J . CHEM. SOC. DALTON TRANS. 1991 2903Temperature-dependent Conglomerate Crystallization ofCarbonatobis( ethylenediamine)cobaIt( 111) Bromide and ItsApplication to Kinetic Optical Resolution tKazuaki Yamanari * and Akira FuyuhiroDepartment of Chemistry, Faculty of Science, Osaka University, To yonaka, Osaka 560, JapanThe binary and ternary solubility diagrams of [Co(CO,) (en),] Br (en = ethylenediamine) have beendetermined at 5-70 "C. It was found that the bromide exists as a racemic compound dihydrate below15.0 "C and as an anhydrous conglomerate above this temperature. The conglomerate is kineticallyresolved by crystallization in the presence of chiral additives, A- [Co(glyO) (en),] Br, (gIy0 = glycinate)and A- [Co(ox) (en),] Br (ox = oxalate), at room temperature.The same resolutions utilizing racemizationof [Co(CO,)(en),]Br were also attempted at 40 "C.As a special method of optical resolution, conglomerates can beoptically resolved with the assistance of 'tailor-made impur-ities'. The mechanism is based on stereoselective adsorption ofa chiral additive and inhibition of crystal growth. We appliedthis method to optical resolution of M'X and M'X, type metalcomplexes where M' and X denote a racemic complex and itscounter ion, respectively.2 For example, in the presence of asmall amount of A-[Co(glyO)(en),]CI, (gly0 = glycinate,en = ethylenediamine), the additive is adsorbed on the Acrystals of [Co(ox)(en),]Cl (ox = oxalate) and inhibits theircrystal growth, leading to different crystallization rates,k , > k,.Therefore, the enantiomer A-[Co(ox)(en),]Clhaving the opposite absolute configuration to that of the chiraladditive crystallizes first in excess. Such resolutions are effectivein a dozen cobalt(Ii1) complex systems., In other words,this resolution procedure becomes a useful criterion to knowwhether the complex to be resolved crystallizes as aconglomerate or a racemic compound because only con-glomerates showed relatively high enantiomeric excesses.Preliminary experiments showed that the bromide of[Co(CO,)(en),] + is effectively resolvable by this procedure.Therefore, the binary and ternary solubility phase diagramshave been reinvestigated at various temperature^,^ and theabove kinetic resolution experiments utilizing racemization of[Co(CO,)(en),]Br at 40 "C, the second-order asymmetrictransformation, in addition to ones at room temperature,at tempted.ExperimentalPreparations of Metal Complexes.-The following cobalt(ir1)complexes used in this study were prepared and/or resolvedaccording to the methods described in the literature and wereconverted into the desired salts on a QAE-Sephadex A-25column: AA-, A- and A-[C0(C0,)(en),lBr,~ -[Co(o~)(en>,]Br;~-[Co(en),]X, (X = C1 or Br),6 -[Co(glyO)(en),]X, (X = C1 orB r ), ' 7 and - [Co( NO ,) ,( en) B r .9Solubility Measuren1ents.- -The solubilities of complexes inwater were determined as previously rep~rted:~.' weighedsaturated solutions were diluted with water to a certain volume.The solubility was determined on the basis of the optical densityand circular dichroism (CD) spectrum.The binary and ternary~ ~t Supplementur! dutcr uruiluhle (No. SUP 56847, 6 pp): solubility data.See Instructions for Authors. J. Chcm. Soc.. Dullon Trrrns., 1991, Issue 1.pp. xkiii -xxii.values in weight (7J are presented in SUP 56847. The solidphases were identified by elemental analysis, infrared.absorption and CD spectra and X-ray powder diffractionpatterns. The absorbances were measured with a Hitachi 330spectrophotometer and the CD spectra with a JASCO 5-500spectropolarimeter. X-Ray powder diffraction was recorded ona Rigaku model RAD-ROC instrument at the X-Ray diffractionService of the Department of Chemistry, using Cu-KY. radiation.Resolution of' Metal Complexes witJj the Assistunce of c'hirulAdditives-The racemic bromide [Co(CO,)(en),]Br and asmall amount of a chiral additive were dissolved in anappropriate amount of water.The mixed solution was allowedto stand at room temperature (L'LI. 22 'C). After partialcrystallization of the bromide, the solution was decanted andthe crystals were dried on filter-paper. All crystals wereredissolved in water, and the crystal yield and enantiomericexcess (e.e.) were determined by measuring the absorption andCD spectra of the solution. The results are collected in Table 1.Racemization-crjstullizution E.yperiments.-A mixed solutionof the racemic bromide and a chiral additive was allowed tostand at 40 "C. After crystallization of most of the carbonatocomplex, spontaneous evaporation was stopped.The crystalsobtained were redissolved in water, and the amount and e.e.were determined by absorption and CD spectral measurements.In the system A- or A-[Co(glyO)(en),]X, (X = C1 or Rr), thesolution of the racemic bromide and the chiral additive wasallowed to evaporate to dryness at 35 or 40 "C. The residueredissolved in water was chromatographed on a column ofSP-Sephadex C-25 and then the optical purity of [Co-(CO,)(en),]Br was determined. In this case, the recovery of thestarting carbonato complex was complete. The results arecollected in Table 2.Results and DiscussionFig. 1 shows the binary solubility curves at 5--70 C; the curvesof the enantiomer above 35 'C cannot be determined because ofthe racemization.The racemate shows polymorphism: it existsas a dihydrate below 15 C and an anhydrate above 15 "C. Theenantiomer exists as an anhydrate over the entire measuredregion 5-35 -C. We previously clarified that the racematecrystallizes as a conglomerate, when the solubility of theracemate is 3 2; times larger than that of the pure enantiomer ina M'X type metal complex.3 The ratio of the solubilities of theanhydrous racemate and the enantiomer S,,: S , is 1.55: 1 a29040.40.3v-I0, xE 0.2fF--.0.1J. CHEM. SOC. DALTON TRANS. 1991---0 ' I0 20 40 60T/"CFig. 1 Solubility curves, where solubility is presented in molality mof anhydrous salt: (1) A-[Co(CO,)(en),]Br and (2) AA-[Co(CO,)-(en),]Br-nH,O, n = 2 ( 6 15.0 "C) or 0 ( 2 15.0 "C)0IFig. 2 Solubility isotherms of the ternary system A-[Co(CO,)(en),]Br(l)--A-[Co(CO,)(en>,]Br (2)-water, at 5.0 and 25.0 "C.- - - - -,Metastable state20 "C and 1.54: 1 at 35 "C and therefore the present systemsatisfies the above condition. To confirm the conglomeratecrystallization, the racemic bromide was crystallized at 2 1and 60°C. Each aqueous solution of one crystal at bothtemperatures showed a CD spectrum. The optical purities werebetween 0.3 and 31% of the pure enantiomer. The infraredspectrum of the anhydrous racemate was the same as that of thepure enantiomer. The same situation was found in the X-raypowder patterns. Thus, it is concluded that the racemate forms aI ' \ I /' II cnYT -..F/Fig.3 Temperature dependence of the invariant points, C, E and F,and the point D corresponding to the raccmic compound in the ternaryisotherm. - - - -, Metastable stateracemic compound below 15 "C and a conglomerate above thistemperature as a stable phase.These results were supported by the ternary solubilitydiagrams at 5 and 25 "C shown in Fig. 2. At 5 "C the racemiccompound corresponding to the line FDE is a stable phase inequilibrium because there are two invariant points F and E, andthe conglomerate corresponding to point C appears only as ametastable state. On the other hand, the situation is reversed at25 "C: the conglomerate becomes a stable phase and the racemiccompound corresponding to point D is a metastable state.Thus,the tempera t ure-dependen t conglomerate crystallization isconfirmed for the present [Co(CO,)(en),]Br system by thebinary and ternary solubility phase diagrams. This is extremelyrare in metal complexes and the only precedent is [Co-The temperature for transition of the racemic compound tothe conglomerate can be determined more accurately using theternary isotherms. Fig. 3 shows only the central part of theternary diagrams at various temperatures. The range of the lineFDE showing the racemic compound dihydrate at 5.0 ' Cgradually diminishes with increasing temperature and is finallysuperimposed on the point C representing the anhydrousconglomerate at 15.0 "C. This temperature corresponds to theintersection between the dihydrate and the anhydrate in thebinary system.Above it the line FDE appears only as ametastable state as shown in the isotherm at 17.5 "C.The racemic compound dihydrate easily effloresces to theanhydrate in air. This anhydrate is clearly different from theanhydrous conglomerate, as is confirmed by their infraredspectra and X-ray powder patterns.(gly O>(en>,l so,.'Kinetic Optical Resolution.-Since the present complex[Co(CO,)(en),]Br forms a conglomerate above 15.0 "C, kineticoptical resolution by a chiral additive which is stereochemicallysimilar to one enantiomer of the conglomerate to be resolvedis possible. Table 1 shows the results of such resolutionexperiments. We found that two kinds of chiral additivesJ. CHEM. SOC. DALTON TRANS.1991 2905Table 1 Resolution of AA-[Co(CO,)(en),]Br in the presence of a chiral additive"Experiment Chiral additive Amount' e.e. (x) Configuration Yield ("/,)123456789100.40.8224255553.76.1122959273727212AAAAAAAAAA35332217171818222951" Crystallization was performed by allowing the solution to stand at room temperature for 1-3 d. In mg of additive per 100 mg of AA-CCo(CO,)(en),lBr.Table 2 Second-order asymmetric transformation of AA-[Co(CO,)(en),]Br in the presence of a chiral additiveExperiment Chiral additive Amount * T/"C Crystallization time/d e.e. (%) Configuration Yield (yo)1 A-[Co(ox)(en),]Br 4 40 22 ~-[Co(glYO)(en),lC1, 4 40 23 A-[Co(glyO)(en),]Br, 16 35 34 A-[Co(glyO)(en),]Br, 20 40 25 A-[Co(glyO)(en),]Br, 40 40 26 A- [ Co( gl y 0)( en),] B r , 7 5 35 3* In mg of additive per 100 mg of Ah-[Co(CO,)(en),]Br.6.2 A 814.6 A 824.5 A 1006.1 A 1007.1 A 1 0011 A 100A (or A)-[Co(ox)(en),]Br and -[Co(glyO)(en),]Br,, are veryeffective. These systems show relatively high e.e.It should benoted that the absolute configuration of the first crop is alwaysopposite to that of the chiral additive and even a small amountof chiral additive works effectively. We previously showed that ahigh e.e. is obtained when the first crystal yield is kept below ca.15% because the second crop having the opposite absoluteconfiguration to that of the first crop will appear at ca. 15-35%crystal yield and compensates the optical purity of the firstcrop., The best e.e.of [Co(CO,)(en),]Br is 59% A (17% yield) inthe case of A-[Co(ox)(en),]Br as a chiral additive, and 37% A(1 80/,, yield) in the case of A-[Co(glyO)(en),]Br,.Second-order Asymmetric Transfurmation.--In contrast tomost resolutions, which employ low temperature to keepracemization to a minimum, the technique of second-orderasymmetric transformation involves warming the solutionduring crystallization in order to induce racemization. Thepresent carbonato complex A (or A)-[Co(CO,)(en),]Brracemizes with a moderate rate above 35 "C withoutdecomposition. The half-life of racemization is t+ = 3.6 h at40 "C under neutral conditions. Therefore, when the abovekinetic resolution is carried out in 2 or 3 d at 4OCC, thecombination of racemization-crystallization is realized, leadingto the second-order asymmetric transformation.The results arecollected in Table 2. Both systems containing A-[Co(ox)(en),]-Br or A (or A)-[Co(glyO)(en),]X, (X = CI or Br) as a chiraladditive are found to be effective. The e.e. values are between ca.5 and 1 1 %. In each experiment, the absolute configuration of thetotal crystals is always opposite to that of the chiral additive,which means the operation of the same resolution mechanismas described above. The realization of a second-orderasymmetric transformation, though the e.e. values are not sohigh, should be noticeable because most other chiral additivessuch as A-[Co(NO,),(en),]Br and A-[Co(en),]Br, cannotinduce any positive effect.Some examples of diastereomercrystallization are known' but the direct crystallization of theenantiomer is extremely rare in metal complexes. 'References1 L. Addadi, J. van Mil and M. Lahav, J. Am. Chem. Soc., 1981, 103,1249; L. Addadi, S. Weinstein, E. Gati, I. Weissbuch and M. Lahav,J. Am. Chem. SOC., 1982, 104,4610.2 K. Yamanari and Y. Shimura, Inorg. Chem., 1985,24,4600.3 K. Yamanari, J. Hidaka and Y. Shimura, Bull. Chem. SOC. Jpn., 1973,46, 3724.4 J. Springborg and C. E. Schaffer, Inorg. Synth., 1973, 14, 64; F. P.Dwyer, A. M. Sargeson and I. K. Reid, J. Am. Chem. Soc., 1963,85,121 5.5 F. P. Dwyer, I. K. Reid and F. L. Garvan, J. Am. Chem. Soc., 1961,83,1285; K. Yamanari, M. Igaki and Y. Shimura, Bull. Chem. Soc. Jpn.,1984,57,3007.6 A. Werner, Chem. Ber., 1912,45, 121; J. A. Broomhead, F. P. Dwyerand J. W. Hogarth, Inorg. Synth., 1960, 6, 186; A. Fuyuhiro, K.Yamanari and Y. Shimura, Bull. Chem. Soc. Jpn., 1980,53, 3577.7 S. Kaki, K. Yamanari and Y. Shimura, Bull. Chem. Soc. Jpn., 1982,55, 769.8 J. Meisenheimer, L. Angermann and H. Holsten, Liebigs Ann. Chem.,1924,438,217.9 J. Springborg and C. E. Schaffer, Inorg. Synth., 1973, 14, 72; F. P.Dwyer and F. L. Garvan, Inorg. Synth., 1960,6, 195.10 Y. Shimura and K. Tsutsui, Bull. Chem. SOC. Jpn., 1977,50, 145.11 A. Werner, Chem. Ber., 19 12,45,3061; F. P. Dwyer and E. C. Gyarfas,Proc. R. SOC. N. S. Wales, 1949, 83, 263; G. Kauffman, N. Sugisakiand I. K. Reid, hzorg. Synth., 1989,25, 139.12 S. Muramoto and M. Shibata, Chem. Lett., 1977, 1499.Received 17th April 1991; Paper ljO1808