It's been pointed out before in this thread that antimatter production is a net energy loss. But the topic deserves more.
An antimatter reactor wouldn't be just what the average person thinks of vaguely as somewhat inefficient, which gives no understanding of how many orders of magnitude are involved here.
Antimatter is currently literally about a *billion* times more expensive per unit of energy than nuclear fuel. After foreseeable future technological development with dedicated antimatter production facilities, it would become on the order of a *million* times more expensive instead.
---------------------
Approximate order-of-magnitude cost illustrations now and for the foreseeable future, outside of extreme situations like a post-singularity uber civilization filling the solar system with self-replicating factories:
- 1a. Total cost of terrestrial nuclear electricity generation (mostly capital and not fuel expense) =~ $0.01 / megajoule (electric)
- 1b. Methods reducing cost are conceivable, especially for a space civilization, such as =~ $0.001 / megajoule (electric)
- 2. Fuel cost of uranium from seawater, without breeding, for thermal energy from fissioning the U235 fraction =~ $0.0001 / megajoule (thermal)
- 3. Energy release from current thermonuclear bombs relative to total production cost =~ $0.0001 to $0.001+ / megajoule (thermal, uncontrolled)
- 4. Fuel cost of lithium-6 deuteride for fusion (if tech using it available) =~ $0.00001 / megajoule (thermal)
versus
- 5. Cost of antimatter production using current particle accelerators not designed for it (1 per 100000 collisions) =~ $100000 / megajoule-antimatter
- 6a. Estimated cost of antimatter with a dedicated production facility (different GeV, 1 per 20 collisions collected) =~ $100 / megajoule-antimatter
- 6b. Optimistic assumptions for a future antimatter factory =~ $1 to $100 / megajoule-antimatter
---------------------
Basis:
1. Several cents per kilowatt-hour (per 3.6 megajoules) can be the total cost for bulk industrial production of nuclear electricity today. Of course, you probably see a higher figure such as $0.10/kWh or more on your electricity bill, but total prices for residential customers are higher wherever one lives, including distribution cost to all the individual end-users, etc.
2. Thermonuclear device cost can be as low as between a fraction of a million dollars and several million dollars per megaton ... per 4 billion megajoules.
3. #3 is based on estimates in prior posts here and here. The cost of current uranium production from mining is also a comparable order of magnitude.
4. Deuterium is currently on the order of $3000 per kilogram (e.g. $600 to $700 per kilogram of heavy water); Li6D =~ $1000 / kilogram. Energy release for such is 64 kilotons per kilogram or 270 million megajoules per kilogram before inefficiencies.
5. Currently antimatter production in particle accelerators is extremely inefficient, e.g. on the order of 1 antiproton per 100000 proton collisions, with an input energy cost on the order of $60+ million per microgram of antimatter with $0.10/kWh or nominally potentially less with cheaper electricity. A microgram of antimatter reacting with a microgram of matter results in annihilation of the two total micrograms involved; E = mc^2, so that is 180 megajoules. (Some discussions refer only to the energy equivalence of the antimatter itself, 90 MJ/µg).
6a. A dedicated production facility could successfully collect an antiproton per 20 collisions of 200-GeV protons against high-atomic-mass material, increasing efficiency so as to reduce antimatter production expense to $25000 per microgram. That's described here and also corresponds approximately to Robert Forward's estimate for a purpose-built antimatter factory.
6b. Optimistically, one could consider a possibility that the assumed electricity input cost might be reduced by an order of magnitude if future power is eventually cheaper. Better than 1 antiproton per 20 collisions might also be obtained, maybe. But the result is still expensive.
---------------------
Applications where an antimatter reactor could be worthwhile are hard to find with its astronomical cost. Plus, storing a really large quantity of antimatter that will all go up upon the slightest contact with matter is about the most impractical method imaginable for power generation from a safety and engineering perspective ... any air or gas leaking in, a sub-microscopic speck of dust, an unanticipated shock or too much vibration, any penetrating damage...
Stationary general power plants running on antimatter are obviously implausible.
What about using antimatter in starship propulsion?
There's the example of the Valkyrie idea, to
have a starship massing 200 tons total accelerating to most of lightspeed. Hypothetically, about 50 tons of antimatter would be stored as solid antihydrogen, apparently thought to have low enough vapor pressure at under 1 Kelvin, to be kept away electrostatically or by magnetic fields from the tank walls. In their idea, some antihydrogen would be ionized and guided out of a magnetic bottle to react with matter, spraying extreme-velocity particle exhaust to propel the ship. The crew compartment would be on a ten-kilometer tether so most of the very penetrating radiation didn't reach it, protected by a small shield.
There are many questionable aspects about that, including whether or not heat transfer versus the limits of materials was analyzed from a quantitative engineering perspective ... not overlooking it but performing calculations. Outside of doubtful assumptions, there's no point in trying for 92% of lightspeed on a journey to a nearby star, not if the time to accelerate without more power than materials could handle would be enough years that a few percent of lightspeed would be more appropriate anyway.
(The only self-propelled starship concepts I know to have been based on proper thermal calculations are various forms of external nuclear-pulse propulsion including inertial confinement fusion, not this one; skipping much longer discussion, just note that this lacks a number of the factors which should make the performance goals of the former workable).
Worst of all, even if antimatter production expense became as efficient as an antiproton per 20 collisions at 200 GeV proton energy as previously discussed, at $25000 per microgram, the ship's 50 tons of antimatter would have an economic and industrial opportunity cost nominally around a *million* trillion dollars.
---------------------
In contrast, for example, consider sending a 1000-ton dry-weight colonization starship at 10% lightspeed to reach a nearby star in several tens of years, sent by a space civilization to set up another on the asteroids and comets of the destination. Neglecting details like whether a magnetic brake could be used for deacceleration, the nominal minimum energy requirement is 500 trillion megajoules before (enormous) inefficiencies.
When energy requirements are so enormous, one likely does not want to use astronomically expensive energy like antimatter but rather cheap energy, more like #3 / #4 than #5 / #6 in the earlier list.
With potential future thermonuclear pulse propulsion, the base fuel cost for Li6D fusion fuel can be $0.00001 / megajoule or less, e.g. the earlier $1000 per 270 TJ being ~ 3.7E-6 $/MJ.
The nominal fuel expense for Li6D would thus be on the order of $2 billion; large inefficiencies would raise that figure greatly, although other factors such as deuterium production expense potentially decreasing below $3000/kg for a space civilization could reduce the equivalent cost. Of course, total starship expense would be more than fuel alone.
However, at least with such nuclear fuel, a colonization starship could reach another star system for a relatively conceivable allocation of resources for a hypothetical future civilization, a plausible number of billions of dollars per thousand tons.
But what if one used antimatter costing a figure like the $100 / megajoule of #6? Then the 500 trillion megajoules becomes a crazy 50000 trillion dollars nominally just for fuel cost even before considering how drive inefficiencies could raise that figure further. Outside of singularity-type assumptions, the equivalent of thousands of years of current world GDP being spent on just that project would be doubtful, to say the least.
*That's* one of the huge disadvantages of antimatter propulsion compared to nuclear pulse propulsion.
(In this post and elsewhere, I often take unpopular positions compared to what's typical in science-fiction, like nuclear pulse propulsion versus antimatter or mass drivers versus space elevators ... but that's because of considering matters quantitatively including real-world factors like plausible economics).
---------------------
If one made the starship instead just a probe returning some scientific data, decreasing mass by orders of magnitude, production expense for the lesser quantity of antimatter fuel would be correspondingly less. However, it'd still be enormous expense, and, for a really small probe, beamed energy is more likely.
(That's even aside from the matter of how much antimatter propulsion could be scaled down without too much penetrating radiation uselessly escaping omnidirectionally when it needs to be directed for the "photon rocket"'s thrust).
Even a cost such as $0.001 to $0.01 per megajoule for electricity from large power stations powering with moderate inefficiency a beam propelling a lightweight probe to fractional lightspeed is likely to be much better than spending $100 per megajoule for antimatter.
For example, nominally, a 100-kilogram probe can be sent to 10% lightspeed with an X inefficiency factor for $X * 0.5 billion of electricity. The exact figure isn't important, but what matters is that it isn't astronomical.
---------------------
A antimatter starship would work better if one could throw economic considerations out the window.
Consider a totally different scenario than almost all sci-fi: an astronomical posthuman post-singularity civilization fully utilizing the solar system's resources with self-replicating factories. Then economic / industrial output millions of times greater than today could allow them to do almost anything they wanted, even build antimatter starships whether small or large. There's still some question of whether such would be preferred to alternatives in that case, though.
---------------------
In the foreseeable future, plausible applications for antimatter involve using exceedingly tiny amounts, as opposed to using it as a bulk source of energy.
For example, as discussed more in a prior post
here, the ICAN II
concept for a high-performance interplanetary spaceship using antiproton-catalyzed microfission/fusion had pellets of 0.8 kg inert propellant mass around a 0.003-kg nuclear fuel capsule, each one releasing 302 billion joules of nuclear energy but with only 6 joules of antimatter used to help ignite its very center, a tiny fraction of a trillionth of a gram of antiprotons used per pellet.
Like matches igniting forest fires, the ship's total antimatter energy equivalent of a few hand grenades indirectly leads to the nuclear energy equivalent of tens of millions of tons of TNT, not all at once but spread out over half-a-million pellet detonations. For example, the basic technology could do a quick journey to Mars or really anywhere in the solar system with some modification.
Currently, Penning traps are used to store tiny amounts of antimatter. At liquid helium temperature and with a magnetic field, a cloud of antiprotons can be stored, a limited amount with their mutual repulsion.
The ICAN II antiproton-catalyzed microfission/fusion spaceship would be based on having 1000 antiproton traps each 0.3 meters in diameter and a meter long. Each trap would
be able to store 10^14 antiprotons without too much loss over a several month period, an improvement over the current HiPAT
device. The total amount of antimatter stored would be 140 nanograms. An accident would present the special concern of penetrating radiation, but there is the energy equivalent of merely several-percent as much as a hand grenade per trap (each the size of a small trashcan but 125 kilograms each).
In such a special application with small amounts needed, antimatter is feasible. Likewise, antiprotons have been used in some medical imaging even today. Merely a
small fraction of a trillionth of a gram of antiprotons is sufficient.
In principle, greater antimatter density than today's Penning traps could be obtained by storing antimatter as antihydrogen. Electrically neutral antihydrogen atoms in a condensate would not repel each other like the antiprotons, if such could be produced someday in sufficient quantity without too many additional inefficiencies and difficulties. In 2002,
for example, CERN produced clouds of tens of thousands of antihydrogen atoms. But a single microgram of antihydrogen would take about 6E17 atoms, trillions of times more.
---------------------
So the most realistic antimatter reactor is no antimatter reactor as such but, rather, only antimatter used in small quantities in specialized applications.
