206
T. Buckup et al.
with high resolution, particularly in excited states, is a major pre-requisite to fully
understand how structural changes in (bio-)chromophores take place [7–10].
Time-resolved spectroscopy of excited potential surfaces is, though, more challenging than the ground state spectroscopy due to two major aspects. The first one
is the number of excited molecules contributing to the signal. Under normal excitation regimes (i.e. well below saturation), the number of excited molecules is much
smaller than the number of molecules in the ground state. Contributions from the
excited state may therefore be weak and difficult to detect. This can be specially
complicated in (self-)heterodyne detection methods if ground and excited state contributions spectrally overlap. In methods using such kind of detection, the intensities
of different contributions add linearly, which can obscure, or even totally cancel,
weak signals. Such an effect can be observed e.g. in transient absorption (TA) when
a negative signal like stimulated emission (SE) overlaps with an excited state absorption (ESA). This contrasts to background-free third-order methods, where all
contributions will be detected with a much higher sensitivity to small variations of
the dynamics in the excited state.
The second aspect is related to the time scale of processes in the excited state.
After excitation, population and coherence in the higher electronic states are not in
equilibrium but evolve in many (bio-)chromophores often on a very fast time scale.
Such time scales range from hundreds of femtoseconds as found for the isomerization of protein-bound retinal [11] to just some tens of femtoseconds as observed for
the internal conversion between excited states of carotenoids [12]. While ultrafast
transient population evolution can in principle be resolved by using state-of-the-art
sub-10 fs pulses [13], following the evolution of the vibrational coherence is not
trivial. Not just the amplitude of oscillatory phenomena contains information on
the deactivation of the excited state, but also the frequency and the phase may offer important insight. Nevertheless, determination of frequency and phase during
an ultrafast relaxation process is demanding, which has received a lot of attention
recently. It requires accurate analysis in order to avoid phase and frequency changes
due to pure analysis artifacts [14, 15].
One way of addressing these central aspects in time-resolved spectroscopy is to
increase the dimensionality of the nonlinear optical interaction. Multidimensional
time-resolved spectroscopy employs additional ultrashort pulses in order to open a
new spectral and/or temporal observation window, which is normally hidden from
lower-order techniques [16]. For example, this allows probing a specific molecular
response without the need of complex fitting algorithms or assumptions [17, 18]. In
this chapter, we show how four wave mixing methods with three beams like coherent
anti-Stokes Raman scattering (CARS) and transient grating (TG) can be combined
with an additional initial pump beam to be selective to excited state dynamics in
the condensed phase (Fig. 9.1). The so called pump-degenerate four wave mixing
(pump-DFWM) technique [19, 20] consists of two interactions with the initial pump
pulse to produce coherences and populations in the excited state, which will be
probed by the consecutive four wave mixing sequence. In this sense, pump-DFWM
can be thought of as a “pump-probe” method, where the probe is not a single pulse,
but is a nonlinear probe sequence of three ultrashort pulses. Due to their bandwidths,
T. Buckup et al.
with high resolution, particularly in excited states, is a major pre-requisite to fully
understand how structural changes in (bio-)chromophores take place [7–10].
Time-resolved spectroscopy of excited potential surfaces is, though, more challenging than the ground state spectroscopy due to two major aspects. The first one
is the number of excited molecules contributing to the signal. Under normal excitation regimes (i.e. well below saturation), the number of excited molecules is much
smaller than the number of molecules in the ground state. Contributions from the
excited state may therefore be weak and difficult to detect. This can be specially
complicated in (self-)heterodyne detection methods if ground and excited state contributions spectrally overlap. In methods using such kind of detection, the intensities
of different contributions add linearly, which can obscure, or even totally cancel,
weak signals. Such an effect can be observed e.g. in transient absorption (TA) when
a negative signal like stimulated emission (SE) overlaps with an excited state absorption (ESA). This contrasts to background-free third-order methods, where all
contributions will be detected with a much higher sensitivity to small variations of
the dynamics in the excited state.
The second aspect is related to the time scale of processes in the excited state.
After excitation, population and coherence in the higher electronic states are not in
equilibrium but evolve in many (bio-)chromophores often on a very fast time scale.
Such time scales range from hundreds of femtoseconds as found for the isomerization of protein-bound retinal [11] to just some tens of femtoseconds as observed for
the internal conversion between excited states of carotenoids [12]. While ultrafast
transient population evolution can in principle be resolved by using state-of-the-art
sub-10 fs pulses [13], following the evolution of the vibrational coherence is not
trivial. Not just the amplitude of oscillatory phenomena contains information on
the deactivation of the excited state, but also the frequency and the phase may offer important insight. Nevertheless, determination of frequency and phase during
an ultrafast relaxation process is demanding, which has received a lot of attention
recently. It requires accurate analysis in order to avoid phase and frequency changes
due to pure analysis artifacts [14, 15].
One way of addressing these central aspects in time-resolved spectroscopy is to
increase the dimensionality of the nonlinear optical interaction. Multidimensional
time-resolved spectroscopy employs additional ultrashort pulses in order to open a
new spectral and/or temporal observation window, which is normally hidden from
lower-order techniques [16]. For example, this allows probing a specific molecular
response without the need of complex fitting algorithms or assumptions [17, 18]. In
this chapter, we show how four wave mixing methods with three beams like coherent
anti-Stokes Raman scattering (CARS) and transient grating (TG) can be combined
with an additional initial pump beam to be selective to excited state dynamics in
the condensed phase (Fig. 9.1). The so called pump-degenerate four wave mixing
(pump-DFWM) technique [19, 20] consists of two interactions with the initial pump
pulse to produce coherences and populations in the excited state, which will be
probed by the consecutive four wave mixing sequence. In this sense, pump-DFWM
can be thought of as a “pump-probe” method, where the probe is not a single pulse,
but is a nonlinear probe sequence of three ultrashort pulses. Due to their bandwidths,