A Cation-Captured Palladium(0) Anion: Synthesis, Structure, and Bonding of [PdBr(PPh 3 ) 2 ] – Ligated by an N -Heterocyclic Phosphenium Cation

Unsaturated N-heterocyclic phosphenium cations ( u NHP) stabilize the [Pd 0 (PR 3 ) 2 X] – anion proposed over the past decade to be the crucial but elusive intermediate in palladium-catalyzed cross-coupling reactions (X = halide). Insertion of metal into the P – Br bond of the precursor mesityl-substituted bromophosphine gives the structurally characterized Pd(0)-phosphenium complex ( u NHPMes)Pd(PPh 3 ) 2 Br, which features a long Pd – Br bond (2.7240(9)Å) and the shortest known Pd – P bond (2.1166(17)Å). The reaction is proposed to proceed by an associative pathway involving a Pd-bromophosphine complex that undergoes P-to-Pd bromide transfer.


Introduction
Palladium(0)-phosphine complexes play a pivotal role in a wide variety of catalytic C-C bond-forming reactions that since the 1970s have become part of the standard lexicon of organic synthesis. The scope, utility, and sophistication of Heck, Suzuki, Sonogashira, Stille and Negishi couplings, among others, have championed the use of palladium as the workhorse for demanding connections, and have spurred the development of a vast array of tailor-made phosphine ligands. 1 The traditional wisdom that oxidative addition of R-X to linear Pd(0)bis(phosphine) complexes generates the requisite organo-palladium(II) intermediates during these catalytic cycles has been challenged over the last decade by Amatore and Jutand, who have argued convincingly, based on kinetic, spectroscopic, electrochemical and computational evidence, that the active species in these additions is the tricoordinate, Pd(0) anion, [Pd(PR3)2X] -(X = halide, acetate, etc.). [2][3][4] In accord with chemical intuition and on the basis of DFT calculations, these researchers have proposed that the stability of [Pd(PR3)2X]should increase with increasing electron-withdrawing character of the phosphine ligands and with increasing electron donating character of the Xanion. 4 However, such putative anionic intermediates have never submitted to isolation and structural characterization.
Our recent investigations have shown that N-heterocyclic phosphenium (NHP) cations are powerful -acids that aggressively sequester electron rich Group 10 metal(0) centres. 5

Results and Discussion
Bromophosphine 1 7 was made using the redox method of Macdonald and coworkers by the reaction of PBr3 with N,N'-dimesityl-1,4-diazabutadiene in the presence of cyclohexene as a scrub for the Br2 byproduct. 8 Its 31 P{ 1 H} NMR spectrum was characterized by a singlet at P 175.1. The molecular structure of 1 is shown in Figure 1.
This compound featured a relatively long P-Br bond of ca. 2.62 Å, which was shorter than the sum of the van der Waals radii of P and Br (3.65 Å), but significantly longer than the bonds in PBr3 (2.22 Å). The P-Br distance in 1 was also longer than the corresponding bonds found in the three crystallographically independent molecules observed by Macdonald and coworkers in the unit cell of ( Mes BIAN)P-Br, which spanned 2.432(3)-2.506(3) Å ( Mes BIAN = 1,2-bis(2,4,6-trimethylphenylimino)acenaphthene).
Gudat's group has structurally characterized chloro analogues having unusually long P−Cl bonds (2.32-2.69 Å) 9 that were attributed to a combination of π(C2N2)−σ*(P−Cl) hyperconjugation and P−Cl bond polarization by intermolecular influences. 9,10 It is likely that similar effects were at play in 1. The C-C and intraanular C-N distances in 1 were in the expected range for double and single bonds, respectively. Compared to the free diimine, the C-C bond was shortened in 1 (1.341 ( Reaction of 1 with Me3SiOTf cleanly gave the known phosphenium triflate [uNHP Mes ]OTf (2) which has been made by Gudat using a different approach. 12 Concomitant with halide abstraction was the expected dramatic downfield shift in the 31 P{ 1 H} NMR spectrum (P 204.7; P 29.6), which indicated only weak contact between the cation and anion in solution. In addition, the "backbone" CH protons gave rise to a peak in the 1 H NMR spectrum of 2 ( 8.30) that was significantly downfield of the corresponding peak in the spectrum of 1 ( 6.67), which was consistent with the 6 aromatic nature of the phosphenium ion. The [uNHP Mes ] + cation has also been isolated by Cowley in both I3and BPh4salts. 13 Our spectroscopic data agreed with those previously reported. 9 features was similar to that which we reported recently for the saturated analogue, OTf; however, in the latter case, the coordinated phosphenium cation gave rise to a peak that was much further downfield (P 260). The solid state structure of 3 is shown in Figure 2. To the best of our knowledge, this compound is the first structurally characterized Pd complex of an NHP cation. The Pd-NHP bond (2.1229(11) Å) was shorter than any other Pd-P contact reported to date in the literature (but slightly longer than that in 4), consistent with double bond character and very strong -backbonding. This description of bonding was supported by computational analyses, which indicated that -backbonding formed 65% of the total orbital interaction energy within the Pd-P bond (see Supporting Information). The near trigonal planar nature of both the NHP P-and Pd-atoms indicated that this cation was accurately described as a Pd(0)-phosphenium rather than a Pd(II)-phosphide, as we have reported for the metal complexes II (Chart 1). 5   (4)     precursors gave square planar Pd(II)-boryl complexes. 16 And for E = P, Dyer reported production of 42-electron tripalladium clusters featuring bridging phosphide ligands by reaction of acyclic bis(dialkylamino)chlorophosphines with Pd2(dba)3. 17 We are not aware of reports of P-X heterolysis without oxidative addition mediated by Pd(0) that predate the present disclosure, but the first example of this general transformation was by Gudat who showed that an unsaturated, cyclic diaminochlorophosphine similar to 1 added to M(0)(CO)4(bipy) (M = Mo, W) to give M(0)-phosphenium complexes. 18  In order to distinguish between these possibilities, we treated 1 and 2 with excess 2,3-dimethyl-1,3-butadiene in parallel reactions under identical conditions. This test relied on our earlier observation of the cycloaddition reaction between saturated NHP cations and dienes to give spirocyclic diazaphospholeniums. 19 Although the reaction was much more sluggish than that of its saturated analogue, the phosphenium triflate 2 underwent the expected cycloaddition with prolonged heating (Scheme 3). The bromophosphine 1 by comparison was completely unreactive. This difference strongly implied that despite its long P-Br bond, 1 did not dissociate bromide in solution and that the formation of 4 probably proceeded initially by an associative reaction between bromophosphine and a low coordinate Pd (0) We have also investigated the reversibility of the formation of 4 by incubating the isolated complex with BruNHP Dipp (Dipp = 2,6-diisopropylphenyl) at rt for several hours. None of the statistically expected complex (uNHP Dipp )Pd(PPh3)2Br was observed by 31 P{ 1 H} NMR spectroscopy, which showed that insertion reaction was essentially irreversible at rt.

Conclusions
We believe that the successful isolation and structural characterization of 4

Supporting Information Available
General experiemental considerations, a tabulated summary of crystallographic data, and full crystallographic data for 1, 3 and 4 as CIF. This information is available free of charge via the Internet at http://pubs.acs.org.