Excelitas (and other) SPAD detectors can show an afterpulse peak several 10s of ns after the main IRF/decay peak. This article describes three ways to suppress the afterpulse using the UniHarp software (MultiHarp/PicoHarp/HydraHarp electronics).
The problem
Raw TCSPC data can contain an afterpulse peak appearing several tens of nanoseconds after the primary detection peak. In one example dataset, the afterpulse appears roughly 28 ns after the first pulse and is clearly visible as a secondary bump in the histogram.
Option 1: Coincidence manipulator
To only count the photon from the first pulse in each laser period, use a coincidence manipulation between the detector channel and the sync channel. Only the first photon in the coincidence window is then counted, which removes the afterpulse from the resulting histogram.
Recommended settings (in the Manipulators > Coincidence panel):
Channels: select the channel with your sync signal and the detection channel (e.g. 0 and 1)
Window Time: set to the laser repetition period (e.g. for a 5 MHz laser, this is 200,000 ps / 200 ns)
Coincidence Mode: Count Once
Coincidence Time: Last
-
Keep Channels: enabled
More on the Coincidence Manipulator functionality can be found in the snAPI documentation: snAPI.Main.Manipulators.coincidence
With this manipulator applied, the resulting histogram no longer contains the afterpulse peak while the main IRF/decay peak is preserved.
Option 2: Programmable dead time
Alternatively, a dead time can be set on the detection channel in the Channel Settings panel (Dead Time [ns] field). If the configured dead time is longer than the temporal separation between the first pulse and the afterpulse, the dead time will suppress the afterpulse while the first pulse remains unaffected.
Option 3: PNR (photon number resolving) manipulator
A third approach combines the Delay manipulator with the PNR manipulator:
Delay the detection channel by 1 ps to create a duplicate/reference timing channel.
Create a PNR manipulator using the sync channel as the reference channel, the (undelayed) detection channel as the X channel, and the delayed channel as the Y channel.
Which option should I use?
The coincidence manipulator (Option 1) is the most direct and commonly used solution, and works well as long as the afterpulse falls within the same laser period.
The dead time approach (Option 2) is simplest to configure but will also block any genuine photons arriving during the dead time window, which can be relevant at higher count rates.
The PNR manipulator (Option 3) offers a more advanced, channel-based way to distinguish real pulses from afterpulses.