An interpretation of the structure of PuO_2.25 (Pu₄O₉) as Pu₄O₈OH

Proposing a hydroxide ion, rather than a central oxide ion, in the structure

The preceding article described how, at temperatures of 50 to 350 degrees Celsius, the reaction of plutonium dioxide (PuO₂) with water yielded hydrogen gas and increased oxygen content in the solid. The composition of the resulting fluorite-related solid phase was described as PuO_2+x with x=0.265 as the oxygen increase limit. The importance of this finding is that plutonium dioxide is suggested to oxidize above a plutonium valence state of (IV) with a commensurate weight gain in the solid material. This finding is profound in that all other attempts to produce a binary oxide of plutonium with a higher valence have failed. That the material can be prepared in a nominal oxidation state higher than plutonium(IV) using water as an oxidant is especially puzzling. (Later articles will discuss the crucial role of water.)

The molecular structure of plutonium dioxide (PuO₂) with an added oxygen ion (yellow) in the center. Oxygen atoms are shown in red, and plutonium in black.

A structure that purportedly contains all the salient features of the PuO_2+x entity is the unit cell description of PuO₂ with an added oxygen ion in the center of the structure. This structure, proposed in a report, "First-Principles Calculations of PuO_2±x," that appeared in Science, Vol. 301, July 25, 2003, and described in the previous article, is drawn with strictly oxygen ions, yielding a unit cell composition of Pu₄O₉. We propose replacing the central oxide in Pu₄O₉ with hydroxide, yielding Pu₄O₈OH.

There are several ways to achieve charge balance over the unit cell. One obvious way is to assign one plutonium atom a valence state of (VI). Alternatively, one could spread the charge over two plutonium core atoms in the unit cell and arrive at Pu₂⁵⁺Pu₂⁴⁺O₉, or possibly spread the charge over the entire unit cell core so that each plutonium atom that shares the charge is balanced equally with an average plutonium valence of +4.5. Experimental data using x-ray absorption near edge structure (XANES) and x-ray photoelectron spectroscopy (XPS) unequivocally demonstrate that the PuO_2+x entity has plutonium(V) in the structure. (This will be explained in depth in the following articles.)

This new material, possessing plutonium with an overall valence greater than (IV)-PuO_2+x, has unit cell dimensions scarcely increased over those of PuO₂-5.404 angstroms () versus 5.3975 -in spite of the suggested placement of the extra oxide ion at body center, enclosed by the close-fitting subcell of eight oxide ions. The dominant x-rayscattering arises from plutonium, indicating that the plutonium lattice positions remain essentially unchanged. Structural determinations, whether performed using x-ray or neutron diffraction, have provided the preponderance of physical data upon which we base our understanding of structure and bonding in the actinide oxides and related compounds. More than half of all reported plutonium structures containing oxygen were determined by William H. Zachariasen before 1954 (see sidebar on Page 14). Coupled with single-crystal data obtained on the 5f fluoride compounds where structural data were available, Zachariasen summarized bond length-bond strength relationships among this class of materials and demonstrated the utility of this approach in predicting bond lengths when bond strengths are known or determining bond strengths when bond lengths are known.

Zachariasen bond length-bond strength formulations

Development of an adequate description of bond strengths and lengths has a long history, going back to 1929 and Linus Pauling's definition of bond strength as ion valence divided by its coordination number and Zachariasen's later derivation of the dependence of bond lengths on coordination number. By the middle of the twentieth century it was recognized that interpreting experimental bond lengths in terms of fixed radii (either ionic or covalent) was problematical and resulted in difficulty in explaining the large variation in bond lengths observed in the first coordination sphere of many crystalline solids that included the oxides and halides.

Based on data from well-established oxygen and halogen compounds of the 3d, 4d, 4d-5f, and 6d-5f elements, equations were formulated relating bond strengths to observed bond lengths. These equations largely supplanted the use of ionic radii sums, which did not account for the observed variation in bond lengths. Although seldom referenced in recent literature, Zachariasen's summary equations provide a useful starting point for interpreting both oxygen and halogen compounds of the d and f elements.

In the Zachariasen model, bond strength is defined such that s_ij = s_ji and relates to the bond between the ith and jth atoms in a structure

Sigma s_ij = v_i and Sigma s_ji = v_j ,

j i

where nu_i and nu_j are the valences and oxidation states of the two atoms. The length of the bond D between the two atoms is thus a function only of the strength of the bond. The great value of Zachariasen's work in this area was that it encompassed in a simple, quantitative equation the common observation that longer bonds are weaker than shorter bonds. The constants in the equation are based on x-ray structural measurements of bond distances in known structures, many provided by Zachariasen.

The goal is to find an empirical function that adequately describes and predicts bond lengths when bond strengths are known or that determines bond strengths when bond lengths are known. The power of this approach is that a preponderance of the actinide compounds containing oxygen can be shown to exhibit very predictable bond lengths. By using data from plutonium(IV)-oxygen, plutonium(V)-oxygen, and plutonium(VI)-oxygen distances (determined from many crystal structures), Zachariasen was able to derive the empirical logarithmic equation relating bond strength and bond lengths: D(s) = D₁ -Bln(s); where D(s) is the bond length at bond strength (s), D₁ is the bond length at unit bond strength s = 1, and B is a scale factor. "-Bln(s)" is the correction factor. It represents the deviation from single bond length, i.e., one of unit bond strength. This logarithmic correction shortens the bond length at bond strengths greater than unity (e.g., double bonds) and increases bond lengths for bond strengths less than a single bond.

For plutonium-oxygen bonds, D₁ values for plutonium(III), (IV), (V), and (VI) are 2.142, 2.094, 2.08, and 2.06 , respectively. For plutonium-oxygen bond strengths of unity or less, B = 0.35 is constant. However, the multiplier "B" increases as 0.35 + 0.12(s -1) for bond strengths greater than unity, notably in structures containing the actinide(V) and (VI) dioxo cations, MO₂⁺ and MO₂²⁺. The function is linear over the range of s = 1-2 and displays curvature at s values below unity.

For plutonium(III)-fluorine bonds, the values for B and D₁ are 0.40 and 1.992, respectively. These values are necessary for the plutonium oxyfluoride (PuOF) structural example described below.

For plutonium and oxygen constituents of a structure, the sum of bond strengths to their neighbors is equal to their respective valences, usually to within 5 percent. In his last paper, published in 19) = D₁ -Bln(s), normally predicts the bond length to 0.02  when s is known."

William H. Zachariasen's correct interpretation of the crystal structure for plutonium oxyfluoride (PuOF). Fluorine atoms are shown in green, oxygen in red, and plutonium in black.

Applying Zachariasen's formulasto simple compounds

By applying Zachariasen's formulas to a number of relatively simple plutonium compounds, we arrive at a more complete and correct interpretation for the bond lengths and connections with valence in specific solid compounds. We begin by discussing the relatively simple compound PuOF and then use similar logic to examine the putative Pu₄O₉ ("PuO_2.25") structure, which has also been described as the hyperstoichiometric PuO_2+x entity.

An example of the utility of Zachariasen's approach is found in its ability to predict the correct bond lengths of oxygen and fluorine in PuOF. In this example, the location of the light atoms, oxygen and fluorine, were erroneously determined when ionic radii were used to assign their respective positions in trivalent PuOF, a compound in which each plutonium atom is coordinated by two sets of four anions at 2.569  and 2.385 . Since the ionic radius of the oxide (O^2-is greater than that of the fluoride (F^1-), the longer distance was originally assigned to plutonium-oxygen.

Zachariasen later interchanged the oxide and fluoride positions based on three arguments. First, the (-2) oxide charge shared by four atoms of plutonium(III) yields plutonium-oxygen bond strengths, s = 0.5, and dictates assignment of the shorter bond distance to oxygen. Second, the (-1) fluoride charge shared equally by four atoms of plutonium(III) yields plutonium-fluorine bond strengths, s = 0.25, and thus the longer plutonium-fluorine bond distance. Third, comparison of the bond strength sums makes the interchange clearly preferable.

Original selection: Pu-4O, 2.569 ; s = 0.2952; Pu-4F, 2.385 ; s = 0.3745

Old bond strength sum: Pu = 2.68; O = 1.18; F = 1.50

Preferred interchanged positions: Pu-4O, 2.385 , s = 0.4997; Pu-4F = 2.569 , s = 0.2363

New bond strength sum: Pu = 2.94; O = 2.00; F = 0.95

The sum of the bond strengths over the elements is noted to be much more consistent with chemical intuition. This example demonstrates the simplicity and utility of the Zachariasen bond strength-bond length relationship and its use in discriminating between plausible ligated species. The plutonium oxyfluoride example is particularly germane because of the crystal chemical similarities between the fluorine and hydroxyl ligands.

It is customary to formally denote the actinide MO₂⁺ and MO₂²⁺ cations as double-bonded. At bond strength s = 2, the plutonium(V)-oxygen bond distance is 1.85  and the plutonium(VI)-oxygen bond distance is 1.73 . Bond distances this short are only observed in discrete molecules and rarely occur in solids. Usually, bonding to other ions in the structure reduces the metal-oxygen bond strength and lengthens the multiple bond, providing a range of longer distances. We retain the double bond formality but recognize that its strength is more often less than 2.

Applying Zachariasen's formulas to Pu₄O₉ ("PuO_2.25") and the body-centered oxide ion

Zachariasen's bond strength-bond length relationships are applied to the Pu₄O₉ (PuO_2.25) entity. The value for the cell edge of cubic PuO_2+x of highest oxygen content is 5.404 , yielding a body diagonal of 9.36 . The added oxygen is placed at the body center, midway along the body diagonal at 4.68 . The plutonium(IV)-oxygen distance along this diagonal is 2.33  at s = 0.5 (as shown in the figure on Page 9 with the central yellow sphere). However, increasing the valence of one of the plutonium atoms to (V) yields s = 1.5 and the smaller value of 1.91  for the plutonium(V)-oxygen bond distance. Allowing 2 x 1.4  for the Van Der Waal (VDW) radii sum of the two oxygens and adding the plutonium(V)-oxygen distance of 1.9  yields 2.8 + 1.9 = 4.7 , which agrees well with the experimental value of 4.68  for half the body diagonal.

Conversion of a corner plutonium(IV) to (V) has an effect: increasing s = 0.5 to s = 1.5 for the plutonium(V)-oxygen bond on the diagonal leaves unchanged its seven other plutonium-oxygen bonds at s = 0.5 for a bond sum of 3.5 + 1.5 = 5. This has the strong implication that the cell dimensions will show little size change from those of plutonium dioxide itself, consistent with the x-ray diffraction results described in Luis Morales' article on Page 3.

Withdrawal of the oxygen 0.42  along the body diagonal from its tetrahedral site lengthens somewhat the plutonium(IV)-oxygen bonds on the threefold axis. This may account for the direction and magnitude of the +0.007  change in cell dimensions (5.3975 to 5.404 ). Oxygen volume assignment in compounds containing plutonium(V) is 18 3 and is applicable for oxide, hydroxyl, and water. Using 18 3 for the nine oxygen atoms per unit cell provides a cell volume for Pu₄O₉ of 162 3 versus 158 3 observed.

However, a surprising result arises with the bond strength sum around the central oxide. Octahedral coordination at the body center by the six face-centered plutonium(IV) ions occurs at distances of half the cell edge, or 2.702 . This plutonium-oxygen bond distance corresponds to s = 0.176 and an oxygen valenceΚsum of 1.06, compared to a plutonium(IV)-oxygen distance of 2.72  calculated for s = 1/6. (Note: the bond strength sum normally agrees to within 0.1 valence units and the bond length to 0.02 when s is known.)

This unit charge further substantiates a central ion and the strong possibility of a central hydroxyl ion instead of an oxide ion. A central oxide ion requires s = 2/6 and a much shorter bond distance of 2.48  (not observed). An oxide ion centered in the cubic array of eight oxygen ions would likely expand the lattice by repulsion. Hydrogen bonding by the hydroxyl ion, as well as its lower charge, would attenuate repulsion.

J. Haschke and others, in their article "Reaction of Plutonium Dioxide with Water: Formation and Properties of PuO_2+x," which appeared in Science, Vol. 287, Jan. 14, 2000, reported PuO_2.265 as the excess "oxide" content of the solid without discussing that it exceeded the O = 2.25 limit imposed by Pu₄O₉. This further supports our proposal for Pu₄O₈OH, with the value of 0.265 for the "oxygen" excess over PuO₂ obtained directly as a result of substituting hydroxyl for an oxygen in Pu₄O₉ (that is 17/16 x 0.25 = 0.2656).

Although a central hydroxyl ion can account for the bulk of the increased oxygen uptake, additional detail is needed to account for the spectroscopic observations described in subsequent articles. In the next article, reduced extended x-ray absorption fine structure (EXAFS) spectroscopy data for PuO_2+x highlight the impact of superstructure (structure over multiple-unit cells) and loss of order over multiple-unit cells. These results, taken with our understanding of how extra lattice oxygen incorporates into the uranium oxide, UO₂, as the UO_2+x entity, provide the framework for describing the complex structural manifestations of the water-corrosion reaction.

Conclusions

The short range, local order (structure) of the Pu₄O₉ entity is better described as Pu₄O₈OH. Bond strength-bond length arguments justify the location of a hydroxide ion, rather than a central oxide ion, in a plutonium dioxide structure, providing an alternative interpretation for the experimental data on plutonium dioxide oxidation by water. A central oxide is rejected on the basis of its bond sum. A central oxide requires PuO_2.25, while substitution of hydroxyl for oxide accounts precisely for the experimental value of "excess" oxygen in PuO_2.265, yielding PuO₂(OH)_0.249.

Increasing the valence to plutonium(VI) is similarly rejected on the basis of the sum about the plutonium(VI)-oxygen component. The bond strength sum around the plutonium(V)-oxygen component is normal and supports minimal cell dimension change. However, plutonium(VI)-oxygen at s = 2.0 requires that the seven remaining plutonium(IV)-oxygen bonds increase from s = 0.5 to 0.571, that is (7 x 0.571 + 2.0 = 6), resulting in cell dimensions not present for a plutonium(VI)-oxygen component.

Zachariasen's bond strength-bond length relationships are based on a multitude of actinide structures that support this contention. This approach has found utility for simple structural bond length and bond strength determinations but is limited in structures that display larger deviations, possibly resulting from superstructure disorder.

It has been firmly established that PuO₂ reacts with water at intermediate temperatures, yielding hydrogen gas simultaneously with an increase in the oxygen content of the solid, for which researchers suggested a central oxide ion in Pu₄O₉. The essence of our current work is that a central hydroxyl fits the data better, yielding the formula Pu₄O₈OH, containing one plutonium(V).

This view of the structure with a hydroxide will be shown to be more consistent with XPS and EXAFS data and will be discussed in the following articles. However, the spectroscopic observations call for more work to be done to unequivocally determine the wealth of structural changes manifested when plutonium dioxide is corroded by water.

This article was contributed by Robert Penneman (retired) and Mark Paffett 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