FROM FULL- TO MICRO-SCALE REACTIONS: UNIFYING SYNTHETIC METHOD DEVELOPMENT, MECHANISTIC STUDIES, AND DATA REPRODUCIBILITY AT THE BENCH AND IN SILICO
Loading...
Degree type
Doctor of Philosophy (PhD)
Graduate group
Chemistry
Discipline
Chemistry
Subject
chemoinformatics
density functional theory
high-throughput chemistry
total synthesis
density functional theory
high-throughput chemistry
total synthesis
Funder
Grant number
License
Copyright date
01/01/2024
Distributor
Related resources
Author
Orzolek, Brandon, Joseph
Contributor
Abstract
- Selective oxidative homo- and cross-coupling of phenolic substrates represents a direct means to access natural products, chiral ligands, biologically active compounds, or other structures of interest. Dependent upon the specific substrate, a variety of outcomes have been observed including C–C coupling (ortho-ortho, ortho-para, para-para), C–O coupling (ortho–O, para–O), mixed C–C/C–O coupling (Pummerer ketone), the formation of either ortho– or para–quinones, as well as the formation of higher order poly(phenolic) species. While a considerable amount of attention has been given to the construction of phenol–phenol C–C, and to a lesser extent, phenol–phenol C–O bonds, little exploration has occurred in the area of mixed C–C/C–O couplings. The synthetic utility of the latter cannot be understated as this class of coupling has the potential to give rise to chemically complex scaffolds in a one-pot environment. Presented are efforts towards the synthesis of (±)-usnic acid, a natural product with potential to arise from such a mixed C–C/C–O phenolic coupling. Also disclosed is the serendipitous discovery of a photocatalytic synthesis of para-peroxyquinones and their corresponding para-quinols in excellent yields. Natural products (±)-stemenone B and (±)-parvistilbine B are synthesized with the newly developed conditions. 2. Protonation of C–M bonds and its microscopic reverse, metalation of C–M bonds, are fundamental steps in a variety of metal-catalyzed processes. As such, studies on protonation of C–M bonds can shed light on C–H activation. Presented is a computational investigation into the mechanism of protodemetalation (PDM) of the model (tbubpy)NiII(o-tolyl)(Cl) complex with five different acids that provides evidence for a concerted, cyclic transition state for the PDM of C–Ni bonds and demonstrates that five-, six-, and seven-membered transition states are particularly favorable. 3. The site-selective functionalization of multiply halogenated (hetero)arenes has been a useful tool in constructing structurally diverse backbones for use in total syntheses and pharmaceutical development. Recent efforts have been focused on the use of identically halogenated substrates due to their increased availability and lower cost compared to their non-identically halogenated counterparts. Without differential reactivities present in the latter (e.g. preferential oxidative addition of C–I bonds over C–Br bonds), reaction conditions must be chosen such that the steric and electronic features of the substrate or the directing capabilities of nearby functional groups afford the desired regioisomer. Herein is described our efforts in generating a large set of experimental data for Pd-catalyzed couplings across a diverse set of ligand classes via high-throughput experimentation. With these data, we are working on the creation of a DFT-based Quantitative Structure Activity Relationship (GQSAR) to pinpoint locations across the ligand surface that control the oxidative addition selectivity and thereby give rise to observed regioselectivity. It is our hope to build a model that can then later predict the regiochemical outcomes of non-tested ligands and that can be used for substrates where selectivity has not yet been achieved. 4. The storage and mining of information regarding chemical transformations for use in machine learning or other higher-order analyses continues to drive efforts in both academic and industrial settings. While repositories of such information (e.g., the Open Reaction Database) draw from a wide variety of sources, innovation in high-throughput screening, and, consequently, more rapid access to large swaths of chemical information will be key to such efforts. The reproducibility of HTE experiments hinge on many of the same concerns as standard bench experiments conducted one-by-one. Namely, reagent quality and incorrect protocols (e.g., forgetting to add a component) will lead to results that do not represent the underlying chemistry. Even though such errors might occur at the same frequency in HTE, the larger number of resultant data points may skew models generated therefrom. In addition, there are a number of potential errors unique to the HTE workflow that could contribute additional errors relative to reactions run that the bench. Thus, we endeavored to assess reproducibility of high-throughput experimentation data collected across three unique institutions with varying levels of automation with a focus on reactions considered robust at the HTE scale. Herein, we describe our efforts towards a unified approach for assessing reproducibility in high-throughput screening data with both supervised and unsupervised statistical learning.
Advisor
Kozlowski, Marisa, C
Schelter, Eric, J
Schelter, Eric, J
Date of degree
2024