Aug. 20, 1959
DOUBLEOPTICALISOMERISM IN COBALT(III) COMPLEXES
which enhances the stability of the fluoride complexes. Thus if S C +has ~ a highly strained, coordinated layer of eight water molecules around it, replacement of a water by the slightly smaller fluoride ion would relieve this strain and lead to an unusually stable complex. It would be necessary that additional strain be relieved as each of the next three replacements of water by fluoride ion occur because the first four complexes are known to be exceptionally stable. This explanation, however, is contrary to the fact that scandium shows only a coordination number of six in its solids2and that the radius ratio2 fits a coordination number of six with oxygen. It is possible that it is the aluminum complex which has unusually low stability relative to the stability expected from an ionic model. Such a decrease in stability could arise from steric repulsion. Aluminum ion is on the borderline between fourfold and sixfold coordination by fluoride ion2;
[CONTRIBUTION FROM THE
4189
consequently, there should be strong steric repulsion for sixfold coordinated aluminum ion. Since the water molecules on the hydrated ion are held considerably less tightly than the fluoride ions, the net effect of steric repulsion would be to decrease the binding of the fluorides and lower the values of the complexing constants. I t is not apparent, however, how such an effect would lead to equilibrium constants for the aluminum complexes which are actually smaller than for scandium. Although some of the above suggestions offer possible, partial explanations of the apparent unusual stability of the scandium fluoride complexes, none of them is really satisfying. The reason for this degree of failure of the simple ionic model is not clear. This work was supported by the U. S.Atomic Energy Commission. BERKELEY, CALIFORNIA
NOYESCHEMICAL LABORATORY, UNIVERSITY O F
ILLINOIS]
The Stereochemistry of Complex Inorganic Compounds. XXIII. Double Optical Isomerism and Optical-Geometric Isomerism in Cobalt(IT1) Complexes BY WILLIAME. COOLEY,CHUI FANLIU .4ND
JOHN
c. BAILAR,JR.
RECEIVED FEBRUARY 16, 1959 Double optical isomerism, in which both the whole ccmplex and one of its ligands are asymmetric, exists in the cis-dinitroethylenediamine-2,3-butylenediamine-cobalt(111) ion. This kind of isomerism exerts a stereospecific effect, and some of the four optically active isomers are less stable against racemization in solution than the others. Optical-geometric isomerism, in which the whole complex is asymmetric and a symmetric ligand is so constituted t h a t i t may be oriented in either of two ways, exists in the cis-dinitroethylenediamine-isobutylenediamine-cobalt( 111)ion. The geometric isomerism modifies the optical isomerism so t h a t this ion also has four optically active isomers, but there is no evidence of preferential stability. The crystals of two of the optical-geometric isomers appear different from those of the other two, b u t all the double optical isomers have like crystals. The absorption spectra of these cis complexes are different from those of the corresponding trans complexes in a way analogous to the behavior of other dinitro complexes of cobalt(II1). A study of the preparation of the optically active complexes led to an improved method of preparing the non-electrolyte trinitrotriammine-cobalt(111).
I n the preparation of complex compounds containing optically active ligands, stereospecific effects often appear, with the result that some of the possible isomers of such compounds occur as small fractions of the total yield, or they do not occur a t all. For example, Hiirlimann2 prepared the dinitro-bis-(propylenediamine)-cobalt(II1)ion, [Co pn2 ( N O Z ) ~ ] . +There are twelve possible isomers of the cis form of this ion, corresponding to the two optical isomers of propylenediamine, three geometric orientations of the two propylenediamine molecules, and two optical forms of the whole complex. The trans form has four possible isomers, corresponding to the two isomers of propylenediamine and two geometric orientations. More isomers of both cis and trans form are theoretically possible, of course, if the two propylenediamine molecules in each complex ion are permitted to have either the same or opposite optical configuration. Concluding that he had prepared (1) T h i s article is based on portions of dissertations submitted in partial fulfillment of t h e requirements for t h e P h . D . degree a t t h e University of Illinois. (2) H u r l i m a n n , Thesis, University of Zurich, 1918. Reviewed in F. M. Jaeger, “Optical Activity a n d High Temperature Measurements,” McGraw-Hill Book Co., I n c . , New York, N. Y., 1930, pp. 1.57-168.
the cis form, Hiirlimann found that only two isomers, ~ l and l ~ d d could , ~ be isolated instead of twelve. Rotatory dispersion studies by O’Brien, McReynolds and Bailar4 have shown that Hiirlimann probably worked with the trans form but, even so, the theoretically possible number of isomers was not realized. Similarly, Smirnoff5 was able to isolate only the ~ l land l Lddd isomers of the tris-(propylenediamine)platinum(1V) ion, [Pt pn3I4+. Smirnoff also showed t h a t the analogous cobalt(II1) complex, [Co pn3I3+,is stable only when all three of the propylenediamine molecules have the same optical configuration. Bailar, Stiegman, Balthis and Huff man6 concluded that “mixed” isomers containing both configurations of propylenediamine may form but that they rearrange to form the two products (3) T h e optically active complexes of this t y p e are designated here a n d l a t e r b y such symbols a s ~ l al n d ~ l i l where , t h e capital letter symbolizes t h e direction of rotation of t h e complex a s a whole, while t h e small letters symbolize t h e number and rotation of t h e optically active ligands. (4) T. D. O’Brien, J. P . McReynolds and J. C . Bailar, Jr., THIS J O U R N A L , 7 0 , 749 (1948). ( 5 ) 0. P. Smirnoff, Helo. Chim.Acta, 3, 177 (1920). (6) J. C . Bailar, Jr., C. A. Stiegman, J. H. Balthis, J r . , and E. H. Huffman, THISJ O U R N A L , 61, 2402 (1939).
41'30
\\'ILLIAM
E.COOLEY, CHUIFANLIU AND JOHN C. BAILAR,JK.
Vol. SI
that Sniirnoff found. The same restrictioii on the bromocamphorsulfonate method. Separate prepaisomers of tris-( 1,2-tra~zs-diaminocyclopentane)-corations with each of the isomers of propylenediamine balt(II1) was observed by Jaeger and Blu~nendal.~made it possible to avoid the presence of all eight Jaeger summarized the effects of optically active isomers in a single batch. The eight separated ligands by saying that two or more such ligands in bromocamphorsulfonates were converted to the the same complex usually have the same configura- bromides, whose four positive optical rotations and tion, and the complex exists in only one configura- four negative rotations matched in four pairs. tion. The sign of rotation of the whole complex is The many steps of crystallization, with their sometimes the same as that of the ligands, sonie- inevitable losses of material, prevented an estimate times opposite. of the relative quantities of the isomers. Thus, A few examples are known of less restrictive Werner and Smirnoff left unanswered the question effects of optically active ligands. LifschitzY of whether some degree of preferential stability prepared tris-(d-a1aninato)-cobalt(II1) and noted operates to produce some of the isomers more that it yields Q- and P-geometric isomers; he re- abundantly than others. If it does, i t would be solved one of these into optical isomers with Dddd interesting to discover whether the effect can be and ~ d d d configurations. This complex is an attributed to the optical isomerism, or to the geoexception to Jaeger's generalization. Sister Mary metric isoiiierism, or both. hfartinette, McKeynolds and Bailar9 have found For these reasons we have studied complex ions that the carbonato-bis-(l-propylenediamine)-cobalt in which one source of isomerism is asymmetry of (111) ion exists in both D and L forins and that the the whole complex and the other is either optical configuration of its salts depends on the method isomerism of the ligand or geometric isomerism of crystallization. Bailar, Jonassen and Gott'" caused by the position of an optically inactive have demonstrated the existence of both d-tar- ligand. The sources of isomerism due to the nature trato-bis-(2-propylenediamine)-cobalt(III)ion and of the ligand have been considered separately, I-tartrato-bis-(1-propylenediamine)-cobalt(II1)ion. instead of together, as in Werner and Smirnoff's In these complexes, however, the two isomers are work. not equally stable; the I-tartrate group is the less We have prepared and resolved the four isomers tightly bound. of the cis-dinitroethylenediamine-aetive-2,3-butylThe isolation of the eight theoretically possible enedianiine-cobalt(II1) ion, cis- [Co en bn ( N 0 ~ ) 2 ] + . isomers of the cis-dinitro-ethylenediamine-propyl-Here the asymmetry stems from the ligand isomenediamine-cobalt(II1) ion was accomplished by erism and the asymmetry of the whole complex. LVerner and Smirnoff and reported by IVerner." The four isomers are Dd, Ld, ~1 and LI, and for This achievement prompted our own work; i t is purposes of reference and comparison we call this especially interesting because it combines geometric kind of isomerism double optical isomerism. There (a+) isomerism with a kind of optical isomerism in is evidence that the four isomers were formed in which preferential stability is not sufficient to re- nearly equal amounts in the preparative reaction, press completely the formation of any of the iso- but there were marked differences in rates of racemers. The optical isomerism of propylenediamine niization, so that not all the isomers were obtained and of the whole complex gives rise t o the four in an optically pure state. The restrictive informs ~ d DI, , ~d and LI. Each of these is capable fluence of an optically active ligand is appareiit, of existing in either of two geometric forms, de- a t least when geometric isonierisni does not occur pending on the position of the methyl group of at the same time. propylenediamine We also have prepared and resolved the four isoniers of the cis-dinitroethylenediamine-isobutylenediamine-cobalt(II1) ion, czs- [Co en ibn (N02)2]+. The asymmetry of this ion depends on geometric isomerism and asymmetry of the whole complex. Geometric isomerism is introduced in the same way as i t is in the propylenediamine complex diagrammed above, but the Urerner arid Smirnoff separated the eight isomers by ligand is not optically active, because of the prtsfractional crystallization of the d-a-bromocaniphor- ence of two methyl groups on one carbon atom. n-sulfonates. Since the a- and 6-forms are geo- The four isomers are DOI, ~ aDP, and LP; we call this metric isomers, their crystalline forms and solu- kind of isomerism optical-geometricisomerism. NO bilities need not be the same. Actually, one series differences in rates of racemization were ob( a ) crystallizes as prisms and the other ( p ) as served, and the matching pairs of rotations of the needles. The solubilities are similar enough so that resolved isomers indicates that they were optically the separation of all four isomers corresponding to pure. As in the preparation of the 2,3-butylenedione form of propylenediamine was made by the amine complex, nearly equal amounts of the four (7) I?. 31.Jaeger and H . B. Blumendal, %. wiut'g. allgriti.
