Secondary Phosphine Oxides as Multitalented Preligands En Route to the Chemoselective Palladium-Catalyzed Oxidation of Alcohols

International audienceSecondary phosphine oxides O = PHR 2 (SPOs) were identified as multitalented preligands for the chemoselective Pd-cata-lyzed oxidation of alcohols by a hydrogen-abstracting methodology. SPOs were found to promote the hydrogen-abstraction step as well as hydrogen transfer to a Michael acceptor by generating a putative active HÀPd species. The catalytic system operates under neutral conditions and was proven to be compatible with various electrophilic and nucleophilic func-tionalities within the substrates as well as water-and air-sensitive functional groups. The surge in transition-metal-based catalysis has unquestionably paved the way for the design of efficient catalytic systems in the field of the oxidation of alcohols, as evidenced by the numerous reports in this regard. [1] In particular, the recent ree-merging interest in using palladium(II)-based catalysts [2] in combination with oxygen [3] has conditioned the design of well-defined complexes for achieving highly selective transformations of primary or secondary alcohols into the corresponding aldehydes or ketones, respectively. [4] Such a Pd II-based approach typically involves a putative Pd–hydride species as the crossroads through alcoholate formation and subsequent b-H elimination. Completion of the catalytic cycle then requires either reoxidation of a Pd 0 species or direct reaction of HPdX with oxygen to regenerate the active catalyst (Scheme 1, pathway a). [5] Despite the clear environmental benefit of using oxygen and great advances achieved thereof, many serious issues such as intolerance of sensitive functional groups and the nature of the ligands may emerge in maintaining catalytic activity and/or regio-and chemoselectivity. [6] Alternatively, oxidizing alcohols in the absence of any formal reoxidant necessarily implies metal-mediated hydrogen transfers from the sub-strate to a hydrogen acceptor such as electron-poor alkenes. [7] This approach was recently termed " abstracting-hydrogen methodology " [8] (AHM) (Scheme 1, pathway b). Surprisingly, scarce literature precedent [9] in which this combination has convincingly been used proves that such a synergism remains to be harnessed. It should be noted that oxidation through hydrogen transfer is well documented within ruthenium chemistry. [7] However, most of the involved substrates are devoid of other functional groups, [10] probably owing to the high oxophi-licity of the metal and the basic conditions, which are usually required. The AHM-based approach is feasible if the HÀPd II species generated by alcohol oxidation reacts with the H-acceptor faster than it suffers from reductive elimination or a degradation process, and the quest for a multitalented preligand L is therefore of paramount importance. First, L should sufficiently provide electron density towards the metal but still maintain moderate steric hindrance. [11] Second, it should not promote intramolecular deprotonation of the PdÀH bond, which thus rules out a good number of bidentate nitrogen ligands. [12] Third, it should operate under neutral conditions to preserve the chemical integrity of both the PdÀH species and the sub-strate. Lastly, L would be synthetically useful only if it is both air and moisture stable and easy to handle. [13] In this context, finding an alternative to ligands that are able to stabilize HÀPd species such as pincer ligands is highly desirable. Our solution to tackle this vexing issue lies in secondary phosphine oxides [14] (SPOs, Scheme 2), which are air stable, [15] unaffected by water, [16] easy to handle, [14a] and good electron-donating preligands. [17] For instance, complex A, easily available from SPOs and Pd(OAc) 2 , [18] would be expected to generate a HÀPd species B through oxidation of an alcohol. Moreover, the six-membered hydrogen-bonded chelate structure of B would enable stabilization of a 14-electron cat-ionic Pd II complex, from which reductive elimination would be rather difficult. This peculiar structure would also allow strong minimization of the steric hindrance, as the edge between the P atom and the hydride could approach 1208. [19] In earlier studies, we noticed that treatment of Pd(dba) 2 (dba = dibenzylideneacetone) with rac-tert-butyl(phenyl)phos-phane oxide (L, 2 equiv.) and acetic acid (2 equiv.) in toluene resulted in the reduction of one C=C bond of the dba ligand (Scheme 3). [20] This result could be attributed to putative species C formed through ligand exchange. The loss of acetic acid would afford Pd II –hydride complex C, which upon loss of AcOH yields D. The latter then reacts with dba and AcOH to deliver E. This observation clearly highlights L as a Pd II ÀH generator preligand class, all the more so given that oxidative addition of AcOH to Pd 0 is known to proceed through an unfavorable equilibrium. [21] [a] Dr