Learning from molecules: Coherent feedback control in liquid phase.

Abstract

Coherent control of light and matter is a major tool in current atomic physics and physical chemistry research. Technological advances in ultrafast laser pulse shaping allow control over the dynamics of molecular systems on sub-picosecond timescales. This thesis demonstrates feedback control over a variety of quantum systems and describes an adaptive genetic algorithm used in the experiments. In the gas phase, we examine control of multi-photon, dissociative ionization in diatomic sodium. The learning algorithm achieves control over the dissociative fraction primarily by adjusting the peak intensity of the laser field. We also investigate control of multi-mode stimulated Raman scattering in liquid phase methanol. Analysis of the results leads to identification of quasi-impulsive Raman coupling as a possible control mechanism. Further experiments, combined with numerical simulations of the interaction, demonstrate that an impulsive coupling of the two modes by the pump pulse can control the Raman gain. We extend the Raman control experiments to ethanol and isopropanol. Finally we apply feedback control techniques in an initial investigation of a liquid-phase chemical reaction. Specifically, we examine the ring-opening conversion of cyclohexadiene into hexatriene. The reaction dynamics occur on an excited state potential energy surface. Through control of the excitation pulse shape, we seek to guide the wavepacket evolution on the excited state to exert control over the reaction branching ratio.