“…Theirs uses the FBI transform in a compact region of phase space and and Weyl-Hörmander calculus with a suitable metric near infinity, while ours works globally using the Wick quantization and some standard Weyl calculus. Hitrik and Sjöstrand attained a similar estimate for certain one-dimensional non-self-adjoint Schrödinger operators [7], with ellipticity assumptions on the potential. Also, Dencker, Sjöstrand, and Zworski showed that for non-self-adjoint semiclassical operators, under suitable assumptions including ellipticity at infinity, the resolvent can be similarly estimated in a small region near a boundary point of the range of the symbol, away from critical values of the symbol [3].…”