106
R. Spesyvtsev et al.
The relationship between pulse duration and spectral bandwidth for Gaussian
pulses is restricted by the inequality,
ν · t ≥ 0.441,
(5.2)
which can be rewritten as,
λ [nm] ≥ 1.471 × 10
−3 λ 2 [nm 2 ]
t [fs]
,
(5.3)
where ν is the spectral width (FWHM), λ is the bandwidth (FWHM) at wavelength λ, and t is the pulse duration (FWHM) of the Gaussian laser pulse. Equation (5.2) stipulates that it is not possible to have infinitely good spectral resolution and time resolution simultaneously. The timescales for non-adiabatic processes
range from tens of femtoseconds to hundreds of picoseconds. From Eq. (5.3), we
see that the bandwidth of a 100 fs pulse at 250 nm is around 1 nm (equivalent to
around 20 meV or 150 cm −1 ) and the bandwidth of a 0.1 fs pulse at 15 nm is around
3 nm (> 10 eV). Thus, planning a TRPES experiment requires careful consideration of the requirements for the photoionisation wavelength alongside the achievable
temporal and spectral resolution.
Another important aspect of a TRPES experiment is the delay between the pump
and probe laser pulses. This is usually controlled by changing the path length between the two pulses using optics mounted on commercial translation stages. In UVUV and UV-VUV/XUV pump-probe experiments, the pump and probe laser pulses
generally both originate from the same oscillator, which keeps the jitter well below
the temporal resolution of the translation stage (typically around 0.01 fs). However,
in pump-probe experiments using XFELs, UV pump and x-ray probe pulses do not
have the same origin and the jitter may be significant [59].
5.3.2 Molecular Sources
In order to make time-resolved photoelectron spectroscopy measurements of isolated molecules, an expansion into a vacuum chamber is usually employed ensuring
a collision-free environment. Such conditions also allow for the stringent requirements of most electron detectors which place an upper limit of around 10 −5 mbar
in the laser-molecule interaction region.
The simplest way to deliver a gas sample into the laser-molecule interaction region is to feed it through a thin needle. The gas will expand into the vacuum chamber
with the highest density close to the needle tip. Such a source is not well collimated
and only a small fraction of the effused molecules overlap with the laser focus. Such
low number densities place a practical limit on the achievable signal-to-noise ratio
in such experiments.
A more efficient way to deliver a gas sample into the laser-molecule interaction
region is with a supersonic molecular beam. The gas is usually delivered through a
R. Spesyvtsev et al.
The relationship between pulse duration and spectral bandwidth for Gaussian
pulses is restricted by the inequality,
ν · t ≥ 0.441,
(5.2)
which can be rewritten as,
λ [nm] ≥ 1.471 × 10
−3 λ 2 [nm 2 ]
t [fs]
,
(5.3)
where ν is the spectral width (FWHM), λ is the bandwidth (FWHM) at wavelength λ, and t is the pulse duration (FWHM) of the Gaussian laser pulse. Equation (5.2) stipulates that it is not possible to have infinitely good spectral resolution and time resolution simultaneously. The timescales for non-adiabatic processes
range from tens of femtoseconds to hundreds of picoseconds. From Eq. (5.3), we
see that the bandwidth of a 100 fs pulse at 250 nm is around 1 nm (equivalent to
around 20 meV or 150 cm −1 ) and the bandwidth of a 0.1 fs pulse at 15 nm is around
3 nm (> 10 eV). Thus, planning a TRPES experiment requires careful consideration of the requirements for the photoionisation wavelength alongside the achievable
temporal and spectral resolution.
Another important aspect of a TRPES experiment is the delay between the pump
and probe laser pulses. This is usually controlled by changing the path length between the two pulses using optics mounted on commercial translation stages. In UVUV and UV-VUV/XUV pump-probe experiments, the pump and probe laser pulses
generally both originate from the same oscillator, which keeps the jitter well below
the temporal resolution of the translation stage (typically around 0.01 fs). However,
in pump-probe experiments using XFELs, UV pump and x-ray probe pulses do not
have the same origin and the jitter may be significant [59].
5.3.2 Molecular Sources
In order to make time-resolved photoelectron spectroscopy measurements of isolated molecules, an expansion into a vacuum chamber is usually employed ensuring
a collision-free environment. Such conditions also allow for the stringent requirements of most electron detectors which place an upper limit of around 10 −5 mbar
in the laser-molecule interaction region.
The simplest way to deliver a gas sample into the laser-molecule interaction region is to feed it through a thin needle. The gas will expand into the vacuum chamber
with the highest density close to the needle tip. Such a source is not well collimated
and only a small fraction of the effused molecules overlap with the laser focus. Such
low number densities place a practical limit on the achievable signal-to-noise ratio
in such experiments.
A more efficient way to deliver a gas sample into the laser-molecule interaction
region is with a supersonic molecular beam. The gas is usually delivered through a