May 20, 195.5
2i47
High Yield Resolution of Tris-(ethylenediamine)cobalt(II1) by a Second-order Asymmetric Process
is termed a second-order asymmetric induction because a second substance, the precipitating ion, is necessary for its promotion. BY DARYLEH. BUSCH I n the instances previously reported, the comRECEIVEDDECEMBER 20, 1954 plexes involved were by nature optically labile, The resolution of tris-(ethylenediamine)-cobalt- and therefore naturally underwent this asymmetric (111) ion was first accomplished by Werner' in process. I t has now become possible to obtain 1912, and, following his procedure, numerous other such an asymmetric induction with a substance investigators and students have been able to obtain ([Co1I1en3]ion) which normally does not tend to this compound in optically pure form. The results racemize or otherwise undergo changes in configurais crystalreported here show that it is possible to obtain yields tion. As the It- [Co*lTenj]Cl(d-tartrate) in excess of 150yoof the amount of the dextro or lized from solution, the remaining I- [Corr1en3]inn is levo isomer originally present, as a result of a sec- catalytically racemized, providing mqre of the dexond-order asymmetric process. This involves cou- tro form, which then crystallizes. The process has pling the conditions for racemization with the condi- also been employed utilizing levo-tartrate with a tions for selective crystallization of one of the di- resulting high yield of 1- [C~~~~ens]CI(l-tartrate). astereoisomeric salts. Experimental Racemization of Tris-(ethylenediamine)-cobaltMaterials.-Tris-(ethylenediamine)-cobalt(III) chloride (111) Ion.-The ethylenediamine complex of co- was prepared and purified by the method of Work.' dbalt(II1) is of great configurational stability, its Tartaric acid was J. T. Baker reagent grade while the Z-tartaric solutions showing no loss in optical activity after acid was obtained from Aldrich Chemical Company. Silver was prepared in the necessary amount for each run. several weeks. Prior to the work reported here, a tartrate The specific rotations reported were calculated from obsingle method for the catalytic racemization of served rotations made on 1% solutions in a 1.01 d m . all tris-(ethylenediamine)-cobalt(II1)ion had been re- glass polarimeter tube, using a Schmidt and Haensch Polarimported. This involved the use of active carbon as a eter, Number 11698 (precision, 3=0.02"). Preparation of [ C ~ ~ ~Cl(d~ e or n ~I-tartrate).-One ] hunheterogeneous catalyst.2 dred and six grams of silver tartrate was prepared by adding The method emplqyed here is based on the elec- 99 g. of silver nitrate, dissolved in 120 ml. of water, to43.5g. tron exchange of tris- (ethylenediamine)-cobalt(III) of d-(or &tartaric acid dissolved in 117 ml. of 5 N sodium ion with tris-(ethylenediamine)-cobalt(I1) ion. The hydroxide, which had been diluted t o 200 ml. The presilver tartrate was filtered onto a 4-inch buchner kinetics of this electrnn exchange have been investi- cipitated funnel and the water was removed by thorough washing with gated. The reaction is bimolecular and essentially acetone. The acetone was then removed under suction and first order in the concentration of both complex the product was used immediately. The silver tartrate was triturated for 15 minutes in a 6ions. Since the cobalt(I1) complex is configurainch mortar with a hot (90") solution of 100 g. of tris-(ethyltively unstable, the path of racemization is consid- enediamine)-cobalt (111) chloride in 400 ml. of water. The ered to involve conversion of the d-[Co111en3]ions precipitated silver chloride was removed by filtration and into d- [Co11en3] ions, which racemize instantly. washed with 100- to 150-ml. portions of hot water (90') unThe simultaneously formed [Co111en3]ions sup- til the filtrate emerging from the funnel was colorless. The volume of the combined filtrate and washings was approxiposedly have an equal probability of being either mately 800 ml. This solution then contained 123 g. of trisdextro or The net effect is racemization of the (ethylenediamine)-cobalt(II1)chloride d (or I-)-tartrate. Resolution.-The solution obtained by the method deactive cobalt(1II) complex by a process of homogeneous catalysis. The racemization process may be scribed above was placed in an evaporating dish and concentrated in a stream of air for approximately 24 hours. terminated at any point by acidifying the solution The crystals of active complex which formed were separated since this destroys the cobalt(I1) complex and by filtration. The mother liquor was again concentrated for approximately 3 hours and cooled. Any additional thereby prevents further electron exchange. solid was then removed by filtration. Yield of crude active Second-order Asymmetric Induction.-Werner5 complex was 60 to 70 g. (This represents more than the theoand Dwyer and Gyarfasa have shown that when one retical yield; however, the excess is due to the presence of of the diastereoisomers of a salt of an optically a considerable amount of racemic material.) Recrystallilabile complex ion is much less soluble than the zation from approximately 200 ml. of hot water (90") other, virtually all of the complex may be obtained yielded 45 to 50 g. of o p $ a l l y p2re complex; reported ] n l O l f , f o u n d [ a ] o + 1 0 2 , -103 . as the less soluble form. This occurs because the [ aThe viscous mother liquor (volume 75 t o 100 ml.) from more soluble isomer racemizes, as the less soluble the separation of the active form was placed in a flask and isomer crystallizes, thus producing more of the less oxygen-free nitrogen was passed through it for I5 minutes. soluble form, which then crystallizes. The process Three and one-half milliliters of 95% ethylenediamine and one gram of cobalt(I1) chloride &hydrate were then added (1) A. Werner, Bcr., 45, 121 (1912). ( 2 ) B. Douglas, THIS J O U R N A L , 76, 1020 (1954). (31 W. B. Lewis, C. D. Coryelland J. W. Irvine, Jr., J. Chem. Sac.,
5386 (1949). (4) The kinetics of racemization of active [Co"'enrl
ion, by this catalytic process, are now under investigation. The results may clarify this point. ( 5 ) A . Werner, Bcr.. 45. 3061 (1912). , , (6) F. P. Dwyer and K . C . Gyarfas, J . Proc. Roy. SOC.N . S . Wales, 83, 203 (1950).
and the solution was maintained under a stream of nitrogen for 12 hours. During this time a considerable amount of the active complex crystallized. Yields of recrystallized c o y plex by second-order process were 25 t o 30 g., [a]+IO1 , - 101 '. T h e mother liquor from this process was placed under nitrogen as rapidly as possible to limit the oxidation of the cobalt(I1) complex. Concentrating and cooling the filtrates from recrystalliza(7) J. B. Work, Inorganic S y n f h e s c s , 2 , 221 (1916).
2748
NOTES
tion of the active fractions obtained by the initial resolution and by the second-order asymmetric process yielded approximately 10 g. of active product: [a]!, +l02", -101". The filtrates from this third active fraction were then concentrated to a small volume (20 to 30 ml.) and added t o the solution in which the asymmetric process was occurring. After standing a n additional 24 hours, this solution yielded a further fraction giving the same rotation as those previously isolated. After recrystallization from a small volume of ht; watero, this fraction amounted to about 10 g., l 4 D +lo2 , -101. * I n order t o obtain a n approximate material balance, the remaining solution was evaporated to dryness and weighed. This material was essentially racemic. Total recovery of material was approxiriiately 95%. Total yield of active complex was 90 to 95 g. or 74 to 77%, based on total amount of complex (148 t o 154%, based on the amount of the isomer originally present.) From 76 to 79y0 of the total material recovered was optically pure complex of one configuration. T h e ultimate yield obtained by the method of resolution described here is determined by the volume of the solution in which the asymmetric process occurs. The quantities of the dextro and levo isomers remaining after as much active complex as possible has been removed are the same, this amount corresponding to a saturated solution of the least soluble diastereoisomer. The volume of the solution is determined by the first step in the resolution. Concentration of the original solution t o facilitate crpstallization of the least soluble form results in a very viscous solution which eventually begins to form a gelatinous mass. This limits the extent t o which the volume may be reduced. I t is, of course, theoretically possible to obtain R 100% yield in a process of this sort; however, the steps which would be necessary in order to obtain the final 30 t o 25% of the complex in the desired active form are not sufficiently convenient to be justified. THEDEPARTMENT OF CHEMISTRY THEOHIOSTATEUNIVERSITY COLUMBUS, OHIO
Vol. 77
characteristics were produced, it was decided to investigate this phenomenon.8 Experimental A . Reagents.--A standard solution of 0.103 Ji copper nitrate was prepared from reagent grade Cu(NOs)2. One drop of C.P. concd. nitric acid was added to 4 liters of the solution t o prevent hydrolysis. The molarity of the solution was determined by electrodeposition. Commercial ethylenediamine after drying over C.P. sodium hydroxide and then over sodium metal, was fractiorially distilled from sodium metal (b.p. 119', lit. b.p. ethylenediamine hydrate, 118"). The purified product was used t o prepare an nqueous 0.1128 Ji solution of ethylenediamirle. The concentration of ethylenediamine was determined by titration with standard HC1, using methyl orange indicator. All solutions of [Cu(en)]Z+and [Cu(en)2I2+were prepared from pipetted volumes of the respective standard solutions. B. Spectrophotometric Studies.-Scanning curves in the visible region were made with a Beckman Model B spectrophotometer using 1-cm. Corex cells. B y the addition of potassium nitrate solution a concalculated volumes of 1 stant ionic strength of 0.5 was maintained in all of the spectrophotometric measurements. The effect of sodium hydroxide concentration on the absorption spectra of the [Cu(en),]*+ion vias observed while varying the base concentration from 0.05 to 1 X. C. Conductometric Titrations.-An Industrial Instrument Co. 1000-cycle conductivity bridge with a dipping typc cell with vertical electrodes was used. The temperature during titration was held at 25 =t1" in a regulated waterbath. The calculated conductances were corrected for dilution.
Results and Discussion Studies.-Jolley9 reported that an increase in sodium hydroxide concentration from 0.1 to 0.9 31 increased the solubility of the C U ( O H ) in ~ ethylenediamine solution because i t appeared to favor the formation of Inorganic Complex Compounds Containing Poly- [ C ~ ( e n ) ] ( 0 H a) ~ t the expense of [Cu(en)z](OH)2. dentate Groups. XI. Effect of Hydroxide Ion on Jonassen and D e ~ t e r however, ,~ showed that the the Bis-ethylenediaminecopper(I1)Ion addition of a slight excess of sodium hydroxide to BY HANSB. JONASSEN, RICHARDE. REEVES A N D LEON [Cu(en)I2+results in the precipitation of Cu(OH)2 and the formation of the [Cu(en),I2+ion. SEGAL~ The effect of sodium hydroxide on the [Cu(en)t]*+ RECEIVED DECEMBER13, 1954 ion is shown in the change in absorption characThe reaction of copper(I1) ions with ethylene- teristics of the [ C ~ ( e n ) ~ ]complex ?+ in increasing diamine has been studied in the past by Traube,2a concentration of sodium hydroxide (Fig. 1). The Chattaway and Drewzb and Job3 who concluded presence of an isosbestic point a t 590 mp clearly that only the [Cu(en)zI2+ ion existed in solution. indicates that a second complex ion is formed. Manda14 doubted the results of Chattaway and Whereas Jolley claimed that [Cu(en)]?+ was obDrew, and in a later paper Carlson, McReynolds tained, the curves of Fig. 1 show that even a t an and Verhoeks accepted the existence of the [Cu- hydroxide ion concentration of 0.G M the resulting (en)I2+and [ C ~ ( e n ) ~ions. ] ~ + Bjerrum and Niel- absorption curve does not approach that of the son6 presented spectrophotometric evidence for [Cu(en)],+ ion (also given in Fig. 1). One possible these two ions, and their existence was conclusively explanation of these data is that under these demonstrated by Jonassen and Dexter7 using extreme conditions of basicity one hydroxide ion spectrophotometric and conductometric titration adds to [Cu(en),]'+ to give a pentacovalent copper studies. (11) complex. Since i t has been observed that upon addition of These conditions are as extreme as those under large excess of hydroxide ion, changes in absorption which Bjerrum and Nielson6 obtained [Cu(en),lz++. (1) Abstracted in part from a dissertation submitted b y Leon Segal Other investigations also seem to favor such a to t h e Tulane University in partial fulfillment of the requirements for pentacovalent structure for copper (11) complexes.lo t h e degree of Doctor of Philosophy. The results of continuous variation and saturation (2) (a) W. Traube, Ber., 54, 3220 (1921); ( b ) F. W. Chattaway and studies were meaningless since the complex is very H. D . K. Drew, J . Chem. Soc., 917 (1937); ( c ) R. hl. Levy and P. unstable. M u 5 a t , P a p e r T r a d e J . , 118, T40 (1944). (3) P Job, A n s . Chim.,[lo] 2 , 113 (1928). B. Conductometric Studies.-In Fig. 2 the first (4) K. L. hlandal, Current Science, 10,7 8 (1911). break in the conductance curve a t a Cu/en ratio ( 5 ) G . A . Carlson, J P. McReynolds and F. H. Verhoek. THIS 67, 1334 (1945). (6) J. Bierriim and E . 1. r i e l s o n , A r l o Chem. S r a n d . , 2 , 297 (1948). (7) H. B. Jonaysen and T. 11. Dexter, T H I S J O U R N A L , 71, 1553 (1949).
JOURNAL,
A. Spectrophotometric
(8) H. B. Jonassen, R. E. Reeves and L. Segal, i b i d . , 7 7 , 2007 (1955). (9) 1,. J Jolley, J . T e x t i l e I n s ! . , SO, T22 (1939). (10) H. B. Jonassen, M . hl. Cook and J. S. Wilson, THISJ O U R N A L , IS, 4683 (1951).
