Structural and Spectroscopic Comparison of Soft-Se vs. Hard-O Donor Bonding in Trivalent Americium/Neodymium Molecules

Article abstract of DOI:10.1002/anie.202017186

Covalency is often considered to be an influential factor in driving An³⁺ vs. Ln³⁺ selectivity invoked by soft donor ligands. This is intensely debated, particularly the extent to which An³⁺/Ln³⁺ covalency diffe

Full text of DOI:10.1002/anie.202017186

Angewandte
Research Articles
Chemie
How to cite: Angew. Chem. Int. Ed. 2021, 60, 9459–9466
International Edition: doi.org/10.1002/anie.202017186
German Edition: doi.org/10.1002/ange.202017186
Actinide Complexes
Hot Paper
Structural and Spectroscopic Comparison of Soft-Se vs. Hard-O Donor
Bonding in Trivalent Americium/Neodymium Molecules
Conrad A. P. Goodwin+, Anthony W. Schlimgen+, Thomas E. Albrecht-Schçnzart,
Enrique R. Batista,* Andrew J. Gaunt,* Michael T. Janicke, Stosh A. Kozimor, Brian L. Scott,
Lauren M. Stevens, Frankie D. White, and Ping Yang*
Abstract: Covalency is often considered to be an influential
factor in driving An3+ vs. Ln3+ selectivity invoked by soft donor
ligands. This is intensely debated, particularly the extent to
which An3+/Ln3+ covalency differences prevail and manifest as
the f-block is traversed, and the effects of periodic breaks
An3+/Ln3+ separation a significant challenge. Differences in
the bonding between Am3+ ions and soft donor ligands vs.
analogous Ln3+ ions have been exploited to affect selectivity
at laboratory-scale,[3] but currently, practical limitations exist
regarding process-scale implementation.[4]
=
beyond Pu. Herein, two Am complexes, [Am{N(E PPh2)2}3]
Early actinides (Th-Pu) have energetically accessible 5f
orbitals as well as 6d and 7p orbitals which have larger radial
extensions and are often significantly involved in bonding.
This allows Th-/U-ligand bonding to be partially covalent due
to overlap of ligand and metal orbitals.[5] For the early
actinides, differential covalency can manifest structurally as
shorter metal–ligand bonds (compared to lanthanide con-
geners) with various soft donors such as nitrogen heterocycles
and thioethers, as determined by X-ray diffraction techni-
ques.[3a,6] These effects can become small for Am[7] where
more sensitive spectroscopic techniques are required to
observe bonding differences to these light atoms.[8] By the
middle of the An series (Am-Cf) there is a significant
contraction of the 5f orbitals, and concomitant lowering of
their energy. Here, there is poor 5f orbital overlap but,
depending on the ligand, there can be favorable energy
matching between ligand and metal valence orbitals: this can
lead to energy degeneracy-driven covalency.[5,9] The two
mechanisms of overlap-driven and energy-degeneracy-driven
covalency are significant for different parts of the An series,
and complicate comparisons of early and middle An ele-
ments.[9c,d,10] Topological methods, such as the quantum theory
of atoms in molecules (QTAIM), describe Am-Cf as pre-
dominantly ionic in their bonding, in accord with conven-
tional thought,[9d,10b,11] while recent experimental approaches
have focused on measuring physical manifestations of elec-
tronic structure differences.[5b, 12]
(1-Am, E = Se; 2-Am, E = O) are compared to isoradial
[Nd{N(E PPh2)2}3] (1-Nd, 2-Nd) complexes. Covalent con-
=
tributions are assessed and compared to U/La and Pu/Ce
analogues. Through ab initio calculations grounded in UV-vis-
NIR spectroscopy and single-crystal X-ray structures, we
observe differences in f orbital involvement between Am–Se
and Nd–Se bonds, which are not present in O-donor congeners.
Introduction
The minor actinides, Am and Cm, are significant contrib-
utors to the long-term radiological burden of used nuclear
fuel, leading to proposals for their separation and trans-
mutation into less hazardous isotopes.[1] However, the triva-
lent oxidation state dominates the chemistry of Am and Cm
and also the lanthanide (Ln) fission products; therefore,
separation based on redox chemistry is difficult.[2] Histori-
cally, both An3+ and Ln3+ ions have been considered to
engage in largely ionic bonding modes with a wide array of
ligands, exhibiting very similar chemical behavior that renders
[*] C. A. P. Goodwin,[+] A. J. Gaunt, M. T. Janicke, S. A. Kozimor,
L. M. Stevens, F. D. White
Chemistry Division, Los Alamos National Laboratory
Los Alamos, NM 87545 (USA)
E-mail: gaunt@lanl.gov
A. W. Schlimgen,[+] E. R. Batista, P. Yang
Theoretical Division, Los Alamos National Laboratory
Los Alamos, NM 87545 (USA)
E-mail: erb@lanl.gov
The comparison of multiple donor-atom systems on
a metal ion across homologous molecular series provides
a systematic approach capable of teasing out subtle bonding
and electronic structure periodic trends in a more convincing
manner than standalone studies of a single metal or donor
type; such experimental studies are extremely rare for
transuranium elements.[13] Some of us have previously
pyang@lanl.gov
T. E. Albrecht-Schçnzart
Department of Chemistry and Biochemistry, Florida State University
95 Chieftain Way, Tallahassee, FL 32306 (USA)
=
reported
tris-imidodiphosphinochalcogenide,
[M{N(E
B. L. Scott
PR2)2}3], complexes for U/Pu, and their isoradial Ln con-
geners La/Ce, where E = S (R = iPr, Ph), Se (R = iPr, Ph), Te
(R = iPr) and M = metal.[13] Analysis of M-E bond lengths in
conjunction with density functional theory (DFT) and
QTAIM computational analysis allowed a clear conclusion
to be drawn that the data was consistent with an enhancement
of covalency in An–E vs. Ln–E bonding. Attempting to
Materials Physics and Applications Division
Los Alamos National Laboratory
Los Alamos, NM 87545 (USA)
[+] These authors contributed equally to this work.
Supporting information and the ORCID identification number(s) for
the author(s) of this article can be found under:
https://doi.org/10.1002/anie.202017186.
Angew. Chem. Int. Ed. 2021, 60, 9459 –9466
ꢀ 2021 Wiley-VCH GmbH
9459

Angewandte
Research Articles
Chemie
understand more subtle Am/Ln differences in a comparable
study has been impeded by a lack of well-developed Am
synthetic chemistry methods along with radiological and
isotope availability considerations. Recent advances in air-/
moisture-free Am chemistry now permit us to compare hard
(43%) yield (Figure 1, left). 1-M complexes have fairly low
solubility even in THF, thus we used a protonolysis layering
protocol rather than salt-metathesis using KLSe to avoid
troublesome separation of 1-Am from KCl. Complex 2-Am
was synthesized in water from KLO[18] and “AmCl3(DME)n”,
prepared under anhydrous conditions. The precipitated
hydrophobic 2-Am was recrystallized in low (26%) yield
(Figure 1, right). Whilst a simplified protocol using dried
O- and soft Se-donor complexes,[12d,14] [Am{N(E PPh2)2}3]
=
(E = Se, 1-Am; O, 2-Am), to their isoradial Nd congeners
=
[Nd{N(E PPh2)2}3] (1-Nd, 2-Nd). We have examined the
bonding and electronic structure in these complexes by single
crystal X-ray diffraction, UV-vis-NIR spectroscopy, multi-
nuclear NMR spectroscopy, and ab initio calculations.
Results and Discussion
243Am is a scarce resource. Each reaction was pre-
optimized with a goal of yielding X-ray quality single crystals
on the first attempt. The rationale for selecting the LSe,
{N(Se PPh2)2}1À, ligand to prepare 1-Am and 1-Nd was that
=
previous studies comparing U3+/La3+ and Pu3+/Ce3+ pairs
exhibited larger An–E vs. Ln–E bond length differences for
=
E = Se and Te than E = S in isomorphous [M{N(E PR2)2}3]
Figure 1. Photographs of the crystals of 1-Am (left) and 2-Am (right).
While 1-Am crystallized uniformly as pale-yellow parallelepipeds, 2-Am
formed two different polymorphs.
(R = Ph, iPr) complexes. The largest difference (0.060(4) Š
shorter An–E vs. Ln–E) was observed for La3+/U3+, where
E = Te and R = iPr.[13d] We hypothesized that if a shorter Am-
E vs. Nd-E bond was to be crystallographically observable,
the probability would be highest using the heavier Se or Te
chalcogen donors. To keep the reaction scale under 5 mg (due
to 243Am availability and radiological hazards) we opted for
the highly crystalline R = Ph variant, rather than R = iPr. The
E = Te variant is not known to be stable for R = Ph so we used
LSe, which has also been shown to exhibit significant An–E
vs. Ln–E differences within the Pu3+/Ce3+ pair.[13d,15] More-
over, there are a limited number of structurally characterized
molecular Am complexes,[16] and LSe offered an opportunity
to explore a new donor type in molecular Am chemistry. To
“Am3+(aq)” residue with KLO was desirable, this was not
successful using Nd3+ and neither were attempts using
“NdCl3(DME)n” prepared from dried acidic residue and
Me3SiCl/ DME in air. Therefore, we opted to use the proven
route to “AmCl3(DME)n”performed under inert atmosphere
conditions. Both 2-M (M = Am, Nd) crystallized in multiple
different crystal systems which could be identified by
morphology (blocks, plates, or rods). Isolated crystals of both
1-M complexes coated in oil are stable to traces of atmos-
pheric air for several hours with no visible signs of degrada-
tion, and solutions in [D]8-THF decolorize slowly over the
course of tens of minutes after exposure to air.
1-Nd and 2-Nd and were prepared similarly (see Support-
ing Information). The ionic radius of Nd3+ (6-coord. 0.983 Š;
8-coord. 1.109 Š) is similar to Am3+ (6-coord. 0.975 Š; 8-
coord. 1.09 Š) and thus it was chosen as the Ln analogue for
structural comparisons.[19] Attempts to synthesize 1-Eu for
spectroscopic comparison to 1-Am (both M3+ ions are f6) were
unsuccessful, with the Eu2+ complex [Eu(LSe)2(THF)2] (3) as
the only isolable product.[20] Such a result was not wholly
unexpected given the “Eu2+-like” electronic structure of
computationally modeled 1-Eu by Kaltsoyannis.[13c]
The structures of both 1-M
provide an anticipated purely ionic comparison we then
turned to LO (LO = {N(O PPh2)2} ) as an O-donor ana-
1À
=
logue of LSe. [M(LO)3] (M = Ln) complexes are generally 6-
coordinate where the N-atom remains unbound,[17] whereas
[M(LSe)3] (M = Ln, An) complexes are 9-coordinate,[13a,d]
with intra-ligand steric repulsion previously invoked to
rationalize these differences.[13a,d]
Complexes 1-Am and 2-Am (Scheme 1) were synthesized
by first converting aqueous acidic 243Am3+ (3.5 mg) into
“AmCl3(DME)n”.[12d] For 1-Am this was used to prepare
“Am(N’’)3” (N’’ = {N(SiMe3)2}) in situ, then layered with
a solution of HLSe, which subsequently formed crystals in fair
and 2-M (M = Am, Nd) were
determined by single crystal
X-ray diffraction (Figure 2
and Figures S16–20) and veri-
fied by collection of full data
sets from two different crys-
tals for each Am structure.
The two data sets feature
statistically indistinguishable
Am–Se bond lengths (3.0625-
Scheme 1. Synthesis of 1-Am and 2-Am (see the Supporting Information). N’’={N(SiMe3)2}1À, LSe={N-
1À
1À
=
=
(Se PPh2)2} , LO={N(O PPh2)2}
.
9460 www.angewandte.org
ꢀ 2021 Wiley-VCH GmbH
Angew. Chem. Int. Ed. 2021, 60, 9459 –9466

Angewandte
Research Articles
Chemie
distances can be subjected to more detailed analysis by
plotting vs. metal ionic radius (Figure S26).[13a,d] While 8- and
9-coordinate radii are not reported by Shannon for U3+/Pu3+,
these values are known for Ln3+ ions and Am3+.[19] Coinci-
dentally a straight line of 8-coordinate radius vs. M–Se
distance for 1-Ln (Ln = La, Ce, Nd) passes through our value
for 1-Am when the experimentally determined 8-coordinate
radius of Am3+ is also used. The M–N distances (1-Am,
2.672(2) Š; 1-Nd, 2.678(3) Š; D = 0.006(4) Š) show a signifi-
cantly smaller An vs. Ln difference than the M–Se bonds. This
is not unexpected for softer Se donors with a better energy
match to the contracted 5f orbitals as compared to the harder
N donors (Table 1).[5,9c,12b,c] Computational work by Kalt-
soyannis has highlighted the minor role that the N-atom plays
in bonding for these and related complexes.[13c] The Am–Se
bond length in 1-Am is also the longest measured Am-E bond
in a molecular system. The closest other examples come from
the Am-S bonds in [nBu4N][Am{S2P(tBu2C12H6)}4], [AsPh4]-
[Am{S2P(OEt)2}4] and [Am{S2C(NEt2)}3(phen)] (phen = 1,10-
phenanthroline; range 2.8015(9)–2.969(4) Š across these
three),[12c,24] and the Am-Br bonds in [AmBr3(THF)4] and
=
[AmBr3(O PCy3)3] (range 2.8221(7)–2.9118(13) Š across
both).[25]
Figure 2. Molecular structures of a) 1-Am and b) the Am1 molecule of
2-Am.[36] Ellipsoids set at 50% probability. H-atoms and lattice solvents
removed for clarity. a) Am1–Se1 3.0625(8) Š, Am1–N1 2.672(2) Š; P1-
N1-P1 144.27(18)8; b) Am1–O1 2.317(3) Š, Am1–O2 2.342(3) Š; P1-
N1-P1 128.9(3)8.
Table 1: A summary of M–Se and M–N distances in reported f-element 1-
M complexes.
M
M–N [Š]
D Ln/An [Š][a]
M–Se [Š]
D Ln/An [Š][a]
La
U
Ce
Pu
Nd
Am
2.706(3)
2.701(3)
2.691(3)
2.668(2)
2.678(3)
2.672(2)
3.1229(3)
3.0869(4)
3.1013(3)
3.0710(2)
3.0801(3)
3.0625(8)
0.005(4)
0.023(4)
0.006(4)
0.0360(5)
0.0303(4)
0.0176(9)
(8) and 3.0615(9) Š), the longer of which is used for the
following discussion (to guard against over-interpretation) as
this represents the smallest difference between 1-Am and 1-
Nd. Complex 1-Am crystallized in the centrosymmetric space
[a] Numerical difference, error calculated as the root sum of the squares.
ꢀ
group R3c, and is isomorphous with the reported U, Pu, La,
Ce congeners, thus comprising a directly comparable series.
The Am center is 9-coordinate with all three LSe ligands
bound k3-Se,N,Se and best described as tri-capped trigonal
prismatic.[21] A comparison of the M–Se distances in 1-Am
(3.0625(8) Š) and 1-Nd (3.0801(3) Š) reveals that the Am–Se
bond is shorter than the Nd–Se bond (D = 0.0176(9) Š) in
a statistically meaningful manner (3s criterion),[22] a note-
worthy feature because often Am–L vs. Nd–L distances are
statistically indistinguishable from each other. While this is
larger than the difference between 6-coord. Nd3+/Am3+ (D =
0.008 Š), it is smaller than the difference in their 8-coordinate
radii (D = 0.019 Š, see Supporting Information for more
discussion). Thus, we see no structural evidence of increased
covalency in this Am3+ complex compared to the Nd3+
analogue. The difference in M–Se bond lengths was larger
for both the U3+/La3+ pair (D = 0.0360(5) Š) and the Pu3+/
Ce3+ pair (D = 0.0303(4) Š). Therefore, it can be suggested
that while the magnitude of the An–Se vs. Ln–Se shortening is
statistically equivalent upon moving from U to Pu, there is
a significant drop in this difference upon moving to Am,
consistent with the traditional view of a transition to more
“lanthanide-like” behavior from Pu to Am.[1b, 23]
ꢀ
Complex 2-Am crystallized as two polymorphs (R3, 2-
a
b
a
ꢀ
Am ; P3, 2-Am ); for brevity only 2-Am is discussed here
(see Supporting Information for discussion on both). The two
Am centers in the asymmetric unit of 2-Ama are 6-coordinate
with three LO ligands bound k-O,O. Each has a distinct
geometry with Am1 best described as octahedral, and Am2 is
almost perfectly trigonal prismatic.[21,26] The presence of both
geometries in the asymmetric unit has been seen previously
with f-element imidodiphosphinate complexes.[17b,18,27] The
same polymorphs (plus two additional ones) were found for 2-
Nd and were selectively grown via appropriate crystallization
conditions. The unique Am O bonds in 2-Am (Am1:
2.317(3) and 2.342(3) Š; Am2: 2.342(3) and 2.355(3) Š) are
broadly the same between the two geometries, and are mostly
indistinguishable from Nd–O bonds in 2-Nda (Nd1: 2.313(3)
and 2.339(3) Š; Nd2: 2.341(3) and 2.340(2) Š).
a
À
=
Prior calculations on a model R = H complex, [Am{N(O
PH2)2}3], predicted an Am–O length of 2.358 Š,[13c] which is
slightly larger than those in 2-Ama, however there are no
other 6-coordinate An3+ analogues using this ligand set to
determine trends. Previous work has computationally deter-
mined M–Se distances for a number of 1-M analogues.[13c] We
found the Am–Se distance in 1-Am to be slightly shorter
As structures of 1-M have now been experimentally
determined where M = La, Ce, Nd, U, Pu, Am, their M–Se
Angew. Chem. Int. Ed. 2021, 60, 9459 –9466
ꢀ 2021 Wiley-VCH GmbH
www.angewandte.org 9461

Angewandte
Research Articles
Chemie
(3.0625(8) Š) than the calculat-
ed value (3.089 Š using a full
R = Ph model) which is also
longer than the value for Nd–
Se in 1-Nd (3.0801(3) Š).
Equivalent computational data
is not available to compare with
1-Nd. That the experimental
bond length in 1-Am is shorter
than calculated is not necessa-
rily a sign that calculations
underestimated the degree of
covalency in the Am–Se bond
which is likely small; rather, it
is likely that the calculations
overestimated the degree of
charge transfer which would
result in longer bond lengths
Figure 3. Highest occupied molecular orbitals (HOMOs) of 1-Am (left) and 1-Nd (right) (contour value
0.02 a.u.). For 1-Am there is significant mixing between the metal f and Se p orbitals, which is mostly
absent in the 1-Nd analogue. The entire f manifold for all molecules, as well as select ligand-centered
orbitals are available in the Supporting Information, Figures S54–S57.
(see below).[13c] Plots of theoretical M–E (M = U, Np, Pu, Am,
Cm; E = O, Te) normalized to U as a zero-point showed that
when O donors were investigated the M–E length decreased
across the 5f row as ionic radius decreases.[13c] In opposition to
this, when the donor was Te instead, the M–Te lengths
increased gradually to reach a maximum at Am before a sharp
drop at Cm. One way of interpreting this is to consider that
the U–Te bond has a much larger covalent contribution than
that of an Am-Te bond and is thus much shorter than it would
be if purely ionic.[28] Simultaneously, calculations predict
larger deviation from formal f orbital population at the Am3+
and 1-Nd. This MO is principally comprised of metal 5f/4f
character (75.43% for 1-Am; 91.14% for 1-Nd), but in 1-Am
the HOMO has a significant contribution from Se p orbitals
that is absent in 1-Nd. Table 2 compares the relative
populations of the f orbitals for 1-M (M = Am, Nd) and 2-
M (M = Am, Nd). In both 2-Am and 2-Nd the frontier f
Table 2: Percent atomic orbital compositions of frontier orbitals for 1-M
and 2-M (M=Am, Nd).[a]
Frontier Orbital
1-Am
Se p 5f
2-Am
O p 4f
1-Nd
Se p 4f
2-Nd
O p
5f
center than that of U3+ in model 6-coordinate [M{N(Se
=
PH2)2}3] and that a similar observation is seen for Eu3+ in the
Ln series.[13c] Therefore, in the presence of these soft donors,
formal Eu3+/Am3+ metal ions appear to behave “M2+-like”,
resulting in bond lengthening.[13c] This also correlates with our
inability to synthesize 1-Eu, and isolation of the Eu2+
complex, 3, instead. Previous Am studies with S-donor
ligands have also yielded Am–S/Nd–S bond length compar-
isons that do not conclusively reflect covalency differences.
For [nBu4N][M{S2P(tBu2C12H6)}4] (M = Am, Nd) the average
Am–S and Nd–S distances were 2.921(9) and 2.941(8) Š,
respectively; within the 3s criterion, as is the case for
[AsPh4][M{S2P(OEt)2}4] (M = Am, Nd).[24] Increases in
Am3+ vs. Nd3+ covalency within the [M{N(E = PR2)2}3] series
of molecules, even with soft Se donors, are not conclusive
based on bond length differences alone when taking into
account the 0.019 Š difference between the 8-coordinate
ionic radii of Am3+/Nd3+. Thus, we turned to computational
and spectroscopic techniques to probe electronic structure
differences.
To allow comparison with previous studies we utilized the
GGA PBE functional to compute the orbital compositions of
the ground states of 1-M and 2-M.[13c,d] We found significant
differences in the electronic structure between 1-Am and 1-
Nd when considering the mixing between the metal f and Se p
orbitals. Furthermore, there are substantial differences be-
tween the orbital compositions of 1-Am and 2-Am. On the
other hand, no significant differences were found in the
electronic structure of 1-Nd vs. 2-Nd. Figure 3 shows the
highest occupied molecular orbital (HOMO) for both 1-Am
HOMO
LUMO
70.82 17.34 87.25 1.18 91.14 0.00 95.80 0.00
75.43 13.34 84.74 1.08 97.38 0.00 92.28 0.00
[a] The compositions of all MOs are also included in the Supporting
Information (Tables S21–S24; Figures S54–S57 provide visualizations).
orbitals show almost no mixing with O p orbitals (see
Supporting Information Figures S56, S57, and Tables S23,
S24), though O-participation is greater in 2-Am than 2-Nd.
This agrees with previous work which showed that heavier
chalcogenide donors resulted in larger metal d participation
with ligand bonding orbitals, but that Ln/An differences are
mostly in the f orbitals.[13c,d] For 1-M (M = Am, Nd) the largest
metal d contribution to the bonding orbitals is around 9.6%,
while for both 2-M complexes the metal d contribution is
negligible. Varying the metal changes the metal f contribution
to bonding orbitals, while varying the chalcogenide changes
the metal d contribution.
Table 3 compares the average atomic spin density for all
molecules herein. Both 2-M (M = Am, Nd) have the expected
Table 3: Average atomic spin density for each species studied from DFT/
PBE.
Molecule
Am/Nd
Se/O
N
P
1-Am
2-Am
1-Nd
2-Nd
6.248
6.059
3.175
3.053
À0.038
À0.014
À0.024
À0.008
À0.011
0.000
À0.001
0.004
À0.007
À0.003
0.000
0.000
9462 www.angewandte.org
ꢀ 2021 Wiley-VCH GmbH
Angew. Chem. Int. Ed. 2021, 60, 9459 –9466

Angewandte
Research Articles
Chemie
metal formal spin (i.e. 6.0 for 2-Am and 3.0 for 2-Nd).
can therefore afford insight into fine electronic structure
properties.[12b,d]
However, the values for 1-M (M = Am, Nd) deviate modestly
indicating extra charge transfer to the metal center which
could indicate some covalent interaction. In the absence of
large structural effects, the spin density analysis and orbital
compositions point to a charge-transfer-like description of the
bonding as previously found by Kaltsoyannis.[9d,13c]
The experimental UV-vis-NIR spectra suggest differences
in electronic structure between the soft donor complexes 1-M
and the harder 2-M analogues (M = Am, Nd), and highlight
Am vs. Nd differences. The solution phase spectrum of 1-Am
(Figure 4a, black line) is dominated by a broad featureless
ligand to metal charge transfer (LMCT) band above
26000 cmÀ1, which appears resolved in the solid-state spec-
trum into peaks at 26214, 27269, 27950, and 28197 cmÀ1
(Figure 4a, red line). Below this band, the solution spectrum
shows two peaks (12050 and 12478 cmÀ1) that could corre-
The ground state of Am3+ (5f6) is predominantly 7F0 which
is formally non-magnetic (j = 0) which leads to sharp signals in
the NMR spectra of 1-Am and 2-Am. The 1H NMR spectrum
of 1-Am in [D]8-THF shows three multiplet resonances at 6.94
(t), 7.13 (t) and 7.69 (q) ppm. The 31P{1H} NMR spectrum of 1-
Am shows a sharp singlet at 125.60 ppm, (c.f. HLSe at
53.29 ppm). The 77Se{1H} NMR spectrum of 1-Am exhibits
a doublet at À98.10 ppm (1JSeP = 604.90 Hz; c.f. HLSe at
7
spond to the 7F0! F6 transition which, due to their proximity,
appear in the solid-state spectrum as a broad feature around
12204 cmÀ1. In the solution spectrum of 2-Am (Figure 4b,
black line) this region appears as three resolved peaks at
12189, 12346 and 12974 cmÀ1; once again the solid-state
spectrum (Figure 4b, red line) is broader with only two
resolved peaks (12326 and 12537 cmÀ1). For the higher
1
À189.14 ppm, JSeP = 800.50 Hz). Complex 2-Am was studied
in [D]2-DCM as 2-M complexes can coordinate Lewis bases
such as THF.[17a] The 1H NMR spectrum of 2-Am showed
three apparent multiplets at 7.11 (dq), 7.25 (t), and 7.63
(dt) ppm, and the 31P{1H} NMR spectrum exhibits a single
resonance at 70.95 ppm. Both 1-Nd and 2-Nd display
7
7
energy F0! F6 transition, neither the solid-state or solution
spectrum of 1-Am shows splitting of the 19448 cmÀ1 peak;
but, for 2-Am in solution this feature appears as four well-
resolved peaks (19026, 19531, 19677 and 19778 cmÀ1) which
in the solid-state are less resolved and present as two slightly
broad peaks at 19013 and 19661 cmÀ1. Table 4 shows these
data along with that for AmX3 (X = Cl, Br, and I) and the
Am3+ aquo ion.[29,32]
4
1H NMR spectra where the I9/2 ground state of Nd3+ (4f3)
leads to small paramagnetic shifts but broadening that
obscures the splitting of the aromatic protons (Supporting
Information).
Figure 4 shows solution UV-vis-NIR absorption spectra in
THF for 1-M (M = Am, Nd) and toluene for 2-M (M = Am,
Nd), and solid-state spectra for 1-Am and 2-Am from single
crystals. For Am3+ there are two intense characteristic
7
In 1-Am the transition to the F6 state shows a notable
nephelauxetic shift compared to 2-Am, similar in size to the
differences between AmCl3 and AmBr3.[29b,33] This suggests
decreased electron-electron repulsion in the f orbitals of 1-
7
Laporte-forbidden f!f peaks which derive from 7F0! F6
5
and 7F0! L6 transitions, respectively.[1a,29] For the Am3+ aquo
ion these appear at 12407 and 19881 cmÀ1.[1a] Splitting of the
higher energy peak is not always observed, but was reported
for [Am(Cptet)3], and was reproduced by SO-CASSCF/
NEVPT2 calculations; the splitting was rationalized by
increased splitting of excited crystal field states.[12d] Other
Am3+ complexes that show comparable peak splitting have O-
donor ligands such as: poly-borate anions for example,
Table 4: Selected peaks from the spectra of 1-Am, 2-Am, AmX3 (X=Cl,
Br, I), and the Am3+ aquo ion in frozen solution.[29,32]
1-Am
2-Am[a]
AmCl3
12376
AmBr3
12285
AmI3
Am3+
(aq)
7F6
5L6
12204
12326;
12537
19013;
19661
11930
12407
19881
{B9O13(OH)4},[30] or neutral O-donor ligands for example,
19448
19627
19500
19250
[25b,31]
=
O PCy3.
It is likely that this splitting derives not just
from ligand field effects, but also solution/solid phase
[a] The sets of two peaks for 2-Am have been assigned by proximity to the
known transition, but potentially arise from a different jj transition.
symmetry differences and vibronic coupling.[29,32] This peak
Figure 4. Normalized solution (black line) and solid-state (red line) UV-vis-NIR spectra for a) 1-Am and b) 2-Am; and c) solution UV-vis-NIR
spectra for 1-Nd (black line) and 2-Nd (red line). Spectra truncated to 8800–28500 cmÀ1 (1136–351 nm) and normalized to compare the f!f
transitions between the LSe and LO complexes.
Angew. Chem. Int. Ed. 2021, 60, 9459 –9466
ꢀ 2021 Wiley-VCH GmbH
www.angewandte.org 9463

Angewandte
Research Articles
Chemie
Am than 2-Am, and potentially greater covalent interaction.
The same general trend in energy shift also appears to be
present for transitions to the 5L6 level, although the
19013 cmÀ1 feature of 2-Am (this transition is split into
multiple peaks for 2-Am) is an outlier to this trend. When this
concept is applied to the peaks of 1-Nd (Figure 4c, black line),
2-Nd (Figure 4c, red line), NdCl3, and the Nd3+ aquo ion
(Table 5) we find that not only is there a very small difference
between the spectra of highly ionic 2-Nd, NdCl3, and the aquo
ion, but that 1-Nd is also essentially the same as these species.
This suggests that all of these Nd3+ compounds/ions are
approximately as ionic as each other. The exception to this is
4
4
the hypersensitive I9/2! G5/2 transition which is strongly
influenced by local differences in polarizability.
Table 5: Selected peaks from the spectra of 1-Nd, 2-Nd, NdCl3, and the
Nd3+ aquo ion in frozen solution.[34]
1-Nd
2-Nd
NdCl3
11438
Nd3+
(aq)
4F3/2
11289;
11390
12333
12429
13245;
13351;
13492
18868
11429
11460
12480
13500
4F5/2
12398
13528
12466
13437
4F7/2, 4S3/2
Figure 5. Solid-state experimental (black lines) and SO-CASPT2 simu-
lated (red lines) UV-vis-NIR spectra of a) 1-Am and b) 2-Am. The 1-Am
simulation utilized 6 electrons in 7 Am f orbitals, while the 2-Am
simulation utilized 10 electrons in 11 orbitals, which include 7 Am f
orbitals, 2 occupied ligand orbitals, plus an Am d and s orbital. No
empirical shift has been applied to align the simulated/experimental
spectra.
2K13/2, 4G7/2
18783
18939
19160
The UV-vis-NIR spectra were simulated using the com-
plete active-space self-consistent field theory (CASSCF)
including correlation effects via second-order perturbation
theory and spin-orbit coupling effects (SO-CASPT2).[35] The
complexes were modelled by replacing Ph groups with Me
groups (Supporting Information). Simulated spectra for 1-
Am and 2-Am are shown in Figures 5a and b, respectively
(red lines, see Supporting Information for discussion on 1-Nd
and 2-Nd). For 1-Am we utilized an active space that included
the Am f electrons and orbitals; for 2-Am we included two
occupied ligand orbitals, along with the Am 7s and one Am 6d
orbital. We searched for an active space with an Am 6d orbital
for 1-Am, but were unable to find a reasonable solution. We
find good agreement in the low energy (9000–15000 cmÀ1)
region, which correspond to f!f transitions that maintain the
spin-polarization of the ground state (S = 7 to S = 7 transi-
tions). The large peaks at ca. 19500 cmÀ1 in the experimental
spectra also appear in the simulations and involve a spin-flip
and change in the spin-symmetry of the system (S = 7 to S =
5). The energy mismatch between experimental and simu-
lated peaks in the 19500 cmÀ1 region for 1-Am (Figure 5a),
has been observed in other Am3+ simulated spectra and likely
derives from residual electron correlation effects not fully
captured in computationally affordable basis sets of active
spaces.[12d] For 2-Am, the simulation accurately describes the
energy of the characteristic ca. 19500 cmÀ1 peak using a larger
active space (Figure 5b). However, we were unable to find
a larger active space that resulted in a comparable improve-
ment for 1-Am. While the larger active space used with 2-Am
includes a metal d and s orbital, the observed transitions are
almost exclusively LMCT states centered on the f orbitals.
Though the 6d and 7s orbitals are too high in energy for
transitions in our experimental range, they play a role in
balancing the wavefunction in the perturbation theory treat-
ment. While there are few apparent LMCT states involving
the Am d orbital in 2-Am, we leave the question open
whether transitions involving the metal d orbitals are
important in the 1-Am spectrum as we were unable to find
a suitable active space that included Am d orbitals. The larger
Am d orbital contribution to the bonding orbitals in 1-Am
compared to 2-Am could result in ligand!d charge transfer
states. We observed similar accuracy and features for 1-Nd
and 2-Nd (Figure S52). We note the simulated spectrum of 2-
Am does not capture the 19013 cmÀ1 shoulder present in the
solid-state experimental spectrum.
The experimental UV-vis-NIR spectra suggest a modest
degree of increased electron delocalization in the orbitals
involved in several key f!f transitions for soft Se-donor 1-
Am over those in hard O-donor 2-Am; a change absent in
both Nd analogues. This is reflected in the MO compositions
of all four complexes (Figure 3, Figures S53–S57), which show
greater Am 5f participation in bonding MOs for 1-Am than 2-
Am; again, the Nd complexes are largely the same as each
other. Finally, good agreement between simulated and
experimental spectra lend credence to this argument that
there is a small observed increase in what could be considered
covalency in 1-Am when compared to 1-Nd, or 2-M (M = Am,
Nd), though all four are still predominantly ionic.
9464 www.angewandte.org
ꢀ 2021 Wiley-VCH GmbH
Angew. Chem. Int. Ed. 2021, 60, 9459 –9466

Angewandte
Research Articles
Chemie
Conclusion
[1] a) L. R. Morss, N. M. Edelstein, J. Fuger, The Chemistry of the
Actinide and Transactinide Elements, Springer, Dordrecht, 2006;
b) J. N. Mathur, M. S. Murali, K. L. Nash, Solvent Extr. Ion Exch.
2007, 19, 357.
The first molecular transplutonium complex featuring an
=
actinide-Se bond in [Am{N(Se PPh2)2}3] (1-Am) has been
synthesized and compared through structural/spectroscopic
[2] a) B. F. Myasoedov, M. S. Milyukova, E. V. Kuzovkina, D. A.
Malikov, N. S. Varezhkina, J. Less-Common Met. 1986, 122, 195;
b) C. J. Dares, A. M. Lapides, B. J. Mincher, T. J. Meyer, Science
2015, 350, 652; c) K. McCann, D. M. Brigham, S. Morrison, J. C.
Braley, Inorg. Chem. 2016, 55, 11971.
[3] a) L. Karmazin, M. Mazzanti, J. PØcaut, Chem. Commun. 2002,
654; b) D. Girnt, P. W. Roesky, A. Geist, C. M. Ruff, P. J. Panak,
M. A. Denecke, Inorg. Chem. 2010, 49, 9627; c) J. F. Corbey,
B. M. Rapko, Z. Wang, B. K. McNamara, R. G. Surbella III,
K. L. Pellegrini, J. M. Schwantes, Inorg. Chem. 2018, 57, 2278;
d) G. J. Deblonde, A. Ricano, R. J. Abergel, Nat. Commun. 2019,
10, 2438; e) T. S. Grimes, C. R. Heathman, S. Jansone-Popova,
A. S. Ivanov, V. S. Bryantsev, P. R. Zalupski, Inorg. Chem. 2020,
59, 138.
=
analysis to a hard O-donor complex, [Am{N(O PPh2)2}3] (2-
Am), along with their Nd congeners (1-Nd and 2-Nd).
Through ab initio and DFT calculations we found that the
energetic/spatial proximity of Se 4p orbitals on the soft-donor
1À
=
{N(Se PPh2)2} ligand to the Am 5f manifold resulted in
a good degree of mixing and Se composition in the frontier
orbitals of 1-Am. This was not matched in 1-Nd, which has
almost no Se composition in the frontier orbitals. Comparing
1À
=
Am–O/Nd–O bonding with the hard-donor {N(O PPh2)2}
ligand, we found that 2-Am and 2-Nd were extremely
comparable, with little to no O composition in frontier
orbitals. Interestingly, the computed electronic structure
variations did not manifest as measurable structural differ-
ences. However, differences are observed in the UV-vis-NIR
spectra of 1-Am and 2-Am that may reflect the more diffuse/
mixed nature of the occupied molecular orbitals in 1-Am,
whereas the spectrum for 2-Am is much closer to that of the
Am3+ aquo ion or AmCl3. Both 1-Nd and 2-Nd spectra are
largely identical, and comparable to the Nd3+ aquo ion or
NdCl3. Hence, we observe a greater soft-donor electronic
structure effect for Am3+ than for Nd3+.
[4] A. Bhattacharyya, P. K. Mohapatra, Radiochem. Acta 2019, 107,
931.
[5] a) M. P. Kelley, J. Su, M. Urban, M. Luckey, E. R. Batista, P.
Yang, J. C. Shafer, J. Am. Chem. Soc. 2017, 139, 9901; b) J. Su,
E. R. Batista, K. S. Boland, S. E. Bone, J. A. Bradley, S. K. Cary,
D. L. Clark, S. D. Conradson, A. S. Ditter, N. Kaltsoyannis, J. M.
Keith, A. Kerridge, S. A. Kozimor, M. W. Loble, R. L. Martin,
S. G. Minasian, V. Mocko, H. S. La Pierre, G. T. Seidler, D. K.
Shuh, M. P. Wilkerson, L. E. Wolfsberg, P. Yang, J. Am. Chem.
Soc. 2018, 140, 17977.
[6] a) M. Mazzanti, R. Wietzke, J. PØcaut, J. M. Latour, P. Maldivi,
M. Remy, Inorg. Chem. 2002, 41, 2389; b) T. Mehdoui, J. C.
Berthet, P. ThuØry, L. Salmon, E. Riviere, M. Ephritikhine,
Chem. Eur. J. 2005, 11, 6994.
[7] a) M. A. Denecke, P. J. Panak, F. Burdet, M. Weigl, A. Geist, R.
Klenze, M. Mazzanti, K. Gompper, C. R. Chim. 2007, 10, 872;
b) Z. Kolarik, Chem. Rev. 2008, 108, 4208.
[8] a) C. Adam, P. Kaden, B. B. Beele, U. Müllich, S. Trumm, A.
Geist, P. J. Panak, M. A. Denecke, Dalton Trans. 2013, 42, 14068;
b) C. Adam, B. B. Beele, A. Geist, U. Mullich, P. Kaden, P. J.
Panak, Chem. Sci. 2015, 6, 1548.
[9] a) M. J. Tassell, N. Kaltsoyannis, Dalton Trans. 2010, 39, 6719;
b) I. Kirker, N. Kaltsoyannis, Dalton Trans. 2011, 40, 124;
c) M. L. Neidig, D. L. Clark, R. L. Martin, Coord. Chem. Rev.
2013, 257, 394; d) N. Kaltsoyannis, Inorg. Chem. 2013, 52, 3407;
e) T. Vitova, I. Pidchenko, D. Fellhauer, P. S. Bagus, Y. Joly, T.
Pruessmann, S. Bahl, E. Gonzalez-Robles, J. Rothe, M. Altma-
ier, M. A. Denecke, H. Geckeis, Nat. Commun. 2017, 8, 16053.
[10] a) S. T. Liddle, Angew. Chem. Int. Ed. 2015, 54, 8604; Angew.
Chem. 2015, 127, 8726; b) A. Kerridge, Chem. Commun. 2017,
53, 6685.
Acknowledgements
We thank the U.S. Department of Energy, Office of Science,
Basic Energy Sciences (DOE-BES), Heavy Element Chemis-
try Program at Los Alamos National Laboratory (LANL)
(A.J.G., E.R.B., P. Y. , B.L.S., S.A.K., F.D.W.; DE-AC52-
06NA25396) for experimental Se donor work and theory, and
the Center for Actinide Science and Technology (CAST), an
Energy Frontier Research Center (EFRC) funded by DOE-
BES for experimental O donor work (A.J.G., T.E.A.S.; DE-
SC0016568). C.A.P.G. thanks a distinguished J. R. Oppen-
heimer Postdoctoral Fellowship (20180703PRD1) and L.M.S.
the Exploratory Research program (20190091ER) both under
LANL-LDRD (Laboratory Directed Research and Develop-
ment). A.W.S. thanks a G.T. Seaborg Postdoctoral Fellowship
at LANL. During manuscript preparation, we became aware
of complementary solution spectroscopic studies on Am3+
with the LO ligand by Natrajan and Kaden (S. D. Woodall,
Ph.D. Thesis, 2014, The University of Manchester), to be
published separately.
[11] N. Kaltsoyannis, Chem. Eur. J. 2018, 24, 2815.
[12] a) S. G. Minasian, J. M. Keith, E. R. Batista, K. S. Boland, D. L.
Clark, S. D. Conradson, S. A. Kozimor, R. L. Martin, D. E.
Schwarz, D. K. Shuh, G. L. Wagner, M. P. Wilkerson, L. E.
Wolfsberg, P. Yang, J. Am. Chem. Soc. 2012, 134, 5586; b) J. N.
Cross, J. Su, E. R. Batista, S. K. Cary, W. J. Evans, S. A. Kozimor,
V. Mocko, B. L. Scott, B. W. Stein, C. J. Windorff, P. Yang, J. Am.
Chem. Soc. 2017, 139, 8667; c) S. K. Cary, J. Su, S. S. Galley, T. E.
Albrecht-Schmitt, E. R. Batista, M. G. Ferrier, S. A. Kozimor, V.
Mocko, B. L. Scott, C. E. Van Alstine, F. D. White, P. Yang,
Dalton Trans. 2018, 47, 14452; d) C. A. P. Goodwin, J. Su, T. E.
Albrecht-Schmitt, A. V. Blake, E. R. Batista, S. R. Daly, S.
Dehnen, W. J. Evans, A. J. Gaunt, S. A. Kozimor, N. Lichten-
berger, B. L. Scott, P. Yang, Angew. Chem. Int. Ed. 2019, 58,
11695; Angew. Chem. 2019, 131, 11821; e) J. M. Sperling, E. J.
Warzecha, C. Celis-Barros, D. C. Sergentu, X. Wang, B. E.
Klamm, C. J. Windorff, A. N. Gaiser, F. D. White, D. A. Beery,
Conflict of interest
The authors declare no conflict of interest.
Keywords: actinides ·americium ·covalency ·spectroscopy ·
structure
Angew. Chem. Int. Ed. 2021, 60, 9459 –9466
ꢀ 2021 Wiley-VCH GmbH
www.angewandte.org 9465

Angewandte
Research Articles
Chemie
A. T. Chemey, M. A. Whitefoot, B. N. Long, K. Hanson, P.
Kogerler, M. Speldrich, E. Zurek, J. Autschbach, T. E. Albrecht-
Schçnzart, Nature 2020, 583, 396.
Barros, J. M. Sperling, A. N. Gaiser, C. J. Windorff, T. E.
Albrecht-Schçnzart, Inorg. Chem. 2020, 59, 16291.
[25] a) S. S. Galley, J. M. Sperling, C. J. Windorff, M. Zeller, T. E.
Albrecht-Schmitt, S. C. Bart, Organometallics 2019, 38, 606;
b) C. J. Windorff, C. Celis-Barros, J. M. Sperling, N. C. McKin-
non, T. E. Albrecht-Schmitt, Chem. Sci. 2020, 11, 2770.
[26] R. Ketkaew, Y. Tantirungrotechai, P. Harding, G. Chastanet, P.
Guionneau, M. Marchivie, D. J. Harding, Dalton Trans. 2021, 50,
1086.
[27] A. P. Bassett, R. Van Deun, P. Nockemann, P. B. Glover, B. M.
Kariuki, K. Van Hecke, L. Van Meervelt, Z. Pikramenou, Inorg.
Chem. 2005, 44, 6140.
[28] M. P. Kelley, I. A. Popov, J. Jung, E. R. Batista, P. Yang, Nat.
Commun. 2020, 11, 1558.
[13] a) A. J. Gaunt, B. L. Scott, M. P. Neu, Chem. Commun. 2005,
3215; b) A. J. Gaunt, B. L. Scott, M. P. Neu, Angew. Chem. Int.
Ed. 2006, 45, 1638; Angew. Chem. 2006, 118, 1668; c) K. I.
Ingram, M. J. Tassell, A. J. Gaunt, N. Kaltsoyannis, Inorg. Chem.
2008, 47, 7824; d) A. J. Gaunt, S. D. Reilly, A. E. Enriquez, B. L.
Scott, J. A. Ibers, P. Sekar, K. I. Ingram, N. Kaltsoyannis, M. P.
Neu, Inorg. Chem. 2008, 47, 29.
[14] C. J. Windorff, J. M. Sperling, T. E. Albrecht-Schçnzart, Z. Bai,
W. J. Evans, A. N. Gaiser, A. J. Gaunt, C. A. P. Goodwin, D. E.
Hobart, Z. K. Huffman, D. N. Huh, B. E. Klamm, T. N. Poe, E.
Warzecha, Inorg. Chem. 2020, 59, 13301.
[29] a) W. T. Carnall, B. G. Wybourne, J. Chem. Phys. 1964, 40, 3428;
b) R. G. Pappalardo, W. T. Carnall, P. R. Fields, J. Chem. Phys.
1969, 51, 1182.
[30] M. J. Polinski, S. Wang, E. V. Alekseev, W. Depmeier, T. E.
Albrecht-Schmitt, Angew. Chem. Int. Ed. 2011, 50, 8891; Angew.
Chem. 2011, 123, 9053.
[15] M. B. Jones, A. J. Gaunt, J. C. Gordon, N. Kaltsoyannis, M. P.
Neu, B. L. Scott, Chem. Sci. 2013, 4, 1189.
[16] C. R. Groom, I. J. Bruno, M. P. Lightfoot, S. C. Ward, Acta
Crystallogr. Sect. B 2016, 72, 171.
[17] a) M. A. Katkova, M. E. Burin, A. A. Logunov, V. A. Ilichev,
A. N. Konev, G. K. Fukin, M. N. Bochkarev, Synth. Met. 2009,
159, 1398; b) H. V. Stafford, PhD. Thesis, The University of
Manchester, 2019.
[31] G. Tian, D. K. Shuh, C. M. Beavers, S. J. Teat, Dalton Trans. 2015,
44, 18469.
[32] W. T. Carnall, J. Less-Common Met. 1989, 156, 221.
[33] A. Barbanelꢀ, Radiochim. Acta 1997, 78, 91.
[18] S. W. Magennis, S. Parsons, Z. Pikramenou, Chem. Eur. J. 2002,
8, 5761.
[34] W. T. Carnall, P. R. Fields, K. Rajnak, J. Chem. Phys. 1968, 49,
4424.
[19] R. D. Shannon, Acta Crystallogr. Sect. A 1976, 32, 751.
[20] M. Geissinger, J. Magull, Z. Anorg. Allg. Chem. 1997, 623, 755.
[21] M. Llunell, D. Casanova, J. Cirera, P. Alemany, S. Alvarez,
SHAPE 2.1: Program for the Stereochemical Analysis of
Molecular Fragments by Means of Continuous Shape Measures
and Assosciated Tools, http://www.ee.ub.edu, 2013.
[22] W. Clegg, A. J. Blake, J. M. Cole, J. S. O. Evans, P. Main, S.
Parsons, D. J. Watkin, Crystal Structure Analysis, Oxford Uni-
versity Press (2nd ed.) 2009, p. 240.
[23] a) J. H. Burns, J. R. Peterson, J. N. Stevenson, J. Inorg. Nucl.
Chem. 1975, 37, 743; b) M. P. Jensen, A. H. Bond, J. Am. Chem.
Soc. 2002, 124, 9870; c) S. K. Cary, M. Vasiliu, R. E. Baumbach,
J. T. Stritzinger, T. D. Green, K. Diefenbach, J. N. Cross, K. L.
Knappenberger, G. Liu, M. A. Silver, A. E. DePrince, M. J.
Polinski, S. M. Van Cleve, J. H. House, N. Kikugawa, A.
Gallagher, A. A. Arico, D. A. Dixon, T. E. Albrecht-Schmitt,
Nat. Commun. 2015, 6, 6827.
[35] a) B. O. Roos, P. R. Taylor, P. E. M. Sigbahn, Chem. Phys. 1980,
48, 157; b) P.-Š. Malmqvist, B. O. Roos, Chem. Phys. Lett. 1989,
155, 189; c) C. Angeli, R. Cimiraglia, S. Evangelisti, T. Leininger,
J. P. Malrieu, J. Chem. Phys. 2001, 114, 10252; d) C. Angeli, R.
Cimiraglia, J.-P. Malrieu, Chem. Phys. Lett. 2001, 350, 297; e) C.
Angeli, R. Cimiraglia, J.-P. Malrieu, J. Chem. Phys. 2002, 117,
9138; f) F. Neese, J. Chem. Phys. 2005, 122, 034107; g) F. Neese,
WIRES Comput. Mol. Sci. 2018, 8, e1327.
[36] Deposition Numbers 2020987, 2020988, 2020989, 2020990,
2020991, 2020992, 2020993, 2020994, 2020995, 2020996,
2020997, and 2020998 contain the supplementary crystallo-
graphic data for this paper. These data are provided free of
charge by the joint Cambridge Crystallographic Data Centre and
Fachinformationszentrum Karlsruhe Access Structures service
www.ccdc.cam.ac.uk/structures.
[24] a) J. N. Cross, J. A. Macor, J. A. Bertke, M. G. Ferrier, G. S.
Girolami, S. A. Kozimor, J. R. Maassen, B. L. Scott, D. K. Shuh,
B. W. Stein, S. C. Stieber, Angew. Chem. Int. Ed. 2016, 55, 12755;
Angew. Chem. 2016, 128, 12947; b) R. D. M. Greer, C. Celis-
Manuscript received: December 29, 2020
Accepted manuscript online: February 2, 2021
Version of record online: March 10, 2021
9466 www.angewandte.org
ꢀ 2021 Wiley-VCH GmbH
Angew. Chem. Int. Ed. 2021, 60, 9459 –9466