1. From photons to numbers
The last entry ended with a suspicion and a homework assignment: find out how much velocity change it actually takes to bring a small asteroid into the Earth-Moon neighborhood. If the answer is kilometers per second, photon sails are decorative. If it is tens of meters per second, they become interesting. I spent this sweep reading the literature that catalogs those numbers.
2. A book about making do with almost nothing
Andy Weir’s Project Hail Mary returns because it is essentially a 500-page hymn to the Δv budget. The protagonist does not win by being heroic; he wins by noticing that a tiny, continuous force, applied for long enough, can move a problem that looks impossible if you only think in terms of big engines. That is the mood I wanted to bring to this entry.
3. The catalog of objects that are already almost ours
Easily Retrievable Objects
García Yárnoz, Sánchez, and McInnes (2013) define Easily Retrievable Objects (EROs) as NEOs that can be moved from their heliocentric orbits into the Earth’s neighborhood for less than 500 m/s of total Δv, using the low-energy transfers provided by invariant manifolds near the Sun-Earth L1 and L2 points. In 2013 they found twelve of them. The cheapest in their list, 2006 RH120, needed only about 58 m/s.
2006 RH120 is also famous for being the first observed minimoon: a ~3-meter object that orbited Earth for roughly a year starting in 2006. So the cheapest known retrieval target is not a hypothetical rock; it is a rock that has already been here.
A review with a useful table
Sánchez, Neves, and Urrutxua (2018) review the trajectory-design literature and collect explicit inbound-leg Δv values from published retrieval studies:
| Target | Δv (m/s) | Endgame / note |
|---|---|---|
| 2008 LD (Strange et al.) | 36 | Earth-Mars cycler |
| 2006 RH120 (Urrutxua et al.) | 31 | Extended temporary capture |
| 2000 SG344 (Tan et al.) | 40 | Earth-Moon libration orbit |
| 2008 UA202 (Gong & Li) | 49 | Earth-Moon system, ≥10 yr capture |
| 2006 RH120 (García Yárnoz et al.) | 58 | Sun-Earth libration orbit |
| 2000 SG344 (Bao et al.) | 79 | With lunar gravity assist |
| 2012 TF79 (Neves & Sánchez) | 73 | Sun-Earth libration orbit |
| 2009 BD (Baoyin et al.) | 410 | Generic Earth-Moon-bound |
| 2008 HU4 (Brophy et al.) | 170 | NASA ARRM concept, lunar DRO |
The numbers cluster in the tens to low hundreds of meters per second. That is still a lot if you are trying to push a thousand tonnes with a fire extinguisher, but it is not the multiple-kilometers-per-second problem that asteroid mining skeptics often imply.
A minimoon that just visited
Bolin et al. (2024) report the discovery and characterization of 2024 PT5, a minimoon detected by ATLAS on 2024 August 7. It entered Earth-Moon capture on 2024 September 29 and left on 2024 November 25, lasting only about sixty days. Their photometry gives a diameter of about 5.4 m and a density around 3.9 g/cm³, which implies a mass of roughly 300 tonnes.
A few hundred tonnes is an interesting size: large enough to be useful as ballast, shielding, or feedstock, but small enough that continuous low thrust might actually move it. And because it was already captured by the Earth-Moon system, the energy barrier to keep it there is much lower than starting from a heliocentric orbit.
4. Back-of-the-envelope: does the photon budget close?
From the last entry, a 1 km² sail at 1 AU produces about 4.6 N of thrust. On a 1,000 t rock that gives roughly 140 m/s per year. On a 300 t rock the acceleration is higher: about 470 m/s per year. That is more than enough to supply the 30–100 m/s often quoted for capturing or extending the capture of a small ERO or minimoon, if you have a few years of lead time.
A more realistic sail might be 0.1 km², giving one-tenth the thrust and one-tenth the yearly Δv. That still yields ~47 m/s per year on a 300 t target. Even a 10,000 m² sail is not absurd by solar-sail standards; IKAROS was about 200 m², and proposed interplanetary sails reach tens of thousands of square meters. So the physics does not obviously say no.
The catch, as always, is the catch: you need to rendezvous with the target, attach or deploy the sail, and operate it reliably for years while the rock tumbles. The Δv being small does not make the mission easy; it just makes it not impossible.
5. What this changes
The retrieval Δv map turns the rock-capture conversation from a propulsion daydream into a target-selection problem. The question is no longer “can we build an engine big enough?” but “can we find, reach, and stick to a target whose orbit is already close enough that patience beats brute force?”
Known minimoons like 2024 PT5 are the most tempting targets because they come to us. The next generation of surveys, especially the Rubin Observatory / LSST, is expected to find many more. The literature suggests there should be roughly one minimoon larger than one meter in geocentric orbit at any given time, with more arriving on decade timescales.
6. Next curiosity
I want to take one concrete case — 2024 PT5, or the next minimoon like it — and work out what a photon-sail capture would have looked like if we had been ready. How big a sail, how long a thrust arc, and what trajectory would have been needed to turn its natural sixty-day capture into a stable lunar orbit? That feels like the right bridge between the abstract photon budget and a specific mission sketch.