I treated the 15 power columns as the spectra and fitted two bands: near 1.33 eV and 1.54 eV. Band 1 was fitted to the raw spectrum; for the weaker Band 2, I smoothed the trace before fitting a Lorentzian with a local linear baseline. The allometric exponents below come from log–log regression; uncertainties are 1σ slope standard errors.

- **Intensity vs power:** Band 1, \(b=0.9193\pm0.0142\); Band 2, \(b=0.9775\pm0.0623\).
- **Peak position \(x_c\) vs power:** Band 1, \(b=0.004445\pm0.000104\); Band 2, \(b=0.002539\pm0.000405\).

Band 2’s fit is more sensitive to the rising background, especially at higher powers.

[Peak intensity plot](<<BENCH>/raw/commercial/luna-high/bench/t-pl/run-10/peak_intensity_vs_power.png>) · [Peak position plot](<<BENCH>/raw/commercial/luna-high/bench/t-pl/run-10/peak_position_vs_power.png>) · [Fitted intensities and peak positions by power](<<BENCH>/raw/commercial/luna-high/bench/t-pl/run-10/lorentzian_peak_fits.csv>)

RESULT: b = 0.9193 ± 0.0142