I. The Synthesis of Spiro(3,5)Nonane. II. The Presumed Synthesis of 1,3-Cyclobutanedicarboxylic Acid by Markownikoff and Krestownikoff. III. The Synthesis of Cis and Trans-1,3-Cyclobutanedicarboxylic Acids Citation Deutsch, Daniel Harold (1951) I. The Synthesis of Spiro(3,5)Nonane. II. The Presumed Synthesis of 1,3-Cyclobutanedicarboxylic Acid by Markownikoff and Krestownikoff. III. The Synthesis of Cis and Trans-1,3-Cyclobutanedicarboxylic Acids. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/K23E-7793. https://resolver.caltech.edu/CaltechTHESIS:10122017-111506286 Abstract Spiro(3,5)nonane was prepared by methods indicative of structure. The historically The historically important acid reported in 1881 by Markownikoff and Krestownikoff and assigned the structure 1,3-cyclobutanedicarboxylic acid (a structure which has never subsequently been questioned) was shown to be 1-methyl- 1,2-cyclopropanedicarboxylic acid. The mechanism of the for­mation of the Markownikoff-Krestownikoff acid is discussed. The previously unknown cis- and trans-1,3-cyclobutane­- dicarboxylic acids were synthesized. After numerous unsuc­cessful attempts the synthesis was achieved starting from pentaerythritol. The reaction between the benzylidene der­ivative of 2,2-bis(bromomethyl)-1,3-propandiol and sodium malonic ester gave a condensation product which after hydrolysis and oxidation with nitric acid yielded 1,1,3,3-cyclobutanetetracarboxylic acid. On decarboxylation of the latter, the desired acids were obtained. Item Type: Thesis (Dissertation (Ph.D.)) Subject Keywords: Chemistry and Mathematics Degree Grantor: California Institute of Technology Division: Chemistry and Chemical Engineering Major Option: Chemistry Minor Option: Mathematics Thesis Availability: Public (worldwide access) Research Advisor(s): Buchman, Edwin Raphael Thesis Committee: Unknown, Unknown Defense Date: 1 January 1951 Funders: Funding Agency Grant Number United States Government UNSPECIFIED State of California UNSPECIFIED Record Number: CaltechTHESIS:10122017-111506286 Persistent URL: https://resolver.caltech.edu/CaltechTHESIS:10122017-111506286 DOI: 10.7907/K23E-7793 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 10510 Collection: CaltechTHESIS Deposited By: Benjamin Perez Deposited On: 12 Oct 2017 19:40 Last Modified: 27 Apr 2023 23:13 Thesis Files Preview PDF - Final Version See Usage Policy. 19MB Repository Staff Only: item control page

PART I THE SYNTHESIS OF SPIRO(J,5)NONANE PART II TEE PRESUMED SYNTHESIS OF l,J-CYCLOBUTANE- DICARBOXYLIC ACID BY MARKOWNIKOFF AND KRESTOWNIKOF'F PART III THE SYNTHESIS OF CIS AND TRANS-1,,J-CYCLO- BUTANEDICARBOXYLIC ACIDS Thesis by Daniel Harold Deutsch In Partial Fulfillment of the Requirements For the Degree of Doctor of Philosophy California Institute of Technology Pasadena, California 19.51 ACKN OWLEDG ENiENT The author wishes to express his appreciation and indebtedness to the United States Government and the State of California for their financial assistance which made this work possiblej to the Chemistry Department at t he Institute wlwse courses it was a privilege and a pleasure to attend; and particularly to Dr. Edwin R. Buchman, whose patience, knowledge of organic chemistry, and constructive critic ism were a constant inspiration to accomplish finer work. He also wishes to acknowledge his indebtedness to Drs. James Conley, William Finnegan, and Hershel Herzog for their criticism and suggestions regarding much of the work in this thesis. We are indebted to Dr. Adalbert Elek lLos Angeles, Calif.) for the micro-analyses (Philadelphia, Pa.) spiro(J,5)nonane. and to Samuel P. Sadtler & Son, Inc. for the infrared absorption data on ABSTRACT Sp iro(J ,5)nonane was prepared by methods indicative of structure. The historicall y important acid reported in 1881 by Markownikoff and Kres tovmikoff and assigned the structure l,J-cyclobutanedicarboxylic acid ( a structure which has never subsequently been questioned) was shown to be l-methyl1,2-cyclopropanedicarboxylic acid. The mechanism of the formation of the Markownikoff-Krestownikof'f acid is discussed. The previously unknown cis- and trans-1,J-cyclobutanedicarboxylic acids were synthesized. After numerous unsuccessful attempts the synthe~s was ac~ieved starting from pentaerythritol. The react ion between the benzylidene derivative of ~,2 -bis\bromomethyl)- l,J-propandiol and sodiwn malonic ester gave a condensation product hydrolysis which af ter and oxidation with nitric acid yielded 1,1,J,3- cyclobutan etetracarboxylic acid. On decarboxylation of the latter, the desired acid s we re obtained. 1 THE SYNTHESIS OF SPIR0(3,5)NONANE (1) The synthesis of spirane hydrocarbons containing the cyclobutane ring is a part of the general investigation of cyclobutane chemistry now in progress in this laboratory. Spiro(J,J)heptane (I), the first cyclobutane Spiro-hydrocarbon to be described, was recently synthesized in tbis laboratory (2). In continuing this work spiro(J,5 )nonane (II) was synthesized by two different methods. I II Relatively few p reparative routes exist for the practical synthesis of spiranes containing the cyclobutane ring . Fecht's (3) early synthesis of 2, 6-spiro(3,J)heptanedicarboxylic acid (III), an extension of the Perkin (4) synthesis, illustrates one mode of spirane for ma tion; C(CH2Br)4 +2CH 2 (C02CsH11) 2 + 4Na0 CsH11 > Hydrolysis 2 C0 2 H III A variety of l,J-spiro-cyclobuta.nedicarboxylic acids hare been prepared through a Guareschi - type .i mide (5) . Paul (6) synthesized l,J-spiro(J,5)nonanedicarboxylic acid (IV) by the following reactions; NH3 NaOEt =0 + ) 2H 2 C(CN)C0 2 Et ~ NC CH 2 I 2 ' Ito Hydrolysis > ) NaOEt ~o ( C0 2 H )z A 0 > ( 0 0 2 H)z IV Kon (7) prepared spiro(J,5)nonan-J-one, in very poor yield, by the reaction . Na or K Xylele v 3 The synthesis of ethyl l,2-spiro(J,S)nonan-3-onedicarboxylate (VI), in very small yield, was reported by Chatterjee (8) ; C0 2 Et I Na C0 2 Et CH - I Xylene CH 2 I "0 C0 2 Et C0 2 Et VI Poly-spiro cyclobutanes have been prepared in this laboratory (9) by ketene dimerization (10} followed by Wolff-Kishner or mercaptol reduction; Wolff-Kishner or Mercaptol Red. ( l-15 C2 ):;N ---> 2Et0 2 C ) =C=O 4 0 II II 0 Wolf'f-Kishner ) Et0 2 C C0 2 Et Fujimoto (1), in this laboratory, prepared spiro~3,5)nonane (II), in low yield, by the following r eac- tions; COCl 'Nolff-Kishner 0 II The synthesis oi' spiroU,5 )nonane (II), effected by the writer, is outlined in Table I . Bis(hydroxymethyl)cyclohexane (IX) was prepared by the method of Boord and co-workers (11). Conversion of (IX) to bis {bromomethyl)cyclohexane (XI) through the tosylat.e (X) gave yields of l.I\ 2 ~CH 2 HC' Ht ~ CH KOH ' 7 AgN0 3 ' 7 + 2 CH II ---7 CH 0t osyl Cti2 0to s yl 2 CH-CHO x XIII J:VI VII NaBr ' 7 CHO CH 0 2 '7 XI Na OH H2S04 ' 7 Brz ~ Table 1 VIII CH 2 0H Hz Nt ' 2H2 Cl C0 2Cs H 11 L>- Br C0 2 H Ni 7 H2 , 7 C0 2 H 6 , 7 2 NaOC 5 H 11 XIV CH 2 Er CH Br The Synthesis of Spiro(J,5)nonane C0 2 K C0 2 K C02Ag J:VII IX CH 2 0H 'CH2 OH XII x:v II Tosyl Cl ) C0 2 H 6 50-70jb, compared to the 27% yield reported by Boord \11) using phosphorus tribromide. All the subsequent steps were carried out in good yield. Isoamyl 2,2-spiro(3,5)nonanedicarboxylate (XII) , resulting from the malonic ester synthesis carried out in isoamyl alcohol (cf.3), was s aponi fi ed a nd deca r boxylated in the usual manner. The 2-spiro(3,5 )nonanecarboxylic acid (XV) was degraded to 2-bromospiro(3,5)nonane (XVII) by the Hunsdiecker reaction \ 12). Spiro ~ 3 ,5 )nonane ( II) was subsequently prepared by catalytic hydrogenation of the bromide (XVII) in sodium methoxide-methanol (13). The physical constants of the spiro(3,S)nonane lII) prepared by t hese two different routes agreed as did the analysis. The i dentity of' these two samples was conclusively demonstrated by comparing their infra-red a bsorption sp ectra (Fig . 1). ·-·· 1 , _::=-- I 1 _, i 2.0 0 I I - j - -:.f-c-:1 .~"'l= ~= I . _J n "I: I I 1" ~ I I i--,-m -!4=; H-! ui '"i I 1 I I 11 1'° I :J:g~t:::1:: ~ 1 3.0 o.o e.o 7.0 o.o H. ·· - I I I , °'l" ' Fig. 1 Wave Length in Microns o.o 10.0 11 .0 ~-== --- --= - -- IOOO I 12.0 eoo 13.0 ' 14.0 ' "i'° 0.01 mm. Courtesy Samuel P. Sadtler & Son, Inc. lPhiladelphia, Pa. ). Infrared absorpti on spectrum of spiro(J,5)nonane. Cell thickness, 4.0 - -- I ' E-o= -=- 1~F=-o ~- 0=+-+ J_ 0-0- n I r-+ 1-= 1-! , ~· F:l=-:- - e=:-i= - _1= -- -- I~'-"-' i::r~1~~t= ' cl ._ . -:=="-:'-:~t- · ,- r-'ft:-1::-- :i~=.i ~ ~- · -_-- .. : __ .:i_,~-.LL! •l Cl :t ' ~ = --'= "" '·"""-°'>=~~=· I Wave Numbers in cm - ' Lfl _j_l.= h-U:8 - I- ~-. :~ -~ !: Y :1 I I , ___-,4 -·:-_-,_,' 1·33_ 'ri tt:-·: I --L: -· , , ~ I 12 1 i ~· - E'- ~~ _=1-~ :-- ~l-.:fV_-· - ; f.T -. -=-'f~fi-*l~~tJ~i;:- ~k*1:_-"'=g ~ -= I 2000 I , Ji! ' I .tL ;--. - 7= _ =l ·I cl l _cc __ 'i': I I ".. .n. : . . I --- - - I 10 , 4r, ~ 00 •ooo 25 1 - 15.0 == ---=-_:;-..:::;. I 16.0 = _,..-, °'P' 7 EXP ER I MEN TAL 3-Cyclohexene-1-carboxaldehyde (VII) The procedure used was patterned after that described by Chayanov (14) and Boord and co-workers (11). Five moles each of acrolein (Eastman Kodak Co.) and 1,3-butadiene were placed in a one-liter bomb (American Instrument Co. ) previously cooled overnight in a dry ice chest. After the addition of 5 g. of hydroquinone, the bomb was sealed and heated as rapidly as possible to 115 ° . The temperature rose slowly to 130°, then jumped suddenly to 230°. After the bomb was cooled, t he contents were distilled through an 18-inch column. The product (VII), b.p. 87° at 75 mm., was obtained in y.ields of 65-75'j~. 4.4-Bis(hydroxymethyl)-l-cyclohexene (VIII) Following the method of Whitmore and co-workers (15), 520 g. (4.73 moles) of 3-cyclohexene-1-carboxaldehyde (VII) was mixed with one liter of 4 07~ formaldehyde and sufficient a lcohol to form a homogeneous solution. Over a one-hour period this solution was run into 600 g. of sodium hydroxide in 2.5 1. of '.1570 alcohol in a 5-1. threenec .ked flask provided wi th a mechanic a l stirrer, reflux condenser, and dropping funnel. After refluxing for 20 hours, the alcohol wa s stripped off on a steam bath, and the residual black solution was continuously extracted with ether for four days. The ether was stripped off from the extract and the product (VIII) was distilled (b.p. 130° at 3-4 mm. ) after the addition of two grams of' Raney nic ke l . The pure white diol (VIII) contained traces of 8 oil (m.p. 5;; 0 ) . After one crystallizat i on from me thyl aleohol it melted at ;;0.5° ~literature values (11) 92.0°i ll6) 92.5°). The yield of (VIII) was 67~. 1,1-Bis(hydroxymethyl)-cyclohexane (IX) 4,4-Bis(hydroxymethyl)-l-cyclohexene was dissolved Iim methyl alcohol (40 g. per 100 ml.) and placed in a steel bomb along with 10~ Raney nickel (17). Hydrogenation was carried out at 1000 p.s.i. a nd 70-80°. When the hydrogenati on was completed, the solution was filtered free of Raney nickel; the methyl alcohol was distilled off; and the diol (IX) was dried in vacuo over sodium hydroxide; m.p. 97-98° (literature value (11) 98.5°), yie ld, 88~. 1,1-Bis(tosyloxymethyl)-cyclohexane (X) Three hundred sixty-eight grams (2.55 moles) of 1,1-bis(hydroxymethyl)-cyclohexane (IX) was dissolved , , in 2100 ml. of pyridine (Barrett Co., redistilled over barium oxide) and then cooled for 30 minutes in an ice bath. To this solution 1075 g. (6.1 moles) of tosyl chloride lEastman Kodak Co .. practical, reerystallized from petroleum ether) wa s then added over a twentyfive-minute period while the solution was kept cold in an ice bath and swirled occasionally. When the reaction subsided, t he mixture was allowed to stand overnight at room temperature. The product was then poured into di lute hydrochloric acid. A light yellow liquid sepa r a ted and quickly solidified. The tosylate (X) was filt e red, washed with dilute hydrochloric acid, and crystallized twice from methyl alcohol; large prisms, m.p. 74.0-74.5°, yield,778 g. 9 (67 ~ ). On smaller scale (using Eastman Kodak Co. white l abel tosyl chloride) nearly qua ntitative yields were obtained. Anal. Galea. for C2 2 H2 e06S2; C, .58.38; H, o.24. Found; C, 58. 54; H, 6.39. 1,1-Bis( bromomethyl)-cyclohexane (XI) A mixture of 778 g. ll.72 moles) of tosylate (X), ' 4 26 g. ( 4 .13 moles, 2 O;c, excess) of sodium bromide ( reagent), and 12 00 ml. of di ethylene glycol \Eastman Kodak Co., yellow label) was heated for three hours at 150170° in a 5-1. flask, provided with a very efficient Hershberg stirrer. Considerable frothing took place. After the mixture was cooled to room temperature 2.5 was 1. of water/added. The lower organic layer was separated and the aqueous phase was extracted four times wi th carbon tetrachloride . The combined carbon tetrachloride extracts and dibromide were washed with water and dried over sodium sulfate. Af ter the solvent was stripped off under a water pump, the residue was distilled through an 18-inch Vigreux column. The dibromide (XI), b.p. 110° at 5-6 mm. ,nl) 5 1. 5 355 2S \literature values (11), b.p. 117°at o ~Jn., nu'l.5310), was obtained in 7'J;o yield ( 366 g. ) • The yield of di bromide (XI) based on dial wa s 53/a• When Eastman Kodak Co. whit e label tosyl chloride was used on a smaller scale, the yield of dibromide wa s 65 -70~ based upon dial. Isoamyl 2,,_2-spiro( 3 . 5 )nona4~di.carboxylate \XIIJ. A solution of 85. 0 g. ( 3. 7 moles, 25)~ excess) of sodium in t hree liters of isoamyl alcohol \previously 10 dried by distilling from magnesium turnings) was prepared in a 5-1. flask. To this, 594 g. (3.7 moles) of ethyl rnalonate (Eli Lilly Co., dried by ciistillation) was added and the mixture was refluxed for one hour, during which time ethyl alcohol was distilled off. The flask was cooled and 404 g. ( 1. 48 moles) of the d i bromide ~XI) and a hand- ful of porous chips were added. The mixture was refluxed gently for 44 hours. After the flask was cooled, ice water was poured in, and the solution was neutralized with hydrochloric acid. The organic phase was separated, and the aqueous phase was extracted with ether. The combined ether extract and organic phase were dried over sodium sulfate, and the bulk of the ether and isoamyl alcohol was distilled off under reduced pressure from a steam bath. Fractionation of the residue, nearly 2 1., gave 241 g. of' material, b.p. 165° at 4mm. to 180° at 1-2 mm., n[i 5 1.45>'0. Redistillation gave 151.3 g. of l30io ) 22 ester (XII), b.p. 1.:15° at 4-5 rrUT1., n}) 5 1.45o4, d4 O.:;t611. Anal. Calcd. for C21 H32 04 ; C, 71.55; H, 10.2:;. Found; C, 71.65; H, l0.33. 2 J..2-Spiro ( 3 ,5 )nonanedj.:.carboxy:J.:.i.c Acid (XIV) To 133.5 g. (0.377 mole . ) of the ester (XII) in a + 1-1. flask , 84 g. (1.5 moles) of potassium hydroxide dissolved in 500 ml. of absolute alcohol (filtered to clarify) +The hydrolysis coul d be more conveniently carried out in a beaker. 11 was add ed all at once with swirling. A heavy white precipitate formed after a few minutes. The mixture was heated on a steam cone for twenty minutes and cooled to room temperature. The resultant cake was broken up and transferred to a sintered glass funnel. After as much of the liquid as possible was pressed out, the precipitate was air dried overnight in an open beaker. The potassium salt (XIII) had a slight yellow color and smelled strongly of isoamyl alcohol. The product was then tri tura t ed with absolute alcohol, fil t ered, and dried on a steam cone to constant weight, 1 02 g. (white and odorless). This salt (XIII) was dissolved in 200 ml. of cold water and 38 g. (0.4 mole . ) of sulfuric acid, diluted with an equal volume of water, was added slowly with stirring. 2,2-Spiro(J,5)nonanedicarboxylic acid (XIV) precipitated as a white solid. After being filtered and washed with cold water, the product was dried overnight in vacuo over sodium hydroxide; crude yield, bj.6 g. ( theory 88.5 g. ). A small portion of the acid was crystallized twice from water; m.p. 2oou (dee.). Anal. Calcd. for C11 H16 0 4 ; C, 62.29; H, 7.oO. Found: C, 6 2. 35; H, 7. 76. 2-Spiro(3,5)nonanecarboxylic Acid (XVl For decarboxylation, 97.5 g. of the crude acid (XIV) was placed in a 500-ml. Claisen flask and heated to 220°. When the evolution of carbon dioxide ceased (ten minutes), the bath temperature was lowered and the 2-spiro(J,S)nonane24 carboxylic acid was disti ll ed; b.p.163° at 15 mm., d4 1.0363, 12 25 nD 1.47 ~ 3, m.p. 21°; yield, 45 . 6 g. 6 7/o based upon the ester (XII)). Anal. Calcd . for C1 0 H1602; C, 71039; H, 9.59 . Found ; C, 71. 4 2 ; H, 9 • 5 3. The amide of 2-spiro(3,5)nonanecarboxylic acid (XV) was prepared in the usual manner f'ro:rr. thionyl chloride, the acid, and ammonium hydroxideo After tl1ree crystallizations from water-methanol the amide melted at 156.0-156.5°. Anal. Calcd. for C10 H1 70N; C, 71.81; H, 10.25. Found: c, '71.87; H, 10.27. Silver 2-spiro\3.5)nonanecarboxylate (XVI) A solution of 41. 6 g. ( O. 248 mole ) of' the spirane acid in JOO mL of water containing two drops of phenolphthalein was prepared. To this, a 15/'o aqueous potassium hydroxide solution was added until the solution just turned red. The solution was then made slightly acid with one drop of the spirane acid (XV); whereupon, the color faded completely. After this potassium salt solution was heated to 50°, it was slowly added to a mechanically stirred solution o.f 42.2 g. (0. 248 mole ) of s ilver nitrate in 168 ml. of wa ter heated to 40°. The silver salt (XVI) came down as a fine white precipitate. After t he mixture was cooled, the silver salt was filtered and washed with 150 ml. of water in which i .n it was quite insoluble. (XVI ) was suspended/and filtered from acetone, met hyl alcohol and absolute ether in that order. The product was air- dried overnight, powered in a mortar, and transferred to a 300-ml. round- bottom flask. The flask was 13 evac"l].a ted with an oil pump; and after the gas evolution ceased, the flask wa s heated on a steam cone in vacuo for 70 minutes. After 15 minutes the visible gas evolution ceased. The yield of silver salt (XVI) was 65.0 g. (nearly quantitative). 2-Bromospiro(3.5)nonane (XVII) Forty grams of bromine (reagent), previously dried over sulfuric acid, was dissolved in 200 ml. of carbon tet.rac hloride, previously dried over calcium sulfate. The bromine solution was placed in a dry 5 00-ml. t hree-nec ked 1·1ask provided with a reflux condenser (protected with a calcium chloride tube), a mercury-sealed Hershberg stirrer, and a short length of wide rubber tubing for the addition of the silver salt. After the flask was cooled in an ice-bath, 65.0 g. of the silver salt (XVI) was slowly added over a 45-minute period. A vigorous evolution of carbon dioxide ensued. The ice-bath was removed and the mixture was heated to reflux for 15 minutes. After t Le product was cooled, the solid silver bromide was filtered off and washed with carbon tetrachloride until the filtrate came through colorless (approximately 200 ml.). The combined filtrates (deep red) were washed with aqueous sodj_um bisulfite (color faded), and dried over sodium sulfate. After the solvent was stripped off, the 2-bromospiro(J,5)nonane (XVII) was distilled, b.p. 103-104° at 1325 2 l 14 mm., nD 1.5015, d 4 1.2544; yield, 57%.• Anal. Calcd. for C~H 15 Br; C, 53.22; H, 7.44. Found: C, 53.58; H, 7.6':). 14 Spiro(3,5)nonane (II) To an ice cold solution of 4. 6 g. ( O. 2 mole · ) of sodium disso lved in 150 ml . of absolute methyl alcohol, there was added 28.3 g. (0.13)! mole . ) of 2-bromospiro(3,5)nonane \X"VII). Fifteen grams of Raney nickel (17) was added and hydrogen was passed in at atmospheric pressure and room temperature. The hydrogen uptake was :i2/o of theory over a six-hour period. Some water was inadvertently allowed to enter the flask during the hydrogenation, but apparently it had n o effect other than slowing down the rate of hydrogenation. Additional p ortions of 1-2 g. of catalyst were added from time to time during the hydrogenation. The resulting solution was filtered an asbestos mat, and the residue was washed three times ~hrough with ether. After tlrn filtrate was washed three times with water and once with 6l'J hydrochloric acid, the ether-hydrocarbon solution was dried over potassiw11 carbonate and then over sodium for 18 hours. The resulting solution was filtered and the ether was stripped off under reduced pressure. Distillation of the residue at atmospheric pressure (750 mm.) gave ll.83 g. of product b.p. 150-157°, which did not decolorize an aqueous or acetone solution of potassium per- mangenate. Redistillation of this material yielded 7.70 g. 23 of pure spiro(3,5)nonane (II), b.p. 15 7. 5- 15 7. 7°, d 4 0.8512, 25 nD 1. 45 80. Fow~d; C, 86. ~4; H, 13.ll. 15 REFERENCES (1) Pres ented at the ll) th Meeting of the American Chemical Society, San Francisco, California, March, 1.949, E. R~ Buchman, Do H. Deutsch, and G. Fujimoto. (2) E. R. Buchman, J. C. Conley and G. Fujimoto, unpublished. (3) H. Fecht, Ber., 40, 3883 (1907). Also see H. J. Backer and H. B. J. Shurink, Rec. trav. chim., 50, )26 (1}31); S. E. Janson and W. J. Pope, Chem. and Ind ., 51, 316 ( 1.:13 2 ~; Proc. Roy o Soc. , Al5 4., 5 3 ( 17 36 ) • \4 ) w. H. Perkin, jun.' Ber., 16' 17 '.:13 ll883). (5 ) J. Guar eschi, Atti. accad. sci. Torino, Jl±., )' 28 (ld:79). (6 ) P. K. Paul, J. I ndian Chem. Soc., 8, 717 \ 1.:131). (7) G. A. R. Kon, J. Chem. Soc., 1 21, 515 ,.5 2 0 ( 1.9 2 2 ) • \ 8) l\J. (9) Chem. Abstr., J..Q, 5747 ll.:136). H.M. Walborsky, E. R. Buchman, /W. Fi1lllegan and H. Herzog, unpublished. N. Chatterjee, J. Indian Chem. Soc., 14, 127 \1:137); (10) W. E. Hanford and J. C. Sauer, The Preparation of Ketenes and Ketene Dimers, Organic Reactions, Volume III, John Wiley and Sons, Inc., New York City . (11) W. R. Shortridge, R. A. Craig, K. W. Greenlee, J. M. Derfer and C. E. Boord, J. Amer. Chem. Soc., 70, :746 ( 194 ~). (12) J. Cason and R. L. Way, J. Org. Chem., 14, 31 (1749); J . Kleinberg, Chem. Rev., 40, J81 (l;iL17 ). (13) Cf. 1. Ruzicka, P. A. Plattner and H. Wi ld, Helv. Chim. Acta.,.£.§., 395 (1745). (14) N. A~ Chayanov, J. Gen. Chem. (U.S.S.R. ), ~, 460 (1;;38). 16 ll5) Fo Co Whi tmore, A. H. Popkin, H. I. Bernstein and J. P. Wi lkins, J. Amer. Chem. Soc.,~' 124 (1941). ll6) H. E. French and D. M. Gallagher, J. Amer. Chem. Soc., 64, 14i7 ~1942). (17) A. A. Povlic and H. Adkins, J. Amer. Chem. Soc., .§._§_, 14'71 ( l;i 46 ) • 17 THE PRESUMED SYNTHESIS OF l,J-CYCLOBUTANEDICARBOXYLIC ACID BY MARKOWNIKOFF AND KRES TOWNIKOFF(l)+ cis- and trans-1,J- Oyclobutanedicarboxylic acids ( I) and (II) are of considerable interest in this laboratory as possible intermediates for the synthesis of cyclobutadiene (III). Several possible routes from (I) or (II) to (III) are given in Table I. Acids (I) and (II) are also potential sources of such interesting compounds as cyclobutan-1,J-dione (IV) , cyclobutan-1,J-diol (V), and l,J-dimethylenecyclobutane (VI). IV v An acid which was assigned the structure trans-1,J-cyclobutanedicarboxylic acid (II) was reported by NJarkownikoff and Krestownikoff (1) in 1881. This preparation has been accepted as the first correctly interpreted alicyclic synthesis (2). The Niarkownikoff-Krestownikoff synthesis was especially important to this laboratory since it had pre- viously been shown here (3) that other syntheses of 1,3cyclobutanedicarboxylic acid described in the literature did not furnish this compound and thus the Iv'f..arkownikoffKrestownikoff synthesis appeared to be the only route to the desired acid~·::--~~~~~~~~~~~~~~~~~~~ ,... D. H. Deutsch and E. · R. Buch/nan, in press. .-l co H2N0 lHoCh NO 1rn2 CH20 ./ ... Schn1id t 7 or Curtius Ag OH 7 \.. HOzC Table 1 ~ ~ Nao. 9 Cl Cl Ag OH " BrO Br ( H3 C ), NO C02Na NaOH COCl Hunsdiecker Cl Cl 0 ClOC "'-... SOz Clz COzH or I I O I III 6 r I N (CH, lo OH ,/ 0 . l HO( H, c loll O O(H,C)z~ N(CH,J,0 ~02 N, CH,), 1 I N(CH3)3I Br N( CH3 )oBr (CH3)dN Proposed Syntheses of Cyclobutadiene from 1,3-Cyclobutanedicarboxylic Acid CH3I I( H3C),NO 19 As a consequence oi' the work presented in this section of the thesis, it was discovered that the acid prepared by Markownikoff and I\:res townikoff ( 1) was not trans-1, 3- cyclobutanedicarboxylic acid (II) but its isomer, trans-1-methyl+ 1,2-cyclopropanedicarboxylic acid (VII) • It is indeed sur- II VII pris.ing that (despite the his tori cal importance of this acid) this incorrect formulation had been accepted for nearly seyenty years. It may be of' interest to the reader to be given a resume of the literature pertaining to the MarkownikoffKrestownikoff acid. In 1881 Markownikoff and Krestovmikoff (1) reported that, by the action of alcohol free sodium ethoxide on ethyl ~ 0 chloropropionate, a small yield of ester, formulated as (VIII), was obtained which on hydrolysis gave an acid melting at 170°. On the basis of its saturated properties and elementary analysis they assigned this acid the structure (II). These authors postulated that (VIII) resulted from +Historical Note. As a result of these findings it follows that the first correctly interpreted synthesis of an alicyclic compound was Freund's (4) synthesis of' cyclopropane (1882), and that ethyl 1,1-cyclobutanedicarboxylate (Perkin, jun., (5) (1883)) was the first cyclobutane derivative to be synthesized. 20 the symmetrical dimerization of' ethyl acrylate formed f'rom ethyl ~ -chloropropionate. H3C-CHC1-C0 2 Et + NaOEt H2 C=CH-C0 2 Et > + Et0 2 CVIII At a later date, Markownikof'f (6) reported the conversion of his trans acid into a cis anhydride and thence into cis-1,2-cyclobutanedicarboxylic acid (IX), melting point 138-139°. Authentic (IX), having a melting point of' 138°, had been synthesized a few years earlier by Perkin , jun. (7). Haworth and Perkin jun. (8) in 1898 reinvestigated the ma tter and agreed that Mark~wnikoff's trans acid wa s represented by (II), but showed that his cis acid was not identical with (IX) since their anhydrides melted 25° apart. From this they concluded that no position isomerization had occurred and that Markownikoff' s cis acid was cis-1, 3-cyclobutanedicarboxylic acid (I) . The scientific world accepted the conclusions of Perkin, jun. and no furthe r communication relating to the Markownikof'f-Krestownikoff' acid appeared until 1:125. In connection with his studies on the relative ease of formation of small-ring compounds, Ingold (9) investiga ted the products of the reaction of ethyl ti..-bromo-y -methylglutarate + (X) with potassium hydroxide. Among these were found cis+The reactions of the iodo and chloro analogs of (X) were also investigated. 21 and trans-l-methyl-1,2-cyclopropanedicarboxylic acids (XI) and ~VI I) with melting points 142° and 168° respectively and a search was made for 1,3-cyclobutanedicarboxylic acids. x IX XI The following is a quotation from I ngold's article; 11 It remains to be added that no trace of 1,3-cyclobutanedicarboxylic acid could be detected, although its formation from the bromo-ester (IV) might be expected to take place according to the sche~ -C0 2 Et KOH ) 0 IV If the mec)lanism of' the formation of cyclobutane-1:3 dicarboxylic ester from o<..- chloropropionic ester by the action of sodium ethoxide ( Markovmikoff and Krestovmikoff·, Annalen 1 881,208 333) is, as might naturally be supposed, the following; H3 C-CHC1-C0 2 Et ~ Et0 2 C- CHC1-CH3 / CHC1-C0 2 Et NaOEt H2 C NaOEt ) ) ' CH-CH 3 Et0 2 C;- XXII -C0 2 Et 0 Et0 2 C- XXIII for the hypothetical intermediate product ~XXIIl is the chloro analogue of the bromo-ester JlY)_. Moreover , the chlorine compound (XX II) has been prepared synthetically, by the chlorination of oe. -methylglutaric acid, and all at - temp t s to convert it into the cyclobutane ester (XXIII) by 22 the agency of sodium ethoxide have hitherto failed. It has been thought right to mention these negative results here, as, although the mechanism of the MarkownikoffKrestownikoff reaction is at present ur1der investigation, the matter is of such complexity that it may well baffle enquiry for a considerable time. ~ In the same year that the above article appeared, Goss and Ingold (10) wrote a stimulating article, ~The Possible - II Enhanced Activity of Newly Formed Molecule s • These authors had noted that numerous cases have arisen where chemists have been unable to confirm apparently obvious reaction mechanisms by preparing supposed intermediate products and subjecting them to the conditions of the original experiment: the substances remain unaltered or behave differently from expectation. In this connection Goss and Ingold cited: '' Two mechanisms suggest themselves for · the formation of the cyclobutane ester l1Y.l ( Markovmikov and Krestovmikov, Annalen 1880, 208 334) from chloropropionic ester (I) and sodium ethoxide: the elimination of hydrogen chloriCT9-"is either inter-molecular, in which case o<.- chloro-'\- methyl glutaric ester l..I.U is the intermediate product, or intra.-molecular, ethyl acrylate (III) being first formed The chloro-ester .l.ill has been prepared and subjected to the action · of sodium ethoxide under the same conditions. It gave no detectable quantity of the cyclobutane ester, but on the other hand, yielded the cyclopropane ester J::!.J together with the lactone ester UU..), the unsaturated ester (VII) and the 23 ethoxy-ester (VIII). -C0 2 Et Et0 2 C-CHMe-CB=CB-C0 2 Et (VII) Et0 2 C-CBMe-CB2 -CH(OEt )-C0 2 Et (VIII ) Acrylic ester (III) was treated in a similar method. Again no cyclobutane ester could be detected, but only its unsaturated isomeride, ~ -methyleneglutaric ester \IX) which was the chief product apart from ~ thoxypropionic ester J.Kh'" \ '"' The experimental section of this paper, consisting of the following three parts, is particularly noteworthy; (a) Repetition of the original Markownikoff-Krestovmikoff procedure gave the suppo ffid trans-1, 3-cyclobutanedicarboxylic acid (II ) , melting point 1 70°. (b) Treatment of ethyl "'-.-chloro- y -methylglutarate under the Markovmikoff-Krestownikoff conditions gave trans-l-methyl-1,2-cyclopropanedicarboxylic acid (VII), melting point 168°, together with «.-methyleneglutaric acid and o<, -ethoxy- Y-methylglutaric acid. (c) Ethyl acrylate under the Markownikoff-Krestownikof'f conditions gave 107; of c:<.,- methyleneglutaric acid and YO% of @- ethoxypropionic acid. How the i dentifications of' trans-l,3-cyclobutanedicarboxylic acid (II) and trans-1- 24 methyl-1,2-cyclopropanedicarboxylic acid (VII) were arrived at was not stated. The last published wo r k on the Marlrnwnikoff-Krestownikoi'f acid came from this laboratory when Buchman., Reims and Schlatter (3) corroborated Markovmikoff and Krestownikoff' ' s synthesis and assumed the product to be 1,3-cyclobutanedicarboxylic acid. In the hands of this author, the preparation of the Markownikoff acid proved very unsatisfactory . Modifications in the Markownikoff-Krestownikoff procedure were carried out with the hope of devising a better synthesis of 1,3cyclobutanedicarboxylic acid through an elucidation of the reaction mechanism. Despite the fact that ethyl acrylate under the Markownikof.f-Krestownikoff' conditions gave no four-member ring compound ( 10) and ethyl d.. -chloropropiona te gave only 550 of the Ma rkownikoff-Krestownikof'f' acid, the writer considered it likely that an equimolar mixture of t.hese two esters under the Ma rkownikoff-Krestownikoff conditions might give a higher yield of the desired acid. The result of' this experiment was a 30% yield of an acid identical with that obtained by the original Ma rkovmikoff-Krestovmikoff' procedure. Dr. Saul Winstein of the University of California at Los Angeles had proposed (11) that the intermediate in the Markownikoff-Krestowni koff reaction, resulting from the condensa ti on of ethyl a, _ chloropropiona te and ethyl acrylate, (formed from ethyl "' -chloropropionate by the action of sodium 25 ethoxide) might be ethyl d..-chloro-~-methyl glutarate (XII). Although a satisfactory mechanism for the reaction ; Cl Et02C- ~-CH 2 -CH2 -C02Et ~H3 NaOEt ) XII A.Et Et~ VIII could not be devised, the experiment just described strengthened the case for (XII) as an intermediate in the Mar kovmikof.f-:Kr es townikoff reaction. Since it was not clear how (XII) could give (VIII), I proceeded to re-examine whether the structure of the Markownikof'f'-Krestownikoff acid had been correctly assigned. It immedia tely became apparent , .from a close study of the literature that no absolute structure proof had ever been adduced. In fact, all the previous work on the MarkownikoffKrestownikoff acid gave no positive indication that a fourmembered ring was present. A previously knovm finding which favored a three-membered ring over a .four-membered ring was the instability of the cis form of the Markownikoff-Krestownikoff acid noted by Haworth and Perkin, jun. (8), and verified by Bucbman, Reims and Schlatter (3) during attemp ts to convert the supposed cis-l,3-cyclobutanedicarboxylic acid to the trans form by heating with concentrated hydrochloric acid at 180° for several hours. An important additional consideration led the writer to the trans-1-methyl-l,2-cyclopropanedicarboxylic acid (VII) structure as a probable 1·ormulation of' the Markownikoff-Krestownikof'f acid. This was that it was possible in an ortho- 26 dox mann er to explain t he formation of (VII) from the Markownikof'f-Krestownikoff starting materials. 'l'he course oi' the reaction may be represent ed as follows; H3 C-CHC1-C0 2 Et -.. CH2 - CH-C0 2 Et XIII A search of the literature substantiated the above formulation. The reported melting point (169°) of trans-l-methyl1, 2-cyclopropanedicarboxylic acid (VII) · (9 ,12,.14,15) agreed with the melting point (170-173°) of the Markownikoff-Krestownikoff acid (l,3,6,8,lO); the melting point \141-142°) . reported for cis-l-methyl-1,2-cyclopropanedicarboxylic acid (XI) (9,12,13) coincided with the melting point \142-143°) of the cis form of the Markownikoff-Krestownikoff acid U, cf. 6,B)j and the melting point (3 7-42°) reported for the anhydride of cis-l-methyl-1,2-cyclopropanedicarboxylic acid ll2) was in reasonable agreement with the melting point (47.5-48.0°) found for the anhydride of the cis form of the Markownikoff-Krestow:nikoff acid (J,8). The Markownikoff-Krestownikoff acid and the acid obtained from ethyl acrylate and ethyl a< -chloropropionate under the Markownikoff-Krestownikoff conditions were conclusively demonstrated to be identical with trans-l-methyl-1,2-cyclopropanedicarboxylic acid (VII). Authentic (VII) was unambig- 27 uously synthesized from ethyl methacrylate and diazoacetic ester followed by a cid hydrolysis (cf. 12,13). CH3 I Et0 2 C-C - - -CH --)- I '\ CH-C02Et N;: ·_==N/ 6.. ) XIV VII The trans-l-methyl-1,2-cyclopropanedicarboxylic acid (VII) thus obtained melted at 172.2-172.7° and gave no depres- sion when mixed with either of the samples of the Markownikoff-Krestowni koff a cid. The p-toluidide of each of t hese samples of acid ,pad .. melting points of 205-206.5° and 21121 2° + From an understanding of the Markownikof'f-Krestownik of'f reaction mechanism it follows that a l;l mixture of ethyl methacrylate and ethyl chloroa cetate under the IvlarkownikoffKrestownikoff' reaction conditions could li k ewise give ethyl trans-l-methyl-1,2-cyclopropanedicarboxylat e (XIV) by t h e following reaction; +Two enantiotropic forms of the p-toluidide exist. 28 CH CH3 I H2 C ==C-C0 2 Et + ClH 2 C-C0 2 Et I 3 > I :: NaOEt CH2 -C-CO"'Et ~ ~ + Cl CHC1-C0 2 Et xv XIV This was experimentally realized. The lower yield ( 1010) obtained by this last method resulted from the difference in structure of the intermediate ions (XIII) and (X:V). In - ion (XIII) the chlorine is on a tertiary carbon and hence ' is considerably more react ive than in ion (XV) where the corresponding carbon is secondary. Also, the carbon bearing the negative charge in (XIII) is secondary and hence will form more easily than the corresponding ion (X:V) where the electron donating methyl group tends to repel the negative charge on the adj oining o<.carbon atom. An attempt to synthesize (XIV) by subjecting ethyl - o<.-chloro- oC-methylglutarate (XII) to the MarkownikoffKrestownikoff conditions was unsuccessi'ul, perhaps due to the inabilj_ty to obtain (XII) in a pure state (cf. 16,17 ). Also, it should be noted that the ion (XIII), postulated as an intermediate in the Markownikoff-Kres townikoff ,reaction, may not be formed from (XII). 29 EXPERIMENTAL trans-l-Methvl-1,2-cyclopropanedicarboxylic acid \VII) From ethyl acrylate and ethyl OC: -chloropropiona te Twenty- six grams ( 1.13 mole s ) of' sod.ium was dissolved in 500 ml. of absolute alcohol, and the excess alcohol was distilled off under reduced pressure. The last trace of alcohol was removed by heating at 180° for forty-five minutes under the ful l vacuum of an oil pump. After the flask was cooled to 50° , nitrogen was admitted. A total reflux partial ta ke- off head and a dropping funnel were connected to the sodium ethoxide flask, and a mixtur e of 100 g. ~1.0 moles) of ethyl acrylate and 13 7 g. ( 1. 0 mole ) of ethyl o( -cbloropropiona te was slowly added to the sodium ethoxide from the dropping funnel. By maintaining the reaction f l ask bath temperature at 12 0-130°, the reaction proceeded briskly. Ethyl alcohol distilled out (the tem1Jera ture at the top of the column was 78-80°) as the mixed esters were added. After the addition was completed (thirty minutes), the reaction mixture was maintained at 130° for an additional thirty minutes. The product was cooled in an ice bath and then poured into 400 ml. of water containing 20 ml. of 12N Lydrochloric acid. The organic phase which separated was washed with aqueous sodium bicarbonate and combined with the ether extract of the aque- ous phase. After the ether s olution was dried over sodium sulfate, tLe ether was stripped off, and the residue was d istilled . The following fractions were collected; 30 -r A 55-115° at 20 mm. B 115-124° at 20 mm. nD c 124° at 20 mm. (center cut) D 124-128° at 20 mm. no 26 1.436 6 2.8 g. nn 26 1.437 4.1 g. 25 1. 436 16. 8 g. Fractions B and D distilled mainly at 124° at 20 mm. The yield of crude ethyl 1-methyl-l,2-cyclopropanedicarboxylate was 83.7 g. (41.8; & ). Fractions .B, C, and D were combined and hydrolyzed by refluxing with 1 1. of 6N hydrochloric acid for sixteen hours. The resulting solution was boiled with Norite for ten minutes to r·emove the color, filtered, and evaporated down to dryness on a steam cone under reduced pressure. Two recrystallizations of the white crystalline residu e from hot water gave 41. 4 g. ( 2 8. 5%) of pure trans-l-methyl-1, 2-cyclopropanedicar- boxylic acid (VII), m.p. 172-173°. The moderate amount of 9is isomer (XI) which remained in the mother liquors was not worked up. The p-toluidide of (VII) was prepared by refluxing 0.5 g. of (VII) with 1.5 g. of p-toluidine for seventy-five minutes, bath temperature 210°. After cooling, the residue was triturated with dilute hydrochloric acid and recrystalli zed three times from absolute alcohol, clusters of thin prisms. The p-toluidide had two enantiotropic modifications o\,and 'iS m.p. 205.5-206.0° and 211.0-212.0° respectively. When the rate of heating was one degree per minute or less,modification t:A.was observed ; but when the r ate of heating was three degrees or more per minute •. modification @wa s noted. At an I -r 4'('. and @-ethoxypropionic e sters were formed a s by-p roducts. 31 intermediate rat e of heating, softening was noted at the melting point of d. • Anal. Calcd. for C20 H22 02N2; C, 74.50; H, 6.68. Found; C, 74.46i H, 6.87. From ethvl methacryl ate and diazoacetic ester (cf. 12, 13) A mixture of 57 g. 52 g. ~0.5 (0.5 mole ) of ethyl methacrylate and moles) of diazoacetic ester (Org. Syn., 24, 56 ) was slowly added to 110 ml. of refluxing xylene. The nitrogen evo- lution became vigorous after approximately one-third of the mixture was added. Refluxing was continued for three hours a.fter the addition was completed. After the bulk of the xylene was stripped, the esters of (VII) and (IX) were distilled, b.p. 120-125° at 20 mm.; yield, 34.5 g. (34. 5%). Basic sponifica ti on of these ester with lOO;;o excess of alcoholic potassium hydroxide, followed by acidification and ether extraction gave a mixture of the cis and trans acids (XI) and \VII). Repeated recrystciLJization from wa ter and then from acetone-benzene yielded the pure trans-l-methyl-1,2-cyclopropanedicarboxylic acid (VII), m. p. 172.2-172.71). The melting point of a sample of this acid showed no depression when mixed with a sample of (VII) which was prepared from ethyl acrylate and ethyl ciL-chloropropiona te. Anal. Calcd. for C6 H8 0 4 ; C, 50.00j H, 5.6 0. Found; C, 50.10; H, 5.5 7. The p-toluidide of (VII) was prepared as above, m.p. 205.0- 205. 7° and 211.5° respectively for the d. and ~ i'orms. Anal. Calcd. for Found; C, 74. 36 ; H, c 20 H 22 0 2 N 2 ~ 6.~6. C, 7h.) O; H, o.88. 32 From ethyl chloroacetate and ethyl methacrylate Starting from ethyl chloroacetate and ethyl methacrylate, (VII) was prepared by the same procedure as was employed when starting from ethyl ol -chloropropionate and ethyl acrylate. 'l'he yield of (VII), m.p. 172.5-172.7°, was 10% \estimated). The mixed melting point of this sample of tl1e trans acid and one prepared from ethyl methacrylate and diazoacetic ester showed no depression. From ethyl c:( -chloropropiona te ( 1) Starting from ethyl o<. - chloropropiona te, \VII) was prepared by essentially the same procedure as was employed starting from ethyl o( - chl.oropropionate and ethyl acrylate. This is the Marlrnwnikoff-Krestownikoff syn thesis of (VII). The yield of (VII), m.p. 172-173° was 2-4%. The p-toluidide of (VII) was prepared as des cribed above. The (;(. and ~ forms melted at 206.5° and 211.6° respectively. Anal. Ca.led. for C2 0 H22 0 2 N2 ; C, 74.50; H, 6.88. Found: C, 74.42; H, 6.86 . The attempted preparation of ethyl o<. -chloro-4(-methylµlutar-_ ate ( XIIJ.. A 25 0-ml. flask provided with a reflux condenser, a calcium chloride tube, and a hydrogen c hloride trap wa s cooled in a dry ice bath. To the flask were added 150 ml. of dry chloroform, 21.4 g. \0.147 mole ) of .r-carboxy- y-valerolactone+ (m.p. 70.5-71.5°) and 62.5 g. of phosphorus pentachloride. +( -Carboxy- "(-valerolactone was prepared by the procedur·e of J. Block, K. Kreckler, a nd B. Toll ens ( 18, cf. 17). 33 The dry ice bath was removed, and tl:e temperature was slowly raised until tl:e chlo roform refluxed. A vigorous evolution of hydrogen chlorid e ensued, and all of' the phosphorus pentachloride dissolved. After the gas evolution ceased, the chloro1·orm, phosphorus oxychloride, and excess phos phorus pentachloride were removed in vacuo. 'I'he residue was cooled to 0°, and 50 ml . of absolute alcohol was added. After standing overnight, the solution was poured into ice-cold aqueous sodium bicarbonate whe reupon a li ght yellow liquid separated. The aqueous phase was extracted twice with etller; the ether extract combined with the product, was dried over sodium sulfate and the ether was stripped off in vacuo. Distillation of ihe residue through an 8-inch column gave 13 . l g. of material, b.p. '.:18- 100° at 0.5 mm., nD 5 1.4460. Redistillation yielded two fractions; A, b.p. 102-105° at O.S mm., nD25 1.4481 and B, b.p. 105- 107° at 0.5 mm., 25 fl .,.., LJ 1.4471. A portion of frection B was analyzed~ Anal. Calcd. for C10 H17 0 4 Cl: C, S0 .74; H, 7. 24; Cl, 14.)18. Found: C, 51.04j H, 6.55i Cl, 10 .8 1. 34 REFERENCES (1) Wo Markownikoff and A.Krestkownikoff, Ann.., 208, 333 ll881). (2) ' R. C. Fuson, H. Gilman's Organic Chemistry, Vol. I, 2nd ed., J·ohn Wiley and Sons, Inc., New York, 1943, p.67. (3) E.R. Buchman, A. O. Reims and M. J. Schlatter, J. Amer. Chem. Soc . , .21!., 2703 (1942). (4) A. Freund, Monatsh., J, 626 (1882). (5) W. Ho Perkin, jun., Bera, 16, 1793 (1883). t6) W. Markownikof'f, Ber., .DlL 432 (1890). (7) 'N. H. (8) E. Haworth and W. H. Perkin, jun., ibid., lJ., 332 (1898). (9) C. K. Ingold, ibid., 127, 387 (1925). Perkin, jun., J. Chem. Soc., 65, 5 72 (1894) (10) F.Ro Goss and C. K. Ingold, ibid., 127, 2776 (1925)0 (11) Private communication from Dr. Saul Winstein. Cf. B. Keilin, Thesis for the degree of Master of Science, Calif. Inst. of Tech., 1945, p.22. l 12) K. v. Auwers and F. Konig, Ann.' 4JO, 276 ( 19 32). r (13) K. Auwers and E. Gauer, ibid., 470, 304 (1:12 9 ) • n \ 14) H. Staudinger, o. Mu).wyler, L. Ruzicka and S • . Seibt, Vo Helv. Chim. Acta, l, 401 (1924). (15) K. E. Wilzbach, F. R. Mayo and R. Van Meter, Jo Amero Chem. Soc., 70, 4069 (1948). (16) J. Brent, Ber., 19, 513 (1886). (17) J. W. Baker, Proc. Leeds Phil. and Lit. Soc., Sci. Sec. , 1_, 115 ( 19 30). (18) J.Block, K.Kreckeler and B.Tollens, Ann.,£]§_, 287 ll887). 35 THE SYNTHESIS OF CIS- AND TRANS-l,3-CYCLOf.U'.I'ANEDICARBOXYLIC ACIDS The synthesis of cis- and trans-1,3-cyclobutanedicarboxylic acids (I) and (II) became a problem of prime i mportance once it was established that the Markownikoff acids were cisand t.rans-l-methyl-1,2-cyclopropanedicarboxylic acids (IV) ~cf. preceeding section). ~III) and Hence authentic 1,3-cyclo- butanedicarboxylic acid had not previously been prepared. A d HH{J C0 2 H H~ -C0 2 H -C0 2 H C0 2 H II I III H~ -C0 2 H H0 2 C- IV number of probable routes to (I) and (II) were extensively investigated while other less promising routes were only sup e rficia l l y a nd ~II) studied. A highly satisfactory synthesis of ~I) was eventu ally developed. An obvious paper synthesis of ethyl l,l,3-cyclobutanetricarboxyla te (XI) involves t he condensation of ethyl ~' ~ 1 dibromoisobutyrate (VII) with sodium malonic ester . Hydrolysis and d ecarboxylati on of (XI) would give (I) and (II). I Since ethyl ? , ~ -di bromoisobutyra te (VII) has not been reported in the literature, it was prepared as outlined in Table 1. ~ , f3~-Diiodoisobutyric acid had be en synt.hesized by Glattfeld and Schneider (1 ), bvt t heir me thod vvas not adapt/ a ble to the pr epa ration of fl , ~ -dibromoisobutyric acid (VI). VV'elch ( 2) prepared bis ( hydroxymethyl )-malonic ester ( V) and 36 Table 1 HBr ) v BrH 2 C-C-C0 2 H II ;- CH2 . l VIII VI EtOH H2 S04 BrH 2 C-C-C0 2 Et H CH2 IX VII ' H2 C( C0 2 Et ) 2 H2 C( C0 2 Et)2 NaOEt NaOEt 1 H2 C( C0 2 Et )z NaOEt (Et0 2 C) 2 CH-CH2 - C-C0 2 Et JI CH2 x XI I II Attempted Syntheses of 1,3-Cyclobutanedic a rboxylic Acid 37 reported that hea ting lV) with hydriodic . ac id gave ~ -liodomethyl)-acrylic a cid- t h e iodo analog of (VIII). The synthesis of large quantities of bislhydroxymethyl)-malonic ester \V) was a cr1ieved by improving the procedure of Gault and Roesch l 3). 1reatmsnt of (V) with hydrobromic acid yielded a mixture of I ? , ? -dibromoisobutyric acid (VI) and <J'.. - ~ bromomethyl )-acrylic acid (VIII). Upon esterif'ication with ethyl alcohol and sulfuric acid, (VI) and (VIII) gave the corresponding esters lVII) and (IX). When the esterification of (VIII) was carried out with alcoholic hydrogen bromide, hydrogen bromide added to the double .bond and the ester (VII) was produced. Both (VII) and \IX) might be expected to yield ethyl 1,1,J-:cyclobutanetricarboxylate (XI) when treated with sodium malonic ester, the former by a Perkin reaction l4), the latter by a malonic ester condensation followed by an internal Michael condensation. Numerous cases of internal Michael condensations leading to cyclobutane compounds are discussed at length by Ingold and co-workers (5). When (VII) was treated . with sodium malonic ester, it presumably underwent d ehydrohalogenation giving (IX) . as the first step in the reaction. This was borne out by the fact that both (VII) and (IX) gave the same quaternary picrate when treated with trimethylamine followed by lithium picrate• 38 Br- CH2 + fl l (H 3 C) 3 N-CH 2-C-C02Et VII BrH 2C-C-C0 2Et L.i picra te + (picrate) (H 3 C) 3 N-CH 2 -C-C0 2 Et II II CB2 CH 2 IX XII An analogous case of dehydrobromination was noted by Perkin, jun. and Simonsen (6); BrH2C-CHBr-CH2Br + H2C(C0 2Et) 2 NaOEt ) H2 C=CBr-CH2-CH(C0 2Et) 2 From both (VII) and (IX) with sodium malonic ester only traces of distillable ma terial could be obtained . The ester (X) preswnably underwent a series of inter- mo lecular Michael condensations leading to non-volatile polymers. In t Le hope t ha t milder conditions would give the desir ed cyclobutane ester (XI), magnesiwn malonic ester was employed in place of sodium malonic ester. Both (VII) and lIX) gave the same product· etl1yl cl-.- carbethoxy- '< -methylene6lutara te \X). All attempts to induce cyclization of lX) to (XI) by the action of' magnes ium ethoxide, sodium ethoxide, or piperici ine failed. The structure of (X) was verified by hydrolysis and deca rboxylation to o\ - methyl eneglutaric acid . and by hydrogenation, hydrolysis, and de carboxylation of (X) to o<-methylglutaric acid. 39 Although the addition of hydrogen bromide to (X) would yield ethyl oC. - l bromomethyl )- y -carbethoxyglutara te, which upon elimination of hydrogen bromide might give ethyl 1,1,3- cyclobutanetricarboxylate (XI), the f'act that Michael condensations are reversible led the author to believe that (XI) would not be stable in basic media. It is thought that the ring would open giving (X) followed by inter-molecular Tuachael condensations leading to polymers. Starting from bis(hydroxymethyl)-malonic ester (V),other routes to 1,3-cyclobutanedicarboxylic acid were open which eliminated the difficulties noted above. With this in mind the tosylate of (V), (XIII), was prepared in the usual manner and subjected to a malonic ester condensation. The only identif'ied product that could be isolated was a small amount of o<. -methyleneglutaric acid. At tempts to prepare ethyl bis ( bromomethyl )-rnalona te (XIV) from the tosylate (XIII) by an exchange reaction gave tars and unreacted starting material. By refluxing (XIII) with sodium iodide for six days at 100°, only one tosylate group could be replaced by iodine. XIII XIV Gutzeit and Dressel (7) had reported the synthesis of methyl 1,1,3,3-cyclobutanetetracarboxylate (XV) by the following steps: 40 base + ) NaOEt xv Although the ester (XV) gave proper analytical results, attempts by Gutzeit and Dressel to convert it to 1,3-cyclobutane- dicarboxylic acid did not give a pure product. Repetition of' this work by Schlatter (8) and by this author were unsuccessful.. N_o pure (XV) could be obtained. Although a modification of the Kerr (9 ) synthesis of norpinic acid ( 2,2 -dimethyl- 1,3- cyclo butanedicarboxylic acid) (XVI) would be expected to yi eld 1,3-cyclobutanedicarboxylic NH NH NaOEt NC NC hydrolysis ) ) X:VI acid, such a synthesis did not prove feasible. Attempts to 41 repeat Hi gson and Thorpe's (10) synthe sis of ethyl o<. , y - dicyanoglutara te (XVII) were uniformly unsuccessful. (XVII) polymerEt02C-CH(CN)-CH2 -CH(CN)-C0 2Et XVII ized in the column and condenser during distillation (cf.11) . Since it was not possible to prepare 1,3-cyclobutanedicarboxylic acid by direct methods, the writer·'s attention turned to methods of' preparing compounds of the typ e (XVIII) and (XIX) where Y is a methylene or hydroxymethyl group which y O C0 2 Et C0 2 Et XVIII XIX could be oxidized to a carboxylic acid after the cyclobutane ring was formed. ~-(Chloromethyl)-allyl ly chloride (XX) would theoretical- give ethyl 3-m~thylene-1,1-cyclobutanedicarboxyla te (XXI) which could be converted to 1,3-cyclobutanedicarboxylic acid. H2 C(C0 2 Et)z NaOEt xx <> C0 2 Et H2 C= C0 2 Et XXI (XX) was prepared by the method of .Mooradian and Cloke ( 1 2) starting from pentaerythritol. Treatment of' (XX) with sodium malonic ester did not give any detectable amount of (XXI). 42 Presumably the methylene group in (XX) increased the bond angles (109° to 120°) sufficiently to prevent ring closure. A somewhat similar alternate route to 1,3-cyclobutane- dicarboxylic acid would start from YCH 2 -CH( CH2 Br )z (XXII) which might be prepared as follows; LiAlH4 HBr ) ) YCH2 -C ( CH 2 Br )z XXII In this case the proposed method failed when the tris(hydroxymethyl)-me thane (L'CIII), although it presumably formed, could + not be isolated from the inorganic salts • 'l'he eventual synthesis of' cis- and trans-l , 3-cyclobutanedicarboxylic acids (I) and (II) was achieved as outlined in table 2 ++ • 2,2-Bis(bromomethyl)-l,3-propandiol (XXIV) was prepared by the method of' Beyaert and Hansens (14), Zelinsky and Krawetz's (15) synthesis being unsuitable for large-scale preparative purposes. Following the meager directions given in Chemical Abstracts, the author obtained a 25- 35/o yield of \ XXIV) in contrast to the reported 80~10 of Beyaert and Hans ens ( 14) . After this work was completed a copy of the original TAfter t his work was discontinued an alternate route to (XXIII) appeared in the literature (13) starting from formaldehyde and acetaldehyde; 2CH 2 0 OH ) (HOH 2 C) 2 CH-CHO H2 --? (HOH 2 C) 2 CH XXIII ++A somewhat analogous scheme was suggested to the writer by Dr. - James Conley. Cf. Conley, Ph.D. Thesis (Propositions), Cal. Inst. of Tech., 1950. 0\ c::t ' 7 C0 2 C5H11 H0 2 C Ac0H C CH2 Br AcOH2 C' ' CH2 Br ) Na0C 5 I-h1 KOH / 6) HOH 2 C O H+ Et OH XXIX I ' / -C0 2 H H c HOH2 C, / CH2 Br + XXIV o- o- XXVII H+ ' ./ II O Hy CHO ,C0 2 CsH11 'co2CsH11 A H xxx HOH 2 C~COzH HOH 2 C~ 0 2 H HOH 2 C/ ' CH2 Br 0 2K HOH2 C~0 2 K H2 C(C02CsH11 )z C02H H0 2 C~C0 2 H XXVIII 2 ' c/ Table 2 The Successful Synthesis of l,,J-Cyclobutanedicarboxylic Acid HBr 0 XXVI - = >CH2 Br CH2 Br CH2 0H HOH2 C, c' HOH 2C HOH2 C~C02CoH11 0 HO Ac 7 HOH2 C/ ' cH 2 0H H+ H2CJ HN'0 3 ~ XXXI ' 7 44 article was obtained. Since this journal, Natuurwetenschap- pelijk Tijdschrift , is quite obscure and the article is written in Flemish, Beyaert and Hansens' procedure is given in detail in the experimental section. Because (XXIV) could not ·+ be used directly in the malonic ester condensation, tl1e f'ree l1ydroxyl groups were protected by preparing the benzylidene derivative (XXVI). (XXVI), when subjected to a malonic ester condensation,gave isoarnyl benzylidene-1,1-bis(hydroxymethyl)3,J-cyclobutanedicarboxylate (XXVII). Acid hydrolysis liber- ated isoamyl l,l-bis(hydroxymethyl)-3,3-cyclobutanedicarboxylate (XXVIII). Basic saponification produced potassium 1,1-bis- (hydroxyrnethyl)-3,3-cyclobutanedicarbo.xylate (XXIX) which was oxidized directly with nitric acid to l,l,3,3-cyclobutanetetracarboxylic acid (XXXI). Alternatively, the free l,l-bis\hydroxymethyl)-3,3-cyclobutanedicarboxylic acid (XXX) was from its potassium salt (XXIX) liberated 2nd tl1en oxidized to the tetra ' acid (XXXI). The latter method has the advantage that no salts are present to interfere with the isolation of the product (XX.XI). Decarboxylation of (XXXI) gave a mixture of cis- and trans-1,3-cyclobutanedicarboxylic acids (I) and (II). The pure cis acid (I) was prepared via its anhydride from the mixed cis and trans acids (I) and (II). By fractional crystall~- +2,6-Dioxaspiro(3,3)h~ptane (X.XV) is formed under these . conditions. Cf. Backer and Keuning (16). base XXIV ) · xxv 45 zation of the mixed cis and trans acids from benzene-dioxane and acetone the pure trans acid (II) was isolated. 46 EXPERIMENTAL Ethyl bis(hydroxymethyl)-malonate (V). In a 5--1. round-bottom flask, provided with a powerful Hershberg stirrer, were placed 250 g. of sodium carbonate, 2 1. of water, and 1 kg. of ice. After the temperature had fallen to 0°, 5 70 ml. ( 6. 3 moles) of 3 7'}~ aqueous formaldehyde was added all at once 1 and then 480 g. (3.0 moles ) of ethyl malonate was slowly run in (thirty to forty-five minutes). The mixture was maintained at 0-5° by the occasional addition of ice and vigorously stirred during the addition and for one hour thereafter. Technical grade ammonium sulfate was added until the solution was saturated- two liquid phases formed. One liter of ether was added: and the lighter organic phase wa s separated. The ether was stripped off under a water aspirator at room temperature for eighteen hours and then at 50-60° for thirty minutes. Upon cooling to 0°, the entire residue solidified, weight 544 g. Crystal.lization from 1 1. of carbon tetrachloride gave 315 g. (48%) of nearly pure ethyl bis(hydroxymethyl)malonate (V), m.p. 49° (Welch (2) reported m.p. 52° ). A second crop of less pure product separated upon standing, 63 g. ( 9%). The tosylate of ethyl bis(hydroxymethyl)-malonate (XIII) was prepared by the method of Sekera and Marvel tion ~17). A solu- of 71 g. (0.37 mole ) of the dial (V) in 350 ml. of pyridine was cooled to 0° and 155 g. (O. B mole ) of tosyl chloride (Eastman Kodak Co., white label) was added portion- 47 wise during one hour with occasional stirring. After standing over night,, the reaction mixture was poured into 1 1. of 20)o sulfuric acid. An oil separated which quickly solidified. 'l'he air dried product crystcllized slowly from methyl alcohol as splendid rhombohedral prisms~ yieldw nearly quantitative. After two recrystcill.izations from methyl alcohol the tosylate melted sharply at i8° Anal. Calcd. for C;.,; 3 H1 eS 2 010: C, 52.14; H, 5.34. Found; C, 52.31; H, 5.40. ~ , e''-Dibromoisobutyric acid (VI) and ~(. bromomethyl )-acrylic acid. (VII) . To l 1. of' 48/o 11ydrobromic acid in a 2-1. round-bottom flask was added 220 g . (1. 0 mole ) of ethyl bis\hydroxymethyl)malonate (V). The solution was refluxed for exactly one hour under a total reflux partial take-off condenser; during tl1is time 100 ml. o.r distillate was collected (mostly etliyl brom- ide). The light yellow residue was transferred t o a beaker and cooled to -12°. A crystfil_ine precipitate separated whicl1 was filtered off on a sintered glass funnel and then dissolved in hot petroleum ether (b . p. 100-130°). The aqueous phase which remained was combined with i:,he original filtrate , refluxed for an additional hour, cooled to room temperature 4 and extracted twice with 100-ml. portions of benzene. These extracts were combined with the above petroleum etber solution, solvents were stripped off, and the residue was di s til led ., Two main fractions were col'l ected. Fra ction A, 5 7 g . ( 32 /o ), b. p. 108° at 3 nun., vms crystallized from petroleum ether l b.p.100- 48 130° ); m.p. 73.5°. Fraction B, 24 g. l>'/b), b.p. ll+0-142° at 3-4 mm., was crystallized twice from water; m.p. "79.2-'.19.7°. The analyses agreed with the formula ti on of A as d... - ( bromoI methyl)-acrylic acid and B as ~ ,, ~ -dibrornoisobutyric acid. Anal. (A) Calcd. for C4 H5 02 Br; C, 21.12; H, 3.06. Found: C, 29. 20; H, 3.14. Anal. (B) Calcd. for C4H60 2 Br 2 ; C, 19.53; H, 2.46. Found: C, 19.68; H, 2.)2. Ethyl c:{ - ( brornomethyl )-acrylate (IX). A mixture of 5 7 g. (0.345 mole ) of o<. -(bromornethyl)' acrylic acid (VIII), 50 ml. of absolute alcohol, 150 ml. of carbon tetrachloride and 1 ml. of sulfuric acid was refluxed for four hours during which time the water that formed was removed by means of an automatic liquid separator. '1 .he 1 carbon tetrachloride solution was washed with aqueous sodium bicarbonate and dried over magnesium sulfate. Distillation of the carbon tetrachloride solution gave 4 7.8 g. of the crude ester, h.p. 80° at 10 mm. Redistillation through an 18-inch column gav e 33.4 g. (66.5 7; ) of pure ethyl d. -(bromomethyl)-acrylate 25 26 (IX), b.p. 80° at 10 mm., nD 1 .4 734, d4 ~ 1.301. Anal. Calcd. for C 6 H~0 2 Br: C, 37.33; H, 4o70. Found: C, 37.45; H, 4.79. 49 . I Ethyl 0 .@-dibromoisobutyrate (VII). Crl)de d.. - ( bromomethyl )-acrylic acid, 24 g. ( 0.145 mole ) , - was dissolved in 100 ml. of absolute alcohol, and the solution was saturated with gaseous hydrogen bromide at 0°. After standing overnight at 4° the solution was poured into 1 1. of water whereupon a brown oil separated. The aqueous phase was extracted with 100 ml. of chloroform, and the extract was combined with the bulk of the ester. After being washed with wa ter, aqueous sodium bicarbonate, and again with water, the chloroform solution was dried over magnesium sulfate. Distillation_ of the chloroform solution gave 17.6 g. l44%) of the dibromo-ester (VII), b.p. 120-121° at 10 mm., nz5 D 1.4738, d~ 1.711; M.RD calcd., 47.07, found, 46.95. 3 Anal. Calcd. for C5H1 0 0 2 Br 2 : C, 25.70; H, 3.68. Found: C, 26. 05; H, 3.78. ,, This ester (VII) was also prepared from ~ , ~-dibromoisobutyric acid (VI) by direct esterification as in the case of ethyl ol-(bromomethyl )-acrylate (IX). ""- - ( Chloromethyl )-acrylic acid. A mixture of SS g. (0.25 mo le ) of ethyl bislhydroxymethyl )-malona te (V) and 200 ml. of 12N hydrochloric acid was refluxed for one hour. UiOjo) of Upon being cooled to -10°, 12 g. ol. - ( chloromethyl )-acrylic acid crystallized out as large blades. After three recrystalli zat ions i'rom petroleum ether (b.p. 100-130°), the acid melted at 60.0-61.2°. Anal. Calcd. for C4 H5 0 2 Cl; C, 39.8); H, 4. 18 . Found : C, 39.94; H, 4.2 7. 50 Ethyl G( _ ( chlorornethyl )-acrylate. A solution of 3 7. 6 g. ( 0. 313 mole ) of' ci- ( chloromethyl )acrylic acid, 0.5 ml. of sulfuric acid, 100 ml. of absolute a lcohol, and 100 ml. of carbon tetrachloride was refluxed fo~ five hours. The water that formed during the esterifica- tion was removed from the system by means of an automatic liquid separator. After being cooled the solution was poured into water; the organic phase was separated, washed with aqueous sodium bicarbonate, dried over n:agnesium sulfate, and the solvent was distill ed off. Distillation of the residue gave 26.8 g. (56. 5/o ) of ethyl o<.-(chloromethyl )-acrylate, b.p. 70-75° at 16 mm., n~ 5 1.446, d! 6 1.08 74. The attempted preparation of ethyl 1 ,1, 3-cyclobutanetricarbo:x:ylate (XI). Me thod A. To a solution of o.o g. ( O. 2 86 mol e . ) of sodium in I' / 200 ml. of absolute alcohol was added 23 g. l0.143 mole ) of Upon the slow addition of 3:1.4 g. (0.143 dry e t byl malonate. ( mole ) of e t hyl t3 ,~ -dibromoisobutyrate (VII) to the above hot solution, a vigorous reaction took place and sodium bromide precip i t a ted . After the mixture was refluxed for t wo hours, the prec ipitate was filtered off and washed with absolute alcohol, and the combined filtrates were reduced to a small volume on a steam cone under reduced pressure. The residue was washed twice with water, dried over magnes ium sulfate, and distilled in va cuo. The following fractions were collected: 51 below 150° at 10 mm. 2.5 g. to 150° at 1-2 mm. 4.7 g. to 22 0° at 1-2 nun. 8.6 g. residue, estimated 25 g. None of the fractions was homogeneous. The estimated boiling point of lXI) was 1 25 -135° a t 3 ll'Jll. This experiment was repeated three times with essentially t he s ame results- tars and high boiling products. Method B. '.I'o a solution of S. 0 g. l 0. 205 mole ) of sodium in 200 ml. of absolute alc ohol was added 26.3 g. (O. l o4 mole : of dry ethyl malonate. After the solution was cooled to 0°, 31. 7 go ( 0 . 164 mole ) of ethyl d.. - ( bromomethyl )-acrylate l IX) was added over a five minute period. Thi s gave rise to a vigorous reaction; heat was evolved, and sodium bromide precipitated. The reaction mixture was then heated to rei'lux for two hours, cooled to room temperature , and poured into dilute hydrochloric acid. Af ter the organic phase was separated, t he aqueous phase was extracted twice with ether. The combined ether extract and organ ic product were dried over magn esium sulfate, and the ether was stripped off under a water pump. Dist1llation of the resi due gave only small arr,ounts of high boiling ma terials and non- d istillable tars. Ethyl o<. -mel:hyl ene- ( - carbethoxyglutara te l X). :Magnesium turnings, 12.5 g. lO. ) mole ), were refluxed with 300 ffil. of absolute alcohol, a cry stal of iodine, and 40 g . l0 .25 mole ) of dry ethyl malonate. After the reaction subsided, 200 ml. of purified isopropyl ether was addedj the ffi ixture was refluxed for twenty hours after which all of the magne sium turnings had dissolved, and a. large white precipitate had formed. The mixture was cooled to room tempe rature, I and 71 g. (0.25 mole was added ensued of ethyl ~ , (3 -dibromoisobutyra te (VII) all at once with stirring . A vigorous reaction which was moderated with a water bath. When the reac- tion had subsided, t he mi xture wa s refluxed for ninety minutes, and then the bulk of the alcohol and ether were di stilled off. The residue was poured into water, and treated with hydrochloric acid until all of the solid salts had dissolved. The organic mate rials were then taken up in isopropyl ether. After t.te aqueous solution was extracted with isopropyl ether, the combined isopropyl ether solutions were washed with a queous ferrous sulfate and dried over magnesium sulfate. The isopropyl ether was removed on a steam cone under reduced pressure. Upon fractionation through a ten-inch column t.he residue yielded 36 g. (53;0) of e thyl o<. - me thylene- r -carbe thoxyglutarate (X), b.p. 130° at 3 mm., nf) 5 1.!+li08. Anal. Calcd. for C13 H20 0 6 : c, 5 7.Jw. ; H, 7. l.+ O. Found : C, 5 7 • 40 ; H, 7 • 43 • Ethyl °' - me thylene- '(-carbethoxyglutara te (X) vms a lso prepared from ethyl o<. -(bromomethyl)-acrylate (IX) in a similar manner. 53 A solution of 2. 99 g. of ethyl c( -methylene- Y-carbethoxyglutarate (X) in 35 ml . of absolute alcohol was prepared, and 20 mg. of Adams'. catalyst was added. On hydrogenation , until no further amount of hydrogen was absorbed, 323.S ml. (102%) of hydrogen wa s taken up. The hydrogenated product was hydrolyzed wi th lOO;G excess of alcoholic po tassium hydroxide. The alcohol was removed in vacuo, and the residue was acidified with hydrochloric acid. Af ter the solution was evaporated to dryness, the residue was deca r boxylated by heating to 220° for twenty minutes. The product was extracted three times with ether; the ether was evaporated off from the extract, and the material which remained wa s dis solved in a little hot water. vYhen this solution was cooled to 0°, c;{. -methyl- glutaric acid crystallized out, m.p. 76.0- 76.5 ° (literature values; (18), 77-78°; (19), 77.5°) . Three grams of ethyl d... -methylene- \'" - carbethoxyglutara te lX) was refluxed overnight with 5 0 ml. of 6N hydrochloric acid. After the solution was evaporated down to dryness, the residue, ~ -methy l eneglutaric acid, was recrystallized twice from water (Norite); m.p. lJ0.0-130.5 °. A mixed melting point of this material with an authentic sample of' ~ -methyleneglutaric acid showed no depress iona l ~ -Carbethoxyallyl) tr:t_m~thy]-ammonimn pi_c r_atL.UQ_Lh Approximately O. 3 g. of ethyl oC. - \ chloromethyl )-acrylate was treated with an excess of trimethylamine in benzene. The precipitate which formed inunediately was centrifuged, washed with a little benzene, and dried in vacuo. After the salt was dissolv ed in a little water, an excess of saturated aqueous 54 lithium picrate wa s added; there resulted the i rmned i a te crystalliza tion of the yellow-orange picrate (A). \A) was recrystallized twice from water; m. p. 130.0-130.2°. Anal. Calcd .. for C15 H20 0sN4; C, 45.00; H, 5.04. Found: C, 44.9 7; H, 5.04. A picrate 130.0°~ (B), m.p. ~-(bromomethyl)-acrylate was obtained from ethyl by the above procedure. The mixed melting point of lA) and (B) was 130.0-130.2 ° . A picrate (C), m.p. 129.S-130.5°, was obtained from / ethyl ~ , ~ -dibromoisobutyrate by the a bove procedure. 1'he mixed melting point of (A) and (C) was 130.0-130.5°. Ethyl (iodome thyl)- (tosyloxymethyl)-malonate. , A solution of 52.8 g. (0.10 mole ) of ethyl bis \tos yl- \XIII) oxymethyl)-malonate / _ and 60 g. (O.Li mole) of' sodium iodide in 500 ml. of butanone was refluxed for six days; during this time sodium tosylate precipitated out, and the solution turned black (free iodine). After the precipitate was filtered off, the butanone was stripped off, and the residue was poured into water. The product was taken up in benzene, washed with aqueous sodiwn bisulfite, and dried over magnesium sulfate, and the benzene wa s stripped off. Recrystallization of the residue three times from absolute a l cohol gave pure ethyl (iodomethyl)-{tosyloxyrnethyl)-malonate, Found ; C, J}.77; H, 4.51. 55 ~2-Bis ( bromomethyl )-1, 3-propandi ol (XXIV). A solution of 246 g. (3.0 moles) of gaseous hydrogen bromide,1530 g. (9.0 moles) of 48~ hydrobromic acid, and 408 g. l3.0 mole s) of pentaerythritol in 1440 g. of glacial acetic acid was refluxed for rorty hours. After being cooled to room temperature, the solution was diluted with water to 10 1., whereupon the diacetate of \XXIV) separated as a brown oil. Further small amounts of this ester were obtained from ~he aqueous solution by ether extraction after the free acid was neutralized. The crude diacetate of (XXIV) was refluxed with 500 ml. of absolute alcohol under a partial take-off total reflux condenser. After all of the ethyl acetate had distilled out and the boiling point of 77° had been reached (six hours), the excess ethyl alcohol was removed on a steam cone under reduced pressure. Repeated extraction of the oily residue with hot water gave 265 g. of the crude crystalline diol (XXIV). The air dried product was recrystallized from benzenej yield, 225 g. l32 fo ), m.p. ' 102-104° (Beyaert and Hansens (ll-1-) reported m.p. 111°). An additional 55 g. ( 8/b) of less pure product was obtained by further extraction. Jhe procedure of Beyaert and Hansens \14) for ara tion of l XXIV). the pr~ Glacial acetic aci.d, 120 ml., 20 g. of pentaerythritol, and 1 ml. of 66~ hyd~obromic acid w~re placed in a 500-ml. two-necked flask which was provided with a reflux condenser and a dropping funnel. The mixture was refluxed until the solid d issolved; 41 ml. of 66/u hydrobrornic acid was then added slowly, and the solution was refluxed 56 for an additional twenty hours. The reaction mi x ture was reduced to dryness wi th moderate heating in vacuo. The residue, SO ml . of absolute alcohol, and 3 ml. of 66/b hydrobromic acid were heated in a dis tilling flask which was provided with a two-meter fractionating column. Etbyl acetate ( 25 ml. ) distilled over slowly at 71°. next 10 ml., the - During the distillation of the temperature rose slowly to 7 8~. An addi- tional 50 ml . of absolute alcohol was added to the pot and slowly distilled out. The resi du e wa s concentrated in vacuo, evaporated down with 5 0 ml. of toluene, hea ted to 115 ° in vacuo, and again evaporated down with 5 0 ml. of toluene . After the residue was boiled with 100 ml . of benzene, the benzene soluti on was poured into a clean beaker. As the solution cooled the bromide lXXIV) crystallized out. The solution and precipitate were set in an ice box for one hour, and the product wa s then filtered off and washed with a li t tle benzene; weight 31 g .. , m.p. lOJ-105°. One recrystallization 1'rom water yielded 29.6 g . of (XXIV), m.p. 111°. Benzylidene-2,2-bis(bromomethyl)-l,3- oropandi ol (XXVI). A mixture of 225 g. (0. 86 mole ) of 2,2-bis( bromome thyl)l,J-propandiol (XXIV) , 100 g. (0.95 mo l e ) of benzaldehyde, and 3 g. of p-toluenesulfonic acid was refluxed with 1 1 . of benzene. In the course of t hree hours distilled. out 19 ml. of water and wa s collected in an automatic liquid sepa- r a tor. The resj_due was washed with aqueous sodium carbonate, dr ied over anhydrous potassium carbonate, and stripped of ben zene . Crysta llization of the residue from methyl alcohol 57 yielded 217 g. (727; ) of the benzylidene derivative (:X..XVI); m.p. 67.6-68.3°. Recrystallization from methyl alcohol gave colorless clusters of prisms; moP• 68.0-68.3 ° . Anal. Calcd. for C1 zH 14 02Br2: C, 41.1 7; H, 4.03. Found: C, 41.59i H, 3.99. Jsoamyl benzylidene-l,l-bis(hydroxymethyl)-3.3-cyclobutanedicarboxylate (XXVII). A 2-1. standard-tapered, thre&necked flask was provided with a "" tru-boren stirrer, a partial take-off total reflux condenser, a tube for the addition of solids, and a 500 ml . graduated dropp i ng funnel-the stem of which extended to the bottom of the flask. Isoamyl alcohol, 1200 ml., was put in the flask,and a small f'orerun distilled off. After 22.l g. (0.916 mo l e ) of sodium was dissolved in the alcohol, the solution was cooled somewhat,and 500 ml. of' this solution was forced up into the graduated dropping funnel by a stream of nitrogen. Dry ethyl malonate, 88.5 g. (0.547 mole ) , was added to the reaction flask . The solution was heated to reflux and 160 g. (0.458 mole ) of benzylidene-2,2-bis(bromomethyl)l,3-propandiol (XXVI) and the 500 ml. of' sodium isoamylate solution contained in the dropping funnel were added in portions as follows: Each of the above addenda was divided into eight equal parts ,and a t thirty minute intervals one part of each was added to the reaction flask. During the addition , 200 ml. of ethyl and isoamyl alcohols were distilled out from the reaction flask ,and the distilling temperature 58 reached 123°. Refluxing was continued for sixteen hours. After the bulk of the isoamyl alcohol was distilled off, the mixture was cooled to room temperature and washed with 200 ml. of 2N hydrochloric acid. Ether extraction of the aqueous washing gave further small amounts of product. This residue con- taining the isoamyl benzylidene-l,l-bis(hydroxyrnethyl)-3,3' cyclobutanedicarboxylate (XXVII) was used directly in the following reaction. lJl-Bis(hydroxymethyl)-3,3-cyclobutanedicarboxylic acid (.XXX). The crude condensation product (XXVII) was hydrolyzed by refluxing with 350 ml. of water and 20 ml. of 12N hydrochloric acid. Water, benzaldehyde, and isoamyl alcohol were distilled off until the distilla-te was clear. To the residue (two phases) was added 112 g. of potassium hydroxide (100/o excess) dissolved ' in 400 ml. of ethyl alcohol. A vigorous reaction ensued and a large white precipitate formed. After the mixture was heated on a steam bath for one hour, the bulk of the alcohol and water were removed under reduced pressure. The residual pasty mass was repeatedly shaken with absolute alcohol until the supernatant liquid was colorless. The precipitate was filtered, washed with ether, and air dried. To this potassium salt 100 ml. of 12N hydrochloric acid was added, and the resultant precipitate of potassium chloride was filtered off and washed with acetone. Combination of the hydrochloric acid solution and the acetone washings gave an additional precipitate of potassium chloride. The total weight of' potassium chloride 59 was 82 g. After the acetone was removed by disti llation, the pH of the solution was adjusted to the acid side of Congo red. This solution was then extracted with ether using a special emulsification stirrer + and a down-draft condenser. The ether extract yielded 7 g. of dark oily residue upon evaporation of the solvent. The aqueous solution was further a cidified with 90 ml. of 12N hydrochloric acid and again extracted for twenty hours. The ether flask contained two phases ~ (l,l-bis(hydroxymethyl)-3,3-cyclobutanedicarboxylic acid (XXX) is very hygroscopic and nearly insoluble in ether). Distillation of' the ether left 20 g. of' light brown oily material. Upon further extraction for tllree days an addi ti.onal 40 g. of similar material was obtained, presum.ably a concentrated aqueous solution of the hydroxy-acid (XXX). l,l,3,3-Cyclobutanetetracarboxylic acid lXXXI). ' One portion, 20 g., of the crude aqueous solution of l,l-bis(hydroxymethyl)-3,3-cyclobutanedicarboxylic acid (XXX) was mixed with 0.2 g. of ammonium vanada te in a 1-1. flask provided with a reflux condenser. In one portion, 2 00 ml. of 70fa nitric acid was added. A vigorous reaction started after a few minutes; ni tm-g;en oxides were copiously evolved, and the solution refluxed. Af ter the reaction had subsided ( ten minutes), the mixture was refluxed for another ten minutes, and then placed in an ice bath for thirty minutes . The nicely crv stalline precipitate which separated was filtered off and + A special stirrer was employed which consisted of a glass tube, sealed at the center and at both ends, and having two holes three inches from one end and a lso three short curved glass tubes sealed perpendicular to the shaft at this same end. In operation tlie ether wa s sucked in through the holes aud sprayed out in the lower aqueous phase. This greatly increased the efficiency of the extraction. 60 dissolved in water. The bulk of tbe water and nitric acid was removed on a steam bath under reduced pressure, and the residue was dried at 60° in vacuo. Crystallization of the product from acetone-benzene gave 11.8 g. of t be tetra-acid (XXXI). By distilling off the nitric acid from the mother liquor, an additional 5.7 g. of less pure tetra-acid was obtained. The estimated overall yield for the malonic ester condensation, acidic B.nd basic hydrolyses, and the nitric acid oxidation was 50%. A portion of the tetra-acid (XXXI) was recrystallized from ether-benzene; m.p. 205° (dee. ) . Anal. Cal ed. for C8 He0 8 ; C, !+l. 38; H, 3. 4 7. F'ound; C, 41. 35; H, 3.)l. cis-1,3-Cyclobutanedicarboxylic acid,I). l,l,3,3-Cyclobutanetetracarboxylic acid OCXXI) was decarboxylated to a mixture of cis- and tr.§12§.-l,3-c yclobutanedicarboxylic aci.ds (I) and (II) by heating to 210-220 ° for . thirty minute s . The resultant ~ixture was refluxed with ten_ times its weight of acetyl chloride for two and one-half hours. After the excess ace tyl c.hloride and acetic anhydride were dis tilled off, the anhydride of cis-l,3-cyclobutanedicarboxylic acid was sublimed at 2- 3 mm. lbath temperature 125-130°). The anhydride crystallized from benzene as long needles; m.p. 131-132° . Anal. Calcd. for C6 H6 0 0 ; C, 57.14; H, 1+.80. Found: C, 5 7.14; H, 4.76 . Hydrolysis of the cis-anhydride by refluxing with di. lute 61 hydroc hloric acid gave the free cis-l,3-cyclobutanedicarboxylic acid (I). After the solution was evaporated to dryness, the cis-acid was crystallized from acetone-benzenej m.p. 131-132 ° . Found: C, 49.79; H, 5.7'}. Methyl cis-l,3-cyclobutanedicarboxylate was prepared by the method of Clinton and Laskowshi l20). A mixture of 5. 0 g. ( O. 034 mole . ) of cis-acid (I), 6. 7 g. ( 0.14 mole ) of methyl a lcohol, 20 ml. of ethylene dichloride, and 0.2 ml. of sulfuric acid was re.fluxed overnight. The resulting liquid was washed with w~ ter , aqueous sodium bicarbonate, and again with water. After the solvent was stripped off, the residue was distilled giving 3.22 g. (54%) of pure methyl ester, b.p. 132.5 ° at 30 rnm. and 0. '.1 0 g. (l) ;; ) of' less pure ester, b.p. 1 25.5-132.5° at 30 mm. Anal. Calcd. for CeH 12 0 4 : C, 55.~0; H, 7.03. Found: C, 55.50; H, 7.23 trans-1,3-Cyclobutanedicarboxylic acid (II). ' The mixed cis- and trans-1,J-cyclobutanedicarboxylic acids (I) and lII) (see above ) were fractionally crystallized twice from acetone a t-2 5° and then twice from dioxane-benzene. By this means pure trans-acid (II) was isolated in approximately 20fa yield; m.p. 189.8- 190.3°. Anal. Calcd. for C6 H8 0 4 : C, 50.00 ; H, 5.60. Found: C, 5 0. L3 ; H, 5 • 73. 62 REFERENCES \1) W.E. Glattfeld and J. Ivl. Schneider, J. Amer. Chem Soc., 6 0 , 415 (19 3 8 ) • (3 ) K. N. Welch, J . Chem. Soc., 257 (1930). M. Gault and A. Roesch, Compt . rend., 199, 613 (1934). ( 4) W. H. Perkin, jun., Ber., 16, 1793 (1883)0 (5 ) C. K. Ingold, E. A. Perren and J . F. Thorpe, J. Chem. (2) Soc., 121, 1765 (1922),; C. K. Ingold and W. J. Powell, ibid., 119, 1976 (1921); W. J , Powell and J. F. Thorpe, ibid ., 75, 52 (l8;i9), (6) W. H. Perkin, jun., and J . L. Simonsen, ibid., 91, 817 (190 7 ). (7) M. Gutzeit and O. Dressel, Ann., 256, 1)19 (1 869). (8) M. J. Schlatter, unpublished. (9) C. A. Kerr, J , Amer. Chem. Soc., 51, 614 \1929) • {10) A. Higson and J . F. Thorpe, J . Chem. Soc., 89, 1458 .. (1906). \11) K. v . Auwe rs and J . F . Thorpe, Ann., 285, 322 (1895 ). (12) A. Mooradian and J. B. Cloke, J . Amer. Chem. Soc., §.1., (13) >142 (1945 ) . s. Fujii, Patent, Japan 1)3;25; Chem. Abstr. l±J., 3447 (1949) (14) M. Beyaert and M. Hansens, Natuurwentenschappelij k tijdschrift, 22, 249 ( l ;i40 ); Chem. Abstr., J]_, 5373 \l.>144 ) . \15) N. Zelinshy and W. Krawetz, Be r ., 46, 1 63 \1713). 63 (16) H. J. Backer and K. J. Keuning, Rec. trav. chim. , .iJ., (17) 812 (l;,134). V. C. Sekera and C. S. Marvel, J. Amer. Chem. Soc., .52_, 345 (l:i33); Org. Syn., iO, 5 0 (1940). (18) K. v. Auwers, Ann., 2J2, 210 (1896). \19) A. Fr2nke and M. Kohn, Monatsh., _gj_, 742 (l;i02). (20) R. O. Clinton ands. C. Laskowshi, J. Amer. Chem. Soc., lQ, 3135 (l)l48). o4 PROPOSITIONS 1. The synthesis of a wide variety of cyclopropane compounds could be easily achieved by the basic condensation of an unsaturated compound, suitably activated (-C0 2 Et, -CN, -N0 2 , etc.), with an Cl( halo compound similarly activated. Cf. part II of this thesis. 2. An acid which was assigned the structure 3-methyl1,2-cyclopropenedicarboxylic acid (I) was reported by Feist. ~I) I propose that Et0 2 C-CH=CH-CH 3 + NaOEt ) Et0 2 9-CHC1 2 (1) could be prepared by the following sequence; CH~ ~ Et0 2 C - C0 2 Et F. Feist, Ber.j 12_, 747 (lbj3); Ann., 60 ~1:106); J.!±2., ibid., i.d6, 125 (1924). (2) C. K. Ingold, J. Chem. Soc., 121, 26 76 (1922). l3) F. R. Goss, C. K. Ingold, and J. F. Thorpe, ibid., 11..3., 327, 3342 (1923); 127, 460 (1:125). (4) 3. G.A.R. Kon and H.R. Nanji, ibid., 2557 ll932). d and 1 trans-1,2-Cyclopropanedicarboxylic acids could be degraded to the optically active d and 1 trans-1,2dibromocyclopropanes by the Hunsdiecker reaction. The latter compounds are of' particular int erest since Ki.rkwood' s Theory of Optical Rotatory Power permits the exact calculation of both the magnitude and sign of their optical rotation. The correspondence between an optically active compound and its 65 physical model would then be unambiguously established for the first time. J . G. 4. Kirkwood, J. Chem. Physics, .2.., 479 (1;137 ). The ring cleavage of' d and 1 trans-1,2-dibromo- cyclopropane (by such agents as hydrogen or hydrogen bromide) could each lead to only one optically active product, and the configuration of the remaining optically active carbon atom in each case would be retained. By relating these open chain products to d and 1 lactic acids the validity of Fisher's convention could be confirmed or refuted. ). (a) I propose that the suffix carbone be adopted as the designation of the =C=O group in ketenes. Examples; propan-2-carbone cyclobutan-l,3-dicarbone (b) The preparation of poly~spiro-cyclobutane hydro- carbons could be effectively achieved through the cross polymerization of cyclobutancarbone with cyclobutan-1,J-dicarbone or spirotJ,J)nonan-2,6-dicarbone followed by a Wolf'f-Kishner or mercaptol reduction. 66 o. I propose that the absolute configuration of an optically active compound such as o(- ethyl- o<. -aminostearic acid (I) ac.id could be experimentally determined by the following method. Saturate a mixture of water and benzene with (I). '.l'he rnmo-layer of lI) which is formed at the inter- face will be orientated so that the pola r ends of the molecules will be in the aqueous phase. Apply an electrostatic field parallel to the interface of {I) thus aligning the dipoles in the direction of the field. The interfacial boundries at the junction of the liquid interface with the walls of the container in the two regions perpendicular to the direction of the field will be different due to the f'ixed orientation of the asymmetric molecules (I). By corre- lating the interfacial tension of various surface active agents in a water-benzene system with the meniscus distort i on described above, the absolute configuration of (I) could be ascertained. 7. One of -che major problems of bio-chemistry is the elucidation of enzymatic specificity. I propos e that cyclo- propane and cyclobutane derivatives are ideally suited for the study of enzyme reactions because these compounds possess rigid molecules of known or calcula ble dimensions. o7 8. Vinylglycine, which has been postulated as a meta- bolic intermediate, could be synthesized by the following series of reactions: hexamethylene- NBS H2 C=CH- CHBr-CN tetramine (1) Org. Syn., Col. Vol. I, 2nd ed., p. 46. (2) C. (3) A. Galat and G. Elion, J.Amer. Chem. Soc., Djerassi, Chem. Rev., .!±J., 271 (l:JJ..J.8). 61, 3585 ( 1 939 ). 9. I propose that chemistry jour·nals be printed on less expensive paper for curr ent use and micro-card copies be issued for per manent files. This would be more economical for all parties concerned and eliminate the enormous storage problem facing libraries and individuals today. 10 . I propose that t h e Chemistry Depart ment offer a year course in the mechanism of organic reactions. The basis for such a course might be E. R. Alexander's book Principles of Ionic Organic Reactions, John Wiley and Sons, New York llY50). 68 11. I propose that a chain comnound could be prepared by the following series of reactions; Br NH Br -r- N Br - LiAlH 4 Na ~ ) 2 steps I I :_ ( CH2 ) 7C0 2 Et - - (CH 2 )?- NCH 3 I I I I Cope and Fenton, Abstrac t of papers presented at the Division of Organic Chemistry, American Chemical Society Me eting , Chicago, Sept., 1950. 69 12. It has not been possible to determine the angle of contact of water on solids such as asphalt because the angle of contact depends upon whether the liquid is contracting or expanding. I propose that this angle of contact could be determined indirectly as follows; A series of' liquids of varying surface tension \preferably pure liquids) which are insoluble in both water and asphalt are tested on the asphalt surface in the order of decreasing surface tension. From the known surface tension_ of' the liquid sample which spreads out completely on the surface and the liquid of slightly higher surface tensj_on which does not spread out completely on the surface, the angle of contact between water and asphalt could be calculated by Dupre's Equation.