SymPhoTime can reconstruct an Instrument Response Function (IRF) mathematically from the fluorescence decay data itself, without needing a separately measured IRF. This is sometimes called a calculated or reconstructed IRF.
If the reconstructed IRF looks irregular — showing multiple peaks or oscillations — this is almost always a noise artefact, not evidence of a hardware problem. Importantly, an irregular-looking reconstructed IRF does not necessarily mean your fit results are wrong. There is a subtle but important reason for this: the oscillations typically appear precisely in the region where the reconstruction problem is mathematically ill-conditioned (the transition between the IRF and the "IRF-free" decay). This same ill-conditioning means that noise in this region has little influence on the actual fit results. In other words, the fit can be perfectly valid even when the reconstructed IRF looks noisy. That said, a noisy IRF is obviously not ideal.
How the calculated IRF works (in SymPhoTime)
The reconstruction algorithm derives the IRF from the rising edge of the fluorescence decay histogram — the steeply ascending part at the beginning of the decay. This region contains information about the instrument timing response. However, the algorithm is inherently sensitive to noise in this transition region, because the problem is mathematically ill-conditioned near the boundary between the IRF and the "IRF-free" part of the decay.
Why the IRF may look noisy or irregular
The most common cause is insufficient photon statistics in the decay — for instance, because the measurement was too short or the count rate was low. Even a difference of roughly one order of magnitude in total photon counts can be the difference between a clean and a noisy reconstructed IRF.
Solutions
- Measure longer: Acquiring more photons improves the statistical quality of the rising edge and typically yields a cleaner reconstructed IRF. This is the most reliable fix if a clean-looking IRF is important (e.g. for figures).
- Apply smoothing: Within the TCSPC analysis in SymPhoTime, you can smooth the decay before the IRF reconstruction step. This reduces noise-driven oscillations at the cost of some temporal resolution.
When to use a measured IRF instead
If the calculated IRF is consistently irregular despite acquiring sufficient photons, or if your sample has a very short lifetime (close to the instrument timing resolution), it is generally preferable to measure the IRF directly rather than reconstruct it.