Skip to content

General 1-D outgoing-pole cavities

Extend the Quan–Lončar outgoing-pole optimizer from Zernike ellipses to arbitrary quasi-periodic nanobeam cavities, then reproduce and improve three published devices.

Campaign notebook Engine

Objective

  1. Represent Quan 2011, Sawfish 2022, and Fryett 2017 in one operator.
  2. Recover high-\(Q\) poles near the published wavelengths.
  3. Raise model \(Q\) by at least \(2\times\) with implicit \(dQ/dp\).
  4. Confirm each gain with Tidy3D (≤15 flexcredits).

Trust level

Analytical ≥2× on all three families; FDTD ≥2× only on Quan. Production Quan (P2-z1 / y3-z1) raises model \(Q\) 7722 → 16835 (\(2.180\times\)) and matched 12 ps Tidy3D 15877 → 56909 (3.58×). Sawfish and Fryett analytical 2× do not raise 3-D \(Q\) (0.76× / 0.98× / 0.996× / 0.993× / 0.968×). Paper-ellipse y1-z2 seed model \(Q=167\) is comparable to FDTD \(Q=207\); unique-\(r_y\) and unique-center 2× are null directions of 3-D \(Q\). Ledger 0.888 / 15. See the campaign notebook.

Distinct observables

Analytical \(Q\) from \(\tilde\omega\) and FDTD ringdown \(Q\) are never substituted for one another. The Quan campaign already showed that directions can transfer while magnitudes drift by more than \(10\times\). This engine keeps that contract.