Differences in actinide metal-ligand orbital interactions: Comparison of U(IV) and Pu(IV) β-ketoiminate N,O donor complexes

Article abstract of DOI:10.1039/c1cc12409a

Syntheses and characterization of UCl₂(^Aracnac) ₂, UI₂(^Aracnac)₂, and PuI ₂(^Aracnac)₂ are reported (^Aracnac denotes a bis-phenyl β-ketoiminate ligand where Ar = 3,5- ^tBu₂C₆H₃). Structural analyses and computations show significant metal-ligand orbital interaction differences in U(iv) vs. Pu(iv) bonding.

Full text of DOI:10.1039/c1cc12409a

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
ChemComm
Cite this: Chem. Commun., 2011, 47, 7647–7649
www.rsc.org/chemcomm
COMMUNICATION
Differences in actinide metal–ligand orbital interactions: comparison
of U(IV) and Pu(IV) b-ketoiminate N,O donor complexesw
David D. Schnaars,ac Enrique R. Batista,*b Andrew J. Gaunt,*c Trevor W. Hayton,*a
Iain May,c Sean D. Reilly,c Brian L. Scottd and Guang Wua
Received 25th April 2011, Accepted 20th May 2011
DOI: 10.1039/c1cc12409a
Syntheses and characterization of UCl2(Aracnac)2, UI2(Aracnac)2,
and PuI2(Aracnac)2 are reported (Aracnac denotes a bis-phenyl
b-ketoiminate ligand where Ar = 3,5-tBu2C6H3). Structural
analyses and computations show significant metal–ligand orbital
interaction differences in U(IV) vs. Pu(IV) bonding.
ultimately facilitate design of ligand sets for enhanced selective
separations of actinides in the same oxidation state (e.g. An(IV)),
and are of relevance to proposed group actinide extractions.5
b-Ketoiminate ligands have previously found utility in
stabilizing UVO2 species through steric saturation of the
+
equatorial plane.6 Their ability to form stable complexes with
uranyl(V), a relatively weak Lewis acid, indicates that they
should readily coordinate to An(IV) ions which are much stronger
Lewis acids. We believe that Aracnac complexes are interesting
candidates for comparison of U(IV) vs. Pu(IV) bonding para-
meters because they contain two different donor atoms within
the same chelating ligand (O and N, respectively). These
ligands would allow us to look for differences in metal–ligand
interactions arising from 5f and 6d orbital participation. Our
synthetic approach was to first demonstrate that b-ketoiminates
can form stable U(IV) complexes and then attempt to isolate
isostructural Pu(IV) complexes.
Comparative bonding studies of molecular compounds across
the actinide series are rare, with the exception of Th vs. U,
which have long-lived low specific activity radioisotopes readily
available for synthetic chemistry. For transuranic elements,
significantly higher radiotoxicity hazards require specialized
radiological facilities to allow safe manipulation.1 Despite
these practical challenges, establishing bonding trends and
differences across the 5f series over a wide range of ligand
types is important to provide underpinning chemical knowledge
for nuclear fuel cycle applications.2
Historically, the majority of transuranic coordination chemistry
reported is with hard O donor chelates for which ionic
bonding to the metal is believed to be the only relevant
consideration (the obvious exception being the terminal dioxo
actinyl moieties which contain multiple bond character).3 Such
donors were suitable for early actinide separation processes
that utilized differences in actinide redox chemistry to access
their markedly different chemical properties.3 However, advanced
nuclear fuel cycles require alternative separations. For example,
soft donors (N and S) are being studied for Am(III)/Cm(III)
separation from the chemically similar Ln(III) ions by harnessing
increases in An(III) vs. Ln(III) complex stability (likely due to
subtle covalency increases).4 Here, we are interested in exploring
subtle variations in 6d and 5f orbital interactions, which may
Treatment of 2 equivalents of the b-ketoiminate ligand
Na(ArNC(Ph)CHC(Ph)O) (Ar = 3,5-tBu2C6H3) (generated
in situ) with UCl4 or UI4(OEt2)2 forms UCl2(Aracnac)2 (1)
and UI2(Aracnac)2 (2), respectively (Scheme 1). Preliminary
attempts to synthesise the analogous Pu(IV) complexes,
PuX2(Aracnac)2 (X = Cl or I), were complicated by the lack
of readily accessible non-aqueous organic soluble Pu(IV) starting
materials, with the only well established example being various
salts of the [PuCl6]2À anion.1,3 Reaction of [PPh4]2[PuCl6] with
two equivalents of Na(Aracnac) did not lead to isolation of
PuCl2(Aracnac)2.7 As an alternative strategy, the use of Pu0 as
a precursor to a low valent Pu b-ketoiminate complex was
examined. We initially employed reaction conditions designed
to isolate a Pu(III) complex because prior observations
had suggested that direct oxidation to Pu(IV) was unlikely
(excess I2 or Br2 alone does not oxidize Pu0 to PuIV under inert
atmospheric/anhydrous conditions).8 However, oxidation of
a Department of Chemistry and Biochemistry,
University of California Santa Barbara, Santa Barbara, CA 93106,
USA. E-mail: hayton@chem.ucsb.edu; Tel: +1 805 893 3392
b Theoretical Division, Los Alamos National Laboratory, Los Alamos,
NM 87545, USA. E-mail: erb@lanl.gov; Tel: +1 505 667 8177
c Chemistry Division, Los Alamos National Laboratory, Los Alamos,
NM 87545, USA. E-mail: gaunt@lanl.gov; Tel: +1 505 667 3395
d Materials Physics and Applications Division,
Los Alamos National Laboratory, Los Alamos, NM 87545, USA
w Electronic supplementary information (ESI) available: Syntheses of
1–3, IR, NMR, UV/vis/nIR, cyclic voltammetry, DFT calculation
details, and CIF’s for 1–3. CCDC 824025–824027. For ESI and
crystallographic data in CIF or other electronic format see DOI:
10.1039/c1cc12409a
Scheme 1 General preparative routes to 1–3.
Chem. Commun., 2011, 47, 7647–7649 7647
c
This journal is The Royal Society of Chemistry 2011

View Article Online
Pu0 with 1.6 equivalents of I2 in THF followed by treatment
with 2 equivalents of Na(Aracnac) (in situ) afforded several
single crystals of PuIVI2(Aracnac)2, rather than the anticipated
Pu(III) complex.9 In light of this surprising result we increased
the initial stoichiometry of I2 to two equivalents providing
isolation of PuIVI2(Aracnac)2 (3) in a low, but pure (single
crystals), 17% yield (Scheme 1). It is likely that access to the
Pu(IV) oxidation state is made possible by the strong electron
donating ability of the Aracnac ligand shifting the Pu(III)/Pu(IV)
redox potential within the range of I2.
although the difference of 0.046 A is perhaps slightly larger
than might be anticipated (the ionic radii for the 6-coordinate
metal ions differ by 0.03 A; 0.89 A for U and 0.86 A for Pu).10
Thirdly, considering the An–O bonding to the Aracnac ligand,
the U–O distance in 2 is 2.163(4) A and the Pu–O distance in 3
is 2.1727(17) A. These bonds differ by just 0.01 A, are within
statistical error of each other, and thus do not follow the
expected trend of a similar shortening of the U–O vs. Pu–O
bond in the way that the U–N vs. Pu–N bond shortens. The
fact that the O and N donor atoms in 2 and 3 are contained
within the same ligand (Aracnac), yet do not exhibit similar
changes in bond lengths upon moving from U(IV) to Pu(IV)
(i.e. An–O remains the same while An–N shortens), suggests
that the bonding changes in these complexes across the An(IV)
series are not adequately explained by the actinide contraction
and ionic bonding models alone.
1
The H NMR spectra of 1–3 exhibit paramagnetic shifting
and contain resonances consistent with the formulation of the
proposed complexes (Fig. S6–8w). Interestingly, the addition
of 1 equiv of THF to a C6D6 solution of 1 results in a change
of the resonance profile, suggesting possible coordination to
the complex (Fig. S6A–Bw). Additionally, in the 1H NMR
spectrum of 2 in CD2Cl2, the broad resonances at À1.10 ppm
and À3.84 ppm are consistently present (independent of
the synthetic method) and we have tentatively assigned them
The only other U(IV)/Pu(IV) discrete molecular isostructural
comparison that we are aware of in the literature is the
polyoxometalate anion [An(a2-P2W17O61)2]16À (An = Th, U,
Np, Pu, Am), which afforded an isostructural comparison of
An–O bonds across the An(IV) series.11 In that study, the
average Pu–O distance is shorter than the average U–O
distance by 0.03 A, and was attributed to the effect of the
different ionic radii of the 8-coordinate metal centres (1.00 A
for U and 0.96 A for Pu).10 The scarcity of literature comparisons
and the unexplained bond length trends in 2 and 3 motivated
us to probe the metal–ligand interactions further.
t
as the two inequivalent Bu environments arising from the
presence of a minor cis isomer (Fig. S7w).
In the solid state, 1–3 are geometrically isostructural consisting
of a six coordinate An(IV) metal centre with a ligand arrange-
ment best described as distorted octahedral (Fig. 1, S3–5w). In
each case, the halides sit trans to one another in axial positions
and the two bidentate Aracnac ligands occupy the equatorial
plane in a trans configuration. In addition to being geometri-
cally isostructural, the bond lengths and angles between
uranium and the Aracnac ligand in 1 and 2 are identical within
statistical errors (Table 1), thus the identity of the coordinated
halide has no significant effect upon the metrical parameters
associated with the coordination of the Aracnac ligand to
uranium.
We turned to hybrid DFT modeling, demonstrating reasonable
agreement between the experimental and calculated structures,
and although actinide–ligand bond distances tend to be over-
estimated by DFT,12 the larger variation in the An–N (0.04 A)
distances vs. the An–O distances (0.01 A) between 2 and 3
is accurately captured (Table 1). Natural orbital analysis
indicates a clean unpaired spin density with ground state
configurations triplet U(5f2) for 2 and quintet Pu(5f4) for 3
(Fig. S17–18w). Further 5f and 6d electronic density from the
U and Pu (0.7 and 1 eÀ, respectively) is involved in bonding
interactions with the I, O, and N ligands. Natural orbitals
indicate the 5f density is involved in three s-type bonds
between U or Pu and the three ligand types. The 6d density
participates in s bonding with the I and p bonding with the
O and N atoms, however, it is challenging to deconvolute all
three interactions as the ligands interact with one another. We
then performed a Natural Bond Orbital (NBO) analysis. The
metal was identified as a fragment by itself in both 2 and 3 and
so was the iodine, while the oxygen and nitrogen fragments are
represented as core and lone-pair orbitals (LP) plus bonding
orbitals with the neighboring carbon. The level of interaction
and energy stabilization via charge transfer from occupied natural
atomic orbitals was obtained by second order perturbation
analysis in that basis (Table S2–3w). The largest interaction
corresponds to dative bonds from the occupied oxygen
and nitrogen lone pairs to the virtual lone pairs of the metal
(Fig. S19–24w show all LPs).
The structural similarity of 2 to 3 offers a very rare
opportunity for a direct comparison of U(IV) vs. Pu(IV) solid
state molecular structures (Table 1). Firstly, considering the
An–I bonding, the U–I distance in 2 is 3.0288(5) A, whilst the
Pu–I distance in 3 is 0.043 A shorter, with a length of 2.9859(3) A.
This represents the first example of a Pu(IV)–I bond. Secondly,
considering the An–N bonding to the Aracnac ligand, the U–N
distance in 2 is 2.409(5) A and the Pu–N distance in 3 is
2.3635(19) A. As expected from the actinide contraction and
increasing effective positive charge of the An(IV) ions across
the series, the Pu–N bond is shorter than the U–N bond,
For the An–N interactions, the overall strength is very
similar for U(IV) vs. Pu(IV) and there are two main contri-
butions in both cases: in the Pu complex (3) the stabilization
energies are 21 and 7 kcal molÀ1, and in the U complex (2)
they are 25 and 8 kcal/mol. In both cases the interacting
Fig. 1 Solid-state structure of 3 with 50% probability ellipsoids.
H atoms and solvent have been omitted for clarity. 1 and 2 are
geometrically isostructural. Lattice solvation also differs between 1–3.
c
7648 Chem. Commun., 2011, 47, 7647–7649
This journal is The Royal Society of Chemistry 2011

View Article Online
Table 1 Selected bond distances (A) and angles (1) for 1–3
Calculated
1
M = U, X = Cl
2
M = U, X = I
3
M = Pu, X = I
2
3
M–X
M–O
M–N
X–M–X
O–M–N
2.6315 (12)
2.173 (3)
2.406 (3)
180
3.0288 (5)
2.163 (4)
2.409 (5)
180
2.9859 (3)
2.1727 (17)
2.3635 (19)
180
3.073
2.185
2.462
180
3.045
2.198
2.422
180
71.19 (11)
71.06 (15)
73.29 (7)
72.67
75.58
orbitals are almost the same: the N lone pair (25% s + 75% p)
couples with a 100% 5f orbital in the smaller contribution, and
in the larger contribution the N lone pair couples with a metal
virtual LP of 6d and 7s orbital composition. Analysis of the
An–O interactions indicates that the U–O orbital interactions
are stronger than the Pu–O orbital interactions. In contrast to
the An–N interactions, the interacting pairs in the An–O bonds
are different between 2 and 3 (see Fig. 2 for representations of
the strongest An–O interactions in 2 and 3). Three dominating
pairs are found in the U–O bond: the first of energy 23 kcal molÀ1
involving the LP1 in oxygen (40% s + 60% p) and a virtual LP
in the U (70% 6d + 30% 7s), the second of energy 15 kcal molÀ1
involving the LP2 in oxygen (16% s + 84% p) and the same
virtual LP in U, and the third of energy 13 kcal molÀ1 involving
the LP2 in oxygen and a virtual LP in U (100% 5f). There are
another three minor interactions of 5.5, 4.7 and 4.3 kcal molÀ1
for the U–O bond. Turning to the Pu–O bond, there is
one main interaction of 16.5 kcal molÀ1 between LP1 of the
oxygen (46% s + 54% p) and a 100% 6d orbital of Pu. The
rest of the Pu–O interaction comprises four minor contri-
butions of 8.9, 6.5, 5.7, and 5.4 kcal molÀ1 all involving the
same oxygen LP2 (12% s + 88% p) with a 100% 6d orbital, a
100% 5fp, a 52% 6d + 48% 7s orbital, and a 100% 5fj orbital,
respectively. A fifth minor interaction of 6.1 kcal molÀ1 is
between the oxygen LP3 (100% p) and a 100% 6d LP in Pu.
We interpret the increased sum of the orbital interactions
for the U–O bond (65.5 kcal molÀ1) vs. the Pu–O bond
(49.1 kcal molÀ1) as an explanation for the observed shortening
of the U–O bond with respect to the Pu–O bond. Since the
U–N vs. Pu-N interaction compositions and total strength (33 and
28 kcal molÀ1, respectively) are similar, we attribute the experi-
mentally observed shorter Pu–N vs. U–N bond to the effect of
the actinide contraction. This finding is significant because,
upon considering the role of covalency in f-element separations,
it is usually the softer donors (N or S atoms) which are
anticipated to exhibit differences that can result in selectivity
for a particular actinide ion. NBO analysis of the relative charges
on the O vs. N atom in 2 and 3 confirm that the O atom is
‘harder’ than the N atom in the Aracnac ligand (À0.744 and
À0.757 on O in 2 and 3, compared to À0.617 and À0.632 on N;
about 17% smaller) and yet it is this ‘hard’ oxygen donor which
is the origin of metal–ligand orbital overlap differences rather
than the relatively ‘soft’ N atom. However, further comparisons
of isostructural O-donor actinide complexes are required to
facilitate development of how these bonding differences can be
exploited in ligand designs for advanced nuclear fuel cycle
separation processes.
We thank the University of California, Santa Barbara, and
the University of California Laboratory Fees Program. D.D.S.
thanks the Seaborg Institute Summer Research Fellowship
Program at LANL. A.J.G./S.D.R. thank the U.S. Department
of Energy, Office of Science, Early Career Research Program
(contract DE-AC52-06NA25396). E.R.B./I.M. thank the U.S.
Department of Energy, Office of Science, Basic Energy
Sciences, Heavy Element Chemistry Research Program.
Notes and references
1 A. J. Gaunt and M. P. Neu, C. R. Chim., 2010, 13, 821.
2 Spent fuel reprocessing options, IAEA-TECDOC-1587, 2008,
http://www-pub.iaea.org/MTCD/publications/PDF/te_1587_web.pdf.
3 The Chemistry of the Actinide and Transactinide Elements,
ed. L. R. Morss, N. M. Edelstein, J. Fuger and J. J. Katz, Springer,
Dordrecht, The Netherlands, 2006.
4 See, for example (and references therein): A. J. Gaunt, S. D. Reilly,
A. E. Enriquez, B. L. Scott, J. A. Ibers, P. Sekar, K. I. M. Ingram,
N. Kaltsoyannis and M. P. Neu, Inorg. Chem., 2008, 47, 29;
M. P. Jensen and A. H. Bond, J. Am. Chem. Soc., 2002, 124,
9870.
5 E. Aneheim, C. Ekberg, A. Fermvik, M. R. S. Foreman,
T. Retegan and G. Skarnemark, Solvent Extr. Ion Exch., 2010,
28, 437.
6 T. W. Hayton and G. Wu, Inorg. Chem., 2009, 48, 3065.
7 Method of synthesis of [PPh4]2[PuCl6] provided by S. A. Kozimor,
et al., and is to be reported elsewhere.
8 A. J. Gaunt, S. D. Reilly, A. E. Enriquez, T. W. Hayton,
J. M. Boncella, B. L. Scott and M. P. Neu, Inorg. Chem., 2008,
47, 8412.
9 X-ray structural data was low quality but clearly established atom
identities and molecular connectivity.
10 R. D. Shannon, Acta Crystallogr., Sect. A: Cryst. Phys., Diffr.,
Theor. Gen. Crystallogr., 1976, 32, 751.
11 M. N. Sokolova, A. M. Fedosseev, G. B. Andreev, N. A. Budantseva,
A. B. Yusov and P. Moisy, Inorg. Chem., 2009, 48, 9185.
12 T. W. Hayton, J. M. Boncella, B. L. Scott, E. R. Batista and
P. J. Hay, J. Am. Chem. Soc., 2006, 128, 10549.
Fig. 2 The strongest An–O orbital interaction in 2 compared to 3: a
23 kcal molÀ1 U–O couple with 30% 7s and 70% 6d character for the
metal contribution (left), and a 16.5 kcal molÀ1 Pu–O couple with
100% 6d character for the metal contribution (right).
c
This journal is The Royal Society of Chemistry 2011
Chem. Commun., 2011, 47, 7647–7649 7649