Controlling uranyl oxo group interactions to group 14 elements using polypyrrolic Schiff-base macrocyclic ligands

Heterodinuclear uranyl/group 14 complexes of the aryl- and anthracenyl-linked Schi ﬀ -base macrocyclic ligands L Me and L A were synthesised by reaction of UO 2 (H 2 L) with M{N(SiMe 3 ) 2 } 2 (M = Ge, Sn, Pb). For complexes of the anthracenyl-linked ligand (L A ) the group 14 metal sits out of the N 4 -donor plane by up to 0.7 Å resulting in relatively short M ⋯ OUO distances which decrease down the group; however, the solid state structures and IR spectroscopic analyses suggest little interaction occurs between the oxo and group 14 metal. In contrast, the smaller aryl-linked ligand (L Me ) enforces greater interaction between the metals; only the Pb II complex was cleanly accessible although this complex was relatively unstable in the presence of HN(SiMe 3 ) 2 and some organic oxidants. In this case, the equatorial coordination of pyridine-N -oxide causes a 0.08 Å elongation of the endo UO bond and a clear interaction of the uranyl ion with the Pb( II ) cation in the second donor compartment.


Introduction
Uranyl [U VI O 2 ] 2+ is the most stable and prevalent form of uranium in the environment. 1,2 The redox properties of uranyl are of interest in order to establish chemical routes to separate and immobilise actinide radioactive wastes. 3 Reduced [U V O 2 ] + species are inherently unstable to disproportionation under aqueous conditions forming [U VI ] and [U IV ] products, with the latter being insoluble in aqueous waste streams and therefore immobilised by this process. 4 Oxidised U VI O 3 is also poorly soluble in water and T-shaped U-trioxo species have recently been proposed, alongside cisoid-UO 2 complexes, to be important in oxo-transfer processes in aqueous media. 5 As such, the synthesis of these motifs in molecular species bound within defined ligand environments can improve our understanding of actinide bonding and help us to predict uranyl speciation.
We have studied extensively the reduction of uranyl incorporated within a Schiff-base macrocyclic ligand with two aryl-linked coordination pockets (Scheme 1(a)). 2,[6][7][8] Incorporation of two uranyl units into this ligand environment results in reduction to U V and oxo group rearrangement to form the homodinuclear 'butterfly' complex (Me 3 SiOU V ) 2 {µ-(O) 2 }L Me which contains bridging oxo groups and acute OUO angles. 2 We have seen similar reduction during the formation of mixed-metal complexes from U VI O 2 (H 2 L Me ). Coordination of an electropositive metal within the vacant compartment of the macrocycle was shown to activate the uranyl towards reduction so forming a range of stable mixed-metal, [U V O 2 ] + com-plexes. 7,8 In contrast, incorporation of a less reducing metal (e.g. Fe(II), Mn(II)) did not promote reduction and the corresponding uranyl(VI) U VI O 2 M(L Me ) complexes were isolated. 9 In this ligand, the proximity of the uranyl endo-oxygen to the lower macrocyclic pocket facilitates interaction with the second metal. More recently we have developed the chemistry of an anthracenyl-linked analogue of this ligand and synthesised mono-and dinuclear uranyl complexes (Scheme 1(b)). 10,11 The anthracenyl linker enforces both a greater distance between the N 4 -coordination compartments and a greater degree of coplanarity.
We have now explored and compared the reactivity of the uranyl complexes of these two ligands, UO 2 (H 2 L) (L = L Me , L A ) towards group 14 metal silylamides (M{N(SiMe 3 ) 2 } 2 , M = Ge, Sn, Pb). These latter metals have been used in order to target macrocyclic complexes of U VI O 3 , with the chalcophilic group 14 metal potentially charge balancing through oxidation from M II to M IV . The interaction between the two metal ions in the resulting complexes and any resulting activation of the UO 2 bonds has been assessed spectroscopically, as has the reactivity of these complexes towards oxo-transfer reagents.

Results and discussion
Complexes of the anthracenyl-hinged macrocycle L A The reaction of a dark green solution of UO 2 (H 2 L A ) in THF or pyridine with M{N(SiMe 3 ) 2 } 2 (M = Sn, Pb) results in the immediate formation of deep red solutions and the appearance of new sets of resonances in the 1 H NMR spectra which are consistent with the formation of 1(Pb) and 1(Sn) (Scheme 2). In contrast, reaction of UO 2 (H 2 L A ) with Ge{N(SiMe 3 ) 2 } 2 in pyridine takes 6 h at room temperature to reach full conversion, forming 1(Ge)-py, and in THF elevated temperatures for ca. 8 h are required to form 1(Ge)-THF cleanly. Removal of solvent and workup yielded 1(M)-sol for M = Sn, Ge, and Pb.
X-ray quality crystals of the THF solvate of each complex were grown by either slow cooling concentrated THF/C 6 D 6 solutions (1(Pb)-THF) or diffusion of hexanes into THF (1(Ge)-THF and 1(Sn)-THF) ( Fig. 1(a): 1(Pb)-THF; Fig. S1: † 1(Ge)-THF and 1(Sn)-THF). While the three structures are very similar there are some subtle differences (Table 1). Intriguingly all of the group 14 metal ions sit above the N 4 -donor plane within the molecular cleft, so moving closer to the uranyl endooxygen. We have previously demonstrated that the flexibility of the N 4 -donor pocket in L Me allows the coordinated metal to sit either above or below the plane (by up to ca. 0.6 Å) to accommodate the steric requirements of various co-ligands. 12 The structures show that descending group 14 causes the metal to move further out of the plane and closer to the endo-oxygen. This promotes a Pb⋯OUO distance of as little as 3.06 Å, well within the range for bridging M-O-Pb bonds of 2.22-3.33 Å, although slightly longer than the MvO⋯Pb interactions that have been characterised (all ca. 2.6 Å). 13 However, the question of whether the metal interacts with, and induces any weakening of the UO 2 bonding is less clear. The X-ray crystal structures show only a very slight lengthening of the UvO endo bond on descending the group (ca. 0.01 Å) and no significant change in the UvO exo bonds. Similarly, the X-ray crystal structure of the uranyl(VI) complex UO 2 (LiHL Me ) (1.794(3) Å) 7 also does not show significant lengthening of the UvO bonds relative to its precursor UO 2 (H 2 L Me ) (1.790(4) Å) 14 despite there being a significant interaction between the lithium and the endo-oxygen as evidenced in the solid state IR spectra of these compounds (899 vs. 908 cm −1 respectively). The solid state IR spectra (Charts S1 and S3 †) of the three THF adducts of 1(Sn), 1(Ge) and 1(Pb) are consistent with a very slight lengthening of the UO 2 bond down the group; in contrast, the solution state IR spectra (Chart S6 †) are incon-clusive with multiple bands appearing in the [U VI O 2 ] region (ca. 890-930 cm −1 ). These data may suggest the proximity of the group 14 metal to the uranyl oxo group is simply a product of crystal packing effects with the metal moving out of the N 4donor plane due to the increasing ionic radius of the metal down the group.
The 119 Sn NMR spectrum of 1(Sn)-THF shows a resonance at −459 ppm which is deshielded from that of Sn 2 (L A ) at −527 ppm; however, with the large anthracenyl-linked molecular cleft we have previously seen endo-solvent coordination in homodinuclear complexes which may shield the metal. 15 Endo-coordination of solvent would not be possible in 1(Sn)-THF due to the presence of the linear UO 2 group meaning the tin cation would appear relatively deshielded in the mixed metal complex.
Reaction of 1(Pb) with one equivalent of pyridine N-oxide yields a mixture of compounds in both THF and pyridine solvent which does not change upon heating. However, addition of an excess of pyridine N-oxide (>8 eq.) allowed conversion to one new product by 1 H NMR spectroscopy which was assigned as 2(Pb). No free pyridine was observed in the 1 H NMR spectrum suggesting no oxidation of the Pb II centre has occurred and instead a broadened doublet of doublets at 6.26 ppm indicates coordination of one pyridine N-oxide molecule.
Single crystals of 2(Pb) were grown by slow diffusion of hexane into a pyridine or THF solution containing excess pyridine N-oxide. The refined structures ( Fig. 1 Despite the lower oxidation potential of Sn II (−0.13 V cf. Pb II −1.8 V) 16 the reaction of 1(Sn) with pyridine N-oxide similarly results in formation of the uranyl(VI) adduct 2(Sn) which was found to be in equilibrium with solvent coordinated 1(Sn) in the 1 H NMR spectrum. The 119 Sn NMR spectrum of 2(Sn) shows a single resonance at −480 ppm which is very slightly shielded relative to 1(Sn).
Ge II is a reducing oxidation state compared with Sn II and Pb II but treatment of 1(Ge) with pyridine N-oxide yields a mixture of compounds in the 1 H NMR spectrum. This may suggest that the relatively electropositive Ge centre may coordinate pyridine-N-oxide in a similar manner to uranium giving a competitive equilibrium between the two metals or even one that displaces a Ge II cation from the macrocycle.   Crystals suitable for X-ray diffraction of both THF and pyridine solvates of 3(Pb) were grown by diffusion of hexane into pyridine or THF solutions (Fig. 1(c) and (d) respectively). The solid state structures of these adducts are different to each other with the Pb cation coordinated by a solvent molecule in 3(Pb)-py but unsolvated in 3(Pb)-THF. This results in a significantly longer Pb-N py bond (2.58(1) Å) compared to 2(Pb)-py (2.480(5) Å). It is interesting to note that the Pb⋯OUO distance in 3-py of 2.9 Å is very close to those in 1(Pb) and 2(Pb) demonstrating the ability of the macrocyclic ligand pockets to distort to optimise metal geometry and separation.
In 3(Pb)-THF, a short Pb⋯O bond is present and instead of THF coordinating to the Pb cation, an η 5 -interaction with the pyrrole π-system of an adjacent molecule exists ( Fig. 1 and S3 †). Both structures show bonding with the uranyl endo oxygen although the THF adduct contains a slightly lengthened UvO endo bond (1.817(9) Å) relative to the pyridine adduct (1.77(1) Å). The solid state IR spectra (Charts S8 and S9 †) show a UO 2 asymmetric stretch for 3(Pb)-py at 907 cm −1 while that for 3(Pb)-THF is seen at 896 cm −1 . The former is comparable with that for ( py)UO 2 (H 2 L Me ) (908 cm −1 ) while the latter is closer to that for (THF)UO 2 Li(HL Me ) (899 cm −1 ) which contains a strong UvO⋯Li interaction, known to affect the reduction chemistry of the resulting complex.
The treatment of 3(Pb) with excess pyridine N-oxide resulted in the formation of the py-O adduct 4(Pb)-py. Single crystals were grown by vapour diffusion of hexane into a concentrated pyridine solution containing excess pyridine N-oxide. The molecular structure (Fig. 2) shows the expected coordination of pyridine N-oxide to the uranium centre while the Pb centre is solvated by pyridine.  17 (1.821(6) Å). This demonstrates that activation of the linear UO 2 bonds is facilitated by the presence of strong oxo-donors in the equatorial plane as well as by Lewis acidic metal cations coordinating to the axial uranyl oxygen atoms. The elongation of this bond by 0.08 Å is less than seen for other uranyl U VI complexes in which a strong Lewis acid binds to an oxo group and may reflect the relatively weak Lewis acidity of Pb II ; the activated UO bond (1.890(4) Å) in OU(OB{C 6 F 5 } 3 )( Ar acnac) 2 (Ar = 3,5-tBu 2 C 6 H 3 ) is significantly longer (0.14 Å) than for UO 2 ( Ar acnac) 2 (1.755(5) Å), 18 and in OU(OB{C 6 F 5 } 3 )(NCN) 2 (NCN = {Me 3 SiN}CPh{NSiMe 3 }) at 1.898(3) Å the U-OB bond is elongated (0.15 Å) relative to 1.750(4) Å in UO 2 (NCN) 2 . 19 The IR spectrum of 4(Pb)-py (Chart S10 †) shows a UO 2 asymmetric stretch at 893 cm −1 , significantly shorter than for 3(Pb)-py (908 cm −1 ) and comparable to LiOU V OLi(HL Me ) (893 cm −1 ) and some other reduced U V O 2 complexes such as ({(py)R 2 AlOU V OH 2 L Me } and {(py) 3 MOU V OH 2 L Me }; R = Me, i-Bu, M = Li, Na, K) 20 which all appear within the region 891-894 cm −1 .
Attempts to oxidise the Pb II centre with alternative oxidising agents, e.g. o-iodylanisole, p-iodosotoluene, bis-trimethylsilyperoxide, and trimethylamine-N-oxide resulted in the formation of UO 2 (H 2 L Me ) and minor degradation products including a highly symmetrical set of resonances which were assigned as the dinuclear Pb complex 5 (eqn (1)).
The reasons behind the degradation of 3(Pb) upon reaction with strong oxygen-atom oxidants are unclear; however, Pb II has a relatively large ionic radius (1.2 Å) 16 and therefore may be unstable within the cleft of this smaller ligand. We have previously shown that complexes of larger metal cations such as Ca II adopt an alternative, bowl-shaped coordination mode with L R (R = Me, Et). 15 As such, 5 was synthesised directly by treatment of H 4 L Me with two equivalents of Pb{N(SiMe 3 ) 2 } 2 in pyridine (eqn (1)).
The solid state structure of 5 (Fig. 3) shows the expected 'bowl' coordination geometry in which the macrocycle folds at the meso-carbon instead of the aryl linker. This provides a larger cleft which can accommodate the two Pb centres without the significant distortion out of the N 4 -plane seen in the solid state structure of the other, Pacman-shaped Pb complexes in this work. The N pyrrole ⋯N pyrrole distance of 5.25 Å is only slightly larger than the equivalent N imine ⋯N imine distance for 3(Pb)-py of 5.15 Å and is smaller than that in 1(Pb)-THF (5.27 Å); however, the larger N pyrrole -Pb-N pyrrole angle for 5 (152°cf. N imine -Pb-N imine angle of 130°for 3(Pb)-py and 138°f or 1(Pb)-THF) allows more room for the metal to sit within the N 4 -donor plane and as a result the Pb atoms in 5 sit only 0.3 Å out of their respective N 4 -pockets. In addition, π-interactions between the lead centre and the aryl ring of the adjacent ligand or the pyridine solvent (see ESI Fig. S4 †) add stability to this structure.
The reactions of UO 2 (H 2 L Me ) in THF or pyridine with various equivalents of Sn{N(SiMe 3 ) 2 } 2 at different temperatures yield a complex mixture of diamagnetic and paramagnetic species as evidenced by 1 H NMR spectroscopy. While slight activation of the UO 2 2+ unit is suggested by the IR spectrum of 3(Pb)-THF, it is likely that significantly more interaction occurs with Sn due to its increased Lewis acidity compared to Pb, resulting in the activation of UO 2 2+ towards reduction by a second molecule of Sn{N(SiMe 3 ) 2 } 2 . We have studied extensively the reactions of UO 2 (H 2 L Me ) with electropositive metal amides (e.g. LDA, Ln{N(SiMe 3 ) 2 } 3 , and Mg{N(SiMe 3 ) 2 } 2 ) and have shown that coordination of these metals within the vacant compartment of the macrocycle results in strong interactions with the endo-oxo of the uranyl and activates this group towards reduction to U V . In contrast, 'softer' polarisable metal amides (e.g. Fe{N(SiMe 3 ) 2 } 2 , Mn{N(SiMe 3 ) 2 } 2 ) yield diamagnetic transamination products. In contrast to the above reactions with Sn II , no reaction between UO 2 (H 2 L Me ) and Ge{N(SiMe 3 ) 2 } 2 occurs at ambient temperatures in either THF or pyridine. Upon heating to 80°C a number of minor paramagnetic resonances appear suggesting that reduction of the uranyl centre occurs, whilst heating to 120°C in pyridine results in the formation of the previously reported oxo-silylated complex [(Me 3 Si)OUO(H 2 L Me )] after 24 h as the major product along with multiple minor paramagnetic species. This may indicate that coordination of Ge{N(SiMe 3 ) 2 } 2 to the exogenous uranyl oxo occurs, resulting in reductive silylation as seen with ZnCl{N(SiMe 3 ) 2 }. 8

Conclusions
Mixed-metal uranyl/group 14 complexes of two different Schiff base macrocyclic ligands have been synthesised by transamination from UO 2 (H 2 L) (L = L Me and L A ). With the larger anthracenyl-linked macrocycle Ge II , Sn II and Pb II all coordinate in the vacant macrocyclic pocket without reduction of the uranyl. In these complexes the metal ion sits above the N 4 -donor plane, closer to the uranyl endo-oxygen suggesting a possible bonding interaction. However, it is clear from IR spectroscopy and solid-state structural analysis that only minor activation of the uranyl oxo groups occurs, and so this U-O-M interaction is likely a result of crystal packing forces. In contrast, similar reactions between group 14 silylamides and the uranyl complex of a smaller aryl-linked macrocycle results in reduction to multiple paramagnetic species for Sn and Ge, likely due to the closer interaction between the metal and UvO endo enforced by the ligand framework. Only using Pb were we able to isolate the expected mixed-metal uranyl-Pb complex which showed similarly minor activation of the uranyl bonding in the solid state. These data suggest that the proximity of the second metal to uranyl is not an overriding factor in the reduction of uranyl(VI) to uranyl(V) and in the formation of uranyl(VI) oxo-metal bonds. However, the exchange of the equatorial ligand from THF or pyridine in 3(Pb) with pyridine N-oxide to form 4(Pb)-py results in significant elongation of the endo-oxo group and the formation of a clear OUO-Pb bonding interaction, similar to those seen in simple uranyl-perfluoroborane Lewis acid-base adducts. This is only the case when using the more constrained macrocycle environment provided by L Me , and supports the premise that the formation of uranyl (VI)-oxo Lewis acid-base interactions requires a ligand environment (i.e. the Pacman macrocycle) that not only defines the approach of the Lewis acid but also the equatorial coordination sphere of the uranyl, in this case the weak-field macrocycle N 4 -donor set plus the strongly donating pyridine oxide.

General details
All manipulations were carried out under a dry, oxygen-free dinitrogen atmosphere using standard Schlenk techniques or in a glove box unless otherwise stated. Solvents (toluene, n-hexane, diethyl ether and tetrahydrofuran (THF)) were dried by passage through activated 4 Å molecular sieves or activated alumina towers and stored over activated 4 Å molecular sieves. Pyridine was distilled from potassium under dinitrogen in a solvent still prior to use. Deuterated solvents were refluxed over potassium, freeze-pump-thaw degassed three times and vacuum transferred prior to use. 1 H NMR spectra were recorded at 298 K unless otherwise stated on either a Bruker AVA400 spectrometer at 399.90 MHz, or AVA500 spectrometer at 500.12 MHz. 13  1(Ge)-THF. To a solution of UO 2 (H 2 L A ) (500 mg, 0.42 mmol) in THF (10 mL) was added a solution of Ge{N(SiMe 3 ) 2 } 2 (180 mg, 0.46 mmol) in THF (10 mL) and the solution was stirred for 8 h at 80°C before the solvent was concentrated by half and hexane added, resulting in the precipitation of microcrystalline 1(Ge)-THF (340 mg, 0.25 mmol, 61%). Deep red X-ray quality crystals were grown by slow diffusion of hexane into a concentrated solution of 1(Ge)-THF in THF. 1