Indicadores de Desempeño

The major problem associated with these reactions was the removal of tin contaminants from the product mixtures at the end of the reactions. As outlined above, the best results were obtained with a partitioning of the crude product mixture between acetonitrile, hexane and cyclohexane with the hydrocarbon solvents removing most of the tin contaminants. Other tin purification methods were also attempted during the course of this work. Curran describes a tin removal procedure involving diazabicycloundecene (DBU) used in the purification of atom transfer cyclisations of a-iodo esters.90 The procedure involves addition of DBU to a diethyl

Chapter 5: P-Lactam synthesis via 4-exo amidyl radical cyclisations

solution. They formulated that molecular iodine converts both hexaalkylditins and trialkyltin hydrides into trialkyltin iodides. The DBU hydrolyses the tin halides to tin hydroxides which are removed by a short column. This procedure was unsuccessful with our experiments. This is presumably because the main by-products involved in this work, tin benzoates, cannot be converted into tin halides in this way. Curran also reports use of a fluorinated tin reagent [tris(2-perflourohexyl)ethyl)tin hydride] which

can be removed from the product mixture by simple phase separation techniques. 119

Adamantyl bromide was reduced with this reagent in (trifluoromethyl)toluene [which acted as a mixed, (part hydrocarbon, part fluorocarbon) solvent]. Evaporation of the solvent followed by phase separation between perfluoromethylcyclohexane (PFMC) and dichloromethane separated the tin products from adamantane. We encountered difficulties in the three step preparation to form the fluorinated tin hydride and therefore did not attempt to use this method in our cyclisations. However instead we looked at the use of a fluorinated precursor. A precursor for the 5-membered ring cyclisations presented in Chapter 3 was used. Hence 128 was prepared simply from the methyl hydroxamic acid, 81g and perfluorobenzoyl chloride in 91% yield (Scheme 5.13). After treatment with Bu3SnH and AIBN the crude 'H NMR showed

evidence o f cyclised product but partitioning between PFMC/dichloromethane and PFMC/acetonitrile showed little removal of the tin by-products. As no improvement was seen this procedure was not developed further. A critical amount o f fluorine substitution is necessary for organic fluoro compounds to be partitioned into fluorinated solvents and the lack of success here is presumably due to the relatively low amount o f fluorination in the resulting tin perfluorobenzoate.

Chapter 5: p-Lactam synthesis via 4-exo amidyl radical cyclisations

C6F5COCI Et3N, CH2CI2

Scheme S.13

Catalytic use of tin hydride has proved very successful with radical cyclisations involving halide precursors where the tin halides produced are reduced in situ by NaBH4.27-29 However, all attempts to reduce tributyltin benzoates failed and so this

system was not compatible with the O-benzoyl hydroxamic acid derivatives used here. There have been various reports in the literature on using polymer- or carrier- supported tin reagents.23-26 The idea is to bind the tin reagent in a very stable position

to an insoluble porous polymer which can be separated after the reaction simply by filtration. These supported reagents have been used for both reductions and cyclisations of organic halides and in most cases show good regenerative ability of the active tin hydride by reduction with diisobutylaluminium hydride. Neumann also reported use of his polystyrene-supported, regenerable tin hydride in performing dehydroxylation of secondary alcohols and deamination o f secondary or tertiary amines.2 5 120 The dehydroxylation reactions are carried out as Barton-type

deoxygenations which makes the leaving group similar to that used in this work (Scheme 5.14). Use o f a polymer-supported tin hydride could prove to be a useful method to overcome our tin purification system but there was insufficient time to investigate the technique during the course of this work.

Chapter 5: P-Lactam synthesis via4-exo amidyl radical cyclisations

Ph()C(S)CI

(p)-d^-C H 2C H 2S nB u2-H

Toluene, AIBN. 80°C, 18 h

Scheme 5.14

5.3 Conclusions and future work

The low yields obtained in the 4-exo amidyl radical cyclisations and the great purification problems encountered suggest that this is not yet a viable route for forming azetidinones. The reactions are important however in showing that amidyl radicals can undergo 4-exo-trig radical cyclisations to form p-lactams. By altering the way in which the radical is generated or the conditions under which the reaction is conducted a more successful, general method for this transformation may be found.

Use of the diphenyl precursor, 121e gave the best results and so further experiments involving biphenyl precursors with different jV-alkyl groups would hopefully improve the yields of all the cyclisations. The cyclisations could also be attempted with chiral N-R groups and with substituents attached to the chain in the 2-position. This might allow further functionality to be introduced for the synthesis o f more complex

Chapter 5: p-Lactam synthesis via 4-exo amidyl radical cyclisations

molecules. Varying the concentration at which the reactions are conducted would also be important in optimising the ratio of cyclised to reduced and rearranged compounds.

Ishibashi121 recently performed carbon-centred 4-exo radical cyclisations to form p-

lactams and used phenylsulfenyl substituents to stabilise the cyclic radicals produced. Thus, 129 was cyclised (by reaction with Bu3SnH and AIBN) to the P-lactam 130 in 56% yield as a mixture of two diastereoisomers (Scheme 5.15).

PMB = p-methoxybenzyl

Scheme 5.15

Performing the reaction in the absence of the phenylsulfenyl group at the terminus of the W-vinylic bond resulted in only low yields o f p-lactam formation along with the reduction product and the 5-endo cycliscd product. Clearly the phenyl group alone was insufficient to stabilise the radical intermediate. With this in mind adapting our procedure to involve precursors with one or two phenylsulfenyl substituents at the terminal end o f the double bond may well result in improved yields of cyclised products. It is likely also to have an effect on the ease of formation of the corresponding rearrangement products found in these reactions.

Chapter 6: Rearrangements of O-benzoyl hydroxamic acid derivatives

Chapter 6

Rearrangements of O-benzoyl hydroxamic acid derivatives