Synthesis and Characterization of a Linear, Two-Coordinate Pt(II) Ketimide Complex.

Herein we report the synthesis and characterization of a linear, two-coordinate Pt(II) ketimide complex, Pt(N=C t Bu 2 ) 2 ( 1 ), formed via reaction of PtCl 2 (1,5-COD) with 2 equiv of Li(N=C t Bu 2 ). Also generated in the reaction is the bimetallic complex, Bu)C(Me) 2 C H 2 )Pt(N=C t Bu 2 )] ( 2 ). Both complexes 1 and 2 have been characterized by NMR spectroscopy and X-ray crystallography. Notably, complex 1 exhibits short Pt-N distances (av. Pt-N = 1.815 Å) and an unusually deshielded 195 Pt chemical shift ( δ Pt = –629 ppm) with a

1 J( 195 Pt,14 N) coupling constant (537 Hz). These data, in combination with a detailed DFT electronic structure analysis, reveal the presence of highly covalent Pt=N multiple bonds (with a formal bond-order of 2.5) formed by a combination of -donation, -donation, and backdonation.

Introduction
Two-coordinate complexes have come under increased scrutiny for their high reactivity and unique magnetic properties. 1 Their low coordination number renders them highly reactive, owing to the ease of substrate access to the metal ion. Consequently, these materials are of interest for catalysis, small molecule activation, and as precursors to nanomaterials. [2][3][4][5][6] These properties are perhaps best exemplified by the M(N{SiMe3}2)2-type complexes (M = Mn, Fe, Co). 1 For example, Fe(N{SiMe3}2)2 was found to be an effective pre-catalyst for both alkene hydrogenation and carbonyl hydrosilylation. 7,8 With respect to magnetism, the highly anisotropic ligand field and unquenched orbital angular momentum extant in two-coordinate complexes leads to large magnetic moments and high barriers for magnetic reversal, which make these complexes promising single molecule magnets (SMMs). [9][10][11] For example, the two-coordinate Fe(II) amide complex, Fe(N t Bu2)2, features a high magnetic moment and a large internal magnetic field. 12 Similar results are observed for [Fe(C{SiMe3})2] 0/-. 9,13,14 For the lanthanides,  [15][16][17] Similarly, [Dy(O t Bu)2(py)5][BPh4], which contains two axial alkoxide ligands, features a remarkably high barrier to magnetic relaxation. 18 While these examples reveal the inherent promise of two-coordinate complexes in catalysis and for magnetic materials, further development in this area would benefit from the identification of new ligands that can enforce the desired 2-coordinate geometry. In this regard, the ketimide ligand, [R2C=N] -, may be a suitable candidate. Like amides and alkoxides, ketimides are strong donors, and should be able to generate the required anisotropic ligand field for SMMs. And like amides and alkoxides, ketimides are easily tailored and can feature a broad range of steric profiles and donor abilities. [19][20][21][22] Our research group has been exploring the chemistry of the ketimide ligand with a variety of transition metals. For example, we recently reported the syntheses of the homoleptic transition metal ketimides, M(N=C t Bu2)4 (M = Fe, Co), 23,24 which both feature the relatively rare +4 formal oxidation state for these metals. 23 They also both possess squashed tetrahedral geometries. 24,25 These unusual properties are thought to be a consequence of the interplay between the strong π-donating and π-accepting abilities of the linear ketimide ligand. 26 These strong donor properties suggested to us that ketimides could generate a strong axial ligand field in a two-coordinate complex.
Several other homoleptic M(N=C t Bu2)4-type complexes are also known, including examples containing M = Ti, V, Nb, Ta, Cr, Mo, W, and Mn. [24][25][26][27][28][29][30] Homoleptic ketimide complexes are also known for copper, boron, aluminum, uranium, and cerium. [31][32][33][34][35] In this context, the absence of a homoleptic group 10 ketimide is notable, especially given the foundational role that these elements have played in the development of organometallic chemistry. 36 To rectify this oversight, we explored the reactivity of a series of Pt(II) salts with Li(N=C t Bu2). Herein, we describe our first foray into group 10 ketimide chemistry, specifically the synthesis and characterization of a linear Pt(II) coordination complex, Pt(N=C t Bu2)2 (1), along with an analysis of its electronic structure and 195 Pt NMR spectrum using relativistic DFT calculations.

Results and Discussion
Synthesis and Characterization. The reaction of PtCl2(1,5-COD) with 2 equiv of Li(N=C t Bu2) in THF results in immediate formation of yellow-orange solution, which gradually changes to red-brown over 90 min. Work-up and crystallization from pentane results in the deposition of a mixture of small deep red-brown blocks of Pt(N=C t Bu2)2 (1) and [( t Bu2C=N)Pt(-N,C- (2) on the vial walls. 1 H NMR analysis of this mixture revealed the presence of 1 and 2 in an approximate 1:1 molar ratio (Scheme 1). Because of their similar solubilities, crystals of 1 and 2 always co-deposited, and because of their similar colors, they could not be manually separated. Thus, they were characterized as a mixture. They are both highly soluble in pentane, hexanes, Et2O, benzene, toluene, and THF, and somewhat soluble in MeCN. A few bright yellow crystals of a third product were also isolated from this reaction.
Analysis by X-ray crystallography revealed these to be [Pt(N=C t  The N=C bond lengths of the ketimide ligands in 1 (1.256(6), 1.258(2) Å) are typical of carbonnitrogen distances in other ketimide complexes, [24][25][26]43 and the Pt-N-C bond angles in 1 (179.7(4), 178.4(4)°) are consistent with sp hybridization of the N atom. [24][25][26]43 Finally, we note that the two ketimide ligands in 1 are co-planar, suggesting that the nitrogen lone pairs on the two [ t Bu2C=N]ligands interact with the same Pt(5d) orbital (see further discussion below). Complex 2, which was isolated along with 1, crystallizes in the monoclinic space group P21/n ( Figure 2). It features two Pt(II) centers, each ligated by a terminal ketimide ligand, and each bridged by a ketimide ligand that has also been deprotonated at a methyl carbon. The nitrogen of the modified ketimide ligand is ligated to both Pt centers, while its methylene group is only ligated to Pt2. Similar to 1, all three ketimide ligands in 2 are co-planar and the N-Pt-N angles (166.1(4) and 170.6(3)) approach linearity. The Pt-N distances for the terminal ketimide ligands (1.825(8) and 1.85(1) Å) are similar to those observed in 1, while their Pt-N-C angles (175.8 (8) and 164.7(8)) are close to 180. Both features are suggestive of a strong degree of π-donation and/or π-backdonation from/to the ketimide ligand. The Pt-Pt distance (2.5951(6) Å) is much shorter than those reported for other platinum(II) complexes with bridging amido ligands, 44,45 but is in line with those seen in Pt(II) acetate. 46 Finally, the Pt-C distance (2.08(1) Å) is similar to those of other structurally characterized Pt(II) complexes with C-H activated t Bu groups. [47][48][49][50]  While we were unable to isolate pure samples of 1 or 2 (their crystals always co-deposited; see Experimental Section), we were able to fully assign the 1 H NMR spectrum of the combined solids. The spectrum in C6D6 features of a sharp singlet at 1.11 ppm, which is assignable to 1 ( Figures S5 and S6). Additionally, there are 3 singlets, at 1.46, 1.30, and 1.28 ppm, each integrating for 9 protons, and one singlet at 1.22 ppm, integrating for 18 protons, which are assignable to the 5 magnetically inequivalent t Bu groups (with two overlapping) of 2.
Additionally, the spectrum features a singlet at 1.18 ppm, which integrates for 6 protons, and a singlet at 3.25 ppm, which integrates for 2 protons (and features Pt satellites with 2 JPtH = 88 Hz). Moreover, for a mixture of 1 and 2 that was slightly enriched with 1, we detected in C6D6 at 25 C a 195 Pt NMR resonance at -663 ppm that is assignable to 1 (Figure S10 in SI; 195 Pt NMR shifts in ppm vs. Na2PtCl6(aq)). Upon increasing the temperature to 50 °C, the signal shifted downfield (δPt = -629 ppm) and became a well-resolved quintet due to couplings with two equivalent 14 N nuclei (I = 1), with 1 J( 195 Pt, 14 N) = 537 Hz (Figure 3). The 195 Pt nucleus in 1 is remarkably deshielded as compared to other Pt(II) complexes with nitrogen-based ligands, which normally feature δPt within the range -2700 to -1700 ppm (cf. Table S7 in SI). In addition, the 1 J( 195 Pt,14 N) value is the largest coupling constant reported to date for a Pt-N bond, consistent with the short distance and multiple bond character in 1. The assignment of these NMR parameters to  Complex 3 can be independently synthesized by reaction of PtCl2(1,5-COD) with 1 equiv of Li(N=C t Bu2) in THF (Scheme 2). When generated in this fashion it can be isolated as an orange powder in 46% yield. Complex 3 has been characterized by X-ray crystallography and by 1 H and 13 C{ 1 H} NMR spectroscopy ( Figures S7, S8, and S15 in SI). In the solid state, it features a Pt-N bond length of 1.961(4) Å, which is notably longer than those of 1, and is suggestive of a significantly lesser degree of π-donation and π-backdonation between the Pt center and ketimide ligand ( Figure S15 in SI). Consistent with this hypothesis, the Pt-N-C angle in 3 (143.8(4)) deviates significantly from linearity.

Scheme 2. Synthesis of complex 3.
We also explored the suitability of 3 as a precursor to complexes 1 and 2. Thus, reaction of 3 with 1 equiv of Li(N=C t Bu2) in THF results formation of both 1 and 2 in an approximately 1:1 ratio, according to a 1 H NMR spectrum of the crude reaction mixture ( Figure S9 in SI). Several other minor resonances are also present in the reaction mixture, which we have tentatively assigned to complex 4. A small amount of HN=C t Bu2 is also present in the sample. Given that this route did not appear to offer any advantages over the initial method of preparation, it was not pursued further.
Computational Analysis. In an effort to explain the unique structural features of this twocoordinate Pt complex, we performed a computational bonding analysis of 1 and its truncated model complex, Pt(N=CH2)2 (1′). DFT calculations accurately reproduce the short Pt-N bond lengths, linear coordination geometry, and co-planar arrangement of ketimide ligands in both 1 and 1′ (Table 1). These features remain virtually unaffected upon replacing tert-butyl groups in the ketimide ligand with hydrogen atoms, demonstrating the minimal role of attractive London dispersion forces and/or steric repulsion between bulky alkyl groups on the structure.
A molecular orbital (MO) diagram for 1′ in its singlet (S = 0) ground-state is depicted in Figure 4 (selected frontier MOs of 1 are given in Figure S17 in SI and give qualitatively the same picture as presented here). As expected, the metal-based atomic orbitals dxy (HOMO-3) and dx2-y2 (HOMO-4) are basically non-bonding (they do not interact with the orbitals of the R2C=N fragments), the HOMO is predominantly metal-based 5dz2-6s in character, and the LUMO involves the 5dyz atomic orbital and has a clear π*(Pt-N) antibonding character. While it is sometimes a challenge to classify the nature of a bonding interaction from the canonical MOs, 53 the HOMO-5 (dyz) and HOMO-7 (dxz) orbitals are best described as the Pt-N π donation and Pt-N π back-donation interactions, respectively. This assessment is confirmed by inspection of the corresponding Naturally Localized Molecular Orbitals (NLMOs) 54 ( Figure S20) and Adaptive Natural Density Partitioning (AdNDP) 55 analysis ( Figure 5), which provide a chemically more intuitive description of bonding, exploiting the concept of Lewis structures with shared pairs of electrons between adjacent atoms (see discussion below). 53 In particular, the NLMO Pt-N bonds (shown in Figure S20 Table 2 for NPA charges and electron configurations at the metal center for a series of pertinent complexes). Our oxidation-state assignment is also confirmed by using the localized orbital bonding analysis (LOBA) method (Table 2), 59 which accurately produces the chemically intuitive oxidation state for challenging systems with redox non-innocent ligands. 60,61   (Table 1 and Figure S20) and becomes more covalent, along with the σ(M←N) component, on moving down group 10, which can be attributed to the more diffuse valence d-orbitals of the heavier metals (see also   Table S4 for results of EDA analysis, which provide similar trends). The lack of an analogous backdonation interaction in the group 10 M(N t Bu2)2 bis(amide) series is likely responsible for their longer M-N bonds, as compared to ketimides (Table S3). For instance, the Pt-N bondlength in the hypothetical bis(amide) Pt(N t Bu2)2 is computed to be 1.892 Å, which is 0.09 Å longer than that of 1.
As discussed above, the experimental structure of 1 exhibits a co-planar (eclipsed) arrangement of its ketimide ligands. We therefore explored computationally the alternative "allene-like" geometry wherein the ketimide ligands feature a staggered conformation. The structure of 1 with a staggered arrangement of the two ketimide ligands is computed to have a triplet ground-state (S = 1) and is energetically disfavored by 75.2 kJ/mol in comparison to the experimentallyobserved co-planar structure featuring a singlet ground-state (Tables 2 and S5). The energetic preference of the co-planar configuration permits the maximum constructive overlap between the -symmetry orbitals on the ligand and the relevant Pt(5d) orbitals. Specifically, a p-orbital from each nitrogen interacts with the "in-plane" 5dyz orbital and a *(C=N) orbital from each ligand interacts with the "out-of-plane" 5dxz orbital ( Figure S23). In the structure with the orthogonal configuration, however, the two ketimide ligands must interact individually with the Pt center, via two different Pt(5d) orbitals. As a result, the two orthogonal N(2p) orbitals cannot be involved in π-bonding; instead they are each occupied by an unpaired electron (cf. Figure S23 and Table S5 in SI for NLMO analysis and corresponding Lewis structures). Although these unpaired electrons are delocalized over the Pt and N atoms (cf. Figure S24 for the spin-density distribution), the metal center in the structure of 1 with an orthogonal arrangement of ketimide ligands is best characterized as a Pt(0) complex with two σ(Pt←N) bonding and two π(Pt→N) backbonding interactions. This is also reflected in the reduced atomic charge at the Pt center in contrast to the experimentally observed co-planar structure of 1 (S = 0), with its clear Pt(II) character (Table 2).

Summary
We have prepared and characterized the linear Pt(II) complex, Pt(N=C t Bu2)2 (1), expanding the scope of Pt(II) coordination chemistry beyond the more common square planar and T-shaped geometries. While many two-coordinate Pt(0) and Pt(I) complexes are known, 2-6 complex 1 is the first two-coordinate Pt(II) complex to be reported. 1 Its unusual coordination geometry can be rationalized by its exceptionally covalent M-N interactions, a consequence of the strong -donor and -acceptor properties of the linear ketimide ligand, as revealed by a detailed computational DFT analysis. These interactions result in short, highly covalent Pt-N bonds, which stabilize this formally unsaturated, 16especies. This covalency is also evident in its highly deshielded 195  C6H3-2,6(C6H2-2,4,6-i Pr3)2), 62 are also known, but these tend to be paramagnetic, with high-spin d 8 configurations at the Ni center. This change in the electronic ground-state is caused by two factors: first, the greater spatial extent of the 5d vs. 3d orbitals, and thus their better overlap with donor atoms on the ligands (i.e., larger M-L bond covalency); and, second, the strong π-donating and π-accepting properties of the linear ketimide ligand, which produces a larger crystal field than that provided by an amide ligand.
The unusual electronic properties imparted to Pt suggest that other linear [M(ketimide)2] n+ complexes (M = transition metal or lanthanide) would also feature interesting electronic structures and could potentially possess unique magnetic properties. The generation of species of this type will likely require the use of substantially bulkier ketimide ligands, by analogy to the bulky alkyl, aryl, and amido ligands previously used to stabilize two-coordinate transition metal complexes; 1,9,12,62 however, because of the relatively straightforward synthesis of the ketimide fragment, this should be easy to achieve.

ASSOCIATED CONTENT
Experimental, computational, and crystallographic details (as CIF files) for complexes 1 -4.