Small molecule activation by uranium tris(aryloxides): Experimental and computational studies of binding of N2, coupling of CO, and deoxygenation insertion of CO2 under ambient conditions

ARTICLE

pubs.acs.org/JACS

Small Molecule Activation by Uranium Tris(aryloxides): Experimental

and Computational Studies of Binding of N₂, Coupling of CO, and

Deoxygenation Insertion of CO₂under Ambient Conditions

Stephen M. Mansell,^†Nikolas Kaltsoyannis,*^,‡and Polly L. Arnold*^,†

^†EaStCHEM School of Chemistry, University of Edinburgh, Joseph Black Building, The King’s Buildings, Edinburgh, EH9 3JJ, U.K.

^‡Department of Chemistry, University College London, 20 Gordon Street, London, WC1H 0AJ, U.K.

S Supporting Information

b

ABSTRACT: Previously unanticipated dinitrogen activation is ex-

hibited by the well-known uranium tris(aryloxide) U(ODtbp)₃,

U(OC₆H₃-Bu^t₂-2,6)₃, and the tri-tert-butyl analogue U(OTtbp)₃,

U(OC₆H₂-Bu^t₃-2,4,6)₃, in the form of bridging, side-on dinitrogen

complexes [U(OAr)₃]₂(μ-η²:η²-N₂), for which the tri-tert-butyl N₂

complex is the most robust U₂(N₂) complex isolated to date.

Attempted reduction of the tris(aryloxide) complex under N₂gave

only the potassium salt of the uranium(III) tetra(aryloxide) anion,

K[U(OAr)₄], as a result of ligand redistribution. The solid-state structure is a polymeric chain formed by each potassium cation

bridging two arenes of adjacent anions in an η⁶fashion. The same uranium tris(aryloxides) were also found to couple carbon

monoxide under ambient conditions to give exclusively the ynediolate [OCCO]^2ꢀdianion in [U(OAr)₃]₂(μ-η¹:η¹-C₂O₂), in direct

analogy with the reductive coupling recently shown to aﬀord [U{N(SiMe₃)₂}₃]₂(μ-η¹:η¹-C₂O₂). The related U^IIIcomplexes

U{N(SiPhMe₂)₂}₃and U{CH(SiMe₃)₂}₃however do not show CO coupling chemistry in our hands. Of the aryloxide complexes,

only the U(OC₆H₂-Bu^t₃-2,4,6)₃reacts with CO₂to give an insertion product containing bridging oxo and aryl carbonate moieties,

U₂(OTtbp)₄(μ-O)(μ-η¹:η¹-O₂COC₆H₂-Bu^t₃-2,4,6)₂, which has been structurally characterized. The presence of coordinated N₂

in [U(OTtbp)₃]₂(N₂) prevents the occurrence of any reaction with CO₂, underscoring the remarkable stability of the N₂complex.

The di-tert-butyl aryloxide does not insert CO₂, and only U(ODtbp)₄was isolated. The silylamide also reacts with carbon dioxide to

aﬀord U(OSiMe₃)₄as the only uranium-containing material. GGA and hybrid DFT calculations, in conjunction with topological

analysis of the electron density, suggest that the UꢀN₂bond is strongly polar, and that the only covalent UfN₂interaction is π

backbonding, leading to a formal (U^IV)₂(N₂)^2ꢀdescription of the electronic structure. The NꢀN stretching wavenumber is

preferred as a metric of N₂reduction to the NꢀN bond length, as there is excellent agreement between theory and experiment for

the former but poorer agreement for the latter due to X-ray crystallographic underestimation of r(NꢀN). Possible intermediates on

the CO coupling pathway to [U(OAr)₃]₂(μ-C₂O₂) are identiﬁed, and potential energy surface scans indicate that the ynediolate

fragment is more weakly bound than the ancillary ligands, which may have implications in the development of low-temperature and

pressure catalytic CO chemistry.

’ INTRODUCTION

produce alkanes and alkenes for use as fuels and bulk chemicals.^9,10

The catalysts used are heterogeneous and are based on cobalt or

iron, and the process requires temperatures (200ꢀ350 ꢀC) and

pressures signiﬁcantly (20ꢀ44 atm) above ambient.⁹

Carbon monoxide and dinitrogen are the two diatomic

molecules which have the strongest bonds found in the Periodic

Table (1076.5 ( 0.4 kJ mol^ꢀ1and 945.33 ( 0.59 kJ mol^ꢀ1

,

respectively).¹Despite this, dinitrogen is the source of nitrogen

atoms for both biological systems^2,3and the chemical industry,⁴

while carbon monoxide is an important C₁feedstock for con-

version into larger molecules in industrial processes.⁵In nature,

dinitrogen is reduced by nitrogenase enzymes involving active

sites that include a MoFe metal core as the active site.^3,6In

industry, the well-known HaberꢀBosch process produces 100

million tons of ammonia per year, a scale that is thought to be

similar to that observed in nature,⁷using heterogeneous catalysts

based on iron or ruthenium and at a huge energetic cost

(350ꢀ550 ꢀC and 150ꢀ350 atm).⁸The FischerꢀTropsch

process uses carbon monoxide and dihydrogen as feedstocks to

Industrially viable homogeneous HaberꢀBosch or Fischerꢀ

Tropsch processes that work under mild conditions have yet to

be realized,^10,11although there have been some signiﬁcant

advances in the further reactivity of metal-ligated dinitrogen

and carbon monoxide.¹²Examples of N₂reactions with group 4

metals include the reaction of bridging, side-on bound dinitrogen

coordinated to zirconium with dihydrogen and primary silanes,¹³

the reaction of a C₅Me₄H zirconium N₂complex with H₂to form

NH₃,¹⁴a Hf analogue with CO to form addition products,¹⁵and

Received: March 8, 2011

Published: May 17, 2011

r

2011 American Chemical Society

9036

dx.doi.org/10.1021/ja2019492 J. Am. Chem. Soc. 2011, 133, 9036–9051

|

Journal of the American Chemical Society

ARTICLE

the catalytic reduction of N₂to NH₃occurring with a few

turnovers at a single Mo center.^8,16f-Element dinitrogen chem-

istry remains poorly developed in comparison. The highly

contracted nature of the 4f orbitals leads to almost no orbital

contribution to lanthanide coordination chemistry, and therefore

binding of classical ligands used in transition metal chemistry,

such as N₂and CO, was not anticipated. The complex

[(C₅Me₅)₂Sm] was the ﬁrst f-element complex shown to bind

N₂, but the N₂is easily displaced, and in solution the N₂-bound

and non-N₂bound species are in equilibrium.¹⁷Since then, more

reducing Ln(II) complexes have been used, aiding N₂binding by

partially reducing the N₂ligand.^18,19The combination of LnA₃

(A = N(SiMe₃)₂, (C₅Me₅), (C₅Me₄H), OC₆H₃-Bu^t-2,6) and a

strong reductant such as Na or KC₈leads to dinitrogen reduction

if the Ln does not have an available þII oxidation state, aﬀording

{A₂(THF)Ln}₂(μ-η²:η²-N₂) in yields that can vary because

other reduction products are formed in some cases.^20,21Particu-

In f-element chemistry, CO has been reductively homologated

by the strongly reducing La^IIIand Sm^IIcomplexes, [La(η-

Cp*)₂]₂(μ-N₂) and [Sm(η-Cp*)₂(thf)₂], aﬀording a ketene

carboxylate dianion from three CO molecules at 90 psi pressures

in moderate isolated yields.^17,38Binding of CO to uranium(III) was

ﬁrst demonstrated in [(Me₃SiC₅H₄)₃U(CO)] with v_CO1922

cm ,

^{ꢀ1 39}and since for other tris(cyclopentadienyl) derivatives,^39ꢀ41

and for the chelating aryloxide complex [{(^tBuArO)₃tacn}U]₂(CO)

({(^tBuArO)₃tacn} = {C₆H₁₂[NCH₂C₆H₂Bu^t-3,5-O))₃}),⁴²but re-

ductive coupling has been demonstrated only in the past few years,

and for three systems: The sterically congested, trivalent, organo-

metallic uranium complexes U(η-C₈H₆{SiPrⁱ₃-1,4}₂)(η⁵-Cp^R)]

(Cp^R= C₅(CH₃)₅or C₅(CH₃)₄H) can reductively homologate

CO to form the deltate dianion,⁴³squarate dianion,⁴⁴or if exactly

1 equiv of CO is used per U, the ynediolate dianion.⁴⁵We recently

showed a CO coupling reaction that exclusively forms the ynediolate

dianion from a simple uranium starting material, U{N(SiMe₃)₂}₃,

using an excess of CO at ambient temperature and pressure, and that

heating the resultant complex [U{N(SiMe₃)₂}₃]₂(μ-η¹:η¹-OCCO),

resulted in the formal addition of a silylmethyl CꢀH bond across

the CC triple bond.⁴⁶

f-Element complexes are also capable of carbon dioxide reduction

and activation. Most commonly, insertion into a MꢀEσ-bond occurs,

forming esters, carbonates, or carbamates, according to the nature of E

(C, O, N). For reducing f-element complexes, deoxygenation is

also regularly seen, converting M^IIIcomplexes to M^IVꢀOꢀM^IVcom-

plexes,⁴⁷with the CO usually not identiﬁed in the products, except in

one example in which a particularly encumbered ligand system allows

both [{(^RArO)₃tacn}U]₂(η¹-CO) (R = Bu^t) and the μ-O adduct to

be isolated,^42,48and the unusual linear, O-bound CO₂molecule,

[{(^AdArO)₃tacn}U(η¹-OCO) (Ad = adamantyl), is formed when an

even larger ligand was used in the reaction with CO₂.⁴⁹

3ꢀ

larly notable was the isolation of the unusual N₂as a ligand

from the reaction of Y{N(SiMe₃)₂}₃with K under N₂or the

reaction of DyI₂þ 2 KOAr under N₂.^22,23Even a scandium N₂

complex [(C₅Me₄H)₂Sc]₂(μ-η²:η²-N₂) has recently been

reported.²⁴

The 5f orbitals are not as contracted as the 4f orbitals, which leads

to greater orbital participation in actinide bonding with the potential

for actinide compounds to access diﬀerent chemistry to the

lanthanides. Only four molecular dinitrogen uranium complexes

have been reported, although side-on binding is increasingly found

in electropositive metal systems.²⁵The side-on binding of N₂in the

seminal tren complex [UN(NCH₂CH₂NSiBu^tMe₂)₃]₂(μ-η²:η²-

N₂) occurs at ambient pressure and is reversed by freezeꢀpumpꢀ

thaw degassing of solutions;²⁶5 psi pressures are required to

crystallize [U(η-Cp*)(η-C₈H₄{SiⁱPr₃-1,4}₂)]₂(μ-η²:η²-N₂)

despite the formal reduction to diazenido (N₂^2ꢀ),²⁷and 80 psi

pressures of N₂are required to stabilize U(η-Cp*)₃(N₂),²⁸

whereas additional reduction from Mo^IIIallows the isolation

of {C₆H₃Me₂-3,5(Bu^t)N}₃U(μ-η¹:η¹-N₂)Mo{N(Bu^t)Ph}₃.²⁹

Gambarrota and co-workers have also reported the crystal

structure of a bridging dinitrido U^IV/U^Vcomplex from the reduc-

tion of a uranium(III) tetrapyrrole compound under dinitrogen

using potassium naphthalide as an external reducing reagent.³⁰

Interestingly, it has been observed that uranium is an eﬃcient

catalyst in the HaberꢀBosch process,³¹and hence these initial

results are of great potential interest.³²

While many metal complexes are capable of inserting CO into a

MꢀE bond, very few can mediate the CꢀC coupling of CO. With

transition metal systems, [M(CO)₂(dmpe)₂Cl] (M = Nb, Ta) can

be reduced with magnesium inthepresenceof Cp₂ZrCl₂or with Na/

Hg followed by addition of Me₃SiCl to aﬀord coupled and silylated

[M(Me₃SiOCtCOSiMe₃)(dmpe)₂Cl] (M = Nb, Ta; dmpe =1,2-

bis{dimethylphosphino}ethane).³³The Ta compound can also be

protonated to give coordinated HOCtCOH.³⁴Insertion of 1 and 2

equiv of CO into the weak RhꢀRh bond in the rhodium octaethyl-

porphyrin (oep) dimer [Rh(oep)]₂has been observed, and 12 atm

pressures of CO were used to force the equilibrium reaction over to a

double insertion product characterized by NMR spectroscopy as

[(oep)Rh(CO)(CO)Rh(oep)].³⁵A ditantalum hydride with a sup-

porting tridentate, trisaryloxide ligand was recently found to reduc-

tively couple six CO molecules, giving a chain of six carbon atoms

containing alternate single and double bonds in [Ta(2,6-(CH₂-3-^tBu-

5-Me-2-OPh)(thf)]₄(C₆O₆).³⁶It is known that alkali metals react

with CO to form ill-deﬁned MOCCOM materials, but these are

shock-sensitive and can be thermally unstable.³⁷

In light of the exciting chemistry displayed by the U^IIIsily-

lamide, we decided to explore the chemistry of some other

simple U^III‘starting materials’. The uranium tris(aryloxides)

U(ODtbp)₃(ODtbp = 2,6-di-tert-butylphenoxide) and [U-

(ODipp)₃]₂(ODipp =2,6,-diisopropylphenoxide) have been

previously reported from the reaction of U{N(SiMe₃)₂}₃with

HODtbp and HODipp, respectively, but their small molecule

reactivity remained relatively unexplored,^50,51and only the

molecular structure of the dimeric Dipp compound had been

reported.⁵²Herein, we show the CO coupling reaction can be

generalized to tris(aryloxide) complexes, both U(ODtbp)₃and

the new tri-tert-butyl analogue U(OTtbp)₃(OTtbp = 2,4,6-tri-

tert-butylphenoxide), and that these systems exhibit other small

molecule binding, namely the binding of N₂. We also report the

reactivity of related alkyls and amides U{CH(SiMe₃)₂}₃and

U{N(SiPhMe₂)₂}₃toward CO, and of U(ODtbp)₃, U(OTtbp)₃,

and U{N(SiMe₃)₂}₃toward CO₂. To yield further insight into

the Uꢀsmall molecule interactions, key features of the electronic

and geometric structures are explored using DFT and QTAIM

calculations.

’ RESULTS AND DISCUSSION

Dinitrogen Binding and Activation by U(OAr)₃Com-

plexes. The synthesis of U(ODtbp)₃1 has previously been

described;⁵²made from UN⁰⁰(N⁰⁰= N(SiMe₃)₂), and in the

3

absence of an X-ray-determined crystal structure, was attributed a

monomeric structure, in contrast to the structure of {U(ODipp)₃}₂

made by the same authors, which forms a dimer with η⁶-arene

coordination to U^III, depicted in Scheme 1.⁵²In our hands, the

9037

dx.doi.org/10.1021/ja2019492 |J. Am. Chem. Soc. 2011, 133, 9036–9051

Journal of the American Chemical Society

ARTICLE

Scheme 1. Dinitrogen Binding by U(OAr)₃

Figure 1. Thermal ellipsoid plot for (a) [U(ODtbp)₃]₂(μ-η²:η²-N₂) 3 crystallized from hexane; probability 50%, Me groups and lattice solvent omitted

for clarity. (b) Thermal ellipsoid plot for U(ODtbp)₃1 crystallized from toluene; probability 50%. Only one independent molecule in the asymmetric

unit (out of the three for 3, and out of two for 1) is shown. Symmetry operator used to generate symmetry generated atoms for 3: ꢀx, ꢀy þ 1, ꢀz þ 1.

reaction of di-tert-butylphenol with UN⁰⁰₃in hexane yields a dark,

green-black precipitate as described in the original manuscript, and a

saturated brown hexane supernatant, from which a small number of

red crystals were obtained after it had been decanted into a separate

vessel. An X-ray crystallographic study of these red crystals revealed

the complex to be a dinitrogen-coordinated dimer [U(ODtbp)₃]₂-

(μ-η²:η²-N₂) 3, Figure 1a, rather than the anticipated monomeric

U(ODtbp)₃complex 1, in which the general structural parameters

are consistent with the reduction of the side-on bound N₂molecule

to a diazenido unit by the two U^IIIcations, resulting in a nominally

five-coordinate U^IV(N₂)^2ꢀdimer.

This observation caused us to study further the reaction that

aﬀords U(ODtbp)₃. The less soluble, dark green-black material

which precipitates during the synthesis was recrystallized from

either n-hexane or toluene solution and identiﬁed as monomeric

U(ODtbp)₃1 by crystallography, Figure 1b. Investigation of

the same bulk product of the reaction using powder diﬀraction

(see Supporting Information), prior to this recrystallization

conﬁrmed that monomeric U(ODtbp)₃is the major product

that crystallizes from the reaction mixture, as originally suggested

by Sattelberger et al. No evidence of arene-binding has been

observed for 1, but the synthesis and characterization of the solvate

(thf)U(ODtbp)₃were described in the original manuscript.

Complex 1 is isostructural from either solvent, containing

nominally three-coordinate U^III(Figure 1b), but with three close

UꢀC distances to methyl groups on the ortho-^tBu substituents with

the average UꢀC distance (3.272 Å) suggesting a stabilizing

UꢀCH agostic interaction, similar to that shown by the pyramidal

U{N(SiMe₃)₂}₃⁵³and U{CH(SiMe₃)₂}₃(3.05 and 3.09 Å UꢀC

distances to agostic methyl carbon atoms, respectively).⁵⁴

The pseudo-C₃arrangement of the three aryloxides forms a

shallow UO₃pyramid in which the U^IIIcenter sits 0.81 Å away

from the O₃plane, an arrangement also seen in uranium tris-

(amides), and is explained by the polarized ion model.^53,55The

UꢀO bond lengths range from 2.149(4) to 2.165(3) Å (av.

2.159 Å), longer than the UꢀO distance in the U^IVcomplex

[U(ODtbp)₄] (2.135(4) Å).⁵¹Selected distances and angles

are collected in Table 1.

Thus, although the N₂-bound product is only a minor

component, this is still an unusual case of N₂binding, particularly

9038

dx.doi.org/10.1021/ja2019492 |J. Am. Chem. Soc. 2011, 133, 9036–9051

Journal of the American Chemical Society

ARTICLE

Table 1. Selected Bond Lengths (Å) for the Crystallographically Characterized Uranium Aryloxide Compounds U(ODtbp)₃1,

[(DtbpO)₃U]₂(N₂) 3, and [(TtbpO)₃U]₂(N₂) 4

U(OAr)₃1

U(ODtbp)₃from hexane 1 U(ODtbp)₃from toluene 1

(μ-η²,η²-N₂) 3, 4

[(DtbpO)₃U]₂(N₂) from hexane 3

[(TtbpO)₃U]₂(N₂) [(TtbpO)₃U]₂(N₂)

molecule

A

B

A

B

A

B

C

from hexane 4

from toluene 4

UꢀO_Ar

2.151(3)

2.157(3)

2.164(3)

2.157

2.161(3)

2.165(3)

2.164

2.149(4)

2.156(5)

2.161(4)

2.155

2.158(5)

2.161(5)

2.164(4)

2.161

2.120(6)

2.165(6)

2.176(7)

2.154

2.123(7)

2.156(7)

2.166(6)

2.148

2.144(6)

2.150(10)

2.175(9)

2.156

2.095(6)

2.146(6)

2.170(6)

2.137

2.110(2)

2.144(2)

2.151(2)

2.135

UꢀO_{Ar (av.)}

UꢀN₁

UꢀN₂

UꢀN_(av.)

NtN

2.159

2.152

2.393(11)

2.393(10)

2.393

2.389(10)

2.440(10)

2.415

2.396(10)

2.430(10)

2.413

2.413(9)

2.410(9)

2.386(3)

2.423(3)

2.407

2.411

2.410

1.163(19)

1.189

1.204(17)

3.050

1.201(19)

3.067

1.190(18)

1.236(5)

NtN_(av.)

U

C

3.262

3.284

3.400

3.315

3.272

3.200

3.236

3.245

3.227

3.267

3.280

3.394

3.313

3.200

3.249

3.255

3.234

3.164

3.088

3.123

3 3 3

U

C_(av.)

3 3 3

3.093

3.088

3.123

this product, grown from either n-hexane or toluene, were both

conﬁrmed to be the side-on bound N₂adduct [U(OTtbp)₃]₂-

(μ-η²:η²-N₂) 4 by single crystal X-ray diﬀraction experiments,

Figure 2 (with selected distances and angles in Table 1), with a

molecular structure similar to that seen for 3. The isolated yield

of [U(OTtbp)₃]₂(μ-η²:η²-N₂) 4 is 80%, and N₂activation

appears to be quantitative according to a range of spectroscopies,

even though the product is dried under dynamic vacuum.

The structures of dinitrogen-bound 3 and 4 (both solvates)

are very similar, showing dinitrogen bound side-on between two

uranium centers with each bound to three aryloxide ligands and

with an inversion center situated between the N atoms. The

NꢀN distances in the three independent molecules of 3 are

1.163(19), 1.204(17), and 1.201(19) Å (av. 1.189 Å), while in

two molecules of 4 they are 1.190(18) Å (from n-hexane) and

1.236(5) Å (from toluene). The data with the smallest uncer-

tainty in bond lengths (4 crystallized from toluene) gave

statistically signiﬁcant results which are consistent with a large

degree of reduction from free N₂(NꢀN 1.0975 Å) to diazenido

Figure 2. Thermal ellipsoid plot for [U(OTtbp)₃]₂(μ-η²:η²-N₂) 4

crystallized from toluene; probability 50%, lattice solvent and most

Me groups omitted for clarity. Symmetry operator used to generate

symmetry-generated atoms for 4: ꢀx þ 1, ꢀy þ 1, ꢀz þ 1.

2ꢀ

N₂(PhNdNPh 1.255 Å).⁵⁶Interestingly, a close UꢀC

contact between the uranium center and one CꢀH group is

seen (av. 3.09 Å to C41 in 3, 3.18 Å to C8 in 4.toluene) which is

shorter than the three contacts in U(ODtbp)₃.

since the reactions were carried out at ambient N₂pressures, and

without any particular precautions taken to retain the N₂. We

thus investigated the scope of dinitrogen binding to uranium

tris(aryloxides) by the synthesis of the new tri-tert-butylphen-

oxide analogue U(OTtbp)₃, which has allowed further study of

dinitrogen activation.

It is instructive to compare the NꢀN distance in 3 and 4 to

the four molecular actinide dinitrogen complexes that have

been reported: the side-on bound N₂in the tren complex

{UN(NCH₂CH₂NSiBu^tMe₂)₃}₂(μ-η²:η²-N₂) A with a Nꢀ

N distance 1.109(7) Å; {U(η-Cp*)(η-C₈H₄{SiⁱPr₃-1,4}₂)}₂-

(μ-η²:η²-N₂) B, NꢀN distance 1.232(10) Å; end-on terminal

bound [{U(η-Cp*)₃(N₂)] C, NꢀN distance 1.120(14) Å; the

Mo-stabilized [{C₆H₃Me₂-3,5}₂(Bu^t)N}₃U(μ-η¹:η¹-N₂)Mo-

In an analogous manner to U(ODtbp)₃, the reaction of UN⁰⁰

3

with 3 equiv of HOTtbp was carried out, Scheme 1. Under argon,

yellow-brown benzene solutions of U(OTtbp)₃2 are formed

quantitatively overnight from UN⁰⁰₃and 3 equiv of HOTtbp; 2 is

the same color as 1 and does not coordinate arenes. When the

reaction is repeated under N₂in hexane solution, dark yellow

microcrystalline material is formed in high yield. Single crystals of

{N(Bu^t)Ph}₃] D, NꢀN distance 1.232(11) Å. The NdN bond

2ꢀ

length in 4 is closest to those in B and D, consistent with N₂

,

while A and C have NꢀN distances indistinguishable from free N₂.

9039

dx.doi.org/10.1021/ja2019492 |J. Am. Chem. Soc. 2011, 133, 9036–9051

Journal of the American Chemical Society

ARTICLE

Table 2. Selected Experimental (4) and Computational Data

for [U(OPh)₃]₂(μ-η²:η²-N₂) and Free N₂(B3LYP unless

stated)^a

r(NꢀN)/Å

expt

1.190, 1.236

1.303

calcd PBE

calcd

1.255

calcd (free N₂)

expt

1.104

σ(NꢀN)/cm^ꢀ1

1451

calcd PBE

calcd

1237

1486

calcd (free N₂)

expt

2454

r(UꢀN) ave/Å

2.407

calcd

2.398

F (electron density)

UꢀN av.

UꢀO av.

NꢀN

NꢀN (free N₂)

UꢀN av.

UꢀO av.

NꢀN

0.073

0.112

0.461

0.661

Figure 3. UVꢀvisꢀNIR spectrum (TtbpO)₃U(N₂)U(OTtbp)₃4 in

toluene (red line) and of U(OTtbp)₃2 (blue line) formed from heating

the solution of (TtbpO)₃U(N₂)U(OTtbp)₃to 80 ꢀC. The spectrum of

U(ODtbp)₃1 (green line) is also overlaid. Lines arising from incomplete

solvent subtraction around 1700 cm^ꢀ1have been omitted.

2

r F (electron density Laplacian)

0.214

0.489

ꢀ0.974

ꢀ2.021

ꢀ0.010

ꢀ0.003

ꢀ0.558

ꢀ1.057

NꢀN (N₂)

UꢀN av.

UꢀO av.

NꢀN

H (energy density)

Comparable lanthanide-N₂systems are Cp*₂Ln(N₂)LnCp*₂, with

NꢀN bond lengths of 1.259(4) Å (Ln = Tm),¹⁸and 1.088(12) Å

(Ln=Sm).¹⁷We return to the signiﬁcance of the NꢀNbondlength

in the discussion of our calculations below.

NꢀN (N₂)

2

^aF, r F, and H data obtained at bond critical points.

There are notable diﬀerences in the position of the equilibrium

of N₂-coordination by 3 and 4. For the other f-element N₂

complexes described above, equilibria in solution are generally

observed and these are aﬀected by increasing the pressure of N₂.

However, no loss of N₂from 4 is observed after freezeꢀ

pumpꢀthaw degassing a solution, and N₂coordination is main-

tained under vacuum in the solid state, or when THF is added

to the sample, which is most surprising. If the heterobimetallic

Mo-containing complex is discounted, 4 appears to be the most

robust uranium N₂complex yet observed. The ¹H NMR

spectrum of the N₂-bound compound contains approximately

ﬁve very broad resonances between þ22 and ꢀ26 ppm, which

would be reasonable if there is a slow exchange between

magnetically diﬀerent aryl environments at ambient temperature.

A low-temperature ¹H NMR spectrum of a toluene solution of 4

contains 13 resonances attributable to an asymmetric low

temperature structure like that found in the X-ray study (see

the Supporting Information for a fuller interpretation), but upon

heating to 80 ꢀC (even under an atmosphere of N₂), sharp

resonances for the non-N₂bound complex are seen as N₂is

released. The ¹H NMR spectrum of base-free tri-tert-butyl

complex 2 is very similar to that for 1, containing one sharp

set of ligand resonances in the region þ17 to ꢀ7 ppm. This con-

version to non-N₂bound complex was conﬁrmed by UVꢀvisꢀ

NIR spectroscopy: heating a toluene solution of 4 in a gastight

UVꢀvis cell to 80 ꢀC under N₂for 20 min results in a modest

change in color from dark yellow to brown, but a signiﬁcant change

in the spectrum occurs as 4 is converted to 2 (Figure 3), a spectrum

which overlays very well with that of N₂-free 1. Storage of the

sample of 2 for 1 week shows no further change in the spectrum.

Raman spectroscopy was used to determine the extent of

reduction to the dinitrogen ligand in 4. A strong band at

1451 cm^ꢀ1is assigned as the N₂stretch; see Supporting

Information. The ¹⁵N₂analogue of 4 was also made, and the

Raman spectrum of ¹⁵N-4 showed a stretch at 1404 cm^ꢀ1very

similar to the predicted stretch of 1402 cm^ꢀ1for a harmonic

oscillator. According to the literature, no Raman active band

could be identiﬁed for other side-on bound uranium N₂com-

plexes, but an N₂stretch of 1425 cm^ꢀ1was measured for

[{(Me₃Si)₂N}₂(thf)Y]₂(μ-N₂) in agreement with the assigned

NN bond order of two.²²We attribute the signiﬁcant stability of

the N₂complex possibly to the increased electron donation from

the third tert-butyl group of the aryloxide enhancing the reducing

power of the U^IIIcenter, but possibly more importantly to the

neatly packed structure made by 4 in solution as well as in the

solid state. Solutions of 4 do not react with the other, more reactive

small molecules discussed below, namely CO and CO₂, until

solutions are heated above the temperature at which N₂dissociates

(80 ꢀC in toluene). For example, single crystals of 4 toluene were

3

grown from a saturated toluene solution of the reaction mixture

between 4 and excess CO at room temperature, attributed to

incomplete reaction of the N₂compound, and no reaction with

CO₂was observed for 4 unless the mixturewas heated (see below),

but 2 reacts to completion with CO within hours at room

temperature (still notably faster than UN⁰⁰₃reacts with CO).

In order to probe any parallels with the LnA₃/K systems

described by Evans (A = N⁰⁰, OAr, Cp), that can display further

N₂reduction chemistry, U(ODtbp)₃was treated with an excess

of K metal in toluene under N₂at ambient temperature with

rigorous exclusion of donor solvents such as THF. A red

crystalline solid formed in the resulting brown solution that

could not be separated from the excess potassium and was

insoluble in hydrocarbon solvents. The identity of the product

was determined by X-ray crystallography to contain no coordi-

nated N₂, but instead the U^III‘ate’ complex K[U(ODtbp)₄],

which forms a polymeric chain in the solid state, in which

9040

dx.doi.org/10.1021/ja2019492 |J. Am. Chem. Soc. 2011, 133, 9036–9051

Journal of the American Chemical Society

ARTICLE

Table 3. Selected Bond Lengths (Å) for the Crystallographically Characterized Uranium Aryloxide Compounds K[U(ODtbp)₄]_n

5, [(DtbpO)₃UOC]₂6, [(TtbpO)₃UOC]₂7, and (TtbpO)₄U₂(μ-O)(μ-O₂COTtbp)₂9

K[U(ODtbp)₄]_nfrom

[(DtbpO)₃UOC]₂from

[(TtbpO)₃UOC]₂from

(TtbpO)₄U₂(μ-O)(μ-

O₂COTtbp)₂9

toluene 5

toluene 6

benzene 6

toluene 7

UꢀO_Ar

2.244(4)

2.251(4)

2.274(4)

2.321(4)

2.273

2.098(5)

2.108(5)

2.112(5)

2.102(3)

2.111(3)

2.115(3)

2.104(6)

2.132(6)

2.146(7)

UꢀO_{Ar (av.)}

UꢀOCC

CtC

2.106

2.109

2.104

2.139

2.120(5)

1.147(15)

1.307(8)

2.112(3)

1.185(9)

1.298(6)

2.132(9)

1.27(2)

1.237(14)

CꢀO

UꢀμꢀO

UꢀO_carboxylate

2.095(3)

2.315(7)

2.371(7)

1.258(12)

1.253(13)

1.336(12)

UOꢀC_carboxylate

CꢀOAr_carboxylate

K1 C_(av.)

3.137

2.806

3.087

2.749

3 3 3

K1ꢀAr_centroid

K2 C_(av.)

3 3 3

K2ꢀAr_centroid

[U(ODtbp)₄] anions are linked through two η⁶-arene interac-

tions to a potassium cation on each side. The molecular structure

is shown in the Supporting Information, with selected distances

and angles collected in Table 2. The uranium center is four-

coordinate with UꢀO bond lengths ranging from 2.244(4) to

2.321(4) Å (av. 2.273 Å) which are signiﬁcantly longer than those

seen in 1 (av. 2.159 Å) and in the uranium(IV) analogue

[U(ODtbp)₄] (2.135(4) Å).⁵¹The potassiumꢀarene interac-

tion is symmetrical with respect to the center of the arene rings

(av. KꢀC 3.112 Å, Kꢀarene_centroid2.806 and 2.749 Å), a little

shorter than other KꢀC contacts in η⁶-arene coordination

previously observed (mean 3.285 Å).⁵⁷This result is similar to

the reduction of [U{N(SiMe₃)₂}₃] with potassium in thf that

generated [K(thf)₆][U{N(SiMe₃)₂}₄].²¹

Previous DFT calculations by one of us identiﬁed UfN₂

backbonding as the only covalent UꢀN₂interaction in A,^58,59

and we have here performed further DFT calculations on

[U(OPh)₃]₂(μ-η²:η²-N₂) as a model for 3 and 4, within the

constraint of C_isymmetry. Initially, three diﬀerent spin states were

investigated using the PBE functional; quintet, septet, and nonet,

corresponding to formal uranium oxidation states of IV, III, and II,

5

respectively. The A_gstate (i.e., (U^IV)₂(N₂^2ꢀ)) is clearly the

ground state, 155.8 kJ mol^ꢀ1and 285.6 kJ mol^ꢀ1, respectively,

more stable than the ⁷A_uand ⁹A_gstates. The NꢀN distance of the

PBE ⁵A_gstate (1.303 Å, Table 2), however, is ca. 0.1 Å longer than

that observed experimentally, and the NꢀN stretching wavenum-

ber (1237 cm^ꢀ1) smaller than that identiﬁed in the Raman

5

spectrum of 4, and hence the A_gstate was recalculated at the

Figure 4. Representations of the four highest occupied R spin molec-

ular orbitals in [U(OPh)₃]₂(μ-η²:η²-N₂) (B3LYP). H atoms omitted

for clarity.

B3LYP level. Table 3 shows that, while the NꢀN stretching

wavenumber and the UꢀN_avedistance are now almost exactly as

found experimentally, the NꢀN distance remains a little long.

Interestingly, however, it is exactly the same as in PhNdNPh.⁵⁶

Figure 4 shows the four highest occupied R spin MOs of the

no N₂contribution, but the other two (HOMO-1 and HOMO-3)

are π backbonding from the uranium into the NꢀN antibonding

π_glevels. This interaction is very similar to that found previously

5

B3LYP A_gstate. All four clearly have signiﬁcant uranium 5f

content, as well as OPh ring π character. Two of the orbitals have

9041

dx.doi.org/10.1021/ja2019492 |J. Am. Chem. Soc. 2011, 133, 9036–9051

Journal of the American Chemical Society

ARTICLE

for a model of the tren N₂system and accounts for the

lengthening of the NꢀN distance and reduction in vibrational

wavenumber from the values for free N₂(1.104 Å and 2454 cm^ꢀ1

,

respectively, at the present level of theory). The Hirshfeld charge

of the N atoms is calculated as ꢀ0.33, in agreement with a π

backbonding interaction.

Our previous calculations on the model trenꢀN₂system^59,60were

performed at a time when geometry optimizations of diuranium

organometallic species were not tractable. We were puzzled by the

short experimental r(NꢀN), as the π backbonding interaction

indicated by the calculations suggested strongly that the dinitrogen

unit should be signiﬁcantly perturbed from free N₂. We speculated

that the interactions between the bulky tren ligands in the experi-

mentally characterized compound were preventing the two ends of

the molecule from moving closer together, as would be expected

upon NꢀN lengthening and UꢀN shortening. Subsequent DFT

study of [U₂(μ²-N₂)(η⁵-C₅H₅)₂(η⁸-C₈H₈)₂],⁶¹in which geometry

optimizations were performed, resulted in a distinct lengthening of

the NꢀN distance from its free value, and we characterized the

system as containing two U(IV) f²centers bridged by a N₂^2ꢀligand.

This is highly reminiscent of the present calculations, and the weight

of evidence now indicates that the experimental NꢀN distances in

this type of compound are probably underestimated, to a greater or

lesser extent, by the diﬀraction experiment, whose data are based on

electron density rather than atomic position, so likely to err on the

short side for a multiply bonded diatomic fragment. When comparing

experiment with theory, we suggest that it is better to focus on NꢀN

stretching wavenumber, as we have done for the ﬁrst time here. As

Figure 5. Molecular graph of [U(OPh)₃]₂(μ-η²:η²-N₂) (B3LYP). U,

yellow; N, blue; O, red; C, gray; H, white; bond critical points, green;

ring critical points, purple; bond paths, gray.

UꢀN interaction as being strongly polar is consistent with

population analysis of the HOMO-1 and HOMO-3, which shows

that the orbitals have contributions of, respectively, 54% U f, 12%

N p and 58% U f, 4% N p.

Discussion of N₂Complexes. All the experimental and

computational evidence collected for the N₂complexes

[U(OAr)₃]₂(μ-N₂) indicates a formal oxidation of the uranium

centers to U^IVwith concomitant reduction of the dinitrogen

ligand to N₂^2ꢀ. The ligand exhibits a long N₂bond length, a

Raman stretch greatly reduced from free N₂, and the complexes

contain weak absorptions in the UVꢀvisꢀNIR spectrum con-

sistent with a formal U^IVoxidation state. The remarkable stability

of 4 can then be partially attributed to these factors although the

existence of 3 predominantly in its three-coordinate non N₂

bound form points to the influence of the specific aryloxide

ligand as being a highly important factor. We attribute the

significant stability of the N₂complex possibly to the increased

electron donation from the third tert-butyl group of the aryloxide

enhancing the reducing power of the U^III, but possibly more

importantly to the neatly packed structure made by 4 in solution

as well as in the solid state. Solutions of 4 must be heated to

eliminate N₂forcibly before other small molecules (CO and

CO₂, see below) can be activated. It is increasingly clear that

experimental NꢀN distances in this type of compound, i.e., one

containing a multiply bonded diatomic ligand, are probably

underestimated by the X-ray diffraction experiment, whose data

are based on electron density, giving slightly shorter NꢀN

separations than in reality. Here, we have focused on NꢀN

stretching wavenumber when comparing experiment with theory

and propose that the NꢀN bond stretch should be taken as a

more accurate metric for reduction than the distance. Although

5

noted above, although the present B3LYP calculation of the A_g

ground state overestimates the experimentally determined NꢀN

distance by ca. 0.07 Å for 3 and 4 hexane (but slightly less for

3

4 toluene), ν(NꢀN) matches to within 35 cm^ꢀ1

.

3

We have very recently begun to apply the quantum theory of

atoms-in-molecules (QTAIM)⁶²to f-element organometallic

complexes^63,64and have found it to be very useful in quantifying

the relative roles of ionicity and covalency. QTAIM tells us that there

is one bond critical point (BCP) between each pair of atoms that are

bonded to one another, and that chemical bonding interactions may

be characterized and classiﬁed according to the properties of the

BCPs;⁶⁵we have focused on the electron density F, itsLaplacianr F,

2

and the energy density H. Values of F greater than 0.2 are typical of

2

covalent bonds, and r F is generally signiﬁcantly less than zero for

such interactions, reﬂecting the concentration of electron density

along the bond path linking the bonded atoms. H at the BCP is

negative for interactions with signiﬁcant sharing of electrons, its

magnitude reﬂecting the ‘‘covalence’’ of the interaction. Strongly polar

bonds are characterized by low values of F at the BCP, together with a

2

local charge depletion (r F > 0).

BCP data are collected for selected bonds in ⁵A_g

[U(OPh)₃]₂(μ-η²:η²-N₂) in Table 2 (which also contains data

for N₂), and the molecular graph is shown in Figure 5. This

reveals BCPs between all atoms that would be expected to have a

bond between them, together with ring critical points at the

centers of the phenyl rings and each UꢀN₂triangle. The NꢀN

BCP data (Table 2) are typical of covalent bonds, and it is

5

the present B3LYP calculation of the A_gground state here

2

noticeable that the magnitude of F, r F, and H are all less in

overestimates the experimentally determined NꢀN distance by

ca. 0.07 Å, the ν(NꢀN) matches to within 35 cm^ꢀ1

.

complexed N₂than free N₂. This is consistent with the lengthen-

ing and weakening of the NꢀN bond on complex formation. By

contrast, the UꢀN and UꢀO BCP data indicate much less

covalency, being typical of strongly polar (or indeed ionic)

interactions, and very much in keeping with other Uꢀp-block

element bonds we have characterized.⁶³The description of the

CO Binding and Coupling Using Uranium Tris(aryloxides),

and Studies of CO Binding with Other Simple Homoleptic

U^IIIStarting Materials. Exposure of a degassed yellow-brown

toluene solution of U(ODtbp)₃1, or the tris(tri-tert-butylaryloxide)

analogue, U(OTtbp)₃2, to an atmosphere of carbon monoxide

9042

dx.doi.org/10.1021/ja2019492 |J. Am. Chem. Soc. 2011, 133, 9036–9051

Journal of the American Chemical Society

ARTICLE

Scheme 2. CO Binding and Coupling by U(OAr)₃and U(NR₂)₃

(an excess) over 16 h results in the precipitation of bright yellow

microcrystals of the U^IVynediolate complex, [U(ODtbp)₃]₂-

(μ-η¹:η¹-C₂O₂) 6, or darker orange [U(OTtbp)₃]₂(μ-η¹:η¹-

C₂O₂) 7, Scheme 2. Both complexes have been fully characterized,

and the formation is essentially quantitative. There is no indication

of the formation of higher homologues of [CO]_n^2ꢀin the products.

It is notable however that the more stable N₂adduct 4 must be

heated briefly in solution to ensure that the N₂is fully decoordinated

in order to allow the reductive CO coupling reaction to occur,

because samples of 4 have been isolated from room temperature,

12 h reactions between 4 and CO in toluene. The tri-tert-butyl

complex 7 is much less soluble than 6 in common solvents. The

central alkyne carbon in 6 resonates at 0 ppm in the solution ¹³C

NMR spectrum, confirmed by comparison with that of 6a

[U(ODtbp)₃]₂[μ-O¹³C¹³CO] made analogously from ¹³CO. We

recently reported the synthesis of U{N(SiPhMe₂)₂}₃, an analogue

6 benzene, respectively) but shorter in 7 1.237(14) Å although

3

the reason for this is not readily apparent. In 6 toluene, the

3

UꢀOAr bond lengths average to 2.106 Å, in 6 benzene they

3

average to 2.109 Å, and in 7 they average to 2.104 Å; the

UꢀOCC bond lengths are the same length within error for 6

(2.120(5) and 2.112(3) Å) and similar to 7 (2.132(9) Å). The U

atom sits out of the O_Arplane by 0.51 Å in 6 toluene, 0.42 Å in

3

6 benzene, and 0.48 Å in 7. The parameters for 6 are very similar

3 2

3

to those measured in [UN⁰⁰] (C₂O₂) (N⁰⁰= N(SiMe₃)₂) which

has a CtC bond length of 1.183(7) Å, an OꢀC bond length of

1.301(4) Å, and a UꢀO bond length of 2.102(2) Å.⁴⁶

Pleasingly, both ynediolate complexes are very thermally

3 2

robust. Unlike [UN⁰⁰] (C₂O₂), a sample of ynediolate 6 heated

to 80 ꢀC for ﬁve days neither decomposes nor reacts intramo-

lecularly, paving the way for other intermolecular reactions of the

C₂product.

of UN⁰⁰, which has the potential to form a protective pocket from the

DFT calculations, without symmetry constraints, have been

performed on [U(OPh)₃CO] and [U(OPh)₃CO]₂, simpliﬁed

from the experimental systems by the removal of tert-butyl

groups. Cloke and Green suggested a mechanism in which two

CO molecules bound to two uranium centers couple to form a

zigzag transition state which can either react with a further

molecule of CO, or rearrange to the linear ynediolate form which

is no longer reactive. Likewise here, four intermediates have been

located on a possible reaction pathway to form an ynediolate IV

from a monomeric carbonyl I. I has three unpaired electrons,

which mimics the ﬁrst step in the experimental procedure in

which neutral CO binds via C to a U^IIIf³center. The dinuclear

molecules IIꢀIV each have four unpaired electrons, representing

electron transfer from both U centers to form U^IV. The geome-

tries of the (UCO)_ncomponents of IꢀIV are shown in Figure 7,

and their relative energies and CO vibrations are collected in

Table 4. I has approximately C₃symmetry while IIꢀIV are close

to C_i. The UꢀOꢀCC angle and UꢀO distance in IV (174ꢀ and

2.147 Å) are extremely close to those obtained experimentally in

6, and the calculated CꢀC distance is intermediate between

those found for 6 and 7. The structures in Figure 7 are rather

similar to those calculated by Cloke and Green, although the

CꢀC distance in II is ca. 0.1 Å longer than for [U(η-COT)-

(η-Cp)CO] while that in III is ca. 0.1 Å shorter. In agreement

with Cloke and Green, we ﬁnd that only the monomer would be

3

three silyl phenyl groups around a U-coordinated small molecule;⁵⁵

the solid-state structure of the amide suggested an agostic interaction

between the ipso-carbon on the SiPh group with the U. In arene

solution, no reaction was observed between U{N(SiPhMe₂)₂}₃and

an excess (atmospheric pressure) of CO, even after heating to 65 ꢀC

for 24 h. We also studied the CO reactivity of the uranium tris(alkyl)

complex U{CH(SiMe₃)₂}₃which has similar agostic interactions to

U{N(SiMe₃)₂}₃between U and three silyl methyl CH groups. Upon

addition of CO to a blue solution of U{CH(SiMe₃)₂}₃, the solution

turned orange and ¹H NMR spectroscopy revealed the presence of

H₂C(SiMe₃)₂as the decomposition product.

The ynediolates 6 (two solvates) and 7 have been structurally

characterized by single crystal X-ray diﬀraction, and the molec-

ular structures of 6 and 7 are shown in Figure 6, with selected

distances and angles collected in Table 3. First, it is clear that

there is some variability in these structures with the central

UOCCOU unit linear in 6 toluene (C1AꢀC1ꢀO1 angle

3

179.4(13)ꢀ, U1ꢀO1ꢀC1 angle 177.3(5)ꢀ) and 7 (C1Aꢀ

C1ꢀO1 angle 180ꢀ, U1ꢀO1ꢀC1 angle 180ꢀ), but slightly bent

in 6 benzene (C1AꢀC1ꢀO1 angle 175.9(7)ꢀ, U1ꢀO1ꢀC1

3

angle 167.8(3)ꢀ). The CtC bond is 1.147(15) Å in 6 toluene

3

and 1.185(9)ꢀ in 7 benzene, consistent with an alkyne triple

bond but it is signiﬁcantly longer in 7 (1.27(2) Å), whereas the

3

CꢀO bond lengths are 1.307(8) and 1.298(6) (in 6 toluene and

3

9043

dx.doi.org/10.1021/ja2019492 |J. Am. Chem. Soc. 2011, 133, 9036–9051

Journal of the American Chemical Society

ARTICLE

Figure 6. (a) Thermal ellipsoid plot for [U(ODtbp)₃]₂(μ-η¹:η¹-C₂O₂) 6 grown from C₆D₆. Symmetry operator used to generate symmetry generated

atoms: ꢀx þ 2, ꢀy, ꢀz þ 1. (b) Thermal ellipsoid plot for [U(OTtbp)₃]₂(μ-η¹:η¹-C₂O₂) 7; from toluene. Symmetry operators used to generate

symmetry-generated atoms: #1, ꢀx þ y þ 1, ꢀx þ 1, z; #2, ꢀy þ 1, x ꢀ y, z; #3, ꢀx þ 4/3, ꢀy þ 2/3, ꢀz ꢀ 1/3. Ellipsoid probability 50%; lattice

solvent and Me groups omitted for clarity.

Figure 8. Relative energy of IV upon elongating either one UꢀOAr

distance or one UꢀOC distance.

Discussion of CO Complexes. The CO coupling reaction at

U^IIIthat results in the formation of a U^IV[(CO)_n]^2ꢀU^IVfrag-

ment appears to require a fine balance between steric protection

and coordinative unsaturation. Only loosely coordinated CO

complexes are formed by cyclopentadienyl complexes such as

[(C₅H₅)₃U(CO)]³⁹and [(C₅Me₅)₃U(CO)].⁴¹The presence of

agostic interactions between the U^IIIcenter and ligand CꢀH

bonds in [U{N(SiMe₃)₂}₃] and U(ODtbp)₃does not block the

CO coordination and subsequent reductive coupling, but the

η⁶ꢀareneꢀU interaction in [U(ODipp)₃]₂and η¹ꢀareneꢀU

interaction in [U{N(SiPhMe₂)₂}₃] seems sufficient to block CO

Figure 7. Calculated distances (Å) for the UCO units in IꢀIV.

Table 4. Relative Energies and CO Vibrations of IꢀIV

I

II

III

IV

E/kJ mol^ꢀ1

σ/cm^ꢀ1

0

ꢀ97

1721, 1665

ꢀ149

1356

ꢀ117

1407

1993

expected to have an identiﬁable CO IR band. We also concur

that the product IV is less stable than the intermediate III.

Cloke and Green suggested that this may arise from neglect of

the bulky substituents on the aromatic rings. Unfortunately,

calculations incorporating the full aryloxide ligands are

intractable.

In order to gain an estimate of which type of UꢀO bond in IV

is the stronger, we have performed scans of the potential energy

surfaces for lengthening one UꢀOAr bond and one UꢀOC

bond from their equilibrium value, and the results are shown in

Figure 8. In each case, no relaxation of the rest of the molecule

was allowed, so the curves shown in Figure 8 are upper limits to

the actual potential wells. Nevertheless, and although SCF

convergence failures precluded extending these bonds by much

more than 2 Å, it is clear that the potential well for the UꢀOAr

interaction is much deeper, strongly suggesting that the UꢀOAr

bond is stronger than the UꢀOC bond.

binding. The aryloxides couple CO noticeably faster than UN⁰⁰

3

couples CO, forming ynediolate complexes with less steric

pressure, because the aryloxides can pack more effectively than

silylamido ligands. The aryloxides that form [(CO)₂]^2ꢀcomplexes

that are thermally stable, and not susceptible to ligand degradation, as

was seen in the addition chemistry of the proximal silyl methyl group

in the silylamide complexes. We cannot rule out the possibility that

these latter two molecues also have lowered U^III/IVreduction

couples, but this seems unlikely to be the cause. We suggest that

the ligands on the tris(alkyl) complex U{CH(SiMe₃)₂}₃are

too reactive toward CO insertion to allow a clean reductive coupling

of CO in our hands; no evidence of the formation of acyl ligands

from CO insertion was observed, only CH₂(SiMe₃)₂, the usual

product of the simple thermal decomposition of U{CH(SiMe₃)₂}₃.

The solvent THF blocks CO coupling for both [U{N(SiMe₃)₂}₃]

and U(ODtbp)₃simply by metal coordination.

9044

dx.doi.org/10.1021/ja2019492 |J. Am. Chem. Soc. 2011, 133, 9036–9051

Journal of the American Chemical Society

ARTICLE

Scheme 3. Reactions of Uranium Tris(aryloxides) with Car-

bon Dioxide

Figure 9. Thermal ellipsoid of (OTtbp)₂U(μ-O)(μ-O₂COTtbp)₂U-

(OTtbp)₂, 9 grown from benzene solution; probability 50%, lattice

solvent and tert-butyl groups omitted for clarity. Symmetry operator

used to generate symmetry-generated atoms: y, x, ꢀz.

The absence of further intramolecular reaction in the

[U(OAr)₃]₂(C₂O₂) complexes is pleasing in that it suggests

further intermolecular reactions may now be accessible for the

ynediolate dianion. Particularly, as shown in Figure 8, the fact

that the UꢀOCt bond is weaker than the UꢀOAr bond

suggests a catalytic cycle could be accessible, based on the dis-

placement of the [OCtCO]^2ꢀion and reduction of the released

U^IVtris(aryloxide) byproduct. It has been observed that the

organic diether acetylenes, ROCtCOR, are diﬃcult to synthe-

size, a matter which is hampered by lower thermal stability for

derivatives with smaller R groups,⁶⁶so alternative synthetic

routes would be advantageous.

Reactions of Uranium Tris(aryloxides) with Carbon Diox-

ide. Reactions of uranium(III) complexes with carbon dioxide

have produced a number of interesting products including a

terminally bound CO₂complex⁶⁷and two-electron reduction of

CO₂to CO,⁴²as well as the formation of oxalate from two

molecules of CO₂by a Lu N₂complex.⁶⁸Thus, it is of interest to

identify potential CO₂chemistry with these simple tris(ligand)

systems. Exposure of a benzene solution of U(ODtbp)₃1 to 1 bar

of CO₂resulted in a lightening of the color of the solution, but ¹H

NMR spectroscopy indicated that [U(ODtbp)₄] 8 was the major

product arising from oxidation of the uranium center followed by

ligand redistribution, Scheme 3. Again, surprisingly no reaction

was observed between [U(OTtbp)₃]₂(N₂) and CO₂in benzene

solutions until the temperature of the solution was elevated to

allow the coordinated N₂to be displaced. At this point, green

crystalline material formed in the bottom of the flask, which

proved to be very poorly soluble, similarly to the dimeric 7. A

single crystal X-ray crystallographic study revealed the identity of

the complex as (OTtbp)₂U(μ-O)(μ-O₂COTtbp)₂U(OTtbp)₂

9, a symmetrical U^IVdimer resulting from U^IIIreduction of CO₂

and incorporation of the abstracted oxo atom. A further 1 equiv

of CO₂has inserted into one of the three OTtbp ligands, forming

an aryl carbonate; thus, 1.5 equiv of CO₂has been consumed per

uranium. The structure is shown in Figure 9, with selected

distances and angles collected in Table 3.

C₅Me₄H)(η-C₈H₄{SiⁱPr₃-1,4}₂)] and CO₂was found to give a

carbonate ligand which bridged between two uranium centers in a

μ-η¹:η²-CO₃fashion. Disorder in the carbonate unit prevents a

comparison of CꢀO bond lengths but the η²UꢀO distances are

slightly longer than those seen in 9 (2.422(10) Å).⁶⁹

Finally, it is noted that the reaction of UN⁰⁰₃with CO₂had not

been discussed previously. In our hands, exposure of dark blue

purple hexane solutions of UN⁰⁰to CO₂(excess, atmospheric

3

pressure) results in the formation of a pale green solution, and the

precipitation of a pale green material from which toluene recrys-

tallization aﬀords a solid which is determined to be U(OSiMe₃)₄,

although the ¹H NMR spectrum contains four magnetically inequi-

valent SiMe₃resonances, suggesting a more complex molecular

structure in solution. FTIR spectroscopic analysis of the volatiles

showed a strong absorption at 2185 cm^ꢀ1which correlates with the

formation of the isocyanate OdCdNSiMe₃as an elimination

product as the silylamido ligand is converted completely to a siloxide

ligand, via a carbamate insertion product.

Discussion of CO₂Complexes. The formation of a bridging

oxo group has precedent in the two-electron reduction of CO₂seen

previously in uranium(III) chemistry,⁴⁷with the assumed release of

CO as a byproduct. Insertion of CO₂into polar MꢀZ bonds is well-

known, and hence the complex has displayed two modes of

reactivity with CO₂. The lack of further CO₂insertion into the

two other Uꢀaryloxide bonds has not been observed previously and

is most likely attributable to the fact that the product crystallizes out

of solution at this point. The two-electron reduction of CO₂implies

that a molecule of CO is released in the reaction. However, the flask

still contains an excess of CO₂which must react in preference to the

substoichiometric CO. The conversion of silylamide to silanolate

ligands via CO₂insertion has precedent in other s- and d-block

Lewis acidic metal complexes.⁷⁰

’ CONCLUSIONS

The UꢀOAr bonds are 2.132(6) and 2.146(7) Å, similar to

the other distances seen. The UꢀμꢀO distance is shorter

(2.095(3) Å) and the UꢀOꢀU angle is not linear (140.4(5)ꢀ).

The carbonate moiety bridges somewhat asymmetrically (UꢀO

2.315(7) and 2.371(7) Å) between the two U centers, and the

bond lengths within the group indicate delocalization of the charge

(UOꢀC 1.258(12) and 1.253(13) Å). The reaction between [U(η-

U^IIIcomplexes which have previously been considered too

sterically protected by hydrocarbyl ligand groups to coordinate

further small molecules show binding and reduction of both N₂

and CO at ambient pressures, forming [U(OAr)₃]₂(μ-N₂) and

[U(OAr)₃]₂(μ-C₂O₂) complexes. The kinetic inertness of the

TtbpꢀN₂complex is remarkable and attributed mainly to high

9045

dx.doi.org/10.1021/ja2019492 |J. Am. Chem. Soc. 2011, 133, 9036–9051

Journal of the American Chemical Society

ARTICLE

steric congestion; this is particularly notable in comparison to 3,

in which the bound and activated N₂is signiﬁcantly less readily

captured. Metrics for di- and tri-tert-butylaryloxide N₂adducts do not

suggest signiﬁcant diﬀerences in the reducing capability of the U^III

centers in 1 and 2. Uranium complexes continue to produce a range

of interesting small molecule activation reactions,^32,71particularly

when the strongly reducing U^III/IVcouple is invoked.⁷²Computa-

tional analysis provides strong evidence that the N₂ligand in

[U(OAr)₃]₂(μ-N₂) has indeed been reduced, and that the

(U^IV)₂(N₂^2ꢀ) formulation is appropriate. The UꢀN₂binding is

found to be strongly polar overall; the only covalent UꢀN₂interac-

tion is π backbonding which leads to a lengthening and weakening of

the N₂bond. We have, for the ﬁrst time, been able to resolve the

discrepancy between theoretical and X-ray crystallographic data for

r(NꢀN) noted in the very limited number of previous studies of

uranium dinitrogen complexes, concluding that the experimental

distance is an underestimation. Rather, we prefer to rely on the NꢀN

stretching wavenumber as a better metric; the agreement between

theoretical and experimental ν(NꢀN) is excellent and entirely in

keeping with the (U^IV)₂(N₂^2ꢀ) description.

Our experimental results show that N₂coordination is not

always easily identiﬁed and may well be prevalent in many more

systems, even highly crowded f-block metal centers, than pre-

viously assumed. This would provide a pleasing explanation for

the wealth of N₂-reductive functionalization shown by

[Ln{N(SiMe₃)₂}₃]/K mixtures, in which precoordination of

N₂to the Ln center could render the N₂reducible by potassium.

Despite the fact that no N₂coordination has been observed in

solution for these and related f-block complexes, the subtle

diﬀerences in agostic interactions and the presence of donor

solvent clearly has signiﬁcant eﬀects on the reaction outcome, as

demonstrated by the diﬀerence in CO coupling reactivity by

U{N(SiRMe₂)₂}₃which is successful for R = Me but not for R =

Ph. The reaction of the U(OAr)₃complexes with CO₂is more

facile than that with CO and leads to a combination of CO₂

reduction and insertion at the same metal center.

resonances calibrated against external TMS. Infrared spectra were recorded

on Jasco 410 spectrophotometers. Solutions for UVꢀvisꢀNIR spectro-

photometry were made in a nitrogen-filled glovebox or on a Schlenk line

under argon, and spectra were recorded in a Teflon-tapped 10 mm quartz

cell on a Unicam UV1 spectrophotometer. Raman spectroscopy was

performed using a LabRam instrument equipped with a 50 mW HeꢀNe

laser of wavelength 632.8 nm. Powder diffraction samples were made up in a

glovebox, and the glass capillaries were flame-sealed prior to measurements

being taken at room temperature on a Bruker D8 diffractometer.

Uranium(III) Aryloxides. Slightly Improved Synthesis of U-

(ODtbp)₃1 and Isolation of the Adduct [U(ODtbp)₃]₂(μ-η²,η²-N₂)

3. a. Under an Argon Atmosphere, 1. Under an atmosphere of argon,

hexane (3 cm³, freezeꢀpumpꢀthaw degassed and stored under argon)

was added to a mixture of HO-2,6-^tBu₂C₆H₃(161 mg, 0.78 mmol, 3.1

equiv) and [U{N(SiMe₃)₂}₃] (181 mg, 0.25 mmol, 1 equiv) and the

solution was stirred for 16 h. The suspension was allowed to settle, and a

green/black precipitate was isolated from the brown supernatant by

cannula filtration, washed with hexane (3 cm³), and dried under vacuum

to yield base-free U(ODtbp)₃1 (174 mg, 0.20 mmol, 80% yield). The

C₆D₆NMR solvent was freezeꢀpumpꢀthaw degassed and refilled with

argon before transfer onto the solid sample of 1 which was then sealed in

a Youngs Teflon-valved NMR tube under argon. Characterization data

agree with the literature.^{52 1}H NMR (400 MHz, 298 K, C₆D₆) δ (ppm)

16.65 (d, ³J_HH= 7.9 Hz, 6 H, meta Ar-H), 13.75 (t, ³J_HH= 7.2 Hz, 3 H,

para Ar-H), ꢀ6.21 (s, 54 H, ^tBu).

b. Under a Dinitrogen Atmosphere, 1 and 3. Under an atmosphere of

dinitrogen, a solution of HO-2,6-^tBu₂C₆H₃(1.797 g, 8.71 mmol, 3.1 equiv) in

hexane (10 cm³) was added to a solution of [U{N(SiMe₃)₂}₃] (2.020 g,

2.81 mmol, 1 equiv) in hexane (10 cm³) and the solution was stirred for 16 h.

Stirring was then ceased, the dark black/green precipitate was allowed to

settle, and the supernatant brown solution was removed by cannula filtration.

The precipitate was then washed with hexane (10 cm³) at ꢀ10 ꢀC, and

the product, base-free U(ODtbp)₃1 was dried under vacuum (2.051 g,

2.40 mmol, 85% yield). Single crystals suitable for an X-ray diffraction

experiment were grown from a solution of 1 in hexanes at room temperature

and also from a solution of 1 in toluene at ꢀ30 ꢀC. Powder diffraction

measurements showed that the bulk material which precipitated from hexane

is pure U(ODtbp)₃. ¹H NMR (400 MHz, 298 K, C₆D₆) δ (ppm) 16.68

(d, ³J_HH= 8.3 Hz, 6 H, meta Ar-H), 13.78 (t, ³J_HH= 8.2 Hz, 3 H, para Ar-H),

ꢀ6.22 (s, 54 H, ^tBu). No change was observed after freezeꢀpumpꢀthaw

degassing of the sample. ¹³C NMR (101 MHz, 298 K, C₆D₆) δ (ppm) 210.4

(ipso O-Ar), 182.3 (ortho Ar), 138.9 (meta Ar), 129.2 (para Ar), 8.4

We suggest that the CO coupling to form the OCCO

ynediolate dianions may also be formed by many other strongly

reducing f-block or early d-block complexes, opening up the

coupling reaction to further chemistry of the C₂product. Since

the computational analyses of the bond strengths in the aryloxide

derivatives [U(OAr)₃]₂(μ-C₂O₂) show the ynediolate fragment

to be more weakly bound than the ancillary ligands, work is in

progress to identify catalytic routes to functionalize and remove

the OCCO fragment from the metal and close a cycle.

1

(CMe₃), ꢀ60.8 (CMe₃). H NMR (400 MHz, 298 K, C₆D₁₂) δ (ppm)

16.53 (d, ³J_HH= 8.3 Hz, 6 H, meta Ar-H), 13.49 (t, ³J_HH= 8.1 Hz, 3 H, para

t

Ar-H), ꢀ6.61 (s, 54 H, Bu). No change upon freezeꢀpumpꢀthaw

degassing. μ_eff(Evans NMR method) 3.3 μ_B. Mp: 195 ꢀC.

A small number of red crystals grown from the supernatant liquors of

reactions in hexanes under dinitrogen were also isolated and identiﬁed as

the N₂adduct [U(ODtbp)₃]₂(μ-η²,η²-N₂) 3.

’ EXPERIMENTAL SECTION

General Details. All manipulations were carried out under a dry,

oxygen-free dinitrogen, or argon where noted, atmosphere using stan-

dard Schlenk techniques or in MBraun Unilab or Vacuum Atmospheres

OMNI-lab gloveboxes unless otherwise stated. THF, toluene and

hexane were degassed and purified by passage through activated alumina

towers prior to use. Deuterated benzene, toluene, and cyclohexane were

boiled over potassium, vacuum transferred, and freezeꢀpumpꢀthaw

degassed three times prior to use. The compounds U{CH(SiMe₃)₂}₃,⁷³

Synthesis of U(OTtbp)₃2 and the Adduct [U(OTtbp)₃]₂(μ-η²,η²-N₂)

4. a. Under an Argon Atmosphere, 2. Under an atmosphere of argon,

C₆D₆(0.7 cm³, freezeꢀpumpꢀthaw degassed and stored under argon)

was added to HO-2,4,6-^tBu₃C₆H₂(20 mg, 0.077 mmol, 3.1 equiv) and

[U{N(SiMe₃)₂}₃] (18 mg, 0.025 mmol, 1 equiv) in a Youngs Teflon-

valved NMR tube and the solution was stored for 16 h, during which

time dark brown/yellow [U(OTtbp)₃] 2 was observed to form by ¹H

1

NMR spectroscopy along with HN(SiMe₃)₂quantitatively. H NMR

52

U{N(SiMe₂Ph)₂}₃,⁵⁵U{N(SiMe₃)₂}₃,⁷⁴and [U(ODipp)₃]₂were

(400 MHz, 298 K, C₆D₆) δ(ppm) 16.82 (s, 6 H, meta Ar-H), 5.12 (s, 27 H,

was made as previously described in the literature, and U{N(SiMe₃)₂}₃

was sublimed prior to use. The phenols HODtbp and HOTtbp were

sublimed prior to use while all other reagents were used as purchased

without further purification. ¹H and ¹³C NMR spectra were recorded on

Bruker AVA 400 or 500 MHz NMR spectrometers at 298 K. Chemical

shifts are reported in parts per million and referenced to residual proton

para ^tBu), ꢀ5.87 (s, 54 H, ortho ^tBu).

b. Under a Dinitrogen Atmosphere, 4. Under an atmosphere of

dinitrogen, a solution of HO-2,4,6-^tBu₃C₆H₂(1.640 g, 6.25 mmol, 3.1

equiv) in hexane (10 cm³) was added to a solution of [U{N(SiMe₃)₂}₃]

(1.459 g, 2.03 mmol, 1 equiv) in hexane (10 cm³) and the solution was

left for 16 h. The brown supernatant solution was removed by cannula

9046

dx.doi.org/10.1021/ja2019492 |J. Am. Chem. Soc. 2011, 133, 9036–9051

Journal of the American Chemical Society

ARTICLE

filtration, the dark yellow crystalline solid was then washed with hexane

(2 ꢁ 10 cm³), and the product, identified as [U(OTtbp)₃]₂(μ-η²,η²-

N₂) 4, was dried under vacuum (1.682 g, 0.81 mmol, 80% yield).

Single crystals suitable for an X-ray diffraction experiment were grown

from a solution of 4 in either hexane or toluene at room temperature. ¹H

NMR (400 MHz, 298 K, C₇D₈) δ (ppm) very broad signals: 21.9 (bs),

ꢀ1.2 (bs), ꢀ26.9 (bs). ¹H NMR (400 MHz, 333 K, C₇D₈) δ (ppm) 10.7

(bs), 2.5 (bs), ꢀ7.0 (bs). Small resonances for U(OTtbp)₃were also

observed, and by 353 K they represent the only species in solution. IR

(nujol mull) cm^ꢀ1(intensity): 1361 (m), 1274 (w), 1219 (s), 1190 (s),

1114 (s), 916 (w), 875 (w), 849 (m), 833 (s), 820 (m), 770 (w), 748

(m), 722 (w), 537 (m). Raman (powdered sample) cm^ꢀ1(intensity):

1600 (1184); 1451 (1099); 1245 (870); 1190 (794). The strong band at

1451 is assigned as the N₂stretch. Mp: 165ꢀ167 ꢀC. Anal. Calcd for

C₁₀₈H₁₇₄N₂O₆U₂: C, 62.59; H, 8.46; N, 1.35. Found: C, 62.10; H, 8.37;

N, 1.43.

(meta Ar), 35.0 (CMe₃), ꢀ0.37 (CtC) ꢀ5.7 (CMe₃). μ_eff(Evans NMR

method) 4.6 μ_Bper molecule, 2.3 μ_Bper U center. IR (nujol mull), cm^ꢀ1

(intensity): 1406 (s), 1262 (w), 1210 (s), 1191 (s), 1122 (m), 1099 (w),

864 (s), 820 (m). Mp: > 250 ꢀC. Anal. Calcd for C₈₆H₁₂₆O₈U₂: C, 58.56;

H, 7.20; N, 0. Found: C, 57.13; H, 7.33; N, 0.20. Single crystals suitable

for an X-ray diffraction experiment were grown from a solution of 5 in

toluene at ꢀ30 ꢀC.

U(ODtbp)₃reacts similarly with CO/H₂(1:2) to give the same product,

i.e., no reaction with H₂was observed at ambient pressures. There was no

change upon heating a solution of (DtbpO)₃UOCCOU(ODtbp)₃for 5

days at 80 ꢀC in C₆D₆.

Synthesis of (DtbpO)₃UO¹³C¹³COU(ODtbp)₃¹³C-6. In a Youngs

Teflon-valved NMR tube, a brown/yellow solution of U(ODtbp)₃

(24 mg, 0.028 mmol) in C₆D₆(0.7 cm³) was degassed using the

freezeꢀpumpꢀthaw method and ¹³C labeled carbon monoxide (95%

¹³C, atmospheric pressure) was admitted and the NMR tube was sealed.

¹³C NMR (101 MHz, 298 K, C₆D₆) δ (ppm) 232.6 (ipso O-Ar), 180.8

(ortho Ar), 134.5 (para Ar), 131.1 (meta Ar), 35.0 (CMe₃), ꢀ0.04

(strong, CtC) ꢀ5.6 (CMe₃).

Freezeꢀthaw degassing a toluene solution of 4 resulted in no loss of

N₂according to ¹H NMR spectroscopy, and the addition of THF to the

sample also resulted in no change, i.e., no binding of THF and no

displacement of N₂.

Synthesis of (TtbpO)₃UOCCOU(OTtbp)₃7. A brown/yellow solution

of (TtbpO)₃U(N₂)U(OTtbp)₃4 (304 mg, 0.147 mmol) in toluene

(20 cm³) was degassed using the freezeꢀpumpꢀthaw method, carbon

monoxide (>99.9% purity, atmospheric pressure, in excess) was

admitted, and the vessel was sealed and heated to 80 ꢀC briefly, before

being stirred for 4 days, to maximise precipitation of the product forming

a clear yellow solution with yellow crystalline solid. The solid was

isolated by filtration, washed with hexane (2 ꢁ 10 cm³), and dried under

vacuum. Yield: 174 mg, 0.083 mmol, 57%. The product is insoluble in

benzene or chloroform, and only very broad ¹H nuclear magnetic

resonances were observed in pyridine solution. Single crystals suitable

for an X-ray diffraction experiment were grown from a solution of 7 in

toluene cooled from 80 ꢀC to room temperature.

Synthesis of [U(OTtbp)₃]₂(μ-η²,η²-¹⁵N₂) ¹⁵N-4. HO-2,4,6-^tBu₃C₆H₂

(354 mg, 1.35 mmol, 3.1 equiv) and [U{N(SiMe₃)₂}₃] (313 mg, 0.435,

1 equiv) were weighed into an ampule, and degassed hexane (10 cm³)

was condensed onto the solids at ꢀ196 ꢀC. ¹⁵N-labeled dinitrogen

(98%þ, in excess) was then added to the vessel, and the purple solution

was allowed to warm to room temperature, changing the color to brown.

This was left to stand for 16 h, and the yellow-brown crystalline material

was washed with hexane (5 cm³) and dried under vacuum to afford

¹⁵N-4 (292 mg, 0.141, 65%). Raman (powder) υ (intensity) (cm^ꢀ1):

1404 (336).

Formation of U(OTtbp)₃2 by Loss of N₂. For NMR spectroscopic

analysis in C₆D₆, samples of 4 were heated in benzene to 80 ꢀC in order

to completely dissolve the samples, which resulted in loss of coordinated

N₂and formation of 2 in solution. ¹H NMR (400 MHz, 298 K, C₆D₆)

δ (ppm) 16.84 (s, 6 H, meta Ar-H), 5.14 (s, 27 H, para ^tBu), ꢀ5.90 (s, 54

H, ortho ^tBu). ¹³C NMR (101 MHz, 298 K, C₇D₈) δ (ppm) 207.6 (ipso

O-Ar), 182.0 (ortho Ar), 150.1 (para Ar), 135.2 (meta Ar), 38.9 (para

CMe₃), 36.8 (para CMe₃), 31.0 (ortho CMe₃), ꢀ60.4 (ortho CMe₃). μ_eff

(Evans NMR method) 3.4 μ_Bper molecule.

Synthesis of K[U(ODtbp)₄]_n5. A solution of U(ODtbp)₃1 (146 mg,

0.17 mmol) in toluene (8 cm³) was added to a vessel sealed with a

Young’s tap containing a potassium mirror (in excess), and the mixture

was sealed under nitrogen for 48 h. A brown solution was observed along

with red crystals which adhered to the potassium mirror and were

insoluble in noncoordinating solvents. The structure of the crystalline

product was determined to be K[U(ODtbp)₄]_n5 by an X-ray crystal-

lography study on a sample of these red crystals.

CO-Coupled Uranium(IV) Aryloxides. Synthesis of (DtbpO)₃

UOCCOU(ODtbp)₃6. A brown/yellow solution of U(ODtbp)₃(1.004 g,

1.17 mmol) in toluene (25 cm³) was degassed using the freezeꢀ

pumpꢀthaw method and carbon monoxide (>99.9% purity, atmospheric

pressure, in excess) was admitted to the vessel, and the mixture was sealed

and stirred for three days. Over 16 h, a bright yellow precipitate of

(DtbpO)₃UOCCOU(ODtbp)₃6 formed as the solution lightened to

bright yellow, and the reaction was stirred for a further 56 h. The precipitate

was isolated by cannula filtration, washed with hexane (10 cm³), and dried

under vacuum, yielding (DtbpO)₃UOCCOU(ODtbp)₃6 as a bright

yellow powder (468 mg, 0.265 mmol, 45%). A further crop of product

was isolated from the supernatant solution by reducing the volume to ca.

8 cm³under reduced vacuum and storage at ꢀ30 ꢀC for one week (171 mg,

0.36 mmol, total yield: 61%). ¹H NMR (400 MHz, 298 K, C₆D₆) δ (ppm)

13.91 (d, ³J_HH= 8.4 Hz, 12 H, meta Ar-H), 10.87 (t, ³J_HH= 8.4 Hz, 6 H,

para Ar-H), ꢀ7.35 (s, 108 H, ^tBu). ¹³C NMR (101 MHz, 298 K, C₆D₆)

δ (ppm) 233.1 (ipso O-Ar), 181.0 (ortho Ar), 134.6 (para Ar), 131.1

The reaction proceeds only slightly if stored at room temperature for

16 h, without initial heating. Heating the solution to 50 ꢀC for 50 min

results in a signiﬁcant conversion to product, but the reaction requires

heating to 80 ꢀC to achieve full conversion of all the starting material to 7

1

in a reasonable (sub-24 h) time period. H NMR (400 MHz, 298 K,

NC₅D₅) δ (ppm) 5.7 (very broad), ꢀ12.2 (extremely broad). (400

MHz, 298 K, C₇D₈) δ (ppm) 14.12 (s, 2 H), 3.78 (s, 9 H), ꢀ7.11 (s, 18

H). IR (nujol mull), cm^ꢀ1(intensity): 1423 (s), 1361 (s), 1220 (s), 1192

(s), 1115(s), 918 (w), 878 (w), 839 (s), 773 (w), 751 (m), 728 (m), 540

(m). Mp: > 250 ꢀC.

Lack of Reaction of the Solvate (thf)U(ODtbp)₃and CO. In a Youngs

Teflon-valved NMR tube, to a brown/yellow solution of U(ODtbp)₃

(21 mg, 0.025 mmol) in C₆D₆(0.7 cm³) was added thf (0.1 cm³, excess),

forming (thf)U(ODtbp)₃. This solution was freezeꢀpumpꢀthaw de-

gassed, an atmosphere of carbon monoxide was admitted (an excess),

and the tube was sealed and allowed to stand at room temperature. NMR

spectroscopy revealed no reaction even after one week. ¹H NMR (400

MHz, 298 K, C₆D₆) δ (ppm) 16.00 (d, ³J(HꢀH) = 7.9 Hz, 6 H, meta Ar-

H), 13.32 (t, ³J(HꢀH) = 7.2 Hz, 3 H, para Ar-H), ꢀ1.51 (s, 54 H, ^tBu).

Characterization data agree with the literature.⁵²Coordinated thf reso-

nances not observed due to exchange with unbound thf.

Lack of Reaction between [U(ODipp)₃]₂and CO. A purple solution

of [U(ODipp)₃]₂(9.8 mg, 0.006 mmol) in C₆D₆(0.7 cm³) was degassed

using the freezeꢀpumpꢀthaw method, carbon monoxide (>99.9%

purity, atmospheric pressure, in excess) was admitted, and the vessel

was sealed. ¹H NMR spectroscopy revealed no change after 16 h or after

heating the sample to 60 ꢀC for 72 h. ¹H NMR (400 MHz, 298 K, C₆D₆)

δ (ppm) 10.98 (d, ³J_HH= 7.5 Hz, 6 H, meta Ar-H), 9.22 (t, ³J_HH= 7.5

Hz, 3 H, para Ar-H), 2.07 (bs, CHMe₂), ꢀ1.83 (s, CHMe₂). Char-

acterization data for 9 agree with the literature.⁵²

Lack of Reaction between [U{N(SiPhMe₂)₂}₃] and CO. In a Youngs

Teflon-valved NMR tube, a brown/yellow solution of [U{N(SiPhMe₂)₂}₃]

9047

dx.doi.org/10.1021/ja2019492 |J. Am. Chem. Soc. 2011, 133, 9036–9051

Journal of the American Chemical Society

ARTICLE

Table 5. Selected Crystallographic Details

U(ODtbp)₃1 from

U(ODtbp)₃

[(DtbpO)₃U]₂(N₂)

[(TtbpO)₃U]₂(N₂)

hexane

1 from toluene

3 from hexane

4 from hexane

4 from toluene

CCDC number

chemical formula

803905

803906

803909

803910

803911

C₄₂H₆₃O₃U

3(C₈₄H₁₂₆N₂O₆U₂)

C₁₀₈H₁₇₄N₂O₆U₂

C₆H₁₄

2(C₆H₁₄

)

3(C₇H₈)

3

formula mass

853.95

5293.96

triclinic

18.4831(2)

19.2548(3)

22.2299(3)

77.9200(10)

79.0490(10)

89.1150(10)

7592.87(18)

100(2)

2244.90

2348.95

triclinic

15.3745(6)

15.5496(7)

15.7297(9)

119.494(5)

90.110(4)

107.466(4)

3071.1(3)

171(2)

P1

crystal system

monoclinic

11.2943(2)

31.7024(6)

11.6570(2)

90.00

monoclinic

11.2779(7)

31.7231(17)

11.6467(5)

90.00

triclinic

a/Å

13.7019(4)

15.5528(4)

16.3640(5)

62.492(3)

76.539(3)

86.768(2)

b/Å

c/Å

R/deg

β/deg

105.185(2)

90.00

105.190(5)

90.00

γ/deg

unit cell volume/Å³

4028.11(13)

171(2)

P2₁

4021.3(4)

171(2)

P2₁

3002.72(15)

100(2)

P1

temperature/K

space group

P1

no. formula units/unit cell, Z

radiation type

absorption coeﬃcient, μ/mm^ꢀ1

no. of reﬂections measured

no. of independent reﬂections

R_int

4

1

Mo KR

4.063

Mo KR

4.070

Cu KR

7.896

Mo KR

2.684

9.245

45752

44614

152225

30135

57635

11845

0.0771

0.0805

0.1631

0.1087

0.1852

43366

17542

16048

14066

0.0570

0.0663

0.0452

0.1205

ﬁnal R₁values (I > 2σ(I))

ﬁnal wR(F²) values (I > 2σ(I))

ﬁnal R₁values (all data)

ﬁnal wR(F²) values (all data)

Flack parameter

0.0238

0.0466

0.0866

0.0372

0.0599

0.0576

0.2271

0.0635

0.0260

0.0698

0.1109

0.0503

0.0618

0.0670

0.2710

0.0672

0.462(4)

0.556(5)

[KU(ODtbp)₄]_n

[(DtbpO)₃UOC]₂

[(TtbpO)₃UOC]₂

(TbpO)₄U₂(μ-O)

(μ-O₂COTtbp)₂9

5 from toluene

6 from toluene

6 from benzene

7 from toluene

CCDC number

chemical formula

814203

803907

804073

803908

814204

C₅₆H₈₄KO₄U

C₈₆H₁₂₆O₈U₂

C₁₁₀H₁₇₄O₈U₂

C

₁₁₀H₁₇₄O₁₁U₂

2(C₆H₆)

3

1/2(C₇H₈)

1143.43

triclinic

11.814(2)

12.271(3)

20.781(4)

93.065(16)

99.900(16)

105.924(19)

2837.9(10)

171(2)

2(C₇H₈)

5(C₆H₆)

3(C₇H₈)

formula mass

1948.20

monoclinic

14.1939(6)

21.5647(9)

18.0645(7)

90.00

2154.47

monoclinic

14.1571(4)

27.4870 (7)

14.3715 (4)

90.00

2376.95

trigonal

23.8357(12)

18.9486(17)

90.00

2304.77

trigonal

16.3129(4)

38.4536(15)

90

crystal system

a/Å

b/Å

c/Å

R/deg

β/deg

95.800(4)

90.00

105.898 (3)

90.00

90

γ/deg

unit cell volume/Å³

120.00

9323.2(11)

171(2)

R3

120

5501.0(4)

171(2)

P2₁/n

5378.6 (3)

171(2)

P2₁/n

8862.0(5)

171(2)

P3₂21

temperature/K

space group

P1

no. formula units/unit cell, Z

2

3

radiation type

Mo KR

2.975

Mo KR

2.985

Mo KR

3.06

Mo KR

2.654

Mo KR

2.792

absorption coeﬃcient, μ/mm^ꢀ1

no. of reﬂections measured

no. of independent reﬂections

R_int

33520

51145

57034

79453

90056

10713

11241

10979

4217

11216

0.0818

0.0254

0.041

0.0452

0.0764

0.1539

0.0822

0.1572

0.0665

0.0590

0.1313

0.0662

0.1356

0.095(11)

ﬁnal R₁values (I > 2σ(I))

ﬁnal wR(F²) values (I > 2σ(I))

Final R₁values (all data)

ﬁnal wR(F²) values (all data)

Flack parameter

0.0486

0.0432

0.0408

0.0950

0.1608

0.0895

0.0724

0.0710

0.0588

0.1071

0.1666

0.0992

9048

dx.doi.org/10.1021/ja2019492 |J. Am. Chem. Soc. 2011, 133, 9036–9051

Journal of the American Chemical Society

ARTICLE

(26 mg, 0.024 mmol) in C₆D₆(0.7 cm³) was freezeꢀpumpꢀthaw

degassed, an atmosphere of carbon monoxide was admitted (an excess),

and the tube was sealed and allowed to stand at room temperature for 16 h.

NMR spectroscopy revealed no reaction even after heating to 65 ꢀC for

24 h.

Atoms-in-molecules analysis was performed using the AIMALL

program,⁸⁰version 10.12.16, using a formatted G09 checkpoint file

as input.

Crystallographic Details. Crystals of (DtbpO)₃U(N₂)U(ODtbp)₃

hexane, 3 hexane, and (TtbpO)₃U(N₂)U(OTtbp)₃hexane, 4 hexane,

3

Reaction between U{CH(SiMe₃)₂}₃and CO. In a Youngs Teflon-

valved NMR tube, a blue solution of U{CH(SiMe₃)₂}₃(7 mg, 0.01

mmol) was degassed using the freezeꢀpumpꢀthaw method, carbon

monoxide was admitted (an excess), and the tube was sealed. ¹H NMR

spectroscopy revealed disappearance of the signals due to U{CH-

(SiMe₃)₂}₃and appearance of signals identified as H₂C(SiMe₃)₂arising

from decomposition. ¹H NMR (400 MHz, 298 K, C₆D₆) δ (ppm) 0.06

(s, 18 H, SiMe₃), ꢀ0.36 (s, 2 H, H₂C).

suitable for X-ray diffraction analysis were grown from saturated hexane

solutions which had been cannula filtered away from the precipitate

produced in the reaction mixture and had been stored at room temperature

for several days. These were mounted in an inert oil and then transferred to

the cold gas stream of an Oxford diffraction four-circle Supernova

diffractometer employing Cu KR radiation (λ = 1.54178 Å). Crystals of

U(ODtbp)₃1 were grown from either a saturated hexane solution at room

temperature or a toluene solution at ꢀ30 ꢀC. Crystals of (TtbpO)₃U(N₂)-

U(OTtbp)₃toluene, 4 toluene, were grown from a saturated toluene

Reaction of U(ODtbp)₃with CO₂. Synthesis of [U(ODtbp)₄]. A purple

solution of [U(ODtbp)₃] 1 (12.7 mg, 0.015 mol) in C₆D₆(0.7 cm³) was

degassed using the freezeꢀpumpꢀthaw method, carbon dioxide

(>99.9% purity, 1 bar pressure, in excess) was admitted, and the vessel

was sealed. After 16 h, ¹H NMR spectroscopy revealed the formation of

3

solution at room temperature (surprisingly, under an atmosphere of CO).

Red crystals of K[U(ODtbp)₄]_ntoluene, 5 toluene, were grown directly

3

from the toluene reaction mixture of U(ODtbp)₃with potassium, crystals of

(DtbpO)₃UOCCOU(ODtbp)₃toluene, 6 toluene, from a saturated

3

1

[U(ODtbp)₄] 8 as the major product. H NMR (400 MHz, 298 K,

toluene solution at ꢀ30 ꢀC, crystals of (DtbpO)₃UOCCOU(ODtbp)₃

C₆D₆) δ (ppm) 10.57 (d, ³J_HH= 8.3 Hz, 8 H, meta Ar-H), 8.36 (t, ³J_HH

=

benzene, 6 benzene, from a saturated benzene solution at 20 ꢀC, and

3

8.3 Hz, 4 H, para Ar-H), ꢀ0.97 (s, 72 H, CMe₃). Characterization data

crystals of (TtbpO)₃UOCCOU(OTtbp)₃toluene 7 toluene from a satu-

3

for [U(ODtbp)₄] agree with the literature.⁵¹

rated toluene solution cooled slowly from 80 ꢀC to room temperature. Green

crystals of (TtbpO)₂U(μ-O)(μ-O₂COTtbp)₂U(OTtbp)₂benzene, 9 ben-

Synthesis of (TtbpO)₂U(μ-O)(μ-O₂COTtbp)₂U(OTtbp)₂9. A brown/

yellow solution of [(TtbpO)₃U]₂(N₂) 4 (152 mg, 0.073 mmol) in

toluene (8 cm³) was degassed using the freezeꢀpumpꢀthaw method,

carbon dioxide (>99.9% purity, 1 bar pressure, in excess) was admitted, and

the vessel was sealed and heated to 100 ꢀC briefly, before being left to stand

for 4 days, forming a pale brown solution with green solid. The solid was

isolated by filtration, washed with hexane (2 ꢁ 10 cm³), and dried under

vacuum. Yield: 45 mg, 0.021 mmol, 29%. The product is insoluble in

benzene or thf. Single crystals suitable for an X-ray diffraction experiment

were grown from an analogous reaction in benzene.

3

zene, were grown from the benzene reaction mixture of U(OTtbp)₃and

CO₂. Diffraction experiments on these samples were carried out on an Oxford

diffraction Excalibur four-circle diffractometer employing Mo KR radiation

(λ= 0.71073 Å).⁸¹The structures were solved by direct or Patterson methods

and refined by least-squares on weighted F²values for all reflections.⁸²All

hydrogen atoms were constrained to ideal geometries and refined with fixed

isotropic displacement parameters. Refinement proceeded to give the

residuals shown in Table 5. Complex neutral-atom scattering factors were

used.⁸³Crystal structure data are available in cif format from www.ccdc.ac.uk,

codes 803905ꢀ803911, 804073, 814203, and 814204.

Reaction of U{N(SiMe₃)₂}₃with CO₂. Synthesis of U(OSiMe₃)₄. To a

freeze-pump-thaw degassed solution of UN⁰⁰(0.30 g, 0.42 mmol) in

3

The molecular structure of U(ODtbp)₃1 determined from crystals

grown from both hexane and toluene solutions showed the uranium

atoms to be disordered above and below the plane formed by the oxygen

atoms. There were no systematic absences for a c-glide, and examination

of the structure revealed that P2₁is the correct space group, as the space

group P2₁/c would map the higher occupancy U position onto the low

occupancy U position of the other molecule in the asymmetric unit. The

structure of (DtbpO)₃U(N₂)U(ODtbp)₃hexane 3 hexane revealed three

toluene (10 mL) was added CO₂(1 atm) to afford a clear, pale green

solution. The volatiles were removed in vacuo and recrystallization from

toluene at ꢀ20 ꢀC afforded a pale green powder characterized as

U(OSiMe₃)₄by empirical formula. ¹H NMR (C₆D₆, 500 MHz):

38.36 (9 H, s), 29.24 (3H, s), 8.67 (9 H, s), ꢀ42.86 (6 H, s), ꢀ52.02

(9 H, s) ppm. Anal. Found (calcd for C₁₂H₃₆O₄Si₄U₁): C, 24.18

(24.23); H, 6.03 (6.10). IR (nujol mull) cm^ꢀ1(intensity) 2185 (s).

Reaction of U{N(SiPhMe₂)₂}₃with CO₂. As above, the reaction of a

benzene solution of [U{N(SiPhMe₂)₂}₃] with carbon dioxide at atmo-

spheric pressure results in the precipitation of the uranium-containing

products as a pale green solid. No further analyses were carried out.

Computational Details. Density functional theory calculations

were carried out on the CO complexes using the TPSSh hybrid

functional,⁷⁵as implemented in the Gaussian 09 Rev. A.02 (G09)⁷⁶

quantum chemistry code. For the N₂complex, both PBE⁷⁷and B3LYP⁷⁸

functionals were employed, as discussed in the main text. A (14s 13p 10d

8f)/[10s 9p 5d 4f] segmented valence basis set with StuttgartꢀBonn

variety relativistic effective core potential was used for U.⁷⁹Dunning’s cc

pVDZ basis sets were employed for the non f elements. Spin-unrest-

ricted calculations were performed on all target molecules. Wave

function stability was confirmed, and spin-contamination was verified

as minimal (all deviations from the ideal values of ÆS²æ were less than

0.001). Energies quoted are SCF energies, including zero point energies

and thermal corrections to 298 K. The ultrafine integration grid was

employed, as were the default geometry and SCF convergence criteria,

unless indicated otherwise in the Supporting Information. All stationary

point structures had no imaginary wavenumbers, unless indicated

otherwise in the Supporting Information. As the largest of these is less

than 6i cm^ꢀ1, it is assumed they arise from incompleteness in the

integration grid and do not represent genuine transition state structures.

3

molecules in the asymmetric unit. One molecule of hexane solvent was

modeled successfully, but the SQUEEZE routine in the PLATON suite of

software identiﬁed a number more in a solvent-accessible void and was dealt

with accordingly. The structure of (TtbpO)₃U(N₂)U(OTtbp)₃hexane,

3

4 hexane, has a disordered molecule of hexane which was successfully

3

modeled. The structures of (TtbpO)₃U(N₂)U(OTtbp)₃toluene, 4 to-

luene, and (TtbpO)₃UOCCOU(OTtbp)₃toluene, 7 toluene, both had

molecules of toluene disordered over inversion centers. The structure of

3

K[U(ODtbp)₄]_ntoluene, 5 toluene, has a disordered molecule of toluene

3

and disorder of two tert-butyl groups which were successfully modeled. The

structure of (DtbpO)₃UOCCOU(ODtbp)₃toluene, 6 toluene, has a

3

disordered molecule of toluene, successfully modeled, and SQUEEZE

revealed no further disordered solvate molecules even though solvent-

accessible voids were identiﬁed. The structure of (TtbpO)₂U(μ-O)-

(μ-O₂COTtbp)₂U(OTtbp)₂benzene, 9 benzene, had a disordered ben-

3

zene solvate molecule in the asymmetric unit.

’ ASSOCIATED CONTENT

S

Supporting Information. Full crystal structure and space-

b

ﬁlling drawings, further details for the complexes described, addi-

tional spectra and characterization data, additional computational

9049

dx.doi.org/10.1021/ja2019492 |J. Am. Chem. Soc. 2011, 133, 9036–9051

Journal of the American Chemical Society

ARTICLE

details and atomic coordinates, and complete ref 76. This material is

available free of charge via the Internet at http://pubs.acs.org.

(21) Evans, W. J.; Lee, D. S.; Rego, D. B.; Perotti, J. M.; Kozimor, S. A.;

Moore, E. K.; Ziller, J. W. J. Am. Chem. Soc. 2004, 126, 14574–14582.

(22) Evans, W. J.; Fang, M.; Zucchi, G.; Furche, F.; Ziller, J. W.;

Hoekstra, R. M.; Zink, J. I. J. Am. Chem. Soc. 2009, 131, 11195–11202.

(23) Fang, M.; Bates, J. E.; Lorenz, S. E.; Lee, D. S.; Rego, D. B.;

Ziller, J. W.; Furche, F.; Evans, W. J. Inorg. Chem. 2011, 50, 1459.

(24) Demir, S.; Lorenz, S. E.; Fang, M.; Furche, F.; Meyer, G.; Ziller,

J. W.; Evans, W. J. J. Am. Chem. Soc. 2010, 132, 11151–11158.

(25) (a) Spencer, L. P.; MacKay, B. A.; Patrick, B. O.; Fryzuk, M. D.

Proc. Natl. Acad. Sci. U. S. A. 2006, 103, 17094–17098. (b) Gardiner,

M. G. Materials 2010, 3, 841.

’ AUTHOR INFORMATION

Corresponding Author

polly.arnold@ed.ac.uk; n.kaltsoyannis@ucl.ac.uk

’ ACKNOWLEDGMENT

We thank Mr. S. Hunter and Prof. C. R. Pulham for the

collection of Raman data, Drs. Ian J. Casely and Zoe R. Turner for

the data for U(OSiMe₃)₄, the UK EPSRC (fellowship for P.L.A.,

funding for S.M.M.), and the University of Edinburgh for

funding. We are also grateful for computing resources via the

UCL Research Computing “Legion” cluster and “Unity” facility,

and the UK National Service for Computational Chemistry

Software (http://www.nsccs.ac.uk).

(26) Roussel, P.; Scott, P. J. Am. Chem. Soc. 1998, 120, 1070–1071.

(27) Cloke, F. G. N.; Hitchcock, P. B. J. Am. Chem. Soc. 2002,

124, 9352–9353.

(28) Evans, W. J.; Kozimor, S. A.; Ziller, J. W. J. Am. Chem. Soc. 2003,

125, 14264–14265.

(29) Odom, A. L.; Arnold, P. L.; Cummins, C. C. J. Am. Chem. Soc.

1998, 120, 5836–5837.

(30) Korobkov, I.; Gambarotta, S.; Yap, G. P. A. Angew. Chem., Int.

Ed. 2002, 41, 3433–3436.

(31) Monreal, M. J.; Diaconescu, P. L. Nat. Chem. 2010, 2, 423–423.

(32) Fox, A. R.; Bart, S. C.; Meyer, K.; Cummins, C. C. Nature 2008,

455, 341–349.

(33) (a) Bianconi, P. A.; Vrtis, R. N.; Rao, C. P.; Williams, I. D.;

Engeler, M. P.; Lippard, S. J. Organometallics 1987, 6, 1968–1977.

(b) Bianconi, P. A.; Williams, I. D.; Engeler, M. P.; Lippard, S. J.

J. Am. Chem. Soc. 1986, 108, 311–313.

(34) Vrtis, R. N.; Rao, C. P.; Bott, S. G.; Lippard, S. J. J. Am. Chem.

Soc. 1988, 110, 7564–7566.

(35) Coﬃn, V. L.; Brennen, W.; Wayland, B. B. J. Am. Chem. Soc.

1988, 110, 6063–6069.

(36) Watanabe, T.; Ishida, Y.; Matsuo, T.; Kawaguchi, H. J. Am.

Chem. Soc. 2009, 131, 3474.

’ REFERENCES

(1) Activation of Small Molecules, 1st ed.; Tolman, W. B., Ed.; Wiley:

Weinheim, 2006.

(2) Burgess, B. K.; Lowe, D. J. Chem. Rev. 1996, 96, 2983–3011.

(3) Howard, J. B.; Rees, D. C. Proc. Natl. Acad. Sci. U.S.A. 2006,

103, 17088–17093.

(4) Erisman, J. W.; Sutton, M. A.; Galloway, J.; Klimont, Z.;

Winiwarter, W. Nat. Geosci. 2008, 1, 636–639.

(5) Zheng, C.; Apeloig, Y.; Hoﬀmann, R. J. Am. Chem. Soc. 1988,

110, 749–774.

(6) Hoﬀman, B. M.; Dean, D. R.; Seefeldt, L. C. Acc. Chem. Res. 2009,

42, 609–619.

(37) (a) Ellis, J. E.; Fjare, K. L.; Hayes, T. G. J. Am. Chem. Soc. 1981,

103, 6100–6106. (b) Weiss, E.; Buchner, W. Helv. Chim. Acta 1963,

46, 1121. (c) Weiss, E.; Buchner, W. Z. Anorg. Allg. Chem. 1964,

330, 251–258. (d) Buchner, W.; Weiss, E. Helv. Chim. Acta 1964,

47, 1415.

(38) Evans, W. J.; Lee, D. S.; Ziller, J. W.; Kaltsoyannis, N. J. Am.

Chem. Soc. 2006, 128, 14176–14184.

(39) Brennan, J. G.; Andersen, R. A.; Robbins, J. L. J. Am. Chem. Soc.

1986, 108, 335–336.

(7) (a) Howard, J. B.; Rees, D. C. Chem. Rev. 1996, 96, 2965–2982.

(b) Schrock, R. R. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 17087–17087.

(8) Schrock, R. R. Nat. Chem. 2011, 3, 95.

(9) Khodakov, A. Y.; Chu, W.; Fongarland, P. Chem. Rev. 2007,

107, 1692–1744.

(10) Herrmann, W. A. Angew. Chem., Int. Ed. Engl. 1982, 21, 117–

130.

(11) Miller, A. J. M.; Labinger, J. A.; Bercaw, J. E. J. Am. Chem. Soc.

2008, 130, 11874–11875.

(40) (a) Conejo, M. D.; Parry, J. S.; Carmona, E.; Schultz, M.;

Brennan, J. G.; Beshouri, S. M.; Andersen, R. A.; Rogers, R. D.; Coles, S.;

Hursthouse, M. Chem.—Eur. J. 1999, 5, 3000–3009. (b) Parry, J.;

Carmona, E.; Coles, S.; Hursthouse, M. J. Am. Chem. Soc. 1995,

117, 2649–2650. (c) Maron, L.; Eisenstein, O.; Andersen, R. A. Orga-

nometallics 2009, 28, 3629–3635.

(41) Evans, W. J.; Kozimor, S. A.; Nyce, G. W.; Ziller, J. W. J. Am.

Chem. Soc. 2003, 125, 13831–13835.

(42) Castro-Rodriguez, I.; Meyer, K. J. Am. Chem. Soc. 2005,

127, 11242–11243.

(12) (a) Ballmann, J.; Munhꢀa, R. F.; Fryzuk, M. D. Chem. Commun.

2010, 46, 1013–1025. (b) Werkema, E. L.; Maron, L.; Eisenstein, O.;

Andersen, R. A. J. Am. Chem. Soc. 2007, 129, 2529–2541.

(13) Fryzuk, M. D.; Love, J. B.; Rettig, S. J.; Young, V. G. Science

1997, 275, 1445–1447.

(14) Pool, J. A.; Lobkovsky, E.; Chirik, P. J. Nature 2004, 427, 527–

530.

(15) (a) Knobloch, D. J.; Lobkovsky, E.; Chirik, P. J. Nat. Chem.

2010, 2, 30–35. (b) Knobloch, D. J.; Lobkovsky, E.; Chirik, P. J. J. Am.

Chem. Soc. 2010, 132, 10553–10564.

(16) (a) Yandulov, D. V.; Schrock, R. R. Science 2003, 301, 76–78.

(b) Arashiba, K.; Miyake, Y.; Nishibayashi, Y. Nat. Chem. 2011, 3, 120.

(17) Evans, W. J.; Ulibarri, T. A.; Ziller, J. W. J. Am. Chem. Soc. 1988,

110, 6877–6879.

(18) Evans, W. J.; Allen, N. T.; Ziller, J. W. J. Am. Chem. Soc. 2001,

123, 7927–7928.

(19) (a) Evans, W. J.; Allen, N. T.; Ziller, J. W. Angew. Chem., Int. Ed.

2002, 41, 359. (b) Evans, W. J.; Zucchi, G.; Ziller, J. W. J. Am. Chem. Soc.

2003, 125, 10–11.

(20) (a) Evans, W. J.; Lee, D. S.; Johnston, M. A.; Ziller, J. W.

Organometallics 2005, 24, 6393–6397. (b) Evans, W. J.; Lee, D. S.; Lie,

C.; Ziller, J. W. Angew. Chem., Int. Ed. 2004, 43, 5517–5519. (c) Evans,

W. J.; Lee, D. S.; Ziller, J. W. J. Am. Chem. Soc. 2004, 126, 454–455.

(d) Evans, W. J.; Rego, D. B.; Ziller, J. W. Inorg. Chem. 2006,

45, 10790–10798. (e) Schmiege, B. M.; Ziller, J. W.; Evans, W. J. Inorg.

Chem. 2010, 49, 10506–10511.

(43) Summerscales, O. T.; Cloke, F. G. N.; Hitchcock, P. B.; Green,

J. C.; Hazari, N. Science 2006, 311, 829–831.

(44) Summerscales, O. T.; Cloke, F. G. N.; Hitchcock, P. B.; Green,

J. C.; Hazari, N. J. Am. Chem. Soc. 2006, 128, 9602–9603.

(45) Frey, A. S.; Cloke, F. G. N.; Hitchcock, P. B.; Day, I. J.; Green,

J. C.; Aitken, G. J. Am. Chem. Soc. 2008, 130, 13816–13817.

(46) Arnold, P. L.; Turner, Z. R.; Bellabarba, R. M.; Tooze, R. P.

Chem. Sci. 2011, 2, 77–79.

(47) Berthet, J. C.; Lemarechal, J. F.; Nierlich, M.; Lance, M.; Vigner,

J.; Ephritikhine, M. J. Organomet. Chem. 1991, 408, 335–341.

(48) Castro-Rodriguez, I.; Meyer, K. Chem. Commun. 2006, 1353–1368.

(49) Castro-Rodriguez, I.; Nakai, H.; Zakharov, L. N.; Rheingold,

A. L.; Meyer, K. Science 2004, 305, 1757–1759.

(50) (a) Van der Sluys, W. G.; Sattelberger, A. P. Chem. Rev. 1990,

90, 1027–1040. (b) Van der Sluys, W. G.; Sattelberger, A. P. Inorg. Chem.

1989, 28, 2496–2498.

9050

dx.doi.org/10.1021/ja2019492 |J. Am. Chem. Soc. 2011, 133, 9036–9051

Journal of the American Chemical Society

ARTICLE

(51) Berg, J. M.; Clark, D. L.; Huﬀman, J. C.; Morris, D. E.;

Sattelberger, A. P.; Streib, W. E.; Vandersluys, W. G.; Watkin, J. G.

J. Am. Chem. Soc. 1992, 114, 10811–10821.

(52) Van der Sluys, W. G.; Burns, C. J.; Huﬀman, J. C.; Sattelberger,

A. P. J. Am. Chem. Soc. 1988, 110, 5924–5925.

(53) Stewart, J. L.; Andersen, R. A. Polyhedron 1998, 17, 953–958.

(54) Van der Sluys, W. G.; Burns, C. J.; Sattelberger, A. P. Organo-

metallics 1989, 8, 855–857.

(55) Mansell, S. M.; Perandones, B. F.; Arnold, P. L. J. Organomet.

Chem. 2010, 695, 2814–2821.

(56) Fryzuk, M. D.; Johnson, S. A. Coord. Chem. Rev. 2000,

200, 379–409.

(57) Search of the CSD Version 5.31, update 1.

(58) Kaltsoyannis, N.; Scott, P. Chem. Commun. 1998, 1665–1666.

(59) Roussel, P.; Errington, W.; Kaltsoyannis, N.; Scott, P.

J. Organomet. Chem. 2001, 635, 69–74.

(60) Kaltsoyannis, N.; Scott, P. Chem. Commun. 1998, 1665.

(61) Cloke, F. G. N.; Green, J. C.; Kaltsoyannis, N. Organometallics

2004, 23, 832–835.

(62) Bader, F. Atoms in Molecules: A Quantum Theory; OUP: Oxford,

1990.

(63) (a) Tassell, M. J.; Kaltsoyannis, N. Dalton Trans. 2010, 39, 6719–

6725. (b) Arnold, P. L.; Turner, Z. R.; Kaltsoyannis, N.; Pelekanaki, P.;

Bellabarba, R. M.; Tooze, R. B. Chem.—Eur. J. 2010, 16, 9623–9629.

(c) Kirker, I.; Kaltsoyannis, N. Dalton Trans. 2011, 40, 124.

(64) Saleh, L.; Birjkumar, K.; Protchenko, A.; Schwartz, A.; Aldridge,

S.; Jones, C.; Kaltsoyannis, N.; Mountford, P. J. Am. Chem. Soc. 2011,

133 (11), 3836–3839 .

(65) Matta, C. F.; Boyd, R. J. In The quantum theory of atoms in

molecules; Matta, C. F., Boyd, R. J., Eds.; Wiley-VCH: Weinheim, 2007,

pp 1ꢀ34.

(66) (a) Gleiter, R.; Werz, D. B.; Rominger, F.; Zhutov, E.; Zeﬁrov,

N. S.; Proskurnina, M. V. Eur. J. Org. Chem. 2007, 5834–5839.

(b) Gleiter, R.; Werz, D. B. Chem. Rev. 2010, 110, 4447–4488.

(67) Castro-Rodriguez, I.; Nakai, H.; Zakharov, L. N.; Rheingold,

A. L.; Meyer, K. Science 2004, 305, 1757–1759.

(68) Evans, W. J.; Lorenz, S. E.; Ziller, J. W. Inorg. Chem. 2009,

48, 2001–2009.

(69) Summerscales, O. T.; Frey, A. S. P.; Cloke, F. G. N.; Hitchcock,

P. B. Chem. Commun. 2009, 198.

(70) Phull, H.; Alberti, D.; Korobkov, I.; Gambarotta, S.; Budzelaar,

P. H. M. Angew. Chem., Int. Ed. 2006, 45, 5331–5334.

(71) (a) Monreal, M. J.; Diaconescu, P. L. Organometallics 2008,

27, 1702–1706. (b) Ephritikhine, M. Dalton Trans. 2006, 2501–2516.

(c) Lam, O. P.; Anthon, C.; Meyer, K. Dalton Trans. 2009, 9677–9691.

(72) Evans, W. J.; Kozimor, S. A. Coord. Chem. Rev. 2006, 250, 911–935.

(73) Van der Sluys, W. G.; Burns, C. J.; Sattelberger, A. P. Organo-

metallics 1989, 8, 855–857.

(74) Clark, D. L.; Sattelberger, A. P.; Andersen, R. A. Inorg. Synth.

1997, 31, 307.

(75) Tao, J. M.; Perdew, J. P.; Staroverov, V. N.; Scuseria, G. E. Phys.

Rev. Lett. 2003, 91, 14.

(76) Frisch, M. J.;Gaussian 09, Revision A.2; Gaussian, Inc.: Wallingford,

CT, 2009.

(77) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1996,

77, 3865.

(78) Becke, A. Phys. Rev. A 1988, 38, 3098.

(79) Cao, X. Y.; Dolg, M. J. Mol. Struct. (THEOCHEM) 2004,

673, 203.

(80) Keith, T. 2011, AIMALL version 10.12.16, http://aim.tkgristmill.

com/.

(81) Agilent Oxford Diﬀraction Ltd., Yarnton, England, 2009.

(82) Sheldrick, G. M. Acta Crystallogr., Sect. A: Found. Crystallogr.

2008, 64, 112–122.

(83) International Tables for Crystallography; Kluwer: Dordrecht,

1992; Vol. C.

9051

dx.doi.org/10.1021/ja2019492 |J. Am. Chem. Soc. 2011, 133, 9036–9051

Article Doi

Small molecule activation by uranium tris(aryloxides): Experimental and computational studies of binding of N2, coupling of CO, and deoxygenation insertion of CO2 under ambient conditions

DOI: 10.1021/ja2019492

Source and publish data:

Authors:

Article abstract of DOI:10.1021/ja2019492

Full text of DOI:10.1021/ja2019492

Products guided by the article

R&D Labs maybe for 1307903-81-6

Relevant to this article

Hot Product