Calibrate a probe and refine R²
DiffAtOnce Legacy · web

00:00 — Choose a clean reference
The probe study is independent of the active analysis. In this simulation, we search near four point seven parts per million using a fixed local window and monitor each spectrum's original maximum.

00:13 — Compare all ramps
Six groups remain visible: three usable ramps and three diagnostic rows. Calculate all slopes lets us compare the fits while each acquisition retains its conditions. Review the chosen calibration row without averaging different diffusion times or ramps.

00:31 — Inspect the proposal
Refine R² targets 0.9999 while retaining enough points and gradient range. This example proposes twenty-two points instead of twenty-three, excluding point eleven. The proposed selection still needs your review.

00:48 — Apply to this row only
Apply selection changes the included points for this row and invalidates its previous calibration. Measured values remain unchanged. Review the slope and conditions, then evaluate the reference again before saving.

01:03 — Restore the selection
Undo refinement restores the previous selection. This screenshot follows the real undo action, with all twenty-three original points included again. An optimized R² does not demonstrate a physically valid probe calibration.

01:19 — Evaluate a compatible reference
The simulation uses pure water at twenty-five degrees Celsius and a compatible literature diffusion coefficient. Evaluation derives the gradient from the reviewed fit and conditions. This demonstration does not save or share an experimental probe.

Synthetic data demonstration. This is not experimental validation or a confirmed molecular identification.