Graded by the challenger

★★★★☆ 4 / 5

It's mostly right, but dismissed the core concept of using particles with alternating charges to reduce the net repulsion force, because it's unreasonably difficult to engineer that mechanism. Storing only positrons doesn't double the density in a useful way, it just makes it unreasonably difficult in a different direction. Either way, it's completely science fiction for now.

AskSmith.ai Hall of Reveals

A reveal from the AskSmith.ai Hall

Bottled Positrons

Challenger: Joshua Category: Research Difficulty: Tricky

The challenge was confirm the multicell trap is 1000 times denser. The trick: the gain comes from the cells, not the electrons.

What the AI did: worked the Brillouin limit, space charge, relativistic drift and cell geometry in Python (numpy), checked the result against the real positron trap literature (Surko, ATRAP, ALPHA, the multicell prototypes), and built a cryogenic heat budget for a 1 cubic meter tank inside a 10 meter equipment envelope. The calculator at the end recomputes payload, energy and idle heat live.

Hard science fiction support, not an engineering design. These are order of magnitude estimates. Physics limits are firm; engineering terms are labelled assumptions you can change.

The brief, verbatim

“Peer review hard-scifi methods for storing positrons as fuel for a reactor. A penning trap is the default mechanism, but even hypothetically storing nanograms in a reasonable volume is pushing physics. A multi cell trap alternating electrons and positrons can hypothetically be denser, when not accounting for equipment. List any unique problems which only one of those machines would have. Confirm the hypothetical payload of a 1 meter cubed volume, with up to ten meters of external equipment. Estimate the waste heat for idle operation. Note: I am not trying to be a smartass, this is an actual thing.”

What he expects

“Both machines have extreme problems, but AFAIK neither has a unique problem. It should report about 1000x the density of a penning trap, if everything was absolutely perfect.”

Verdict

One expectation fails, one half holds

Disagree

“Neither has a unique problem”

Expectation 1

They share the big ones (vacuum, cryogenics, the magnet, gamma shielding). But each has problems the other cannot have. A single trap's space charge grows with the square of its whole radius: holding Brillouin density in a half meter column takes about 1.4 × 1011 volts and stores 34,000 times more electrostatic energy than the fuel itself, and the plasma edge would have to move faster than light. The alternating array has positronium: anywhere an electron meets a positron, fuel turns into gamma rays in under a microsecond. Opposite polarity end caps side by side, counter rotating neighbors and hundreds of millions of microcells are also its alone. details

Partly right

“About 1000x the density of a Penning trap, if perfect”

Expectation 2

1000x is not a constant of nature; it is a geometry ratio, and it depends on what you compare against.

  • Against the Brillouin limit (the usual “perfect Penning trap” ceiling): 0.45x. No gain. Every cell still obeys the same limit, and half the cells hold electrons.
  • Against one perfect 1 m³ trap limited only by physics (drift below light speed): about 33 × B in tesla. That is 170x at 5 T, 330x at 10 T, and 1000x only near 30 T, a field no 1 meter bore magnet makes.
  • Against a real voltage limited trap: 1000x arrives with cells about 1 cm across, which is exactly the regime of Surko's multicell design. Microcells give millions.

And the electrons add nothing to density. Positron only cells separated by grounded walls (the published multicell concept) hold twice as much. details

Physics ceiling, 1 m³, 10 T443 ngsingle species Brillouin, 80 MJ
Ideal alternating array, 10 T166 ng30 MJ, about 8 kWh
Buildable style (1 kV, 1 cm cells)0.75 pg136 J of fuel
Idle waste heat, wall plug3 to 235 kWsee the heat budget

The identity that decides it. At the Brillouin limit, the annihilation energy of the stored positrons is nB · 2mc² = B²/μ0, exactly twice the energy of the magnetic field holding them. A perfect 10 T tank stores 80 MJ per cubic meter, about 400 times less than gasoline by volume, and the Iseult class magnet that could hold it stores 338 MJ on its own. The magnet holding the fuel is a bigger battery than the fuel.

On the payload: his instinct is right for real machines (femtograms to picograms), and slightly pessimistic for pure physics. The Brillouin ceiling allows about 110 ng at 5 T and about 600 ng at 11.7 T in a cubic meter. Idle operation would spend the whole stored energy in two to thirty minutes even in the ideal case.

1 · The hard ceiling

Payload of one cubic meter

A cloud of one charge sign can only be held by a magnetic field up to the Brillouin density, where the magnetic force just balances the cloud's own repulsion. S1 S2 No trap design gets around it for a single species plasma, because it comes from radial force balance alone.

nB = ε0B² / (2m) = 4.86 × 1018 m−3 × (B / 1 T)² · energy per positron = 2mc² = 1.022 MeV

Single species Brillouin limit filling 1 m³ (each positron annihilates with one ordinary electron)
FieldPositronsMassEnergykWhTNTRotationA 1 m bore magnet?
1 T4.9 × 10184.4 ng0.80 MJ0.220.19 kg14 GHzYes, routine
5 T1.2 × 1020111 ng19.9 MJ5.54.8 kg70 GHzYes
10 T4.9 × 1020443 ng79.6 MJ2219 kg140 GHzYes, at the frontier
11.7 T6.7 × 1020606 ng109 MJ3026 kg164 GHzIseult: 90 cm warm bore, 5 m long, 5 m across, 132 t, 338 MJ stored S14
20 T1.9 × 10211.8 µg318 MJ8876 kg280 GHzOnly as a fusion test coil (HTS, 2021) S17
45 T9.8 × 10219.0 µg1.6 GJ448385 kg630 GHzNo: the record 45 T hybrid has a 32 mm bore and draws 30 MW S15
100 T4.9 × 102244 µg8.0 GJ2,2101.9 t1.4 THzNo: 15 mm bore, 15 ms pulses S16

Reading the ten meter envelope. We assume “up to ten meters of external equipment” means the magnet, cryostat, cryocoolers, vacuum and power electronics fit inside a box about 10 m across around the 1 m³ tank. That admits an Iseult class whole body MRI magnet (about 11.7 T in a 90 cm bore, 5 m by 5 m), so roughly 10 to 12 T is the honest field. The 45 T and 100 T rows are physics curiosities: no magnet makes them over a meter.

Rotation column: a plasma at the Brillouin limit spins at half the cyclotron frequency. Today's rotating wall drives work at megahertz on plasmas a million times thinner S7; nobody has driven a plasma at 70 to 160 GHz.

2 · Machine one

Why one big Penning trap stalls

A real Penning trap for many particles is a Penning-Malmberg trap: a long stack of ring electrodes in a solenoid. The magnet handles radial confinement, end electrodes handle axial confinement, and the end voltage must exceed the plasma's own space charge potential. S3

φcenter = e n R² / (4ε0) · Ewall = e n R / (2ε0) · Uelectrostatic = λ²L / (16πε0)

At the Brillouin limit, 5 T, R = 0.5 m

  • Space charge potential: 1.4 × 1011 V, with 5.5 × 1011 V/m at the wall.
  • Net charge: 19.5 coulombs in one column.
  • Electrostatic energy: 6.7 × 1011 J, about 34,000 times the 2.0 × 107 J the positrons would release. At 10 T the ratio is 134,000.
  • The E × B drift at the edge would exceed the speed of light. Keeping drift below c caps density at 2ε0Bc/(eR): only 0.27% of Brillouin at 5 T (0.14% at 10 T), still needing 3.7 × 108 V.

So the Brillouin ceiling is unreachable in a meter scale single column even with perfect electrodes. That limit is unique to this machine and is the whole reason subdividing helps.

Voltage limited, the way real traps run

Payload of a single 1 m³ trap by end electrode voltage
End voltagePositronsMassEnergy
1 kV8.8 × 10110.8 fg0.14 J
10 kV8.8 × 10128 fg1.4 J
100 kV8.8 × 101380 fg14 J
1 MV8.8 × 10140.8 pg145 J

For scale: the largest positron store reported is ATRAP's 4 × 109 positrons (about 3.6 attograms) at 1.2 K and below 6 × 10−17 Torr S8. ALPHA's antihydrogen positron plasmas hold a few million at about 1014 m−3 S9. The brightest slow positron beam delivers about 109 per second S21, so filling even 1 ng would take about 35 years.

3 · Machine two

The alternating multicell, worked out

Split the volume into many thin columns. Each column's potential scales with its own radius squared, so at a fixed electrode voltage the density per cell rises as (R/r)². This is the idea behind the multicell Penning-Malmberg trap proposed by Surko and Greaves S4: 95 cells at 5 T and 1 kV, aimed at 1012 positrons S5. A second prototype has now held electron plasmas in two off axis cells at once S6. Joshua's version alternates electron and positron cells so the far fields cancel.

Cross sections of the two machines Left: one large positron column inside a single electrode. Right: a honeycomb of small cells, alternating positron and electron columns, each inside its own electrode wall. + Single column: φ grows as R² Alternating cells: φ grows as r²
Acid columns hold positrons, lilac columns hold electrons, dark rings are electrode walls. Wall thickness and the electron cells are what the alternating design pays in volume.

Gain over one trap at the same voltage

G = (R / r)² × fill × share (share = ½ if half the cells hold electrons)

Gain by cell radius at 1 kV and 5 T, walls 20% of the cell radius (fill 0.75)
Radiusn (m−3)Alternatinge+ onlyWall E
10 mm2.2 × 1015940x1,900x0.2 MV/m
5 mm8.8 × 10153,700x7,500x0.4 MV/m
1 mm2.2 × 101794,000x190,000x2 MV/m
0.1 mm2.2 × 10199.4 million x19 million x20 MV/m

1000x needs R/r ≈ 52: cells about 1 cm in radius in a half meter tank. The gain does not depend on the voltage, only on geometry, so it keeps growing as cells shrink until something else binds.

What stops the gain

Brillouin, per cell. Each cell is still a single species plasma, so each is capped at nB. At 5 T and 10 kV that happens at r ≈ 0.14 mm; neighbors cancel only the far field, not the field inside a column.

Wall field. Ewall = 2V/r. Capping walls at an assumed 30 MV/m, Brillouin cells must be about 27 µm in radius at 5 T (about 7 µm at 10 T), and the wall field pulls electrons out of the walls of every positron cell.

Wall thickness and electron cells. Even with zero thickness walls the best honeycomb fill is 0.907, and half of it is electrons, so positrons fill at most 45% of the volume at Brillouin density.

Electrode count. With 30 µm cells, which reach Brillouin density at 10 T with about 2 kV, a 1 m³ array has about 2 × 108 cells. Even with row and column addressing that is roughly 105 wires into a 4 K space.

Mixing. If the walls are dropped so the species can neutralize each other, the Brillouin limit vanishes (a neutral plasma is capped only by pressure, about 3.6 × 1028 m−3 at 5 T and 10 K). That is the tempting version, and it destroys itself: at 1020 m−3 and 10 K, direct annihilation alone empties a mixed plasma in about 2 ms, and positronium formation is faster still S11 S20.

Instabilities. Diocotron modes per column, two stream if species ever stream through each other, and field errors that torque every column S2.

Is ~1000x achievable if everything is perfect? Only by choosing the baseline. Perfect on both sides (Brillouin cells versus a meter column held below light speed drift): G ≈ 0.45 × eBR/(4mc) ≈ 33 × B[T], so 1000x needs about 30 T across a meter. Perfect array against a real 10 kV trap: millions. Perfect array against a perfect Brillouin trap: 0.45. The densest version is not alternating at all: positron only cells beat it by 2x everywhere.

4 · The peer review

Problems only one machine has

Unique means the other design does not have it at all, not that it has it less.

Unique Single Penning-Malmberg

  • Global space charge. Potential and wall field scale with the whole radius squared: 1011 V class at meter scale.
  • Light speed drift. E × B drift at the edge exceeds c above about 0.3% of Brillouin at 5 T, so the ceiling cannot be reached in principle.
  • Self energy larger than the fuel. 6.7 × 1011 J of electrostatic energy to store 2 × 107 J of positrons.
  • One rotor carrying everything. All angular momentum sits in one column; any field error torque expands the entire store, and rotating wall compression must hold the whole charge at once S7.
  • Single point of failure. One instability or one bad electrode dumps the whole payload.

Unique Alternating multicell

  • Positronium. Electrons are inside the machine by design. Any leak, transfer error or wall failure makes positronium, which annihilates in 125 ps (para) or 142 ns (ortho) S20. Recombination speeds up steeply as temperature falls S11, the opposite of what cryogenic storage wants.
  • Opposite polarity plugs side by side. Positron cells need positive end caps, electron cells negative ones: adjacent caps differ by twice the confining voltage.
  • Counter rotation. E × B rotation reverses with charge sign, so neighbors spin opposite ways; rotating wall drives need alternate phase per cell.
  • Charge balance bookkeeping. The far field cancels only if both species are matched cell by cell; any mismatch returns the global space charge.
  • Electron field emission into positron cells at microcell wall fields.
  • Wiring. 108 cells, 105 class lines into 4 K, and alignment of every column with the field, which the prototypes found essential S6.

Shared Both machines

  • Vacuum. Positrons annihilate on residual gas: about 33 days at 10−10 Torr of H2, about 9 years at 10−12 Torr, far longer in a 1.2 K cryogenic bore S13 S8. Some hydrocarbon contaminants annihilate positrons thousands of times faster.
  • Cryogenics. Cold electrodes, cryopumping and a superconducting magnet all live at 1.8 to 4 K.
  • Magnet quench. An Iseult class magnet stores 338 MJ, more than the ideal fuel load S14.
  • Cyclotron radiation. Useful cooling (0.1 s at 5 T), but all plasma heating ends up as millimeter wave power on 4 K walls.
  • Gamma shielding. A 1020 positron store with a one year lifetime is a 100 curie class 511 keV source while idle.
  • Filling. No source on Earth can fill either one in a human lifetime S21.

What real e+e− work tells us. Holding both species together is a real research goal (the APEX pair plasma project) S12, and nested Penning traps already hold opposite charges in adjacent wells to make antihydrogen S10; ATRAP used electrons to cool positrons during its record accumulation S8. Those are dilute, short lived or deliberately recombining systems, not dense long term storage.

5 · Idle operation

Waste heat at idle

At idle nothing is supposed to happen, so almost all the heat comes from keeping things cold. Every watt that lands at 4 K costs many watts at the wall plug, and all of it leaves as waste heat.

Cost of one watt at 4 K

  • Carnot limit: (300 − 4.2)/4.2 ≈ 70 W per watt lifted.
  • Large helium plants: the best reach 25 to 30% of Carnot S18: about 250 W/W.
  • Commercial cryocoolers: over 1 kW/W S18; a standard 4 K Gifford-McMahon unit lifts 1.5 W at 4.2 K for about 7 to 8 kW of compressor power S19, about 5,000 W/W.

Heat sources modeled

  • Radiation from a 50 K shield to 4 K: about 0.15 W.
  • Supports, magnet cryostat and leads (assumed 0.5 W/m² of cold surface): about 4 W.
  • Wiring: 0.1 mW per line at 4 K, 5 mW at 50 K.
  • Plasma heating (assumed 1 eV/s per positron), radiated as cyclotron emission to the walls: 29 W for 1.8 × 1020 positrons.
  • Annihilation gammas with a one year lifetime, half absorbed cold: 0.5 W.
  • Electronics, vacuum gauges and ion pumps 1.5 kW; rotating wall drives at 30% efficiency plus 0.1 W per line.
  • Magnet in persistent mode draws nothing; above 20 T the calculator switches to a resistive estimate.
Idle wall plug power for 1 m³
Case4 K loadCryocoolersPlant efficiency
Single trap, 10 kV, 5 T (1.4 J stored)4.3 W23 kW2.8 kW
Surko style alternating, 1 kV, 1 cm cells, 5 T (136 J)4.4 W24 kW2.8 kW
Ideal alternating at Brillouin, 10 T, 30 µm cells (30 MJ)43 W235 kW29 kW

Order of magnitude answer: tens of kilowatts for any machine that could actually be built, rising to a few hundred kilowatts for the ideal dense array on commercial cryocoolers. Even the ideal tank spends its entire 30 MJ of fuel on idle power in about 2 minutes (cryocoolers), 18 minutes (a large plant's efficiency) or 30 minutes (Carnot).

The two biggest terms in the dense case, plasma heating and wiring, are assumptions; no experiment has held plasmas anywhere near these densities. Move them in the calculator and the conclusion holds across orders of magnitude.

6 · Try it

Positron tank calculator

Everything runs in your browser with the same formulas as above. Physics limits always apply (Brillouin, light speed drift); engineering limits switch off under “Everything perfect”.

Design the tank

Change anything; results update instantly.

Machine

1 to 100 T. About 12 T is the practical ceiling for a 1 m bore.

What you ask each cell to hold; the limits below may cut it.

Cells

Finite limits also cap the wall field at 30 MV/m.

Positron payload0
Annihilation energy0
Density vs Brillouin0
Gain over one trap, same limits0
Idle waste heat, wall plug0
Idle power eats the fuel in0

Where the idle power goes

    7 · Side by side

    The two machines compared

    Single Penning-Malmberg trap versus alternating multicell trap, 1 m³
    PropertySingle Penning-MalmbergAlternating multicell
    Density ceilingBrillouin in principle; about 0.3% of it at 5 T in a half meter column (light speed drift)Brillouin per cell; 45% of it averaged (half the cells are electrons, honeycomb fill 0.907)
    Ideal payload, 10 T0.6 ng (drift limited)166 ng to 200 ng
    Practical payloadAbout 1 fg to 1 pg at 1 kV to 1 MVAbout 0.75 pg with 1 cm cells at 1 kV
    Confining voltage at Brillouin1011 V classAbout 2 to 5.5 kV with 30 to 100 µm cells
    What sets the gainNot applicable(R/r)² × fill × ½; the ½ is the electron cells' cost
    Signature failureWhole store expands or dumps at oncePositronium at any species contact; gamma flash
    Electrode countTens108 cells, about 105 addressed lines
    Rotation and compressionOne rotor, rotating wall over the whole chargeCounter rotating neighbors, alternate phase drives at tens of GHz
    Vacuum, cryo, magnet, gammasShared: below 10−12 Torr or a cryogenic bore, 1.8 to 4 K, a 10 T class magnet storing more energy than the fuel, 100 Ci class gamma shielding
    Idle wall plug, 1 m³About 3 to 23 kWAbout 3 to 24 kW practical; 16 to 235 kW ideal dense
    Better alternativePositron only multicell with grounded walls: same physics, twice the payload, no positronium risk by design

    8 · Fine print

    Assumptions

    1. The tank is a cylinder of radius 0.5 m and length 1.27 m (1 m³), filled end to end by long plasma columns; end effects are ignored.
    2. Each stored positron releases 2mc² = 1.022 MeV when it later meets an electron in the reactor. Production energy, which is many orders of magnitude larger, is not counted.
    3. Space charge uses the long uniform column result with the plasma filling its electrode. The light speed bound uses Eedge < cB, a nonrelativistic shortcut for the relativistic Brillouin flow.
    4. Cells sit in a hexagonal honeycomb; fill = πr² / ((√3/2)(2r + t)²). Alternating designs give half the cells to electrons.
    5. Engineering limits under finite voltage: wall field capped at 30 MV/m (a rule of thumb for clean cryogenic electrodes, not a measured value).
    6. Heat model: 50 K shield with emissivity 0.05; 0.5 W/m² static load and 1 W/m² shield load over 1.5 × the cylinder surface; 6√(cells) + 20 wires; 1 eV/s of heating per positron; half of annihilation gamma power absorbed at 4 K; 1.5 kW electronics; persistent magnet below 20 T, otherwise a resistive estimate scaled as B² × bore from the 45 T hybrid.
    7. Annihilation on gas uses πr0²c·Zeff·ngas with Zeff ≈ 14.6 for H2 S13. Mixed plasma lifetime uses the Sommerfeld enhanced free annihilation rate only.

    Not modeled: magnet quench protection, radiation damage to electronics, loading and transfer losses, and how a reactor would extract the fuel without a gamma flash.

    9 · Checked

    Sources

    Each source was located and opened on Oct 16, 2026.

    1. L. Brillouin, “A theorem of Larmor and its importance for electrons in magnetic fields,” Phys. Rev. 67, 260 (1945). link.aps.org
    2. D. H. E. Dubin and T. M. O'Neil, “Trapped nonneutral plasmas, liquids, and crystals (the thermal equilibrium states),” Rev. Mod. Phys. 71, 87 (1999). link.aps.org
    3. J. R. Danielson, D. H. E. Dubin, R. G. Greaves and C. M. Surko, “Plasma and trap-based techniques for science with positrons,” Rev. Mod. Phys. 87, 247 (2015). link.aps.org
    4. C. M. Surko and R. G. Greaves, “A multicell trap to confine large numbers of positrons,” Radiat. Phys. Chem. 68, 419 (2003). sciencedirect.com
    5. J. R. Danielson, T. R. Weber and C. M. Surko, “Plasma manipulation techniques for positron storage in a multicell trap,” Phys. Plasmas 13, 123502 (2006): 95 cells, 5 T, 1 kV, target 1012 positrons. positrons.ucsd.edu (PDF)
    6. M. Singer, J. R. Danielson, S. König, T. Sunn Pedersen, L. Schweikhard and E. V. Stenson, “Multi-cell trap developments towards the accumulation and confinement of large quantities of positrons,” J. Plasma Phys. 89, 935890501 (2023). doi.org
    7. J. R. Danielson and C. M. Surko, “Torque-balanced high-density steady states of single-component plasmas,” Phys. Rev. Lett. 94, 035001 (2005). link.aps.org
    8. ATRAP Collaboration (D. W. Fitzakerley et al.), “Electron-cooled accumulation of 4 × 109 positrons for production and storage of antihydrogen atoms,” J. Phys. B 49, 064001 (2016). ADS
    9. ALPHA Collaboration (R. Akbari et al.), “Be+ assisted, simultaneous confinement of more than 15000 antihydrogen atoms,” Nat. Commun. 16, 10106 (2025). nature.com
    10. G. Gabrielse, S. L. Rolston, L. Haarsma and W. Kells, “Antihydrogen production using trapped plasmas,” Phys. Lett. A 129, 38 (1988), the nested Penning trap. inspirehep.net
    11. F. Robicheaux, “Atomic processes in antihydrogen experiments: a theoretical and computational perspective,” J. Phys. B 41, 192001 (2008). doi.org
    12. M. R. Stoneking, T. Sunn Pedersen et al., “A new frontier in laboratory physics: magnetized electron-positron plasmas,” J. Plasma Phys. 86, 155860601 (2020). cambridge.org
    13. G. F. Gribakin, J. A. Young and C. M. Surko, “Positron-molecule interactions: resonant attachment, annihilation, and bound states,” Rev. Mod. Phys. 82, 2557 (2010). arXiv
    14. N. Boulant, L. Quettier et al., “Commissioning of the Iseult CEA 11.7 T whole-body MRI,” MAGMA 36, 175 (2023): 11.72 T, 90 cm warm bore, 1.8 K, 338 MJ, 132 t. pmc.ncbi.nlm.nih.gov
    15. National MagLab, 45 Tesla, 32 mm bore hybrid magnet: 30 MW. nationalmaglab.org
    16. National MagLab, 100 tesla multi shot pulsed magnet: 15 mm insert bore, 15 ms. nationalmaglab.org
    17. MIT News, “MIT-designed project achieves major advance toward fusion energy” (20 T large bore HTS magnet), Sept 8, 2021. news.mit.edu
    18. A. Perin, B. Bradu, J. Bremer, S. Claudet, D. Delikaris and F. Ferrand (CERN), “Cryogenic efficiency and sustainability aspects for particle accelerators and detectors” (talk). indico.cern.ch (PDF)
    19. SHI Cryogenics cryocooler catalogue: RDK-415D, 1.5 W at 4.2 K; F-50 class compressors 6.5 to 8.5 kW. catalogue (PDF)
    20. R. Y. Shopa et al., positronium imaging in J-PET (2023): para and ortho positronium vacuum lifetimes 125 ps and 142 ns. bamsjournal.com
    21. MLZ Garching, NEPOMUC positron source: above 109 moderated positrons per second. mlz-garching.de

    Share this reveal

    Email Bluesky X LinkedIn Facebook

    Got a task you think AI can't touch? AskSmith.ai is invite only. Join the waitlist