Room-temperature ionic liquids

Unusual solvent system shows potential to revolutionize separation and purification technologies

Room-temperature ionic liquids (RTILs) represent an intriguing and highly unusual solvent system in which to carry out basic research and develop advanced separation and purification technologies for the actinide elements. The ionic liquids themselves are the subject of intense worldwide interest, as much for their curious properties as for their potential in revolutionizing chemical synthesis, catalysis, separations, and electrochemistry.

Part of the optimism surrounding ionic liquids rests on their improved safety characteristics as compared to volatile organic solvents and their ability to solubilize both nonpolar organic solutes as well as highly charged inorganic ions. In many cases ionic liquids combine the solvating properties of polar solvents such as water with those of non-polar organic solvents such as benzene.

The ionic liquids are characterized as having negligible vapor pressure, thereby eliminating health effects through airborne exposure, and significantly reducing flammability hazards compared to traditional organic solvents. Additionally, these solvent systems offer a novel chemical environment that may uniquely influence the course of chemical reactions as compared to traditional molecular solvents.

By way of definition, the ionic liquids used in our work are low-melting organic salts, distinct from other classes of solvents in that they are composed of separate and discrete organic cations and weakly coordinating anions. These solvents are nonproton donating and are not related to the more familiar, self-ionizing solvents such as water (2 H₂O = H₃O⁺ + OH^–; ionic product = 10^–14), anhydrous hydrofluoric acid, or the various "super acids." Classical inorganic salts consist of infinite three-dimensional arrays of close packed spherical ions and are characteristically brittle solids with high melting points (for example, the melting point of NaCl is 801 °C). Low-melting ionic liquids are obtained through two molecular design principles. First, coulombic interactions are minimized by diffusing charge over several atoms in a molecule. In the case of ionic liquid cations, the 1,3-dialkylimidazolium unit (1) is a common example that spreads the positive charge over the five atoms of the heterocyclic ring. Weakly coordinating anions delocalize charge through induction to highly electronegative fluorine atoms or by resonance as shown explicitly for the –N(SO₂CF₃)₂ anion. More-complex and higher molecular weight cations and anions have also been explored, but the greater number of intermolecular contacts serve to raise melting points and increase the viscosity in such systems. For convenience in routine use as solvents, lower molecular weight constituents are preferred. The second design principle in ionic liquid synthesis is molecular asymmetry. All things being equal, low-symmetry cations and anions reduce packing efficiency in the crystalline state and lower melting points. A dramatic example is provided by the comparison of the PF₆– and –N(SO₂CF₃)₂ salts of cation 1, which melt at 58 °C and –3 °C, respectively.

In many respects, RTILs are most closely related to classical molten salts. In theory, some 10 trillion different cation/anion combinations can be used to produce binary ionic liquid systems. This flexibility affords the potential to tune or "design" the ionic liquid solvent for specific applications. In practice, the requirements for low viscosity and ease (low cost) of synthesis have limited the choice of cation to relatively low molecular weight organic salts complemented by weakly coordinating inorganic anions.

Recent work has employed imidazolium or quaternary ammonium salts of the bis(trifluoromethanesulfonyl)imide anion (N(SO₂CF₃)₂ abbreviated as –NTf₂) because they are easily prepared, chemically and thermally robust, and yield low-viscosity/high-conductivity anhydrous fluids. Because of their good solubilization characteristics and wide electrochemical window, these solvents are an ideal choice for studying the basic electrochemical behavior of soluble actinide complexes with an eye toward developing redox-based separations technology. Our research has shown that high-quality electrochemical data can be obtained in these ionic liquids on par with what is possible in conventional electrochemical solution media.

We have also focused on understanding the solvent (i.e., coordinating) properties of these ionic liquids. Using a combination of Lewis-acid- and Lewis-base-specific solvatochromic probe molecules that change color as a function of chemical environment, we find that the Lewis-acid properties of the ionic liquids depend on the organic cation and are generally comparable to aliphatic alcohols such as ethanol or butanol. The Lewis basicity, determined by the –NTf₂ anion, lies between very weakly coordinating solvents such as nitromethane and more-coordinating solvents such as THF.

The discovery that ionic liquids are simultaneously weakly Lewis acidic and weakly Lewis basic is consistent with the fact that these salts are liquids at room temperature. This observation indicates that they are characterized by very low lattice energies determined by weak coulombic interactions between the cations and anions.

Salts composed of stronger Lewis acids/bases are more likely to exist as solids at room temperature. These studies highlight the differences between ionic liquids and classical high-temperature molten salts. The solvating properties of the ionic liquids more closely resemble weakly coordinating organic solvents, with the distinct difference that they are highly conductive and the ionic constituents can move independently to stabilize charged solutes.

A comparison in the behavior of two simple U(IV) chloride complexes illustrates the unique chemistry observed in ionic liquids as compared with traditional molecular solvents. Tetrabutylammonium salts of U(IV) hexachloride, [NBu₄]₂UCl₆, readily dissolve in the ionic liquid solvent to give pale blue solutions with a solubility limit of about 0.1 M. Probing the solutions via cyclic voltammetry experiments indicates that reduction of the U(IV) complex to U(III) or oxidation to U(V) is chemically reversible and that the six chloride ligands remain tightly bound to the metal center throughout these electron-transfer reactions.

In a similar fashion, the coordinatively unsaturated uranium tetrachloride complex, UCl₄, also dissolves with gentle warming to give deep-green solutions with a solubility limit of about 0.4 M at room temperature. It is known that UCl₄ dissolves in coordinating organic solvents such as acetonitrile or THF to give eight coordinate adducts of the form, UCl₄(L)₄ (L = donor solvent). In contrast, we have discovered that when dissolved in the ionic liquid, UCl₄ undergoes a facile chloride redistribution reaction to give mixtures of [UCl₆]^–2 and additional species stabilized by five or fewer chloride ions. Well-formed x-ray-quality single crystals of [cation]2UCl₆ readily separate from these solutions, where the cation identified in each case is derived from the particular ionic liquid used in the experiment.

The –NTf₂ anion provides a range of coordination modes to accommodate the particular electronic and steric preferences of a metal ion. Softer metals, represented by low-valent transition metal complexes, display coordination to the nitrogen atom as h1-N or h₂-N,O. High-valent transition metal and actinide complexes prefer oxygen coordination as h1-O or h₂-O,O, depending on the availability of one or two open coordination sites.

Ligand redistribution of this type is surprisingly common in actinide coordination chemistry, driven by the stabilization afforded by maximizing steric saturation. For example, while UCl₄(DMSO)₃ appears to be a reasonable complex, in fact this system exists as a complex salt of the form [UCl₂(DMSO)₆][UCl₆] (DMSO = dimethylsulfoxide). In –NTf₂-based ionic liquid systems, coordinately unsaturated uranium species are no doubt stabilized by interactions with the anion. We have shown that these interactions are weak by adding two equivalents of chloride ion (as the tetrabutylammonium salt) to yield pure [UCl₆]^–2 solutions.

Given the unexpected complexity of the chloride system and the implication that the –NTf₂ anion is playing an active role in the coordination chemistry of unsaturated metal species, we have set out to document the fundamental binding properties of this anion and to help understand how these interactions determine resultant redox properties.

A series of coordinatively unsaturated metal species was generated either through the reaction of metal alkyl complexes with HNTf₂ or through reaction of metal halide complexes with AgNTf₂, as illustrated for the (C₅H₄Me)₃U fragment shown in the following equation.

Single-crystal x-ray diffraction experiments for a number of different metal complexes revealed that four different binding modes were possible for the –NTf2 anion. The anion can bind in a monodentate fashion either through the nitrogen atom or through one of the sulfonyl oxygen atoms. For less sterically constrained metal species, two different bidentate binding modes were also identified, either via two oxygen atoms from each end of the bis(sulfonyl)imide moiety or via a more acute nitrogen-oxygen interaction. Electronically the –NTf2 anion behaves as a weakly coordinating anion that donates limited electron density to the metal center. The redox consequences of –NTf2 ligation are clearly illustrated by comparison of the 4+/3+ redox couples for (C5H5)2TiCl2 and (C5H5)2Ti(NTf2)2. While the titanium dichloride system is reversibly reduced at about –1.43 V versus Ag/Ag+ reference electrode, the (C5H5)2Ti(NTf2)2 complex is reduced at –0.49 V, a stabilization of 0.94 V. These are very large effects that suggest the potential to significantly stabilize low-valence metal species in the ionic liquid medium. The possibility of directly electroplating very active actinide metals is also suggested by the stabilization of low-valence complexes. We recognized these observations as an opportunity to directly study the simplest of metal systems by introducing “bare” metal ions into the ionic liquid solvent in the absence of any competing ligands. In the case of uranium, we found that reaction of UCl4 with four equivalents of AgNTf2 or anodic oxidation of a uranium electrode immersed in the ionic liquid solvent produced deep green U(IV) solutions that were identical when characterized by UV/vis spectroscopy. These results suggest that the U(IV) ion is likely stabilized by ligation to four –NTf2 anions, each coordinated in a bidentate fashion through the oxygen atoms. Such a binding mode is reminiscent of well-characterized actinide ß-diketonate complexes.

Similar to the titanium complex, the effect of removing the chloride ligands from the U(IV) metal center is profound. In the case of [UCl6]–2 the 4+/3+ reduction potential is measured at –1.98 V versus Ag/Ag+, while for “U(NTf2)4” this reduction occurs at –0.24V, a stabilization of 1.74 V. Additionally, despite the extreme electrophilicity (strongly electron-attracting) indicated by this shift in reduction potential, no evidence was found for decomposition of the ionic liquid in the presence of such a potentially reactive U(IV) ion. The behavior of the electrochemically produced U(III) ions at more negative electrode potentials is under active investigation. The study of actinide elements in ionic liquid solvents presents a number of challenges in the areas of product isolation and structural characterization. These solvents do, however, allow facile spectroscopic and electrochemical characterization opportunities for the study of novel actinide coordination chemistry. The study of the actinides in ionic liquids is therefore complementary to parallel investigations in aqueous and organic solvents, as well as high-temperature molten salts. Ultimately, the prospect of preparing new classes of compounds and discovering new reactivity patterns that might be exploited in advanced separations technology guides this work. Given the complementary roles of aqueous/organic solvent extraction, ion exchange chromatography, and high-tempera-ture molten salt electrorefining in current actinide processing and purification practice, we believe that ionic liquids might be used to develop new technologies that combine these disparate tasks into one integrated unit operation with possible advantages in production efficiency, safety, and space requirements.

U(IV) cations can be introduced into the ionic liquid solvent without the complication of additional halide anions by electrochemically dissolving a sacrificial uranium anode in a two-compartment cell. An electrochemically etched uranium anode produced by this process is shown in the photo. Spectroscopic characterization of the resulting green solutions indicate that U(IV) is the dominant oxidation state. The uranium ions are likely to be stabilized by the –NTf2 anions that compose the ionic liquid and the proposed structure of U(NTf2)4 is indicated in the top right. A comparison of the electrochemical behavior of [UCl6]2– and U(NTf2)4 is shown in the lower right. The 4+/3+ reduction potential is increased by 1.74 V upon elimination of the six chloride anions from the metal’s coordination sphere. This article was contributed by Los Alamos researchers Warren J. Oldham of the Chemistry Division and Michael E. Stoll and David A. Costa of the Nuclear Materials Technology 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