Beyond Neptune's orbit lies an ocean of icy islands that has remained nearly unchanged for four billion years. These islands, known as trans-Neptunian objects, hold the recipe for planet formation. Until recently, astronomers could only see the large ones, while the small, city-sized ones stayed invisible. Now the James Webb Space Telescope has peered into this darkness deeper than anyone before and has even counted the ten-kilometer ice floes.
For this census, they used the NIRCam instrument, which imaged the same patch of sky three times at five-day intervals. Over that time, a distant object shifts against the background stars, like a slow raft among stationary buoys. But catching such faint light, comparable to a candle flame at the distance of the Moon, is no easy feat. The astronomers aligned the images according to each candidate's expected motion, boosting its signal, then ran a million algorithm detections through neural networks trained on artificially injected sources. This weeded out false tracks, leaving only genuine ice islands.
The result surprised them: the distribution of objects by brightness, that is, the luminosity function, turned out to be equally shallow for both the cold and hot populations. Cold ice floes follow calm, circular orbits, while hot ones travel on inclined and elongated paths. If the hot ones had endured violent collisions, their distribution would be steeper, but the data show no such thing. Instead, both groups follow a single power law with a slope of about 0.29, with no downturn at small sizes.
This discovery changes how we think planets are born. Previously, it was thought that at small sizes an exponential cutoff should appear, as predicted by the model of dust clumping into dense aggregates. But Webb has shown that ice floes continue to appear down to ten kilometers, as if nature makes no distinction between cold and hot outskirts. Perhaps collisions were weaker than assumed, or the surfaces of the two populations reflect light differently, masking a difference in size.
The census of icy islands has only just begun. Future surveys, such as LSST, will increase the sample but will not reach the faint end. To see even smaller ice floes, we will need new deep fields from Webb or dedicated missions. For now, the ocean beyond Neptune has become a little less mysterious, and its inhabitants have told us that planets are born the same way everywhere.
The faintest reliably detected object, JPB24, is 100 times fainter than the brightest in the survey, 2015 GK56, and its brightness is comparable to a candle flame at the distance of the Moon.