COMMUNICATIONS Angew. Chem. Int. Ed. 2002, 41, No. 13 ¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002 1433-7851/02/4113-2317 $ 20.00+.50/0 2317 Ion-Pair-Mediated Asymmetric Synthesis of a Configurationally Stable Mononuclear Tris(diimine)±Iron( ii ) Complex** David Monchaud, JonathanJ. Jodry, Didier Pomeranc, Vale ¬rie Heitz, Jean-Claude Chambron, Jean-Pierre Sauvage, and Je ¬ro √me Lacour* In memory of Andre¬Collet Chiral mononuclear divalent tris(diimine) complexes of first-row transition metals–keystones of coordination chem- istry–are notoriously known for their high chemical but low configurational stability. [1] The D and L enantiomers–right- and left-handed propellers, respectively–can be isolated in good enantiomeric purity by efficient resolution proced- ures. [1d, 2] However, once dissolved, these derivatives racemize rapidly when no other source of chiral information is present on the ligands [3] or in the reaction medium. [4] Recently, the synthesis of a bis(1,10-phenanthroline) ligand L that forms octahedral complexes with a well-defined axis was reported: simple treatment of [Fe(dmbp) 3 ][PF 6 ] 2 ([1][PF 6 ] 2 , 1.0equiv, dmbp  4,4'-dimethyl-2,2'-bipyridine) with L (1.0 equiv) in refluxing 1,2-dichloroethane (DCE) afforded [Fe(dmbp)(L)][PF 6 ] 2 ([2][PF 6 ] 2 ) in high yield and purity [Eq. (1); for b1 and b2, see Figure 1]. [5] Here we report on the unusual configurational stability of 2, which can be resolved by simple preparative thin-layer chromatography (TLC), and on its direct asymmetric synthesis by using TRISPHAT anions (see below) as noncovalent chiral auxil- iaries (diastereomeric ratio, d.r. > 20:1). The ease of synthesis of 2 and the rapidity of its formation led us to assume a high chemical stability for [2][PF 6 ] 2 . It was then debatable whether an improved chemical stability would also mean an increased configurational stability, so that the D and L enantiomers might be inert and separable from each other. Previously, the synthesis and resolution of the D 3 - symmetric tris(tetrachlorobenzenediolato)phosphate( v) an- ion (3), known as TRISPHAT, was reported. [6] In association with mononuclear ruthenium( ii ) or iron( ii )tris(diimine) com- plexes, it is an efficient NMR chiral-shift, resolving, and asymmetry-inducing agent. [7] It was therefore foreseen that anion 3 could behave as a NMR chiral-shift reagent for the structurally related complex 2 and possibly lead to its resolu- tion. Racemic complex 2 was stud- ied in combination with a TRISPHAT salt. In an NMR tube, [Et 4 N][D-3] [8] was added as a solid to a solution of [rac-2][PF 6 ] 2 in 10% [D 6 ]DMSO/CDCl 3 (Figure 1, spectra (a) and (b)). Efficient separation of the signals of 2 was achieved with small amounts of chiral-shift reagent (1.0 ± 2.5 equiv). Protons H(b1) and H(b2) [see Equation (1)] were most easily monitored, and a rather large difference in chemical shift (Dd max  0.15 ppm) was observed. A 1:1 ratio of D and L enantiomers could be measured by direct integration of the respective signals. This 1:1 ratio of the diastereomers [D-2][D-3] 2 and [L-2][D-3] 2 indicated a possible configurational stability of cation 2. Indeed, it was recently shown that anions 3 act as effective asymmetry inducers on [Fe(dmbp) 3 ] 2 (1); when associated with the labile cationic guest, they control its configuration with high diastereoselec- tivity (d.r. > 49:1 in CDCl 3 in favor of [D-1][D-3] 2 ). [7b] Two hypotheses could then explain the lack of asymmetric induction observed in the NMR titration experiment (Fig- ure 1, spectrum (b)): poor chiral recognition by 3 or high configurational stability of 2. We thus decided to attempt an ion-pair chromatographic resolution, as the physical separation of ion pairs [D-2][D-3] 2 and [L-2][D-3] 2 would prove the second hypothesis to be correct. [9] Under previously reported conditions, [7d, e] solutions of [cinchonidinium][D-3] (2.5 equiv) in acetone and of [rac- 2][PF 6 ] 2 in CH 2 Cl 2 were prepared, mixed, and adsorbed on silica gel plates. Development by elution with CH 2 Cl 2 showed a much reduced affinity of salts [2][D-3] 2 for silica gel, as they were retained to a much lesser extent than the PF 6  precursor (R f  0). [10] Two well-separated bands were obtained (R f  0.94 and 0.84 in analytical TLC), abraded from the glass [*] Prof. J. Lacour, D. Monchaud, J.J. Jodry De ¬partement de Chimie Organique Universite ¬ de Gene ¡ve quai Ernest Ansermet 30, 1211 Gene ¡ve 4 (Switzerland) Fax: ( 41)22-328-73-96 E-mail: jerome.lacour@chiorg.unige.ch D. Pomeranc, Dr. V. Heitz, Dr. J.-C. Chambron, [] Dr. J.-P. Sauvage Laboratoire de Chimie Organo-Mine ¬rale UMR CNRS n8 7513, Faculte ¬ de Chimie Institut Le Bel, Universite ¬ Louis Pasteur 4, rue Blaise-Pascal, 67070 Strasbourg-Cedex (France) [  ] Present address: Universite ¬ de Bourgogne Faculte ¬ des Sciences Gabriel, LIMSAG (UMR 5633) 6, Boulevard Gabriel, 21100 Dijon (France) [**] We are grateful for financial support of this work by the Swiss National Science Foundation, the Federal Office for Education and Science (COST D11), the Fondation de Famille Sandoz (JL), the EC, the CNRS, and the MENRT (DP). Supporting information for this article is available on the WWW under http://www.angewandte.org or from the author.