A Catalytic Cross‐Olefination of Diazo Compounds with Sulfoxonium Ylides

Abstract A ruthenium‐catalysed cross‐olefination of diazo compounds and sulfoxonium ylides is presented. Our reaction design exploits the intrinsic difference in reactivity of diazo compounds and sulfoxonium ylides as both carbene precursors and nucleophiles, which results in a highly selective reaction.

acceptor diazo compounds can be cross-coupled selectively to acceptor-substituted diazo compounds by rhodium catalysis.S tereoselectivity was generally high in favour of the E olefin. [7a] Although the scope was extended by Sun and coworkers,t his method requires both an electron-withdrawing and an electron-donating group adjacent to the diazomethane moiety of at least one of the reaction partners. [7a, c-e] An additional procedure by Wang and co-workers relies on the use of cyclopropenes as carbene precursors,w hich are then coupled to diazo compounds. [6d] Thesame group published an interesting coupling of diazo compounds with in situ generated difluorocarbene. [6e] Sulfonium and sulfoxonium ylides have witnessed ar enaissance in contemporary catalysis beyond the well-known Johnson-Corey-Chaykovsky reactions, [8] namely in CÀH functionalisation, [9] NÀHi nsertion, [10] and cycloisomerisation reactions. [11] Their popularity is owed to the fact that they are easy to prepare,r eadily purified, and considerably safer to handle than their diazo counterparts.
Although sulfoxonium ylide dimerization has been observed indirectly,i th as never been used for the effective synthesis of olefins.T his is probably also due to the fact that the generated products (electron-poor olefins) are also good substrates for Johnson-Corey-Chaykovsky cyclopropanation under the reaction conditions. [12] Indeed, sulfoxonium ylides are generally better nucleophiles than their diazo analogues, but tend to undergo decomposition to the metal carbene at much lower rates than their diazo counterparts,r endering ah ypothetical catalytic sulfoxonium ylide (cross)-coupling adifficult prospect. [13] This led us to speculate that the catalytic cross-coupling of ad iazo compound with as ulfoxonium ylide should be possible.W es urmised that formation of am etal carbene should take place faster from the diazo compound precursor, and that the resulting electrophilic carbene would be attacked preferentially by the more nucleophilic sulfoxonium ylide. However,wewere uncertain whether the coupling product (a Michael acceptor) would be prone to conjugate addition by the sulfoxonium ylide.
In our first trials,arange of iridium(I) and rhodium(II) catalysts were investigated for the cross-olefination of diazoester 1a and sulfoxonium ylide 2a,o wing to their welldocumented proficiency in metallocarbene formation. [7a,14] Those preliminary experiments ( Table 1, entries 1a nd 2; see the Supporting Information for further experiments) led to low but promising yields of the desired product, together with diethyl maleate/diethyl fumarate resulting from homodimerization of diazoester 1a as the main side product. Importantly,h omodimerization of the sulfoxonium ylide 2a was virtually absent, corroborating our initial hypothesis. Unreacted sulfoxonium ylide could be removed completely along with the catalyst during workup,resulting in aclean and easy-to-analyse crude 1 HNMR spectrum. Cyclopropanation side products were never observed. [15] In the course of catalyst screening,w ef ound that the cheap ruthenium complex [Ru(p-cymene)Cl 2 ] 2 [16] displays the highest efficiency for this cross-olefination. Further optimization of the conditions led to good isolated yields above 70 %w ith a Z/E ratio of 9:1 (Table 1, entry 5).
[c] Yields determined by 1 HNMR spectroscopy using mesitylene as an internal standard.
[d] Yield of isolated product.

Scheme 2.
Sulfoxonium ylide substrate scope. The Z/E ratios were determined by 1 HNMR analysis of the crude product mixture. All yields are for pure, isolated material unless indicated otherwise.
[a] 1 HNMR yield determined using mesitylene as an internal standard.
[b] DMF was used as acosolvent.
Ad irect comparison of the cross-olefination procedure reported herein with the cross-olefination of two diazo compounds reveals that yield and selectivity are considerably higher when sulfoxonium ylides are employed (Scheme 5). Moreover,w hile the cross-olefination of sulfoxonium ylides and diazoesters delivers the product of homodimerization of the ester moiety (product 3ma;s ee Scheme 5) as the only undesired side product in small amounts,cross-olefination of two diazo compounds results in amixture of all three possible coupling products with virtually no selectivity.( In the event, the desired cross-coupled product 3aa is not even the major product.) During the optimization studies,several quenching agents were investigated. While pyridine,p yrimidine,a nd dimethyl sulfide shut down the reaction, triphenylphosphine had an additional effect: Z/E diastereomeric mixtures were converted completely into the E isomer when substoichiometric amounts of PPh 3 were added to the reaction mixture.Further studies showed that this isomerization [21] takes place not only under the reaction conditions but also in solutions of isolated products (Scheme 6).
In conclusion, an ovel ruthenium-catalysed cross-olefination of diazo compounds and sulfoxonium ylides has been presented. Our reaction design exploits the intrinsic difference in reactivity of diazo compounds and sulfoxonium ylides as both carbene precursors and nucleophiles,r esulting in ah ighly selective reaction that nicely complements known, often less selective diazo-diazo coupling reactions.T his results in the generation of olefin products with high Z selectivity.