Hello Naresh,

In time-correlated single-photon counting, the decay signal that you measure, I(t), is typically not the information you want it.  The  fluorescence decay, f(t), the information you want, is convoluted with the instrument response funtion, g(t), of your system.  The instrument response function of your system can be measured using a scattering solution in place of the sample. In an ideal world this signal would be a Dirac delta function, however uncertainties arise do to the pulse width of your excitation, geometry of your cuvette, rise time of your detector, bandwidth of your amplifier, etc.  The relationship between the measured signal and the fluorescence decay is given by a convolution integral

                             I(t) = integral (limits from 0 to t) {g(t-t')*f(t')}dt'

Usually one has a software package that allows one to do an iterative reconvolution in the fitting routine to determine f(t). 

If g(t) has a temporal width much greater than f(t), the uncertainty introduced to the parameters obtained from your fitting routine will be significantly large; in other words your amplitudes and lifetimes will be highly questionable.  Hence, if you are measuring a time course with a 100 femtosecond duration, your instrument response function should have a width that is much shorter (say a few femtoseconds).

References:

"Chemical Applications of Ultrafast Spectroscopy", Graham Fleming, Oxford Press.

"Time-Correlated Single-Photon Counting", Desmond O'connor, Accademic Press.

Best,

John

Thank you, sir.