contents

Modeling Plutonium on Anion Exchange Resins Helps Separation Efficiency

"Docked" configuration of the plutonium hexanitrato dianion and a model bifunc-tional dication (nitrogen atoms in blue).

Anion exchange in nitric acid is a frequently used process for the recovery of plutonium from a wide range of impure materials. Plutonium can be selectively removed from dissolved residues (for example, salt cakes from electrorefining) because the large complexation sphere and high charge/radius ratio of Pu(IV) enables it to form anionic complexes in nitric acid where few metals form competing species. This unique chemical behavior has long been exploited at the Plutonium Facility at Los Alamos National Laboratory where Reillex HPQTM, a macroporous polymer of N-methylated 4-vinylpyridine, is the resin of choice. We would like to improve upon the anion-exchange process for two main reasons: the rate at which the Pu(IV) complex sorbs onto the resin is unusually slow, and we would also like to remove americium, which does not form anionic nitrato complexes as readily as plutonium, from the waste solutions. To this end, we are developing models of the molecular-level interactions of actinide anions with the resin sites. Accurate models will help us to understand the mechanisms of sorption and ultimately allow us to design resins for a variety of anions under a variety of conditions.

Anion-exchange resins are polymers that contain positively-charged (cationic) functional groups bound in the solid matrix. Under typical anion-exchange conditions, sorption of a dianionic species requires complexation to two separate cationic resin sites. The orientation and availability of these sites for cooperative interactions with a dianion is not well controlled. Spectroscopic studies suggest that plutonium sorption onto the resin may occur via a process in which an uncharged tetranitrato complex in solution is converted to a dianionic hexanitrato complex at the resin surface, acquiring two nitrate groups in the process. We hypothesized that a resin that could facilitate the uptake process, for example by positioning the two nitrate groups in the proper configuration, could provide superior binding properties and selectivity for plutonium nitrato complexes, exhibit enhanced kinetics for plutonium uptake from solution, and encourage the formation of weaker nitrato complexes.

To test our "facilitated uptake" hypothesis, we synthesized and evaluated a series of bifunctional resins in which the structure of the anion-receptor site is well-defined; the two anion-exchange sites are separated by a fixed distance in a fixed orientation. These resins are synthesized via modification of poly(4-vinylpyridine) resins with a second cationic site such that the two anion-exchange sites are linked by "spacer" arms of varying length and flexibility. Plutonium sorption data from nitric acid media indicate that this controlled geometry of the anion-exchange sites has a positive impact upon the sorption of the plutonium dianion. Most notably, a "spacer" length of 4-5 methylene units generally provides the best plutonium uptake conditions regardless of the functionality of the second cationic site.

Modeling the dianion/dication interactions is a complex, multi-stage process. The first step of this modeling is a proper description of the actinide complex. We recently developed refined MM2 parameters for Pu(IV), U(IV), Np(IV) and Th(IV) hexanitrato complexes. Ideally, we like to describe actinide complexes with as few parameters as possible, which allows for faster structural optimization times and/or the ability to model larger systems in the same amount of time. Our parameters are optimized to allow maximum structural flexibility for the dianion in order to determine what structural distortions may occur upon "docking" of the dianion with cationic sites. For this work, we used solution EXAFS (extended x-ray absorption fine structure) data aquired by LANL researchers to establish initial structural parameters for the plutonium hexanitrato dianion. Estimated van der Waals radii were used to extend the model to the analogous thorium, uranium, and neptunium complexes. The excellent agreement between the models and experimental single-crystal x-ray structures of the four complexes gives us some confidence in the efficacy of this method, and we plan to extend our modeling to other systems that lack solid-state structural data.

Determination of the distribution of charges for each atom of the dianion and dication is critical to the calculation of their electrostatic attraction. The simplest method is to use "formal charges" for all the atoms. However, formal charges are really just a way to keep track of the total charge and do not accurately reflect the actual charge distribution of a complex molecule. To calculate the partial-charge distribution for the large, unwieldy plutonium hexanitrato dianion, we had to use a theoretical neutral "compound," the triradical Pu(NO3). To model the bifunctional resin sites, we replace the polyvinylpyridine "backbone" with a pyridine molecule. There are many different ways to calculate partial charge distribution for organic molecules, all providing very different answers. To some extent, we rely on "chemical intuition" and internal consistancy to help us decide which method provides the best charge metrics.

Once we decide what set of charges to use, we determine the optimized anion/cation configuration for each of the "docked" ion pairs using molecular mechanics and calculate the net force by summing the attractive (opposite charges) and replusive (same charges) forces between the anion/cation pair. This "stickyness factor" (SF) is then correlated with the experimental plutonium distribution coefficients (Kd) found for the corresponding resin. We have found that the strongest (most attractive) SF tends to correspond to the highest experimental Kd. Using formal-charge metrics, the models accurately predict that a 4-5 atom "spacer" between the cationic sites is the best for complexation of the plutonium hexanitrato dianion. We are currently refining our models to incorporate more chemically realistic partial charges. Our models cannot yet predict a Kd for a specific system, but they can determine trends within a system, and we would like to develop a method for a priori prediction of distribution coefficients.

The principal developers of this project are Mary E. Barr, Eddie Moody, and Gordon Jarvinen (NMT-6); S. Fredric Marsh (NMT-6 emeritus); and Richard A. Bartsch (Texas Tech 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 1996 - Disclaimer 26 June 1996