An apparatus for thermal lens spectrophotometry of liquid samplesIsconshucted udngapuleednhgen laserexdtstkn source a m l a w laser probe. Both laser beams are passed through a fused a k a optlcal fiber prlor to bdng Hghtly focused Into the sample cell. The optical flber k used to reduce mode and pointtng varlatlons in the laser outputs. A dlgltal correlation filter is developed and used for estlmatlon of the thermal lens signal. T h k dlgltal estbnatlon procedure exhlblts a consMerably enhanced slgnal-to-nolse ratio over that of other estknatlon methods. Moreover, It is relatively Immune to background Interference and rapM In estlmatlon time. Finally, the effectlve sample path kngth for focused beams In pulsed laser exclted thermal l e n sis examlned from a theoretical b a a . It is found that the slgnal will be very near maximum for cell path lengths on the order of 4 times the Raylelgh range.Pulsed laser excited photothermal spectrophotometry is an ultrasensitive technique for concentration estimation of analytes that do not fluoresce subsequent to absorption of light at the excitation wavelength ( I ) . There are several potential advantages to pulsed laser excitation, over that of continuous wave (CW) excited photothermal lens spectrophotometry (TLS), that have been discussed in a number of recent publications (2-8). The pulsed laser excitation techniques have large theoretical enhancement factors at a small excitation beam radius (2-4), the signal magnitudes are not dependent on solution flow at moderate rates (7,8), in contrast to the related CW laser excited techniques which are very sensitive to flow (!+I1 1, and there is a potentially higher signal-to-noise ratio (SNR) due to the high-frequency components that make up the rapid rise time of the TLS signal (6).However, pulsed laser are in general noisy sources. The pulse energy can vary substantially, and mode or pointing noise can result in a large errors. Mode noise is due to transverse mode fluctuations that will result in different focused beam spatial profiles in the sample, and pointing noise will result in minor shifts in the focus position. The errors associated with the latter two noise sources are due to the sensitivity of the photothermal spectrophotometric signal to the spatial overlap of the pulsed excitation laser beam and that of the CW probe. Pulse energy variations can be corrected for in a simple fashion by monitoring this energy, but the mode and pointing variations are very difficult to detect and the corrections would not be simple. Pointing noise in the pump and probe beams has been found to be a major limiting factor in the accuracy of the pulsed laser excited photothermal deflection technique (6) and appears to be one of the limiting factors in the CW TLS technique as well (12).