As identical fermions in identical nuclear spin
states, these molecules were forbidden by quantum mechanical symmetries from scattering via
l = 0 partial waves and so scattered according to
angular momentum l = 1. However, on changing
the nuclear spin for half the molecules, this symmetry no longer held, and the molecules could
again collide head-on with l = 0. Because there
was no centrifugal barrier associated with this
channel, the reaction rate was enhanced by almost two orders of magnitude. This was therefore the first instance of choosing the partial
wave degree of freedom in relative motion, thus
completing the task of controlling the reactants
at all possible degrees of freedom of the collision.
Further control of the molecules could be implemented by engineering the spatial configuration of the trap to restrict molecular motions in
low spatial dimensions (41) and by introducing
external fields to control the molecular orientation and dipolar moment in the laboratory frame
(42). These experiments pave the way for manipulating polar molecules in an optical lattice to
form synthetic quantum materials, as discussed
in the second part of this Review.
Having prepared the reactants, it is necessary to
observe and perhaps even control the products of
reaction. This development, too, has a long history
in beam experiments. The principal ambition of
product investigation is to ascertain which chemical species emerge, in which internal states, and
moving in which direction. The first two factors are
often accessible by means of standard spectroscopic
techniques, such as laser-induced fluorescence.
To measure the velocities of outgoing products,
techniques such as velocity map imaging (VMI)
are used (43). An outgoing molecule is first selectively ionized with little perturbation of its
momentum. An electric field then guides the ion
to a multichannel plate, which records the location of its arrival. The mass and charge of the ion
and the shape of the electric field connect the
original velocity of the molecule to the impact
location on the plate.
To capitalize on the velocity distribution and
extract the angular distribution of products
requires that the velocities be referred to a well-defined incident axis of the collision. VMI techniques are thus most useful when the incident
beams are as well controlled as possible, using
techniques described above. This type of control
was achieved, at least for inelastic scattering,
by the van de Meerakker group, who crossed a
pulsed beam with a Stark decelerator. This experiment succeeded in measuring narrow diffraction oscillations in the differential cross sections
of NO scattering from rare-gas atoms (44).
In the coldest of samples, residing in traps,
measurements of differential cross sections are
problematic precisely because there is no well-defined initial direction for collisions, given that
the molecules are traveling in many different directions. Still, identifying the products and their distribution into internal states can prove valuable.
Many trapped-molecule experiments are not well
suited to measuring products, at least at present.
Part of the problem is that the products of reaction
can emerge with large kinetic energies, on the order
of electron volts (corresponding to many thousands of kelvins) and therefore are not themselves
trapped for further study. At least two types of
experiments can circumvent this issue, however.
The first of these was recently demonstrated
by the Weinstein group, studying the reaction
Li + CaH → LiH + Ca (45). This experiment was
conducted not in a trap but in a helium buffer
gas cell at a few kelvins, where both the reactants
and products were stored. Despite releasing 0.9 eV
of energy in the reaction, the products quickly
cooled to the ambient temperature and were
available for interrogation. Rotational levels cooled
quickly, so their nascent distribution after re-
action was undetermined. However, vibrational
relaxation is notoriously slow in a buffer gas,
meaning that vibrational populations produced
from the reaction were still available for viewing.
A second option for harnessing the products is
provided by ion traps. Owing to the comparative
strength of electromagnetic field confinement,
molecular ions can be trapped even after they
react, if the product is also electrically charged.
Ions in such a trap arrange themselves into a
“Coulomb crystal,” where they can be individually
observed by fluorescence. When a slowed reactant
beam flows through the crystal, reactions can
be observed one by one as the ions vanish from
the crystal, providing the ultimate number resolution of a chemical reaction.
If the products happen also to be ionic, they
can be sympathetically cooled by the original ions
and join the Coulomb crystal themselves. This
has occurred, for example, in the experiments of
the Softley group (46), where certain calcium ions
would disappear from the crystal during the reaction Ca+ + CH3F → CaF+ + CH3. The CaF+ ions
were also trapped and made their presence known
by perturbing the Ca+ crystal, even though CaF+
was not directly observed. Knowing it is there,
though, makes it ripe for exploration.
Careful preparation of reactants and observation
of products often yield no progress toward the
even greater challenge of observing or manipulating intermediates during the reaction. Generally, following the reaction in transition is the
job of theory. However, the reaction itself contains
useful handles by which it can be manipulated.
These handles appear in the form of resonant
states or transition states of the collision complex. That is to say, if reactants A and B (each of
which may consist of many atoms) collide, they
may form various configurations of all the atoms
in the AB cluster, before finding their way to the
products of reaction. This situation can lead to a
resonant enhancement of the cross section at the
energy of the AB cluster.
One such set of cluster states constitutes the
van der Waals states, where clusters are held together by the weakly attractive van der Waals
interaction between molecules but prevented
initially from reacting by an energetic barrier
(47). More intricate resonant states, termed transition state resonances, may exist behind the barrier.
Observing and characterizing these resonances
can shed light on the way in which the atoms
share energy and pass it between them, giving
hints about the chemical process. Cold, controlled
collisions of reactants may allow the energy resolution necessary to observe these resonances.
Moreover, if the resonant states happen to have
magnetic or electric moments different from
those of the reactants, then it is possible to apply
a corresponding field to scan one or more resonances through the collision energy. This is a
process long known and extremely useful in
Fig. 3. Experimental progress since 2008
toward production of a quantum gas of
bialkali molecules. G R