Time-autocorrelated two-photon counting technique for time-resolved fluorescence measurements

We describe a new instrumental technique for the excitation, acquisition, a nd analysis of fluorescence decays from a variety of substances, in the pre sent case plastic scintillators. The fluorescence is excited by beta partic les from a radioactive source (100 mu Ci Sr-90). A random photon from the r esulting fluorescence decay provides a trigger pulse to start a time-to-amp litude converter (TAC), while another random photon from the same beta-exci tation event provides the stop pulse. The optical components and geometry f or detecting these two photons, i.e., the two photomultipliers (PMT), the f ilters, and the pulse counting system, are identical. As a consequence, the measured fluorescence signal is the autocorrelation function of the fluore scence decay from the sample. A delay line of 50 ns is inserted between the "stop" signal PMT and the TAC so that those "stop'' pulses which arrive be fore "start pulses'' also are recorded. Thus the acquired fluorescence sign al versus time is symmetric about the delay time and contains twice as many counts as without delay. We call the new technique the "time-autocorrelate d two-photon counting technique'' (TATPC) in distinction to the conventiona l "time-correlated single-photon counting technique'' (TCSPC). We compared both techniques with the same equipment and scintillators, where in the TCS PC case, a beta particle is used for the start of the TAC instead of a rand om photon in the TATPC technique. We find that under similar experimental c ircumstances, the signal count rate with TATPC is about 50 times larger tha n with TCSPC. The new method is well suited for obtaining fluorescence deca y times from plastic scintillators, which we use in this article to exempli fy the technique. More generally, beta-particle excitation in combination w ith TATPC should prove useful for materials with high energy levels or band gaps, which cannot be excited with pulsed lasers in the visible region. Th e length of our excitation pulse is less than 20 ps and is negligible compa red to the temporal response of about 1 ns of the rest of the apparatus. By employing mathematical deconvolution, we are able to measure fluorescence decays from the subnanosecond range and to longer times. (C) 1999 American Institute of Physics. [S0034-6748(99)03201-3].