contents

A unique role for 5f orbitals in bonding

Unusual tetraoxo coordination in heptavalent neptunium

The light actinide elements—uranium, neptunium, plutonium, americium, and curium—in their highest oxidation states (V–VII) have an unusually high affinity for formation of metal-oxygen bonds and demonstrate a high degree of covalency and metal-ligand multiple bonding. This is particularly true for oxo complexes, where, for example, the U=O multiple bond in the UO22+ ion is remarkably short, and the mean bond strength is similar to that of the C=O bond in CO2.

These unusually strong bonds can be traced to the ability of the uranium atom to use a combination of 5f, 6d, and 6p atomic orbitals in chemical bonding. Understanding the relative roles of the actinide 5f, 6d, and 6p atomic orbitals in forming such strong bonds has been a fundamental challenge in actinide chemistry since the recognition of the presence of the 5f series, and a good deal of the fundamental chemistry and physics of light actinides has been devoted to understanding the nature of these bonding interactions. Historically, the relative roles of 5f/6d/6p atomic orbitals in forming chemical bonds within the linear dioxo ion (AnO22+) of hexavalent actinides generated a healthy scientific debate. Much of this debate originated at Los Alamos, as exemplified by the seminal arguments on why ThO2 is bent in the gas phase and UO22+ is always linear. It took the clever application of two-photon spectroscopy and oxygen k-edge x-ray absorption spectroscopy by Robert Denning of Oxford University to sort it all out.

The remarkable strength of actinide-oxygen bonds:

a comparison of gas-phase bond enthalpies

UO2 710 kJ/mole

UO22+ 701 kJ/mole

MoO2 587 kJ/mole

CO2 802 kJ/mole

What we learned from those studies has changed our fundamental thinking about valence orbitals in heavy-element chemistry. The radial distributions of the early actinide 6s and 6p atomic orbitals are not buried within the core of the atom but lie in the valence region and must be considered to be active in chemical bonding. The ability of the 6p orbitals to hybridize with the 5f can lead to unusually strong σ bonding, and this type of 6p-5f hybridization was shown to be a predominant contributor to the strong covalent bonds in linear UO22+ ions.

More recently, we have turned our attention to light actinide ions in even higher oxidation states to look for evidence of increased covalency and 5f orbital participation in bonding. Our early studies have focused on developing the basic understanding of the molecular structure of actinide ions in the heptavalent state (oxidation state VII) based on x-ray diffraction and absorption, and assessing the relative bond strengths as determined by nuclear magnetic resonance and vibrational spectroscopy.

Molecular structure determination of the NpO4(OH)23- ion

The molecular structure of the NpO4(OH)23- ion (top) in the solid state was determined by x-ray crystallography. This thermal ellipsoid drawing emphasizes the pseudo-octahedral coordination geometry about the central neptunium ion. This ion has four short Np–O bonds (1.88 Å) perpendicular to the page and two long Np–OH bonds (2.33 Å) in the plane of the page. The structural parameters for the NpO4(OH)23- ion (bottom) were determined by XAFS spectroscopy. This figure shows the Fourier transform of the XAFS spectrum (solid black line) and the theoretical fit (dashed red line). The components of the fit, shown beneath the spectrum with negative amplitudes, correspond to individual shells of atoms. Note that there are no atoms at 2.72 and 3.76 Å. The peaks labeled "ms," which are routinely observed in XAFS data of linear ions, are due to multiple scattering of a photoelectron off the oxygen atoms in the linear O=Np=O unit. The radii of the coordination shells observed in the XAFS (1.88 and 2.30 Å) match the solid-state structure, giving high confidence that the NpO4(OH)23- anion is present in solution.

The heptavalent oxidation state of actinide ions is rare, but has been known since the late 1960s due largely to the pioneering efforts of Russian scientists V.I. Spitsyn and N.N. Krot. X-ray diffraction data show that heptavalent neptunium can exist in the familiar trans dioxo form, as illustrated by the structure of CsNpO4 shown in I, or in an unusual tetraoxo form with a square planar arrangement of O atoms, as indicated by the structure of Li5NpO6 shown in II, though we note that the identity of this structure has been recently questioned by L.R. Morss and coworkers at Argonne National Laboratory.

To understand the nature of chemical bonding in the unusual square planar tetraoxo ion, we have been studying the physical and spectroscopic properties of discrete molecular forms of heptavalent neptunium and plutonium. For example, we have determined the molecular structure of the discrete NpO4(OH)23- ion by x-ray diffraction in the solid state and by XAFS spectroscopy in both the solid state and solution. Using a technique pioneered by Krot and coworkers at Moscow’s Institute of Physical Chemistry, we isolated the NpO4(OH)23- ion from alkaline solution, using a rather large Co(NH3)63+ cation, to give a solid compound of formula [Co(NH3)6][NpO4(OH)2]·2H2O.

Solid-state structure of CsNpO4. This structure has a puckered layer of NpO4 sheets, with axial Np=O bonds above and below the layer. Np atoms are green, O atoms are red, and Cs atoms are purple.

Solid-state structure of Li5NpO6. This structure has an isolated unit, with four equatorial Np=O bonds above and below the plane of metal atoms, and two short Np–O bonds within the plane. Np atoms are green, O atoms are red, and Li atoms are purple.

The solid-state structure determined by x-ray diffraction is shown in the figure at upper left. The tetraoxo ion of formula NpO4(OH)23- displays a highly unusual geometry with four oxo ligands in an equatorial plane and two trans OH- ligands, one above and one below this plane. The four planar O ligands have an average Np–O distance of 1.878(5) Å, and the two Np–OH bonds are 2.238(4) Å. XAFS spectroscopy was used to determine structural details of the Np(VII) ion both in highly alkaline solution and in the solid state. The figure at lower left shows a representative Fourier transform (without phase corrections) of the k3-weighted XAFS spectra, the theoretical fit to the data, and illustrates the single-shell contributions to the fit. These bonding parameters are nearly identical to those found in the crystal structure determination described above.

It is very unusual for a metal ion to have four coplanar oxo ligands as indicated in III. For transition metals in their highest oxidation states, the tetraoxo ions of general formula MO4- (Mn, Tc, Re) or MO42- (Cr, Mo, W) are always tetrahedral as shown in IV. In the tetrahedral geometry, all five metal d orbitals can be used to maximize the M–O bonding. The unusual planar arrangement of O atoms in NpO4(OH)23- can be traced to the presence of 5f orbitals and the use of a combination of 5f, 6d, and 6p orbitals in chemical bonding.

The 1.88 Å Np=O bonds in the planar NpO4(OH)23- ion are rather long when compared with other Np=O bonds in dioxo species, which average around 1.75 Å for NpO22+ ions. Transition metal tetraoxo ions are even shorter, as evidenced by the comparison with the ReO4- ion, which shows four Re=O bonds at 1.69(2) Å. Thus, the long Np=O bonds observed in NpO4(OH)23- suggest that the Np=O bond is weakened relative to the Np=O bond in NpO22+ ions.

To better understand the relative strengths of the Np=O bond in NpO4(OH)23- we turn to other forms of spectroscopic characterization. Raman spectroscopy can provide sensitive insights into the strength of the M=O bonds and has found wide utility in the study of actinide oxo complexes. The idealized molecular structure of NpO4(OH)23- possesses D4h symmetry, and this tetraoxo fragment I should display two separate vibrations in the Raman spectrum—a totally symmetric mode of A1g symmetry and an asymmetric mode of B1g symmetry, as shown in V.

Raman spectra recorded for Np(VII) solutions at varying hydroxide concentrations and in the solid state always showed a single invar-iant Raman peak. Since it was possible that the two vibrational modes were coincidentally overlapping one another, depolarization ratio studies were used to assess whether the mode we were observing was the totally symmetric (A1g) vibration, the asymmetric (B1g) vibration, or an overlap of the two. Totally symmetric vibrations give rise to polarized Raman lines, whereas nontotally symmetric vibrations give Raman lines that are depolarized.

Polarized Raman spectra of the NpO4(OH)23- ion identify the vibrational symmetry

A totally symmetric vibrational mode gives rise to a polarized Raman line, and a vibration with lower symmetry is depolarized. When polarized laser radiation is employed, we can monitor the polarization of the scattered Raman line to determine the symmetry of the vibration. This figure shows a solution spectrum of NpO4(OH)23- recorded with perpendicular and parallel polarization of the exciting laser. The observed frequency is clearly polarized, and can be assigned to the totally symmetric A1g vibration.

In the figure at upper right, a marked decrease is seen in the intensity of the Raman band when observed in a perpendicular as opposed to parallel orientation, thus identifying a symmetric A1g vibration. The decrease in intensity of the A1g band also serves to demonstrate that there was not a B1g band masked underneath this mode, since the B1g band should still be present at full intensity in the figure at lower left. Raman spectroscopy on single crystals also revealed a single A1g vibrational mode identical to that observed in the solution. The reason that we don't observe the asymmetric vibration in our studies is under evaluation, but it is possible that we are observing resonance enhancement of the totally symmetric mode. Additional studies where we vary the frequency of the exciting laser are in progress.

For single-crystal samples, the symmetrical Np=O vibrational mode is observed at the extremely low frequency of 716 cm-1, while in 2.2 M LiOH solution this value increases slightly to 735 cm-1. These low vibrational frequencies confirm that the Np=O bond is weakened relative to the more common dioxo ions, which show typical Np=O stretching modes in the higher-frequency (stronger bonds) range 914–952 cm-1. The longer, weaker bonds observed in the tetraoxo ion are an indication that there is much less π-electron donation from the oxo ligands to the metal center in tetraoxo versus dioxo ions. This assessment is supported by the observation of an 17O NMR chemical shift at δ 1470 ppm, which can be compared to the value of 1125 parts per million observed for UO22+ under the same solution conditions.

The theory of NMR chemical shifts is quite complicated, but in general, a lower electron density on an oxygen atom leads to a greater downfield shift. Our data are therefore consistent with the notion that the NpO4(OH)23- ion has more electron density on the oxygen atom relative to the transdioxo ions. The combination of this data-a long Np=O bond, a low (Np=O) vibrational frequency, and a high-field 17O resonance-all point to weaker bonding in the Np=O bonds of tetraoxo ions relative to dioxo ions. Our current efforts are focused on quantifying the relative contributions of σ and π bonding and the relative roles of the 5f, 6d, and 6p orbitals in forming these bonds in the square planar NpO4 unit in NpO4(OH)23-.

This article was contributed by Los Alamos researchers David L. Clark of the Nuclear Materials Technology Division; Phillip D. Palmer, C. Drew Tait, and D. Webster Keogh of the Chemistry Division; Steven D. Conradson, of the Materials Science and Technology Division; and Robert J. Donohoe of the Biosciences Division.

The relative roles of 5f, 6d, and 6p atomic orbitals

Why UO22+ is linear and isoelectronic ThO2 is bent

The UO22+ ion is well known in actinide chemistry and displays a linear geometry in every known example of its complexes. In contrast, the isoelectronic ThO2 molecule is bent in the gas phase with an O–Th–O angle of 122°. The theoretical understanding of these differences has taken nearly 20 years to understand fully and hinges on a detailed understanding of the relative roles of the atomic 5f, 6d, and 6p orbitals in forming the metal oxygen bond in these two systems.

In the early 1980s, there were two points of view: one based on U–Oπ bonding (Bill Wadt at Los Alamos) and another based on U–Os bonding (Roald Hoffmann at Cornell University). The Los Alamos view was that for U, the 5f levels are significantly lower in energy than the 6d, while for Th, the 6d level is lower than the 5f. Theoretical calculations showed that 6dπ bonding was important for both Th and U, but that U could also use the 5f orbitals for π bonding. A bent geometry was preferred for Th due to the dominance of the 6dπ bonding, and a linear geometry was preferred for U because a linear geometry could use both 6d and 5f orbitals for bonding. The Cornell point of view, also based on theoretical calculations, was that U could use a combination of 5f and 6p (normally considered to be in the core) orbitals to form an unusually strong s bond in the linear geometry.

Twenty years later and after the application of very sophisticated spectroscopy (two-photon laser and oxygen k-edge x-ray absorption) and more modern theoretical calculations based on the hybrid density functional approach, we have found a resolution of these two points of view. In the final analysis, both s and π bonding are found to be important. Both fπ and dπ bonding occur, but the dπ bonding is significantly stronger. Moreover, in a strange and unexpected way, we also learned that the fs hybridizes with 6ps, and the resulting f-p U–-Os bond is much stronger than the U-–Os bond formed by the d orbitals. Therefore, the dominant covalent interactions in linear OUO2+ are due to the strong 5f-6ps bonding and 6π bonding.


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