Design and performance analysis of a picosecond-pulsed laser Raman spectrometer for fluorescence rejection in Raman spectroscopy
Many attempts have been made to reduce fluorescence backgrounds in Raman spectra. A critical appraisal of fluoresence rejection techniques reveals that while many techniques are available which improve the Raman/fluorescence ratio (R/F), very few actually increase the spectral signal/noise (R/N), which is the most important parameter. Temporal-resolution of Raman and fluorescence photons was investigated in this laboratory, using a picosecond-laser system and gated photon detection. Two detection methods were evaluated. The first, an intensified diode array detector (DAD), could be gated "on" for periods of ca. 5 ns, at rates of up to 5kHz. This gave a 5-fold increase in R/F, but a slight reduction in R/N, for a fluorescor with τ(_f) ̴̱ 1O.5 ns. The R/N degradation arose as a result of the low laser output intensity at kHz pulse rates, rather than inefficiency in fluorescence rejection. The second method used a continuously-operated photomultiplie tube (PMT), and time-correlated photon counting with ca. 1 ns timing-resolution. This yielded R/F and R/N improvements of ca. 15 and 3 respectively (τ(_f) ̴̱ 12 ns).Although efficient fluorescence rejection was obtained with each system, the corresponding R/N enhancements were not practically significant. However, the development of theoretical models describing the performance of each system has identified modifications which should give valuable improvements. These include the use of a laser with MW peak powers at kHz pulse rates (DAD system), and use of a microchannel-plate PMT with 50 ps timing resolution. When these (and other) modifications are made, significant R/N enhancements (ca. 7 and 13 (DAD and PMT systems respectively)) are expected, thus enabling the study of the majority of "real world" samples. In addition, the limiting theoretical and practical performance of time-resolved rejection is considered, and several hitherto unreported aspects of the behaviour of the laser and detection systems are discussed. Other techniques were also evaluated, in particular utilising the differing Raman and fluorescence response to variations in laser intensity. While the non-linear fluorescence responseto intensity variations of cw lasers has been previously exploited, simple calculations indicate that the use of high-powered pulsed sources could allow discrimination at ca. 100-fold lower average powers. However, a satisfactory test of the calculations requires the construction of apparatus not presently available in this laboratory.