Rise time and fall time of a analog pulse

It is observed that a pulse generated from a detector (scintillator detector with PMT) usually have short rise time i.e. sharp while rising and longer fall time. Does anyone know the reason behind this ?

Because your sensing-node impedance is low when the
photoconductive / photovoltaic event is active, and high
once the carriers have been swept out. So you have two
different time constants at play. This will be magnified by
any loading that attends the probe or front end elecronics.

For purposes of illustration, consider a photodiode which
sources 10uA with 10V back-bias, and leaks 10pA, looked
at with a 100fF FET-probe. Assume I-V linearity, although
this is seldom entirely true.

10V/10uA=1Mohm 1Mohm*100fF=100nS rising
10V/10pA=1Gohm 1Gohm*100fF=100uS falling

You could improve the falling edge with a shunt conductance
without badly affecting the rising edge sensitivity - a 10Mohm
shunt, if it adds trivial capacitance, would pull in your fall
time to 1uS while costing you 10% of your input signal
current amplitude.