The field of optical microscopy experienced significant gains in resolution and speed following the introduction of lasers. Unfortunately, these gains came at the expense of sample degradation caused by the continuous flux of intense light. Taking advantage of the two-photon absorption process, Webb and Denk implemented a microscope based on the use of near-IR light pulses capable of causing simultaneous multiple fluorophore excitation. Two-photon microscopy is now widely applied in the biomedical imaging field due to the absence of out-of-focus photobleaching and reduced photodamage and fluorescence scattering. These advantages are brought about collectively by the inherent instantaneous peak intensity and narrow focal plane of excitation. Given that peak intensity increases with decreasing laser pulse duration, one would expect extensive use of available ultrashort (sub-10 fs) pulse laser systems in the field of biomedical imaging. However, most two-photon microscopes still use the same pulse duration that Webb and Denk used in 1990 (≈150 fs).
In principle, replacing a laser producing 150 fs pulses with one that produces 10 fs pulses should result in a factor of 15 improvement in signal. However, experimentally one would observe a net decrease of 30% in signal, with the loss of signal being caused by chromatic dispersion introduced by the high numerical aperture microscope objective lens (Figure 1). It is a common practice to distinguish between group velocity dispersion (GVD) and a third order dispersion (TOD). GVD causes different frequency components of the pulse to arrive at the sample at different times, effectively increasing the pulse duration, while TOD breaks the pulse into sub-pulses. A simple prism pair can compensate GVD. Such correction would cause a 2.2× increase in signal when using 10 fs instead of 150 fs pulses; unfortunately, the prism pair introduces a significant amount of additional TOD. Only by correcting both GVD and TOD, ensuring transform limited (TL) pulses (i.e. pulses with no dispersion), would the expected 15× improvement in signal for two-photon microscopy and an impressive 225× improvement for third-harmonic generation microscopy be observed.
For a given laser system with a fixed set of optics, it is possible to design a static optical device to eliminate GVD and TOD. However, practical correction for GVD and TOD for microscopy must take into account that the laser parameters and objective lenses are usually changed during imaging. Therefore, the ideal device must be able to measure and correct the femtosecond pulses at the focus of the microscope. Here we present the use of such an adaptive system capable of measuring and correcting all orders of dispersion, thus delivering TL pulses at the focal plane.