Skip to main content

RATE_TOLERANCE

Constant RATE_TOLERANCE 

Source
pub const RATE_TOLERANCE: f64 = 0.01;
Expand description

The tolerance of the rate refinement (#339) and of the rate gate of lindhard-cli/examples/inelastic_low_energy.rs (#291), fixed there before any value was measured: the largest relative error of the interpolated 1/λ over RATE_WINDOW_EV above the Fermi level must not exceed 1 %. It is the statistical error of the most precise committed default δ entries, rounded up: in validation/experiments/se_yield_results.json the Poisson floor of Al default is 0.0585 on δ = 6.85 at 600 eV (0.85 %) and 0.0571 on δ = 6.52 at 400 eV (0.88 %). Other configurations are more precise: Al barrier-off reaches 0.60 % (600 eV), and at 400 eV the Al barrier-off, cutoff-band-bottom, phi-low and phi-high entries are 0.63 to 0.90 %; the Au, Cu and Si default entries at 400 eV are 1.4, 1.7 and 2.8 %. A grid error below the tolerance cannot be told from the noise of those tables if δ responds at most in proportion to the rate in that window; that response is not measured here.