Advertisement

Topics in Current Chemistry

, 376:35 | Cite as

Multidimensional Vibrational Coherence Spectroscopy

  • Tiago Buckup
  • Jérémie Léonard
Review
Part of the following topical collections:
  1. Multidimensional Time-Resolved Spectroscopy

Abstract

Multidimensional vibrational coherence spectroscopy has been part of laser spectroscopy since the 1990s and its role in several areas of science has continuously been increasing. In this contribution, after introducing the principals of vibrational coherence spectroscopy (VCS), we review the three most widespread experimental methods for multidimensional VCS (multi-VCS), namely femtosecond stimulated Raman spectroscopy, pump-impulsive vibrational spectroscopy, and pump-degenerate four wave-mixing. Focus is given to the generation and typical analysis of the respective signals in the time and spectral domains. Critical aspects of all multidimensional techniques are the challenges in the data interpretation due to the existence of several possible contributions to the observed signals or to optical interferences and how to overcome the corresponding difficulties by exploiting experimental parameters including higher-order nonlinear effects. We overview how multidimensional vibrational coherence spectroscopy can assist a chemist in understanding how molecular structural changes and eventually photochemical reactions take place. In order to illustrate the application of the techniques described in this chapter, two molecular systems are discussed in more detail in regard to the vibrational dynamics in the electronic excited states: (1) carotenoids as a non-reactive system and (2) stilbene derivatives as a reactive system.

Keywords

Ultrafast laser spectroscopy Multidimensional spectroscopy Raman Vibrational spectroscopy Coherence spectroscopy Excited states Vibronic coupling Photoisomerization 

1 Introduction

The dynamics and function of complex molecular systems can be understood as resulting from the interaction between three different interacting sub-units composed of the electronic, vibrational, and environmental degrees of freedom. Focusing particularly on photoreactions—i.e., excited state reactions—they may be described as (non radiative) transitions between distinct electronic states of one or several molecules. Depending on the nature of the coupling responsible for such a transition in a given case, the photoreaction may correspond to a charge or energy transfer, or an internal conversion, an intersystem crossing, etc. Within the Born–Oppenheimer (BO) approximation, electronic states are characterized by multidimensional electronic potential energy surfaces (PESs) representing the energy of the electronic subsystem as a function of all internal nuclear coordinates (3 N-6 for N nuclei) treated as fixed, external parameters. The vibrational degrees of freedom, which drive the molecular system along the photoreactive path from the Franck–Condon state (i.e., initially produced by the photon absorption) to the photoproduct, contribute to the reaction coordinate. All other molecular (and solvent, in condensed phase) vibrational degrees of freedom constitute a large thermal bath—the environment—responsible for very fast energy relaxation and dissipation. There is not only fundamental interest in understanding how these three sub-units interact to perform a given photoreaction and govern its quantum yield, but also in understanding how such interactions shape the molecular functionality.

The mechanism of photoreactions may be elucidated by identifying the conformations and vibrational dynamics of transient electronic states successively populated, from the Franck–Condon state to the vibrationally and thermally relaxed photoproduct. Vibrational spectroscopies (i.e., Raman and IR spectroscopies) have long been exploited to reveal vibrational activity and conformations of stationary or transient molecular states. In this contribution, we will review recent experimental developments, which implement in various ways stimulated Raman scattering to monitor vibrational dynamics along the course of a photoreaction, with a time resolution typically below 100 fs, therefore allowing to resolve the vibrational activity accompanying ultrafast photoreactions. In condensed phases, the vibrational energy relaxation and dissipation to the environment occurs on the 0.1–10 ps time scale, which is in many cases faster than the photoreaction itself. We shall consider as “ultrafast” the photoreactions occurring on a similar time scale or faster. One difficulty inherent to their investigations comes from the fact that there is no time scale separation between the various relevant processes.

To investigate the dynamics and functions of molecular systems on ultrashort time scales, time-resolved UV–VIS spectroscopy has been used since the advent of femtosecond laser light sources. Time-resolved transient absorption—or so-called pump-probe spectroscopy—exploits the non-linear response of the complex system described above upon interaction with coherent laser light pulses. More generally, in the regime of weak-field light–matter interaction, the use of short, coherent laser pulses enables the preparation of controlled, coherent superpositions of molecular (i.e., electronic and vibrational) quantum states. The quantum evolution of these initial coherent states may be followed spectroscopically, i.e., interrogated by the interaction with another laser pulse, until decoherence (interaction with the bath) occurs on typical time scales of a few femtoseconds (electronic decoherence) up to a few picoseconds (vibrational decoherence) [1, 2, 3, 4]. After electronic decoherence has occurred, the time evolution of the laser-induced populations can be further observed spectroscopically.

Over the last decade, the investigation of coherence in molecular processes occurring in the condensed phase has become a frontier research topic in molecular quantum physics [5, 6, 7, 8, 9, 10, 11]. In this chapter, we will describe an application of femtosecond coherent multidimensional spectroscopy which engineers vibrational coherence in molecular systems and uses it as a spectroscopic tool. More precisely, the goal is to follow in time the vibrational coherence imprinted in the electronic excited states by the non-linear interaction with coherent laser light and exploit the peculiar spectroscopic signatures of such vibrationally coherent molecular states. This type of spectroscopy can be performed under several distinct experimental implementations, which have been named differently. Here we will summarize all these implementations under the name multidimensional vibrational coherence spectroscopy, or multi-VCS. This contribution does not deal, however, with other kinds of vibrational spectroscopies based on so-called rephasing mechanisms like 2D Raman or 2D infrared spectroscopies. These topics are reviewed in different contributions in this collection.

We will first shortly describe how VCS results from the third-order interaction (and higher-order for multi-VCS) of the molecular system with femtosecond light pulses. This initial description will assist the reader in understanding how the higher-dimensional versions of VCS are able to report on structural dynamics in excited states. The main experimental implementations in the time-domain (pump-Impulsive Vibrational Spectroscopy-pump-IVS, pump-Degenerate Four Wave Mixing-pump-DFWM, and Population-controlled IVS) and in the frequency-domain (Femtosecond Stimulated Raman Scattering-FSRS) will be discussed and compared. Finally, we will illustrate the success of multi-VCS at revealing a mechanistic understanding of ultrafast photoreactions in a selection of molecular systems: carotenoids and stilbene derivatives.

2 Introduction to Vibrational Coherence Spectroscopy (VCS)

The principle of vibrational coherence spectroscopy was demonstrated in molecules as soon as picosecond and then femtosecond laser pulses became available. In particular, laser pulses which are shorter than the period of nuclear motions in a molecule have a spectrum larger than the corresponding vibrational level spacing. With such a laser pulse, coherent superposition of vibrational levels, also referred to as vibrational wavepackets, may be produced impulsively in essentially any molecule. The spectroscopic signature of such vibrational wavepackets allows tracking molecular structural dynamics accompanying ultrafast photoreactions in molecules. Early investigations of vibrational wavepacket signatures in simple molecules in gas phase have pioneered the so-called field of femtochemistry [12].

The physical mechanism of VCS may be introduced by discussing a conceptually simple pump-probe experimental scheme. Let us consider a particular example where a short-enough, resonant pump pulse impulsively excites a molecular system, thus producing a non-stationary population, and the probe pulse is used subsequently to measure the absorbance of this pump-induced population. Considering the interaction with the pump laser as a perturbation of the molecular system, a very general result of the perturbation theory is the following. At the first order of the perturbative expansion, a coherent superposition of the two states coupled by the perturbation is produced. Such a superposition is called a “coherence”. Instead, a “population” (e.g., depopulation of the ground S0 state and population of the excited S1 state) is produced at the second order only. With the vocabulary of non-linear optics, this is rephrased as: A first interaction with the pump field (denoted by its wavevector k1) creates an electronic coherence while a second interaction (k2) creates an electronic population. Since the pump pulse is spectrally broad, both interactions (i.e., with k1 and k2) may occur with distinct spectral components of the laser spectrum, resulting in the population of distinct vibrational levels in the same electronic state, as illustrated in Fig. 1a, b. Under impulsive excitation (i.e., the pump pulse duration is shorter than the vibrational period and dephasing), a coherent superposition of vibrational states is produced, which in this case is labeled vibrational coherence. In addition, the second interaction may act on the ground-state component of the coherence and couple it to the excited state (Fig. 1a), but it may also act on the excited state component of the coherence and couple it back to the ground state (Fig. 1b). As a result, the pump-induced vibrational coherence is a vibrational wavepacket in the S1 electronic state in the first case, or in S0 in the latter case. Since these two cases correspond to two terms of the perturbative expansion at the same order, a short-enough laser pump pulse will in general produce vibrational wavepackets both in the ground and excited states [13, 14, 15]. Particular chirp (i.e., time-ordering of the spectral components within the short pump pulse) may be engineered to favor S1 or S0 wavepacket formation [16, 17, 18, 19, 20, 21]. The impulsive formation of a ground-state wavepacket upon interaction with an ultrashort resonant laser pulse is a Raman process, named “resonant impulsive stimulated Raman scattering” RISRS [22, 23]. It also operates with a non-resonant laser pulse (ISRS) [24].
Fig. 1

Generation and probing of vibrational wavepackets in molecules with impulsive pump-probe spectroscopy (see e.g., Ref. [25] for a detailed discussion). Provided the laser spectrum is broader than the vibrational level spacing, a vibrational wavepacket is produced at the second order of the perturbation theory (i.e., two “interactions”) either a in the excited state (i.e., one interaction with wave k1 on the “ket” side and the other with wave − k2 on the “bra” side of the density matrix element |0 > <0| representing the initial S0 population, thus generating a population |1 > <1| in S1) or b in the ground state (both interactions on the bra side, generating a vibrational wavepacket in S0). c A vibrational wavepacket produced by the first two interactions, for instance in S0 via an impulsive stimulated Raman process, is probed after a waiting time τ via a third interaction with the “probe” wave k3. This third interaction generates a third-order coherence, which radiates a fourth wave (wavy arrow) in the direction k1 − k2 + k3, imposed by the phase matching condition

In the sequential pump-probe scheme discussed here (i.e., the probe pulse does not overlap temporally with the pump pulse), the first two interactions with the pump field are followed, after a “waiting” time τ, by a third interaction with the probe field, as illustrated in Fig. 1c. The latter couples (i.e., creates a “coherence” between) the pump-induced population and another electronic state. The resulting (third-order) polarization of the molecular system radiates a field that builds up constructively in a fourth wave (to be detected), which propagates in a specific direction k1 − k2 + k3 imposed by the phase matching condition [25]. In the present case, where the first two interactions occur with the same pump pulse k1 = k2, the fourth wave propagates collinearly with the probe pulse k3. Hence, one detects the co-propagating third and fourth waves, which means detecting the fourth wave heterodyned by the incident probe field. Finally, this experiment may be described as the measurement of the linear absorbance of the probe beam (i.e., first order in the probe field), by the non-stationary population induced (at the second order of perturbation) by the pump [26].

The vibrational wavepackets produced by the quantum superposition of vibrational states results in the classical oscillation of the vibrational degrees of freedom (bonds elongations, torsions, etc.). These oscillatory molecular motions induce an oscillatory modulation of all the linear optical properties of the system such as absorbance, dichroism, birefringence, etc. [15, 22]. Hence, the pump-induced change in the spectrum or polarization state of the transmitted probe oscillates accordingly as a function of the waiting time τ. The sequential pump-probe experiment described here and performed with ultrashort laser pulses, so as to generate and probe vibrational wavepackets both in the ground and excited states, is referred to as impulsive vibrational spectroscopy (IVS). Fourier transformation of the oscillatory signals reveals the Raman activity of the system. For a ground-state wavepacket—RISRS or ISRS—the resulting vibrational spectrum coincides with the molecular Raman spectrum as measured directly in the frequency domain by conventional Raman spectroscopy [27, 28, 29]. The limited bandwidth of the laser pulses may, however, result in the attenuation of the relative intensity of the highest-frequency vibrational modes or even prevent from detecting them. In practice, 10–12 fs pulses are short enough to trigger and detect vibrational activities up to the 3000 cm−1 range. Hence, VCS of ground-state wavepackets is a time-domain equivalent of frequency-domain Raman spectroscopy. For excited-state wavepackets, time–frequency representations of mode-specific coherent oscillations was recently proposed as a spectroscopic tool for detecting CInt’s [30].

In fact, a broad variety of four-wave mixing spectroscopies have been developed for VCS, as depicted schematically in Fig. 2, and the IVS scheme introduced above to illustrate the physical mechanism at the origin of VCS is only one possible (Fig. 2a) implementation of VCS. Enhanced experimental control on the first two interactions (k1 and k2) may be achieved by engineering these two interactions with two distinct, non-collinear laser pulses instead of one (Fig. 2b). In this three-beam experiment, the signal is still generated in the k1 − k2 + k3 direction which is, however, no longer parallel to the probe beam k3. Consequently, the signal is detected on a dark background, in the so-called homodyne detection scheme. Note that homodyne and heterodyne detection schemes have been briefly reviewed in the first contribution of this collection and explained in detail in the literature [31]. This implementation has been named “transient grating”, because the signal can be understood as originating from the diffraction of the probe beam by the non-stationary population grating imprinted in the sample by the interaction with the non-collinear, interfering, first two beams. Two types of experimental realizations have been devised and named time-resolved coherent anti-Stokes Raman scattering (CARS) when pulses with different spectra are used, [32] or degenerate four-wave mixing (DFWM) [33] when all three laser pulses are derived from the same initial femtosecond, spectrally broad pulse.
Fig. 2

Multiple implementations of VCS have been demonstrated in the time-domain with two (a) or three (b) laser pulses, or in the frequency domain (c), as used in FSRS. The waiting time τ is the time delay between the second and third interactions in cases a and b, while τ refers to the pump-pulse duration in case (c). In all cases, τ defines the vibrational coherence observation time window

Alternatively, VCS may also be performed directly in the frequency domain according to an experimental scheme implementing stimulated Raman scattering and illustrated in Fig. 2c. The major difference is that the pump pulse is temporally significantly longer (typically a few ps) and spectrally narrower than the probe pulse, and both pump and probe pulses overlap temporally, in contrast to the sequential scheme discussed above. The typical non-linear process at work in this implementation is the following. The vibrational wavepacket is produced by one interaction with the pump pulse (k1) and one with the probe pulse (k2). This vibrational wavepacket is subsequently interrogated by another interaction with the long, spectrally narrow pump pulse (k3). The signal is here again detected in the direction of the probe (i.e., self-heterodyned). The result of this interaction mechanism is that the probe spectrum is amplified at frequencies that correspond to the difference between the pump frequency and the frequencies of the vibrational mode initially populated with the probe pulse. This coherent stimulated Raman technique reveals in the frequency domain the spectroscopic signatures of vibrational wavepackets in both the ground and excited state, [34] thus providing, in principle, the same spectroscopic information as IVS does in the time domain. In this frequency-domain VCS approach, the Raman pump-pulse duration is the observation window which limits the achievable spectral resolution, in the same way as the time window considered for Fourier transformation with respect to the waiting time τ limits the spectral resolution of Raman spectra recorded in the time domain.

A fundamental difference between time-domain and frequency-domain VCS is that in the latter case, the three light–molecule interactions are no longer engineered sequentially as in IVS or DFWM, but the interaction with the probe wave may occur at any moment before, between or after both interactions with the pump pulse. In fact, these various time-ordering options correspond to distinct terms in the third-order perturbative expansion, and all of them contribute to the third-order polarization [35] while in the sequential scheme by experimental design, only those where the probe interaction is the last one contribute. Hence, a drawback of the frequency-domain VCS is the background signal generated by contributions other than the stimulated Raman signal of interest. The features and challenges of each technique will be presented in more detail in the next section.

3 Multidimensional Vibrational Coherence Spectroscopy (Multi-VCS)

All the VCS approaches introduced above engineer a third-order light–molecule interaction which simultaneously reveals the vibrational activity of both electronic states coupled by a resonant pump laser field, i.e., the ground and the Franck–Condon excited state. This may pose a challenge to distinguish between vibrational signatures specific of each electronic state. Above all, this does not generally give access to the vibrational signatures of other possible transient states produced along the course of a photoreaction. In this regard, VCS has been further developed to be sensitive specifically to the excited states and to successive transient states by adding an “actinic” pulse. The role of this actinic pulse is to trigger a photoreaction prior to generating and probing a vibrational wavepacket by a subsequent third-order VCS scheme applied after a given time delay T. The vibrational activity can thus be monitored along the successive structures and electronic states achieved by the molecular system during the course of its photoreaction. This leads to the multidimensional character of the VCS, where usually one axis displays Raman frequencies while the other axis shows the photoreaction time delay T.

Several implementations of multidimensional VCS (multi-VCS) have been proposed, called “transient CARS” [36], “pump-IVS” [37], “pump-DFWM” [38, 39] and, in the spectral domain, femtosecond stimulated Raman spectroscopy (FSRS) [40], ultrafast Raman loss spectroscopy (URLS) [42, 43]. Some of these are schematically introduced in Fig. 3. Here, it is important to mention a potential problem in the semantics of the word “pump”: The actinic pulse has received different names by different groups and techniques along all years, e.g., “pump”, “initial pump”, or simply “excitation pulse”. The word “pump” in front of the techniques names, e.g., “pump-IVS” or “pump-DFWM” denotes the actinic pulse which triggers the photoreaction, prior to the IVS or DFWM third-order scheme introduced in the previous section. This should not confound the reader with the pump pulse present in the IVS or DFWM techniques themselves, which triggers the vibrational coherence to be probed.
Fig. 3

Pulse scheme of multidimensional vibrational coherence spectroscopies. In all schemes, the actinic pulse is used to trigger a photoreaction by populating a transient excited state, which is then probed by VCS, after a waiting time T. From top to the bottom: FSRS is also often called time-resolved FSRS due to the ability of detecting stimulated Raman spectra in dependence of the delay T between the actinic pulse (black) and the Raman pump-probe pulse pair (red and green, respectively). Pump-IVS is a three-pulse experiment with an actinic pulse (black) delayed by T from a “repump” pulse (red), followed by a white light probe pulse (multicolor). Pump-DFWM uses an actinic pulse (black) delayed by T from two spectrally degenerate broadband pulses (called pump and Stokes; both in red), followed by an equally spectrally degenerate broadband probe pulse (also in red). Population controlled pump-IVS is an extension of pump-IVS, which adds an additional depletion pulse (blue) between the repump (red) and probe pulses (multicolor)

Independent of the technique, the actinic pulse always excites the electronic population via two interactions with its electric field. Since these first two interactions originate from a single actinic beam, they do not affect the phase matching of the subsequent VCS signal: kmulti-VCS = ka − ka + k1 − k2 + k3. Strictly speaking, multi-VCS engineers a fifth-order light–molecule interaction [44], but again within a sequential scheme, the initial actinic pulse first triggers a photoreaction, i.e., creates a non-stationary excited-state population, which is subsequently probed by a third-order VCS scheme revealing the Raman activity as a function of the photoreaction time T. Importantly, the transition dipole moment, which operates in the T-delayed VCS scheme, does not necessarily involve the ground state anymore. This is a major difference between VCS and multi-VCS. Indeed, with multi-VCS, vibrational activity may be triggered and probed in a transient state by using laser spectra on resonance with one specific electronic transition of this very state, which may not be resonant with the ground state absorption. This allows taking advantage of the so-called resonant enhancement for generating a vibrational wavepacket specifically in this transient state.

For completeness, we note that another implementation of multi-VCS uses what we call here the actinic pump to specifically trigger a vibrational wavepacket in the ground state (by ISRS), followed by the subsequent three-pulse VCS scheme to generate and monitor an excited-state vibrational wavepacket from this non-stationary ground state population. This enables correlating the vibrational activity on the excited state to that of the ground state in a so-called two-dimensional resonance Raman (2DRR) spectroscopy. Another chapter in this collection is dedicated to this technique. Therefore here, we limit the discussion to a scheme where the effect of the actinic pulse can be simply understood as initiating a photoreaction by populating an excited state.

3.1 Femtosecond Stimulated Raman Scattering (FSRS)

FSRS is a frequency-domain spectroscopy based on stimulated Raman scattering generated after an actinic pulse (Fig. 3, top panel). The signal is generated in the same direction as the probe beam, leading to a self-heterodyne detection geometry. The delay T between the probe pulse and the actinic pulse is scanned and Raman spectra are detected in dependence of this delay. The first implementation of FSRS in its three-beam configuration used a 10-Hz laser and had a Raman resolution of only 76 cm−1 [45]. After almost 25 years of experimental development, the state-of-the-art FSRS setup nowadays offers a Raman resolution of about 10 cm−1 while using actinic pulses with durations of less than 100 fs, and covering the whole ultraviolet and visible spectral range [47, 48, 49, 50].

The probe pulse is spectrally broad and usually red-shifted with respect to the narrow spectrum of the Raman pump. The advantage of a broadband implementation is that a full Raman spectrum, i.e., typically from 200 to 3000 cm−1, can be directly recorded and the high peak intensity of femtosecond pulses enhances the efficiency of the stimulated Raman process generating vibrational wavepackets [40]. While the actinic pulse is electronic resonant with the ground-state absorption, the Raman pump-probe pair is usually off-resonant in most experiments. The ability of tuning the spectrum of the Raman pump-probe pair allows to probe specific electronic transitions, other than those involving the ground state [51]. Thus, all Raman modes in the excited state or subsequent transient states can be measured, not only those Franck–Condon active. Being a spectral domain technique where only the scan of the actinic pulse T delay is required, the acquisition of several spectra is intrinsically faster than time domain techniques where an additional probe τ delay must be scanned (Fig. 3). While FSRS has been already demonstrated with a single laser shot acquisition [40], typical acquisition times of FSRS transients are only a few minutes [52].

A major challenge in FSRS is the separation of the stimulated Raman spectrum from the non-coherent background spectrally overlapping with the signal. The lack of automatic methods to extract the excited-state Raman spectrum from the ground-state Raman spectrum, probe background, and transient absorption signal has been a major obstacle in the practical development of FSRS as an analytical tool. Several experimental and numerical approaches have been developed in this regard. Since FSRS is usually performed with kHz laser sources, the chopping of the actinic and Raman pulses at different frequencies (Fig. 4) has been shown to separate to some degree the different overlapping signal contributions [53]. This allows for the subtraction of the transient absorption (TA) baseline from the FSRS raw data (“unprocessed”, see Fig. 5a), but does not eliminate all baseline distortions or solvent contributions in FSRS (Fig. 5b) [54]. The intensity modulation of the spectrally narrow Raman pulse has been identified as a major artifact in FSRS, often requiring ad hoc scaling of baselines (see e.g., Fig. 5a FSRS-baseline fit). This artifact can be corrected by a factor numerically calculated by including transient absorption changes measured under similar experimental intensities [55]. More recently, the use of carefully crafted Raman spectra in form of watermarks has been used to easily identify Raman resonances in the raw signal [56]. The last word in the correction of the baseline distortions in FSRS has not been spoken yet and it is an active research focus in the field of multidimensional VCS.
Fig. 4

Scheme of the FSRS experimental setup and pulse-chopping scheme. All three beams are focused and overlapped on the sample, while only the actinic pulse (Excitation) is delayed (T). The chopping of the actinic beam and Raman pump beam with different frequencies allows automatic subtraction of the signal background. In this scheme, the phase 1 of the chopping detects only the probe background. In phase 2, the stimulated Raman signal of ground state (stationary), including solvent, is detected. In phase 3, a transient absorption signal at delay T is detected. Finally, at phase 4 all signals are collected together. The FSRS signal is calculated from signal 4 – (3 + 2). See Fig. 5 for more details. Reprinted from Ref. [53] with permission of Springer

Fig. 5

a Typical baseline correction of FSRS difference spectrum. “TA” stands for “transient absorption”. b Negative features originating from depleted ground-state Raman signals contribute to the FSRS difference spectrum. This feature is removed by adding back an appropriately scaled background Raman spectrum, optimized for one of the solvent modes. Correction of this feature of the FSRS spectrum yields the FSRS signal of interest

Reprinted with permission from Ref. [54]. Copyright 2017 American Chemical Society

3.2 Pump-IVS

Pump-IVS is based on the combination of impulsive stimulated Raman scattering (ISRS) with an actinic excitation. As FSRS, it can be easily implemented since the signal is generated in the direction of the probe beam. However, it requires scanning two time delays, namely the photoreaction time T between the actinic pulse and the “repump” pulse (or Raman pump, or even impulsive pump, simply called “pump” hereafter), and the waiting time τ between the pump and probe pulse, during which the wavepacket dynamics evolve. Typical acquisition times for pump-IVS are of several tens of minutes for laser systems with kHz repetition rates. The acquired data must then be post-processed by Fourier transformation along τ to reveal the T-dependent impulsive Raman spectrum. The spectral range of the Raman spectrum depends exclusively on the bandwidth of the pump spectrum and the length of the transients measured along the τ delay. The pump bandwidth limits the highest upper Raman frequency and the length of the τ scanning interval defines the lowest detectable Raman frequency. Raman spectra from as low as few tens of wavenumbers up to 3000 cm−1 have been demonstrated with pump-IVS [19, 57, 58]. Detection of very low frequency Raman modes (< 200 cm−1) is another central advantage of time-domain methods in comparison with spectral domain methods like FSRS, which are constrained by optical filters to spectrally cut the Raman pump from the Raman spectrum. Moreover, since the probe spectrum in pump-IVS is normally spectrally resolved via a grating spectrometer, Raman spectra can be obtained at several detection wavelengths, depending only on the bandwidth of the probe spectrum.

Experimentally, pump-IVS setups are based on non-collinear optical parametric amplifiers (nc-OPA) [59, 60] to generate the broadband spectra of the actinic and pump pulses. Pulse durations below 10 fs have been successfully used [19, 57, 58]. The probe pulse is usually a white light supercontinuum with a very broad spectrum spanning from about 300 nm up to 750 nm or more (depending on the crystal used for white light generation). The pump spectrum is usually spectrally resonant with excited-state absorption bands. Typically, the transients along the τ delay are measured for different T delays with respect to actinic pulse with and without the pump pulse by using a chopper (Fig. 6). This assists in the separation of the transients originated in the excited-state manifold from the ground-state or solvent contributions (induced by the actinic pulse alone). The extraction of the excited-state signal induced by the VCS pump pulse can be easily obtained if the pump spectrum is off-resonant with any ground-state absorption. Pump spectra resonant (or near-resonant) with the ground-state absorption will lead instead to contamination of the pure “excited-state” signal since the pump pulse will also efficiently generate vibrational wavepackets in the ground electronic state as well as in the electronically excited state. Under these conditions, a way to partially mitigate this effect is by additionally chopping the actinic pulse to subtract the vibrational wavepackets induced without the actinic pump.
Fig. 6

Scheme of the pump-IVS signal analysis procedure. From left to right Two transients are detected, one with pump pulse (“P2 on”) and one with the pump pulse blocked (“P2 off”). The transients are subtracted from each other (“P2 on-P2 off”). The residual of this subtraction is finally Fourier transformed to obtain the Raman spectrum

Reprinted from Ref. [57] with the permission of AIP Publishing LLC

3.3 Pump-DFWM

Pump-degenerate four-wave mixing is based on the same concept as pump-IVS, but the two interactions with the second pulse in pump-IVS (repump) is split in two pulses. This four-beam geometry offers several advantages with respect to the three beams used in pump-IVS. One of them is the delay between the two “repump” field interactions (called here “pump” and “Stokes”), which can be used for selective nonlinear response preparation and to separate optical beating artifacts from molecular vibrations [20]. The four-beam geometry generates the signal in a background-free configuration (homodyne detection), i.e., not in the direction of the probe as in pump-IVS, which typically leads to a superior signal-to-noise ratio and shorter acquisition times [58].

Similar to pump-IVS, pump-DFWM has been usually performed with kHz laser sources. A typical setup consists of two non-collinear OPAs to generate the broadband spectra of the actinic and DFWM pulses independently. Like in pump-IVS, time-resolved signals are recorded as a function of the actinic pulse delay (T) and of the probe delay (τ). In pump-DFWM, however, the transients along the τ delay are not subtracted from any reference, e.g., with and without actinic pump or by any other method, due to the homodyne detection (see below). Beyond that, the analysis of each transient at a given T delay containing oscillatory and non-oscillatory contributions is essentially the same as for pump-IVS, and illustrated in Fig. 7.
Fig. 7

The signal analysis scheme of pump-DFWM is very similar to pump-IVS. At each T delay a kinetic trace is recorded, which contains oscillatory and non-oscillatory contributions. They can be separated by a polynomial fitting performed for probe delays τ > 100 fs, typically. Vibrational spectra are then obtained by fast Fourier transformation (FFT) of the oscillatory contribution along the τ delay after zero-padding, windowing and apodization

Adapted from Ref. [170]

As briefly mentioned above, pump-DFWM differs from pump-IVS and FSRS in the signal detection and the dependence on the molecule concentration in the sample. The pump-DFWM signal is detected in a homodyne configuration. This contrasts to the self-heterodyne detection used in the other two techniques, where the signal created in the sample is generated in the same direction as the (Raman) probe beam. (Self-)heterodyne detection offers amplification of weak optical signals (see the contribution “Introduction to State of the Art Multidimensional Time-resolved Spectroscopy Methods” in the collection), but its signal-to-noise ratio often suffers from local oscillator fluctuations, inserting random noise as well as signal distortion that can only be corrected under e.g., specific balanced detection and averaging schemes [30, 61]. Pump-DFWM is not prone to such signal distortions, instead it may suffer from interferences (see below on the Challenges in Multidimensional VCS) leading to an ambiguous assignment of (low) frequency modes and artificially broadened Raman lines [20]. The second major difference between pump-DFWM and other techniques is the nonlinear concentration dependence: while pump-IVS and FSRS depend linearly on the probed concentration, pump-DFWM (as any other homodyne nonlinear technique) shows a quadratic dependence. This nonlinear dependence leads to a distortion of the evolution of the amplitude of Raman modes during the T delay, which can only be linearized by using a heterodyne detection [58]. A major drawback of such nonlinear concentration dependence is the difficulty of detecting minor components in a multi-component sample, since the contribution of major component(s) will dominate the signal.

4 Challenges in Multidimensional VCS

A fundamental challenge of multi-VCS in the detection of vibrational coherence in the electronically excited manifold is the small amount of excited-state population compared to the ground state or other non-actinic activated molecules (like solvent or buffer molecules). The amount of excited molecules can be changed by the actinic pulse energy, but in general is well below 20–30% to avoid saturation or other nonlinear effects. Usually, it is the interplay of the involved transition dipole moments (i.e., resonant or non-resonant with the ground state or with the small excited-state population) and the spectral overlap with optical signals originated by other states, that will determine whether a VCS signal of a given electronic state is detectable or not. In practice, very seldom a VCS signal will contain the vibrational signature of only one single electronic state. The only exception for that is electronically nonresonant experiments involving only the electronic ground state (no actinic pump). In spectrally resonant experiments, for example, the VCS spectrum is often resonant with distinct transitions during the course of a photoreaction, since the involved transition dipole moments and FC overlap with higher lying states are continuously changing. This leads inevitably to a modulation of the VCS signal due to a modification of the resonance enhancement conditions. The overlap of VCS signals originated by different electronic states is one of several causes for the interpretation challenges in VCS experiments and has been the motivation of new techniques to disentangle them.

One approach to assist the assignment of a Raman mode to the ground- or excited-state manifold is directly related to the properties of the oscillatory signal itself. Ground-state vibrational modes often show a longer-living dynamics than transient excited states, where the dephasing may be faster, due to e.g., a photoinduced reaction or internal conversion to the ground state [63, 64]. In time-domain techniques, this leads to vibrational wavepackets with shorter dephasing times (or broader Raman peaks in spectral domain techniques like FSRS) normally associated to excited state modes, while longer dephasing times are taken as coming from the ground state. Of course, longer and shorter are relative quantities, which can only correctly be interpreted when taking into account the duration of other vibrational modes and electronic population times of the involved electronic states. An additional approach often used to identify the nature of vibrational wavepackets is to rely on the phase of the oscillatory signal [13, 66, 67, 68]. For example, in spectrally dispersed time-resolved experiments, probing at wavelengths red- and blue-detuned from the center of the absorption spectrum often leads to oscillatory signals with a π-phase difference due to vibrational wavepacket dynamics. This information, when combined with the ground- and excited-state absorption bands, may assist in the assignment of a Raman mode to the respective electronic state (see e.g. [30]).

Separation and assignment of contributions originating from different excited states or electronic transitions can be acutely challenging in multi-VCS, especially when the respective excited-state absorption spectra overlap. However, when they do not overlap, the assignment can be facilitated by exploiting the tuneability of VCS laser spectrum. Importantly, when the VCS pump pulse(s), i.e., k1 and k2 in Figs. 2 or 3, are resonant with different electronic transitions from the same electronic state, different Raman activities may be observed. This has been demonstrated in pump-DFWM applied to all-trans-retinal (ATR) and retinal Schiff Base (RSB) in solution to obtain the Raman active modes at different excited-state electronic transitions (Fig. 8) [69]. In this example, the DFWM spectrum was tuned to different electronic transitions in the excited state absorption (ESA) bands of ATR (Fig. 8a) and RSB (Fig. 8b). While ATR displays a single ESA and an excited state stimulated emission (ESE) band (Fig. 8a), RSB shows two ESA bands above 17,000 cm−1 and a shallow ESE below 15,000 cm−1. These two bands observed for RSB are due to different transitions from the S1 state to two different high-lying electronic states, which are allowed only for RSB. The Fourier transformed signal of pump-DFWM at different DFWM detuning shows how different the Raman active modes are in each case (Fig. 8c–e): For example, in ATR, the C=C bond structure at 1510 and 1580 cm−1 shows a strong shift and amplitude change when the DFWM spectrum is detuned from resonance with the ESA band at about 19,000 cm−1 (Fig. 8c) to the red-shifted spectrum resonant with the ESE band (Fig. 8d). In RSB, the tuning of the DFWM spectra from one ESA band to a red-shifted one (Fig. 8b), shows that the C=N bond observed around 1700 cm−1 is active at only one electronic transition (Fig. 8e, f).
Fig. 8

a Ground-state absorption of ATR in ethanol (black), UV excitation (25,000 cm−1, violet), DFWM spectra for ATR (cyan and red), and transient absorption (TA) spectrum (orange) of ATR in ethanol at T = 510 fs after UV excitation. b Ground-state absorption of RSB in ethanol (black), UV excitation (violet), DFWM spectra (blue and green), and TA spectrum of RSB (orange) in ethanol at T = 510 fs after UV excitation. Pump-DFWM data was acquired at T = 1 ps for ATR c (cyan DFWM spectrum) and d (red DFWM spectrum). Pump-DFWM data were acquired at T = 2 ps for RSB e (blue DFWM spectrum) and f (green DFWM spectrum). Vibrational bands assigned to solvent dynamics are indicated with asterisks

Figure adapted with permission from Ref. [69]. Copyright 2013 American Chemical Society

Detuning of the VCS laser spectrum can be also exploited in the absence of the actinic pulse to assist in the assignment of ground- and excited-state modes in multi-VCS. In this case, the detuning of the VCS spectrum from non-resonant, over near-resonant, up to completely electronically resonant leads to different degrees of vibrational coherence in the excited state. While electronically non-resonant excitation cannot induce any vibrational coherence in the excited state, near resonant excitation, for example, can induce low-frequency vibrational coherence only. Complete resonant excitation is able to induce high- as well as low-frequency coherence in the excited state. This dependence on the spectral overlap between an absorption spectrum and VCS laser spectrum is different for the groundstate, where low- as well as high-frequency modes will be always induced. By comparing how the amplitude of specific vibrational modes decreases when the excitation becomes non-resonant, it is possible to pinpoint which vibrational modes are present only in the excited state or in both states, and how specific vibrational modes are being activated (via direct laser interaction or via coherent excitation from other vibrational mode) [10, 19, 28, 70]. For example, this has been applied to retinal protonated Schiff base (RPSB) to show that low-frequency modes are not active in the electronic ground state but are coherently activated by vibrational energy redistribution from high-frequency modes directly excited in the electronic excited state [71]. In FSRS, the detuning of the Raman pump wavelength was exploited to record the excited-state Raman spectrum in the absence of the actinic pump (Fig. 9) [34]. By tuning the Raman pump wavelength from red to blue towards the S0 → S1 resonance, the S1 modes become stronger (at about 200, 300, 650, and 850 cm−1), while the S0 contributions (peaks with negative amplitude in Fig. 9b) remain effectively unchanged. This method has also been recently applied to record unambiguously for the first time the low frequency Raman modes of the S2 state of all-trans-β-carotene (see Sect. 5.2) [51].
Fig. 9

FSRS difference spectra of trans-azobenzene in n-hexane measured without actinic pump. a Raw data and FSRS difference spectra measured with a Raman pump at 511 nm. b FSRS difference spectra measured as function of the Raman pump detuning. Negative bands correspond to the conventional S0 signal, while positive peaks originate from S1. The S1 contributions vanish when the Raman pump is detuned farther from the S0 → S1 resonance, i.e., from blue to red detuning

Reproduced from Ref. [34] with permission from AIP Publishing

An additional central aspect in the signal interpretation of VCS-detected signals is interferences between signal contributions. Several kinds of interferences can be experimentally observed which may overlap with the vibrational coherences of interest [73, 74, 75]. For example, interference beats originate from the interference between vibrational modes of two different electronic states, and are an intrinsic beat of electronically resonant techniques, since vibrational coherence will be excited on both electronic states involved. Polarization beats originate from interference between nonlinear polarization contributions at the detector and are thus a pure optical effect. The origin of the polarization beating can be e.g., the emission, by different kinds of molecules, of polarizations oscillating at different vibrational frequencies and interfering at the detector. We refer the reader to references [20, 31] for more technical details.

In general, interference between oscillatory signals resulting in additional low- or high-frequency oscillations hampers the correct interpretation of VCS experiments. Polarization beating, for example, is of special importance when the low-frequency region (< 800 cm−1) in FFT spectra of transient dynamics is interpreted [20]. In the spectral domain, band positions of molecular normal modes and signal contributions from polarization beating between molecular high-frequency contributions can overlap, hampering an unambiguous interpretation and assignment. Polarization beating is an interference effect present in all types of VCS methods, with homodyne and, to a lesser degree, (self-)heterodyne detection. The discrimination between molecular normal modes and polarization beating is possible in techniques where the vibrational wavepacket generation originates from two different pulses like DFWM and CARS, where the delay τ12 between pump and Stokes pulses can be exploited. The separation of different response pathways has been demonstrated for several polyatomic molecules in solution using chirped, spectrally resolved DFWM (Fig. 10) [20]. Two organic dye molecules, rhodamine B and S-9, show strong oscillatory beating centered at τ12 = 0 when no chirp is applied (Fig. 10a, c, respectively). When chirp is applied, the maxima of interference contributions are shifted with a given amount depending on the chirp applied, while beating due to vibrational wavepackets are not shifted with the delay τ12 between pump and Stokes pulses (Fig. 10b, d).
Fig. 10

Disentanglement of real molecular vibrational bands from polarization interferences depicted for a, b Rh-B and for c, d S-9 in methanol. The disentanglement is possible by comparing the maximum of a given frequency in respect to τ12 when the excitation chirp (ϕ″) is varied. By using positively chirped (ϕ″ > 0) DFWM pulses, oscillatory contributions due to polarization beating will appear for slightly delayed pump and Stokes pulses (τ12 > 0), while real molecular vibrational bands will not be shifted in τ12 in comparison to the non-chirped excitation (ϕ″ = 0). While for Rh-B (a, b), several frequencies can be assigned to molecular modes, in particular at 205 and 620 cm−1, for S-9 (c, d), only two weak contributions at about 505 and 555 cm−1 are real molecular vibrational normal modes

Reprinted from Ref. [20] with the permission of AIP Publishing

Finally, interference beating can be also addressed by applying higher-order nonlinear techniques. Population-controlled pump-IVS is an extension of pump-IVS to suppress undesired vibrational coherence contributions to the optical signal (Fig. 3, bottom panel) [77, 78]. It is based on the same three-beam geometry to excite impulsively Raman vibrations, but adds an additional pulse named “dump” to interact with the sample between the pump and the probe pulses, formally resulting in a seventh-order nonlinear time-resolved method. PC-pump-IVS follows the same approach as pump-depletion-probe experiments in visible [80, 81, 82, 83, 84] and infrared [85, 86] to disentangle population relaxation pathways, and applies to the depletion of the vibrational coherence in the excited state manifold. In PC-pump-IVS, the interaction of the narrowband dump pulse has the effect of only changing the population of the excited state and decreasing the respective Raman signal. By measuring the transients along the τ delay with different combination of pulses (Fig. 11a), it is possible to subtract the contribution of the ground-state and solvent vibrational coherence (Fig. 11b). A key aspect of this technique is the bandwidth of the dump pulse: The spectrally narrow dump pulse with a pulse duration of few hundreds of femtoseconds is not able to induce any additional vibrational coherence, leading to a pure depletion of electronic population. PC-pump-IVS has been demonstrated with dump pulses with a duration of 200 fs (i.e. about 75 cm−1 FWHM spectral width) [77].
Fig. 11

Scheme of the population-controlled pump-IVS; a experimental setup and b pulse chopping scheme. By chopping all pulses but the probe pulse, it is possible to extract the pure signal of the excited state S1 in this scheme

Reprinted from Ref. [77] with permission of ACS https://pubs.acs.org/doi/abs/10.1021%2Fjp5075863

5 Application of Multidimensional VCS

5.1 How can Multidimensional VCS Help a Chemist?

The most natural way of describing how multi-VCS can assist a chemist is by comparing it to Raman spectroscopy. In its essence, multi-VCS delivers a sequence of Raman spectra obtained after an actinic pulse interacts with the sample. In Raman spectroscopy, structural and interaction information comes from the energy, width, and amplitude of Raman bands, which assist in the identification of specific molecular species and conformations. Multi-VCS provides the same information and goes beyond by measuring how Raman frequencies shift and amplitudes evolve in time (for example, Refs. [88, 89]). Hence, it reveals time-resolved structural information on transient molecular species along the course of a photoreaction. In this regard, multi-VCS follows the same conceptual approach as transient spontaneous Raman measurements with picosecond pulses [91, 92, 93] pioneered by Lauberau et al. more than four decades ago [94]. Multi-VCS differs from transient picosecond Raman spectroscopy, however, in many aspects. The first one is the signal intensity of the spontaneous Raman signal detected in the transient picosecond Raman spectroscopy, which is intrinsically much weaker than the coherent/stimulated signal detected in multi-VCS methods. The second one is related to the Fourier relation between the spectral resolution and pulse duration. Transient picosecond Raman spectroscopy is limited to molecular processes much slower than picoseconds due to the intrinsic time duration of the narrow-band Raman pulse, if a spectral resolution of few wavenumbers is desired. For example, a 1-ps Raman probe pulse leads to an intrinsic band broadening of more than 10 cm−1. It is also interesting to note that the vibrational content in multi-VCS can be very similar to the one detected by transient infrared absorption spectroscopies, in particular for complex systems where symmetry rules are relaxed and vibrations can be probed by Raman as well as infrared interactions.

Generally, following how the frequency of Raman bands change in dependence of the actinic pulse time delay T gives information on how bond strengths are modified during a photochemical reaction [37, 89, 95]. Frequency blue-shifts indicate stiffening of the vibrational motion or relaxation within an anharmonic potential, while red shifts may hint at elongation of a given chemical bond. On the other hand, the lack of any shift or amplitude changes is usually taken as the signature of non-reactive coordinates. The changes in the amplitude of Raman bands in multi-VCS report on the formation and breaking of chemical bonds after interaction with the actinic pulse. When correlated with frequencies changes, they become the central feature in the mapping of structural changes during photochemical reactions. Moreover, the evolution of the amplitude of a band does not need to follow a first-order kinetics and decay or grow exponentially, it can also show very complex dynamics. For example, modulation of the amplitude of a Raman band with the frequency of another Raman band has been identified as the signature of anharmonic vibrational coupling between the two corresponding modes [88]. Information on how vibrational energy redistributes can be retrieved from the width as well as frequency changes of Raman bands. Strong coupling between molecular vibrations leads to strong dephasing and damping of specific bands and can be exploited to record how e.g., intramolecular vibrational relaxation (IVR) and vibrational cooling—e.g., solute–solvent interaction—take place in complex systems [97, 98]. Initially, very fast photoreactive systems evolve within few tens or hundreds of femtoseconds away from the Franck–Condon region, leaving very broad multi-VCS bands in the Raman spectrum.

Also, the ability of multi-VCS to extract information on the molecular dynamics is enhanced when the evolution of Raman bands is combined with other techniques [99]. This is particularly interesting when methods sensitive to other molecular degrees-of-freedom are considered, like transient absorption and the detection of the electronic population dynamics. For example, rise or decay times of specific Raman bands can be correlated to respective times observed in the electronic population evolution to assign Raman spectra to specific electronic states or assist the assignment of electronic states by using known Raman spectra, as will be further illustrated below. Finally, the ability of multi-VCS to reveal detailed structural information on excited electronic states makes it a very powerful experimental approach to assess the accuracy of state-of-the-art computational developments targeting quantum chemical modeling of electronic structures and of molecular photoreactivity in general (Table 1).
Table 1

Selected molecular systems investigated with multidimensional VCS in the last decade

Molecular system

Technique

References

Based on carotenoids

Pump-CARS

Pump-DFWM

Pump-IVS

FSRS

PC-pump-IVS

[39, 55, 58, 78, 101, 102, 103, 104, 105]

Based on retinal

Pump-DFWM

Pump-IVS

FSRS

PC-pump-IVS

[69, 71, 77, 107, 108, 109]

Based on stilbene

Pump-IVS

FSRS

PC-pump-IVS

[34, 37, 48, 77, 110, 111]

Dye molecules in solution

Pump-IVS

FSRS

[21, 112]

TIPS-pentacene

Pump-IVS

[113]

Aromatic amino acid residues

Pump-IVS

FSRS

[114, 115]

Green fluorescence protein

Pump-IVS

FSRS

[88, 109, 116]

Photoactive yellow protein

Pump-IVS

FSRS

[99, 117]

Bis(phenylethynyl)benzene

URLS

[42, 118]

Tetraphenylethylene

URLS

[43]

Photoactive flavoproteins

FSRS

[97, 119]

Myoglobin

FSRS

[96]

Azobenzenes

FSRS

[34, 89]

Cyclohexadiene derivative

FSRS

[95]

A seminal example in multi-VCS, which depicts many of the points mentioned above, is the pioneering FSRS investigation of the photoinduced, sub-ps structural evolution of the protonated Schiff base of retinal (PSBR) in the rhodopsin protein, the visual sensor [120]. Upon light excitation to S1, the C11=C12 bond of PSBR extends and acquires a single-bond character. This enables a very fast torsional motion to drive the system within ~ 100 fs to a CInt [121], where it decays to S0, with a predicted C11=C12 bond twist of ~ 90°. By implementing FSRS, the vibrational frequencies associated to the hydrogen out-of-plane motions (in the 800–900 cm−1 range) of both hydrogen atoms linked to the H–C11=C12–H isomerizing bond could be observed as a function of time, after decay to the S0. This experiment revealed a rapid (300-fs time scale) blue shift of the H wagging frequencies by ~ 100 cm−1, which is the signature of a large structural reorganization of PSBR occurring on the ground state PES S0. More specifically, the blue shift is shown to reveal the stiffening of these vibrations resulting from the planarization of the PSBR backbone, i.e., completion of the C11=C12 isomerization (torsional) motion towards the first vibrationally relaxed photoproduct intermediate, called bathorhodopsin.

5.2 Ultrafast Vibrational Dynamics in the Excited States of Carotenoids

Carotenoids have been an important class of molecules investigated by multi-VCS due to their central role in several biological functions [122]. As chromophores in light-harvesting complexes (LHC), for example, carotenoids are involved in the initial absorption of light and energy transfer to other chromophores (e.g., bacteriochlorophylls). The very strong absorption from the ground state S0 is due to the π–π* transition to the second electronic state labeled S2. The transition from the S0 to the first electronic S1 state is not one-photon allowed, making the S1 state a “dark” electronic state. Upon light absorption to S2, the electronic relaxation to S1 is very fast (within 100–200 fs for all-trans-β-carotene and lycopene). Further electronic relaxation from S1 to S0 takes place in the picosecond time scale, varying with the number N of effective conjugated C=C double bonds. For example, the S1 state in all-trans-β-carotene (N ~ 10.5) decays with about 9 ps, while in lycopene (N = 11) decays faster with about 4.1 ps [123]. The S2 to S1 relaxation mechanism has been intensely debated in the last decades (see for example Ref. [124, 125, 126, 127, 128]). One source of discussion has been, for example, the nearly identical lifetime of the S2 state for several open-chain carotenoids with different numbers N of effective conjugated C=C bonds, which has been interpreted as the result of additional electronic dark states between the S2 and S1 states. The experimental identification of these dark states via transient absorption, nevertheless, has been extremely challenging due to the high spectral overlap of the respective absorption bands, the ultrashort time scales involved, but also due to the couplings [130, 131] between these states.

Multi-VCS has been applied to carotenoids to assist in the elucidation of the excited state manifold. In general, the excited-state active vibrations of carotenoids are typical of polyenes consisting of strong high-frequency modes like methyl deformation (~ 1000 cm−1), C–C stretching (~ 1140–1200 cm−1), and C=C stretching (~ 1500–1550 cm−1) (Fig. 12). All these vibrational modes are active in all electronic states. However, since carotenoids are polyenes with C2h symmetry, an additional frequency at about 1800 cm−1 is observed for the S1 state, due to the adiabatic coupling between the S1 and S0 states [129, 130]. The mapping of the Raman activity, frequency shifts, amplitude rising, and dephasing of these vibrational modes in dependence on the actinic pulse delay has been the main focus of multi-VCS.
Fig. 12

2D mapping of the vibrational coherence for all-trans lycopene, upon photoexcitation in the S2 state. Being a polyene, lycopene shows very strong C=C and C–C stretching modes at 1564 and 1180 cm−1, respectively. The formation of the S1 C=C mode at 1783 cm−1 as the S1 state is populated can be clearly seen within the initial 300 fs. THF: tetrahydrofuran

Reprinted with permission from Ref. [102]. Copyright 2011 American Chemical Society

One of the first experimental observations of the evolution of the totally symmetric C=C stretching mode at about 1800 cm−1 has been pioneered by Hashimoto et al. using an actinic pulse to promote all-trans-β-carotene to S2 and followed by a stimulated Raman scattering scheme [132]. With a temporal resolution of 300 fs, the energy flow between the S2 and the S1 states was followed. The initial vibrational relaxation until the v = 1 level of the totally symmetric C=C mode of the S1 state was found to be very fast, while further relaxation from v = 1 to v = 0 was much slower than the internal conversion to the ground state S0. This has been explained by the presence of an additional electronic dark state, which assisted the vibrational relaxation to v = 1 of the S1 state. The presence of an electronic dark state being populated within 20 fs in the decay between S2 and S1 states has been also proposed in lycopene based on extensive modeling of the spectroscopic signal observed by pump-DFWM [102].

The vibrational relaxation during the internal conversion between the S2 and S1 states for all-trans-β-carotene has been also explained by pure vibrational cooling, challenging the presence of any additional electronic dark state assisting this relaxation. The first FSRS measurements of all-trans-β-carotene showed that the S1 C=C stretching mode at 1798 cm−1 relaxes via a two-step process identified by two distinct frequency up-shift kinetics [133]. A fast time constant (200 fs) was related to a strong coupling followed by a slower equilibration process (450 fs) to the complete set of vibrational normal modes. Although these first FSRS measurements were not able to resolve the initial 250 fs of the dynamics, more recent comprehensive studies using pump-DFWM and pump-IVS with improved temporal resolution corroborated the ultrafast initial step of the vibrational cooling model of the S1 C=C. A similar ultrafast vibrational cooling step was also observed for the C=C 1525 cm−1 mode for all-trans-β-carotene [101] and in other carotenoids [58]. Cooling of the S1 C=C mode was also observed for carotenoids in a light-harvesting complex by applying FSRS to spirilloxanthin in the native LH1 of Rhodospirillum rubrum [134]. After the actinic excitation, the frequency at 1740 cm−1 evolved to 1767 cm−1 with a time constant of 300 fs.

The vibrational relaxation after deactivation of the S2 state has been recently further addressed for a series of open-chain carotenoids (like lycopene) with increasing conjugated length N by using pump-DFWM [104]. The simple picture of vibrational cooling, accounted for by a bi-exponential frequency up-shift as discussed in the previous paragraph for all-trans-β-carotene, is actually only observed for longer open-chain carotenoids (N = 11 and 13). Short open-chain carotenoids (N = 9 and 10) show a down-shift of the C=C stretching mode from about 1580 to about 1510 cm−1, which has been explained as a further indication of coupling between the S2 and an additional electronic dark state (between the S1 and S2 states).

A very intriguing observation in the vibrational dynamics of carotenoids is the additional observation of two vibrational bands in the 1800 cm−1 spectral region in several experiments. These bands have been observed with e.g., FSRS at 1770 and 1800 cm−1 as well as at 1770 and 1790 cm−1 [132, 135], and at 1740 and 1785 cm−1 with pump-IVS and pump-DFWM [58] for all-trans-β-carotene. In the latter case, they have been explained as the result of vibrationally hot C=C levels of the typical S1 C=C mode contributing to the signal (agreeing with the above picture). More recently, FSRS has been applied to spirilloxanthin (N = 13) and two bands at 1743 and 1771 cm−1 have been also observed [56]. However, the much longer life time (3 ps) detected for 1771 cm−1 in comparison to the S1 lifetime (1.5 ps) has been interpreted as the signature of an electronic state of another carotenoid conformer present already in the ground state and also excited by the actinic pulse. Whether these two bands are due to inhomogeneous S0 conformational distributions or to vibrationally hot bands in the S1 manifold is still unclear.

The first attempt to follow the evolution of the high-frequency Raman spectrum of the S2 state directly was done by FSRS using a Raman pump and probe spectrally resonant with the S2 ESA of all-trans-β-carotene [136]. A very broad spectrally unresolved Raman band was disentangled with two bands at 1654 and 1739 cm−1. While the 1739 cm−1 band was assigned to the C=C mode in S1, the mode at 1654 cm−1 was due to the C=C in the S2 state. A similar frequency for the C=C in the S2 state at 1660 cm−1 was also observed for another kind of carotenoid (trans-apo-8′-carotenal) with FSRS [137]. It is important to note that, more recently, Kennis et al. showed that the FSRS signal in all-trans-β-carotene measurements is distorted by the transient absorption signal of the molecule and careful correction must be performed to extract the correct Raman signal [55]. Non-optimal background subtraction may lead to false vibrational bands. For example, contributions from the conventional transient absorption were taken into account, and the FSRS was corrected, indicating that the spectra observed earlier for S2 [138] were strongly modulated by the S2 transient absorption signal. A more careful study of the Raman spectrum of the S2 state with FSRS and of the S0 Raman activity by non-resonant stimulated Raman scattering recently showed that all high-frequency Raman bands initially between 800 and 1600 cm−1 previously assigned to the S2 state are due to the dispersive ground-state vibrational bands contaminating the excited state signal (Fig. 13, in particular b) [51, 135].
Fig. 13

FSRS on all-trans-β-carotene a raw FSRS spectrum. b Background-corrected FSRS signal showing positive signal from the S2 state and negative signal from the S0 ground state

Reprinted with permission from Ref. [51]. Copyright 2018 American Chemical Society

Finally, low-frequency contributions have also been investigated in carotenoids [20, 51, 58]. Low-frequency bands are notoriously complicated to detect in any multi-VCS, in particular for short-lived electronic states as the S2 state of carotenoids like all-trans-β-carotene and lycopene. Nevertheless, low-frequency modes of the S2 state were reported by applying resonant stimulated Raman scattering, i.e., without an actinic pump for all-trans-β-carotene [51]. The bands at 200, 400, and 600 cm−1 in the S2 have larger amplitude compared to ground state modes, hinting at strong anharmonicities and mixing of low frequencies in the S2 state (Fig. 13b).

In spite of being investigated by several techniques, the ultrafast vibrational dynamics in the excited states of carotenoids still poses a challenge for multi-VCS and will require in the future additional experimental and theoretical work to clarify it. Nevertheless, the application to carotenoids illustrates the ability of multi-VCS at revealing the signatures of vibrational dynamics involved in ultrafast, non-reactive, internal conversion. We will describe below another example of application of multi-VCS to the investigation of photoreactive vibrational dynamics.

5.3 Photoisomerization of Stilbene and Derivates

Multidimensional VCS performs vibrational (Raman) spectroscopy of molecular excited states, from the FC state to the photoproduct formation along the photoreaction pathway. Hence, it reports on the time evolution of the molecular structure, which is particularly informative when monitoring photoreactions involving large amplitude motions and structural changes such as C=C double bond photoisomerizations. Here we will illustrate the recent, successful use of multi-VCS for the investigation of the isomerization reaction of stilbene and derivatives (Fig. 14). These photoreactions are not only models for ultrafast C=C photoisomerization but also prototypes of light-to-mechanical energy conversion in well-known synthetic rotary motors [139, 140].
Fig. 14

Chemical structures of stilbene (left), stiff-stilbene (middle), and a fluorene-based rotary motor (right)

Both the cis and trans isomers of stilbene, named “c” and “t”, respectively, undergo C=C double bond photoisomerization via a common, so-called “phantom” dark state [142, 143], which is a perpendicular S1 transient structure named p*, from which further pyramidalization rapidly drives the system to decay to S0 via a conical intersection (CInt) [145, 146] in a similar way in both cases. Upon photoexcitation of the planar t isomer, the formation of the p* transient state from the t* Franck–Condon state takes ~ 100 ps due to a significant S1 energy barrier. Photoexcitation of the non-planar c isomer leads either (1) to ultrafast, further C=C bond torsion and sub-ps formation of the same transient p* state, or (2) to planarization enabling a cyclization reaction and dihydrophenanthrene (DHP) formation. Accurate time-resolved structural information along the S1 reactive paths has long been sought after in order to decipher the photoreaction mechanism in condensed phase [37, 77, 148, 149].

The time-resolved, spontaneous, and stimulated Raman spectroscopy of t* and isotopomers has been investigated with outstanding detail and accuracy, offering precious opportunities to benchmark computational methodologies for modeling the excited-state electronic structure and anharmonic PES landscapes [147, 150, 151, 152, 153, 154]. Because of its high solubility and strong Raman activity, trans-stilbene has also been used to demonstrate the remarkable sensitivities of state-of-the-art multi-VCS experimental setups producing very high quality spectra [46, 77], as illustrated in Fig. 15. Time-resolved Raman signals observed in the spectral domain [149] or in the time domain [77] were obtained for t* after actinic excitation of t in the UV. Their intensities decay on a time scale corresponding to the t* lifetime (i.e., ~ 80 ps in n-hexane). While the S0 trans-stilbene (t) Raman spectrum (Fig. 15, bottom) is characterized by the prominent 1639 and 1596 cm−1 modes, respectively, assigned to motions dominated by central C=C stretch and phenyl rings stretch [155], in the S1 state (t*) however, the same spectral range is dominated by a single mode at 1570 cm−1 and the accurate assignment of the t* vibrational modes remains challenging [149]. Multi-VCS therefore appears to be a powerful experimental approach for assessing the accuracy of state-of-the-art computational methodologies for excited-states modeling. The peak position of some t* Raman peaks (in particular the ~ 1570 cm−1 mode) are seen to shift on the 10-ps time scale by up to 5 cm−1 [147], depending on the excess vibrational energy [149], due to vibrational cooling in the excited state. This was exploited to investigate the mechanism of solute–solvent energy dissipation [98]. Sub-ps decay kinetics of the t* Raman peak intensities reported with FSRS were eventually identified to be controlled by experimental conditions (Raman depletion effect) [46], and indeed not confirmed by population-controlled pump-IVS [77].
Fig. 15

Trans-stilbene Raman spectra measured in n-hexane, in the t* excited state (S1) by PC-IVS (top, in black) or FSRS (middle, in brown), and in the t ground state (S0) by FSRS (bottom, in brown). Both S1 Raman spectra were recorded after actinic excitation at 325 or 326 nm. In the PC-IVS experiment [77], the vibrational coherence was subsequently induced with a 9-fs pulse centered at 550 nm and resonant with the t* excited state absorption (ESA), and probed in the time domain before Fourier transformation. In the FSRS experiment [149], a 645-nm, 2-ps-long Raman pump pulse is used, which is pre-resonant with the same ESA. The ground-state Raman spectrum is also acquired by FSRS, without actinic pulse (or at negative time delays) with the same 615-nm Raman pulse, which is off-resonant with respect to the ground state absorption

Reprinted with permission from Ref. [77]. Copyright 2014 American Chemical Society https://pubs.acs.org/doi/abs/10.1021%2Fjp5075863

Multi-VCS of cis-stilbene was also carried out similarly, but its very short c* lifetime increases the experimental challenge [37, 148, 149], as portrayed above for the excited states of carotenoids. A seminal pump-IVS experiment demonstrated the ability of multi-VCS to reveal quantitative structural information along the ultrafast isomerization pathway from c* to the CInt [37]. Following actinic UV excitation, a 620-nm, 11-fs pulse (repump) on resonance with ESA and SE of c* was used to impulsively trigger and subsequently probe a S1 vibrational wavepacket. A dominant oscillation is detected with a frequency around 240 cm−1, which down-shifts by up to 30 cm−1 on the ps time scale as a function of the waiting time between actinic and repump pulses. This mode, previously observed in the form of a Franck–Condon activated vibrational wavepacket in a conventional pump-probe experiment [156] and by ps-resolved Raman spectroscopy [148], is attributed to a spectator mode whose frequency shift reveals the anharmonicity and gradual topography change of the excited state PES as the system evolves from the FC point towards the CInt. FSRS experiments performed with a Raman pump on resonance with the c* ESA band [149] revealed Raman peaks broadened by the very short c* lifetime, but the background subtraction (inherent to FSRS data, see above) is in this case particularly challenging, leading the authors to consider the data with caution.

While both c and t photoisomerizations have long been postulated to occur via the same transient perpendicular p state [141], its experimental evidence and spectroscopic characterization in condensed phase came four decades later [143]. Recently, further characterization of the p* state was made by investigating stilbene derivatives where the formation and decay time of the p* state vary by more than two orders of magnitude, while its characteristic UV absorption band is invariably observed at 350 ± 20 nm depending on the substitutions [111, 158, 159]. The dependence of the p* lifetime on the solvent polarity is interpreted as the signature of its zwitterionic nature. Multi-VCS was performed on the trans- 1,1′-dicyanostilbene, which features a nearly 30-ps-long-lived p* state in n-hexane, thus facilitating the elucidation of some of its vibrational spectroscopic signatures thanks to a very sensitive FSRS set-up [111]. Three Raman peaks of a few μOD signal amplitude and characterized by a 20–30 ps decay kinetics were attributed to the transient p* species, including a 1558 cm−1 mode which was tentatively assigned to the phenyl quadrant vibrations, in the absence of any computational prediction of the p* vibrational modes.

The synthetic molecular rotary machines developed by Feringa’s group are based on so-called crowded derivatives of stilbene and stiff-stilbene [139, 160], such as the fluorene-based compound displayed in Fig. 14. The photoisomerization of such compounds also involves a transient excited state species. For stiff-stilbene, a sub-ps, viscosity-independent excited state relaxation is observed but the UV absorption signature of a stilbene-like perpendicular “phantom” state is not detected [161]. While the S0 Raman spectrum of stiff-stilbene also features the same double peak around 1600 cm−1 as in trans stilbene (see above), its S1 Raman spectrum shows a dominating peak at 1500 cm−1, i.e., downshifted by 70 cm−1 as compared to the 1570 cm−1 mode of the t* state of stilbene. For the fluorene-based rotary machine investigated by the Meech’s group, the 0.1-ps, viscosity-independent formation of a transient excited state is also observed, which is characterized as a dark state, and proposed to result from the ultrafast relaxation along a volume-conserving coordinate, such as pyramidalization of one of the carbon atoms of the isomerizing bond [54, 162]. For both stiff-stilbene and the fluorene-based machine, the subsequent decay to the ground state is instead viscosity-dependent like it is for trans-stilbene. This is expected for the torsion around the central C=C double bond in these compounds, since it is a large amplitude motion displacing a significant volume of solvent. Recently, Meech’s group applied multi-VCS to follow the structural relaxation of the transient dark state until the formation of the ground state isomer on the 1.6-ps time scale [119]. FSRS employing a Raman pump resonant with the dark state absorption around 550 nm revealed the Raman spectrum of the dark state dominated by 1345 and 1430 cm−1 Raman peaks, strongly downshifted with respect to the ground state Raman double peak at 1560 and 1585 cm−1 (see Fig. 16). These excited-state Raman signatures are discussed as being related to the isomerizing C=C double bond by analogy to the case of stiff-stilbene. They decay on the same time scale as the dark state population, to give rise to the Raman signature of the ground state photoproduct (so-called “unstable rotor”, see Fig. 16) dominated by the C=C stretch peak pair at 1510 and 1550 cm−1. The large frequency downshift of the C=C Raman signature of the dark state is tentatively attributed to the elongation of the central isomerizing bond in the excited state, in line with computational predictions [163].
Fig. 16

a Time-resolved Raman spectra of the excited, so-called “dark” state of the fluorene-based molecular rotor displayed in Fig. 14 right, recorded in cyclohexane. The blue spectrum is attributed to the Raman signature of the dark state. It decays on the same time-scale as the dark state lifetime to produce the Raman spectrum (yellow and brown spectra) attributed to the photoproduct, called “unstable rotor”. b Computed (DFT) and experimental (FSRS) ground state Raman spectra of the reactant ("GS stable rotor”) and photoproduct ("GS unstable rotor")

Adapted with permission from Ref. [54]. Copyright (2017) American Chemical Society

Finally, another common feature of substituted trans-stilbene derivatives is the existence of ground state, sub-populations of rotamers resulting from the thermally activated phenyl ring rotations [164]. While the various rotamers may not easily be distinguished by their steady-state UV–Vis spectroscopic signatures, they may feature distinct photoreaction kinetics and vibrational signatures. Multi-VCS was recently applied to the investigation of a family of di-fluorinated stilbene compounds [165]. This work demonstrates the efficiency of multi-VCS at discriminating the vibrational signatures and photoreaction kinetics of distinct subpopulations. Chemical and/or structural heterogeneity is a common feature of complex molecular systems in condensed phase. One present challenge of time-resolved non-linear spectroscopy and of physical chemistry is to resolve such heterogeneity, and multi-VCS is a promising spectroscopic tool for that.

6 Conclusions

The detection of Raman spectra as a function of photoreaction time is one of the most natural ways to investigate structure changes and interactions at the molecular level. Multi-VCS has achieved this goal by detecting stimulated Raman scattering (SRS) after the interaction with an actinic pulse triggering the photoreaction of interest. SRS is a third–order non-linear spectroscopy technique which exploits coherent light–matter interactions to prepare vibrational wavepackets in one (or two) electronic states and to probe their spectroscopic signatures either in the spectral domain or in the time domain. The examples discussed in this contribution portrayed how such stimulated Raman spectra are acquired by Multi-VCS and used to (1) map structural dynamics along a photoreaction, (2) identify transient molecular species, and (3) reveal chemical/structural heterogeneity of complex molecular systems in condensed phase. In particular, the ability to directly follow Raman shifts of only a few wavenumbers as well as Raman amplitudes during a photoreaction is a central feature of time- as well frequency-domain Multi-VCS methods.

The recent success of multi-VCS in clarifying the dynamics of a myriad of molecular system is mainly due to advances in key optical technologies of ultrashort pulse generation and the ability of tuning the excitation and probing spectra. On the one hand, multi-VCS in the time domain requires short pulses, i.e., below 15 fs to effectively induce high-frequency vibrational coherence. On the other hand, spectral tuneability has been shown to be a central piece when disentangling e.g. ground from excited-states vibrational dynamics as well as addressing different chromophores. These two aspects have become experimentally more accessible in recent years by the combination of commercial non-collinear optical parametric amplifiers (nc-OPA) and broadband chirped mirrors. As also shown in this contribution, this has enabled the application of multi-VCS to a wide range of different chromophores absorbing from the UV, over the visible and up to the near-infrared spectral region. Present experimental developments, e.g., towards microscopy, are the promise for turning multi-VCS into a so-called “high-content” analytical technique [167, 168]. In particular, the combination of e.g., multi-VCS with super-resolution microscopy exemplifies the ability to identify chemical compounds with spatial resolutions beyond the diffraction limit without any chemical labeling [169].

In spite of the success of multi-VCS, there are still several ongoing research topics. One of these topics is the extraction of the pure excited state vibrational dynamics. The complexity of molecular signal extraction is present in all multi-VCS methods and can be due to different causes (spectral overlap between ground- and excited-state absorptions, signal distortion due to other optical signals contributions, etc.). The development of cautious data post-processing, for example, to isolate reliably the specific stimulated Raman spectra of interest is still an intense research topic.

Another ongoing research topic in multi-VCS is the calculation of the evolution of Raman spectra in excited molecular states. Compared to other multidimensional techniques like 2D electronic or 2D infrared spectroscopies, the application of theoretical methods to calculate the evolution of Raman spectra in excited molecular states still is in its infancy, and in the overwhelming majority of the experimental cases, analysis of multi-VCS spectra is still done in a very qualitative way. From the theoretical point-of-view, models (like sum-over-states (SOS) and multimode Brownian oscillator) and nonlinear signal calculation techniques are well known, but there have been very few examples where the experimental optical signal in multi-VCS has been completely numerically simulated [102]. A major challenge for accurate modeling of third-order non-linear spectroscopy signals (VCS or UV–Vis 2DES) obtained in complex molecular systems still resides in the accurate quantum chemical simulation of excited molecular states and of their photoreaction dynamics. The contribution by Segarra-Martin et al. in this collection reviews the present state-of-the-art theoretical developments for the simulation of 2DES spectroscopy signal in complex molecular systems. Rapid progress in theoretical development is opening new horizons towards quantitative modeling of experimental signal based on realistic models of complex systems.

In conclusion, multi-VCS is now established as one of the state-of-the-art non-linear spectroscopy techniques for the investigation of ultrafast photoreactivity in molecules. More specifically, it is a technique of choice for monitoring vibrational dynamics in excited states with unparalleled time resolution. It is our belief that the combination of multi-VCS methods with recent and rapid theoretical developments will further enhance the impact of multi-VCS in unraveling ultrafast photoreaction mechanisms.

References

  1. 1.
    Joo TH, Albrecht AC (1993) Vibrational frequencies and dephasing times in excited electronic states by femtosecond time-resolved 4-wave-mixing. Chem Phys 173(1):17–26CrossRefGoogle Scholar
  2. 2.
    Weiner AM, Desilvestri S, Ippen EP (1985) 3-Pulse scattering for femtosecond dephasing studies—theory and experiment. J Opt Soc Am B 2(4):654–662CrossRefGoogle Scholar
  3. 3.
    Hwang H, Rossky PJ (2004) Electronic decoherence induced by intramolecular vibrational motions in a betaine dye molecule. J Phys Chem B 108(21):6723–6732.  https://doi.org/10.1021/jp037031b CrossRefGoogle Scholar
  4. 4.
    Vos MH, Jones MR, Martin JL (1998) Vibrational coherence in bacterial reaction centers: spectroscopic characterisation of motions active during primary electron transfer. Chem Phys 233(2–3):179–190.  https://doi.org/10.1016/S0301-0104(97)00355-8 CrossRefGoogle Scholar
  5. 5.
    Collini E, Wong CY, Wilk KE, Curmi PMG, Brumer P, Scholes GD (2010) Coherently wired light-harvesting in photosynthetic marine algae at ambient temperature. Nature 463(7281):U644–U669.  https://doi.org/10.1038/nature08811 CrossRefGoogle Scholar
  6. 6.
    Fuller FD, Pan J, Gelzinis A, Butkus V, Senlik SS, Wilcox DE, Yocum CF, Valkunas L, Abramavicius D, Ogilvie JP (2014) Vibronic coherence in oxygenic photosynthesis. Nat Chem 6(8):706–711.  https://doi.org/10.1038/nchem.2005 CrossRefPubMedGoogle Scholar
  7. 7.
    Huelga SF, Plenio MB (2013) Vibrations, quanta and biology. Contemp Phys 54(4):181–207.  https://doi.org/10.1080/00405000.2013.829687 CrossRefGoogle Scholar
  8. 8.
    Romero E, Augulis R, Novoderezhkin VI, Ferretti M, Thieme J, Zigmantas D, van Grondelle R (2014) Quantum coherence in photosynthesis for efficient solar-energy conversion. Nat Phys 10(9):677–683.  https://doi.org/10.1038/nphys3017 CrossRefGoogle Scholar
  9. 9.
    Scholes GD, Fleming GR, Olaya-Castro A, van Grondelle R (2011) Lessons from nature about solar light harvesting. Nat Chem 3(10):763–774.  https://doi.org/10.1038/nchem.1145 CrossRefPubMedGoogle Scholar
  10. 10.
    Gueye M, Manathunga M, Agathangelou D, Orozco Y, Paolino M, Fusi S, Haacke S, Olivucci M, Leonard J (2018) Engineering the vibrational coherence of vision into a synthetic molecular device. Nat Commun 9:313.  https://doi.org/10.1038/s41467-017-02668-w CrossRefPubMedPubMedCentralGoogle Scholar
  11. 11.
    Scholes GD, Fleming GR, Chen LX, Aspuru-Guzik A, Buchleitner A, Coker DF, Engel GS, van Grondelle R, Ishizaki A, Jonas DM, Lundeen JS, McCusker JK, Mukamel S, Ogilvie JP, Olaya-Castro A, Ratner MA, Spano FC, Whaley KB, Zhu XY (2017) Using coherence to enhance function in chemical and biophysical systems. Nature 543(7647):647–656.  https://doi.org/10.1038/nature21425 CrossRefPubMedGoogle Scholar
  12. 12.
    Rose TS, Rosker MJ, Zewail AH (1989) Femtosecond real-time probing of reactions. IV. The reactions of alkali halides. J Chem Phys 91(12):7415–7436CrossRefGoogle Scholar
  13. 13.
    Pollard WT, Fragnito HL, Bigot JY, Shank CV, Mathies RA (1990) Quantum-mechanical theory for 6 fs dynamic absorption spectroscopy and its application to Nile blue. Chem Phys Lett 168(3):239–245.  https://doi.org/10.1016/0009-2614(90)85603-A CrossRefGoogle Scholar
  14. 14.
    Fragnito HL, Bigot JY, Becker PC, Shank CV (1989) Evolution of the vibronic absorption spectrum in a molecule following impulsive excitation with a 6 fs optical pulse. Chem Phys Lett 160(2):101–104.  https://doi.org/10.1016/0009-2614(89)87564-5 CrossRefGoogle Scholar
  15. 15.
    Dhar L, Rogers JA, Nelson KA (1994) Time-resolved vibrational spectroscopy in the impulsive limit. Chem Rev 94(1):157–193.  https://doi.org/10.1021/cr00025a006 CrossRefGoogle Scholar
  16. 16.
    Ruhman S, Kosloff R (1990) Application of chirped ultrashort pulses for generating large-amplitude ground-state vibrational coherence: a computer simulation. J Opt Soc Am B 7(8):1748–1752.  https://doi.org/10.1364/josab.7.001748 CrossRefGoogle Scholar
  17. 17.
    Bardeen CJ, Wang Q, Shank CV (1998) Femtosecond chirped pulse excitation of vibrational wave packets in LD690 and bacteriorhodopsin. J Phys Chem A 102(17):2759–2766.  https://doi.org/10.1021/jp980346k CrossRefGoogle Scholar
  18. 18.
    Malkmus S, Dürr R, Sobotta C, Pulvermacher H, Zinth W, Braun M (2005) Chirp dependence of wave packet motion in oxazine 1. J Phys Chem A 109(46):10488–10492.  https://doi.org/10.1021/jp054462g CrossRefPubMedGoogle Scholar
  19. 19.
    Kahan A, Nahmias O, Friedman N, Sheves M, Ruhman S (2007) Following photoinduced dynamics in bacteriorhodopsin with 7-fs impulsive vibrational spectroscopy. J Am Chem Soc 129(3):537–546.  https://doi.org/10.1021/ja064910d CrossRefPubMedGoogle Scholar
  20. 20.
    Kraack JP, Motzkus M, Buckup T (2011) Selective nonlinear response preparation using femtosecond spectrally resolved four-wave-mixing. J Chem Phys 135(22):224505CrossRefGoogle Scholar
  21. 21.
    Wand A, Kallush S, Shoshanim O, Bismuth O, Kosloff R, Ruhman S (2010) Chirp effects on impulsive vibrational spectroscopy: a multimode perspective. Phys Chem Chem Phys 12(9):2149–2163CrossRefPubMedGoogle Scholar
  22. 22.
    Chesnoy J, Mokhtari A (1988) Resonant impulsive-stimulated Raman scattering on malachite green. Phys Rev A 38(7):3566–3576CrossRefGoogle Scholar
  23. 23.
    Pollard WT, Dexheimer SL, Wang Q, Peteanu LA, Shank CV, Mathies RA (1992) Theory of dynamic absorption spectroscopy of nonstationary states. 4. Application to 12-fs resonant impulsive Raman spectroscopy of bacteriorhodopsin. J Phys Chem 96(15):6147–6158.  https://doi.org/10.1021/j100194a013 CrossRefGoogle Scholar
  24. 24.
    Ruhman S, Joly AG, Nelson KA (1988) Coherent molecular vibrational motion observed in the time domain through impulsive stimulated Raman scattering. IEEE J Quantum Electron 24(2):460–469.  https://doi.org/10.1109/3.146 CrossRefGoogle Scholar
  25. 25.
    Mukamel S (1999) Principles of nonlinear optical spectroscopy, vol 6. Oxford University Press on Demand, OxfordGoogle Scholar
  26. 26.
    Pollard WT, Lee SY, Mathies RA (1990) Wave packet theory of dynamic absorption spectra in femtosecond pump–probe experiments. J Chem Phys 92(7):4012–4029.  https://doi.org/10.1063/1.457815 CrossRefGoogle Scholar
  27. 27.
    Johnson AE, Myers AB (1996) A comparison of time- and frequency-domain resonance Raman spectroscopy in triiodide. J Chem Phys 104(7):2497–2507.  https://doi.org/10.1063/1.470998 CrossRefGoogle Scholar
  28. 28.
    Liebel M, Schnedermann C, Wende T, Kukura P (2015) Principles and applications of broadband impulsive vibrational spectroscopy. J Phys Chem A 119(36):9506–9517.  https://doi.org/10.1021/acs.jpca.5b05948 CrossRefPubMedGoogle Scholar
  29. 29.
    Tanimura Y, Mukamel S (1993) Temperature-dependence and non-condon effects in pump-probe spectroscopy in the condensed-phase. J Opt Soc Am B 10(12):2263–2268.  https://doi.org/10.1364/josab.10.002263 CrossRefGoogle Scholar
  30. 30.
    Brazard J, Bizimana LA, Gellen T, Carbery WP, Turner DB (2016) Experimental detection of branching at a conical intersection in a highly fluorescent molecule. J Phys Chem Lett 7(1):14–19.  https://doi.org/10.1021/acs.jpclett.5b02476 CrossRefPubMedGoogle Scholar
  31. 31.
    Hamm P, Zanni MT (2011) Concepts and methods of 2d infrared spectroscopy. Cambridge University Press, CambridgeCrossRefGoogle Scholar
  32. 32.
    Siebert T, Schmitt M, Gräfe S, Engel V (2006) Ground state vibrational wave-packet and recovery dynamics studied by time-resolved CARS and pump-CARS spectroscopy. J Raman Spectrosc 37(1–3):397–403.  https://doi.org/10.1002/jrs.1441 CrossRefGoogle Scholar
  33. 33.
    Motzkus M, Pedersen S, Zewail AH (1996) Femtosecond real-time probing of reactions. 19. Nonlinear (DFWM) techniques for probing transition states of uni- and bimolecular reactions. J Phys Chem 100(14):5620–5633.  https://doi.org/10.1021/jp960265t CrossRefGoogle Scholar
  34. 34.
    Dobryakov AL, Quick M, Ioffe IN, Granovsky AA, Ernsting NP, Kovalenko SA (2014) Excited-state Raman spectroscopy with and without actinic excitation: S1 Raman spectra of trans-azobenzene. J Chem Phys 140(18):184310.  https://doi.org/10.1063/1.4874854 CrossRefPubMedGoogle Scholar
  35. 35.
    Sun Z, Lu J, Zhang DH, Lee S-Y (2008) Quantum theory of (femtosecond) time-resolved stimulated Raman scattering. J Chem Phys 128(14):144114.  https://doi.org/10.1063/1.2888551 CrossRefPubMedGoogle Scholar
  36. 36.
    Payne SA, Hochstrasser RM (1986) Picosecond coherent anti-stokes Raman scattering from the excited states of stilbene and benzophenone. J Phys Chem 90(10):2068–2074CrossRefGoogle Scholar
  37. 37.
    Takeuchi S, Ruhman S, Tsuneda T, Chiba M, Taketsugu T, Tahara T (2008) Spectroscopic tracking of structural evolution in ultrafast stilbene photoisomerization. Science 322(5904):1073–1077CrossRefPubMedGoogle Scholar
  38. 38.
    Motzkus M, Pedersen S, Zewail AH (1996) Femtosecond real-time probing of reactions.19. Nonlinear (DFWM) techniques for probing transition states of uni- and bimolecular reactions. J Phys Chem Us 100(14):5620–5633CrossRefGoogle Scholar
  39. 39.
    Hauer J, Buckup T, Motzkus M (2007) Pump-degenerate four wave mixing as a technique for analyzing structural and electronic evolution: multidimensional time-resolved dynamics near a conical intersection. J Phys Chem A 111(42):10517–10529CrossRefPubMedGoogle Scholar
  40. 40.
    McCamant DW, Kukura P, Mathies RA (2003) Femtosecond broadband stimulated Raman: a new approach for high-performance vibrational spectroscopy. Appl Spectrosc 57(11):1317–1323CrossRefPubMedPubMedCentralGoogle Scholar
  41. 41.
    Mallick B, Lakhsmanna A, Umapathy S (2011) Ultrafast Raman loss spectroscopy (URLS): instrumentation and principle. J Raman Spectrosc 42(10):1883–1890.  https://doi.org/10.1002/jrs.2996 CrossRefGoogle Scholar
  42. 42.
    Roy K, Kayal S, Kumar VR, Beeby A, Ariese F, Umapathy S (2017) Understanding ultrafast dynamics of conformation specific photo-excitation: a femtosecond transient absorption and ultrafast Raman loss study. J Phys Chem A 121(35):6538–6546.  https://doi.org/10.1021/acs.jpca.7b03893 CrossRefPubMedGoogle Scholar
  43. 43.
    Kayal S, Roy K, Umapathy S (2018) Femtosecond coherent nuclear dynamics of excited tetraphenylethylene: ultrafast transient absorption and ultrafast Raman loss spectroscopic studies. J Chem Phys 148(2):024301.  https://doi.org/10.1063/1.5008726 CrossRefPubMedGoogle Scholar
  44. 44.
    Tokmakoff A, Lang MJ, Larsen DS, Fleming GR, Chernyak V, Mukamel S (1997) Two-dimensional Raman spectroscopy of vibrational interactions in liquids. Phys Rev Lett 79(14):2702–2705.  https://doi.org/10.1103/PhysRevLett.79.2702 CrossRefGoogle Scholar
  45. 45.
    Yoshizawa M, Hattori Y, Kobayashi T (1994) Femtosecond time-resolved resonance Raman gain spectroscopy in polydiacetylene. Phys Rev B 49(18):13259–13262.  https://doi.org/10.1103/PhysRevB.49.13259 CrossRefGoogle Scholar
  46. 46.
    Kovalenko SA, Dobryakov AL, Ernsting NP (2011) An efficient setup for femtosecond stimulated Raman spectroscopy. Rev Sci Instrum 82(6):063102.  https://doi.org/10.1063/1.3596453 CrossRefPubMedGoogle Scholar
  47. 47.
    Rhinehart JM, Challa JR, McCamant DW (2012) Multimode charge-transfer dynamics of 4-(dimethylamino)benzonitrile probed with ultraviolet femtosecond stimulated Raman spectroscopy. J Phys Chem B 116(35):10522–10534.  https://doi.org/10.1021/jp3020645 CrossRefPubMedGoogle Scholar
  48. 48.
    Weigel A, Ernsting NP (2010) Excited stilbene: intramolecular vibrational redistribution and solvation studied by femtosecond stimulated Raman spectroscopy. J Phys Chem B 114(23):7879–7893.  https://doi.org/10.1021/jp100181z CrossRefPubMedGoogle Scholar
  49. 49.
    McCamant DW, Kukura P, Yoon S, Mathies RA (2004) Femtosecond broadband stimulated Raman spectroscopy: apparatus and methods. Rev Sci Instrum 75(11):4971–4980CrossRefPubMedPubMedCentralGoogle Scholar
  50. 50.
    Dietze DR, Mathies RA (2016) Femtosecond stimulated Raman spectroscopy. ChemPhysChem 17:1224–1251.  https://doi.org/10.1002/cphc.201600104 CrossRefPubMedGoogle Scholar
  51. 51.
    Quick M, Dobryakov AL, Kovalenko SA, Ernsting NP (2015) Resonance femtosecond-stimulated Raman spectroscopy without actinic excitation showing low-frequency vibrational activity in the S-2 state of all-trans beta-carotene. J Phys Chem Lett 6(7):1216–1220.  https://doi.org/10.1021/acs.jpclett.5b00243 CrossRefPubMedGoogle Scholar
  52. 52.
    Frobel S, Buschhaus L, Villnow T, Weingart O, Gilch P (2015) The photoformation of a phthalide: a ketene intermediate traced by FSRS. Phys Chem Chem Phys 17(1):376–386CrossRefPubMedGoogle Scholar
  53. 53.
    Laimgruber S, Schachenmayr H, Schmidt B, Zinth W, Gilch P (2006) A femtosecond stimulated Raman spectrograph for the near ultraviolet. Appl Phys B Lasers Opt 85(4):557–564CrossRefGoogle Scholar
  54. 54.
    Hall CR, Conyard J, Heisler IA, Jones G, Frost J, Browne WR, Feringa BL, Meech SR (2017) Ultrafast dynamics in light-driven molecular rotary motors probed by femtosecond stimulated Raman spectroscopy. J Am Chem Soc 139(21):7408–7414.  https://doi.org/10.1021/jacs.7b03599 CrossRefPubMedGoogle Scholar
  55. 55.
    Kloz M, van Grondelle R, Kennis JTM (2012) Correction for the time dependent inner filter effect caused by transient absorption in femtosecond stimulated Raman experiment. Chem Phys Lett 544:94–101.  https://doi.org/10.1016/j.cplett.2012.07.005 CrossRefGoogle Scholar
  56. 56.
    Kloz M, Weissenborn J, Polivka T, Frank HA, Kennis JTM (2016) Spectral watermarking in femtosecond stimulated Raman spectroscopy: resolving the nature of the carotenoid S-star state. Phys Chem Chem Phys 18(21):14619–14628.  https://doi.org/10.1039/c6cp01464j CrossRefPubMedGoogle Scholar
  57. 57.
    Kuramochi H, Takeuchi S, Tahara T (2016) Femtosecond time-resolved impulsive stimulated Raman spectroscopy using sub-7-fs pulses: apparatus and applications. Rev Sci Instrum 87(4):10.  https://doi.org/10.1063/1.4945259 CrossRefGoogle Scholar
  58. 58.
    Kraack JP, Wand A, Buckup T, Motzkus M, Ruhman S (2013) Mapping multidimensional excited state dynamics using pump-impulsive-vibrational-spectroscopy and pump-degenerate-four-wave-mixing. Phys Chem Chem Phys 15(34):14487–14501CrossRefPubMedGoogle Scholar
  59. 59.
    Cerullo G, De Silvestri S (2003) Ultrafast optical parametric amplifiers. Rev Sci Instrum 74(1):1–18.  https://doi.org/10.1063/1.1523642 CrossRefGoogle Scholar
  60. 60.
    Riedle E, Beutter M, Lochbrunner S, Piel J, Schenkl S, Sporlein S, Zinth W (2000) Generation of 10–50 fs pulses tunable through all of the visible and the NIR. Appl Phys B Lasers Opt 71(3):457–465.  https://doi.org/10.1007/s003400000351 CrossRefGoogle Scholar
  61. 61.
    Feng Y, Vinogradov I, Ge NH (2017) General noise suppression scheme with reference detection in heterodyne nonlinear spectroscopy. Opt Express 25(21):26262–26279CrossRefPubMedGoogle Scholar
  62. 62.
    Lanzani G, Cerullo G, Brabec C, Sariciftci NS (2003) Time domain investigation of the intrachain vibrational dynamics of a prototypical light-emitting conjugated polymer. Phys Rev Lett 90(4):047402CrossRefPubMedGoogle Scholar
  63. 63.
    Nagasawa Y, Yoneda Y, Nambu S, Muramatsu M, Takeuchi E, Tsumori H, Miyasaka H (2014) Femtosecond degenerate four-wave-mixing measurements of coherent intramolecular vibrations in an ultrafast electron transfer system. Vib Spectrosc 70:58–62.  https://doi.org/10.1016/j.vibspec.2013.11.006 CrossRefGoogle Scholar
  64. 64.
    Song Y, Hellmann C, Stingelin N, Scholes GD (2015) The separation of vibrational coherence from ground- and excited-electronic states in P3HT film. J Chem Phys 142(21):212410.  https://doi.org/10.1063/1.4916325 CrossRefPubMedGoogle Scholar
  65. 65.
    Ruhman S, Joly AG, Nelson KA (1987) Time-resolved observations of coherent molecular vibrational motion and the general occurrence of impulsive stimulated scattering. J Chem Phys 86(11):6563–6565CrossRefGoogle Scholar
  66. 66.
    Kumar ATN, Rosca F, Widom A, Champion PM (2001) Investigations of ultrafast nuclear response induced by resonant and nonresonant laser pulses. J Chem Phys 114(15):6795–6815CrossRefGoogle Scholar
  67. 67.
    Kumar ATN, Rosca F, Widom A, Champion PM (2001) Investigations of amplitude and phase excitation profiles in femtosecond coherence spectroscopy. J Chem Phys 114(2):701–724CrossRefGoogle Scholar
  68. 68.
    Cina JA, Kovac PA, Jumper CC, Dean JC, Scholes GD (2016) Ultrafast transient absorption revisited: phase-flips, spectral fingers, and other dynamical features. J Chem Phys 144(17):175102CrossRefPubMedGoogle Scholar
  69. 69.
    Kraack JP, Buckup T, Motzkus M (2013) Coherent high-frequency vibrational dynamics in the excited electronic state of all-trans retinal derivatives. J Phys Chem Lett 4(3):383–387CrossRefPubMedGoogle Scholar
  70. 70.
    Schnedermann C, Liebel M, Kukura P (2015) Mode-specificity of vibrationally coherent internal conversion in rhodopsin during the primary visual event. J Am Chem Soc 137(8):2886–2891.  https://doi.org/10.1021/ja508941k CrossRefPubMedGoogle Scholar
  71. 71.
    Kraack JP, Buckup T, Motzkus M (2011) Vibrational analysis of excited and ground electronic states of all-trans retinal protonated Schiff-bases. Phys Chem Chem Phys 13(48):21402–21410CrossRefPubMedGoogle Scholar
  72. 72.
    Ohta K, Larsen DS, Yang M, Fleming GR (2001) Influence of intramolecular vibrations in third-order, time-domain resonant spectroscopies. II. Numerical calculations. J Chem Phys 114(18):8020–8039.  https://doi.org/10.1063/1.1359241 CrossRefGoogle Scholar
  73. 73.
    Erland J, Balslev I (1993) Theory of quantum beat and polarization interference in four-wave mixing. Phys Rev A 48(3):R1765–R1768CrossRefPubMedGoogle Scholar
  74. 74.
    Koch M, Feldmann J, von Plessen G, Göbel EO, Thomas P, Köhler K (1992) Quantum beats versus polarization interference: an experimental distinction. Phys Rev Lett 69(25):3631–3634CrossRefPubMedGoogle Scholar
  75. 75.
    Faeder J, Pinkas I, Knopp G, Prior Y, Tannor DJ (2001) Vibrational polarization beats in femtosecond coherent anti-stokes Raman spectroscopy: a signature of dissociative pump–dump–pump wave packet dynamics. J Chem Phys 115(18):8440–8454.  https://doi.org/10.1063/1.1412253 CrossRefGoogle Scholar
  76. 76.
    Weigel A, Sebesta A, Kukura P (2015) Shaped and feedback-controlled excitation of single molecules in the weak-field limit. J Phys Chem Lett 6(20):4032–4037CrossRefPubMedPubMedCentralGoogle Scholar
  77. 77.
    Wende T, Liebel M, Schnedermann C, Pethick RJ, Kukura P (2014) Population-controlled impulsive vibrational spectroscopy: background- and baseline-free Raman spectroscopy of excited electronic states. J Phys Chem A 118(43):9976–9984CrossRefPubMedGoogle Scholar
  78. 78.
    Liebel M, Schnedermann C, Kukura P (2014) Vibrationally coherent crossing and coupling of electronic states during internal conversion in beta-carotene. Phys Rev Lett 112(19):198302CrossRefPubMedGoogle Scholar
  79. 79.
    Wohlleben W, Buckup T, Hashimoto H, Cogdell RJ, Herek JL, Motzkus M (2004) Pump-deplete-probe spectroscopy and the puzzle of carotenoid dark states. J Phys Chem B 108(10):3320–3325CrossRefGoogle Scholar
  80. 80.
    Buckup T, Savolainen J, Wohlleben W, Herek JL, Hashimoto H, Correia RRB, Motzkus M (2006) Pump-probe and pump-deplete-probe spectroscopies on carotenoids with N = 9–15 conjugated bonds. J Chem Phys 125(19):194505CrossRefPubMedGoogle Scholar
  81. 81.
    Larsen DS, Papagiannakis E, van Stokkum IHM, Vengris M, Kennis JTM, van Grondelle R (2003) Excited state dynamics of beta-carotene explored with dispersed multi-pulse transient absorption. Chem Phys Lett 381(5–6):733–742.  https://doi.org/10.1016/j.cplett.2003.10.016 CrossRefGoogle Scholar
  82. 82.
    Papagiannakis E, Vengris M, Larsen DS, van Stokkum IHM, Hiller RG, van Grondelle R (2006) Use of ultrafast dispersed pump-dump-probe and pump-repump-probe spectroscopies to explore the light-induced dynamics of peridinin in solution. J Phys Chem B 110(1):512–521.  https://doi.org/10.1021/jp053094d CrossRefPubMedGoogle Scholar
  83. 83.
    Thaller A, Laenen R, Laubereau A (2006) The precursors of the solvated electron in methanol studied by femtosecond pump-repump-probe spectroscopy. J Chem Phys 124(2):024515CrossRefPubMedGoogle Scholar
  84. 84.
    Draxler S, Brust T, Eicher J, Zinth W, Braun M (2010) Novel detection scheme for application in pump-repump-probe spectroscopy. Opt Commun 283(6):1050–1054CrossRefGoogle Scholar
  85. 85.
    van Wilderen LJGW, Clark IP, Towrie M, van Thor JJ (2009) Mid-infrared picosecond pump-dump-probe and pump-repump-probe experiments to resolve a ground-state intermediate in cyanobacterial phytochrome cph1. J Phys Chem B 113(51):16354–16364CrossRefPubMedGoogle Scholar
  86. 86.
    Bradler M, Werhahn JC, Hutzler D, Fuhrmann S, Heider R, Riedle E, Iglev H, Kienberger R (2013) A novel setup for femtosecond pump-repump-probe IR spectroscopy with few cycle CEP stable pulses. Opt Express 21(17):20145–20158CrossRefPubMedGoogle Scholar
  87. 87.
    Buckup T, Weigel A, Hauer J, Motzkus M (2010) Ultrafast multiphoton transient absorption of beta-carotene. Chem Phys 373(1–2):38–44CrossRefGoogle Scholar
  88. 88.
    Fujisawa T, Kuramochi H, Hosoi H, Takeuchi S, Tahara T (2016) Role of coherent low-frequency motion in excited-state proton transfer of green fluorescent protein studied by time-resolved impulsive stimulated Raman spectroscopy. J Am Chem Soc 138(12):3942–3945.  https://doi.org/10.1021/jacs.5b11038 CrossRefPubMedGoogle Scholar
  89. 89.
    Hoffman DP, Mathies RA (2012) Photoexcited structural dynamics of an azobenzene analog 4-nitro-4[prime or minute]-dimethylamino-azobenzene from femtosecond stimulated Raman. Phys Chem Chem Phys 14:6298–6306.  https://doi.org/10.1039/C2CP23468H CrossRefPubMedGoogle Scholar
  90. 90.
    Gustafson TL, Roberts DM, Chernoff DA (1983) Picosecond transient Raman-spectroscopy—the photo-isomerization of trans-stilbene. J Chem Phys 79(4):1559–1564.  https://doi.org/10.1063/1.446027 CrossRefGoogle Scholar
  91. 91.
    Gustafson TL, Roberts DM, Chernoff DA (1984) The structure of electronic excited-states in trans-stilbene—picosecond transient stokes and anti-stokes Raman-spectra. J Chem Phys 81(8):3438.  https://doi.org/10.1063/1.448068 CrossRefGoogle Scholar
  92. 92.
    Hashimoto H, Koyama Y, Hirata Y, Mataga N (1991) S1 and T1 species of beta-carotene generated by direct photoexcitation from the all-trans, 9-cis, 13-cis, and 15-cis isomers as revealed by picosecond transient absorption and transient Raman spectroscopies. J Phys Chem Us 95(8):3072–3076.  https://doi.org/10.1021/j100161a022 CrossRefGoogle Scholar
  93. 93.
    Weaver WL, Huston LA, Iwata K, Gustafson TL (1992) Solvent solute interactions probed by picosecond transient Raman-spectroscopy—mode-specific vibrational dynamics in S1 trans-stilbene. J Phys Chem Us 96(22):8956–8961.  https://doi.org/10.1021/j100201a047 CrossRefGoogle Scholar
  94. 94.
    Laubereau A, Vonderli D, Kaiser W (1972) Direct measurement of vibrational lifetimes of molecules in liquids. Phys Rev Lett 28(18):1162.  https://doi.org/10.1103/physrevlett.28.1162 CrossRefGoogle Scholar
  95. 95.
    Valley DT, Hoffman DP, Mathies RA (2015) Reactive and unreactive pathways in a photochemical ring opening reaction from 2D femtosecond stimulated Raman. Phys Chem Chem Phys 17(14):9231–9240.  https://doi.org/10.1039/C4CP05323K CrossRefPubMedGoogle Scholar
  96. 96.
    Ferrante C, Pontecorvo E, Cerullo G, Vos M, Scopigno T (2016) Direct observation of subpicosecond vibrational dynamics in photoexcited myoglobin. Nat Chem 8:1137CrossRefPubMedGoogle Scholar
  97. 97.
    Weigel A, Dobryakov A, Klaumunzer B, Sajadi M, Saalfrank P, Ernsting N (2011) Femtosecond stimulated Raman spectroscopy of flavin after optical excitation. J Phys Chem B 115:3656–3680CrossRefPubMedGoogle Scholar
  98. 98.
    Iwata K, H-o Hamaguchi (1997) Microscopic mechanism of solute- solvent energy dissipation probed by picosecond time-resolved Raman spectroscopy. J Phys Chem A 101:632–637CrossRefGoogle Scholar
  99. 99.
    Kuramochi H, Takeuchi S, Yonezawa K, Kamikubo H, Kataoka M, Tahara T (2017) Probing the early stages of photoreception in photoactive yellow protein with ultrafast time-domain Raman spectroscopy. Nat Chem 9:660–666.  https://doi.org/10.1038/nchem.2717 CrossRefPubMedGoogle Scholar
  100. 100.
    Namboodiri V, Scaria A, Namboodiri M, Materny A (2009) Investigation of molecular dynamics in beta-carotene using femtosecond pump-FWM spectroscopy. Laser Phys 19(2):154–161.  https://doi.org/10.1134/s1054660x09020029 CrossRefGoogle Scholar
  101. 101.
    Buckup T, Hauer J, Mohring J, Motzkus M (2009) Multidimensional spectroscopy of beta-carotene: vibrational cooling in the excited state. Arch Biochem Biophys 483(2):219–223CrossRefPubMedGoogle Scholar
  102. 102.
    Marek MS, Buckup T, Motzkus M (2011) Direct observation of a dark state in lycopene using pump-DFWM. J Phys Chem B 115(25):8328–8337CrossRefPubMedGoogle Scholar
  103. 103.
    Liebel M, Kukura P (2013) Broad-band impulsive vibrational spectroscopy of excited electronic states in the time domain. J Phys Chem Lett 4(8):1358–1364CrossRefPubMedGoogle Scholar
  104. 104.
    Miki T, Buckup T, Krause MS, Southall J, Cogdell RJ, Motzkus M (2016) Vibronic coupling in the excited-states of carotenoids. Phys Chem Chem Phys 18(16):11443–11453.  https://doi.org/10.1039/c5cp07542d CrossRefPubMedGoogle Scholar
  105. 105.
    Takaya T, Anan M, Iwata K (2018) Vibrational relaxation dynamics of b-carotene and its derivatives with substituents on terminal rings in electronically excited states as studied by femtosecond time-resolved stimulated Raman spectroscopy in the near-IR region. Phys Chem Chem Phys 20(5):3320–3327CrossRefPubMedGoogle Scholar
  106. 106.
    Kraack JP, Buckup T, Hampp N, Motzkus M (2011) Ground- and excited-state vibrational coherence dynamics in bacteriorhodopsin probed with degenerate four-wave-mixing experiments. ChemPhysChem 12(10):1851–1859CrossRefPubMedGoogle Scholar
  107. 107.
    Kraack JP, Buckup T, Motzkus M (2012) Evidence for the two-state-two-mode model in retinal protonated schiff-bases from pump degenerate four-wave-mixing experiments. Phys Chem Chem Phys 14(40):13979–13988CrossRefPubMedGoogle Scholar
  108. 108.
    Liebel M, Schnedermann C, Bassolino G, Taylor G, Watts A, Kukura P (2014) Direct observation of the coherent nuclear response after the absorption of a photon. Phys Rev Lett 112(23):4.  https://doi.org/10.1103/PhysRevLett.112.238301 CrossRefGoogle Scholar
  109. 109.
    Hoffman DP, Mathies RA (2016) Femtosecond stimulated Raman exposes the role of vibrational coherence in condensed-phase photoreactivity. Acc Chem Res 49:616–625.  https://doi.org/10.1021/acs.accounts.5b00508 CrossRefPubMedGoogle Scholar
  110. 110.
    Barclay MS, Quincy TJ, Williams-Young DB, Caricato M, Elles CG (2017) Accurate assignments of excited-state resonance Raman spectra: a benchmark study combining experiment and theory. J Phys Chem A 121(41):7937–7946.  https://doi.org/10.1021/acs.jpca.7b09467 CrossRefPubMedGoogle Scholar
  111. 111.
    Quick M, Dobryakov AL, Ioffe IN, Granovsky AA, Kovalenko SA, Ernsting NP (2016) Perpendicular state of an electronically excited stilbene: observation by femtosecond-stimulated Raman spectroscopy. J Phys Chem Lett 7(20):4047–4052.  https://doi.org/10.1021/acs.jpclett.6b01923 CrossRefPubMedGoogle Scholar
  112. 112.
    Han FY, Liu WM, Zhu LD, Wang YL, Fang C (2016) Initial hydrogen-bonding dynamics of photoexcited coumarin in solution with femtosecond stimulated Raman spectroscopy. J Mater Chem C 4(14):2954–2963.  https://doi.org/10.1039/c5tc03598h CrossRefGoogle Scholar
  113. 113.
    Musser AJ, Liebel M, Schnedermann C, Wende T, Kehoe TB, Rao A, Kukura P (2015) Evidence for conical intersection dynamics mediating ultrafast singlet exciton fission. Nat Phys 11(4):352–357.  https://doi.org/10.1038/nphys3241 CrossRefGoogle Scholar
  114. 114.
    Kuramochi H, Fujisawa T, Takeuchi S, Tahara T (2017) Broadband stimulated Raman spectroscopy in the deep ultraviolet region. Chem Phys Lett 683:543–546.  https://doi.org/10.1016/j.cplett.2017.02.015 CrossRefGoogle Scholar
  115. 115.
    Harris MA, Mishra AK, Young RM, Brown KE, Wasielewski MR, Lewis FD (2016) Direct observation of the hole carriers in DNA photoinduced charge transport. J Am Chem Soc 138:5491–5494CrossRefPubMedGoogle Scholar
  116. 116.
    Oscar BG, Liu WM, Zhao YX, Tang LT, Wang YL, Campbell RE, Fang C (2014) Excited-state structural dynamics of a dual-emission calmodulin-green fluorescent protein sensor for calcium ion imaging. Proc Natl Acad Sci USA 111(28):10191–10196.  https://doi.org/10.1073/pnas.1403712111 CrossRefPubMedGoogle Scholar
  117. 117.
    Creelman M, Kumauchi M, Hoff WD, Mathies RA (2014) Chromophore dynamics in the PYP photocycle from femtosecond stimulated Raman spectroscopy. J Phys Chem B 118(3):659–667.  https://doi.org/10.1021/jp408584v CrossRefPubMedGoogle Scholar
  118. 118.
    Roy K, Kayal S, Ariese F, Beeby A, Umapathy S (2017) Mode specific excited state dynamics study of bis(phenylethynyl) benzene from ultrafast Raman loss spectroscopy. J Chem Phys 146(6):064303.  https://doi.org/10.1063/1.4975174 CrossRefPubMedGoogle Scholar
  119. 119.
    Hall CR, Heisler IA, Jones GA, Frost JE, Gil AA, Tonge PJ, Meech SR (2017) Femtosecond stimulated Raman study of the photoactive flavoprotein AppA(BLUF). Chem Phys Lett 683:365–369.  https://doi.org/10.1016/j.cplett.2017.03.030 CrossRefGoogle Scholar
  120. 120.
    Kukura P, McCamant DW, Yoon S, Wandschneider DB, Mathies RA (2005) Structural observation of the primary isomerization in vision with femtosecond-stimulated Raman. Science 310(5750):1006–1009CrossRefPubMedGoogle Scholar
  121. 121.
    Polli D, Altoe P, Weingart O, Spillane KM, Manzoni C, Brida D, Tomasello G, Orlandi G, Kukura P, Mathies RA, Garavelli M, Cerullo G (2010) Conical intersection dynamics of the primary photoisomerization event in vision. Nature 467(7314):U440–U488CrossRefGoogle Scholar
  122. 122.
    Frank HA (1999) The photochemistry of carotenoids, vol 8. Advances in photosynthesis. Kluwer Academic, DordrechtGoogle Scholar
  123. 123.
    Polivka T, Sundstrom V (2004) Ultrafast dynamics of carotenoid excited states—from solution to natural and artificial systems. Chem Rev 104(4):2021–2071.  https://doi.org/10.1021/cr020674n CrossRefPubMedGoogle Scholar
  124. 124.
    Buckup T, Savolainen J, Wohlleben W, Herek JL, Hashimoto H, Correia RRB, Motzkus M (2006) Pump-probe and pump-deplete-probe spectroscopies on carotenoids with N = 9–15 conjugated bonds. J Chem Phys 125(19):194505.  https://doi.org/10.1063/1.2388274 CrossRefPubMedGoogle Scholar
  125. 125.
    Jailaubekov AE, Song SH, Vengris M, Cogdell RJ, Larsen DS (2010) Using narrowband excitation to confirm that the S* state in carotenoids is not a vibrationally-excited ground state species. Chem Phys Lett 487(1–3):101–107.  https://doi.org/10.1016/j.cplett.2010.01.014 CrossRefGoogle Scholar
  126. 126.
    Hauer J, Maiuri M, Viola D, Lukes V, Henry S, Carey AM, Cogdell RJ, Cerullo G, Polli D (2013) Explaining the temperature dependence of spirilloxanthin’s S* signal by an inhomogeneous ground state model. J Phys Chem A 117(29):6303–6310.  https://doi.org/10.1021/jp4011372 CrossRefPubMedPubMedCentralGoogle Scholar
  127. 127.
    Ehlers F, Scholz M, Schimpfhauser J, Bienert J, Oum K, Lenzer T (2015) Collisional relaxation of apocarotenals: identifying the S* state with vibrationally excited molecules in the ground electronic state S-0*. Phys Chem Chem Phys 17(16):10478–10488.  https://doi.org/10.1039/c4cp05600k CrossRefPubMedGoogle Scholar
  128. 128.
    Balevicius V, Abramavicius D, Polivka T, Pour AG, Hauer J (2016) A unified picture of S* in carotenoids. J Phys Chem Lett 7(17):3347–3352.  https://doi.org/10.1021/acs.jpclett.6b01455 CrossRefPubMedPubMedCentralGoogle Scholar
  129. 129.
    Tavan P, Schulten K (1987) Electronic excitations in finite and infinite polyenes. Phys Rev B 36(8):4337–4358.  https://doi.org/10.1103/PhysRevB.36.4337 CrossRefGoogle Scholar
  130. 130.
    Orlandi G, Zerbetto F (1986) Vibronic coupling in polyenes—the frequency increase of the active C=C Ag stretching mode in the absorption-spectra. Chem Phys 108(2):187–195.  https://doi.org/10.1016/0301-0104(86)85040-6 CrossRefGoogle Scholar
  131. 131.
    Nagae H, Kakitani Y, Koyama Y (2009) Theoretical description of diabatic mixing and coherent excitation in singlet-excited states of carotenoids. Chem Phys Lett 474(4–6):342–351.  https://doi.org/10.1016/j.cplett.2009.04.039 CrossRefGoogle Scholar
  132. 132.
    Yoshizawa M, Aoki H, Hashimoto H (2001) Vibrational relaxation of the 2A(g)(−) excited state in all-trans-beta-carotene obtained by femtosecond time-resolved Raman spectroscopy. Phys Rev B 63(18):180301.  https://doi.org/10.1103/PhysRevB.63.180301 CrossRefGoogle Scholar
  133. 133.
    McCamant DW, Kukura P, Mathies RA (2003) Femtosecond time-resolved stimulated Raman spectroscopy: application to the ultrafast internal conversion in beta-carotene. J Phys Chem A 107(40):8208–8214CrossRefPubMedPubMedCentralGoogle Scholar
  134. 134.
    Yoshizawa M, Nakamura R, Yoshimatsu O, Abe K, Sakai S, Nakagawa K, Fujii R, Nango M, Hashimoto H (2012) Femtosecond stimulated Raman spectroscopy of the dark S-1 excited state of carotenoid in photosynthetic light harvesting complex. Acta Biochim Pol 59(1):49–52PubMedGoogle Scholar
  135. 135.
    Quick M, Kasper MA, Richter C, Mahrwald R, Dobryakov AL, Kovalenko SA, Ernsting NP (2015) Beta-carotene revisited by transient absorption and stimulated Raman spectroscopy. ChemPhysChem 16(18):3824–3835.  https://doi.org/10.1002/cphc.201500586 CrossRefPubMedGoogle Scholar
  136. 136.
    Kukura P, McCamant DW, Mathies RA (2004) Femtosecond time-resolved stimulated Raman spectroscopy of the S-2 (1B(u)(+)) excited state of beta-carotene. J Phys Chem A 108(28):5921–5925CrossRefPubMedPubMedCentralGoogle Scholar
  137. 137.
    Kardas TM, Ratajska-Gadomska B, Lapini A, Ragnoni E, Righini R, Di Donato M, Foggi P, Gadomski W (2014) Dynamics of the time-resolved stimulated Raman scattering spectrum in presence of transient vibronic inversion of population on the example of optically excited trans-beta-apo-8′-carotenal. J Chem Phys.  https://doi.org/10.1063/1.4879060 CrossRefPubMedGoogle Scholar
  138. 138.
    Shim S, Mathies RA (2008) Development of a tunable femtosecond stimulated Raman apparatus and its application to beta-carotene. J Phys Chem B 112(15):4826–4832CrossRefPubMedGoogle Scholar
  139. 139.
    Koumura N, Zijlstra RWJ, van Delden RA, Harada N, Feringa BL (1999) Light-driven monodirectional molecular rotor. Nature 401:152CrossRefPubMedGoogle Scholar
  140. 140.
    van Leeuwen T, Lubbe AS, Štacko P, Wezenberg SJ, Feringa BL (2017) Dynamic control of function by light-driven molecular motors. Nat Rev Chem 1:0096CrossRefGoogle Scholar
  141. 141.
    Saltiel J (1967) Perdeuteriostilbene. The role of phantom states in the cis-trans photoisomerization of stilbenes. J Am Chem Soc 89:1036–1037CrossRefGoogle Scholar
  142. 142.
    Sension RJ, Repinec ST, Szarka AZ, Hochstrasser RM (1993) Femtosecond laser studies of the cis–stilbene photoisomerization reactions. J Chem Phys 98:6291–6315.  https://doi.org/10.1063/1.464824 CrossRefGoogle Scholar
  143. 143.
    Kovalenko SA, Dobryakov AL, Ioffe I, Ernsting NP (2010) Evidence for the phantom state in photoinduced cis–trans isomerization of stilbene. Chem Phys Lett 493:255–258.  https://doi.org/10.1016/j.cplett.2010.05.022 CrossRefGoogle Scholar
  144. 144.
    Quenneville J, Martínez TJ (2003) Ab initio study of cis–trans photoisomerization in stilbene and ethylene. J Phys Chem A 107:829–837.  https://doi.org/10.1021/jp021210w CrossRefGoogle Scholar
  145. 145.
    Minezawa N, Gordon MS (2011) Photoisomerization of stilbene: a spin-flip density functional theory approach. J Phys Chem A 115:7901–7911.  https://doi.org/10.1021/jp203803a CrossRefPubMedGoogle Scholar
  146. 146.
    Harabuchi Y, Keipert K, Zahariev F, Taketsugu T, Gordon MS (2014) Dynamics simulations with spin-flip time-dependent density functional theory: photoisomerization and photocyclization mechanisms of cis-stilbene in ππ* states. J Phys Chem A 118:11987–11998.  https://doi.org/10.1021/jp5072428 CrossRefPubMedGoogle Scholar
  147. 147.
    Iwata K, H-o Hamaguchi (1992) Picosecond structural relaxation of S1 trans-stilbene in solution as revealed by time-resolved Raman spectroscopy. Chem Phys Lett 196:462–468.  https://doi.org/10.1016/0009-2614(92)85721-L CrossRefGoogle Scholar
  148. 148.
    Kwok WM, Ma C, Phillips D, Beeby A, Marder TB, Thomas RL, Tschuschke C, Baranović G, Matousek P, Towrie M, Parker AW (2003) Time-resolved resonance Raman study of S1cis-stilbene and its deuterated isotopomers. J Raman Spectrosc 34:886–891.  https://doi.org/10.1002/jrs.1070 CrossRefGoogle Scholar
  149. 149.
    Dobryakov AL, Ioffe I, Granovsky AA, Ernsting NP, Kovalenko SA (2012) Femtosecond Raman spectra of cis-stilbene and trans-stilbene with isotopomers in solution. J Chem Phys 137:244505.  https://doi.org/10.1063/1.4769971 CrossRefPubMedGoogle Scholar
  150. 150.
    Sakamoto A, Tanaka F, Tasumi M, Torii H, Kawato K, Furuya K (2006) Comparison of the Raman spectrum of trans-stilbene in the S1 state calculated by the CIS method and the spectra observed under resonant and off-resonant conditions. A collection of papers presented at the 3rd international conference on advanced vibrational spectroscopy (ICAVS-3), Delavan, WI, USA, 14–19 August 2005—Part 1 42:176–182.  https://doi.org/10.1016/j.vibspec.2006.04.001 CrossRefGoogle Scholar
  151. 151.
    Tsumura K, Furuya K, Sakamoto A, Tasumi M (2008) Vibrational analysis of trans-stilbene in the excited singlet state by time-dependent density functional theory: calculations of the Raman, infrared, and fluorescence excitation spectra. J Raman Spectrosc 39:1584–1591.  https://doi.org/10.1002/jrs.2095 CrossRefGoogle Scholar
  152. 152.
    Angeli C, Improta R, Santoro F (2009) On the controversial nature of the 1 B1u and 2 B1u states of trans-stilbene: the n-electron valence state perturbation theory approach. J Chem Phys 130:174307.  https://doi.org/10.1063/1.3131263 CrossRefPubMedGoogle Scholar
  153. 153.
    Ioffe IN, Granovsky AA (2013) Photoisomerization of stilbene: the detailed XMCQDPT2 treatment. J Chem Theory Comput 9:4973–4990.  https://doi.org/10.1021/ct400647w CrossRefPubMedGoogle Scholar
  154. 154.
    Orlandi G, Garavelli M, Zerbetto F (2017) Analysis of the vibronic structure of the trans-stilbene fluorescence and excitation spectra: the S 0 and S 1 PES along the C e [double bond, length as m-dash] C e and C e–C ph torsions. Phys Chem Chem Phys 19:25095–25104CrossRefPubMedGoogle Scholar
  155. 155.
    Meić Z, Güsten H (1978) Vibrational studies of trans-stilbenes—I. Infrared and Raman spectra of trans-stilbene and deuterated trans-stilbenes. Spectrochim Acta Part A 34(1):101–111.  https://doi.org/10.1016/0584-8539(78)80193-7 CrossRefGoogle Scholar
  156. 156.
    Ishii K, Takeuchi S, Tahara T (2004) A 40-fs time-resolved absorption study on cis-stilbene in solution: observation of wavepacket motion on the reactive excited state. Chem Phys Lett 398:400–406.  https://doi.org/10.1016/j.cplett.2004.09.075 CrossRefGoogle Scholar
  157. 157.
    Berndt F, Dobryakov AL, Quick M, Mahrwald R, Ernsting NP, Lenoir D, Kovalenko SA (2012) Long-lived perpendicular conformation in the photoisomerization path of 1,1′-dimethylstilbene and 1,1′-diethylstilbene. Chem Phys Lett 544:39–42.  https://doi.org/10.1016/j.cplett.2012.07.007 CrossRefGoogle Scholar
  158. 158.
    Dobryakov AL, Quick M, Lenoir D, Detert H, Ernsting NP, Kovalenko SA (2016) Time-resolved photoisomerization of 1,1′-di-tert-butylstilbene and 1,1′-dicyanostilbene. Chem Phys Lett 652:225–229.  https://doi.org/10.1016/j.cplett.2016.04.060 CrossRefGoogle Scholar
  159. 159.
    Dobryakov AL, Quick M, Richter C, Knie C, Ioffe IN, Granovsky AA, Mahrwald R, Ernsting NP, Kovalenko SA (2017) Photoisomerization pathways and Raman activity of 1,1′-difluorostilbene. J Chem Phys 146:044501.  https://doi.org/10.1063/1.4974357 CrossRefPubMedGoogle Scholar
  160. 160.
    Pollard MM, Meetsma A, Feringa BL (2008) A redesign of light-driven rotary molecular motors. Org Biomol Chem 6:507–512.  https://doi.org/10.1039/B715652A CrossRefPubMedGoogle Scholar
  161. 161.
    Quick M, Berndt F, Dobryakov AL, Ioffe IN, Granovsky AA, Knie C, Mahrwald R, Lenoir D, Ernsting NP, Kovalenko SA (2014) Photoisomerization dynamics of stiff-stilbene in solution. J Phys Chem B 118:1389–1402.  https://doi.org/10.1021/jp411656x CrossRefPubMedGoogle Scholar
  162. 162.
    Conyard J, Addison K, Heisler IA, Cnossen A, Browne WR, Feringa BL, Meech SR (2012) Ultrafast dynamics in the power stroke of a molecular rotary motor. Nat Chem 4:547–551.  https://doi.org/10.1038/nchem.1343 CrossRefPubMedGoogle Scholar
  163. 163.
    Kazaryan A, Kistemaker JCM, Schäfer LV, Browne WR, Feringa BL, Filatov M (2010) Understanding the dynamics behind the photoisomerization of a light-driven fluorene molecular rotary motor. J Phys Chem A 114:5058–5067.  https://doi.org/10.1021/jp100609m CrossRefPubMedGoogle Scholar
  164. 164.
    Mazzucato U, Momicchioli F (1991) Rotational isomerism in trans-1,2-diarylethylenes. Chem Rev 91:1679–1719.  https://doi.org/10.1021/cr00008a002 CrossRefGoogle Scholar
  165. 165.
    Quick M, Dobryakov AL, Ioffe IN, Berndt F, Mahrwald R, Ernsting NP, Kovalenko SA (2018) Rotamer-specific photoisomerization of difluorostilbenes from transient absorption and transient Raman spectroscopy. J Phys Chem B.  https://doi.org/10.1021/acs.jpcb.7b09283 CrossRefPubMedGoogle Scholar
  166. 166.
    Ploetz E, Laimgruber S, Berner S, Zinth W, Gilch P (2007) Femtosecond stimulated Raman microscopy. Appl Phys B 87(3):389–393.  https://doi.org/10.1007/s00340-007-2630-x CrossRefGoogle Scholar
  167. 167.
    Schnedermann C, Lim JM, Wende T, Duarte AS, Ni L, Gu Q, Sadhanala A, Rao A, Kukura P (2016) Sub-10 fs time-resolved vibronic optical microscopy. J Phys Chem Lett 7(23):4854–4859.  https://doi.org/10.1021/acs.jpclett.6b02387 CrossRefPubMedPubMedCentralGoogle Scholar
  168. 168.
    Czerwinski L, Nixdorf J, Florio GD, Gilch P (2016) Broadband stimulated Raman microscopy with 0.1 ms pixel acquisition time. Opt Lett 41(13):3021–3024.  https://doi.org/10.1364/OL.41.003021 CrossRefPubMedGoogle Scholar
  169. 169.
    Silva WR, Graefe CT, Frontiera RR (2016) Toward label-free super-resolution microscopy. ACS Photonics 3(1):79–86.  https://doi.org/10.1021/acsphotonics.5b00467 CrossRefGoogle Scholar
  170. 170.
    Buckup T, Motzkus M (2014) Multidimensional time-resolved spectroscopy of vibrational coherence in biopolyenes. In: Johnson MA, Martinez TJ (eds) Annual review of physical chemistry, vol 65., pp 39–57.  https://doi.org/10.1146/annurev-physchem-040513-103619 CrossRefGoogle Scholar

Copyright information

© Springer Nature Switzerland AG 2018

Authors and Affiliations

  1. 1.Physikalisch-Chemisches InstitutUniversität HeidelbergHeidelbergGermany
  2. 2.Université de Strasbourg, CNRS, Institut de Physique et Chimie des Matériaux de Strasbourg, UMR 7504, and Labex NIEStrasbourgFrance

Personalised recommendations