Ԫ The Actinide Research Quarterly: 1st Quarter 2004 contents

Homoleptic nine-coordinate An(III) and Ln(III) complexes

Classic high-symmetry species that advance our understanding of f-element structure, bonding, and dynamics

Understanding the behavior of actinide and lanthanide ions in water and other solvent media remains one of the great experimental and theoretical challenges in f-element chemistry. The behavior of actinide ions in aqueous solution is highly relevant to issues in nuclear material processing, concerns about actinide-ion transport in groundwater, and hydrological implications of spent fuel storage strategies. The strong Lewis acidity of the actinide ions leads to complex, highly pH-dependent equilibria in aqueous media. (Lewis acids are substances with a strong affinity for accepting electron pairs, but which do not necessarily directly affect hydrogen ion concentration).

Understanding the nature of f-element solvent interactions is also important for fuel cycle processes such as actinide/lanthanide separations and require an understanding of the intimate coordination sphere of solvent molecules about actinide and lanthanide ions, of the interactions of the primary coordination complexes with the bulk solvent that surround them, and of the dynamical processes between the coordinated and bulk solvent molecules.

Successful modeling of the behavior of lanthanide and actinide ions in solution begins with a detailed understanding of the direct interactions between the ions and solvent molecules as ligands. The level of understanding necessary for such modeling can greatly benefit from studying the bonding, electronic structure, spectroscopy, and chemical properties of well-defined coordination complexes of the early actinides. Given the importance of isolating the important ion-solvent interaction, the strategy of using noncoordinating or very weakly coordinating counterions was adopted; exemplary is the trifluoromethane sulfonate anion (CF3SO3, commonly called triflate).

This thermal ellipsoid representation shows the ideal tricapped, trigonal-prismatic structure of the Pu(III) aquo complex (the triflate anion is not shown). The six symmetry-equivalent prismatic oxygen ligands (red) define the vertices of the trigonal prism. The three capping oxygen ligands (blue) are positioned outside the rectangular faces of the prism. The plutonium atom (green) is in the center.

Triflate salts of Pu(III) and Am(III) aquo complexes--[An(OH2)9][OTf]3--were prepared, in which the oxygen atoms of the nine-coordinated water molecules adopt a highly symmetric tricapped, trigonal-prismatic (TCTP) geometry about the Pu3+ and Am3+ ions. Lanthanide triflates having the same compositions and analogous geometries are well characterized, allowing for a direct comparison between the structure and bonding in An(III) and Ln(III) complexes.

Because the valence electronic structure of actinide ions are dominated by atomic 5f and 6d orbitals, whereas that of the lanthanide ions involves the 4f and 5d orbitals, the comparison of these two classes of complexes might provide valuable insight into how electronic structure affects geometric structure and reactivity.

The TCTP structure has six symmetry-equivalent prismatic ligands at the vertices of a trigonal prism. The remaining three ligands are capping ligands positioned outside the rectangular faces of the trigonal prism. The capping ligands are equivalent to one another by symmetry, but are unique from the six prismatic ligands. As such, the metal-ligand distances for prismatic and capping ligands can be different as demonstrated in the Pu(III) aquo structure, which contains Pu-OH2 bond distances of 2.574(3) Å for the capping waters and 2.476(2) Å for the prismatic waters. Bond distances for the Am(III) complex are slightly shorter, correlating well with the decrease in ionic radius across the actinide series.

We are engaged in computational modelling of these actinide species in solution. Our theoretical studies of solvated f-element ions (those chemically bound to solvent molecules such as water) will ultimately use a combination of quantum mechanics and molecular dynamics. The electronic structural description of the direct interactions of the solvent molecules with the f-element ions requires relativistic quantum chemical techniques.

Once the primary coordination sphere (first layer of solvent molecules surrounding the central dissolved ion) is correctly described, molecular dynamics will be used to model the interaction of the first-coordination-sphere complex with the bulk solvent. The crystalline experimental complexes, which isolate the first coordination sphere, thus provide an opportunity to test the quantum chemical description of the metal-ligand interactions. In the present studies, density functional theory (DFT) has been used to perform this assessment. Scalar relativistic effects were included via the zero-order regular approximation (ZORA) method.

Theoretical studies have shown that the vacant 6d atomic orbitals of the actinide atom (in this case plutonium) are the principal sites of ligand electron acceptance. This contour plot of Pu(H2O)93+ shows the water ligands donating primarily into the plutonium 6d orbitals.

The structure of isolated Pu(H2O)93+ was calculated using relativistic DFT with the ion constrained to a rigorous TCTP geometry. These calculations will suffice for discussing the s donation from the H2O ligands into appropriate acceptor orbitals on the plutonium center. The calculated Pu-O bond distances (PuOcap = 2.619 Å and PuOpris = 2.540 Å) are slightly longer than those observed in the crystal structure. Interactions involving the second lone pair of each H2O ligand will involve subtle PuO effects. bonding describes a particular type of overlap between the atomic orbitals of bonding electrons, differing from the more symmetric s electron-density overlap.

It is important to note that the calculations predict correctly that the PuOcap distance is longer than the PuOpris distance. Relaxing the symmetry to allow the water molecules greater rotational freedom, which will allow them to maximize their bonding to the Pu3+ center, could lead to some shortening of the Pu-O bonds and improve the agreement with the observed crystal structure. Overall, there is clearly good agreement between experiment and theory on these complex systems under the constraints of high symmetry and without specific consideration of anion effects.

The calculations on Pu(H2O)93+ provide insights into the nature of the Pu-OH2 bonding. As expected, the water ligands act as electron-pair donors (Lewis bases) that donate electron density to the Pu3+ ion, which serves as an electron acceptor (Lewis acid). Our previous theoretical studies of organoactinide complexes demonstrated that the vacant 6d atomic orbitals of the actinide atom served as the principal sites of ligand electron acceptance, whereas the 5f orbitals were used to house any metal-based electrons. The same bonding paradigms appear to be present in Pu(H2O)93+. The water ligands donate primarily into the plutonium 6d orbitals.

To study structure and bonding properties of actinides, researchers dissolve Pu-239 metal strips (shown in the vial above left) or metal powder in water or acetonitrile and place them under an argon atmosphere. The resulting solids (shown in the photo at right and in this case plutonium strips dissolved in acetonitrile) are analyzed by a number of methods, including single-crystal x-ray diffraction.

Because a neutral plutonium atom has eight valence electrons, the Pu3+ ion is expected to have five metal-based electrons. Our research indicates that this is indeed the case and that those electrons reside in nearly pure plutonium 5f orbitals. This dichotomy in the roles of the actinide 5f and 6d orbitals will change somewhat for lanthanide complexes because of the different spatial extents of the atomic orbitals between the two rows of the periodic table.

From experimental studies, we can compare the plutonium and americium aquo structures with those for lanthanides of similar ionic radii, Nd(III) and Sm(III). The capping waters have nearly identical bond lengths, while the An(III) prismatic waters are approximately 0.03 Å shorter than those of the Ln(III) with the same ionic radii. Electronic structure calculations aimed at explaining these differences are under way.

To further explore the structure and bonding properties, we have recently prepared nonaqueous trivalent actinide complexes that are nearly isostructural with the aquo complexes and contain weaker s-donor ligands coordinated to the metal. For example, [Pu(NCMe)9][PF6]3ĢMeCN was prepared by treating an acetonitrile suspension of Pu-239 metal turnings under argon atmosphere with three equivalents of either AgPF6 or TlPF6. Compared with the PiO in the aquo complex, the average PuN distance (2.572 Å) is 0.047 Å longer. The geometry of the acetonitrile complex is distorted rather than ideal TCTP, perhaps because the PF6 ion is more strongly coordinating than the triflate anion.

Optical absorbance and diffuse reflectance spectra suggest that Pu3+ has very similar coordination geometries in the solution and solid states of the acetonitrile and aquo complexes. This in turn suggests that the ion is nine-coordinate in solid and solution forms and in both solvent environments. The acetonitrile spectra are shifted by about 15 nm relative to the aquo spectra, reflecting the differing Lewis basicity of the ligands. The top two spectra are diffuse reflectance spectra of the solids. The bottom two spectra are the absorbance spectra of the solutions.

Diffuse reflectance spectra obtained on ground crystals of the solids and optical absorbance spectra of the solutions for the aquo and acetonitrile complexes are nearly superimposable, but each feature in the spectrum of the acetonitrile complex is red-shifted by approximately 15 nm, providing spectroscopic evidence of the difference in bond strengths.

In contrast to the plutonium complex, the uranium aceto-nitrile adduct we prepared displays nearly ideal TCTP geometry. The six prismatic UN distances are 2.60 Å. The difference between the average prismatic MN distances of Pu and U complexes is 0.026 Å, which is close to the 0.025 Å difference in the atomic radii of U and Pu. The three 2.65 Å capping UN interactions in the U complex are longer than the prismatic interactions by 0.05 Å, and 0.087 Å longer than the average PuN capping interactions in the plutonium acetonitrile complex. This UN capping distance is 0.062 Å longer than that predicted by atomic radii differences between U(III) and Pu(III).

Compared to nine-coordinate acetonitrile Ln(III) tricapped trigonal prismatic adducts (Ln = La, Sa, and Pr), the three capping LnN distances are on average slightly shorter than the six prismatic ones in the crystal structures of La(Davg = 0.016 Å) and Pr (Davg = 0.004 Å). This trend reverses in the Sa complex (Davg = 0.008 ). However, the difference between the average prismatic and average capping distance in a particular Ln(NCMe)93+ compound is small.

Published DFT calculations of the nine-coordinate acetonitrile Ln3+ (Ln = Eu, Yb, and La) solvates predicted that the three capping distances of Ln(NCMe)93+ would be longer than the six prismatic distances. Qualitatively, the uranium complex follows the trend predicted by DFT, but the plutonium complex follows the trend of the experimentally determined lanthanide structures.

We will continue to combine experimental and theoretical methods to explore the rich chemistry of f-element complexes in solution. Such judicious combination of synthetic and structural chemistry with modern computational modelling provides understanding of a very complex and relevant topic in current f-element chemistry.

This article was contributed by Los Alamos researcher Mary P. Neu of the Chemistry Division and Jason L. Sonnenberg and Bruce E. Bursten of the Department of Chemistry, The Ohio State University.


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

Questions? - Copyright © UC 1998-2000
- For conditions of use, see Disclaimer