New perspectives from organometallic chemistry

Electronic structure and bonding in f-element complexes

The electronic structure in discrete molecular systems containing actinide metals is dominated by states made up from the seven 5f orbitals on the actinide ion. These 5f orbitals possess a host of interesting properties. From a chemical and electronic structure perspective, one of the more interesting aspects of the 5f orbitals is their radial extension from the nucleus. This is a very important consideration in the formation of chemical bonds. Only those orbitals that extend to the periphery of the atom are actively engaged in direct chemical bonding.

Because of this increased spatial extension of the 5f orbitals, electrons in these orbitals are much more strongly influenced by the presence of neighboring atoms in the actinide-containing molecules than, for example, the 4f electrons in the molecules of the lanthanide series metals. Thus, the measurable properties of the electrons in these 5f orbitals will provide direct information on the other atoms, ions, and/or molecules in the immediate coordination environment of the actinide metal. In fact, the spatial extension of the 5f orbitals provides an opportunity for them to overlap directly with the orbitals on these neighboring atoms. This type of orbital overlap, referred to as a covalent bonding interaction, has been the subject of debate in actinide chemistry for decades.

To help us achieve our objectives in understanding electronic structure and bonding in the actinides, we are creating and investigating an entirely new class of actinide molecules derived from organometallic chemistry. Here we show four different types of structures that we have developed and are investigating. All of these are based on the (C₅Me₅)₂U core (represented by the two pentagons connected to the U metal center). The simplest molecules are typified by (4), which contains two methyl (CH₃) groups bound to the metal. The more interesting systems contain ligands that can bind to the metal in unusual ways—like the ketimide ligands (7) that possess a metal-nitrogen bond of ~1.5, the hydrazonato ligands (8) that each have two nitrogen atoms bound to the metal, and the imido ligands (9) that have a bona fide double bond between the metal and the nitrogen atom on the ligand.

Much of our understanding of the relationship between electronic structure and chemical bonding in actinide molecules comes from investigations of classical coordination compounds such as the halide salts (e.g., UCl₄). These studies have been invaluable and have laid the foundation for interpretation of spectral data. However, to further our understanding of the more subtle aspects of the electronic/molecular structure link and to explore in greater detail the existence of direct f-orbital participation in bonding, we need to access chemical systems that allow greater control over and variability in the molecular structure.

Organometallic actinide chemistry provides an opportunity to exercise precise control over the structure of the coordination environment around the actinide ion, while also allowing the introduction of a broader range of ligands than possible in classical coordination chemistry. Thus, in addition to classical ligands like the Cl- ion, unusual ligands can be introduced that are capable of participating in both s and π bonding with the metal center. We recently investigated the electronic structure and bonding in a series of tetravalent organouranium complexes of the general formula (C₅Me₅)₂U(R)(R') where R and R' are ancillary ligands including halide, triflate, alkyl, imido, hydrazonato, and ketimido.

Two of the most useful experimental probes of electronic structure in discrete molecular systems are electronic absorption spectroscopy and electrochemistry, in particular cyclic voltammetry. The former technique, also commonly referred to as UV-visible-near-infrared or optical spectroscopy, provides information about the energies and intensities of transitions between the electronic states in a molecule. For the organoactinide complexes, we will be interested in transitions between the electronic states derived from the 5f orbitals and between states derived from electron excitations from these 5f orbitals to other molecular orbitals, both metal- and ligand-based.

Voltammetry is based on the addition or removal of electrons from molecules at a solution/solid electrode interface. Because these electron transfer processes occur from the highest occupied orbitals (oxidation) or to the lowest unoccupied molecular orbitals (reduction), this technique gives a direct measure of the stability of the oxidation state of the parent complex and the energy to change that oxidation state by ± one or more electrons. Ideally, the trends in the data provided by both experimental probes can be related back to trends in structural variations to obtain a deeper understanding of the influence of molecular structure on electronic structure.

Cyclic voltammetry provides information about the energy required to add or remove electrons from molecules by scanning the potential in a cyclic fashion and measuring the current that results. Peaks in the current response (positive for addition of electrons, negative for removal of electrons) signal the onset of electron transfer, and the potential at which these peaks occur provides a relative measure of the energy of the process. Here we show voltammetry for three molecules. Complex 1 (green) only exhibits a single wave (the terminology for a combined positive and negative peak that are slightly offset from each other in potential) at ~-1.8 V assigned to the U(IV) / U(III) transformation in this complex. Complex 4 (blue) has similar features but at more negative potential values. (Note that voltammograms are plotted with more negative potentials to the right.) Complex 7 (green) exhibits a wave at ~-2.5 V for U(IV) / U(III) and a wave at ~-0.5 V. This additional wave is assigned to a U(IV) / U(V) oxidation process. This new wave signals the onset of new bonding features associated with the nitrogen-containing ligands.

Redox properties from cyclic voltammetric studies

To investigate the influence of the ancillary ligands (R, R') on the energetics of the metal-based redox couples, we focused on the redox waves attributable to the uranium metal-based processes. All of the complexes can be classified into only two categories: Category 1 includes complexes where R and R' are chloride, alkyl, or triflate groups; Category 2 includes complexes where at least one of the R or R' groups is a hydrazonato, imido, or ketimido nitrogen donor ligand.

The ancillary ligands in the Category 1 complexes are all principally s donors, and the voltammetry in this category is dominated by a single reduction process from U(IV) to U(III) that occurs between ~ –1.8 and –2.7 V versus [(C₅H₅)₂Fe]^+/0. A useful comparison is the variability in the U(IV)/U(III) redox potential for differing ligand sets. For example, it has been shown that the uranium metal center becomes more difficult to reduce in the series (C₅H₅)₃UCl < (C₅H₅)₄U < (C₅Me₅)₂UCl₂. This corresponds to the stabilization of the tetra- valent oxidation state due to the electron-donating ability of the ligands, with donor strength decreasing from C₅Me₅^- to C₅H₅^- to Cl^-. For the Category 1 members, the reduction potential reflects the more strongly electron-donating nature of the alkyl groups relative to the chloride ion. No complexes in the first category show any evidence for a metal-based oxidation process (i.e., U(IV) to U(V)).

All complexes in Category 2 contain one or two nitrogen-containing ligands that have the ability to interact with the metal center in both s and π modes via the nitrogen lone pair orbitals and/or the π orbital system on the ligand (for the hydrazonato complexes). All these Category 2 complexes exhibit a reversible reduction wave at potentials between ~–2 and –2.6 V that is attributable to a metal-based U(IV)/U(III) process. The distinguishing feature of these Category 2 complexes is the appearance of an oxidative process in the region from ~+0.1 to –0.6 V. The U(V) species appears to be stable on a voltammetric time-scale, with no evidence for chemical reactivity such as disproportionation.

The nitrogen ligands in the Category 2 complexes provide the same measure of stabilization of tetravalent uranium as do the alkyl ligands in the bis and tris alkyl complexes in Category 1. Thus, the potential of the U(IV)/U(III) couple is shifted to quite negative potentials compared, for example, to the bis halide. However, in the case of the nitrogen ligand complexes, the metal center becomes so electron rich that the one-electron oxidation process is also readily accessible and yields a stable pentavalent complex. The difference between these Category 2 complexes and those in Category 1 is the ability of the nitrogen ligands to engage in π bonding with the metal center. Undoubtedly, in the complexes with imido, ketimido, or hydrazonato ligands, the stability of the pentavalent oxidation state derives from the additional π interaction between these nitrogen ligands and the metal center, stabilizing the high valent oxidation state in a way that simple s-donor ligands alone cannot.

Electronic absorption spectra provide us with important information about the energies required to promote electrons between various states within the molecules. For the actinide molecules studied here, there are two important yet distinct types of transitions that occur in different energy regions. The first, shown in green, are associated with changes in electronic states derived from the actinide f orbitals. The second, shown in blue, derive from transition in which the electron is promoted from an orbital localized on one portion of the molecule (for example, the metal) to another. We are currently attempting to develop theoretical justifications for the observed differences in molar absorptivity for these transitions in the structurally different complexes (4, 8, and 7). We believe these intensity differences are a direct indicator of the extent of covalent interaction between the uranium metal center and the ligands.

One of the more interesting trends in the redox energetics is that in the Category 2 complexes the potential separation between the U(V)/U(IV) couple and the U(IV)/U(III) couple remains nearly constant over the entire series of complexes (~2.1 V). A similar potential separation has been found in most of the other published voltammetric studies of U(IV) metallocene complexes having both U(V)/U(IV) and U(IV)/U(III) couples. Thus, it appears that this potential separation is consistent across most U(IV) cyclopentadienyl complexes having high electron density at the metal center and is not a strong function of the nature of the ancillary ligands.

Electronic structure and properties from optical spectroscopies

The UV-visible-near-infrared electronic absorption spectra for these (C₅Me₅)₂U(R)(R') complexes all have two different energy regions of interest. The first region, which lies in the energy region below ~15,000 cm^-1, contains the electronic transitions between the states derived from the 5f² electronic configuration (f-f bands). The second region lies at energies above ~15,000 cm^-1. In general, this region will contain transitions derived from promotion of a 5f electron into higher energy metal-based 6d orbitals as well as transitions to ligand-based nonbonding and antibonding orbitals. These latter transitions are referred to as metal-to-ligand charge transfer (MLCT) because the electron is transferred from a metal-based orbital onto an orbital localized on a ligand.

These electronic spectral data also fall into distinct classes related to the nature of the ancillary ligands on the (C₅Me₅)₂U core. The first class of complexes includes those that have ancillary ligands of essentially s-donor character. All of the complexes in this class have a large number of narrow bands in the f-f region at similar energies. The most important distinguishing characteristics of the spectra for this class of complex are the relative weakness of both the f-f bands and the less-well-resolved, higher-energy bands associated with f-d and MLCT transitions.

The second class of complexes contains a hydrazonato ligand. For complexes in this class, there is also a high degree of similarity in the number and energy of the bands in the f-f region of the spectra, as well as in the broad, poorly resolved bands in the high-energy region. For spectra in this class, however, the intensity of the bands in the f-f region is on average about 50% greater than for those in the class discussed above, and for the broad band in the higher energy region, the intensity has increased dramatically.

The final class of complexes includes the imido and ketimido complexes. For spectra in this last class there is an apparent decrease in the number of f-f bands in the low-energy region. The higher-energy region becomes significantly more structured, although the bands remain unresolved. The f-f bands are even more intense in this class, and the bands in the higher-energy region are comparable in intensity to those in the second class. Notably, the higher-energy bands extend to much lower energy in the spectra for this class than in those of the previous two classes.

The intensity variation across these three classes of spectral data sets is clearly the most interesting and significant result. The intensities of the f-f bands for the first class of complexes are nominal for actinide complexes having a 5f2 electron configuration, and reflect the Laporte spin-forbidden character of these transitions. The substantial increase in intensity observed in the f-f bands for the second and third classes suggest the presence of a new intensity-generating mechanism. The existence, intensity, and energetic proximity of the broad, unstructured, higher-energy bands in these spectra appear to hold the key to this unusual trend in intensity behavior in the f-f bands.

We propose that the increase in intensity in the f-f bands for the second and third class of complexes is a direct result of an intensity-stealing mechanism in the f-f transition moment integral due to coupling of these electronic states to the higher-energy MLCT states in these nitrogen-containing complexes. This type of intensity-stealing mechanism reflects a significant degree of covalency in the bonding between the metal and ligand responsible for the charge-transfer transition. The extent to which the mechanism adds intensity to the f-f transitions is also related to the energetic separation between the metal-based states and the charge-transfer states. The closer in energy the states, the stronger the coupling and the intensity-stealing process.

The key feature of the present series of uranium metallocene complexes that has made it possible to readily identify this interesting electronic phenomenon is the incorporation of nitrogen-bearing ligands. These ligands introduce low-energy, ligand-based orbitals that engender the lower energy charge-transfer transitions to facilitate coupling. The ligands also provide a mechanism for enhanced covalent interaction with the metal center because of the presence of π orbitals of proper symmetry to mix with the metal f orbitals. Thus, they provide a direct experimental observable (the f-f band intensities) for the degree of covalent interaction between the metal f orbitals and the ancillary ligands. The increase in intensity in the f-f bands in the third class of complexes reflects the fact that the MLCT transitions for these complexes extend to much lower energy and can therefore interact more strongly with the f-f states.

Conclusions

Important trends have been identified in both the electrochemical and the spectral data for these organouranium complexes. Furthermore, these trends are very much related to each other. The voltammetric data illustrate that the nitrogen-bearing ligands result in greater electron density at the metal center, destabilizing the U(III) oxidation state and significantly stabilizing the U(V) oxidation state. These same ligand systems also induce increased intensity in the metal-based f-f transitions by providing an intensity-stealing mechanism that reflects stronger coupling to the charge-transfer excited states and indirectly demonstrates a greater degree of covalency in the metal-ligand bonding than is observed in the s-donor ligand systems. Ultimately, these studies demonstrate that these important trends can be most readily identified when synthetic methodologies are available to provide a broad range of coordination environments to investigate.

This article was contributed by Los Alamos researchers David E. Morris and Jacqueline L. Kiplinger of the Chemistry Division.

NMT | LANL | DOE
Phone Book | Search | Help/Info

L O S  A L A M O S  N A T I O N A L   L A B O R A T O R Y
Operated by the University of California for the US Department of Energy