H. H . Elrzss, 'ho!-.
I iSCi
DE ~ R I E SAND
H. H.
JAFFE
VOl. 65
Discussion
it is interesting t o note that in the compounds IC4Si(CN)4 and K&l(CN)4 the central atom of 1 lit expermienti &scribed hexe show that the complex ion exhibits an eflectizle atomic number liquid ammonia solutions of potassium and potas( E. A. N . j equal to that of the rare gas a t the end wxn tetracyanopalladate( 11) react in the ratio of 2 1 t o form moderately soluble potassidm tetra- o i the period in which the element is placed in the c*yanopalladateiOt which 1s more soluble and less periodic classification: i. e., E. A. K . of Xi in = 36 ( K r ) ; E. A. S.of Pd in Pd$table than the corresponding compound of nickel. CSj,==== 54 (Xe). Hence, these complex ions are 130th cyanopalladatei0) and the cyanonickelatc(0) reduce azobenzenc to hydrazobenzerie and to he compared to the known carbonyl of nickel, d \ e r and mercuric tons to the metal.; -In- ;Si(CC))4,and the expected carbonyl of palladium, Inonium cyanopalladatc ( 0 ) exists in aininonia Pd(C!))+ The difference lies in the fact that the iolution but has not i m n isolated In contrast to CN - ion, although isoelectronic with the CC] iiicke1, palldium appeat 5 to form tin compound of molecule, carries with it a negative charge which the type of K2€'d(CA)d At least the composition must be balanced by a corresponding charge on a of the precipitate is the same throughout the pre- positive ion. Hence, the same pattern of comcipitation. It sfems likely that lfanchot and hination leads t ) a neutral molecule in the one SitCO)d, and to a complex ion in the other, Schmid actually prepared a iolution of K41Y(CS)1 case, y iI , C y ,) =_ (,or Pd(CS)4===) .I3 . m c l not K:Pd(C'U)3 as they supposed. Summary Experience in this Laboratory has no1 confirmed the experimce of Eastes and Burgess in their t n 1 . Potassium tetracyanopalladate(0) , K4Pdnbility to prepare K&II C h ) + by the actioii of CN)J, has been prepared by the reaction between potassiurri on solution5 ( ~ t &Si(CN)d Chly potassium and potassium cyanopalladate(l1) in occasionally dit1 the reductio:i stop at the K2Si- liquid ammonia solution. tCS), stage. A possible cause U i these apparent 2 . Same of the properties of KJ?d(CiV)4 have mconsistencirs ma> lie Y 11 the phvsical character been detennined and compared with those of of the initial precipitate. Sirice the reaction KINi (CK ).$, KJJIICS)~ 4 E; KCX S4hi(C-U),requires ( 1 3 1 For a discussion of t h e relation between E. A. N. and stability complex ions see A. A. Blanchard, Chrm. Reo., 26, 409-422 (1940) the reaction oi two wltible materials with a pre- of . q r i m c e . 94. 311-317 (1941) cipitate, a firm coating of more insoluble material cpries were by M. E. Bailey and H. B. Hass, THIS J O U R N A L , 63,1989 1\14 I j , and by E. Shapiro and R. F. N e w t o n . rbid , 66, 777 !I%?,) ,L'; .\lcC'oy Research Fellow, 1941-1842. :i: A . Fi-chgold a n d R. Ammon, Biocizenz. Z . , 234, 39 (1951 I . & , :A - > i r n l l mil /i Hofman. j! p h
PURDUE 1,'NIVERSITYI
JAFF~?
isolysergic acid by chromatographic adsorption of the d-isopropanolamide on columns of alumina. Jamison a i d Turner5 have recently reported a sniall separation oi the 1-menthyl dl-mandelates by selective adsorption on aluinina. The methods of chromatographic adsorption have also been used to resolve stereoisomers. The first separation of that kind was claimed by Winterstein and Stein.6 Cook and others7 were able to separate cis- and trans-azobenzene oii !5: & M I.. Jamison and E . E. T u r n e r , J . C h e m SOL.,811 (1942). (ti) -1.Winterstein and G Stein, Z . p h y s i o l Chem., 230, 2 4 i
19XB'I. ( 7 ) (a) A . €1. Cook, J . C h e w Soc.. 876 (1938); (b) L.Zechmeister, (3. Frehden a n d P Fischer-Jorgensen, ~Vat~~rmisreizscknf/ri,. 26, 49ii t
>!>:jxj
Aug., 1943
CHROMATOGRAPHIC ADSORPTION OF ENANTIOMORPHS
1487
TABLE I SUMMARY OF RESOLUTION EXPERIMENTS Adsorbent
Aluminum oxide Calcium sulfate Carbon Carbon Carbon Carbon Carbon Fuller’s earth Magnesium oxide Magnesium oxide
Solvent
Size of column, cm.
Amount used, g. and ml.
A. dZ-a-Phenylethylammonium bi-d-tartrate 60yG ethanol 1 X 13 1.7/50 60% ethanol 1 X 13 1.7/50 Water 1 X 13 20/lOO €107~ acetone 1 X 13 20/100 607, ethanol 1 X 13 1.7/50 60% ethanol 2.2 X 50 20/100 60% ethanol 50/250 5.2 X 120 60% ethanol 1 X 13 1.7/50 1 x 10 20/100 Water 5.6/100 ethanol 1 X 13
Typical resultsa RD
-0.03
+ .02 + .04 + .16 + .29 + .07
+ -
-
.05 .02 .05 .02
B. Brucine &mandelate Benzene 2.8 X 55 Calcium sulfate 35/350 f0.05 35/200 2.2 X 50 Benzene .16 Carbon 2.2 X 50 Benzene - .02 40/350 Cotton 105/1000 5.2 X 120 Benzene .18 Dextrose 39/200 Chloroform - .05 5.2 X 120 Dextrose Carbon tetrachloride 2.2 X 50 35/350 .02 Dextrose Dextrose %butanol 2.2 x 50 35/300 - .23 2.2 X 50 35/350 - .03 Fuller’s earth Benzene Magnesium oxide Benzene 2.2 X 50 35/300 .o a Maximum observed rotation of a-phenylethylamine in ether and mandelic acid in water recovered from the four to nine liquid chromatograms. See Table I1 for further details. See Table 111.
+
+ -
*
[ . I D
-0.3 +0.2 +0.3 +1.4 f1.7
$1.2 +7.2 -0.4’ -0.4’ -0.2 +4.4 +2.7 0.0
$4. le -0.3 0.0 -0.7 -1.7 0.0 best of
(California Chemical Co.) was mixed with two parts of hyflo-supercel (Johns-Manville Co.) to increase the rate of filtration. Similarly, fuller’s earth (Lloyd’s Reagent, Eli Lilly and Co.) was used, mixed with an equal weight of hyflo-supercel. Carbon (Norit A) when used was first extracted with hydrochloric acid and washed with water. The addition of hyflo-supercel did not appreciably increase an impractically slow filtration rate. Hydrated calcium sulfate and anhydrous akminum oxide (Brockmann) were -dZ-a-Phenylethylamine was prepared according to the used for several experiments. Finally, cotton and dexmethod of Ingersollg from ammonium formate and aceto- trose were also used. With most of the columns it was phenone. The final product was analyzed and found to be necessary to apply nitrogen pressures up to three atmossufficiently pure for this investigation. The amine was pheres to obtain a practical rate of flow. added to an equivalent amount of nearly saturated dAnalysis.-To determine the efficiency of resolution, the tartaric acid solution. The mixture solidified nearly desired compound was recovered from the solution and its completely within a few hours. The salt was collected on rotation measured for the sodium D line in 1-dm. tubes a filter and dried a t room temperature; m. p. 159-162”. with a Schmidt and Haensch, model 52B, polarimeter, Anal. Calcd. for ClzHlrOsN: C, 53.13; H, 6.32. with monochromator attached. The a-phenylethylamine Found: C, 52.3;H , 6.3. was extracted with ether for the bi-tartrate solution after Brucine dl-Mandelate.-One-twentieth mole of brucine it had been concentrated and made alkaline with sodium was dissolved in 300 ml. of hot benzene and an equivalent hydroxide solution. The rotation of the ether solution amount of dl-mandelic acid (U. S. P. grade, Merck & Co.) was measured to *O.0lo. The concentration of the amine was added. Solutions of brucine dl-mandelate in 2- was determined by adding a standard hydrochloric acid butanol and in carbon tetrachloride were prepared in the and back-titrating with sodium hydroxide, with methyl same manner. Chloroform solutions were prepared by red as indicator. Mandelic acid was recovered by extracting the benzene, dissolving the constituents in cold solvent. Solutions with hexane could not be prepared. Attempts to prepare chloroform or carbon tetrachloride solutions of its brucine 1-nicotine dl-mandelate and I-nicotine dl-a-bromopropio- salt with one normal sulfuric acid (stronger acid decomposed the brucine). The aqueous layer would then connate were unsuccessful. Only pasty tars resulted. Adsorben&.-A variety of adsorbents were used. For tain brucine sulfate and mandelic acid. Three successive example, one part of “Micron magnesium oxide, 2641” extractions with 50-ml. portions of ether removed the mandelic acid and possibly traces of brucine. The latter (8) G . H. Coleman, A. F. Farnham and A. Miller, THISJOURNAL, was removed by extraction with 10 ml. of water. The 64, 1501 (1942). mandelic acid in the 150 ml. of ether was transferred to 10 (9) A. W. Ingersoll, “Organic Syntheses,” I T , 76 (1937).
alumina. Coleman and his co-workersa separated sugar azoates using “magnesol” and silicic acid as adsorbents. I n this investigation a study was made of the separation of diastereoisomers by the well-known methods of chromatographic adsorption. Materials. dZ-~-Phenylethylammonium Bi-d-tartrate.
H. 13. HASS,THOS.DE VRIESAND H. H. JAFPE
1-48s
till, of water by evaporation of the ether layer over a layer of water. The rotation of the aqueous solution was measured. The concentration of the mandelic acid was determined by titration with s i andard sodium hydroxide Tvith phenolphthalein as indicator. The absence of brucine was checked with colicelit rated nitric acid, which gives a dark red color with even traces of brucine. Procedure.-Columns of four difiererit sizes were used. The smallest was 1 cm. iii diameter and 15 cm. loiig and the largest was 5 by 120 c i u T*arious solveiits were u s r t l iiich as benzene, 2-hutano1, carbori tetrachloride, chloroform, water and aqueous solutions of acetotie or c.than01 (40',; water). It. was 11 ssary t o us12rnotlrrately concentrated solutions of the d tereoisomers in order t o obtain, even at best, some iignificant poiarinieter readings for the iromatogranis. The coiiceii trations arid the volume from 50 to 1000 rnl., depending on the size of thc culunin. The chroniatog~aliis were developed with a sufficient volume of thf, mlvetir u n ti1 the desired results vitrc obtained. Results.-In Table I are summarized the experimental details and results, The following are t w o typical experiments described in some detail dl-cu-Phenylethylamine.----~~~enty grams of dl-a-phenylethylammonium bi-d-tartrate dissolved in 100 ml. of water was filtered through a culuriin (10 K 100 mm.) packet1 with magnesium oxide mixed with two parts of hyflosupercel. 'The chroiiiatograin was developed with 800 Inl. of water. Four successive liquid chromatograms werc' collected (their volumes are listed in the second column of Table 11). The amine was recovered in ether solutions aiid the rotations measured to i.O.01' in a 1-dm. tube. The [(.ID of pure s-phenylethylamine is i-39.9, hence the results show that practically no resolut.ion was obtained. Similarly there was little resolution when other solvetiti arid adsorbents were used. The results of 21 experiment. have been summarized in 'Table I. 'The Id salt was ad -orbed less than the dd salt on all adsorbents with thc tlx crptioti of carbon 'rABLE 11 RISJLLTIOX O F d l - o ( - P H E N ~ Z E T N I L A ~ i I N I :
Filtrate
I 2 3 3.
Mi.
40 100 200 200
Concn. of amine in ether, g./ml. (1
118 11x
.(iOi8 000;
nu
-- 0.03 ,113
-+
.o , 0:i
LalD
-0 4 --
4
.0 140)
Vol. 65
dl-Mandelic Acid.-One hundred five grams of brucinedl-mandelate in 1000 ml. of benzene was filtered through a column ( 5 . 2 X 120 cm.j packed with dextrose. The chromatogram wa5 developed with beiizene. The mandelic acid was recovered from five liquid chroiiiatogratns arid ita rotation was measured in aqueous solution. The / a ] uof pure d-mandelic acid in aqueous solution is 156.5. lleiice the results tabulated in Table I11 show that the rnaxiniuiii rcsoluLiou was about 3 ( 2 . Some resolution was
TABLI-: JII
also obtaiiied on fuller's earth, carbon and calcium sulfate. The d salt was adsorbed less than the I salt. No reproducible results could be obtained on magnesium oxide or on cotton with benzene as solvent, or on dextrose with chloroform or carbon tetrachloride as solvents. With 2butanol as solvent and dextrose as adsorbent, the [.ID of mandelic acid recovered from four 300-ml. filtrates were -0.71, 0.0, +7.1 and +lF", respectively. The results of I7 experiments are summarized in Table I.
The authors are grateful to Dr. H. N. McCoy who generously furnished the funds for this inves1igation through the Purdue Research Foundation, and to X r . M. H. Danzig for the organic analyses.
Summary The techniques of chromatographic adsorption were applied to the resolution of racemic compounds after converting them to soluble diastereoisomers. d2-a-Phenylethylammonium bi-d-tartrate aiid brucine dl-mandelate were prepared and only partial resolutions were obtained on magnesium oxide, fuller's earth, carbon, calcium sulfate, cotton and dextrose. XVFS'I r,;;ZFAYLTTE, IXDIAV.1
RECEIVED APRIL1, 194'3
