Transannular Interactions. IV. Products and Rates of Solvolysis of cis

Transannular Interactions. IV. Products and Rates of Solvolysis of cis- and trans-5-Cyclodecen-1-yl p-Nitrobenzoate in Aqueous Acetone1. Harlan L. Goe...
0 downloads 0 Views 975KB Size
Aug. 20, l O G l

SOLVOLYSIS OF

5-CYCLODECEN-1-YL ~NITROBENZOATES

10% 1,2,3-tris-(2-cyanoetlioxy)-propaneo n Cclite a t 170" showed only a single peak. Anal. Calcd. for CloHleO: C, '78.89; H , 10.60. Found: C, 78.87; H , 10.69. The 2,4-dinitrophenylhydrazone, prepared in the usual manner, melted a t 177.5-179.0" dec. (ethyl acetate-petroleum ether). Anal. Calcd. for Cl&00(: C, 57.82; H, 6.07. Found: C, 58.09; H , 5.79. The semicarbazone nielted a t 210-211' dec. (ethanol); A 9ma; 5 9 LtoH . 229.8 mp, emHX11,100. Anal. Calcd. for CllH18N30: C, 63.13; H , 9.15; N, 20.08. Found: C,63.32; H,9.10; N,20.17. The oxime melted a t 110.0-110.6" (ethanol-water). dlnaZ.Calcd. for CloH1,XO: C, 71.81; H , 10.25. Found: C, 71.90; H, 10.12. cis-5-Cyclodecenone (VIII) was reduced to cis-5-cyclodecenol as follows. A solution of 0.53 g. (3.48 mmoles) of VI11 in 30 ml. of ether was slowly added to a stirred slurry of 0.3 g. ('7.9 mmoles) of lithium aluminum hydride in 35 ml. of ether. When addition was complete, the mixture wds stirred for an hour. After the excess hydride was destroyed with saturated aqueous ammonium chloride, the mixture was extracted with ether in a continuous extractor for 18.5 hr. Evaporation of the ether yielded 0.537 g. (3.48 mmoles) of a solid alcohol, m.p. 43-44'. This product had an infrared spectrum (CCld solution) identical with thdt of authentic czs-5-cyclodecenol. Oxidation of trans-5-Cyclodecen-l-ol.-trans-5-Cyclodecen-1-ol(I1, X = OH) was oxidized by the method described above for the cis isomer. A mixture of uroducts (isomers) was obtained in 71% yield, b.p. 99-103' (15 mm.), nz6D 1.4918-1.4923. Anal. Calcd. for C10H160: C, 78.89; H, 10.60. Found: C, '78.50; H, 10.87. Gas chromatographic analysis of this material with a 6-ft. column of l0Yc 1,2,3-tris-(2-cyanoethoxy)-propaneon Celite ( l i 0 ' ) shoivcd it to consist of a t least three com[ COXTRILlLTION

PROM THE

3511

ponents. The major component comprised about 80% of the total material. This was identified as trans-5-cyclodecenone (VII) by its infrared spectrum. A semicarbazone derivative was prepared by treating the crude ketone with semicarbazide hydrochloride and sodium acetate in aqueous 232.3 ethanol; m.p. 179.5-180.5' (acetonitrile), Xgz mp, emsx 11,000. Anal. Calcd. for CIIHIDNaO:C, 63.13; H, 9.15. Found: C, 63.22; H, 8.94. Attempts to prepare the 2,4dinitrophenylhydrazone and the p-nitrophenylhydrazone failed. Pure trans-5-cyclodecenone (VII) was obtained from the crude ketone (oxidation product) by gas chromatography (1,2,3-tris-(2-cyanoethoxy)-propane on Celite a t 170"), followed by low temperature recrystallization from pentane.lB Capillary gas chromatography (Ucon) showed this material to be over 99%.pure. The infrared and ultraviolet spectra also indicated this material was homogeneous. Anal. Calcd. for CloHleO: C, '78.89; H, 10.60. Found: C, 78.80; H, 10.58. Reduction of 0.521 g. (3.42 mmoles) of the crude ketone with lithium aluminum hydride by the method described for the reduction of the cis isomer VI11 gave an almost quantitative yield of colorless oil. The infrared spectrum of this material was slightly different from that of trans-5cyclcdecenol. The material was dissolved in 15 ml. of olefin-free petroleum ether and shaken with 10 ml. of 10% aqueous silver nitrate. The two phases were separated and the petroleum ether solution evaporated, yielding 0.145 g. of material which was not investigated further. The silver nitrate solution was added to 20 ml. of icecold ammonium hydroxide and the resulting emulsion was extracted with ether. The yield of crude unsaturated alcohol was 0.220 g. (42%). This material was converted t o 0.280 g. of p-nitrobenzoate, m.p. 108.0-109.0°, which was identical with authentic trans-5-cyclodecenyl p-nitrobenzoate. (18) We are indebted to h l r James C. Gross for carrying out this experiment.

DEPARTMENT OF CHEMISTRY, THEUXIVERSITY

OF

\VISCOXSIN, MADISON 6 , WIS.]

Transannular Interactions. IV. Products and Rates of Solvolysis of cis- and frans-5-Cyclodecen-1-yl @-Nitrobenzoatein Aqueous Acetone BY HARLAN L. GOERINCAND WILLIAMD. CLOSSON~ RECEIVED MARCH20, 19G1 Traiisaiiiiular part icipation by the double bond is involved in the solvolysis of cis- (Ia) and trans-5-cyclodccen-1-yl p Iiitrobenzoate (11) ili aqueous acetone. In each case the first-order rate is enhanced (dkyl-oxygen cleavage is involved) and only bicyclic products are formed. At 120' the trans-p-nitrobenzoate I1 is about 300 times more reactive than the cis isomer Ia which in turn is more reactive than the saturated analog, cyclodecyl p-nitrobenzoate. The cis isomer Ia is converted to cis-cis-1-decalol(V). Solvolysis of the trans-p-nitrobenzoate I1 is accompanied by an isomeric rearrangement (ca. 18yo)to an unreactive product. The solvolysis product is trans-trails-I-decalol ( I V ) and the rearrangement product is mainly trans-cis-1-decalyl p-nitrobenzoate (111). The rearrangement product I11 formed from carbonyl-O'*--trans-pnitrobcnzoate I1 has two-thirds of the label in the carbonyl position.

d n investigation of the solvolysis of cis-5-cycloOzCCsHdNOZ decen- 1-yl p-toluenesulfonate (Ib) was reported in an earlier paper.3 I n that work it was found that I1 a t 30' the unsaturated cis-p-toluenesulfonate I b Ia, X = OzCCsHa~02 ethanolizes ten times faster and acetolyzes seven b, X = OTs times faster than the saturated analog cyclodecyl c, X = OH p-toluenesulfonate. From the solvolytic behavior and the formation of bicyclic solvolytic and reTransannular interactions involving carbon-cararrangement products, it was concluded that ioniza- bon double bonds have been observed in othcr tion probably involves "transannular''~ participa- medium-sized ring systems. Pertinent examples tion by the double bond. (4) Apparently the term "transannular" has not been define ) A . Aebi. D. H. R .H a r t o n a n d A . S Lindsey, .T. C ' / I C V I . Soc., 3 1 2 % k , / k , , can be determined readily from the observed and calculated "infinity" titers (;.e., kr;'k, is the A . S i c k o n , J . .4iii. C h e w 4oc.. 7 7 , 1I'JO (lU.7.7). difference between the calculated and observed ( 7 ) L). H. R. Barton, 0 . C. Brjckman and P. de brayo, J . Chcm. titers divided by the observed titer). Values of .So(., 2269 (1960). 23) A. C . Cope and P. E . Peterson, J . A l i i . Cheiii. .So( ., 81, lti43 k,,'k, for various conditions are given in Table 11. (1959); A . C Cope, J. hZ. Grisar a n d P. E. Peterson, i b i d . , 8%, 4299 These were determined from the calculated titers i 1960) and those observed a t about ten half-periods. (9) H. I,. Goering, A C. Olson a n d H. €3. E s p y , ibiJ , 78, 6371 Each value is an average (anti average deviatiol~) (1956). (10) (a) J. B. Hendrickson, Tetrahedron, 7 , 82 (1439); (h) D. H. It. of two independent experiinerits. Barton and P. d e Mal-o, ( l n o r l . Rms., 11, 189 (I9J7); L. Ruzicka, The data show that kr'k, is not affected by the Erpcuirntia, 9 , 3.57 (1933). presence of sodium p-nitrobenzoate or p-nitrobeii(11) H. L. Goering, W. D. Clossun aud A . C. Olson, J . AJU. Chcin. zoic acid. The ratio decreases slightly with teniSOL.,89, 3507 (1961). i . c j . ,

+

Aug. 20, 1961

SOLVOLYSIS OF 3-CYCLODECENl - n p-NITROBENZOATES

3513

Thus k for these compounds is a conservative upper limit for the carbonium-ion process (alkyl-oxygen cleavage). Solvolysis Products.-The products resulting Temp., Added solute, from the-simultaneous solvolysisand rearrangemen'i "C. 102 1M kr/ksa of 0.03 M truns-5-cyclodecen-l-yl p-nitrobenzoate 90% aqueous acetone (11) in 90% acetone a t 100" were isolated after 118.6 0.226 0.001 twelve half-periods. Vapor phase chromatography 99.6 ,203 ,002 (v.p.c.) indicated that a volatile product, which .224 i ,001 118.6 2 . 4 7 SaOPNBb was not identified, is formed in 10-20% yield. 99.6 2 . 4 7 SaOPNBb .I91 i. ,009 The product was separated into an alcohol (60118.6 2 . 5 4 HOPNB' .235 f .001 69% yield) and a p-nitrobenzoate (15-18% yield) 99.6 2.54 I-IOPNB" .205 f ,006 fraction by either column chromatography or sub80% aqueous acetone limation. The alcohol fraction was identified as 118.6 0.170 i 0.001 trans-trans-l-decal01 (IV) by comparison of the in92.7 0.164 4 0.001 frared spectrum and p-nitrobenzoate derivative Average value and average deviation of two independent with those of an authentic sample. Vapor phase experiments. Sodium p-nitrobenzoate. P-Nitrobenzoic chromatography showed that this fraction conacid. tained l-2Y0 of a contaminant that had the same perature (rearrangement is more sensitive to tem- retention time as trans-cis-l-decalol. perature changes than solvolysis) and is lower for Recrystallization of the p-nitrobenzoate fraction 80% acetone than for 90% acetone. (the rearrangement product) gave pure trans-cis-lA discrepancy between the observed and cal- decalyl p-nitrobenzoate (111) which was identified culated infinity titers would result from contamina- by comparison with an authentic sample. Saponition of the reactant with an unreactive isomer as fication of the p-nitrobenzoate fraction gave a well as from formation of such an isomer during binary mixture of alcohols consisting of 85% transsolvolysis. In the present case the unreactive iso- cis-1-decalol (v.p.c.) and 15% of a material which mer is formed during the solvolysis. This was dem- was not identified but had the same retention time onstrated by showing that when I1 is contami- as trans-trans-l-decal01 (IV). nated with isotopically labeled (OIa) trans-cis-l-deControl experiments showed that from 93-9570 calyl p-nitrobenzoate (1111,which is the major com- of the alcohol and 97-99% of the p-nitrobenzoate ponent of the unreactive rearrangement product, are isolated by the techniques used in the product the contaminant is completely removed by the studies. It was also found that ester interchange methodll (recrystallization) used to purify I1 for the does not occur under the conditions of the solvolysis kinetic experiments. Also, the fact that k,/k; is reaction. solvent and temperature dependent and reproducible The solvolysis product of cis-5-cyclodecen-l-y1 pshows that the unreactive isomer is formed during nitrobenzoate (Ia) was also isolated. In this case a solvolysis. solution of Ia in 90% aqueous acetone was Rate constants for the solvolysis of cis-5-cyclode- 0.03 cen-l-yl p-nitrobenzoate (Ia)," a,y-dimethylallyl heated to 120' for 545 hours. This corresponds to p-nitrobenzoate, l 2 truns-cis-l-decalyl p-nitrobenzo- about 3670 reaction. The product (alcohol) and ate (111) and cyclodecyl p-nitrobenzoate in 90% unchanged +-nitrobenzoate were isolated and aqueous acetone a t 120' are included in Table I. separated in the manner described above; 65y0of I n the case of a, y-dimethylallyl p-nitrobenzoate the the original I a was recovered. The solvolysis prodreaction was followed to over 80% completion; no uct obtained in 33% yield (based on the original trends were observed and the infinity titer was in amount of Ia) was identified as cis-cis-l-decalol (V) by comparison of v.p.c. retention time, the phenylgood agreement with the calculated value. The kinetic experiments with cis-5-cyclodecen-1- urethan derivative and infrared spectrum with vl p-nitrobenzoate (Ia) were complicated by the those of an authentic sample. Vapor phase chroiow solubility and reactivity in 90% aqueous ace- matography indicated that this fraction was homotone. The measurements were made with 0.01 M geneous. When cis-3-cyclodecen-1-01 (IC) was heated in solutions and the reactions were followed to 30-50% 905% aqueous acetone under the conditions of the completion. The rate constants, computed using the calculated infinity titers, did not show any solvolysis reaction it was recovered unchanged. trends. Since pure cis p-nitrobenzoate Ia can be Thus IC is not a solvolysis product and evidently isolated in good yields after partial solvolysis, it cis-cis-l-decalol (V) is the initially formed product. appears that solvolysis is not accompanied by It is apparent from the product that the solvolysis isomerization. The solvolyses of cyclodecyl and of I a involves alkyl-oxygen cleavage. Rearrangement of Carbonyl-0 truns-5-Cyclotrans-cis-l-decalyl p-nitrobenzoates (111) were so slow that only approximate values for the first- decen-l-yl p-Nitrobenzoate (II).-To obtain addiorder constants were obtained. These were cal- tional information about the rearrangement of I1 culated from the small amounts (a few per cent.) to truns-cis-decalyl p-nitrobenzoate (111) which of acid produced after long periods of time. In the occurs during solvolysis, the relative positions of latter two cases the reaction probably involves the carboxyl oxygen atoms in the reactant I1 and acyl-oxygen rather than alkyl-oxygen cleavage. product I11 were established. Carbonyl-Ol*-II was solvolyzed in 90% acetone a t 100" and the (12) H. L. Goering a n d h i . hi. Pombo, J . A M . Chcln. Soc., 82, 2515 (lUCj0). major rearrangement product, 111, was isolated TABLE IT RATIOO F :REARK4NGEMENT ( k , ) TO SOLVOLYSIS ( k a ) FOR 0.02 M trUnS-5-CYCLODECEN-l-YL @-NITROBESZOATE IN AQUEOUSACETONE

* *

HARLANL. GOERINCAND WILLIAMD. CLOSSON

3514

VOl. 83

after ten half-periods. The distribution of OL8 in I11 (and also in IT) was determined by saponification (acyl-oxygen cleavage), followed by reconversion to the p-nitrobenzoate with unlabeled p-nitrobenzoyl chloride. The 0 I 8 content of the ester derived in this way corresponds to that of the ether-oxygen position of the original ester. The difference in 0 ' 8 content before and after saponification and re-esterification corresponds to the 0 I 8 content of the carbonyl position of the original ester. The results of the 0l8experiments are summarized in Table 111. Experiment 1 shows the OI8 content of the reactant 11. In the second experiment unreacted TI was isolated after one half-life to determine if any scrambling occurs prior to reaction. Clearly there is little if any mixing of oxygen atoms in I1 during solvolysis. In experiment 3 the reaction was carried out in the presence of 0.05 -11sodium p-nitrobenzoate. This experiment, in complete agreement with the kinetic results (i.e., k r / k s is not affected by p-nitrobenzoate ion), shows that the rearrangement is intramolecular and that none of the label is lost under the conditions of the rearrangement. Experiments 4 and 5 are duplicate experiments and show the Ols distribution in IT1 derived from carbonyl-018 I1 (3.95 atom 7 0 excess OIg). These experiments show that the 111 has about one-third of the OI8 in the alkyl-oxygen position.

those of p-nitrobenzoates of saturated secoiidary alcohols. Presumably the unassisted ionization rates for l a and TI would be substantially lower than the rate of solvolysis of cyclodecyl p-nitrobenzoate. Thus the present data indicate that the ionization rates of Ia and I1 arc ncce!erated by factors of over 5 and 103, respectively. For several reasons these estimates are lower limits. In the first place, for reasons that will be presented in the Experimental section, the rate constant for cyclodecyl p-nitrobenzoate in Table I is a rough upper limit and in fact may be too high. Secondly, solvolysis of cyclodecyl p-nitrobenzoate may well involve acyl-oxygen cleavage in which case the solvolysis rate exceeds the ionization rate. Finally, the unassisted ionization rate for Ia and I1 would be expected to be slower than that €or the cyclodecyl system. Introduction of a double bond in a ten-membered ring should relieve "medium ring" strain which presumably is responsible for the relatively high reactivity of the cyclodecyl systemcyclodecyl p-toluenesulfonate acetolyzes about 500 times faster than cyclohexyl p-toluenesulfonatc. ' The inductive effect of the double bond would also tend t 3 lower the unassisted ionization rate soniewhat. In connection with the large driving force observed with the tram-p-nitrobenzoate I1 i t is iiiteresting to note that the ultraviolet spectrum of the corresponding ketone, trans-5-cyclodecenoiie,' TABLE I11 contains a solvent-dependent band (A, 2 14.5 OxYcm-18 DATA FOR COSVERSIOSOF CARUONYL-O'~ mp, E 2300, in 2,2,3,3-tetrafluoropropanol)which ti.UlLS-5-CYCLODECEN-1-YL p-NITROBENZOATE (11) TO & r U l l S has been assigned to the photodesmotic transition':

(111) &-~-DECALYL p-NITROBENZOATE ACETONE AT 99.6" Expt.

Compound

Position

IN

90% AQUEOUS

Atom

70 excess

0'8

1

11" Carbonyl-0 3.98 =t0.04 p I1 illkyl-0 0.020 f ,012 3" I11 Total 3.99 f .06 4 I11 Alkyl-0 1.37 f .01 5 I11 Alkyl-0 1.33 f . O 1 a Starting material. The material was recovered from the solvolysis mixture after one half-life a t 99.6'. The experiment was carried out in the presence of 0.05 A l sodium p-nitrobenzoate.

Discussion The data in Table I show that in OOyoaqueous acetone a t 120' trans-5-cyclodecen-1-yl p-nitrobenzoate (11) solvolyzes over 300 times faster than the cis isomer I a which in turn is about 5 times inore reactive than the saturated analog, cyclodecyl p-nitrobenzoate. It is interesting to note that the trans-5-cyclodecen-1-yl system is even more reactive than the e,y-dimethylallyl system by a factor of over ten. It was shown earlier that solvolysis of a,y-dimethj.lallyl p-nitroberizoate in aqueous acetone involves alkyl-oxygen cleavage. l 2 < l 3 The high reactivities of Ia and I1 relative to cyclodecyl p-nitrobenzoate indicate that their rates of ionization (solvolysis) are accelerated by transannular participation by the double bond as illustrated for the two isomers by 11' and Ia'. In the absence of participation hydrolysis by acyloxygen cleavage would probably take over and the rates would be expected to be coinparable with (13) €I. L. Goering and R . W. Greiner, J . A m . Chcm. Soc., 79, 3464 (1957).

b.:.6 c-.

@

The ultraviolet spectrum of cis-3-cyclodeceiic7ii~' does not contain this band.16 The possibility of a correlation between absorption resulting lrom such interactions and anchimeric acceleration observed with derivatives of the corresponding alcohols has both the spectra been noted p r e ~ i o u s l y . ~Thus ~ of the isomeric ketones and the relative anchimeric accelerations observed with the isomeric p-nitrobenzoates (Ia and 11) indicate that stereoelectronic factors are more favorable for transannular interactions between the double bond and Cl in thc trans-5-cyclodecen-1-yl system than in the cis isomer. Solvolysis of cis-(Ia) and trans-5-cyclodecen-1-yl p-nitrobenzoate (11) results in bond forination between C1 and & a s illustrated by 11' and Ia'. I t is interesting that these cyclizations are stereospecific. The trans isomer I1 gives the transdecalyl system and the cis isomer Ia gives the cisdecalyl system. l8 (14) R. Heck and V. l'relog, lfelu. Chim .4cliz, 38, 1 5 i l ( l U , i . j ) (15) E. M. Kosower, W. D Closson. 11. L. Goering and J C. Gross, J . A m . Chem. Soc., 8 3 , 2013 (l9G1). (16) W. D. Closson, unpublished results. (17) For leading references see S. Winstein, L. de Vries and K. Orloski, J . A m . Chem. Soc., 83, 2020 (1961). tl1.il (18) I t is interesting t h a t this is t h e same stereochemi;try involvcd in (a) t h e transannular cyclization of pyrcthrusin to cycI,,pyrethrosin7 (conversion of a cis-cyclodecenc system t o a cis-clcciilin derivative) a n d (b) t h e proposed biogenesis of eudesmol*ua(transuu-

Aug. 20, 1961

SOLVOLYSIS OF 5-CYCLODECEN- 1-YL

p-NITROBENZOATES

3515

In each case, solvolysis involves a trans addition to the rate of ionization. Xnternal return to give of C1 and solvent across the Cs,C6-doublebond. On rearrangement product III results in some oxygen the other hand, the isomeric rearrangement of II to mixing and if only one intimate ion pair intermedi111 which accompanies solvolysis amounts to a cis ate is involved in the ionization-dissociation procaddition of C1 and the migrating p-nitrobenzoate ess, internal return to reactant I1 should also regroup across the double bond. The isomeric re- sult in mixing of the oxygen atoms. Internal rearrangement involved in the solvolysis of the cis-p- turn has been observed to result in equilibration of toluenesulfonate I b is stereochemically analogous carboxyl oxygen atoms during the solvolysis of to the conversion of I1 to 111; I b isomerizes to cis- allylic p-nitrobenzoates12f21s22a and benzhydryl ptrans-1-decalyl p-toluenesulfonate during solvolysis nitrobenzoate.22b However, more than one inter(ethanolysis and acetolysis). nal ion pair intermediate may be involved and thus the absence of oxygen equilibration during H H solvolysis does not rule out the possibility of internal return. One feature of the stereochemistry of the transannular cyclizations of cis- and trans-5-cyclodecenI-yl derivatives remains obscure. This concerns H 11’ H the stereochemical change at C1. If the reactive I 111, Ar = p-C6H1 ~ ~ , conformations of the reactants are those illustrated H by 11’ and Ia’, displacement of the leaving group a t c1 by c6 proceeds with inversion of configuration. This arrangement, which allows for nucleophilic attack by the double bond from the rear, seems to be the best one for accounting for the rather large driving force-anchimerically accelerated ionizations evidently invariably involve back-side internal displacements. However, there is another possibility. This is illustrated for the trans isomer by 11”. If this conformation is the one involved in the cyclization, C1 is substituted (by C,) with retention of configuration. Although there are many precedents for intramolecular front-side displacements, e.g., SNi reactions,24these usually occur when there is no Evidciitly the rearrangemcnt of I1 to I11 involves alternative. Moreover, these probably often, if ”ion pair return,” probably “internal return.”1g not generally, involve internal return from ion pair That the process is ionic is indicated by the fact intermediates in which case front-side attack is not the rate shows about the same sensitivity toward simultaneous with rupture of the original bond. varying the ionizing power of the solvent as does It is not clear to what extent, if any, front-side thc rate of solvolysis (as); cf. k r / k s for 80% and 90% nucleophilic attack can facilitate departure of a acetone. I t is clear that the rearrangement is in- leaving group. Thus one is left with the dilemma tramolecular because k r / k s is not affected by the that conformation 11’ nicely accounts for the acaddition of p-nitrobenzoate ion and the OI8 experi- celerated ionization, but makes i t difficult to rationinents show that rearrangement does not result in alize the migration of the ester group from C, to C5 exchange with either p-nitrobenzoic acid or p-nitro- without complete equilibration of the carboxyl benzoate ion. The fact that rearrangement results oxygen atoms. On the other hand, 11” can nicely and 0l8results of in only partial equilibration of the oxygen atoms account for the ~tereocheinistry~~ analogous conformation (the carbonyl oxygen atom retains its identity in the rearrangement-the part) suggests that the process involves internal was used to rationalize the isomerization of I b to it is return. Return from a “solvent separated”19 ion cis-trans-1-decalyl p-toluenesulfonate3-but pair is believed to result in complete equilibration not clear if such an arrangement can accommodate of the oxygen atoms in the anion portion of the the rate enhancement. I n another problem we are molecule.20 On the other hand, cases are known in attempting to establish the relative configurations which internal return results in no or only partial of C1 in the reactant I1 and C,, in the solvolysis IV equilibration of the oxygen atoms in a p-nitro- and rearrangement products I11 to settle this point. The large driving force indicates that ionization benzoate. 1 2 ~ 2 0 ~ 2 1 The observation that carbonyl-O1* trans-5- of the trans-p-nitrobenzoate I1 results in the direct cyclodecen-1-yl p-nitrobenzoate (11) remains dis(22) (a) H. L. Goering and J. T. Doi, J. A m . Chem. Soc., 8 2 , 6850 cretely labeled throughout the solvolysis indicates (1960); (b) J. F. Levy, unpublished results. that the reactant is not reformed by internal return, (23) For leading references see (a) A. Streitwieser, Jr., Chem. Rem., i.e., i t appears that the rate of solvolysis corresponds 66, 571 (1956), and (b) D. J. Cram, “Steric Effects in Organic Chemis-

CPCOC,H,NO,

nular cyclization involving a rvans double bond t o give a Irons-decalin system). (19) S. Winstein, E. Clippinger, A. H . Fainberg, R . Heck and G. C. Kotiinson, J . A m . Chem. Sac., 1 8 , 328 (19.56). ( 2 0 ) S. Winstein and G . C . Robinson, i b i d . , 80, 169 (1958). (21) 31. h l . Pombo and E;. D. hlcMichael, unpublished work.

try,” M. S. Newman, ed., John Wiley and Sons, Inc., New York, N. Y.,1956, Chapt. 5. (24) E. L. Eliel, ibid., Chapt. 2. (25) The tro,rs-l-decalyl isomer resulting from an intramolecular isomeric rearrangement of either 11’ or 11” would be expected to give the observed product -11, i . e . , if the reaction is intramolecular the product resulting from cis addition wuuld be expected in any evcnt.

3516

HARLAN L. GOERING AND WILLIAMD. CLOSSON

h

/

H

11"

formation of an ion pair in which the cation is the hans-1-decalyl carbonium ion VI or the non-classical ion VII. Similarly, the initially formed cation from the cis-p-nitrobenzoate is either the classical "rearranged" ion VI11 or the bridged non-classical ion IX. The anchimeric accelerations and probably also the stereochemistry are explicable in terms of either the classical or non-classical ions.

H

VI

€!VI1 I

I;I

H

I t is of interest in this connection that nitrous acid deaminations of trans-trans- and cis-cis-l-decalylamine give the corresponding alcohols as the only substitution products,26i.e., in both cases there is complete retention of configuration. Presumably these reactions involve the same carbonium ions involved in the solvolysis-product forming step in the present work. The results of the deamination reactions have been interpreted in terms of the classical carbonium ions VI and VIII.2' The reason proposed27for the retention of configuration was that solvent attacks the carbonium ion along the path of easiest approach, the approach to form equatorial alcohol being less hindered than that t o If this interpretation is corform axial rect the classical carbonium ions can similarly accommodate the present results. On the other hand, the non-classical structures for the cis-(IX) and tvans-1-decalyl carbonium ions (VII) account for the products derived from these ions in an obvious way. Indeed, intermediates of this type (i.c., bridged ions) are generally used to rationalize stereospecific trans addition reactioris (which are observed with both isomers in the present case) and replacement reactions which involve retention of c o n f i g u r a t i ~ n . ~I n~ other problems we are investigating the structures of bicyclic carbonium ions of this type.

Experimental Material%-The cis-(Ia) and trans-5-cyclodecen-1-yl P-nitrobenzoates (11) used in the present work are described in the preceding paper." Cyclodecyl p-nitrobenzoate, m.p. 114.5-115.0" (ethanol) (lit.28m.p. lie'), was prepared" (2fi) W. G. D a u b e n , R . C 'l'writ and C. hlannPrsknniz. .l. ( ' h e m . Soc., 76, 4420 (1954). (27) A. Streitwieser, J r . , J . O r & . (,'hr~ii, 2 2 , X l i l ( I ! K ) i ) .

:liii.

Vol. 83

from cyclodecanolg in Y3cl/b yield. tiuns-cis-1-Decalyl p nitrobenzoate (III), m.p. 116.0-116.5' (aqueous methanol) (lit.26m.p. 116O), w?s prepared" from trans-cis-l-decalol,3 m.p. 48.0-48.5", in 727, yield. a,y-Dimethylallyl-pnitrobenzoate, m.p. 53-54' (lit.Z9 m.p. 5 4 O ) , was prepared from the pure alcohol1* in 32% yield. p-Nitrobenzoic acid, m.p. 240.3-240.9', was purified by recrystallization from methanol. Sodium p-nitrobenzoate was prepared by shaking a solution of 0.119 mole of p-nitrobenzoic acid in 200 ml. of ether with 50 ml. of 2.0 M aqueous sodium hydroxide. The aqueous layer containing the salt was extracted with ether in a continuous extractor t o remove any traces of excess acid and then evaporated until the salt crystallized. The salt was separated by filtration and dried to constant weight at 140" under reduced pressure. An aqueous solution of this salt had a PH of about 8. Aqueous acetone was prepared by mixing pure acetone (fractionated from calcium chloride) and conductivitz water. The pure components were equilibrated at 25 prior t o mixing. Solvent compositions are based on the volumes of the pure components a t 25". Kinetic Experiments.-Standard solutions of the substrates in aqueous acetone were prepared and distributed into ampules. Ester concentrations of 0.020 M t o 0.025 i l l were used except in the case of cis-5-cyclodecen-I-y1 p-nitrobenzoate (Ia) where the limited solubility necessitated lower concentrations. All concentrations and volumes were measured at 25 '. Aqueous acetone solutions develop an acid titer slowly when heated in the presence of oxygen at high temperat u r e ~ . For ~ ~ example, the titer of a 0.01 -Ti' solution of pure p-nitrobenzoic acid in 90y0 acetone increased about 357, in 200 hr. a t 119' and then decreased slowly. The rate of production of acid by this process was too slow t o affect the rate studies of the more reactive esters, e.g., a,?-dimethylallyl p-nitrobenzoate and I1 which a t 119' have half-periods of 29 and 2.5 hr., respectively. However, it is obvious that it would seriously interfere with the less reactive esters. It was found that the acid-producing side reaction could be essentially eliminated by flushing the ampules several times with nitrogen before sealing them.31 Titers of 0.01 M solutions of p-nitrobenzoic acid under nitrogen in ampules increased less than 1% when heated for 333 hr. at 119'. I n all of the kinetic experiments, except those with the aforementioned reactive esters, ampules were flushed with nitrogen prior t o sealing. Even so, it is clear that the acid-producing side reaction may be significant in experiments with cyclodecyl and trans-cis-ldecalyl p-nitrobenzoate (111), which at 119' react only t o the extents of 7.4y0 in 432 hr. and 2.4y0in 528 hr., respectively. Because of this complication the approximate constants in Table I for these two esters may be too high and thus should be considered t o be rough upper limits. The rate of acid formation was measured by potentiometric titration of aliquots with standard 0.025 Jd aqueous sodium hydroxide. X Beckman model G pH meter w a s used for the potential measurements. The first-order rate constants presented in Table I were calculated from the rate of acid formation in the usual manner. Standard solutions of the trans-p-nitrobenzoate I 1 uiiticr nitrogen in sealed ampules were heated for 10-12 half-lives to obtain the data presented in Table 11. Solvolysis Products. ( A ) trans-5-Cyclodecen-1-yl p Nitrobenzoate (II).-A solution of 1A00 g. (4.61 mmoles) of I1 in 135 ml. of 90% aqueous acetone in a glass bomb was heated a t 99.6" for 188 hr. ( ( a . 12 half-lives). The contents were diluted with 750 ml. of water and made basic t o phenolphthalein with dilute sodium carbonate solutioi; and the resulting solution was extracted with 40-60 petroleum ether in a continuous extractor for 71 hr. Gas chromatographic analysis of the petroleum ether solution with a 12-foot column of 20'34 saturated silver nitrate solution in polyethylene glycol (Carbowax 400) on Celite at 80" indicated the presence of olefin in amounts corresponding to 10-207, yield. The solution was concentrated and I___._

..-

(28) hf. Kobelt, P. Barman. 1'. Prelog a n d I,, K i i z i c k a , I l r l a . Chiin . d ( t a , sa, z ~ f (i 1 ~ 4 4 ) . (2!3) 11. If, P o m h o . Ph n 'l'hesis. Ilniveriity { b f Wiscnn.in. I O l i O (:30)'This ha5 also heen noted by h l . Silver, private cmnmunicnli