Metal‐Mediated Oligomerization Reactions of the Cyaphide Anion

Abstract The cyaphide anion, CP−, is shown to undergo three distinct oligomerization reactions in the coordination sphere of metals. Reductive coupling of Au(IDipp)(CP) (IDipp=1,3‐bis(2,6‐diisopropylphenyl)imidazol‐2‐ylidene) by Sm(Cp*)2(OEt2) (Cp*=1,2,3,4,5‐pentamethylcyclopentadienyl), was found to afford a tetra‐metallic complex containing a 2,3‐diphosphabutadiene‐1,1,4,4‐tetraide fragment. By contrast, non‐reductive dimerization of Ni(SIDipp)(Cp)(CP) (SIDipp=1,3‐bis(2,6‐diisopropylphenyl)‐imidazolidin‐2‐ylidene; Cp=cyclopentadienyl), gives rise to an asymmetric bimetallic complex containing a 1,3‐diphosphacyclobutadiene‐2,4‐diide moiety. Spontaneous trimerization of Sc(Cp*)2(CP) results in the formation of a trimetallic complex containing a 1,3,5‐triphosphabenzene‐2,4,6‐triide fragment. These transformations show that while cyaphido transition metal complexes can be readily accessed using metathesis reactions, many such species are unstable to further oligomerization processes.


Introduction
[3][4][5][6] Cyanide-containing organic molecules, or nitriles, are also highly important in materials chemistry as monomer precursors to triazines, which are privileged scaffolds in reticular chemistry for the construction of framework materials, e.g.9][10][11] While the oligomerization of nitriles is achievable by several methods, including acid catalysis and at elevated temperatures, [12][13][14] the predominant oligomerization path-way for the cyanide ion is by oxidation (to afford cyanogen). [15]More uncommonly, the cyanide ion will afford a tetrameric C 4 N 4 4À moiety, [16,17] or the reduced dimer C 2 N 2
We recently reported a general synthetic route to cyaphide-containing molecules using a magnesium(II) cyaphide transfer reagent Mg( Dipp NacNac)(dioxane)(CP) (A; Dipp NacNac = CH{C(CH 3 )N(Dipp)} 2 ; Dipp = 2,6-diisopropylphenyl), enabling the isolation of reactive cyaphido transition metal complexes. [37]In contrast to the relatively inert behavior of cyanide ligands, metal cyaphide complexes are significantly more reactive due to the weak nature of the C(2p)À P(3p) π-bond.While this presents some challenges for the isolation of cyaphido metal complexes, it also provides an opportunity for the generation of longer chained, conjugated organophosphorus anions that may serve as linkers between metal atoms, for example in MOFs.Further, through the study of these oligomerization processes we can gain understanding of the factors that govern the stability of transition metal cyaphide complexes, which will inform the design of cyaphide-containing materials such as analogues of Prussian Blue.
In this work, we report a pattern of reactivity of the cyaphide ion toward oligomerization to form di-, tri-, and tetra-anionic cyaphide oligomers.
The solid-state structure of 1 contains a reduced cyaphide dimer, C 2 P 2 4À , in which the two halves of the molecule are joined by a single bond between the phosphorus atoms (Figure 2).This is in contrast to the behavior observed for the cyanide ion which forms a CÀ C bond on reductive coupling. [18]The P1À P1' bond length is 2.306(3) Å, and the C1À P1 bond lengths are 1.623(6) Å.The central C 2 P 2 4À moiety is bonded to the two gold centers by its carbon atoms, and to two samarium centers by both carbon and phosphorus.The Au1À C1 bond length is 2.031 (5) Å, the Sm1À C1 bond length is 2.461(6) Å, and the Sm1À P1 bond length is 2.894(2) Å.The C1À P1À P1' bond angle is 110.4(2)°, and the Au1À C1À P1 angle is 118.2(3)°,suggesting sp 2 hybridization of the carbon atom. 1 is isostructural to a previously reported 2,3-diphosphabut-diene-1,4-diide ion, [20] and is also structurally similar to reported carbene-stabilized moieties. [42]t room temperature B does not oligomerize spontaneously despite poor steric protection of the C�P triple bond.However, we have found that the attempted syntheses of other sterically unprotected cyaphide complexes frequently results in uncontrolled reactivity.We hypothesized that this could be due to poor electronic stabilization of the cyaphide ion, which in turn allows for oligomerization or polymerization to occur.With the aim of investigating this possibility, we targeted the synthesis of cyaphide complexes featuring a range of different transition metals.
The reaction of A with Ni(SIDipp)(Cp)Cl results in the formation of the nickel(II) cyaphide complex Ni(SIDipp)-(Cp)(CP) (2) (Scheme 2). [43]Compound 2 exhibits a singlet resonance in its 31 P{ 1 H} NMR spectrum at 181.2 ppm corresponding to the cyaphide phosphorus atom, as well as a doublet in its 13 C{ 1 H} spectrum at 224.02 ppm ( 1 J C-P = 10.4Hz) corresponding to the cyaphide carbon atom.The C�P stretching frequency of 2 appears at 1307 cm À 1 in its infra-red (IR) spectrum (cf.1342 cm À 1 for B) (Figure S4). [37]o our surprise, attempts to grow X-ray quality crystals of 2 at room temperature over 1 week (or at À 35 °C over several

Research Articles
weeks) resulted in the crystallization of an asymmetric bimetallic complex, {Ni(SIDipp)(Cp)}{Ni(Cp)}-{μ 2 -(SIDipp)C 2 P 2 } (3), the product of spontaneous dimerization of 2. A closely related complex has been invoked by Wolf as an intermediate in the nickel-mediated dimerization of tert-butyl phosphaalkyne to afford a diphosphatetrahedrane. [26]he 31 P{ 1 H} NMR spectrum of 3 reveals an AX spin system with doublets at 130. The mechanism of the dimerization of 2 was investigated by density functional theory (DFT) calculations (see the Supporting Information for details). Th mechanism proceeds via an initial rate-limiting migratory insertion of the cyaphide ion into the nickel carbene bond (Figure 4), forming a η 2 -phosphaallenide intermediate. Rlated migratory insertions of isocyanides on palladium(II) have previously been reported.[45] This insertion step can be compared to the reverse reaction of the recently reported oxidative addition of the sp 2 -sp CÀ C bond of ArÀ CP (Ar = 2,4,6-trimethylphenyl) by the heavier group 10 complex LPt (L = 1,2-bis(dicyclohexylphosphino)ethane).[34] The transient η 2 -phosphaallenide reacts with another equivalent of 2 in a [2+2] cycloaddition reaction, followed by a η 2 -η 3 hapticity change to afford 3. Alternate plausible dimerization products with η 1 : η 1 and η 2 : η 2 hapticities were both found to be higher in energy than 3.
The solid-state structure of 5 reveals three Sc(Cp*) 2 centers coordinated to a central C 3 P 3 3À ion (Scheme 3).The C 3 P 3 3À ring is disordered over two chemically equivalent positions in the solid-state, preventing reliable discussion of its experimentally determined bond metrics.However, DFT

Research Articles
calculations are in good agreement with the modeled disorder components and show that (in the gas phase) they are related by a low energy ring-rotation barrier.
In comparison to the gold(I) cyaphide complex B, which has only been observed to oligomerize on reduction, the spontaneous trimerization of 4 stands in stark contrast.Neither B nor 4 afford the cyaphide ion any significant steric protection.However, the two complexes are opposites in terms of their MÀ C bond character.Scandium(III), an early 3d metal, is expected to form bonds with highly ionic character, contrasting the highly covalent MÀ C bonds formed by the late 5d metal gold(I).The ScÀ CP bond in 4 has a substantially lower covalency than the AuÀ CP bond in B, as shown by their QTAIM delocalization indices (δ- ) = 0.46, δ(AuÀ C) = 1.13; Figure S17-S20).Their Natural Localized Molecular Orbital (NLMO) bond orders follow a similar pattern with a value of 0.31 for ScÀ CP and a value of 0.41 for AuÀ CP (Table S2).In addition, Natural Bond Order analysis identifies a AuÀ CP covalently bonded Lewis structure for B, but instead describes the ScÀ CP bond as a dative bond from a carbon-centered lone pair (Table S2).These data suggest that the difference in reactivity between B and 4 may be due to the higher degree of electronic stabilization of the cyaphide ion by covalent bonding in B when compared to 4, which is also reflected in a larger C�P π-π* orbital energy gap (Figure S16).In addition, the coordinative assembly of 4 into Lewis acidbase adducts in the calculated trimerization mechanism may further activate the C�P bond.This suggests that strong Lewis acids would, in general, facilitate oligomerization.
Complex 2 falls between B and 4 in terms of MÀ C covalency.Nickel(II), a late 3d-transition metal, is likely capable of forming stable cyaphide complexes given a suitable coordination environment (we recently reported an isolable cobalt(I) cyaphide complex supported by a NNNpincer ligand). [37]However, in the case of 2 a different oligomerization pathway is available in which migratory insertion yields a transient η 2 -phosphaallenide responsible for the subsequent cyclization to 3 (Figure 4).This observation, in combination with the recent work of Görlich et al., [34] implies that (at least for the group 10 metals) ligands coordinated to the metal through an sp 2 -C should be avoided if the intention is to form stable MÀ CP complexes.

Conclusion
To conclude, we have shown that the cyaphide ion is highly reactive in transition metal complexes, readily oligomerizing to form C n P n xÀ units (n = 2, x = 2, 4; n = 3, x = 3).This mode of reactivity shows that the cyaphide ion can act as a synthetic precursor to larger conjugated organophosphorus molecules.The propensity for cyaphide oligomerization is heavily dependent on electronic stabilization, proceeding most readily for the ionic scandium(III) complex 4. Importantly, this shows that the cyaphido ligand can remain reactive in the coordination sphere of metals with low steric protection, which will be of consequence in the design of cyaphide containing materials such as heavy analogues of Prussian Blue.

Figure 1 .
Figure 1.Top: Oligomerization reactions of cyanides; Bottom: Oligomerization reactions of the cyaphide ion reported in this work.

Scheme 1 .
Scheme 1. Synthesis of 1 by reductive coupling of B.

Figure 6 .
Figure 6.Quantum Theory of Atoms-In-Molecules (QTAIM) analysis of 5 (full details provided in the Supporting Information).

Figure 7 .
Figure 7. Calculated mechanism for the formation of 5 from 4 (ZORA-ωB97X-D3/ZORA-def2-TZVP).The barrier for C 3 P 3 3À ring rotation relating the two disorder components in the solid-state structure of 5 is also shown.