Hubble reveals a dark side to Uranus’s moons

While Mercury, Venus, Mars, Jupiter, and Saturn have all been known since prehistoric times, it wasn’t until 1781 that humanity discovered Uranus. Just six years later, in 1787, the largest two Uranian moons, Titania and Oberon, were discovered, followed by Ariel and Umbriel in 1851. The next largest one, Miranda, wasn’t discovered until 1948, and the one after that, Puck, wasn’t found until Voyager visited it in the mid-1980s. While Voyager 2 imaged all of these Uranian moons (as well as discovering several other, smaller ones) during its fly-by in 1986, we weren’t able to acquire full 360° images of these worlds: just a few key snapshots, using 40+ year old technology, from afar.

Each of these moons, however, is tidally locked to Uranus, meaning that the same side of each moon always faces toward its parent planet, while the opposite side always faces away. These large moons all make complete orbits much more quickly than our own moon revolves around the Earth: with the four largest, Ariel, Umbriel, Titania, and Oberon completing one Uranian revolution in 2.5, 4.1, 8.7, and 13.5 days, respectively. Because they’re so close to Uranus, and because rapidly-rotating Uranus possesses such a strong magnetic field, astronomers had long suspected that these moons would be preferentially darkened on one side — on their trailing sides — relative to the other.

Here in 2025, critical observations of these four Uranian moons were taken in the ultraviolet, with the capability of finally answering the question of whether this predicted darkening actually occurred. Instead, astronomers found the exact opposite: the darkening is real for 3 out of 4 of these moons, but in exactly the opposite fashion to what was predicted. Here’s what we’ve just learned.

The five largest moons of Uranus, in order from the innermost to the outermost, are Miranda, Ariel, Umbriel, Titania, and Oberon, with the latter two being the largest and first-discovered among Uranus’s moons. All of these moons and the innermost one rotate within a single degree of Uranus’s orbital plane except for Miranda, which is inclined by 4.3 degrees.

The initial theory was simple and straightforward. Uranus, one of the gas-rich worlds of our Solar System with about 14.5 times the mass of Earth, is unique among the known planets for rotating on its side: inclined at 98° to its orbit. Whereas Earth, tilted at 23° relative to its orbit, has a magnetic field that’s almost perfectly aligned with our rotational axis — off by a mere 7° — Uranus’s magnetic field is wildly different from its rotational direction, tilted by 59° from its orbital plane, or different from the rotational axis by a full 39°. At the same time, Uranus’s four largest satellites, Ariel, Umbriel, Titania, and Oberon, all of which are over 1000 km in diameter, are inclined at just ~0.1° to Uranus’s orbital plane.

With a combination of factors at play:

• a heavily tilted, rapidly spinning Uranus,
• with a steeply inclined, strong magnetic field,
• consisting of four large, tidally locked, orbiting moons,
• all with mean orbital distances between 190,000 and 600,000 km, comparable to the Earth-Moon distance,

astrophysicists calculated that Uranus’s magnetosphere would accelerate and trap charged particles, such as electrons, and would cause a phenomenon known as radiation darkening on the trailing sides of these moons.

This model shows two views of the Uranian magnetic field, as measured during Voyager 2’s fly-by of the planet, from both edge-on and face-on views from the perspective of its orbit. The scale of the magnetic field is shown with respect to its five largest moons, at right, while the difference between Uranus’s rotational axis and magnetic axis is shown at left.

Radiation darkening isn’t something most of us encounter in our daily lives, but the principle behind it is simple. When you strike any organic (and some inorganic) materials with energetic radiation, it leads to an effect where the radiation makes the material more absorptive and less reflective, causing it to become darker (and often discolored) in appearance. Radiation darkening has been observed to occur under many terrestrial conditions:

• in the optics of cameras,
• within optical fibers,
• in human skin that undergoes radiation therapy,
• and when various glasses are exposed to radiation-rich environments, including in space.

However, it’s an effect that’s also expected to occur when icy systems, including the surfaces of the large Uranian moons, are exposed to radiation-rich environments. Since Uranus’s magnetosphere ought to be extremely efficient at capturing electrons and should preferentially transport those electrons onto the trailing sides of the moons that orbit it, many astronomers supported the idea that Uranus’s moons might be two-toned. In particular, the trailing side (facing away from the direction of their orbital motion) was expected to be darker than the leading side (facing in the direction of motion) of these four major moons.

From top to bottom, the Uranian moons of Puck, Miranda, Ariel, Umbriel, Titania, and Oberon are shown as imaged several times, from approach to departure, by Voyager 2 during its fly-by in 1986. A complete map of these moons and their features has not yet been possible from the existing data.

Unfortunately, the only close-up, high-resolution images ever taken of these large Uranian moons came all the way back in 1986: when Voyager 2 flew by the seventh planet. Many fascinating pictures of the four largest moons were captured, but full coverage of the surfaces of those worlds was still lacking. Although many missions to Uranus have been proposed in all the time since then, none have ever been approved and flown. As a result, the concept of radiation darkening as applied to the trailing hemispheres of Uranus’s largest moons has long been considered as a possibility, but has never been put to the critical test.

In addition, different models made vastly different predictions for just how much the amount of radiation darkening should be, as well as which Uranian satellites should be affected most severely. However, what all the models did agree on was the simple fact that Uranus and its magnetic field lines must rotate far more quickly than the Moons of Uranus orbit the planet, implying that the field lines are constantly “sweeping” past the moons: from the trailing side to the leading side. Therefore, the thought went, that the magnetosphere of Uranus must interact with its moons, and that any charged particles swept up by those field lines should preferentially land on the trailing side. The big questions astronomers were asking were:

• what the effect of radiation darkening would be,
• which moons would be most affected,
• and by how much?

These maps of the six largest moons of Uranus were constructed from the full suite of Voyager 2 images as it made its closest approach to Uranus on January 24, 1986. Although Ariel and Miranda display the most high-resolution features, only a fraction of each of these worlds was able to be imaged during Voyager 2’s fly-by.

Without a dedicated mission to Uranus, you might think that we’re stuck: unable to make maps of the hemispheres of these moons, and thus, to observe differences between them. However, there is a clever way to look for hemispheric differences in brightness, and astronomers found a way to exploit it. Even though these moons are tidally locked to Uranus, with one side always facing toward the parent planet and the opposite side always facing away from it, the orbital positions of these moons are constantly changing with respect to the Sun. Therefore, if you can observe:

• the leading hemispheres of these moons when they’re fully Sun-lit,
• and the trailing hemispheres of these moons when they’re fully Sun-lit,

then you can compare the brightness/darkness of the two hemispheres at any wavelength you dare to observe them: ultraviolet, visible, or infrared, for example.

That was precisely what a team led by Richard Cartwright attempted to do: taking advantage of the world-class capabilities of the Hubble Space Telescope and the unique ultraviolet capabilities of its STIS (Space Telescope Imaging Spectrograph) instrument. The proposal called for studying these four largest moons — in order: Ariel, Umbriel, Titania, and Oberon — and breaking the light up into its individual wavelengths so that features between the leading side and trailing side could be compared more directly. Being the four largest moons meant that they’d appear as the four brightest moons of Uranus, making them ideal targets for such a study.

These four graphs show the spectra of Uranus’s four largest moons, Ariel, Umbriel, Titania, and Oberon, for both the leading (cyan) and trailing (magenta) hemispheres of each world. In defiance of the expectation of radiation darkening of the trailing hemisphere, instead there was evidence seen for dust darkening of the leading hemisphere.

When the first spectra were acquired, the initial data looked good. At short wavelengths — wavelengths shorter than about 270 nanometers — there was no signal at all; all of that light was being absorbed. At longer wavelengths, however, the reflectivity changed as a function of the properties of the planet, leading to a series of clear absorption and emission features that must have been tied to the composition and properties of the surfaces of these worlds. In some cases, bright, spike-like lines appeared: clear evidence of atomic and/or molecular excitations. At other wavelengths, deep absorption features were present.

But when the two hemispheres — the leading and the trailing ones of these worlds — were compared, the findings, shockingly, were not at all consistent with what was expected. Sure, the features found on the leading hemispheres were also present on the trailing hemispheres, but the expected darkening of the trailing hemisphere wasn’t seen on any of these worlds. Instead:

• the innermost of these four moons, Ariel, showed a trailing hemisphere that was about 5% brighter than the leading hemisphere,
• the next moon out, Umbriel, showed identical brightnesses between its leading and trailing hemispheres,
• even farther out, Titania displayed just a 1% difference between leading and trailing hemispheres, again with the trailing hemisphere being brighter,
• and finally, the farthest moon out of all, Oberon, had the greatest differences of all, with the trailing hemisphere being a full 20% brighter than the leading hemisphere.

This annotated Hubble view of Uranus shows its rings and its five largest moons, along with the shadow of Ariel on Uranus itself. While Hubble has been able to detect brightness differences between the leading and trailing hemispheres of Oberon, Ariel, and slightly for Umbriel, the ultimate cause of this darkening has only been suspected to be dust, not confirmed as such.

The expectation, that the trailing hemisphere would be darker than the leading hemisphere of these moons due to radiation darkening, was clearly way off. Even though this could have been an effect that darkened the trailing side, the fact that the trailing sides are observed to be just as bright or brighter than the leading sides implied that radiation darkening was negligible. Furthermore, it taught us that another effect must be at play — and dominant — instead: something that darkened the leading hemisphere rather than the trailing hemisphere. This isn’t unheard of, as it’s a phenomenon known to afflict several of Jupiter’s and Saturn’s moons, with Saturn’s Iapetus being the most prominent: dust darkening.

For Iapetus, there’s an outermore moon known as Phoebe that’s an incredible source of dust: Phoebe generates its own large, diffuse dust ring that was discovered in infrared light. As Iapetus moves ahead in its orbit, it plows into this dust ring: accumulating this darkened material on its leading hemisphere but not on its trailing hemisphere. Even though ices form on Iapetus continuously, the darker material heats up when it’s in direct sunlight, sublimating those ices and causing them to preferentially be deposited on the lighter-colored trailing side. As a result, Iapetus’s leading side is six times more reflective than its trailing side: the most severe difference known in all the Solar System.

The striking color difference on Iapetus is most clearly visible if you divide Iapetus into its leading and trailing hemispheres, where the leading hemisphere looks very much like the windshield of an enormous vehicle that’s plowed into a swarm of oncoming insects. In this case, it’s the dust generated by an outermore moon, Phoebe, that’s responsible for the dust darkening of Iapetus’s leading hemisphere.

This idea becomes particularly appealing when we attempt to apply it to the Uranian system. Although these STIS observations have been mainly focused on Uranus’s four largest moons — Ariel, Umbriel, Titania, and Oberon — it’s worth pointing out that Uranus has a whopping 28 moons that are known:

• 13 inner moons,
• the 5 major moons (including Miranda, which is exterior to the 13 inner moons but interior to Ariel),
• and then 10 outer, irregular moons.

This is important, because there are no additional moons between Ariel and Oberon other than Umbriel and Titania; there are no small, low-gravity moons that could be copious sources of dust production in that orbital region.

That means there’s a plausible scenario where one or more of the small, low-mass inner moons could be generating large amounts of material for the inner major moons, like Miranda and Ariel, to plow into. It means there’s a plausible scenario where one or more of the small, low-mass outer moons could be generating large amounts of material for the outermost major moon, Oberon, to plow into. That material would then preferentially darken the leading hemisphere while leaving the trailing hemisphere of those worlds unaffected, while the in-between worlds — Umbriel and Titania — were only negligibly affected by that material.

This schematic view, to-scale, shows the Uranian ring system along with the known moons that orbit and shepherd the rings around Uranus along with it. The dense inner rings, from epsilon on inward, are the ones clearly imaged by JWST. The “mu” ring, meanwhile, may be the culprit behind the observed leading hemisphere darkening on Ariel, and may also darken Miranda and Puck, neither of which has been observed to sufficient precision to know for certain.

Instead of the “magnetosphere story,” which was the story that astronomers anticipated these moons of Uranus would tell, we now have a “dust transfer story” for the Uranian moons, particularly for Oberon and Ariel. It’s the leading sides of these moons that gets darkened, not the trailing sides, and it’s likely due to the accumulation of darker, dusty material, not due to radiation darkening. In the words of Richard Cartwright, the principal investigator of the study that used Hubble’s STIS instrument to acquire these key measurements:

“Based on these findings, we suspect that Uranus’ magnetosphere may be fairly quiescent, or it may be more complicated than previously thought. Perhaps interactions between Uranus’s moons and magnetosphere are happening, but for some reason, they’re not causing asymmetry in the leading and trailing hemispheres as researchers suspected. Instead, we think that dust from some of Uranus’ irregular satellites is coating the leading side of Titania and Oberon. The answer will require further investigation into enigmatic Uranus, its magnetosphere, and its moons.”

Ideally, the way we’d investigate Uranus next would be with a dedicated orbiter mission to it, enabling us to go up close and make detailed maps of the full globes of its major moons, just as we’ve done in the past for Saturn’s Iapetus with the Cassini mission.

This diagram shows an illustrated cutaway, along with the suspected composition of the four largest moons of Uranus: Ariel, Umbriel, Titania, and Oberon. Even here in 2025, the best data we have about the surfaces of these worlds comes from the fly-by observations of Voyager 2 back in 1986; in all the time since, no more than a single pixel has ever been acquired of these worlds.

Unfortunately, no such mission has ever been approved and brought to the final design stage: a prerequisite for construction, launch, calibration and commissioning, and finally the science operation stage. However, there is the possibility of conducting follow-up observations that could indeed teach us additional key pieces of information about these four major moons, and the Uranian system in general, with existing technology.

• We could use Hubble, or potentially even the world’s largest ground-based optical telescopes, to acquire visible light images and spectra of each hemisphere of these four major moons, and see if the leading/trailing trends hold in the visible as well as the ultraviolet.
• We could use JWST in the near-infrared to acquire leading/trailing hemisphere spectra of Ariel, Umbriel, Titania, and Oberon to see what sort of brightness differences exist at a variety of different wavelengths that the NIRSpec instrument is capable of covering.
• And we could even use JWST’s mid-infrared instrument (MIRI) to search the areas interior to Ariel and exterior to Oberon for evidence of faint, diffuse, dusty rings, similar to the Phoebe ring discovered by NASA’s Spitzer around Saturn back in 2009.

At the present time, however, the only relevant observing program that’s been greenlit is to acquire near-infrared data of Ariel (alone) using JWST, meaning that we’ll have to wait a little bit longer to test the new leading theory, and to find the ultimate culprit for this surprising, leading-hemisphere-darkening phenomena.