Expanding the Family of Octahedral Chiral-at-Metal Cobalt(III) Catalysts by Introducing Tertiary Amine Moiety into the Ligand

: Chiral metal-templated complexes are attractive catalysts for organic synthetic transformations. Herein, we introduce a novel chiral cobalt(III)-templated complex based on chiral trans -3,4-diamino-1-benzylpyrrolidine and 3,5-di- tert -butyl-salicylaldehyde which features both hydrogen bond donor and Brønsted base functionalities. The obtained complexes were fully characterized by 1 H, 13 C NMR, IR-, UV-vis, CD-spectroscopy and by a single X-ray diffraction analysis. It was shown that chlorine anion is connected with amino groups of the complex via a hydrogen bonding. DFT calculations of charges and molecular electrostatic potential of the cobalt(III) complex showed that the basicity of the complex is certainly diminished as compared with the routine tertiary amines but the acidity of the conjugated acid of the complex should be increased. Thus, the catalytic potential of the complex may be much greater as a chiral acid than a chiral base. We believe that this work opens a new way in chiral bifunctional catalyst design.

The pioneering work of Gladysz et al. was the first example of such catalysis based on chiral Werner salts [4], although with a moderate asymmetric efficiency [5]. Later, a novel generation of such catalysts proved to be much more efficient [17]. Particularly striking was the efficiency of a chiral iridium complex developed by the Meggers group in asymmetric conjugate reduction of nitroalkenes with high enantioselectivity at parts-permillion loadings [19].
The formation of the complexes occurred stereospecifically with only one diastereoisomer, Λ(R,R)-1 or ∆(S,S)-1, formed from the corresponding diamines. The NH 2 groups of the ligands were coordinated to the metal ion and their acidities and their hydrogen bonding donor (HBD) abilities were greatly boosted by the coordination. Expectedly, complexes 1 (Figure 1a) proved to be efficient and recyclable hydrogen bond donor (HBD) catalysts for several important asymmetric transformations [6,8,10,13,18,26,28]. The transformation included trimethylsilylcyanation of benzaldehyde [6] and asymmetric alkyl halide alkylation of glycine derivatives [8,28], leading to a general protocol for enantiomerically enriched amino acid synthesis. In addition, Michael addition to the same glycine derivative [10] allowed the preparation of a series of enantiomerically enriched glutamic acid derivatives. For example, the activation of O'Donnell's substrate and enantiocontrol by the complex 1 via multiple hydrogen bonding interactions is depicted in Figure 1c [10]. The epoxidation of chalcones under phase-transfer conditions [13], as well as the synthesis of cyclic carbonates from epoxides and carbon dioxide [13,18] were other examples of the performance of the catalytic family. The catalysts were even able to promote the redox reaction of naphthol couplings [26]. (c) The optimized structure of the ion pair formed between the Z-configured enolate of O'Donnell's substrate and the cobalt(III) catalyst 1 (Reprinted with permission from Ref. [10]. Copyright © 2014 Wiley-VCH Verlag GmbH and Co., KGaA).
We believed that the introduction of other basic or acidic groups into the ligand skeleton of the complex 1 would open new vistas of their catalytic performance.
Herein, we report the synthesis of the first-generation of the family exemplified by the chiral cobalt(III)-complex 2 ( Figure 1b). Its ligand was prepared from (R,R)-and (S,S)trans-3,4-diamino-1-benzylpyrrolidine and 3,5-di-tert-butyl-salicylaldehyde. The formed complex, in addition to the HBD capable NH 2 groups, features a tertiary amine moiety-Brønsted base functionality ( Figure 1b) and simultaneously potentially conjugate acidic site. In addition, we describe the single crystal X-ray structures of Λ(R,R)-2 and ∆(S,S)-2 and DFT calculations of the basicity (conjugate acidity) of the tertiary amine moieties of the complexes.

Synthesis of the Complexes 2
The chiral cobalt(III)-templated complex 2 was prepared similarly to the earlier described [8] general procedure (Scheme 1). Accordingly, Λ(R,R)-2 was obtained in a 13% yield. The complex was isolated by a flash chromatography on SiO 2 (Scheme 1). Simi-larly, a complex ∆(S,S)-2 was prepared starting from (S,S)-3,4-diaminpyrrolidine. Notably, both diastereoisomers of 2 were prepared stereoselectively without any amount of the alternative isomers formed in the reaction.

Characterization of the Complexes 2
Complexes Λ(R,R)-2 and ∆(S,S)-2 were characterized by 1 H, 13 C NMR, IR-, UV-vis, CD-spectroscopy and elemental analysis. Besides, the structures of 2 were unambiguously determined by the single crystal X-ray diffraction analysis ( Figure 2). Unfortunately, we failed to obtain suitable single crystals of both diastereoisomerically pure Λ(R,R)-2 and ∆(S,S)-2. Still, a 1/1 mixture of the diastereoisomers could be crystallized from a solvent system-MeCN/hexanes. The mixture of 2 crystallized into PĪ space group and the crystal data was collected at 120 K. Since one tert-butyl group (attached to a salicylaldehyde) was disordered over two positions, the refined occupancies were 0.54/0.46. Two tridentate NNO ligands are coordinated to cobalt ion forming a regular octahedral geometry. The Λor ∆-configurations were attributed according to left or right helical mutual orientations of their tridentate ligands relative to the C 2 symmetry axis [49][50][51]. The three positive charges of cobalt(III) central ion were compensated by the two ionized phenolic oxygen atoms and the halide atom. Atoms O1 and N2 lie in axial position due to their longer distances (3.882 Å) in comparison with O2···N5 (3.875 Å) and N1···N4 (3.807 Å). NH 2 groups are hydrogen bonded to the chlorine anion as indicated by the distances of N2-H···Cl1 and N5-H···Cl1 equal to 2.336 Å and 2.384 Å correspondingly, which is shorter than the same type of bonds in the analogous complexes derived from 1,2-diaminocyclohexane (2.448 Å) [6,8,52]. Besides, the chlorine anion coordinated to the third NH 2 group of the second molecule of the complex in the unit cell, and this bond is a little bit longer (2.446 Å). The hydrogen bonding between the amino group and the chlorine anion was also found in CH 2 Cl 2 solution which was testified by the IR spectra where the N-H···Cl stretching vibrations were observed in the region of 3200-3100 cm −1 (see Figure S3  . The UV-vis spectrum of the cobalt complex Λ(R,R)-2 was recorded in CH 2 Cl 2 (see Figure S4 in the Supplementary Materials). The spectrum showed an intense band at 408 nm, which could be associated with the ligand-to-metal charge transfer transition, and confirmed the coordination of the imine and amine nitrogen atoms to the central Co(III) atom. A less intense band at 500 and 700 nm could be assigned to d→d transitions (see Figure S4 in the Supplementary Materials).

Catalytic Activity of the Complex 2
The catalytic properties of the Co(III) complexes 2 could be expected to come from the hydrogen bonding of the NH groups and the tertiary amino group of the ligand skeletons. The group in its turn could be basic enough to affect the ionization of some CH acids or function as acidic site in case of the group protonation.
We evaluated the catalytic performance of the complex 2 in benchmark asymmetric reactions [13] (Scheme 2, for details see Supplementary Materials). Although the complex 2 could catalyze the epoxidation of chalcone (Scheme 2a) and kinetic resolution of racemic chalcone epoxide with CO2 (Scheme 2b), its activity was poor in terms of both enantioselectivity of the products and the conversion of the starting material as compared with the performance of 1 [13]. Next, we tested 2 in Michael addition reaction of ethyl

Catalytic Activity of the Complex 2
The catalytic properties of the Co(III) complexes 2 could be expected to come from the hydrogen bonding of the NH groups and the tertiary amino group of the ligand skeletons. The group in its turn could be basic enough to affect the ionization of some CH acids or function as acidic site in case of the group protonation.
We evaluated the catalytic performance of the complex 2 in benchmark asymmetric reactions [13] (Scheme 2, for details see Supplementary Materials). Although the complex 2 could catalyze the epoxidation of chalcone (Scheme 2a) and kinetic resolution of racemic chalcone epoxide with CO 2 (Scheme 2b), its activity was poor in terms of both enantioselectivity of the products and the conversion of the starting material as compared with the performance of 1 [13]. Next, we tested 2 in Michael addition reaction of ethyl malonate to trans-β-nitrostyrene. Unfortunately, the preliminary results showed that the Brønsted basicity of the complex 2 was not enough to catalyze this reaction, even under heating (see Supplementary Materials).

DFT Calculations of Charges and Molecular Electrostatic Potential of the Complex 2
In order to access the possible basicity of the tertiary amino groups of the ligands, we performed the DFT calculations of the parameters. For this purpose, we have optimized the cation using PBE0/def2-TZVP level of theory and calculated charges and molecular electrostatic potential (MEP) on its surface.
Whereas the NPA charges [53] that are often used to describe reactivity are closely similar for the tertiary amine nitrogen atoms N3 and N6 in 2 (−0.40 e) and in the triethylamine molecule (−0.42 e), the CHELPG charges [54] that are fitted to molecular electrostatic potential (−0.31 and −0.26 for N3 and N6) are significantly smaller by magnitude than in the case of triethylamine (−0.66). The distribution of the MEP on the molecular surface ( Figure 4) shows that the regions close to the N atoms are either characterized by significantly positive electrostatic potential that prevents the protonation of the side chain amino groups or are strictly inaccessible. Thus, the basicity of the complex is certainly diminished as compared with the routine tertiary amines but the acidity of the conjugated acid of the complex should be increased. Thus, the catalytic potential of the complex 2 may be much greater as a chiral acid than a chiral base.

Materials and Methods
Proton nuclear magnetic resonance ( 1 H NMR) spectra and carbon nuclear magnetic resonance ( 13 C NMR) spectra were recorded on a Bruker Avance III-400 NMR spectrometer (BRUKER DALTONIK GmbH, Bremen, Germany) (operating at 400/101 MHz respectively, referring to 1 H/ 13 C nucleus). Chemical shifts are reported in ppm relative to the residual solvent peak (CDCl 3 : δ = 7.26 ppm for 1 H NMR, δ = 77.2 for 13 C NMR). NMR data are reported as follows: chemical shift, multiplicity, coupling constant, integration, and nucleus. X-ray diffraction data were collected on a Bruker SMART APEX CCD diffractometer (BRUKER DALTONIK GmbH, Bremen, Germany) (λ(MoKα) = 0.71072 Å, ω-scans, 2θ < 58 • ). Optical rotations were measured on a Perkin-Elmer 341 polarimeter (Perkin-Elmer, Waltham, MA, USA) in a 10 cm cell. UV-vis spectrum was measured in CH 2 Cl 2 on Cary50 spectrometer (Varian Medical Systems, Palo Alto, MO, USA) using CaF 2 cells. IR spectrum of the compound was measured in CH 2 Cl 2 on a Shimadzu IR Prestige 21 FT-IR equipped with MCT detector (Shimadzu, Kyoto, Japan) in CaF 2 cells. Circular dichroism (CD) curves were recorded on a Universal SKD-2MUF (Institute of Spectroscopy RAS, Troitsk, Russia). Elemental analyses were carried out in the Laboratory of Microanalysis of the A.N. Nesmeyanov Institute of Organoelement Compounds of the Russian Academy of Sciences (INEOS RAS).

Synthesis of the Complexes 2
(R,R)-and (S,S)-3,4-diamino-1-benzylpyrrolidine (486 mg, 2.5 mmol, 1 equiv.) and NH 4 Cl (133 mg, 2.5 mmol, 1 equiv.) were dissolved in dry MeOH (20 mL) by stirring during 5 min, and the solvent was removed on a rotary evaporator. The resulting salt was dissolved in EtOH (30 mL) and 3,5-di-tert-butylsalicylaldehyde (525 mg, 2.5 mmol) was added to the residue, and after 5 min stirring, the reaction solution became a deep yellow. Then, Na 3 [Co(CO 3 ) 3 ]·3H 2 O (460 g, 1.3 mmol, 0.6 equiv.) was added to this solution. The resulting dark-brown reaction mixture was heated at reflux for 4 h after which time the mixture was allowed to cool to room temperature. After removal of the solvent on a rotary evaporator, the resulting precipitate was dissolved in CH 2 Cl 2 (20 mL), and the aqueous solution of sodium chloride (5 g in 20 mL) was added. After 1 h stirring, the organic layer was separated, and the aqueous layer was washed with CH 2 Cl 2 (3 × 15 mL). The combined organic layers were dried over Na 2 SO 4 , the solvent was evaporated on a rotary evaporator, and the resulting residue was purified by silica column chromatography (eluent: CHCl 3 /acetone 5:1 → CHCl 3 /acetone 1:1 → MeOH) to separate from impurities, and further purified on LH-20 (eluent: benzene/EtOH 3:1) and, then the complexes Λ(R,R)-2 or ∆(S,S)-2 were isolated as brown powders (150 mg, 0.17 mmol, 13% yield).

X-ray Diffraction Study
Single crystal X-ray study of the complex 2 was carried out in Center for Molecule Composition Studies of INEOS RAS. Dark-brown single crystals of the cobalt(III) complexes 2 were obtained by slow evaporation from acetonitrile/hexane mixture. The structures were solved by direct method and refined in anisotropic approximation for non-hydrogen atoms. Hydrogen atoms of methyl, methylene and aromatic fragments were calculated according to those idealized geometry and refined with constraints applied to C-H and N-H bond lengths and equivalent displacement parameters (U eq (H) = 1.2U eq (X), X-central atom of XH 2 group; U eq (H) = 1.5U eq (Y), Y-central atom of YH 3 group. Analysis of residual electron density and displacement parameters in 2 has shown that tert-butyl group is disordered. All structures were solved with the ShelXT program [56] and refined with the ShelXL program [57]. Molecular graphics was drawn using OLEX2 program [58]. Experimental details and crystal parameters are listed in Table S1 (see Supplementary Materials). CCDC 2050585 contains the supplementary crystallographic data for complex 2. These data can be obtained free of charge from The Cambridge Crystallographic Data Centre.

DFT Calculations
Geometry optimization of the cation was performed with ORCA [59] software package using the density functional theory. The combination of PBE0-D3 [60,61] functional and def2-TZVP [62] basis set was used. NPA charges were calculated with Janpa [63] program; the calculation of CHELPG charges and MEP was performed using Multiwfn program [64]. The molecular surface for MEP calculation was defined by an electron density isosurface with a value of 0.001 au.

Conclusions
In conclusion, here we reported a novel chiral cobalt(III)-templated complex featuring tertiary amine moiety (Brønsted base functionality) which expands the family of octahedral chiral-at-metal cobalt(III) complexes for applications in asymmetric catalysis. Compared to the previously reported chiral cobalt(III) hydrogen bond donor catalysts, the tertiary amino groups in the ligand sphere of the complex can be basic enough (Brønsted base center) to affect the ionization of some CH acids or function as an acidic site in case of the group protonation making the complex 2 a bifunctional asymmetric catalyst. The DFT calculations of the parameter showed that the basicity of the complex is certainly diminished as compared with the routine tertiary amines but the acidity of the conjugated acid of the complex should be increased. Thus, the catalytic potential of the complex 2 may be much greater as a chiral acid than a chiral base. We believe this work will open a new chapter in the field of chiral catalyst design. The application of the new chiral cobalt(III)-templated complex in challenging asymmetric transformations are underway in our laboratory.