# Mirror-count doubling ablation

<div class="status-banner"><strong>COMPLETE</strong> · paired 7 ps FDTD · completed 2026-08-21T14:29:48+00:00</div>

This is a controlled test of whether the selected curved cavity was limited by finite Bragg-mirror length. The original 8 fronts per side are preserved exactly; 8 additional fronts at the terminal period, duty, curvature, and aperture extend each mirror to 16. No defect, inner mirror, slab, material, grid, source, y extent, or fitting rule changes between the two solves.

| Metric | control | doubled mirror | ratio |
|---|---:|---:|---:|
| fitted Q | 21.068 | 23.67 | 1.123× |
| atom-frequency LDOS | 0.020774 | 0.023863 | 1.149× |
| near-atom peak LDOS | 0.020896 | 0.025049 | 1.199× |
| total guided β | 8.197% | 2.405% | 0.293× |
| guided Purcell | 0.0017028 | 0.0005738 | 0.337× |
| fractional detuning | 0.551% | 0.073% | 0.132× |

**Interpretation:** Yes: the added terminal mirrors raise fitted Q by 12.3% and atom-frequency LDOS by 14.9%. They are not a complete device improvement, because guided beta changes by -70.7%.

The matched physical interior is 16.80 × 5.70 × 3.40 µm. The y extent remains exactly 5.70 µm. The x extent is enlarged only if needed to keep the longer mirror and terminal guide away from the x-PML; both control and extended devices are replayed in that identical box.

![Literal control and doubled-mirror geometries](../_static/generated/curved_mirror_doubling_geometry.png)

![Matched LDOS spectra and atom-frequency Ez fields](../_static/generated/curved_mirror_doubling_evidence.png)

## Integrity checks

- original fronts preserved exactly: `True`;
- y extent unchanged: `True`;
- source control reused in an identical domain: `False`;
- single-etch 2-D extrusion and hard 1 µm all-air gap retained;
- grid spacing: 25 nm;
- LDOS / outer-radiation frequency counts: 237 / 25;
- symmetry: `(1, 1, -1)`;
- source winner: `final_audit_5d2d8b7899`.

{download}`Machine-readable paired result <../_static/generated/curved_atom_mirror_doubling_v1.json>`