OrbitalHub

Mariner 9 arrived at Mars on November 14, 1971, the first spacecraft ever to orbit another planet. A planet-girdling dust storm forced the mission to wait, and a planetary scientist named Joseph Veverka suggested putting the cameras on Phobos and Deimos in the meantime. Phobos came into focus on November 29 from 14,440 km, then again on December 1 from 5,720 km. The little 27-kilometer rock was already strange: a near-potato shape scarred by a 9-kilometer crater called Stickney, a long axis pointed permanently at Mars, and a charcoal-black surface that did not match the warm reddish dust it was orbiting (NASA/JPL, 1971–1972).

Two hypotheses have run in parallel since Asaph Hall discovered both moons in August 1877. The capture hypothesis says Phobos and Deimos are asteroids from the outer belt, drawn in by aerodynamic drag through an extended early-Martian atmosphere and circularized into their present orbits (Hunten 1979; Pollack et al. 1979). The color match between Phobos’s surface and the so-called D-type asteroids, dark, organic-rich bodies, has carried that story.

The giant-impact hypothesis argues the moons formed from debris thrown up when a roughly Mars-sized protoplanet slammed into early Mars, much as our own Moon is thought to have formed from a glancing blow on Earth (Craddock 2011; Rosenblatt & Charnoz 2012). If true, Phobos is made from re-accreted Martian mantle mixed with impact glass, with up to ~30% porosity to explain the low density.

The distinction isn’t trivia. If Phobos is a captured D-type, MMX returns a bona fide sample of outer-belt material with all the volatile- and organic-rich implications that follow. If Phobos is impact debris, MMX effectively becomes the first Mars sample return mission, a piece of the early Martian mantle frozen in place because there was never enough gravity to melt it (Burns 1992; Lainey & Jacobson 2017).

The first attempts were Soviet. Phobos 1 launched in July 1988 and died 51 days later of a software bug. Phobos 2 reached Mars in January 1989, swung past Phobos, and went silent 30 m above the surface during its second landing rehearsal, returning 37 images. Phobos-Grunt launched in November 2011 on a Zenit-2, failed to leave Earth orbit, and burned up over the Pacific Ocean, taking its piggybacked Chinese Yinghuo-1 smallsat with it.

Between those failures, orbiters gave us what we had. ESA’s Mars Express has now completed more than 130 dedicated Phobos flybys; Mars Reconnaissance Orbiter (HiRISE) returned the highest-resolution images of the surface, down to ~6 m/px; Perseverance and Curiosity imaged Phobos eclipsing the Sun from the ground. None was a sample return.

JAXA committed. MMX was approved in 2019 with Kawakatsu Yasuhiro as Project Manager, deliberately built on Hayabusa and Hayabusa2 heritage. It launches in JFY 2026 on H3 from Tanegashima. It carries a Coring Sampler, a NASA-built pneumatic sampler, and the DLR–CNES IDEFIX rover, which is to be released from below 100 m altitude and operate for ~3 months. After regolith capture, MMX lifts off Phobos, swings out of orbit, and heads home. The sample capsule lands in Australia’s Woomera desert in JFY 2031 (JAXA, 2026).

Phobos’s specifics dictate the engineering. It is tidally locked to Mars and orbits at 9,376 km from the planet’s center, inside synchronous. Its mass is only 1.064 × 10¹⁶ kg. Surface gravity is 0.0057 m/s² and escape velocity a casual 11.4 m/s. Dust kicked up by landing takes a long time to settle because there is essentially no atmosphere to grab it (Wikipedia, citing Jacobson 2010; Lainey et al. 2024).

The body itself is also a study in low-cohesion regolith. Stickney, the dominant crater, is nearly half the mean diameter of the moon — close to the threshold where a same-sized impactor should shatter the parent rather than scar it. The long, shallow grooves running away from Stickney’s rim are still debated: some teams interpret them as fractures propagating from the basin-forming impact (Asphaug & Melosh 1993), while others argue for ejecta-drape patterns or secondary craters from Mars-thrown debris (Murray & Heggie 2007). Mars Express and MRO photometry has refined the question but not closed it; the grooves sit at the wrong angle to simple Stickney-radial fracture models, leaving loose ends that ground truth will resolve.

Sampling strategy reflects the same facts. The primary sampler is a Coring Sampler (C-Sampler): a titanium-alloy tube pressed against the surface under low main-engine thrust, intended to pull up to a 2-cm core from the uppermost regolith. Secondary is the Pneumatic Sampler (P-Sampler), contributed by NASA: a short burst of nitrogen gas lifts loose regolith into a chamber. Multiple sites will be visited across lit and shadowed hemispheres, with a combined sample target of more than 10 g (JAXA, 2026 mission specification).

In-situ discrimination happens before the sample ever leaves Mars. NASA’s MEGANE instrument (Mars-moon Exploration with GAmma Rays and NEutrons) is a high-purity germanium γ-ray spectrometer paired with a ³He-neutron spectrometer, designed to map iron, silicon, oxygen, hydrogen, thorium, and potassium across Phobos’s surface (NASA / APL, 2024). D-type asteroidal material and basaltic Martian-mantle-derived material produce very different signatures in such an instrument, particularly in the Fe/H and Th/K ratios. Knowing the answer in orbit lowers the risk of returning rock that doesn’t actually answer the central question.

Return physics is unforgiving. The Δv budget to enter Martian orbit from a direct Hohmann transfer is on the order of 1 km/s, plus a few hundred m/s for Phobos orbit acquisition and the touch-and-go capture. Per Tsiolkovsky’s rocket equation: Δv = v_e × ln(m_0 / m_f), with v_e ~ 320 s for the bipropellant thrusters. The mass fraction is what makes the Sample Return Capsule, with its ~10 g payload, only a small fraction of the ~4,200-kg launch stack. The capsule’s heat shield has to survive entry into Earth’s atmosphere at ~12 km/s after a direct Mars return.

H3 is Japan’s first fresh-built orbital launcher since the 1980s. Its LE-9 engines run an expander bleed cycle, each chamber delivering ~1,422 kN of vacuum thrust. H3 sent up Michibiki 7 earlier this month, and MMX is the next major LE-9 passenger on the manifest (JAXA / MHI, 2026).

JAXA’s MMX is going to do something the United States and Russia combined have tried and failed to do three times: land briefly on Phobos, grab a few grams of regolith, and bring that bag of rock back to a clean lab on Earth.

Three outcomes are plausible. If the rocks look like D-type carbonaceous material with deep hydration features, the capture hypothesis wins and Phobos becomes the first outer-belt material ever returned to Earth. If they are dry, fine-grained basalts with high calcium and iron and no hydration, the impact hypothesis wins and we have, in effect, a sample of early Mars. If they are a messy mix, we’ll learn something about Phobos’s regolith history that neither hypothesis prepared us for.

The capsule should come down near Woomera in JFY 2031. JAXA’s Sample Curation Facility at Sagamihara will be where scientists first crack open the sealed container. The instruments they run on that small sample — mass spectrometers, electron microprobes, atom-probe tomographs — will quietly close out a 150-year-old argument about where Mars got its moons.

For three decades, “habitable zone” was the label that mattered most when astronomers tried to pick out Earth-like worlds. A planet at the right distance from its star, where water could stay liquid, made the shortlist. Now there is a second gate, and the radio sky is starting to open it.

In the second half of September 2026, two pieces of work landed within a week of each other. The first, by Kevin N. Ortiz Ceballos and colleagues at the Center for Astrophysics | Harvard and Smithsonian, used South Africa’s MeerKAT array to record auroral radio emission from Beta Pictoris b and, by fitting that emission, measured the planet’s magnetic field at no less than 1,250 gauss where the radio waves were produced, more than a hundred times the field strength near Jupiter. The second, a chapter in the SKA Observatory’s 2026 science book Advancing Astrophysics with the SKA II (AASKAII), lays out how the larger Square Kilometre Array will turn that lone detection into a population survey running across the radio sky (Kavanagh et al., arXiv:2607.11507, July 13 2026).

The Ortiz Ceballos paper, posted to arXiv on September 19, 2026, reports “the first direct detection of auroral radio emission from an exoplanet” (arXiv:2609.16720). The Kavanagh chapter anticipates “thousands of detections of [ultracool] dwarfs within a few hundred parsecs”. Read together, they mark the moment exoplanet magnetism stops being a curiosity of one star system and starts to look like a measurable physical quantity across the Milky Way.

Why this matters takes a paragraph to lay out and a century of physics to back up. A magnetic field around a rocky planet is a shield: it deflects the stellar wind, it keeps the upper atmosphere from being sandblasted into space, it channels incoming charged particles into auroral rings rather than letting them rain onto the surface. The stripped atmosphere of Mars, the thin envelope around present-day Venus, and the divergent atmospheric histories of Earth and Venus over four billion years trace in significant part to whether the planets kept dynamos. Exoplanet climate models that include magnetic fields produce a more Earth-like picture for many candidate worlds; models that ignore magnetic fields systematically over-predict atmospheric retention. For three decades the field had to be inferred from atmospheric escape rates or from star-planet interaction signatures baked into transit spectra. Both methods were indirect.

What changed in September 2026 was that MeerKAT picked up repeating low-frequency bursts from the Beta Pictoris system and the team could place the source unambiguously on the planet, not the star. The trick was a calibration borrowed from very long baseline interferometry: pin the array’s coordinate frame to background quasars so distant they barely move on human timescales, then re-grid every radio image against that fixed sky. When the team overlaid their burst positions on the quasar-locked grid, the radio centroid sat precisely on the orbital position of Beta Pictoris b at the time of the observation rather than on its parent star (Ortiz Ceballos et al., 2026). Past exoplanet radio claims had been ambiguous because the array’s beam was wider than the planet-star separation; MeerKAT’s combination of baseline length, up to 8 km in the Karoo array, and the quasar-locked astrometry pushed that ambiguity below the planet’s projected offset.

The MeerKAT emission corresponds to an electron cyclotron maser instability operating in the planet’s magnetosphere. Energetic electrons trapped along open magnetic field lines near the planet’s magnetic poles radiate coherently at harmonics of the local electron-cyclotron frequency f_ce = (e*B) / (2*pi*m_e). The useful relation between frequency and field strength is f_ce (MHz) ~ 2.80 * B (gauss). Plugging in the 1,250 gauss figure gives a peak cyclotron frequency near 3.5 GHz, which matches where the team reports the most luminous bursts. The mechanism is the same one that lets Jupiter’s decametric emission light up ham radio receivers every few decades when Earth passes through Io’s plasma torus, and that lets the giant planets of our solar system outshine their host stars in the low-frequency sky. Beta Pictoris b’s 1,250 gauss is roughly 100 times Jupiter’s 4 to 13 gauss surface field. The authors note that figure matches theoretical predictions for young, massive giant planets whose dynamos are driven by efficient heat transport through a still-cooling interior rather than core convection alone.

If you stopped reading here, the story would already be a first. The AASKAII chapter by Robert D. Kavanagh and sixteen coauthors sketches what comes next. The chapter’s premise is that the SKA, when it reaches full operation, will deliver both an order of magnitude more sensitivity at low frequencies through SKA-Low in Western Australia (50 to 350 MHz) and astrometric precision in the microarcsecond regime at higher frequencies through SKA-Mid in the Karoo (0.35 to 15.4 GHz, baselines extending to 150 km in Phase 1). That combination does two new things at once. It opens up giant exoplanets whose magnetic field strengths put their cyclotron emission in the tens-of-megahertz range that is now too faint for any current array to detect. And it allows the team to monitor the radio centroids of ultracool dwarfs, substellar objects spanning the mass range from Jupiter to brown dwarfs that have been known to emit coherent auroral radio for two decades, over multi-year baselines to detect the wobble induced by a planet of a few Earth masses. Astrometric monitoring turns a planet’s magnetic field into a planetary-mass measurement at the same time; the same observation that confirms a magnetic field also weighs the planet, including any unseen companions whose slight gravitational pull makes the host wander.

What changes across the next twenty years is the kind of work that becomes routine. Detection on a single target will give way to a sample size in the thousands; the Kavanagh chapter suggests SKA-Mid’s first decade of ultracool-dwarf monitoring could yield enough astrometric wobble detections to populate a mass-versus-magnetic-field diagram comparable in coverage to the radial-velocity planet catalogues of the 2010s. Statistical statements about magnetic field strength against planetary mass, stellar activity, and orbital distance will replace single-source case studies. The chapter does not promise that an Earth-mass rocky planet’s magnetic field will be measurable before 2040, because those fields project emission at kilohertz frequencies the ionosphere absorbs. It does argue that the SKA’s first decade of operation will make magnetic field a primary observable for the gas giants and ice giants that are the plausible hosts for life in our galactic neighbourhood. By the time biologists get serious about looking for biosignatures on rocky planets, the SKA programme will have spent a generation taking the magnetic-field census that makes those biosignatures physically credible. A radio antenna in the Karoo just gave that census its first datapoint.

A star only 0.55 light-years from Sagittarius A* is still making water.

Astronomers working on the Mid-Infrared Characterisation of Nearby Iconic galaxy Centres programme (MICONIC) used Webb’s Mid-Infrared Instrument (MIRI) to take the first continuous mid-infrared spectrum of IRS 3, an oxygen-rich asymptotic giant branch star that sits inside the most punishing radiation environment in the Milky Way. The spectra carry clear silicate dust features and, for the first time in this object, a water signature inside the star’s outflowing envelope. The result, published on 11 August 2026 in Astronomy and Astrophysics, shifts a long-standing assumption about galactic centres: stars there can keep producing the kind of molecular and dusty material that future planetary systems are built from, even when they are essentially inside the black hole’s sphere of influence.

Why this matters comes down to where the chemistry of galaxies gets done. A normal disk galaxy gets most of its interstellar dust and much of its recycled gas from evolved, mass-losing stars like IRS 3. These stars puff off their outer layers as they age, and the ejecta cool into grains that seed the next generation of star and planet formation. Galactic centres are different in kind, not just degree. The supermassive black hole at the centre of every large galaxy pumps out enough ultraviolet and X-ray radiation to break water apart in seconds in unprotected gas, and its gravity roars past anything orbiting closer than a few parsecs. If stars near galactic centres cannot produce dust and molecules, the inner few hundred light-years of a galaxy should be remarkably dust-poor, with a different chemistry than the disk. That in turn changes how astronomers read the spectra of distant galactic cores taken by mid-infrared space observatories, since the same chemistry that runs in our own galactic centre runs in every other large galaxy’s centre. The MICONIC programme was created to test exactly that question, and IRS 3 is its flagship target.

The Webb observation tells the IRS 3 story in detail. Lead author Florian Peissker of the University of Cologne and his team, including Macarena Garcia Marin of ESA (the MICONIC principal investigator), pointed MIRI at IRS 3 during the 2025 Guaranteed Time Observations campaign (programme ID 1266) and collected a continuous spectrum across the mid-infrared. The MIRI medium-resolution spectroscopy mode spans wavelengths where cold dust and molecular gas both radiate, so a single exposure covers the silicate bands, the water rotational-vibrational lines, and the broader dust continuum that anchors the dust mass estimate. Two strong silicate features in the data reclassify the star: earlier studies had flagged it as a candidate carbon-rich AGB star, but the new spectrum is unambiguously oxygen-rich, with the silicon-oxygen grain signature that is the standard fingerprint for a dusty red giant shedding mass in an oxygen-dominated envelope. Layered on top of the silicate features is a water signature inside the envelope, which is the first clear molecular detection for this object and the headline result. The reclassification is important on its own terms: carbon-rich AGB chemistry would have implied a different set of grain types, and confirming that IRS 3 follows the oxygen-rich path is the precondition for the water finding being meaningful.

The numbers behind that headline make it sharper. IRS 3 lies about 0.55 light-years from Sgr A*, roughly 3,300 astronomical units, which puts it well inside the sphere where the black hole’s tidal forces are a factor in the star’s mass-loss budget. The dust envelope itself extends some 10,000 AU from the photosphere, with temperatures falling from roughly 1,200 Kelvin close to the star to around 100 Kelvin in the outer regions; the spectrum is consistent with a shell-like, layered distribution rather than a smooth outflow, which is useful because layered shells are exactly the structures models expect when an AGB star loses mass in pulses. Stellar modelling puts IRS 3 at about six solar masses and roughly 72 million years old, near the end of its AGB phase. Both the geometry of the envelope and the chemistry of its grains point to a star that is still very much in the business of enriching its surroundings, as the ESA/Webb release weic2617 lays out.

The deep dive into why the water survives is where the physics gets interesting. Water in space is fragile because ultraviolet photons break the molecule into hydrogen and oxygen on timescales of days in an unprotected environment, and the same harsh UV field also drives chemistry off the silicate grain surfaces that protect newly formed water in cooler clouds. Around Sgr A*, that radiation field is unusually intense because hot stars in the central parsec and the black hole’s own outflows pile ultraviolet photons on top of one another. The MICONIC result shows two things at once: the silicate dust itself has survived, which it would not have done if the radiation had sputtered the grains apart, and water has formed and persisted inside that dust. The standard interpretation is that the water molecules are locked onto grain surfaces in the cooler outer regions of the envelope, the same way water ice is locked onto silicate grains in dense molecular clouds in the solar neighbourhood. The fact that those ices survive this close to a supermassive black hole is the result that makes the paper more than a detection.

The takeaway for galactic evolution is the part that will move through the field. If IRS 3 is doing this, other AGB stars in the same neighbourhood almost certainly are, and the central parsec of the Milky Way likely contains a steady drizzle of newly formed silicate and icy grain material rather than the chemically barren zone some models had assumed. That material feeds the mini-spiral of gas and dust that orbits Sgr A*, the reservoir that has been indirectly imaged by the Event Horizon Telescope in 2022 and by GRAVITY at the Very Large Telescope in subsequent years. Over millions of years, even a small per-star dust contribution adds up, and the chemistry of the inner galaxy starts to look more like the chemistry of the disk than the chemistry of a void. The MICONIC programme is extending the same MIRI observations to other evolved stars in the inner parsec; the answer to how many more IRS 3s are out there is the next observational push, and the TechTimes coverage traces the working assumption that the survey will turn up several similar cases.

What we have learned is that proximity to a supermassive black hole is not, by itself, a death sentence for water or for silicate dust in evolved-star envelopes, and the next step is observing the same kind of spectroscopy on the next dozen AGB candidates inside the central parsec to see whether IRS 3 is representative or unusually resilient. The next MICONIC data release, plus parallel JWST spectroscopy of the dusty objects orbiting within a few arcseconds of Sgr A*, will pin that down.

NASA announced on Wednesday, September 23, 2026 that its next great space observatory will be PRIMA, the PRobe far-Infrared Mission for Astrophysics, and that the $1.2 billion project has been advanced into Phase B development on a path toward a 2033 launch (NASA news release). PRIMA is the first in a new mission line NASA is calling Probe Explorers, a billion-dollar class that sits between the agency’s smaller Explorer-class satellites (such as SPHEREx and NICER) and its flagship observatories like the newly launched Nancy Grace Roman Space Telescope (SpaceNews). The observatory’s job is to do something neither Roman nor the James Webb Space Telescope was designed to do well: stare through the cold, dusty universe at wavelengths from 24 to 235 micrometers, the far-infrared window where most of the radiant energy from planet birth, black hole accretion, and interstellar dust actually emerges (Caltech announcement).

The reason far-infrared astronomy needs its own flagship is straightforward and a little humbling. Roughly half the energy released since the Big Bang has been absorbed by interstellar dust and re-radiated in the far-infrared, a band that is essentially opaque from the ground and that earlier infrared telescopes have only been able to nibble at. ESA’s Herschel Space Observatory, the last major far-infrared mission, ended in 2013. NASA’s Spitzer, the predecessor JPL ran for Caltech, ran from 2003 to 2020. SPHEREx, launched in March 2025, maps the whole sky in near-infrared colors but is not optimized for the deep, narrow surveys a 1.8-meter cryogenic far-infrared telescope can deliver (IPAC project page). JWST, for its part, runs from 0.6 to 28 micrometers and is fundamentally a near-and-mid-infrared observatory, leaving the long-wavelength half of the infrared spectrum to PRIMA (Live Science). The Astro 2020 Decadal Survey, led by Caltech physicist Fiona Harrison, asked NASA to fill that gap, and the question of how Earth’s water arrived is one of the targets the survey explicitly named as ripe for a far-infrared answer (NASA news release).

PRIMA is being designed to be, in Caltech President Ray Jayawardhana’s words, “orders of magnitude more sensitive than previous far-infrared space missions” (Caltech announcement). The trick is the same one Herschel and Spitzer used, but pushed harder: a 1.8-meter primary mirror and a detector array cooled to about 4.5 kelvin, so that thermal noise from the telescope itself stops drowning out the faint signals from distant galaxies and cold dust. At the back of the optics sit two science instruments. FIRESS, the Far-Infrared Enhanced Survey Spectrometer, is a four-band grating spectrometer covering the full 24 to 235 micrometer range, with a Fourier Transform Module that can be inserted to push spectral resolution as high as 20,000 (PRIMA instrument page). PRIMAger, the imager, splits into two sub-instruments: the Hyperspectral Imager (PHI), which scans 24 to 84 micrometers at a spectral resolution of about 8, and the Polarimetric Imager (PPI), which uses four filters from 92 to 235 micrometers to measure the I, Q, and U Stokes parameters of incoming light, the first dedicated far-infrared polarimeter ever flown on a NASA mission. Both instruments read out through superconducting kinetic inductance detectors (KIDs), a technology Jonas Zmuidzinas at Caltech and Rick LeDuc at JPL first sketched out in 1999 over coffee near campus and that has spent the intervening decades maturing in ground-based and balloon-borne instruments (Caltech announcement).

What those instruments let PRIMA actually do falls into three themes the 2020 Decadal Survey highlighted. First, the chemistry of planet formation. Far-infrared spectroscopy captures the cold molecular fingerprints (HD rotational lines, water vapor bands, OH, neutral carbon, nitrogen, oxygen, neon) that trace the disk midplanes where gas giants and ice giants are still accreting, the wavelengths JWST’s mid-infrared instruments struggle to reach from inside dust cocoons (NTRS PRIMA concept paper). Second, the co-evolution of galaxies and their central supermassive black holes. Far-infrared dust continuum tracks the obscured side of accretion that X-ray and optical surveys miss, and the new PRIMA Hyperspectral Imager is built to map dust temperature and star-formation rate across cosmic time in single pointed observations. Third, the build-up of dust and heavy elements from the first galaxies to today, the link between stellar nucleosynthesis and the cold universe that JWST and ALMA have been hinting at but cannot survey with the same speed (NASA news release). Twenty-five percent of the five-year prime mission goes to the PI-led science program; the remaining seventy-five percent is open to the broader community through a General Observer call, modeled on the same time-allocation structure Hubble and JWST use today (IPAC project page). The engineering challenges are also unusual for a Probe-class mission: holding a 1.8-meter telescope and two large instruments near 4.5 K for five years, in a Sun-Earth L2 halo orbit similar to Roman’s and JWST’s, requires a multi-stage mechanical cooler chain with no flight-qualified NASA heritage at that scale. The mission also leans on international partners more heavily than most Probe-class concepts typically do: CNES (France) is co-leading PRIMAger with the Laboratoire d’Astrophysique de Marseille and CEA, SRON (Netherlands) is co-developing both instruments, the Max-Planck Institute in Heidelberg built the beam-steering mirrors, and the Canadian Space Agency, DLR, ASI, KASI and KASA, JAXA, and the UK Space Agency all contributed to the Phase A studies (NASA news release). NASA has not yet chosen a launch vehicle, and the $1.2 billion cost cap explicitly excludes launch and other non-project costs, which is why the total mission price tag will be higher than the headline number.

PRIMA’s two-year Phase B is now the gating step. The team has to lock the optical design, demonstrate the cryocooler chain to NASA review board standards, and pass the confirmation review that lets the project enter Phase C implementation, with the same kind of cost-and-schedule scrutiny that pushed the Roman mission through development (SpaceNews). If everything lines up, a 2033 launch puts PRIMA into operation in the mid-2030s, after Roman’s prime five-year survey ends and as JWST enters its second decade of fuel-limited life. The selection also marks the formal opening of the Probe Explorer line, with the next call for proposals expected within the next two NASA budget cycles (NASA news release). In the words of Astrophysics Division director Shawn Domagal-Goldman, the goal is “a pipeline that will consistently have missions of this caliber ready to go,” one Probe Explorer per decade, each opening a wavelength band the current fleet cannot reach. The choice to start that pipeline with the far-infrared is not accidental. Roughly half the photons the universe has ever produced live in that band, and PRIMA is the first NASA mission in more than a decade built specifically to count them.

At 06:34 UTC on August 5, 2026, a hollow aluminum cylinder weighing roughly 3,900 kilograms crossed the last few hundred kilometers between Earth and the Moon at 2.43 km/s and stopped existing. The spent second stage of a SpaceX Falcon 9, cataloged as 2025-010D, slammed into the lunar surface near Einstein crater on the western limb and excavated a crater somewhere between 17 and 30 meters across. Almost no one on Earth saw it. The flash lasted less than a second, the impact site was on the sunlit side of the Moon during morning twilight for European and American observers, and the Sun-corona glare swamped any amateur telescope hoping to catch a thermal signature.

Eleven days later, the picture is clearer. South Korea’s Danuri orbiter published the first before-and-after images of the crater. Astronomers at Boston University working with the European Southern Observatory’s Very Large Telescope in Chile pulled a sodium and lithium line out of the ejecta plume. NASA’s Lunar Reconnaissance Orbiter is planning a follow-up pass that should image the fresh crater at half-meter resolution. For the first time in lunar science, the community has a controlled, predicted, multi-instrument observation of an artificial impact on the Moon.

The Chang’e 5-T1 third stage hit the Moon in March 2022 and the community was surprised by the double crater it left, because the shape implied more mass at the front of the rocket than the published payload would account for. The Falcon 9 stage is the second confirmed unintentional lunar impact, and the first that was on the near side, on the sunlit side, and watched in advance. About 100 lunar missions are planned over the coming decade, and each one leaves a spent upper stage in cislunar space unless the operator deliberately disposes it. SpaceX’s pre-impact statement was honest: on a translunar injection, almost all the propellant goes to the payload, and a controlled deorbit burn is not always possible. Solar radiation pressure and Earth-Moon perturbations then nudge the stage for months or years. The Falcon 9 stage launched in January 2025 carrying Firefly’s Blue Ghost-1 and ispace’s Hakuto-R Resilience; it took nineteen months for the Moon to catch it.

Most of the cameras were not pointed at the right pixel at the right fraction of a second. Jose Maria Madiedo at the Astrophysics Institute of Andalusia in Spain did publish ground-based video the next day, showing a faint plume rising above the lunar limb, but the impact point was hidden behind the Moon’s edge from his vantage. The VLT sodium and lithium detection came from Carl Schmidt, a Boston University planetary astronomer whose thesis work was on Mercury’s sodium exosphere. He had the right high-resolution spectrograph on the right telescope at the right time, and he caught a falloff of sodium and lithium emission extending tens of kilometers above the impact for five to ten minutes. The lithium is the smoking gun: the upper stage’s aluminum-lithium alloy tankage is the only plausible source.

The more satisfying story is Danuri’s. The Korean orbiter made a deliberate maneuver in mid-July to clear a worrying close approach with the incoming stage, then returned to its nominal orbit in time to take a pre-impact image at 3:01 p.m. KST on August 5, thirty-three minutes before the strike. Seven more imaging sessions followed over the next fifteen hours. The before-and-after comparison, released through the Korea Aerospace Administration on August 12, shows an unambiguous change in surface texture around the impact coordinates and a darker patch of ejecta extending some distance from the new crater. The LUTI camera has a resolution of a few meters per pixel, and the new crater is roughly the size of a single pixel or slightly larger.

The finer imaging will come from LRO. Mark Robinson, the LRO Camera principal investigator, told Inside Outer Space that the orbiter’s clean pass over the impact region on August 11 closed the acquisition cycle, with the crater in view at the LROC Narrow Angle Camera’s native 50-centimeter resolution. The high-resolution LROC imagery is expected to be published once the data is processed.

The science starts with the energy budget. A 3,900-kilogram object hitting the Moon at 2.43 km/s carries KE = 0.5 * m * v^2 ≈ 11.8 gigajoules of kinetic energy, equivalent to roughly 2.8 tons of TNT. By the standard point-source crater scaling for competent regolith, this corresponds to a transient crater about 17 meters in diameter and a final crater in the 20-30 m range once the walls have slumped, matching the predictions published by Fernando and colleagues (arXiv:2607.14625) and by Jo and colleagues at UT Austin (arXiv:2607.23904). The latter paper predicted an ejecta curtain extending 15-20 km above the surface, a central spike reaching 75-100 km, and a lateral spread of about 183 km. The central spike is the part that briefly exceeded the lunar escape velocity and would have scattered fine dust into cislunar space.

The interesting engineering question is what the hollow impactor did to the standard crater-scaling law. A spent Falcon 9 stage is mostly an empty propellant tank with a single Merlin Vacuum engine hanging off one end. Schmidt’s lithium detection is direct evidence that the stage broke up during entry rather than punching through as a coherent cylinder, because the lithium is locked inside the aluminum-lithium alloy of the tank bulkhead and would have vaporized before reaching the surface if the stage had stayed intact. The Chang’e 5-T1 double crater in 2022 told a similar story: a hollow rocket body, a mass distribution front-to-back that nobody had published, and a crater shape that surprised the modellers. Artificial impactors are not point masses.

The cislunar tracking problem is the bigger engineering gap. Ground-based optical and radar observations lose accuracy past about 0.1 AU because the radar return on a tumbling aluminum cylinder is weak and the optical signal is dominated by the Earth’s background. The Falcon 9 stage’s final weeks were tracked primarily by amateur astronomers using the Bill Gray Project Pluto pipeline and a handful of university observatories. The successful prediction and observation of this impact is a useful stress test of that network, and it came through. The next test, with about a hundred more lunar missions in the queue, will not be optional.

The Moon now has a second confirmed artificial crater, a fresh cislunar debris benchmark, and the first multi-instrument dataset of its kind. The VLT sodium and lithium detection is the cleanest engineering fingerprint of an artificial impact ever recorded, and Danuri’s images give crater-scaling models a calibration point they have not had before. LRO’s high-resolution pass on August 11 is expected to close the loop once the imagery is processed.

The less comfortable lesson is that cislunar traffic management is still a volunteer effort. The Falcon 9 stage was tracked, modeled, and observed by a coalition of amateur astronomers, two national space agencies, and academic research groups. There is no equivalent of the FAA for the volume of space between Earth and the Moon. As lunar launch cadence rises, that will have to change, because the next artificial impact will probably not be a quiet one with a known stage and a predicted trajectory. It will be a piece of hardware nobody realizes has drifted into the wrong orbit, and the only way to notice it in time is to be looking.

On 17 March 2026, somewhere past the orbit of Mars, a European spacecraft fired its hydrazine thrusters for the second time since launch. The burn lasted long enough to dump 123 kilograms of propellant and shift the probe’s heliocentric velocity by 367 metres per second, the equivalent of accelerating from a standstill to just above the speed of sound (ESA, Hera mission update, 17 March 2026). Two months later, in Thessaloniki, the entire Hera science team gathered for what the project’s own newsletter called “the last time before the probe finally arrives on site” (heramission.space, 17 May 2026). That arrival is now eleven weeks out.

ESA’s Hera, launched 7 October 2024 from Cape Canaveral on a Falcon 9, is closing on the binary asteroid 65803 Didymos. It will reach the system in early November 2026, almost exactly four years after NASA’s Double Asteroid Redirection Test slammed into Dimorphos, the smaller of the two bodies, and changed its orbit by 33 minutes (NASA/APL, 26 September 2022). What Hera will do once it gets there is, in a phrase the mission team keeps using, turn a grand-scale experiment into a repeatable planetary-defence technique. That phrase is doing a lot of work. It is the entire reason a 1,214-kilogram spacecraft is at this moment drifting toward two rocks that, until 2022, almost no one outside planetary science had heard of.

Dimorphos is small. Before DART, its diameter was estimated at roughly 160 metres, the height of a 50-storey building, and its mass was known only to within a factor of two. That uncertainty is the central problem Hera was built to solve. The first measurement of how much momentum DART delivered to the moonlet, the so-called momentum enhancement factor beta, was published by Rivkin and colleagues in 2023. They reported beta = 3.61 with a 1-sigma uncertainty of about 0.2, anchored to an assumed Dimorphos bulk density of 2,400 kilograms per cubic metre (Rivkin et al., Nature 2023). A weaker, more porous asteroid gives a higher beta. A denser, more cohesive one gives a lower beta. Without measuring Dimorphos’s actual mass, every planetary-defence scenario that uses the DART result inherits that uncertainty.

That matters because planetary defence is no longer hypothetical. The 2024 PDC25 exercise at Johns Hopkins Applied Physics Laboratory used beta = 3.61 as its baseline. The number is now embedded in the operating procedures of NASA’s Planetary Defense Coordination Office. A 0.2-point error in beta, applied to a 200-metre asteroid on a 10-year warning, translates into roughly a thousand kilometres of miss distance. That is the difference between a near miss and an ocean impact.

There is a quieter reason to care. Hera is going to make the first internal radar scan ever performed on an asteroid. The instrument is called JuRa, and it sits on a CubeSat the size of a shoebox.

NASA’s DART was the kinetic-impact end of an experiment called AIDA, the Asteroid Impact and Deflection Assessment. ESA’s Hera was always supposed to be the diagnostic half. The architecture is not unusual in physics: you hit something, then you measure what happened. In this case, the impactor flew in 2022. The diagnostician is arriving on time.

Dimorphos is in a 12-hour orbit around Didymos, a roughly 780-metre primary. Both are near-Earth asteroids in the rubble-pile class. DART hit Dimorphos head-on at about 6.6 kilometres per second and shortened its orbital period by 33 minutes, far more than the 7-minute change the spacecraft alone would have produced. The extra came from ejecta, the plume that flew off the surface in the opposite direction and acted as a natural afterburner.

Hera will do the part ground telescopes could not. It will map Dimorphos’s shape from hyperbolic arcs at 20 to 30 kilometres, then closer, down to 1 to 2 kilometres during the experimental phase in April and May 2027 (ESA Hera operations plan, 2026). Two CubeSats will be deployed starting 13 January 2027. Juventas carries JuRa, a 50 to 70 megahertz radar that will sound the asteroid’s interior to a depth of roughly 100 metres with 10 to 15 metres of vertical resolution. Milani carries ASPECT, a hyperspectral imager covering 0.5 to 2.5 micrometres, and VISTA, a dust sensor. Juventas will attempt a soft landing on Dimorphos in late January 2027, the first controlled touchdown on a body this small.

The 367 m/s maneuver tells you what the mission was actually doing during cruise. The full delta-v budget from launch through arrival is on the order of 1.4 kilometres per second, and Hera carries roughly 600 kilograms of hydrazine for chemical propulsion. That is a small budget by interplanetary standards, smaller than OSIRIS-REx’s, smaller than Hayabusa2’s. The reason it works is the Mars gravity assist of 12 March 2025, which bent the trajectory and added, for free, several hundred metres per second. Without that flyby, Hera would not have arrived.

What the chemical system has to do is trim. Each burn is small. The cumulative effect, over five maneuvers across two years, is to align the spacecraft with Didymos’s slow drift around the Sun so that, in late October 2026, the probe can begin hyperbolic approach arcs instead of flying past.

The autonomy stack is the harder problem. At 20 kilometres, Didymos subtends roughly the same angle as a person standing 100 metres away. At 4 kilometres, Dimorphos, at 160 metres across, occupies about 2.3 degrees of sky. Both bodies move visibly against the star background during a single exposure. The Asteroid Framing Camera, two redundant 1020-by-1020 CMOS sensors built by Jena-Optronik, has to track Didymos in one field of view while Dimorphos arcs through the same frame. At 1 kilometre altitude, AFC delivers roughly 10 centimetres per pixel. That is enough to resolve the rim of the DART crater and build a digital terrain model accurate to about half a metre (Jena-Optronik AFC documentation, 2024).

JuRa is the more exotic instrument. Low-frequency radar in deep space is rare because the antenna that works at 50 to 70 megahertz is, by radio-physics standards, large: on Juventas, the antennas deploy to about 1.5 metres tip-to-tip. The bandwidth is wide enough to see the boundary between regolith and a more competent interior, if such a boundary exists. Dimorphos is thought to be a rubble pile, a gravitationally bound aggregate of fragments from centimetres to tens of metres, with bulk porosity between 30 and 50 percent. Until now, no one has measured which kind of object DART actually hit.

The momentum transfer itself is described by an equation that hides a great deal of physics. The change in Dimorphos’s orbital velocity is

Δv = (m_s × v_s + m_e × v_e) / (m_s + m_e + m_t)

For DART, m_s was about 580 kilograms, v_s about 6.6 kilometres per second, giving a momentum of about 3.8 megagram-metres per second. The orbital change was 2.7 millimetres per second, implying an effective momentum delivery roughly 3.6 times larger than the spacecraft alone would have produced. That factor, beta, is what Hera will nail down by weighing Dimorphos.

Hera will not prevent an asteroid impact. That is not the job. The job is to convert the most famous kinetic impact experiment in history from a single dramatic event into a calibrated engineering method. By mid-2027, planetary-defence planners will know, to within a few percent, how much velocity change a given kinetic impactor delivers to a rubble-pile asteroid of a given size and porosity, from direct measurement rather than extrapolation.

The November 2026 arrival is the moment the field stops being theoretical.