Solvable lattice models for metals with Z2 topological order

The authors study a lattice model of spin-singlets coupled to fermionic dimers, defined on the triangular lattice. Building on a previous study concerning a similar model on the square lattice, an exactly solvable point in parameter space is identified, and its physical properties are thoroughly analyzed. The resulting many-body state, known as a fractionalized Fermi-Liquid, violates Luttinger's theorem through the appearance of Z2 topological order. The exactly solvable point is massively degenerate due to its flat-band spectrum. Perturbing away from this point lifts the degeneracy, giving rise to a non-trivial dispersion. A crucial feature of the triangular lattice model, which distinguishes it from all previous studies, is the stability of a Z2 deconfined phase. This property is in sharp contrast to the finely-tuned RK point on the square lattice. The authors make some connections to recent experimental observations in twisted bilayer graphene. In the appendix, results for a similar model defined on the Kagome lattice are presented.

The paper is very well written and contains a concise yet self-contained presentation of the main results. Identifying an exactly solvable model exhibiting a topologically ordered metallic state is a fundamental topic with direct links to experiments in strongly correlated metals. Therefore, I recommend publication in SciPost.

Below, I list a few points which and I would like the authors to consider.

1. In contrast to the square lattice case that without fine-tuning is always confining, on a triangular lattice, a confinement-deconfinement transition is allowed. Could the authors comment on the nature of such a transition? In particular, its effect on the fermionic degrees of freedom.

2. Could the authors provide a physical understanding of the electronic dispersion? In particular, could the band minimum at the M point be explained through symmetry-based arguments?

3. I find that the discussion on possible connections to twisted to bi-layer graphene physics is slightly unclear. While I understand that inevitably such a link is heuristic, it might be worth devoting a full section to make the discussion more precise.