The peer-reviewed findings were published in Communications Earth & Environment by a collaboration between Diamond Light Source and the ISIS Neutron and Muon Source. Lead author and Diamond beamline scientist Liam Perera said the inclusions could act as microscopic capsules containing material from the ocean beneath Enceladus’s icy crust. If biological material or other biosignatures are present in that ocean, natron could potentially trap and preserve them, although the experiment did not detect life on Enceladus or demonstrate that preserved material survives under the moon’s surface conditions.

Enceladus is of particular interest because a global subsurface ocean is thought to lie beneath its frozen surface and because material from the interior is ejected through massive cryovolcanic plumes at the moon’s south pole. The ocean is buried beneath about 20 km of ice, making direct drilling impractical. Plumes nevertheless provide a natural route for material to reach the surface.

Perera said planned NASA and ESA flagship missions will need to look for geochemical signatures representing fresh, unprocessed material from the subsurface. To investigate how such material might be incorporated into minerals, the researchers prepared a sodium chloride–sodium carbonate–water solution as an analogue for Enceladus material. They cooled the sample while collecting synchrotron X-ray computed microtomography, or μCT, and X-ray diffraction measurements on Diamond’s DIAD beamline.

Combining these techniques allowed the team to follow the development of ice and salt phases during freezing and relate the mineral phases to their three-dimensional internal structures. The experiment examined mineral formation in a terrestrial laboratory analogue; it did not analyse a sample collected from Enceladus. DIAD has two independently configurable X-ray beams, enabling spatially registered imaging and diffraction measurements.

Tomography provided detailed views of the sample’s internal structure, while diffraction identified its mineral composition. The researchers said this combination was central to distinguishing the trapped fluid pockets in situ. Perera said the method allowed the salts, which are hundreds of micrometres in size and formed within ice at 205 K, to be examined non-destructively as crystallisation progressed.

The bright synchrotron source also enabled the researchers to watch salt crystallisation in real time as the Enceladus-type solution froze. The final frozen sample contained three major phases: hexagonal ice, a eutectic mixture comprising ice, natron and hydrohalite, and individual natron crystals. Natron accounted for around 21% of the imaged volume.

Correlating the diffraction data with the crystals’ three-dimensional microstructures revealed that natron enclosed pockets of fluid from which it crystallised. The study therefore provides direct laboratory evidence for a mechanism through which ocean-derived material could become sealed within a mineral, although the experiment did not establish how frequently such inclusions form on Enceladus or whether they are present there. Fluid inclusions matter because they can preserve fluid at the moment a mineral forms.

On Earth, inclusions in salt minerals have been used to reconstruct the chemistry of ancient environments, and some salt-hosted inclusions have been found to contain organic biological signatures and microorganisms. The Enceladus experiment did not reproduce those terrestrial biological findings. Instead, it showed that an Enceladus-like salty solution can generate natron with enclosed fluid, providing a plausible mechanism by which ocean-derived chemistry, and potentially biological material if present, might be trapped within crystals.

Based on the results, the team suggests that future spacecraft should be equipped to identify crystalline natron and other hydrated carbonate minerals at Enceladus’s surface. The researchers recommend looking for natron in fresh deposits close to plume sources, where larger salt-rich particles may be more likely to contain inclusions. They also identify near-infrared spectroscopy as a possible method for identifying hydrated carbonate deposits and helping select promising sampling locations.

These are proposed mission considerations arising from the laboratory analogue study, not evidence that natron or biologically meaningful inclusions have been detected on Enceladus. Perera said future landers should prioritise sites with signs of natron, while future spacecraft should have the capability to identify the mineral remotely. Understanding how minerals form and preserve fluids at extremely low temperatures may also help scientists interpret observations and samples from icy worlds, subject to confirmation that the proposed minerals and structures exist under Enceladus conditions.

The study highlights the potential value of synchrotron science for planetary research. Its central achievement is a laboratory demonstration that natron can trap fluid when an Enceladus-like solution freezes, together with a non-destructive method for identifying the inclusions inside crystals. The result narrows a possible mineralogical target for future missions without establishing that the ocean contains life, that natron is abundant on Enceladus, or that inclusions preserve biosignatures after material leaves the ocean.

For now, the experiments establish natron as a potential astrobiological target and show how synchrotron imaging can reveal the hidden structure of cold, salty materials that may inform the design and interpretation of future missions.

Subject of Research: Laboratory study of natron fluid inclusions in a frozen Enceladus-like solution

Article Title: Diamond Light Source scientists identify natron as a key target for future Enceladus missions

Article References: Diamond Light Source scientists identify natron as a key target for future Enceladus missions. Original publication (Bibliographic details unavailable.)

Image Credits: AI Generated

DOI: Not provided

Keywords: Enceladus, natron, fluid inclusions, subsurface ocean, synchrotron X-rays, Diamond Light Source, planetary mineralogy, cryovolcanism, astrobiology, icy moons, Diamond, Light

The findings, published in the Annales Geophysicae as part of a special issue devoted to Juice’s first-ever lunar and Earth flybys, are remarkable for what they accomplish without a single laser pulse. Laser altimeters measure distance by timing how long a laser pulse takes to travel to a planetary surface and back. To convert those timings into accurate footprint positions on a moon’s surface, scientists must know precisely where the instrument’s line of sight points relative to the spacecraft’s reference frame. Launch vibrations and the thermal environment of space can shift that alignment after launch, so in-flight calibration is standard practice for such instruments. The plan for GALA had been to use the lunar flyby of the Lunar Earth Gravity Assist maneuver in August 2024, during which JANUS, Juice’s scientific camera, might even have caught flashes of GALA’s laser reflected off the lunar nightside. The instrument reboot ended that hope.

The Earth-farewell campaign on 9 September 2024 offered a creative alternative. Juice stared at Earth between roughly 09:39 and 10:08 UTC, adjusting its pointing intermittently without rotating, while JANUS captured 168 images of the planet through all of its filters. GALA, too far away to receive any return pulses, was instead operated in a passive mode, functioning as a single-pixel telescope. In this configuration, the instrument’s noise level, recorded as the root-mean-square of electrical signals in each range window, serves as a sensitive proxy for the amount of sunlight reflected from Earth that reaches its avalanche photodiode detector. Because the noise contribution from incoming photons grows with the square root of the photon flux, every shift in the telescope’s field of view across the bright disk of Earth, its clouds, continents, and oceans, leaves a distinct signature in the noise record.

The technical challenge was to connect two very different kinds of measurements. GALA detects photons in a narrow band of roughly three nanometers around 1064 nanometers, while JANUS’s filter-12, the best match, spans a much broader range from 949 to 1081 nanometers. The team computed a conversion ratio between the radiance measured by JANUS and the radiance collected by GALA by incorporating the transmittance of the GALA bandpass filter, arriving at a value of 0.0326. Tests with realistic Earth spectra for continents, seawater, and clouds showed that this flat-spectrum assumption introduces systematic errors of at most about 9 percent, small compared with the enormous natural variation in Earth’s reflectance, where clouds can be an order of magnitude brighter than seawater. The team then translated the photon flux arriving at the detector into a predicted noise level using a theoretical model that accounts for the avalanche photodiode’s quantum efficiency, multiplication gain, bandwidth, and transimpedance, alongside the constant contributions of surface and bulk dark current.

One further obstacle was the coarse temporal resolution of the spacecraft attitude data, which were available only at one-minute intervals during the campaign. To bridge the gaps, the team tested three interpolation methods for the quaternions defining the spacecraft’s orientation: linear interpolation, Piecewise Cubic Hermite Interpolating Polynomial interpolation, which preserves monotonicity, and Akima interpolation, which estimates derivatives locally from neighboring samples. Other spline methods were rejected because they produced unrealistic overshooting between coarse attitude points. The team also propagated measurement errors by adding random perturbations with a standard deviation of 15 microradians to the attitude samples before interpolation.

The comparison between observation and simulation produced a striking result. If GALA still pointed where pre-launch measurements said it should, the noise level should have risen and fallen at specific times as the field of view slid across the Earth. Instead, the observed noise rose gradually over about 20 seconds but dropped rapidly within about 5 seconds, an asymmetry reflecting faster attitude motion at the end of the overlap, and none of the simulated timings under any interpolation method matched the observations. At the end of the overlap, every simulation predicted the noise should fall more than 13 seconds earlier than it actually did. Earth’s atmosphere could not explain the discrepancy: the bright atmospheric halo visible in images from satellites such as Himawari-8/9 alters the overlap timing by less than 0.3 seconds, far too small to matter. The absolute magnitude of the noise also disagreed, with simulations based on the pre-launch alignment predicting levels about 0.1 millivolts too high, and the fine structure of the noise disagreed too, with observed local maxima around 09:39 to 09:40 and 09:42 to 09:43 where the simulation predicted local minima.

To find where the instrument actually pointed, the team searched systematically across possible boresight directions, comparing simulated and observed start and end times of the noise increase, the absolute noise magnitude, and the correlation between the temporal patterns of noise variation. The most-likely solutions clustered more than 700 microradians away from the pre-launch boresight vector, with estimation errors below 100 microradians within each method. Notably, the alignment of JANUS itself cannot be blamed for the shift: the Earth’s position predicted in the JANUS image matches its observed position with an error of less than 2 pixels, corresponding to only about 30 microradians, and the predicted boresight offset for JANUS from thermal effects is an order of magnitude smaller than the GALA discrepancy. Nor can known pre-launch behavior easily account for the result. Thermal-vacuum cycling tests before launch showed boresight offsets below 100 microradians across operational temperatures from minus 7 to plus 41 degrees Celsius, and launch vibration effects were expected to be smaller than 700 microradians by roughly an order of magnitude.

Such large shifts are not unprecedented. The LIDAR instrument on Japan’s Hayabusa2 asteroid mission showed variations of approximately one milliradian in its line of sight when comparing different calibration methods, crossover analysis of the NEAR Laser Rangefinder revealed boresight shifts exceeding one milliradian, and the Lunar Orbiter Laser Altimeter on the Lunar Reconnaissance Orbiter exhibited comparable variations between dayside and nightside operations caused by thermal distortion. The authors caution, however, that their conclusion for GALA is not yet definitive. Discrepancies between the best simulations and the observations remain, and the limited attitude sampling during the staring observation means the influence of interpolation and attitude uncertainty cannot be fully excluded. Additional cruise-phase data will be needed to confirm the offset and identify its cause, which remains unidentified at present.

The study delivers more than a diagnosis. It establishes a radiometric calibration of the GALA receiver, an empirical relationship between the photon flux incident on the detector and the observed noise level, which is a prerequisite for GALA’s planned passive reflectance measurements of Jupiter and the Galilean moons. The authors estimate that during icy moon flybys the GALA noise level should rise by 21 to 66 percent above its background of 1.12 millivolts, large enough to serve both as a boresight check under the cold operating conditions near Ganymede and as a signal for surface reflectance measurements, though Jupiter’s harsh radiation environment may increase the detector’s dark noise and complicate interpretation. Future opportunities loom soon: Earth Gravity Assist maneuvers in September 2026 and January 2029, and lunar observations around those flybys, where the Moon’s predictable brightness and lack of atmosphere make the noise-based method especially clean, could push the expected noise increase above 110 percent. The framework, the authors note, extends beyond GALA, offering an independent alignment-calibration approach for other planetary laser altimeters, including the BepiColombo Laser Altimeter headed to Mercury, and complementing traditional crossover and topographic methods as Juice continues toward its 2031 arrival in the Jovian system.

Subject of Research: In-flight receiver calibration and boresight alignment of the Ganymede Laser Altimeter aboard the Juice spacecraft using passive Earth observations.

Article Title: In-flight receiver calibration of the Ganymede Laser Altimeter (GALA) by passive Earth observations

Article References: Nishiyama, G., Stark, A., Matz, K.-D., Hüttig, C., Wickhusen, K., Agostini, L., Althaus, C., Binger, J., Enya, K., George, J., Lingenauber, K., Kimura, J., Kobayashi, M., Macrì, G., Neumann, W., Penasa, L., Portyankina, G., Tubiana, C., Willner, K., … Hussmann, H. (2026). In-flight receiver calibration of the Ganymede Laser Altimeter (GALA) by passive Earth observations. Annales Geophysicae, 44(2), 1075-1092. https://doi.org/10.5194/angeo-44-1075-2026

Image Credits: AI Generated

DOI: 10.5194/angeo-44-1075-2026

Keywords: GALA, Juice, laser altimeter, Ganymede, in-flight calibration, boresight offset, JANUS, Earth flyby, avalanche photodiode, radiometric calibration, icy moons, Jupiter system

The motivation for the study stems from a growing realization among planetary scientists that many of the smaller moons in the outer Solar System may not be the originals. Several lines of evidence have led researchers to question whether the moons we observe today around Saturn and Uranus, particularly those with radii below 1,000 kilometers, are primordial bodies that formed alongside their parent planets. Instead, a compelling hypothesis holds that most or all of these smaller moons were blown apart in large collisions at some point in their history, with the moons now visible to spacecraft reassembling later from the orbiting debris of their shattered predecessors. For Saturn’s system, some models have proposed disruptive collisions occurring remarkably recently — even within the last 100 million years. That possibility raises an immediate and consequential question: if the candidate ocean moons we see today are, in effect, second-generation bodies stitched together from collisional wreckage, can they still host oceans at all?

To tackle this problem, the research team assembled a sophisticated two-part modeling framework. The first component was a smoothed particle hydrodynamics, or SPH, model, a computational technique widely used in astrophysics to simulate the behavior of fluids and deformable materials under extreme conditions. SPH allowed the team to render catastrophic collisions in fine detail, tracking how a moon’s icy and rocky components scatter, mix, and reaccumulate when struck by an impactor of comparable or substantial size. The second component was a thermal-structural evolution model, which simulated how a moon’s interior evolves over time following such an event — how heat generated by the impact redistributes, how melted material refreezes, and how the boundary between ice and rock migrates as the body cools. By coupling these tools, the scientists could compare three scenarios side by side: a moon before an impact, the same moon after a disruption, and an otherwise identical moon that never experienced a collision at all.

The team’s definition of a disruptive impact was deliberately stringent. As Rhoden explained, the threshold required that the largest surviving fragment be less than half the size of the original target moon. In other words, these were not glancing blows or crater-forming events, but true catastrophic disruptions — collisions energetic enough to break a moon apart entirely, dispersing some fraction of its material into space while the remainder gravitationally reconsolidates into what is essentially a brand-new moon built from the debris. Such events, the researchers note, are far from exotic. Disruptive impacts were common throughout the Solar System’s formative era, and they would fundamentally rework a moon’s interior structure rather than merely scarring its surface.

The intuitive expectation might be that pumping enormous quantities of energy into a frozen moon would be a recipe for ocean creation. After all, a sufficiently violent impact should melt ice, and melted ice is, by definition, liquid water. Rhoden framed the question directly: imagine a small, frozen, geologically quiet moon. If something slams into it with enough force to cause a major collision, would that impart enough energy to generate an ocean? The models delivered a decisive answer — it is actually incredibly difficult. In most simulated scenarios, a small moon that experiences a disruption either loses an existing ocean or fails to form one in the first place. The counterintuitive explanation lies in the physics of heat loss at small scales. Rufu noted that although introducing energy into the system could in principle melt ice and create oceans, that energy dissipates very quickly, which works against ocean formation rather than for it.

Rhoden offered an everyday analogy to illustrate the effect: baking a whole potato versus baking French fries. The smaller pieces heat up and cool down far faster than an intact potato. The same principle applies to a moon blown apart by an impact. Once the body is fragmented into countless pieces of ice and rock, each fragment radiates its heat into space rapidly, and any meltwater generated during the collision freezes again before it can accumulate into a lasting ocean. The very act of breaking the moon apart, which concentrates energy in the moment, undermines the long-term thermal conditions needed to sustain liquid water. The net result, the study found, is that large-scale collisions affect the thickness and longevity of existing oceans but do not contribute meaningfully to creating new ones.

Disruptive collisions do, however, leave a lasting structural imprint on the moons that survive them. In the aftermath of a catastrophic impact, the reassembling body briefly experiences widespread melting, which triggers a process known as ice-rock differentiation. A disrupted moon may begin as a jumbled mixture of rock and ice, but when the ice melts, the denser rocky material sinks toward the center while the lighter water rises. As the water refreezes, it forms a thicker coating of ice around a more substantial core. As Rufu described it, when everything consolidates after a collision, the end product is a moon with a larger core and a thicker ice shell than it had before. This differentiation matters for future exploration, because the thickness of a moon’s ice shell and the structure of its interior strongly influence how scientists interpret gravity measurements, surface features, and potential plume activity observed by spacecraft.

The fate of an existing ocean, meanwhile, depends heavily on the moon’s size. The simulations showed that when moons possessed subsurface oceans prior to impact, larger bodies — those with radii of 1,000 kilometers or more — might retain their oceans through a disruption and reassembly. Smaller moons, by contrast, tend to lose their oceans entirely. This size-dependent outcome carries significant implications for how scientists interpret the moon systems of Saturn and Uranus. If a small moon is suspected of harboring an ocean today, and if it also shows signs of having been disrupted and reassembled, the new results suggest the two histories may be difficult to reconcile — unless the disruption occurred long enough ago, or under conditions the models have not yet fully explored.

That caveat points to the remaining uncertainties in the study. Outer Solar System moons span a wide range of physical and orbital characteristics, and the parameters governing disruptive collisions — the frequency, scale, and timing of such events — are not well constrained by current observations. The authors emphasize that more work is needed to determine whether any set of conditions could promote ocean formation in smaller reassembled moons. The modeling framework itself, combining SPH collision simulations with long-term thermal evolution, offers a template for such follow-up studies, and future refinements could probe a broader range of impact scenarios, moon compositions, and orbital environments.

For now, the findings offer a measure of reassurance for scientists excited about some of the most famous ocean world candidates. Rhoden noted that it seems unlikely that moons like Saturn’s Enceladus or Dione — both thought to harbor oceans today — would have been disrupted and reassembled within the past 100 million years. Enceladus, in particular, has captivated the planetary science community since NASA’s Cassini spacecraft observed plumes of water vapor and icy particles erupting from its south polar region, providing direct evidence of a subsurface sea in contact with a rocky core. If those oceans are to be taken as long-lived and potentially habitable environments, the new simulations suggest their host moons must have avoided catastrophic disruption over geologically recent timescales — a conclusion that, if confirmed, strengthens the case that these small, icy worlds are genuine ocean worlds rather than transient products of violent collisions.

Subject of Research: The effects of disruptive impacts on the formation and longevity of subsurface oceans in icy moons of the outer Solar System

Article Title: SwRI simulations indicate large impacts may limit ocean formation in icy moons

Article References: SwRI simulations indicate large impacts may limit ocean formation in icy moons. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: icy moons, subsurface oceans, disruptive impacts, Saturn, Uranus, Enceladus, smoothed particle hydrodynamics, ice-rock differentiation, Nature Astronomy, Southwest Research Institute, ocean worlds, planetary science

The study combines two types of numerical modelling that are rarely used together for this problem. First, the researchers simulate disruptive collisions with a smoothed-particle hydrodynamics, or SPH, model. In this approach, a moon is represented by a vast collection of computational particles that interact according to the laws of fluid dynamics, gravity and material physics. The method is well suited to impacts because it can follow the extreme deformation, fragmentation, heating and redistribution of ice and rock during an event in which a target may be partially or almost completely destroyed. The resulting debris and reaccreted moon are then passed into a thermal-structural evolution model, which tracks how the rebuilt body cools, differentiates and changes internally over geological time.

The central question is deceptively simple: can an impact create an ocean where none would otherwise exist, or erase an ocean that would have formed naturally? The answer from the simulations is largely negative. The models did not produce a moon that remained frozen in its collision-free history but later developed a lasting ocean because of the collision itself. Neither direct impact heating, the heat generated as debris falls back together, nor collision-induced changes in orbital conditions consistently supplied enough long-term energy to transform such a moon into an ocean world. The findings suggest that impacts are more effective at modifying pre-existing ocean environments than at creating entirely new ones.

That conclusion is important because a major collision can inject enormous energy into a moon. During an impact, kinetic energy is converted into shock heating, deformation and motion. Ice may melt locally or globally, rock can be heated and redistributed, and fragments can escape before some of the material returns to orbit and reaccretes. It might seem inevitable that such violence would produce a deep, persistent sea beneath the surface. But an ocean is not defined only by whether ice melts once. It must survive against freezing, heat loss and changes in the moon’s internal structure. If the energy is released too quickly, much of it can radiate away or remain concentrated in regions that do not sustain a global liquid layer.

The simulations show a striking size-dependent pattern. In reaccreted moons with radii approaching 1,000 kilometres, an ocean can survive a disruptive collision and become substantially thicker than it would have been without the impact. Larger moons retain heat more effectively because their greater volume gives them a lower surface-area-to-volume ratio. They may also develop stronger internal pressure and more substantial rocky and icy layers, allowing liquid water to persist for longer periods. When a large target is broken apart and rebuilt, the resulting internal rearrangement can place heat and materials in configurations that favour a thicker subsurface ocean, particularly when the collision occurs late in the moon’s thermal evolution.

The story is different for smaller bodies with radii near 500 kilometres. In these moons, an ocean that would otherwise develop can disappear from the post-impact history because the collision promotes ice–rock differentiation. Differentiation occurs when the moon’s materials separate according to density, with rock sinking toward the centre and ice moving outward. This process changes the distribution of radioactive elements, thermal insulation and latent heat—the energy absorbed or released when ice melts or freezes. Instead of maintaining a mixed, slowly evolving interior that supports a liquid layer, the collision can accelerate the separation of materials and redirect the moon’s thermal evolution. In some cases, the result is a frozen world where an ocean would have existed without the impact.

The researchers also tested whether orbital changes caused by collisions could indirectly generate oceans through tidal heating. A disruptive impact can alter a moon’s mass, spin, shape or orbit, potentially changing its gravitational interactions with a planet or neighbouring moons. Those changes could drive stronger tides, and tidal flexing converts orbital energy into heat inside the moon. This mechanism powers or may power internal activity on several icy worlds in the Solar System. However, in the modelled scenarios, collision-induced orbital changes did not create a post-impact ocean in a moon that otherwise would have remained frozen. The tidal response was either insufficient, too brief or unable to deliver heat in the right way and for long enough to maintain global liquid water.

The strongest ocean-enhancing outcome requires a specific combination: a large, roughly 1,000-kilometre-class target, a late disruptive impact and a collision energetic enough to substantially reshape the moon without simply removing its volatile materials. “Late” is crucial because the timing of an impact determines how much internal heat remains and how the moon’s layers have already evolved. An early collision occurs while the young Solar System is still dynamically violent, but the rebuilt body then has a long time to cool and freeze. A later collision can refresh the thermal state of a large moon at a stage when its interior is otherwise losing the ability to sustain liquid water.

That timing also limits how relevant the effect may be to moons in the modern outer Solar System. The researchers conclude that late disruptive impacts onto large icy moons are unlikely in recent Solar System history. The most spectacular collision-driven extension of an ocean may therefore be physically possible but statistically uncommon. For the moons we observe today, the existence and persistence of an ocean are more likely to be controlled by initial composition, radiogenic heating, insulation by ice layers, pressure-dependent melting behaviour and long-term tidal interactions than by a single late catastrophe. Impacts remain important, but usually as modifiers rather than originators of ocean worlds.

The study offers a more nuanced picture of planetary habitability and the evolution of icy satellites. A collision does not simply switch an ocean on or off; its effects depend on body size, internal composition, impact timing, reaccretion and the competition between heating and cooling. Large moons may gain thicker or longer-lived oceans after rare late impacts, while smaller moons may lose oceans through accelerated differentiation. Yet the simulations found no example in which a collision reliably created a lasting ocean inside a moon that would otherwise have stayed frozen. This result narrows one proposed pathway to ocean formation while highlighting how violent events can still reshape hidden environments beneath alien ice. For future missions investigating worlds such as Europa, Ganymede, Callisto, Enceladus or other ocean candidates, the geological record of impacts may reveal not only when a moon was struck, but also how that strike changed the depth and duration of its concealed water.

Subject of Research: The effects of large disruptive impacts on the formation, thickness and longevity of subsurface oceans in icy moons.

Article Title: The role of disruptive impacts on ocean generation and longevity in icy moons

Article References: Neveu, M., Rufu, R., Rhoden, A. et al. “The role of disruptive impacts on ocean generation and longevity in icy moons.” Nature Astronomy (2026). https://doi.org/10.1038/s41550-026-02955-x

Image Credits: AI Generated

DOI: https://doi.org/10.1038/s41550-026-02955-x

Keywords: icy moons, subsurface oceans, planetary impacts, disruptive collisions, smoothed-particle hydrodynamics, thermal evolution, ice–rock differentiation, tidal heating, ocean longevity, outer Solar System