Organometallic uranium(V)-imido halide complexes: From synthesis to electronic structure and bonding

Article abstract of DOI:10.1021/ja711010h

Reaction of (C₅Me₅)₂U(=N-2,4,6- ^tBu₃-C₆H₂) or (C₅Me ₅)₂U(=N-2,6-ⁱPr₂-C₆H ₃)(THF) with 5 equiv of CuX_n (n = 1, X = Cl, Br, I; n = 2, X = F) affords the corresponding uranium(V)-imido halide complexes, (C ₅Me₅)₂U(=N-Ar)(X) (where Ar = 2,4,6- ^tBu₃-C₆H₂ and X = F (3), Cl (4), Br (5), I (6); Ar = 2,6-ⁱPr₂-C₆H₃ and X = F (7), Cl (8), Br (9), I (10)), in good isolated yields of 75-89%. These compounds have been characterized by a combination of single-crystal X-ray diffraction, ¹H NMR spectroscopy, elemental analysis, mass spectrometry, cyclic voltammetry, UV-visible-NIR absorption spectroscopy, and variable-temperature magnetic susceptibility. The uranium L_III-edge X-ray absorption spectrum of (C₅Me₅)₂U(=N-2,4, 6-^tBu₃-C₆H₂)(Cl) (4) was analyzed to obtain structural information, and the U=N_imido (1-97(1) A), U-Cl (2.60(2) A), and U-C₅Me₅ (2.84(1) A) distances were consistent with those observed for compounds 3, 5, 6, 8-10, which were all characterized by single-crystal X-ray diffraction studies. All (C ₅Me₅)₂U(=N-Ar)(X) complexes exhibit U ^V/U^IV and U^VI/U^V redox couples by voltammetry, with the potential separation between these metal-based couples remaining essentially constant at ～1.50 V. The electronic spectra are comprised of π→π* and π→nb_5f transitions involving electrons in the metal-imido bond, and metal-centered f-f bands illustrative of spin-orbit and crystal-field influences on the 5f¹ valence electron configuration. Two distinct sets of bands are attributed to transitions derived from this 5f¹ configuration, and the intensities in these bands increase dramatically over those found in spectra of classical 5f¹ actinide coordination complexes. Temperature-dependent magnetic susceptibilities are reported for all complexes with μ^eff values ranging from 2.22 to 2.53 μ_B. The onset of quenching of orbital angular momentum by ligand fields is observed to occur at ～40 K in all cases. Density functional theory results for the model complexes (C ₅Me₅)₂U(=N-C₆H₅)(F) (11) and (C₅Me₅)₂U(=N-C₆H₅)(I) (12) show good agreement with experimental structural and electrochemical data and provide a basis for assignment of spectroscopic bands. The bonding analysis describes multiple bonding between the uranium metal center and imido nitrogen which is comprised of one σ and two π interactions with variable participation of 5f and 6d orbitals from the uranium center.

Full text of DOI:10.1021/ja711010h

Published on Web 03/26/2008
Organometallic Uranium(V)-Imido Halide Complexes: From
Synthesis to Electronic Structure and Bonding
Christopher R. Graves, Ping Yang, Stosh A. Kozimor, Anthony E. Vaughn,
David L. Clark, Steven D. Conradson, Eric J. Schelter, Brian L. Scott,
J. D. Thompson, P. Jeffrey Hay,* David E. Morris,* and Jaqueline L. Kiplinger*
Los Alamos National Laboratory, Los Alamos, New Mexico 87545
Received December 11, 2007; E-mail: kiplinger@lanl.gov
Abstract: Reaction of (C5Me5)2U(dN-2,4,6-tBu3-C6H2) or (C5Me5)2U(dN-2,6-iPr2-C6H3)(THF) with 5 equiv
of CuXn (n ) 1, X ) Cl, Br, I; n ) 2, X ) F) affords the corresponding uranium(V)-imido halide complexes,
(C5Me5)2U(dN-Ar)(X) (where Ar ) 2,4,6-tBu3-C6H2 and X ) F (3), Cl (4), Br (5), I (6); Ar ) 2,6-iPr2-C6H3
and X ) F (7), Cl (8), Br (9), I (10)), in good isolated yields of 75–89%. These compounds have been
characterized by a combination of single-crystal X-ray diffraction, 1H NMR spectroscopy, elemental analysis,
mass spectrometry, cyclic voltammetry, UV–visible-NIR absorption spectroscopy, and variable-temperature
magnetic susceptibility. The uranium LIII-edge X-ray absorption spectrum of (C5Me5)2U(dN-2,4,6-tBu3-
C6H2)(Cl) (4) was analyzed to obtain structural information, and the UdNimido (1.97(1) Å), U-Cl (2.60(2)
Å), and U-C5Me5 (2.84(1) Å) distances were consistent with those observed for compounds 3, 5, 6, 8–10,
which were all characterized by single-crystal X-ray diffraction studies. All (C5Me5)2U(dN-Ar)(X) complexes
exhibit UV/UIV and UVI/UV redox couples by voltammetry, with the potential separation between these metal-
based couples remaining essentially constant at ∼1.50 V. The electronic spectra are comprised of πfπ*
and πfnb5f transitions involving electrons in the metal-imido bond, and metal-centered f-f bands illustrative
of spin–orbit and crystal-field influences on the 5f1 valence electron configuration. Two distinct sets of
bands are attributed to transitions derived from this 5f1 configuration, and the intensities in these bands
increase dramatically over those found in spectra of classical 5f1 actinide coordination complexes.
Temperature-dependent magnetic susceptibilities are reported for all complexes with µeff values ranging
from 2.22 to 2.53 µB. The onset of quenching of orbital angular momentum by ligand fields is observed to
occur at ∼40 K in all cases. Density functional theory results for the model complexes (C5Me5)2U(dN-
C6H5)(F) (11) and (C5Me5)2U(dN-C6H5)(I) (12) show good agreement with experimental structural and
electrochemical data and provide a basis for assignment of spectroscopic bands. The bonding analysis
describes multiple bonding between the uranium metal center and imido nitrogen which is comprised of
one σ and two π interactions with variable participation of 5f and 6d orbitals from the uranium center.
Introduction
broad knowledge base for the elucidation of the aforementioned
fundamental questions in actinide chemistry. The role that
The complexes of the light actinide elements present an
important challenge toward the understanding of electronic
structure, bonding, and reactivity, as they often deviate from
the chemical principles established for either the transition
metals or the lanthanides, and while bonding in f-element
complexes has generally been considered to be ionic, the issue
of covalency remains an important subject of debate.1–3 Recent
work in our group has focused on the organometallic chemistry
of both thorium and uranium,4–9 with the goal of understanding
unprecedented structures and reaction pathways, to develop a
oxidation state plays in bonding is also of great interest in these
studies. Compared to the more common tetravalent and hexava-
lent organouranium complexes,10,11 pentavalent systems are rare,
presumably as a consequence of their instability. Indeed, most
of our current knowledge regarding the structures, electrochemi-
cal behavior, spectroscopic properties, and magnetic susceptibil-
(6) Pool, J. A.; Scott, B. L.; Kiplinger, J. L. Chem. Commun. 2005, 2591–
2593.
(7) Pool, J. A.; Scott, B. L.; Kiplinger, J. L. J. Alloys Compd. 2006, 418,
178–183.
(1) Diaconescu, P. L.; Arnold, P. L.; Baker, T. A.; Mindiola, D. J.;
Cummins, C. C. J. Am. Chem. Soc. 2000, 122, 6108–6109.
(2) Mazzanti, M.; Wietzke, R.; Pecaut, J.; Latour, J.-M.; Maldivi, P.;
Remy, M. Inorg. Chem. 2002, 41, 2389–2399.
(8) Kiplinger, J. L.; Scott, B. L.; Schelter, E. J.; Pool Davis Tournear,
J. A. J. Alloys Compd. 2007, 444–445, 477–482.
(9) Schelter, E. J.; Yang, P.; Scott, B. L.; Thompson, J. D.; Martin, R. L.;
Hay, P. J.; Morris, D. E.; Kiplinger, J. L. Inorg. Chem. 2007, 46,
7477–7488.
(10) Burns, C. J.; Eisen, M. S. Organoactinide Chemistry: Synthesis and
Characterization. In The Chemistry of the Actinide and Transactinide
Elements, 3rd ed.; Morss, L. R., Edelstein, N. M., Fuger, J., Eds.;
Springer: The Netherlands, 2006; Vol. 5, pp 2799–2910 and references
therein.
(3) Meyer, K.; Mindiola, D. J.; Baker, T. A.; Davis, W. M.; Cummins,
C. C. Angew. Chem., Int. Ed. 2000, 39, 3063–3066.
(4) Schelter, E. J.; Morris, D. E.; Scott, B. L.; Kiplinger, J. L. Chem.
Commun. 2007, 1029–1031.
(5) Schelter, E. J.; Yang, P.; Scott, B. L.; Da Re, R. E.; Jantunen, K. C.;
Martin, R. L.; Hay, P. J.; Morris, D. E.; Kiplinger, J. L. J. Am. Chem.
Soc. 2007, 129, 5139–5152.
(11) Ephritikhine, M. Dalton Trans. 2006, 2501–2516.
9
5272 J. AM. CHEM. SOC. 2008, 130, 5272–5285
10.1021/ja711010h CCC: $40.75
2008 American Chemical Society

Uranium(V)-Imido Halide Complexes
A R T I C L E S
ity for this oxidation state comes from classical coordination
complexesofthehalides,(e.g.,UX5,whereX)halide).12–15While
pentavalent uranium complexes containing metal–nitrogen and
metal–oxygen bonds have been reported in recent years,16,17 very
few examples of organometallic uranium(V) complexes have been
documented and thoroughly characterized,18,19 and the development
of a general synthetic strategy for the preparation of stable
pentavalent organouranium complexes is desirable.
Our recent explorations into electrochemical trends of ura-
nium organometallic complexes20 revealed a surprisingly ac-
cessible, chemically reversible one-electron metal-based UV/
UIV redox couple for (C5Me5)2U(dN-2,4,6-tBu3-C6H2) (1)21
(E1/2 ) –0.73 V vs [(C5H5)2Fe]+/0), demonstrating that the
electrochemically generated UV species of 1 is stable on the
voltammetric time scale. Analogous behavior is observed for
the sterically less-congested (C5Me5)2U(dN-2,6-iPr2-C6H3)
(THF) (2)19,22 (E1/2 ) –0.86 V vs [(C5H5)2Fe]+/0).23 These
voltammetric observations suggested that isolable organometallic
UV-imido complexes are viable, provided a suitable chemical
oxidant can be identified. We recently communicated a simple
synthetic procedure for the one-electron oxidation of the
UIV-imido complexes 1 and 2 with copper(I) salts to give the
corresponding UV-imido(X) complexes (X ) I, SPh, OTf) in
good isolated yield.23 This reaction pathway has allowed for
the systematic study of a series of pentavalent uranium
complexes. The in-depth synthetic protocol, as well as the
crystallographic, electrochemical, spectroscopic, and magnetic
characterization of the UV-imido halide series (halide ) F, Cl,
Br, I), paired with a computational study to investigate the
electronic structure of this new class of uranium complexes, is
presented herein.
Results and Discussion
Synthesis and Structural Characterization. Equations 1 and
2 display the synthetic protocol utilized and yields obtained in
the preparation of the UV-imido halides. Treatment of a toluene
solution of either 1 (eq 1) or 2 (eq 2) with excess Cu-halide
oxidant24 at room temperature generates the corresponding
UV-imido halide complexes 3–10 as dark brown solids in
75–89% isolated yields. Following workup by filtration through
Celite and crystallization, all complexes were reproducibly
isolated as analytically pure solids and were characterized by a
1
combination of H NMR, electrochemistry, UV–visible-NIR
spectroscopy, elemental and mass spectrometric analyses,
magnetic susceptibility, and X-ray crystallography.25
(12) Bagnall, K. W. In ComprehensiVe Coordination Chemistry; Wilkinson,
G., Gillard, R. D., Mc Cleverty, J. A., Eds.; Pergamon: Oxford, 1987;
Vol. 3, p 1129.
(13) For a review on the synthesis, stability, and properties of pentavalent
uranium, see: Selbin, J.; Ortego, J. D. Chem. ReV. 1969, 69, 657–
621.
(14) For recent advances in the chemistry of pentavalent uranyl complexes,
see: (a) Natrajan, L.; Burdet, F.; Pecaut, J.; Mazzanti, M. J. Am. Chem.
Soc. 2006, 128, 7152–7153. (b) Burdet, F.; Pecaut, J.; Mazzanti, M.
J. Am. Chem. Soc. 2006, 128, 16512–16513. (c) Berthet, J.-C.; Siffredi,
G.; Thuery, P.; Ephritikhine, M. Chem. Commun. 2006, 3184–3186.
(d) Arnold, P. L.; Patel, D.; Wilson, C.; Love, J. B. Nature 2008,
451, 315–318.
(15) Ryan, J. L. J. Inorg. Nucl. Chem. 1971, 33, 153–177.
(16) For representative examples of UV amide and alkoxide stabilized
complexes, see: (a) Zalkin, A.; Brennan, J. G.; Andersen, R. A. Acta
Crystallogr. 1988, C44, 1553–1554. (b) Roussel, P.; Hitchcock, P. B.;
Tinker, N. D.; Scott, P. Inorg. Chem. 1997, 36, 5716–5721. (c) Castro-
Rodriguez, I.; Olsen, K.; Meyer, K. Inorg. Chem. 2003, 125, 4564–
4571. (d) Salmon, P.; Thuery, P.; Ephritikhine, M. Polyhedron 2007,
26, 631–636.
(17) Castro-Rodriguez, I.; Olsen, K.; Gantzel, P.; Meyer, K. J. Am. Chem.
Soc. 2003, 125, 4565–4571.
Although silver(I) salts are routine oxidants for organometallic
complexes and have been utilized in the chemical oxidation of
uranium compounds,26,27 reports of analogous Cu(I) oxidations
are quite rare, with only a few examples in the literature limited
to the oxidation of electron-rich transition metal species.28,29
Furthermore, as Ag(I)-promoted oxidations are often difficult
to predict and control,27 the use of inexpensive, readily available,
(18) For examples of pentavalent organouranium complexes, see: (a)
Boisson, C.; Berthet, J.-C.; Lance, M.; Nierlich, M.; Vigner, J.;
Ephritikhine, M. J. Chem. Soc., Chem. Commun. 1995, 543–544. (b)
Gourier, D.; Caurant, D. Inorg. Chem. 1997, 36, 5931–5936. (c)
Ephritikhine, M.; Berthet, J. C.; Boisson, C.; Lance, M.; Nierlich, M.
J. Alloys Compd. 1998, 271-273, 144–149. (d) Boaretto, R.; Roussel,
P.; Alcock, N. W.; Kingsley, A. J.; Munslow, I. J.; Sanders, C. J.;
Scott, P. J. Organomet. Chem. 1999, 591, 174–184. (e) Arliguie, T.;
Rourmigue, M.; Ephritikhine, M. Organometallics 2000, 19, 109–
111.
(24) While CuCl, CuBr, and CuI are all readily available from commercial
sources, CuF is not. The present oxidation protocol is also applicable
to the use of CuX2 as the one-electron oxidant, and CuF2, which is
available, acted as the appropriate oxidant in the synthesis of both 3
and 7.
(19) Arney, D. S. J.; Burns, C. J. J. Am. Chem. Soc. 1993, 115, 9840–
9841.
(20) Morris, D. E.; Da Re, R. E.; Jantunen, K. C.; Castro-Rodriguez, I.;
Kiplinger, J. L. Organometallics 2004, 23, 5142–5153.
(21) Arney, D. S. J.; Burns, C. J. J. Am. Chem. Soc. 1995, 117, 9448–
9460.
(25) For the characterization data for 3-5 and 7-9, see the Supporting
Information. For the characterization data for 6 and 10, see ref 23.
(26) For examples of the oxidation of uranium complexes utilizing Ag(I)
salts as the oxidant, see: (a) Burns, C. J.; Smith, W. H.; Huffman,
J. C.; Sattelberger, A. P. J. Am. Chem. Soc. 1990, 112, 3237–3239.
(b) Boisson, C.; Berthet, J.-C.; Lance, M.; Nierlich, M.; Viger, J.;
Ephritikhine, M. J. Chem. Soc., Chem. Commun. 1995, 543–544. (c)
Maynadié, J.; Berthet, J.-C.; Thuery, P.; Ephritikhine, M. Organome-
tallics 2007, 26, 4585–4591.
(22) Herein, we also report the structural characterization of complex 2,
with crystalline product being obtained from the slow evaporation of
a concentrated pentanes solution at room temperature, yielding 86%
of the recovered product. See the Supporting Information for a thermal
ellipsoid representation and crystallographic experimental parameters,
as well as a table of bond lengths and angles.
(23) Graves, C. R.; Scott, B. L.; Morris, D. E.; Kiplinger, J. L. J. Am.
Chem. Soc. 2007, 129, 11914–11915.
(27) Connelly, N. G.; Geiger, W. E. Chem. ReV. 1996, 96, 877–910.
9
J. AM. CHEM. SOC. VOL. 130, NO. 15, 2008 5273

A R T I C L E S
Graves et al.
paramagnetic trivalent and tetravalent uranium systems,32 and
while the exact reasons for this phenomenon have been widely
discussed and contested,33–35 in the uranium systems it has been
proposed that the trend results from variation in the amount of
π donation from the halogen atom to the metal center across
the series.32a,c In effect, the better the π donor (F > Cl > Br >
I), the more electron-rich the uranium center, manifesting in a
larger shielding of the auxiliary protons and an upfield shift of
the resonances for those protons. This reasoning is in accord
with electrochemical findings, which suggest that the fluoride
derivatives 3 and 7 are more electron-rich at the metal center
than their Cl/Br/I counterparts (Vide infra).
X-ray diffraction studies confirmed the identity of the
UV-imido halides. Representative structures are provided for
complexes 3 and 8 in Figure 3, and selected geometric
parameters for complexes 3, 5, 6, 8–10 are presented in Table
1. Single crystals could not be obtained for 4 and 7, but their
formulations are consistent with 1H NMR, mass spectrometry,
and elemental analysis, as well as electrochemical, spectroscopic,
and magnetic characterization data. The crystallographic pa-
rameters for 3, 5, 6, 8–10 are provided in the Supporting
Information.
Figure 1. 1H NMR spectrum of (C5Me5)2U(dN-2,4,6-tBu3-C6H2)(Br) (5).
The molecular structures for both the tBu and iPr UV-imido
halide series feature a typical bent-metallocene framework with
the imido and halide ligands contained within the metallocene
wedge, similar in constitution to the related (C5Me5)2U(dO)(O-
2,6-iPr2-C6H3) UV-oxo alkoxide complex.19 All of the UV-
imido halide complexes have nearly linear UdNsCAr angles
(169.6(4)-172.2(9)°) and short UdNimido bond distances
(1.958(6)-1.975(6) Å), with no obvious trends present on the
basis of donating ability of the halide. These metrical parameters
compare well with those reported for other high-valent uranium
(UIV-UVI)-imido compounds. The UV-imido structures (C5H4-
Me)3U(dN-C6H5)36 and [(Me3Si)2N]3U(dN-4-Me-C6H4)37 have
UdNimido bond distances of 2.019(6) and 1.940(6) Å and
UdNsCAr angles of 167.4(6)° and 171.4(5)°, respectively.
UVI-imido complexes show similar geometric parameters:
(C5Me5)2U(dN-C6H5)238 has UdNimido bond distance of 1.952(7)
Å and UdNsCAr angle of 177.8(6)°, while (C5Me5)2U(dN-
Figure 2. 1H NMR spectrum of (C5Me5)2U(dN-2,6-iPr2-C6H3)(Cl) (8).
and diverse Cu(I) salts as oxidants provides a versatile and useful
synthetic advancement.
These paramagnetic uranium(V) complexes were character-
ized by 1H NMR spectroscopy. Complexes 3–6 exhibit a broad30
signal at ∼4–6 ppm, corresponding to the C5Me5 ligand protons.
Further, there is an inequivalency of the ortho tBu groups, which
is not observed in the UIV-imido starting material.21 Similar
1H NMR spectra were observed for the UV-iPr-imido halides
7–10: a broad signal at ∼4–6 ppm, corresponding to the C5Me5
(32) This trend in chemical shift for like protons as a function of halide
has been seen for other paramagnetic uranium systems. (a) [(1,3-
R2C5H3)2UX]2 (R ) SiMe3, CMe3; X ) F, Cl, Br, I): Lukens, W. W.,
Jr.; Beshouri, S. M.; Stuart, A. L.; Andersen, R. A. Organometallics
1999, 18, 1247–1252. (b) (C5Me5)3UX (X ) F, Cl, Br): Evans, W. J.;
Nyce, G. W.; Johnston, M. A.; Ziller, J. W. J. Am. Chem. Soc. 2000,
122, 12019–12020. (c) (1,3-R2C5H3)2UX2 (R ) SiMe3, CMe3; X )
F, Cl, Br, I): Lukens, W. W., Jr.; Beshouri, S. M.; Blosch, L. L.; Stuart,
A. L.; Andersen, R. A. Organometallics 1999, 18, 1235–1246. The
1H NMR behavior of the (C5H5)3U-X system has also been
extensively studied, (Fisher, R. D. NMR-Spectroscopy of organo-f-
element and pre-lanthanoid complexes. In Fundamental and Techno-
logical Aspects of Organo-f-Element Chemistry; Marks, T. J., Fragala,
I. L., Eds.; NATO ASI Series C: Mathematical and Physical Sciences;
D. Reidel Publishing Co.: Dordrecht, 1985; Vol. 155, p 294), not
only for X ) halide but also for other donor groups, such as OR,
NR2, SR, PR2, etc.
i
1
ligand protons and inequivalent Pr groups. Representative H
NMR spectra for the UV-tBu-imido halides and the
UV-iPr-imido halides are shown in Figures 1 and 2, respectively.
Interestingly, there is a relationship between the chemical shift
for analogous protons (i.e., the C5Me5 resonances) and the
identity of halogen atom in the paramagnetic (C5Me5)2U(dN-
2,6-iPr2-C6H3)(X) structure, with the chemical shift of the C5Me5
group moving upfield with variation of the halogen from I f
Br f Cl f F.31 Similar trends have been observed for other
(33) Dhingra, M. M.; Ganguli, P.; Mitra, S. Chem. Phys. Lett. 1974, 25,
579–581.
(34) Drago, R. S.; Wayland, B. B. Inorg. Chem. 1968, 7, 628–630.
(35) LaMar, G. N.; Fischer, R. H.; Horrocks, W. D., Jr. Inorg. Chem. 1967,
6, 1798–1803.
(28) Leigh, J. S.; Whitmire, K. H.; Yee, K. A.; Albright, T. A. J. Am.
Chem. Soc. 1989, 111, 2726–2727.
(29) Rabier, A.; Lugan, N.; Mathieu, R.; Geoffroy, G. L. Organometallics
1994, 13, 4676–4678.
(36) Brennan, J. G.; Andersen, R. A. J. Am. Chem. Soc. 1985, 107, 514–
516.
(30) For example, complexes 6 and 10 have ∆ν1/2 values of 93 and 125
Hz, respectively.
(37) Stewart, J. L. Tris[bis(trimethylsilyl)amido]uranium: Compounds with
Tri-, Tetra-, and Pentavalent Uranium. Ph.D. Dissertation, Univeristy
of California, Berkeley, CA, 1988.
(31) The chemical shifts for the (C5Me5) protons in the 1H NMR spectrum
for 7-10 are as follows: δ 3.94 ppm (7, X ) F), 4.92 ppm (8, X )
Cl), 5.31 ppm (9, X ) Br), and 5.78 ppm (10, X ) I).
(38) Arney, D. S. J.; Burns, C. J.; Smith, D. C. J. Am. Chem. Soc. 1992,
114, 10068–10069.
9
5274 J. AM. CHEM. SOC. VOL. 130, NO. 15, 2008

Uranium(V)-Imido Halide Complexes
A R T I C L E S
Figure 3. Molecular structures of complexes 3 and 8, with thermal ellipsoids projected at the 50% probability level. Hydrogen atoms have been omitted
for clarity.
Table 1. Selected Metrical Parameters for Complexes 3, 5, 6, and 8–10
3
5
6
8
9
10
(X ) F)
(X ) Br)
(X ) I)
(X ) Cl)
(X ) Br)
(X ) I)
UsX (Å)
UdN (Å)
NsCAr (Å)
NsUsX (°)
UdNsCAr (°)
2.122(5)
1.965(8)
1.415(11)
97.0(3)
2.7744(10)
1.958(6)
1.424(9)
96.11(18)
169.8(5)
3.0116(6)
1.975(6)
1.418(8)
97.20(16)
169.7(5)
2.6209(15)
1.963(4)
1.404(7)
105.79(13)
169.6(4)
2.789(3)
1.969(7)
1.40(2)
105.3(2)
172.2(9)
3.0385(7)
1.974(7)
1.406(10)
106.6(2)
170.7(6)
171.0(7)
2,6-iPr2-C6H3)()O)19 has corresponding values of 1.988(4) Å
and 170.5(4)°. These geometric parameters demonstrate that
UdN bond distances and UdNsC angles in uranium-imido
complexes are insensitive to oxidation state. In fact, the
UdNimido distances and UdNsCAr angles for the UIV-imido
starting materials (1,21 1.952(12) Å, 162.3(10)°; 2,22 2.006(5)
Å, 172.6(5)°) also fall within the same range as for the
UV-imido halides reported in the present work.
As expected on the basis of the ionic radius of the halogen
atoms,39 the U-X bond length increases down the series, with
U-F < U-Cl < U-Br < U-I.40 At 2.122(5) Å, the U-F
bond distance observed in 3 compares well with those for other
known pentavalent uranium fluoride complexes41 and falls
within range of distances reported for other high-valent uranium
compounds that have been structurally characterized containing
a U-F bond. Uranium(V) complexes containing a U-Cl bond
have been previously structurally characterized, with typical
bond lengths ranging from 2.4 to 2.7 Å.4243,44 At 2.6209(15)
Å, the U-Cl bond length observed for 8 fits in this range and
also compares well with metrical parameters observed for other
organometallic uranium complexes exhibiting U-Cl bonds.45
Several pentavalent uranium complexes are also known contain-
ing a U-Br bond, with distances ranging between 2.6 and 2.9
Å,46,47 a range that includes those distances found for both 5
(2.7744(10) Å) and 9 (2.789(3) Å). To the best of our
knowledge, there are no structurally characterized pentavalent
iodide complexes; however, the distances reported for 6
(3.0116(6) Å) and 10 (3.0385(7) Å) fall in the range of other
structurally reported U-I distances.48,49
EXAFS Analysis of (C5Me5)2U(dN-2,4,6-tBu3-C6H2)(Cl).
While (C5Me5)2U(dN-2,4,6-tBu3-C6H2)(Cl) (4) was isolated in
high yield (75%) in a crystalline form, the crystals were not
suitable for analysis by single-crystal X-ray experiments.
(42) U-F distances in the range of ∼2.1–2.4 Å have been reported for
tetra- and hexavalent uranium complexes. For examples, see: (a) Ryan,
R. R.; Penneman, R. A.; Kanellakopulos, B. J. Am. Chem. Soc. 1975,
97, 4258–4260. (b) Burns, C. J.; Smith, W. H.; Huffman, J. C.;
Sattlelberger, A. P. J. Am. Chem. Soc. 1990, 112, 3237–3239. (c)
Lukens, W. W., Jr.; Beshouri, S. M.; Blosch, L. L.; Stuart, A. L.;
Andersen, R. A. Organometallics 1999, 18, 1235–1246. (d) Evans,
W. J.; Nyce, G. W.; Johnston, M. A.; Ziller, J. W. J. Am. Chem. Soc.
2000, 122, 12019–21020.
(43) Smith, G. S.; Johnson, Q.; Elson, R. E. Acta Crystallogr. 1967, 22,
300–303.
(44) Wedler, M.; Noltemeyer, M.; Edelmann, F. T. Angew. Chem., Int.
Ed. 1992, 31, 72–75.
(45) For representative examples of U-Cl distances for other uranium
oxidation states, see: (a) Zalkin, A.; Beshouri, S. M. Acta Crystallogr.
1989, C45, 1080–1082. (b) Hall, S. W.; Huffman, J. C.; Miller, M. M.;
Avens, L. R.; Burns, C. J.; Arney, D. S. J.; England, A. F.; Sattelberger,
A. P. Organometallics 1993, 12, 752–758. (c) Peters, R. G.; Scott,
B. L.; Burns, C. J. Acta Crystallogr. 1999, C55, 1482–1483.
(46) Bohrer, R.; Conradi, E.; Mueller, U. Z. Anorg. Allg. Chem. 1988, 558,
119–127.
(39) Cotton, F. A.; Wilkinson, G.; Bochmann, M.; Murillo, C. AdVanced
Inorganic Chemistry, 6th ed.; John Wiley & Sons, Inc.: New York,
1998; p 548.
(40) For a detailed list of various UV halide complexes, see: Grenthe I.;
Drozdzynski, J.; Fujino, T.; Buck, E. C.; Albrecht-Schmitt, T. E.; Wolf,
S. F. Uranium. In The Chemistry of the Actinide and Transactinide
Elements, 3rd ed.; Morss, L. R., Edelstein, N. M., Fuger, J., Eds.;
Springer: The Netherlands, 2006; Vol. 1, pp 501–529.
(47) Levy, J. H.; Taylor, J. C.; Wilson, P. W. J. Inorg. Nucl. Chem. 1978,
40, 1055–1057.
(41) R- and ꢀ-UF5 have U-F bond distances in the range of 2.18–2.30 Å
(Zachariasen, W. H. Acta Crystallogr. 1949, 2, 296–298), while (2,2′-
bipyridyl)UF5 has U-F bond lengths of 2.04–2.06 Å (Arnaudet, L.;
Bougon, R.; Ban, B.; Lance, M.; Nierlich, M.; Vigner, J. Inorg. Chem.
1994, 33, 4510–4516). For detailed crystallographic parameters for a
series of anionic [X][UF6] complexes, see:(a) Brown, D. The Halides
of the Lanthanides and Actinides; Wiley Interscience: New York, 1968.
(b) Penneman, R. A.; Ryan, R. R.; Rosenzweig, A. Struct. Bonding
(Berlin) 1973, 13, 1–52.
(48) While no structurally characterized pentavalent uranium iodide
complexes have been reported, bond lengths between ∼3.0 and 3.2 Å
have been noted for other uranium oxidation states. For representative
examples, see: (a) Avens, L. R.; Burns, C. J.; Butcher, R. J.; Clark,
D. L.; Gordon, J. C.; Schake, A. R.; Scott, B. L.; Watkin, J. G.; Zwick,
B. D. Organometallics 2000, 19, 451–457. (b) Crawford, M. J.; Mayer,
P. Inorg. Chem. 2005, 44, 5547–5549. (c) Maynadié, J.; Berthet, J. C.;
Thuéry, P.; Ephritikhine, M. Organometallics 2006, 25, 5603–5611.
9
J. AM. CHEM. SOC. VOL. 130, NO. 15, 2008 5275

A R T I C L E S
Graves et al.
Figure 5. Occupied and virtual MO energy levels of (C5Me5)2U(dN-
C6H5)(I) (12) and plots of orbitals with significant 5f or UdN bonding
character (spin-up levels are shown with unpaired electron in HOMO).
Figure 4. Raw uranium LIII-edge k3-weighted EXAFS data (top inset),
Fourier transform without phase corrections of the k3-weighted EXAFS
(middle), and curve-fitting analysis of the Fourier-transformed spectrum
(bottom inset) for (C5Me5)2U(dN-2,4,6-tBu3-C6H2)(Cl) (4).
nitrogen atom with a UdN distance of 1.97(1) Å and is
consistent with the formulation of a UdNimido bond. This
distance is similar to the UdNimido distances observed for
compounds 3, 5, 6, 8–10, which were determined by single-
crystal X-ray crystallography. The second wave, which was fit
with a single chloride atom, has a distance of 2.60(2) Å and is
similar to the U-Cl distance (2.621(2) Å) in complex 8. The
final wave was attributable to the ring carbons of the penta-
methylcyclopentadienyl ligand and fit with a U–C5Me5 distance
of 2.84(1) Å. This average distance in 4 is significantly longer
than the 2.75(4) Å average U–C5Me5 distance observed in
(C5Me5)2U(dN-2,6-iPr2-C6H3)(Cl) (8) and is likely a manifesta-
tion of the steric demands of the (dN-2,4,6-tBu3-C6H2) ligand
compared to the (dN-2,6-iPr2-C6H3) ligand.52 The amplitude
of the U-C wave was only approximated to half of that
expected from the total number of ring carbons, which may be
attributable to variation in the U–C5Me5 bond distances in 4,
resulting in an anharmonic distribution and the loss of amplitude.
This is consistent with metrical parameters observed for
compounds 3, 5, 6, 8–10, whose U–C5Me5 distances range from
2.70 to 2.90 Å. Although the data were of high enough quality
toobtainbonddistancesfortheuranium-imido,uranium-chloride,
and U–C5Me5 interactions, the methyl groups on the C5Me5
ligand and the carbon atoms on the imido ligands were not
resolvable.
Density Functional Calculations. The structures of the ground
states of the model systems (C5Me5)2U(dN-C6H5)(F) (11) and
(C5Me5)2U(dN-C6H5)(I) (12) were calculated using density
functional theory (DFT) approaches. As can be seen in Table
2, there is good overall agreement (within 0.05 Å and 10°)
between the calculated and experimental structures. The cal-
culated U-X bond lengths for 11 and 12 are 2.117 Å (expt )
2.122(5) Å) and 3.064 Å (expt ) 3.0116(6), 3.0385(7) Å), and
the UdN distances are 1.969 Å (expt ) 1.965(8) Å) and 1.944
Å (expt ) 1.975(6), 1.974(7) Å), respectively. Likewise, 11
and 12 have NsUsX angles of 107.8° (expt ) 97.0(3)°) and
However, solid-state structural information was obtained by
analyzing the uranium LIII-edge X-ray absorption spectrum.50
The background-subtracted k3-weighted extended X-ray absorp-
tion fine structure (EXAFS) spectrum of 4 and the calculated
fit are shown in Figure 4, as are the Fourier transform modulus
and theoretical fit of the k3-weighted EXAFS data, which were
generated without phase corrections. The theoretical EXAFS
modeling code, FEFF7,51 was employed to calculate the
backscattering phases and amplitudes of the individual neigh-
boring atoms, using (C5Me5)2U(dN-2,6-iPr2-C6H3)(Cl) (8) as
a model compound, and both the k3-weighted ꢁ(k) and ꢁ(R)
indicate a reasonable fit in both phase and amplitude.
The Fourier-transformed spectrum contains two large peaks,
and although the first peak was fit with a single shell, the curve-
fitting analysis revealed the second peak to be asymmetric, and
it was fit with two shells. The first shell is attributed to the
(49) Compound 10 crystallized with two independent yet similar structures
in the unit cell. Selected geometric parameters for one molcule are
presented in Table 1, with those for the other being as follows: UsI
) 3.0482(7) Å; UdN ) 1.960(7) Å; NsCAr ) 1.431(10) Å; NsUsX
) 107.0(2) °; UdNsCAr ) 165.9(6)°. For a complete list of bond
distances and angles for complex 10, see ref 23.
(50) For recent examples where EXAFS has been used to provide structural
information, see: (a) Allen, P. G.; Bucher, J. J.; Clark, D. L.; Edelstein,
N. M.; Ekberg, S. A.; Gohdes, J. W.; Hudson, E. A.; Kaltsoyannis,
N.; Lukens, W. W.; Neu, M. P.; Palmer, P. D.; Reich, T.; Shuh, D. K.;
Tait, C. D.; Zwick, B. D. Inorg. Chem. 1995, 34, 4797–4807. (b) Allen,
P. G.; Shuh, D. K.; Bucher, J. J.; Edelstein, N. M.; Reich, T.; Denecke,
M. A.; Nitsche, H. Inorg. Chem. 1996, 35, 784–787. (c) Clark, D. L.;
Conradson, S. D.; Donohoe, R. J.; Keogh, D. W.; Morris, D. E.;
Palmer, P. D.; Rogers, R. D.; Tait, C. D. Inorg. Chem. 1999, 38, 1456–
1466. (d) Lukens, W. W., Jr.; Allen, P. G.; Bucher, J. J.; Edelstein,
N. M.; Hudson, E. A.; Shuh, D. K.; Reich, T.; Andersen, R. A.
Organometallics 1999, 18, 1253–1258. (e) Jiang, J.; Renshaw, J. C.;
Sarsfield, M. J.; Livens, F. R.; Collison, D.; Charnock, J. M.; Eccles,
H. Inorg. Chem. 2003, 42, 1233–1246. (f) Catalano, J. G.; Brown,
G. E., Jr. Geochim. Cosmochim. Acta 2005, 69, 2995–3005. (g)
Nikitenko, S. I.; Hennig, C.; Grigoriev, M. S.; Le Nauor, C.; Cannes,
C.; Trubert, D.; Bossé, E.; Berthon, C.; Moisy, P. Polyhedron 2007,
26, 3136–3142.
(51) Ankudinov, A. L.; Rehr, J. J. Phys. ReV. B: Condens. Matter 1997,
56, R1712–R1715.
(52) Evans, W. J.; Kozimor, S. A.; Ziller, J. W. Inorg. Chem. 2005, 44,
7960–7969.
9
5276 J. AM. CHEM. SOC. VOL. 130, NO. 15, 2008

Uranium(V)-Imido Halide Complexes
A R T I C L E S
Table 2. Comparison of Calculated Geometric Parameters for Model Compounds (C5Me5)2U(dN-C6H5)(X) (X ) F (11), I (12)) with Complexes 3, 6,
and 10
(C5Me5)2U(dN-C6H5)(F) (11)
(C5Me5)2U(dN-C6H5)(I) (12)
t
expt, Bu (3)
t
expt, Bu (6)
i
expt, Pr (10)
calcd
calcd
UsX (Å)
UdN (Å)
NsCAr (Å)
UsC5Me5(cent) (Å)
NsUsX (°)
UdNsCAr (°)
2.122(5)
1.965(8)
1.415(11)
2.505
97.0(3)
171.0(7)
2.117
1.969
1.376
2.505, 2.502
107.8
176.9
3.0116(6)
1.975(6)
1.418(8)
2.528
97.20(16)
169.7(5)
3.0385(7)
1.974(7)
1.406(10)
2.459, 2.454
106.6(2)
3.064
1.944
1.380
2.512, 2.510
97.22
175.2
170.7(6)
Table 3. MO Components for the f1 Complexes
(C5Me5)2U(dN-C6H5)(F) (11) and (C5Me5)2U(dN-C6H5)(I) (12)
97.22° (expt ) 97.20(16)°, 106.6(2)°) and UdNsCAr angles
of 176.9° (expt ) 171.0(7)°) and 175.2° (expt ) 169.7(5)°,
170.7(6)°), respectively. Presumably, the small overall deviations
in bond angles can be largely attributed to the unfavorable steric
interactions in the experimental structures with Bu and Pr
substituents on the imido groups that are not present in the
calculated model complexes.
(C5Me5)2U(dN-C6H5)(F) (11) (C5Me5)2U(dN-C6H5)(I) (12)
MO
E (eV)
U6d U5f
N
F
MO
E (eV)
U6d U5f
N
I
t
i
L+6
L+5
L+4
L+3
L+2
L+1
LUMO
0.27
1
85
68
92
94
77
92
L+6
L+5
L+4
L+3
L+2
L+1
LUMO
0.15
3
87
80
70
94
85
90
42
-1.38
-1.72
-2.02
-2.07
-2.10
-2.29
-1.68
-2.12
-2.32
-2.40
-2.60
-2.64
8
The electronic structure of the UV-imido halide complexes
arises from the interactions of the pentamethylcyclopenta-
dienyl, halide, and imido ligands with the uranium metal
center. The multiple bond formed between the imido ligand
and the uranium center can be viewed by considering a formal
dianionic [N-C6H5]2- fragment interacting with the uranium
6d and 5f orbitals. A primary σ donation from the [N-C6H5]2-
fragment to the metal center, coupled with two π interactions,
results in a nearly linear (UdNsCAr ) 176.9° and 175.2°,
calculated) bonding interaction between the imido ligand and
the uranium metal center. The first imido nitrogen π lone
pair is conjugated with the imido aryl group to form a UdN
π bond, while the second lone pair lies in the plane of the
aryl ring forming a UdN π| bond. This interpretation is
similar to that from calculations53 determined for the model
system (C5H5)2U(dN-C6H5)2 utilized in describing the bond-
ing for the previously reported UVI-bis(imido) complex,
(C5Me5)2U(dN-C6H5)2.38 A more recent investigation by
Belkhiri and co-workers on the bonding analysis of the
actinide-imido bond in (C5H5)2An(dN-C6H5)2 (An ) Th,
U) and related systems also noted the importance of the
participation of 5f orbitals in σ and π interactions to form
the coplanar UdN-C6H5 group.54 In the same report, it was
also observed that the transition metal-imido bond in Mo(VI)
complexes cannot achieve the analogous triple bond nature
utilizing only d-orbitals, and as a result the ModNsCAr angle
is close to 150°.
HOMO -5.18
33 17
HOMO -5.39
3
H-1
H-2
H-3
H-4
H-5
H-6
H-8
H-14
H-16
H-21
H-23
-5.35
-5.49
-5.59
-5.99 10 13
-6.11 31
-6.43 14 14 49
-7.87
-9.02
-9.19
-9.84
-9.93
11
73
3
2
H-1
H-2
H-3
H-4
H-5
H-6
H-7
H-9
H-10
-5.56
-5.70
-5.96
-6.06
-6.16
-6.20
-6.44 11
-6.92 15
-7.26 12 18 14 48
-8.12
-9.23
-9.40
-10.66
37 13
5
4
3
9
2
19
14
37
7
8
5
9
21
15
20
71
58
1
7
8
4
4
2
2
4
7
2
3
1
1
2
2
3
2
30
7
10
3
2
5
25 39
39 H-11
53 H-17
61 H-19
13 H-27
4
6
2
4
5
1
2
2
31
4
12
15
2
4
8
H-26 -10.35
H-27 -10.57
Table 4. Net Spin and Charge Distribution for (C5Me5)2U(dN-C6H5)(F)
(11) and (C5Me5)2U(dN-C6H5)(I) (12) and the Corresponding Anions
([11]- and [12]-) and Cations ([11]+ and [12]+)
U
N
C
X (I, F)
Net Spin
Charges
11
+1.38
+2.18
+1.44
+2.20
-0.19
-0.11
-0.20
-0.11
+0.04
+0.01
+0.04
+0.02
-0.01
-0.01
-0.02
-0.01
[11]-
12
[12]-
11
+1.17
+1.14
+1.14
+0.95
+0.87
+1.02
-0.64
-0.58
-0.69
-0.63
-0.58
-0.68
+0.21
+0.20
+0.22
+0.21
+0.19
+0.21
-0.47
-0.43
-0.52
-0.46
-0.26
-0.66
[11]+
[11]-
12
The occupied and virtual MO energy levels are depicted
schematically in Figure 5, where selected orbitals are plotted,
and analysis of the highest occupied orbitals in the complexes
is presented in Table 3, where the amount of 5f and 6d
admixture on the uranium metal center is given along with the
[12]+
[12]-
the fluoride ligand lone pairs occur at much lower energies
(–10.3 to –9.8 eV in 11 vs –7.2 to –6.2 eV in 12).
Nimido and halide components. The HOMO contains the unpaired
The calculated spin densities (Table 4) reflect the 5f1
configuration on the uranium(V) metal center in these com-
plexes. The Mulliken analysis shows 1.38 unpaired electrons
on the uranium atom for 11 and 1.44 for 12, somewhat higher
than the formal spin of one unpaired electron. Small amounts
of opposite spins arise on the nitrogen atom involved in forming
a multiple bond with the metal. The halide atom and the aryl
group on the imido ligand exhibit no spin density. These results
show that the spin properties are not wholly described by the
HOMO (or SOMO) from the DFT calculations, which has only
33% (11) and 42% (12) metal character, but that there are
5f electron, but has a considerable ligand component. The
HOMO-2 for 12 shows strong UdN (π ) π bonding character
and 5f (19%) and 6d (5%) admixture. Below this orbital are
four occupied π orbitals corresponding to two pairs of the set
from the C5Me5 ligands as well as the three iodide lone pair 5p
orbitals. The MOs show considerable mixing of the various
contributions from different ligands, and, as a result, it is difficult
to isolate one metal–ligand combination in any particular orbital.
The UdN (π|) bond is particularly evident in HOMO-9, where
some U-I bonding character also exists. The resulting energy
level diagram for the fluoride complex 11 is similar, except that
9
J. AM. CHEM. SOC. VOL. 130, NO. 15, 2008 5277

A R T I C L E S
Graves et al.
Table 5. Flow of Charge upon Oxidation or Reduction of (C5Me5)2U(dN-C6H5)(F) (11) and (C5Me5)2U(dN-C6H5)(I) (12) from Mulliken Analysis
U
X
N
Ph
C5Me5
sum
11 f [11]+
11 f [11]-
-0.022
-0.031
+0.042
-0.049
+0.056
-0.052
+0.206
-0.231
+0.360, +0.358
-0.313, –0.324
+1.000
-1.000
12 f [12]+
12 f [12]-
-0.088
+0.033
+0.209
-0.185
+0.047
-0.049
+0.172
-0.208
+0.330, +0.329
-0.299, –0.291
+1.000
-1.000
Table 6. Comparison of Calculated Geometric Parameters for (C5Me5)2U(dN-C6H5)(X) (X ) F (11), I (12)) and Their Corresponding Cations and
Anions
(C5Me5)2U(dN-C6H5)(F) (11)
charge ) +1
(C5Me5)2U(dN-C6H5)(I) (12)
charge ) +1
charge ) 0
charge ) -1
charge ) 0
charge ) -1
UsX (Å)
UdN (Å)
2.117
1.969
2.084
1.926
2.170
2.022
3.064
1.944
2.958
1.909
3.212
1.994
NsCAr (Å)
1.376
1.380
1.358
1.380
1.384
1.365
UsC5Me5(cent) (Å)
NsUsX (°)
UdNsCAr (°)
2.505, 2.502
107.8
176.9
2.449, 2.449
101.1
172.2
2.582, 2.588
104.1
179.6
2.512, 2.510
97.22
175.2
2.468, 2.471
93.9
173.4
2.581, 2.580
100.0
178.6
additional contributions to the net spin polarization from other
occupied orbitals.
electrons, for [11]- and [12]-, respectively (Table 4). Compared
to the neutral complexes (11 and 12) the spins of the negative
ions have added about 0.8 electron to the metal, with the
remainder of the unpaired spin (0.2 e-) delocalized throughout
the C5Me5 rings. As can be seen in Table 5, it is also interesting
to note that the Mulliken charges on the uranium center and
imido functional groups are essentially constant, regardless of
whether an electron is added or removed. For the case of adding
an electron, the additional charge goes partially to the halogen
(0.05 electron for fluoride [11]-, 0.2 electron for iodide [12]-),
with ∼0.3 electron to each C5Me5 ligand and 0.2 electron onto
the imido ligand. Similar results were provided by the NBO
analysis of natural charges.55
The next six orbitals of the virtual set (LUMO, LUMO+1 to
LUMO+5) correspond to the other 5f orbitals, thereby complet-
ing the manifold of the f levels. The first orbital beyond this
group (LUMO+6) corresponds to the π* orbital of the imido
aryl group. In terms of the electronic absorption spectra, this
leads to the qualitative picture of low-lying f-f states as the
lowest excitations for the molecule. Other excitations could
involve ligand-to-metal 5f excitations and, at higher energies,
ligand-based excitations involving the virtual π* orbitals.
Preliminary spectroscopic assignments are discussed below in
the Electronic Spectroscopy section. In order to get a qualitative
picture of the energies involved in the f-f states including
spin–orbit coupling, a simple ligand field calculation was
performed. The preliminary results show a set of three closely
spaced Kramers doublet levels between 0 and 2800 cm-1, and
To evaluate the relative UV/UIV redox potentials for the
UV-imido complexes, calculations were performed for their
negative ions as gas-phase species. The electron affinity of 12
is greater than that for 11, as shown by the relative energy of
the anion [12]- (-2.20 eV) compared to that of the anion [11]-
(-1.73 eV). This result is consistent with the experimental
observations that the UV-imido iodide complexes have more
positive reduction potentials than the corresponding UV-imido
fluoride complexes (see Table 7). The calculations for the cations
[11]+ and [12]+ show essentially no difference between the UVI/
UV redox potentials for 11 and 12 (within 0.03 eV). Inclusion
of zero-point energy corrections at 298 K, enthalpy corrections,
and solvation effects has little effect on the potential of the UVI/
UV redox couples; the two potentials remain within 0.02 eV
difference. Experimentally, the UVI/UV redox potentials show
much less variation across the halide series than do UV/UIV
redox potentials (Vide infra).
Electrochemistry. Cyclic and square-wave voltammetric data
were collected for all UV-imido halide complexes in ∼0.1 M
[Bu4N][fluoroarylborate]/tetrahydrofuran solution ([fluoroaryl-
borate] ) [B(C6F5)4]- or [B(3,5-(CF3)2-C6H3)4]-). Most data
were collected at room temperature; however, complex 7 was
found to decompose within tens of minutes in these solvent/
electrolyte solutions. Thus, investigations of the scan-rate-
dependent behavior in the cyclic voltammetry for this complex
were conducted at approximately –50 °C. Typical cyclic
voltammograms are illustrated in Figure 6. Each complex
a second set of four doublets between 6200 and 10000 cm-1
with the first two levels occurring at 6200 and 7100 cm-1
,
,
suggesting that two peaks in this region may be arising from
distinct electronic origins (Vide infra).
Cation and anion calculations were performed for both
(C5Me5)2U(dN-C6H5)(F) (11) and (C5Me5)2U(dN-C6H5)(I) (12)
in order to understand their structural, electronic, spin, and redox
properties. The metrical parameters for the corresponding cations
([11]+ and [12]+) and anions ([11]- and [12]-) are essentially
identical to their neutral counterparts (Table 6). While the bond
lengths are slightly shorter for the cationic species and margin-
ally longer for anionic complexes compared to those of the
corresponding neutral parent species, the geometrical parameters
are generally unaffected by the removal or addition of one
electron. Note that this finding agrees well with experiment, as
noted above, with the UdN bond length in particular being
noticeably insensitive to metal oxidation state.
The formation of the anions [11]- and [12]– corresponds to
a 5f2 configuration on the uranium metal, giving UIV. The added
electron is largely localized in the f-orbital on the uranium center
and comprised entirely of 5f character, which was verified by
investigating the ground-state natural orbitals (NOs). The
calculations indicate a higher spin, 2.18 and 2.20 unpaired
(53) Hay, P. J. Faraday Discuss. 2003, 124, 69–83.
(54) Belkhiri, L.; Lissillour, R.; Boucekkine, A. THEOCHEM 2005, 757,
155–164.
(55) Glendening, E. D.; Badenhoop, J. K.; Reed, A. E.; Carpenter, J. E.;
Bohmann, J. A.; Morales, C. M.; Weinhold, F. NBO 5.0; Theoretical
Chemistry Institute, University of Wisconsin: Madison, WI, 2001.
9
5278 J. AM. CHEM. SOC. VOL. 130, NO. 15, 2008

Uranium(V)-Imido Halide Complexes
A R T I C L E S
Table 7. Summary of Redox Potential Dataa for
(C5Me5)2U(dN-Ar)(X) Complexes 3–10 in ∼0.1 M
[Bu4N][fluoroarylborate]b/THF Solution at Room Temperature
process. The only other voltammetric activity seen in these
systems occurs at ∼1 V (vs [(C5H5)2Fe]+/0) and is ascribed to
irreversible C5Me5-based oxidation, as noted in previous
investigations.9,20 In particular, there are no clear indications
of any oxidation processes attributable to the coordinated halide
ligands, although the irreversible nature of the nominal C5Me5-
based processes could easily mask additional halide-based redox
transformations. Finally, there were no additional metal-based
reduction processes (e.g., UIV/UIII as seen in previous studies
of uranium(IV) metallocenes9,20) resolved for the UV-imido
complexes within the negative potential range of these solvent/
electrolyte systems. The redox potential data derived from the
voltammetric experiments are summarized in Table 7.
UV
complexes
electrolyte
anionb
E
1/2(UVI/UV) E1/2(UV/UIV)
(V) (V)
∆E1/2
(V)
Ar
X
3
4
5
2,4,6-tBu3-C6H2
F
BArF
-0.19 -1.78
0.04 -1.50
0.04 -1.43
0.04 -1.25
-0.14 -1.81
0.03 -1.52
0.07 -1.44
0.11 -1.37
1.59
1.54
1.47
1.29
1.67
1.55
1.51
1.48
2,4,6-tBu3-C6H2 Cl BArF
2,4,6-tBu3-C6H2 Br B(C6F5)4
6
2,4,6-tBu3-C6H2
2,6-iPr2-C6H3
2,6-iPr2-C6H3
2,6-iPr2-C6H3
2,6-iPr2-C6H3
I
F
BArF
BArF
7c
8
Cl BArF
Br BArF
I
9
10
B(C6F5)4
|∆E1/2 (av)
|
1.51 ( 0.04
The observed redox behavior of these complexes provides
direct insight into the generic influence ligation of the halide
ion (F-, Cl-, Br-, or I-) to the (C5Me5)2U(dN-Ar) core plays
on the redox energetics in these bent metallocene complexes.
Further, it enables assessment of the more subtle impact of
changing the halide ion from fluoride through iodide in these
structures. In comparison to their UIV-imido precursors, the
UV/UIV couple for all the UV-imido halide complexes is shifted
to more negative potentials (destabilized) by ∼1 V, and, in
contrast to the behavior of the UIV-imido precursors and other
tetravalent uranium bent metallocene complexes,9,20,23 the UVI/
UV couple emerges within the available potential window,
indicating a significant stabilization of the UVI oxidation state
in the halide complexes. These two gross perturbations in the
redox behavior across the entire suite of UV-imido halide
complexes are the principal manifestations of halide ion
complexation on the (C5Me5)2U(dN-Ar) core.
UIV
precursors
electrolyte
anionb
E
1/2(UV/UIV)
(V)
E
1/2(UIV/UIII) ∆E1/2
Ar
(V)
(V)
1d
2e
2,4,6-tBu3-C6H2
2,6-iPr2-C6H3
B(C6F5)4
BArF
-0.73
-0.86
-2.61
-2.40
1.88
1.54
a All E1/2 values are versus [(C5H5)2Fe]+/0 and were determined from
the peak position in a square-wave voltammogram or from the average
of the cathodic and anodic peaks in
a cyclic voltammogram.
b Electrolyte anion was either [B(3,5-(CF3)2-C6H3)4]- (BArF) or [B-
(C6F5)4]- (B(C6F5)4). c Although scan-rate-dependent behavior was
explored at ∼-50 °C, potential calibration data were collected for this
complex at room temperature using a freshly prepared solution. d Data
from ref 20. e Data from ref 23.
exhibits one oxidation wave, attributed to the UVI/UV redox
process, and one reduction wave, attributed to the UV/UIV redox
The variations in redox potentials that are specific to the
identity of the halide ion across this series of complexes are
found to be more subtle, and, as seen in Table 7, they are more
pronounced on the E1/2 values for the reduction process than
those of the oxidation process. Nonetheless, for both the tBu-
i
and Pr-imido series, there are consistent trends in the E1/2
values for both the oxidation process and the reduction process
according to E1/2(F-) < E1/2(Cl-) < E1/2(Br-) < E1/2(I-), with
the exception of the oxidation process for the tBu-imido series,
for which the half-wave potential remains constant for X ) Cl,
Br, and I. These results are remarkably consistent with those
predicted from the DFT calculations for the model complexes
(C5Me5)2U(dN-C6H5)(F) (11) and (C5Me5)2U(dN-C6H5)(I)
(12). Specifically, theory predicts that 11 should be more difficult
to reduce (have a lower electron affinity) than 12, while the
oxidation potential (ionization potential) is predicted to be
roughly equal for the two halide complexes. This trend in redox
potentials versus halide ion ligation at first might seem coun-
terintuitive on the basis of the relative electronegativities of the
halide ions and the inductive influence this might be expected
to exert on the electron density at the redox-active metal center.
We attribute the observed trend to a greater degree of polariz-
ability of the halide orbitals with increasing atomic number of
the halide that more readily accommodates the addition of
electrons in the principally metal-based orbitals in the heavier
halide complexes.
A final observation on the thermodynamic properties of these
UV-imido halide complexes revealed in the voltammetric data
is the nearly constant spacing between the metal-based oxidation
and reduction processes across this entire series of eight
complexes: ∆E1/2 ) 1.51 ( 0.04 V (Table 7). A similar
consistency in separation has been found between metal-based
oxidation (UIV/UV) and reduction (UIII/UIV) processes for a
Figure 6. Cyclic voltammograms for (C5Me5)2U(dN-Ar)(X) complexes
4, 5, 7, and 10 in ∼0.1 M [Bu4N][fluoroarylborate]/THF solution at 200
mV/s scan rate and at room temperature for all complexes except 7, for
which T ≈ -50 °C. The low-current signal marked with an asterisk in the
scan of 7 is believed to be due to a small amount of 8 present as an impurity.
Vertical arrows indicate the initial (rest) potential, and horizontal arrows
indicate the initial scan direction.
9
J. AM. CHEM. SOC. VOL. 130, NO. 15, 2008 5279

A R T I C L E S
Graves et al.
broad range of (C5Me5)2UIVL and (C5Me5)2UIVL′2 (L ) imide,
L′ ) ketimide, hydrozanate) complexes, but in these tetravalent
systems this consistency in separation between oxidation and
reduction waves is on average ∼2.1 V.9,20 The constancy in
these average values implies that attractive (electron-nucleus)
and repulsive (electron–electron) interactions, not specific
perturbations introduced by the ligand sets in the metallocene
wedge, are the dominant factors in determining the relative redox
energetics, but we are unaware of evidence for these trends in
other actinide systems.
The cyclic voltammetric experiments provide one additional
interesting result. As illustrated in Figure 6 for cyclic scans
collected at 200 mV/s and confirmed in the more detailed scan-
rate-dependent behavior (not shown), the UV-imido complexes
all exhibit markedly different heterogeneous electron-transfer
rate constants (ket) for the metal-based oxidation vs reduction
processes. In particular, the oxidation process is always found
to be more sluggish than the reduction process, as evidenced
by the greater separation between the anodic and cathodic peaks
for the oxidation wave compared to the reduction wave. No
attempt was made to quantify these electron-transfer rate
constants because of inherent problems in completely accounting
for contributions from uncompensated resistance in these
relatively low dielectric solutions. Nonetheless, it is clear even
from the data in Figure 6 that the effect varies from complex to
complex, with the difference in ket(UVI/UV) versus ket(UV/UIV)
being greatest for the fluoride complexes (3 and 7) and least
for the iodide complexes (6 and 10). The DFT results described
above (Table 6) do not indicate any substantially greater change
in the coordination environment about the metal ion for the
oxidation step versus the reduction step to which one could
attribute a difference in the rate constants associated with inner-
sphere structural reorganization. However, we believe this effect
is a manifestation of the greater structural compensation required
to accommodate metal ion size reduction (UVfUVI) than that
of metal ion size expansion (UVfUIV) in these complexes.
Figure 7. UV–visible-NIR absorption spectra of (C5Me5)2U(dN-Ar)(X)
complexes 3–6 (top) and 7–10 (bottom) in toluene solution at room
temperature. Dashed lines are component bands from Gaussian fit of spectra
for complexes 4 and 8.
The room-temperature electronic absorption spectra over the
entire accessible energy range (determined by properties of the
toluene solvent) for all eight UV-imido halide complexes are
illustrated in Figure 7. These spectra are clearly dominated by
broad, relatively intense bands over the entire spectral range,
but weaker and narrower bands typical of f-f transitions can
be readily seen in the lower energy region. To provide some
gauge of spectral band components in the broadband portions
of the spectra, the spectral region above ∼7000 cm-1 has been
fit using the minimum number of Gaussian profiles required to
adequately reproduce the observed spectra. Examples of the
component bands from these fitting results are highlighted for
complexes 4 and 8 in Figure 7. It is noteworthy that the total
spectral band profile over the entire energy range is quite similar
Electronic Spectroscopy. The organometallic UV-imido
halide complexes provide two important opportunities to explore
electronic structure. The first pertains to the relatively uncom-
mon 5f1 valence electronic configuration possessed by these
systems. Systems with a single 5f-electron (Pa4+, U5+, Np6+
)
constitute the simplest examples to explore spin–orbit versus
crystal-field effects since there are no electron–electron repulsion
terms to consider and the f-orbital density of states is reduced
to only two manifolds, a 2F5/2 ground-state manifold and a 2F7/2
excited-state manifold, whose individual multiplet levels will
split in response to the symmetry and strength of the field
induced by the ligands. Thus, for these (C5Me5)2U(dN-Ar)(X)
complexes, the electronic transitions between states derived from
these manifolds (f-f or ligand-field transitions) provide direct
insight into ligand-field strengths from both imido and halide
ligands for an approximately fixed spin–orbit coupling interac-
tion associated with uranium(V) in this coordination environ-
ment. The second opportunity derives from the existence of
occupied and virtual ligand-based orbitals derived from the
imido ligand that lie close in energy to the metal valence orbitals,
as discussed in the DFT section. This engenders the formation
of new molecular electronic states of ligand-localized and
metal–ligand charge-transfer character similar to those found
in early transition-metal-imido complexes and from which the
degree of covalent metal–ligand bonding interactions can be
inferred from electronic spectral data involving transitions
between these states.
t
i
for all Bu-imido complexes (3–6) and for all Pr-imido
complexes (7–10), although there are changes in the molar
absorptivities that certainly exceed the uncertainty in these
measurements.56
Specific and detailed assignments of all the broad bands
reflected in the Gaussian fits to these spectra are not possible
on the basis of existing information. However, the DFT results
described above and published data for early transition-
metal-imido complexes can be used to suggest plausible generic
assignments for this portion of the spectra. While we know of
no reports of spectroscopic data for d1 imido systems to which
we can compare our results, there are numerous d0 imido
systems for which detailed assignments are available.57,58 The
lowest-energy electronic transitions seen in transition-metal-imido
complexes such as (dme)Cl3Ta(dN-R) (dme ) dimethoxy-
(56) Uncertainties in the molar extinction coefficients are determined from
the accuracy in weighing the complexes (milligram quantities) in the
glovebox to prepare sample solutions and are estimated to be (20%.
(57) Heinselman, K. S.; Hopkins, M. D. J. Am. Chem. Soc. 1995, 117,
12340–12341.
(58) Williams, D. S.; Korolev, A. V. Inorg. Chem. 1998, 37, 3809–3819.
9
5280 J. AM. CHEM. SOC. VOL. 130, NO. 15, 2008

Uranium(V)-Imido Halide Complexes
A R T I C L E S
ethane; R ) alkyl or aryl) have been assigned to the electric-
dipole-allowed singlet and triplet components of the πMdNfnbd
(nbd ) nonbonding metal d orbital) charge-transfer transition.
DFT results described in Figure 5 suggest that, among the set
of valence orbitals for these uranium(V)-imido complexes,
there are also high-lying occupied orbitals comprised principally
of πMdN character (HOMO-1) and low-lying unoccupied
orbitals of nonbonding f-orbital character (LUMO) that could
give rise to similar πMdNfnb5f charge-transfer transitions. In
the case of these 5f1 systems, the spin assignments for these
charge-transfer transitions would be quartet and doublet (instead
of triplet and singlet), but the relative intensities in the pair of
transitions [I(triplet or quartet) < I(singlet or doublet)] should carry over
to the 5f1 case. Thus, consistent with both DFT calculations
and in analogy to transition-metal-imido systems, we propose
that the bands at ∼10000 and ∼14000 cm-1 are 4(πMdNfnb5f)
and 2(πMdNfnb5f) transitions, respectively. The relatively low
oscillator strengths in these charge-transfer transitions are
entirely consistent with the expected relatively small orbital
overlap between the donor (πMdN) and acceptor (5f) charge-
transfer orbitals compared to that found in the transition-metal
analogues.
The bands that lie to higher energy of the putative πMdNfnb5f
charge-transfer transitions have substantially greater oscillator
strength, suggesting that a different electronic transition is
operative. Again, with recourse to the DFT results, we note that
there is a virtual π*Ph orbital (LUMO+6) that lies to higher
energy than the nb5f orbital manifold. Thus, it is plausible that
at least some the higher-energy bands in these spectra can be
attributed to transitions of πMdNfπ*Ph character. Several such
transitions, occurring at slightly different energies but having
this same orbital description, would be possible for these
complexes because of the existence of two discrete πMdN bonds
(Vide supra), although the intensities should differ in a manner
that reflects the orthogonality of the two πMdN bonds and the
impact this has on the allowedness of the πfπ* transition.
One further important observation from the broadband
spectral data is that there does not appear to be any significant
sensitivity to the identity of the halide ion. Thus, other than
changes in molar absorptivity noted above, the spectra for the
entire series of tBu-imido complexes (3–6) look very similar,
as do the spectra for all the iPr-imido complexes (7–10). Fully
allowed halide ligand-to-metal charge-transfer transitions are a
common feature in the spectroscopy of uranium halide com-
plexes.59 Thus, the insensitivity in the spectral data to the identity
of the halide indicates that any electronic transitions involving
the halide ions, e.g., the halide-to-metal charge-transfer band(s),
must lie to higher energy than probed in our experiments or
have insufficient oscillator strengths to impart distinct features
on top of the existing spectral bands.
Figure 8. NIR absorption spectra of (C5Me5)2U(dN-Ar)(X) complexes
3–6 (top) and 7–10 (bottom) in toluene-d8 solution at room temperature.
Insets show expansions of vertical scale to illustrate vibronic hot bands.
from this operation are indicated in Figure 8 for all eight imido
halide complexes. Note that data in this spectral region were
collected in toluene-d8 in 1 mm path length cells to minimize
interference from the strong solvent vibrational overtone and
combination bands. The resulting spectra are characterized by
what appear to be two sets of bands, and there is now a very
clear sensitivity in the spectra to the identity of both the halide
ion and the imido ligand.
While there is not a large body of literature relating to
spectroscopic characterization of 5f1 systems, there have been
several detailed treatments of results for octahedral or pseudo-
octahedral hexahalide complexes, particularly of UV and
PaIV.15,60–65 These reports provide a framework within which
we can begin to interpret the f-f spectral data for these
UV-imido systems. As noted above, spin–orbit coupling
interactions will split the f-orbital manifold in an f1 system into
2
2
two subsets: a F5/2 ground-state manifold and a F7/2 excited-
state manifold. The influence of an octahedral or pseudo-
octahedral crystal field as in the hexahalide complexes will be
to further lift the degeneracy within each manifold, giving rise
to a total of five states in rigorous octahedral symmetry (Γ7,
Γ8, Γ7′, Γ8′, and Γ6 from lowest to highest energy using the
conventional double-group notations in the Oh point group), two
of which (Γ8 and Γ8′) are orbitally two-fold degenerate. A total
of four electronic transitions would thus be expected in Oh
The remaining important elements in these spectral data are
the low-energy, low-intensity, narrow-band features that are just
discernible in the data shown in Figure 7 and are ascribed to
the f-f or ligand-field transitions. To accentuate this spectral
region, the lowest-energy Gaussian bands from the fits described
above were used to define the sloping tails that underlie these
f-f bands. These Gaussian bands were subtracted from the total
spectral envelopes in the energy region from ∼9000 to 5000
cm-1 to isolate the contributions from the f-f bands. Results
(60) Edelstein, N.; Brown, D.; Whittaker, B. Inorg. Chem. 1974, 13, 563–
567.
(61) Karraker, D. G. Inorg. Chem. 1964, 3, 1618–1622.
(62) Penneman, R. A.; Sturgeon, G. D.; Asprey, L. B. Inorg. Chem. 1964,
3, 126–129.
(59) Liu, G.; Beitz, J. V. Optical Spectra and Electronic Structure. In The
Chemistry of the Actinide and Transactinide Elements, 3rd ed.; Morss,
L. R., Edelstein, N. M., Fuger, J., Eds.; Springer: The Netherlands,
2006; Vol. 3, pp 2013–2111.
(63) Reisfeld, M. J.; Crosby, G. A. J. Mol. Spectrosc. 1963, 10, 232–234.
(64) Reisfeld, M. J.; Crosby, G. A. Inorg. Chem. 1965, 4, 65–70.
(65) Selbin, J.; Ortego, J. D.; Gritzner, G. Inorg. Chem. 1968, 7, 976–982.
9
J. AM. CHEM. SOC. VOL. 130, NO. 15, 2008 5281

A R T I C L E S
Graves et al.
symmetry, and for crystalline Cs[UF6], for which the most
attention has been devoted, these electronic transitions (and/or
their non-totally symmetric vibronic components, since pure
electronic transitions are dipole forbidden in rigorous Oh
symmetry) have been identified to lie at ∼4600, 6900, 12700,
relationship describing the energy separation between these
excited states and the ground state is a complicated function of
both the spin–orbit coupling strength and the crystal field
strength,60,64 and on the basis of measurements of spacing
between only these three levels (ground-state plus two excited-
state levels), it is not possible to determine experimental values
for the spin–orbit coupling parameter and the crystal-field
splitting parameters. Further efforts to extract these parameters
using a combination of theoretical calculations and experimental
data are underway.
and 14200 cm-1 64
.
The (C5Me5)2U(dN-Ar)(X) complexes 3–10 have at best Cs
symmetry. This Cs crystal field environment results in complete
lifting of the orbital degeneracies in the spin–orbit manifolds,
yielding a total of seven states from which six electronic
transitions would be expected. In addition, in this low-symmetry,
non-centrosymmetric environment, the electric dipole selection
rules are relaxed, and one might expect that the pure electronic
transitions are able to carry substantial intensity. As noted in
the theory discussion above, a simple ligand-field calculation
that explicitly incorporates spin–orbit coupling for the model
complexes (C5Me5)2U(dN-C6H5)(X) (X ) F (11), I (12)) results
in the prediction of two groups of energy levels: the first is
comprised of the ground-state level and two additional levels within
∼2800 cm-1 of the ground state, and the second is comprised of
the four remaining levels that lie in the region between 6000 and
10000 cm-1 above the ground state. Notably, the two lowest-lying
levels in this second group were predicted to be 6200 and 7100
cm-1 above the ground-state level.
As can be seen in Figure 8, there appear to be two sets of
bands. The first set is comprised of an intense band in the region
of ∼6200 cm-1 for all complexes, the peak energy of which
varies with halide ion (Vide infra), and several much weaker
bands to higher energy that are likely vibronic in character. This
first set of bands also contains several very weak features on
the low-energy side of the most intense band (inset, Figure 8).
Data collected for these systems at 77 K, which will be presented
in more detail in a forthcoming paper that will include additional
discussion of the vibronic structure in these spectra, clearly show
that these weak features at lower energy disappear at low
temperature, indicating that they are vibronic hot bands. Thus,
we propose that the most intense feature in this first set, at
∼6200 cm-1, is an electronic origin transition. The energy of
this band is in excellent agreement with that predicted from
simple ligand-field/spin–orbit calculations noted above. The
second set of bands contains greater vibronic structure than the
first set, and the low-temperature data are less helpful in
identifying what the likely electronic origin band might be.
Nonetheless, it is notable that this collection of bands occurs in
approximately the same energy range (∼7000 cm-1) as the
simple theoretical calculations predicted for the next higher-
lying electronic transition. Of the remaining four electronic
transitions that should be found in these complexes under this
low-symmetry crystal field, we note that the two that are
predicted to lie below ∼3000 cm-1 would not be observed in
our experiments because of strong solvent interferences, and
the remaining two, predicted to lie at energies in the range from
∼7000 to 10000 cm-1, could be masked by the more intense
charge-transfer bands that tail well into this NIR spectral range.
The final important observation from these electronic spectral
data concerns the intensities in these f-f bands. The few reports
of molar extinction coefficients for pentavalent uranium com-
plexes pertain to complexes having octahedral or pseudo-
octahedral symmetry, for which electric dipole transitions are
forbidden to a first approximation and transitions gain intensity
through vibronic mechanisms involving non-totally symmetric
modes. Thus, for example, the more intense band normally
assigned as a component in the Γ7fΓ7′ electronic transition at
∼7000 cm-1 has an extinction coefficient, ꢀ (in M-1 cm-1), of
- 15
∼5 for UF6-, ∼14 for UCl6-, and ∼22 for UBr6
.
In stark
contrast, for 3-10, the molar extinction coefficients are in the
range from ∼100 to 400 M-1 cm-1 for the more intense bands.
As noted previously, the lower symmetry of these complexes
leads to a relaxation of the electric-dipole selection rules, and
this certainly contributes to the dramatic increase in the observed
intensities here. However, it is likely that these f-f bands are
also gaining some intensity from coupling to the nearby
π
MdNfnb5f charge-transfer transitions. A similar intensity-
stealing mechanism has been observed for UIV-ketimide
complexes,9,20 and the origin is similar in both the ketimide
and imido series of complexes. Namely, there are metal f-orbitals
that are common to both transitions that facilitate the mixing
of charge-transfer character into the f-f transitions.
Magnetic Susceptibility. The paucity of organometallic UV
complexes has similarly resulted in few reports on their
temperature-dependent magnetic behavior. The 5f1 electronic
configuration for UV results in a nominal electronic ground term
of 2F5/2 from S ) 1/2, L ) 3, and J ) 5/2, and the predicted ꢁT
value for the UV ion follows as 0.80 emu K mol-1 (2.54 µB)
from the Russell–Saunders coupling scheme. As discussed
2
above, ligand-field (LF) splitting of the F5/2 term in a low-
symmetry environment results in three Kramers doublets, with
the ground term having a net magnetic moment. Reduction of
the ꢁT product with decreasing temperature results from thermal
depopulation of these LF states. Figure 9 shows magnetic
susceptibility data for complexes 6 and 10, which are repre-
sentative for 3–6 and 7–10, and Table 8 lists the observed
magnetic moments for the set of complexes obtained from linear
fits of the high-temperature susceptibility data. These values
are lower than the expected value but are consistent with those
reported for UV complexes.13,14d,17,66–69
The energy of the most intense f-f band in these spectral
data, at ∼6200 cm-1, exhibits an interesting and highly
(66) Edelstein, N. M.; Lander, G. H. Magnetic Properites. In The Chemistry
of the Actinide and Transactinide Elements, 3rd ed.; Morss, L. R.,
Edelstein, N. M., Fuger, J., Eds.; Springer: The Netherlands, 2006;
Vol. 4, pp 2241–2247.
–
–
conserved trend: E(F ) ) 6172 (3), 6142 (7) < E(Cl ) ) 6241
–
–
(4), 6178 (8) < E(Br ) ) 6253 (5), 6190 (9) < E(I ) ) 6257 (6),
6198 (10) (all energies in cm-1). It is also clear qualitatively
that the splitting between the two excited-state energy levels
that give rise to the two electronic transitions seen in these data
is significantly larger for the fluoride complexes than for the
other halide complexes for both imido ligand systems. The
(67) Miyake, C.; Hirose, M.; Ohya-Nishiguchi, H. Inorg. Chim. Acta 1989,
165, 179–183.
(68) Selbin, J.; Ahmad, N.; Pribble, M. J. J. Chem. Soc., Chem. Commun.
1969, 759–760.
(69) Selbin, J.; Ahmad, N.; Pribble, M. J. J. Inorg. Nucl. Chem. 1970, 32,
3249–3258.
9
5282 J. AM. CHEM. SOC. VOL. 130, NO. 15, 2008

Uranium(V)-Imido Halide Complexes
A R T I C L E S
2
regime, all the LF states of the F5/2 term are populated, such
that the ꢁT product is expected to reach the predicted value of
0.80 emu K mol-1sthough it clearly falls short in all the
complexes here. The slopes of the linear regimes are due to the
presence of temperature-independent paramagnetic (TIP) com-
ponents from admixture of paramagnetic excited states into the
ground state. The precipitous decrease of the LT regime is due
2
to lifting of the F5/2 term into its three ligand-field Kramers
doublets and quenching of orbital angular momentum. On the
basis of the magnetic data, it can be concluded that the total
LF splitting energies are on the order of ∼40 K in these
complexessconsistent with data reported for other UV-imido
complexes.66 This energetic splitting is significantly smaller than
the result obtained from the simple spin–orbit LF calculation
described above, which would appear to overestimate the
splittings between the three doublets arising from the 2F5/2 term.
In future studies, we shall report additional spectroscopic and
magnetic studies along with more sophisticated theoretical
models to obtain a more detailed picture of the lowest energy
levels of these UV complexes.
Figure 9. Magnetic susceptibilities and ꢁT products of (C5Me5)2U(dN-
2,4,6-tBu3-C6H2)(I) (6, orange) and (C5Me5)2U(dN-2,6-iPr2-C6H3)(I) (10,
black) from 2 to 350 K.
In the context of the small amount of literature available,
previous reports of UV paramagnetism have asserted that the
generally low magnetic moments for UV complexes (compared
to the L-S predicted value of 2.54 µB) arise from covalent
character of the metal–ligand interaction that presumably results
from siphoning of spin density from the metal to the ligand
with an associated reduction in the orbital magnetism.17,66 This
would result in large differences in the room temperature
magnetic moment for the complexes, dependent on geometry
and the extent of metal–ligand overlap. However, negative
correlation exists between complexes expected to show such a
reduction due to large covalent metal–ligand interactions (4,
µeff ) 2.51 µB) and those whose interactions are primarily
through σ donation (UCl5, µeff ) 1.42 µB).13 Accompanying
large differences in temperature-dependent susceptibilities, i.e.,
sensitivity of LF state energies to the ligands, are also not
evident.66 As such, it appears that simple conclusions on the
extent of covalency as determined by room temperature
magnetic moments for UV complexes are not possible.
Table 8. Magnetic Susceptibility Data for (C5Me5)2U(dN-Ar)(X)
Complexes 3–10
Ar ) 2,4,6-tBu3-C6H2
µeff (µB/U)
Ar ) 2,4,6-iPr2-C6H2
µeff (µB/U)
X
X
3
4
5
6
F
2.46
2.51
2.30
2.53
7
8
9
10
F
2.22
2.42
2.42
2.34
Cl
Br
I
Cl
Br
I
Conclusions
The series of UV-imido halide complexes reported here have
provided a valuable entry into the synthesis and characterization
of pentavalent uranium organometallic complexes. The protocol
is applicable to the full suite of halides, allowing for the
synthesis of (C5Me5)2U(dN-Ar)(X) (where Ar ) 2,6-iPr2-C6H3
or 2,4,6-tBu3-C6H2, and X ) F, Cl, Br, I) from the corresponding
UIV-imido precursors, (C5Me5)2U(dN-Ar). Structural analyses,
electrochemical and spectroscopic studies, and DFT calculations
provide a coherent description of the electronic structure of these
systems and indicate that both the UsX and UdN bonds exhibit
covalency. The qualitative picture of the energy levels that
emerges from both absorption spectra and the calculations shows
that the f-f excitations occur at low energies and both charge-
transfer and ligand-based excitations arise at higher energies,
where the latter transitions involve excitations from MdN
bonding orbitals to either 5f orbitals or ligand-based π* orbitals.
The electrochemical analysis shows that the ligand framework
can stabilize both the UIV and UVI oxidation states, and the DFT
results provide insights into the charge redistribution and
structural changes occurring in these redox processes. While
the halide series is not broad enough to allow us to make
generalizations, studies involving the reactivity of these
Figure 10. Temperature-dependent ꢁT data for (C5Me5)2U(dN-2,6-iPr2-
C6H3)(X) (7–10) from 2 to 350 K. Red, 7 (X ) F); blue, 8 (X ) Cl);
green, 9 (X ) Br); black, 10 (X ) I).
The temperature dependence of the ꢁT products for 7–10 is
shown in Figure 10.70 The data are generally similar in
appearance, indicating no major changes in magnetic behavior
based on the identity of the halide ion in these complexes. Their
temperature dependences can be analyzed in terms of two
general regimes of behavior: the high-temperature regime, from
∼40 to 300 K, which is linear with large slopes, and the low-
temperature regime, <40 K, where the ꢁT products decrease
precipitously. There are several instructive points to take from
the appearance of these two regimes. In the high-temperature
(70) For a plot of the temperature dependence of the susceptibilities for
3-6, see the Supporting Information.
9
J. AM. CHEM. SOC. VOL. 130, NO. 15, 2008 5283

A R T I C L E S
Graves et al.
UV-imido halides for the synthesis of non-halogenated
UV-imido complexes (C5Me5)2U(dN-Ar)(L) (L * F, Cl, Br,
I) are currently underway. It is anticipated that expansion of
this series to include non-halogenated derivatives will provide
greater insight into the electronic structure of these pentavalent
uranium systems and allow for the development of a compre-
hensive theoretical model that is consistent with the data
obtained from electrochemical, spectroscopic, and magnetic
measurements.
When considered in the broader context of other recent results
for organouranium complexes, the results presented here for
the UV-imido complexes provide an important bridge between
the data for UVI-imido systems (e.g., (C5Me5)2U(dN-Ar)2),
which have long been regarded as possessing substantial
covalency in the metal–ligand multiple bonds, and the more
recent data for UIV-ketimide systems (e.g., (C5Me5)2U(sNdC-
Ar2)2), which possess metal–ligand bonding intermediate be-
tween single and double bonds on the basis of structural data.
All three systems, (C5Me5)2U(dN-Ar)2, (C5Me5)2U(dN-Ar)(X),
and (C5Me5)2U(sNdC-Ar2)2, exhibit similar experimental
hallmarks for covalent metal–ligand bonding. While it is not
possible to make quantitative comparisons regarding the degree
of covalency from the existing data, the emerging picture is
that all three uranium oxidation states are able to support
covalent bonding in these organometallic complexes.
metric data were analyzed using Wavemetrics IGOR Pro (version
4.0) software on a Macintosh platform.
Magnetic susceptibility data were collected using a Quantum
Design Superconducting Quantum Interference Device (SQUID)
magnetometer at 5 T from 2 to 350 K. The samples were sealed in
a 5 mm Wilmad 505-PS NMR tube along with a small amount of
quartz wool, which held the sample near the tube center. Contribu-
tions to the magnetization from quartz wool and the NMR tube
were measured independently and subtracted from the total
measured signal. Diamagnetic corrections were made with the use
of Pascal’s constants.
For the X-ray absorption measurement, (C5Me5)2U(dN-2,4,6-
tBu3-C6H2)(Cl) (4) was loaded under a helium atmosphere into two
nested aluminum sample holders equipped with Kapton windows.
One set of windows was glued on, and one set was sealed with
indium wire. The sample holder was shipped under He in a sealed
container that was opened at the beam-line, immediately attached
to the coldfinger of a liquid N2 cryostat, and quickly evacuated
(10-5 Torr). The cryostat was cooled with liquid N2, and the X-ray
absorption fine structure (EXAFS) spectrum was measured at the
Stanford Synchrotron Radiation Laboratory, under dedicated operat-
ing conditions (3.0 GeV, 5%, 100 mA) on end station 10-2. A single
energy was selected from the white beam with a liquid-N2-cooled
double-crystal monochromator utilizing Si[220] crystals. The
crystals were run fully tuned, and a flat Pt-coated mirror, tilted to
have a cutoff energy of 20–22 keV, was used to eliminate
harmonics. The energy of the first inflection point of an Y foil,
whose spectrum was measured within the same scan as the sample
using the transmitted portion of the beam, was defined as 17032.08
eV. The spectra were measured in the fluorescence mode using a
30-element Ge detector and digital amplifiers. The signals from
the different detector channels were examined individually and
averaged by weighting them to maintain a constant signal:noise
ratio. Dead times were set at ∼1.2 µs.
Experimental Section
Instrumentation and Sample Protocols. Electronic absorption
spectral data were obtained for toluene or toluene-d8 solutions of
complexes over the wavelength range 300–2500 nm on a Perkin-
Elmer model Lambda 950 UV–visible-NIR spectrophotometer.
Data were collected in 1 cm and 1 mm path length cuvettes loaded
in a recirculating Vacuum Atmospheres NEXUS model inert
atmosphere (N2) drybox equipped with a 40CFM Dual Purifier NI-
Train and run versus the appropriate toluene solvent reference.
Samples were typically run at multiple dilutions to optimize
absorbance in the UV–visible and near-infrared regions, respec-
tively. Spectral resolution was typically 2 nm in the visible region
and 4–6 nm in the near-infrared.
Voltammetric data were obtained in the Vacuum Atmospheres
drybox system described above. In addition, data for complex 7
were obtained in a Schlenk-line electrochemical cell immersed in
a dry ice/2-propanol bath at ∼-50 °C. All data were collected using
a Perkin-Elmer Princeton Applied Research Corp. (PARC) model
263 potentiostat under computer control with PARC model 270
software. All sample solutions were ∼1–2 mM in complex with
0.1 M [Bu4N][B(3,5-(CF3)2-C6H3)4] or [Bu4N][B(C6F5)4] supporting
electrolyte in THF solvent. All data were collected with the positive-
feedback IR compensation feature of the software/potentiostat
activated to ensure minimal contribution to the voltammetric waves
from uncompensated solution resistance (typically ∼1 kΩ under
the conditions employed). For experiments at ambient temperature,
solutions were contained in PARC model K0264 microcells
consisting of a ∼3 mm diameter Pt disk working electrode, a Pt
wire counter electrode, and a Ag wire quasi-reference electrode.
For the low-temperature experiment, a Schlenk cell was employed
consisting of Pt wire working and counter electrodes sandwiching
a Ag wire quasi-reference electrode. Scan rates from 20 to 5000
mV/s were employed in the cyclic voltammetry scans to assess the
chemical and electrochemical reversibility of the observed redox
transformations. Half-wave potentials were determined from the
peak values in the square-wave voltammograms or from the average
of the cathodic and anodic peak potentials in the reversible cyclic
voltammograms. Potential calibrations were performed at the end
of each data collection cycle using the ferrocenium/ferrocene couple
as an internal standard. Electronic absorption and cyclic voltam-
Using an in-house program, the averaged spectrum of 4 was
normalized to a per-atom basis by offsetting the spectrum so that
the value of a polynomial fit through the pre-edge region equaled
zero at 17185 eV and then scaled so that the value of a polynomial
fit through the post-edge region equaled unity at this same energy.
The EXAFS were extracted from the spectrum by initially subtract-
ing out the absorption edge and part of the white line determined
by curve-fitting the edge region with the sum of an arctangent and
asymmetric Gaussian fit to the absorption edge and peak. The
smooth atomic-like background, which was approximated with a
spline polynomial whose knots were set to minimize the area of
the Fourier transform modulus of the spectrum over the region R
) 1–1.1 Å, was then subtracted. Finally, the data were divided by
the absorption falloff calculated from the McMaster tables.71,72 The
Fourier transform metrical parameters were extracted from the
EXAFS by nonlinear least-squares curve-fitting of the ꢁ data using
amplitudes and phases calculated by FEFF7.51 The ionization
energy was constrained to be the same for each shell of the fit.
Errors were calculated by determining the contribution of each shell
to the overall error of the fit and then finding the shift in this
parameter that resulted in the error being higher by 10% of this
contribution.
Synthesis. The synthesis of complexes 3–5 and 7–9 is similar
to that previously reported for the preparation of 6 and 10.23
A
generic procedure is reported here for complexes 3–5 and 7–9: To
a stirring solution of 1 or 2 (0.50 g, 0.65 mmol) in toluene (50
mL) was added CuXn (3.25 mmol; X ) F, n ) 2; X ) Cl or Br,
(71) McMaster, W. H.; Del Grande, N. K.; Mallett, J. H.; Hubbel, J. H.
Lawrence LiVermore National Laboratory Report UCRL-t0174, Section
II ReVision 1; Lawrence Livermore National Laboratory: Livermore,
CA, 1969; p 70.
(72) Hubbell, J. H.; McMaster, W. H.; Del Grande, N. K.; Mallett, J. H.
In International Tables for X-Ray Crystallography; Ibers, J. A.,
Hamilton, W. C., Eds.; Kynoch Press: Birmingham, 1974; Vol. 4, pp
47–70.
9
5284 J. AM. CHEM. SOC. VOL. 130, NO. 15, 2008

Uranium(V)-Imido Halide Complexes
A R T I C L E S
n ) 1) as a solid. The reaction vessel was sealed, and the solution
was stirred at room temperature. After 12 h, the reaction was filtered
through a Celite-packed coarse-porosity fritted filter, the filtrate was
collected, and the volatiles were removed under reduced pressure.
The resulting dark brown solid was taken up into hexanes and
refiltered through a Celite-packed coarse-porosity fritted filter.
Solvents were removed from the filtrate to give the product as a
dark brown solid collected in isolated yields ranging from 75 to
89%. (See Supporting Information for additional details.)
slightly shorter for the RECP model (0.02 Å) compared to the all-
electron results. Bond angles differed by 2°, and very little
difference was observed in the orbital composition or energies of
the MOs. Unless otherwise specified, the results in the paper are
reported from the all-electron calculations, which were all carried
out using the Gaussian03 suite of codes.76
To estimate the effects of spin–orbit coupling on the 5f manifold,
a simple ligand field model was employed. The calculated f energy
levels and orbitals were obtained from a DFT calculation in which
one electron was averaged over the 5f manifold. Using a one-center
spin–orbit coupling operator with λSO ) 1800 cm-1, the 14 × 14
spin–orbit matrix was diagonalized to obtain the seven Kramers
doublet levels.
Caution:Depleted uranium (primary isotope 238U) is a weak
R-emitter (4.197 MeV) with a half-life of 4.47 × 109 years; manipula-
tions and reactions should be carried out in monitored fume hoods or
in an inert atmosphere drybox in a radiation laboratory equipped with
R- and ꢀ-counting equipment.
Acknowledgment. For financial support of this work, we
acknowledge LANL (Director’s PD Fellowship to C.R.G. and
Frederick Reines PD Fellowship to E.J.S. and S.A.K.), the
LANL G. T. Seaborg Institute (PD Fellowships to C.R.G. and
P.Y.; summer graduate student fellowship to A.E.V.), the
Division of Chemical Sciences, Office of Basic Energy Sciences,
Heavy Element Chemistry program, and the LANL Laboratory
Directed Research & Development program. The authors thank
Kevin S. Boland, Carol J. Burns, Juan S. Lezama, Marianne P.
Wilkerson, and Laura E. Wolfsberg (all LANL) for assistance
with the EXAFS measurement on complex 4.
Computational Methods. The B3LYP hybrid density func-
tional was employed to optimize the equilibrium molecular
structures of the model complexes (C5Me5)2U(dN-C6H5)(F) (11)
and (C5Me5)2U(dN-C6H5)(I) (12).73 The calculations incorporate
pentamethylcyclopentadienyl (C5Me5) ligands rather than adopt-
ing the computationally simpler strategy of using C5H5 as a
C5Me5 surrogate. The substituted imido groups (dN-2,4,6-tBu3-
C6H2) and (dN-2,6-iPr2-C6H3) were modeled as a phenyl group
(dN-C6H5) for simplicity. The Stuttgart RSC 1997 effective core
potential (ECP) was employed for uranium, which incorporates
scalar relativistic effects and replaces 60 core electrons.74 The
valence electrons are represented as [6s/6p/5d/3f]; 6-31G* basis
sets were used for carbon, hydrogen, and nitrogen. The fluorine
and iodine atoms were represented by 6-311G* basis sets. For the
iodide complex 12, calculations using a relativistic ECP (RECP)
for iodide using (Lanl2dz)75 basis plus one polarization function
were also compared with the all-electron calculations. The geom-
etries from the all-electron and RECP calculation were very similar,
with the main difference being that the U-I bond length was
Supporting Information Available: General experimental
consideration; full characterization details for 3-5, 7-9;
crystallographic data (CIF) for 2, 3, 5, 8, 9; XYZ data for 11,
12; and complete ref 76. This material is available free of charge
via the Internet at http://pubs.acs.org.
JA711010H
(73) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652.
(74) Kuechle, W.; Dolg, M.; Stoll, H.; Preuss, H. J. Chem. Phys. 1994,
100, 7535–7542.
(76) Frisch, M. J.; et al. Gaussian 03; Gaussian, Inc.: Wallingford, CT,
(75) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270–283.
2004.
9
J. AM. CHEM. SOC. VOL. 130, NO. 15, 2008 5285