as a dark red crystalline solid in 3 % yield.

Scheme 4: Synthesis of [(U(8-C8H6{SiiPr3-1,4}2)(5-CpMe4H))2(-2:2-13C4O4)] (2.5)

The reaction between 2.1 and an overpressure of CO is reported to produce [(U(8- C8H6{SiiPr3-1,4}2)(5-CpMe4H))2(-2:2-C4O4)] in a 66 % isolated yield);15b there is no

recorded yield for 2.5 that pre-dates this work. The extremely low yield of 2.5 seen here, is due in part to the difficulty of isolating crystalline material. Complex 2.5 is sparingly soluble in Et2O and very soluble in THF, but it was found that an 8:1 ratio of

Et2O to THF was effective. It is quite possible the yield of 2.5 is a result of the lower

reaction pressure. There is a significant pressure disparity between the larger scale reaction at ca. 1 atm and the smaller scale reaction with labelled 13CO. It is not economic to ‘waste’ 13

CO in doing these reactions, as the squarate dianion is the product of the 2 molecules of CO per uranium(III) centre. Doubling the amount of 13CO used (in mmHg) did not improve the yield of 2.5.

It is significant that the maximum pressure the Toepler pump can pressurise to is 0.85 atm. The reaction of [U(8-C8H6{SiiPr3-1,4}2)(5-CpMe5)] with a sub-stoichiometric 0.9

CpMe5)]2(-1:1-13C2O2)], also yields small amounts of the deltate complex38 and

indeed the analogous reaction to form the labelled deltate complex also yields the ynediolate complex.39 The relationship between the specific pressure of the low- pressure system and the yields of the various carbocycles has not been fully rationalised. Pressure of the reaction vessel was seen to be important in the isolation of the coupled CO product in the reaction of [Sm(5-CpMe5)2(THF)2] with CO.40

The 13C{1H} NMR spectrum of 2.5 in d8-thf room temperature exhibits the

characteristically broad singlet assigned to the enriched squarate fragment at  –106.2 ppm and is shown in

Figure 5. The 1H NMR spectrum of 2.5 displays no evidence for restricted rotation on the NMR timescale and the CpMe4H-CH resonance cannot be identified (see Section 2.1.5). As

Figure 5 shows, labelling of the C4O42- unit enables the observation of the 13C4O42- unit

in 2.5 by 13C{1H} NMR in a very few scans and therefore provides a very useful spectroscopic handle for the reactivity of 2.5 with organic halides and pseudohalides.

Figure 5: 13C{1H} NMR spectrum of 2.5, 11 mgs in d8-thf, 100.45

MHz, 30 ºC, 64 scans, lb = 5.

2.2.3 Reactivity of [(U(8-C8H6{SiiPr3-1,4}2)(5-CpMe4H))2(-2:2-13C4O4)] (2.5)

with SiR3X (R = Me, Ph, X = I, Cl, OTf):

The complex [(U(8-C8H6{SiiPr3-1,4}2)(5-CpMe5))2(-1:2-13C3O3)] was shown to

react with SiMe3Cl to give [(U(8-C8H6{SiiPr3-1,4}2)(5-CpMe5)(Cl)] by 1H NMR but

the functionalised 13C3O3 unit was not detected by 13C{1H} NMR. The reaction of

[(U(8-C8H6{SiiPr3-1,4}2)(5-CpMe4H))2(-2:2-13C4O4)] (2.5) in d8-thf with 2.5

equivalents of SiMe3Cl at room temperature resulted in the complete disappearance of

by 13C{1H} NMR at  = 189.1, 189.7, 190.4, 195.7, 196.3 and 197.0 ppm and the observation of the chloride complex [(U(8-C8H6{SiiPr3-1,4}2)(5-CpMe4H)(Cl)] by 1H

NMR.17 No fully 13C-labelled squarate derivatives have been synthesised other than that reported by Summerscales et al15,17 however a range of substituted squarate derivatives have been synthesised and their 13C{1H} NMR shifts recorded.41 The 13C{1H} NMR spectrum of C4O2(OMe)2 displays two resonances at  188.9 and 184.3 ppm.42 In the 13

C NMR spectrum the resonance at  184.3 ppm is split into a quartet JCH = 4 Hz and

this was assigned to the -carbon. In C4O2(OSiMe3)2 no coupling was observed by

NMR between the ring carbons and the SiMe3.43 In all examples of substituted squarate

derivatives the carbonyl resonance is found upfield from that of the -carbon and the difference in chemical shift between the two resonances depends on the substituent groups on the oxygen atoms bonded to the -carbon.

The NMR scale reaction between 2.5 and 2.5 equivalents of SiMe3I (stored at -20 ºC) in

d6-benzene at room temperature resulted in a darkening of the solution from orange to

red on addition and red solids were visible. The 13C{1H} NMR spectrum was run within 10 min of addition and showed labelled broad multiplets at  194.0 and 188.1 ppm in the region expected for the 13C4O2(OSiMe3)2 product.17,41,43 When complex 2.5 was

reacted with 1 equivalent of SiMe3OTf, in an analogous manner to the reaction of 2.5

with SiMe3I, the physical observations were the same and the 13C{1H} NMR spectrum

showed more intense labelled muliplets for the 13C4O2(OSiMe3)2 product at the same

chemical shifts. The reactions of 2.5 with SiMe3I and SiMe3OTf were undertaken in d6-

benzene, to avoid side reactions with a more coordinating solvent. As 2.5 is sparingly soluble in d6-benzene and solids were present, the samples were placed in a NMR

heating block set at 70 ºC. The 13C{1H} NMR spectra after heating each displayed a small amount of 2.5 but neither resonances in the product region nor other enriched resonances could be reliably identified.

The combination of the limited solubility and the incomplete nature of the reaction of SiMe3I and SiMe3OTf with 2.5 prevent any unambiguous conclusions from being

drawn. The 13C{1H} NMR data indicates that the reaction does occur but why it does not go to completion and why the product resonances disappear on heating remains unclear, the 13C4O2(OSiMe3)2 product should be thermally stable up to 170 ºC.43 There

are several factors to consider; C4O2(OSiMe3)2 was found to undergo rapid

intermolecular silyl migration by VT 13C{1H} NMR and the observation of the crossover product by MS EI after the mixing of C4O2(OSiR3)2 (R = Me, CD3) at room

temperature, though the mechanism is unknown and also the T1 of the squarate carbons

in C4O2(OMe)2 are long.41 It is also not clear what effect being in solution with the

paramagnetic U(IV) centre will have on the 13C{1H} NMR spectrum of

13

C4O2(OSiMe3)2.

To avoid the difficulty associated with measuring small quantities of volatile liquids and to further test the stoichiometery of the reaction the bulkier SiPh3Cl was reacted with 2.5. To a solution of 2.5 in d8-thf 2 equivalents of SiPh3Cl was added in an NMR tube.

The solution was shaken manually and allowed to react overnight, after which the colour of the solution was observed to lighten from dark red to orange. The initial

13

C{1H} NMR spectrum indicated that incomplete reaction had taken place and an overnight 13C{1H} spectrum of the 13C4O2(OSiPh3)2 (2.6) product shown in Figure 6.

The NMR tube was heated at 50 ºC to encourage further reaction but no change to the spectrum was observed. However, the addition of excess Ph3SiCl (10 equivalents) did

result in the complete disappearance of bound squarate peak of 2.5.

The incomplete reaction of 2.5 with 2 equivalents of Ph3SiCl suggests the reaction does

not proceed stoichiometrically. It is unclear why this should be, although this was also the case in the reaction of 2.5 with SiMe3Cl, which required at least 2.5 equivalents of

SiMe3Cl to go to completion.17 Larger scale reactions would be needed to test the

stoichiometry of the reaction, most likely using the unlabelled complex [(U(8- C8H6{SiiPr3-1,4}2)(5-CpMe4H))2(-2:2-C4O4)] and GC-MS rather than NMR

spectroscopy.

Scheme 5: Synthesis of 13C4O4(SiPh3)2 (2.6)

The 13C{1H} spectrum of 2.6 (Figure 6) shows second order effects resulting from the coupling of the 13C labelled carbons, result in an AA’BB’ system of 12 lines.44 The second order nature of the spectrum can be seen in the roofing of the outer signals. The

13

C{1H} spectrum of 2.6, appears as two apparent triplets of doublets at  = 187.1, 187.7, 188.4, 193.4, 194.0 and 195.7 ppm, however, owing to the mixing of spin-states that results from the values  and the J being similar in magnitude, the apparent multiplicity in the second-order spectrum of 2.6 cannot be assigned. The chemical shifts in 2.6 are very similar to those observed for 13C4O2(OSiMe3)2, at  = 189.1, 189.7,

190.4, 195.7, 196.3 and 197.0 ppm, the 13C{1H} spectrum of which displays the same second order pattern.17

In 13C4O2(OSiR3)2 (R = Me, Ph) the value of the coupling constants will depend on the

amount of s-character in the hybridisation of the bonding orbitals45, however, couplings in conjugated systems46 or between 13C-labelled atoms47 can be much larger than predicted, as they result from a combination of -character and contributions from other coupling mechanisms.48 Data on 13C-13C coupling constants is not available for direct comparison in most cases49 and the solution of the specific contributions of the 1JCC, 2

JCC and 3JCC or the sign of 2JCC cannot be experimentally determined because the

experimentally observed couplings are the product of the mixing of spin states. The spectrum may be able to be either simulated, either from experimental data or from a calculated structure. The 13C{1H} NMR spectrum of 13C4O2(OSiMe3)2 was simulated

from the calculated structure by Nilay Hazari at the University of Oxford and is shown in Figure 7.

Figure 6: 13C{1H} NMR spectrum of 13C4O2(OSiPh3)2 (2.6) (d8-thf, selected data,

100.45 MHz, 30 ºC, 10000 scans, lb = 5)

Figure 7: Simulated 13C{1H} spectrum of 13C4O2(OSiMe3)2. JAA’ = 93.4 Hz, JBB’ = 90.0

Hz, JAB = JA’B’ = 49. 1 Hz and JAB’ = JA’B = 48. 3 Hz.

195.0 194.0 193.0 192.0 191.0 190.0 189.0 188.0 187.0 Ph₃SiO O O Ph₃SiO 13 13 13 13 A A' B B'

2.2.4 Reaction of [(U(8-C8H6{SiiPr3-1,4}2)(5-CpMe4H))2(-2:2-C4O4)](2.5) with

other halogenated reagents:

The reaction of 2.5 with 2 equivalents of benzyl chloride in d6-benzene showed only

starting material in the 13C{1H} NMR, even after heating at 70 ºC for 3 days. This is presumed to because the formation of a carbon-oxygen bond provides an insufficient driving force to break the uranium-oxygen bonds, even in combination with the formation of the uranium-halogen bond. For this reason 2.5 was reacted with reagents containing heteroatoms known to form strong bonds to oxygen.

Complex 2.5 did not react with an excess (7 equivalents) of isopropylphenylchloro