JWST discovers how we’re able to see the Universe at all

Although it’s shown us the Universe as we’ve never seen it before — deeper, earlier, and at longer wavelengths — it’s important to remember that the main goals of JWST didn’t have anything to do with the quality of pictures it would acquire, but rather for the science questions that they’d reveal the answer to. One of the puzzles was simple: we know that the Universe is full of stars and galaxies today, whose light we can easily see, but that early on, there were no stars and galaxies, and the Universe was instead filled with neutral atoms. How, then, did all of those neutral atoms become ionized once again, enabling us to see the Universe and all the starlight generated within it?

The simple answer, of course, would have to be stars. It must be that the Universe formed stars in sufficient numbers, eventually, to produce enough high-energy (e.g., ultraviolet) light so that all of those once-neutral atoms then became ionized, allowing starlight to pass through space unimpeded. But where were those stars located? What types of galaxies housed them? And when, exactly, did all of those stars form to drive this process, known as reionization?

That was one of the main JWST science goals, and for the three years since it began science operations, we’ve been closing in on the answer. Now, at last, with a new study led by Isak Wold of the UNCOVER collaboration, we’ve got the answer: tiny, modest, but common galaxies are the culprit behind how the Universe became transparent to starlight. Here’s the cosmic story of what happened, and how we know.

This low-resolution image shows the full field of the COSMOS-Web survey conducted with JWST. Spanning 0.54 square degrees in the sky, or nearly three full Moons’ worth of area, this represents the largest, deepest wide-field view of the Universe ever acquired. Space, as we can clearly see here, is full of luminous sources of light in practically all directions.

When most of us think about the distant Universe, we think about images like the one you see above: deep field images, acquired with our most powerful space telescopes, including Hubble and JWST. It appears that what we’re seeing is stars and galaxies everywhere, limited only by the amount of time we spend observing and the capabilities of our instruments. But that’s not quite true; there aren’t stars and galaxies absolutely everywhere, as even with infinite amounts of observing time, we wouldn’t see starlight coming from all regions of space.

• One reason is that stars don’t form everywhere, but only in those regions where enough matter accumulates to gravitationally collapse and trigger the formation of stars. This occurs in a web-like fashion, where more dense regions accrue the matter from their less-dense surroundings.
• Another reason is that we can only see the objects that we can pick up in the wavelength ranges our instruments are designed for. For Hubble, for example, it cannot see light longer than about two microns in wavelength, which means that it isn’t capable of seeing any light from the most distant, early galaxies of all.
• And another very important — but less commonly discussed — limitation is that these early stars will form from neutral atoms, where 100% of the atoms in the Universe become neutral within the first half-million years after the Big Bang. Neutral atoms are incredibly good at absorbing starlight, and so we can only see starlight once there’s a clear path from those stars to our eyes.

This animation shows the Bok globule Barnard 68 in a variety of visible and infrared wavelengths. As the longer wavelengths reveal, this is not a hole in the Universe but simply a dusty cloud of gas, where the longer (redder) wavelengths of light penetrate and pass through the dust. As dust clouds form and dissipate, the dust density can be revealed by examining the light blocked and transmitted by fixed, background objects.

Think about that last point for a minute. The Universe, initially after the Big Bang (i.e., before we form any stars at all) is made of neutral atoms. As gravitation begins to draw more and more matter into the most initially overdense regions, more and more neutral atoms accumulate there: contracting, heating up, and drawing even more matter into those regions. Eventually, these overdense regions will fragment, heat up sufficiently to ignite nuclear fusion in their cores, and go on to form stars: the first stars in the Universe. These stars will live-and-die — giving rise to heavier elements like carbon, oxygen, neon, sulfur, silicon, iron, and more — and from their ashes, along with more hydrogen and helium, subsequent generations of stars will form.

But starlight will struggle to get out, across the Universe, from those early generations of stars. These stars will be surrounded by neutral atoms, and those neutral atoms will be very efficient at absorbing much of that starlight. In fact, for as long as neutral atoms persist, which observationally appears to be until the Universe has aged for about ~550 million years after the Big Bang on average (and up to ~1 billion years in places), at least some of that initially emitted starlight gets blocked. In fact, the only reason we can see the Universe as clearly as we can, in all directions, is because of the existence of a large enough number of high-energy, ultraviolet photons: enough photons to ionize literally every neutral atom in intergalactic space.

Initially, at left, the Universe is filled with neutral, light-blocking matter back before any stars have formed. When stars begin to form, however, they create ionizing ultraviolet photons, which lead to pockets that behave as though they’re transparent to visible light, as shown in red. Over time, as we move to the right, more and more of the Universe becomes reionized, until reionization completes around 550 million years after the Big Bang.

If we want to understand, “how did the Universe become transparent to starlight?” then the real question we have to answer is, back during those early stages (i.e., in the first ~550 million years of cosmic history), where did all of those ultraviolet photons that reionized all of those atoms come from?

That’s what we’re attempting to study when we’re looking at the cosmic question of reionization: what made all of those ultraviolet photons that reionized the Universe?

That’s also one of the main science goals of JWST. If Hubble was the telescope that showed us what the Universe looked like, then JWST, launched a full 31 years after Hubble, is the telescope that is showing us how the Universe grew up to be the way that it is today. After all, JWST has capabilities that exceed Hubble’s in a variety of important fashions.

• JWST is larger than Hubble: 270% as large in terms of the diameter of its primary mirror, giving it superior resolution to Hubble.
• JWST has more light-gathering power than Hubble, with around seven times Hubble’s collecting area. That means that JWST can gather more light, faster, than what Hubble can. What Hubble would take a full month to observe, JWST can accomplish in just four days.
• JWST is more distant from Earth than Hubble, and is kept much colder: at about ~40 K instead of about ~200 K. This allows JWST to probe much more distant wavelengths than Hubble, and to see more redshifted, fainter, and more cosmically distant objects.
• And, again, although it isn’t as thoroughly appreciated as the other JWST advances, JWST also has superior instrumentation: more wavelength filters covering a greater range with greater throughput/sensitivity and the ability to hone in on specific features.

Preliminary total system throughput for each NIRCam filter, including contributions from the JWST Optical Telescope Element (OTE), NIRCam optical train, dichroics, filters, and detector quantum efficiency (QE). Throughput refers to photon-to-electron conversion efficiency. By using a series of JWST filters extending to much longer wavelengths than Hubble’s limit (between 1.6 and 2.0 microns), JWST can reveal details that are completely invisible to Hubble, including of doubly ionized oxygen even for ultra-distant, high redshift galaxies.

There’s a big, physical reason why that last point matters. Why is it so important that JWST has the ability not to merely measure light over a larger wavelength range, but to be able to hone in on specific emission line features?

Well, for understanding the galaxies that reionized the Universe, the galaxies that emit the greatest amounts of ultraviolet light are specifically starburst galaxies: galaxies that are forming stars so incredibly rapidly that the entire galaxy is behaving like a star-forming region. When enormous starbursts occur in our Universe, we don’t just see the signature of ionized hydrogen — hydrogen atoms that have had their single electrons stripped away by ultraviolet radiation — but of doubly ionized oxygen atoms as well.

In fact, regions of space that are known to have the emission signatures of doubly-ionized oxygen are:

• extremely hot,
• emit large amounts of ionizing light,
• and have extremely high “escape fractions” for their ultraviolet light,

which means they’re excellent not only at creating the light that can ionize neutral atoms, but that a large fraction of that light escapes into intergalactic space. If we hope to find the sources of light that reionized the Universe, it would make a tremendous amount of sense to use JWST’s capabilities to try to find these types of galaxies, with doubly ionized oxygen as emission signatures, in the early Universe at these critical, young times.

Around a variety of stellar corpses and dying stars, doubly-ionized oxygen atoms produce a characteristic green glow, as electrons cascade down the various energy levels when heated to extreme temperatures often exceeding ~50,000 K. Here, the planetary nebula IC 1295 shines brilliantly. These conditions are present in intense star-forming regions (including in the early Universe) and around stellar corpses, where the green phenomenon also helps color the so-called “green pea” galaxies, as well as Earth’s aurorae.

Here in the nearby Universe, we can see doubly ionized oxygen signatures in many extremely hot regions of space. We can also test and measure doubly ionized oxygen atoms in the lab, allowing us to determine that it emits lines at a series of wavelengths: primarily at 500.7 nanometers, and then secondarily at 495.6 and 493.1 nanometers. In fact, they weren’t initially even discovered in laboratory experiments, but in the ionized regions around dying, Sun-like stars way back in the 1860s. These wavelengths are specific and important, and allow us to identify these signatures in space both nearby, where they have these same wavelengths, and also in the distant Universe, where they get stretched by the expansion of space to longer, more redshifted wavelengths.

That’s where the power of JWST comes into play.

JWST has not just these wide-band or medium-band filters, but also narrowband filters, enabling it to hone in on a specific set of light signatures. One important signature comes from this doubly ionized oxygen at a rest-frame wavelength of 500.7 nanometers, which then gets redshifted (or stretched to longer wavelengths) by the expansion of the Universe. By looking for galaxies at a redshift of ~7 (or just over), we can make this doubly ionized oxygen fall into a specific wavelength filter that its NIRCam instrument posseses: the F410M (medium-band) filter.

The galaxies that compose Pandora’s Cluster, Abell 2744, are present within the three separate cluster components easily visually identifiable, while the remaining background sources are scattered all throughout the Universe, including many from the first ~1 billion years of cosmic history. This field of view is now known to contain many of the earliest galaxies ever found, as well as the youngest protocluster of galaxies ever discovered to date: just 650 million years after the Big Bang.

Where was the best place to do this?

In the field behind a very rich galaxy cluster, like Abell 2744, shown above. Abell 2744 was the central target of a number of early JWST science surveys, including UNCOVER and JADES: two of the galaxy surveys that have found many of the most distant galaxies (7 of the top 8, as of June 2025) known at present. The galaxy cluster itself isn’t of interest here; it’s fairly close by: at a distance of ~4 billion light-years. But the galaxy cluster itself, with huge amounts of mass distributed over a relatively large area of the sky, possesses an enormous amount of gravity, and that gravity bends, distorts, and importantly magnifies the light from objects that happen to be located behind it, as seen from our perspective.

That’s what makes this field-of-view so important. What we can do, with JWST, is:

• look at this region of sky in a large number of different wavelength ranges,
• identify candidate objects that have the signatures of a galaxy, even an intrinsically faint one, from the first ~550 million years of cosmic history (when the Universe became reionized),
• and then focus in on those objects that have a strong emission feature that corresponds to doubly ionized oxygen,
• and perform follow-up observations on those objects to see what their properties are.

After all, if there are enough of these objects, emitting enough light, early enough in cosmic history, we could determine whether these are the types of galaxies that could have reionized the Universe.

In this field of view, which includes massive galaxy cluster Abell 2744, a total of 83 young galaxies were found from within the first ~800 million years of cosmic history. By combining the gravitational lensing enhancements of the foreground clusters with deep JWST imagery, scientists identified 83 candidate young galaxies that were doubly ionized oxygen emitters. The 20 located inside the white diamonds were chosen for follow-up with NIRSpec, with all 20 being confirmed as young, star-forming galaxies.

If you’ve heard about what JWST has found in the ultra-distant Universe, you probably heard about an abundance of bright, early galaxies that exceeded what most astronomers expected. These little red dot galaxies, as they’re now known, wound up teaching us a lot about the Universe, but even though there were many of them seen at very great distances and early times, there weren’t close to as many as we’d need for them to explain reionization. In fact, most estimates concluded that those bright, early galaxies only contributed around ~5% of the necessary ultraviolet photons for reionizing the Universe, with an upper limit of around ~20% given the most generous assumptions. Something else had to be the culprit.

That made these low-mass, faint, but far more abundant galaxies a very attractive candidate for being responsible for reionizing the Universe. The UNCOVER collaboration put out a preprint last July, in 2024, where they identified many of these faint, star-forming galaxies with tentative signatures of doubly ionized oxygen emission. This research was expanded and published earlier in 2025, and the results were presented by lead author Isak Wold at the American Astronomical Society’s 246th meeting in June of 2025. Remarkably, in just this small field-of-view, a whopping total of 83 galaxies from the first ~550 million years of cosmic history were found with this characteristic signature of doubly ionized oxygen emission.

This view shows a region of space around galaxy cluster Abell 2744 with three individual galaxies highlighted in green diamonds: labeled 41038, 41028, and 41006, respectively. All corresponding to a cosmic age of ~790 million years, their estimated stellar masses are shown, along with their emission line signatures, redshifts, and lensing magnifications.

These galaxies turn out to be incredibly important for reionization for a number of reasons. First off, there are a great many of them: they’re abundant in precisely the numbers that we’d expect if they were responsible for ~100% of the ionizing, ultraviolet photons needed to make the Universe transparent to starlight. Second, these doubly ionized oxygen emitters have a very high escape fraction (of ~25% or more) for all ultraviolet photons generated, far greater than from galaxies that lack this signature, including little red dot galaxies. And third, although these galaxies are small and compact, gravitational lensing enhances their brightness, enabling us to spectroscopically test many of them to see if they are what they appear: of the 20 tested galaxies in this way, all 20 were confirmed, teaching us that we aren’t fooling ourselves.

Spectroscopy was also able to detect signatures of other elements in these galaxies: hydrogen, helium, neon, nitrogen, and even different ionization states of oxygen. The authors of the study — members of the UNCOVER collaboration — took the doubly ionized emission signatures from oxygen observed in this gravitationally lensed sample of galaxies, combined it with spectroscopic properties from JWST’s NIRSpec instrument, and was able to estimate how much ionizing, ultraviolet light was produced by the galaxies they found. They estimated the escape fraction of those ultraviolet photons by using galaxies found nearby with similar properties, and calculated how many total ultraviolet photons were produced by these galaxies (and galaxies like them) at early times: within the first ~800 million years of cosmic history.

And what they found, profoundly, was that these early galaxies provide ~100% of the light needed for reionization.

Although galaxies like 41028, shown here, are very hard to image directly due to their intrinsic faintness and great distances, gravitational lensing can enhance the brightness of galaxies like this to make them visible to JWST’s eyes. With an estimated stellar mass of just ~2 million Suns, comparable to large star-forming regions (like 30 Doradus) found in the Local Group, these low-mass but quite numerous galaxies, starbursting early on in cosmic history, provide the majority of ultraviolet light needed to reionize the Universe.

This is a remarkable achievement when you take a step back and consider just what was accomplished. Early on in the JWST era, we realized that we couldn’t count on the brightest, most massive galaxies to reionize the Universe; as big and bright and abundant as they were, there simply weren’t anywhere near enough of them to create the ionizing light needed to make the Universe transparent to starlight. Then as we started looking to fainter galaxies at those early times, we started seeing a new candidate for reionization: the low-mass, but very hot and very rapidly star-forming galaxies that should have been far more numerous and abundant.

Early studies, however, couldn’t go faint enough (or deep enough) to see how many galaxies there were at the fainter end of the galactic spectrum, and this new study — with the aid of gravitational lensing — was finally able to get there. It’s only because of those dual capabilities:

• the novel observatory capabilities that came along with JWST,
• aided by the enhancement of gravitational lensing provided by the gravity of the foreground cluster, Abell 2744,

that we were able to finally solve the puzzle of how the Universe became transparent to starlight. It wasn’t the brightest, biggest, rarest galaxies that did it, but rather the more common, mundane, early Milky Way analogues that, all together, produced enough light to reionize the Universe. That’s why, ultimately, we can see it as we do today.