Article pubs.acs.org/IC
Cite This: Inorg. Chem. XXXX, XXX, XXX−XXX
Electrochemical Studies of Selected Lanthanide and Californium Cryptates Matthew L. Marsh,† Frankie D. White,† David S. Meeker,† Carla D. McKinley,† David Dan,† Cayla Van Alstine,† Todd N. Poe,† Danielle L. Gray,‡ David E. Hobart,† and Thomas E. Albrecht-Schmitt*,† †
Department of Chemistry & Biochemistry, Florida State University, Tallahassee, Florida 32306, United States Department of Chemistry, University of Illinois at Urbana−Champaign, Champaign, Illinois 61820, United States
Downloaded via GUILFORD COLG on July 18, 2019 at 11:14:50 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
‡
S Supporting Information *
ABSTRACT: Efforts to quantitatively reduce CfIII → CfII in solution as well as studies of its cyclic voltammetry have been hindered by its scarcity, significant challenges associated with manipulating an unusually intense γ emitter, small reaction scales, the need for nonaqueous solvents, and its radiolytic effects on ligands and solvents. In an effort to overcome these impediments, we report on the stabilization of CfII by encapsulation in 2.2.2cryptand and comparisons with the readily reducible lanthanides, Sm3+, Eu3+, and Yb3+. Cyclic voltammetry measurements suggest that CfIII/II displays electrochemical behavior with characteristics of both SmIII/II and YbIII/II. The °E1/2 values of −1.525 and −1.660 V (vs Fc/Fc+ in tetrahydrofuran (THF)) for [Cf(2.2.2crypt)]3+/2+ and [Sm(2.2.2-crypt)]3+/2+, respectively, are similar. However, the ΔE values upon complexation by 2.2.2-cryptand for CfIII/II more closely parallels YbIII/II with postencapsulation shifts of 705 and 715 mV, respectively, whereas the shift of SmIII/II (520 mV) mirrors that of EuIII/II (524 mV). This suggests more structural similarities between CfII and YbII in solution than with SmII that likely originates from more similar ionic radii and local coordination environments, a supposition that is corroborated by crystallographic and extended X-ray absorption fine structure measurements from other systems. Competitiveion binding experiments between EuIII/II, SmIII/II, and YbIII/II were also performed and show less favorable binding by YbIII/II. Connectivity structures of [Ln(2.2.2-cryptand)(THF)][BPh4]2 (Ln = EuII, SmII) are reported to show the important role that THF plays in these redox reactions.
■
INTRODUCTION
systematic, although nonmonotonic, lowering in energy of 5f orbitals across the actinide series; eventual energy degeneracy of 5f orbitals with filled, ligand orbitals that promote chargeand perhaps electron-transfer from ligands to the metal cations; relativistic reorganization and potential mixing of frontier orbitals (5f, 6p, 6d, 7s, 7p), decreased e−···e− repulsion in 5f versus 4f orbitals leading to less energetic cost for reduction, and specifically at No2+, fully filled 5f14 orbitals. Ostensibly, this leads to the transition from 5fn6d1 (Bk3+ and earlier actinides) → 5fn6d1 ≈ 5fn+16d0 (Cf3+ to Md3+) → 5fn+16d0 (No2+).12−14 However, while this view has some utility as a first approximation, the magnitude of spin−orbit induced splitting/mixing (∼2 eV) of the spinors is so large in the actinide series that the ground states are often both multireference and multiconfigurational.15−18 Notably, the relativistic stabilization of the 7p1/2 spinor, and indirect destabilization of the 6d spinors, becomes so large at
3+
The successful prediction of the elution order of Md (Z = 101) during column chromatography experiments represents both the first and last time a new actinide element’s solution properties were successfully predicted by extrapolating from a lanthanide that was thought to possess an equivalent ionic radius.1 The origin of the subsequent failed prediction of the elution order of No3+ (Z = 102) from α-hydroxybutyrate cation-exchange columns was not understood at the time, but later experiments showed that it originates from the unanticipated fact that No2+, instead of No3+, is the only stable oxidation of nobelium in oxygenated, aqueous media.2,3 The combination of these observations with postcurium electrochemical data and standard heats of formation of the oxides led to the understanding that the An2+ oxidation state becomes increasingly stable starting at americium, becomes metastable from californium through mendelevium, and finally is the only stable state at nobelium.4−11 Even today, the origins of this phenomenon are not fully understood, but some of the underlying factors are the © XXXX American Chemical Society
Received: March 29, 2019
A
DOI: 10.1021/acs.inorgchem.9b00920 Inorg. Chem. XXXX, XXX, XXX−XXX
Article
Inorganic Chemistry
conditions and ligands for preparing isolable CfII complexes using lanthanide analogues as much-needed guides for manipulating isotopes that require significant precautions for safe handling.
lawrencium (Z = 103) that it possesses a different ground state than earlier actinides (5f147p1/21 instead of 5f146d1).19,20 For comparison, these features are not observed in the late lanthanides, and Yb2+ (4f14; °E1/2 = −1.05 V) is not stable with respect to oxidation in water.3 Thus, these electronic factors are not purely academic and have substantial effects on the chemistry and physical properties of the 5f elements that clearly differentiate them from the 4f series. Significant caution is needed when extrapolating from the electronic structure of elements in pure states (either as atoms in the gas phase or as bulk pure substances, i.e., metals), because upon ionization many of these electronic structure effects are lost. This is exemplified quite commonly by differences between first-row and second-row transition metals with classic examples like nickel (Ni:[Ar]4s23d8) versus palladium (Pd:[Kr]4d10). Although the neutral atoms in the gas phase possess different ground states, in solution both are typically found as M2+ ions, and these differences are no longer apparent. However, the metastability of An2+ cations that starts from californium is observed in all states. It is found in the neutral atoms and ions in the gas phase,21,22 manifests again in the metal where both divalent and trivalent metals have been observed,9,23−26 reappears in the standard reduction potential of Cf3+ versus earlier actinides, and is evident again in effective magnetic moments and electronic excitations in the form of reduced magnetic moments and long-lived, ligand-to-metal charge-transfer states in most californium complexes.27−29 While this discussion seems to indicate well-established periodic trends in heavy elements, nothing could be farther from the truth. In fact, few series have been established that extend from earlier actinides to those that lie beyond curium. These examples are largely restricted to halides,30−33 triflates,34−39 cyclopentadienylides,40 dithiocarbamates,29,41 iodates,17,42−44 hydroxypyridonates,18,45 and borates,46 and even in these cases some of the compounds are quite historic and do not even vaguely meet modern standards of characterization. The situation is even more pronounced for complexes containing An2+ cations beyond uranium.47 These are currently limited to a handful of Np 2+ and Pu 2+ organometallic complexes and the binary AnX2 compounds (An = Am, Cf; X = Cl, Br, I).48−50 In short, there are no AmII or CfII molecular complexes for which single-crystal structures have been reported. There is, however, reason to believe that these complexes are isolable, because recent gas-phase experiments have shown that Cf2+ molecules exist and that their properties bear some similarities with those observed from Sm2+ complexes.22 The standard reduction potentials for Cf3+/2+ and Sm3+/2+ are quite similar.51 However, these periodic trends, especially the standard reduction potentials, are almost exclusively obtained from aqueous media, and some of these data again would not withstand contemporary scrutiny. Finally, even though many of these values are available in compendia that long ago gathered dust,52−56 a careful examination of postcurium data shows that much of it is not actually known and is based on rudimentary extrapolations.50,57 To rectify this situation and place the electronic and structural features of AnII complexes on firmer footing we report on the electrochemistry of californium cryptates and compare these results with the readily reducible lanthanides, SmIII, EuIII, and YbIII. Cryptands are good ligands for reductive studies, because they show considerable redox stability. These studies represent early steps in establishing appropriate
■
EXPERIMENTAL SECTION
Caution! 249Cf (t1/2 = 351 y, specif ic activity = 4.1 Ci/g) presents signif icant radiolytic challenges because of a and abundant and unusually energetic γ emission that reaches energies up to 388 keV. The 388 keV line is, in fact, the most abundant line for 249Cf. This requires different biological monitoring than is necessary with most actinide isotopes, choreographing of experiments to minimize exposure times, and signif icant amounts of thick lead shielding of both the samples and the researchers. All manipulations were conducted in a Category II nuclear hazard facility with constantly monitored radiation levels that includes air sampling. All f ree-f lowing solids were handled in glove boxes connected to air f iltration systems. General Procedure. All manipulations were performed in an argon atmosphere with rigorous exclusion of air and water. Anhydrous tetrahydrofuran (THF; Sigma-Aldrich; butylated hydroxytoluene (BHT) as inhibitor ≥99.9%) was stored over sodium metal for 24 h. Acetonitrile was obtained from a pressurized silica column followed by storage over 4 Å molecular sieves for 24 h. All other reagents were purchased from Sigma-Aldrich without any further purification. Synthesis. As detailed further in the Supporting Information, the preparation of [Ln(2.2.2-cryptand)][OTf]3 (Ln = Eu, Sm, Yb) and [Cf(2.2.2-cryptand)][OTf]3 involved equimolar additions of either Ln(OTf)3 or Cf(OTf)3 and 2.2.2-cryptand mixed together in a 6 mL glass vial in THF (Figure S1). [Eu(2.2.2-cryptand)(THF)][BPh4]2 (10.6 mg, 0.026 mmol) and [Sm(2.2.2-cryptand)(THF)][BPh4]2 (19.5 mg, 0.05 mmol) were prepared by dissolving LnI2 in THF. To the stirred solution was added 2 equiv of tetrabutylammonium tetraphenylborate (23.1 mg, 0.05 mmol for Eu, 56.2 mg, 0.1 mmol for Sm). This mixture was stirred for 1 h, to which a 1 equiv solution of 2.2.2-cryptand in THF was added (0.026 mmol for Eu, 0.05 mmol for Sm). For Sm, the mixture turned from blue to dark green and was followed by the precipitation of the product (Figures S2 and S3). The vial was then centrifuged, and the supernatant was removed. The precipitate was then dissolved in 3 mL of acetonitrile, in which the solution turned red (Figure S4). For europium and ytterbium, a similar procedure was followed. [Eu(2.2.2-cryptand)(THF)][BPh4]2 forms a colorless solution, and the ytterbium complex did not precipitate from the original THF solution, but electrochemistry experiments were able to be conducted from THF/CH3CN mixtures. Ultimately, all of the solutions were transferred into a 6 mL glass vial for the electrochemical experiments. As discussed later, YbII forms a different compound under these conditions. Electrochemistry Measurements. Cyclic voltammetry experiments were performed using a CH Instruments 600E potentiostat. A three-electrode configuration was selected for use with a Pt working electrode (2 mm), a Pt auxiliary electrode, and a Ag wire pseudoreference electrode. Frequent working electrode polishing was conducted before and between the experiments, and ferrocene was added at the conclusion to determine an accurate internal reference potential. Additionally, the electrolyte tetrapropylammonium tetrakis[3,5-bis(trifluoromethyl)phenylborate] (or NPr4BArF4 at 1.0 M) was synthesized according to the prior literature and used accordingly.58,59 When we used solvents with particularly low dielectric constants, both IR compensation and increases in electrolyte concentration were implemented. Before each new run, the solvent/electrolyte underwent background scans before generating the complexes in situ. Complexation was determined by shifts in redox potential consistent with the literature. Runs were generally conducted at scan rates between 10 and 1000 mV/s and were expanded from the initial potential windows until degradation of the analyte/electrolyte/solvent. For californium, four different sets of data varying switching potential (larger window), polarity (anodic), and concentration (dilution) were taken to gain a better understanding of these variables. Time stamps were recorded to try and calibrate the B
DOI: 10.1021/acs.inorgchem.9b00920 Inorg. Chem. XXXX, XXX, XXX−XXX
Article
Inorganic Chemistry
among all three Ln species. [Sm(2.2.2-cryptand)]III/II has the second-strongest current transduction. Because of its larger potential, the [Sm(2.2.2-cryptand)]III/II CV has both the largest quasi-reversibility value (696 mV) and Ipc/Ipa ratio (2.64). This demonstrates that, despite greater stabilization of its 2+ state, the [Sm(2.2.2-cryptand)]2+ generated is still relatively unstable and undergoes a partial reoxidation before the return scan. Faster scan rates of 1000 mV/s also did not change this observation. The [Sm(2.2.2-cryptand)]III/II scans also show broad features like that of [Eu(2.2.2-cryptand)]III/II. Finally, features of the [Yb(2.2.2-cryptand]III/II CV depart markedly from those of the two earlier lanthanides. Whereas values for ΔE are nearly the same and the order of quasi-reversibility scales similarly in [Eu(2.2.2-cryptand)]III/II and [Sm(2.2.2-cryptand)]III/II, the ytterbium complex has a larger ΔE (+705 mV) and shows much greater reversibility (90 mV). The origin of the greater stabilization potential perhaps lies in solvent dynamics with the ytterbium ion. Since Yb3+ has a higher charge density relative to the earlier lanthanides, its potential to be stabilized within a cryptand molecule in a relatively nonpolar solvent such as THF probably exceeds similar values to SmII and EuII despite their divalent ions’ preference (Sm2+ = 1.27 Å and Eu2+ = 1.25 Å vs Yb2+ = 1.14 Å)65 with respect to fitting the cryptand diameter (1.42 Å).66 It may also be possible that there is a structural change that affects the mobility of the complex, as observed in other macrocyclic examples.67 If 1:2 adducts of the complex are formed, for example, this might explain the drastic reduction in current transduction observed as well as play a role in the reversibility. Notably, the lack of cathodic decay for this couple would most likely suggest some type of electric double layer capacitive (EDLC) behavior. The buildup of Yb complexes on the surface of the electrode could be plausible considering the combination of its poorer fit in the cryptand, instability in solution, and better reversibility despite its possibly slower diffusion characteristics. Its Ipc/Ipa values being consistently less than unity would also support this conclusion. The cyclic voltammogram of [Cf(2.2.2-cryptand)]III/II is shown to be quite similar to that of [Sm(2.2.2-cryptand)]III/II, as can be expected from their similar standard reduction potentials. The similarity is perhaps most exemplified by almost identical behavior in the reoxidation of the 2+ species. However, the potential shift for CfII (+715 mV) ranges much more closely to YbII (+705 mV). The origin for this greater stabilization in each of these is most likely different. For CfII, the 5f orbitals extend farther out and present opportunities for covalent interactions with 2p orbitals of etheric oxygen atoms on the cryptand molecule. The nonpolar solvent dynamics also favors bonding of this type. The summation of these energetic values amounts to a ΔE = 195 mV stabilization of CfII with respect to its closest electrochemical LnII analogue, SmII, for this solution matrix. Furthermore, values that represent reversible behavior in Table 1 indicate that [Cf(2.2.2cryptand)]III/II has more reversible behavior than [Sm(2.2.2cryptand)]III/II. Further analysis shows that the 2+ species produced for all complexes is kinetically unstable with respect to the trivalent counterpart. For example, a plot of peak current ratios as a function of scan rate (Figure 2) shows that, while peak current ratios remain constant for most of the scan rate range, slower scan rates (
