Interstellar comet 3I/ATLAS sheds metals unevenly on its journey past the Sun

Following the metals outward

Damien Hutsemékers and colleagues studied the cloud of gas and dust surrounding 3I/ATLAS, known as its coma, using the UV-Visual Echelle Spectrograph on the European Southern Observatory’s Very Large Telescope. Their new observations ran from December 4, 2025, to February 22, 2026, as the comet moved outward from 1.88 to 4.35 astronomical units from the Sun. One astronomical unit is approximately the Earth–Sun distance.

Combined with revised measurements from the inward journey, the spectra reveal how neutral nickel and iron changed around perihelion, the comet’s closest solar approach. The labels NiI and FeI denote electrically neutral atoms, not ions or solid grains.

Iron emission lines were last detected at 3.50 astronomical units and were not detected at 3.85 or in subsequent spectra. This does not establish that iron was physically absent. Nickel remained measurable at 3.85; at 4.35, only two barely detectable lines remained, insufficient for a reported production rate.

A different outward journey

At comparable solar distances, inferred metal production rates were higher after perihelion and declined more gradually as the comet receded. The difference was particularly strong for iron: around 2.5 astronomical units, its outward production rate was approximately ten times its inward rate.

Near two astronomical units, the combined nickel and iron production rate was at least an order of magnitude above that of typical Solar System comets in the authors’ comparison. This concerns atoms entering the coma, not the comet’s total metal inventory or its original planetary system’s composition.

Nickel exceeded iron throughout the new observations where both could be measured. The unusually high nickel-to-iron ratio seen early on the inward journey had fallen toward values found in Solar System comets near two astronomical units. After perihelion, the ratio changed less strongly with distance. A separate observation incorporated into the analysis, closer to perihelion at 1.51 astronomical units, showed slightly more iron than nickel.

Turning light into release rates

The researchers compared individual emission-line brightnesses with a multilevel fluorescence model. This calculates how atoms absorb and re-emit sunlight, accounting for their energy levels and the comet’s motion relative to the Sun. Earlier measurements were revised using the same atomic-transition list, including infrared transitions that improve the nickel calculations.

Converting inferred numbers of atoms into production rates required an assumed ejection speed. If the atoms travel at half the adopted speed, the calculated rates would also be halved. Absolute flux calibration adds a systematic uncertainty of approximately 0.1 dex, a multiplicative factor of about 1.26. Differences found against an independent fluorescence model were also included in the error budget.

The nickel-to-iron ratio is less sensitive to shared calibration and velocity assumptions. Nevertheless, atomic ejection speeds and their variation with solar distance remain uncertain. Spatial brightness profiles indicate a compact source near the nucleus rather than predominantly extended production across the coma.

Testing a chemical explanation

The proposed parents are nickel tetracarbonyl, Ni(CO)4, and iron pentacarbonyl, Fe(CO)5. These molecules could pass directly from ice into gas at temperatures far below those needed to vaporize ordinary metal-bearing grains. The model assumes sunlight then rapidly breaks them apart, releasing neutral atoms.

Nickel carbonyl is more volatile than iron carbonyl. Within this scenario, that difference and the relative effective areas releasing the compounds can explain nickel’s earlier appearance and the nickel-rich gas without requiring an extraordinary bulk nickel abundance.

A model with adjustable temperature profiles reproduced the observations. Required temperatures generally varied more slowly with solar distance than the simple equilibrium relation, in which temperature falls with the inverse square root of distance. Temperature and effective emitting area can compensate for each other, however, so successful fits do not uniquely determine physical conditions or carbonyl abundances.

A plausible mechanism, not a detection

Using a more detailed nucleus thermal model required substantial empirical changes: a temperature ceiling around 200–210 kelvin near perihelion, sharply reduced active areas through perihelion, and transient warming to reproduce the early nickel excess. The added warming peaked at 13 kelvin when a layer’s temperature was around 120 kelvin and was applied deeper than five centimeters.

The authors explicitly call these modifications ad hoc. Crystallization of amorphous water ice could supply additional heat, but was not established as its source. Explanations for the temperature ceiling and active-area reduction are also speculative.

Subsurface carbonyl sublimation is an interpretation within this modified model, not an observed layer. It could suggest earlier moderate heating that depleted volatile material near the surface, but that inference depends on validating the mechanism.

Carbonyls remain undetected in cometary comae, and their formation in cometary ices is poorly understood. Alternatives, including release from superheated metallic nanoparticles, remain viable. Targeted searches for carbonyl signatures around 5 and 18 micrometers, laboratory tests, and combined thermal and compositional modeling could help distinguish the mechanisms.

Publication status and source limits

The arXiv version 2 record, revised November 6, 2026, states that the study is accepted for publication in Astronomy & Astrophysics. The original arXiv submission was May 8, 2026; neither date establishes final journal publication. This account uses the supplied landing record and complete paper-text extraction, which omits figure images and may distort equations, tables, and layout.