The Stereochemistry of Metal Chelates with Polydentate Ligands. Part I

Part I. By BasudebDas Sarma and. John C. Bailar, Jr. Received April 25, 1955. This paper describes bis-salicylaldehyde-triethylenetetramine, the nitro...
0 downloads 0 Views 703KB Size
Download PDF
BASUDEBDASSARMA AND JOHN C. BAILAR,JR.

5476

ally are considerably weakened by the interposition of even a single methylene group between charged site and reaction site. The charge on an N + might orient an approaching dipolar molecule of butyl bromide into an unfavorable position for reaction with a neighbor, or possibly the field of a nearby N+can stabilize the 3 N.BrBu complex in its transition state, so that decomposition into > N + B u and Br' becomes less probable. Finally, the intense field (megavolts/cm.) of a quaternized nitrogen might increase the density of the highly polar solvent in its neighborhood by electrostriction, and

[CONTRIBUTION FROM

THE

Vol. 77

thereby make access to the region by alkyl halide more difficult. One interesting consequence of the decelerating effect of partial quaternization should be mentioned. If there were no interaction, the distribution of quaternized sites among the chains would be a Bernouilli distribution around an average degree of quaternization. The interaction sharpens the distribution, because chains with more quaternized sites than the average will react more slowly with further alkyl halide, while those with less than the average will react more rapidly. NEW HAVEN,COSN.

NOYES LABORATORY O F CHEMISTRY, UNIVERSITY O F ILLINOIS]

The Stereochemistry of Metal Chelates with Polydentate Ligands. Part I BY BASUDEB D.4s SARMA AND JOHN C. SAILAR, JR. RECEIVED APRIL25. 1955 This paper describes bis-salicylaldehyde-triethylenetetramine,the nitrogen analog of Dwyer and Lions' sexaclentate 1,8-bis-(salicylideneamino)-3,6-dithooctane, as a chelating agent. The Schiff base offers six polnts of attachmelit and gives well defined, stable compounds with Cot3, CO'~, Fe+3, Fe+2,Al+3, Cu+*, Xif2, P d f 2 and the ions of other transition , Al+3 have been obtained in their optically active forms, only two of the metals. The octahedral complexes of C O + ~Fe+3, eight possible isomers being obtained. The copper compound, which is tetra-coordinated and planar, has been obtained in optically active forms through a method of asymmetric synthesis. The tentative structures of the complexes have been described from the steric point of view and their infrared spectra The Fe+3compound is covalent with d2sp3bonding and this has been resolved directly into the optical isomers.

Since the work of Dwyer and Lions' there has ally planar. If we assume that the two protons been a number of publications on complexes with are replaced from the hydroxyl groups by cupric sexadentate chelating agents.? The sexadentate ion (like those of other Schiff bases, the compounds formed with a bivalent metal are non-elecligand (I) HOC6H4CH=r\'CHzCHzNHCH2CH2NHCHzCHzX=CHC8H40H trolytes), and two of the four remaining nitro(1) (2) (3) (4) (5) (a) gens are coordinated to it, the structure may span the octahedral positions around a metal ion in the following ways I

CH?

I

Cnt'

CH?

I

';

I

1 5,

ii

f P kHCH2CH?hTT "CH-

l i

I4

03 CJ14 /

b 5

L)

rj3

together with the mirror images of each, as they are all asymmetric structures. But Fisher-Hirschfelder models show that the structures B, C and D are greatly strained and the models for them cannot be made. The angle N-N-0 tends toward 90°, as is the case with A. In all probability, therefore, these compounds have the structure represented by A. The picture is slightly changed when all of the six coordinating positions are not utilized. For example, the cupric ion is normally reluctant to extend its coordination number of four, and is gener(1) F. P. Dwyer and F. Lions,THISJ O U R N A L , 69, 2917 (1947). ( 2 ) F. P. Dwyer, F. Lions and co-workers, ibid., 7 2 , 1545 (1950); 73, 5037 (1950); 7 4 , 4188 (1952); 7 5 , 2443 (1953). G. Schwarzen-

bach and H. Ackermann, H e l v . Chim Acta, 31, 1029 (19481. Busch and J. C. Bailar, Jr., THTS JorrnLYa~,7 5 , 457-1 (14533 Mnkherjee, ,?&we awd C x i i t , r r , 19. 107 flR54)

D. €1. .4. K .

is still not symmetrical as the two ben7erie rings cannot lie in the plane of the complex. This gives rise to two active forms and of course, a meso-form. The classical method of resolution cannot be applied here, the chelate being a non-electrolyte. But if the compound is made from an optically active copper complex, such as the d-tartrate, it is obtained in optically active form. \lTe have not yet been able to prepare the meso-form. There is an alternate structure for this compound, with copper coordinated to NBand N4 instead of N? and N6. This will require the stabilization of two nine-membered rings, instead of one large ring and two six-membered rings. This alternate structure is more difficult to make with a model. The infrared spectra of these compounds (Table I) show that when the secondary amine nitrogens, N3 and N,, are coordinated, there is a multiplicity of the N H stretching vibrations in the region 3100 Xi00

STEREOCHEMISTRY OF METALCHELATES WITH POLYDENTATE LIGANDS

Yov. ri, 1955

5477

cm.-', due to a salt-like character of the secondary splitting of the ether band has been observed with C-0-C-C-0-C in glycol dialkyl ethers. There is a amine. grouping in all these compounds Unlike the spectra of the complexes of C O + ~ , C-N-C-C-N-C group in the tris-salicylaldeFe+3, Al+3, those of the CuII and Co'I complexes and a -N-C-C-N\ / show only one band in the same region, indicating C that these nitrogens are not coordinated in the hyde base. The relative intensities of the three latter series. The infrared spectra of the complexes and that bands seem to show some correlation with the of the tris-salicylaldehyde-triethylenetetraminewere type of compound. In the Schiff base TS3H3 the band a t 1123 recorded with an Perkin-Elmer auto-recording infrared spectrophotometer using a sodium chloride ern.-' is the strongest, while it is the weakest one crystal grating. The samples were run in nujol in all of the complexes. The one a t 1192-1207 emulsions. From a study of them and the existing cm --I is strongest for the definitely hexa-coordiliterature, some bands have been assigned to spe- nated complexes (Cof3, F e 9 , is almost of the same intensity in Al+3 and Ni+2 complexes, and cific groups, as shown in Table I. weaker than the band a t 1152-1159 cm.-' for the TABLE I tetra-coordinated complexes ( C U + ~C, O + ~ ) . The C-0 stretching vibration band a t 1265 ern.-' INFRARED BAXDS SHOWN BY TRIS-SALICYLIDENEAMINOshifts t o a higher wave number in the complexes, TRIBTHYLENETETRAMINE A N D SOME METALDERIVATIVES OF RIS-SALICYLIDESEAMINOTRIETHYLENETETRAMINEindicating possibly a slightly stronger than normal single bond in C-0, presumably by a resonance in the ring C=N (str.) 0-H (str ) (phenolic) x-H (str )

C-X (str.)

C-0 (str.) 0-substd. benzene C=C

C-H (Wag.)

C=C (aromatic)

..

..

..

3340

..

..

.,

3620 3480 3408 3318 3160 3100 1205 1155 1131 1314 1605 1542

3610 3510 3380 3170 3100

3485 3150 3220 3430 3300 3285 3120

1205 1154 1132 1300 1602 1542

1207 1156 1133 1315 1607 1553

1182 1152 1123 1265 1615 158.5 1503 761 762 755 755 1488 1475

1197 1153 1131 1350 1604 1538

1192 1152 1129 1344 1602 1535

758 757 755 743 1475 1480 1475 1473

As our knowledge of the infrared spectra of coordination compounds is still meager, the above assignments should be accepted as tentative. But some observations in this series are interesting. The C=N band of the Schiff base is very little changed on coordination of a metal to the nitrogen, which, however, is not surprising in view of the stronger force existing in the C=N group. But while the N-H stretching band is definitely multiplied in the hexa-coordinated compounds as compared to those where the secondary amine nitrogen is not coordinated, the band a t 1152 cm.-', assigned to the H2CN stretching vibration, remains practically unchanged om coordination to various metals. In the region 1100-1200 em.-', there are two bands in addition to 1152-1159 cm.-' assigned to C-N stretching vibration. I t has been shown by Bergmann and Pinchas3 that there are three bands in the region 1063-1190 cm.-' in the acetals and ketals in place of a strong band for C-0-C in ethers a t 1130 cm.-l. This is accounted for by the three types of vibrations of C-0-C-0-C grouping in acetals and ketals. According to them, no such (3) 12. D. Bergmann and S. Pinchas, Rec. f r a u . c h i m . , 71, 161 (1951).

I t has been shown already4 that salicylaldehyde reacts with triethylenetetramine t o give tris-salicylideneaminotriethylenetetramine (11)

CH.CsHa.OH

which, however, loses a molecule of salicylaldehyde when it coordinates to metal ions. Examination of models shows that when positions (3) and (4) are coordinated to a metal, the salicylaldehyde across them has to split off. Salicylaldehyde is actually liberated by the reaction of a cobalt salt and the tris-salicylidene compound in the presence of air. In the case of four-coordinate metals like Pd+2, C U +and ~ Co+?, the third salicylaldehyde may remain in place if positions (l), ( 2 ) , (5) and (6) are utilized in coordination, but the structure is still somewhat strained. The reaction product with Pd+2 could not be purified, but the product approaches the calculated composition for a molecule containing triethylenetetramine condensed to three salicylaldehyde molecules. Infrared spectra for this product indicate that there is no NH band and the presence of OH is exhibited.

Experimental 11.-As

Mukherjeezd has found, attempts to prepare the sexadentate ligand I by the condensation of triethylenetetramine ( 1 mole) and salicylaldehyde ( 2 moles) in different solvents and under different conditions lead only to product 11. I n the presence of sodium acetate, however, another product was obtained, which did not contain any nitrogen and could not be identified. I1 can be obtained in best yield by the slow addition of 40 g. of 71 .5y0triethylenetetramine (0.2 mole) in 100 cc. of cold methanol to a cooled solution of 73.2 g. of salicylaldehyde (3 moles) in 300 cc. of methanol. When the mixture is gradually warmed to room temperature, yellow crystals separate. The crude yellow (4) B. Das Sarma and J. C. Bailar, Jr., THISJOURNAL, 76, 4051 (1954).

5478

R ~ s r : n s n DASSARMA AND

JOFIN

C. BATLAR, JR.

1'01. 7 ;

Calcd. for [ C O ~ ~ ~ ( C ~ O H ~ ~ N ~ C, O ~49.91; ) ] ~ S 11, .~H~O: product can be recrystallized from TI arm absolute alcohol or dilute methanol. I t melt, .it lotj-106", i i highly soluble 6.23; IT, 11.64. Found, C, 49.92; H, ;i.8R; If, 11.11. in acetone, c&-ly ~ ) I i i b l ri i i itictliciiiul .iiid prcictically i n Even on heating on the steam-bath in the presence d c\soluble in mater; yield 30%. cess sulfide, very little cobalt sulfide was formed. Resolutions into Optically Active Forms.-The rotations Anal. Calcti. for C2,HarIOlS4:C, 70.7'4; H, 6.35; S, recorded in this paper were taken on aqueous solutions ii? a 12.23. Foulid: C , 70.3ti; I I , (j.47; S , 12.32. Attempts to prcpur-c I I,?. t1ie rcmclval o f cop1,cr sulfidc one-dm. tube, using a high-precision polarimeter reatling from a meth:inol solution of tlie corrcspoiidiiig copper c'otn- to 0.001". All readings were taken a t the D-line of sodium unless otlicrwise stated. pound were urisuccessful. The resolution was attempted tlirough d-tartrate, d-caiii111. Bis-salicylaldehyde -triethylenetetraminecobalt( 11). pliorsulfonate, d-bromocamphorsulfonate and by adsorption [CoirT&].-Cobalt( 11) bis-salicylaldehyde was prepared by d-quartz. the addition of 11.9 g. cobalt chloride hexahydrate in 100 on There was no effective resolution as the d-tartrate or dcc. of water to 12.2 g . of salicylaldehyde in 100 cc. of meth- camphorsulfonate. anol, followed by the dropwise addition of 5.6 g. of caustic With d-Bromocamphor-r-sulfonate.-A sample of 0.5 g. of potash in 50 cc. of cold water (with constant stirring). The the complex cobalt chloride was dissolved in methanol (20 yellow brown product was washed with water and recrys- cc.) and ground with 0.43 g. of silver d-bromocamphtrrsultallized from hot dilute ethanol and dried. fonate. The silver chloride was filtered and the filtrate Seven and a half grams of cobalt( 11)-bis-salicylaldehyde was concentrated in 'L'UCUO. When the volume was 5-6 cc. (0.025 mole) was dissolved in 30 cc. of hot methanol and 50 a dark oil separated. The solution was decanted, evapoCC. of a 7.3% solution of trimethylenetetramine (0.025 mole) and taken up with a minimum quantity of water ( a ) . was added to it. The flask containing the mixture was im- rated The oil obtained previously was also dried in vacuo and dismediately closed and the contents shaken well. The light solved in water (b). The least soluble fraction from ( b ) yelloivish-brown product which separated was collected by the most soluble fraction from (a) were dissolved i i i filtration, washed thoroughly with water and recrystallized and and precipitated with a saturated solution of potasfrom warm ethanol, taking care to avoid excessive exposure water sium iodide, filtered, washed with cold water and dried; t o air. A little of i t was oxidized during recrystallization, iodide from the least soluble fraction (0.01% solution), 01 but this did not contaminate the product, for the oxidized -0.023 f 0.003", [ a ] D -230"; iodide from the most compound is soluble in both water and ethanol and was solublefraction, 01 +0.022 rt 0.003", [ a ]+220". ~ filtered off as a red brown solution. The product was dried With d-Antimonyl Tartrate.-To a solution of 1.0 g. of the in a vacuum at room temperature. I t is insoluble in water complex chloride in 50 cc. of methanol was added 1 g . of and soluble in alcohols. silver d-antimonyl tartrate. The mixture was stirred for Anal. Calcd. for [Co11(C2~H:~4SI02)] 0.5HzO: C, 57.13; an hour and filtered. On concentration of the filtrate, the H, 5.95; A-, 13.33; Co, 14.03. Found: C, 57.54; H, 6.14; first fraction contained a little silver salt and was rejected. N,13.29; Co, 14.15. The second fraction (about 0.3 9.) was collected and dissolved in water. On fractional crystallization, this gave The magnetic susceptibility, measured with a modified about 0.1 g. of ~-[CoTS*]-d-SbOtart. For the dextro fracCurie balance, was found to be 2.01 B.M. tion of the complex, the filtrate, after collecting the first IV. Bis-salicylaldehyde-Trimethylenetetraminecobalt fraction from methanol solution, was concentrated until (111) Salts, [ C O ~ ~ ~ T S J X . - - T can ~ ~be S ~prepared in three 0.8-0.9 g. of the complex separated. The filtrate from this different ways. A . T o a suspension of 4.1 g. of the cobalt (11) complex operation was evaporated to dryness, dissolved in water and (0.01 mole) in 100 cc. of water 10 cc. of iV hydrochloric acid precipitated as the iodide; least soluble fraction as iodide -250", most was added gradually while air n a s passed through the sus- (0.01% solution), (Y -0.025 f 0.003", [.ID soluble fraction, as iodide (0.01% solution), N $0.026 =t pension (12 hours). During this time the Co(II1) complex went into solution. The deep red-brown solution was fil- 0.003", [ a ] D $260". With d-Quartz.-A solution of the complex chloride in 50 tered and concentrated at room temperature. The deep brown-black hexagonal prisms which separated were col- cc. of water was stirred for 15-20 minutes with 1 g. of finely powdered d-quartz. The mixture was filtered and the lected by filtration and dried in I'QCUO, yield 4 g. B. The following were mixed in the order given in quick filtrate was diluted until its rotation could be read in the succession: (1) 10 cc. of 71.5% triethylenetetramine (0.05 polarimeter. The concentration, as obtained by estimation mole) in 50 cc. cold methanol; (2) 12.1 g. of salicylaldehyde of the chlorine content mas about 0.01%; 01 -0.030 k (0.1 mole) in 50 cc. cold methanol; (3) 5.6 g. of caustic 0.003', [ o ~ ] D -300". The quartz was then shaken with methanol. The solupotash (0.1 mole) in 20 cc cold nater; ( 4 ) 11.9 g. of cobalt tion gave a rotation of $0.015 + 0.002". chloride hexahydrate (0.05 mole) in 75 cc. of cold mater. V. Bis-salicylaldehyde-Triethylenetetraminecopper(I1~. The mixture was then oxidized by passing a strong current of air through it for one hour, after which 10 cc. of 1 : 1 HC1 [ C U T S ].-The copper compound was made by the addition of triethylenetetramine ( 4 g. of a 71.5% solution) t o a \vas added. The aeration xias continued for 12 hours and suspension of 6 g. of copper bis-salicylaldehyde in a mixture the dark solution was filtered, concentrated and cooled. Dark brown-black crystals separated and ITere dried as be- of 100 cc. of water and 100 cc. of methanol. The bluish green solution was concentrated and cooled. The bluish fore; yield 18.9 g. C. A solution of 9.4 E. of I 10.02 mole) in 100 cc. of ace- green prismatic needles which separated wexe recrystallized tone was added to a sol;tion of 4.8 g . of'CoClz,6H?O (0.02 from mater containing about 25% methanol. The product mole) in 50 cc. of water. A little activated carbon \vas melted a t 78-79', but lost some water in vacuo a t room temadded. The mixture was oxidized with a current of air for perature to yield green crystals, which decomposed. levo-Compound.-To a solution of 8.5 g. of copper chloride 5-6 hours and 20 cc. of 2 N caustic soda was added. Aeration was continued for another two hours, the mixture was dihydrate (0.05 mole) in 100 cc. of water was added 34.5 cooled in an ice-bath, and filtered. The complex chloride, g. of sodium d-tartrate (0.15 mole). Copper tartrate pretogether with carbon, was obtained as the residue. This cipitated. The mixture mas cooled in ice arid cold 2 -11 was treated with 100 cc. of hot water and filtered. On con- KOH was added with constant stirring until the green cot>centration and cooling, the filtrate gave brown-black crys- per tartrate went into solution with a deep blue color. Til tals of the complex chloride. These products were a11 this solution was added a cold solution of 12.2 g. of sulic>.laldehyde (0.1 mole), 10 cc. of 71.5y0 triethylenetetrantine found to be diamagnetic as expected. (0.05 mole) and 5.6 g. of caustic potash (0.1 mole) i i i 150 d n u l . Calcd. for [CoT1r(C~ilH2aSr0*)1C1~2.5H~0: C , cc. of methanol and 25 cc. of water. The mixture ~ Y X ; 49.00; €1, 5.90; IT, 11.43; Co, 12.01; C1, 7.21. Found: stirred and kept in ice for two hours and was then filtered C, 49.79, 49.73; H, 5.35, 5.86; N, 11.49, 11.13; Co, from the separated sodium and potassium salts. Three 12.03, 12.06; C1, 7.15, 7.28. hundred cubic centimeters of methanol was added to tlii. The material melts a t 240". I t has a red-brown color in filtrate, which was filtered again if a precipitate formed. concentrated aqueous solution, and greenish brown in When the solution was concentrated by a stream of air, the dilute solution. It was found to be quite stable in aqueous copper compound separated as greenish blue crystals. The solution, as an xtternpt to rlecmlposc the comnlex with C Y product was collected by filtration, nashrd with a little cold cess of aqueous sodium sulfide gave practically tin cobalt water and recrystallized from 25Yc i n c t h a ~ ~ o l It . t~~r~ictl sulfide, but the complex sulfidc crystallized out. green in a desiccator.

Nov. 5 , 195.7

STEREOCHEMISTRY OF R f E T A L C H E L A T E S W I T H POLYDENTATE L I G A N D S

A n d . Calcd. for [CUII(C~OHNN~OZ)] : C, 57.76; H , 5.78; N, 13.47; CuO, 19.13. Found: C, 58.23; H , 6.16; S ,18.13; CuO, 19.24. A 0.04% solution in water gave a rotation of -0.020 f 0.002" against the solution of the racemic compound as blank; [CY]D-50". Another fraction of the copper complex made from disodium copper( 11) bis-levo-glutamate, instead of copper dtartrate, gave a rotation of f0.032 f 0.003' for a 0.05% solution; [CY]D$64'. It is interesting to note that I-glutamic acid apparently has a structure opposite to that of dtartaric acid and, by these asymmetric processes, give the copper isomers of opposite rotations. &Tartaric acid, dtartrates and copper &tartrate all have dextrorotation. levo-Glutamic acid is dextrorotatory.

5479

or L-[A~TSZ] d-SbOtart was found to give dextrorotation by reaction with dilute hydrochloric acid. The difficulty experienced in obtaining a higher activity of the dextro fraction was due to the low solubility and slow rate of dissolution. Repeated fractionation, therefore, could not be carried out. Both isomers of [AlTS2]NO3 undcrwent rapid raccmization in mcthanol (half-life 1.5 to 4 hours depending oti the solvent and other factors). A solution of rac-[AITSz]-d-SbOtart in methanol was found to change rotation with time, while D- and ~-[AlTS?]-d-SbOtartwere found to be very slow in racemization. In the solid form, they were found to retain about 5 0 4 0 % of their activity after seven months. The rate of racemization of all the aluminum compounds in aqueous solution was too rapid to follow. The solids as VI. Bis-salicylaldehyde-Triethylenetetraminealuminum- well as the solutions slowly turn yellowish on exposure to air (111) Iodide, [AlTS2]I.-As in the preparation of the co- and possibly also to light. The solutions of the active nitrates should be kept in the dark and out of contact with air. balt( 111) compound (method B), the following were quickly If the slight excess of silver salts present in these solutions mixed in the order given: (a) 20 g. of 71.5% triethylenetetramine (0.1 mole) in 75 cc. of methanol; (b) 24 g. of salicyl- be removed there is no need to keep them in the dark. The aldehyde (0.2 mole) in 100 cc. of cold methanol; (c) 11.2 g. details of the rate and possible mechanism of racemization of caustic potash (0.2 mole) in 50 cc. of cold water; (d) 37.5 will be published in a separate paper. VI. Bis-salicylaldehyde-Triethylenetetramineiron(111) g. of Al(N03)3,9H~O(0.1 mole) in a mixture of 200 cc. of Compounds [Fe'I'T&]X.--These were made in the same water and 100 cc. of methanol. way as the corresponding aluminum and cobalt(II1) comThe mixture was stirred for one hour and the white prepounds (method B), by the reaction of ferric nitrate hexacipitate filtered off. To the cooled filtrate was then added a saturated solution of potassium iodide, whereupon the com- hydrate (25 g. in 100 cc. of water), and a cold mixture of triethylenetetramine (10 g. of 71.5% solution in 100 cc. of plex iodide separated. This was filtered and washed with 50% methanol), salicylaldehyde (12.2 g. in 100 cc. of methacetone. The compound was purified by redissolving in warm water (the solution should not be boiled as there is a anol) and potassium hydroxide (5.6 g. in 20 cc. of cold slow hydrolysis) and reprecipitated by the addition of po- water). On cooling and concentration, dark purple plates of the complex nitrate separated. This compound was retassium iodide solution and cooling in ice. The product from warm water and dried in vacuo. It was was washed first with cold water and then with acetone and crystallized dried in vacuo. It forms white crystals, with a greenish found to be more soluble in alcohol than in water. Anal. Calcd. for [Fe(CzoH~tNdO~)] N03.HzO: C, 49.18; tinge when the crystals are large. The iodide was found to be fairly soluble in water, very soluble in methanol and H, 5.33; N, 14.34; Fez03, 16.40. Found: C, 49.04; H , 5.23; N, 13.76; Fez03, 16.11. slightly soluble in acetone. To a solution of the nitrate in warm water was added a Instead of precipitating the less soluble iodide with potassaturated solution of potassium iodide. On cooling in ice, sium iodide, the nitrate can be isolated as hexagonal prisms by slow concentration a t room temperature. Some of the shining, almost black crystals of the complex iodide sepacrystals obtained in this way were as large as 0.5 cm. in rated. This material was washed with cold water and dried in air. It gave a deep purple solution. length. Anal. Calcd. -for [F~(C~OH~PN~OZ)]I.~.~HZO: C, 42.70; Anal. Calcd. for [Al(C Z O H Z ~ N ~1.0.5H20: OZ)] C, 46.6; H , 4.85; N, 10.87; Al, 5.24. Found: C, 46.59; H , 4.77; H,4.80; N,9.99; Fe203, 14.14. Found: C, 42.99; H , 5.38; N, 9.99; FezOa, 14.14. S, 10.87; Al, 5.13. The iodide melts a t 117-118'. The magnetic susceptiResolution of the Aluminum Complex.-One gram of the Genbility was found to be 1.81 Bohr magnetons at 27'. complex iodide in 50 cc. of methanol was allowed to react erally FelI1 complexes give a much higher value6 compared with 1 g. of silver-d-antimonyl tartrate, filtered, washed to the theoretical 1.73 for one unpaired electron. Only in with methanol and the filtrate and washings concentrated zn vacuo. There was a separation of solid (0.2 g.) when the rare cases like Fe"'-heme is the value so close to theory. This suggests a strong covalent d%ps bonding. volume was reduced to 20 cc. This is D-[AlT&]-d-SbOtart. Resolution of most optical isomers of Fer" compounds A 0.25% solution in methanol gave CY +0.112 f 0.002"; cannot be obtained directly,6 but only through oxidation of [a]D +44.8'. F e + 2 complexes. This complex, howAnal. Calcd. for [ A ~ ( C ~ ~ H ~ ~ N ~ O Z ) C,] 43.30; S ~ C ~ Hthe ~ Ooptically ~has : beenactive ever, resolved into optical isomers directly. H , 4.21: N, 8.42; AbO3 Sb& 29.60. Found: C, One gram of th'e complex iodide was dissolved in 100 cc. 43.25; H , 4.81; N, 8.76; A1203 Sbz03, 29.30. of methanol and 1 g. of silver d-antimony tartrate was added D-[AlT&] Nos, obtained from the methanol solution of to it. The mixture was stirred for 0.5 hour, filtered and the complex antimonyl tartrate and slightly greater than the concentrated in vacuo. The first fraction (A) was collected equivalent quantity of silver nitrate, gave [ a ] ~ +20' as shining black crystals (0.7 g.), rejecting the scales formed (0 167% solution in methanol). on the sides of the vessel. No more solids separated when The filtrate, after separation of the first fraction, was the volume of the filtrate was concentrated to 20 cc. This concentrated to 1-2 cc. in vacuo, a t e r e d , diluted with 25 cc. was then evaporated to dryness in vacuo (residue B). of methanol and again concentrated to 2 cc. This was re( A ) and ( B ) were dissolved separately in 50-cc. portions peated twice. The final filtrate was evaporated to dryness of methanol, and the least soluble fraction from (A) and i n vacuo. most soluble fraction from (B) collected. 0.086 g. of this ~-[AlTSz]d-SbOtartwas dissolved in 20 ( A ) Least Soluble Fraction, L-[FeTS]-dSb0tart.-A cc. of methanol and the rotation observed; CY -0.255 f 0.005% solution in methanol gave a rotation of -0.005 =k 0.002' (concentration 0.43%), [ a ]-59.3'. ~ 0.003', [ a ]-100'. ~ Calcd. for [Fe(CznH24N10~)]SbC4H401:C, 41.5; Anal. Calcd. for [Al(C ? O H Z ~ N ~ O Z ) ] S ~ C ~ H ~C, O T H ~Anal. O: H, 4.04; N, 8.07. Found: C, 41.87; H, 4.17; N, 7.03. 42.16; H , 4.39; N, 8.21; AlZO3 Sb,Os, 28.82. Found: C, 41.81; H , 4.43; N, 7.92; AlzOs SbzOs,28.47. L - [ F ~ T & ] N Oobtained ~, by the action of silver nitrate on the solution of the complex antimonyl tartrate (0.!033% Silver nitrate (0.022 g.) was dissolved in a drop of water and the methanol solution (20 cc., 0.43%) of the complex in methanol), gave a rotation of -0 017 f 0.003 , [CU]D -515'. antimony tartrate was added to it. The volume was made (B) Most Soluble Fraction, ~-[FeTs~]-d-SbOtart.-A up to 25 cc., the solution was shaken well and filtered di0.005% solution in methanol gave a rotation of +0.022 frectly into the polarimeter tube. The observed rotation (15 minutes after mixing) was -0.285 f 0.003'. The con- 0.004", Or [ a ] D +440'. centration of the soluiion in terms of [ A ~ T S I N O Iwas ( 5 ) M. Calvin and C. H.Barkelew, THISJOURNAL, 68, 2287 (194fi); 0.229%; [a10 -124.4'. R. S. Nyholm, J . Chem. Soc., 851 (1950). The precipitate obtained with silver nitrate from either D( 6 ) F.P. Dwyer and E. C.Gyarfas, THISJ O U R N A L ,74, 4099 (1962).

+ +

+ +

BASUDEB DASSARMA AND

5480

A 0.0033% solution of D - [ F ~ T & ] N Oobtained ~, as before, gave (Y $0.018 & 0.003", [ a ] +545". ~ Further work on this and other polydeiitate ligands is in progress.

[CONTRIBUTIOV FROM

THE

JOHN

C. BAITAR,JR.

VOl. 77

The authors' best thanks are due to the University of Illinois and t o the National Science Foundations for providing funds t o one of them (Das Sarma). URBANA, ILLINOIS

NOYESLABORATORY OF CHEMISTRY, UNIVERSITY OF ILLINOIS]

The Stereochemistry of Complex Inorganic Compounds. XVIII. A New Method for the Preparation of Inorganic Complexes in their Optically Active Forms' BY BASUDEB DASSARMX A N D JOHN C. BAILAR,JR. RECEIVED ~IPRIL 25, 1955

It has been observed that when a compound of the type c ~ s - [ C o " ~ A CC1 l ~(where ] A is equivalent to two bidentate amines or a tetradentate amine), is treated with silver d-antimonyl tartrate, there is an induced asymmetric synthesis of the cobalt complex, t o give ~-[CoA(d-SbOtart)~](d-SbOtart) during the re:hxrnent of chloride by the d-antimonyl tartrate. When the d-antimonyl tartrate is replaced subsequently, therc is obtained an L- [CoTTTAX2] X, which has been isolated where X is OH, 1/&03or C1. The purity of the optical isomer obtaixci depends on the stability of the complex and the mode of operation. L-[COI" enpCl~]CIZ has been obtained as 2 5 3 0 % resolved, whereas the L-[COI" trien Cln]C1 seems t o be optically pure, though there is no way to compare the purity, as [Co trim C111CI could not be resolved by the classical method.

It has recently been shown3 that there is a preferential coordination of the isomer of d-configuration when tartaric, chloropropionic, or lzctic acid is introduced into [Co(l-propylenediamine'i2CO,]C1. The present authors have seen that the method is quite general in the following scheme D-

+

__

or L-[MA,] 21n(dZ-B) D- or ~ - [ h l A - ~ ( or d - I-B),]

+ n2-4 + nz(2- or d-B)

where M = CoIII, Curl, xi11 A = (d or I)-propylenediamine, d-tartaric acid, dgluconic acid, I-glutamic acid B = propylenediamine, alanine, tartaric acid

It has further been observed that when A is not optically active, but the complex [hlAn]is, a similar process occurs which results in active [MA,-,. B n ] and B.4 The presence of an optically active ligand does not completely fix the configuration of the complex of which i t is a part, but it is known that the stabilities of D-[M(~-A).]and L- [M(/-A).] are generally quite different, and a forced asymmetry is induced. The same effect of asymmetric induction is also observed when the environment is asymmetric in nature or there is an exchange between asymmetric groups in the coordination sphere.: I n the present investigation, cis-dichlorotriethylenetetraminecobalt(II1) chloride was treated with silver d-antimonyl tartrate. This replaced the chloride, both inside and outside the coordination sphere, by d-antimonyl tartrate. The resulting (1) Presented a t t h e Cincinnati meeting of t h e American Chemical Society, April 5, 1955. (2) T h e following abbreviations have been used in this article: en = ethylenediamine; pn = propylenediamine; trien = triethylmetetramine; d-SbO t a r t = d-antimonyl tartrate. (3) A. D. G o t t and J. C . Bailar, Jr., T m s JOURNAL, 74, 4820 (1952), J. C . Bailar, Jr., H. B. Jonassen and A. D. G o t t , ibid.. 74, 3131 (lQ32). (4) D-[COI" E D T A ] - 4- 3en + D-[COens]+a by F. P. Dwyer. EDTA]G!dl-pn) I.private communication. L- [Co"' [Co(l-pn)rli* B(d-pn) w-here E T D A = ethylenediamine tetraacetate. s. Kirschner, Thesis, University oi I!linoic, 1954. I,IC0 enzC121 %l-pn -+ L-[COenzd-pnl +a I-pn b y present authors, t u be communicated. ( 5 ) H. B. Jonassen, J. C. Bailar, Jr., and E . H. H u f f m a n , Triis J O U R N A L , 7 0 , 256 (1948); h-.R. Davies and F. 1 '. Dwyer, J . Prtic. R o y . SOC.,A'. S . W a l e s , 86, 64 (1953); B . 1):~s Snrmn and J o l i i i C. Bailar, Jr., Tms J o r J R N a r . , 7 6 , 4&?1 (1954).

+

+

+

-

solution gave no indication of separation of optical isomers on fractional crystallization or precipitation. On the other hand, the specific rotation ( [ a ]$-I10 ~ to 1%') decreased rapidly to a very low value (0 to +so) when the aqueous solution was heated on the steam-bath with accompanying increase in acidity. As the d-antimonyl tartrate is highly dextrorotatory there must be some levorotatory component present to reduce the rotation almost to zero, as the racemization of antimonyl tartrate is highly improbable under these conditions. Treatment with barium hydroxide con\ w t s the product presumably into dihydroxytriethylenetetraminecobalt(II1) hydroxide, which in turn can be converted into carbonato- or dichlorotriethylenetetraminecobalt(II1) ion. These are all levorotatory ( [ a ]=~ -80 to -150"). Thus asymmetry is induced by the presence of the d-antimonyl tartrate which can then be replaced and the active complex containing only optically inactive ligands can be isolated easily. ,1 siniilar approach has been attempted on a series of cobalt(II1) compounds of the type [CoACIX], where A is a tetradentate amine or two moles of a bicientate amine and X is either C1 or "3. The results indicate that there is an induced asymmetric synthesis only if both C1 and X are replaced by dantimonyl tartrate. cis-Dichloro-bis-ethylenediaminecobalt(II1) ion gives the same results as the triethplenetetramine analog. The dichloro-bispropylenediamine complex gives a doubtful result ; in this case the matter is more complicated due to the presence of d- and 1-propylenediamine in the complex. TVith compounds like [Co"I en2NHsCI]CI2 there is no asymmetric synthesis, because there is only one replaceable group. In order to see whether the effect is the same with other active components in the place of dantimonyl tartrate, silver d-bromocamphorsulfonate and silver d-camphorsulfonate were employed. The induced synthesis of the active isomer is very feeble, if it exists a t all. The reason here may be the inability of the camphorsulfonates to replace the chloride ion in the cotjrdination sphere in the :Lqueoiis system.