Regioselectivity in the transfer-dehydrogenation of alkanes by two pincer-iridium-based catalysts

The activation of C-H bonds by ligated transition metal complexes has captivated the attention of catalytic chemists since the work of Bergman, Graham, Jones and others in the early 1980's. Perhaps the most important aspect of this work was the discovery that such complexes showed selectivity in oxidative addition toward stronger C-H bonds (1° > 2° > 3°). In accord with this stoichiometric selectivity, it was later found that PCP-pincer iridium complexes selectively dehydrogenated n-alkanes at the terminal position. More recently this has been exploited in the catalytic metathesis of n-alkanes by tandem systems based on pincer-iridium-catalyzed dehydrogenation and olefin metathesis by Schrock-type Mo catalysts. In some cases, such systems give selectivity for C(2n-2) n-alkane plus ethane, from two mol C(n) n-alkane, e.g. n-decane and ethane from two mol n-hexane. The two pincer ligands most used in this work to date have been the bisphosphine, ^tBuPCP, and the bis phosphinite, ^tBPOCOP. Only complexes of the ^tBPCP ligand, however, have demonstrated selectivity for the C(2n-2) n-alkane; ^tBPOCOP-ligated catalysts (and others) typically give n-alkane product with a broad distribution of molecular weights. A priori, this different overall regioselectivity could result from rapid isomerization of terminal olefins by (^tBPOCOP)Ir to give internal olefins (prior to metathesis), or initial dehydrogenation of n-alkanes at internal positions by (^tBPOCOP)Ir. In this paper we will present and analyze evidence for both of these scenarios, and attempt to explain the origin of the subtle but critical differences between complexes of these two seemingly similar ligands.