Non-Directed Allylic C-H Acetoxylation in the Presence of Lewis Basic Heterocycles

Abstract

We outline a strategy to enable non-directed Pd(II)-catalyzed C-H functionalization in the presence of Lewis basic heterocycles. In a high-throughput screen of two Pd-catalyzed C-H acetoxylation reactions, addition of a variety of N-containing heterocycles is found to cause low product conversion. A pyridine-containing test substrate is selected as representative of heterocyclic scaffolds that are hypothesized to cause catalyst arrest. We pursue two approaches in parallel that allow product conversion in this representative system: Lewis acids are found to be effective in situ blocking groups for the Lewis basic site, and a pre-formed pyridine N-oxide is shown to enable high yield of allylic C-H acetoxylation. Computational studies with density functional theory (M06) of binding affinities of selected heterocycles to Pd(OAc)₂ provide an inverse correlation of the computed heterocycle-Pd(OAc)₂ binding affinities with the experimental conversions to products. Additionally, ¹H NMR binding studies provide experimental support for theoretical calculations.

Figure 1

Figure 2

Figure 3

Figure 4

Reaction conditions A: Pd(OAc)₂/PhS(O)C₂H₄S(O)Ph (10 mol %), p-benzoquinone (2 equiv), AcOH (10 equiv), 1,4-dioxane (0.3 M), 45 °C, 48 h; B: Pd(OAc)₂ (10 mol %), 4,5-diazafluorenone (10 mol %), p-benzoquinone (2 equiv), NaOAc (40 mol %), AcOH (16 equiv), 1,4-dioxane (0.3 M), 60 °C, 48 h. Compiled results of the high-throughput heterocycle screen. Heterocycles are separated into three tiers for both reaction A and B; the colors represent the following for each reaction: green (50% to 100% product conversion), yellow (16% to 49% conversion), and red (≤15% product conversion). Conversion is defined as % product formation by HPLC relative to remaining starting material.

Figure 5

Validation of the potential predictive qualities of the high-throughput heterocycle screen employed (see Figure 4). Five representative heterocycles from the high-throughput screen were incorporated within substrates that were subjected to Pd(II)-catalyzed allylic C–H acetoxylation. Reaction conditions A: Pd(OAc)₂/PhS(O)C₂H₄S(O)Ph (10 mol %), p-benzoquinone (2 equiv), AcOH (10 equiv), 1,4-dioxane (0.3 M), 45 °C, 48 h; B: Pd(OAc)₂ (10 mol %), 4,5-diazafluorenone (10 mol %), p-benzoquinone (2 equiv), NaOAc (40 mol %), AcOH (16 equiv), 1,4-dioxane (0.3 M), 60 °C, 48 h. Isolated yields reported.

Figure 6

Theoretical binding affinities of heterocycles to Pd(OAc)₂ vs. product conversion (%) observed in reaction A (a) and reaction B (b); Theoretical binding affinities of heterocycles to Pd(OAc)₂ to form a 2:1 complex vs. product conversion (%) observed in reaction A (c) and reaction B (d). Binding affinity is defined as −ΔH for the reaction L + Pd(OAc)₂ → LPd(OAc)₂ (a and b) or 2 L + Pd(OAc)₂ → L₂Pd(OAc)₂ (c and d). The horizontal red line indicates the binding affinity of propene. Geometries were optimized at the B3LYP/SDD–6–31G(d,p) level with single-point energies computed at the M06/SDD(f)–6–311++G(2d,p) level with SMD solvation model (1,4-dioxane); see the Supplementary Information for data with acetic acid. Product conversions and corresponding reaction conditions A and B are based on the high-throughput heterocycle screen (Figure 4). Each data point represents a specific heterocycle fragment from Figure 4.

Figure 7

¹H NMR binding values of heterocycles to Pd(OAc)₂ vs. product conversion (%) observed in reaction A (a) and reaction B (b). Binding values were determined by quantitating disappearance of parent ligand by ¹H NMR with 1:1 ligand/Pd(OAc)₂. Product conversions and corresponding reaction conditions A and B are based on the high-throughput heterocycle screen (Figure 4). Each data point represents a specific heterocycle fragment from Figure 4.