= Femtochemistry
<Femtochemistry> studies chemical change on the time scale of atomic motion, using measurements with femtosecond resolution, where $1\,\mathrm{fs}=10^{-15}\,\mathrm s$. Its central question is how the atoms move while <chemical bonds> break, form or rearrange, rather than only how much product remains after a <chemical reaction>. <Ahmed H. Zewail>’s primary Nobel lecture explains the development of these experiments and their connection to motion on <potential energy surfaces>.<Femtochemistry: Atomic-Scale Dynamics of the Chemical Bond Using Ultrafast Lasers by Ahmed H. Zewail>[Introduction and discussion of molecular motion and transition-state dynamics]{ref}
In an optical <pump-probe spectroscopy> experiment, a short pump <optical pulse> prepares an excited molecular ensemble and a delayed probe <optical pulse> measures a selected optical response. Scanning the delay constructs a time-dependent signal. The preparation and detection need not amount to a direct photograph of individual nuclei: the observed <absorption> or <fluorescence> depends on which molecular states the probe can access. Different probe wavelengths can therefore emphasize different regions of a reaction pathway.
The <sodium iodide> experiments discussed in the lecture make the distinction especially vivid. The excited molecular motion could revisit a region where ionic and covalent electronic characters interact, giving recurring structure in the probe signal rather than a simple monotonic rise of separated products. A <wave packet> on the coupled <potential energy surfaces> provides a description of that motion. The lecture’s NaI/NaBr example connects these time-dependent observations to the 1988 experiments.<Femtochemistry: Atomic-Scale Dynamics of the Chemical Bond Using Ultrafast Lasers by Ahmed H. Zewail>[The NaI Discovery, Figure 15; experiments reported in 1988]{ref} Resolving such recurrences reveals dynamics that a single fitted <reaction rate constant> would discard.
A scale calculation shows why very short <optical pulses> matter. For a hypothetical <molecular vibration> with spectroscopic wavenumber $\widetilde\nu=300\,\mathrm{cm}^{-1}$, its ordinary <frequency> is $f=c\widetilde\nu$, so
$$
T=\frac1{c\widetilde\nu}
=\frac1{(2.998\times10^{10}\,\mathrm{cm\,s^{-1}})(300\,\mathrm{cm^{-1}})}
\approx111\,\mathrm{fs}.
$$
A $1\,\mathrm{ps}$ response averages over roughly nine such periods. For Gaussian pump and probe intensities each with $30\,\mathrm{fs}$ full width at half maximum, the ideal cross-correlation width is $\sqrt{30^2+30^2}\approx42\,\mathrm{fs}$, before other broadening. This estimate concerns time resolution; it does not guarantee that every <molecular vibration> is visible in the chosen optical observable.
A convincing experiment must distinguish molecular evolution from the instrument response and effects present only while the <optical pulses> overlap. Delay calibration, the pulse durations at the sample, probe wavelength and sample conditions belong to the interpretation. An oscillating signal is useful evidence only together with a model connecting that observable to the underlying <chemical reaction>.
Back to article page
