June 20, I0.X
PREPARATION OF 1 ,'?-PROPY LENEDI,~~IINETETRA;\CETIC ACID
Acknowledgment.-The authors gratefully acknowledge financial aid from the Radiation Project of the University of Notre Dame, supported in part under IJ. S . Atomic Energy Commission
2933
Contract AT( 11-1)-38 and from a research grant H-2218 (C) from the Department of Health, Education and Welfare, Public Health Service. NOTRE DAME, I Y D I A N A
[CONTRIRUTION P R O M 1IIE DEPARTMENT O F IXORCANIC CHEMISTRY, JOHS C U R T I N SCHOOL O F >IEDICAL -4USTRALIAN NATIONAL USIVERSITY]
RESEARCH,
The Preparation of 1,Z-Propylenediaminetetraacetic Acid and its Resolution through the Cobalt (111) Complex BY FRANCIS P. DWYER AND FRANCIS L. GAR VAN^ RECEIVED DECEMDER 17, 1058 Racemic propylenediaininetetraacetic acid and t h e opticnlly active forms have been prepared in good yield by the condensation of the d l and active bases with sodium chloruacetate a t 20". T h e optical isomers gave [ a ] D& 4 i 0 for the monohydrate. T h e pure optical forms were also obtained froin the Co(II1) complex which had been resolved through the active cis-dinitrobis-(etliS-lenedinrnine)-c[)~alt( 111) salt.
Introduction The only references in the literature to 1,2propylenediaminetetraacetic acid, (HhPDTA), are to patents?-9 in which the preparation of the acid and some of the esters and sodium salts are quoted in relation to a number of general preparative reactions for aminopolycarboxylic acids and esters. KO metal complexes have been isolated. The preparation of the racemic acid, and its resolution, have been undertaken in order to carry out ligand exchange studies with metal complexes, in which the ligand is labeled by its optical activity, and to investigate the principle of stereospecific limitation in complexes containing a single optically active organic molecule. This work will be described later. The detailed method of preparation'" of ethylenediaminetetraacetic acid when applied to H4PDTA gave only a viscous sirup which could not be induced to crystallize.11 The difficulty of crystallization, and the low yields of aminopolycarboxylic acids after acidification of the reaction mixture, have been ascribed to enhanced solubility due to dissolved inorganic and organic compounds and reaction by-product^.^ It has been found that dl-propylenediamine condenses readily a t room temperature with sodium chloroacetate in strongly alkaline solution, but low yields result when the reaction mixture is heated." Slow crystallization of the amino acid ensues on acidification, but ultimately a good yield results. Unlike the racemic compound, the optical forms are quite soluble in water and do not crystallize when the reaction mixture is acidified. Great difficulty was found in separating the active acid from the extraneous sodium salts. After addition of hydro(1) Rev. Brother B. L. G a r v a n , i?S C I € , General hlotors-Holden's L t d . post-graduate research fellow. (2) U. S.P a t e n t 2,884,816 (1945). (3) U. S . P a t e n t 2,387,735 (1945). (4) U. S . P a t e n t 2,407,045 (1946). ( 5 ) U. S. P a t e n t 2,428,353 (1947). (0) British P a t e n t 601,816 (1948). (7) British P a t e n t F01,817 (1948i. ( 8 ) U. S. P a t e n t 2,461,519 (1949). (9) British P a t e n t 710,705 (1954). (10) R. S m i t h , J. L. Bullock, F. C. Bersworth a n d A E . Martell, J . Ore C h ~ m . 1, 4 , 3.5.5 (1049). ( 1 1) A 11 Sarge-on, private communication
chloric acid to p H 3.3, the solution was evaporated to one-third volume and sodium chloride and a sodium salt of the acid precipitated by a large volume of ethanol. The latter salt was dissolved in glacial acetic acid, in which sodium chloride is insoluble and precipitated with benzene. Finally it was converted to the optically active acid itself by the use of the hydrogen form of a strong cation exchange resin. Some retention on the resin lowered the yield.12 The exchange column was heated to prevent crystallization. Strong cation exchange resins of the sulfonic acid type adsorb F e f Svery strongly, and as supplied usually contain this ion. In order to avoid contamination of the acid, the iron was removed from the column by prolonged washing with hot hydrochloric acid1s rather than with the more efficient disodium salt of ethylenediaminetetraacetic acid. l 4 After passage through the column the active acids crystalfrom a lized as the monohydrates, ( [ a ] &Aio), ~ small volume of water. The specific rotation decreases with increasing deprotonation. The respective values of [ a ] D in 0.570 solution in the presence of 1 , % , 3and 4 equivalents of potassium hydroxide were *39, *37, *30 and A22O. The rotation rose to *42' in the presence of 15 equivalents of alkali. The infrared spectra of the dl and active acids in Nujol show differences in the 1500 to 1800 cm.-' region, the former having bands at 1700 cm.-' (strong) and 1020 cm.-' (weak). The 1700 cm.-' band is due to carboxyl groups. Evidently hydrogen bonding has decreased the frequency as double bond character is lost for the COOH group. The optically active acids (monohydrates) show bands a t 1727 cm.-' (strong), l(i.70 cm.-' (weak) and 1570 cm.-l(strong). The carboxyl groups show the normal frequency, revealing less effective hydrogen bonding. The band a t 1570 cm.-' can be assigned to OH bending from the water of crystallization. The OH stretching frequency is also shown by a strong 8-1-70 cm.-l band. The (12) R o h m and Haas Co., "Amberlite I R C 50," Technical Bulletin, 1954, p . 8. (13) " P e r m u t i t Ion Exchange Resins in t h e I,aboratory," T h e Perm n t i t C o . . P t g L t d . , London. 1955, p 8. ( 1 4 ) R . A ICrlge, C h ~ i n z c tA v a i y s t , 41, 7 2 ( l l ) S 8 \ .
differences in hydrogen bonding are possibly due to different crystal lattices in the dl and active acids. The Co(II1) complex of the dZ-acid was best prepared by oxidation of the Co(I1) coiiipleues iii alkalitie solution iii the presence of charcoal aiid hydrogen peroxide in a stream of air. Keduced yields due to decompositiou resulted when the method describedi5 for the Co(II1) complex of etliylenediaininetetraacetic acid was used. Resolution was effected as the salt of the active (csdinitrobis-(ethyle~~edia~nine)-cobalt(III) cation. Iiecovery of the optically active acid was carried out by conversion of the [Co PDTAI- anion to the [Co (CN)E],'- anion a i d removal of the latter as the copper(I1) salt. Excess copper ion was bound in the anion [Cu PDTA]?-. Extraneous cations were exchanged for hydrogen ions with a cationexchange resin, and the acid H2[Cu PDTA] isolated by evaporation of the solution. After removal of the copper as the sulfide, the pure acids were crystallized by evaporation and gave [a111 547". The resolution of tmns-l,2-cycloliexancdiaminetetraacetic acid through the Co(111) complex will be described in a later paper.
Experimental All rotations were measured in a 2 din. polarimeter tube at 20' unless otherwise stated. T h e melting points recorded :ire uncorrected. dl-Propylenediaminetetraacetic Acid.-Monochloroaceti~ acid 1280 E., 3 moles) in water (150 ml.) was cooled t o 10 and acold-solutioii of sodium hydroxide (240 g., 6 moles) in water (500 ml.)added slowly from a dropping funnel with good stirring, so t h a t the temperature did not exceed 20'. dI-Propylenediamine (37 g., 0.5 mole) was then added and the mixture allowed t o stand for six days at 20". After filtration, the cooled, well stirred mixture was treated witli concentrated sulfuric acid (65 nil. j added dropwise. Thc sides of the vessel were scratched with a rod and the niixturc allowed t o crystallize for 16 lir. at Y t o 10". The fine crystalline powder, sometimes contaminated with sodium sulfate, was washed by stirring with ccild water (500 ml.j, and after filtration washed with water, ethanol and ether and tlried a t 100'; yield 100 g. The mother liquor, treated with concentrated hydrochloric acid (30 nil.) and allowed to crystallize overnight in the refrigerator gave a further 30 g. The crude material i n water (150 m1.j as treated, with stirring, with solid sodium hydroxide (31.7 g.), aiid the clear solution of the sodium salt acidified with sulfuric acid (9 N , 92 ml.). The acid crystallized slowly on cooliiig, and after washing with water, aqueous ethanol ar!d acetoiie, was tlried a t 100'; i r i . p . 23G0, yield 126 g., 82',o. A n d . Calcd. for CIIHIsN?OS: C, 33.13; IT, 5.92; N, 9.15. Found: 43.16; €1,6.06; N, 9.16. I-Propylenedianiinetetraacetic Acid 1-Hydrate.--The initial part of the procedure was as above. T h e l-propylenediamine, ( [ L Y J D in benzene, -34.1') was prepared as described previously.16 The reaction mixture, after 6 days, was brought t o p H 3.5 by the additioii of concentrated hydrochloric acid (160 nil.) and the solution evaporated on the steam-bath in a current of air t o a volume of 600 nil. After removal of the solid sodium chloride, the filtrate was poured into vigorously stirred ethanol ( 2 1.). The gum which deposited initially, crystallized overnight and was filtered off and washed with ethanol (200 ml.), ether (100 ml.) and air-dried. It was suspended in glacial acetic acid (600 ml.) a t 40' and shaken t o dissolve out the sodium salt of the active acid. T h e undissolved sodium chloride was removed and waslied with a further (30 inl. j portioii of glacial acetic acid. The filtrate and washings added t o benzene (2 1.) deposited a viscous gum which became crystalline b y pour-
c,
iiig off the benzeiie and adding :icetrine (2 I.). The acetow was poured off,the solid ground u p to a file powder and a fresh portitill of acetone (2' 1.) added :ind allrJwed to stand ovrrnight t o tlissolve out trnccs of acetic acid. The solid \vas filtered off, sucked dry, dissolveti iii water (I500 nil.) : ~ t i devapciratc,ti t r i o i i c litcr. 111 tliis W I J . tile Ixst ~ ~ : I C I .ofS acetone aiid iiiucli ( i f the acetic acid were reriiovetl. 'Tlir. palc y c ~ l l i i ~solution v ( P I 1 4.6) was passed tlirciugli the Iiydrogen form of a strong c:itioii exchange resiii of the si~lfoiiic acid type (Zeri Kxrh or .\nibcrlitc I I< 120, 1.(iI., bet1 volunie). The rcsiri ( f(ii-111) \vas firht washtd with IiydnJc-lihiric xicl (3 iV,8 1. 70" i i i tlie crilumii t o remove Fe ' 3 ion, at an rlutioti rate of 1 l./lir., :i11(1the escess acid thcn rrmovcd with water ((j I . ) in 1.5 hr. The S ~ l ~ U t i O of i l tlic. sotliiini snlt o f tile active ~ ~ r ~ ~ ~ ~ y l e i i e t l i a r i i i i i c t eacid tr~~:~c~~tic \vas followed b)- cult1 water a t tlie rate o f 50 nil./inin. TI lirst 2 liters of cfAucnt were pale yell(iw aiid coiitained trac of acetic acid b u t iiiily iirgligilile arnouiits o f optie:illy ncti\ material. This fraetioii w;is rejected. T h e column was then licated to 70" with hcatitig tapes or infrared lamps antl tlie eluting water was adtl(d nt 115' at the rate of 80 nil./iiiin. About 15 liters of water were required t o remove the aciil from the coluniii. This was dcterrniiied by periodic iiie:isiirciiicnt of tlic rot:ition of the effluent. The volume of water was reduced t o 900 n i l . I)y evaporation under rctluccd pressure. The sides of t h e flask were scratched aiid t l i c s ~ i l u t i ~was ~ n cooled ill ice atid allowed to crystallize at 8" ovcriiight. The first fractioii (7'0 g.) was washed with icewater (150 nil.) and then ethanol and acetonr and air-drirtl. T h e aqueous washings and the filtrxte were evaporated t o 120 nil. and yielded a further 1 U g. of material. A third crop (5 9.) was obtained o i i evaporation to 20 nil. (total yield Y5 g., 52','&). S ) m e uf the acid appeared to bc retained b y the resin. All fractions gavc [ O ] D -47" i i i 0.5';; aqueous solution. T h e anliydrous acid, ([CY]D -50" j , was ol~tainrd Iiy heating at 130" for 6 lir. The 1ii.p. (191") w:is raised t o 198' after reeryst~illiz:itioiifroin w:tfer I Jevalmration ~ to :i small volume. A d . Calcd. for C1lIT~oX1.O~: C, 10.71; I T , 6.22; N, 8.64; 1 1 2 0 , 5.5ii. F ~ i i r i d : C, H1.9-k; 11, G.22; N, 8.-10; 1 1 2 0 , 5.63.
d-Propylenediaminetetraace was prepared in the s:tme ~ ~ y l e i i c t l i n n ~ iused i i e ~ ~liad [ C Y ] D ;iriiuuiit of iiiactive material, i'iiial solution W:LS evaporated t o 900 inl. and cooled t o riit)iii t e t n p e r ~ t u r e ,was rernovetl before cooling in ice overnight. The nionohydrate gave [ U ~ D+4io for a 0.576aqueous solut i t ~ i i ,and the aiiliydrous :wid liad [ a ]+50°. ~ ,:inai. I;oullti: 11, 6.11; K, ~ 3 8 I; r a , 5.5;. dl-Potassium (Propy1enediaminetetraacetato)-cobaltate(111) I-Hydrate.-dl-Pr~ipylenetliamiiietetraaeetic acid (31 g.) was dissolved b y shakillg in water (600 nil.) coritainiiig potassium acetate (50 g.) and cobalt (11) acetate 4hytlratc (25 g.) in water (200 nil.) added. Xfter addition of activated charcoal (10 g.) a n d hydrogen peroxide (:
