Stereochemistry of Optically Active Transition Metal Compounds

Stereochemistry, long the backbone of coordination chemistry, is still a major area for investigation and discovery. The methods of studying it have changed, and perhaps because of that, the emphasis on the topics under investigation has changed. Synthetic chemistry has been displaced by, or better, augmented by, physical methods. Both approaches are important and both contribute to the growth of our knowledge. It is most appropriate that we should hold a symposium on the stereochemistry of complexes and compare notes on what is being done in the countries which are represented at this meeting. In introducing the symposium, I should like to recount some developments of the nearly fifty years that I have been involved in this field. Most of the work which I shall describe was done by my students -- not that I think that their work is better than that of others, but because I know it

In this paper an overview will be given on the structural studies of optically active transition metal complexes and their interaction with other fields in coordination chemistry. A quarter of a century has passed away since the first determination of the absolute configuration of a transition metal complex, [Co(en)₃]³⁺, was carried out by anomalous scattering of X-rays (1). Since that time the number of transition metal complexes whose absolute configuration have been determined by X-ray method has been growing at an increasing rate and at this time it has exceeded 130. In addition to this, numerous optically active organometallic compounds have been studied and their absolute configurations established.

At an early stage of the development, the structural information was rather fragmentary. Nowadays, however, the accumulation of structural data for isomers has enabled us to understand structural principles and the optical properties of chelate complexes in considerable detail. In this connection,

The chiroptical properties of optically active transition metal complexes have played an enormously influential role in the stereochemical and electronic structural characterization of metal coordination compounds. Werner's early work (1) in resolving the optical isomers of bis- and tris-chelated transition metal complexes containing achiral ligands established the octahedral structure of hexa-coordinated complexes and posed the problems of molecular stereochemistry and absolute configuration of metal coordination compounds. In the 1930's, the optical rotatory properties of Werner's complexes were studied extensively by Jaeger (2), Mathieu (3), and Kuhn (4, 5, 6) with the objective of relating these properties to specific stereochemical features (most notably, the absolute configuration) of the systems. Attempts were also made to construct a theory of optical activity in transition metal complexes which would permit systematic correlation of the observed optical rotatory (and circular dichroic) properties with absolute configuration (4,5). These latter attempts employed a classical representation of the

The problem of distinguishing Tweedledum from Tweedledee is not peculiar to Alicœ(1). Determination of absolute configurations of chiral metal complexes, apart from being of basic chemical significance, has proved an intriguing challenge to a range of techniques. Circular dichroism in particular, because of the ease of measurement, has enjoyed considerable popularity as a stereochemical probe, albeit with incommensurate success. The major difficulty lies in the interpretation of the CD spectrum in terms of some theoretical model of the complex, so that the reliability of the technique is determined largely by the validity or relevance of the theoretical model.

Empirical correlation rules and a complete molecular orbital approach may be considered to constitute the two extremes of the model approach. The many empirical rules (see, e.g. Discussions and references in Hawkins' book (2)) cannot be considered modelindependent in that they make the implicit assumption that the complexes to which they apply

The relationship between the absolute configuration and CD spectrum has been widely discussed, but mostly for octahedral complexes. (1) Asymmetric coordination of η²-olefins was first demonstrated by Cope et al. (2) with trans-[PtCl₂(l-phenylethylamine)(trans-cyclo-octene)] and extended by Paiaro and Panunzi (3) by the preparation of a pair of diastereoisomers, such as transdichloro(R or S-α-phenylethylamine) (trans-2-butene)platinum(II) , trans-[PtCl₂(l-phenylethylamine) (tbn)]. Since then several complexes with such an asymmetry have been prepared, and the relationship between the absolute configuration and CD pattern has been discussed for platinum(II) complexes. (4) Scott and Wrixon (5) reported that S,S-η² and R,R-η² configuration give CD peaks with positive and negative signs in the d-d transition region at ca. 27,000 cm^-l. Less information is available for the complexes with other metal ions, and only palladium(II) (5) and iron(0) (6) complexes were discussed.

A change in the kind of olefin does not cause significant changes in absorption spectra, so long

Over the past twenty years the conformations and steric effects in chelate ring systems have been examined in considerable detail. Structural studies along with equilibrium measurements, some kinetic studies and conformational analyses have given us a better insight into the steric interactions within the chelates and between the chelates (1). For five membered chelates such as coordinated 1,2 diaminopropane the axial-equatorial nature of the substituents on the C atoms is sufficiently pronounced(2) that the axial methyl groups interact with axial substituents on the metal ion and this conformation is rarely if ever observed (1,2). In this way the conformation of the chelate is controlled as λ for Co(R)(-)pn (1,2). Tris chelate systems of this type show a marked preference for the isomer where the C-C axes of the individual chelates are near parallel to the C₃ axis of the complex ion, e.g. Δ[Co(R) (-)pn₃]³⁺ (1e1₃) is ~15 fold more stable

Siderophores are naturally occurring chelate compounds synthesized by microorganisms and used to sequester the biologically essential but sparingly soluble ferric ion from their solution environment (1, 2, 3). The close relationship between microbial pathogenicity and iron metabolism is now well established (4, 5, 6). For example, in salmonellosis (7), infantile enteritis by E. coli (8), and mycobacterial infections such as tuberculosis and leprosy (4) a clear-cut effect of iron-binding on pathogenicity has been shown.

The low-molecular-weight siderophores are very powerful chelating agents which typically use hydroxamate, thiohydroxamate, or catecholate groups to encapsulate iron(III) ion in an octahedral, high-spin complex (see Figure 1). These bidentate chelating groups are usually linked to a linear or cyclic peptide of three to six amino acids. The transport of siderophores across cell membranes evidently can be quite conformation dependent, since in some cases the ferric complexes are rapidly passed into the cell while metal-free siderophores

Asymmetric catalysis has great potential for synthetic organic chemistry, since optically active molecules may be prepared without the need for resolution. The hydrogenation of prochiral olefins is particularly attractive, since in principle many diverse types of tertiary asymmetric center may be introduced in a single step. Following the discovery of efficient rhodium catalysts for homogeneous hydrogenation by Wilkinson and co-workers, (1),complexes derived from phosphines of the type R₁R₂R₃P* were prepared and evaluated. The results were invariably disappointing, and optical yields of hydrogenated products rather poor (2). Two developments led to considerable improvement. First, the use of dehydroamino acids led to efficient and stereoselective reduction (3), and secondly chelating chiral phosphines (4) were found to be effective ligands. Progress between 1971 and 1977 has been reviewed (5,6,7) and now optical yields of more than 90% in the hydrogenation of dehydroamino acids or their derivatives are commonplace.

There have been considerable developments

Circular dichroism (CD) has played an important role in our studies on the modification of enzymes and hormones with Co(III). The objective of these studies has been to incorporate selectively substitution inert metal ions at specifically modified sites in proteins as probes of biological function. Significant information concerning the catalytic mechanism of carboxypeptidase A (CPA) (1) has been obtained from a site specific modification of tyrosine 248 with Co(III) (2). The method developed for CPA has been extended to other enzymes and hormones in order to develop an improved method for incorporating stable radioisotopes (⁵⁷Co) into proteins. The substitution-inertness of Co(III) provides the necessary stability in these derivatives (3).

The method involves conversion of tyrosines and histidines on the protein of interest into chelating agents by reaction with a diazonium salt followed by incorporation of cobalt as shown in Figure 1. It has been found that specific incorporation occurs if

Since Wymore and Bailar (1) first prepared trans-[CoX₂{1,2-bis(diethylphosphino)ethane}₂]X (X = Cl^-, Br^-, and I^-) in 1960, many cobalt(III)-phosphine complexes have been reported (2). However, no optically active cobalt(III)-phosphine complex has ever been prepared. This paper deals with the preparation, characterization and circular dichroism (CD) spectra of octahedral cobalt(III) complexes containing aminophosphines of the type, NH₂CH(R)CH₂PR′R″. The aminophosphine is an intermediate ligand between a diamine and a diphosphine. The optical activity of such aminophosphine complexes can be compared with that of diamine complexes studied extensively.

Preparation of Ligands

a) NH₂CH₂CH₂P(C₆H₅)₂ (N-C-C-P). This ligand was prepared by a modified method of Issleib et al. (3).

b) (S)-NH₂CH(CH₃)CH₂P(C₆H₅)₂ (N-C*-C-P). This ligand was prepared from (S)-alanine by the following method;

NH₂CH(CH₃)COOH LiAlH₄→THF NH₂CH(CH₃)CH₂OH HCl, SOCl₂→CHCl₃ NH₂CH(CH₃)CH₂Cl·HCl NaP(C₆H₅)₂→liq. NH₃ N-C*-C-P

c) (S)- or (R)-NH₂CH₂CH₂P(n-C₄H₉) (C₆H₅) (N-C-C-P*). This ligand was prepared from N-C-C-P by the following method;

N-C-C-P Na→liq. NH₃n-C₄H₉Cl→

Since before the turn of the century it has been known that the optical activity of some chiral compounds is solvent dependent (1). For example, in 1877 Landolt (2) reported that the specific rotations of (+)-camphor, (-)-nicotine, (+)-diethyl tartrate, and (-)-turpentine varied with solvent and concentration. In the last decade there has been renewed interest in this solvent dependence. A number of different types of organic compounds has been investigated and the results have been interpreted in terms of variations in conformer populations that have resulted from either the effect of the dielectric on coulombic interactions between dipolar groups in the molecule (3) , or from hydrogen-bond interactions between the solvent and the chiral solute (4).

Recently papers have appeared dealing specifically with the effects of solvents on the CD of metal complexes (5,6). Work in this field in our laboratory has been directed at elucidating the particular solvent-solute interactions

The Pfeiffer Effect (1) is defined as the change in optical rotation of an optically active system (usually a solution of one enantiomer of an optically active compound, called the "environment substance", dissolved in an optically inactive solvent) upon the addition of a racemic mixture of a dissymmetric, optically labile coordination compound. Much work has been done on this Effect (2 - 8) and several mechanisms have been proposed to explain it, which are described in a review by Schipper (9). It is of interest to note that the Effect can occur with racemic mixtures of certain optically labile complex cations (e.g., D.L-[Zn(o-phen)₃]²⁺) whether the environment substance is anionic (d- α -bromo-camphor-π-sulfonate), neutral (levo-nicotine), or cationic (d-cinchoninium), The most frequently used solvent for the Pfeiffer Effect is water (10), although the Effect is known to occur in other solvents as well (3,4,6).

Since the magnitude of the Effect is proportional

Additivity of Circular Dichroism Contributions

The major contribution to the rotational strength of optically active complexes of transition metals is usually the chiral arrangement of chelate rings. The additivity of the contributions to the rotational strength was demonstrated (1) for complexes of the type [Co(en)₂(aa)]²⁺ (aa = amino acid anion). The Δ- and Δ- isomers of [Co(en)₂(S-pala)]²⁺ are diastereoisomers, not enantiomers. The CD curves (Figure 1) are not mirror images. The sum of the 2 CD curves is the same as the CD curve for the unresolved complex, an active racemate. The contributions for the Δ and Δ arrangements of the chelate rings should cancel for this "vicinal effect" curve, giving twice the contribution of the coordinated S-phenylalaninate ligand. Subtraction of one-half of this CD curve from the curve for either Δ- or Δ- [(Co(en)₂-(S-pala)]²⁺ gives a curve which agrees well with the CD curve of one of the isomers

Sources of dissymmetry in optically active metal complexes can be classified as: (a) inherent dissymmetry within the metaldonor atom coordination cluster, (b) configurational dissymmetry due to the chirality of the chelate system, (c) conformational dissymmetry due to the individual chelate ring conformations, and (d) vicinal dissymmetry due to asymmetric sites upon the ligands (1). For many complexes this set of contributions to the CD spectrum can be reduced to the configurational effect, the conformational effect and the vicinal effect (2,3). The most pragmatic approach to follow generally is to assume that the CD spectrum of a metal chelate complex is simply a summation of a vicinal effect and a configurational effect where the term vicinal effect retains its meaning as before and the other contributions are included in the conf igurational term (4,5,6,7).

The independent systems/perturbation model, as carried to second order in perturbation theory in Schipper's AICD (associate-induced circular

Most CD spectral studies of cobalt(III) complexes have been undertaken to investigate various sources of optical activity such as distribution of chelate rings, conformation of chelate rings, vicinal effect due to asymmetric carbon in an optically active ligand, and vicinal effect due to an asymmetric donor atom. Extensive reviews on these subjects have been written by Fujita and Shimura (1), Hawkins (2) , and Mason (3).

In contrast to those investigations, much less work has been reported on complexes whose optical activity results from a certain arrangement of unidentate ligands. The reason for the delay in work is considered to be the lack of useful syntheses for such chiral complexes.

The cis-cis-cis-isomer of the [Coa₂b₂c₂]-type, where a, b and c denote unidentate ligands, is asymmetric and exhibits two enantiometric forms I and II in Figure 1. The cis-cis-isomer of the [Coa₂b₂CC]-type, III in Figure 1, also belongs to the same

It is not easy to obtain both enantiomers of non-charged chiral complexes in optically pure form. Chromatographic separation by the use of asymmetric adsorption on an optically active adsorbent(for example, quartz, starch and d-lactose etc.) has been regarded as almost the only means for this purpose. However, usually such procedures lead to partial separation of enantiomers. This is probably because asymmetric adsorption is every weak. To make matters worse, we do not know the stereochemistry of asymmetric adsorption, nor how to improve the efficiency of separation. We attempted to attack this problem by extending the technique and the ion association model for ion-exchange chromatography. Here, we describe the mechanism of chromatographic separation of enantiomers of facial and meridional tris(aminoacidato)cobalt(III) chelates (See Figure 1).

Complete Resolution of fac-[Co(β-ala)₃]

The facial isomer of [Co(β-ala)₃] has three carboxyl oxygen atoms in a triangular face of an octahedron and three amino groups in the

Curtis and coworkers first reported the amine-imine linkage formation of the type shown in Figure 1 by condensation of two molecules of acetone with ethylenediamines in the presence of nickel(II) or copper (II) ions (1, 2, 3). It was also shown that this type of reaction generally occurs with several carbonyl compounds (4, 5). Later, the mechanism of the linkage formation and the stereoisomerism and some reactions of the metal complexes with macrocycles were investigated in detail (5, 6, 7).

Further, it should be noted that the same macrocycle as in Figure 1 could be obtained by using condensation dimers of acetone, such as diacetone alcohol or mesityl oxide (8). It was also found that several kinds of α,β-unsaturated ketones and β-hydroxyketones may be used to prepare macrocyclic diaminodiimines (8). This observation is very important, because by using suitable ketones it becomes possible to introduce desired substituent(s) into the three

In recent years there has been considerable interest in the complexes of transition metals with α, β-unsaturated β-ketoamines (1), 1, especially in their stereochemical properties (2, 3, 4, 5). Because the β-ketoaminate ligand is unsymmetrical, metal tris(β-ketoaminates) may exist as fac and mer isomers. Previous studies of V(III) and Cr(III) tris(β-ketoaminates) (5, 6) have revealed the presence of only one stereoisomer, which can be assigned the mer-configuration on the basis of NMR studies of the paramagnetic V(III) complexes (5). Evidently the fac-isomer is destabilized by steric interactions between the three N-alkyl or N-aryl groups, which project, in the fac-isomer, from a single octahedral face (5, 6). A similar situation obtains for tris(N-R-salicylaldimine) and tris(N-R-pyrrole-2-aldimine) complexes; these complexes also exist exclusively as the mer-stereoisomer (5, 7).

When R = H, steric interactions should not be a problem and both geometric isomers should exist. Indeed, in some unpublished work about 15 years

The reaction of aqueous cyanide with a tris(diimine)iron(II) chelate

Fe(diimine)₃²⁺ + 2CN^- → Fe(diimine)₂(CN)₂ + diimine (1)

shows interesting stereochemical changes, especially when the diimine is 1,10-phenanthroline, phen. The reaction produces an optical inversion throughout the visible (charge transfer) and ultraviolet (π-π*) spectral regions (1) as shown in Figure 1.

The reaction of aqueous cyanide with such tris(diimine)iron-(II) chelates was originally of interest because of the apparent bimolecular nature of the reaction (2,3,4). Initial stereochemical interest in the cyanide reaction (1) was based on the observation (5,6,7) that stereochemical changes in substitution reactions of d⁶ octahedral complexes show an inverse correlation to ligand field spectrochemical splittings provided that both the reactant and the product(s) are spin-paired; i.e., inert to further stereochemical reactions which would mask the initial stereochemistry. The cobalt(III) Bailar inversion reactions (8,9,10) are the classical examples of such changes and occur only for cobalt(III) chelates with four

The advantages of using crystalline solids (particularly single crystals) for optical absorption and linear dichroism measurements are well known (1). Less widely appreciated perhaps are the advantages (and limitations) of extending this technique to the measurement of optical activity. Although many of the points below apply equally well to magnetically induced "optical activity", the primary emphasis of our work has been on the measurement of "natural" optical activity as a technique for studying conformation and absolute configuration. These data can serve as important bench marks since independent (X-ray) determination of absolute configuration and conformation is frequently possible (2). The relatively low dipole strengths of d-d and certain other low lying, highly forbidden electronic transitions makes solid state optical activity measurements particularly applicable to transition metal ion complexes.

For intrinsically chiral species that are inert enough to be resolved conventionally, the measurement of natural optical activity in crystals has the same

Abstract

A brief summary of the stereodescriptive terms cis, trans, fac, mer, endo, exo, etc., currently employed in the chemical literature is presented with a discussion of their limitations. The development of ligand index numbering of skeletal positions and its relationship to the IUPAC system of locant designators is outlined. The extension of the Cahn-Ingold-Prelog (CIP) priorities to tetracoordinate through nonacoordinate systems, introduced into the Indexes to CHEMICAL ABSTRACTS (CA), is given with a comparison to the ligand indexing systems. The stereochemical relationships of more complex coordination compounds are considered, including (a) mononuclear complexes with two or more chiral centers (relative and absolute stereochemical descriptors); (b) organic systems with other than tetrahedral atoms; (c) the stereochemical notation problems of hapto systems, and (d) polynuclear coordination compounds. These systems are presented with illustrative examples and some of the apparent advantages and shortcomings of the CIP priority designators and ligand index notations are discussed.