Setup
  (layers from the top down: the bottom one, the substrate, is the first laid down)
  layers:
    cladding       half-space   air, n 1.000000
    slab           252 nm       Si less its holes (air), n 3.474139
    substrate      half-space   air, n 1.000000
  shapes:
    on slab: a triangular lattice of a = 420 nm, holes of radius 121.8 nm (0.29a), 21 × 15 sites less three in a row (L3); the end holes moved out by 0 and 0.15a
  wavelength: silicon taken at 1.57 µm; the pulse at a/λ = 0.264 ± 0.04, an E_y current at the centre
  window: the lattice and a lattice constant beyond, 0.7a of air above and below, CPMLs of 12 cells
  grid: 8 cells a lattice constant, subpixel smoothing (diagonal), Courant number 0.9; the record 300 µm/c after the pulse
  solver: 3D FDTD; the mode's frequency and Q by harmonic inversion
  tolerance: the measured 1.565 to 1.58 µm widened by 2 % (1.5 % at 12 cells and more); Q's rise and the shifted Q within a factor of 2; the body's volume to 1e-7
  materials:
    Si: a table of 35 wavelengths, 1.2 to 14 µm, between them natural cubic splines; valid 1.2 to 14 µm
      source: H. H. Li, J. Phys. Chem. Ref. Data 9, 561 (1980), doi:10.1063/1.555624, at 293 K
      crystalline silicon; Table 1, 293 K, checked against the paper; uncertainty ±2e-3 (Li)
      n 3.474139 at 1.57 µm

Akahane et al.'s L3 cavity: a = 0.42 um, slab 0.6a, holes of radius 0.29a, 21 x 15 sites, the end holes moved by 0 and 0.15a
the cavity as one body (exact: no grid): its volume from its fills over a/4 cells 8.3e-9 off T(|A| - N pi r^2); 0 of 1887 points on the wrong side
  volume, relative difference            0.000000   fills 0.000000   ± 1e-7      ok
  points on the wrong side                      0   paper 0                      ok
3D FDTD, 8 cells a lattice constant (52.50 nm cells, 200 x 135 x 40), the record 300 um/c after the pulse:
  unshifted: a/lambda 0.2635, lambda 1.5941 um, Q 5574
  shifted 0.15a: a/lambda 0.2618, lambda 1.6042 um, Q 56156 (10.1 times)
  lambda unshifted (um), 8 cells         1.594057   paper 1.5725     ± 3.895000000000006e-2 ok
  lambda shifted (um), 8 cells           1.604180   paper 1.5725     ± 3.895000000000006e-2 ok
  log2 of Q's rise, 8 cells              3.332762   paper 3.321928094887362 ± 1e0 ok
  log2 of Q shifted, 8 cells            15.777164   paper 15.45763738099176 ± 1e0 ok
all 6 within tolerance of the paper
