Systematic studies of early actinide complexes: Thorium(IV) fluoroketimides

Article abstract of DOI:10.1021/ja0686458

Reaction of (C₅Me₅)₂Th(CH ₃)₂ with 2 equiv of N=C-Ar_F gives the corresponding fluorinated thorium(IV) bis(ketimide) complexes (C ₅Me₅)₂Th[-N=C(CH₃)(Ar _F)]₂ (where Ar_F = 3-F-C₆H ₄ (4), 4-F-C₆H₄ (5), 2-F-C₆H ₄ (6), 3,5-F₂-C₆H₃ (7), 3,4,5-F ₃-C₆H₂ (8), 2,6-F₂-C ₆H₃ (9), 2,4,6-F₃-C₆H₂ (10), and C₆F₅ (11)). The complexes have been characterized by a combination of single-crystal X-ray diffraction, cyclic voltammetry and NMR, and UV-visible absorption and low-temperature luminescence spectroscopies. Density functional theory (DFT) and time-dependent DFT (TD-DFT) results are reported for complexes 5, 11, and (C₅Me₅) ₂Th[-N=C(Ph)₂J₂ (1) for comparison with experimental data and to guide in the interpretation of the spectroscopic results. The most significant structural perturbation imparted by the fluorine substitution in these complexes is a rotation of the fluorophenyl group (Ar _F) out of the plane defined by the N=C(C_ME)-(C _ipso) fragment in complexes 9-11 when the Ar_F group possesses two Ortho fluorine atoms. Excellent agreement is obtained between the optimized ground state DFT calculated structures and crystal structures for 11, which displays the distortion, as well as 5, which does not. In complexes 9-11, the out-of-plane rotation results in large interplanar angles (φ) between the planes formed by ketimide atoms N=C(C_Me)-(C_ipso) and the ketimide aryl groups in the range φ = 49.1-88.8 °, while in complexes 5, 7, and 8, φ = 5.7-34.9 °. The large distortions in 9-11 are a consequence of an unfavorable steric interaction between one of the two ortho fluorine atoms and the methyl group [-N=C(CH₃)] on the ketimide ligand. Excellent agreement is also observed between the experimental electronic spectroscopic data and the TD-DFT predictions that the two lowest lying singlet states are principally of nonbonding nitrogen p orbital to antibonding C=N π* orbital (P_N→π*_C=N or nπ*) character, giving rise to moderately intense transitions in the mid-visible spectral region that are separated in energy by less than 0.1 eV. Low-temperature (77 K) luminescence from both singlet and triplet excited states are also observed for these complexes. Emission lifetime data at 77 K for the triplet states are in the range 50-400 μs. These emission spectral data also exhibit vibronic structure indicative of a small Franck-Condon distortion in the ketimide M-N=C(R₁)(R₂) linkage. Consistent with this vibronic structure, resonance enhanced Raman vibrational scattering is also observed for (C₅Me₅)₂Th[-N=C(Ph)(CH ₂Ph)]₂ (2) when exciting into the visible excited states. These systems represent rare examples of Th(IV) complexes that engender luminescence and resonance Raman spectral signatures.

Full text of DOI:10.1021/ja0686458

Published on Web 03/30/2007
Systematic Studies of Early Actinide Complexes: Thorium(IV)
Fluoroketimides
Eric J. Schelter, Ping Yang, Brian L. Scott, Ryan E. Da Re, Kimberly C. Jantunen,
Richard L. Martin, P. Jeffrey Hay,* David E. Morris,* and Jaqueline L. Kiplinger*
Los Alamos National Laboratory, Los Alamos, New Mexico 87545
Received December 2, 2006; E-mail: kiplinger@lanl.gov
Abstract: Reaction of (C5Me5)2Th(CH3)2 with 2 equiv of NtC-ArF gives the corresponding fluorinated
thorium(IV) bis(ketimide) complexes (C5Me5)2Th[-NdC(CH3)(ArF)]2 (where ArF ) 3-F-C6H4 (4), 4-F-C6H4
(5), 2-F-C6H4 (6), 3,5-F2-C6H3 (7), 3,4,5-F3-C6H2 (8), 2,6-F2-C6H3 (9), 2,4,6-F3-C6H2 (10), and C6F5 (11)).
The complexes have been characterized by a combination of single-crystal X-ray diffraction, cyclic
voltammetry and NMR, and UV-visible absorption and low-temperature luminescence spectroscopies.
Density functional theory (DFT) and time-dependent DFT (TD-DFT) results are reported for complexes 5,
11, and (C5Me5)2Th[-NdC(Ph)2]2 (1) for comparison with experimental data and to guide in the interpretation
of the spectroscopic results. The most significant structural perturbation imparted by the fluorine substitution
in these complexes is a rotation of the fluorophenyl group (ArF) out of the plane defined by the NdC(CMe)-
(Cipso) fragment in complexes 9-11 when the ArF group possesses two ortho fluorine atoms. Excellent
agreement is obtained between the optimized ground state DFT calculated structures and crystal structures
for 11, which displays the distortion, as well as 5, which does not. In complexes 9-11, the out-of-plane
rotation results in large interplanar angles (φ) between the planes formed by ketimide atoms NdC(CMe)-
(Cipso) and the ketimide aryl groups in the range φ ) 49.1-88.8 °, while in complexes 5, 7, and 8, φ )
5.7-34.9 °. The large distortions in 9-11 are a consequence of an unfavorable steric interaction between
one of the two ortho fluorine atoms and the methyl group [-NdC(CH3)] on the ketimide ligand. Excellent
agreement is also observed between the experimental electronic spectroscopic data and the TD-DFT
predictions that the two lowest lying singlet states are principally of nonbonding nitrogen p orbital to
antibonding CdN π* orbital (pNfπ*CdN or nπ*) character, giving rise to moderately intense transitions in
the mid-visible spectral region that are separated in energy by less than 0.1 eV. Low-temperature (77 K)
luminescence from both singlet and triplet excited states are also observed for these complexes. Emission
lifetime data at 77 K for the triplet states are in the range 50-400 µs. These emission spectral data also
exhibit vibronic structure indicative of a small Franck-Condon distortion in the ketimide M-NdC(R1)(R2)
linkage. Consistent with this vibronic structure, resonance enhanced Raman vibrational scattering is also
observed for (C5Me5)2Th[-NdC(Ph)(CH2Ph)]2 (2) when exciting into the visible excited states. These
systems represent rare examples of Th(IV) complexes that engender luminescence and resonance Raman
spectral signatures.
Introduction
complexes of the type (C5Me5)2AnR2 to afford bis(ketimide)
complexes (C5Me5)2An[-NdC(Ph)(R)]2 (where An ) Th, R
Organometallic actinide complexes containing metal-ligand
multiple bonds continue to provide a useful platform for
probing the involvement of 5f and 6d orbitals in bonding and
the effects of such bonding on electronic structure. Bond orders
greater than one engender tunable metal-based electronic
properties due to strong interactions of ligand and 5f- and/or
6d-actinide valence orbitals.1-3 Recently, we reported that
benzonitrile inserts into the actinide-carbon bonds in
) Ph, CH2Ph, CH3; An ) U, R ) CH2Ph, CH3) and determined
through structural, electrochemical, and spectroscopic studies
that the An-N interactions in these complexes possessed bond
orders greater than one.3-5 As few systematic studies on the
electronic structures across a homologous series of actinide
complexes have been reported,6-9 we were interested in
(4) Kiplinger, J. L.; Morris, D. E.; Scott, B. L.; Burns, C. J. Organometallics
2002, 21, 3073.
(5) Da Re, R. E.; Jantunen, K. C.; Golden, J. T.; Kiplinger, J. L.; Morris, D.
E. J. Am. Chem. Soc. 2005, 127, 682.
(6) Durakiewicz, T.; Joyce, J. J.; Lander, G. H.; Olson, C. G.; Butterfield, M.
T.; Guziewicz, E.; Arko, A. J.; Morales, L.; Rebizant, J.; Mattenberger,
K.; Vogt, O. Phys. ReV. B: Condens. Matter 2004, 70, 205103.
(7) Cramer, R. E.; Roth, S.; Edelmann, F.; Bruck, M. A.; Cohn, K. C.; Gilje,
J. W. Organometallics 1989, 8, 1192.
(8) Lukens, W. W.; Allen, P. G.; Bucher, J. J.; Edelstein, N. N.; Hudson, E.
A.; Shuh, D. K.; Reich, T.; Andersen, R. A. Organometallics 1999, 18,
1253.
(1) Burns, C. J.; Eisen, M. S. In The Chemistry of the Actinide and
Transactinide Elements, 3rd ed.; Morss, L. R., Edelstein, N. M., Fuger, J.,
Eds.; Springer: Dordrecht, 2006.
(2) Ephritikhine, M. Dalton Trans. 2006, 2501.
(3) (a) Jantunen, K. C.; Burns, C. J.; Castro-Rodriguez, I.; Da Re, R. E.; Golden,
J. T.; Morris, D. E.; Scott, B. L.; Taw, F. L.; Kiplinger, J. L. Organome-
tallics 2004, 23, 4682. (b) Morris, D. E.; Da, Re, R. E.; Jantunen, K. C.;
Castro-Rodriguez, I.; Kiplinger, J. L. Organometallics 2004, 23, 5142. (c)
Schelter, E. J.; Morris, D. E.; Scott, B. L.; Kiplinger, J. L. Chem. Commun.
2007, 1029.
9
10.1021/ja0686458 CCC: $37.00 © 2007 American Chemical Society
J. AM. CHEM. SOC. 2007, 129, 5139-5152
5139

A R T I C L E S
Schelter et al.
exploiting the (C5Me5)2An bis(ketimide) framework to prepare
a series of complexes from fluorinated nitriles to tune electronic
properties of the resulting complexes.
Our previous reports on the thorium bis(ketimide) complexes
(C5Me5)2Th[-NdC(Ph)(R)]2 R ) Ph (1), CH2Ph (2), or CH3
(3) describe their intense yellow-to-orange colors.3a
engender resonance-enhanced Raman spectra when exciting into
the visible electronic transitions.
Experimental Section
Instrumentation and Sample Protocols. Electronic absorption
spectral data were obtained for toluene solutions of all complexes over
the wavelength range 300-1600 nm on a Perkin-Elmer model Lambda
950 UV-visible-near-infrared spectrophotometer. All data were
collected in 1-cm path length cuvettes loaded in a recirculating Vacuum
Atmospheres model HE-553-2 inert atmosphere (N2) drybox with a
MO-40-2 Dri-Train and run versus a toluene reference. Spectral
resolution was typically 2 nm.
Emission spectral data were collected on either a Spex Industries
Fluorolog 2 system or a Photon Technology International (PTI) model
QM-4 system. The former is comprised of a 450 W Xe arc lamp CW
source, a 0.22-m single-stage excitation monochromator, a 0.22-m
double emission monochromator, both with 1200 grove/mm gratings,
and a thermoelectrically cooled Hamamatsu R928 photomultiplier
detector. The PTI system comprises 0.2-m single-stage emission and
excitation monochromators with 1200 grove/mm gratings. It can be
implemented in either a continuous-wave excitation mode using a 75
W Xe arc lamp source, or a pulsed (time-resolved) excitation mode
using a Xe flashlamp. Both CW and time-resolved detectors employ
Hamamatsu R928 photomultiplier tubes. Spectral resolution in all
emission scans was <2 nm. Emission spectral samples were typically
1-10 mM in toluene or THF and were prepared in the Vacuum
Atmospheres drybox system described above; data were collected on
samples in flame-sealed medium-walled NMR tubes (Wilmad 504-PP)
at 77 K as glasses or ices using a liquid-nitrogen-cooled optical dewar.
Resonance Raman spectra were obtained using excitation (λex) from
a Spectra Physics 2045 Ar-ion laser or an Ar-ion-pumped Coherent
Radiation 599 dye laser charged with trans-stilbene. Scattered light
was collected and directed into a Jobin Yvon HR640 spectrometer
operated as a single-stage spectrograph equipped with an 1800 grooves/
mm grating; the single-stage spectrograph requires that the scattered
light be passed through a holographic notch filter (Kaiser) to reduce
Toluene solutions of these complexes were found to have
molar extinction coefficients in the visible region of ∼500 M-1
cm-1; such intensities in the visible region are quite unusual
for complexes of the 6d05f0 Th(IV) ion.10 The electronic
transitions responsible for the colors of the thorium bis(ketimide)
species were subsequently assigned to ligand-centered non-
bonding pNfantibonding π*CdN (nπ*) transitions, but the
orbitals involved in these states were determined to contain
contributions from metal-based valence orbitals.11 Interestingly,
this signature nπ* electronic transition for 1 was found to be
significantly red-shifted from that observed for 2 and 3. This
observation suggests that a strong substituent effect exists upon
changing the R group of the ketimide ligand from the similar
-CH3 and -CH2Ph groups to -Ph, an effect that may be
attributed to modulation of the thorium ketimide bonding
interaction and/or the result of changes in the symmetry of the
ketimide ligand itself.
Given these significant changes in the ligand-based electronic
transitions for the thorium bis(ketimide) complexes 1-3 on
changing from alkyl to phenyl substituents, we prepared a series
of complexes designed to systematically perturb their electronic
structure, perhaps even to the extent of inverting the ordering
of energetically close d and f states of the Th(IV) ion12,13 by
virtue of the electron-withdrawing nature of the R substituent,
thereby inducing new ground- and excited-electronic states
derived from thorium f-orbital parentage. Herein we report a
series of ketimide complexes possessing fluorinated aromatic
groups with an eye toward establishing trends in their respective
electronic structures based on electron-withdrawing ability.
Detailed structural, electrochemical, and spectroscopic data have
been obtained for this series to explore and quantify the
manifestations of substituent-induced changes in electronic
effects arising from the ketimide ligands. Results from density
functional theory calculations are also reported for a subset of
these complexes to guide the interpretation of the structural and
spectroscopic data. Finally, we demonstrate that these thorium
bis(ketimide) species are rare examples of tetravalent actinide-
containing complexes that exhibit luminescence, with measured
emission lifetimes at 77 K in the range of 50-400 µs, and also
contributions from Rayleigh scattering. Typical resolution was ∼5 cm-1
.
For all excitation wavelengths, scattering intensities were recorded using
a Princeton Instruments liquid-nitrogen-cooled CCD detector. A
polarization scrambler was placed at the entrance slit of the spectrograph
to minimize distortions in the observed scattering intensities due to
wavelength-dependent instrument response to polarized light. Typical
excitation powers and integration times were 20-50 mW and 15-120
min, respectively. Raman shifts were calibrated against an external
reference of 4-acetamidophenol, and spectra were corrected by subtrac-
tion of a spectral baseline obtained with a spline-fitting procedure. The
dispersion characteristics of our instrumentation typically require the
acquisition and concatenation of several spectral windows to encompass
the entire Raman shift range from ∼150 to 1700 cm-1 (e.g., two
windows are required for 457.9 nm excitation); windows were selected
such that adjacent windows contained sufficient overlap of Raman bands
to derive a continuous trace. Peak positions were determined by fitting
band profiles. Samples for Raman studies were prepared using the same
protocols described above for the emission spectral samples.
Cyclic voltammetric data were obtained in the Vacuum Atmospheres
drybox system described above. All data were collected using a Perkin-
Elmer Princeton Applied Research Corporation (PARC) model 263
potentiostat under computer control with PARC model 270 software.
All sample solutions were ∼5 mM in complex with 0.1 M [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 minimize contribution to the voltammetric
waves from uncompensated solution resistance (typically ∼1 kΩ under
the conditions employed). 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 silver wire quasi-reference
(9) Gaunt, A. J.; Scott, B. L.; Neu, M. P. Inorg. Chem. 2006, 45, 7401.
(10) Katzin, L. I.; Sonnenberger, D. C. In The Chemistry of the Actinide
Elements, 2nd ed.; Katz, J. J., Seaborg, G. T., Morss, L. R., Eds.; Chapman
& Hall: New York, 1986.
(11) Clark, A. E.; Martin, R. L.; Hay, P. J.; Green, J. C.; Jantunen, K. C.;
Kiplinger, J. L. J. Phys. Chem. A 2005, 109, 5481.
(12) Kaltsoyannis, N.; Bursten, B. E. J. Organomet. Chem. 1997, 528, 19.
(13) Li, J.; Bursten, B. E. J. Am. Chem. Soc. 1997, 119, 9021.
9
5140 J. AM. CHEM. SOC. VOL. 129, NO. 16, 2007

Studies of Early Actinide Complexes
A R T I C L E S
electrode. Scan rates from 20 to 5000 mV/s were employed to assess
the chemical and electrochemical reversibility of the observed redox
transformations. Potential calibrations were performed at the end of
each data collection cycle using the ferrocenium/ferrocene couple as
an internal standard.
Results and Discussion
Structural Chemistry. Equation 1 displays the synthetic
methods used and the yields obtained in the preparation of the
fluorinated thorium(IV) ketimide complexes. Treatment of a
colorless toluene solution of (C5Me5)2Th(CH3)2 with 2 equiv
of nitrile at room temperature gives the corresponding thorium-
(IV) bis(ketimide) complexes 4-11 as yellow or orange solids
in 44-95% isolated yields. Following workup and crystallization
from pentane solution, all complexes were reproducibly isolated
Electronic absorption/emission, Raman, and cyclic voltammetric data
were analyzed using Wavemetrics IGOR Pro (version 4.0) software
on a Macintosh platform or KaleidaGraph (version 3.5) on a Windows
platform.
Synthesis. The synthesis of compounds 4-11 is similar to that
previously reported for 1-3. A generic procedure is reported here for
complexes 4-11. For all cases, in an inert atmosphere drybox, toluene
solutions of fluorinated nitrile were added dropwise to stirred colorless
toluene solutions of (C5Me5)2Th(CH3)2 contained in a 125-mL flask.
The excess nitrile was then rinsed into the reaction using the toluene
mixture. Within 30 s of stirring, the reaction mixture turned bright
yellow or orange and was stirred at room temperature for 14 h. The
volatiles were then removed under reduced pressure, producing an oily
residue or solid, which was dissolved in pentane and filtered hot through
a Celite-padded coarse frit in all cases. Pure product was obtained either
by cooling the concentrated filtrate to -30 °C overnight and isolating
the crystalline product by decanting the supernatant or by leaving the
filtrate to evaporate to dryness at RT for ∼1 week. The product was
further dried under reduced pressure and collected with isolated yields
ranging from 44 to 95%. See Supporting Information for additional
details.
1
in pure form and were characterized by a combination of H/
13C/19F NMR and UV/vis/NIR absorption and emission spec-
troscopy, electrochemistry, elemental analysis, and X-ray crys-
tallography. The modest yields for some complexes reflect the
high solubility of the bis(ketimide) complexes imparted by the
fluorinated ligands.
Computational Methods. Three complexes were selected as
representatives to study the fluorine substitution effects by density
functional theory: (C5Me5)2Th[-NdC(Ph)2]2 (1), (C5Me5)2Th[-Nd
C(CH3)(4-F-C6H4)]2 (5), and (C5Me5)2Th[-NdC(CH3)(C6F5)]2 (11).
-
The calculations incorporate pentamethylcyclopentadienyl (C5Me5
)
ligands rather than adopting the computationally simpler strategy of
using C5H5 as a C5Me5 surrogate. The B3LYP hybrid density functional
was employed to optimize the equilibrium molecular structures of all
complexes.14 The Stuttgart RSC 1997 effective core potential (ECP)
was employed for thorium,15 where 60 core electrons are replaced to
account for scalar relativistic effects. The valence electrons are
represented as [8s/7p/6d/4f]; 6-31G* basis sets were used for carbon,
hydrogen, nitrogen, and fluorine. All calculations were carried out with
the Gaussian03 suite of codes.16 The excited states were calculated
using the time-dependent density functional theory (TD-DFT) approach
at optimized ground state geometries.17-21 The characterization of
excited particle and hole interactions for a given excited state was
described by natural transition orbitals (NTOs).22
In contrast to those involving transition metals, synthetic
strategies for the preparation of actinide organometallic fluo-
rocarbon complexes are still rare. This has been commonly
attributed to the fluorophilicity of the electropositive actinide
metal centers, leading to fluoride abstraction chemistry and
ultimately preventing more desirable reaction pathways.23 The
system reported here clearly demonstrates that organometallic
complexes of actinides containing fluorinated ligands are viable,
contrary to the convention that suggests that the strength of
An-F bonds would drive their decomposition.23-25
Single-crystal X-ray diffraction studies on 5-11 confirmed
their formulation as thorium(IV) bis(ketimide) complexes and
were consistent with the NMR data in all cases (Figure 1 and
Supporting Information). Experimental crystallographic details
are listed in Table 1, while salient geometric parameters are
shown in Table 2. Complexes 4-11 display typical bent-
metallocene geometries with the ketimide ligands residing within
the metallocene wedge. Previous reports of actinide ketimide
complexes have stressed the importance of the An-NdC angle
and An-N bond distance in discussion of bond order and
electronic structure effects. Nitrogen lone pairs on the ketimide
ligand have the potential to interact with the metal ion in both
σ- and π-modes, resulting in an intermediate An-N bond order
between 1 and 2.3,4 Interestingly, lanthanide ketimide complexes
also exhibit the nearly linear M-NdC angles previously
associated with this two-fold metal-ligand interaction in
Calculations were also performed for the ketimide ligand anions
[-NdC(CH3)(ArF)]-. These were modeled after the ketimide ligands
found in the corresponding thorium bis(ketimide) complexes. Full
geometry optimizations were carried out for the closed shell negative
ions of these ligands. Additionally, calculations were performed for
lithium and potassium salts of the ketimide anions, but the bare ligand
anion results were the most chemically meaningful.
(14) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.
(15) Kuchle, W.; Dolg, M.; Stoll, H.; Preuss, H. J. Chem. Phys. 1994, 100,
7535.
(16) Frisch, M. J.; et al. Gaussian03, Revision C.02, Gaussian, Inc.: Wallingford,
CT, 2004.
(17) Jamorski, C.; Casida, M. E.; Salahub, D. R. J. Chem. Phys. 1996, 104,
5134.
(18) Casida, M. E.; Jamorski, C.; Casida, K. C.; Salahub, D. R. J. Chem. Phys.
1998, 108, 4439.
(19) Bauernschmitt, R.; Ahlrichs, R.; Hennrich, F. H.; Kappes, M. M. J. Am.
Chem. Soc. 1998, 120, 5052.
(20) Stratmann, R. E.; Scuseria, G. E.; Frisch, M. J. J. Chem. Phys. 1998, 109,
8218.
(23) Cuellar, E. A.; Marks, T. J. Inorg. Chem. 1981, 20, 2129.
(24) Weydert, M.; Andersen, R. A.; Bergman, R. G. J. Am. Chem. Soc. 1993,
115, 8837.
(25) Dorn, H.; Murphy, E. F.; Shah, S. A. A.; Roesky, H. W. J. Fluorine Chem.
1997, 86, 121.
(21) Petersilka, M.; Gossmann, U. J.; Gross, E. K. U. Phys. ReV. Lett. 1996,
76, 1212.
(22) Martin, R. L. J. Chem. Phys. 2003, 118, 4775.
9
J. AM. CHEM. SOC. VOL. 129, NO. 16, 2007 5141

A R T I C L E S
Schelter et al.
falling in between Th(IV)-N amide (2.299(7)-2.398(3) Å) and
Th(IV)dN imide (2.045(8) Å) bond distances in structurally
related complexes.28-32 The Th-N ketimide bonds in complexes
4-11 span the range 2.250(2)-2.286(5) Å. Although the high
end of this range begins to approach the Th-N bond distances
observed for structurally relevant Th(IV)-amide complexes,
overall the Th-N π-bonding interactions in thorium ketimide
complexes produce a shortening of ∼0.1 Å compared to Th-N
bond distances in amide systems.
Within the series of complexes, the bond distances and angles
do not reveal any particular structural changes that track with
changes in their physicochemical properties. Such metrics as
the Th-N bond distances are generally within 3σ of each other
across the series. The lack of structural and spectroscopic trends
(Vide infra) implies subtlety and competing effects in the global
determination of energy states in these complexes; however,
electronic structure calculations identified a single primary effect
leading to the observed geometries of the complexes. Fluorine
substitution at the ortho position on the ketimide aryl group
imposes an unfavorable steric demand that forces the aryl group
to be rotated orthogonal to planes formed by the corresponding
ketimide NdC(CMe)(Cipso) atoms. This out-of-plane rotation can
be observed by comparing the molecular structures for com-
plexes 11 and 8 in Figure 1. As demonstrated in Figure 2, the
amount of rotation can be quantified by analyzing the torsion
angles that describe the smallest interplanar angle formed by
atoms N(1)-C(21)-C(23)-C(24/28). Looking down the C(21)-
C(23) axis, out-of-plane distortions of the aryl ring result from
either clockwise or counterclockwise rotation of the group along
the C(21)-C(23) axis. The interplanar angles (φ) range from
0° (aryl groups planar to ketimide atoms NdC(CMe)(Cipso)) to
90° (orthogonal aryl groups). As shown in Table 2, two distinct
structure types emerge from this analysis: complexes 5, 7, and
8 exhibit interplanar angles in the range φ ) 5.7-34.9° while
6, with its single site of ortho-fluorination, has φ ) 24.7 and
48.0°. Compounds 9-11 exhibit significantly larger values in
the range φ ) 49.1-88.8°.
Figure 1. Thermal ellipsoid plots of complexes 11 (top) and 8 (bottom)
projected at the 30% probability level. Torsion angles formed by atoms
N(1)-C(11)-C(13)-C(14) quantify the relative planarity of the ketimide
ligands in 8 and nonplanarity in 11 as a result of the fluorine substitution
at the ortho positions.
The structure of 7 and one of the two molecules in the
asymmetric unit of 9 display an additional type of distortion in
which a ketimide ligand is rotated nearly 180° around its Th-N
bond axis such that the ketimide methyl group points ap-
proximately “inward” toward the aryl group of the second
ketimide ligand (see Supporting Information). Because both of
these rotational conformations are present in the structure of 9,
this suggests that a low rotational energy barrier exists for the
ketimide ligand, which is consistent with the ambient-temper-
ature NMR spectroscopy data. The observation of this rotational
variation in the structures of 7 and 9 is presumably due to crystal
packing forces; the rotational conformers adhere to the ketimide-
aryl group planarity/nonplanarity described above. In total, the
structural data imply an overall destabilization in compounds
9-11 due to disruption of conjugation of the fluorinated
aromatic groups with the aza-vinyl moieties in the ketimide
ligands.
Figure 2. Views orthogonal to (top) and along (bottom) the C(21)-C(23)
bond axis of complex 9. As highlighted in red, the torsion angle N(1)-
C(21)-C(23)-C(24) ) -121.4 ° defines the interplanar angle φ ) 58.6°.
actinide complexes.26,27 Given the small radial extension of 4f
orbitals and the absence of π-bonding modes in Ln-ketimide
complexes, the most important metric for f-element ketimide
complexes is decidedly the metal-nitrogen distance.
(28) Gilbert, T. M.; Ryan, R. R.; Sattelberger, A. P. Organometallics 1988, 7,
2514.
The Th-N distances in complexes 4-11 are comparable to
the values reported for 1 and 2 and show a similar shortening,
(29) Barnhart, D. M.; Clark, D. L.; Grumbine, S. K.; Watkin, J. G. Inorg. Chem.
1995, 34, 1695.
(30) Korobkov, I.; Gambarotta, S.; Yap, G. P. A. Angew. Chem., Int. Ed. 2003,
42, 814.
(26) Evans, W. J.; Forrestal, K. J.; Ziller, J. W. J. Am. Chem. Soc. 1998, 120,
9273.
(27) Hou, Z. M.; Yoda, C.; Koizumi, T.; Nishiura, M.; Wakatsuki, Y.; Fukuzawa,
S.; Takats, J. Organometallics 2003, 22, 3586.
(31) Haskel, A.; Straub, T.; Eisen, M. S. Organometallics 1996, 15, 3773.
(32) Straub, T.; Haskel, A.; Neyroud, T. G.; Kapon, M.; Botoshansky, M.; Eisen,
M. S. Organometallics 2001, 20, 5017.
9
5142 J. AM. CHEM. SOC. VOL. 129, NO. 16, 2007

Studies of Early Actinide Complexes
A R T I C L E S
Table 1. Crystallographic Experimental Parameters for 5-11
5
6
7
8
9
10
11
formula
a (Å)
b (Å)
C36H44F2N2Th
12.9201(6)
14.0054(7)
18.2113(9)
90
98.677(1)
90
3257.6(3)
4
C36H44F2N2Th
10.180(5)
17.880(8)
18.364(8)
90
90
90
3343(3)
4
C36H42F4N2Th
15.568(2)
11.3893(17)
18.841(3)
90
90.668(3)
90
3340.3(9)
4
C36H40F6N2Th
13.087(3)
13.927(3)
18.783(5)
90
99.272(3)
90
3378.7(14)
4
C36H42F4N2Th
10.7290(5)
17.7108(8)
19.3565(9)
108.7590(10)
90.1920(10)
102.9020(10)
3383.5(3)
4
C36H40F6N2Th
14.566(2)
16.206(2)
15.119(2)
90
94.027(2)
90
3560.1(8)
4
C36H36F10N2Th
16.066(3)
16.393(3)
26.997(5)
90
90
90
7110(2)
8
c (Å)
R (deg)
â (deg)
γ (deg)
V (Å3)
Z
FW
space group
T (K)
774.77
C2/c
141(2)
0.71073
1.580
4.615
774.77
P212121
203(2)
0.71073
1.540
4.498
0.0359
0.0794
810.76
P21/c
203(2)
0.71073
1.612
4.513
0.0413
0.0860
846.74
P21/c
141(1)
0.71073
1.665
4.474
0.0145
0.0372
810.76
P1h
141(2)
0.71073
1.592
4.455
0.0442
0.1092
846.74
C2/c
203(2)
0.71073
1.580
4.246
0.0309
0.0790
918.71
Pccn
141(2)
0.71073
1.717
4.275
0.0249
0.0594
λ (Å)
D
calc (g‚cm-3
)
µ (mm-1
)
R1 (I > 2σ(I))
wR2 (all data)
0.0204
0.0578
Table 2. Selected Metrical Parameters for Complexes 1, 2, and 5-11 from X-ray Diffraction Data
1
2
5
6
7
8
9
10
11
Th-N (Å)
2.259(4)
2.265(5)
2.256(8)
2.250(2)
2.260(5)
2.269(5)
2.263(5)
2.286(5)
2.2727(19)
2.255(5)
2.261(5)
2.255(5)
2.263(5)
1.272(8)
1.280(8)
1.262(8)
1.273(8)
171.4(5)
177.1(5)
171.9(5)
167.3(5)
107.35(19)
104.8(2)
56.2
2.263(3)
1.253(5)
174.7(4)
2.268(3)
2.273(3)
NdC (Å)
1.259(7)
1.272(7)
1.264(13)
161.2(8)
102.9(4)
1.259(4)
175.5(2)
1.265(8)
1.248(8)
1.245(7)
1.268(7)
1.256(3)
170.5(2)
1.253(4)
1.251(4)
Th-NdC (deg)
174.0(4)
179.4(5)
176.6(5)
174.7(5)
172.8(4)
170.8(4)
176.0(2)
177.1(2)
N-Th-N (deg)
φ (deg)a
108.95(17)
108.83(13)
8.4
108.65(19)
107.37(17)
109.72(11)
5.7
108.2(2)
72.4
105.96(13)
104.90(13)
68.5
24.7
48.0
34.9
27.7
88.8
75.6
49.1
57.7
a See Figure 2 for definition.
state structure and data obtained from the crystal structure for
11. Excellent agreement is obtained between the experimental
and calculated structures as shown in Figure 3. For complexes
5 and 11, the most significant difference in the optimized
structures is the orientation of the ketimide aryl ring; it lies in
the NdC(CMe)(Cipso) plane for complex 5, while it is rotated
out of the plane (φ ) 67.7° calc versus 75.6 and 68.5° expt)
for complex 11. As discussed in detail below, this appears to
arise principally from unfavorable steric interactions between
the ortho fluorine atoms on the aryl ring and the ketimide ligand
methyl group [-NdC(CH3)]. Finally, in the optimized structure
for complex 1, one phenyl ring lies in this plane, while
unfavorable steric interactions rotate the second phenyl ring out
of the plane.
Figure 3. Overlapping atomic coordinates obtained from DFT optimized
ground-state structure (blue) and the experimental crystal structure (yellow)
for complex 11. The representation demonstrates the excellent agreement
between calculated and observed ground-state geometries for this complex.
In terms of the canonical molecular orbitals derived from the
ground-state properties, several observations can be made
regarding the highest occupied molecular orbitals (HOMOs).
In addition to the two orbitals that are predominantly nitrogen
lone pair in character, there are four other orbitals in the same
energy region arising from the π-systems of the pentamethyl-
cyclopentadienyl (C5Me5) ligands. These comprise the two pairs
of degenerate π-orbitals that are the HOMOs in the two isolated
pentamethylcyclopentadienyl ligands. Overall, the set of six
orbitals have energies differing by only 0.8 eV (0.03 au) as
Computational Studies. The structures of the ground states
for the three thorium bis(ketimide) complexes (C5Me5)2Th[-
NdC(Ph)2]2 (1), (C5Me5)2Th[-NdC(CH3)(4-F-C6H4)]2 (5), and
(C5Me5)2Th[-NdC(CH3)(C6F5)]2 (11) were calculated using
DFT approaches. Figure 3 shows overlapping projections of the
atomic coordinates obtained from the DFT optimized ground-
9
J. AM. CHEM. SOC. VOL. 129, NO. 16, 2007 5143

A R T I C L E S
Schelter et al.
Table 3. Mulliken Analysis of the Contributions (%) to the
Molecular Orbitals in Complexes 5 and 11
eV compared to (C5Me5)2Th[-NdC(Ph)2]2 (1). Similarly, a very
small deviation (0.04 eV) of the calculated singlet excitation
energies is observed between (C5Me5)2Th[-NdC(CH3)(4-F-
C6H4)]2 (5) and its model complex Cp2Th[-NdC(CH3)(4-F-
C6H4)]2 (5a).
Th
d
CdN
orbital
E (au)
sp
f
N
C
ArF
C5Me5
(C5Me5)2Th[-NdC(CH3)(4-F-C6H4)]2 (5)
As noted, both S1 and S2 states arise from promotion of an
electron from a nonbonding nitrogen p orbital to an antibonding
CdN π* orbital (pNfπ*CdN).11 While this is less evident from
the MO parentages in the TD-DFT calculation where each state
has four to six major contributions, it becomes more easily
visualized when viewed in terms of natural transition orbitals
(NTO), which give the most compact particle-hole description
of each state.22 For complex 1, this S1 representation collapses
to a single excitation involving an occupied “hole” orbital (H)
with predominantly nitrogen character and a virtual “particle”
(P) orbital involving the π-systems of the ketimide (CdN) and
aryl groups, both localized on a single ketimide ligand. The
corresponding S2 state involves a similar excitation on the
second ketimide ligand. This result is very similar to earlier
calculations on the model complex 1a.11 For complexes 5 and
11, a parallel picture emerges with the exception that the particle
and hole orbitals are delocalized over both ketimide ligands
(Table 5). In contrast to 1 and 1a, complexes 5 and 11 possess
a mirror plane of symmetry that passes between the two ketimide
ligands, and the particle and hole orbitals are either symmetric
(S) or antisymmetric (A) with respect to this mirror plane. For
the S1 states, the combination involves occupied to virtual orbital
AfA excitations in both complexes 5 and 11. However, for
the S2 states, the combination involves an occupied to virtual
orbital AfS excitation for 5 and SfA excitation for 11. These
lowest excited states are mainly localized on the ketimide ligands
with small contributions from pentamethylcyclopentadienyl
ligands rings based on the NTO representations. As a conse-
quence, the experimental spectroscopic data for the complexes
are expected to show some significant variability for the different
ketimide ligands.33
LUMO+1 -0.045 3.1
8.8 18.2 16.4 29.8 17.4
2.8
7.6
LUMO
HOMO
-0.050 0.0
-0.187 0.3
9.3 21.6 14
29.2 14.8
0.1
4.0
2.2
9.0
9.1 33.8
0.6 25.2
3.5 29.7
1.4
1.0
1.4
1.4
2.6
1.6
6.0 46.0
3.8 56.6
7.4 48.2
0.6 80.9
5.0 70.9
7.0 41.1
HOMO-1 -0.192 4.8
HOMO-2 -0.193 4.3
HOMO-3 -0.202 0.4
HOMO-4 -0.203 0.1 11.4
HOMO-5 -0.217 0.0 17.2
4.0
1.9
3.2
6.8
0.7 28.5
(C5Me5)2Th[-NdC(CH3)(C6F5)]2 (11)
6.0 12.1 16.4 24.0 34.2
6.7 14.9 14.4 24.2 31.2
LUMO+1 -0.043 2.2
1.8
5.4
LUMO
HOMO
-0.049 0.0
-0.193 2.1
0.5
3.5
4.7
6.7
5.7
5.8
0.2
0.6
1.8
1.8
2.8
1.8
0.4 84.3
1.2 74.1
6.0 47.6
4.8 43.2
0.8 79.4
8.8 16.6
HOMO-1 -0.198 4.5
HOMO-2 -0.206 1.4
HOMO-3 -0.211 1.5
0.8 13.2
5.7 29.6
4.9 33.2
HOMO-4 -0.213 0.0 12.5
HOMO-5 -0.225 0.2 17.6
1.8
0.7 48.4
1.6
shown in Table 3. For (C5Me5)2Th[-NdC(CH3)(4-F-C6H4)]2
(5), there is much more nitrogen character in the HOMO
(33.8%), while in (C5Me5)2Th[-NdC(CH3)(C6F5)]2 (11) the
HOMO has only 5.8% nitrogen character. In contrast, there is
only 46% contribution from the C5Me5 ligands to the HOMO
in complex 5 versus 84% in 11. The lowest two unoccupied
orbitals (LUMOs) predominantly consist of the NdC and aryl
ring π* orbitals with small contributions from thorium 6d- and
5f-orbitals. For (C5Me5)2Th[-NdC(CH3)(4-F-C6H4)]2 (5), this
latter contribution (9% 6d and 18-22% 5f) is larger than that
found in (C5Me5)2Th[-NdC(CH3)(C6F5)]2 (11) (6% 6d and
12-15% 5f).
To assess the role of fluorine substitution on the ketimide
aryl groups and the impact of the sterically induced structural
distortion on the energies of the electronic states, TD-DFT
studies were performed on complexes 1, 5, and 11 for
comparison with spectroscopic results discussed below. The
calculated excited-state energies from TD-DFT calculations at
the optimized ground-state geometries are summarized in Table
4. There is a pair of nearly degenerate singlet states for each
molecule (for example, 2.69 (S1) and 2.76 eV (S2) for complex
5). In addition, the calculated S1 excitation energies quantita-
tively mirror the experimental increase (vide infra) as one goes
along the series of three complexes: complex 1, 2.56 eV calc
versus 2.48 eV expt; complex 5, 2.69 eV calc versus 2.62 eV
expt; complex 11, 3.03 eV calc versus 2.96 eV expt. We also
note that the predicted S1 and S2 states for the model complex
Cp2Th[-NdC(Ph)2]2 (1a)11 were uniformly higher by only 0.13
In order to better understand the overall trends in the
excitation energies of the fluorinated bis(ketimide) complexes
4-11, the excited states of isolated ketimide ligands were also
systematically examined using DFT techniques. In lieu of time-
intensive excited-state calculations on all the complexes in the
series, isolated ketimide ligand calculations were performed to
obtain direct information on the S1 energies for comparison
without perturbation from metal-based orbitals. The closed shell
negative ions of the ligands, [-NdC(CH3)(ArF)]-, were chosen
as the most chemically meaningful set for this purpose, after
examining other alternatives such as alkali metal-ligand
compounds. The results are displayed in Figure 4, where the
Table 4. Excitation Energies from TD-DFT Calculations for Complexes 1, 1a, 5, 5a, and 11
expta
calc
calc
(C5Me5)2Th[-NdC(Ph)2]2 (1)
(C5H5)2Th[-NdC(Ph)2]2 (1a)
S1
S2
2.48
2.64
2.56
2.61
2.69b
2.75b
(C5Me5)2Th[-NdC(CH3)(4-F-C6H4)]2 (5)
(C5H5)2Th[-NdC(CH3)(4-F-C6H4)]2 (5a)
S1
S2
2.62
2.85
2.69
2.76
2.65
2.79
(C5Me5)2Th[-NdC(CH3)(C6F5)]2 (11)
S1
S2
2.96
2.97
3.03
3.05
-
-
a
b
From Gaussian fits of ambient temperature electronic absorption spectral data. See text.
9
5144 J. AM. CHEM. SOC. VOL. 129, NO. 16, 2007

Studies of Early Actinide Complexes
A R T I C L E S
Table 5. Singlet Natural Transition Orbitals for Complexes 1, 5, and 11
calculated S1(L) excited-state energy of the [-NdC(CH3)-
(ArF)]- ligand at the optimized ground-state geometry is
compared with that of the average energy of the S1 and S2
transitions [Save(expt)] from the corresponding thorium bis-
(ketimide) complex absorption spectroscopy data. The results
are presented with the [-NdC(CH3)(ArF)]- ligands in order
of roughly increasing Save(expt) energy for the corresponding
thorium bis(ketimide) complexes.
of [-NdC(CH3)(2,4,6-F3-C6H2)]- lying somewhat lower com-
pared to the energies of the rest of the series. The biggest
difference arises for the three ketimide ligands [-NdC(CH3)-
(2,6-F2-C6H3)]-, [-NdC(CH3)(2,4,6-F3-C6H2)]-, and [-Nd
C(CH3)(C6F5)]- where there is an unfavorable steric interaction
between one of the two ortho fluorine atoms and the ketimide
moiety [-NdC(CH3)]. In these three structures the fluorinated
aryl ring is rotated out of the NdC(CMe)(Cipso) plane to
accommodate this repulsive interaction (see Supporting Infor-
mation for pictorial representations). As a result of these
nonplanar ligand structures, the conjugation of the π-orbitals is
The overall trend for the thorium complexes of increasing
excitation energy is also qualitatively followed for the free
[-NdC(CH3)(ArF)]- ketimide ligands, with the S1(L) energy
9
J. AM. CHEM. SOC. VOL. 129, NO. 16, 2007 5145

A R T I C L E S
Schelter et al.
Figure 5. Cyclic voltammetric data for ∼5 mM (C5Me5)2Th[-NdC(R1)-
(R2)]2 (where R1, R2 ) Ph (1); R1 ) Ph, R2 ) CH2Ph (2); R1 ) Ph, R2 )
CH3 (3); R1 ) CH3, R2 ) 3,5-F2-C6H3 (7); R1 ) CH3; R2 ) 3,4,5-F3-C6H2
(8); R1 ) CH3, R2 ) 2,4,6-F3-C6H2 (10)) in ∼0.1 M [Bu4N][B(C6F5)4]/
THF at room temperature. All scans were collected at 200 mV/s. Vertical
arrows indicate the rest (initial) potential for the scans and the horizontal
arrows indicate the initial scan direction.
electron into each of two nearly degenerate ketimide-based
orbitals consistent with the theoretical results for the LUMOs
described above. The apparent absence of multiple reduction
waves in some of these thorium bis(ketimide) complexes is
likely a reflection of close spacing of multiple waves that
precludes resolution of discrete waves in the voltammetry and/
or the proximity of these processes to the negative potential
limit in this supporting electrolyte/solvent system that prohibits
observation of analyte behavior. In general, this electrochemical
behavior parallels the spectroscopic data described in detail
below, indicating that there are two discrete, nondegenerate,
low-lying virtual orbitals (∆E ≈ 0.1 eV) available to accept
the added electrons.
Figure 4. Comparison of DFT calculated transition energies for ketimide
ligand anions (green line), (C5Me5)2Th[-NdC(Ph)2]2 (1) and select (C5-
Me5)2Th[-NdC(CH3)(ArF)]2 complexes (blue points), and averaged ex-
perimental electronic absorption data for the two, nearly degenerate
transitions S1 and S2 (red line).
also disrupted in the corresponding three thorium complexes
(9-11). In the case of a single ortho fluorine substituent, the
ground-state structure for [-NdC(CH3)(2-F-C6H4)]- adopts a
planar arrangement with the aryl group rotated such that the
F-CAr carbon atom is trans to the ketimide nitrogen atom with
respect to the NC-Cipso bond.
Electronic Absorption Spectroscopy. As noted in previous
investigations the unusual intense color of the thorium bis-
(ketimide) complexes in the solid state and in solution derives
from the existence of a moderately intense, broad electronic
transition in the mid-visible spectral region.3,11 The electronic
spectral data for complexes 1-11 are shown in Figure 6. Clearly
all 11 complexes exhibit this same hallmark spectral feature,
although the energy and intensity varies significantly across the
series, reflecting electronic perturbation by the fluorinated aryl
substituents. TD-DFT calculations have predicted this significant
variability across the ketimide derivatives with the lowest-energy
electronic absorption spectral features in the thorium bis-
(ketimide) complexes derived from transitions best described
as pNfπ*CdN, with two nearly degenerate transitions (S1 and
S2) of this description. In fact, all the spectra shown in Figure
6 can be quite adequately fit using a sum of three Gaussian
bands; one for the unresolved very high-energy feature and two
for the mid-visible features under discussion here. The two-
band model for this mid-visible feature is a necessary and
sufficient condition for fitting of all 11 spectra and is consistent
with the TD-DFT calculations. Examples of the resultant fits
Electrochemistry. Typical cyclic voltammograms for the
thorium bis(ketimide) complexes in 0.1 M [Bu4N][B(C6F5)4]/
THF are shown in Figure 5. Redox potential data for all
complexes are provided in Table 6. As shown in Figure 5 and
noted in previous studies, the principal redox activity is limited
to reduction processes that take place at potentials near the
negative limit of this supporting electrolyte/solvent system.3b
These processes have been ascribed to ketimide ligand-based
reductions because nearly identical reduction waves are observed
in the analogous uranium bis(ketimide) complexes and because
metal-based reduction of the closed-shell Th(IV) ion is expected
to lie to much more negative potentials than could be accessed
under these experimental conditions. The characteristics of the
voltammetric waves are for the most part consistent with
irreversible electron-transfer steps, although there is a wide
range of observed behavior as reflected in the traces shown in
Figure 5. In addition, coupled chemical processes are observed
in most cases as revealed by the appearance of new anodic peaks
in the potential region from ∼-0.5 to -2.0 V that are not
observed if the potential is first scanned in the positive direction
(Figure 5).
It is noteworthy that many complexes exhibit two discrete
reduction waves that are separated by ∼100-400 mV (Table
6). This behavior is attributed to the sequential addition of an
(33) The preceding discussion refers to the dominant component of the NTO
analysis for a given state. Additional details are given in the Supporting
Information.
9
5146 J. AM. CHEM. SOC. VOL. 129, NO. 16, 2007

Studies of Early Actinide Complexes
A R T I C L E S
Table 6. Summary of Physical Chemical Data for the Thorium Bis(ketimide) Complexes, (C5Me5)2Th[-NdC(R1)(R2)]2
room-temperature electronic absorption dataa
reduction potentials
0
complex
(V vs [(C5H5)2Fe]+/
)
high-energy band
Emax
low-energy band
Emax
77 K electronic emission datab
osc. strengthc
osc. strengthc
Σ
(osc. strength)
E00(S1)
E00(T1)
τmeas
1
1
1
1
R1
R2
E1 2(1)
E1 2(2)
/
(cm-
)
(x 103)
(cm-
)
(x 103)
(x 103)
(cm-
)
(cm-
)
(µs)
/
d
1
2
3
4
5
6
7
8
9
Ph
Ph
Ph
Ph
CH2Ph
CH3
-2.78
-2.84
-2.94
-2.90
-2.53
-2.99
-2.70
-2.83
-2.94
-3.08
-2.84
21310
23030
22830
22160
22890
24620
21830
22340
25350
25240
23950
5.9
0.3
0.9
2.1
3.0
5.4
3.3
2.5
1.8
1.7
1.8
19900
21620
21170
20480
21180
22520
20170
20700
23410
23440
23840
1.6
1.8
1.0
0.5
1.1
5.1
0.6
1.5
6.9
5.8
6.6
7.5
2.1
1.9
2.6
4.1
10.5
3.9
4.0
8.7
7.5
8.4
19900
20380
22100
22110
21800
21440
20500
20640
20760
16460
18130
18330
19100
19600
18290
18660
18760
18440
18700
18150
50
-2.99
-3.12
410
400
340
140
400
290
290
350
290
360
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
3-F-C6H4
4-F-C6H4
2-F-C6H4
3,5-F2-C6H3
3,4,5-F3-C6H2
2,6-F2-C6H3
2,4,6-F3-C6H2
C6F5
-2.99
d
-3.08
-3.11
-2.92
d
d
10
11
22850
d
-3.02
a
Metrical parameters derived from Gaussian fits. See text for details. b E00 values estimated from apparent position of highest energy vibronic component.
c
d
Calculated from spectral fit data according to Eqn. 4.2 in ref 34. Not observed.
to these putative nπ* transitions are illustrated in Figure 6 for
complexes 1 and 5. The metrical data derived from these fits
are compiled in Table 6. Note that the oscillator strengths34 in
these transitions (∼10-3) are wholly consistent with their
assignments, being spin-allowed but orbitally forbidden in
character, and agree well with previously reported TD-DFT
calculations.11
complexes for which no substantial structural distortion is
observed. This is a clear manifestation of the disruption of the
π* orbital network across the -NdC(CH3)(ArF) framework that
has the net effect of raising the energy of the terminal π* orbital
in the nπ* transition.
Electronic Emission Spectroscopy. Perhaps the most inter-
esting aspect of the spectroscopic properties of these thorium
bis(ketimide) complexes is that they exhibit luminescence. The
emission was first observed in the course of surveying their
Raman spectral behavior using blue laser excitation in which
the emission, even at room temperature, could be clearly seen
emanating from the laser excitation volume. Subsequent, more
detailed studies confirmed that this luminescent behavior is
generic to all the thorium bis(ketimide) complexes reported here;
it is quite intense and long-lived at 77 K for all complexes, the
range of energies parallels that seen in the electronic absorption
data, and the decay pathways are both surprising and complex,
particularly in light of the presence of such a high Z metal in
intimate proximity with the emitting chromophore. Indeed, while
luminescence is fairly common in d0 transition-metal complexes,
including Cp2Zr[-NdC(Ph)2]2 (12) (see Supporting Informa-
tion), it is typically ligand-to-metal charge transfer in nature as
observed in systems like Cp2Ti(NCS)2.38-43 There appear to be
only a few examples of Th(IV) systems for which optical
emission has been reported, and most of these examples are
based on porphyrin-type complexes for which the emission is
also ligand-based.44-48 However, it has been pointed out that
high Z metals like Th can actually be used to induce preferential
phosphorescence decay over fluorescence in organic ligands as
a result of enhanced spin-orbit coupling.46
The assignment of the lowest excited states to ketimide-
localized nπ* transitions is also supported by experimental
observations and theoretical calculations for the structurally and
electronically analogous Cp2Zr[-NdC(Ph)2]2 (12) for which a
very similar mid-visible spectral feature is observed.5,11 In fact,
77 K electronic absorption spectral data for both Cp2Zr[-Nd
C(Ph)2]2 (12) and (C5Me5)2Th[-NdC(Ph)2]2 (1) clearly show
vibronic resolution in this lowest-energy absorption feature (see
Supporting Information) as might be expected for an nπ*
transition in which there is a net change in the ketimide ligand
bonding between the ground and excited states. The observed
vibronic progressions (∼1300 cm-1 for 12; ∼1200 cm-1 for 1)
do not correspond to real vibrational modes in these systems;
this is the familiar “MIME” effect in systems having multiple
Franck-Condon active vibrational modes.35-37 However, the
observation of this resolved vibronic structure entailing several
Franck-Condon active modes in the absorption spectra suggests
(1) that the structural distortions in these excited states are not
simply localized on the CdN fragment and (2) that these
thorium ketimide systems are poised to yield resonance-
enhanced Raman spectra discussed below.5
Finally, the trend in the energies of these two singlet nπ*
transitions across the series of ketimide ligands is wholly
consistent with predictions from the TD-DFT calculations
described above for the entire series of anionic ketimide free
ligands and for the three complexes for which full calculations
were carried out (Figure 4). Of particular note is the apparent
influence of the sterically induced structural distortion in
complexes 9-11. As illustrated in Figure 4, the average value
for the two singlet nπ* transitions for all three of these distorted
complexes is significantly greater than that for all other
(38) Pfennig, B. W.; Thompson, M. E.; Bocarsly, A. B. Organometallics 1993,
12, 649.
(39) Patrick, E. L.; Ray, C. J.; Meyer, G. D.; Ortiz, T. P.; Marshall, J. A.; Brozik,
J. A.; Summers, M. A.; Kenney, J. W. J. Am. Chem. Soc. 2003, 125, 5461.
(40) Tonzetich, Z. J.; Eisenberg, R. Inorg. Chim. Acta 2003, 345, 340.
(41) Yam, V. W. W.; Qi, G. Z.; Cheung, K. K. J. Organomet. Chem. 1997,
548, 289.
(42) Yam, V. W. W.; Qi, G. Z.; Cheung, K. K. Organometallics 1998, 17, 5448.
(43) Yam, V. W. W.; Qi, G. Z.; Cheung, K. K. J. Chem. Soc., Dalton Trans.
1998, 1819.
(44) Bilsel, O.; Rodriguez, J.; Milan, S. N.; Gorlin, P. A.; Girolami, G. S.;
Suslick, K. S.; Holten, D. J. Am. Chem. Soc. 1992, 114, 6528.
(45) Knor, G.; Strasser, A. Inorg. Chem. Commun. 2002, 5, 993.
(46) Kunkely, H.; Vogler, A. Chem. Phys. Lett. 1999, 304, 187.
(47) Martarano, L. A.; Wong, C. P.; Horrocks, W. D.; Goncalves, M. P. J. Phys.
Chem. 1976, 80, 2389.
(34) Lever, A. B. P. Inorganic Electronic Structure and Spectroscopy, 2nd ed.;
Elsevier: Amsterdam, 1984.
(35) Tutt, L.; Tannor, D.; Schindler, J.; Heller, E. J.; Zink, J. I. J. Phys. Chem.
1983, 87, 3017.
(36) Tutt, L. W.; Zink, J. I.; Heller, E. J. Inorg. Chem. 1987, 26, 2158.
(37) Tutt, L.; Tannor, D.; Heller, E. J.; Zink, J. I. Inorg. Chem. 1982, 21, 3858.
(48) Tranthi, T. H.; Dormond, A.; Guilard, R. J. Phys. Chem. 1992, 96, 3139.
9
J. AM. CHEM. SOC. VOL. 129, NO. 16, 2007 5147

A R T I C L E S
Schelter et al.
Figure 6. UV-visible electronic absorption spectra for (C5Me5)2Th[-NdC(R1)(R2)]2 in toluene solution at room temperature. Spectral deconvolutions of
the constituent bands in the low-energy region are shown in dashed lines for 1 and 5. Wavelength markers in top panel denote excitation wavelengths used
for Raman spectra of 2, shown in Figure 9.
Typical 77 K emission and emission excitation data are
illustrated in Figure 7 for complexes 1 and 5. The room-
temperature electronic absorption spectral data are superimposed
on these emission results, and the correspondence in energy of
the low-lying nπ* absorption transitions described above with
the manifold of emission bands suggests that the luminescence
originates from these same ketimide ligand-localized states.
Note, however, that the emission spectral behavior in particular
is inconsistent with a simple organic fluorescent-like behavior.
The substantial shift (∼4000 cm-1) observed between the
emission maximum and the excitation maximum in the data
for 1 is more in keeping with emission from a relaxed triplet
state following excitation into the singlet manifold. The more
surprising result is the appearance of distinct multiple emission
bands in the case of complex 5 as seen from the emission
spectrum obtained using CW excitation (Figure 7, bottom, green
trace). The contribution of this higher-energy band to the total
emission envelope disappears under time-resolved excitation/
detection conditions (red trace), demonstrating that the lifetime
of this state is much less than that of the lower-lying, longer-
lived state. This multiple-state emission behavior is a common
feature of all complexes reported here except for 1 and 11.
There are several different scenarios that can give rise to
multiple-state emission behavior. The two most common are
(1) spatially isolated emission in which the emissive states are
localized in spatially distinct, electronically uncoupled (or very
weakly coupled) orbitals on the molecule49,50 and (2) singlet
and triplet emission from states having a common orbital
parentage.45 We believe the latter is the more precise description
of the operative mechanism in the thorium bis(ketimide)
(49) DeArmond, M. K.; Myrick, M. L. Acc. Chem. Res. 1989, 22, 364.
(50) DeArmond, M. K. Acc. Chem. Res. 1974, 7, 309.
9
5148 J. AM. CHEM. SOC. VOL. 129, NO. 16, 2007

Studies of Early Actinide Complexes
A R T I C L E S
Figure 7. 77 K Emission spectral data obtained for toluene glasses of complexes (C5Me5)2Th[-NdC(R1)(R2)]2 1 (top) and 5 (bottom). Spectral parameters
used for 1 were as follows: CW emission: λex ) 450 nm (22220 cm-1), CW excitation: λem ) 600 nm (16670 cm-1), TR emission: λex ) 450 nm (22220
cm-1), delay time ) 20 µs, gate width ) 1 ms. Spectral parameters used for 5 were as follows: CW emission: λex ) 365 nm (27400 cm-1), CW excitation:
λem ) 575 nm (17400 cm-1), TR emission: λex ) 360 nm (27780 cm-1), delay time ) 60 µs, gate width ) 1 ms.
complexes. Note that the higher-energy, shorter-lived emission
band (Figure 7, bottom) corresponds more closely in energy to
that of the two states in the singlet manifold from the absorption
spectrum. This is prima facie evidence that the lowest-energy
absorbing and higher-energy emitting states are the same,
namely the lowest-lying singlet state(s). It is also possible to
observe the identical lower-energy emission band in all these
multiple-emitting complexes by pumping directly into the region
of the higher-energy emission band (not shown in Figure 7).
This demonstrates that these two emissive states are strongly
coupled as expected for singlet/triplet pairs of the same orbital
parentage. Finally, we note that even the nitrile precursors
exhibit multiple-state emission from both singlet and triplet
manifolds. This behavior is well documented in the literature
for benzonitrile for which the ratio of singlet to triplet emission
intensity is strongly medium dependent.51,52 This observation
has been confirmed for all the nitrile precursors used in this
study (Supporting Information). Note, however, that the excited
states in the nitrile precursors are ππ* in character in contrast
to the nπ* character in the ketimide complexes. On the whole,
the evidence seems to best support the assignment of the two
observed emitting states in the thorium bis(ketimide) complexes
as being ketimide nπ* singlet (S1) and triplet (T1) in character.
While emission from the S1 state is not directly observed in
complexes 1 and 11, it is nonetheless possible to locate this
state in the emission excitation spectrum (Figure 7, top) as the
lowest-lying state from which the triplet emission is excited. A
complete summary of the S1 and T1 state energies from the 77
K emission data is provided in Table 6. All emission and
emission excitation bands are vibronically resolved. However,
the extent of the resolution at this temperature is not high, so
precise locations of the S1 and T1 electronic origins (E00) are
not possible. The data in Table 6 are derived from the apparent
(52) Mordzinski, A.; Sobolewski, A. L.; Levy, D. H. J. Phys. Chem. A 1997,
101, 8221.
(51) LeBel, G. L.; Laposa, J. D. J. Mol. Spectrosc. 1972, 41, 249.
9
J. AM. CHEM. SOC. VOL. 129, NO. 16, 2007 5149

A R T I C L E S
Schelter et al.
Figure 8. Time-resolved emission spectra from the nominal T1 excited-state for (C5Me5)2Th[-NdC(R1)(R2)] complexes in toluene glass at 77 K. All
spectra were collected using a 1-ms gate width and a gate delay as indicated in the legends.
energy of the highest-energy vibronic band in each manifold.
In the previous TD-DFT study, the energies of the lowest triplet
states were also determined for the model complex 1a, and the
calculated values (nearly degenerate pair at 2.4 eV or 19350
cm-1) are in good agreement with those observed for (C5Me5)2-
Th[-NdC(Ph)2]2 (1).
A complete summary and comparison of the triplet emission
spectra for complexes 1-11 are provided in Figure 8. The
spectra are all derived from time-resolved data collection
at 77 K. As with absorption spectra shown in Figure 6, these
triplet emission data illustrate the range of electronic energies
found in this series of ketimides. Measured emission lifetimes
have also been determined for this complete series of
complexes by collecting emission decay data on the red edge
of these triplet spectral bands while exciting at energies
greater than the S1 energy for each system.53 In all cases, the
measured decay curves were adequately fit using a two-exponent
function comprising a very short lifetime (attributed to the
instrument response function) and a long lifetime assigned to
the measured lifetime of the complex. The lifetime data are
summarized in Table 6. Note that there is very little dispersion
(∼150-400 µs) in these measured lifetime values except
for complex 1, which has a much shorter (∼50 µs) lifetime.
This shorter lifetime may simply be a consequence of the
energy-gap law54 since the emission energy for 1 is so
much lower than that of the other thorium ketimide complexes.
Even for a ligand-based emission, these measured lifetimes are
(53) These samples are all rigorously oxygen-free since they were prepared in
an inert atmosphere drybox using deoxygenated solvents. Thus, excited-
state lifetimes are not impacted by quenching from oxygen.
(54) Forster, L. S. Coord. Chem. ReV. 2002, 227, 59.
9
5150 J. AM. CHEM. SOC. VOL. 129, NO. 16, 2007

Studies of Early Actinide Complexes
A R T I C L E S
Figure 9. Resonance Raman vibrational spectra for ∼10 mM solutions of (C5Me5)2M[-NdC(R1)(R2)]2 complexes in THF at room temperature. Top panel:
Excitation-wavelength dependence in the resonance-enhanced Raman spectra of 2 in the region of the ketimide CdN and phenyl CdC modes. Solvent
bands are labeled “(S)”. Bottom panel: Raman spectral comparison among bis(ketimide) complexes. Complex 2 at 458 nm, 13 at 514 nm, 14 at 514 nm,
and 12 at 458 nm. Modes dominated by CdN stretching are indicated with arrows. Band positions (in cm-1) are indicated for νCdC and νCdN modes. Values
in parentheses are from DFT calculations (ref 3). Data for 12-14 are taken from ref 3.
surprisingly long; particularly so since a high Z metal is in
intimate contact with the emitting chromophore. Presumably
the long lifetimes are at least in part a consequence of both
orbital and spin-forbidden nature of the 3nπ* state, and as noted
above, the Th atom may actually serve to facilitate the excited-
state relaxation. We were unable to measure comparable
emission lifetimes while monitoring emission within the putative
S1 band because the temporal resolution of our instrument (∼20
µs) was insufficient to capture this very short lifetime of this
state.
The triplet spectra also exhibit varying degrees of vibronic
structure. For those systems for which this structure is most
evident, it appears that most of the intensity resides in the first
(highest-energy) vibronic band. As noted above, the apparent
9
J. AM. CHEM. SOC. VOL. 129, NO. 16, 2007 5151

A R T I C L E S
Schelter et al.
vibronic spacings in these triplet emission spectra do not
correspond to real vibrational modes of the complexes. Instead,
they represent a time-domain weighted average of mode energies
and displacements in the excited electronic states.35-37 However,
the observation that the greatest intensity appears in the first
apparent vibronic band is consistent with the conclusion that
the distortion between the ground state and T1 along the normal
mode(s) responsible for this vibronic structure is not particularly
large, again consistent with nonbonding to antibonding (nπ*)
character of the excited state.
Resonance Raman Spectroscopy. The final noteworthy
spectral property observed for the thorium systems is resonance
enhanced Raman scattering. This behavior is perhaps anticipated,
given the vibronic structure seen in the low-temperature
absorption and emission spectral data. Such vibronic structure
is a hallmark of Franck-Condon activity in one or more
vibrational modes between the ground and excited electronic
states, and this is the primary mechanism giving rise to
enhancement in the Raman spectrum. This novel resonance
Raman behavior has been previously noted for several U(IV)
and Zr(IV) bis(ketimide) complexes,5 but there is no known
precedent for this behavior in the literature of Th(IV) chemistry.
The observation of resonance Raman scattering is a much greater
experimental challenge in the case of the thorium complexes
than in the analogous uranium systems, because all the thorium
bis(ketimide) compounds are intensely emissive in the same
spectral region that is probed in the resonance Raman experi-
ment.55 Thus, our studies to date have been limited to complex
2, which is a somewhat weaker emitter than the other thorium
bis(ketimide) complexes.
Typical resonance Raman spectral data for ∼10 mM complex
2 in THF solution are illustrated in Figure 9. The top panel
shows the intensity profile in the region of the CdN and CdC
stretching modes as the excitation wavelength is tuned into
resonance with the nπ* singlet manifold and the higher-lying
ππ* state(s). These data also confirm the resonance enhanced
nature of the scattering since (1) the intensities in the complex
modes are equal to or exceed those of the THF modes (indicated
by “(S)” in Figure 9, top) for a solution that is ∼10 mM in 2
but ∼12 M in solvent (i.e., ∼103 difference in concentration of
scatterers), and (2) the intensity in the complex modes increases
as the excitation is tuned further into the absorption envelope
(Figure 6).
A comparison of this same spectral region for 2 and several
U(IV) and Zr(IV) bis(ketimide) complexes is provided in the
lower panel of Figure 9. Most notable in this comparison is the
fact that the energies of the two principally CdN stretching
modes for 2 have shifted to significantly lower energy than those
observed for the same two modes in (C5Me5)2U[-NdC(Ph)-
(CH2Ph)]2 (13), while the energies of the principally CdC
(phenyl) stretching modes remain at about the same energy
between the two complexes (Figure 9B, complex 2). This
general trend in the energy of these two sets of modes (νCdN
and νCdC) was also obtained in previously reported DFT
calculations.11 This observation indicates that the CdN bonds
in 2 are weakened relative to the same bond strengths in the
uranium analogue, and the implication is that there may be
stronger metal-nitrogen bonding in the thorium complex than
in the analogous uranium complex that pulls some of the
electron density from the CdN bonds, leading to their weaken-
ing. Such an effect could easily be envisioned to arise from
explicit involvement of the metal 6d orbitals in ligand bonding.
This involvement should be greater in thorium than uranium
because of the relative lower energy of the 6d orbitals in the
thorium system. The proposed strengthening of the Th-N bond
should also be manifest in the vibrational data, but the energy
of these metal-ligand stretching modes are expected to be quite
low, and we are unable to unambiguously identify them in our
spectral data.
Conclusions
This systematic investigation of the structural and electronic
effects of fluorine substitution on the phenyl ring (ArF) in the
ketimide ligands of (C5Me5)2Th[-NdC(CH3)(ArF)]2 complexes
has revealed a substantial impact on both the structural chemistry
and the physical data (spectroscopy and electrochemistry) that
report on the electronic structure in this series of complexes.
We do not observe a strong correlation among the spectroscopic
or electrochemical data for this series that would suggest a
dominant influence from the variation in the fluoride substitution
pattern. Nonetheless, we do find excellent agreement in both
the structural chemistry and the electronic energy levels between
experiment and density functional theory calculations. Clearly,
there are multiple factors at play in this series of complexes,
most notably the structural distortions that accompany the
fluoride substitution at both ortho positions on the aryl ring for
complexes 9-11, that likely mask any simple trends resulting
from inductive effects.
Acknowledgment. For financial support of this work, we
acknowledge the LANL Glenn T. Seaborg Institute for Trans-
actinium Science (Postdoctoral fellowships to E.J.S., P.Y., and
R.E.D.; summer research fellowship to K.C.J.), LANL (Direc-
tor’s Postdoctoral Fellowship to E.J.S.), the LANL Laboratory
Directed Research and Development Program, and the Division
of Chemical Sciences, Office of Basic Energy Sciences, Heavy
Element Chemistry program. This work was carried out under
the auspices of the National Nuclear Security Administration
of the U.S. Department of Energy at Los Alamos National
Laboratory under Contract No. DE-AC52-06NA25396.
Supporting Information Available: Full synthetic details and
characterization data for complexes 4-11, X-ray structural data
and tables for complexes 5-11, details of the DFT calculations
and absorption and emission data for complexes 1 and 12, and
emission data for some nitriles. This material is available free
of charge via the Internet at http://pubs.acs.org.
(55) The analogous uranium complexes do not emit, even at 77 K, presumably
because the lower-lying manifold of ligand field states are very efficient
at non-radiatively relaxing the otherwise emissive ligand-based excited
states.
JA0686458
9
5152 J. AM. CHEM. SOC. VOL. 129, NO. 16, 2007