I fitted two peaks in each spectrum with a Lorentzian plus a local linear background. The allometric fits use \(y=aP^b\); the reported uncertainties are standard errors on \(b\).

| Quantity | Peak 1 near 1.33 eV | Peak 2 near 1.54 eV |
|---|---:|---:|
| Intensity exponent \(b\) | 0.9233 ± 0.0121 | 0.6631 ± 0.0588 |
| Position exponent \(b\) | 0.00447 ± 0.00008 | 0.00235 ± 0.00031 |

The second peak fit is less certain, particularly at low power, because it is broad and close to the high-energy upturn. The per-power fitted intensities and positions are in [lorentzian_peak_fits.csv](<BENCH>/raw/commercial/luna-high/bench/t-pl/run-21/lorentzian_peak_fits.csv).

Plots: [intensity vs power](<BENCH>/raw/commercial/luna-high/bench/t-pl/run-21/intensity_vs_power.png) · [peak position vs power](<BENCH>/raw/commercial/luna-high/bench/t-pl/run-21/peak_position_vs_power.png)

RESULT: b = 0.9233 ± 0.0121