SPHEREx and JWST reveal what comet 3I/ATLAS is… and isn’t

When it comes to finding out about the Universe in a scientific fashion, most of us think about performing experiments here on Earth under laboratory controlled conditions. That’s all well and good for many different scientific fields — experimental physics, chemistry, biology, etc. — but for astronomy and astrophysics, our laboratory is the depths of space, and whatever the Universe happens to show us while we’re looking at it. For most of human history, the only objects we ever discovered in our own Solar System were native objects: objects that are and always were members of our Solar System, just like we are.

That changed in 2017, when interstellar interloper ‘Oumuamua arrived: our first known object within the Solar System to originate from outside of it. Then, 2019 brought with it another interstellar object: Borisov, which was larger and appeared much more like a conventional comet than ‘Oumuamua did. Starting in early July of 2025, we got our first glimpse of humanity’s third identified interstellar interloper: 3I/ATLAS, which initially appeared to be larger, brighter, and faster-moving than either ‘Oumuamua or Borisov was. As this new object speeds towards perihelion, it isn’t just ground-based telescopes that have taken a look at it now, but space-based ones as well. Here’s what we know about it today, with plans to find out so much more over the next few months.

A team of scientists used the TTT3 (Two-meter Twin Telescope) to image interstellar object 3I/ATLAS from July 02 to July 03, 2025, revealing the extent of the object, its central nucleus, and strong hints of outgassing activity.

Back in early July, the best images we were getting of 3I/ATLAS looked like the image you see above: faint, blurred views of what appeared to be more than a point-like object, but rather an extended source surrounded by a halo-like appearance. With a source such as this, you can still measure the object’s velocity and position very well, enabling you to determine its orbital trajectory both past and present: including a very accurate determination of its origin. However, its size cannot be well-determined, as there’s no observed delineation between the central nucleus and the diffuse coma.

An object whose eccentricity is less than 1.0 is gravitationally bound to our Sun; an object with an eccentricity greater than 1.0 makes a hyperbolic orbit, and will exit the Solar System. While gravitational kicks from massive objects, like Jupiter or Neptune, might enable eccentricities up to 1.05 or so; any value beyond that indicates an interstellar origin for such an object. As of July, we knew of three interstellar objects, with the following stats:

• ‘Oumuamua was first, with a size of about 100 meters, an eccentricity of 1.119, and an entrance speed of 26 km/s with respect to the Solar System.
• Borisov was second, with a size of just under 1 kilometer, an eccentricity of 3.36, and an entrance speed of 32 km/s with respect to the Solar System,
• and 3I/ATLAS, the third interstellar object, was estimated to be large (between 10-30 kilometers) but with huge uncertainties, an eccentricity of an unprecedented 6.2, and an entrance speed of 58 km/s with respect to the Solar System, the fastest ever.

This diagram shows the trajectory of interstellar comet 3I/ATLAS as it passes through the Solar System. It will make its closest approach to the Sun in October: when Earth is on the opposite side of the Sun from the object but where it’s relatively close to the location of planet Mars.

That very fast speed, coupled with the measured trajectory of 3I/ATLAS, has enabled astronomers to pinpoint the most likely origin of this object: from the thick disk of the plane of the Milky Way, which just happens to be the location where the greatest numbers of new stars and star systems form in our galaxy. Stars and other objects in our galaxy generally move rapidly with respect to one another, with relative speeds of 20-30 km/s being typical for objects in the same location in our galaxy and even greater relative speeds — up to a couple of hundred km/s — for objects separated by thousands of light-years. Even though 3I/ATLAS is in the same location as our Sun, today, it almost certainly didn’t originate nearby.

Although 3I/ATLAS is still getting closer to the Sun, it’s actually moving away from the Earth right now. When this interstellar object reaches perihelion, its closest approach to the Sun, which it will do near Halloween, Earth will actually be on the opposite side of the Sun from the comet, making observations impossible. That means we’re getting our best views of it now, while there’s still a good line-of-sight to it that doesn’t pass near the Sun, and won’t get good views of it again until November or December. While most of the images and studies that have come out of 3I/ATLAS have relied on ground-based observatories, there are three space-based observatories that have taken a look at it: Hubble, SPHEREx, and JWST.

This Hubble image, from July 21, 2025, shows the interstellar comet 3I/ATLAS, with the telescope tracked on the (moving) comet to keep it stationary in-frame, while the stars in the background blur behind it. Hubble’s imaging reveals a relatively small size for the central nucleus, with a diameter no greater than 5.5 kilometers, while revealing an enormous extend to the coma, or halo, surrounding it.

Above, you can see the Hubble image of 3I/ATLAS, which was taken on July 21, 2025: when the comet was still 365 million kilometers away from Earth, or 2.4 times as far away as the Earth-Sun distance. You can clearly see that there’s a bright nucleus that isn’t completely resolved (i.e., you can’t tell where the coma begins and the nucleus ends) embedded within a larger, diffuse coma, or halo-like puffy cloud that surrounds the nucleus. You can see strong evidence for this coma, but no evidence for either a dust tail or an ion tail, which are normally features common to comets found in our own Solar System.

From the Hubble data, which has much, much better resolution than either SPHEREx or any ground-based observatory that can look at 3I/ATLAS, you can conclude that the nucleus, at most, is about 2.7 kilometers in radius: comparable to the radius of the asteroid that is suspected to have extincted the non-avian dinosaurs on Earth some 65 million years ago.

Notably, this measured size is much smaller than earlier estimates of 10-30 km, because those estimates couldn’t differentiate between the nucleus and the coma at all; they often assumed that the brightness would be wholly or significantly impacted by reflected sunlight off of the nucleus. With these Hubble observations, we now know that’s not the case; we have a large, bright coma that’s responsible for most of the light coming from 3I/ATLAS, and that although there is a nucleus present, it’s relatively small.

These flux map show the spectrally integrated flux for 3I/ATLAS observed with JWST’s NIRSpec instrument, showing scattered light from coma dust at 1.2 microns (panel a), carbon dioxide at 4.3 microns (panel b), water at 2.7 microns (panel c), and carbon monoxide at 4.7 microns (panel d). Note the weak, consolidated nature of the water and carbon monoxide, compared to the bright, diffuse carbon dioxide signal.

The JWST data was just released for everyone to look at, and the full preprint has just been made available to all. What happened was that JWST observed 3I/ATLAS back on August 6th of 2025, acquiring spectra of the object with the NIRSpec Integral Field Unit, which can provide spatially resolved imaging spectroscopy over an area that’s 3 arc-seconds by 3 arc-seconds in size, with 0.1 arc-second resolution in each dimension. Because that data was only taken with a spectrograph, not with an imager, we can tell what this object is composed of as far as molecules that are exhibiting emissions, but we can’t use it to determine the size of the comet’s nucleus any better than Hubble could.

Just a few days before the JWST preprint came out, however, a different space telescope’s collaboration beat them to the punch: SPHEREx released their imaging and spectroscopy data. SPHEREx, whose architecture is shown below, is a near-infrared telescope that’s significantly smaller than JWST and that has much worse resolution than even Hubble (by about a factor of 10), but which can perform spectroscopy out to infrared wavelengths that Hubble isn’t sensitive to. That means that, unlike Hubble or most optical (or ultraviolet) observatories, SPHEREx can be sensitive to the signatures produced by different volatile, ice-rich compounds. In particular, it can distinguish between water-ice, carbon dioxide-ice, and carbon monoxide-ice with the wavelength capabilities that it possesses.

This image shows the SPHEREx observatory and its nested metallic cones, which passively radiate excess heat back out into space. SphereX will be at a frosty 45 K as far as its operating temperature goes, which is why it required being launched into space to be successful.

This is pretty remarkable, because SPHEREx wasn’t really designed to be an imager of Solar System objects; it was designed, as you can see above, to be a wide-field imager that can take spectroscopic data of a wide variety of light sources — i.e., galaxies — located at an enormous variety of distances from us. SPHEREx is designed to map out the large-scale structure of the Universe, and in particular to see how galaxies are clustered across cosmic time. From a cosmic perspective, we’re trying to measure things like:

• how clumpy or clustered is the Universe,
• how likely are you, if you find a galaxy, to find another galaxy located a certain distance away from it,
• how much clustering power exists on a variety of cosmic scales, both large and small,
• and how does that clumpiness/clustering power evolve as the Universe expands.

With the power of SPHEREx, augmented by observatories that perform similar tasks like DESI, Euclid, Vera C. Rubin, and the future Nancy Roman Telescope, we’re hoping to see how clustering, large-scale structure, and inferred properties like dark matter and dark energy evolve over cosmic time. Those are the main science goals of SPHEREx, and it’s well on its way to uncovering data-driven answers to these existential cosmic questions.

But with its infrared, spectroscopic capabilities, SPHEREx is also capable of going beyond its main science mission, and can perform these critical measurements even for nearby objects. Fortunately, From August 8 to August 12 of 2025, SPHEREx did exactly that, and now a paper showcasing SPHEREx’s first science results is out. It’s not about cosmology at all, but rather about 3I/ATLAS, with fascinating implications for anyone who’s curious about its nature. (And the question of whether it’s alien-related or not.)

Using a stacking technique, scientists working on the SPHEREx mission were able to produce a deep image of the nucleus plus the scattered light by dust of the flux from 3I/ATLAS. There is no significant extension beyond a single point-source identified at these short wavelengths by the SPHEREx data.

Above is the kind of image that you might have expected: one taken at short wavelengths of 3I/ATLAS. Over a wavelength range from 0.75 to about 1.5 microns, SPHEREx saw something that was indistinguishable from a single point. This tells us that there’s no obvious jet or tail structures coming from 3I/ATLAS, and that the nucleus is compact: as compact as SPHEREx can resolve. (Which isn’t very good, honestly, at about a radius of ~23 km; the Hubble limits of ~2.7 km in radius for the nucleus is superior to anything SPHEREx can tell us.) In these short wavelengths of light, the object looks like a point, which indicates that it’s largely composed of some type of reflective ices.

However, at longer wavelengths, SPHEREx can still perform imaging and spectroscopy, and those longer wavelengths can distinguish between the three main forms of ice found on comets and comet-like objects:

• water-ice, or H2O, which exhibits strong absorption signatures between 2.6 and 3.5 microns along with water-gas emissions at 2.7 and 3.0 microns,
• carbon dioxide ice, or CO2, which exhibits strong absorption signatures between 4.1 and 4.4 microns along with CO2-gas emissions at 4.25 and 4.27 microns,
• and carbon monoxide ice, or CO, which would exhibit strong absorption signatures between 4.6 and 4.8 microns along with CO-gas emissions at 4.7 microns.

SPHEREx is kind of remarkable for this, as it’s optimized to find extended ice and gas structures on the sky, with the explicit capabilities to distinguish between CO2 and CO: carbon dioxide, which remains solid up to temperatures of 90 K or upwards, and carbon monoxide, which sublimates away into the gaseous phase at much lower temperatures.

As you can see, below, there is no detection of features that would be associated with sublimating water-ice; there is a non-detection of H2O in the coma of this object. Similarly, where a carbon monoxide signature could/should be, there’s also nothing. But where a carbon monoxide signature should be, you can see approximately 100% of the flux from 3I/ATLAS.

These three side-by-side graphs show the gas coma of 3I/ATLAS as imaged in three different wavelength ranges: one of which is sensitive to H2O (at left), one sensitive to CO2 (center), and one sensitive to carbon monoxide (right). The signatures for water and carbon monoxide are non-existent, but the signature from carbon dioxide is remarkably strong.

This is consistent with an interpretation where 99% or more of the measured continuum flux, at least from SPHEREx, is arising from coma dust, and where that coma dust is consistent with being made of CO2 ices that appear in relatively large fragments. This is really interesting, because it allows us to infer some facts about 3I/ATLAS, which itself is just the third object ever detected within our own Solar System to not arise from our own Solar System.

‘Oumuamua was very small, and offered no hint of CO, CO2, or H2O. It was almost as though all of the volatile, easily sublimated-away ices were already gone from it. On the other hand, Borisov was much larger than ‘Oumuamua, and resembled a very young comet from our own Kuiper belt; it was volatile rich, and was even particularly rich in CO. Meanwhile, for 3I/ATLAS, which is still approaching perihelion (getting closer to the Sun), we find no carbon monoxide and no signs of water-ice, but it’s very rich in CO2, which it’s presently offgassing.

Now, consider this: carbon monoxide, carbon dioxide, and water-ice are the three “big” ices found in Solar System comets, making up around 99% of comet ices. (Others include methane ice and nitrogen ice.)

• It only takes low temperatures to evaporate off carbon monoxide, nitrogen, and methane ice; 40K will easily do it, which means that even for objects located beyond the orbit of Neptune, these ices can be sublimated away.
• Water-ice is very difficult to boil or sublimate away, requiring temperatures at hundreds of kelvin. Objects need to be much closer than Jupiter to begin evaporating/sublimating water-ice away: at about one-and-a-half times the distance to Mars from the Sun.
• And carbon dioxide is in the middle: requiring temperatures of ~100 K to sublimate away, or distances from the Sun comparable to the planet Saturn.

This image shows the directly imaged cometary nuclei of Comets Tempel 1 and Hartley 2, as imaged in 2005 and 2010, respectively, by NASA’s Deep Impact spacecraft. Hartley 2 is an old comet that is still fragmenting and emitting chunks of CO2, but has run out of the light ices nitrogen, methane, and carbon monoxide.

Above, you can see two objects that were explored up close from the Deep Impact spacecraft: Tempel 1, which was visited in 2005, and 103P/Hartley 2, which was visited in 2010. (Deep Impact ended its mission in 2013.) The reason it’s important to look at these images is because Comet 103P/Hartley 2 was observed to have precisely the same properties that 3I/ATLAS displays. Hartley 2 showed lots of carbon dioxide, no carbon monoxide and water only when it was sufficiently close to the Sun, and was fragmenting and spitting out large chunks of itself: pristine material that showed it originated from our own Kuiper belt. Those large chunks, of centimeter-to-decimeter sizes, had never been observed before Deep Impact’s rendezvous with Hartley 2.

Why doesn’t Hartley 2 have carbon monoxide? And why don’t all outgassing objects that emit carbon dioxide also show signatures of water?

Because Comet Hartley 2 is old, has spent a lot of time relatively close to the Sun, only emits water-gas when temperatures are high enough to sublimate water-ice, converting it into the gaseous phase. Put simply: Kuiper belt-like objects, like comets, don’t last forever when they either get too hot or travel through space too rapidly; they begin to erode. Continued heating will sublimate away all of the most easily-destroyed ices (methane, carbon monoxide, nitrogen, etc.), so that it’s only the young or more pristine objects that maintain and still exhibit them. The oldest, smallest, most eroded objects might be mostly water-ice, having offgassed all of their other volatiles. But before that stage, there will be objects that have a mix of water-ice and carbon dioxide ice, where only carbon dioxide ices get released at temperatures of between about 90-200 K. That’s what Hartley 2 is, and that’s what the data is very, very strongly hinting that 3I/ATLAS actually is.

This graph shows the combination of SPHEREx data points (with error bars), SPHEREx upper limit constraints (red arrows), IRTF data (green), and the reference spectrum of objects from the Kuiper belt in our own Solar System (pink). As the data quite clearly shows, 3I/ATLAS bears remarkable similarities to Kuiper belt objects, and in particular to older, CO-free Kuiper belt objects.

In fact, using all of the data acquired from SPHEREx (with ancillary data from NASA-IRTF SpeX folded in as well), as you can see above, the spectrum very closely matches the line in pink, which is a combined model for Kuiper belt objects originating from our own Solar System. Those strong emission features, above, that you see in yellow? The SPHEREx team was cautious about them, but that was something that the JWST data, shown below, was very definitive about. In addition to a big spike in carbon dioxide, JWST also made a very sensitive, low-level detection of solid water-ice: precisely what the SPHEREx data suggested but couldn’t make a definitive detection of for a comet-like object still located an impressive 3.2 astronomical units from the Sun.

What’s specifically remarkable about 3I/ATLAS may be its water-ice to carbon dioxide ratio, with a large excess of CO2 (especially that strong gas emission) compared to the amount of water present, suggesting that perhaps this object either had an unusual origin in terms of its composition for a Kuiper belt object, or that it’s more differentiated in terms of composition than a standard Kuiper belt object. Only a small amount of carbon monoxide is present, but unlike SPHEREx, JWST is sensitive enough to see it.

How did it get to be this way? We can’t be sure with these observations alone. After all, as we’re observing it now, we’re seeing 3I/ATLAS as it is after potentially billions of years of evolution and travel through interstellar space, all while we remain relatively ignorant about its origins and ejection/collision history.

This graph shows the JWST NIRSpec spectrum, using the NIRSpec prism with the sky background subtracted, plotted with a logarithmic flux scale. Features related to water (H2O), carbon dioxide (CO2), and carbon monoxide (CO) are labeled and highlighted. Water exists in the ice phase only, carbon dioxide gas emits an extremely strong signature, and a small but present carbon monoxide signature appears.

When we put all of this data together, we can begin comparing the three known interstellar interlopers, ‘Oumuamua, Borisov, and 3I/ATLAS, to one another. Whereas ‘Oumuamua looked very old, like it was in the final stages before evaporating completely, and Borisov looked young, like a fresh Kuiper belt-like object just approaching a star system’s interiors for the first time, 3I/ATLAS, like 103P/Hartley 2, looks like an in-between object. It’s most likely:

• an ice-rich object,
• formed as part of an alien star system,
• that was ejected and hurled into interstellar space,
• where it traveled for at least hundreds of millions of years, and maybe even many billions of years,
• having been eroded so that all of its “light” ices have been boiled away already,
• but where both water-ice and carbon dioxide ice remain,
• where the CO2 ices are sublimating now and breaking off in big chunks (rather than little motes),
• but which won’t begin offgassing water until it either gets hotter or makes an even closer approach to our Sun.

Despite unfounded speculations that 3I/ATLAS could be an alien artifact, the data instead shows that it is alien-like in the sense that it originated from an alien star system in the Milky Way, but naturally: the same way Kuiper belt objects exist in our own Solar System. Moreover, 3I/ATLAS isn’t just comet-like, it’s a special type of comet-like, akin to the evolved comets (like Hartley 2) that we find in our own Solar System: the old, semi-fragmenting short period comets.

As they orbit the Sun, comets and asteroids typically break up over time, with debris between the chunks along the path of the orbit getting stretched out to create debris streams. These streams cause meteor showers when the Earth passes through that debris stream. This image taken by Spitzer along a comet’s path shows small fragments outgassing, but also shows the main debris stream that gives rise to the meteor showers that occur in our Solar System. Note that comets, and comet-like objects, evaporate the most quickly when they’re at their hottest, and when they approach a star at the smallest separation distances.

To be sure, though, while Hartley 2 will likely be gone completely in the next ~1000 years or so, the one close pass of 3I/ATLAS through our Solar System won’t significantly change its composition. After passing by our Sun, it will exit the Solar System, likely encountering many other stars over at least hundreds of millions of years further. There are likely at least sextillions of these objects flying through interstellar space, far outnumbering the stars and planets of our galaxy.

Rather than scream “aliens!” every time we find a new one, we’re instead entering an era where we’re identifying, measuring, and learning about the origins of these leftover materials from the primordial formation of stellar and planetary systems beyond our own. Our imaginations may be limitless, but the very enterprise of science demands that we constrain them to be consistent with both

• the full suite of data that we collect about the Universe
• and the fewest number of unnecessary, extraneous assumptions

in order to explain what we observe. 3I/ATLAS is definitely comet like, but potentially a slightly different comet, with possibly different water-to-carbon dioxide ratios, than the comets of Kuiper belt origin we’re more familiar with.

Most importantly, this isn’t the end of the story, but rather our first steps into categorizing, identifying, and studying objects of interstellar origin that pass through our Solar System. With the dawn of the Vera Rubin Observatory era of astronomy now upon us, we’re bound to detect many more of these interstellar interloper objects in the years to come, improving our understanding of how these objects that permeate the interstellar reaches of our galaxy come to be, evolve, and eventually perish amongst the stars.

The author acknowledges Carey M. Lisse for extremely helpful discussions about SPHEREx data and the nature of 3I/ATLAS.