Cool horizons lead to information loss
Borun D. Chowdhury
TL;DR
The work argues that preserving unitarity in black hole evaporation requires invertibility of the initial-to-final state map, which in turn forces order-one deviations from the Unruh vacuum at the horizon throughout evaporation. By demonstrating that generic black holes can be formed in arbitrary microstates (with Schwarzschild failing and D1-D5-P enabling state preparation), the paper shows that smooth horizons cannot be maintained generically, favoring fuzzball/firewall structures. It critiques mainstream proposals (e.g., Susskind-Papadodimas-Maldacena models) for not satisfying invertibility, and it connects horizon structure to thermo-field doubling and analytic continuation, while reconciling entropy counting with semi-classical gravity through AdS/CFT perspectives. The key takeaway is that the information paradox cannot be resolved by purifying the final state alone; a non-smooth, information-bearing horizon (fuzzball) appears necessary for unitary evolution of generic black holes.
Abstract
There are two evidences for information loss during black hole evaporation: (i) a pure state evolves to a mixed state and (ii) the map from the initial state to final state is non-invertible. Any proposed resolution of the information paradox must address both these issues. The firewall argument focuses only on the first and this leads to order one deviations from the Unruh vacuum for maximally entangled black holes. The nature of the argument does not extend to black holes in pure states. It was shown by Avery, Puhm and the author that requiring the initial state to final state map to be invertible mandates structure at the horizon even for pure states. The proof works if black holes can be formed in generic states and in this paper we show that this is indeed the case. We also demonstrate how models proposed by Susskind, Papadodimas et al. and Maldacena et al. end up making the initial to final state map non-invertible and thus make the horizon "cool" at the cost of unitarity.
