Novel Applications of Dynamic NMR in Organic Chemistry

ERKKI KOLEHMAINEN    Department of Chemistry, University of Jyväskylä, Finland

1 INTRODUCTION

The available literature dealing with dynamic phenomena studied by various NMR techniques is enormous and too large to be covered in one review article. Even the subject itself “Dynamic NMR” is somewhat flexible and dependent on the phenomena studied, measuring techniques and nuclei involved. From 1985 when M. Oki's textbook “Applications of Dynamic NMR Spectroscopy to Organic Chemistry”1 was published the variety of experimental NMR and calculational methods has enlarged remarkably. However, as it has been argued by Keith G. Orrell in his recent review on Dynamic NMR in Inorganic and Organometallic Chemistry2 the basic one-dimensional band-shape analysis can be used over slow, intermediate and fast exchange regimes but it is most suited to intermediate rates (10–103/s). Exchange rates (typically 0.1–10/s) are most accurately measured by magnetization transfer experiments and especially by T1-relaxation times while fast dynamic processes (rates >103/s) can be obtained by spin–spin relaxation times, T2, or via T1-based estimates for molecular correlation times. In fact, all NMR data have always been more or less an average of time or frequency domains. Only the time scale of the dynamic processes involved determines whether we observe only one time averaged signal or two or more signals originating from different contributors participating in various dynamic equilibria. Traditionally it has been thought that a characteristic limiting factor in applying NMR to dynamic studies is its slow time scale when compared with some other techniques such as pulsed laser spectroscopy where the time resolution can be on the femtosecond (10−15 s) time scale. In order to adjust the time scale of a dynamic process suitable for NMR, variable temperature runs, solvent changes or some additives have been frequently used. All these techniques tend to slow down the dynamic processes and freeze some or all of their contributors to allow them to be observed separately. When we know the NMR parameters joint with the individual contributors it is often straightforward to calculate the statistical average of all signals which one can observe in the fast exchange regime. A more comprehensive approach can be obtained using various band shape simulators3,4 developed for a two-site exchanging system, which can produce the observed spectra in slow, intermediate and fast exchange regimes. There exist also more sophisticated calculational procedures for multi-spin systems and random two-site exchange processes as reviewed by Orrell2 but the detailed descriptions of those methods fall beyond the scope of this review. Similarly, dynamics in biological macromolecules, organometallics and inorganics as well as solid state NMR studies are not considered in this review.

Novel theoretical methods such as density functional theory (DFT)/gauge included atomic orbitals (GIAO) procedure which can reproduce experimental NMR chemical shifts in reasonable accuracy are, however, changing the role of NMR in dynamic studies remarkably. Now, it is possible to estimate by reasonable computational time and resources the NMR parameters of all, or at least of the major, contributors involved in dynamic processes. As a very first example can be mentioned DFT/GIAO studies of the 1H, 13C and 15N NMR chemical shifts in aminopyrimidines and aminobenzenes and their relationships to electron densities and amine group orientations.5 It has been shown that in aniline the ring atom chemical shifts and 2pz electron densities at ortho and para (but not meta) positions are quite sensitive to the orientations of the amine groups which are pyramidalized as the result of the balance between delocalization with the ring and the use of strongly directed sp3-orbitals at the nitrogen. The calculated results show that the barriers to amine group torsional and inversion motions are low, but averaging the chemical shifts over these appears to be relatively unimportant.

In addition to the above mentioned review by Orrell and the books catalogued therein, there exist excellent new review articles such as “Dynamic NMR Studies of Supramolecular Complexes” by Pons and Millet,6 “Recent Advances in Studying Tautomerism in Solution and in the Solid State” by Kleinpeter7 and “15N NMR Spectroscopy in Structural Analysis” by Marek and Lyčka.8

In the review on supramolecular complexes with 231 references by Pons and Millet6 the investigations of the subject structures by 1D dynamic and 2D EXSY techniques are reported. The dynamic aspects of many supramolecular oligomers such as melamine-cyanuric acid rosettes, self-assembling glycoluril-based capsules, noncovalent calixarene-based capsules, cavitands, carcerands, hemicarcerands, helicates, catenanes and rotaxanes are discussed. The dynamic features associated with the binding of small guests (molecules or ions) to large hosts are also treated.

In Kleinpeter's review7 on tautomerism with 88 references the various NMR methods are classified with respect to the rate constants of the present tautomeric equilibria. Detailed discussions on temperature dependencies of chemical shifts, heteronuclear coupling constants, NOEs, deuterium-induced 13C and 15N chemical shifts, and temperature dependent line shape variations due to tautomeric equilibria are included. For solid state studies, both the CP-MAS 13C NMR spectroscopy and X-ray diffraction are considered. The review is completed by a critical survey on semiempirical and ab initio MO calculations relating to tautomerism; the calculation of chemical shifts and their application to estimate tautomeric equilibria is included as well.

In their very recent review from 2002 with 198 references Marek and Lyčka8 discuss a variety of 15N NMR studies on tautomerism such as azo–hydrazone equilibria, prototropic tautomerism with ring opening of nitrogen heterocycles, tautomerism of nitrazoles, natural chlorins, porphyrins, 2-phenacylpyridines and 1,8-bis(dimethylamino)naphtalenes. Further, they have included studies on conformational changes and rotational barriers for benzenediazonium salts with aliphatic nitro compounds, substituted triazines and N-acyl-N-alkyl-substituted amino acids.

2 NOVEL ROLE OF DYNAMIC NMR

As mentioned above novel calculational capabilities have changed the role of dynamic NMR studies because many dynamic processes can be simulated reliably with PC installed programs. One should remember that experimental dynamic NMR studies themselves (especially at natural abundance of 13C and 15N nuclei) often are spectrometer time invasive requiring measurements under different conditions (temperature, solvent etc.). In addition, owing to the complexity of the line shape equation derived even for the simplest nonbiased AB-system, the interpretations of these data most often are based on heavy approximations. Although the sources of errors joined with these approximations has been discussed by Harald Günther in his well-known textbook9 an organic chemist often is unsure of the validity limits of the method which he or she has used. This is true even in case of the most frequently used Eyring equation although it is recently familiarized for organic chemists by Goodman et al.10 In spite of the fact that the calculations of NMR parameters at different levels of theory have been for a long time as one focus of research by NMR spectroscopists the reliability of the NMR chemical shift calculations (especially 13C) using DFT/GIAO procedure has gained a special attention very recently as reported in 1999 by Smith et al.11–13

Although dynamic NMR studies cover almost all branches of chemistry and a huge variety of dynamic processes, in this review the main audience is thought to be organic chemists and only recent applications in prototropic tautomerism and conformational equilibria are included. In addition, especially novel calculational methods as an aid in the interpretation of dynamic NMR data are considered.

3 PROTOTROPIC TAUTOMERISM

Prototropic tautomerism such as its azo–hydrazone, imino–amino and keto–enol variants are very typical dynamic processes met...