Extracting Information from Adaptive Control Experiments

In the study of photochemical reactions laser pulses serve as both reagent and measuring device. The role of measuring device has been exemplified by ultrafast spectroscopies that exploit broadband optical pulses to probe molecular dynamics with femtosecond time resolution. More sophisticated multi-dimensional spectroscopies are using optical pulse-shaping to take measurement of molecular level quantum coherence to a new level of sensitivity and selectivity. The role for laser pulses as reagents is less explored, but has the potential to address one of the fundamental goals of chemistry: control of chemical reactivity. While in most spectroscopic studies the role of the incident “pump” pulse is limited to that of a tunable energy source, the field of optical coherent control focuses on designing optical fields that will drive the light–matter interaction toward a particular chemical outcome. Owing to over a decade of technical advances in the development of broadband lasers and optical manipulation, it is now possible to create complex pulses in the laboratory with commercial pulse shapers. A typical scheme using a spatial light modulator or acousto-optic modulator to manipulate broadband optical pulses is illustrated in Figure 1. These shaped pulses may be used as smart reagents, or smart catalysts, to control molecular dynamics and chemical reactivity by producing specific quantum states or by steering a molecule through a series of light–matter interactions. The majority of optical control experiments have used shaped visible or near-IR pulses. The capability for optical pulse shaping has recently been extended to the ultraviolet region of the spectrum using an acousto-optic modulator in a 4-f arrangement (Figure 1), an acoustooptic programmable dispersive filter (AOPDF, Dazzler, Fastlite), or a microelectromechanical system (MEMS) array. The ability to manipulate and characterize ultraviolet pulses greatly extends the range of chemical systems available for spectroscopic investigation and control using shaped pulses. Many important small molecule systems absorb between 200 and 400 nm, and may now be studied, including both cyclohexadiene ringopening reactions and cis-stilbene isomerization reactions. Identifying shaped optical pulses that will act as chemical reagents is a challenge. The space of potential pulse shapes — potential reagents — is enormous, and the identification of optimally effective reagents or catalysts Abstract : Optical control of chemical reactivity is achieved through the use of photonic reagents, that is, “shaped” ultrafast optical pulses created using a pulse shaper. It has been demonstrated in a number of molecular systems that these pulses can effectively guide the system into a desired final state. Effective pulses are often found through an experimental search involving thousands of individual measurements. An examination of the pulses tested in these experiments can reveal the pulse features responsible for control and also the underlying molecular dynamics. In this article we review attempts to extract information from optical control experiments using adaptive learning algorithms to search the available parameter space, and we discuss how these kinds of experiments can be used to achieve and understand multiphoton optical control.