect; the intense radiation helps. (green/magenta strands)

Together, the field lines twist into helices. In the simulation the field peaked at tens of gigagauss and settled at about 6 GG.

In preliminary runs with linearly polarised light (no twist) at the same amplitude, the authors got a similar pair plasma, with only the wrapping field.

Illustrative field shapes, coloured like the paper’s Fig. 2b: red/blue = wrapping around the axis (+/−), green/magenta = along the axis (+/−).

How strong is “6 gigagauss”? log scale: each step is 10× stronger

Everyday and astrophysical values are typical reference figures; the plasma values are from the simulation. 1 GG = 10⁹ gauss; 1 tesla = 10⁴ gauss.

5 · Catch them if you can

Now you hold the bottle

You’re looking straight down the cone’s axis at a fresh burst of positrons and electrons , born near the centre and flying out at almost light speed. The gold ring is a wall 5 µm from the axis.

A magnetic field bends each particle into a circle — positrons one way, electrons the other. The hotter the particle, the wider its circle. Circles too wide for the room hit the wall, and those particles are lost.

Radiation cooling: charges circling at nearly light speed shed energy as light (synchrotron radiation, the yellow flickers). In the simulation the pairs cooled from a temperature of about 59 MeV to about 3 MeV in roughly 120 fs. Cooler particles make tighter circles. (MeV, a million electronvolts, is an energy unit; an electron’s own rest energy is 0.511 MeV.)

At the paper’s end-of-run values — about 6 GG and 3 MeV — the circles are only about 50 nanometres in radius: hundreds of times smaller than the cloud.

positronelectron⊙ field points out of the screeninset: magnifier ×30

Your goals

Keep more than half for 200 fsnot done

Find the weakest field that still works with cooling onnot done

Turn cooling off at 1 GG — what changes?not done

Your bursts (kept after 200 fs)

Toy model: uniform field in one cross-section; real fields vary in space and time and particles also drift slowly along the axis. Circle size from the paper’s Eq. (7); cooling follows the temperatures fitted in Fig. 5; energies are spread around 3× the temperature.

Now compare with real simulated positrons

Paths of a small sample of positrons from the full simulation, 140–290 fs after focus. Left: the same end-on view as your game, with each dot trailing its last 10 fs. Top right: how far each has gone round the axis (in full turns). Bottom right: its position along the axis. The dashed line is “now”.

What to notice: each positron makes small, tight loops around the local magnetic field (the circles from your game) while, on average, following the field lines that wrap round the axis, so the group slowly circles it, roughly 1–2 µm out. Over these 140 fs most shift by only a few micrometres along the axis. The authors read these paths as further evidence that the plasma is strongly magnetized.

6 · Is it a real plasma?

Big enough to act together, tight enough to stay put

A cloud of charged particles only behaves as a plasma — moving and waving as a whole — if it is much bigger than two natural lengths:

Skin depth : how far electric and magnetic disturbances reach into the plasma before its particles screen them out.

Debye length : the distance over which a single charge is hidden by the crowd around it.

And it is strongly magnetised if its particles’ circles (the Larmor radius ) are tiny compared with the cloud. Zoom in to see all three side by side.

Collective: cloud ≈ 20 × 10 × 10 µm vs skin depth ≈ 50 nm — about 400 times bigger. The Debye length is smaller still.

Strongly magnetised: Larmor radius ≈ 50 nm at 6 GG, again hundreds of times smaller than the cloud.

Field in charge: β ≈ 0.68 — the particles’ pressure is only about two-thirds of the magnetic pressure.

End-of-run estimates from the simulation (≈ 200 fs after focus; density ≈ 10²³ per cm³, temperature ≈ 3 MeV).

The whole positron cloud: about 20 µm long and 10 µm wide, sitting inside the cone.

Dots and motion are illustrative; the lengths are the paper’s estimates.

Fast seed particlesnot yet cleared

Dense, collective pair cloudnot yet cleared

Held together by its own field for hundreds of femtosecondsnot yet cleared

7 · What the simulations showed

Inside the virtual experiment

Everything here comes from computer simulations — no experiment has been done yet. The authors ran full-scale 3D particle-in-cell simulations, which follow huge numbers of sample particles and the electromagnetic fields between them, with quantum effects (QED) for gamma-ray emission and pair creation. They used the VLPL code on the JUWELS supercomputer.

Reference run (Table I, entry 12). 1 PW (petawatt) = 10¹⁵ watts.

Where does the laser’s energy go?

Traced from Fig. 4a, approximate. Shares of the laser’s initial energy; “laser light” = electromagnetic energy still inside the simulation box; “other” = light that has left the box plus tracing error.

8 · How big a laser?

An enormous laser — for now

The scheme needs a pulse far beyond today’s lasers; the most powerful ones reach around 10 petawatts. Across runs with the same cone, the results grew steeply with pulse power P: the trapping score as P3.4, the number of pairs as P2.2, the density as P1.6. Halve the power and the trapping score drops about tenfold.

The smallest pulse that still trapped a significant share in these runs was 185 PW (3 kJ in 16.2 fs). The authors write that even this is probably beyond the reach of the most ambitious planned laser facilities, so the next step is to adapt the scheme to more moderate lasers. Preliminary runs with linearly polarised light at the same amplitude — half the power — gave a pair plasma with similar parameters.

Trapping score ν/S₀: the share kept in the core (ν) divided by S₀, a number combining target density and laser strength, used to compare runs with different targets.

trapping score ν/S₀ ∝ P3.4

number of pairs ∝ P2.2

peak density ∝ P1.6

individual simulation runs

Fitted power laws from Fig. 7 (runs with tan θ = 0.5); valid only across the simulated 82–431 PW range.

9 · Back to the stars

A speck of pulsar-like plasma, made on Earth

A pair plasma that outlives the flash that made it could be watched evolving on its own — on its natural timescale, where it behaves like a magnetised fluid. That would open lab tests of processes thought to shape the most extreme places in the Universe: waves, instabilities, collisionless shocks and magnetic reconnection, in a strongly magnetised plasma like those around pulsars and black holes.

For now it is a design on a supercomputer, waiting for a laser big enough — or a cleverer version that needs less.

One pulse, three hurdles — in simulation

Seed particles: electrons accelerated along the grazing cone walls.not yet cleared

Dense, collective pair cloud: a cascade where the reflected pulse meets the electrons — 1.9 × 10¹³ positrons in a cloud hundreds of times wider than its skin depth.not yet cleared

Held together: a self-made field of about 6 GG keeps a quarter of the positrons in a small core for hundreds of femtoseconds.not yet cleared

Source

Alexander Samsonov and Alexander Pukhov, “Production and magnetic self-confinement of e−e+ plasma by an extremely intense laser pulse incident on a structured solid target,” Matter and Radiation at Extremes 10 , 057202 (2025). doi.org/10.1063/5.0260941 · article page

What is simulated, what is illustrated

From the paper (simulation results): all quoted numbers — pair counts, field strengths, temperatures, trapped shares, energy shares, power scalings, length scales — the four figures and the authors’ three supplementary videos.

Traced by eye from figures (approximate): the time curves behind the hero readouts and the energy-flow bar (Fig. 4).

Illustrative: the 3D cone animation, collider, cone cross-section, field shapes, the trapping game (a toy model using the paper’s Larmor-radius formula and cooling fits) and the zoom view.

No experiment has been done; the scheme is a proposal tested in simulation.

Licence and credits

Article and figures © 2025 the authors, published under a Creative Commons Attribution (CC BY 4.0) licence; figures cropped and recaptioned here. Supplementary Videos 1–3 © the authors, published with the article under a Creative Commons Attribution-NonCommercial (CC BY-NC 4.0) licence (doi.org/10.60893/figshare.mre.c.7933274); resized and re-encoded here, otherwise unchanged. Reference values for everyday and astrophysical magnetic fields and today’s laser powers are general background, not from the paper.

Typefaces: Instrument Serif, IBM Plex Sans and IBM Plex Mono, under the SIL Open Font License.