In the October 16 space digest, JAXA’s MMX launch to Phobos got a single line. Yet this is a “first in history” kind of event: on Tuesday, October 20 at 04:41 Tokyo time (Monday evening in Europe and America), the first mission in history to land on another planet’s moon, collect a sample, and return it to Earth leaves Tanegashima. Not from an asteroid—the Hayabusas have already done that twice. From a moon, and not even ours.
And unlike the Mars longreads on this blog, where all the drama happened long ago (transits, occultations, Phobos’s slow death), here the drama is unfolding right now, live: the H3 rocket, with two failures in nine launches over three years, has to send a spacecraft on a five-year expedition with minute-level launch precision.
Window. October 20, 2026, 04:41:03 JST—that is 21:41 UTC on Monday, October 19, and 23:41 Kyiv time. The backup period runs from October 21 to November 7. Yoshinobu Launch Complex, Tanegashima. The launcher is H3, tenth flight (F10).
The stakes. The MMX spacecraft (Martian Moons eXploration) weighs roughly 4,200 kg with propellant and consists of three modules: propulsion, exploration (the four-legged lander), and return (with its own 500-newton main propulsion system and high-gain dish for the journey back). 13 scientific instruments, plus a rover. Partners: NASA (MEGANE, the pneumatic sampler, DSN, 15% of the sample), ESA (transponder and Estrack ground support), CNES (the MIRS spectrometer, navigation support), DLR (test infrastructure, the rover jointly with CNES).
Schedule. Arrival at Mars: August 2027. From September, three years in a quasi-satellite orbit around Phobos (the spacecraft “hangs” above one point on the moon in its own reference frame, like a geostationary satellite above Earth). In early 2029, the Idefix rover is deployed. In 2030, landing and sample collection. In 2031, the sample capsule lands in Australia, from where it will be taken to Sagamihara, to the ISAS laboratory outside Tokyo.
The target sample is a modest minimum of 10 grams. For comparison: Hayabusa2 brought back 5.4 grams from Ryugu, and that was enough for hundreds of papers. MMX aims to collect more than 10 g at once, in a single landing, with two different devices. And this is not about “how much.” It is about where from: Phobos, under the leading hypothesis, is a candidate for the role of an “archive of Martian ejecta”—meteorite strikes on the neighboring planet have dusted it with Martian material over billions of years, more than any other object nearby (by JAXA’s estimate, ~100 times more than distant Deimos). Those grams may contain a record of ancient oceans, organics, and—the dream ceiling—biosignatures.
Why land on a tiny potato 27 km across rather than on Mars itself? Because Phobos physics is physics taken to absurdity:
| Parameter | Phobos | Our Moon |
|---|---|---|
| Diameter | ~27 km | 3,474 km |
| Surface gravity | 1/1700 of Earth’s | 1/6 of Earth’s |
| Escape velocity | 11.4 m/s (41 km/h) | 2.38 km/s |
| Distance to planet | 6,000 km from the surface | 384,000 km from Earth |
| Orbital period | 7 h 39 min | 27.3 days |
Eleven meters per second. That is a sprinter’s speed at the finish of the 100 meters—at his peak, Usain Bolt reached 44.7 km/h. Standing on Phobos, you can toss a rock by hand hard enough for it to leave the moon forever and enter orbit around Mars. A person in a spacesuit, stumbling on a slope, could fly off the moon forever—when escape velocity is 11 m/s, caution stops being a virtue and becomes a condition of survival.
The 25-kg Idefix rover weighs 15 grams on Phobos’s surface—less than a chicken egg. CNES and DLR tested it in laboratory microgravity simulators, because “driving” there on wheels is effectively impossible: at the slightest acceleration, the rover lifts off the surface. Idefix will move by hopping, and the Franco-German team frankly calls it their riskiest element.
That very physics dictates the mission architecture. The spacecraft does not hover over Phobos under thrust—the propellant expenditure would be fatal to the budget. Instead, it takes up a quasi-satellite orbit, where the combined gravity of Mars and Phobos lets it “stand” above a chosen point. And landing here is not “bleeding off speed against gravity,” as it is on Mars, but a ballet of inertia: matching the surface, which itself races around its orbit at 2.1 km/s relative to Mars, then touching down at a relative speed close to zero.
But low escape velocity has a flip side, and it is humiliating for engineers: dust kicked up by the landing engines will not settle back down. It will enter orbits around the moon itself, smear it and any future spacecraft, and only partly return. JAXA and NASA spent years modeling how to land without blinding their own optics or fouling the very spot from which they will take a sample a couple of hours later. That is why MMX’s landing engines will run at minimum thrust right up to contact, and why the landing itself is planned with almost no downward “fountain.”
There are two collection devices aboard, and their pairing is not duplication but insurance against two different scenarios.
C-SMP (Japan, JAXA). A robotic arm with a cylindrical corer 25–30 mm in diameter. The planned depth is about 2 cm below the surface: that is what scientists need, because the top centimeter of Phobos has been “dusted” with fresh material for billions of years, while the “old” soil lies beneath. The corer punches into the regolith and seals the sample in a tube. The technology grew directly out of Hayabusa, but for MMX it had to be rewritten: the sample-mass requirements are more than a hundred times higher.
P-SMP (United States, NASA/Honeybee Robotics). Pneumatics: high-pressure compressed nitrogen fires short pulses into the surface and raises a cloud of surface regolith straight into a container. The whole operation takes fractions of a second.
Why duplicate? Because nobody knows what Phobos’s surface is. If it is loose ash, pneumatics will work perfectly, while the corer may simply fail to hold onto a column of soil. If it is a sintered crust, the corer will knock out a sample, while the pneumatic system will have nothing to lift. Two independent mechanisms on one landing leg—that is the wager: “we do not know where we are arriving, so we are taking both tools.” From an engineering standpoint, it is the soundest hedge possible: a main parachute and a reserve.
The story of the P-Sampler deserves a line of its own: Chris Zacny of Honeybee Robotics conceived pneumatic sampling back in graduate school, more than 20 years ago, carried it through NASA grants, the Planetary Society, and the pandemic, and in October 2024 the Flight Model was handed over to Mitsubishi Electric in Kamakura. Twenty years from an idea to a “leg” on an alien moon.
H3 deserves a word of its own. It is JAXA and Mitsubishi Heavy Industries’ flagship rocket, and over three years its story looks like a textbook case of “fell—tore it apart—fixed it”:
That makes seven successes out of nine, with both failures in the same second stage but through different mechanisms. JAXA has methodically addressed both causes. And still, the MMX wager—years of work and hundreds of millions of dollars, together with its partners—is on the ninth flight of a rocket that is still far from a veteran. This is not bravado but calculation: the backup window runs through November 7, and the spacecraft itself has enough propellant to compensate for insertion inaccuracies.
The mission’s scientific core turns on one question: what is Phobos, anyway? Two camps, two scenarios:
More than the fate of two rocks depends on the answer. If Phobos is impact-born, its “Martian dust” may preserve a record of the era when rivers flowed on Mars: traces of water, organic chemistry, at the limit, biosignatures. If Phobos was captured, its composition will tell us about the migration of water and organics into the inner Solar System—about where the water on worlds like Earth came from. Tom Statler, MMX program scientist at NASA, puts it this way: knowing Phobos’s origin will “shed light on that early epoch when Earth and other inner planets may have acquired much of their water and organic compounds.”
MEGANE (“eyeglasses” in Japanese; that is the name of the gamma-ray and neutron spectrometer from Johns Hopkins APL flying aboard) will map the chemical composition of the surface—both to choose a landing site and to provide context for the sample. The instrument catches gamma rays and neutrons knocked out of the soil by cosmic rays: every chemical element has its own “signature.” Then comes a whole battery: TENGOO (a telescopic camera with ~40 cm resolution), OROCHI (a wide-angle multispectral camera), MIRS (near-IR, CNES)—all working to choose a site where the sample will say the most.
A side story pleasing to any history buff. In Gulliver’s Travels (1726), the scholars of the flying island of Laputa discovered two moons of Mars with orbits at 3 and 5 Martian diameters and periods of 10 and 21.5 hours. The real Phobos and Deimos, discovered by Asaph Hall only in 1877: 1.4 and 3.5 diameters, periods of 7.66 and 30.35 hours. Swift missed the numbers, but guessed the fact itself and the broad picture—small, close, fast. Historians still argue about where it came from: deduction by analogy (Earth has 1 moon, Jupiter has 4, so Mars “must” have 2—which means they must be close and tiny, otherwise they would be visible from Earth) or something else. Voltaire picked up the idea in Micromégas (1752). That is why Deimos has the Swift and Voltaire craters, while Phobos has the Laputa region, the Lagado plateau, and the Clustril, Flimnap, and Grildrig craters: almost all of Phobos’s place names are Gulliver characters.
In that sense, MMX closes a literary loop: a mission that will bring back the first rock from a moon neighboring craters named after a book that wrote about those moons before astronomers did.
What I like about MMX is its engineering honesty. The mission does not pretend it knows everything: instead of one “confident” sampler, there are two incompatible principles; instead of an exact model of the surface, three years of orbital reconnaissance before landing; instead of a promise to “bring back half a kilo,” a modest 10 grams, guaranteed by two independent mechanisms. Every element is backed up, and the whole plan is built on the assumption that something will surely go wrong—which is why there are two samplers, why there is a scout rover before the landing, why there is a quasi-satellite orbit instead of a perilously low one.
And the second thing: MMX is the first test of a full-cycle “to Mars and back” technology. Not “get there and transmit data,” but “get there, land, take, leave, bring it home.” Every future Mars sample-return mission—and the entire logic of “bring Mars into an Earth laboratory”—will inherit from this flight. If everything goes smoothly on Tuesday, by 2031 humanity will have a sample from the first alien moon and a proven “Mars → Earth” route with cargo. If not, Phobos can wait: by cosmic standards it has 30–50 million years left (sooner or later tidal forces will tear it apart into a ring around Mars), but the first chance to get to know it better is right now.
Tuesday at 04:41 Tokyo time—a livestream on JAXA YouTube. If H3 decides to surprise us again, I will write an epilogue to this post.