Using NMR Spectroscopy to Evaluate Metal–Ligand Bond Covalency for the f Elements

Abstract

Understanding f element-ligand covalency is at the center of efforts to design new separations schemes for spent nuclear fuel, and is therefore of signficant fundamental and practical importance. Considerable effort has been invested into quantifying covalency in f element-ligand bonding. Over the past decade, numerous studies have employed a variety of techniques to study covalency, including XANES, EPR, and optical spectroscopies, as well as X-ray crystallography. NMR spectroscopy is another widely available spectroscopic technique that is complementary to these more established methods; however, its use for measuring 4f/5f covalency is still in its infancy. This Account describes efforts in the authors’ laboratories to develop and validate multinuclear NMR spectroscopy as a tool for studying metal–ligand covalency in the actinides and selected lanthanide complexes. Thus far, we have quantified M–L covalency for a variety of ligand types, including chalcogenides, carbenes, alkyls, acetylides, amides, and nitrides, and for a variety of isotopes, including ¹³C, ¹⁵N, ⁷⁷Se, and ¹²⁵Te. Using NMR spectroscopy to probe M-C and M-N covalency is particularly attractive because of the ready availability of the¹³C and ¹⁵N isotopes (both I = 1/2), and also because these elements are found in some of the most important f element ligand classes, including alkyls, carbenes, polypyridines, amides, imidos, and nitrides.

The covalency analysis is based on the chemical shift (δ) and corresponding nuclear shielding constant (σ) of the metal-bound nucleus. The diamagnetic (σ_dia), paramagnetic (σ_para), and spin–orbit contributions (σ_SO) to σ can be obtained and analyzed by relativistic density functional theory (DFT). Of particular importance is σ_SO, which arises from the combination of spin–orbit coupling, the magnetic field, and chemical bonding. Its magnitude correlates with the amount of ligand s-character and metal nf (and (n+1)d) character in the M–L bond. In practice, Δ_SO, the total difference between calculated chemical shift for the ligand nucleus including vs excluding SO effects, provides a more convenient metric for analysis. For the examples discussed herein, Δ_SO accounts primarily for σ_SO in an f-element complex, but also includes minor SO effects on the other shielding mechanisms and (usually) minor SO effects on the reference shielding. Δ_SO can be very large, as in the case of [U(CH₂SiMe₃)₆] (348 ppm), which is not surprising as the An–C bonds in this example exhibits a high degree of covalency (e.g., 20% 5f character). However, even small values of Δ_SO can indicate profound bonding effects, as shown by our analysis of [La(C₆Cl₅)₄]⁻. In this case, Δ_SO is only 9 ppm, consistent with a highly ionic La–C bond (e.g., <1% 4f character). Nonetheless, the inclusion of SO effects in the calculation are necessary to achieve good agreement between the predicted and experimentally determined chemical shifts. Overall, the examples discussed herein highlight the exquisite sensitivity of this method to unravel electronic structure in f element complexes.