cis-and tra~zs-3,5-Dimethyl-3-carbo11~ethoxy-A~-pyraoline (IV and V) have been separated, and , . product analysis for the pyrolysis and photolysis of these compounds has been completed.I he cyclopropane products are formed with some change in the relative coniig~iration of the cis-trans groups in both reactions. 'The a,@-unsaturated olefir1 prod~rct is formed stereospecifically with IV, yielding methyl 2-methyl-cis-2-pentelloate (XI), and with V, yielding methyl 2-methyl-tra?zs-2-pentenoate 0 ; ) . The results of the pyrolysis and photolysis of cis,tra?ts-3,4,5-trimethyl-3-carbornethoxy-A1-pyrazoline are also described. T h e product s t~~d i e s and some kinetic s t~~d i e s are the basis of a new ~nechanisrn proposed for the pyrolysis of pyrazolines.Recent studies on the pyrolysis of pyrazolines have reopened the question regarding the stereocheinistry of the conversion of pyrazolines into cyclopropanes (1-3). This reactio~l was reported by von Auwers (4) to be essentially stereospecific and this was the accepted view on the reaction for many years (stereospecific here is used t o mean t h a t the substituents on the cyclopropane ring have the same cis-trans relationship as on the initial pyrazoline). \Ian Aulcen and Rinehart (1) showed that the pyrolysis of a pair of isomeric pyrazolines with inethyls cis and trans on Cg and Cq (c0111pounds I and 11) gave cyclopropanes with effective randoinization of this cis-trans relationship. On the other hand, Overberger and Anseline (2) have found that the pyrolysis of trans-3,5-diphenyl-a'-pyrazoline (111) gave stereospecifically the trans-1,2-diphenyl~~rclopropane, and thus the stereocheinistry of this reaction is not clear.\lie have prepared some new pyrazolines, \vhich shoxv a cis-trans relationship a t positions Cs and Cj, in an effort to find out more about the pyrolysis reaction. T h e preparation of a cis-trans mixture of 3,s-dimethyl-3-carboinethoxy-A1-pyrazoline and its pyrolysis was reported earlier ( 5 ) . This mixture has now been separated by distillation through a spinning-band coluilln a t a temperature slightly below the pyrolysis temperature. T h e lower-boiling isoiner has been assigned the structure with the methyl groups cis (I\/) and the higher-boiling isoiner t h a t ~v i t h the methj.1 groups trans (V). T h e preparation and ex m, x = c o , c~, 11, X = COCH, N +

The pyrolysis of 3-1nethyl-3-acetyl-4~-pyrazoline (IV) and cis-and trans-3,s-dimethyl-3-a~etyl-4~-pyrazoline (V and VI) has been studied. Pyrolysis cf IV gave 2,3-dimethyl-4,5-dihydrofuran (XI) as a ~ninor product, indicating sollie ring closure to oxygen in these reactions. Pyrolysis of V gave 2,3,5-tri111ethyl-4,5-dihydrofuran (XVII) as 23y0 of the product, whereas VI yielded only a trace of S V I I . This fact and other differences in the product analysis from isomers V and VI are discussed in terms of the possible mechanism fcr the pyrolysis reaction.

A number of 4,4-diall~yl-3-cya11o-3-carbomet1i0xy-~-pyra~oli1es have been synthesized.Pyrolysis of these pyrazolines yieldcd cyclopropane and olelin products. The formation of olefin products by rearrangement of an allql group from Cq to C: of the pyrazoline system suggests that positive charge is developed on C: ill the transition state. Ionic character in the transition state was also indicated by the fact that the rate of pyrolysis is faster in polar than in nonpolar solvents.A mechanism xi~lrhich has been considered by a number of workers for the pj'rolysis of 3-carbalkox).-A'-pyrazolines (I) and related systems has followed a forin similar to the sche~ne shoxvn in eq.[a] (1, 2). The products consist of cyclopropa~le derivatives and a,P-and P,r-unsaturated olefins. The ring-opened intermediate I1 has been considered, since rotation around the bonds in I1 would explain the experimental fact (2) that cyclopropane products in which the stereocl~en~istry in the cyclopropane is different froin that in the pyrazoline are formed. As a starting point for further study me have therefore assumed I1 to closely represent the intemlediate or transition state for this reaction. In this paper and the next paper in this series (3) we have looked for rearrangements which are characteristic of ionic compounds and xvhic11 might be expected for intermediates with the ionic character shown in 11.Structure I1 suggests that positive charge is built up a t C 5 of the pyrazoline systein in the transition state. If this is the case, then rearrangeinents characteristic of carboniu~n ions should be possible for this system. Wagner-AIeerwein rearrangements and related reactions clearly demonstrate that alkyl groups can readily migrate to a positive center. Hoxvever, the migration of allcyl groups to a radical center or an anion center are predicted to be especially unfavorable (4). The occurrence of allcyl migration xvould be diagnostic of positive charge a t C j in the transitioil state for the pyrolysis reaction.In the pyrolysis of the inany pyrazolines to date, there has been no observation of rearrangement of an allcyl or aryl group from C4 of the pyrazoline ring to C5 (see eq. [b]). This has been due, presumably, to the facile conlpetitive lnigration of hydrogen in those pyrazolines analogous to compound 111. T o observe alkyl migration we therefore need two alkyl groups a t C4.The pyrazoline from ~nesityl oxide is known ( 5 , 6); however, heating this A?-pyrazoline does not give an identifiable procluct. Attenlpts in this laboratory to prepare the correspondiilg methyl ester by reaction of diazomethane and methyl senecioate were unsuccessful. The presence of the t x o 6-methyl groups on the acrylate systeln sterically hinders the system to such an extent that no reaction occurred when the reaction was carried out in a sealed tube for several days. Successful preparation of a series of 4,4-dialli~l-A'-pyrazolines was achieved by the For personal use only.

Irradiatiorl of a n ether solutioll of the title compounds with a medium-press~~re luercury arc gave in each case a renrral~gement reaction, with formation of a n acetyl cyclopropane isomer as the major product. This rcactio11 is similar to that induced thermally a t GO0.2-~~Ietl1yl-4,5-clil~~~drof~irans are formed as minor procl~~cts in the pyrolysis and photoljrsis of 3-acet)rl-A'-pyrazolines (1). These derivatives \\lere found to be stable to light filtered by a pyrex vessel; lie\\-ever, in a silica-glass flask they were rearranged ~i n d e r irradiation by a mercur)-lamp to cyclopropane derivatives.When 2,3-dirnethyl-4,5-dih-)rdrof~1rm (1) (A,,,:,, = 218 111~) \\;as irradiatecl in an ether solution, the n-ork-up of the reaction after 86% of I had reacted gave a mixture which contained 1-methyl-1-acet~~lcyclopropane (11) as the major component. Other products identified only by their retention tiilles in the vapor chromatogram \\.ere trans-3-methyl-3-buten-2-one (111), cis-3-metI~y1-3-bute1~-2-one (IV), and 3-methyl-4-buten-2-one (V) (see eq.[a]). Similarly, photolysis of 2-meth)-I-4,5-dihydrofuran (A,,,!,, = 210 inp) gave methyl cyclopropyl l;etone. This latter reaction is complicated by extensive polymerization.T h e photorearrang-eiileit of dihydrofurans is analogous t o the thernially induced rearrangement (2-4). For example, 2-inethyl-4,S-dihyclrof~1ran, \\;hell heated to 450" in a flo\\, system, is convertecl t o n~e t h y l cj-clopropyl ketone and 3-b~iten-2!-one (2, 3).I t is of interest that the product ratio for compounds 11-V in the photorearrange~l~erlt of I (90:5:3:2 respectively) is close to that obtaiilecl for the photolysis of 3-acetyl-3-n~etl~yl-A'-p~~razolii~e (88:6:3:3 respectively) (1). Although this may be fort~iitous, it inay nlean that a similar internlediate is involvecl in the t\\ro reactions.T h e photolysis of vinyl ethers in the gas phase has been found to give t\vo prinlary processes, the first caused by cleavage of the allcyl group -oxygen bond and the second caused by rearrangement and frag~llentation to give an aldehyde and c t h~~l e n e (5) (see ecl.[O]). In the dihydrofuran systenl analogous cleavage \\~ould give a diradical intermediate \\:hich, in a dipolar form, is similar t o that proposed for the pj.rolysis of pyrazolines (6) (see eq.[c]). I-Io\\rever, the possibility of a illolecular mechanisin must also be considered, since stereochcmical results in the pyrolysis and photo1~-sis of pyrazolines suggested a molecular mechanism is possible (7). Boykin and Lutz (8) have observed that the photorearrangement of 2,3,5,;5-tetraphenyl-4,5-dil1ydrofuran gi\:es the valence tautonler l-benzoyl-l,2,2-tripl1ei1ylcyclo-propane (these results were published \\rhile the present paper was in press). 'Their systeill is interesting in that the dihydrofuran can be reforined from the c)rclopropane by treatnlent \\:it11 acid, and more drastic conditions are needed to give ring-opened products. This inight suggest also that ring opening is not involved in the dihydrofuran-cyclopropane tauto...