A campfire that doesn't go out all at once but smolders for hours because the wood inside still holds the heat of the past flame. That's roughly how the memory effect works for black holes. Normally, a tiny black hole with the mass of a mountain should evaporate in fractions of a second, scattering particles like sparks. But if it 'remembers' everything it has swallowed, the process slows down, and the hole lives for billions of years, surviving to this day.

Such survivors are excellent sources of neutrinos, ghostly particles that barely interact with matter. As a hole evaporates, it produces neutrinos of staggering energies — millions of times higher than anything the most powerful Earth-based accelerator can achieve.

Detectors like IceCube, a cubic kilometer of Antarctic ice studded with sensors, are trying to catch them.

The authors calculated what signal memory-holes would leave in IceCube and future facilities like IceCube-Gen2 and GRAND200k. It turns out that what matters is not just the average mass of the holes, but also the spread around it. If masses are broadly distributed, like the heights of people in a crowd, then light holes at the edge of the distribution produce an especially powerful neutrino flux.

This means such populations can be constrained more tightly than if all holes had the same mass.

Moreover, if future detectors see about 30 neutrino events, one could try to distinguish whether the hole population is uniform or spread out in mass. It turns out that proving a spread is easier than proving its absence. This all feels like detective work: we peer into ice and radio waves to understand the invisible 'embers' of a long-extinguished cosmic fire.