2
A. Rouzée et al.
periments have become possible, which operate either in the frequency or in the
time domain. The latter type of experiment has been particularly informative. Using pump-probe approaches, where a first “pump” laser pulse triggers a structural
change in a molecule, and a second “probe” laser pulse interrogates the molecule after it has evolved for some time, detailed questions can be asked that are pertinent to
chemical reactivity. The importance of this new research field of “femtochemistry”
was recognized by the Nobel Prize in Chemistry that was awarded in 1999 to Prof.
Ahmed Zewail (Caltech) [1].
In femtochemistry experiments, information about an evolving molecular structure is typically inferred by measuring how the molecular absorption spectrum (or
a related quantity that can be measured, such as a photoelectron or Raman spectrum) changes as a function of pump-probe delay. If it is known how the molecular
absorption spectrum depends on the molecular structure, then measuring its timedependent changes in a pump-probe sequence can inform us about time-dependent
structural changes that occur in the molecule. It follows however, that femtochemistry experiments become very challenging when wavelength-dependent spectral
features are not very pronounced, or if the relation between the spectrum and the
structure is not known ahead of time. Correspondingly, the level of detail that can
be extracted from femtochemistry experiments is reduced when the complexity of
the molecule increases.
In the last few years a number of new ideas (summarized in Fig. 1.1) have been
put forward that aim to remove the above-mentioned limitations of present-day femtochemistry experiments. The common denominator in all these ideas is that they
base themselves on diffraction rather than absorption, so that the requirements on
pre-existing knowledge of the electronic spectroscopy of the molecule under investigation are significantly relaxed. In a diffraction experiment structural information is encoded in interference patterns that result from the way that an electron
or light wave scatters. In the case of light diffraction (see Fig. 1.1a), the required
wavelength to resolve interatomic distances is in the X-ray regime. Time-resolved
X-ray diffraction was first developed at X-ray synchrotrons, making use of the intrinsic X-ray pulse duration of about 100 ps at typical facilities [2], and was significantly improved by the implementation of slicing facilities where time resolution
into the femtosecond regime was accomplished, at the expense of a very significant
reduction in the available X-ray fluence [3]. Alternatively, laser-plasma based X-ray
sources have been developed that allow performing X-ray diffraction experiments
with a time resolution around 100 fs [4]. Finally, time-resolved X-ray diffraction is
one of the main driving forces behind the development of X-ray free electron lasers
(FELs) like the LCLS at Stanford (which became operational in the fall of 2009
[5]), the SACLA X-ray FEL in Japan and the future European X-ray Free Electron
Laser (XFEL) that is under construction in Hamburg. At LCLS, several remarkable
results illustrating the potential of coherent diffractive imaging using X-ray FELs
have already been achieved [6].
As an alternative to X-ray diffraction, the diffraction of fast electrons can be
used. In doing so, an important advantage is the fact that in order for electron
wavelengths to match interatomic distances significantly lower electron kinetic en-
A. Rouzée et al.
periments have become possible, which operate either in the frequency or in the
time domain. The latter type of experiment has been particularly informative. Using pump-probe approaches, where a first “pump” laser pulse triggers a structural
change in a molecule, and a second “probe” laser pulse interrogates the molecule after it has evolved for some time, detailed questions can be asked that are pertinent to
chemical reactivity. The importance of this new research field of “femtochemistry”
was recognized by the Nobel Prize in Chemistry that was awarded in 1999 to Prof.
Ahmed Zewail (Caltech) [1].
In femtochemistry experiments, information about an evolving molecular structure is typically inferred by measuring how the molecular absorption spectrum (or
a related quantity that can be measured, such as a photoelectron or Raman spectrum) changes as a function of pump-probe delay. If it is known how the molecular
absorption spectrum depends on the molecular structure, then measuring its timedependent changes in a pump-probe sequence can inform us about time-dependent
structural changes that occur in the molecule. It follows however, that femtochemistry experiments become very challenging when wavelength-dependent spectral
features are not very pronounced, or if the relation between the spectrum and the
structure is not known ahead of time. Correspondingly, the level of detail that can
be extracted from femtochemistry experiments is reduced when the complexity of
the molecule increases.
In the last few years a number of new ideas (summarized in Fig. 1.1) have been
put forward that aim to remove the above-mentioned limitations of present-day femtochemistry experiments. The common denominator in all these ideas is that they
base themselves on diffraction rather than absorption, so that the requirements on
pre-existing knowledge of the electronic spectroscopy of the molecule under investigation are significantly relaxed. In a diffraction experiment structural information is encoded in interference patterns that result from the way that an electron
or light wave scatters. In the case of light diffraction (see Fig. 1.1a), the required
wavelength to resolve interatomic distances is in the X-ray regime. Time-resolved
X-ray diffraction was first developed at X-ray synchrotrons, making use of the intrinsic X-ray pulse duration of about 100 ps at typical facilities [2], and was significantly improved by the implementation of slicing facilities where time resolution
into the femtosecond regime was accomplished, at the expense of a very significant
reduction in the available X-ray fluence [3]. Alternatively, laser-plasma based X-ray
sources have been developed that allow performing X-ray diffraction experiments
with a time resolution around 100 fs [4]. Finally, time-resolved X-ray diffraction is
one of the main driving forces behind the development of X-ray free electron lasers
(FELs) like the LCLS at Stanford (which became operational in the fall of 2009
[5]), the SACLA X-ray FEL in Japan and the future European X-ray Free Electron
Laser (XFEL) that is under construction in Hamburg. At LCLS, several remarkable
results illustrating the potential of coherent diffractive imaging using X-ray FELs
have already been achieved [6].
As an alternative to X-ray diffraction, the diffraction of fast electrons can be
used. In doing so, an important advantage is the fact that in order for electron
wavelengths to match interatomic distances significantly lower electron kinetic en-