Dedicated to Professor Philip P. PowerThe terminal, anionic niobium phosphide (P 3À ) complex Na [P Nb(N[Np]Ar) 3 ] (Na-1; Np= neopentyl, Ar = 3,5-Me 2 C 6 H 3 ) has served as a platform for the construction of multiply bonded, phosphorus-containing moieties as both complexed ligands and free entities. [1][2][3][4][5] Exemplifying the latter is the synthesis of phosphaalkynes (RCP) and [O Nb{N(Np) [5] We envisioned that a similar method could be employed to synthesize new transition-metal phosphide complexes, given the availability of a suitable reaction partner for Na-1. We recently reported an attractive candidate for this application: the tungsten oxide/chloride complex, [O(Cl)W{N(iPr)Ar} 3 ] (3).[6] In a reaction similar to Equation (1), complex 3 is prepared by treating the corresponding nitride complex [NW{N(iPr)Ar} 3 ] (4) with pivaloyl chloride (tBuC(O)Cl), and is quantitatively obtained as a blood-red, highly lipophilic solid of sufficient purity for further synthetic studies. We hypothesized that in the presence of Na-1, 3 would behave as an "inorganic acid chloride", engaging Na-1 by elimination of NaCl. With subsequent intermetal exchange of P and O ligands, the known oxidoniobium complex 2 would be generated as well as a new terminal phosphide complex [P W{N(iPr)Ar} 3 ] (5). Several examples of installing multiply bonded ligands (oxide, imide, and alkylidene) by intermetal ligand exchange have been reported; [7][8][9][10][11][12][13][14][15][16] herein, we extend this concept to the terminal phosphide functional group.Our hypothesis was tested by treating a yellow-orange, ethereal solution of Na-1 with a red, ethereal solution of 3 at À35 8C, following in situ preparation of 3 from 4 and removal of tBuCN. Upon stirring for two hours at 22 8C, an orangebrown, homogeneous solution was obtained. Following solvent removal under reduced pressure the product mixture was dissolved in C 6 D 6 . [17]Separation of 5 from NaCl was trivial; however, separation of 5 from coproduct 2 at first proved difficult, owing to their similar solubility properties. This was initially overcome by applying the Pasteur method, [18] whereby mixtures of crystalline 2 and 5 were manually separated.A more efficient method of separating 5 from coproduct 2 was achieved by in situ conversion of 2 to the bistriflate complex [(TfO) 2 Nb{N(Np)Ar} 3 ] (6, Tf = CF 3 SO 2 ), [5] a complex that is only sparingly soluble in common hydrocarbon solvents. Accordingly, treatment of a 1:1 mixture of 2 and 5 in Et 2 O with neat triflic anhydride (Tf 2 O, 1 equiv) at room temperature resulted in a color change from orange-brown to yellow-brown over several minutes as a yellow precipitate formed. Following solvent removal under reduced pressure, the product mixture was dissolved in C 6 D 6 . The 1 H NMR spectrum of the product mixture revealed clean and selective production of 6, with 5 remaining unperturbed. The 31 P{ 1 H} NMR spectrum displayed only the downfield signal associated with 5, thus confirming its role as a spectator in this react...