The classic work on the stereoclremistry of metal complexes first gave formulation to the geo- metric structures of coordination compounds at the turn of the century. The work revolved around the acido- amines of cobalt(III), [Co(NH,),(X)6-,](z-3'-, and those of Pt(II), [Pt(NHa),(X)4-,](z-2)-. These works depended heavily on studies on isomerism, both optical and geometrical, but also reflected the application of what physicochemical tools were availahle. The suh- ject has been reviewed excellently in a number of sources, of which the outstanding treatments are to be found in the books by Bailar (1) and Grinberg (8). Many recent observations are included in the review by Wil- kins and Williams (3). Although early studies were certainly elegant and decidedly thorough in many cases, a number of compelling questions escaped solution or even modest progress until quite recently.

Early techniques were adequate only in dealing with complexes that are kinetically stable. Thus, the im- portance of Co(II1) and Pt(I1) to inorganic chemistry during the early decades of this century derives more from the fact that compounds containing these ions are relatively inert in the kinetic sense rather than from any intrinsic significance. The two ions have been of equal importance for each has served to epitomize the relationships centering on a particular geometric array of coordinated groups. What of equally abundant complexes of less accommodating ions? Those proving to form complexes of similar kinetic stability have been shown to conform to the isomer patterns consistent with assignable geometries: Rh(III), Ir(III), Ru(II), Os(II), Pt(1V) are octahedral; Pd(II), Au(III), Rh(1) are planar. However, with rare exceptions, knowledge of the geometry of the remaining ions had been almost totally restricted to the solid state where definitive X-ray studies were possible. With regard to the metal carbonyls and certain other volatile species, electron diffraction and detailed infrared investigations have provided valid structural assignments. Paradoxi- cally, many of the most familiar metal ions have offered the greatest difficulty from the structural standpoint. For example, the relationship between structure, color, and magnetism, among such ions as Mn(II), Fe(II), Co(II), ;R;i(II), and Cu(II), remained contradiction- ridden and questionable until the development of modern crystal field (or ligand field) theory. To be sure, a few structures were proved early by definitive X-ray studies, and optical isomerism was observed

Daryle H. Burch The Ohio State University

Columbus

This article is based upon a paper presented as s, pert of the Symposium on Three-Dimensional Chemistry before the Division of Chemical Education at the 144th Meeting of the American Chemical Society, Los Angeles, California, April, 1963.

The Stereochemistry of

Complex Inorganic Compounds

among some six coordination complexes of divalent ions. Indeed, geometric isomerism was known in the case of planar Ni(I1). Yet as one viewed the overall chemistry of an ion such as Ni(I1) or Co(II), he was confronted with a bewildering array of possibilities with only moderate probability of finding a criterion adequate to settle the structure, short of complete X-ray work.

Applications of Modern Theory

The applications of crystal field theory, and its more sophisticated molecular orbital counterpart, may be illustrated with a very few examples (4-8). The simple idea of a crystal field stabilization energy (CFSE) (9), differing according to atomic number, number of d-electrons, and nature, number, and geometric arrangement of ligands, leads to the prediction that in those cases where the difference in CFSE is a de- termining factor, ions with 3 or 8 d-electrons would show a marked preference for octahedral coordination as compared to a tetrahedral arrangement of ligands. This is illustrated in Table 1 in the column marked A CFSE. The interpretation of the data is that tetra- hedral should be rare for da and d8 and most common for dl, dZ, d5, d6, rF, dlo. A most important consequence of this prediction was the realization that the many nickel(I1) complexes long believed to contain tetrahe- dral coordination spheres were most probably octs, hedral.

Table 1. CFSE and Relative Stabilities of Tetrahedral and Octahedral Structures

4 F S E ( D q ) - Config- uration Octahedron Tetrahedron A CFSE(Dq)

This has been amply demonstrated and it is a by- product of these deductions that it has been intriguing to discover and study true tetrahedral complexes of Ni(I1) (10). Perhaps the prime example of such species is KiC142-.

In a still more approximate manner, the modern theories clearly agree with the relatively frequent oc- currence of square planar structures among ions of

Volume 47, Number 2, February 7964 / 77

rF, da, and d9 configurations, as illustrated (after Pear- son (9)) in Table 2.

Table 2. CFSE and Relative Stabilities of Octahedral and Square Planar, Strong Field Complexes

-CFSE(Dq)- Config- Square uration Octahedron Plane ACFSEIDa)

Concurrent with the emergence of these predictions has been the development of more detailed interpreta- tions of magnetic moments and electronic spectra as useful, if not necessarily definitive, criteria for geometric structure. This is illustrated in Tables 3 and 4, where the usual magnetic and spectral features of planar, tetrahedral, and octahedral nickel(I1) complexes are compared. Under favorable conditions, the number, intensity, and positions of the d-d electronic absorption bands provide a distinction between these structures. The "strong field" configurations and experimental ranges of magnetic moments for the three classes of nickel(I1) complexes, as given in Table 4, often provides strong corroborate evidence as to stereochemistry. S i l a r summaries may be assembled from available data on other metal ions.

Table 3. Variation of Magnetic Moment of Nickel(ll) Comdexes with Geometric Structure

Octahedral Orbitally nondegenectte - 1 1 eo re,, = 2.9-3.2 Bohr Magnetom 11 1 ?1 t.0 - Examples: Ni(H20)eP+, Ni(NH&'+,

Ni(dipy)2+ Tetrhedral

Orbitally degenerate 1111 13

p,!, = 3.4-1.0 Bohr Magnetom IL li e Examples: NiCl,z-, NiBr,+, NiBr2. -

WCeHsh Square planar - dz--">

Spin paired 11 dux p.11 = 0 " 2 Examples: Ni(CN)&¶-, Ni(DMG)l 3 11 &.;4..

Table 4. Spectral Properties and Stereochemistry of Ni(ll)

-Octahedral- --Tetrahedral- ---Planar- I/Amxx '/Amax e

800&13000 5-50 6900- 20-60 -17000 -100-600 7000 2W00 -10&F00

1400&19000 5-50 -15000 100-250 (multi-

25000-30000 5-50 plet)

Although the geometric isomers of the cobalt(II1) tetramines have been known for many years, the appli- cation of modern measurements and symmetry con- siderations has provided techniques that may, in some cases, serve to distinguish between the cis and trans

forms relatively easily. The trans isomer of Co- (NH3)&12+ has d4,, symmetry while the cis isomer is only c2, in symmetry. The reduction in symmetry from Oh to either that of the cis or trans isomer results in the splitting of the orbitally degenerate exicted states; however, the number of states produced by the splitting is greater in the case of the cis isomer while the separation between individual states produced by the splitting is sometimes greater in the case of the trans isomer. Thus, in the optical spectrum of the cobalt(II1) complexes, the bands of the trans isomer are less symmetrical or are sometimes resolved into two components (3). On the other hand the group vibra- tional frequencies of coordinated amines are commonly split in the case of the cis isomer and not in the trans case. Infrared spectra may he utilized to distinguish the isomeric structure of such carbonyl derivatives as Cr(CO)4B2 ( I t , 13).

Optical Activity

Although the optical activity of metal complexes has received much attention since its discovery by Werner more than half a century ago, a number of very basic issues remain open to investigation. It cannot be claimed that the ultimate source of the rotational power has been interpreted on a broad and detailed theoretical base. A number of investigators have considered the problem, particularly as regards the tris bidentate complexes such as the ethylenediamine complex of cobalt, C ~ ( e n ) ~ ~ + ; however, most of these efforts have fallen short of their objectives (14). The most promising theoretical model has not yet been applied to a valid experimental case, but is generally consistent with fact. The earlier efforts were based on the electrostatic model while the latest theoretical treatment involves molecular orbital theory. It ap- pears likely that the rotation cannot simply be ascribed to the perturbations associated simply with the D3 symmetry of the immediate environment of the metal ion. Liehr associates the optical rotation with an angle of cant, a, between the ligand donor orbitals and the metal acceptor orbitals. This implies a strong sensitive dependence of the strength and, perhaps, sign of rotation on ring size and other steric factors (for a given metal ion) as well as on the absolute direction of spiral of the helically disposed bidentate ligands.

Absolute Configuration

One of the principal needs of the area relates to the establishing of methods for the determination of the absolute configuration of complexes. Perhaps the most significant experimental result concerning the optical

activity of metal complexes during the past decade is the establishing of the absolute configuration of (+),- Co(en).2+ ion by a refined X-ray study (15). As shown in structure I, the so-called d-isomer, which has a specific rotation of +440 degrees circular

78 / lournol o f Chemical Education

a t the mercury green line, has its ligands so arrayed that the structure is similar to that of a left-handed, three-bladed propeller. Thus, the isomer long re- ferred to as dextro actually has a natural screw-form that is left handed according to mechanical conventions. Recognizing the difficulties of nomenclature that so often occur in treating the subject of absolute confignra- tion, Piper (16) has proposed that absolute right-handed helical molecules or ions be called delta (A) and the enantiomers be called lambda (A). The significance of this determination lies in the fact that the availability of a single species for which the absolute configuration is known opens the possibility of the rapid establish- ment of the absolute configurations of many other re- lated substances through relatively simply techniques for the determination of relative configuration. Such techniques have been considered a t length by Basolo (1 7, 18).

The use of rotatory dispersion measurements has received renewed attention recently in attempts to relate the configurations of complex ions (19, $0). Despite the difficulty of developing detailed theories, the fact remains that the metal ion in an optically ac- tive octahedral complex generally constitutes a chromo- phore congruent with the "center of dissymmetry." Consequently the absorption bands of the metal ion are optically active. In the case of ( + ) - C ~ ( e n ) ~ ~ + , ab- sorption bands occur at 470, 340, and 218 mw. Of these only the first and last contribute materially to rotatory power. The rotatory dispersion curves for cobalt(II1) complexes have long been suspected of containing sufficient information to relate the configura- tions of many complexes. Because of the relative ease of measurements, attention has focused on the long

:;p -2

-3

-4

-5

500 600 700 Wavelength, mjt

Figure 1. The rolotory dispersion of d-I+)-Colenlss+.

wavelength band. The rotatory dispersion curve of (+)-C~(en) ,~+ is given in Figure 1. The enantiomer exhibits a rotatory dispersion that is the mirror image of this curve. For very similar complexes, it is not difficult to accept the assignment of relative configura- tion on the basis of corresponding rotatory dispersion curves. For example,

(+ 1-ICo(en)da+, ( + I - 0 - )(pnklB+, (+)-KO(- )(pn)(enhIP+, (+ 1-Go(- )(pn)Aen)la+, and (+)[Co(en)z(NH&la+

almost certainly have the same configuration (19).

In all these cases, the spectral band ('A,, -+ 'TI,) occurs a t essentially the same position with very little alteration in its shape. Except for the last example, all contain 3 five-membered chelate rings. The simi- larities in the rotary dispersion curves of (+)-[Co- (en)3Iw+, (-)-[Ir(en)3lW, (-)-IRh(en)31a+, and (+)- [Cr(en)s]a+ suggest that the configurations correspond. Certainly, all the ions except Cr(II1) are d q n confignra- tion and the transition involved is 'Al, - 'TI,.

In the two particular series cited, add~tional evidence supports the conclusion that the configurations are all congruent. The relative solubilities of diastereoiso- meric salts corresponds. (+) - [Co(en)a]CI(+) -tar- trate.5Hz0 and (-)-[Rh(en)a]C1(+)-tartrate.4H20 are least soluble, as are the (+)-nitrocamphor salts of (+)-[Cr(en),]", (-)-[Rh(e&Is+, and (-)-[Ir(en)a]a+. All the ions reputed to be of the same configuration exhibit Cotton effects of the same sign. Finally, a third point of evidence is found in the method of active racemates (18) due to Delepine. In some crystals of racemic materials the enantiomeric sites are distinct and present in stoichiometric ratio. When this is true an isomorphic complex ion of another metal ion may replace one of the enantiomers. Thus, in D(+), L(- ) [C~(en)~]Br~ , the D(+) isomer may be replaced by (-)-[Rh(en)8]a+. This means that (+)-[Co(en)#+ and (-)-[Rh(en)Js+ have the same configuration.

Similar series may be set up in the case of other li- gand types (21); however, it is relatively difficult to connect the various series to one another.

On the basis of similar R D curves, corresponding solu- bilities, and relative toxicities of enantiomers (H), the following series have been assigned.

(A) ( + I [Ni(phen)da+, (+)IRu(phen)~ls+, (-) [Oa(phen)rlS+ ( - ) [Fe(phen).lz+ (1)

(B) (-1 ICo(C*O+)d-, (+)[Rh(G03daf (+) [Ir(C,O,)d-, (+ )[Cr(C,O&l8-

Some of the difficulties and nncertainties of studies in this area become evident even with these relatively ideal, rather selected series. The solubility criteria are based on a complex phenomenon. In some cases (e.g., [Ni(dipy),](tart).6H20) rapid crystallization with excess anion will bring down one diastereoisomer while slow crystaIlization from stoichiometric solutions will yield the other.

Many more interesting examples exist among com- plexes of still lower symmetry, where nncertainties mnl-

I . L.. 650 600 550 500 450 (cry,)

Figure 2. Rotatory dispersion curves for l+).[Co(en)rNH8CI]'+---, (+ )- [Co(cn)rNHaH~O15'-.-, (+I-[ColenlnlNHn)~lS+-.., and I+)-[CdednNHr NO1]'+. . . ..

Volume 41, Number 2, Februarv 1964 / 79

clarification of the question of the ligand-ligand inter- actions which depends on absolute configuration (31). The reaction of interest is given below.

[Pt( - )(pn)Cld + 2(+)pn.- D,L- Pt(- )(pn(+ )(pn)d4+

This and D,L-[Pt(+)pn(-) ( ~ n ) ~ ] ~ + have been resolved into D and L forms. The compounds are extremely sta- ble and unreactive. The isomer yields prepared in this way are essentially k'metically determined since the D:L ratio for P t ( - ) ( ~ n ) , ~ + prepared in this way is 3/2 while the equilibrium ratio is 3 7/1.

A classic paper by Bailar and Carey (1959) provides the basis for understanding many of the observations reported on stereospecificity among complexes (52). This represents the first substantial application of conformational analysis to the problem represented by chelate rings.

In C~(en)~ ,+ , strain-free rings are formed in which the conformation is a staggered or gauche form. How- ever, two such forms are possible and the repulsions between the atoms of the three rings differ according to the form present. The two forms produce different relationships among the skeletal atoms of the struc- ture. For a d-isomer, in the so-called k-rings, the C-C axis is parallel to the major 3-fold axis of the octahe dron. The second or k'-ring is slanted or staggered with respect to the major 3-fold axis. This is shown in Figure 3. Further, the k-ring is predicted by calcula- tion to be more stable by about 0.6 kcal/mole than the k'-ring, in a dextro tris-octahedral structure.

Actually, C ~ ( e n ) ~ ~ + could be kkk, kkk', kk'k', or k'k'k'. The analysis shows the kkk form to be favored by about 1.8 kcal/mole or 0.6 kcal per ring on a molar basis. This selectivity occurs because of H-C interac- tions which appear to account for some 95+% of re- pulsion. This conclusion is sustained by the crystal structure of the compound.

These results may be applied immediately to the pro- pylenediamine complexes since the absolute configura- tion of pn has been determined (55, $4). L(+) pn has the absolute configuration given in structure IV.

In order to minimize repulsions, the methyl group must occur in an equatorial orientation, structure V. Fur- ther, for a given isomer of pn,

the placing of the CH, group in the equatorial position is equivalent to specifying whether the ring is k or k' in configuration. The more stable form of C ~ ( p n ) ~ ' + should certainly have the k conformation in its rings. These relationhips are shown in detail in Figure 3 and structures I, IV, and V. We can now understand clearly why D(+)-c~(+)(pn) ,~+ is the more stable isomer and more abundant in synthetic systems a t equilibrium.

This has been further demonstrated by studies on the equilibrium mixture of products from systems con- tauimg en and (-)pn, and is summarized below ( W .

kCA r I w kr,k',k', rings C-C axis parallel to 3-fold axis C-C oxis staggered

Figure 3. Absolute conflgurdion and ring conformotion.

Ratio A F r.-[Co(en)d - )pnl S+/~-[Co(en)l(- )pnl

k' k' kt kk k' 211 0.4

Figure 4. Enantiomarphic ring conformations.

It is perhaps not obvious that a single k-ring and a corresponding k'-ring are merely mirror images of each other. This is illnstrated in Figure 4. Clearly, one ring enantiomer is better suited to minimize repulsions for a given tridihedral, octahedral enantiomer. The conformational analysis considers:

(1) H-H repulsion (nonbonded interaction); (2) H-C repulsion for H(axia1) of NH2 and C of different ring; (3) H-H repulsion of NH2 groups (coulombic).

As seen, the kt-ring isomer favors the L (or A) octahedral isomer roughly in proportion to its abundance.

The requirement that the bulky group occupy an equatorial position is well supported by equilibrium data on complexes of C-substituted ethylenediamines. For the reaction

the value of log K , for en, pn, and d, l-bn (bn = NH, CH(CH3)CH(CH3)NH2), is 20.05, 19.77, and 20.39, respectively (56). In these cases, it is possible for all methyl groups to assume equatorial orientations. However, in meso-bn and in i-bn (NH2C(CH3)2CH2- NH2) one of the methyl groups must be axial. I n the latter cases, log K , is greatly reduced, being 16.74 and 15.98 respectively. In the case of tetramethyl-

Volume 4 1 , Number 2, February 1964 / 81

ethylenediamine, the repulsions are so great that the tris complex is not formed.

Probably the most dramatic example of stereospeci- ficity in the course of reactions of complexes is that oc- curring when the cohalt(II1) complexes of EDTA and PDTA react with ethyleuediamine (3'6,3'7).

(+ )-[Co(EDTA)I- + 3en - (+)-[Co(en)ala+ + EDTA4-

(-)-[Co(EDTA)XIP- + 3en - (+)-[Co(en)rl" + EDTA*-X-

For first equation and second, when X = CI, Br, the yields of (+) isomer of product is < 63%; when X = NO?-, a racemate is produced. ( f )-[Go(- )PDTAld + 3en- (- )-[Co(en)a13+ + ( - ) PDTA4-

Yields of excess isomer is 100%. As a result of stereochemical studies and detailed in-

vestigation of the kinetics of these fascinating processes, mechanisms and complete unequivocal stereochemical paths have been deduced (3'7). The kinetically de- rived mrchauism is shown below.

k, Co(EDT.4- + en F=== *[Co(EDTA)enl - (monodentate en)

k - I

k* *[Co(EDTA)en-HI2-- [Co(EDTA)en-HIZ- (bidendate en)

When corrections are made for competing hydrolysis reactions, the equation representing this mechanism is obeyed over a range of [H+] that is 3160 fold and an [en] range of 575 fold.

The critical kinetic intermediate involves monoden- tate coordination of the first molecule of en. This species is assumed present in accord with steady state approximations. It is capable of initiating rapid sucessive substitution steps via an SNlcb mechanism (Q), a proposal in excellent accord with the data. The form of the rate law is such that en must enter the mechanism before OH-. Further, though [OH-] is in the rate law, that species functions entirely as a catalyst.

Although monodentate complexing of ethylenedi- amine is required in the critical kinetic intermediate, the manner of chelation of the first molecule of en is critical to the stereospecificity of the reaction. Only readily acceptable assumptions need be made in this regard. (1) The multiply connected ligand will pre- vent rearrangement. (2) Cis groups, not trans groups, will be replaced. (3) No group is replaced if it is still bound in two chelate rings. Proceeding in this manner,

three stereochemical intermediates of similar, or pos- sibly identical, probability may be formed with the first molecule of en. As the additional molecules of en are

added, two of the stereochemical intermediates produce racemic Co(en)2+. The intermediate showing selec- tivity involves replacing two equivalent carboxyl groups from the plane of the three fused chelate rings (structure VI). This intermedite produces C ~ ( e n ) ~ ~ + of the same absolute configuration as the starting ma- terial. Further, the simple conversion

(+)-[Co(EDTA)Xlz- - (-)-[Co(EDTA)I - + X-

provides a check on the relationships given. Also, the fact that (+) [Co(EDTA)KOz] yields only race- mic Co(en)P is of much significance, for the KOo- group is held very tightly and most probably occupies one of the positions that must be freed to produce the intermediate that causes stereospecificity. The ab- solute stereochemical relationships proved in this way are shown in Figure 5.

Figure 5. Absolute configurations of EDTA complexes (signs refer to rolo- tian at mercury green line).

In the PDTA case, the absolute configuration of the ligand can be deduced from that of pn, while that of the complex follows from Co(EDTA)-. In conformance with expectation, the methyl group in the only known isomer is oriented equatorially. This ligand is per- haps the most extreme in its determination of the geom- etry of an octahedral complex, the second diastereo- isomer does not exist in detectable amounts (58).

It should be recalled that the reaction of (+)-[Co- (PDTA)]- with en produces Co(en)2+ in a totally resolved form and with the opposite configuration to that found in the case of (+)-[Co(EDTA)]- (57). The Emetics of the reactions are essentially the same so that the mechanism must be unaltered except as regards the relative importance of the critical stereo- chemical intermediates. Since none of the C ~ ( e n ) ~ ~ + is of the same configuration as the starting material, the two carboxyl groups coplanar with the S-C- C-K ring of PDTA are not replaced by the same en molecule. The presence of the CH, group has caused the othelwise pairwise identical carboxyl groups to be distinct. In fact, the two groups attached to chelate rings in closest proximity to the CH3 group would be expected to be most reactive. Assuming these two groups to be replaced by the first molecule of en and proceeding as before, it is seen that inversion is favored. It is to be expected that the combination of kinetic, structural, and conformational techniques will lead to continued advances in the area of stereospecificity.

82 / Journal of Chemical Educafion

Linkage Isomers

The existence of isomeric forms of complexes con- taining monodentate ligands bound in different ways was first observed in the case of cobalt(II1) complexes of nitrite ion in 1893; however, the phenomenon has been extended to include rhodium(III), iridium(III), and platinum(I1) complexes only recently (39). The second example of ligand manifesting linkage isomerism was reported by Basolo, Burmeister, and Poe in 1963 (40). The reaction of K2Pd(SCN)4 with 2,2'-bipyri- dine (N-N) in alcohol a t -7S°C leads to the forma- tion of the sulfur-bonded complex (structure VII).

When this substance is heated to 150°C, it undergoes rearrangement, forming the nitrogen bonded isomer (structure VIII).

N\ /NCS

N/ \NCS VIII

A rather closely replated phenomenon of much stereo- chemical significance, but not involving isomerism has been termed flexidentate chelation. Polyfunctional molecules are often capable of coordinatining in a vari- ety of ways, depending on which of their donor groups are bound to a metal ion. The phenomenon is per- haps best illustrated with the simplest amino acid, glycine. In the more familiar complexes, the glycinate anion forms bideutate derivatives as in [Co(KH2- CHzC0z)3]. However, under certain conditions mono- dentate coordination occurs, e.g., [CoCla(NHzCHzCOr C02H),] (41). Similarly, the anion of EDTA has been found to bind in stable complexes in which it is sexa- dentate (Fig. 5) (48) pentadentate (Fig. 5) (43), tet- radentate (structure IX) (44), and bidentate (struc- ture X) (44).

The significance of this phenomenon may be seen in the fact that a number of naturally occurring amino acids (e.g., cystine, ornithine) undergo such variations in mode of chelation. Further the interaction of metal ions with complex natural substances in biological sys- tems or with polyfunctional medicinals is almost cer- tainly complicated by behaviors of this sort. This is emphasized by X-ray structural results (@) that in- dicate that in neutral media, peptide units of glycyl- glycylglycine bind to copper(I1) ion through a terminal amino group and the carbonyl oxygen atom, while in basic systems binding is by means of terminal and ion- ized amido nitrogen atoms.

Electronic Isomers

For the electronic configuration of free metal ions d4, dS, d6, 8, and d8, both high-spin and low-spin com-

plexes have been known for many years. Modern theory suggests that the right balance of structural parameters for a given ion (having such a configura- tion) might lead to the coexistence of the two spin states. This would represent an isomerism dependent principally on electronic arrangement alone.

It is an interesting fact that relatively detailed the- oretical treatments (46) were developed for the case of tetragonal nickel(II), but that the suspected examples have all been shown to involve variations in molecular as well as electronic structure.

Eaton, Phillips, and Caldwell (47) have established that the unusual isomerism of the nickel(I1) complexes of aminotroponeimine derives from a confonnational equilibrium involving diamagnetic planar and paramag- netic tetrahdral forms of the four-coordinate complexes. A more complex behavior has been observed among the N-substituted salicylaldimines (48, 49). Depend- ing on the nature of the alkyl or aryl group on the imine nitrogen, varying behaviors are observed. A tetrahe- dral (paramagnetic)-planar (diamagnetic) conforma- tional equilibrium has been found. Also, with low steric requirements on the part of the substituents, po- lymeric, paramagnetic octahedral forms also occur.

In contrast to the behavior of nickel, compounds have been found with iron and cobalt, for which the possi- bility of electronic isomerism remains, as yet, un- complicated by indications of variations in coordination number or conformation. George and co-workers (50) are able to account quantitatively for the anoma- lous magnetic and spectral properties of the hydroxides of iron(II1) hemoglobin and myoglobin on the assump- tion that the substances are compo~ed of isomers of high-spin and low-spin forms. Stoufer, Busch, and Hadley (51) have found that certain six coordinate cobalt(I1) complexes mostly with a-triimine structures anologous to 2,2',2'-terpyridine, exhibit anomalous magnetic moments whose temperature dependences sug- gest the existence of high and low spin forms in equlib- rium. For example, the room temperature magnetic moment of his(2,6-pyridindialdihydrazone)cobalt(II) iodide (ligand structure XI) is 3.04 Bohr magnetons

while the usual high-spin and low-spin values are ap- proximately 5 and 1.9 Bohr magnetons, respectively. Although any structural differences that exist between the electronic isomers of the cobalt(I1) complex are almost certainly more subtle than in the nickel(I1) case, the probability remains that the bond distance and bond angles differ substantially. Indeed, in view of the nonequivalent electronic distributions, identical atom arrangements seem unlikely for electronic isomers even in those instances where coordination number and symmetry remain unaltered.

Template Reactions

The presence of a metal ion during the course of an organic reaction may have a profound effect on the stereochemistry of the process. The formation of cy- clic chelates about metal ions provides the best ex-

Volume 4 1, Number 2, February 1964 / 83

amples of such processes. In these cases the metal ion may be considered to serve as a template, effectively organizing the reaction either by stereoselective arrange- ment of reactants or by favoring the formation of a product ideally suited for chelation.

Perhaps the classic example is that of phthalocyanine formation, essentially from phthalonitrile and the me- tal ion. However, this complicated reaction is not fully understood and is carried out under conditions not well suited to detailed study.

A number of other fully cyclic chelates have recently been reported. The simplest example is that due to Schrauzer (52) and Thierig and Umland (65). In this case, the proton from the hydrogen bridge in bis(di- methyglyoxime) nickel(I1) has been replaced by the function BX2+, where X is For CeHs (structure XII).

A similar process is found in the reaction of tetra- dentate biacetylbis(mercaptoethylimino)nickel(II) with alkyldihalides (54).

The dissolution of tris(ethylenediamine)nickel(II) per- chlorate in acetone leads to a most remarkable reaction, forming a cyclic ligand (55).

The most recent example of a cyclic ligand has been obtained by the self-condensation of o-aminobenzal- dehyde in the presence of Ni(I1) or Cu(I1) (56). The product of the condensation has been formulated ac- cording to structure (XIII).

These unusual substances (57) exhibit remarkable stabilities, many of them resisting boiling concentrated mineral acids. Further, such reagents as sulfide, di- methyl glyoxime, and, in certain cases, even cyanide fail to remove the metal ion from its encycled coordina- tion site. The fact that the metal ion is instrumental in the synthesis of these cyclic ligands requires special emphasis. In this regard, the planar metal ions, copper(I1) and nickel(II), understandably have dom- inated studies leading to the synthesis of new tetra- dentate macrocycles. Metal ions tending more gen- erally to favor other stereochemistries might confer different structures on ligands synthesized under their directive influence. As an indication that this is true, Curry and Busch (58) have reported the synthesis of a macrocyclic sexadentate ligand in the presence of iron salts.

Recent advances in the stereochemistry of metal complexes have been made in many ways on many sub- jects in addition to the several summarized here. Al- though a number of alternate subjects are of equal or greater importance, the author has chosen to emphasize areas not necessarily so widely reviewed in the context of stereochemistry as others, or those of his own in- terest.

Acknowledgment

The financial assistance of the National Science Foundation and the Xational Institutes of Health is gratefully acknowledged.

Literature Cited

(1) BAILAR, J. C., JR., "Chemistry of the Coordination Com- pounds," Reinhold, New York, 1956, chap. 2.

(2) GRINBERG, A. A,, "An Introduction to the Chemistry of Complex Compounds," (English transl.) Pergamon, New Yark, 1962.

(3) WILKINS, R. G., AND WILLIAMS, M. J. G., "Modern Coor- dination Chemistry," Interscience, New York, 1960, chap. 3.

(4) SUTTON, L. E., J. CHEM. EDUC., 37, 498 (1960). (5) PEARBUN, R. G.,J. CHEM. EDWC., 38,164(1961). (6) MANCH, W., AND FERNLEIUS, W. C.. J. CHEM. EDUC.. 38.

192 (1961). (7) LIEHR, A. D., J. CHEM. EDUC., 39,135 (1962). (8) CARLIN, R. I,., J. CHEM. EDUC., 40,135 (1963). (9) BASOLO, F., AND PEARSON, R. G., "Mechmiams of Inor-

ganic Reactions,'' John Wiley, New York, 1958. (10) COTTON, F. A., GOODGAME, D. M. L., AND GOODGAME, M.,

J. Am. C h m . Soe., 84,4161 (1961). (11) MORRIS, M. L., AND BUSCH, D. H., J. Am. Chem. Sue, 82,

1521 (1960). (12) ANQELICI,R. J., BASOLO, F., AND FOE, A. J., J. Am. Chem.

Soc.. 85. 2215 (1963). . . (13) KILN&, M., thesis, U. of Nottingham, 1963.

84 / Journol of Chemical Education

(14) These are summarized in s. paper soon to appear: LIEHR, (36) DWYER, F. P., AND GARVAN, F. L., J . Am. Chem. Soc., 80, A. D., J . Phys. Chem., in press.

(15) SNTO, Y., NAKATSU, K., SHIRO, M., AND KIJROYA, H., Bull. Chem. Soc. Japan, 30, 795 (1957); Acta Cryst., 8, 729 (1955).

(16) PIPER, T . S., J . Am. Chem. Sac., 83,3908 (1961). (17) BASOLO, F., Chem. Rev., 52,459 (1953). (18) BA~OLO, F.: "Chemistry oi the Coordination Compounds."

J. C. BAILAFC, JR., Editor, Reinhold, New York, 1956, chap. 8.

(19) MACDERMOTT, T . E., AND SARPESON, A. M., Australian J . Chem., 16, 334 (1962); I ~ R ~ S ~ M I L L E R , J. G., AMMA, E. L., AND DOUGLAS, B. E., J . Am. Chem. Soe., 84, 3227 (1962).

(20) WOLDBYE, F., Ree. Chem. Pmgiess, in press. (21) DWYER, F. P., (deceased) and SARGESON, A. M., private

communication. (22) BASOLO, F., STONE, B. D., A N D PEARSON, R. G., J . Am.

Chem.Soc.,75,819 (1953). (23) TAUBE, H., AND MLTRMANN, 12. K., J. Am. Chen~. Soc., 78,

4886 (1956). (24) GELMAN, A. D., AND ESSEN, L. N., PTVC. Awd. Sci. U.S.S.R.,

(English trans%), 108,309 (1956). (25) LIFSCHITZ, I , Z . Physik. Chem., 114,493 (1925). (26) BAILAR, J. C., JR., AND MCREYNOLDS, J. P., J . dm. Chem.

Sac., 61,3199 (1939). (27) DWYER, F. P., AND GARVAN, F. L., J. Am. Chem. Soe., 81,

1043 (1959). (28) DWYER, F. P., AND GARVAN, F. L., AND SHULMAN, A,, J.

Am. Chem. Soe., 81,290 (1959). (29) JAEGER, F. M., AND BLUMENDAL, H. B., Z. an07g. Chem.,

175, 161 (1928). (30) DWYER, F. P., AND SAROESON, A. M., J . Am. Chem. Soe., 81,

5269 (1959). (31) DWYER, F. P., A N D SARGESON, A. M., J . Am. Chem. Soe., 81,

5272 (1959). (32) COREY, E. J . , AND BAILAR, J . C., JR., J . Am. Chem. Sac.,

81,2620 (1959). (33) SCHNELL, S.. AND KARRER. P.. Helv. chim. A&. 38. 2036 . . , .

(1955): '

(34) SAITO, Y., AND IWASAEI, H., Bull. Chem. Soe. Japan, 35, 1131 (1962).

(35) BASOLO, F., C R E N , ~ . T., AND MURMANN, R. K., J. Am. C h . Soe., 76,956 (1954).

4480 (1958). (37) Buscu, 1). H., SWAM~NATHAN, K., AND COOKE, DEAN W.,

Inorg. Chem., 1,260 (1963). (381 IM. Y . A.. AND BUSCH. D. H.. J . Am. Chem. Soc.. 83. 3362 . . . .

(1961).' (39) BASOLO, F., AND HAMMAKER, G. S., Inarg. Chem., 1, 1

1lRR2I \ - - - - , .

(40) BASOLO, F., BURMEISTER, J. L., AND POE, A. J., J. Am. Chem. Sac., 85,1700 (1963).

(41) OGOLEVA, V . D., T d g Dagestan SelSkokhoz., Inst., 7 , 130 I 1 0 ~ C I ,I"Y",_

(421 WEAKLEIM. H . A,. AND HOARD. J . L.. J . Am. C h m . Sac., 81. . . . . 550(1959). '

(43) SMITH, G. S., AND HOARD, 3. L., J. Am. Chem. Soc., 81, 556 119.591. ~ - ~ . ~ ~

(44) BUSCH, D. H., AND BArLAR, J. C., JR., J . Am. C h m . Sac., 78,716 (1956).

(45) FREEMAN, H. C., LOVELL, F. N., ROBINSON, G., SCHOONE, J . C., AND SIME, J . G., PIOC. 7thI,CCCCCCC Stock- holm, 1962, p. 115.

(46) BALLHAUSEN. C. J.. AND LIEHR. A. D.. J. Am. C h m . Sac.. . ,

(47) EATON, D. R., PHILLIPS, W. D., AND CALDWELL, D. J., J . Am. Chem. Sac., 85,397 (1963).

(48) SACCONI, L., PA OLE^, P., AND CIAMPOLINI, M., J . Am. Chem. Soe., 85,411 (1963).

(49) HOLM, R., AND SWAMINATHAN, K., Zna~g. Chm., 2 , 181 (1963).

(50) GEORGE, P., BEETLESTONE, J., AND GRIFFITH, J. S., "Hema- tin Enzymes," Pergsmon, New York, 1961.

(511 STOUFER. R. C.. BUBCH. D. H.. AND HADLEY. W. B.. J. Am. . . hem. Boc., 83,3732 ( i96 i ) . '

(52) SCHRAUZER, G. K., C h m . Ber., 95,1438 (1962). (53) THIERIG, D., AND UMUND, F., Angm. Chem., 74, 388

(1962). (54) THOMPSON, M . C., A N D BUSCH, D. H., Chem. Eng. News, 57,

Sept. 17,1962. (55) BLIGHT. M. M.. AND CURTIS. N. F.. J. C h m . Soc., 1062, . .

3016.' (56) MELSON, G. A,, A N D BUSCH, D. H., Proc. Chem. Soc., 1063,

222. (57) Buscn: D. H., Ree. Chem. Progress, in press. (58) CURRY, J. D., AND BUSCB, D. H., J . A m . Chem. Soc., inpress.

Volume 4 1, Number 2, February 1964 / 85