Syntheses and Structures of Magnesium and Zinc Boraamidinates: EPR and DFT Investigations of Li, Mg, Zn, B, and In Complexes of the [PhB(N t Bu) 2 ]  Anion Radical

The first magnesium and zinc boraamidinate ( bam ) complexes have been synthesized via metathetical reactions between dilithio bam s and Grignard reagents or MCl 2 (M = Mg, Zn). The following new classes of bam complexes have been structurally characterized: the heterobimetallic spirocycles {(L)  -Li[PhB(  -N t Bu) 2 ]} 2 M ( 6a or 6b , M = Mg, L = Et 2 O or THF; 6c , M = Zn, L = Et 2 O), bis(organomagnesium) complexes {[PhB(  3 -N t Bu) 2 ](Mg t Bu) 2 (  3 -Cl)Li(OEt 2 ) 3 } ( 8 ) and {[PhB(  3 -N t Bu) 2 ](MgR) 2 (THF) 2 } ( 9a , R = i Pr; 9b , R = Ph), and the mononuclear complex {[PhB(  -NDipp) 2 ]Mg(OEt 2 ) 2 } ( 10 ). Oxidation of 6a or 6c with iodine produces persistent pink ( 16a , M = Mg) or purple ( 16b , M = Zn) neutral radicals {L x  -Li[PhB(  -N t Bu) 2 ] 2 M}  (L = solvent molecule), which are shown by EPR spectra supported by DFT calculations to be C s -symmetric species with spin density localized on one of the bam ligands. In contrast, characterization of the intensely colored neutral radicals {[PhB(  -N t Bu) 2 ] 2 M}  ( 5c , M = In, dark green; 5d , M = B, dark purple) reveals that the spin density is equally delocalized over all four nitrogen atoms in these D 2d -symmetric spirocyclic systems. Oxidation of the dimeric dilithio complex {Li 2 [PhB(  4 -N t Bu) 2 ]} 2 with iodine produces the monomeric neutral radical {[PhB(  -N t Bu) 2 ]Li(OEt 2 ) x }  ( 17 ), characterized by EPR spectra and DFT calculations. These findings establish that the bam anionic radical [PhB(N t Bu) 2 ]  can be stabilized by coordination to a variety of early main-group metal centers to give neutral radicals whose relative stabilities are compared and discussed.


Introduction
The coordination chemistry of the monoanionic amidinate (am) ligand [RC(NR')2]  (Chart 1) has been extensively studied. 1  Until 1993, reports on p-block bam derivatives were restricted to Groups 1416, 11a,11c,11e,13 with the exception of {[-Mes*B(NMe)2 2 N,N']AlMe}2 for which a crystal structure has not been reported. 13i Early investigations of the coordination chemistry of the bam dianion with transition metals were limited to Group 4,11a,14 but recently, Nocera and co-workers have described complexes of Groups 5 and 6 including octahedral tris-bam complexes and a paramagnetic vanadium(IV) complex. 15 16,17 and showed that the oxidation of 4a and 4b with iodine produces stable 18 neutral radicals {[PhB(-N t Bu)2]2M}  (5a, M = Al; 5b, M = Ga), which were characterized in the solid state by X-ray crystallography.
A combination of EPR spectroscopic analyses and DFT calculations showed that spin delocalization is uniform over both bam ligands in 5a and 5b and that the spirocyclic structure is retained in solution (Scheme 1). 17 Scheme 1.

Results and Discussion
Syntheses and Spectroscopic Characterization of Mg and Zn bams. The reaction of Li2[PhB(3-N t Bu)2] (1a) and a Grignard reagent in a 1:1 stoichiometry was attempted in an effort to generate the mixed-metal Mg/Li mono-bam 7 (Scheme 2, Method 1). However, the reaction with t BuMgCl affords the spirocyclic dilithio magnesium bis-bam complex {(Et2O)-Li[PhB(-N t Bu)2]}2Mg (6a, Scheme 2). The proposed pathway involves the initial production of 7, which then undergoes redistribution to produce 6a (89 %) with the elimination of Mg t Bu2 that was detected by 1 H NMR spectroscopy (~ δ 0.2). The generality of Method 1 (Scheme 2) was demonstrated in the reactions of 1a with a variety of Grignard reagents in diethyl ether, including MesMgBr, n BuMgCl, and MeMgBr, all of which produce 6a. Similarly, the THF derivative {(THF)-Li[PhB(-N t Bu)2]}2Mg (6b) is obtained from the reactions of the Grignard reagents PhMgCl or i PrMgCl in THF and 1a in a 1:1 stoichiometry.

EPR and DFT Investigations of Early Main-Group Element and Zn bam
Radicals. In a preliminary communication, we described the one-electron oxidation of the spirocyclic anions 4a and 4b to give the stable neutral radicals 5a (dark red) and 5b (dark green) (Scheme 1). 17,18 In this work, we discuss the characterization of the indium and boron analogues, 5c and 5d, respectively.
The oxidation of 11c with one-half of an equivalent of iodine generates a dark green solution of complex 5c ( Figure 5) that, in contrast to 5a and 5b, is unstable at ambient temperature. Thus the oxidation was carried out at 78C, and the paramagnetic species formed were investigated by variable-temperature (VT) EPR spectroscopy. The bright purple solutions of complex 5d ( Figure 5) formed by air oxidation of the products of the reaction of 1a with BF3·OEt2 in a 2:1 molar ratio (vide supra) were also investigated by EPR spectroscopy. Whereas the intense dark green color of 5c lasts only for minutes at room temperature, 27 diethyl ether solutions of radical 5d remain dark purple and exhibit similar EPR spectra even after two years.  Figures 6b and 7b for 5c and 5d, respectively, were obtained by using the hyperfine coupling (hfc) constants given in     The experimental EPR spectrum obtained at room temperature for 16b is shown in Figure 10a; the spectrum for 16a exhibits a similar pattern owing to the low natural abundance of the spin-active isotopes for magnesium and zinc. Excellent simulations ( Figure 10b) were obtained by using the hfc constants given in Table 6. The experimental and calculated hfc values are in close agreement. Hence, for both radicals, the EPR spectra support the prediction from DFT calculations that the unpaired electron is localized on one bam ligand. The EPR spectra for the zinc radical 16b reveal minor interactions with the adjacent bam ligand, as well as some low-intensity satellites due to coupling of the unpaired electron to the 67 Zn nucleus.   The first indications of the formation of paramagnetic early main-group metal bam complexes came from investigations of lithium derivatives of the type 1. 11c It was observed that the initially colorless solutions of 1ac became red upon exposure to air. In preliminary EPR studies, only coupling to two equivalent 14 N nuclei (a five-line pattern) was resolved. In the absence of resolved coupling to 7 Li centers, it was not possible to speculate on the state of aggregation of the radical species. 11c We have now carried out the oxidation of 1a with one-half of an equivalent of iodine in diethyl ether at room temperature (Scheme 8). The bright pink species so formed is unstable at room temperature. Consequently, the solutions used for VT EPR studies were prepared at 78 C. The EPR spectrum obtained at 40 C is shown in Figure 11a. This spectrum is best simulated (see Table 7) with hyperfine interactions of the unpaired electron with one lithium atom ( 7 Li, I = 3 /2, 92.41 %; 6   The plausibility of the assignment of the experimental EPR spectrum to the monocyclic radical was supported by DFT calculations. Figure   For the spirocyclic systems, the higher stability of the B, Al, and Ga-containing radicals compared to the Mg and Zn systems can be attributed to the delocalization of the unpaired electron over both bam ligands in the Group 13 complexes. The high stability of the boron-centered radical 5d in solution is noteworthy. In contrast to 5a and 5b, however, it has not yet been possible to devise a synthetic route that will allow the isolation of 5d in the solid state. The lower stability of 5c may result from a combination of stronger intermolecular, i.e., radical-radical, interactions as a result of the larger In center and weaker In-N bonds compared to 5a, 5b, and 5d. Intermolecular association to form larger aggregates is likely the reason for the low stability of the mononuclear Li  16 were prepared by the literature methods.
Solvents were dried with appropriate drying agents and distilled onto molecular sieves before use. All reactions and the manipulation of moisture-and/or air-sensitive reagents or products were carried out under an atmosphere of argon or under vacuum using standard Schlenk techniques or a glovebox. All glassware was carefully dried prior to use.

Preparation of {(THF)-Li[PhB(-N t Bu)2]}2Mg (6b). A solution of PhMgCl in
THF (2.0 M, 0.33 mL, 0.66 mmol) was added to a solution of 1a (0.16 g, 0.66 mmol) in toluene (25 mL) at room temperature producing a cloudy pale yellow mixture that was stirred for 18 h. The reaction mixture was filtered and the solvent was removed in vacuo.
The residue was redissolved in THF and concentration of the solution followed by storage at 15 C for 3 d afforded colorless crystals of 6b (0.15 g, 0.23 mmol, 71 %).

Preparation of {(THF)-Li[PhB(-N t Bu)2]}2Zn (6c).
A solution of 1a (0.27 g, 1.12 mmol) in Et2O (100 mL) was added to solid ZnCl2 (0.08 g, 0.56 mmol) at 78 C producing a dark burgundy reaction mixture, which was allowed to warm to room temperature over a period of 20 min then set to reflux for 18 h. The mixture was filtered and the solvent removed in vacuo. The residue was washed several times with and cold n-hexane to give white amorphous 6c (0.20 g, 0.29 mmol, 51 %). Anal. Calcd for C36H66B2Li2N4O2Zn: C,62.86;H,9.67;N,8.15. Found: C,62.28;H,9.98;N,8.12
The reaction mixture was filtered to give a pale yellow filtrate from which the solvent was removed in vacuo. The residue was redissolved in Et2O and concentrated to afford 9a as colorless thin needle crystals (0.21 g, 0.41 mmol, 37 % cooled to 78 C producing a bright purple reaction mixture. After 15 min., the stirred reaction mixture was allowed to reach room temperature affording a clear colorless solution. After an additional 15 min. at room temperature, a cloudy pale yellow reaction mixture was produced that was stirred for 8 h. The resulting mixture was filtered to remove LiCl. Removal of solvent in vacuo and addition of cold n-hexane gave a pale yellow precipitate of 12 (0.41 g, 1.11 mmol, 60 %) that was washed twice with cold nhexane. Anal. Calcd for C14H23AlClBN2: C,58.95;H,9.07;N,7.64. Found: C,58.42;H,9.25;N,7.72 This solution was kept at 80 C prior to recording EPR spectra.

Computational Details
All calculations were done with the Gaussian 03 program. 29 The structures of all compounds were optimized in their ground states using density functional theory. Hybrid PBE0 exchange-correlation functional 30 and Ahlrichs' TZVP basis sets 31