Defense Date


Graduation Date

Summer 8-13-2022


One-year Embargo

Submission Type


Degree Name



Chemistry and Biochemistry

Committee Chair

Jeffrey D. Evanseck

Committee Member

Thomas D. Montgomery

Committee Member

Bruce D. Beaver

Committee Member

Jeffrey J. Rohde


computational chemistry, organic chemistry, tautomerism, acetone, water clusters, solvent models, Grotthuss mechanism, DFT, MP2, hydrogen-bonding, Pople diagram


Keto-enol tautomerization (KET) is a fundamental process impacting a range of molecular phenomena in organic and biochemistry. However, the accurate computation of solution-phase KET energies remains a challenge, even for prototypical acetone.

In Part I, keto-enol tautomers of acetone were incorporated into solvent clusters that interact via uninterrupted, cyclic hydrogen-bonding (UCHB) networks. An empirical model was created to predict accurate KET energies, Etaut, of simple carbonyl compounds. Based on the availability of experimental data and structural simplicity, acetone was selected as a prototype. A discrete-continuum strategy was employed – accounting simultaneously for local noncovalent interactions and bulk-phase effects – wherein acetone, bound to water clusters of size (H2O)n n = 1–4, was paired with implicit solvent, represented by the polarized continuum model (PCM). Geometry optimizations and harmonic vibrational frequency calculations were completed using the B3LYP, ωΒ97Χ-D, and M06-2X hybrid density functionals paired with Dunning’s correlation-consistent basis sets cc-pV[D,T]Z with maug-, jul-, and aug- diffuse functions. Results were compared to second-order Møller–Plesset perturbation theory (MP2) and the experimental value of Etaut = -11.36 ± 0.04 kcal/mol determined by spectrophotometric bromination. In the context of the dissertation, chemical accuracy is defined to be 1.0 kcal/mol. When using the pseudo-aromatic solvent model, MP2 predicted Etaut within chemical accuracy using the maug-cc-PVTZ basis set, and all density functionals with maug-cc-pVTZ achieved errors below 1.50 kcal/mol for Etaut.

In Part II, the solvent model of tautomerization was extended to transition structure (TS) complexes of acetone and explicit water. The same density functionals and MP2 with the same basis sets were employed. Utilizing the concept of a Grotthuss mechanism to mediate proton transfer, models with a single Grotthuss chain provided reductions in activation energies from 60 kcal/mol to less than 40 kcal/mol while UCHB solvated models provided further reductions to less than 35 kcal/mol. Both the addition of Grotthuss chains and incorporation into UCHB solvent networks were necessary to achieve stabilization. In efforts to model acid-catalyzed tautomerism, an excess proton was added to the UCHB structures at three separate sites on acetone. Inclusion of acetone into a solvent cluster allowed for delocalization of the proton defect along the UCHB network. At the B3LYP/jul-cc-pVDZ level of theory, an activation energy of 21.8 kcal/mol was predicted, accounting for 92% of the experimental value of 23.6 kcal/mol, reported by Kresge and coworkers. Nonetheless, the influence of a cyclic solvent network was essential in allowing redistribution of the proton defect.

Incorporation of the carbonyl group on acetone into a solvent network both stabilized the keto and enol forms to achieve chemical accuracy and stabilized the TS of protonated Grotthuss chains to improve agreement with experimental values. Results indicate the necessity of UCHB networks for calculating KET energies and the emergence of a novel mechanism for acid-catalyzed enolization of ketones, namely Grotthuss tautomerism. My dissertation demonstrates the necessity of extending Pople’s diagram of computational chemistry by adding the physical model as a third dimension. In doing so, the new paradigm for computational chemistry forges agreement between computation and experiment and sets the foundation to expand my work to complex organic functionality and ultimately to a generalized theory of aqueous solvation.