I. The Synthesis of Potential Antimaterials. Some 2-Substituted 8-(3-Diethylaminopropylaminol)-Quinolines. II. Isomorphism in Relation to Serological Specificity. III. A Study of the Hammick Reaction Citation Mislow, Kurt Martin (1947) I. The Synthesis of Potential Antimaterials. Some 2-Substituted 8-(3-Diethylaminopropylaminol)-Quinolines. II. Isomorphism in Relation to Serological Specificity. III. A Study of the Hammick Reaction. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/60T0-BG46. https://resolver.caltech.edu/CaltechETD:etd-06142004-140844 Abstract No Abstract Submitted. Item Type: Thesis (Dissertation (Ph.D.)) Subject Keywords: Chemistry and Physics Degree Grantor: California Institute of Technology Division: Chemistry and Chemical Engineering Major Option: Chemistry Minor Option: Physics Thesis Availability: Public (worldwide access) Research Advisor(s): Pauling, Linus Thesis Committee: Unknown, Unknown Defense Date: 1 January 1947 Record Number: CaltechETD:etd-06142004-140844 Persistent URL: https://resolver.caltech.edu/CaltechETD:etd-06142004-140844 DOI: 10.7907/60T0-BG46 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 2579 Collection: CaltechTHESIS Deposited By: Imported from ETD-db Deposited On: 15 Jun 2004 Last Modified: 16 Mar 2023 20:44 Thesis Files Preview PDF (Mislow_km_1947.pdf) - Final Version See Usage Policy. 4MB Repository Staff Only: item control page

Part I The Synthesis of Potential Antimalarials, Some 2-Substituted 8-(3-Diethylaminopro- pylamino)-quinolines. Part II Isomorphism in Relation to Serological Specificity. Part III A Study on the Hammick Reaction. Thesis by hurt M. Mislow In Partial Fulfillment of the hequirements for the Degree of Doctor of Fhilosophy. California Institute of Technology Pasadena, California, 1947 Gontents Acknowledgement Part I the Synthesis of Potential Anti- malarials. Some 2-Substituted 8-(3-Diethylaminopropylamino)- quinolines. introduction I. 2-Nethoxy-8- (5-diethylamino- propylamino)-quinoline II, 2,6-Dimethoxy-8- ($-diethyl- aminopropylamino)-quinoline and 8-(3-Diethylaminop ropyl- amino)-2,6-quinolinediol. III. 6-Methoxy-8-(3-diethylami- nopropylamino)-carbostyril. Sumiary References Part II Isomorphism in Relation to Sero- logical Specificity Introauction I. Techniques and Apparatus for the determination of Phase Dia- grams. A. Apparatus B. Preparation of the mixtures (a) Preparation of the pure components (bo) Preparation of samples of the mixtures Cc, Observational Procedure (a) Beginning of melting (thawing point) (ob) End of melting (mel- ting point (c) Decomposition page 56 47 48 50 54 55 oS page D. Reliability of nesults 55 i. Example of a ‘Typical vetermination D7 II, Phase Relationships of Pairs of bifferent Carboxylic Acids. 60 A. Data 61 B. Interpretation of Data 86 Discussion of hesults 30 - Summary oO 96 neferences 97 Part III A Study of the Hammick Reaction 102 Propositions 123 ACKNOWLEDGEMENT I wish to express my sincere appreciation to the mem- bers of the staff who assisted me with the work described in This Thesis. I am especially grateful to Professor Linus Pauling and Dr. Joseph B. hoepfli for their inspi- ration, encouragement and interest in the work. I also wish to thank Dr. Chester lcCloskey, with whom it has been my privilege to cooperate. parr it The Synthesis of Potential Antimalarials. Some 2-Substituted $-(3-Diethylaminooropylamino)-quinolines. (This Kesearch has been published in the J. Am. Chem. Soc. 68, 1555(1946)) 1 This work was done under contract, recomended by the Committee on Medical Research, between the Office of Scientific Research and Development and the California institute of Technology. INPHOLUCLTION From the time of the elucidation of the structure of the cinchona alkaloids (by Habe in 1908) until the First World War, the major effort of chemists in their search for improved quinoline-type antimalarials was directed towards modifying some features of the quinine molecule while keeping the general structure intact. Ihe inability of Germany during 1914-1918 to import the naturally-occur- ring alkaloids instigated intensive investigations with the object of finding a synthetic substitute. The cul- mination of these researches was the synthesis, in 1926, of pamaquine (or plasmochin) by chemists of the I.G. Far- benindustrie. During the Second World War, this country found itself in a position similar to that of Germany in the First World War. The supply of naturally occurring cinchona alkaloids had been cut off, and the toxicity of pamaquine prevented its large-scale distribution as a drug. For that reason a program was begun under the Committee for Medical Research in the hope of finding a drug with the desirable antima- larial properties of pamaquine but without its high toxi- city. In e recent investigation (1), an in vitro degradation peoduct of quinine (..DP) was isolated which was less toxic, altnough also less active, than quinine itself. Later evidence (2) indicated that QDP was 2'-hydroxy-~-6'-methoxy- S-vinylruban-9-ol, a carbostyril derivative. It seemed of interest, therefore, to prepare some pamaquine analogues having the carbostyril or 2-alkoxyquinoline nucleus, and in this way obtain less toxic pamaquine-type antimalarials. The preparation of certain 2-hydroxy-4-methyl-8- (dial- kylaminoalkylamino)-quinolines as potential antimalarials had previously been: reported (3). Since these compounds had a lepidine nucleus they differed radically from the pamaquine-type of structure. The compounds prepared in This Laboratory had, except in the case of 2-methoxy-8- (diethylLaminopropylamino)-quinoline, a hydroxyl or meth- oxyl group in the 6-position, in view oi the fact that quinine, as well as pamaquine, owes. its - activity in part to the presence of the methoxyl group in the 6-position. It had originally been intended in this work to attach to the é6-aminoquinoline nucleus the side-chain of pamaquine, but recent evidence (4) indicates that the use of this side-~ chain may give rise to highly toxic isomers. Hence the 5- diethylaminovropyl side-chain was chosen for this purpose, since no isomerization is known to occur during its prepar- ation. The compounds which are described here are therefore of the plasmocide, rather than the pamaquine, type. Al- though this meant a departure from the original purpose of synthesising pamaquine derivatives, it permitted the comparison of the toxicity of various plasmocide deriva- tives without the danger of complications from an isomer- ized noval side-chain. The compounds which were prepared in this series are the following. NHCHCHoCHN (CoHs ) 5 I. 2-Methoxy-8- (3-diethylaminopropylamino )-guinoline Ry - OCHg Ro - H II. 2,6-Dimethoxy-8- (3-diethylaminopropylamino)-quinoline III. 8-(3-Diethylaminopropylamino)-2, 6-quinolinediol Ry ~ OH Rp - OH IV. 6-Methoxy-8- (3-diethylaminopropylamino)-carbostyril R, - OH Ro ~ OCHs Their preparation is indicated in the accompanying general scheme, where R represents -H or ~OCH.« c No be 2) a) c + been + tad | ; \_. NOs Chg Oy FS Waar VY wp WL ao db at ry 7 Wit (Gi. wav LE . 2S 2°62 ( tal g* ors! 2 i. 2-Methoxy-8- (S-diethylaminopropylamino)-quinoline 8-Nitroquinoline was prepared in 62 per cent yield from o-nitroaniline by the Knuppel modification of the Skraup synthesis (6). Arsenic oxide was used instead of arsenic acia, but that procedure did not improve the yield. The N-methyl iodide derivative (7), prepared in 84 per cent yield, was then oxidized in 68 per cent yield to l-methyl-8-nitro-2-quinoline (8). This oxidation proceeded more smoothly wnen 30% hydrogen peroxide, instead of potas- siun ferricyanide, was used. This method has been used by Ing(9) in the preparation of a different 2-quinolone. l-wethyl-S-nitro-2-quinolone was converted into é- chloro-8-nitroquinoline in 85 per cent yield by heating the quinolone with phosphorus pentacnioride (3) according to the method of Fischer (10). Subsequent hydrolysis in 20% hydrochloric acid gave 8-nitrocarbostyril (8, 11) in 88 per cent yield. Miethylation of the carbostyril with methyl sulfate (7, 12) gave a 79 per cent yield of e- methoxy-8-nitroquinoline. This compound was first prepared by Decker (12), who believed it to be l-methyl-8-nitro-2- quinolone, although in a later publication (7) he recognized its true nature. Since no details are mentioned on its preparation, it seems advisable to give them in this work. An attempt to prepare it by reacting 2-chloro-~8-nitroquin- oline with sodium methoxide was unsuccessful. Reduction of 2-methoxy-8-nitroquinoline with hydrogen in the presence of platinum oxide gave a 90 per cent yield of 2-methoxy-8-aminoquinoline, and by condensing this com- pound with 1-diethylamino-3-chloropropane hydrochloride, 2-methoxy-8- (3-diethylaminopropylamino)-quinoline was obtained in 78 per cent yield. The base is very unstable to air and was submitted for vesting as the @ihydariodide, which, unlike the dihydrochloride, is not hygroscopic. Experimental Part 8-Nitroquinoline (6), 880 g. of glycerol, 400 g. of o- nitroaniline (E.k. practical), 254 g. of arsenic oxide and 800 g. of sulfuric acid were refluxed for 3 hours and the desired product was isolated from the reaction mixture following the procedure described by Knuppel (6). The yield was 316 g. (62%). It was found necessary to purify the compound by extracting the crude product with boiling benzene (13). 8-Nitroaquinoline Methiodide (7). The proceaure used by Decker et al. (7) was found to give excellent yields. A mixture of 109 g. of 8-nitroquinoline and 119 ml. of methyl sulfate was heated on the steam bath for 3s hours. The solution was diluted with 157 ml. of water and chilled to i All melting points are corrected; microanalyses by Dr. G. Oppenheimer and G. A. Swinehart. 15°. 210 g. of potassium iodide was then added while the solution was stirred vigorously, and the orange-red methi- odide precipitated. The yield was 166 g. (84%). 1-NMethyl-8-nitro-2-quinolene (8). 195 g. of 8-nitroquino- line methiodide was suspended in 1 lt. of 96% ethanol and 450 ml. of 20 wt potassium hydroxide was added. When all of the base had been added, the methiodide dissolved to give a deep red solution which, on standing for a few minutes, set to a crystalline paste, probably due to pre- cipitation of the N-methyl hydroxide. While the paste was stirred vigorously and kept below 45° in an ice bath, 740 ml. of 30% hydrogen peroxide (Dupont "Albone") was added in 50 ml. portions. The reaction was not very vig- orous. When all of the peroxide had been added, the sus- pension was allowed to stand overnight, filtered and the residue washed with 2 1t. of water. 85 g. (68%) of lemon yellow crystals were obtained, which melted at 132-133° (reported (8) 133-1349). 2-Chloro-S8-nitroquinoline (8). This compound was prepared by heating amixture of 72 g. of 1l-~methy1-8-nitro-£-quino- lone and 72 g. of phosphorus pentachloride for 15 min. at 15045° until effervescence had ceased and refluxing the resulting liquid for 1 hour. From the reaction mixture, 63 g. (85%) of light”>tan product was isolated, which melted at 151-152° (reported variously at 152° (8) and 149° (11)). LO 8-Nitrocarbostyril (8, 11). 63 g. of 2-chloro-8-nitro- quinoline was refluxed for 2 hours in 500 ml. of 20% hydrochloric acid. The solution was neutralized with 40 wt.% sodium hydroxide, made just acid with 10% hydro- chloric acid and cooled. The resulting suspension was filtered and the residue dried to give 50.5 g. (88%) of light yellow crystals of 8-nitrocarbostyril. 2-Methoxy-8-nitroquinoline (7, 12). 40 g. of 8-nitro- carbostyril was dissolved in 900 ml. of hot 10% sodium hydroxide. While the solution was kept hot on the steam bath, 610 ml. of methyl sulfate was added in 100 ml. portions. Between each addition the suspension was stir- red vigorously and made alkaline with solid sodium hydrax- ide. Finally the alkaline solution was cooled, filtered and the residue washed with water until the washings were colorless. The product was recrystallized from ethanol to give 55.8 g. (79%) of colorless crystals, mp. 122-123° (reported (7, 12) 124-125°). | 2-Methoxy-S-aminogquinoline. 24.9 g. of 2-methoxy-8-nitro- quinoline was suspended in 500 ml. of absolute ethanol and shaken with hydrogen at 1 atm. and room temperature in the presence of 0.25 g. of platinum oxide. After 4 hour, 9.2 1t. of hydrogen had been taken up, and the rate of hydrogen uptake had decreased to 1 ml./min. Theory ams required 9.2 1t. The solution was filtered from the cat- alyst, the solvent removed and the residual oil distilled in vacuo. The amine came over at 125~-129°/o.2 mm. and crystallized readily to give 19.0 g. (90%) of colorless needles, which were recrystallized from ethanol and mel- ted at 75-76°, Anal. Calcd. for Cy oly (ON, (174.2): C, 68.95; H, 5.79; . N, 16.08. Found: C, 68.89; H, 5.793; N, 1606. i-Diethylamino-3-chloropropane hydrochloride. The bulk of this material was supplied by Columbia University. | Fart of it was also prepared by reacting 271 g. (1 mole) of C.P. phosphorus tribromide with 189 g. (2 moles) of trimethylene chlorohydrin to give 178 g. (56%) of tri- methylene chlorobromide (14). 100 g. (0.63 moles) of this compound was reacted with 129.6 ml. (1.77 moles) of di- ethylamine to give 34.5 g. (57%) of 1-diethylamino-3- chloropropane (15). The hydrochloride was obtained by passing dry hydrogen chloride through an ethereal solu- tion of the base and removing the solvent (16). The hydro- chloride is extremely hygroscopic and has to be kept over sulfuric acid. 2-Methoxy-8- (3-diethylaminopropylamino)-quinoline. The modified (17) procedure of Magidson et al. (18, 19) was followed in this condensation. A solution of 7.0 g. (0.04 mole) of 2-metnoxy-S-aminoquinoline, 16.4 g. (0.2 mole) of sodium acetate and 7.4 ge (0.04 mole) of 1-diethylamino-3-chloropropane hydrochloride in 150 ml. of 66% ethanol was refluxed on the steambath for 5 days, with an additional 7.4 g. of 1-diethylamino-3-chloro- propane hydrochloride added each day. At the end of that time the light brown solution was diluted to 300 ml. with water, made alxaline with 150 ml. of 25% sodium hydroxide, saturated with potassium carbonate and extracted three times with 50 ml. portions of ether. The ether extracts were dried over potassium carbonate, the solvent removed and the residue distilled in vacuo under nitrogen. After the unreacted l-diethylamino-3-chloropropane had distil- led at 60-80°/0.5-1.5 mm., the desired base came over as a viscous, yellow oil, b.ep. 190-192°/0.15 mm. The yield was 9.0 g. (78%) based on 2-methoxy-S-aminoquinoline. The base, which is unstable in air, was dissolved in 56 ml. of ethanol and a solution of 8.5 ml. of hydriodic acid (sp. g. 1.7) in 38 ml. of ethanol was added. The solution was diluted with ether until slightly turbid and allowed to stand at 5° overnight. The colorless, erystalline dihydriodide which had come out of solution was filtered, recrystallized from ethanol and dried. It melted at 140-142° (dec.). The yield was 9.6 g. (56.5%) 13 from the base or an overall yield of 4445 based on 2- methoxy-S-aminoquinoline. Anal. Calcd. for Cy Ho PN 1, (543.2): C, 37.58; H, 5.01; N, 7.74. Found: C, 357.84; H, 5.15; N, 7.57. 14 Il. 2,6-Dimethoxy-8- (3-diethylaminopropylamino )-quinoline and 8-(3-Diethylaminopropylamino) -2,6-quinolinediol The preparation of 2,6-dimethoxy-~8- (3-diethylamino- propylamino)-quinoline was analogous to that of 2-methoxy- 8-(3-diethylaminoprooylamino)-quinoline. The overall yield of the 2,6-dimethoxy base was 17 per cent as compared to 25.6 per cent of the 2-methoxy analogue, based on 6- methoxy-8-nitroquinoline and 8-nitroquinoline respectively. As a whole, the 2-position of this series of compounds was found to be considerably stabilized by the substitution of methoxyl in the 6-vosition, and more strenuous conditions were required for the preparation of the intermediates. 2, 6-Dimethoxy-8- (3-diethylaminopropylamino)-quinoline is even more unstable in air than the 2-methoxy analogue, and it was submitted for testing as the dihydriodide. This derivative was chosen because, unlike the dihydro- chloride, it is a non-hygroscopic salt. 8-(S-Diethylaminopropylamino)-2,6-quinolinediol was obtained in 26 per cent yield by hydrolysis of the 2,6- dimethoxy analogue in 20% hydrochloric acid. The base was a high melting solid and was stable in air, in contrast to the 2-methoxy and 2,6-dimethoxy analogues. Experimental Parte 6-Methoxy-G-nitrogquinoline Methiodide. A mixture of 408 g. (2 mole) of 6-methoxy-8-nitroquinoline (Winthrop Tech.) and 380 ml. (4 mole) of methyl sulfate was heated on the steam bath for 4 hours. The hot oil was diluted with 500 ml. of water and the solution cooled and ex- tracted twice with 500 ml. portions of ether. the aqueous layer was heated on the steam bath to expel most of the dissolved ether and cooled to room temperature. 600 g. of sodium iodide was added to the solution while stirring vigorously and cooling in an ice bath, and after the addition was completed, the suspension was stirred for some time, filtered and the residue dried to give 662 g. (96%) of red needles. It was found unnecessary to recrystallize the bulk of the material for the next step. A sample was recrystallized from butanol, and melted at 149° (dec.). Anal. Caled. for 04H, ,0gNoI (546.1): C, 58.17; H, 5.20; N, 8.10. Found: 0, 58.26; H, 5.06; N, 8.51. 6-Methoxy-l-methyl-8-nitro-2-guinolone. 242 g. of crude 6-methoxy-8-nitroquinoline methiodide was suspended in 1150 ml. of 96% ethanol in a 4 1t. beaker and 460 ml. of tall melting points are corrected; microanalyses by Dr. G. Oppenheimer and G. A. Swinehart. 16 20 wt.% potassium hydroxide was added. When all of the base had been added, the methiodle dissolved momentarily to form a deep red ‘solution, which almost immediately set to an orange, crystalline slurry. While the mixture was stirred vigorously and kept at 30-50° in an ice bath, 920 ml. of 30% hydrogen peroxide (Dupont "Albone") was added in 50 ml. portions. The reaction was violent and very exothermic, and it was necessary to wait about 5 minutes between each addition. When all of the peroxide had been added, the suspension was allowed to stand over- night, filtered and the residue washed with 14 1t. of water. The residue was recrystallized from methanol to give 95 g. (58%) of long, Lemon-yellow needles, m.p. 186-187°. Anal. Caled. for C,H, ,0,N, (254.2): , 56.40; H, 4.303 N, 11.96. Found: C, 56.76; H, 4.21; N, 12.21. 2-Cnloro-6-methoxy-8-nitroguinoline. A mixture of 169 g. of 6-methoxy-l-methyl-8-nitro-2-quinolone, 169 g. of phos- phorus pentachloride and 3 ml. of phosvhorus oxychloride was heated under reflux on an oil bath at 180+5° for 54 hours. A calcium chloride seal was found to be sufficient protection from moisture. During the heating process, the reaction mixture remained solid. The resulting cake was hydrolyzed by the addition of 1 1t. of water, the suspension 17 made alkaline with ammonium hydroxiae and filtered. The residue was dried, pulverized (10 mesh) and suspended in lL it. of celliosolve. ‘The suspension was boiled for 1 min- ute, chilled to 20° and filtered. The residue was washed with ethanol and dried to yield 133 g. (77%) of colorless crystals, m.p. 225-2269, recrystallized from methyl cel- losolve. | Anal. Caled. for C4 Ha O2NoCl (238.6): C, 50.32; H, 2.96; . N, 11.74. Found: G, 50.30; H, 3.00; N, 11.74. 6-Methoxy-S-nitrocarbostyril. A suspension of 125 g. of é-chloro-6-methoxy-G-nitroquinoline in 1200 ml. of 2070 hydrocnloric acia was refluxed for 24 hours. Since con- siderable foaming took place, ml. of capryl aliconol was used as a depressant. The suspension was neutralized with 40 wt.% sodium hydroxide and made just acid with 10% hydrochloric acid. It was cooled, filtered and the resi-~ due suspended in 1 1t. of boiling water. 40 wt.% sodium hydroxide was added until the solid had dissolved. The hot solution was filtered through Celite, acidified with 10% hydrochloric acid and chilled. The yellow precipitate was filtered and recrystallized from methyl cellosolve to give 96 g. (83%) of golden platelets, m.p. 210~211°. Anal. Caled. for Cy oHg4No (220.2): C, 54.54; H, 5.66; N, 12.735. Found: C, 54.33; H, 3.50; N, 12.87. 18 2, 6-Dimethoxy-S-nitroguinoline. This compound was prepared by the same method as 2-methoxy-8-nitroquinoline. 95.4 g. of 6-methoxy-G-nitrocarbostymil in &£ 1t. of hot 10 wt.% sodium hydroxide and 800 ml. of methyl sulfate gave 76.8 g. (76%) of light yellow erystals. The product was recrystal- lized from ethanol and melted at 149-150°. Anal. Caled. for 6,,H,,0 (234.2): C, 56.403 H, 4.30; aNe N, 11.96. Found: C, 56.29; H, 4.30; N, 12.03. 2,6-Dimethoxy-t-aminoquinoline. A suspension of 70 g. of 2,6-dimethoxy-S-nitroquinoline and 0.4 g. of platinum oxide in 1 1t. of absolute ethanol was shaken with hydrogen at 1 atm. and room temperature. After 1 hour 21.0 1t. of hy- drogen had been taken up, and tne rate oi hydrogen uptake had decreased to 1 ml./min. Theory required 2e.4-1t. The solution was filtered from the catalyst, concentrated to 100 ml., diluted to 1 1t. with water and filtered. ‘the residue was recrystallized from ethanol to give 435.8 g. (72%) of silvery crystals, m.p. 135-136°. Anal. Caled. for C1, Hy o05No (204.2): C, 64.71; H, 5.95; N, 13.72. Found: ©, 64.62; H, 5.88; N, 135.95. 2,6-Dimethoxy-8- (3-diethylaminovropylamino)-quinoline. A solution of 8.2 g. (0.04 mole) of 2,6-dimethoxy-8-amino-~ quinoline, 16.4 g. (0.2 mole) of sodium acetate and 7.4 g. 19 (0.04 mole) of 1l-diethylamino-3-chloropropane hydrochloride in 150 ml. of 66% ethanol was refluxed on the steam bath for 5 days with an additional 7.4, g. of 1l-diethylamino-3- chloropropane hydrochloride added each day. At the end of that time the solution was diluted to 300 ml. with water, made alkaline with 150 ml. of 25 wt.% sodium hydroxide, saturated with potassium carbonate and extracted three times with 50 ml. portions of ether. The ether extracts were dried over potassium carbonate, the solvent removed and the residue distilled in vacuo under nitrogen. After the unreacted 1-diethylamino-3-chloropropane had distilled at 60-80°/0.5-1.5 mm., the desired base came over as a viscous, yellow oil, b.p. 216-218°/0.15 mm. The yield was 11.0 g., 86% of the theory based on 2,6-dimethoxy-8-amino- quinoline. | The base was immediately dissolved in 45 ml. of ethanol. To this solution was added a solution of 9.4 ml. of hydriodic acid (sp.g. 1.7) in 43 ml. of ethanol. The red solution was warmed at 60-70° for a few minutes and chilled at 5° overnight. ‘the dihydriodide which had precipitated over- night was filtered and washed with ethanol to ylela 15.2 g. (76%) of colorless needles, recrystallized from ethanol, M.ep. 159-140° (dec.). Anal. Caled. for Gj —Ho90.Nzl, (573.3): C, 37.70; H, 5.10; N, 7453. Found: C, 37.91; H, 4.98; N, 7.28. 20 &-(3-Diethylaminopropylamino)-2,6-quinolinediol. 36.5 g. of 2,6-dimethoxy-8- (3-diethylaminopropylamino )-quinoline was dissolved in 182 ml. of 20% hydrochloric acid and the solution refluxed for 4$ hours. The red solution was cooled, neutralized with ammonium hydroxide, and the green gum which precipitated was collected by centrigugation. By stirring and scratching this gum with a few ml. of boil- ing ethanol, it was possible to convert it into a greenish yellow solid. ‘he solid was recrystallized from methyl cellosolve to give 8.5 g. (26%) of lemon yellow prisms, m.ep. 208-210° (dec.) Anal. Caicd. for Cy gHoz05N. (289.4): C, 66.40; H, 8.015 N, 14.52. Found: C, 66.14; H, 8.243 N, 14.60. el Ill. 6-Methoxy-8- (3-diethylaminopropylamino)-carbostyril Considerable difficulty was encountered in finding a workable synthesis of this compound. The first method of attack was based on the series of reactions: ~ S n7 OH XN H 4 7 ; ' However it was found that 6-methoxy-S-aminocarbostyril, obtained in a 91 per cent yield by the catalytic reduction of 6-methoxy-8-nitrocarbostyril, was an extremely high melting compound which was only slightly soluble in most organic solvents. The condensation with 1-diethylamino- 3-chloropropane hydrochloride therefore had to be attempted under heterogeneous conditions, and the starting matevrials were recovered quantitatively. The second attempt was based on the assumption that the 2-methoxy group in 2, 6-dimethoxy-8- (3-diethylamino- propylamino}-quinoline was more labile than the 6-methoxy group, and that the former could be hydrolyzed preferen- tially to the latter. However this assumption provea to -be fallacious, since even during very short (i hour) periods of refluxing in 20% hydrochloric acid the 6-methoxy group was hydrolyzed, resulting, in every attempt made, in the ee formation of 8-(3-diethylaminopropylamino)-2, 6-quino- linediol. It was also anticipated that, even if a pro- cedure of hydrolysis could eventually be worked out resulting in at least partial formation of the 2-hydroxy- 6=-methoxy derivative, the separation of the 2, 6-dihydroxy- and the 2-hydroxy-6-methoxy- derivatives would involve a rather complicated procedure, based on the relative solubilities of the compounds in sodium hydroxide, and probably resulting in a very low yield of the desired product. Therefore this line of attack was abandoned. The problem was finally solved by making use of the well known hydrogenolysis of benzyloxy compounds to the corresponding alcohols, using palladium catalysts (20,21). € oxen 2(qH) —~ (ars HOR © This iethod, though finding considerable application in sugar chemistry (22,23), had never been applied to pyri- dine or quinoline benzyloxy derivatives, but it proved to be excellent when it was applied to the present ser- ies of compounds. 6-Methoxy-S-nitrocarbostyril was benzylated in alka- line solution with benzyl chloride, giving a 68 per cent yield of the 2-benzyloxy derivative. The latter was reduced with iron and a 0.2% solution of hydrochloric 20 acid in ethanol to give a 64 per cent yield of 2-benzyl- oxy~6-methoxy-S-aminoquinoline. The primary amine was then condensed with 1l-diethylamino-3-chloropropane hydro- chloride in the usual fashion to give an 81 per cent | yield of 2-benzyloxy-6-methoxy~8- (3-diethylaminopropyl- amino)-guinoline, a compound which was unstable in air, though to a much lesser extent than the corresponding e~ methoxy analogue. Attempts were subsequently made to reduce the benzyl- oxy base as the dihydriodide derivative, but all proved unsuccessful. The catalysts used were palladium black. (24), palladium on charcoal (25) and palladium oxide. In each case 1 g..of the dihydriodide in 100 ml. of absolute | ethanol was shaken with an equivalent of 100 mg. of pal- ladium (e.g. 115 mg. of palladium oxide) under hydrogen at 1 atm. and room temperature for 1 hour. In. order to determine the amount of hydrogen required by the catalyst itself, the same amount of catalyst. suspended in 100 ml. of absolute ethanol was then shaken under the same condi- tions. Finally 166 mg. of benzyl alcohol, an amount which theoretically would absorb the same volume of hydrogen as 1 g. of the dihydriodide, was shaken with the same amount of catalyst in 100 ml. of absolute ethanol under the same conditions. It was found that each catalyst reduced benzyl alcohol 24 quantitatively: palladium on charcoal in 35 minutes, pal- ladium black in 60 minutes and palladium oxide in 11 min- utes after a 5 minute lag. When the dihydriodide was used, the hydrogen uptake decreased to 0.25 ml./min. after the amount of hydrogen had been absorbed which was required py the catalyst itself. It may be noted that Raney nickel was also used, but no hydrogen was taken up at all. Platinum catalysts were not used since the possibility existed of reducing the benzene ring in the benzyloxy side-chain, thus preventing hydrogenolysis (20). It was concluded that the iodide ion was responsible for these failures, and the free base was therefore employed in the following attempts at hydrogenolysis. Palladium oxide, since its rate of reducing benzyl alcohol wes the highest, was used as a catalyst. It was found: that, when absolute ethanol was the medium, the hydrogen uptake vir- tually stopped after the amount required by the free cat- alyst had been absorbed. However when a 4% solution of glacial acetic acid in absolute ethanol was used, complete reduction took place in 80 minutes. It did not seem feasible to isolate the free base as such from the reaction mixture, which contained an. equiva- lent amount of toluene, and the compound was therefore isolated as the hydriodide. The dihydriodide could not 25 be preparea, orobably because the quinoline nitrogen and the @-hydroxyl group tautomerize as lactam-lactim, thus destroying tne capacity of the nitrogen to accept protons. it remained to be shown whether benzyl chloride had benzylated: the oxygen or the nitrozen of tne carbostyril nucleus, that is, whether the benzylated intermediates were of one or the other isomeric forms I ana II: CHs0 ( * CHz0 ( . y? ott yO FEO (I) (Iz) where Y is -NOo, NHo and -NH.CH, CH, ot N(CoH.).- isomer would be expected to be caoable of reduction by CH Either palladium oxide and hydrogen, the reaction CS-origtty + 2 (4) —» Cons HNRo being as general (21) as \ . . €_ \-c1ig0% + 2(H) —= ( \-ons+ HOR By comsaring the absorption spectra of 6-methoxy-1l- mnethyl-8-nitro-2-auinolone, 2,6-dimethoxy-S-nitroaguinoline and the "unknown" benzylated compound where Y was -NOg, the isomeric form I was indicated. Experimental Part= 6-hiethoxy-8-aminocarbostyril. A suspension of 67.5 g. of 6-methoxy-8-nitrocarbostyril and 1.0 g. of platinum oxide in 500 ml. of absolute ethanol was shaken with. ‘hydrogen at 1 atm. and room temperature. After 4 hours 21 it. of hydrogen had been taken up, and the rate of hydrogen uptake had decreased to 1 ml./min. heory required 23.0 1t. The suspension was diluted with 400 ml. of water and filtered. The residue was dissolved in 1 lt. of hot 20% hydrochloric acid, the suspension filtered hot and the filtrate brought to pH 8 with concentrated ammonium hydroxide. The precipitate was filtered and dried to give 53.0 g. (91%) of 6-methoxy-8-aminocarbostyril. The com- pound was a greyish brown amorphous powder, which could be recrystallized only with difficulty from putanol or methyl cellosolve to give brown crystals, melting above 300°. Anal. Caled. for 01 Hy 90gN5 (190.2): CG, 63.14; H, 5.350; N, 14.735. Found: ©, 65.06; H, 5.72; N, 14.75. 2-Benzyloxy~6-methoxy-8-nitroquinoline. 83.7 g. of 6- methoxy-8-nitrocarbostyril was dissolved in 1670 ml. of All melting points are corrected; microanalyses by Dr. G Oppenheimer and G. A. Swinehart. 27 LO wt.% sodium hydroxide, and 480 ml. of benzyl chloride was added to the hot solution. The mixture was stirred vigorously on the steam bath for 8 hours, cooled and allowed to stand overnight. Two layers had formed. “The lower, light yellow aqueous layer was syphoned off, and the organic layer, which had set to a black semisolid, was boiled up with 14 1t. of ethanol. The ethanolic solution was chilled, filtered and the residue recrystal- lized from ethanol to give 80 g. (68%) of faintly yellow crystals, mep. 139-140°. Anal. Calcd. for 03 7H, 494No (510.3): C, 65.80; H, 4.55; N, 9.05. Found: C, 65.89; H, 4.73; N, 8.88. The results of the absorption spectra which were taken to indicate the structure of this compound are summarized in the following Table (I). The solvent used was ethanol, and the concentration of the different compounds in that medium was 6-methoxy-l-methyl-8-nitro-2-quinolone: 0.216 millimole/1t. 2, 6-dimethoxy-8-nitroquinoline “ee 0.216 2-benzyloxy~6-methoxy-8-nitroquinoline: 0.187 283 Table (I) Compound Maxiymum Minimum d (np) Bnolal ) bap Bolal 6-methoxy-l-methyl-8-nitro-2-quinolone 373 4860 507 14350 2, 6-dimethoxy-8-nitroquinoline 549 5530 297 1310 2-benzyloxy-6-methoxy-8-nitroquinoline | 351 | 4110 | 297] 1000 All measurements were made on a Beckman spectrophotometer. 2-Benzyloxy-6-methoxy~-8-aminoquinoline. 61.0 g. of 2- benzyloxy-6-methoxy-8-nitrogquinoline and 225 g. of iron by hydrogen were suspended in a solution of 6.7 ml. of 20% hydrochloric acid in 670 ml. of 96% ethanol. The suspension was refluxed with constant mechancial stirring for 6 hours. At the end of that period the suspension was filtered and the residue washed with 350 ml. of 96% ethanol. The filtrate and washings were combined and the solvent stripped off in vacuo at 100°. The residual oil was dark brown and solidified on cooling. to yield 45.5 g. (82%) of crude product. It was purified by distilling it in vacuc under nitrogen. The pure product came over at 210-212°/o.2 mm. as a light yellow oil which solidified readily to yield 35.2 g. (64%) of faintly yellow crystals. One recrystal- lization from 70% ethanol gave colorless needles which were analytically pure and melted at 86-87°. 29 N, 9.99. Found: C, 72.95; H, 5.90; N, 10.58. 2-Benzyloxy-6-methoxy-8- (3-diethylaminopropylamino)-quinoline. A solution of 35.2 ge (0.125 mole) of 2-benzyloxy-6-methoxy- 8-aminoquinoline, 51.6 g. (0.63 mole) of sodium acetate and 23.3 g. (0.125 mole) of 1-diethylamino-3-chloropro- pane hydrochloride in 470 ml. of 66% ethanol was refluxed on the steam bath for 5 days, with an additional 23.3 g. of 1l-diethylamino-3-chloropropane hydrochloride added each day. At the end of that time the solution was concentrated to 500 ml. and diluted to 1 1t. with water. It was made alkaline with 150 ml. of 25 wt.% sodium hydroxide, satur- ated with potassium carbonate and extracted three times with 75 ml. portions of ether. The ether extracts: were dried over potassium carbonate, the solvent removed and the residual oil distilled in vacuo under nitrogen. 40.2 g. (81%) of condensation product came over at 244-2469/0.06 mm. The base was dissolved in 124 ml. of ethanol and a solution of 28.5 ml. of hydriodic acid (sp.g. 1.7) in le4 ml. of ethanol was added. The solution warmed up and became red, while crystals appeared almost at once. The suspension was chilled at 5° for 2 hours and filtered. The residue was recrystallized from ethanol to give 45.5 g. (68.5%) of colorless needles, m.p. 124-1259 (dea.). 50 Anal. Caled. for CogHs,09N,1, (649.4): C, 44.593 H, 5.12; N, 6.47. Found: C, 44.46; H, 5.41; N, 6.65. 6-lwethoxy-8- (3-diethylaminopropylamino)-carbostyril. 11.5 g. of 2-benzyloxy-6-methoxy-8- (3-diethylaminopropylamino)- 1 was dissolved in 200 ml. of absolute ethanol. quinoline to the solution was added 8 ml. of glacial acetic acid and 1.2 g. of palladium oxide*, ‘The suspension was shaken under hydrogen at 1 atm. and room temperature. After 80 minutes, 0.95 lt. of nydrogen nad beeii susorhed, and the pate of hydrogen uptake had decreased to lL ml./min. It may be noted as characteristic of this catalyst that there was a 4 minute lag initially, and that in the following 7 minutes 0.4 1t. of hydrogen were absorbed. Theory re- quired 0.4 1t. of hydrogen for the catalyst and 0.73 Lt. for the base, a total of 1.16 lt. The solution was filtered from the catalyst and 15.5 ml. oi hydriodic acid (sp.g. 1.7) in SLO ml. of absolute ethanol was uaded. ‘ithe solution was allowed to stand ab 5° for 24 hours. At the end of that time the yellow crys- tals were filtered and recrystallized from ethanol to yield 1 his experiment was performed 8 hours after the base had been distilled from the condensation mixture. 2 Prepared by Dr. M. Rapport of these laboratories. OL 11.5 g. (91%) of faintly yellow crystals, m.p. 234-235° (dec.), of 6-methoxy-8- (3-diethylaminopropylamino)-car- bostyril hydriodide. — Anal. Caled. for Cy7Hpg0gNzI (431.4): C, 47.333 H, 6.07; . N, 9.74. . Found: C, 47.19; H, 5.99; N, 9.91. On shaking the hydriodide with 6N sodium hydroxide, a colorless oil was obtained which redissolved in N sodium hydroxide. The free base was not investigated further. of SUMMARY Four compounds related to pamagquine have been _ synthesized in order to determine the effect of a. e-substituent in the quinoline ring on the anti- malarial activity and toxicity of this type of drug. (14) (15) (16) (17) REPL AEWORS Kelsey, Geiling, Oldham, and Dearborn, J. Pharmacol. 80, 391(1944) ead and soepfli, d. Biol. Chem. 154, 507 (1944)- Johnson and Hamilton, J. Am. Chem. Soc. 65, 2867 (1941) Liderfield, Columbia University: Private Communication Knuppel, Ber. 29, 705(1896) Decker, Ber. 38, 1149,1150(1905) Decker and Stavrolopoulus, J. prakt. Chem. 68, 100, 101 (1903) Ing, J. Chem. Soc. 1951, 2202 Fischer, Ber. 31, 609(1898) Fischer and Guthmann, J. prakt. Chem. 93, 383,384(1916) becker and Pollitz, J. prakt. Chem. 64, 92(1901) Bradley ana Robinson, J. Chem. Soc. 1952, 1260 Cloke, Anderson, Lachmann and Smith, J. Am. Chem. Soc. 53, 2794(1951) liarxter, Helv. Chim. Acta 24, 214 (1941) Hass and Huffman, J. Am. Chem. Soc. 63, 1234(1941) Head, Columbia Univergtty: Private Communication hagiason and Strukov, Arch. Pharm. 271, 569(1933); C.A. 28, 1770(1934) Magidson, Strukov, Bobuishev and Torf, J. applied Chem. (U.S.S.8.) 9, 321(1936); C.A. 30, 5582(1956) Hichtmeyer, J. Am. Chem. Soc. 56, 1633(1934) 54 Baltzly and Buck, J. Am. Chem. Soc. 65, 1984(1943) Freudenberg and Plankenhorn, Ann. 536, 261(1938); Ber. 73, 624(1940) Freudenberg and Hull, Ber. 74, 237(1941) Tausz and Putnoky, Ber. 52, 1576(1919) Ott and Schroter, Ber. 60, 653(1927) PART II. Isomorphism in Relation to Serological Specificity. 55 INTRODUCTION The specificity of physiologically active substan- ces has long been recognized in such fields as immuno- logy, enzymatic catalysis and chemotherapeutic action. _ In immunology particularly, the systematic investigations of K. Landsteiner and others (1) have gradually given rise to the concept that the size and shape of such sub- stances is of fundamental importance. Under the insti- gation of L. Pauling, stress has now been laid upon the importance of intermolecular forces in biological re- actions, and an important parameter of this aspect of the problem is the size and shape of the reacting mole- cule or the biological agent (2). Granting that biological agents are hignly specific towards differences in size and shape of a molecule with which they react, the next questions which arise are: can a biological agent differentiate between mole- cules of the same size and shape, regardless of their chemical nature ? and how much can the shape of a mole- cule be varied before the biological agent is able to " ? In answering these questions, "notice the difference it must be remembered that the specificity is confined only to that part of a molecule which is involved in a biological reaction. The answer to the first question is 56 then: a biological agent generally is incapable of such differentiation, provided the binding forces between the biological agent and the molecules is of the same nature (i.e. mainly van der Waals attraction, or coulombic at- traction, etc.). The second question cannot be answered definitely, since the specificity varies from system to system, but in general small variations in the shape of a molecule change its activity considerably. Thus in the investigations of Landsteiner, which were later supplemen- ted by quantitative studies in These Laboratories (3), it was shown that the substitution of a methyl for a hydrogen in a hapten may lower a hapten inhibition cons- tant to one tenth of the original value (4). The activi- ties of the vitamin nicotinamide and of the analeptic agent Coramin are destroyed by the introduction of 2-methyl groups into their pyridine nuclei (5,6). On. the other hand, 2-methylnaphthoquinone (7) and 2emethylthio= _naphthoquinone (2) both have vitamin K activity. Here the absence of the phytyl group present in the vitemin itself shows that this part of the molecule is not involved in the piological reaction; yet the substitution of a thio- phene ring for a benzene ring did not destroy the activi- ty. Indeed the replaceability of the benzene ring by the thiophene ring, and the pyridine ring by the thiazole ring, in various biological systems has been demonstrated in 08 numerous cases (2, 8-12). In this connection it is interesting to note that camphor and pinacol, despite their different chemical nature, have approximately similar odors, because they have in comuon a quasi-globular shape (13). | This dependence of biological activity on shape, its variation with change in shape, and the fact that substan- ces with similar shape react similarly with biological agents (vide supra) (all regardless of chemical structure provided we restrict our discussion to non-polar parts of a molecule) sugsest that the criterion of isomorphism (see below) may be applied to such substances. In other words, substances which replace each other in one another's crystal lattice (forming solid solutions) because of si- milarity in external shape should be expected to act si-~ milarly as biological agents, by the same eriterion. The idea has been the subject of some investigation, princi- pally by Erlenmeyer and Lettré. However the questions which must first be answered are : is solid solution formation observed, provided the components have similar size and shape, regardless of the chemical nature of the components? and: how much can a molecule be varied before solid solution formation is no longer attainea?t These questions were purposely phrased This question composes the study of morphotropys— so-called by Groth in 1870 (16). 39 in terms analogous to those on page 56. In the discussion of these problems, we shall have to use the term isomorphous, which in the past has been used with perhaps more connotations and interpretations than any other. For our purposes, we shall define it as follows:. Two substances which, because of certain metri= cally and energetically similar lattice planes and plane spacings, form solid solutions? are called isomorphous”. The definition is almost cyclical, since we are consider- ing a priori only those pairs where analogy is expected from consideration of the molecular size and form. How- ever it is only these pairs which are of real interest to us, and we shall avoid artificial classifications (17) and terms like anomalous isomorphism (14, 17-19). Such vairs would by implication have similerly sized molecules and polarization properties. Hor a study of morphotropy, organic comoounds are best suited, since non-polar constituents can be gradually varied to assume various sizes and shapes. One of the first of such studies was performed by Ciamician (22, 25), who concluded that organic ring sys- tems with the same number of rings form isomorphous pairs, 1 Other criteria which we could have chosen are overgrowth and perallel growth. Neuhaus (14) called this "partielle Isomorphie", Kofler (15) "Wassenisomorphie", Weygand (20, 21) "Correspondenz". 40 i.e. the pairs benzene-thiophene and fluorene-phenan- threne. His experimental results are not to be trusted, however ( Jaeger (38); This Thesis). Grimm (24, 25) proposed in his "Hydridverschiebungs- satz" that isomorphously replaceable groups ("pseudoatoms") could be arranged periodically as the hydrides of non- metallic elements; thus Nt and CH= ; O+,NH- and.CHp=, etc., “are isomorphously equivalent. This theory, while being supported by an abundance of experimental evidence, is conjectural and has many exceptions; nevertheless it was the basis for the arguments used by Erlenmeyer (8-11) in his attempts to correlate structure and physiological activity. Other workers (26-52) have shown that the valence angle of pseudoatoms is of importance in considerations of iso- morphism ( thus diphenylmethane and diphenyl ether are not isomorphous whereas fluorene and diphenylene oxide are). Lettré (55), in a study of the isomorphous replace- ability of different groups in substituted benzoic acids, found that -H is not isomorphous with -Cl, -Br, -I, -CHz, -OH, -OCHz, or -NOp;+ that -CHs is isomorphous with -Cl and -Br but not with -I; that -Cl, -Br and -I are iso- morphous; and that -NOo, -OH and-OCHz are not isomorphous with any group”. This isomorphous replaceability is res- 1 put benzoic acid is isomorphous with p-fluorobenzoic acid (54). 2 exceptions:the pairs o-nitrobenzoic acid vs. o-toluic, -chlorobenzoic and o-bromobenzoic acids form solid so- 41 tricted to the benzoic acids, and in another type of compound these groups may or may not be isomorphous. The abnormal behavior of the -NO5, -OCHs and -OH groups (cf. also footnote 2, p.40) may be attributed to inter- | molecular hydrogen bonding . | Finally it may be mentioned that stereoisomers of the same compound usually are not isomorphous (16a, .35). | The questions concerning morphotropy which we have asked above may now be answered. Solid solution formation is generally observed, provided the components have ap- proximately similar lattice planes end spacings, and re- gardless of their chemical nature. The extent of morpho- tropy is definitely a functions of the type of molecule in question, but, in general, small variations may pre- clude the formation of solid solutions. Thus slmost complete analogy is attained between the two problems which we are attempting to correlate: iso~ morphism and serological specificity. Evidence for such a correlation is abundant. Yo mention a few examples: thiophene end benzene are isomorphous and can generally replace each other in biological systems (9, 12). 8-hydroxyquinoline (I) and 4-hydroxybenzothi- azole (II) are isomorphous (39) and thiazole and pyridine S \ S Y N N on * OH (I) (IT) 42 can generally replace each other in biological sys- tems (10)+, Globular molecules of the camphor type are isomorphous and have similar odors (13). Pyramidone (I) and isopropyl antipyrine (II) are isomorphous and both CHSC = GN(CHs) 2 CH30= CCH(CH;) p CH3N CsO CHsN G=0 SW (1) a! (1) \ Ces ~C6Hs5 powerful antipyretics (36). Phosphoric and arsenic acid salts are isomorphous and antigens with arsonic and phos- phonic acid haptens crossreact (3). Finally, the sub- stituted benzoic acids which are isomorphous (33) cross= react to about the same extent (4).° It must be kept in mind that: a) in biological re- actions only part of the molecule may be involved (and specific) whereas in the crystal the same molecule is completely surrounded by others. It is therefore iso- morphism which is the more delicate criterion. b) In vivo , the molecule is in solution, and thus may have a different effective form than in the crystal. The hy- The phase relationship of the pair thiazole-pyridine is not known. While nicotinic acid and 5-thiazolecarboxylic acia form solid solutions (This Thesis), the following pairs are not isomorphous: nicotinamide vs. 5-thiazole- carboxamide (40); 2,2'-dithiazolyl vs. 2,2'-dipyridyl (41); 4,4'-dithiazolyl vs. 2,2'-dipyridyl (41); 2,2'-dithiazo- lyl vs. 4,4'-dithiazolyl(41). 2 exceptions:the o-nitrobenzoic acid systems (footnote 2, p40) 435 drogen bonding and charge of a molecule (especially an acidic molecule) may cause the criterion of isomorphism to break down when applied to biological specificity. It is therefore not surprising that many substances which crossreact biologically? are not isomorphous; and that the absence of isomorphism does not mean the ab- sence of crossreactivity. We may however state the hy- pothesis: Two substances which are isomorphous will generally crossreact (and its corollary: the absence of biological crossreactivity generally means the absence of isomorphism). | As stated, the hypothesis is too qualitative to be. more than a working rule. It would be of interest to correlate sSemi-quantitatively the mutual solubility (in the solid state) of two substances, and their hapten in- hibition constants. Since, however, such a program would involve the preparation of many antisera, and since, furthermore, morphotropy usually causes a change from complete solubility to insolubility® it was decided that the working rule (vide supra) should first be investigated more systematically. For this purpose we chose the group of simple aromatic heterocyclic systems, such as furan, thiazole etc. The 1 the term being used in its broadest sense, i.e. sub-. stances which have similar biological activity (inhi- bition to that activity included). 2 For a fairly complete list of phase relationships, cf. (37). 44 “efforts of Erlenmeyer and others (8-12) to correlate iso- morphism and biological activity were restricted mainly to the systems benzene-thiophene and pyridine-thiazole. Landsteiner and Scheer (42) showed that 2-thiophenecarbox- ylic acid inhibited benzoic acid extremely well; that. ni- -cotinic acid, furoic acid, picolinic acid and 2-pyrrole- carboxylic acid all inhibited benzoic acid to an.inter- mediate extent. The substances chosen for this research were carbox- ylic acids. Not only are these substances soluble at se- rological pH, but their melting points lie in a range (100-300°) convenient for the ready determination. of phase relationships. Specifically, it was our intention that this research should investigate the following problems, which are of interest to the study of morphotropy as well as immuno- logy: a) Are the carboxy derivatives of the heterocyclic ring systems benzene, furan, thiophene, pyrrole, pyridine, pyrazine and thiazole isomorphous ? Of the simple rings, only the pairs benzene-thiophene (solid solution (43)) and benzene-pyridine (normal (45)) have been investigated. Of the derivatives of these ring systems (not counting condensed ring systems), the inves- tigations have been restricted to the derivatives of py- ridine and thiazole, mentioned on p. 42 (footnote 1) and 45 to a number of systems containing the pyrrole, the furan and the thiophene ring (22, 23)+, Tis latter investiga- tion is highly unreliable (38). b) What is the effect of changing the position of the carboxyl group on the ring, i.e. are the changed and ori- ginal molecules isomorphous ? Voerman (46) reported that 2- and 3-thiophenecarbox~ ‘ylic acids formed partial solid solutions. It would be interesting to investigate the isomorphism of 2-, 3- and 4-pyridinecarboxylic acids. The results of this investi- gation are part of This Thesis. c) If the sulfur atom in thiophene is sterically and electronically analogous to a vinyl group (e.g. thio- phene vs. benzene), then it would be of special interest to test the phase relationship of the pairs p-toluic acid vs. 5-methyl-2-thiophenecarboxyliz acid, m-toluic acid vs. 4-methyl-2-thiophenecarboxylic acid and o-to- luie acid vs. S=methyl-2-thiophenecarboxylic acid. ls- pecially crucial would be the second of these pairs, along with the pair p-toluic acid vs. 4-methyl-2-thio- ohenecarboxylic acid: COQOH OOH COOH oO 6 ="CHs Hs CHs This last experiment was unfortunately made impossible 1 ef. also (20) concerning chalcones, thiophene~chal- cones and furan-chalcones, 46 ‘for the time being, since the small yield of 4-methyl- z-thiophenecarboxylic acid could not be obtained in sufficient purity to warrant the determination of phase equilibria. | | d) Keeping the heterocyclic ring constant, will substituents which are usually isomorphous retain that property ? This was tested on the -Br and -CH, deriva- tives of 2-thiophenecarboxylic acid. e) If these heterocyclic acids were to be used as haptens in inhibition experiments against anti-X,, anti-X, and anti-X,, would there be a dependence of the inhibition constant on the results of a)-d), thus giv- ing added weight to the hypothesis that substances which are isomorphous will crossreact ? The hapten inhibition experiments which are necessary were rendered partly superfluous by the general absence - of isomorphism in pairs containing benzoic acid as one of the components (cf. Discussion of Results). Other hapten inhibition experiments which were contemplated have been held up for reasons beyond the control of the author. 47 I. Techniques and apparatus for the determination of phase diacrams The three principal methods of determining temp- erature-composition diagrams are 1) the cooling curve method, 2) the capillary method, and 3) the hot stage method. An adequate review of the scope of these me thods has been published (47). The relative advantages of the three methods have been compared by Grimm et al. (24) in a fairly exhaus- tive study of forty-five pairs of substances. His find- ings show that method 1) gives extremely uncertain re- sults at the s-phase boundary. This fact, coupled with the fact that thermal analysis requires relatively large amounts of material (of the order of one gram) and hence is useless as a micromethod, leaves for consideration only the other two methods. In the present investigation, a combination of both was used: the capillary rested on a hotstage. This has the following advantages of both methods: (a) Decomposition and sublimation: can be satisfactorily observed, and at the same time a fairly accurate deter- mination of the thawing point can be made despite initial sublimation. (b) No temperature correction is necded for a properly constructed setup.(c) The initial heating may be a rapid, and may be controlled with great ease. (d) The use of a microscope permits detailed observation of the melting prosess. A polarizing attachment facili- - tates observing the change from anisotropy to isotropy. 48 The microtechnique developed by Kofler (48-52), while more elegant than the hotstage~capillary method, requires photographic attachments and considerable prac- tice. We therefore thought it more expedient to use the. more conventional method for this research. A. Apparatus. The apoaratus consists essentially of a melting-point ‘block mounted on a polarizing microscope. The block was patterned after the one described by Vold and Doscher (53). I wish at this point to express my indebtedness to Dr. Doscher for valuable advice personally given me on the construction of the block and modifications thereof. The block is machined from "g41-0-Cel C-22" Celite- Magnesia bricks (Marine Eng. and Supply Co., Los Angeles); details of its construction may be found in reference (53). Minor modifications include a) the use of a rectangular (1"x 9/16"x1/2") copper strip, instead of the circular strip described in the paper (53), in order to heat the capillary more uniformly; b) the outside of the block was painted with 40 Be. sodium silicate solution and dried in the oven at 130° for two hours. This prevents. cracking or chipping of the otherwise fairly brittle S41-0-Cel. Heating was accomplished electrically by means of a 49 no.26 Nichrome wire, looped through the block and em- bedded in asbestos fiber. Heating was controlled by means of a variac, in series with a 2O ohm heating ele- ment. The samples rested next to the junction of a no.30 B&S gage iron-constantan thermocouple, on top of a small glass plate. The e.m.f. (0.05 mv./ °¢) was measured by ‘means of a Leeds and Northrup student potentiometer and a box-type galvanometer with a sensitivity of 0.05 Ma/nm., permitting readings of the accuracy of 0.2°C. | The thermometer was looped twice within the air-space, in order to avoid temperature corrections. The temperature was computed from standard calibration tables. The apparatus is usable from room temperature to 295°C (corresponding to 16 mv.), the cold junction being at 0°C.+ | The 32 mm. (4x) objective, in conjunction with . an Bxeyepiece, was found to give ample magnification. Observation of the melting process was facilitated by the use of a selenite compensator. | | In order to permit a controlled temperature rise at any given temperature, a series of temperature-time curves were plotted for a number of variac voltages. 1 At higher temperatures, the capillary was immersed in an oil bath, and the melting point observed by the conventional method. 50 B. Preparation of the mixtures (a) Preparation of the pure components The following reagents (EKC white label) were re- crystallized from water to constant melting point. p-Toluic acid, m.p. 177-178° m-Toluic acid, m.p. 110-111° o-Toluic acid, mep. 104-105° p-Bromobenzoic acid,m.p.e 252-253° Nicotinic acid, m.p. 233° Benzoic acid, m.p. 121-122° Pyrazinecarboxylic acid, m.p. 224-226° (dec.)- Picolinic acid, m.p. 136-1579 Furoic acid, mp. 132-1339 The following acids were prepared by different series of reactions. “Isonicotinic acid was prepared by passing hydrogen sulfide through a suspension of its copper salt (kindly supplied by Dr. D. R. Howton). m.p. 312-314 mee(sealed tube). Reported: 317° (55); 315° (56). | 5-Thiazolecarboxylic acid was prepared as follows. Bthyl formate was condensed with ethyl chloroacetate in the presence of potassium ethoxide (57) to give a 39% yield of formyl-chloroacetic acid ethyl ester (58). The condensation product was reacted with thioformamide (59) to give a 25% yield of 5+carbethoxythiazole (60). ol Hydrolysis of the ester in ethanolic potassium hydrox- ide yielded the desired 5-thiazolecarboxylic acid (60), mep. 218-220° (dec.) (reported 218° (61)). 2-Thiophenecarboxylic acid was prepared by iodinat- ing thiophene (Socony-Vacuum) in the presence of mercu- ric oxide (62) and converting the resulting 2-lodothio-~ © phene into the desired acid by the Grignard synthesis (63), the overall yield being 15%. Successive recrystallizat- fons from water and ligroin-benzene gave colorless necdles, m.p. 127-128°. 3-Thiophenecarboxylic acid was prepared by the fol- lowing series of reactions. Ethyl crotonate (64) was prepared by the esterification of crotonic acid (EKO practical) with ethanol in the presence of sulfuric acid, in a 56% yield. Treatment of the ester with po- tassium cyanide and barium hydroxide resulted ina 70% yield of methylsuccinic acid (65). The sodium salt of this acid was heated with phosphorus trisulfide to. give a 28% yield of 3-methylthiophene (66). The pro- cedure for the ring closure was based on the prepara- tion of thiophene by Phillips (67) and comments by Steinkopf (68). Oxidation of 3-methylthiophene with alkaline potassium permanganate (69, 70) gave a 3.7% yield of 3-thiophenecarboxylic acid, Successive re- 52 erystallizations from water, ligroin-benzene and subsequent sublimation gave colorless platelets, m.p. 137-1389, 3-Methyl-2-thiophenecarboxylic acid (71) was pre- pared by iodinating 3-methylthiophene (cf. p. 51), pre- paring the Grignard reagent of the resulting 2-iodo-5- methylthiophene, forming the carbon dioxide adduct, and hydrolyzing, to give the desired acid (m.p. 146-147° efter three recrystallizations from water). Phe yield on the basis of 3-methylthiophene was 44%. | 5-Bromo-2-thiophenecarboxylic acid , MePe 140- 141° was prepared in a 34% yield by the direct bromi- nation of 2-thiophenecarboxylic acid in glacial acetic acid (72). 5-Methyl-2-thiophenecarboxylic acid was prepared ‘as follows. Levulinic acid was heated with phosphorus trisulfide to give a 8.5% yield of 2-methylthiophene (75), the procedure of Phillips (67) being employed. Iodination (cf. (62)) of that compound gave a 80% yield of 5-i0do-2-methylthiophene (74). This compound gave a 55% yield of 5-methyl-2-thiophenecarboxylic acid (74), mp. 138-1399, by the Grignard synthesis. 4-Nethyl-2-thiophenecarboxylic acid could not be purified sufficiently to warrant its use as a component. 53 It was prepared as follows. Ethyl 2-bromopropionate (75) was prepared in a 77% yield by refluxing 2-bromoproplo- nic acid and ethanol in the presence of sulfuric acid. The condensation of this ester with acetoacetic ester in the presence of sodium ethoxide resulted in a 66% yield of ethyl 2-aceto-3-methylsuccinate (75). The hydrolysis of the condensation product by SN hydrochloric acid gave a 73% yield of -methyllevulinic acid (75) which, on heating with phosphorus trisulphide in the usual fashion, gave a 18.5% yield of 2,4-dimethylthiophene (71, 76). Oxidation of this compound with alkaline potassium per- manganate (71, 76) gave a 4.7% yield (0.45 g.) of crude 4-methyl-2-thiophenecarboxylic acid, m.p. 95-108°. Re- crystallization from benzene, sublimation and four re- erystallizations from ligroin raised its melting point to 110-119° (reported (76) 118-119°). Since only 0.1 g of material was left, further attempts at purification were abandoned. 2-Pyrrolecarboxylic acid was prepared as follows. Pyrrole was prepared in a 351% yield by the pyrolysis of ammonium mucate (77). 2-Pyrrolecarboxylie acid (78, 79) was obtained in a 24% yield by treating pyrrole with methyl magnesium iodide, forming the carbon dioxide ad- duct and hydrolyzing (a procedure desctibed in detail by Fischer (80)). The acid melted at 190° (dec.). y 54 (b) Preparation of samples of the mixtures Weighed amounts of the two components are loosely mixed in 2 ml. pyrex beakers, and then thoroughly ground up with the flattened end of a small stirring rod.+ ‘the average mixture weighs 10 mg. In the case of the acids investigated (many of which sublime and decompose), this method of preparing. samples is superior to the melting-solidifying-crushing tech-- nique (for a critical analysis of preparation of samples -ef. (24)). Iwo melting point tubes are then filled with the mix- ture and sealed about 15 mm. from the end. The two samples give a check on the homogeneity of the mixture. | Ce Observational procedure (a) Beginning of melting (thawing point) In this, as in the other cases, the heating rate Was | 241° fain. The thawing point caused much difficulty in the beginning, because transmitted light was used, and the amount of liquid (especially in mixtures of composition close to one of the pure components) was too small to be detected. The technique which finally gave reproducible Care must be taken in weighing the individual components into the beaker, so that the resulting mixtures will be fairly evenly distributed over the composition range. On the average, 9 to 10 such mixtures are prepared per diagram, except when one of the components is available in only very small quantities. 55 results consisted in observing the beginning of melting 7 by brignt reflected light. By observing the tip of the capillary, the thawing point can be very accurately no- ted. Sometimes, as in the case of partial solid solu- tion formation, the first droplets resolidify; this phe- nomenon can, with some practice, be recognized unequivo- cally. (b) End of melting (melting point) This point is readily determined ab the disappear- ance of the last crystal. The polarizing microscope facilitates this observation, transmitted light being used. (c) Decomposition In many cases, decarboxylation, demonstrated by the evolution of bubbles, is observed before all of the ma- terial has melted. This has the effeet of lowering the » liquidus, but it has no effect on the position of the S-phase boundary (cf. Discussion of Results). D. Heliability of results The accuracy of the method, as judged by the repro- ducibility of the results, is+ 2° for the thawing and melting points and+ 4° for the decomposition points (be ginning of decomposition). In order to compare the results of the method with reported results, two typical phase diagrams were plotted, both of which had been reported by Lettré et al. (54). These authers used both the ca- pillary method and the hot stage method, their results 56 being the average of several determinations. In the pre- paration of their samples they used the melting-resoli- difying-crushning technique. The results of the compari- son are reported graphically in figs. I and II. The check is satisfactory. p-CHLOROBENZOIC ACID vs. p-TOLUIG ACID L- —(}— LETTRE et al. (1936) ---(]}--- THIS THESIS 230 210 20 40 60 80 % p-CHLOROBENZOIC ACID Fig. I o7 P-CHLOROBENZOIC ACID vs. p-HYDROXYBENZOIC ACID —C)— eT TRE et ol. (1936) — ---(])--- THIS THESIS 230 2lo ig90 % P-CHLOROBENZOIC ACcIO Fig. Il Ee Example of a typical determination The vesults of all the determinations will be repor- ted in terms of the temperature-wt.% composition coor- dinates of the points only. The following, in tabular form, are the data leading to the final results. The example chosen is the pair Benzoic acid (A) vs. Furoic acid (B). Table I gives the results of the weighing of the mixtures; table II gives the results of the deter- mination of. the various melting ranges corresponding to the different mixtures. 58 Table I Weighings Bea-| Wt. Wt. plus |Wt. plus |Net wttNet wt.| Total Wt ker | empty(g.)A (g.) B (g) A (mg)|B (mg) | wt. Al A nO. | plus. B 1 |1.4969 |1.5059 | 1.5073 | 9.0 1.4 10.4 | 86.5 2 (1.4542 1.4628 | 1.4653 | 8.6 2.5 [1l.1 | 77.5 3 1.3680 1.5751 | 1.3786 | 7.1 3.5 10.6. | 66.9 4 1.3519 1.5585 |1.3635 | 6.6 5.0 11.6 |5'7.0 5 1.5100 1.5153 |1.5209 | 5.3 5.6 10.9 |48.6 6 1.5730 1.5777 |1.5842 | 4.7 | 6.5 11.2° |42.0 7 1.4797 1.4834 [1.4925 | 3.7 9.1 12.8 |28.9 8 1.4832 1.4862 |1.4970 |3.0 10.8 13.8 |21.7 9 1.3942 1.5954 |1.4058 li.e [20.4 11.6 10.4 fable IL Melting ranges Bea4 Tube THAWING POINT MELTING POINT k * no. | wv. [avers] °c | mv. | aver.| 1 2.48 5.50 1 4.48 85el 587 110.8 1 4.42 5260 a 4.48 85.1. D208 105.5 2 4.55 D296 1 4.44 5200 3 4.47 85.0 5.06 96.0 2 4.49 Self . 1 4.45 9200 4 4.47 85.0 5202 9562 2 4.48 5.05 1 4.52 5.89 2 4.48 5267 : - L 4447 6.05 6 4.49 85.5 | 6.08 114.5 2 4.50 6-10 1 4.46 6250 7 | 4.47 85.0 6.46 121.5 2 4.47 6.242 1 4.45 6.60 2 4,50 - 1 4.47 6.81 9 4.52 85.8 6.76 127.2 2 4.55 6.70 59 60 ii, Phase relationships of pairs of different carboxylic acids. The data given on the following pages are plotted as phase diagrams, the coordinates being temperature and wte% composition. The symbols for the points are point on phase boundary. beginning of decomposition. condition (such as melting point after decomposition has set in). —O- --()-- ooint not representing an equilibrium => —OC- initial melting with subsequent reso- lidification. Excepting a few cases, no line has been drawn through points not representing an equilibrium condition. The diameter of the circles and the height of the tri- angles indicates the precision of the measurements. In the tabulations which accompany each phase diagram, the percentage compositions refer to the component indica- ted on the diagram. The abbreviations designate, vespec- tively: Thepe - Thawing point Mep. - Melting point Dec.p.-Decomposition point BENZOIC ACID vs. 2- THOPHENECARBOXYLIC ACID 130 10° 90° O-00—O0-C20—O | L | | | ! 20 40 60 80 % BENZOIC ACID Benzoic acid (A) vs. 2-Thiophenecarboxylic acid % A TheD. M.De % A Th.p. Nop. O 8.4 17.1 20.7 30.8 38.3 41.5 44.5 127 94.6 94.8 94.8 95.4 94.8 95.8 95.2 128 125.8 120.4 118.6 115.2 112.0 110.2 109.2 55.5 71.0 76.7 83.5 86.0 90.3 100 95.4 94,4 94.8 95.6 95.6 97.6 121. 104.8 109.2 111.6 116.2 117.0 120.6 122 62 BENZOIC ACID vs. 2-PYRROLECARBOXYLIC ACID 20d}- 5 0 nr) ® a Oo oo og 160 \ ® — © = A O 140 | A A \ i20° — oO Oo = oO a! oO Oo O- Ole - -q 100 | | | | | | | | | 20 40 60 80 % BENZOIC ACID Benzoic acid (A) vs. 2-Pyrrolecarboxylic acid % Bh Thepe DecePe Mepe Thepe Decep-. 0 10.3 20.6 28.7 38.5 49.0 61.1 70-7 190 116.8 115.6 113.2 115.6 116.4 115.0 116.8 190 158.4 150.6 142.4 140.6 132.4 136.0 129.6 190 180.4 178.4 176.2 172.8 171.2 165.0 158.2 81.4 91.0 100 115.4 115.6 121 133.8 127.0 153.2 127.6 122 635 BENZOIC AC. vs. FUROIC ACID 120° 100° 80°r- 20 40 60 80 % BENZOIC ACID Benzoic acid (A) vs. Furoic acid % A O 10.4 21.7 28.9 42.0 48.6 57.0 Thepe 152 85.6 85.1 85.0 85.5 85.5 85.0 NeDe 153 127.2 124.0 121.5 114.5 109.2 95.2 % A 66.9 77.5 86.5 100 Thep. 85.0 85.1 85.1 121 Mepe 96-0 105.5 110.8 122 140? 120° 1007 BENZOIC ACID. vs. 3-THIOPHENECARBOXYLIC ACID 64 20 40 60 80 % BENZOIC ACID Benzoic acid (A) vs. 3-Thiophenecarboxylic acid. GA Th.epe Mepe % A TheDe MeDe % A Thep. 0 9.5 15.9 25.0 37.8 41.4 42.8 157 102.9 103.2 105.5 105.5 102.6 104.0 138 152.6 151.2 122.6 107.8 105.4 106.0: 47.7 52.0 57.8 61.5 66.2 70.2 76.1 101.0 103.6 101.0 100.6 105.0 105.2 105.0 104.2 107.2 107.0 108.0 108.8 108.0 115.4 79.4 80.0 87.5 90.0 95.0 100 104.1 103.6 104.4 104.2 110.6 121 117.2 116.0 118.2 119.0 120.0 lzZe 65 | BENZOIC ACID vs. 5-THIAZOLECARBOXYLIC ACID b— 210°} . = ® © 0 rot A I70%}— A 150°}— 130°}— toe -O— ae) Oo —O | | | | | | | | | 20 40 60 80 % BENZOIC ACID Benzoic acid (A) vs. 5-Thiazolecarboxylic acid $A 0 7.6 22.0 32.7 50.0. 69.0 78.5 93.3 100 Theps 218 111.4 109.6 111.0 111.2 110.6 111.0 111.6 121 Decep. 218 192.8 183.6 178.6 169.6 167.8 MeDe 220 205.5 199.4 195.4 185.0 172.4 152.0 120.0 122 130° lo 90° BENZOIC ACID vs. PICOLINIC ACID 66. | | jt a {et l 20 40 60 80 % BENZOIC ACID Benzoic acid {A) vs. Picolinic acid & OA O 11.6 18.6 30.3 38.3 46.5 58.8 Thep. 136 88.4 87.6 86.4 87.0 87.2 87.4 Mepe 157 154.2 151.0 123.8 116.0 107.4 102.4 % A 67.2 75.5 86.6 100 Thep. 87.0 87.6 87.6 121 Mepe 107.4 111.4 116.2 122 220° 200 180° 160° 140° 120° BENZOIC ACID vs. NICOTINIC ACID % BENZOIC ACID Benzoic acid (A) vs. Nicotinic acid ba Thepe hi.De van Thepe MieDe 0 960 20.0 28.6 4564 59.5 66.6 200 115.0 115.4 115.4 114.2 115.0 114.6 255 227.6.217.6 211.2 195.8 184.0 169.8 79.8 69.0 96.8 100 114.0 113.6 114.0 121 142.2 117.8 120.4 122 67 BENZOIG ACID vs. ISONICOTINIG ACID oN a TTTTTTTITITTTT ITT TTY 68 % BENZOIC ACID Benzoic acid (A) vs. Isonicotinic acid thePpe HeDe % dece 6) 6.1 356.8 55.9 61.65 °68.4 81.8 $12 118.6 115.4 115.0 115.0 115.0 114.8 514 504% 282.8 255.4 264.2 255.4 225.2 89.4 96.1 100 115.4 115.4 121 199.4 158.8 122 BENZOIC ACID vs. PYRAZINECARBOXYLIC ACID % BENZOIC ACID Benzoic acid (A) vs. Pyrazinecarboxylic acid BA ThePe DecePe MeDe % A Thepe Dec.p e M.epe 0 9.1 18.8 25.8 355.8 41.8 56.6 e224 111.8 114.0 115.0 115.0 114.6 115.6 224 189.8 189.0 188.6 179.8 176.0 170.6 226 205.0 204.2 203.2 198.4 194.2 186.2 62.6 75.0 87.2 95.8 100 114.4 114.6 114.4 114.0 121 167-2 164.2 184.0 175.4 159.4 127.8 122 210° 0) I90%}— = y d A A \ O @ 170°}— A Oo i 4 A . 5O0%— 130°%}— Oe a en ie o> °° © | | | | | | | | 20 40 60 80 170° — I30 70 2-PYRROLECARBOXYLIG ACID vs. 2-THIOPHENECARBOXYLIC ACID Dh Cp sOb= © dy’ 40 60 8 % 2-PYRROLECARBOXYLIC ACID 2-Pyrrolecarboxylic acid (A) vs. 2-Thiophenecarboxylic acid % A O 9.7 19.4 289.0 38.7 48.2 59.4 Thepe 127 121.8 121.8 125.0 126.4 126.0 125.2 Dec.p. 129.4 130.0 131.2 130.6 129.8 150.0 MeDe 128 129.4 140.65 154.6 158.4 164.0 % A 75.5 82.6 100 Thepe 154.2% 141.6% 190 Dec.p. 154.2 141.6 190 MaDe ' 190 % Initially melted at 126.4 , the resolidified. ad oo om ere me care 2-THIOPHENECARBOXYLIC AC. vs. FUROIC AGID = 30 10° 90°|- ce es ee eS GO 20 40 60 BO % 2-THIOPHENECARBOXYLIC ACID 2-Thiophcnecarboxylic acid (A) vs. Furoic acid So AA fe) 7.8 BeS 55.5 87.5 49.2 O7.9 thep. 162 120.5 112.0 102.2 100.9 100.0 Oo €0 8 hede 156 131.0 127.2 118.0 115.2 107.6 106.8 O 85.2 30.8 95.5 100 theo. LOl.ee 101.1 106.1 106.7 111... 12" Rep. 116.0 115.2 122.6 125.4 126.5 128 2ioe I9O°r i70° i5C° Hoe 2-THIOPHENECARBOXYLIC ACID vs. 5-THIAZOLECARBOXYLIC ACID a ® — j ® | 0) | i ® | A ° ! | r Q } —o om O- O om “O— | — 1 | | _| 20 40 60 % 2-THIOPHENECARBOXYLIC ACID 2-Thiophenecarboxylic acid (A) vs. 5-Thiazole- carboxy % A Thepe Dec.p. rlic acid 0 Dee 218 1168.2 2l& 200.4 220 209.4 115.0 114.8 115.4 114.2 114.2 116.0 127 152.6 174.8 165.8 165.4 201.8 192.2 185.6 176.2 159.6 124.6 128 2-PYRROLEGARBOXYLIC ACID vs. FUROIC ACID 15 170° 50° KI 30° DO ® A 20 Se AH GO. | 40 % 2-PYRROLECARBOXYLIC ACID wePyrcolecepborylic eeic (A) yo fh WheDe LECeDe Lie De Oo &.7 LSS 1s. 125.8 Lie. LeT Le7 160 129.2 156..4 144.5 7005 20.0 LOO Oe 190 oO es coe Vey ty cm L127 Le7 145.6 Ls0 ion LE6.6 190 VSe Bupoic ecid NICOTINIC ACID vs. 5-THIAZOLECARBOXYLIC ACID ~ 240} 230 220 wo > gt 2co T => \ LS I90°-— | | | | | | | | | 20 40 60 80 % NICOTINIC ACID a i Nicotinic acid (A) vs. S-thiazolecarboxylic acid an 0 5.9 19.0 34.8 48.5 62.3 77.8 93.% Thede” 218 199.0 200.8 201.4 205.0 210.8 217.4 227.0 liede 220 209.8 206.0 206.6 208.8 217.0 225.0 282 4 ae aw identical with dec.p. (exceot for 100% coordinate) ne 7 7 2k Lu et ie ¢ pt PICCLINIC ACID vs. NICOTINIC ACIL “ C 3G fe) ome’ \& — on 1, re) aS 1) O— | | | | | | | 20 40 60 80 Picolinic scic (a) vs. hicotinie ecid VEC eDe 215.3 109.0 160.0 171.6 164.2 170.5 lieDe 205 250.2 219.2 217.4 201.2 166.0 16.5 jo A 587.5 61.6 98.5 100 VEC eDe neds 157.8 108.2 156.0 187 ISONICOTINIG ACID vs. PICOLINIC ACID ee ee eee ee a | | | 20 40 60 % ISONIGOTINIG ACID Tsonicotinic acid (A) vs. oO 8) e on fe O e He tw IO) ar ji jo #h hep. 15 Oo a) os rae) . oO a) ta te e Ho a oO w UEC ede 155 166 L656 ieDe 157 200.8 2165.8 251 7 3 A 70.€ 80.5 930.5 thepe 152.4 162.4 152.4 Picolinic acid eo 0965 oO 7 eco 1561.2 152.4 166 LES 02 514 80 O 77 i PICOLINIG ACID vs. PYRAZINECARBUs1_46% ACID ro Z ~2/\0 Cr co Cc |) bOK ISQ FE O——————O- oO oO oO = % PICOLINIC ACID ot mae at nts ore Picolinic seid (a) vse Pyragzinecerboxylic acid bs oO he © ae) ay © ® He © e o>) Ht Oo aes © © . en oO © ° en) TheDe Lae 199.4% 129.8 150.5 150.8 150.4 129.5 YEC+D. 224 192.2 180.4 169.6 165.0 158.4 156.0 ueDe vc6 205.8 202.6 200.2 194.8 192.4 186.6 SoA 58.0 80.3 92.4 100 Thed. 150.6 129.6 129.4 126 Oo 157.8 (o; Oy becede LE bt ut) = * initielly melted at 151.4, then resolidificd. > ISONICOTINIG ACID vs. NICOTINIC ACID 320 }—- 300 280 260 240) 220° % ISONICOTINIC ACID lhed. 256 215.8 215.6 214.0 217.0 215.90 215.6 Nie De wD 1.0 Coke 8 208-6 4°27 00 Co0ek 290.0 yo # 70.5 61.2 90.0 150 hed. 297% 897% 300% S14 fy of a > % Gece. #0 £02. NICOTINIC AGID VS. PYRAZINECARBOXYLIC ACID iy <0) < 230 ) Ome é >) , ° — 2lO ~ on" , an 7/\\e ‘ : "> ye’ ---- a o iI90 -— 20 40 60 80 % NICOTINIC ACID Nicotinie acid (A) vs. Pyrazinecarboxylic acid wy & O 8.9 17.1 29.7 57.7 50.9 59.5 ray w won + + ype o rast PSN EE ata Py ~ hiaDe 215.6 214.9 211.6 207.6 207.9 200.6 iheDe 202.6 202.9 215.5 Ro hueDe 215.2 219.8 288.5 2b5 identicel with dec.o. (cceotine the 100. coorcinste) 310° 290}-- ISONICOTINIC ACID vs. PYRAZINECARBOXYLIG ACD 80 270 |— 2301— © ») 210° 20 40 60 80 %ISONICOTINIC ACID Isonicotinic acid (A) vs. Pyrazinecarboxylic acid 6 A QO 10.3 20.2 30.0 38.5 49.0 59.5 Thepe 224 206.8 207.4 208.2 207.0 206.8 209.4 M.pe 226 214.2 230.8 252.2 bs 72.8 66.8 100 Thepe’ 207.6 206.8 312 Mepe 308 314 *sdentical with Cec.eDe 81 5-METHYL-2- THIOPHENECARBOXYLIC ACID vs. 5-BROMO-2-THIOPHENECARBOXYLIC ACID 20 40 60 80 % 5-METHYL-2-THIOPHENECARBOXYLIC ACID 5-Methyl-2-thiophenecarboxylic acid (A) vs. 5-Bromo- e-thiophenecarboxylic acid % A O 11.8 19.7 25.2 28.8 37.9 50.0 52.35 Thep. 140 135.0 128.6 127.4 119.90 118.2 117.2 117.2 Mepe 141 140.2 159.0 157.8 155.4 1835.0 129.0 124.8 % A 60.2 64.0 72.8 75.9 84.8 88.8 100 Thepe 117.2 117.4 117.2 117.0 124.8 128.6 158 Mepe 125.2 127.6 128.8 131.4 155.6 136.0 159 82 220° igo” 160° 140° 20 40 60 80 % p-BROMOBENZOIC ACID p-Bromobenzoic acid (A) vs. 5-Bromo-2-thiophene- carboxylic acid % A O 5.0 18.4 25.7 39.3. 43.4 58.8 Thep. 140 139.0 157.6 137.6 157.8 137.6 137.4 Meps 141 143.8 181.6 187.6 206.6 210.4 228.4 % A 7062 80.6 84.6 100 Thep. 137.8 137.8 137.8% 252 Mep. 237.8 245.6 249.0. 253 # rpesolidifies 0o-TOLUIC ACID vs. 3-METHYL-2-THIOPHENECARBOXYLIC ACID 130° 110° 90°F | ! I | | | | | 20 40 60 80 % o- TOLUIC ACID o-Toluic acid (A) vs. 5-lethyl-2-thiophenecarb- oxylic acid % A 0 9.0 18.8 27.1 35.1 48.1 53.5 Thep. 146 133.2 121.0 115.6 105.6 689.2 88.2 Meps 147 142.0 1355.6 155.4 127.8 124.8 119.6 % A 65.28 81.6 91.5 100 Thepe 88.4 87.6 88.4 104 Mepe 105.6 99.0 102.2 105 p-TOLUIG AC. vs. S-METHYL-2-THIOPHENECARBOXYLIC ACID. 84 160° |— — 140° 120° |—- | | ae | | | | | 20 40 60 80 % p-TOLUIC ACID p-Toluic acid (A) vs. 5-liethyl-2-thiophenecarb- oxylic acid % & O 9.4 19.2 27.9 45.0. 50.0 60.0 Mops 139 135.2 130.4 127-4 144.4 147.0 151.8 $A 70.5 79.5 90.0 100 Thep. 116.2 116.8 154.2 177 Mep. 159.0 165.4 175.4 178 8 nay ome on oa p-TOLUIC ACID vs. 5-BROMO-2-THIOPHENECARBOXYLIC ACID 85 (70 150° 130° % p-TOLUIG ACID p-Toluic acid (A) vs. 5-Bromo-2-thiophenecarb- oxylic acid % A 0 6.6 19.0 28.5 39.8 49.1 58.0 Mop. 141 159.6 138.2 146.2 150.6 157.0 168.0 aa 69.6 86.9 100 Theo. 134.6 154.2 177 Mep. 168.8 175.8 178 86 Be interpretation of data Benzoic acid (A) vs. 2-thiophenecarboxylic acid. A eutectic at 63% A and 95°. There is an indication of solubility at ca. 90% A. Benzoic acid (A) vs. @-pyrrolecarboxylic acid. A eutectic at 93% A and 115°. Decomposition in the region 0-90% A at 2 130° makes the liquidus indeterminate. No indi- cation of solid solution formation. Benzoic acid (A) vs. furoic acid. A eutectic at 60% A and 85°. No indication of solid so- lution formation. Benzoic acid (A) vs. 3-thiophenecarboxylic acid. A eutectic exists at 40% A and 103°, and there is an in- dication of solubility at ca. 90% A. The region 40-80% A does not lend itself& to ready interpretation. However there ‘night be another eutectic at ca. 64% A and 103°, and an ad- dition compound at ca. 53% A (mop. 109°). A 1:1. comoound would be at 49% A, Benzoic acid (A) vs. 5-thiazolecarboxylic acid. A eutectic at 95% A and 111°. Decomposition in the region 0-'70% A at 170° makes the liquidus indeterminate. No indi- eation of solid solution formation. Benzoic acid (A) vs. Picolinic acid. A eutectic at 53% A and 115°, yo indication of solid so- lution formation. 87 Benzoic acid (A) vs. nicotinic acid. A eutectic at 90% A and 115°, No indication of solid solution formation. | Benzoic acid (A) vs. isonicotinic acid. A eutectic at ca. 99% A and 115°. No indication of solid solution formation. Benzoic acid (A) vs. pyrazinecarboxylic acid. A eutectic at ca. 98% A and 114°, Decomposition in the region 0-80% A at2165° makes the liquidus indeterminate. No indication of solid solution formation. 2=Pyrrolecarboxylic acid (A) vs. 2-thiophenecarboxylic acid. A eutectic at ca. 5% A and 124°. Decomposition at the solidus makes the phase boundaries indeterminate. Solid solution formation begins at 60-70% A. 2-Thiophenecarboxylic acid (A) vs. furoic acide. Type III solid solution formation (44) with a min intum at 54% A and 100°. The diagram might also be interpreted as type V solid solution formation with a eutectic at 54% A and 100° and a solubility gap at 58-77% A. 2-Thiophenecarboxylic acid (A) vs. 5-thiazolecarb- oxylic acid. A eutectic at 89% A and 116°. Decomposition in the region 0-70% A at >165° makes the liquidus indeterminate. No indi- cation of solid solution formation. 2-Pyrrolecarboxylic acid (A) vs. furoic acid. A eutectic at 15% A and 124°. Decomposition at the 88 solidus make the phase boundaries indeterminate. No indi- cation of solid solution formation. Nicotinic acid (A) VSe 5-thiazolecarboxylic acid. Type V solid solution formation with a eutectic at 26%. A and 200°, and « solubility gap at 0-35% A. Decomposition at the solidus makes the phase boundaries indeterminate. Picolinic acid (A) vs. nicotinic acid. A eutectic at 87% A and 122°. Decomposition in the re- gion 0-60% A at»165° makes the liquidus indeterminate. There is an indication that solid solution takes place at concentgrations of ¢10% A. Isonicotinic acid (4) vs. picolinic acid. A eutectic at. ca. 2% A and 152°. Decomposition at 2165° makes the liquidus indeterminate. No indication of solid solution formation. Picolinie acid (A) vs. pyrazinecarboxylic acid. A eutectic et 95% A and 130°. Decomposition in the re- gion 0-85% A at %160° makes the Lliguidus indeterminate. Solid solution formation at less than 20% A. (a) Isonicotinic acidavs. nicotinic acid. A eutectic at 18% A and 215°. Decomposition in the re- gion 70-100% A at 2 292° makes the liquidus indeterminate. No indication of solid solution formation. Nicotinic acid (A) vs. pyrazinecarboxylic acid. Type V solid solution formation with a eutectic at 89 55% A and 202°, and a gap at 50-82% A. Decomposition at _ the solidus makes the phase boundaries indeterminate. Isonicotinic acid (A) vs. pyrazinecarboxylic acid. A eutectic at 12% A and 207°. Decomposition at the so- lidus makes the phase boundaries indeterminate. No indi- ‘eation of solid solution formation. (a) o- -Methyl- -2- thiophenecarboxylic ecidavs. 5-bromo-2- thiophenecarboxylic acid. Type V solid solution formation with a eutectic at 56% A and 117°, and a gap at 31-77% A. p-Bromobenzoic acid (A) vs. 5-bromo-2- thiophenecarb- oxylic ecid. A eutectic at 3% A and 138°. Above 80% A, solid solution formation seems to take place, but the large range makes the determination of resolidification and remelting high- ly uncertain. o-Toluic acid (A) vs. 3-methyl-2 ~thi ophenecarboxyl ic acid. Type V solid solution formation with a eutectic at 75% A and 88°, and a gap at 48-100% A. p-Toluic acid (A) vs. 5-methyl-2-thiophenecarboxylic acid. A eutectic at 25% A and 116°, with solid solution ta- king place at concentrations 780% A. p-Toluic acid (A) vs. 5-bromo-2-thiophenecarboxylic acid. A eutectic at 14% A and 134°, with solid solution ta- king place at concentrations 275% A. 90 DISCUSSION OF RESULTS The following conclusions can be reached from a study of the phase relationships investigated in this research. a) Benzoic acid is not isomorphous with any of the he- _terocyclic acids. In the case of 2= and 3-thiophenecarb- - oxylic acids, there does exist an indication of solubili- ty at about 90% benzoic acid, due, probably, to the re- markable physical similarity of benzene and thiophene (9, 12). This insolubility of benzoic acid with heterocyclic acids, as well as the absence of solid solubility amongst most of the heterocyclic acids, would suggest that the spe- cificity of hydrogen bonding (nature, strength) is signi- ficant amongst these acidst. It also follows that sero- logical studies using the various heterocyclic acids as haptens against anti-benzoic acid sera cannot be coprelated satisfactorily with these solubility studies. In agree- ment with our results, however, are the results of Land- steiner and Scheer (cf. pe 44) concerning the inhibition of benzoic acid by some heterocyclic acids. b) While 2-thiophenecarboxylic acid is only ca. 10% soluble in benzoic acid, the 5-substituted (Br, CH) 2- thiophenecarboxylic acids are about 20% soluble in the corresponding para-substituted benzoic acids. In the case 1 ef. in this connection the remarks on p. 41 concerning Lettré's work. 91 of an ortho-substituent, the solubility becomes greater: - o-~toluic acid is about 50% soluble in 5-methyl-2-thio-~ phenecarboxylic acid. This increase in solubility, which is significant, is another manifestation of the "ortho- effect" (81;. This Thesis p.40, footnote 2). | e) The isomorphogeny of Br and CHz is maintained: 5-bromo-2-thiophenecarboxylic acid is 18% soluble in p- ~toluic acid. Furthermore, 5-methyl-2-thiophenecarboxylic acid and 5-bromo-2-thiophenecarboxylic acid are soluble in each other. d) The pyridinecarboxylic and pyrazinecarboxylic acids show an interesting relationship: while isonicotinic acid does not form solid solutions with any of the other acids, picolinic acid forms 10% solid solutions with nicotinic — acid and 20% solid solutions with pyrazinecarboxylic acids in addition, nicotinic acid forms solid solutions with pyrazinecarboxylic acid having but a relatively small so- lubility gap(50-80% nicotinic ecid). | As a first step toward an explanation of these results, let us list some pertinent properties of the acids in duestion (Table I): 92 fable I Picolinic Nicotinic |Isonicotinic Pyrazinecarb. acid acid acid acid hep. 136° 233° 312° 223° . Dec.temp.™ ca. 170° -- ca. 300° Cae 170° Log K(82,83) -5.30 -4.87 -4.97 . . Sol.H,0 , Ves. s.h. seh. Sol. EtOH VeSe Sehe sl.s.h. Sol.Et,0 vesl.S. veslese VeS1.S. The behavior of the acid dissociation constant (which, we propose, is actually related to the base constant of a zwitterion type molecule) cannot be explained unless we assume a positive charge on the nitrogen. This at onee sug- gests a type of intermolecular bonding remtniscent of thet of the simple amino acids (84, 85), i.e. partly coulombic, partly hydrogen bonding. To judge from the melting point, decomposition point and solubility data, this intermolecu- lar bonding vrobably takes place to a much larger extent in isonicotinic acid than in its isomers. Again, the same data suggest that picolinic acid exists as a low order poly- mer, perhaps as a dimer (intramolecular chelation has been suggested (86, 87) to account for its low decomposition temperature. This is not a necessary assumption). % The See (g) for a discussion of decarboxylation. 95 ‘relative positions of the ring nitrogen and the carboxyl may account for the difference in internal arrangement of the erystals. We can thus explain the low solubility of the pyridine carboxylic acids in each other. The fact that pyrazinecarboxylic acid is much more. sol- ‘uble in nicotinic acid than in picolinic acid suggests that “its intermolecular bonding resembles more closely that of nicotinic acid than that of picolinic acid. Strength is lent to this argument by the fact that the melting points of pyrazinecarboxylic acid and nicotinic acid are very close. (e) While 5-thiazolecarboxylic acid does not form solid solutions with 2-thiophenecarboxylic acid, it is soluble in nicotinic acid (40-100% nicotinic acid). This result clearly indicates that interchanging a =CH- and a -S= in a molecule already highly specifically bonded (such as ni- cotinic acid) has little effect on the intermolecular array, wheras it may be sufficient otherwise (as in benzoic acid) to prevent isomorphism. It also shows that the interchange of a =CH-= and an =<N- under those circumstances will pre-= vent isomorphism. The results strongly suggest that, whereas the hetero-N is involved in intermolecular bon- ding, the hetero-S is not. In connection with the isomorphism of thiazole and py- ridine (cf. 9.42) it is of interest to note that many pairs (cf. p42, footnote 1) do not form solid solutions. An explanation of that fact is not obvious. It is also 94 noteworthy that 5-thiazolecarboxylic acid decarboxylates quite readily, although it is formally analogous to nico- tinic acid. (f) The interpretation of the phase diagrams which in- volve as components the 5-membered ring acids (pyrrole-, furan- and thiophenecarboxylic acids) is not straightfor- ward. 2-Pyrrolecarboxylic acid appears not to be isomorphous with: the other two acids although slight solubility is in- dicated with 2-thiophenecarboxylic acid. However 2=thio- phenecarboxylic acid forms solid solutions with furoic acid. This result is unexpected (especially in view of the in- solubility of 2-thiophenecarboxylic acid and benzoic acid) and not readily interpretable. | (g) The decarboxylation of many of the acids offers some interesting results. | The acids for which decomposition is observed are pi- colinic acid, pyrazinecarboxylic acid, 2-pyrrolecarboxyl- ic acid, 5-thiazolecarboxylic acid and isonicotinic ecid. The following are the conclusions drawn from the decompo- sition behavior of these acids. (1) Decarboxylation, as judged by the evolution of gas, takes place only when a liquid phase has been estab- lished. While it is not out of the question that a surface effect is involved, it seems more likely that the acids simply decarboxylate most readily when they are in the 95 liquid state or in solution. (2) At small concentrations of the decarboxylating component, the decomposition begins ab a fairly definite temperature, the decarboxylation temperature. As the con-= centration of the decarboxylating component increases, the “decarboxylation temperature begins to rise also, correspon- ding to the fact that enough liquid has to be present to start decarboxylation. The decarboxylation temperature is independent of the other component and may be above or below the melting point of the decarboxylating component. For the acids in question, we may construct the following Table I. Table I Carboxylic acid meper decarboxylation temperature 2-Pyridine- 155° ca. 170° Pyrazine- 225° ca. 170° 2-Pyrrole- 190° cae 130° 5-Thiazole- 220° ca. 170° 4-Pyridine~ 315° ca. 300° me rounded off The two conclusions elaborated above explain complete- ly the phase diagrams whim involve decomposition. 96 SUMMARY a) In connection with a study on the relation between serological specificity and solid solubility, twenty-five . binary phase relationships have been investigated, the com- “ponents being aromatic carboxylic acids. b) In agreement with the strong inhibition of benzoic acid by 2-thiophenecarboxylic acid, the thioohenecarbox- ylic acids form very limited solid solutions with benzoic acid, which is not isomorphous with any of the other hete- rocyclic acids investigated. Isomorphism is noted with some methyl and bromo substituted benzoic and thiophenecarbox- ylic acids. | c) The phase behavior of pairs involving picolinic, ni- cotinic, isonicotinic and pyrazinecarboxylic acids may be explained on the basis of a difference in hydrogen bonding in crystals of these acids. d) The isomorphism of thiazole and pyridine has been substantiated. — e) The study of the decarboxylation behavior of some of the acids investigated leads to some fundamental conclu- sions, involving the existence of a decarboxylation tem- perature. (1) (4) (5) (6) (7) 97 REFERENCES General references: a) Boyd, “Fundamentals of Immu- nology", Interscience Publishers, New York 1945, | Chs. 2, 3, 6. b) Landsteiner, "The Specifici- ty of Serological Reactions", Harvard University Press, Cambridge 1945. / a) Pauling, Campbell and Pressman, Physiol. Reve. 25, ~ 203(1943). b) Pauling, Chem. Eng. News 24, 1064, 1575(1946) Pressman, Pardee and Pauling, J. Am. Chem. Soc. 67, 1607(1945) and earlier papers in this series. Pressman, Swingle, Grossberg and Pauling, J. Am. Chem. Soc. 66, 1731(1944). Dornow, Ber. 75, 78(1940) Baumgarten and Dornow, Ber. 72, 565(1939) John, Angew. Chem. 54, 209(1941) References (8) to (11) are the principal papers of a study by H. Erlenmeyer on a relation between structure | and physiological activity. The principal coworkers were Berger, Leo, Meyenburg, Ueberwasser and ileiber. ~ (8) (9) (10) Early (general) papers; a) Erlenmeyer et al., Biochem. 2. 282, 22(1952); BOD, 429(1932)3 262, 196 (19335). bv) Erlenmeyer and Berger, Arch. exptl. Path. Pharmacol. L77, 116 (1934). c) Berver and Erlenmeyer, Klin. Wochschr.. 14, 536 (1935). Physiological equivalence (and related subjects) of benzene and thiophene. Erlenmeyer et al., Helv. Chim. Acta 16, 756, 1381 (1933) Physiological equivalence (and related subjects) of pyridine and thiazole. Erlenmeyer et al., Helv. Chim. Acta 20, 204, 510(1957); 21, 709, 1017, 1695(1938); 22, 698, 938(1939); 25, 249, 375, 515, 1066(1942); 29, 1946 (1946). (an) (12) (13) (14) (15) (16) (17) (18) (19) (20) (21) (22) (23) (24) (25) (26) (27) (28) (29) (30) 98 Physiological equivalence of oxazole and thiazole; related subjects. Erlenmeyer et al., Helv. Chim. Acta 20, 1588, 1446 (1957); 21, 111, 1010, 1013(1938). Steinkopf, "Die Chemie des Thiophens", p.29; Verlag Th. Steinkooff, Dresden and Leipzig 1941. Pirsch, Ber. 69, 13235(1936). Neuhaus, Z. phys. Chem. 191, 359(1943) Kofler, Z. phys. Chem. 191, 227(1942) For this and other general discussions of historical interest, see a) Bruni, "Reste Loesungen und Isomorphismus", Akademische Verlagsgesellschaft, Leipzig 1908. b) Arzruni, "Physikalische Chemie der Krystalle", Vieweg & Sohn, Braunschweig 18935. Van't Hoff, Z. phys. Chem. 5, 524(1890) Neuhaus, Z Kryst. 103, 297(1941) Neuhaus, Ber. 76, 531(1943) Weygand and Strobelt, Ber. 68, 1859(1935) Weygand, Ber 76, 5355(1943) Ciamician, Garelli and Ferrati, Z phys. Chem. 13, 1(1894) Ciamician and Garelli, Z. phys. Chem. 18, 51(1895) Grimm, Guenther and Tittus, Z. phys. Chem. 14, 169(19351) Grima, Naturwissensch. 17, 555(1929) | Luettringhaus, Ann. 528, 223(1937) Luettringhaus et al., Ber. 73, 145, 763, 1022(1940) Cullinane and Rees, Trans. Farad. Soc. 36, 507(1940) Cullinane and Flummer, J. Chem. Soc. 1958, 606 Rheinboldt and Mathias, Ber. 73, 4353(1940) (31) (32) (33 (34) (35) (56) (37) (38) (39) (40) (41) (42) (48) (49) (50} (51) (52) (55) 99 cruni, Ber. 73, 763(1940) Pascal end Normand, Bull. Soc. chim. 13, 137, 207(1913) Lettré et al., Ber. 69, 1151(1936); 70, 1410(1937); WL, 416, 1225(1938); 75, 386, 1150(1940) hoopal, Rec. trav. chim. Pays-Bas 34, 115(1915)— Lettré, Ergebn. Enzymforschung 9, 1(1943) | Erlenmeyer and Willi, Helv. Chim. Acta 18, 740 (1935) Landolt-Bornstein, Physikealisch-chemische Tabelle oth ed., Hptw.,6943 Erg.I, 264; Ii, 575; ‘Til, 565 Jaeger, 2. Kryst. 42, 256(1906) Erlenmeyer et al., Helv. Chim. Acta 21, 1695(1938) ibid. 25, 375(1942) ibid. 22, 698, 958(1939) a) Landsteiner and Scheer, J. Exper. lied. 54, 295(1931) b) Landsteiner, J. Immunol. 33, 265(1937) ‘Psakalotos and Guye, J. chim. phys. 8, 540(1910) Roozeboom, Z. phys. Chem. 30, 585(1899) Pickering, Je Chem. Soc. 1895, 998 Voerman, Hec. trav. chim. Pays-Bas 26, 295(1907) Skau and Wakeham, Ch. I, Vol.I of Weissberger's "Phy- sical methods of Organic Chemistry", Interscience Publishers, New York 1945. kofler, Z. phys. Chem. 187, 201, 265(1940)5 l9l, 241 (1942) ~~ ~~ hofler, Z. Elektrochem. 47, 810(1941); 50, 104(1944) hofler, Ber. 75, 998(1942); 76, 391(19435); 77, 110 (1944) Kofler and Brandstatter, Z. phys. Chem. 192, 60, 71, 229 (1943) — Brandstatter, Z. phys, Chem. 191, 227(1942); 192, 76, 82 (1943) Vold and Doscher, Ind. Eng. Chem., Anal. Ed. 18, 154 (1946) (54) (55) (56) (57) (58) (59) (60) (61) (62) (63) (64) (65) (66) (67) 100 Lettré, Barnbeck and Lege, Ber.69, 1151(1936) Skraup, Monatsh. f. Chem. 17, 569(1896) Camps, Archiv @. Pharm. 240, 359(1902) Wislicenus and Silberstein, Ber. 43, 1827(1910) © Wislicenus, Ber, 43, 3530(1910) _ | Private Communication from Dr. D. Rh. Howton Erlenmeyer and hieyenburg, Helv. Chim. Acta 20, -204 (1937) Erlenmeyer and Morel, Helv. Chim. Acta 25,. 1073 (1942) Minnis, in "Organic Syntheses" Coll. Vol.II, p.357 Schlenk and Ochs, Ber. 48, 679(1915) Michael, Ber. 55, 5766(1900) Higginbotham and Lapworth, J. Chem. Soc. 1922, 51 Volhard and Erdmann, Ber. 18, 455(1885) Phillips, in "Organic Syntheses" Coll. Vol.II, pp. 578-579 , Steinkopf, reference (12), 2.34 Damsky, Ber. 19, 5284(1886) Muhlert, Ber. 18, 3003(1885) Rinkes, hec. trav. chim. Pays-Bas 52, 1052 (1955) Steinkopf, Jacob and Penz, Ann. 512, 160(1934) Kues and Paal, Ber. 19, 556(1886) Rinkes, Rec. trav. chim. Pays-Bas 52, 538(1933) Bischoff, Ann. £06, $14 (1881) Zelinsky, Ber. 20, 2021(1887) (77) (78) (79) (80) (81) (82) (85) (84) (85) (86) (87) 101 McElvain and Bolliger, in "Organic Syntheses" Coll. Vol.I, p. 473 Oddo et al., Gazz. 39, 656(1909); 42, 244(1912) Fischer-Orth, "Die Chemie des Pyrrols", Vol.I, p.237, Akademische Verlagsgesellschaft, Leipzig 19354. Fischer, in "Organic Syntheses" Goll. Vol.II, p.-198 Branch and Calvin, "The Theory of Organic Chemistry", pp. 257-265, Prentice-Hall, New York 1941. Ostwald, Z. phys. Chem. 3, 385(1889) Erlenmeyer, Epprecht and Meyenburg, Helv. Chim. Acta 20, 510(1937) | | Albert and Corey, J. Am. Chem. Soc. 61, 1087(1939) Levy and Corey, J. Am. Chem. Soc. 63, 2095(1941) Dyson and Hammick, Jd. Chem. Soc. 1957, 1724 Ashworth, Dafforn and Hammick, d. Chem. Soc. 1959, 809 102 1 ' PART III A Study on the Hammick Reaction 1 . Part of this work was done in collaboration with Dr. C. McCloskey. 103 INTRODUCTION The Hammick reaction (1, 2) proceeds formally according to the scheme 2. where Py-. stands for the X -pyridyl radical. Of a large num- PyCOOH + Ry, -CO-Ro —> PyC(0H)R1Ro + CO ‘ber of decarboxylating agents tried (2), Ashworth et al. found that only picolinic, quinaldinic and isoguinaldinic acids underwent the reaction. Their results, i.e. that the carboxyl must be X to the ring nitrogen, which itself must be tertiary, led them to the conclusion that the, reaction was of the nature of a cyanohydrin addition reaction, i.e. py CV. coon > C3.- + CO, +H yielding a carbanion formally resembling the cyanide ion. the first step Their sparse experimental evidence has since been amply enlarged by thorough experiments done in this Laboratory by Drs. Buchman, Golding and Sargent (3). The following Table I summarizes their main results. Often much ketone, instead of carbinel, was isolated. The third column indi- cates the ratio of carbinol (C) to ketone (K). The symbol C, for example, by itself means that only carbinol was | isolated. The yield is the total yield (C+K) based on picolinic acid. fable I Aldehyde RCHO R- Yield C:K Phenyl 34% ¢ 4-Chlorophenyl | ol GC. 3,4-Dichlorophenyl 22 433 2, 6-Dichlorophenyl 24 521 2,4=-Dichlorophenyl 20 K 1-Naphthyl 25 3:2 4-Chloro-l-naphthyl Cae 2 K 4-Methoxyphenyl 22 8:35 5,4-NMethylenedioxyphenyl Cae 8 235 4-Hydroxy-d5-methoxyphenyl cae O 4-Hydroxyphenyl cae O e~Hydroxyphenyl cae O e-Furyl el 431 3-Nitrophenyl 29 —«K 4-Dimetnyleminophenyl cae 1 K From these data we may draw the conclusion that the Hamnick reaction is simply an electrophilic reaction of the carbonyl carbon: + - . RCH-O + Ga.- —> RCHOH Ry, H 104 105 and that, as more electrons are contributed to the ring by various substituents, the positive charge on the carbonyl carbon decreases, causing the aldehyde molecule to be less reactive with a concurrent decrease in yield. It is inter-— esting to note that, as the yield decreases, the ratio- CiK decreases also, in general. This can be understood if we make the justifiable assumption (4) that, as the+ E ef- fect of the substituents increases, the reduction potential of the resulting pyridyl-arylketones increases. Evidence exists (1, 2) that the yield of carbinol is increased if the reaction takes place under nitrogen’. The Hammick reaction, then, is a new type of carbonyl addition reaction. the effect of ring substituents on the yields of carbinol seems to bear little relation to the stability of the corresponding mandelonitrile (Table II): Table II Aldehyde RCHO Dissociation constant Yield by of substituted mande- Hammnick R= . lonitrile (5) reaction (3) K x 10 — (%) Phenyl - | 0.47 : 34 4-Nitrophenyl- 0.27 29 2-Hydroxyohenyl- 1.67 : fe) 4-Chlorophenyl- 1.81 ~—~él 4~Methoxyphenyl- 512 22 4-Hydroxyphenyl- 7.66 0 4-Diethylaminophenyl- 59.00 6) o wt We may hence infer that air is the oxydizing agent. 106 We therefore See no virtue in the formalism employed by Ashworth et al. (2). | | We propose the following mechanism. I.Decarboxylation RCOOH + B —* RGOO + BH RCOO -» R + COp Then, in the absence of competing electron acceptors R + BH ~ RH + B™ II. Formation of carbinols R + RyR,CO > RRReCO™ RRjRoCO” + BH —* RRYRoCOH + B™ The ionization of the acid as a primary step secms impor-~ tant, to judge from the effect of amine catalysis in many cases of decarboxylation (6). Contrary to Ashworth et al., we do not believe that intramolecular chelation is neces- sary for this purpose. Rather, it is preferable to say that inductive and electromeric effects are at an optimum in picolinic scid (probably a gwitterion) to cause its decarb- oxylation. It is furthermore syperfluous and perhaps mis- leading to emphasize the analogy between pyridyl and cyanide ion. In support of our thesis we have shown experimentally that isonicotinic acid can also be made to undergo the Hammick reaction. ‘To summarize, our modified theory differs in two impor- tant aspects from the theory proposed by Ashworth et al. (2): 107 1) The relative positions of the carboxyl and the basic center do not have to be adjacent as in «-imino acids, and are related only in as far as they affect the stability and reactivity of the carbanion. 2) The formalism of the cyanohydrin analogy is totally unnecessary, and the reaction is treated as an event per se, related to carbonyl additions as a whole, but not to the cyanohyarin addition particularly. In addition to the experiment in which isonicotinic acid was subjected to the Hammick reaction, other, unrelated, ex- periments were performed in an effort to elucidate the struc- ture of some by-products obtained in the Hammick reaction. We had felt for some time that these by-products mignt be benzoins, in view of the fact that the carbinols were, ac-= cording to Ashworth et al. (2), mandelonitrile analogues, and that these might react further with aldehyde to give, finally, the corresponding benzoins (7). We were encouraged in this hypothesis by the isolation of the pyridine analogue of the hypothetical benzoin cyanohydrin, but this compound did not yield benzoin on heating, but rather dissociated into benzaldehyde and pyridylphenylcarbinol. ‘while one of the | otherby-products might be a benzoin, the evidence is not conclusive. 1 Done in collaboration with Dr. C. licCloskey. « 108 A. 1, 2-Diphenyl-1- (2-pyridyl)- ~ethylene glycol. This compound was obtained as a by-product of the benzal- dehyde ~ picolinic acid condensation. Its structure was sug- gested by analysis and lead tetraacetate oxidation, and con- firmed by synthesis. The glycol was unstable at high bem peratures, dissociating into pyridylohenylcarbinol and ben- zaldehyde, and it was found that an equilibrium existed be- tween these compounds and the glycol at elevated temperatures. Although a quantitative study was not contemplated, it was evi- dent that the dissociation of the glycol was nearly complete at temperatures above 200°, The amounts of glycol formed in the Hammick reaction very nearly equaled that of the equi- librium mixture at the same temperature. This dissociation of the glycol is similar to that of benzpinacol, which is re- ported to dissociate into benpyarol and benzophenone, The formation of benzpinacol by a reverse reaction (by heat alone”) has not been shown. a : .o4 Ke) hxoerimental Part ‘ton Lsolation of the glycol from the Hammick reaction mixture. A mixture of 250 g. of dry picolinic acid and 2 kg. of re- distilled benzaldehyde was heated at 1560- 170° for seventeen hours. The resulting solution was diluted with an equal 1 . First noted by Dr. H. Sargent 2 The formation of triphenylglycol, hydrobenzoin end benzpi- nacol by the illumination of a mixture of benzyl alcohol end benzophenone has been reported (10) 3 Melting points corrected; analyses by Dr. Oogenheimer and staff 1LO9 volume of ether and extracted twice with 4 WH hydrochloric acid. The extracts were basified with aqueous ammonia, and an oily precipitate formed. One liter of isopropyl ether was added and the crystals which formed at the interface were filtered off, to give ca. 5 g. of product. On recrystalli- ‘zation from toluene it formed white needles, mp. 191-192° (dec., imzersion temperature 175°)1, The compound is soluble in 2N hydrochloric acid, sparingly soluble in hot. toluene and insoluble in common organic solvents and water. Anal. Caled. for G Hy 7O0oN (291.3): C, 78.343 H, 5.88; WN, 4.81. 19 Found : G, 78.25; H, 5.79; N, 4.60. The glycol forms a hydrobromide hemihydrate which crystalli- zes from acetone in the form of prisms, m.p. 194-196 (dec. ) Anal. Calcd. for Cj gH, 70, NeHBr.2H0 (381.2) C, 59.87; H, 5.02; N, 5.68. _ Found: C, 59.853; H, 4.93; N, 5.74. The base, on quentitative titration with lead tetraacetate in glacial acetic acid (11), was found to use up one equi- valent of this reagent. | From the isoprovyl ether solution there was obtained 108 g. of 2-pyridylphenylcerbinol as previously described (3). An exhaustive investigation of the remaining syrups gave no indication of the presence of benzoin. 1 The melting point varies with thé immersion teniperature and the rate of heating. 110 Synthesis of 1,2-diphenyl-1-(2-pyridyl) ethylene glyco1.? To the Grignard reagent prepared from 32 g. (0.2 mole) of 2=bromopyridine and 10 g. of fine magnesium turnings was added a hot solution of 15.8 g. (@.075 mole) of benzoin dis- solved in 100 ml. of dry thiophene-free benzene. The ‘reaction “mixture was refluxed end stirred mechanically for thirty mi- nutes and then poured unto 50 ml. of saturated emionium chlo- ride solution containing some iee. the reaction flask was washed out with 150 ml. of chloroform and the washings added to the ice mixture. The resulting mixture was stirred well and the solid filtered off, washed with a little alcohol and dried. | The crude material was dissolvea in 250 ml. of boiling pyridine and filtered from the unreacted magnesium. The re- sidue was washed with 50 ml. of hot pyridine and the filt- rates combined. The pyridine solution was heated to voiling, water added until the glycol was just in solution and on cooling the 1,2-diphenyl-1- (2-pyridyl) ethylene glycol crys- tallized out. The crystals were filtered off, washed with a Little. alcohol and dried. There was obtained 8.64 Be (39.6%), mepe 200-201° (rapid heating). A mixed melting point with the compound obtained from the Hammick reaction gave no depression. Anal. Calcd. for C,H 70, Ni GC, 78.543 H, 5.88; N, 4.81. rh | Found : C, 78.16 a 5.91; N, 4.82. 1 Performed by Dr. C. licCloskey. Lil - Equilibrium between e-pyridylphenylcarbinol (A), benz- aldehyde (B) and 1,2-diphenyl-1-(2-pyridyl) ethylene glycol{C). a) Reaction of (A) and (B) A solution of 18.5 g.(0.1 mole) of highly purified (A) and 08 ge (0.5 mole) of redistilled (B) was heated for nineteen hours at 165-170°. The solution was diluted with 60 wl. of ether and then extracted four times with 25 ml. of 2.5 N hydrochloric acid. The extract was basified with aqueous po-= tassium hydroxide, 50 ml. of isopropyl ether was added and the mixture shaken. The solid material was filtered off and washed with isopropyl ether. There was obtained 350 mg. of (C), m.p. 188° on recrystallization from methyl cellosolve. from the ethereal layer of the filtrate was recovered 8.2 g. of (A). A second run at 165-170°,using the same quantities of (A) and (B),gave 450 mg. of (C). . Repeating the run at 190-195° (in a sealed tube) with 6.48 g. of (A) and 17.7 ml. of (B), 1 me. of (c) was obtained. At 130-135°, using 6.48 g. of (A) and 17.7 ml. of (B), . 140 mg. of (C) was obtained. | A reaction of benzoin and picolinic acid at 165° yielded none of the addition compound (C)v b) Dissociation of (C) Ll. A solution of 1.0 g. of (C) in 175 ml. of toluene was refluxed for one hour. There was recovered 0.9 Be of unchanged glycol (C). 2. The glycol (0.25 g.) was heated at 220° for one hour. 112 The odor of (B) was.marked. A little toluene was added and the, solution extracted with 3 N hydrochloric acid. The exe tract was basified and the precipitated oil dissolved ina mixture of ligroin (60- 70° ) and isopropyl ether. On seeding, crystals of (A), mp. 14 75°, formed. 3. A mixture of 5.96 g. of (0) and 4 ml. of (B) was heatea at 165° for eighteen hours. The reaction mixture was cooled to room temperature, an equal volume of ether added. and the solution extracted with two 20 ml. portions of 10% hydro- echlioric acid. The combined extracts were basified with aqueous ammonia, 50 ml. of isopropyl ether added and the mixture shaken. The crystals which formed were, filtered to Give 55 mg. of starting material (C). The isopropyl ether layer was evaporated to a syrup and dissolved in ligroin (60-70°) containing 15% of isopropyl ether. After standing, erystals were deposited which were separated from the mother- liquor to give 0.65 g. of (A), m.p. 72-75? The crystals did not depress the melting point of an authentic sample of (A). The combined filtrates were distilled in vacuo to give 1.5 ge of an oil, b.p. 110-120°/2 mm. , and 0.22 g. of high boiling residue. The oil was reduced by the method of Meerwein and Ponndorf to give 1.14 g. of (A). The total yield of (A) was 1.79 ge (79%). 1135 B. A by-product from the Hammick reaction employing 2, 4~dichlorobenzaldehyde. When 2,4-dichlorobenzaldehyde was condensed with pi- colinic acid in the usual fashion, there was obtained (3) from the neutral fraction, after extraction with base and acid, a small amount of material which was recrystallized from benzene in large, well-formed prisms, m.p. 156-157°. tests indicated the absence of nitrogen in this compound, ana it was tentatively formulated as 2,4,2',4'-tetrachloro- benzoin on the basis of its elementary analysis. gnal. Caled. for 0,,HgC1405: 6, 48.04; H, 2.50., Found.: C, 48.61; H, 1.92. 48.48 1.76 in an attempt to prepare this compound by the benzoin con- densation, 2,4-dichloromandelonitrile was obtained (5). We succeeded in isolating a compound from the benzoin conden- sation which proved identical with the product from the Haranick reaction. This compound analyzed for one active hyarogen per molecule and formed a dinydroquinoxaline with o-phenylenediamine, but it did not reduce Fehling's solution and was recovered unchanged after boiling with concentrated nitric acid. These facts do not allow for a definite assign- ment of structure. A more concentrated attack on the problem has unfortunately been delayed so far, other research programs taking priority. 114 Experimental Part™ Preparation of 2,4,2',4'-tetrachlorobenzoin(?) by the benzoin condensation, An attempt to prepare this compound by conventional methods (12) failed. The following method was finally employed. | | . To a suspension of 27.3 g. of freshly distilled 2,4-d1- chlorobenzaldehyde in 42 ml. of ethanol was added a solution of 5.5 g. of potassium cyanide in 8 ml. of water. The solu- tion was refluxed forty-five minutes and poured into a solution of 10 g. of sodiua bicarbonate in 200 ml. of water. The so- lution was shaken for thirty minutes and then extracted three times with 50 ml. portions of ether. The ether extracts were evaporated, the residue redissolved in benzene and the solution shaken one hour with a saturated solution of sodium bisulfite. The benzene layer was separated from the aqueous phase and the bisulfite addition compound, dried and fractio- nated under nitrogen. The first fraction, b.p. 82-85°/0.15 TMM. y consisted of still unreacted aldehyde. the second fraction consisted of a yellow, viscous oil, b.p. 177-178°/0.15 mm. the yield of this crude 2,4,2',4'-tetrachlorobenzoin(?) was 6.0 ge (22%). The oil slowly solidified, and after four re~ crystallizations from ethanol gabe very pale yellow ..: crystals, M.epe. 156-1579. A sample of this compound did not depress the melting point of a sample obtained from the Hammick reaction. liielting points are corrected; microanalyses by Dr. Oppenheimer and staff. 115 Anal. Caled. for Benzoin, ¢..1 : ¢ .043 H, 2.30, c | or enzoins C1 He0t 0, C, 48.043 Hy 2.50, Calcd. for Benzil, Cy 4HeCl,053 C, 48.52; H, 1.72. Found: 0, 48.15; H, 1.92. Anal. (Zerewitinoff) | Found: M. W. 325 2,3-Bis- (2, 4-dichlorophenyl)-1, 4-dihydroquinoxaline(?). A solution of 0.108 g. of o-phenylenediamine was dissolved in 2ml. of glacial acetic acid, and this solution added to a hot solution of 0.548 g. of 2,4,2',4!-tetrachlorobenzoin(?). After standing overnight, the precipitated dinydroxyquinox- aline(?) was centrifuged off and recrystallized from ethanol to give large colorless hexagonal prisms, meDe 144-145°, With concentrated sulfuric acid, the compound gives a. blood- red color, purportedly (13) a test for quinoxalines. It is not out off the question that the comoound is 2,5-bis- (2,4-dichlorophenyl)-quinoxaline, formed on condensing 2,4,2',4'-tetrachlorobenzil(?) with orthophenylenediamine. 2012244 C, 56.90; H, 2.87; N, 6.64. Anal, Caled. for Dihydroquinoxaline, C..H,N.Cl Caled. for Quinoxaline, CophyoNell 4: C, 57.17; H, 2.40; N, 6.67. Found: 6, 57.22; H, 2.67; N, 6.51. 116 C. The main product from the Hammick reaction employ- ing cinchoninaldehyde. , in this Hammick reaction, the ratio of main product to ketone obtained (no carbinol was obtained) was 20:1 (3). The elementary (C,H,N) analysis of the main product at first indicated it to be 4,4'-quinoloin®, and it formed a hydrochloride which analyzed for 4,4%=-quinoloin di- hydrochloride dihydrate. However, when its dilute acidic solution was made just basic with dilute sodium hydroxide solution, a solid precipitated which, by its analysis and melting point proved to be the reported (15) 1,2-bis- (4-quinolyl)-ethylene. Further investigation of the main product has not so far been attempted. L Experiments performed by Dr. ¢. McCloskey and Dr. E. Buchman et al. (3). “Cinchoninaldehyde undergoes the benzoin condensation with the (unusual) formation of 4,4'-hydroquinealgoin (14). 117 D. The Hammick reaction employing decarboxylating acids other than «-imino acids. The theory of Ashworth, Daffern and Hammick, by which their reaction is restricted tow -imino acids, is based on their conclusions that "... the mere presence of a basic centre in the molecule was insufficient; the de-~ carboxylation of pyridine and quinoline acids other than those containing the carboxyl group x to the nitrogen atom yielded no products of interaction with aldehydes and ketones. Furthermore the nitrogen atom must be ter- tiary..." (2). While giving no experimental evidence, those workers list in a footnote the acids which did not undergo the reaction. It includes, amongst a total of twenty-six acids, 2-methylquinoline-4-carboxylic acid and cinchophen. To test our more general theory of decarbox- ylation - carbonyl addition, we decarboxylated isonico- tinic acid in the presence of excess benzophenone. The acid has its decarboxylation temperature (cf. vart II, this thesis) at ca. 300°, and boiling benzophenone (306°) should therefore be applicable as solvent - carbonyl com- ponent. We obtained from the reaction a 3.5% yield of di- phenyl-4-pyridylcarbinol, which analyzed correctly and had the reported melting point (16). In addition we decarboxylated 2-pyrrolecarboxylic acid (decarboxylation temperature ca. 130°) in benzaldehyde, but only a trace (ca. 5 mg.) of a crystalline product was obtained, along with a large quantity of red amorphous product, which was not investigated further. The crystals dia not analyse for phenyl-2-pyrrylcarbinol or dipyrryl- phenylmethane, and were not investigated further. Exoerimental Part? 'Diphenyl-4-pyridylearbinol. A mixture of 19.8 ge (0.16 mole) of recrystallized iso- nicotinic acid and 250 g. (1.37 mole) of benzophenone was refluxed for 9$ hours, an air condenser being employed. A sealed-in plunger arrangement helped to scrape sublimed crystals of isonicotinic acid back into the reaction mix- ture. After 5 hours, ca. 3600 ml. of carbon dioxide had been evolved (90% of the theory). In the subsequent 4} hours, only 50 ml. more of carbon dioxide was evolved. The dark brown solution was allowed to cool to room temperature and then diluted with 500 ml. of benzene. This solution was extracted with 2.5 N hydrochloric acid, since mére concentrated acid precipitates a tar, which redissolves on dilution with water. The acid extracts were basified and extracted three times with 25 ml. of benzene. About lg. of a brown solid (A) separated at the interface. The solid was removed by centrifugation and the benzene Layer evaporated to dryness. he residue, on cooling, yielded a suspension of crystals in oil, which Was aQilutea with acetone and centrifuged. The centrifugate consisted of crystals (B). The motherliquors, on standing overnignt at lvelting points corrected; microanalyses by Dr. Oppenheimer and staff. 119 5°, did not yield any further crystals and were not worked up further. / Grystals (B) were washed free of motherliquor with acetone, in which they are insoluble. There was obtained - 0.85 ge of tan microscopic prisms, m.p. 192-252°, Four re- ~erystallizations from benzene-pyridine raised its melting ‘point to 237-238°. The pure material consists of. color-— Less microscooic prisms. Anal. Caled. for Cy) gH, .50N (261.5): C, 62.753 H, 5.79; N, 5.56. Found: OC, 82.94; H, 5.79; N, 5.09. The compound is reported to melt at 235° (16). Solid (A) was dried and recrystallized from benzene- pyridine to yield 0.6 gs. of the crude carbinol, m.p. 205°-232°, Thus the total yield of the reaction was 1.45 g. (3.8%). A mixture of 1.3 g. (0.011 mole) of pure c-pyrrolecarb- oxylic acid and 10.0 g. (0.086 mole) of benzaldehyde was heated for three hours at 150+5°. At the end of this time, 290 ml. of carbon dioxide had been evolved (81% of the theory). The pyrrole and 5.2 ml. of benzaldehyde were dis- tilled off, and the remaining red, gummy substance allowed to stand under ethanol overnight. By then the:.gum had soli- dified to a red, amorphous solid, soluble in ethanol. The solution was norited until it was translucent, the ethanol 120 stripped off and the residue heated with benzene. From the resulting solution there precipitated out, on cool- ing, ca. 5 mg. of a crystalline substance, which was @X- tremely soluble in ethanol and insoluble in benzene. Re-- crystallization of the solid from benzene-ethanol gave - colorless platelets, m.p. 192-193°, with a transformation ‘point at 172-173°. “Anal. Caled. for C,1Hy4No0, (7)! C, 56.15; H, 4.72; N, 11.91. Found : C, 55.02; H, 4.76; N, 12.11. el SUMMARY A modified theory of the Hammick reaction has been proposed, and partly substantiated by the successful reaction of isonicotinic acid with benzophenone. Some byproducts obtained in the Hammick reaction have been investigated. The structure of one such by- product has been elucidated, and that of two other byproducts partly, though not conclusively, clarified. (1) (2) (3) (4) (5) (7) (8) {10) {11) (12) L22 References Dyson and Hammick, J. Chem. Soc. 1937, 1724 Ashworth, Daffern and Hammick, J. Chem. Soc. 1959, 809 Unpublished work by Dr. Buchman et al.; Final Report, contract OEM om p7 519. sakins, ana Cor, J. Am. Chem. Soc. 60, 1151(1938) Lapworth et al., J. Chem. Soc. 1928, 2555; 1931, 1976 aharasch, xeinmuth and kayo, J. Chem, Ed. ll, SB (1934). This seems to be the only thorough @ discussion of aecarboxylation and decarbonylation available. Lapworth, J. them. Soc. 19035, 1004 fhorner and Zincke, Ber. 10, 1473(1877) Valeur, Bull. Soc. chim. $1, 1219(1879) Clamician and Silber, Ber. 56, 1576, 1953(1903) Criegee, Ann. 481, 263(1950) Hickinbottom, "keact ions of Organic Compounds", pel6él, 6 Longmans, Green & Co., New York 1956. Heference (12), p.294 Phillips, J. Am. Chem. Soc. 68, 2568(1946) “aplan, J. Am. Chem. Soc. 63, 2654(1941) ischitschibabin and Benewolenskaja, Ber. 61, 547 (1928) | 123 PROPOSITIONS 1. The Hammick reaction (this Thetis) has hithertofore been applied to aldehydes and ketones only. I propose attempts to apply it to esters and nitriles, as a new alcohol and ketone synthesis respectively. . “ 26 (a) it has been shown (this Thesis) that the scope of the Haramick reaction can be extended somewhat. It would be of inter- est’ to generalize it to a greater extent by applying it to other negatively substituted acids which do not contain the pyridine ring. - So (b) A study of the temperature dependence of the rate con- stant of decarboxylation of various substituted picolinic acids would be of theoretical interest. ee 3. (a) The commonly accepted acid strengths of the pyridine carboxylic acids (Ostwald, Z.phys.Chem., 3, 385(1889)). can- not be explained in view of their decarboxylation behavior. I propose that_they actually refer to the acid dissociation constant PyHCOO+PycOO+Hi ae (b) The conclusions drawn by Erlenmeyer et al. (Helv.20, 510(1957)) concerning the effect of the ring on the acid strengths of various heterocyclic ring acids are invalida- ted by the nature of his comparison. - (c) It would be of interest to determine the crystal structures of the three pyridinecarboxylic acids and of. pyrazinoic acid. 4. (a) The conclusions drawn by Ciamician (Z.phys.Chem. 18, - 51(1895)) concerning the isomorphism of some heterocyclic acids with benzoic acid are in conflict with our results (this Thesis). I propose that complete phase diagrams are neces- sary in order to draw such conclusions. (bo) The argument offered by Erlenmeyer and Berger (Bio- chem.eZ. 252, 22(1952)), that a basis for the crossreactivity of diphenyl ether and diphenyl methame may be found in the fact that the two compounds form solid solutions, is in- velid, since evidently these two compounds form a simple eutectic (iuttringhaus, Ann. 528, 225(1937)). 5. (a) The synthesis of 2,4,5-trichloro-3-nitrothiophene, med. 86°(Rosenberg, Ber.19, 651(1886)) and of 3,4,5-tri- chloro-2-nitrothiophene, m.d. 73-76° (Steinkopf et al., Ann. 512, 151(1954)) could not be repeated (Steinkopf et al., Ann. 552, 250(1957)), a product m.p. 70° being obtained in both cases, and neither product depressing the melting point of the other. I propose that this material be submitted to chromatographic analysis, since it may well be the minimum of a series of solid solutions. (b) Many isomeric halo-derivatives of thiophene are iso- 124 morphous and have the same melting point (Steinkopf, "Die Chemie des Thiophens", p. 15ff.). I propose that the lumi- nescence analysis presently employed to arrive at the iden- tity of such compounds be replaced by the criterion of chromatographic behavior. (c) Much difficulty is encountered in finding a substance the derivatives of which lend themselves to a successful © study of morphotropy (this Thesis). I propose the homolo- gous series of 2,35,5-trichloro-4-alkylthiophenes as a pro- mising subject for such a study.’ This, supplemented by an ‘appropriate immunological study, would lead to interesting semi-quantitative results. 6. (a) An electron diffraction study onl, 4-dithiene would be of considerable interest; however the preparation of this compound has met with diff feulties so far (Meyer, "Die Thio- phengruppe", p. 274). The following alternate syntheses are proposed: , Sy \\ ,£-SESP§ CLCH CHO » HSCHSCHO —> " N“ s “ CH2CH (OEt) 3 | a « SUCPS — § (CHgCH(OEt)2)2 —» S(CHgCHO). (bo) Only 2,5-derivatives of 1,4-dithiene are known. A synthesis for 2,6-derivatives is proposed (cf. Baker et aly J.Am.Chem.Soc. 58, 262(1936)). 7. (a) The extreme sensitivity of Pelletierine to oxidation, heat and pH have rendered its synthesis impossible tw date (Spielmann et al., J.Org.Chem. 6, 780(1941)). I propose the ' following synthesis: C Jeon, ,CH(OEt)> 2 Steps ( dexgoze — Pelletierine CHoPh (b) It has been proposed (Spielmann, lo. cite; Beets, Rec. trav. 62, 553(1943)) that Pelletierine exists largely in the bicyclic carbinolamine configuration. This claim could be readily verified by a measurement of its absorption spectrum. 8. I propose a study of the photolysis of W-diazoketones in the presence of water, in order to elucidate somewhat the function of the catalyst in the Wolff rearrangement, and in order to find conditions for a variation of the arndt-EHistert synthesis, in which the rearrangement may take place in the absence of thermal excitation and of catalysts. 125 9. I propose that the measurement of the heat of formation of fluorosilane would constitute a contribution towards our concept of single bond energies (Pauling, "The Nature of the Chemical Bond"). 10. The experimental results of Herz (Z. anorg. allgem. Chem. i187, 379(1950)) concerning the ratio of the temperature coef- ficient of fluidity to the coefficient of thermal expansion of a liquid can be explained on the basis of Batschinski's theory of viscosity (Z.phys.Chem. 84, 643(1913)). ll. Lest visiting lecturers garner a mistaken impression of the soundness of mind of Gates-and-Crellinites, I pro- pose they be forewarned of the idiosyncrasies of The Sisters.