Background

Zeolites and related nanoporous materials play a large role in the chemical and environmental industries. Nearly all the gasoline in the world has been produced using zeolite catalysis. These materials also are widely used in the separation of organic materials and gases, they play a large role in water purification, and are used extensively in ion exchange applications. One of the unique features of these materials is that they have nanometer-sized pores and cavities that allow shape-selective separations to take place. They are ideal for ion exchange applications because the cations are in intimate contact with the crystal lattice and extremely high selectivities can be obtained by judicious choice of the framework.

We are interested in using atomistic simulations in concert with our collaborators experimental and synthesis studies to gain a fundamental understanding of the sorption, diffusion and ion exchange behavior of these materials. In essence the question we seek to answer is: How does the  chemical composition and pore structure affect selectivity, reactivity, and transport through the pores? To answer this,we develop and apply Monte Carlo and molecular dynamics simulations methods and simulate these processes at the atomistic level.

Our current focus is on understanding ion exchange in titanosilicates and polyoxometallate compounds. These materials are being made and tested experimentally by our collaborators May Nyman (Sandia National Lab), Abe Clearfield (Texas A&M) and David Hobbs (Savannah River National Lab). Jim Larentzos and Craig Powers have worked on this in our group.

               Jim Larentzos                                               Craig Powers

Simulations

The figure below shows the primary building unit for these-called CST (crystalline silicotitanate) materials and a periodic cell used in our simulations.

The movie below shows a short piece of a grand canonical Monte Carlo simulation of the hydrogen form of CST and water. The partial pressure of water is gradually increased, and you can observe the water loading increase. Notice that initially water associates strongly with the inner pore surfaces, coordinating with the lattice oxygen atoms. At a certain water partial pressure, the water begins to form hydrogen-bonded clusters in the middle of the pores, and the water loading increases rapidly.

These calculations enable us to track the location of water and ions inside the porous materials to better understand how they associate with each other and the framework. For example, the figures below show the computed positions of water in two different crystallographic sites of H-CST. The figure on the right shows the experimentally determined sites referred to as OW1 (the outer sites closest to the pore walls, and the ones that filled first in the movie above) and OW2 (the sites in the middle of the pores). The figure on the left shows a probability distribution map of water in this material along a pore. The red regions are the highest intensity, and the white spheres are the superimposed experimental sites. The simulations show that the OW1 sites are highly localized and lie right on top of the experimental sites. The simulations predict that the OW2 sites, on the other hand, are highly mobile and there is much interchange between the four distinct sites observed in X-ray crystallography.

Sorption isotherms can be computed from the simulations. The figure below shows the isotherm for water. The saturation loading of water 12 molecules per unit cell is in excellent agreement with experiment.

                                        Water isotherm for H-CST

Applications

Along with our experimental colleagues, we are investigating the use of CST and other materials to remediate radioactive nuclear waste. The figure below is taken from a 2002  article in National Geographic, and shows some of the underground storage tanks at Hanford, Washington that contain large amounts of liquid radioactive waste. The high level waste must be separated from lower level waste and then processed for eventual storage in long-term repositories such as Yucca Mountain.

                                   Tank waste storage at Hanford, WA.

One strategy for performing this separation is to use ion exchange to selectively remove Cs+, Sr2+ and actinides from the tank wastes, while leaving behind other ions such as Na+. This is challenging, because sodium concentrations as high as 6M have been observed, while radioactive ions are at much lower concentrations. This means extremely selective ion exchangers are needed.

The Clearfield group has found that niobium-substituted CST materials are extremely selective for Cs+. We performed simulations on the Nb-substituted CST (lower left) and the ordinary CST (lower right) and found that the Nb-substituted material enabled more water to hydrate the Cs+ ion (shown in green) than was possible with the ordinary CST material. This extra water hydration significantly stabilizes the cesium, thereby leading to greater selectivity. These findings were in excellent agreement with experimental results.

Snapshots of cesium and water inside the pores of the Nb-substituted CST (left) and ordinary CST (right).

Another material we have been investigating is a so-called Keggin-chain or polyoxometallate material. The highly-charged dodecaniobate Keggin ions, [XNb12O40]-16 (X = Si, Ge) and [XNb12O40]-15 (X = P) serves as building blocks to self-assembled, low-dimensional anionic framework materials.  In addition to its high charge, the Keggin ion provides optimal binding geometries which render these materials attractive metal sorbents and ion exchangers. May Nyman and coleagues have made some novel forms of these materials (see Nyman, M.; Bonhomme, F.; Alam, T. M.; Rodriguez, M. A.; Cherry, B. R.; Krumhansl, J. L.; Nenoff, T. M.; Sattler, A. M. Science 2002, 297, 996-998; Nyman, M.; Celestian, A. J.; Parise, J. B.; Holland, G. P.; Alam, T. M. Inorg. Chem. 2006, 45, 1043-1052.)

The figure below shows a poyhedral representation of the sodium form of this material along [001]. The purple TiO6 octahedra and yellow SiO4 tetrahedra are shown, along with sodium cations (pink spheres) and water (red spheres = O and white spheres = H). The ion positions were determined from a simulation.

The computed probability distribution plot of the Na+ ions in this material along the [001] direction is shown below. The simulations found that there are found unique sites for Na+, in agreement with experiment. However, the simulations predict that the cations are highly mobile.

              Probability distribution plot of Na+ ions in the Na-Keggin material.

By carrying out simulations such as these, we are learning how the cations and water associate with the different frameworks and the role substitutions play in the binding strength of ions. We have also developed predictive force fields that enable us to simulate these complex materials. Armed with these methods, we are seeking to perform direct ion exchange calculations to predict selectivities.

To learn more about this work, see:

James P. Larentzos, Abraham Clearfield, Akilesh Tripathi and Edward J. Maginn, “A Molecular Modeling Investigation of Cation and Water Siting in Crystalline Silicotitanates”, Journal of Physical Chemistry B, 2004, 108, 17560-17570.

François Bonhomme, James P. Larentzos, Todd M. Alam, Edward J. Maginn and May Nyman, “Synthesis, Structural Characterization, and Molecular Modeling of Dodecaniobate Keggin Chain Materials”, Inorganic Chemistry, 2005, 44, 1774-1785.

May Nyman, James P. Larentzos, Edward J. Maginn, Margaret E. Welk, David Ingersoll, Hyunsoo Park, John B. Parise, Ivor Bull, and François Bonhomme, “Experimental and Theoretical Methods to Investigate Extraframework Species in a Layered Material of Dodecaniobate Anions”, Inorganic Chemistry, 2007, 46(6) pp 2067 – 2079.

James Larentzos, William F. Schneider and Edward J. Maginn, “A Transferable Force Field for Water Adsorption in Cation Exchanged Titanosilicates”, Industrial and Engineering Chemistry Research, 2007, 46, 5754-5765.