Multi-pulse Interferometric FROG

Abstract

By now you're well aware that in just ten years the measurement of ultrashort light pulses has advanced from practically impossible to nearly indispensable. FROG and related techniques see daily use in labs around the world. Moreover, FROG's self-consistency checks have firmly established it as the gold standard of pulse measurement. In addition, a rarely touted feature of FROG is that it is a type of ultrafast BOXCAR integrator, gating only the pulse and not undesired cw backgrounds. Indeed, if your goal is to characterize a pulse, FROG is undoubtedly the right choice.However, if your aim is to characterize the interaction of a light pulse with something else, e.g., in a pump-probe experiment where you want to time-resolve the optical properties of a laser-excited material, then FROG (or any other self-gating technique) may actually be a poor choice. For example, imagine placing in a beam a I-cm thick piece of fused silica, which only negligibly distorts a 30-fs 800-nm pulse. As a result, the FROG trace of this pulse will be unaltered. However, relative to propagation in air, the pulse is delayed by some 15 ps in time and nearly 6000 wavelengths in phase! Because FROG doesn't measure the zeroth-and first-order spectral phase terms, it doesn't see these effects. Usually this is desirable, but occasionally it isn't. Spectral interferometry (SI) does see these terms, but SI signals are often obscured due to SI's inability to gate a weak signal out from a cw background. As we shall show in this chapter, one can combine the advantages of both FROG and SI by using what we call Multi-pulse Interferometric FROG, or MI-FROG. Before delving into the details of MI-FROG, let's first look at some of the available techniques for materials studies, namely the inter-related techniques of spectral blue-shifting [I], spectral interferometry (SI [2-4]), and FROG [5], and its variants TREEFROG [6] and TADPOLE [7]). Each of these has been successfully used to extract from spectral power density measurements details of ultrafast phase distortions. The first two of these techniques, being linearoptical effects, are powerful tools for measuring constant (or "DC") and slowly varying time-domain phase distortions with extremely high, milliradian, sensitivity. On the other hand, the optically nonlinear FROG accurately recovers only nonlinear variations in spectral phase of an optical pulse and is oblivious to DC and slowly varying terms. Thus FROG provides a complementary diagnostic, yielding only the higher-order phase distortions. One way to combine the advantages of FROG and SI, called TADPOLE [7] and discussed in the previous two chapters, utilizes the time-domain intensity and phase extracted from a standard FROG apparatus to fully deconvolve the frequency-domain