ABSTRACT

Part I. It is important to understand the factors that influence binding. Rigid molecular receptors have been widely studied, with some of these receptors being able to form stable complexes in competitive solvents such as aqueous DMSO. The scope of my research is to study both the binding of ferrocene derivatives to carboxylates in competitive solvents, and the release of these carboxylates when cucurbit[7]uril is added to the system with the aim of identifying more tightly binding hosts to carboxylates in neat water.

In Chapter 1, pincher cationic ferrocene hosts for carboxylate ion guests were synthesized and the binding constants were determined by NMR or UV-vis titrations. These (di)cationic hosts formed tight complexes with benzoate or acetate even in competitive aqueous DMSO solvent. A bis(acylguanidinium) ferrocene dication achieved a remarkable Ka of ~ 106 M-1 to acetate in 9:1 DMSO:H2O and a Ka of 850 M-1 in pure D2O, one of the highest association constants known for a mono-carboxylate complex exploiting only electrostatic interactions in pure water. Density functional theory (DFT) computations of the binding enthalpy were in good agreement with the experimentally determined association constants.

In Chapter 2, association constants of a bis(acylguanidinium) ferrocene dication to various (di)carboxylates in water were determined through UV-vis titrations. Association constant values greater than 104 M-1 were determined for both phthalate and maleate carboxylates to the bis(acylguanidinium) ferrocene salt in pure water. DFT binding enthalpy computations of the rigid carboxylates geometrically complementary to the dication agree well with the experimentally determined association constants. Catch and release competitive binding experiments were done by NMR for the cation-carboxylate ion pair complexes with CB[7], showing dissociation of the ion pair complex upon addition of CB[7].

Part II. Heterolytic bond scission is a staple of chemical reactions. While qualitative and quantitative models exist for understanding the thermal heterolysis of carbon--leaving group (C-LG) bonds, no general models connect structure to reactivity for heterolysis in the excited state.

Time-Dependent Density Functional Theory (TD-DFT) excited-state energy calculations and Complete Active Space Self-Consistent Field (CASSCF) minimum energy crossing (conical intersection) searches were performed to investigate representative systems that undergo photoheterolysis to generate carbocations. Certain classes of unstabilized cations are found to have structurally-nearby, low-energy conical intersections, whereas stabilized cations are found to have high-energy, unfavorable conical intersections. The former systems are often favored from photochemical heterolysis. These results suggest that the frequent inversion of the substrate preferences for non-adiabatic photoheterolysis reactions arises from switching from transition-state control in thermal heterolysis reactions to conical intersection control for photochemical heterolysis reactions. The elevated ground-state surfaces resulting from generating unstabilized or destabilized cations, in conjunction with stabilized excited-state surfaces, can lead to productive conical intersections along the heterolysis reaction coordinate.

From the TD-DFT excited-state calculations, we were able to notice trends and predict if molecules have the potential for a productive conical intersection. To test this experimentally, BODIPY dyes that were shown to have small energy gaps between the ground state and excited state surfaces were synthesized. These dyes were irradiated with a xenon lamp, and the growth of the acetic acid leaving group peak was monitored by NMR over time.