In this case, FLUC data are often simpler to analyze by direct band integration methods (e.g. moment analysis), leading to very straightforward interpretations. For example, one popular application of FLUC is probing solvent relaxations around a chromophore through the observation of picosecond and femtosecond dynamical Stokes shift induced on its fluorescence band. This approach has proven very valuable to analyze solvation dynamics in water [3], and hydration dynamics around proteins, a problem of great biological importance [4, 44].
3.4.4 Kerr‐Based Femtosecond Fluorescence Spectroscopy
Kerr‐based femtosecond fluorescence spectroscopy exploits the optical Kerr effect, through which a material becomes birefringent under the action of the intense electric field provided by a light pulse [48, 49]. As for upconversion method, the sample is pumped by a laser beam and the consequent fluorescence is collected. Unlike FLUC, the emission is focused into an isotropic Kerr medium (glass plate or solvent) arranged between two crossed polarizers intercepting the fluorescence path. In standard conditions, the output after the second polarizer is zero (closed shutter condition). When the gate pulse passes through the Kerr media, it opens the shutter, by inducing a transient change in the refractive indexes of the medium that makes it birefringent. Thereby, the emission traversing the Kerr cell becomes elliptically polarized, a portion of it passes through the second polarizer, and reaches the detector. Changing the delay between the pump and the gate then allows to record fluorescence kinetics. In can be shown that the gate pulse should be polarized at 45∘ with respect to the excitation in order to obtain maximum output [48]. Considering that a phase matching condition is not required in this scheme, a broadband detection is easy and limited only by the optical absorption of the used components and of the detection device.
The temporal resolution is determined by whatever is shorter between the cross‐correlation between excitation and gate and the cross‐correlation between the excitation and the intrinsic response function of the Kerr medium. In fact, if the Kerr medium reacts slowly to the gate modifications, then the temporal resolution is dominated by the latter. For example, with an excitation of ≈50 fs at 475 nm and a gate of 40 fs, the time resolution is ≈120 fs using a 1 mm‐thick fused silica plate as a Kerr medium [49].
3.5 Femtosecond Stimulated Raman Spectroscopy
Femtosecond time‐resolved methods are not limited to electronic spectroscopies, but can also be extended to meet the advantages of vibrationally sensitive techniques, such as infrared absorption and Raman. In this regard, this section addresses ultrafast stimulated Raman spectroscopy, which is one of the most powerful methods to combine the structural sensitivity and chemical selectivity of vibrational methods with the power of femtosecond time resolution [50–55].
3.5.1 The Experimental Method
Classic Raman spectroscopy exploits the Raman effect, discussed in detail in Chapter 5 of this book, that is the inelastic scattering of light associated to the generation or annihilation of vibrational quanta of energy ω V during the interaction of the light beam with the system under study. A typical steady state Raman experiment interrogates the sample by a laser beam at ω, and analyzes the spectral distribution of photons scattered in all directions. In the most common case (Stokes Raman scattering), part of the scattered radiation is found to be red‐shifted to ω′ = ω − ω V, because an energy ω V has been spent to excite phonons of this frequency within the sample. Thus, spectral analysis of the scattered radiation reveals a series of peaks at various ω V reporting on the vibrational mode structure of the system. Vibrational frequencies being much smaller than visible ones, the exciting beam must be narrow enough in frequency to avoid spectral overlap between ω and ω′. Therefore, the vibrational spectral resolution is of the order of the FWHM of the Raman pump beam. Raman spectroscopy is a powerful probe of the vibrational mode structure of any physical system and displays a sensitive dependence on both electronic and molecular structure. Thus, it can be used for accurate chemical recognition and to reveal, with high specificity, any structural changes of the system at study.
On these grounds, femtosecond time‐resolved Raman (TRR) experiments can be conceived as a powerful probe of excited state dynamics of any photoexcited physical system, overcoming the lack of structural sensitivity typical of absorption and fluorescence spectroscopies. In particular, Raman spectra being characteristic of any given state of the system, one should expect strong changes of the vibrational mode pattern whenever the system transitions to an excited state, and subsequently evolves. Following these changes by TRR can be used to extract detailed information on any structural changes or photochemical processes initiated by photoexcitation.
Broadly speaking, TRR spectroscopy is conceptually similar to TA, except, in this case, the observable becomes the change in the Raman spectrum of the sample between a pumped and unpumped situation [52]. TRR is not the only example of femtosecond‐resolved vibrational spectroscopy. An alternative is provided by mid‐infrared TA experiments, which are the time‐resolved counterpart of infrared absorption experiments [56, 57]. Similar to TRR, infrared TA can also probe the vibrational mode structure, but one advantage of TRR spectroscopy is the exclusive use of visible‐ or near‐infrared‐light pulses, which are easier to generate and detect, especially if broadband pulses are needed. Furthermore, due to the higher energy, visible‐light pulses are inherently shorter in time, ultimately providing better time resolution [51]. Another alternative is given by time‐domain techniques founded on TA, such as impulsive stimulated Raman scattering (ISRC) [58], which works best with low‐frequency modes and generally displays lower signal‐to‐noise ratios than state‐of‐the‐art TRR. For these reasons, TRR is generally recognized as the most powerful approach to achieve vibrational spectroscopy with femtosecond resolution, and will be discussed in detail in the rest of this section.
3.5.2 Typical Experimental Setups
The simplest and most straightforward path to implement a TRR experiment is to use a pump–probe approach very similar to TA experiment, except for the fact that the probe beam is used as a Raman pump. Thus, instead of measuring the change in probe transmittance, which is the basis of TA, one can achieve TRR by collecting the isotropic Raman signal produced by scattering of the delayed probe pulse, and analyzing it in a spectrograph as a function of pump–probe delay [54]. However, this method, named probe‐induced Raman scattering, is made very difficult by the very small cross section of Raman scattering, and also brings inherent limitations in time resolution. In fact, considering the time–energy uncertainty relation, the probe beam used as a Raman pump beam must be relatively broad in time to be narrow enough in the spectral domain. For example, a pulse centered at 550 nm, with a FWHM of 1 nm will guarantee a ω V spectral resolution of ∼30 cm−1, which may be adequate to probe a Raman spectrum, but with a transform‐limited pulse duration as long as 450 fs. This problem is fundamentally unsurmountable and limits the time resolution to the picosecond range.
In fact, one of the most effective approaches to carry out a TRR experiment is use a more complex scheme, initially proposed at the end of the 1990s [55], based on what is called the stimulated Raman process. The relation between spontaneous and stimulated Raman process is similar to that between spontaneous and stimulated emission: shortly, the Raman scattering cross section into a given photon mode is strongly enhanced whenever that particular mode is already populated by photons provided by a seed beam. Therefore, stimulated Raman process can be used to probe the same properties as spontaneous Raman process, but with potentially much stronger intensity.
In a typical time‐resolved stimulated Raman experiment [5559–61], such as the one depicted in Figure
