The file is already merged — a single workbook `Book1` with 15 power columns (5–100 µW) plus the eV axis. No Si folder present, so no background subtraction needed. Suggested fit window: **x = 1.23–1.45 eV**.

Now despiking (mandatory before any fit):



![Power Fit]([file link removed])

## Lorentzian power-dependence analysis — `powerdep_manuscript.opj`

Pipeline: inspect → despike (130 spikes removed, σ=4) → Lorentzian fit on all 15 spectra (window 1.23–1.45 eV) → allometric fit. All 15 columns converged to **HIGH** tier (R² = 0.887–0.984, median 0.981).

**Per-power Lorentzian results:**

| P (µW) | Intensity (cnt) | x_c (eV) | FWHM (meV) |
|---|---|---|---|
| 5 | 98.8 | 1.3206 | 62.4 |
| 10 | 189.9 | 1.3244 | 64.4 |
| 20 | 353.8 | 1.3278 | 62.7 |
| 30 | 516.0 | 1.3300 | 61.3 |
| 50 | 818.3 | 1.3328 | 59.0 |
| 70 | 1140.4 | 1.3343 | 57.1 |
| 100 | 1714.6 | 1.3356 | 54.4 |

*(full table in the CSV below)*

**Left panel — Intensity vs Power, allometric fit $I = a\cdot P^{b}$:**
- $a = 16.28 \pm 1.81$
- $b = 1.0048 \pm 0.0261$, R² = 0.995 → essentially **linear (excitonic)** emission, no saturation or bimolecular recombination over 5–100 µW.

**Right panel — Peak position vs Power:**
- x_c shifts from **1.3206 eV (5 µW) → 1.3356 eV (100 µW)**, a monotonic **blueshift of ~15 meV** (~5.0 meV/decade), consistent with band-filling / state filling at higher excitation.

📥 [Download full results CSV]([file link removed])

RESULT: b = 1.0048 ± 0.0261