# London Dispersion versus Intramolecular Hydrogen Bond in Bis‐Pyridines: How Accurate Is DFT for Competing Noncovalent Interactions in the Condensed Phase?

## TL;DR

This study compares how well DFT calculations predict noncovalent interactions in protonated bis-pyridines across different phases.

## Contribution

The paper introduces a systematic test system to evaluate DFT accuracy in predicting competing noncovalent interactions in condensed phases.

## Key findings

- DFT calculations fail to accurately predict relative energies of conformers with competing noncovalent interactions.
- 1H NMR chemical shifts correlate with crystallographic metrics, providing a solution-phase structural readout.
- Dispersion-corrected DFT shows systematic discrepancies in treating noncovalent interactions across phases.

## Abstract

We report a systematic investigation of noncovalent interactions—particularly an intramolecular hydrogen bond and London Dispersion forces—in singly protonated bis‐pyridines, studied across solution and crystalline states. Building on our previous gas‐phase study, we combine variable‐temperature 1H NMR spectroscopy, single‐crystal X‐ray diffraction, and density functional theory (DFT) calculations. The measured 1H NMR chemical shifts of the acidic proton serve as a solution‐phase structural readout, which we correlate with an independent crystallographic metric. By systematically varying the linker (–CH2–, –O–, and –CH2CH2–) and the pendant substituents (H, methyl, tert‐butyl), we examine how increasingly bulky “dispersion energy donors” affect both the intramolecular hydrogen bond and the accessible conformational states. In reference systems, where a single noncovalent interaction governs the geometry, even relatively simple computational models correctly reproduce the experimentally observed structures. However, for molecules featuring two competing noncovalent interactions, the tested, dispersion‐corrected, DFT often fails to predict the relative energies of accessible conformers accurately, highlighting current limitations in predictive accuracy. We briefly discuss broader implications of currently achievable predictive accuracy for homogeneous catalysis.

A test system of 14 singly protonated bis‑pyridine salts, with varying H‑bonded and London dispersion interactions, was designed to systematically study non‑covalent interactions across the gas, solution, and solid phases. A good correlation between the 1H chemical shift and the N‐H–N angle from the crystal structures was established. We observed systematic discrepancies in the treatment of these interactions by DFT‑D3/D4 across the phases.