Fit models¶
This page summarizes the analytical expressions implemented in
TGAnalysis.model1, TGAnalysis.model2, and TGAnalysis.model3.
Notation¶
\(t\): time (ps).
\(t_0\): time-zero shift for the primary decay channel (ps).
\(t_{02}, t_{03}\): optional additional time shifts used by Models 2–3 for extra exponentials (ps).
\(\sigma\): Gaussian instrument-response width (fitted in ps; reported as fs in
params_fit).\(k, k_1, k_2, k_3\): decay rates in ps \(^{-1}\).
\(\tau = 1/k\): decay time (reported in fs in fit summaries).
\(\mathrm{erf}(\cdot)\): error function.
Model 1 (model1)¶
A mono-exponential decay convolved with a Gaussian response plus an offset step term:
Model 2 (model2)¶
Model 1 plus a second exponential decay channel:
where:
Model 3 (model3)¶
A sum of three Gaussian-convolved exponential channels with independent time zeros:
Bounds and reported \(\tau\)¶
Fitted decay rates \(k\) (in ps \(^{-1}\)) are converted to
\(\tau = 1000/k\) fs when filling "tau", "tau2", and "tau3"
entries in params_fit. The Gaussian width sigma is stored in fs in
params_fit. Model 1 and Model 2 also expose "ampoff"
for the convolved offset term; for plotting compatibility across models,
Model 3 stores "ampoff" as NaN.
Practical note¶
Use tg_analysis.TGAnalysis.get_fit_parameters() to fit all loaded scans.
The method stores values and propagated errors in params_fit and reports
reduced \(\chi^2\) and \(R^2\) for each scan.