The Curtius Rearrangement. III. The Decomposition of Substituted

Kinetic study of the Curtius rearrangement of various substituted benzazides was carried out in acidic solventssuch as acetic acid, aqueous acetic aci...
0 downloads 0 Views 709KB Size
April 20, 1959

DECOMPOSITION OF SUBSTITCTED BENZAZIDES IN ACIDICSOLVENTS

2007

Repetition of this experiment using 10% potassium hyacidified potassium iodide solution. The mixture was poured into 75 ml. of water, heated 15-20 min. on a steam- droxide gave a solution an aliquot of which showed 1907, bath and neutralized with 157, sodium hydroxide. Addi- unsaturation on bromide-bromate titration (high unsaturation values are typical of unsaturated sulfonic acids with a tion of 0.5 g. of methone in 10 ml. of 95y0 ethanol precipitated 0.1576 g. (5.390 X mole, !R%) of formaldehyde branch a t the double bond; see ref. 8). Neutralization of the major portion of this solution with hydrochloric acid methone derivative m.p. 187.5-188.5 . A 0.39-9. sample of S-p-chlorobenzvlthiuronium 2,2,3- and evaporation gave 3.0 g. of potassium salts (849;). Coulometric titration'' of a sample gave ilyo unsaturation. trimethyl-3-butene-1-sulfonate was dissolved in methanol 2,3-airnethy1-3-phenyl-l-b~and passed ovcr Dow-X-50 cation exchange resin to liberate S-p-Chlorobenzylthiuronium the free sulfonic acid. T h e column was eluted with metli- tene-1-sulfonate, m.p. 171-172", was obtained in 627, yield m o l . Broniide-bromate titration gave a n unsaturation (after three crystallizations, assunling ;lyOinitial unsnturation) froni the potassium salts. value of 1037,. 4-Methyl-4-hydroxy-2-pentanesulfonicAcid Sultone (X). Anal. Calcd. for C20H2503N&Cl: C, 54.47; H , 5.72. -Evaporation of the solvent from the organic layer (separated from the aqueous layer after hydrolysis of the sulfona- Found: C, 54.25; H , 5.51. tion of 4-methyl-2-pentene) with gentle heat gave only An identical derivative was obtained from the potassium decomposition products. Evaporation without heat gave salt isolated from the "neutral" hydrolysis. a residual amber oil. This was dissolved in ether and satuThe infrared spectrum of the potassium salt from the rated a t 0' with pentane. Cooling t o -78" produced Carbon basic hydrolysis showed no absorption in the 11.2 p region. colorless crystals melting at about -48 to -44". and hydrogen analyses on two different samples gave C, The ultraviolet absorption spectrum of a n aqueous solution 46.76 and 46.83; H , 8.24 and 7.72. The theoretical values showed a maximum at 258 mp (emax 210), typical of a n are: C, 43.89; H, 7.37. Neutral equivalents on different alkylbenzene, and no high intensity absorption at 244 mp, samples ranged from 270-31 1 (calcd. 164). The molecular characteristic of styrenes. 2,3-Dimethyl-4-hydroxyy-2-butene-l-sulfonic Acid Sultone weight, determined cryoscopically in acetic acid, was 141. sulfonation was carried out using 12.0 g. The sultone decomposed at room temperature within a few (XVIII).-The (0.146 mole) of 2,3-dimethyl-1,3-butadiene, 11.4 g. (0.142 hours. 3,4-Dimethyl-4-hydroxy-2-pentanesulfonicAcid Sultone mole) of sulfur trioxide, 13 ml. of dioxane and 200 ml. of (XI).-The two crops of crystals of crude sultnne as pre- ethylene chloride. After hydrolysis, the organic layer was concentrated under vacuuni at room temperature to give cipitated from ether solution with pentane melted at 50Distillation in a Hickman still 53' and 50-58". RecrysJallizstion of the first crop gave a sizable quantity of oil. material melting at 42-43 . After standing for five months a t 100" and 0.3 mm. gave 1.5 g. (6.5yo)of S V I I I , m.p. 40the solid melted at 74-78'. The 42-43' melting material 41 '. Crystallization from ether failed to raise the melting point. on recrystallization from ether-pentane melted at 4343.5', but a second crystallization raised the m.p. to 700Anal. Calcd. for CeHlo03S: C , 44.44; H , 6.22. Found: 73', and a final crystallization gave material, m.p. i6-77 C, 44.95; H , 6.37. Recrystallization of t h e initial material, m.p. 50-58", gave the higher melting form. The infrared spectra of the The aqueous layer contained only 3070 of the sulfur tri4 2 4 3 ' and 76-77' materials were identical except for two oxide used (370as sulfate and 27% as sulfonic acid), sugadditional weak absorption bands at 10.95 and 12.5 p gestinp that the yield of sultone was much higiier than was present in the lower melting form. These appear to be isolated. gimorphic forms of the sultone, rather than cis-trans Repetition of this experiment using pentane to precipiisomers. Hydrolysis of 2,3-Dimethyl-2-phenyl-3-hydroxy-l-butane-t a t e the sultone from the organic layer gave a 16% yield of sultonc. sulfonic Acid Sultone (XII).-Four grams (0.0167 mole) of Thesc expwiments were performed a t a n early stage of the XI1 was hydrolyzed in 400 ml. of water at steam-bath temwork and the precautions later adapted for the hydrolysis perature, and the solution diluted t o 500 ml. Titration of a 20-1111. aliquot with standard base showed the presence stage were not employed. I t seenis likely that the yield could be improved substantially over t h a t realized. of a theoretical quantity of sulfonic acid. Bromide-bromate titration of a 50-ml. aliquot gave a n unsaturation value of 147% ( a small amount of solid precipitated). EVANSTON, ILL.

.

[CONTRIBUTION FROM THE INSTITUTE O F SCIENTIFIC A N D INDUSTRIAL

RESEARCH, OSAKA

UNIVERSITY]

The Curtius Rearrangement. 111. The Decomposition of Substituted Benzazides i n Acidic Solvents, the Acid Catalysis BY YASUHIDE YUKAWA

AND

YUIIO TSUNO

RECEIVED MAY26, 1958 Kinetic study of the Curtius rearrangement of various substituted benzazides was carried out in acidic solvents such as acetic acid, aqueous acetic acid and acetic acid containing sulfuric acid. The rate for a given substituent was increased by the changes of solvents in the order: toluene, acetic anb.ydride, PlcOH, nq. XcOH and AcOH containing H2SO4. I n acetic acid containing sulfuric acid, a linear Hammett correlation was observed. This result on the substituent effect was similar t o that of the Schmidt reaction and was different from t h a t of the Curtius rearrangement in toluene. However, in acetic acid without H~SOI,the linear Hammett plot was not given; the rates of the derivatives containing para electron releasing groups were nearly equal to t h a t of parent compound. The analogous shape of Hammctt plot was obtained in aqueous acetic acid and acetic acid containing lithium chloride. The energy of activation varied considerably by the change of substituent. In acetic acid without sulfuric acid, the change of log PZ was parallel with that of activation energy, whereas the log PZ was constant in the presence of sulfuric acid.

A previous paper' reported the substituent effect of the Curtius rearrangement of the various substituted benzazides in toluene; in the meta-substituted benzazides, the polar nature of the substituents the ratel whereas* in the para-substi(1) Y. Y u k a w a and Y. T s u n o , T H I S JOURNAL, 79, 5530 (1957).

tuted one both electron releasing and attracting substituents in any case retarded the rate as compared with the unsubstituted. These results are somewhat unusual and differ from the Lossen,2 (2) T. F. Bright and C. R. Hauser, ibid., 61, 618 (1939); Renfrow, Jr., and C. R . Hauser, i b i d . . 69, 2308 (1939).

W. B.

2005

1-01, f P O 5 I T I O S O F S L - B S T I T C T E D ment by the comparison of the acid-catalyzed BESZAZIDES I S .\CETIC .\CIn Curtius with the Schmidt rearrangements. Temp.,

Results and Discussion The rates of the decomposition of a number of substituted benzazides were determined in acetic acid, acetic anhydride and acetic acid containing water, sulfuric acid and lithium chloride. In all runs excellent first-order plots were obtained covering over a period of SOYc reaction. The rate constants obtained are listed in Tables I, I1 and 111. From these constants the energies of activation and log PZ were calculated using the method of least squares, and the values are given also in Tables I and 11. X s the solvent varied, the rate for a given substituent was reduced in the order; acetic acid containing sulfuric acid, aqueous acetic acid, acetic acid containing lithium chloride, pure acetic acid and acetic anhydride. I n these solvents the rate was much higher than that in toluene. I n acetic acid containing sulfuric acid (2070 1-01.), electron releasing groups accelerated and electron attracting groups retarded the rate. A plot of the logarithms of the rate constants a t 44.70' against sigma constants gives a straight line, as shown in Fig 1. The regression line is obtained by the sta(3) C R . Hauser a n d R'. B. R e n f r o a , J r . , THISJ O U R N A L , 69, 121 (1937). ( 4 ) L . H. Briggs and J . W. Lyttletone. J Chem. S o r . , 4'71 (1943). ( 5 ) 51 L Gro-sley, R . H . Kienle and C . H. Benbrook, THIS J O U R K A L . 61, 1400 ( I W C ) , C. I;. Ingold, " S t r u c t u r e a n d Mechanism of Organic Chemistry," Cornell Cniversity Press, I t h a c a , S . Y..1953, pp. SO&S02; E. S. L e a i s a n d E. B. Miller, THISJ O U R N A L , 76, 229 (1953) ( 6 ) S G. Cohen a n d C. H. Wang, ibid., 1 6 , 5504 (1953); G. L. Davies, D. H. H e y a n d G. H. Williams, J . Chcm. SoC., 4397 (1956). (7) 51. S. S e K m a n a n d H. L. Gildenhorn, THIS J O U R N A L . 7 0 , 317 f194t3; R . A. Colemmn, M.S.h'ewmnn nad A , B.G a r r e t t , i b i d . , 1 6 , 4634 (10Q.I),

Substituent

p-HO

P-CHjO

P-t-CJIs

Sone

OC.

55.60 6 5 20 70.20 74 50 75.00 54 s5 59.50 65.20 7 5 50 76.00 5,5.00 65.20 70.00 75.50 55.60 62.00 65.20 65.90 70.00 74.50 79.98

10s

x

1;o

'7 i o 5 i 0

73.j i 12. ,56 i 2 0 s =t 21.2 2 2 054 zk 3 41 i 6.634 zk 21 0 i 21 2 I 2 57 2 5.29 i 13.93 f 26.26 i 2.452 I L5 20 d= 7.68 i 8.13 f 12 55 21.57 It 3; 5 i 6.052 f 34.2 i 1 471 i 5 15 i 8 60 i 15 83 i 3 55 i 6 243 d= 13 39 i 21 0 f

+

'-1 I i1.j

11-1 1

7 .Ocl4

.03 016 ,I

.3 07 .02 ,062 ,014 ,013 026 .O5 ,076 .04 01

.3 012 .2 COG4 27 0 p-so3 03 06 0.5 ??l-SO* 02 28 6 006 .02 i6.10 .17 80.00 E The averaged rate constant in reciprocal min. and the averaged standard deciation of rate constants. Xrfn reciprocal rhenius :xtivation energ>. in kcal. mola'L. mind

p-c1

65.20 81,OO 55,oo 65 20 70.00 74.77 65.20 70.00

THE R.ITEF OF

TABLEI11 T H E DECOMPIXITIOS OF BESZAZIDES A T 65 70'

Subst1tuent

THE

p-HO p-CH;O p-t-CIHg S ,)ne p-c1 p-so.

.~cOH-H:O ( 2 0 ~ ;v t ) l , j

p-HO p-CH:O p-t-C;Hr

.JcOH-LiCI

n1-so:

((1

75 .ll)

si.1ne

p-c1

p-so?

?,I-so> p-HO P-CHLO p-t-C;H 9

STBSTITL'TED 1.0'. X-bl trnin. 1 )

So! v e n t

14.2 3 2 0 . 1 13.1 i .1 15 6S i .03 13.44 i .!xi 9 . 4 5 i .04 7 . 9 0 i .06 5 . 4 4 i .02 9,s' i. .20 S . 7 7 6 i ,013 10.30 + . 0 8 9 . 4 1 i .02 6 81 i .06 5 37 i 03 3 . S 6 7 i ,012 3.92 07 3 . 1 0 i .02 4.694 i ,012 4 . 4 2 i .03 3 72 i . 0 2 3 . 7 9 i 09 2 70 i .02

*

AC>O

\-,.)ne

p-c1

p-so. l?l-sO?

tistical treatmentS excluding the point for para. nitro deril-ati\-e. This correlation line is represented by the equation log K , = -1.090

- 2.399

The slope of this regression line, rho value, is - 1.090 0.06s; the standard deviation from the regression line is 0.020 and the correlation coefficient is 0.990. The value of rho, - 1.09, is much higher [more negative value) than that for the reaction in toluene' (-0.33 concerning the metusubstituents only). In acetic acid without sulfuric acid, the rate for a given substituent was markedly increased relative to the rate in toluene. The Hammett plot gave a straight line except strongly electron releasing para groups and para nitro group. The rate of paru-hydroxy and Pam-methoxy groups are nearly equal to that of the parent compound. The rho of this reaction is -0.43, which is slightly higher than the value in the solution of toluene. By the addition of lithium chloride and water to this solvent, the rate was slightly increased, but the consistent pattern of the Hammett plot in pure acetic acid was revealed (Fig. 2 ) . The rho-value was slightly increased, -0.53- - 0.55. In acetic anhydride, para electron-releasing groups showed a remarkable retardation of rate as compared with that of the parent compound. T h e shape of the Hammett plot iii this case is the same as that in toluene. This plot is also given by shifting the shape of plot obtained in toluene along the vertical ordinate. Both reactions are correlated by the equation

*

log

k w e t i e anhydride

= log

ktoluene

+ 0.315

I t would be of considerable interest t o compare with the patterns of these Hammett plots in various solvents. Concerning the substituent effect of the Schmidt rearrangement, Briggs and Lyttletone' reported (e)

2009

DECOJIPOSITIOS OF SCBSTITVTED BESZAZIDES IS.\CIDIC SOLVESTS

-1pril 20, 19.59

H. H. TGAT.. Cham,

sa,i o 1 (iwa),

! WCH\

-2,7!1 : ,

1

-3.20

-0.4

m

-0.2

0.0

0.2

!?.A

\o. \ I !I

0.8

1:

Cl.

Fig. 1.-The Hamniett plot of rate ci-iii~t;lnt~ c)f the rearrangement of benzazides i n 3cetic acid o mt;liriiiig sulfuric acid at 44.70".

-'"'* c t

1

-2.60

U

t-

-0.4

I

-0.2

0.0

0.2

0.4

0.6

0.S

0

Fig. 2.-The Hammett correlations of the rearrangement of the benzazides at 65.20'; ( A ) in sllrc aqueous acetic acid; ( B ) in acetic acid containing LiCl ( I , , M j ; ( C ) in pure acetic acid; (D) in acetic anhydride.

about the half-lives for the reaction of various substituted benzoic acids with hydrazoic acid in the presence of sulfuric acid. .kcording to their results, the rate was favored by electron releasing groups. The reciprocal logarithms of the halflives are found to be almost linear against the Hammett sigma constants. X plausible mechanism could be represented as (9) 1 Fine. "Phy4cnl Organic Che-nktrs-," AIcGraw-Hill Cos,Inc., S e w York, S . Y.,1906, p, aP1,

Boar

R 110-

C'

- T--sZ7S I

-

:icicl

-*['

H

A

0

R

\ I

116

R

1- I H

n 0.

'

c--s--s=s +

-

R

c=s--s=s _

c-s+s,

0/!\

T h e addition of the hydrazoic acid with the osocarbonium cation, R-CO+, rnay be a fast step and the decomposition of this intermediate adduct s e e m to be the rate determining step. -in azide is converted into its conjugate acid by a strong acid such as sulfuric acid. This conjugate acid rnay be formally identical with the adduct nf hydrazoic acid with osocarbonium cation. The acid-catalyzed Curtius rearrangement should be, therefore, similar to the Schmidt rearrangement i n their rate-determining process and also presumablv in their transition state. This leads us to assume that the effect of the substituents on the rate in both cases must be close together. The above assumption is justified by the rate sequence, as shown in Table I and Fig 1, of the rearrangement in acetic acid containing sulfuric acid (2Oq0 vol.). This is the comparable condition with that of the Schmidt rearrangement. The rate of each substituent in this solveiit increased to 50 100 folds of that in toluene. Aforeover, a straight Hanimett correlation line was obtained only in this solvent. This fact does suggest that in this solvent the effect of the polar contribution of substituents to the rate predoniinates over that of the resonance contribution to the bond energy of the breaking N-N bond. I t is similar to the result of the Schmidt rearrangement, i l l contrast to that of the uncatalyzed Curtius rearrangement in toluene. On the other hand, the rate in acetic acid without sulfuric acid, is found to be increased considerably a s compared with that in toluene. The higher rho \ - d u e than t h a t in toluene appears to be an evidence of some increase in polar contribution of substituents. However, the failure of the Hammett relntioii (Fig. 2 ) shows t h a t the increased polar cnntribution of substituent in this solvent is not suf7icicnt to give a regular polar sequence. The smne is true of the results in aqueous acetic acid n n t l acetic acid containing lithium chloride (Fig. 2). Fuliuric acid in acetic acid rnay produce acetic anhydride, which could be expected to influence nn the rate by accelerating the rate or increasing the polar contribution of the substituents. I n this situation, the effect may be significant in acetic anhydride. The result in acetic anhydride showed iio individuality of the anhydride on the substituent effect to the rate. Another possible effect of the sulfuric acid rnay be a salt effect. This effect is also excluded by the experimental evidence that the addition of lithium chloride t o the solution did not exert any remarkable effect on the polar contribution of the substituent to the rate. Furthemore, the salt effect of the st,rong acid or to strongly solvated proton would be be smaller than t h a t of lithium salt.1°

-

(IO) J . E.L e m e r a n d S h i n - ~ u n g L i u , T e I sJOURNAL,?^, 1949 (1950).

In general, as the polarity of solvent is varied, the contribution of the substituent is changed.lI Therefore, i t is iniportant to ascertain whether the similar substituent effect given in toluene prevails in the other solvent of high polarity or not. 'The result in acetic anhydride provides an evidence of this. The rates in other higher solvatiiig solvents were found to be higher than that in toluene or acetic anhydride;'? furthermore, in se\-er:il solvents, such as phenol aIid aniline, higher than iri acetic acid. 'The substituent eiTect i n them, however, is closely similar to that in toluene." Hence the polarity of solvent does not appear to affect the nature of the substituent contribution in this reaction. The participation of the solvent, which has higher solvating power, is presumably distiri guished from t h a t of sulfuric acid. The addition complex of an azide with sulfuric acid or the protonated azide undergoes deco~iipositiori to give a positively charged intermediate. whereas the usual Curtius rearrnngeiiien t of azidc involves an electrically neutral interiiiediate C O I I taining an electron deficient nitrogen atom I I,SO,

CJIS

\ ,/

0'

C--s

CGIIj

\

/

-

---f

\

ITJSOI

4'' S?

/

c--s

+

+ s2

0

The rate of reaction involving the charged reaction center of transition complex must be influenced by the polar nature of substituent, as would presumably fit the Hamniett equation.l 3 In the acid-catalyzed rearrangement, the rate of decomposition would be influenced by the electrical nature of substituents, if the reaction center beconies more positive a t the transition state than the initial state. Hence, the substituent effect of the acid-catalyzed reaction may differ markedly from that of uncatalyzed reaction. The decomposition of benzazides, however, failed to give a regular Hammett sequence in 2576 aqueous acetic acid solution containing sulfuric acid (73%), or in aqueous dioxane solution containing sulfuric acid or hydrochloric acid.14 The linear Hanimett relationship was obtained only in thc presence of anhydrous acid (10 35%). Thus the contribution of sulfuric acid as Lewis acid appears to be effective to give a regular Hammett sequence, and in the protonated azide, substituents influence the rate in the same manner as they influence the rate of rearrangement in toluene. On the energies of activation (Table I and I I ) , there is a clear contrast between the acetic acid containing sulfuric acid and that without sulfuric acid. In acetic acid without sulfuric acid the energies of activation and log PZ are varied in parallel over a range of 25 '19 kcal. mole-' and 13 16 min.-' This phenomenon has been observed in various rate processes, especially in t h e Curtius

-

-

-

D. A . Brown a n d R . F. H u d s o n , J (-hem. Soc., 883 (1933). (I?) Y . Y u k a w a a n d 1 ' . T s u n o . unpublished

(11)

(13) T h i s is suggested essentially b y a Referee, f i e suggests a l w t h a t t h e 0 - h y d r o x y a n d p - m e t h o x y g r o u p s would s h o w greater electronrelease t h a n dewribed by t h e normal sigma c o n s t a n t , in t h e electrophilic reaction. However, t h e deviations of these groups from H a m m e t t relation a p p e a r s t o be involved within experimental u n c e r t a i n t y . ( 1 4 ) Y . Y u k a w a a n d Y.T s u n o . t o be ouhlished. (13) J. E. LeiRer, J . Org. C h i n , 2 0 , 1202 , 193.5)

Xpril 20, 1959

L)ECOMP'USITIOS0 1 7

SUBSTITUTED UENLUIDES

rearrangement, by both solvent change16 and substituent change in t o l ~ e n e . ~The ? ~ ~plots of activation energies against log I'Z for the Curtius reactions in acetic acid gave a straight line lying below those in toluene. On the other hand, in the presence of sulfuric acid, the energy of activation is considerably changed, whereas the log PZ is remained nearly constant (14,s niin.-I). For a given substituent, the energy of activation in this sol\-ent is slightly less (about 1 kcal. mole-') than that in acetic acid without sulfuric acid. Sulfuric acid might rather freeze out the entropy (to give a constant log P Z ) than lower the energy of activation. Then, it is reasonable to conclude that thc rate sequence obtained in acetic acid containing sulfuric acid is attributed to the effect of Lewis acid catalysis by the sulfuric acid, thereby, the dissimilarity in the substituent contribution of the Curtius rearrangcnieiit and the Schmidt rearrangement would be causcd by no intrinsic difference of their transition states but the change in the condition of reaction. Experimental

l f .S . S e w n i a u , S. H. I.ee, Jr., and A . I3. Garrett, ' I H I ~JUUK69, 113 (1947). (17) Y.Yukawa and Y.Tsuno. :bid., 80, G34ti (19.58).

(ICi) N.%.

ACIDIC SOLVENTS

"01 1

I n the cases of solvents, acetic acid, aqueous acetic acid aud acetic acid containing lithium cliluritle, the nieasurcinents were made in 100 ml. of thcm. T h e initial eo~icciitration of azides were about 0.05 mole per liter. For d l runs tlie duplicated measurements, differed in iiiitial CIIIIcentrations within 0.1-0.03 mole per liter, were Inatle a t the same t h e . Tlie results showed t h a t the itiitial C I ) I I C V I I tration did nut give an>-significant effect on the rate n-itliout ones in an aqueous acetic acid. Tlie values listed i n tllc tables and discussed above are the mean values of tlie duplicated results. The reproducibilities of these rate const:ints appcar to be within 2y0,which is demonstrated iii Fig. 3 .

2.00

1.so

s I 8

5 1.60 M

2

t

Materials.-.Ill azides \wrc prepared by tlic iiicthuds I)reviously rcported.l Litliiuin cliloritle uscd \yere a reagent nd dried o\.cr pliuspliurus pcntusidc to a cullstatit Solvents.--.\cetic acid, ivliieli Iiutl Leeti p;trti;tlly frozen and distilled, wis heated uiider reilus \\it11 pliuspliorus peiitosidc for 1.5 liours and c ; d u l l y fractioii:itcd. Tlic middle fraction, b . p . 117.5-8.3°, 12% 1.3700, x i s collcctcd. :Icetic uiliytlride W:IS purified by reflusiiig with calcium carbide for 17 liuurs and fractionally distilled. Tlic iiiitldle fr.ictiun, b.p. 139--40', wrts collected. Tlic stock solvents, 20'; a q U e ( J l l S acetic acid, \ V : L ~ lireIiarctl b y tliluting 200 nil. of water with acetic acid t o 1000 1111. (,it 9 0 0 ) . Tlic acetic :icitl coiit:iining sulfuric acid was prepxrctl b y i i i i atlditioii uf 2i)U nil. uf luu',; sulfuric acid iiito 500 ml. of acetic acid a t U o :iiid the dilutioii witli acetic acid t o 1000 inl. :it 90". Tlie accurate content of sulfuric acid \fits deteriniiietl gr,i\-iinetrie,illy :is lxtriuni sulfate. I t w;is found to lie 1 I i i ~ J l C Sper liter,f = 0.9S9S. The acetic acid used for this solution wets prepared b y reflusiiig arid frxtionating the inisture o f coininercid acetic acid (1300 nil.), wetic :inliytlride (100 nil.) and concd. sulfuric acid 110 ml.), niiddle fr:ictiuii, h.p. 117-11S.5', being cullected. .l solution of litliium chloride in acetic x i d was prepared :is folluws: lithium cliloritle (0.5 inolc) IV,LS tlissolvctl into ,jOO nil. uf acetic acid, the insoluble s d t w i s tiltcrcd off anti 930 tnl. uf this solution \\-as diluted t o 1000 i n l . The accurate eunteiit of the chloride -.I as ileteriiiiiictl as silyer chloride: 0.75 mule per liter,f = 0.9S2G. Rate Determinations.-Tlie rate nicasureiiieiits were carricd o u t b y the same prucedurc a s the previous The temperatures of the reaction bath werc ni:iintaincd t o :L constancy of + 0 . 0 l 0 , and thc uncertainty of tcmperaturc would be within f 0 . 0 3 ' . The reaction mixture was nut stirred, b u t tlie poivderetl glass \vas added t o the reaction flask iii order to prevciit supxsaturation b y nitrogen gas evolved. .\11 reactiuns were followed ovcr a period of a t least three half-lives of azide decomposition. T h e infinity reading, I', , was takeii after an inter-\.a1 of time equal t o 9-10 times a s much as the half-life of re;tctiori. The volume of nitrogen was obtained quantitatively (close t o the theoretical amount). A plot of the log ( Vm - 1') against time gave a n excellent stwiglit line covcring over a period of thrcc half-li\-es of the rwctinn (Fig. 3 ) , and the rate constants were w a l uated from ;in equatinn K , 1 = 2,:3!):3[lug v m - lug ( 1.- - V ) ]

IN

0

h' 1

I

20

40

ti0

1

I

I

YU

100

17U

110

t (iiiiii.).

Fig. 3.-The plots of log ( LTm - )'L i's. time; (.%I viN02Cd14COSaa t 70.00° in AcOII; ( n ) CGIIICOS1at 51.7tj" in AcO€I-I12S01; ( C ) pt-C4HYCJ11COS3 at 7 5 . 3 " iii AcOII. I n acetic acid, tile rates were nicasurcd ;it srvc r,il diffcrent tcmperatiircs. Froin tliehc values, tile IiiicLir .\rrlicniiis plots were ol)t:iiiied. S e w r i i ~ r i irepiirtril ~~ t i l ? r,itc\ for tinsubstituted 1)enznzitlc ;is follows: ( ! . i ! I t 2 4 > [ i i i i i i . - ' ) a t S o ,0.01)828a t 65' and U.!)217 :it 75'. T1iL,\c arc i i i good agrecnient with the values i i i Ttilile 11, Init tlie value a t 65' is snrnc\rh:it higher. From the .\rrliCiiitis pint 01)tainetl by ccimhinitig tlic rate cnnstniit i i i T,ihle I 1 wit11 Scwniaii's, tlie v d u c of tlic rxtc coii't:iiit o f tlie I)resent study, 0.007M a t 05.20°, is more prcferalilv. Tlic sm:ill content of an acetic anhydride in :icetic acid dit1 not xffcct o n tlic rate. The reactions i n an aqrreous acctic acitl siiltitilin tile rxte coiistaiits \rere fluctuxted sensiti\-cly by tliose iiiitixl C O I I cciitrxtinns. The preliniin:try experiments i n 50 inl. (if this solvent (azide, 0 , O X mole) gave t i l e follo\ring r e s u l t h : (subst., kl i n rnin.-' a t 65.20") p - H O , 0.01S0; p - C H n O , 0.015; p-i-C4H8, 0.018.5; n o n , , 0.0175; p-SO:., 0.00871 ant1 nz-SOs, O.l)I)G38. These :ire sliglitly liiglier tlian those in 100 nil. of :in : t q u c n u ~acctic acitl solution. However, the effect of snidl difference in initial conccntratii~n was negligiblc. Tlic measurements, in acetic acid containing sulfuric acid and in acetic anhydride, were carried out in 50 ml. of SOIvents. The initial concentratioris \vere about 0.1 mole per liter. The rates for several substituents in 100 ml. of the latter solrent did not de\-iate from the values in 50 ml. solvent. The duplicated runs of double concentration gave an identical rate constant within a n experimental uiiccrtainty. In the casc of the former solvent, the slight change in the initial concentratinn of azides did not affect tlie rate. All runs in this solvent were of pseudo first order, but a t low temperature, 2 e W 5 0 , the lower initial rates were observed over 30-60 minutes, which is somewhat longer as compared with other solvents. This would be attributed t o the slower dissolution of azides to this solvent. The dissolution of p-t-C4H9derivative was markedly sluw and this showed the longest period of the decrease of the iiiiti:tl rate constant. :\voiding this effect, the solution of azides \vas shaken vigorously for about ten minutes before the measurement was started. T h e reaction products were detccted as acctclnilidLs.

VOl. s1

A. T. BLOMQUIST AND DANIELT. LONGONE

2012

Acknowledgment.-The authors wish to express their sincere appreciation to Professor M. Murakami and Dr. I. Moritani for their invaluable sug-

gestions in this work, and they are also indebted to the Ministry of Education for the partial financial support of this research.

[CONTRIBUTION FROM BAKERLABORATORY OF CHEMISTRY, CORNELL

UNIVERSITY]

Synthesis of Cyclopropane Derivatives. Precursors for Dimethylenecyclopropane and Trimethylenecyclopropane BY A. T. BLOMQUIST AND DANIELT. LONGONE' RECEIVED OCTOBER 11, 1958 Two diamines, trans-1,2-bis-(dimethylaininomethyl)-cyclopropane and tralrs-2,3-bis-(dimetl1ylaminomethyl)-l-methylenecyclopropane, desired as precursors for projected syntheses of dimethylenecyclopropane and trimethylenecyclopropane, respectively, have been obtained v i a the reaction sequence R ( C 0 2 H ) 24 R(COCl)z + R(CONMe2)Z + R(CH;NMe2)z. The acid and trans-lvarious transformations were effected in good yields starting from ~rans-l,2-cyclopropanedicarboxyl1~ methylenecyclopropane-2,3-dicarboxylicacid (Feist's acid). Examination of the infrared spectra of some fourteen cyclopropane derivatives confirmed the view that absorption bands in the 9.8-10.0 and 11.5-11.7/1 regions are not reliable for indicating the presence of a cyclopropane system in a molecule.

The increasing interest in the chemistry of cyclopropane derivatives is due in part to the prediction of non-classical aromatic character for certain unsaturated cyclopropyl compounds. Application of the molecular orbital (LCAO) method in the calculation of the electron delocalization energies, bond orders and free valence indices of a number of small ring compounds predicts delocalization (resonance) energies of about 34, 16 and 29 kcal. for the cyclopropene cation (I), methylenecyclopropene (11) and trimethylenecyclopropane (I1I), respectively.2

I

I1 111 Since the results of these calculations are essentially qualitative due to the known limitations of the method used and the approximations introduced, the only true test of their reliability must eventually lie i n the synthesis and study of the compounds in qucstion. A derivative of I , the tripheiiylcyc1ol)ropeiiyl cation, has recently been ~ y n t h e s i z e d . ~The success in obtaining this relatively stable cation not only supports experinientally the reliability of the theoretical conclusions cited above but also stimulates synthetic effort toward other non-classical aromatic compounds in the cyclopropane series. The cyclopropane compounds of particular interest are those which contain exo or endo double bonds. Besides the naturally occurring cyclopropene sterculic acid4 and the methylenecyclopropane Hypoglycin As the number of, and routes to, such cyclopropyl compounds are meager. For this (1) U. S.R u b b e r Research Fellow, 1957-1958. T h e work reported here was abstracted f r o m p a r t of t h e dissertation presented by Daniel T, Longone in September, 1958, to t h e G r a d u a t e School of Cornell University in partial fulfillment of t h e requirements for t h e degree of Doctor of Philosophy. (2) J. D. Roberts, A. Streitwieser, J r . , a n d C . M. R e g a o . Tms J O U R N A L , 74, 4579 ( l Q 5 2 ) . (3) R . Breslow, ;bid., 79, 5318 (1057). (4) K.L, Rinehart, Jr., W . A . Nilsson and If A . Whaley, i b i d , 80,

,503(19.58). ( 5 ) J. A C a r h o n , W. H iilartin a n d I, I < . S w e l t , tbirl , 80, 1002 (IUOS); R .S.d r R u p p . C L ( 1 1 . . t b i d . , 80, 1004 ( 1 9 5 8 ) .

reason, a careful study of the applicability to cyclopropane systerns of standard classical transformations utilized successfully in the larger, strainless-

Hypoglycin A

sterculic acid

ring homologs to convert dicarboxylic acids to diolefins would be valuable.6 This paper describes the synthesis and characterization of intermediates to be used for such a study. They are derived from two of the more readily accessible cyclopropanedicarboxylic acids, trans-l,2-~yclopropanedicarboxylic acid (XII) and Feist's acid (XX). Derivatives of trans-l,2-Cyclopropanedicarboxylic Acid(XI1).-The dicarboxylic acid X I I , obtained from a-bromoglutaric ester (IV) by the method described by Ingold,' served as the starting material for synthesis of the cyclopropane derivatives given below.

a

'OzH

ELOH,

~ c o z E ~ t COzEt

H ' ,

COzH

I

1, LiAI13,

2, Me$"

2, AGO, AcOH 3, Ac,O

0Lcp:oAACc-

c1

CONMez

xv

VI1

I

LiAIH,+

ml

I

CONMez

I

a CHzOH

MeOH, NaOMe

1,SOCI,

1

~

a

~ C H 2 N W e a CHzNMez

XI

I

Me,N

CH2NMe3Br-J CH2NMe3Br-

+

IX

Cyclization of the bronioester I\' .ilia intramolecular dehydrobromination with methanolic potas. ( 6 ) (a) A . T . Bloin