Article pubs.acs.org/JACS
Covalency in Americium(III) Hexachloride Justin N. Cross,†,§ Jing Su,†,§ Enrique R. Batista,*,† Samantha K. Cary,† William J. Evans,‡ Stosh A. Kozimor,*,† Veronika Mocko,† Brian L. Scott,† Benjamin W. Stein,† Cory J. Windorff,†,‡ and Ping Yang*,† †
Los Alamos National Laboratory, Los Alamos, New Mexico 87545, United States University of California, Irvine, California 92697-2025, United States
‡
S Supporting Information *
ABSTRACT: Developing a better understanding of covalency (or orbital mixing) is of fundamental importance. Covalency occupies a central role in directing chemical and physical properties for almost any given compound or material. Hence, the concept of covalency has potential to generate broad and substantial scientific advances, ranging from biological applications to condensed matter physics. Given the importance of orbital mixing combined with the difficultly in measuring covalency, estimating or inferring covalency often leads to fiery debate. Consider the 60-year controversy sparked by Seaborg and co-workers (Diamond, R. M.; Street, K., Jr.; Seaborg, G. T. J. Am. Chem. Soc. 1954, 76, 1461) when it was proposed that covalency from 5f-orbitals contributed to the unique behavior of americium in chloride matrixes. Herein, we describe the use of ligand K-edge X-ray absorption spectroscopy (XAS) and electronic structure calculations to quantify the extent of covalent bonding inarguablyone of the most difficult systems to study, the Am−Cl interaction within AmCl63−. We observed both 5f- and 6dorbital mixing with the Cl-3p orbitals; however, contributions from the 6d-orbitals were more substantial. Comparisons with the isoelectronic EuCl63− indicated that the amount of Cl 3p-mixing with EuIII 5d-orbitals was similar to that observed with the AmIII 6d-orbitals. Meanwhile, the results confirmed Seaborg’s 1954 hypothesis that AmIII 5forbital covalency was more substantial than 4f-orbital mixing for EuIII.
■
INTRODUCTION Covalency in bonding is one of the most important and fundamental concepts used to explain phenomena in the physical, chemical, and biological sciences. These range from understanding magnetic and superconducting materials1−3 to applications in controlling catalysis,4−8 developing biological and bioinspired processes,9−11 and designing new chemical separations.12−17 While great progress has been made in understanding the role of covalency in these areas,18,19 it has been historically difficult to characterize covalent bonding for the f-elements.20−31 This is especially true for actinides with atomic numbers > 92, i.e., heavier than uranium.32−41 In large part, this stems from difficulties in acquiring sufficient material to study. It is additionally difficult to handle the highly radioactive transuranic analytes, which require formidable capital investment to establish facilities with appropriate safety and engineering controls. On top of these technical challenges are issues associated with actually measuring covalency as well as computational problems associated with modeling actinide orbital interactions while accounting for the significant spin− orbit coupling, the large number of electrons, and the emergence of both 5f- and 6d-orbital contributions to bonding.42,43 Since announcing the discovery of plutonium in 1940, covalency in actinide−ligand bonding has been typically inferred, rather than directly characterized. This practice dates © 2017 American Chemical Society
back to the early 1950s, when Seaborg and co-workers rationalized americium behavior on an ion-exchange resin as evidence of covalent mixing between the Am 5f- and Cl 3porbitals.44 Subsequently, the nodal properties of the 5f-orbitals led Streitwieser and colleagues to predict the existence of actinocenes.45−48 More recently, 5f- and 6d-orbital participation in bonding has been used to explain the existence and stability of many actinide functional groups, i.e. actinide−ligand multiple bonds,20,49−67 actinide−arene delta interactions,24,68−77 and molecular actinide−metal bonds.78−89 The invocation of orbital mixing in these seminal publications motivated our efforts to directly evaluate orbital mixing in an americium−ligand bond. Given the recent advances in spectroscopic technologies and quantum chemical theory, especially the emergence of ligand Kedge X-ray absorption spectroscopy (XAS) and density functional theory (DFT) methods,6,18,21,59,69,90−100 it is now possible to directly interrogate covalency in actinide compounds. Herein, we report the first covalency study on a transplutonium element using ligand K edge-XAS and DFT. In these studies, Am 5f- versus 6d-orbital mixing within the homoleptic AmCl63− anion was evaluated and compared with that of the isoelectronic Eu−Cl interactions in EuCl63−. Using orbital descriptions, metal ligand orbital mixing coefficients Received: April 17, 2017 Published: June 14, 2017 8667
DOI: 10.1021/jacs.7b03755 J. Am. Chem. Soc. 2017, 139, 8667−8677
Article
Journal of the American Chemical Society
of diethyl ether, Et2O, into the MeCN solution (Scheme 2, Figure 1). After crystallization and subsequent characterization
were experimentally and computationally characterized through dipole allowed electronic transitions between Cl 1s-orbitals and virtual (unoccupied) orbitals that resulted from mixing between the Cl 3p- and metal valence orbitals (Scheme 1). Because the Scheme 1. Cartoon Showing Electronic Excitations in a Ligand K-Edge XAS Experiment
Figure 1. Thermal ellipsoid plot of (PPh4)3AmCl6 with ellipsoids drawn at 50% probability. PPh41+ cations have been omitted.
of the AmCl63− sample, care was taken in recovering the 243Am by cation exchange chromatography for future studies using a previously reported method.103 We cannot overstate the importance of establishing effective recycling strategies when conducting fundamental studies on rare and valuable isotopes, like 243Am, t1/2 = 7370(40)y.104 Characterization of (PPh4)3AmCl6. This (PPh4)3AmCl6 compound represented the first homoleptic americium chloride compound structurally characterized by single crystal X-ray diffraction (SCXRD). The hexahalide crystallized in the space group P21 and was isomorphous with the previously reported (PPh4)3CeCl6.26 Around the AmIII ion were six Cl1− ligands that formed a slightly distorted octahedron, such that the AmCl63− anion approached Oh-symmetry. All six Cl1− ligands were crystallographically unique with bond distances ranging from 2.713(3) to 2.752(3) Å and adjacent Cl−Am−Cl angles ranging from 87.67(11) to 92.04(9)°. The average 2.724(2) Å Am−Cl distance (error reported as standard error of the mean) was bracketed between the 2.708(4) and 2.737(7) Å distances in (NEt4)3SmCl6 and (NEt4)3NdCl6, respectively.26 Although these values were consistent with expectations based on six coordinate ionic radii105 (see Figure 2), the Am−Cl distances were significantly shorter than other reported Am−Cl bond lengths. For example, the average Am−Cl distance in AmCl63− was shorter than the SCXRD measurement made on AmCl2(H2O)61+ (2.799(2) Å),106 and the extended X-ray absorption near edge structure (EXAFS) measurement made
core hole was localized on the absorbing Cl1− ligand, the dominant contribution to the transition dipole moment governing the intensity of these electronic transitions was directly related to the amount of Cl 3p-character in the final state.94 Hence, the Cl K-edge XAS experiment provided a direct gauge of covalency in M−Cl bonding (as defined by Heitler and London101). To directly interrogate orbital mixing for transuranic elements, we have developed the first synthesis of tetraphenylphosphonium americium(III) hexachloride, (PPh4)3AmCl6. This compound was characterized by single crystal X-ray diffraction (SCXRD) and electronic absorption spectroscopy (UV−vis−NIR). Moreover, we developed a robust method for making XAS measurements in the tender X-ray regime (2−5 keV) on this highly radioactive sample. Our approach enabled AmCl63− to be characterized at the Am M-edge and at the Cl Kedge. Spectral interpretations were guided by DFT and transition-dipole moment excited calculations that incorporated spin−orbit coupling contributions. Overall, we observed that the Am−Cl interactions were slightly more covalent than the analogous Eu−Cl bonds. The results are discussed in the context of previous indirect analyses of actinide covalency.
■
RESULTS AND DISCUSSION Synthesis of (PPh4)3MCl6 (M = Eu, Am). Single crystals of (PPh4)3AmCl6 were prepared similarly to the recently reported lanthanide and actinide analogues, (PPh 4 ) 3 CeCl 6 and (PPh4)2AnCl6 (An = Th, U, Np, Pu), Scheme 2.102,26 The Scheme 2. Synthesis of (PPh4)3AmCl6
synthesis commenced by generating an AmIIIaq stock solution by dissolving americium(IV) dioxide (AmO2) in hydrochloric acid (HCl; 6 M; Scheme 2). Then, the solution was heated until a solid AmIII residue was obtained, formulated as AmClx(H2O)y·(3 − x)Cl. The solid was suspended in hot MeCN. While maintaining the temperature near 70 °C, addition of three equivalents of tetraphenyl phosphonium choloride, PPh4Cl, almost instantaneously caused the solids to dissolve. From this solution, large pale-pink crystals suitable for single crystal X-ray diffraction were obtained by slow diffusion
Figure 2. Plot of average Z3MCl63− M−Cl bond distance versus ionic radii (Shannon, CN = 6).105 (For M = Pr,130 Z = Ph3PH1+. For Nd,26 Sm,26 Eu,26 and Gd,26 Z = NEt41+. For Ce26 and Am, Z = PPh41+.) 8668
DOI: 10.1021/jacs.7b03755 J. Am. Chem. Soc. 2017, 139, 8667−8677
Article
Journal of the American Chemical Society on AmIII dissolved in aqueous hydrochloric acid solutions (2.8 Å).107,108 We rationalized these longer Am−Cl distances as resulting from a combination of increased coordination number, extensive hydrogen-bonding with the outer-sphere water molecules, and (in the solution-phase measurements) molecular dynamics.109 The Am−Cl distances in AmCl63− were also appreciably shorter (by 0.16 to 0.09 Å) than the two other Am−Cl distances measured to date by SCXRD. These included the Am−Cl distances in AmCl3 where all the chlorides are bridging [Am−Clmean = 2.888(6) Å],110 and the terminal chlorides of Am4[B16O26(OH)4(H2O)3Cl4] [Am−Clmean = 2.81(5) Å], whose higher coordination number leads to a larger ionic radius.111 Variations in the Am−Cl bond distances were not surprising since similar trends reported were for the size matched Nd3+ (ionic radii for six coordinate = 0.975 and 0.983 Å for AmIII and NdIII, respectively). For example, the Nd−Clmean distance in NdCl63− was 0.16 to 0.08 Å shorter than in NdCl3 and NdCl2(H2O)61+.112,113 The UV−vis−NIR absorbance spectrum collected on single crystals of (PPh4)3AmCl6 was dominated by an intense absorption band at high energy (>28 570 cm−1), Figure 3.
low analyte concentrations. Owing to radiological hazards, the AmIII samples were triply contained within three-nested aluminum holders equipped with thin polypropylene windows (4 μm). Prior to conducting the XAS measurements, it was determined that these windows would not fail on beamlines 4− 3 and 14−3 at the Stanford Synchrotron Radiation Lightsource (SSRL) when exposed to X-ray radiation (2400−5000 eV) for 24 h. The samples were placed in a LANL fabricated sample chamber that was separated from the beam pipe by a small air gap (∼0.5 cm). At the furthest point upstream the LANL chamber was sealed with a beryllium window. The chamber was additionally equipped with HEPA filters on gas inlet valves through which helium gas was continually flowed. Downstream from the window was an ion chamber (30 cm) used to monitor the incident radiation (I0). The sample was positioned at 45° from the incident radiation and the sample fluorescence was monitored against I0 using a partially depleted series (PIPS) charged particle detector (see Experimental Section). To confirm the presence of AmIII in the Cl K-edge XAS samples, Am M5,4-edge XAS measurements were made prior to collecting the Cl K-edge XAS data. This AmCl63− spectrum was background-subtracted and normalized (Figure 4) and the
Figure 4. Am M5,4-edge XAS data from (PPh4)3AmCl6.
Figure 3. UV−vis−NIR absorption spectrum from single crystals of (PPh4)3AmCl6 suspended in low absorptivity immersion oil using a microspectrophotometer.
results provided the first XAS data on an americium molecular complex. As far as we are aware, only one other Am M5,4-edge spectrum has been reported from the AmFe2 intermetallic.116 From the perspective of the free ion, peaks at the Am M5,4-edge emerge from electric dipole allowed transitions from Am 3dorbitals to unoccupied states that contain Am 5f-character, e.g., 3d10... 5f6 6d0 → 3d9... 5f 7 6d0. Similar to the published AmFe2 spectrum, the spectrum from AmCl63− contained a pronounced M5-edge peak at 3888.4 eV whose intensity overshadowed that of the small M4-edge peak at 4092.4 eV (energy calibrated externally to the first inflection point of the potassium edge of KCl; 3608.4 eV). The inflection points for the AmIII M5 and M4 features were at 3,886.9 and 4,090.8 eV, respectively. Another important metric characteristic for M5,4-edge data is the intensity ratio. We determined this ratio using a graphical approach based on the integration of the second-derivative spectrum and defined as IM5/(IM5 + IM4), where IM5 and IM4 represent the total areas under the second derivative of the M5 and M4-edge peak intensity ratios, respectively.117 The branching ratio from AmCl63− was 0.79(2), similar to the 0.88 value reported for AmFe2. To directly compare electronic structure and covalent M−Cl bonding in the MCl63− anions, the background subtracted and normalized Cl K-edge XAS spectra from MCl63− (M = Am, Eu) were overlaid in Figure 5. Data presented here from EuCl63− reproduced previous measurements,26 however; the new data was collected under the exact experimental conditions
Superimposed on this charge transfer band were Laporte forbidden 5f → 5f transitions, characteristic of AmIII.114,115 The energies for these transitions were similar to that previously reported on AmCl3.114 In accordance with established interpretations,114,115,103 we describe the AmCl63− absorption spectra from the perspective of the free ion. Hence, spectral features were attributed to excitations from the AmIII 7F0′ ground state to 5L6′ (19 833 cm−1), 7F5′ (12 267 cm−1), and 7 F4′ (9721 cm−1) excited states. We note the term symbols include a prime mark (′) as an indicator of the intermediate coupling schemes, where substantial j−j coupling renders orbital and spin angular momentum eigenvalues (L and S) as no longer appropriate quantum numbers.114,115 Single crystals of (PPh4)AmCl6 did not fluorescence within our spectral window (400−900 nm) upon excitation at 365 or 420 nm. XAS Evaluations of (PPh4)3AmCl6. To evaluate AmIII versus EuIII orbital mixing, MCl63− (M = Am, Eu26) samples were analyzed by X-ray absorption spectroscopy (XAS). Samples were prepared as thin films of fine powders that had been lightly dispersed on tape. Previous screening by Cl K-edge XAS revealed the amount of chlorine in the tape was low and would not interfere with subsequent Cl K-edge XAS measurements. Additionally, using EuCl63−, we established that contributions from self-absorption were minimized (200 μm) of the low Cl tape prevented transmission to the Kapton backing. After shipping to the Stanford Synchrotron Radiation Lighsource (SSRL), the samples were inserted into the LANL Tender X-ray Chamber. The chamber was separated from the beam pipe by a beryllium window and continually flushed with He gas. Sample fluorescence was monitored using a partially depleted series (PIPS) charged particle detector with a 5000 mm2 active area (PD5000-75-500AM). Sample fluorescence was monitored against the incident radiation (I0) that was measured using an ionization chamber (30 cm) downstream from the beryllium window through which He was continually flowed. The X-ray absorption spectra (XAS) were measured at the Stanford Synchrotron Radiation Lightsource (SSRL) under dedicated operating conditions (3.0 GeV, 500 mA) at Beamline 14-3. This beamline is a bending magnet side station equipped with a water-cooled doublecrystal Si[111] monochromator. It employed a flat, bent vertically collimating 1 m Si, Ni-coated focusing mirror (fwhm beam spot size = 5 × 5 μm). The crystals were run detuned by 30% at 3200 eV for Cl Kedge XAS measurements and 3030 eV for Am M5,4-edge measurements. Spectra were collected with 1 mm vertical slits and 5 mm horizontal slits. For Cl K-edge and Am M5,4-edge XAS, energy calibrations were conducted externally using the maxima of the first pre-edge features in Cs2CuCl4 (2820.20 eV)94 and KCl (3608.4 eV), respectively. Cl K-Edge Xas Data Analysis. Data manipulations and analyses were conducted as previously described by Solomon and co-workers.94 Data were analyzed by fitting a line to the pre-edge region, 2702.1− 2815.0 eV, which was subsequently subtracted from the experimental data to eliminate the background of the spectrum. Data were normalized by fitting a first-order polynomial to the postedge region of the spectrum (2836−3030 eV) and setting the edge jump at 2836 eV to an intensity of 1.0. This normalization procedure gave spectra normalized to a single Cl atom or M−Cl bond. A deconvoluted model for the Cl K-edge XAS data was obtained using a modified version of EDG_FIT122 in IGOR 6.0. The pre-edge regions (
